With Special Reference to

the Ammonia Process

A Practical Treatise

By

TE-PANG HOU , Ph. D.

MEMBER, AMERICAN INSTITUTE OF CHEMICAL ENGINEERS;

MEMBER, AMERICAN SOCIETY OF MECHANICAL ENGINEERS;
LIFE MEMBER, CHINESE INSTITUTE OF ENGINEERS; LIFE
MEMBER, THE SCIENCE SOCIETY OF CHINA; ETC ., ETC.

Second Edition, Thoroughly Revised and Enlarged

With 182 Tables and 149 Illustrations
 GENERAL INTRODUCTION
American Chemical Society Series of
Scientific and Technologic Monographs

By arrangement with the Interallied Conference of Pure and Applied Chemistry, which met in
London and Brussels in July, 1919, the American Chemical Society was to undertake the
production and publication of Scientific and Technologic monographs on chemical subjects. At
the same time it was agreed that the National Research Council, in cooperation with the American
Chemical Society and the American Physical Society, should undertake the production and
publication of Critical Tables of Chemical and Physical Constants. The American Chemical
Society and the National Research Council mutually agreed to care for these two fields of
chemical development .The American Chemical Society named as Trustees, to make the necessary
arrangements for the publication of the monographs, Charles L. Parsons, secretary of the society,
Washington, D. C ; the late John E. Teeple, then treasurer of the society, New York; and Professor
Gellert Alleman of Swarthmore College. The Trustees arranged for the publication of the A. C. S.
series of (a) Scientific and (b) Technologic Monographs by the Chemical Catalog Company,
Inc.(Reinhold Publishing Corporation, successors)of New York.
The Council, acting through the Committee on National Policy of the American Chemical
Society, appointed editors (the present list of whom appears at the close of this introduction) to
have charge of securing authors, and of considering critically the manuscripts submitted. The
editors endeavor to select topics of current interest, and authors recognized as authorities in their
respective fields.
The development of knowledge in all branches of science, especially in chemistry, has been
so rapid during the last fifty years, and the fields covered by this development so varied that it is
difficult for any individual to keep in touch with progress in branches of science outside his own
specialty. In spite of the facilities for the examination of the literature given by Chemical Abstracts
and by such compendia as Beilstein’s Handbuch der Organischen Chemie, Richter’s Lexikon,
Ostwald’s Lehrbuch der Allgemeinen Chemie, Abegg’s and Gemlin-Kraut’s Handbuch der
Anorganischen Chemie,Moissan’s Traite’s de Chimie Minerale Generale, Friend’s and Mellor’s
Textbooks of Inorganic Chemistry and Heilbron’s Dic tionary of Organic Compounds, it often
takes a great deal of time to coordinate the knowledge on a given topic. Consequently when men
who have spent years in the study of important subjects are willing to coordinate their knowledge
and present it in concise, readable form, they perform a servic e of the highest value. It was with a
clear recognition of the usefulness of such work that the American Chemical Society undertook to
sponsor the publication of the two series of monographs.
Two distinct purposes are served by these monographs: the first, whose fulfillment probably
renders to chemists in general the most important service, is to present the knowledge available
upon the chosen topic in a form intelligible to those whose activities may be along a wholly
different line. Many chemists fail to realize how closely their investigations may be connected
with other work which on the surface appears far a field from their own. These monographs
enable such men to form closer contact with work in other lines of research. The second purpose is
to promote research in the branch of science covered by the monograph, by furnishing a
well-digested survey of the progress already made, and by pointing out directions in which
investigation needs to be extended. To facilitate the attainment of this purpose, extended
references to the literature enable anyone interested to follow up the subject in more detail. If the
literature is so voluminous that a complete bibliography is impracticable, a critical selection is
made of those papers which are most important.

AMERICAN CHEMICAL SOCIETY

BOARD OF EDITORS

Scientific Series:- Technologic Series:-

William A. Noyes, Editor, HARRISON E. HOWE ,Editor,
S. C. LIND, WALTER A. SCHMIDT,
W. MANSFIELD CLARK, E. R.WEIDLEIN,
LINUS C. PAULING F. W. WILIARD,
L. F. FIESER. W.G.WHITMAN,
C. H. MATHEWSON,
THOMAS H. CHILTON,
BRUCE K. BROWN,
W. T. READ,
CHARLES ALLEN THOMAS.
 FOREWORD
The first edition of “Manufacture of Soda” was written in my home. I count it a privilege,
therefore, to be asked by the author, whom I have known for almost thirty years, to write this
Foreword to the Second Edition of his valuable contribution to chemical technology.
I know of no person so eminently fitted for writing on the subject as Dr. T. P. Hou. Dr. Hou is
not only a scholar by training but also an engineer and technologist with years of experience in
building up China’s chemical industries. To him China owed the first and one of the largest alkali
plants in the Far East. This was located in Tangku near Tientsin. When fighting began at the
Marco Polo Bridge in July 1937,just a short distance from the plant, it had attained a daily
production capacity of some three hundred tons of soda ash, which was a good record on the
Continent of Asia.
To him China also owed her new Synthetic Ammonia industry. Having established the alkali
industry in North China, Dr. Hou turned to the great Yangtze Valley to erect the first and largest
nitrogen fixation plant in China. Dr. Hou personally supervised the building of the four modern
chemical plants at Hsiehchiatien near Nan king, which took three yeas to complete. In the summer
of 1937,the plants were producing 150 tons of ammonia sulfate daily, and the Chinese farmers,
sold on the new fertilizer, had booked orders a year ahead. Soon the War of Invasion broke out,
interrupting the work.
I was so impressed with the transformation of the farming village of Hsiehchiatien into a
humming industrial town, equipped with modern chemical machinery, that on one of my visits to
my homeland I had the American Ambassador, Nelson T. Johnson, and the president of the
Export-Import Bank, Warren Lee Pierson, accompany me on a tour of inspection. The statesman
and the banker were both pleasantly surprised and impressed.
Time marched on…When bombs were falling thick and fast, and when his associates urged
him to leave the plants, Dr .Hou said: “My duty to the farmer is to stay on the job. I want to see it
through.” Dr. Hou caught the last boat to leave Nanking with time to save nothing except a roll of
blue prints of the plants. Dr. Hou’s contributions are continuing, for far into the interior of Free
China he is laying the groundwork for a new chemical industry, despite the difficulties of
transportation over the tortuous Burma Road.
It is against such a background that the author was requested by his publishers to prepare this
Second Edition of “Manufacture of Soda”. It is my privilege to say that he has not spared himself
one bit in the rewriting and revising of this treatise. When the first edition appeared, it was the first
of its kind to appear in the English language. In bringing the book up to date and adding
considerable new material, Dr. Hou has not only enhanced its usefulness but he has shown the
public how an interesting and comprehensive treatise can be written on such a special subject as
the Ammonia Soda industry.
K. C. Li
Woolworth Building
New York, N. Y.
August 1,1941
 Preface to the Second Edition
It is gratifying to see that this monograph has met with favorable reception from the industry
and the public. Exhaustion of the supply of the last edition was the occasion for the revision. More
data have become available since the last edition was published, and much progress in the
ammonia soda industry has been made in such countries as the United States, Russia, Japan, etc. It
is, however, difficult to keep the work entirely up to date because of the rapid growth of the
industry, and because practically no official figures are available outside of the United States. It
has been considered unwise also to include all the latest indications, for fear that events later may
prove to be such that statements made today might be inaccurate a few years hence.
In this edition, typographical errors, as far as detected, have been corrected. The material in
the first edition has been revised, rearranged and brought to date. Almost every chapter throughout
the book has been re-written and enlarged, and five new chapters have been added. So much new
material has been included throughout the text that this edition has virtually the same status as a
new book, as far as its preparation is concerned. For this reason, it cannot be expected that the
present edition will be as free from typographical error as a second edition is usually expected to
be. Some apology may be given for inclusion in this edition of such subject matters as Water
Conditioning and Centrifugal Pumps, for it is believed that such information should be in the
possession of every technical man engaged in the ammonia soda industry because of their
extensive application in the industry.
This work is not meant to take the place of the service of an alkali plant designer or an alkali
expert in the construction of new alkali plants or in the investigation of a new alkali project.
The author’s thanks are due to Mr. Edward N. Trump of Syracuse, N. Y., a great inventor and
pioneer in the early history of American Ammonia Soda Industry, for his interest and suggestions
regarding the history of the ammonia soda industry, in the United States and for making available
certain data concerning the sinking and working of brine wells as practic ed in America.
The author also wishes to express his thanks to Mr. Z. G. Deutsch for his contribution of
certain material in connection with the chapters on the Burning of Limestone, the Generation of
Power, and the Layout and Design of an Ammonia Soda Plant. Mr. Deutsch also has offered some
painstaking criticisms in the text of some other chapters.
Again it has been the good fortune of the anther to receive guidance and encouragement from
Professor Daniel Dana Jackson in the preparation of this edition.
The author wishes to take this opportunity to express his appreciation for the cooperation
accorded him by Mr. G. G. Hawley of Reinhold Publishing Corporation, New York, N. Y., in
handing manuscripts for this edition.
A word may be said conc erning the situation in which the preparation of this edition was
made .At this time of national crisis in China, the author has found it very difficult to attend to this
revision. Manuscripts for the revision were prepared in part but had to be removed from place to
place, swept by the tide of invasion. Consequently, only with considerable effort was he able to
bring this revision to a successful conclusion. The invasion of Manchuria on September
18,1931,started a series of similar acts in other parts of the world. In the beginning some excuse
was offered, but now aggression becomes the order of the day and seems no longer to require an
excuse in the eyes of the aggressors, in such a troubled world, constructive work is impossible and
the author is aware of limitations under these circumstances.
 So much advice and criticism have been generously given the author by his friends, that due
credit must be given to all of them for what merits this book may possess. And the author alone is
responsible for the defects and shortcomings.
T. P. Hou
Chunking, China
December, 1940
 Preface to the First Edition
This work may be considered as an exception to the age-long secretiveness commonly
fostered in the ammonia soda industry. It is probably the first book of its kind published in the
English language covering the subject in such a manner. It is a description of the ammonia soda
industry, giving as detailed an account as is practicable in a work of such character. The work is an
outgrowth of personal notes taken daily from experience and observation in direct contact with
detailed operations in an alkali plant for more than ten years. The author wishes to make an
apology for publishing this detailed practical treatise about an industry considered traditionally so
closed. He feels that manufacturers in general are sufficiently protected by the unique
requirements of the industry, the details of which can be appreciated only through years of
personal contact with the actual operation of a plant. As regards particular features of a plant, it
may be said that with some experimentation one good designer could obtain as satisfactory results
as another, so that there is little necessity for keeping the information regarding the industry
beyond the reach of the general public. This consideration, coupled with the feeling for the need of
research on many aspects of the process and with the idea that the data and experience gained in
this branch of a physicochemical industry dealing with large volumes of liquids and gases, might
be more available to other chemical industries at large, is the basis for the publication of this work.
For it is not to be assumed that little progress has been made in this industry because of the
meager information concerning the details of the process given in the chemical literature or in the
works on chemical technology. Much progress. On the contrary, has been made, but the data have
been withheld from the public by the manufacturers for fear that they might fall into the hands of
their competitors. The patents are so buried and masked that it is hardly possible even for one
engaged in the industry to make out whether or not any such equipment has ever been actually in
use in the competitors plants. In most cases where equipment has been developed and particularly
designed for a certain work, the management prefers not to take any patent on the device, but
seeks to secure protection by not admitting any visitors to the plant. Data concerning manufacture
and detailed drawings concerning individual pieces of machinery are made accessible only to the
most trusted employees who have the complete confidence of the management.
In preparing the work, the author draws the material mostly from his own experience and
investigations. Works of Lunge, Schreiber, Monitor, Kirchner, etc., and articles by Feedstuff,
Jaenecke, Bradburn, Mason, Jurist, etc., have been consulted and critically examined in the light
of present-day operation. Most of the description either give practice that is now antiquated or
show lack in the technical details that are so desirable to those requiring information. In the
following pages, while the subject is approached from the practical side, the author has not
entirely lost sight of the theoretical background. Enough theoretical principles have been given to
enable the treatment to be more in accord with present scientific reasoning .The author asks
indulgence for making certain assumptions in several calculations which, while quite justifiable
from his own experience, may not appear so from some readers’ point of view.
Conditions in different plants vary widely; and, furthermore, practices differ in the existing
plants. Ammonia soda plants in America generally differ so much from European plants that
statements based on the practice in one country may seem prejudicial as viewed from the
standpoint of the other. Also, there is a wide difference sometimes between what should be and
what is as the actual result shows. In the following pages, the figures given represent the best
average of what usually holds under good operating conditions. There may be considerable
disagreement with respect to some of these figures because of the different conditions prevailing
in certain plants. The data and mode of operation given deal necessarily with one particular set of
conditions. They are not expected to cover all installations (e. g., different arrangement or
combinations of apparatus in the layout) or all conditions of operation (e. g., different working
methods on various forms of raw materials available). This fact explains in part why there are
certain apparently conflicting data reported in the existing literature, as considerable variations in
the operating results are possible. For example, in one plant the loss of ammonia may be as low as
3 or 4 kg. Of ammonium sulfate (NH4 )2 SO4 per ton of soda ash, while in another plant this figure
may be as high as 17or18kg.of ammonium sulfate (NH4 )2 SO4 per ton of soda ash. Again, the free
moisture content in the crude bicarbonate from the filters may be as high as 19 per cent H2 O in
one plant, while it can be as low as 12 per cent H2 O in the other. As regards the cost of
manufacture, one plant working with a high efficiency can produce soda ash at as low a cost as
$10 to $12 per ton; while another plant working somewhat less efficiently and possibly having to
haul limestone from a considerable distance may have its cost of production as high as $18 per
ton.
Many details must be worked out in a new plant. No conditions that exist in one plant can be
exactly duplicated by another. Soda works have been built and rebuilt as the days go by, so that no
one plant today remains the same as when it was built. Changes and modifications have been
introduced as defects were gradually discovered during operation. The industry requires a
large-tonnage output, a large capital outlay, and a highly developed organization. It easily costs
more than $15,000 for its construction per ton of soda ash output per day.
In working out examples either for theoretical discussion or for obtaining necessary
calculations, it is to be regretted that our present thermo-chemical data, such as the heat of reaction,
the equilibrium constant, the specific heat, etc., are still in such a shape that close agreement
among different authorities is not possible. Hence there may be a difference of opinion as to which
is the best value to use. Also ionization data, such as activity coefficients, etc., are not available in
many cases for strong electrolytes in rather high concentrations, especially in mixtures of several
electrolytes in solution. Consequently, the results obtained from such calculations are necessarily
approximate, but it is hoped that these results may be of value as a guide for practical operators. In
an attempt to elucidate the underlying principles, the author has not spared any effort to make as
clear and as detailed statements as possible, even to the extent that they may appear to be wordy or
bordering on repetition. To preserve the unity of the matter treated in the book, however, it has
been thought inadvisable to go into details of design and construction of the plants. Such a book as
this cannot be meant to take the place of the services commonly rendered by an alkali designer or
a consulting expert.
The author has taken some pains to give full credit to the previous work quoted in the text.
Wherever possible, special reference is made to the source of data from which the figures have
been taken. He wishes here to express his indebtedness to his former teacher, Prof. Daniel Dana
Jackson, for the encouragement and assistance he received in the preparation of this work. He
wishes also to acknowledge with thanks the assistance of Mr. S. T. Kuo, who has carried out most
of the analytical determinations published throughout the text. Thanks are also due to Mr. Sidney J.
Ballard, who has made some of the more elaborate illustrations for the author. To Miss I. M.
Welch of the Editorial Department of the Chemical Catalog Company, Inc., who has read the
whole manuscript, he wishes to express his appreciation for her cooperation. To his own brother,
Chi-Hsuan Hou, who has aided him in making most of the illustrations and in the preparation of
the index, he wishes to acknowledge assistance.
The author is glad to take the opportunity here to acknowledge the indebtedness to the China
Foundation for the Promotion of Education and Culture, Peeping, China, for a sum granted to
defray a considerable portion of the expenses for the investigation abroad, without which this
work might not have been completed.
Lastly, it is to be added that in a work of such nature where most manufacturing data are not
publicly available, the author is fully aware of his own limitations and would most earnestly
welcome all criticisms and suggestions.
T. P. Hou
Tangku, China, Nov. 1932
 Contents
GENERAL INTRODUCTION ........................................................................................................................... 2

FOREWORD ....................................................................................................................................................... 4

PREFACE TO THE SECOND EDITION ......................................................................................................... 5

PREFACE TO THE FIRST EDITION .............................................................................................................. 7

CONTENTS ....................................................................................................................................................... 10

CHAPTER I INTRODUCTION. HISTORICAL SURVEY OF ALKALI INDUSTRY AND RELATION

OF LEBLANC SODA INDUSTRY TO DEVELOPMENT OF OTHER CHEMICAL INDUSTRIES.... 12

CHAPTER II NATURAL SODA................................................................................................................... 23

CHAPTER III HISTORY OF AMMONIA SODA PROCESS: STATISTICS ......................................... 37

CHAPTER IV PREPARATION OF BRINE: ROCK SALT AND SEA SALT.............................................. 51

CHAPTER V PURIFICATION OF BRINE ................................................................................................. 67

CHAPTER VI BURNING OF LIMESTONE............................................................................................... 73

CHAPTER VII AMMONIATION OF SATURATED BRINE .................................................................. 109

CHAPTER VIII CARBONATION OF AMMONIATED BRINE............................................................. 127

CHAPTER IX WORKING OF CARBONATING TOWERS OR “COLUMNS” ............................... 149

CHAPTER X FILTRATION OF CRUDE SODIUM BICARBONATE (AMMONIA SODA):

COMPOSITION OF THE BICARBONATE ................................................................................................ 156

CHAPTER XI COMPOSITION OF MOTHER LIQUOR FROM CARBONATING TOWERS ......... 172

CHAPTER XII PHASE RULE IN TOWER REACTIONS: GRAPHICAL REPRESENTATION OF

AMMONIA SODA PROCESS........................................................................................................................ 182

CHAPTER XⅢ DECOMPOSITION OF SODIUM BICARBONATE BY CALCINATION .................. 192

CHAPTER XV BY-PRODUCTS FROM DISTILLER WASTE............................................................... 209

CHAPTER XVI ALKALI PRODUCTS OF AMMONIA SODA INDUSTRY......................................... 214

CHAPTER XVII POSITION OF AMMONIA SODA INDUSTRY ............................................................. 222

CHAPTER XVIII MANUFACTURE OF REFINED SODIUM BICARBONATE ................................. 229

CHAPTER XIX MANUFACTURE OF CAUSTIC SODA-CHEMICAL PROCESS ................................ 236

CHAPTER XX MANUFACTURE OF ELECTROLYTIC CAUSTIC, CHLORINE AND CHLORINE

PRODUCTS ..................................................................................................................................................... 259

CHAPTER XXI WET CALCINATION OF SODIUM BICARBONATE................................................... 285

CHAPTER XXII GENERATION OF POWER FOR AMMONIA SODA PLANTS .............................. 292

CHAPTER XXIII BOILER OPERATION AND CONDITIONING OF BOILER FEED AND COOLING
WATER IN AMMONIA SODA PLANTS ...................................................................................................... 310
CHAPTER XXIV SPECIAL REQUIREMENTS OF AMMONIA SODA INDUSTRY............................. 338

CHAPTER XXV CONTROL IN AMMONIA SODA PROCESS ................................................................ 354

CHAPTER XXVI LOSSES AND CONSUMPTION OF RAW MATERIALS IN AMMONIA SODA

PROCESS......................................................................................................................................................... 363

CHAPTER XXVII MODIFICATIONS AND NEW DEVELOPMENTS OF AMMONIA SODA PROCESS

........................................................................................................................................................................... 376

CAPTER XXVIII CHEMICAL ANALYSES AND TESTS IN ALKALI INDUSTRY I. SODA ASH
MANUFACTURE ............................................................................................................................................ 389

CHAPTER XXIX BEHAVIOR OF SODA ASH IN STORAGE .................................................................. 430

CHAPTER XXX LAYOUT, DESIGN AND LOCATION OF AMMONIA SODA PLANT.................... 436

CHAPTER XXXI CENTRIFUGAL PUMPS APPLIED TO AMMONIA SODA INDUSTRY .............. 449
 Chapter I

Introduction. Historical Survey of Alkali Industry and

Relation of LeBlanc Soda Industry to

Development of Other Chemical Industries

We shall first discuss the LeBlanc process and its historical relation to the development of
other chemical industries. In the early part of their work student of elementary inorganic chemistry
come across such process as the Weldon process, the Deacon process, and so forth; but probably
they do not realize the close bearing of these processes on the LeBlanc soda industry. In the
discussion which follows, the development of the LeBlanc soda industry will constitute a
historical treatment of the early chemical industries with the LeBlanc process as their nucleus.
With the passing of the LeBlanc process as an active method for the manufacture of soda, most of
these industries must necessarily also pass into oblivion; but the foundation laid by the LeBlanc
soda industry in paving the way for the more modern chemical industries is well worth
consideration.
The artificial soda industry originated in France, but for its development and application we
must turn to England. France, as a leading nation in Europe in the Eighteenth Century, consumed
considerable quantities of soda annually. While she was engaged with England in the so-called
“Seven Years war” and with practically all Europe in the Napoleonic Wars, the source of Spanish
barilla was closed to her. She had to find some way to get soda. In 1775 her Academy of Science
offered an award of 2400 livers for the invention of a practical process for the manufacture of soda.
It was known in a vague before this time, that soda could be made from common salt, for, in
1773,Scheele was able to get some caustic soda by digesting litharge in strong brine according to
the following reaction:
2NaCl+H2 O+xPbO 2NaOH+【(x-1)PbO】.PbCl2
Where x=2,3,4or5according to the concentration of the brine. The filtrate of brine from the
mixture of PbO and PbCl2 contains some caustic soda, and the conversion decreases as the
temperature employed is raised, owing to the greater solubility of PbCl2 at a higher temperature.
 Among several methods presented, there was one by Nicolas LeBlanc (1742-1806) who
outlined a method of manufacture starting with common salt. His process was so promising that in
1783 the French Academy of science promised to award him the prize .It was however, never
given him. He obtained a patent in 1791, and, being physician to the Duke of Orleans, he received
200,000 livers from him for the construction of a plant in St. Denis, not far from Paris. In 1793,
the French Revolutionists guillotined the Duke. Being desperate to obtain soda, the Committee of
Public Safety of the Revolutionists’ Government compelled LeBlanc to make his patent public
without remuneration except by assigning to him the St. Denis Works from the confiscation of the
Duke’s property (Fig.1) Deprived of his working capital, LeBlanc could not operate the works and

FIG 1 LeBlanc forced by French Tribunal to make his patent public.

was forced to shut it down. He had to spend the latter part of his life in an almshouse where, in his
misery, he committed suicide in 1806.Thus ended the lift of a man whose invention was of such
great value to the world. His original proportion of 100 pts. By weight of salt cake: 100 pts. Of
limestone: 50 pts of coal was only slightly modified to 100 pts. Of salt cake: 100 pts. Of
limestone:351 /2 pts. Of slack coal in the course of the long period of its application. All
improvements that were subsequently made on his process had to do with mechanical equipment.
Rarely has there been a process so nearly perfect in form as that first discovered by LeBlanc. In
spite of the fact that he was the founder of a process so beneficial to mankind and so lucrative to
people who employed his process LeBlanc himself died poor and unknown. His country was not
benefited by his discovery to the extent that it might have been, and his name was for the time
even forgotten by his own countrymen. It was not until 1886,after millions of tons of soda ash had
been made in England and Germany by the LeBlanc process, that they began to erect a statue in
his honor at the Conservatory of Arts and Trades in Paris (Fig.2)
 FIG 2 LeBlanc statue in Paris.
Because of the prolonged wars into which France had been plunged, the soda industry, like
many other industries, could not be developed to any extent. In England it had a better chance,
although the Napoleonic wars left England impoverished like many other European countries.
From the time when peace was declared at the Congress of Vienna in 1815,she had to levy a very
heavy tax of 30 per ton on salt. Under this tax, it would have been impossible for her soda
industry to exist. After eight years, however, the British government wisely decided to repeal the
law, and the year 1823 saw the tax removed. From this time on, the soda industry went ahead by
leaps and bounds quite in contrast to the conditions in France. James Muspratt (1793-1886), an
Irishman, who had served as an apprentice in a fine-chemical factory, first started a sulfuric acid
industry at Liverpool in 1822 and then, taking advantage of the tax exemption on salt, built a
LeBlanc soda plant there in 1823.The undertaking proved profitable. After enlarging his plant to
its fullest extent, he built a second LeBlanc soda plant at Newton, which was later removed to
Widens, and subsequently a third LeBlanc soda plant at Flint. Meanwhile other LeBlanc soda
plants also sprang up at Widens, St. Helens, Runcorn, the Tyne, Glasgow, and in other districts,
where the necessary raw materials, sulfur, salt, limestone and coal, were available in abundance.
Thus, the great development of the LeBlanc soda industry in Great Britain was made possible by
the repeal of the salt tax in 1823.
The period from 1825 to 1890 saw the LeBlanc process occupying a leading position. All
artificial soda was made, and made in large quantities, by this process. In 1861 Ernest Solvay
(1838-1922) rediscovered and perfected an old process now known as the ammonia soda process
in 1863 with his brother Alfred he constructed a works at Couplet, Belgium, near Charleroi, which
began to operate in 1865,using this process. Only after encountering a series of serious difficulties,
which threatened to bring failure to his undertaking, did he finally make a success of it. The works
was gradually improved, enlarged, and remodeled. In 1872 the capacity was brought up to 10 tons
of soda ash a day, which proved to the world the success of his process beyond any doubt. In that
year he designed a large soda plant at Dombasle near Nancy, France, on a scale considered quite
gigantic at that time. Dr. Ludwig Mond, an old LeBlanc soda man, came over to visit Solvay, with
the result that the latter’s process was allowed to be introduced in England. Brunner, Mond & Co.
was then formed and the first works erected at winning ton, near Northwich, England, in 1874.
Like the LeBlanc soda industry, the ammonia soda industry developed by leaps and bounds.
The new soda industry, however, had not only to meet many difficulties inherent in its operation
but also to meet competition from the well-established LeBlanc soda process. However, it was
able to hold its own in the face of competition, because of the simplicity of raw materials needed,
of the low cost of production, and of the very high purity of the soda ash produced. From about
1885 on, the curve of LeBlanc soda production began to take a downward course as the result of
competition from the new process the price of soda ash declined from some 13 per ton to little
more than 4 per ton .The LeBlanc soda manufacturers had to dispose of their soda at a loss and to
make up the loss from the sale of bleaching powder. The Bleaching Powder Association was then
formed in England as an organ of the LeBlanc soda manufacturers to maintain the market price of
bleaching powder. In 1890 this culminated in a then colossal combine of practically all LeBlanc
soda interests in Great Britain into one company known as The United Alkali Co., Ltd., consisting
of forty-eight works (45 chemical and 3 salt works), of which 42 were in England, 4 in Scotland
and one each in Ireland and Wales. From this, some idea can be obtained as to how deadly the
struggle was between the new and the old processes. Heretofore, there had been a certain definite
balance the quantities of soda and bleaching powder produced. With the advent of the ammonia
soda process, only the alkali was produced without its counterpart in chlorine. This balance was
disturbed, and it was difficult for LeBlanc soda manufacturers to adjust themselves to meet the
new situation.
Despite this huge combine, the position of LeBlanc soda was still untenable. To avoid direct
competition, The United Alkali Co. turned its attention to the manufacture of other chemicals,
such as sulfuric acid, monistic acid, bleaching powder, etc., starting from pyrites as raw material.
Meanwhile, Brunner, Mond & Co. grew to a position of mastery in the soda trade of England. Just
as James Muskrat took over Leblanc’s process and developed the LeBlanc soda industry to the
greatest benefit of England, so Ludwig Mond introduced Solvay’s process and developed the
ammonia soda industry to such an extent that Brunner, Mond & Co. (at present merged into
Imperial Chemical Industries, Ltd.) controls the soda market not only in England but also in many
parts of the word, through the export of its large surplus of soda ash.
While the LeBlanc process for soda ash was forced to decline, its use for caustic soda
manufacture possessed some advantages. First, the soda liquor from the lixiviating vats could be
used directly for cauterization, and secondly, the lye as obtained already contained as much as 20
per cent of the causticity of the total alkali present, so that only 80 per cent of the normal amount
of lime was required for cauterization. With the advent of the electrolytic process, however,
another serious blow was dealt to the LeBlanc process. Here in one operation were obtained
chlorine and caustic soda-and these in a pure and concentrated from –while these two main
products of the LeBlanc process were obtained only after a considerable number of intricate steps.
From 1890 on, the LeBlanc process suffered a rapid decline .The Word War (1914-1918)
found the position of the LeBlanc process for soda manufacture still untenable. Because of the
demand for sulfuric acid for the manufacture of munitions, and because of the threatening
shortage of coal, the British government during the War imposed restrictions on the use of these
important raw materials. The LeBlanc soda works was hard hit, and in 1915-1916 The United
Alkali Co. practically stopped in making concentrated sulfuric acid, picric acid, poisonous gases
and other war chemicals for the British Government.
After the war, the chemical industries had to readjust themselves. Competition among the
different nations for the world market was resumed with even greater vigor. Efficiency was the
watchword; and the LeBlanc process, which involved much labor, high cost of manufacture, low
purity, and a complex line of products, had to give way to a newer, more efficient process. In 1923,
The United Alkali Co. decided to scrap the last portion of its LeBlanc soda equipment, and the
death verdict for the Leblanc process for soda ash manufacture was finally delivered. On the
European continent similar conditions occurred, although the remnants of the LeBlanc process in
many places did not even survive up to the War. Born of necessity during the War in France and
superseded because of the necessity caused by the World War I, the LeBlanc process, which had
served mankind for fully a century, had completely fulfilled its mission. We shall now dwell at
some length upon the relationship between the LeBlanc soda industry and the other chemical
industries.
Though the LeBlanc process was discontinued, LeBlanc soda works remained .The United
Alkali Co. in England became engaged in the manufacture of various heavy and fine chemicals.
The change consisted merely in replacing the LeBlanc process for soda manufacture by the
ammonia process at Fleetwood. And now this United Alkali Co. became indeed the largest
manufacturers of sulfuric acid in England, as if the name of the company were a misnomer .The
reason is to be found in the nature of the LeBlanc process. It requires for one of its principal raw
materials salt cake, or sodium sulfate, which in turn has to be made from common salt by
treatment with sulfuric acid. Thus, sulfuric acid indirectly becomes its important raw material .The
LeBlanc soda manufacturers generally made their own sulfuric acid. Thus we find that they were
also large manufacturers of sulfuric acid (chamber acid). From salt and sulfuric acid are obtained
salt cake and hydrochloric acid, which latter is in itself a valuable chemical. Hence we find that
LeBlanc soda manufacturers are necessarily manufacturers of hydrochloric acid .As the muriatic
acid as such did not find much use at the earlier time, most of it was converted to chlorine; from
this, bleaching powder was made which found very extensive application in the textile and paper
industries. Thus the bleaching powder industry was necessarily connected with the LeBlanc soda
industry. The salt cake required by glass manufacturers was also supplied by LeBlanc soda
manufacturers .The CaS residue known as “tank waste” proved to be quite a nuisance, for there
was so much of it that its disposal become a problem. For every ton of soda ash produced there
were more than 11 /2 tons of this waste left. So foul was its odor and so objectionable was the
pollution of surface water by it that its proper disposal became a matter of no small concern to the
community. Added to this was the fact that it contained the entire sulfur product. The high price of
sulfur, largely due to taxation, induced the soda manufacturers to seek a cheaper source of sulfur,
and pyrites were employed, not only for its sulfur but also for the iron and copper contained in it.
Here, then, a group of important products were obtained: namely, sulfuric, acid, hydrochloric acid,
salt cake for the glass industry, bleaching powder for the textile and paper industries, sulfur, and
lastly iron and copper-all clustered together around this LeBlanc soda industry.
Of great interest to students of chemical engineering was the apparatus developed in
connection with the LeBlanc process. The hydrochloric acid fumes from the salt cake process
discharging to the air through the chimney worked havoc on the vegetation in the vicinity. The
British Government had laws known as Alkali Acts (1863,1874), restricting the content to within
2 grains HC1 per cu. ft. A coke-packed washer was invented to scrub the gas. This washer was the
basis of all types of scrubbers now used in other branches of the chemical industry. In the black
ash process, a revolving furnace called the “revolver” took the place of the old hand-operated
furnace. This cut down the strenuous labor required of the furnace workmen and greatly increased
the output. This “revolver” may be called the forerunner of the present rotary kilns employed in
one style or another in so many fur-acing operations. In the final drying of the monohydrate
Na2 CO3 . H2 O for soda ash, and also in the salt cake process, a mechanical furnace with a
revolving bed, called Mc Tear’s furnace, was employed. For evaporating. (Or concentrating) the
black ash liquor, and for crystallizing the monohydrate, an open Thelen pan with revolving
scrapers did away with much difficulty arising from scale formation at the bottom of the pan by
the monohydrate. Without efficient stirring, a hard scale would be formed in the pan, causing the
iron shell to be burned out. This Thelen pan in its closed form, which had oscillating, instead of
revolving scrapers, was later introduced into the ammonia soda industry in England for calcimine
the bicarbonate and for recovery of the CO2 gas. Likewise, Mc Tear’s mechanical furnace was
introduced for finishing soda ash or for making dense ash in the ammonia soda industry. Even up
to the present time, some ammonia soda works are still using these pieces of machinery, especially
in European practice. Finally, the Shank lixiviating system for black ash liquor embodied the
counter-current principle that has been so widely applied in dealing with leaching and extraction
processes in all modern chemical industries. All these important pieces of apparatus and some
minor ones not mentioned above, our young modern chemical industries owe to the old LeBlanc
soda industry.
The LeBlanc soda industry produced a number of brilliant men whose contribution to
chemical technology, either in the discovery of a chemical process or in the invention of chemical
equipment, laid the foundation for many modern chemical industries. These men helped to make
the LeBlanc soda industry what it was. The very difficulties these pioneers encountered and their
methods of attack solved once for all, or offered try. In the early salt cake process, hydrochloric
acid fumes were allowed to discharge into the atmosphere through a chimney, and heavy damages
were collected from these manufacturers. William Goss age in 1836 introduced a coke-packed
scrubbing tower for these fumes. The tower was built of sandstone slabs, boiled in tar and clamped
together with steel straps. Water was sprayed from the top and trickled through the coke to absorb
the up going fumes. The apparatus was primarily devised to prevent the escape of acid fumes to
the air without any idea of recovering any useful product. But it recovered a valuable product,
which was one of the factors that prolonged the life of the LeBlanc process in the face of a strong
competition from the ammonia soda process.
The market for hydrochloric acid at that time was not very strong, but there was a great
demand for bleaching powder. The bulk of the hydrochloric acid had to be converted to bleaching
powder. In doing so it was necessary to convert the hydrochloric acid first to chlorine gas. The
usual method then in use was to oxidize it with manganese dioxide (pyrolusite) according to the
following reaction:
MnO2 +4HCl MnCl2 +Cl2 +2H2 O
It is thus evident that the manganese chloride resulting from the reaction was no longer useful for
the work. Further, one–half of the chlorine was not made available this way. The consumption of
manganese dioxide also was a considerable item of expense. Walter Weldon from 1866 to 1870
devised a process for the oxidation of manganese chloride to manganese dioxide by means of lime,
thus affecting its recovery. The process which bears his name consisted in treating the still liquor
with lime, heating it to 55°-60°C. by steam in the presence of excess lime, and blowing in air.
Thus, almost all the manganese was recovered in a useful form, and strong calcium chloride liquor
resulted, from which calcium chloride liquor resulted, from which calcium chloride powder could
be manufactured rather economically for use in refrigerating plants.
So far as the utilization of chlorine was concerned, oxidation by manganese dioxide made
available only one-half of the total available chlorine as mentioned above, with an efficiency of
only 50 per cent. Besides, manganese dioxide was recovered in the form of calcium magnetite,
CaMn2 O5 , which consumed acid to free the MnO2 from CaO. An improvement on this method was
discovered by Henry Deacon of the firm of Gaskell, Deacon & Co., Whidnes, in 1870, who used
oxygen direct firm the air for oxidation in the presence of copper chloride (CuCl2 ) as a catalyst.
Though the chlorine gas obtained was rather dilute and demanded a special apparatus for the
manufacture of bleaching powder, all the chlorine present in the manganese chloride was made
available, and, further, the excessive alkalinity due to CaO was avoided.
The process for making bleaching powder from chlorine gas and lime, which then took so
much hydrochloric acid from the LeBlanc process manufacturers was discovered by Charles
Tennant, a Scotch merchant, Who, in 1788 first made a solution of “chloride of lime.” Later he
developed a solid powder. The chloride of lime bleached cloth in a few hours and accomplished
what bleaching in the sun’s rays could only do in months. In 1797 Tennant set up a chemical
works in St.Rollox, Glasgow, and in 1799 he received a patent for making solid chloride of lime,
to which he gave the name “bleaching powder.” This so revolutionized the textile industry that
bleaching powder became an important article, and considerable hydrochloric acid from the
LeBlanc process was converted into chlorine for the manufacture of bleaching powder.
In the LeBlanc process the black ash, after leaching. Left a residue known as “tank waste.”
The waste began to accumulate so fast (about 11 /2 tons every ton of soda ash made) that its
disposal became a serious problem. Here came Dr. Ludwig Mond (1862), who made use of air
oxidation to free the sulfur. This is Mond’s process for the recovery of sulfur from tank waste. The
name of Mond is, however, usually associated with the development of the ammonia soda industry
in England, though, he was originally one of the LeBlanc soda men. His work in the development
of the ammonia soda industry has overshadowed all this .At the same time another promising
process was proposed by Schaffer, who used magnesium chloride in conjunction with air:
2CaS + 2MgCl2 + O2 2CaCl2 + 2MgO + S2
The one process that was used successfully on an extensive scale, however, is that known
after its inventors as the Chance-Claus process? In this process the waste was treated with carbon
dioxide gas in a battery of carbonators whereby hydrogen sulfide gas was liberated and calcium
sulfide transformed into calcium carbonate according to the following reactions:
CaS + CO2 +H2 O CaCO3 +H2 S
CaS + H2 S Ca(HS)2
Ca(HS)2 +CO2 +H2 O CaCO3 +2H2 S
The hydrogen sulfide gas was burned to free sulfur by means of a carefully regulated supply of air
in the presence of hydrated iron ore (ferric oxide) as a catalyst, thus:
2H2 S +O2 2H2 O +S2
The sulfur reclaimed from the tank waste is in a very pure from. It is thus possible to reclaim from
the crude sulfur originally present in the pyrites that went to make sulfuric acid, pure free sulfur,
commanding a good price on the market.
In connection with the burning of pyrites for sulfur dioxide gas, an iron oxide rich in copper
resulted. Long maid and Henderson devised a method of extracting copper from it by calcimine it
with salt, leaching out the copper chloride, and precipitating copper with iron scrap. This is known
as the Long maid and Henderson Wet Copper Process.
In an attempt to manufacture soda by Leblanc’s process without starting from sulfuric acid as
one of the raw materials, Hargrove’s in 1870 discovered a method of making sodium sulfate from
sulfur dioxide and common salt .A mixture of sulfur dioxide ,air, and steam is passed through
layers of especially prepared salt blocks at a temperature of 400-500 ., When the following
reaction takes place:
4NaCl + 2SO2 + 2H2 O + O2 2Na2 SO + 4HCl
This is known as the Hargrove’s process. As usual, hydrochloric acid is formed in the reaction, but
without the use of sulfuric acid .In this country at present, because of the large demand for sodium
sulfate for Kraft paper manufacture and because of the introduction of the synthetic method for
manufacturing hydrochloric and nitric acids, certain attempts have been made to revive the
Hargrove’s process for the production of sodium sulfate, which was formerly a by-product from
the treatment of salt or Chile saltpeter with sulfuric acid for the production of hydrochloric or
nitric acid.
All these processes were developed from the necessities of the LeBlanc soda industry, and
their inventors were really pioneers in the application of chemical principles to industry, laying the
foundation of what is now known as the chemical technology of heavy chemicals.
Although the LeBlanc process for soda ash manufacture has now been given up in favor of
the newer ammonia process, the development of the sulfuric acid industry on the one hand, and
the production of hydrochloric acid, bleaching powder, etc., on the other, are the direct outcome of
the LeBlanc process. Heavy–chemical industries in general are lucky enough to have had a start in
the LeBlanc industry. For, unfortunately or fortunately, the LeBlanc process involves the
manufacture of all these chemicals, and what were originally by-products later became the
principal products. It can be truly said that the history of modern inorganic chemical industry owes
its beginning to the LeBlanc soda industry.
For the sake of clearness to the reader, the diagram in Fig.3 is given to show how varied were
the products from the LeBlanc industry. Products enclosed in rectangles were those appearing in
commerce.
In contrast with these older methods, it may be mentioned briefly by way of digression that
modern methods utilize these waste sulfur gases from pyrites roasters for the manufacture of
conc entrated sulfuric acid, not by the chamber process but by the vanadium contact process. For
this purpose, mechanical roasters such as the Heresy off furnace, the Wedge furnace, etc. are used,
nd the gases from these roasters are maintained at 7 per cent or higher SO2 by volume. This
strength of sulfur gas is adequate for efficient conversion by the vanadium mass for the
manufacture of contact acid. The waste gases are settled, cooled, and scrubbed by passing through
an elaborate purification system to remove dust and moisture, led through a wet Cottrell system to
remove SO3 mist, and finally dried through a drying tower. The dried and cleaned gas is then
passed into the vanadium contact mass through a system of heat exchangers. From the vanadium
converters, the gases are led through absorption towers to make 98-99 percent H2 SO4 or 105-109
per cent ileum.

FIG 3 Diagram of LeBlanc soda industries.

If the utilization of such waste gases for the manufacture of sulfuric acid is not sufficient to
take care of all the gases available, due perhaps to the limited demand for sulfuric acid in the
country, the remaining waste gases may be converted to elemental sulfur, not by the LeBlanc
procedure, but by a direct catalytic reduction. This has been done in many modern plants, e.g., in
that of the Consolidated Mining and Smelting Co., Trail, B. C.
For this recovery of sulfur, the present reduction method of SO2 gas to elemental sulfur
proves far superior to the old Chance-Claus process described above. Briefly, there are at present
two main processes for the recovery of sulfur from the waste gas. These are the I.C.I and the
Bolden processes*. The I.C.I process involves first concentrating the SO2 gas from the waste gas,
purifying the gas, and then liquefying it to liquid sulfur dioxide; or, alternatively, reducing it to
elemental sulfur by the water-gas method with the help of a catalyst. Concentration of the gas is
achieved by the absorption of the weak gas in a basic aluminum sulfate solution, having a pH
value of 3.5,or 40 per cent basicity, the solution containing a small percentage of phosphoric acid
as a stabilizer to help keep the Al2 O3 in solution. Regeneration of SO2 gas is then effected by the
liberation of SO2 from the basic aluminum sulfate solution by boiling at about100. C. As the SO2
gas has a tendency to be oxidized to SO3 in the presence of oxygen; pulverized limestone or lime
is used to neutralize the acid formed so as to maintain the proper pH value in the solution.
Gaseous sulfur dioxide is reduced by passing it through a water-gas generator at 1200. ,
whereby many complex many complex products are formed, as shown by the following equations:
1 4C +2SO2 4CO +S2 5 2CO +S2 2COS
2 2C +2SO2 2CO2 +S2 6 C +S2 CS 2
3 C +H2 O CO +H2 7 4H2 S +2SO2 4H2 O +3S2
4 2H2 +S2 2H2 S 8 2SO2 +2CS 2 2CO2 +3S2
9 4CO +2SO2 4CO2 +S2
Catalytic action will convert all these complex products to elemental sulfur. The catalyst consists
chiefly of ferric oxide and aluminum oxide. By this process either liquid sulfur dioxide or
elemental sulfur can be obtained.
The Bolden process, independently worked out by Swedish research chemists, is a direct
reduction of the waste gas into sulfur in one step without concentration or purification of the SO2
gas. This process is particularly suitable for me manufacture of sulfur directly from the waste gas,
where reasonable control of the smelter gas is possible to keep the SO2 content above 5 percent.
These two processes of recovering sulfur from waste SO2 gases are so efficient nowadays
that it is possible to recover sulfur of high purity from the waste gases from roasters or smelters,
comparable in quality with Louisiana and Texas sulfur.
Other processes, such as McCluskey’s process, etc., are all based on the reducibility of SO2 by
solid carbon or carbon monoxide or water gas products according to the following reactions:
2C +2SO2 2CO2 +S2
4CO +2SO2 4CO2 +S2
4H2 +2SO2 4H2 O +S2
CH4 +2SO2 CO2 +2H2 O +S2
2C2 H4 +6SO2 4CO2 +4H2 O +3S2
Along with many side reactions. These reactions are carried out in the presence of a catalyst.
Various catalysts have been suggested, and even solid charcoal in the water gas generator may be
considered as such a catalyst.
In summarizing, we might add that in England and on the continent, LeBlanc’s process for
soda manufacture has now disappeared, while in the United States, soda ash from the very
beginning has been made exclusively by the ammonia process. At present, one part of the LeBlanc
process is left in the form of hydrochloric acid manufacture:
NaCl +H2 SO4 NaHSO4 +HCl
NaHSO4 +NaCl Na2 SO4 +HCl
and another, leaving out the limestone, is used in the manufacture of sodium sulfide:
Na2 SO4 +2C Na2 S +2CO2
A slightly modified form, with the addition of pure quartz sand in place of limestone, is employed
in the water-glass manufacture:
2Na2 SO4 +2C +O2 +2SiO 2 2Na2 SiO2 +2CO2 +2SO2
Na2 SiO3 +xSiO 2 Na2 O. (x +1) SiO 2
But the complete process for the manufacture of soda ash has sunk into oblivion. In the
matter of purity, cost of manufacture, labor involved in handing, and simplicity in the line of
products obtained, LeBlanc’s soda process cannot compete with the ammonia process. And for the
production of caustic soda and bleaching powder, the present electrolytic method possesses every
advantage: nobody nowadays would think of manufacturing bleaching powder from hydrochloric
acid through the many intermediate stages represented by the Weldon or Deacon process. Indeed,
the increased demand for chlorine has greatly spurred electrolytic production to such an extent that
the quantity of caustic soda produced as a joint product from this electrolytic method has now
equaled or exceeded that produced by the lime process, as will be seen in Chapters XIX and XX.
 Chapter II
Natural Soda
Before the advent of the LeBlanc soda proems, all soda came from natural sources, either
vegetable or mineral. Vegetable soda made from the ashes of certain plants or seaweeds must have
been known to the ancients long before recorded history. Primitive people burned the stalks of
certain plants to ashes, which they leached with hot water, obtaining brown-colored lye for
domestic laundering. Gradually soda from these sources became a staple commodity in commerce,
and we hear notably of Spanish barilla at the time of the French Revolution. There were many
such commercial varieties from different localities, and they had widely varying alkali contents.
Undoubtedly in some cases (such as kelp or varies) the alkali content is potash (“Pot-Ash”) rather
than soda. These products went under the names of barilla or bourdon in Spain; Blanquette, sailor
or varies in different parts of the coast of France; and kelp along the Atlantic coast of Great Britain.
Their approximate soda contents are given in Table 1.

Table 1. Natural Soda from Vegetable Souroces (Seaweeds) Before the Days of LeBanc.
Place of Occurrence Name in Commerce Na2 CO3 (%)
Southern Spanish Coast Barilla 25-30
(Alicante, Malaga)
Southern Spanish Coast Bourdin 20-25
(Cartagena)
Northern French Coast Varec 3-8
(Cherbourg)
Southern French Coast Blanquette 4-10
(Argues-Mortes)
Southern French Coast Soude de Narbonne(salicor) 14-15
(Narbonne)
+Western Scotch and Irish Coast Kelp 10-15
* Different authorities give different figures for the soda content. The plant ashes contained from 2 to 40
per cent of their weight of soda.
+ At present found also on the coasts of Japan and on the Pacific cast of the United States.

By far the greatest quantity of natural soda comes from mineral sources-so much so that
natural soda is now understood to be of mineral origin. Natural soda from mineral sources may
usually be identified by its content of more or less varying proportions of sodium bicarbonate and
sodium sulfate; where as artificial (ammonia) soda always contains a small proportion of iron.
Natural soda occurs in many places on the surface of the earth. It is usually found (1) in crystalline
form as sale soda, Na2 CO3.10H2 O, known as matron; (2) in powder form as an efflorescence on
top of the crystalline mass ,monohydrate,Na2 CO3.H2 O, known as thermonatrite; (3) in fine
crystalline as the sesquicarbonate,Na2 CO3 .NaHCO3 .2H2 O, known as Trina in Egypt or urea in
Colombia, Venezuela and Mexico; and (4) in the form of soda brine containing sodium carbonate,
sodium sulfate, and sodium chloride among other salts in solution. The deposits are generally
found in low valleys where there is little rain and where the air is dry.
The origin of natural soda is rather difficult to determine .It seems that many forces in nature
are at work, and in conjunction with local geologic formations, these give rise to peculiar deposits
found in different parts of the world. In general, it seems likely that natural soda owes its origin to
the decomposition of rocks or soil by the agencies of air, moisture, heat, and pressure, followed by
subsequent chemical changes, the products being leached by water, converted to the carbonate by
atmospheric carbon dioxide, and concentrated by natural evaporation. Volcanic activities with
their attendant heat and pressure effects have been considered to be one of the most powerful
agencies responsible for the decomposition of rocks yielding the soluble salts of sodium (silicate,
chloride and sulfate) which are dissolved by water, furnishing raw materials for the subsequent
chemical changes mentioned above .The following is an enumeration of the several ways, in one
or more of which a particular deposit may have been formed:
(1) From the deposition of soda leached out by water from rocks rich in soda, which
were undergoing decomposition (as in the case of the old Laotian Lakes in the State
of Nevada).
(2) From volcanic action and subsequent solution of the volcanic ash by water, giving
rise to soda brine (as in Owens Lake, California).
(3) From the decomposition of sodium silicates by atmospheric carbon dioxide setting
free silica acid. (Some of the Egyptian soda is believed to have been thus formed.)
(4) From sodium sulfate which is reduced by organic matter (such as algae) to sulfide.
Which then is decomposed by atmospheric carbon dioxide forming sodium
carbonate and liberating hydrogen sulfide and sulfur. (The matron lakes of Egypt
are thought to have been formed in this way)
(5) From sodium chloride or sulfate, which, in contact with limestone, is slowly acted
upon by calcium bicarbonate, Ca (HCO3 )2 ,formed from calcium carbonate and
atmospheric carbon dioxide. Double decomposition then takes place, yielding
sodium bicarbonate and calcium chloride or sulfate, the sodium bicarbonate giving
sodium carbonate by decomposition. This accounts for some of the Hungarian
deposits and for the formation of Gay-Lussite,Na2 CO3 .CaCO3 .5H2 O in Taboos-Nor
(Kirin), Loganville (Venezuela), and certain soda lakes in Carson Desert (Nev.)
From the above it will be evident why natural soda is frequently accompanied by more or
less sodium sulfate and sodium chloride. What ever may be the cause; there is seldom one agency,
alone which brings about the formation of these deposits in the different parts of the world. Often
in one locality there is more than one deposit. For example, Gay-Lassie and Trina occur together
in the Loganville Valley, Venezuela.
* T.M.Chatard, Bulletin, U.S.Geol. Survey, No. 60, pp. 89-95.
+ Moses and Parsons, “Mineralogy, Crystallography and Blow-pipe Analysis,” p.425 (1916).
Natural soda contains considerable sodium bicarbonate, mere or less sodium sulfate, some
sodium chloride, potassium chloride, borax, and a small proportion of insoluble matter. Crystals
are gradually formed from soda brine, which becomes concentrated by natural evaporation.
During cold weather, crystals of sale soda separate out from solution. By efflorescence, the
crystals on the exposed layers crumble to the powdered monohydrate, from which by weathering
action under the influence of atmospheric carbon dioxide and moisture, sesquicarbonate or Trina
results. Such processes continue, so that we may find three (Sal soda, monohydrate, and
sesquicarbonate) existing together in the same deposit.

Table 2 Some Better Known Deposits of the World.

Africa Lower Egypt: Wady Natrun

British East Africa: Magadi Lake
Former German East Africa: Moshi
Libya: Fezzan
The Americas United States:
California: Owens Lake (Inyo County ); Searles Lake, Trona
(San Bernardino County ); Borax Lake (Lake County)
Nevada: Ragtown Soda Lakes (near Carson Sink)
Wyoming: Union Pacific Lakes, near Green River and along Union
Pacific Railroad)
Oregon: Abert Lake
Mexico: Lake Tezcoco
Venezuela: Lagunilla Valley
Chile: Antofagasta
Siberia: Chita and Lake Baikal Region, Barnaul, Slavgorod
Armenia: Araxes Plain lakes
India: Lake Looner, etc.
China:
Outer Mongolia: Various “Nors”
Sui-Yusn: Cha-Han-Nor, Na-Lin-Nor, Pa-Yen-Nor
Asia Heilungkiang: Hailar, Tsitsihar
Kirin:Fu-U-Hsien, Taboos-Nor
Liao-Ning:Tao-Nan Hsien (Polishan lake , Tafusu lake , etc.)
Jehol: Various soda lakes
Chahar: Cheng-Lang-Chi
Shansi: U-Tsu-Hsien
Shensi: Shen-Mu-Hsien
Kansu: Ning-Hsia-Hsien
Tibet: Alkali deserts
Europe Russia: Caspian Sea region
Hungary: Szegedin district
The list in Table 2 is by no means exhaustive, but it gives some of the better known representative
deposits.
Regarding the composition of natural soda, analytical results from different sources vary a
great deal .The composition differs with different layers of the same deposit and with varying
climatic conditions. Sampling in extensive deposits (except in the case of soda brine from
individual alkaline lakes) also has a bearing on the results .The results of analyses of natural soda
samples or of soda brine residues are given in Table 3.
 At present natural soda as a whole does not play an important role in supplying the world
market, and probably it will never do so except in a limited area surrounding the locality where the
deposit occurs. With the possible exception of the Malady and California deposits, the difficulties
of transportation have been an unfavorable factor. Then again, because of the low purity of natural
soda due to the presence of large amounts of sodium bicarbonate, more little sodium sulfate,
considerable amounts of sodium chloride, and sometimes also borax, and because of the small
scale on which it can be operated, it cannot compete with ammonia soda in industries, which
require a high-grade product. Any attempt at refining it, e.g., by dissolving, settling, recarbonating,
and fur acing the product, would involve almost the same cost of manufacture as ammonia soda.
In the case of California deposits, however, the operation has held its own, because of the
many valuable by-products obtained. Among those by-products may be mentioned potassium
chloride, borax, sodium sulfate, bromine, etc. It was usual to subject the natural product to a
simple process of crushing, washing, and drying; but this treatment did not remove the many
impurities present. However, a fractional crystallization method has been developed in the plant of
the American Potash & Chemical Crop. Using Sealers Lake underground brine based on the
principles of the phase rule, whereby a double salt known as Bakelite is obtained, from which now
a very pure soda ash is produced. Since 1934, with the improved quality of soda ash produced by
this Corporation comparable to ammonia soda ash, natural soda has become a real competitor in
the soda market on the Western Coast of the United States.
Diversion of the Owens River in Eastern California to Los Angeles aqueduct in 1917 dried
up Owens Lake, to which the river formerly flowed. This lake and the Sealers Lake are now the
most important sources of natural soda in the United States. Numerous other soda lakes exist on
the North American continent, but they are not important sources of supply. The first company for
exploiting natural soda was organized by C. F. S. Wrinkle*, with the financial help of D. O. Mills,
on Owens Lake in the early 1880’s, about the same time as the Solvay process for ammonia soda
manufacture was introduced into America. This concern, under the same of the Into Development
Company, was operating continuously at Keeler until 1920,three years after Owens River was
diverted to Los Angeles aqueduct. For many years since 1885, this company concentrated brine
from Owens Lake by solar evaporation in huge clay vats, covering acres of ground. These vats
(fields) were filled with brine 6-8 inches deep, and the precipitated trona was harvested once a
year in the form of crust about 3 /4 thick. The product was claimed in a reverberate furnace, but
was impure, containing only 95 per cent Na2 CO3 . In 1912, Noah Wrinkle, one of the engineers of
the Into Development Co., established another plant at Keeler on Owens Lake under the name of
Natural Soda Products Co. to exploit the crude soda from the lake. Some bicarbonate of soda and
natural Trina were made by this company. The purity was 97-98 per cent Na2 CO3. Later, natural
Soda Products Co., introduced Oliver filters filters for filtering sodium bicarbonate and rotary
dryers for calcimine it to light ash. Recently this company built a new plant just south of Keeler,
having a daily capacity of 100 tons of soda ash. Much of its output was shipped to Japan and
Sweden for some years.
In the Natural Soda Products Co.’s plant at Keeler on Owens Lake, the brine was
concentrated by solar evaporation and enriched by solid Trina. It was carbonated in wooden
towers, using CO2 gas from limekilns burning native dolomite. Sodium bicarbonate in the slurry
from the towers was filtered in the usual way and claimed light ash having the following
composition:

Table 4. Composition of Light Ash at Keeler. *

Na2 CO3 98.16%
NaHCO3 0.90
Na2 B4 P7 0.33
Na2 SO4 0.33
SiO2 0.25
NaCl 0.05
Fe2 O3 and Al2 O3 0.02
CaCO3 trace
MgCO3 trace
*G. R. Robertson, California Desert Soda, Ind. Eng.Chem.23, 478 (1931)
To make dense ash, the light ash was melted in an oil-fired furnace and atomized by means of
compressed air into an insulated space lined with a fire –proof material. This gave globules of
soda ash which is very dense and particularly favored by glass manufacturers, although the
process is somewhat antiquated. The presence of borax in the soda ash is not unwelcome to glass
manufacturers. A certain quantity of ”water ash ” is also made at this plant by the same method as
used by the ammonia soda manufacturers in the East, and this of course yields a less dance ash. A
combination of the two dense ash products gives several grades of dense ash, having the bulk
densities lying in between.
About 1908, the California Trina Company was organized and established on Sealers Lake.
In 1913, this company was succeeded by the American Trina Corp. In the summer of 1936, the
corporation’s name was again changed to the present American Potash & Chemical Corp. It is
considered to be the largest and the most progressive of its kind.
Before this time F. M. Smith founded the Wear End Chemical Co., on the southwest shore of
Sealers Lake at West End. This company car the brine. Some borax comes down with the
bicarbonate, thus yielding a somewhat lower grade of soda ash, containing borax as one of the
impurities. Borax then is recovered by crystallization from the mother liquor after the precipitation
and filtration of sodium bicarbonate.
 Early in 1917, because of demand for soda ash during the World War I (1914-1918), a group
headed by Great Western Electro-Chemical Co. built a natural soda plant on the southwestern
shore of Owens Lake. The plant was built at Cart ago and was known as the California Alkali Co.
This plant built limekilns and used also wooden towers, rotary vacuum filters and rotary soda
dryers, and produced marketable soda ash both in light and dense form. In 1924, this company
was merged with Into Development Company and became the Into Chemical Co. The Keeler
Plant of the Into Development Company was then leased to Natural Soda Products Co., which
operated it more or less steadily. The Cart ago plant of the California Alkali Company
discontinued operation in 1932.
In 1926, a Los Angeles group built a plant at Bartlett on the west Shore of Owens Lake about
10 miles south of Lone Pine under the name of Pacific Alkali Company. This company pumps
brine through 21 /2 miles of 14 pipes into large open vats which concentrate it by solar
evaporation to 12-14 per cent soda. The soda brine is then carbonated and the slurry drawn off and
centrifuged. The cake from the centrifuges is claimed in a Herresh off furnace into soda ash, and is
screened and packed for shipment. The mother liquor from the centrifuges is cooled to crystallize
out borax and the crude borax crystals are filtered. This filter liquor is returned to the lake. The
plant produces about 1000 tons of soda ash and 2000 tons of borax per season.
In 1917, a Denver group under the name of Chemical Products Co., built a caustic soda plant
on the northwestern corner of Owens Lake at Bartlett, the equipment for which came from an
abandoned soda plant at Green River, Wyoming. The plant was erected but never operated. The
property was later bought by a Boston group, who rebuilt the plant and operated it under the name
of Clark Chemical Co. But the plant was shut down in 1928 and the company dissolved in 1931.
In 1924, a caustic soda plant was built by Great Western Electro-Chemical Co. at Pittsburgh,
Calif., to cauterize natural soda from Owens Lake. In 1928, the same company built another
caustic unit in the plant of the Stauffer Chemical Co. at Los Angeles, Calif.
A natural soda plant, called Sodium products Co. *exists near a town called Wilson Creek
about 100 miles west of Spokane, Wash. The deposit is in a lake occupying about 13 acres, and
consists of sodium carbonate mixed with insoluble silt. Water is piped in from nearby springs to
leach soda from a pit dug in the deposit, and is heated to about, 100°F. by live steam, until the
warm water dissolves enough soda to show a Baum reading of 32 .The solution is then pumped
to settlers, filtered, cooled and allowed to crystallize in a series of crystallizing tanks. The crystals
are given a preliminary drying by warm air and are then heated to a thick consistency on steel belt
conveyers by gas burners. The plastic mass of crystals is then extruded between rubber rolls into
granules, which are dried, to granular ash.
In 1939, the U.S.Geological Survey discovered a large deposit of trona at a depth of from
1300 to 1600 feet in Sweetwater Country, Wyo. These trona beds were found in cores of the John
Hay Oil and Gas Well drilled by Mountain Fuel Supply Co. The trona is associated with small
quantities of northupite (Na2 CO3· NaCl· MgCO3 ) and pirssonite (Na2 CO3· CaCO3 2H2 O) and may
form an important source of natural soda in future. J.J.Fahey, Chemist of the U.S.Geological
Survey Laboratory , identified in the deposit a new mineral in the form of a double carbonate of
sodium and calcium, (Na2 CO3· 2CaCO3 which he named “Shortite” in honor of Dr. Maxwell N.
Short, Professor of Optical Mineralogy at the University of Arizona.
Among the larger manufacturers operating natural soda plants therefore may be mentioned
the American Potash &Chemical Corp. at Trona; the West End Chemical Co. at West End ; the
Pacific Alkali Co. at Bartlett; and the Natural Soda Products Co. at Keeler. The first two named
are located on Searles Lake, Calif., and the last two on Owens Lake, Calif. Of the four companies
mentioned, three are now operating, namely, the American Potash & Chemical Corp., the West
End Chemical Co., and the Pacific Alkali Co. The Natural Soda Products Co. is operating
intermittently, which the Inyo Chemical Co. at Cartago, on Owens Lake, has been shut down since
1933. This plant was sold to the Great Western Electro-Chemical Company. At the present time,
the combined annual production of all the natural soda plants totals well over 100000 tons of soda
ash, being approximately 4 per cent of the total annual production of the ammonia soda plants in
the United States. Some of these products are accepted as on a par with the ammonia soda ash as
regards purity, notably soda ash from the American Potash & Chemical Corp. at Trona on Searles
Lake, as mentioned above.
At Lake Magadi in Kenya Colony in the British East Africa, natural soda was produced by
Magadi Soda Co., Ltd., controlled by the Imperial Chemical Industries, Ltd., of England. Export
of natural soda from Lake Magadi amounted to 65137 tons a year in 1929. The deposit was
exploited since 1911. Because of the decline of natural soda, the Imperial Chemical Industries, Ltd.
(I.C.I) arranged with the Kenya Government for a moratorium on their contract with the Kenya
Government for the annual production of 100000 tons of natural soda from lake Magadi. Lake
Magadi has probably the largest natural soda deposits in the world, being estimated at about 20
million tons of sodium sesquicarbonate. At present, the soda ash output as reported from the
quantities exported from Kenya Colony is as follows:
Table 5 Sodium Carbonate Exported from Kenya Colony.*
Year Long Tons
1929 65,137
1930 49,270
1931 44,171
1932 37,263
1933 43,051
1934 30,832
1935 38,723
1936 46,549
1937 41,330
1938(8 months ) 33,637
1939 40,983
* The Mineral Industry, 1939;1940.
Lake Magadi could be worked to produce 100000 tons of soda ash annually, but the market
could take only about 50000 tons. Formerly, much of the soda was sold to Japan. From about 1937,
export from Magadi has again declined because of the increased production of ammonia soda ash
in Japan. The outlook for the future of the Magadi soda export is not very bright, because both
Australia and India have built their own ammonia soda plants.
In Egypt, natural soda is worked by the Egyptian Salt and Soda Co., Ltd., mainly for caustic
soda, the total output being a little over 3000 tons of caustic soda yearly.
In South Africa at Hamans Kraal near Pretoria, there is a small natural soda plant producing
about 3000 tons yearly. The raw materials are soda brine and trona.
In India, a small quantity of crude natural soda ”Chanio” is recovered by the Sind, the total
production being less than 1000 tons per year.
In China, a great deal of natural soda was obtained from soda lakes (“Nor”) in the Four
Eastern provinces and in the provinces of Inner Mongolia. As a rule, it was made into blocks (see
Soda Blocks, Chapter XVI) for convenience of transportation. Before any ammonia soda ash was
imported or manufactured in the country, as much as 200000 tons of natural soda were supplied
from this source yearly.
In Canada, there has been only a small production from the Canadian deposits. The figures
are as follows:
TABLE 6 Production of Canadian Natural Soda
Year Tons Year Tons
1928 519 1935 242
1929 600 1936 192
1930 364 1937 286
1931 712 1938 252
1932 495 1939 300
1933 559 1940 220
1934 244
Date on natural soda deposits are scattered and very incomplete. Many plants are unreported,
but they are small and not of importance. Some of them operated for a short time and stopped.
Many of these have passed through vicissitudes of fortune, and have suspended operation. It is
likely, however, that large natural soda producers like American Potash & Chemical Corp. will
remain a potential competitor of the ammonia soda manufacturers in the adjoining territories.
Statistics of the production of natural soda in the United States for the period between
1921-1939 inclusive are given in Table 7.

* Mineral Industry (1931), p. 502; Minerals Yearbook (1940).

The figures given above include soda ash, sodium bicarbonate and sesquicarbonate; the 1930-1931 figures also
included sal soda.
Table 8 is a partial list of some natural soda compounds and their crystalline forms:
 AMERICAN NATURAL SODA PROCESSES TODAY
As mentioned above, only three companies are at present operating the natural soda plants on
the West Coast in the United States, namely the American Potash & Chemical Corporation at
Trona on Searles Lake; the West End Chemical Co. at West End, also on Searles Lake; and the
Pacific Alkali Co. at Bartlett on Owens Lake. The Natural Soda Products Co. at Keeler on Owens
Lake, in which Michigan Alkali Co. was interested, suspended operation in about 1937 on account
of dilution of brine due to flood conditions. The older companies, such as Inyo Chemical Co.,
which was sold to the Great Western Electro-Chemical Company in 1931, the Boro-Solvay, started
by the Solvay Process Co. and sold to the American Potash & Chemical Corp., etc., have ceased to
exist. Figure 4 shows the location of these plants.
 FIG 4 Natural soda plants in Californis.

Of the three companies now actually operating, the American Potash & Chemical Corp. is the
largest and has the most complete line of products, working on processes based on scientific
control and physicochemical principles. The American Potash & Chemical Corp. alone produces
500 tons of KC1, 280 tons of borax, 250 tons of Na2 SO4 and 140 tons of soda ash per day, besides
certain quantities of liquid bromine and dilation sodium phosphate. This Corporation produces
soda ash directly in a dense form, and the West End Chemical Co. and the Pacific Alkali Co. both
produce light ash as the initial product.
The Pacific Alkali Co. at Bartlett on Owens Lake carbonates its brine by means of flue gases
from an oil-fired furnace containing about 14 per cent CO2. At this plant, soda in the brine is
partially carbonated to the sesquicarbonate stage in 6’ dia. × 80’ high towers and the sodium
sespuicarbonate crystals are filtered off from the liquor. The filter liquor is then cooled to produce
crude borax crystals. Sodium sesquicarbonate is dried at a low temperature using steam heated air
dryers, or is calcined to soda ash in a furnace of the Herreshoff type.
The West End Chemical Co. at West End on Searles Lake carbonates the brine with lime kiln
gas as in the ammonia soda practice, except that wooden carbonating towers are used. Soda in the
brine is carbonated to sodium bicarbonate, which is filtered off from the slurry by means of rotary
vacuum filters. The filter liquor is chilled to obtain crude borax crystals, but only a portion of the
total available borax is recovered this way. The bicarbonate is calcined to light ash in a rotary
dryer in the usual manner. The borax (sodium tetraborate decahydrate) is dried on steam-heated air
dryers. Some borax comes down with sodium bicarbonate in the towers, and so the soda ash made
is not of the highest quality.
The process used in the American Potash & Chemical Corp. is a complicated and most
scientific one, based on the study of phase-rule solubility diagrams made by Dr. John E. Teeple
and his coworkers and put into execution successfully by Mr. R. W. Mumford and his assistants.
The company was originally organized under the name of the American Trona Corporation in
1913, and the plant started small production in 1914. Up to 1919 only potassium chloride was
produced. From 1919 on, borax was added as another product. In 1926 the name was changed to
the American Potash & Chemical Corporation, and thereafter production rapidly increased. Here
the brine is pumped by vertical turbine-type centrifugal pumps from some 20 8”-bore wells sunk
to between 90 and 100 feet below the surface. The brine*contains, besides small quantities of
other salts,
NaCl 16.35%
NaSO4 6.96%
KCl 4.7%
Na2 CO3 4.7%
Na2 B4 O7 .10H2 O 2.84%
*Gale, W. A., Chemistry of the Trona Process, Ind. Eng. Chem., 30,869 (1938).
and has a pH value of 9.48. It has only traces of calcium and magnesium, and contains a little
tungstate, bromide, and lithium salt. As the processes are based on phase-rule studies, and are both
interesting and instructive, we shall describe them in considerable detail.
The brine is sent by a horizontal centrifugal booster pump into the plant in 12” steel pipes,
some four miles long, and the whole pipe line is lagged with hair and asbestos paper to prevent
excessive heating by the desert sun. The raw brine is clear and has a specific gravity of 1.3. It is
first used as a cooling medium in the condensers of potassium chloride vacuum crystallizers and
for washing various filter cakes. It is thus preheated before it is sent into huge triple-effect
Manistee evaporators (22 feet diameter by about 40 feet high). These evaporators have 4 outside
heaters located at four quadrants and heated by 31 lbs. Pressure (gauge) exhaust steam from
non-condensing turbo-generators. The brine in the evaporators has a strong tendency to foam, and
small drops of coconut oil are introduced to the liquor surface to minimize foaming. The feed to
the evaporators is introduced at the third effect and consists of raw brine and a small stream of end
liquors in the ratio approximately of 3 parts of raw brine to 1 part of the end liquor. The brine in
each effect is circulated by a vertical turbine pump and the slurry is bled to salt traps for the
separation of salt (NaCl), burkeite (2Na2 SO4· Na2 CO3 ) and Li2 NaPO4 crystals. There are two salt
traps in series. Salt crystals, being coarser, are separated from the slimy crystals of burkeite and
dilithium sodium phosphate, by countercurrent flotation in the first set of salt traps; while the finer
burkeite and phosphate crystals settle in the second salt trap and in the Dorr classifier, thus
affording a rough separation of these crystals. The crystals are filtered off separately on rotary
vacuum drum filters. Salt crystals containing about 97 per cent NaCl are produced at the rate of
some 2000 tons daily and are sent to waste, only a small quantity of the salt being utilized in the
process for the separation of burkeite crys tals and for cooling in the Glauber salt crystallizer
condenser. Burkeite and a small quantity of lithium phosphate are worked up in the process as
secondary products.
Operation of these triple-effect evaporators aims by a close chemical control to reach a point
just below the saturation point of potassium chloride at a temperature of 235°F. in the first effect
of the evaporator, a counter-current feed being used in these evaporators to maintain the highest
temperature for the most concentrated liquor in the first effect in order to prevent precipitation of
KCI. The concentrated mother liquor from the filters and from the overflow of Dorr-classifier
(mother liquor NO.1) is then cooled to 100°F. in three-stage vacuum crystallizers with enough
dilution water to keep salt in solution, and the crude KCl crystals are filtered on the centrifuges.
The filter liquor (mother liquor No.2) is then cooled by ammonia refrigeration to 75°F., with
sufficient condensate returned to the liquor to prevent precipitation of KCl, and seeded for
crystallization, when borax crystals (sodium tetraborate decahydrate) are separated in a thickener
and the underflow slurry is filtered, yielding crude borax. The filter liquor and the overflow from
the thickener are returned to the evaporators as the end liquor (mother liquor No.2) together with
the raw brine, as mentioned above.
Both the crude KCl and borax crystals are refined. Chemical grade KCl is obtained by
redissolving the crude KCl crystals, chlorinating the solution to liberate bromine as a by-product,
and recrystallizing to eliminate most of NaCl and NaSO4 impurities. The crude borax crystals are
recrystallized by controlled cooling in vacuum crystallizers, and the mother liquor (mother liquor
No.3) is added to mother liquor No.1 for working up the crude borax. By carefully heating the
crystals to incipient fusion in a circular vertical furnace using oil firing, a porous crystalline
anhydrous borax is obtained, which is sold under the name “Pyrobor” This is a porous crystalline
variety of used borax, but has not the structure of flinty borax glass which is difficult to grind and
to dissolve. The borax crystals are also treated with commercial sulfuric acid, giving boric acid
crystals which are then recrystallized as the chemically pure or U.S.P. boric acid and dried in a
steam-heated rotary dryer. The fine slippery powder is used in making talcum powder and for
medicinal uses.
Since 1934 the plant has also produced soda ash and anhydrous sodium sulfate from burkeite
crystals. The burkeite crystals in the slurry from the evaporators are filtered and re-dissolved and
the sparingly soluble dilithium sodium phosphate is separated as the residue, which is now the
chief source of lithium salts. The solution is warmed to about 28 . and concentrated in vacuum
crystallizes, yielding Glibber salt (Na2 SO4· 10H2 O) containing a small amount of sodium carbonate
monohydrate ((Na2 CO3· H2 O). This Glibber salt is filtered off. To the filter liquor (mother liquor) is
added salt (NaCl), which by common-ion effect displaces burkeite from the solution upon
warming to a temperature of about 53 ., burkeite being more insoluble at the higher temperature.
The burkeite crystals are then filtered off and added to the bulk of burkeite from evaporator slurry
for processing. Filter liquor from the burkeite is cooled down by ammonia refrigeration to around
5 , when sal soda together with a small amount of Na2SO4 and NaCl crystallizes out. The sal
soda is filtered off and melted in its own water of crystallization. The solution is then evaporated
when sodium carbonate monohydrate separates out, leaving most of the Na2 SO4 and NaCl in
solution. The monohydrate crystals are filtered off and dried in an oil-fired rotary dryer to dense
ash directly. This dense ash has an apparent density of about 60 lbs. per cu. ft. and has a purity of
99.50-99.75 per cent Na2 CO3 , 0.3 per cent NaCl, and less than 0.003 per cent Fe2 O3. The mother
liquor from the sal soda crystallization containing mainly sodium sulfate and salt is discarded to
waste. This grade of soda ash compares very favorably with the highest grade of ammonia soda
ash.
Certain users require a bulky ash for their standard packing and therefore object to this dense
ash. Light ash is now produced in the works by wetting the dense ash with sufficient water to
correspond roughly to the hydration of sodium carbonate heptahydrate in a rotary cylinder,
yielding a granular from of porous mass, which is then dried very slowly in a steam-heated tunnel
dryer. The dried mass is then pulverized to about 80 mesh in an impact crusher, and packed as
light ash. In this way, it is possible to reduce the bulk density of soda ash to 23 lb. per cu. ft. By
varying the ratio of water added, or the rate of drying of the granular mass in the hot-air dryer, or
grinding the dried ash to different degrees of fineness, it is possible to effect any degree of
lightness. This process achieves the reverse of what a smaller amount of water does to the
ammonia soda ash in a somewhat different manner.
For certain crops, such as tobacco and pineapple, and for fruit trees, such as orange and
lemon, potassium chloride is not generally used as fertilizer, but potassium sulfate is preferred by
the farmers and fruit growers for such purposes. Potassium sulfate is made in the plant by a double
decomposition between KCl and burkeite Na2 CO3· 2Na2 SO4 when K2 SO4 is separated by fractional
crystallization.
To help the reader to follow this “Trona Process,” a schematic diagram of it is given in Fig.5
No attempt is made to cover all details of modifications frequently made in the plant operation.
The Trona processes are interesting because the brine represents at least a nine-component
system, which is more complex than the Stassfurt salt deposits in Germany, which may be
considered as a five-component system. If the brine were subject to an isothermal evaporation at a
low temperature, such as by solar evaporation, then salt (NaCl), glasserite K3 Na (SO4 )2 , sodium
carbonate heptahydrate (Na2 CO3· 7H2 O) and finally potassium chloride would be precipitated more
or less together, so that no separation of potassium salts from sodium salts would be possible.
However, when evaporation is carried out at a temperature above 100 , as in multiple-effect
vacuum evaporators, only salt (NaCl), burkeite (Na2 CO3· 2 Na2 SO4) and possibly a small amount
of sodium carbonate monohydrate (Na2 CO3· H2 O) will separate out; while at 113 . (235°F) as the
liquor is further concentrated and the temperature rises, it is possible to keep all KCl in solution.
Consequently, after the salt, burkeite, etc. have been filtered off, and the liquor is cooled from 235
to 100°F., potassium chloride then crystallizes out; but because of the great tendency toward
supersaturation on the part of borax crystals, borax remains in solution and only KCl separates.
After KCl crystals have been filtered and the liquor further cooled strongly by liquid ammonia,
then and only then does borax crystallize out.
Another interesting feature is the separation of sodium carbonate from sodium sulfate in the
burkeite. When salt (NaCl) has been roughly separated from burkeite in salt traps and when the
burkeite crystals are redissolved, Glauber salt by vacuum crystallization in the absence of excess
salt separates out from the solution upon cooling. Glauber salt crystals are then filtered off.
 FIG 5 Diagram of Trona process.
When, to the mother liquor, an excess NaCl (about 2 per cent) is added, the remaining burkeite
separates out as such, leaving sodium carbonate in the mother liquor. By cooling further with
ammonia refrigeration, soda then separates out as sal soda (Na2 CO3 . 10H2 O), which necessarily
contains some sodium chloride and sodium sulfate as impurities. Recrystallization from sal soda to
obtain sodium carbonate monohydrate at a higher temperature eliminates most of these impurities
and the monohydrate so obtained yields a very high-grade dense ash upon calcining.
Here the use of NaCl to depress the solubility product of burkeite for the removal of excess
burkeite from the Glauber salt mother liquor before crystallization of sal soda is very instructive;
while the addition of NaCl to the Glauber salt solution to lower its temperature for use as the
cooling medium in the condenser of the Glauber salt vacuum crystallizer, thereby raising the
temperature of the Galauber salt solution for the recrystallization of the anhydrous sodium sulfate
therefrom, is indeed ingenious (see schematic diagram, Fig.5).
The conversion of dense ash to light ash represents an interesting manipulation in that it
accomplishes the opposite results to the manufacture of “water ash ”in the ordinary ammonia soda
process, although the details in the application are somewhat different.
An electrolytic plant using potassium chloride for chlorine and caustic potash may be
advantageously added in this Trona plant to utilize the excess power and to produce chlorine
needed in the bromine recovery, etc. Caustic potash then is again a valuable joint product.
 Chapter III

History of Ammonia Soda Process: Statistics

As early as 1811, a Frenchman, L. Fresnel, conceived the idea of making soda from salt and
ammonium carbonate. A German, A. Vogel, claimed that his father discovered in 1822 the
reaction between ammonium carbonate and brine, whereby sodium bicarbonate was produced. In
1837,an English chemist, John Thom, attempted to make soda, uses the reaction between
ammonium carbonate and salt. He succeeded in manufacturing some 200kg. Of soda form this
reaction. In the year 1838,H.G.Dyer and J. Hemming patented in England the treatment of brine
with ammonium bicarbonate to precipitate sodium bicarbonate, outlining the process with
considerable detail and accuracy. Dyar and Hemming erected a plant at Whitechapel, London, but
it was not a financial success. In 1839, Delaunay patented the process in France on behalf of Dyar
and Hemming. A. A. Canning in 1842 patented the process of carbonation with gaseous carbon
dioxide, and in 1852 Crinus in France attempted to recover carbon dioxide gas from the
calcinations of sodium bicarbonate. In the same year, W. Chisholm patented the process for the
recovery of ammonia form the mother liquor by the distillation of the liquor with lime and the
subsequent absorption of the ammonia in brine. Thus, as early as 1852, all the main reactions in
the ammonia soda process-the formation of sodium bicarbonate, the calcinations of the
bicarbonate with recovery of carbon dioxide gas, the recovery of ammonia by distillation with
lime and absorption of the ammonia gas in brine-all were definitely known. In 1854, T. Schloesing
patented the same process in France and in 1855 built a factory at Poteau near Paris for its
exploitation. By 1858 Closing and Rolland had patented the complete process in France again.
Plant after plant attempting to use the process then failed. The works erected by James
Muskrat, the founder of the LeBlanc soda industry in England, at Newton, Lancashire, in about
1840 and that erected by William Goss age and Henry Deacon at Widens, Lancashire, in 1853,
both were in operation for two years but could not compete with the Le Blanc process. Deacon
then erected a Le Blanc soda works instead. The works erected by Closing and Rolland at Poteau,
priming for a Tim a god prospect of success, lasted also for about tow years. By 1860,nobody
seemed to be able to make the process a success commercially.
In 1861,Ernest Solve, a Belgian, independently discovered the same process and in 1863 a
company called Solvay at Cie was organized in Brussels to manufacture soda, with a capital of
136,000 fr. A plant was erected at Couplet, Belgium, and operation started in 1865. Ernest Solvay
and his brother Alfred devoted their whole life and energy to the improvement of the apparatus. In
the year1866, that plant produced 1- tone of soda ash per day and in 1867 the product was shown
in the Paris exhibit. In 1869 the plant was doubled and the production trebled. By 1872 Solvay
was able to produce 10 tons of soda per day. In the same year (1872) Solvay designed a large plant
erected at Dom Basle near Nancy, France.
In 1872, Ludwig Mond in England had an understanding with Solvay and the outcome was
that in 1874 a plant was erected at winning ton near Northwest, Cheshire. This marked the
beginning of the ammonia soda process in England and the firm of Brunner, Mond & Co., Ltd.
(founded by John Brunner and Ludwig Mond) was then formed. This was also the first company
to use natural brine for the manufacture of soda. This company now has a plant at Winnington,
near Northwich, one at Sandbach, one at Lostock Gralam, one at Middlewich, and a new plant at
Wallerscote, all in Cheshire. The principal works are in Winnington, Lostock Gralam, and
Sandbach with a new works at Wallerscote for future expansion. The total output from these plants
amounts to over 3,500 tons per day, with a maximum daily capacity of more than 4,000 tons.
Besides France and England, Solvay also built plants in Germany, Russia and other principal
countries in Europe, as well as in the United States. Solvay brothers had an interest in all these
plants and these were more or less closely connected. It was thus a huge international syndicate
having its headquarters in Brussels, Belgium. It was through the organizing ability and personality
of Ernest Solvay (Fig.6) that it was possible to mange such a huge international institution,
forming a connecting link between some twenty-six of such member plants throughout the world.
Ernest Solvay also arranged for an interchange of technical information and operating data among
these member plants. Such operating data and results were compiled and circulated monthly
among all such data and had before it the best set of operating results obtainable concerning any
one step of the process in any one of the individual plants during the current month. It was
possible also arrange for an operating engineer from one plant to visit any one of the other plants
to observe certain improvements and study features peculiar to that particular plant with a view to
improving equipment by the experience of the other plants. The center of these interests was in
Brussels. There an International Committee consisting of technical improvements and problems
concerning operation, or to compare notes and discuss progress made in the other parts of the
world.
In America, the ammonia soda industry was started was started by William B. Cogswell, who
went to Belgium to negotiate with Ernest Solvay for the use of his process.

FIG 6. Ernest Solvay (1836-1922)

*Trump, Edward N. “Looking Back at 50 Years in Ammonia Soda, ”Chem . Met. Eng., 40, No. 3, 126
(1933).
For this purpose the Solvay Process Co. of Syracuse, N. Y., was organized in 1881 with Rowland
Hazard, of Providence, R. I., a mine owner, as President; William B. Cogswell as Manager and
Chief Engineer, and O. V. Tracey as Secretary and Treasurer. During the construction period,
Cogswell brought William L. Neill, a chemist, and later also Henry Cooper and Nicholas Bodot, to
the Dombasle Works of Solvay and Co. to study the process. The plant at Syracuse was completed
in January, 1884, when operation was stared. Although the process had been proved successful,
operation at Syracuse was attended with many difficulties.

FIG 7 Edward Needles Trump (1857-….)

A pioneer and ve teran of the American Alkali Industry.
Only three of the employees in the whole plant had been to an ammonia soda plant, and none had
ever operated one or had had any experience with it. Problems arose daily, but through the skill of
Mr. Cogswell assisted by his assistants, among whom was Mr. Edward N. Trump, the output of the
plant was tripled in 1887 and increased tenfold ten years later (Fig.7)
In 1879, a paper entitled “The Manufacture of Soda by the Ammonia Process” was presented
by Oswald J. Heinrich to the Baltimore Meeting of the American Institute of Mining &
Metallurgical Engineers, giving an interesting account of the comparative status between the Le
Blanc process and the ammonia soda process at that time, still not entertaining much hope for the
latter. Meanwhile, Herman Frasch, of sulfur fame, had built and attempted to operate a small plant
at Bay City, Mich.; but the operation was not successful, and plant finally had to be dismantled.
The Solvay Process plant at Syracuse was so successful that in 1898 another plant was built
at Delray, Mich. The Syracuse plant toady is the largest in the country and also the largest single
single ammonia soda plant in the world. During the first World War (1914-1918), the Kansas
Chemical Company’s plant at Hutchinson, Kansas, was leased, remodeled and later bought by the
Solvay Process Co.; but immediately after the war the plant was shut down, because operation was
found to be uneconomical because of small capacity and distance from a limestone supply.
In 1917, Solvay Process Co. of Syracuse, in conjunction with the then Brunner, Mond & Co.,
of England (now Imperial Chemical Industries, Ltd.) built a plant of about 150-ton capacity at
Amherstburg, Ontario, near the mouth of the Detroit River in Canada. The plant was designed and
built by Mr. Trump and his engineers at Syracuse and operated by them. The plant was under the
name of Brunner-Mond (Canada), Ltd., of which Mr. Trump was concurrently president. Later, the
plant was bought over and completely owned by the solvay Process Co. of Syracuse, although the
name has remained Brunner-Mond (Canada), Ltd. This plant was designed with due regard to
future extension and is well situated as regards the supply of raw materials, limestone being
quarried on the plant site and brine wells located only a few miles away.
Among the independent plants, there are several in the United States. These are: Michigan
Alkali Co., Wyandotte, Mich.; Diamond Alkali Co., Fairport, O.; Columbia Alkali Corp.,
Barberton, O.; and Mathieson Alkali Works, Inc., Saltville, Va. Solvay Process Co. is now in
reality also an independent company. Ever since the merger into Allied Chemical & Dye Corp. in
1920, and especially after 1925, it broke all ties with the European Solvay interests, although
about 20 per cent of its stock is still held by the Solvay group across the Atlantic. In addition there
are three or four natural soda plants in California east of the Sierras. (See Chapter II, Natural
Soda.)
In England, there was an independent soda plant at Fleetwood, owned by United Alkali Co.,
Ltd., but this company is now a part of Imperial Chemical Industries, Ltd. (I.C.I., Ltd.), of which
Brunner, Mond & Co., Ltd. Is a controlling member. The latter also controlled Magadi soda
deposits in Kenya Colony, East Africa. (See Chapter II, Natural Soda.)
In the United States, the ammonia soda industry is definitely tending to migrate southward to
the Gulf Coast. Of those who have located soda plants in the South may be mentioned Columbia
Alkali Corp., Mathieson Alkali Works, Inc., and Solvay Process Co. (Allied Chemical & Dye
Corp.) The plant of the Columbia Alkali Corp. is located at Corpus Christi, Texas, and is known as
Southern Alkali Corp. It was designed and built by engineers of the Columbia Alkali Corp. at
Barberton, and is partly owned by the Pittsburgh Plate Glass Co. and by the American Cyanamid
Co. Because of the business depression in the United States, construction was not started until late
in 1933, although the plan had been under consideration for over three years prior to that time.
The first unit of 250-ton capacity of soda ash per day was completed and its operation stared in
October, 1934. Another unit is now installed, so that the total capacity will be about 450 tons per
day.
The Mathieson Alkali Works’ plant is located is located at Lake Charles, La. Construction
was started at the end of 1933 and was completed in the beginning of 1935.The plant has about the
same capacity (450 tons per day), although at the start of the initial operation some trouble was
experienced in burning the oyster shells. The design was made by Mr. Edward N. Trump assisted
by younger engineers.
Solvay Process Co. entered the field somewhat later and located its plant at Baton Rouge, La.
After rapid construction, the Solvay plant was completed and put into operation in June, 1935.
This plant is somewhat larger than the other two. It also does not depend upon the shell deposits
for limestone supply.
All three plants use rock salt from salt domes by simple brine-well methods. The Mathieson
Alkali plant depends entirely on oyster shells for its supply of limestone. Besides, it has a
synthetic “sulfate” of soda plant, which makes soda ash-sulfur sinter by sintering soda ash and
elemental sulfur in molecular proportions. All three plants have caustic soda plants attached, and
two of them also have electrolytic caustic plants. These plants differ from the older, inland plants
only in certain refinements in the apparatus. The Solvay Process plant is unique in that there is no
boiler or turbine plant, since both steam and electric power are advantageously purchased from the
adjacent Louisiana Steam Products Corp.; which supplies power to the mammoth refinery owned
by the Standard Oil Co. of Louisiana at Baton Rouge.The Mathieson Alkali Plant at Lake Charles
is probably in that the entire compression of CO2 gases into carbonating towers, or columns, is
accomplished by means of centrifugal turbo-compressors instead of piston type CO2 compressors.
The plant also burns oyster shells for lime exclusively. The Southern Alkali plant is unique in that
the calcinations of ammonia soda to soda ash is done in high-pressure, steam-heated equipment
rather than in direct-fired rotary dryers. The plant possesses an abundant supply of petroleum and
natural gas, and uses mostly oyster shells for its limestone supply.
The three southern alkali manufacturers were all led to locate alkali production on tidewater
by the economic advantage of delivering to their customers by sea route rather than by rail. Up to
1933, all alkali produced by the ammonia soda process soda process in the United States came
from inland plants. With the outstanding development of artificial sodium nitrate for fertilizer
from soda ash, and the tremendous tonnage thereof manufactured and shipped to foreign countries
form Hopewell, Va., the industry began to appreciate that a large part of the cost of soda ash at the
consumer’s plant was in the freight. The Hopewell nitrate plant converts soda ash to sodium
nitrate with the nitric from the oxidation of the synthetic ammonia. Consumption of soda ash by
this one plant alone is equal to the capacity of a fairly large ammonia soda plant. The ash was
shipped from Syracuse in canal barges and ocean vessels during a relatively short shipping season
over the New York State barge canal.
Investigation as to the possibility of locating alkali plants on the Gulf Coast was made a
number of years ago. Besides good shipping facilities on the Gulf Coast there were many other
considerations. An increased demand, especially for caustic soda, has existed in the Southern
States because of the growing petroleum, rayon, and cotton industries. Abundance of raw
materials for soda manufacture is in evidence. Vast salt domes are found which contain very
high-grade salt. Oyster shell deposits, which may be burned to make a good grade of lime, are
found in abundance along the beach. Fuel in the form of natural gas or oil is plentiful and cheap
everywhere in the region. These naturally provide an ideal location for an ammonia soda industry.
All these, coupled with location on tidewater for shipping to coastal states on the Atlantic
Seaboard, give these plants a decided economic advantage over the older, inland plants. It is
estimated that, given a good capacity production, the cost of manufacture will be as $5.50-$6.00
per ton of soda ash (not including freight or sales expenses).
Recently new ammonia soda plants have been added to certain countries, such as Japan,
China, Australia, India, etc. In Japan, since the remodeling of the Asahi Soda Co.’s plant by the
late Mr. Harald Ahlqvist in 1934, the output of that plant has been almost tripled. New plants have
also been assed: one in Darien, a regular Solvay process plant; one in Konan, Korea, with a
capacity of 100 tons a day designed by Zahn & Co. of Berlin, Germany; and another a small Zahn
plant located in Japan proper. In West China, a new plant is under construction. In Australia, a new
plant has been completed at Port Adela ide, and was placed in operation in 1940 by Imperial
Chemical Industries, Ltd. The plant has about 100 tons of soda ash daily capacity and cost over
1,000,000 for its construction. Salt is obtained from sea water, and limestone is available near
the factory. In India, two new plants are under construction: one at Khewra, Punjab, by Imperial
Chemical Industries, Ltd., and the other at Mithapur, Kathiawad, by Tata Chemicals, Ltd. True,
most of these new plants are small, but they will surely form starting nuclei and grow as the
technique of operation is mastered by the plant operatives.
*Within recent years, there has been perfected a process for converting salt directly to sodium nitrate with
the nitric acid in this plant, chlorine being formed as a joint product.
Considerable work has been done on the ammonia soda process, but what has been published
deals mainly with the theoretical side of the subject, mostly from the application of the phase rule
and mass action principles. Among the classical investigators of this field may be noted such
names as Fedotieff, Jaenecke, etc. Since 1872, a great mass of work has been done to improve the
mechanical operation of the process, although little has been added to the public knowledge. A
large number of patents has also been taken out mostly on the invention of the improvement of
equipment, but this does not mean that such equipment was all in use of stood the test of practice.
Many modifications have been proposed regarding the method of working up the mother liquor
for ammonium chloride.
Recently some fundamental research work on the ammonia soda process was done by several
Russian chemists, and the results were published and made available to the public. In the United
States, similar research was made somewhat later by the Mathieson Alkali Works, Inc. at Saltville,
and the results checked quite closely with the published Russian figures.
STATISTICS
After the plant was established by the Solvay brothers, the annual production between 1864
and 1868 was 300 tons per year. By 1916,the total world output by this process was 3,000,000
tons, of soda ash (all by the ammonia soda process), not including products inn other forms.
TABLE 9. Earlier Production of Soda Ash in the United States (in short tons).*
Year Tons Year Tons
1900 362,806 1914 980,224
1904 600,008 1915 1,175,962
1909 740,455 1916 1,194,183
1910 770,978 1917 1,390,628
1911 854,370 1918 1,390,628
1912 935,612 1919 981,054
1913 899,830
*“Tariff Information Surveys, ”U. S. Tariff Commission, 1921, Washington, D.C.
On the other hand, soda ash by the Leblanc process, which dates back to 1791, reached a
maximum yearly production of about 550,000 tons in 1879-1883.It began to decrease, as the
successful operation of the ammonia process by the effort of the Solvay brothers was gradually
assured. When, in 1902, the world production of soda ash b the ammonia process reached
1,600,000 tons, that by LeBlanc process was decreased to 150,000 tons, about 1/10 of the output
by the ammonia process. At present, practically the whole bulk of soda ash in the world is
manufactured by the ammonia process.
 The earlier production in the United States where no LeBlanc soda was ever produced is
shown in Table 9;more recent production figures are given in Table 10.
TABLE 10. Recent Annual Production of Soda Ash and Allied Products in the United States*(in short tons).
1921 1923 1925 1927 1929 1933
Total soda ash 959,593 1,707,987 1,907,791 2,037,808 2,682,216 2,317,011
Soda ash by ammonia process 929,448 1,674,234 1,850,013 1,974,157 2,586,304 1,582,633
Natural soda and soda by other
Processes 30,145 33,753 57,778 63,651 95,912 68,395
Total caustic soda 238,591 436,619 497,261 573,417 758,800 686,983
Caustic soda by lime process 163,044 314,195 355,783 387,235 524,985 439,363
1921 1923 1925 1927 1929 1933
Caustic soda by electrolytic process 75,547 122,424 141,478 186,182 233,815 247,620
Bicarbonate of soda 129,331 145,316 123,472 121,449 140,234 129,273
Sal soda 69,342 68,802 63,619 55,220 62,062 …….
Modified sodas …… 47,669 47,452 53,866 59,618 21,873

1935 1937 1939

Total soda ash 2,317,011 3,037,421 2,961,632
Soda ash by ammonia process 1,776,470 2,205,006 2,013,264
Natural soda and soda by other
Processes 95,165 118,753 132,897
Total caustic soda 759,381 968,726 1,025,011
Caustic soda by lime process 436,98 488,807 530,907
Caustic soda by electrolytic process 322,401 479,919 494,104
Bicarbonate of soda 136,556 142,161 148,610
Sal soda 39,439 33,064 28,668
Modified sodas 29,103 26,497 32,101
*From “Biennial Census of Manufacturers, ”Dept of Commerce Washington, D. C., up to 1939
+Not including quantities of caustic manufactured and consumed by the wood pulp and textile industries.
Form Table 10, it may be concluded that at present the United States alone (taking into
account all forms of soda made) produces in excess of 3,000,000 tons annually. The production of
electrolytic caustic in this country is increasing very rapidly on account of increased demand for
chlorine. Natural soda is being produced in considerable quantities in California and its influence
will soon be felt. The trade in salt soda, however, is declining consistently.
Further, in the United States as of 1940, the annual production of the five ammonia soda
companies* was 3, 157,000 tons, while that of the natural soda plants was over 100,000 tons, so
that the total annual capacity for the United States is well over 3,000,000 tons, representing 83 per
cent of the total installed capacity existing at the beginning of 1940.
Allocating by plants we obtain the U.S. capacity figures in Table 11.
TABLE 11. United States Ammonia Soda Plant Capacity
PLANTS Capacity in short tons of soda ash per day
1. Solvay Process Co 2,200
Detroit plant 2,000
 Baton Rouge plant 550
Amherstburg plant (Brunner-Mond)
(Canada) Ltd. 230
Hutchinson plant (if rehabilitated) 200
_____
5,180 5,180
2. Michigan Alkali Co.
North plant 1,500
South plant 500
________
2,000 2,000
3. Diamond Alkali Co. 1,800 1,800
4. Columbia Alkali Corp.
Barberton plant 1,000
Corpus Christi Plant 450
__________
1,450 1,450
5. Mathieson Alkali Works, Inc.
Saltville plant 750
Lake Charles plant 450
_________
1,200 1,200
Total _______
11,630 short tons
per day
The foregoing figures indicate that there is a total potential capacity of ammonia soda ash in
America of approximately 4,000,000 tons annually, but that the industry has been operating at
only a fraction of this potential capacity. The increase in production has been rapid and has
averaged about 4 per cent each year for many years, except the last few years. Caustic soda by the
lime process is made by all these words from about one-third of the soda ash produced, although it
is included under and calculated as soda ash. The Solvay process Syracuse plant, the Mathieson
Alkali Saltville plant, and the Michigan Alkali and Diamond Alkali plants also make pure
bicarbonate of soda, washing sodas, calcium chloride and salt.
*Chem. Met. Eng., 48, 93 (1941)
The total world output of soda ash is estimated in Table 12.
TABLE 12. World Output of Soda Ash (in metric tons).
Year production by Production by TOTAL
LeBlane Process Ammonia Process
1800 nil nil nil
1850 150,000 nil 150,000
1863 300,000 nil 300,000
1865 374,000 300 375,000
1870 447,000 2,600 450,000
1875 495,000 30,000 525,000
 1880 545,000 136,000 681,000
1885 435,000 365,000 800,000
1890 390,000 633,000 1,023,000
1895 265,000 985,000 1,250,000
1900 200,000 1,300,000 1,500,000
1902 150,000 1,610,000 1,760,000
1905 150,000 1,750,000 1,900,000
1911 130,000 1,900,000 2,030,000
1913 50,000 2,800,000 2,850,000
1916 Small 3,000,000 3,000,000
1923 nil 3,500,000 3,500,000
1927 nil 4,100,000 4,100,000
1932 nil 5,000,000 5,000,000
1940 nil 7,000,000 7,000,000
*According to Julius Kirchner.
To aid comparison, the curves shown in Fig. 8 have been prepared.
The present annual production of soda ash in each of the various countries in the word is
estimated in Table 13. These figures have been rounded off for the sake of convenience in
comparison even where more accurate figures are available.
Table 13. Recent Ammonia Soda Production in Various Countries.
Approximately Yearly
Production in Terms
Of Soda Ash
Country (Metric Tons)
United States of America 3,000,000
(Natural Soda only 120,000)
Great Britain 1,500,000
(Magadi-Natural Soda only 50,000)
Germany 1,250,000
France 710,000

Approximately Yearly
Production in Terms
Of Soda Ash
Country (Metric Tons)
Russia 570,000
Italy 387,000
Japan 250,000
Czechoslovakia 150,000
Poland 100,000
Belgium 88,000
Canada 83,000
China 80,000
 Jugoslavia 70,000
Spain 50,000
Austria 45,000
Australia 30,000
India 30,000
Switzerland 30,000
Rumania 25,000
Norway 18,000
Holland 15,000
Venezuela 2,000

Table 14 represents an attempt to five a list of ammonia soda plants existing in different
countries in the world together with the year of their construction. The change in geographical
boundaries, especially in warring countries in Europe as the result of world War, may further alter
the distribution of these ammonia soda plants among these countries. It is understood that such an
estimate is approximate as regards the daily capacity of each plant in terms of soda ash.

Tables 13 and 14 are compiled from various sources of which a full list cannot be given here .
The principal references are listed below:
G. Lunge, “Sulfuric Acid and Alkali,”Vol.3, New York, D. Van Nostrand Co.
F. U11mann, “Enzyklopadie der technischen Chemie,” Berlin, Urban and Schwarzenberg.
E. Thorpe, “Dictionary of Applied Chemistry,” London, Longmans, Green &Co
 Solvay & Pennock,” Report of the 5th International Congress of Applied Chemistry.”
“ Minerals Industry,”1892-1939, New York, McGraw-Hill Book Co.
“Minerals yearbook,”1932-1939”, Bureau of Mines, U.S. Dept. of the Interior, Washington,
D.C.
TABLE 14. General Survey of World’s Ammonia Soda Industry.
Daily Capacity
When of Soda Ash
Name and Location Built (Tons)
United States
Solvay Process Co.: Syracuse, N. Y 1881. 2200
Delray, Mich. 1808 2000
Baton Rouge, La. 1935 550
Hutchinson, Kans. (not operating) 1908 200
Michigan Alkali Co., Wyandotte, Mich. 1893 2000
Diamond Alkali Co., Fairport, O. 1910 1800
Columbia Alkali Corp., Barberton, O. 1900 1000
(Southern Alkali Corp.) Corpus Christi, Texas 1934 450
Mathieson Alkali Works: Saltville, Va. 1894 750
Lake Charles, La. 1935 550
Great Britain
I. C. I. (Brunner, Mond & Co., Ltd.): Winnington 1874 1500
Middlewich 1889 350
Sandbach 1875 1100
Wallerscote 1890 700
Lostock Gralam 1925 800
I. C.I. (United Alkali Co., Ltd.) Fleetwood 1890 400
Germany
Deutsche Solvay-Werke: A. G. Bernburg 1883 1100
Rheinberg 1908 400
Wyhlen 1880 200
Kali-Chemie: Heilbronn 1900 200
Aachen … 150
Chem. Fabrik Kallk, Koeln-Kalk …. 150
Henkel & Cie, Duisburg ….. 250
Daily Capacity
When of Soda Ash
Name and Location Built (Tons)

Theodor Goldschmied, Stassfurt 1883 150

I. G. Farbenindustrie, Oppau 1915 300
(And several small establishments)
France
Solvay et Cie: Saaralben 1885 300
 Chateau-Salins 1898 150
Dombasle 1874 1200
Salins de Giraud 1896 200
Kuhlmann, Dieuze 1875 50
Marcheville-Daguin & Cie, La Madeleine 1882 200
St. Gobain, Varangeville 1891 250
Etudes et Produits Chimiques, Mouguerre 1918 50
Russia
Lubimoff-Solvay et Cie: Beresniki 1883 400
Donetz ….. 1000
Slavyansker, Slavyansk ….. 500
Japan
Asahi Glass Co., Ltd., Tobata (Fukuoka) 1916 600
Nippon Soda Co., Ltd., Tokuyama (Yamaguchi) 1918 800
Toyo Soda Co., Tonda (Yamaguchi) 1936 500
TABLE 14. Continued
Approximate
Daily Capacity
When of Soda Ash
Name and Location Built (Tons)
Approximate
South Manchurian Co., Pulantien (Darien) 1937 120
Kyushu Soda Co., Kyushu 1936 100
Nippon Chisso Hyrio, Konan (Korea) 937 100
Dai Nippon Artificial Fertilizer Co., Onoda (Yamaguchi) 1938 75
Italy
Soc. Montecatini, Monfalcone (Trieste) 1913 100
Solvay & Cie, Rosignano 1919 800
Czechoslovakia
Chemische Werke Aussig-Falkenau (I. G. Farben.), Nestomitz 1906 350
Chemische Werke Aussig-Falkenau (I. G. Farben.), Aussig. 1885 150
Synthesia Chemical Works, Ltd., Semtin 1934 10*
Jugoslavia
Solvay et Cie & Aussiger Verein: Lukavac (Bosnia) …… 200
Hrasnica ….. 100
Poland
Solvay et Cie: Montwy (Posen) 1881 200
Podgorze (Galicia) …… 100
Belgium
Solvay et Cie, Couillet 1863 250
Canada
Brunner-Mond (Canada), Ltd., Amherstburg 1917 250
China
Yungli Chemical Industries, Ltd.: Tangku 1 921 250
 Wutungchiao 1941 50
Spain
Solvay et Cie, Torrelavega 1908 320
Austria
Solvay Sodabetriebsgesellschaft, Ebensee 1885 200
Australia
Imperial Chemical Industries, Ltd., Adelaide 1940 100
India
Imperial Chemical Industries, Ltd., Khewra (Punjab), (under construction) 1941 80
Tata Chemicals, Ltd., Mithapur (Kathiawad), (under construction) 1941 120
Rumania
Uzinele Solvay, Turda & Maros Ujvar (2 plants) ….. 150
Switzcrland
Solvay et Cie, Zurzach 1915 100
Norway
Norsk Hydro-Elekt., Heroja 1933 75
Holland
Solvay et Cie, Roermond (Limburg) 1939 50(?)
Venezuela
Comp. Anon. Ind. Quim. Nac., Maiquetia 1935 3
*All as NaHCO3 by a modified Ammonia Soda process. The NaHCOn in used directly in the manufacture of
NaNo3.
For the recent figures, the United States Dept. of Commerce “World Trade Notes on
Chemicals and Allied Products” and more recently “Industrial Reference Service: Part I
Chemicals and Allied Products” published by the Bureau of Foreign and Domestic Commerce,
have proved to bean invaluable aid. Because of the peculiar tradition of this ammonia soda
industry, no official figures are published by the industry nor by the government agencies with the
exception of the United States. Much of the information, therefore, has to be derived from
individual investigations. The foregoing figures represent the best available estimates and no strict
accuracy is claimed for them. A brief summary by Harald Ahlqvist entitled “Alkali Industries
Girdle the Globe” revealed certain figures concerning soda ash production and consumption for
certain countries in the world up to 1936.
In conclusion, it might be added that the International Syndicate of Solvay et Cie suffered
considerable setback immediately after the World War I (1914-1918). Since that war the Russian
Lubimoff plants first broke off relations with the Syndicate. Then followed the American Solvay
Process Co. which severed its ties with the European Syndicate by its merger into the Allied
Chemical and Dye Corporation in 1920.From that time on, all interchange of technical data and
information and all confidential communications between the Russian or American alkali plants
and the Solvay Syndicate and its member plants in the world ceased. Similar change of status had
also to be brought about for the British plants when Brunner, Mond & Co., Ltd. was merged into
the Imperial Chemical Industries, Ltd., inn 1926.How much of the tie is still left between the I.C.I.
and the Solvay Syndicate is a matter of conjecture. However, since that time, the International
Solvay Syndicate seemed to have gradually recovered from this setback and has acquired interest
in several alkali plants in certain European countries, so that today this European syndicate still
owns or has interest in a total of some 26 ammonia soda plants in the world. It is to be noted,
however, that independent firms outside of the Solvay combine have also multiplied very fast and
have built many new plants, particularly in the United States, Japan, China, India, etc.; and the
trend hereafter seems to be distinctly toward increase in the number of the independent alkali
plants the world over.
*Chem. Met. Eng., 43,278-281 (1936).
 Chapter IV

Preparation of Brine: Rock Salt and Sea Salt

Salt employed in the standard ammonia soda process* is in the form of brine exclusively. The
manufacture of soda salt presupposes an abundant supply of cheap salt, for the process is rather
wasteful as far as the utilization of salt is concerned. The best practic e in present-day soda plants
does not utilize more than 75 per cent of salt and oftentimes much less. If therefore salt is not
cheap, it would not be possible to operate the process successfully.
Rock salt or natural brine is generally to be preferred to sea salt on account of its higher
purity; but it is now known that sea brine, after saturation by solar evaporation, can also be used.
Ammonia soda plants in China and Japan are all employing sea salt. A new ammonia soda plant at
Adelaide, Australia, also uses sea salt. The use of sea salt, however, is attended by difficulties, and
means must be provided to get good settling for the ammoniated brine and to handle the
precipitate or ”mud” in the settling system with certain special provisions. Pretreatment of brine is
generally recommended. The impurities (such as magnesium salts) are present in sea brine to a
larger extent than in rock salt; with the result that sea brine leaves a large extent than in rock salt;
with the result that sea brine leaves a large precipitate in the form of MgCO3 in the settling system.
The magnesium “mud” id rather “sticky,” and if means are not provided to keep it in suspension
and draw it out regularly, it will ”set” and build up thick scale, plugging the whole system.
Difficulties of this sort may become so serious that operation may be stopped.
Pretreatment of the brine before sending it to the system for soda ash manufacture has been
employed. The reagents used in the treatment are the usual soda ash and lime. Soda ash
precipitates calcium chloride and calcium sulfate in the form of calcium carbonate, whereas lime,
by causticizing soda ash, yields sodium hydroxide, which precipitates magnesium chloride and
magnesium sulfate as magnesium hydroxide. Purification of brine will be treated in a separate
chapter and hence will not be dwelt upon here.
As we shall see later, the brine in passing through the system first acts as a scrubbing
medium and then as an absorbing medium for ammonia. It finally reacts with ammonia and carbon
dioxide gases, accomplishing exactly the same result as when soda and lime are added. Further the
reaction takes place when the solution has been heated to 70C.Since this reaction results naturally
during the ammoniation of brine, the pretreatment of brine is often omitted for reasons of
economy and expediency.
*A modified ammonia soda process uses solid salt in finely pulverized form.
Typical analyses of rock salt and sea salt are shown in Tables 15, 16 and 17.
TABLE 15 Rock Salt (Brine). (Sp. gr at15 .1. 196).
Per Cent
Grams/liter (on dry basis)
NaC1 297.0 97.95
CaSO4 5.15 1.70
CaCl2 0.55 0.18
MgCl2 0.49 0.16
Suspended matter 0.03 0.01
 TABLE 16. Sea Salt (Crystals).*
Per Cent
Per Cent (on dry basis)
Moisture 5.99
Insoluble matter 0.82 0.890
NaC1 88.73 96.31
KC1 0.143 0.155
KBr 0.024 0.026
KI 0.186 0.202
Fe2 O2 and Al2 O3 Nil Nil
CaSO4 . 2H2 O 1.781 1.531 (CaSO4 )
MgSO4 . 7H2 O 1.249 0.662 (MgSO4 )
MgCl2 . 6H2 O 0.448 0.229 (MgCl2 )
*The sea salt has been more or less purifled as the natural result of partial crystallization by solar evaporation.
The original brine contains more impurities
TABLE 17. Rock Salt from Kingman, Kansas.
Per Cent
NaC1 97.51
CaSO4 1.51
Na2 SO4 0.57
MgCl2 0.10
Fe2 O3 0.11
Insoluble matter 0.20
+F. W. Clarke, “Data of Geochemistry,” U. S. Surv., Bulletin 770,p231.
The average composition of sea water is given in Table 18.
TABLE 18. Average Composition of Solids in Sea Brine.
Per Cent
NaC1 77.76
MgCl2 10.88
MgSO4 4.74
CaSO4 3.60
K2 SO4 2.46
CaCO3 0.34
MgBr 2 0.22

+F. W. Clarke, “Data of Geochemistry,” U. S. Surv., Bulletin 770,p 127.

ROCK SALT
Rock Salt for soda works is generally obtained in the form of saturated brine by pumping
water under pressure to the bottom of the rock salt stratum so that it comes into contact with the
salt deposit and dissolves it to saturated. Saturation is obtained by adjusting the rate of pumping.
This saturated brine is either forced up to the surface by the pressure of the entering water itself, or
is pumped up by means of compressed-air lift or deep-well pumps, if the rock-salt vein has
crevices of is in communication with other wells so that the brine cannot be forced up under
pressure. The depth of the wells varies in different localities.
Rock salt in the United States is found in great beds where it was deposited millions of years
age by the drying up of sea water in shallow lagoons which were once flooded from the sea. The
salt was deposited in layers, and beds frequently covered with floods of muddy water of brine
formed layers of shale, so that these beds of salt are sandwiched between many layers of shale.
Other strata are formed over these beds so that most of them occur below 1000 feet.
In New York State, deposits occur at Tully. The top bed occurs at 1210 feet below the surface
and is 45 feet thick. In the next 300 feet, there occur three other layers, 35,25,and 100feet thick,
respectively. Between these beds are layers of shale of different thicknesses. At Ithaca, N.Y., the
wells are 2300 feet deep, and there are also several beds of salt with layers of shale between them.
In Michigan, at Detroit and Wyandotte, these wells are 1600 feet deep with 400 feet of salt;
whereas at St.Clair, the wells are from 2100 to 2300feet deep. Some beds in the center of
Michigan are said to be as much as 800 feet thick. The whole of Michigan seems to be underpaid
with a great bed of salt, which is perforated by wells at Detroit, Wyandotte, St.Clair, Manistee,
Ludington, and many other places along the lakeshore and in the interior of the state. Impure brine
is found at Bay City and Midland, from which various by-products of salt are recovered.
In Ohio, at Barberton, the wells are about 2800feet deep. Along the Ohio River, the deed is
found at 5000 feet at Pittsburgh, Pa., it is at 6000 feet. A great bed of salt extends under Lake Erie
from Canada under nearly the entire state of Ohio and western Pennsylvania.
In Kansas, at Hutchinson, the wells are about 800 feet deep, and the bed 300 feet thick. This
bed is nearly 100 miles wide, extending into Texas. At Carlsbad, Texas, a bed of potash, which is
being mined, is found with the salt.
In Michigan, Ohio, and West Virginia, natural brine exists, but generally such brine is not
saturated. Natural brine is formed by underground springs passing through the rock salt veins, or it
is formed from the bittern from which beds of salt have been deposited, in which case the brine
contains many impurities which can be worked up into valuable by-products, as in Bay City and
Midland, Michigan. Natural brine often comes up almost to the surface, but sometimes remains at
a great depth requiring deep-well pumps, such as Reda type pumps.
In Texas and Louisiana along the Gulf Coast peculiar deposits of salt are found in the shape
of domes which seem to have been pushed up by pressure from below. These salt domes, or
“plugs,”are of great depth, some having been drilled too 6000 feet and still not penetrated through.
Some are estimated to be perhaps 40,000 feet in thickness. These domes are of various
shapes-round, oval or irregular-from a mile to three or four miles inn diameter, close to the surface
of down to 1000 feet, capped with a layer of limestone with sulfur under it. Frequently petroleum
is found in the beds of sand between layers of clay, which have been pressed upward by the
movement of the dome or “plug” so as to form a reservoir for the oil. These domes have been
located in conjunction with deposits of petroleum found by observations with some geophysical
and electro-magnetic instruments.
 FIG 9 Brine well
The Solvay Process Co. discovered salt at Tully in 1896 and built a 12-inch pipe line 18
miles long from one reservoir at the wells to one at the works. In 1936,an additional pipe line, 20
inches in diameter, was added, which delivers brine by gravity with a fall of 265 feet head. A
water supply from lakes, 500 feet above the highest well, provides ample pressure to elevate brine
so that no pumping is required, except an air pumping plant which was used when groups of wells
were connected and one of them needed repairs.
Brine wells are 6 or 8 inches in diameter, the hole being drilled clear down to the bottom of
salt bed. A10-or12-steel casing is driven through the clay or gravel into a bed of rock to cut off the
surface water, and the hole is drilled inside the casing to a layer of hard rock under the water seam.
The 6-or 8-inch pipe is then seated in the rock and cemented on the outside at the bottom to make
a dry hole. The hole is then continued to the bottom of the salt bed. (See Fig.9.)
A 3-or 4-inch inner tube is suspended inside the well, supported on a casing head screwed on
the top of the 6-or 8-inch casing. Openings are provided into the annular space between the well
and the inner tube, with valves inserted to control the water of air supply. These valves allow the
water of air to be turned into the inner tube or the annular space outside, so that the tube may be
cleared, when necessary, by reversing the flow.
 FIG 10 Tully method.
Several methods are used inn sending water down to dissolve the salt and elevate the brine.
These are the New York or Tully Method, the Detroit of Kansas Method, the Trump Method, etc,
For beds of salt thinner than 200 feet, the Tully method (Fig.10) is used. The water is first turned
down into the center tube and the dilute brine forced up the annular space until gradually a large
cavity is formed. Then the flow is reversed by introducing the water into the annular space so as to
dissolve the salt from the top down.
 FIG 11 Tully method well I peration.
This is the regular production method. (See Fig.11) The cavity so produced gradually assumes the
shape of a cone with the tip of the cone downward in the center below the lower end of the inner
tube. The reason for this is that fresh water introduced through the annular space at the top of the
cavity dissolves the salt much faster at the top portion and the rate of solution decreases as the
water descends, having picked up salt on its way down to the bottom of the center tube, through
which saturated brine is finally forced to the surface. This explains the formation of a shallow
cone-shaped cavity with the base of the cone upward. As the fresh water enters at the center of the
top of the cavity, the water dissolves more salt from the center than from the outer areas; a circular
arch is thus formed at the top of the cavity. This causes the base of the inverted cone to assume a
spherical dome until it gradually extends to the flat shale roof, as the layer has gone to solution
completely.
 FIG 12 Kansas method.

The cavity increases as the salt is dissolved out. When the cavity reaches a diameter such that
the shale roof cannot support itself, it caves in and serious subsidence of ground has sometimes
occurred. The collapse of the cavity often breaks off the center tube. The hole now has to be
predrilled until the tube can be replaced. The frequency of these carvings depend on the character
of the shale roof, the thickness of the salt bed, and the rate of pumping the brine. With a salt bed
about 50 feet thick and ordinary shale roof, caving-in may occur as often as once in six months. At
Tully with a 45-foot bed, caving-in occurred about twice a year, and today only 40 out of some 90
wells drilled remain. Mud often collects at the shallow conical bottom up to 6 inches or more.
Cleaning is an expensive operation and repairs may cost $1500 each time the inner tube is cut off.
If the bed of salt is over 200 feet thick, it is possible to obtain saturated brine by forcing water
down the inner tube to the bottom of salt bed and making brine come up the annular space. This is
the Detroit or Kansas Method (see Fig 12). The pear-shaped cavity will enlarge; but due to the
deposition of mud at the bottom “bianketing” the bottom of the cavity is gradually covered
with soft mud and protected from solution, so that solution of salt takes place much faster at the
top than at the bottom . Consequently an inverted, cone-shaped cavity is again formed. No cavity
will occur for some time; but when a large diameter has been dissolved off at the top, caving-in
will eventually take place.
Deep brine wells sunk into one of the Louisiana salt domes, or “plugs”, give a very long life
per well.
The detailed method of operation of brine wells depends on the nature of the salt stratum
being exploited. Where the strata are relatively thin, a modern technique adds to the life of a well
by avoiding the caving-in of the weaker shale or earth strata overlying the salt beds, by keeping a
strong roof over the salt cavity is thus maintained. If the salt strata are near the surface and there is
difficulty in preventing infiltration of ground waters, the brine is pumped to the surface by an air
lift. By controlling the rate of pumping, the brine may be brought to almost complete saturation.
By the addition of strong brine from other wells, some weaker brine is utilized for the manufacture
of soda. Very weak brine resulting from well development may not be disposed of in fresh-water
streams without endangering fish life in the stream and all the plant life irrigated by it. Such weak
brine is generally utilized in place of fresh water in operation the regular production wells.
Since the capacity of a well depends on the area of salt surface in contact with water, a new
well naturally has a low capacity and is likely to produce weak brine. This is particularly true if
the salt beds are thin. It is therefore necessary to anticipate the life of a well and have a number of
wells under development for use several years hence. The development of new wells in a shallow
salt stratum can be accelerated by circulating abnormally large volumes of water or drilling
several wells in a row abnormally close to each other, so that their cavities may soon join together
and become connected, after which these wells may be operated as a single well with water going
down one hole and brine coming up the other hole at the extreme end.
From the above it should be clear how the cavity becomes cone-shaped in the inverted
position. The cavity first is in the shape of a steep hole, which gradually widens out at the top until
the sides get so flat that the mud in the water and the impurities in the salt settle on the bottom and
blanket it. Consequently, the salt at the bottom will not dissolve except around the edge of the
cavity and on the top. As soon as the cavity reaches 125 or 150 feet in diameter, the rock caves
and still further blankets the floor. This caving may cut off the pipe so that the well has to be
rebuilt through the debris and a new pipe introduced. On thicker beds, 200 feet thick or more,
brine is obtained by putting down water through the central pipe and the water drives the brine up.
This first results in a spherical cavity at the bottom around the pipe, but deposition of the mud
soon covers the bottom. Then the cavity can widen only at the top. This gradually becomes a
shallow conical pit with blanketed flat slopes toward the center and with curved roof over the top.
Thus, again the cavity becomes an inverted cone with a dome-shaped base upward.

.
FIG 13 Trump’s method.
In 1929 Mr. Edward N. Trump, of Syracuse, N.Y., invented a method of undercutting the
beds of salt by holding water at the bottom of the salt bed to a depth of about 4feet above the
bottom, controlled by drilling holes in the center tubes 4 feet from the bottom. Water for
dissolving the salt is sent is sent in through the annular ring as in the Tully method. Air is
introduced with the water, which carries it down in bubbles which are gradually absorbed, while
the surplus of the air escapes through the holes. Water in the salt bed is confined by air which
forms a cushion above it, thus preventing contact between the water and salt at the top.
This makes a flat cavity of a large diameter horizontally, and brine gravitates to the bottom of the
center tube and is forced up with the air through it. Fresh water holds air up to 1.7 per cent by
volume, but saturated brine holds only about 1/4 of this amount. The air dissolved in the water is
thus gradually liberated, as the water picks up salt to saturation, forming bubbles that lighten the
brine and act as a partial air lift. Fig 13 shows progressive stages of undercutting in the Trump
method. The undercut is 4 feet high, and a flat cavity some 300 feet in diameter may be obtained
in 12 months’ time. After the undercut has attained a desired diameter, air is lifted off and the
water level raised, to dissolve the top portion of salt from below. Thus a larger cavity of undercut
with a dome-shaped top is formed, and so on, until a large inverted frustum of cone is obtained,
capable of producing large volumes of saturated brine. Note that as in the Tully method, where
water is admitted at the top, the cavity at the top also assumes a spherical dome shape by the
solution of salt by water from below. With the cavity of 300 feet in diameter t the top, the well will
produce 100 gallons of saturated brine per minute, whereas by the Tully method, the well will
produce only 50 gallons of saturated brine a minute, and may cave in when the top diameter of the
cavity attains 125 feet. It is possible to undercut a still larger diameter than 300 feet by the Trump
method by placing wells in rows at 700 feet apart.
In May 1934, the Solvay Process Co. of Syracuse, N. Y., introduced another method for
dissolving the salt from salt beds. Instead of using a single hole for sending in water and elevating
out brine, two or more holes are drilled; through one of these water is sent down, and through the
other brine is removed either by pumping, air lift, or other means. This is no different from the
method used in the Province of Szechwan China, a number of years ago. At Tawenpao in the
Tzeliutsing district in Szechwan, as many as 36 such wells scattered widely apart are grouped
together. Water is sent in through one of these wells and brine is pumped out by long buckets
operated by a cable hoist from six of the wells in daily rotation.
Natural brine almost up to saturation occurs in Cheshire, England. In southeastern Ohio and
in West Virginia, natural brine is found in a less concentrated form. In Szechwan, China, many
natural brine wells exist, but the brine obtained is of two forms: yellow brine containing from
8-12 per cent NaCl from the shallower wells (1200-1800 feet ), and “dark brine ”containing 12 to
18 per cent NaCl from wells somewhat deeper (1500-2200 feet ) . Salt beds so far are found only
at 3100-3300 feet in the Tawenpao district where the method patented by the Solvay Process Co.
In 1934 has been in use for decades.
In Germany, natural brine is obtained in rather dilute form in many places in central Germany,
notably Bad Duerrenberg, Schoenebeck, etc., where such weak brine is concentrated in rows of
towers packed wit twigs several kilometers long. Weak bring is sprayed over the twigs, allowed to
trickle down from the top and concentrated by evaporation with the aid of solar heat and winds.
Such structure when viewed from afar has all the appearance of a great wall in a Tartar City!
There are a great many salt deposits throughout the world. In England, they exist in Cheshire
and Lancashire Counties. In Germany, there are the famous Stassfurt deposits and others at
Bernburg, etc. In the United States, they occur in Michigan, New York, Ohio, West Virginia,
Kansas, Texas, Louisiana, etc. It is in such places that the ammonia soda industry has been built
up.
In the early days of the American ammonia soda industry, natural brine used to be saturated
with mined rock salt shipped into the plant. Since the development of the deep brine-well
technique, no American ammonia soda plant uses solid salt. Such brine is often transported over a
long distance to the ammonia soda works. The Solvay Process plant at Syracuse, N.Y., pipes its
brine over a distance of 18 miles in 20-inch pipes, as mentioned before. Striking recent examples
are the 60-mile pipe lines for the Southern Alkali Corp. at Corpus Christi, Texas, and the
tremendous Mississippi River crossings for the Solvay Process plant at Baton Rouge, La.
The distribution of Salt production in the United States as of 1938 is shown in Table 19.
Table 19. (1938).* Total production for the year 1938=8,025,768 short tons, including partially
that consumed in the caustic soda manufacture.
Per Cent
Michigan 25.85
New York 21.40
Ohio 18.54
Louisiana 11.93
Kansas 7.46
California 4.36
Texas 4.04
West Virginia 1.62
Utah 0.77
Other States 3.82
* Minerals Yearbook,”1939, Bureau of Mines, U. S. Department of Interior.
Some typical analyses of rock salt brines are shown in Table 20.
Table 20. Typical Analyses of Rock Salt Brines.
Wyandotee, Saltville, Tawenpao, Cheshire,*
Mich. Va. Szechwan, China (Natural Brine)
Sp. gr. 1.02 1.96 at 15 1.199 at 23
NaCl 303.6 297.0 297.84 300.0
CaSO4 4.8 5.15 4.774 4.0
MgCl2 1.6 0.49 1.676 0.6
MgSO4 … … 0.758 0.7
Na2 SO4 nil … nil …
CaCl2 … 0.81 nil …
CaCO2 … 0.15 … 0.2
* Stanley Smith, Martin’s “Industrial Chemistry, Inorganic”, Part II, Vol. I, p, 304, London.
Crosby, Lockwood & Son.
SEA SALT

Sea salt is obtained by solar evaporation of the weak brine from the sea. Flat grounds near the
seacoast which are low enough so that the tide can be conducted in by canals, and which have
dense clay beds, make good evaporation bottoms. The sea water is concentrated on the flat open
fields or clay vats, generally very shallow, by the agencies of solar heat and sea breezes. Localities
on the seacoast with a high rate of evaporation of water, e.g., 5 to 10mm. for 24 hours, are best
suited for solar evaporation. To circulate the sea water from evaporating fields to the creek and
back from the creek to the fields to hasten evaporation, pumps usually driven by windmills are
employed. When brine has been thus concentrated up to the crystallization point, it is pumped to
crystallizing fields or pans (similar to evaporating fields, only shallower), where, on further
evaporation, crystals of salt separate out on the bottom. On these crystallizing fields, the brine is
less than 6 inches deep. The salt crystals are raked out by hand from the smooth bottom of such
fields. If the salt is not contaminated by dirt through handing, etc., there is seldom over 1 /2 per cent
of clay or insoluble matter in it when received at the works. The amount of the impurities included
in the crystals will depend upon the size of crystals and the method of preparation.
The procedure according to one practice is as follows: Sea water is conducted in at high tide,
or sent in by means of pumps, through canals leading from the coast to the salt fields. The sea
water is stored in one or more large reservoirs or ponds, and a system of canals is provided
connecting these ponds with a number of concentrating fields which are about 80 feet wide by 120
feet long by 6 inches deep, and here the brine is brought to a about 18 Be. From the
concentrating fields, the brine is led in small canals to the crystallizing fields which are about 40
feet wide by 60 feet long by 5 inches deep. This transfer of brine from the reservoirs to the
concentrating fields and from the concentrating fields to the crystallizing fields is effected usually
by windmills. In the crystallizing fields, the brine is brought to 26°Bé and allowed to crystallize.
The crystals at the bottom are raked off with long wooden rakes, collected, and stored in a
common spot where they are allowed to stand so that impurities of magnesium chloride, etc. are
drained off. By washing the salt crystals with saturated brine in the fields and draining, most of the
impurities can be washed off and the salt thus obtained contains 90-92 per cent NaCl.
For the location of such solar evaporation salt fields, the following requirements must be met:
(1) The sea adjoining the land must be in the shape of a bay or a gulf, not directly subject to
any ocean currents, and into which no large rivers or fresh-water bodies from the land
empty, so that local sea water may possess a high salt content.
(2) The coastal plan must be flat, open, low and not hilly, affording large areas for
concentrating the brine by solar evaporation and having more or less constant breezes to
aid evaporation.
(3) The locality must not be subject to heavy precipitation or high humidity, so that a high
rate of evaporation may be maintained, especially during the warmer seasons.
With the forgoing pre-requisites, it is not difficult to arrange for large clay vats or concentrating
fields having a dense smooth bottom for holding the brine and to install numerous windmills for
pumping and circulating the brine.
Besides salt (NaCl), sea brine contains many impurities. Fortunately, crystallization of salt by
solar evaporation is also a means of purification, i.e.; freeing the sodium chloride from impurities.
Therefore, it is to be expected that the bittern left after crystallization would have a higher main
impurities contained in a somewhat diluted bittern.
T ABLE 21. Composition of Bittern.
Per Cent
By Weight
NaCl 14.16
MgCl 24.64
CaSO4 0.51
 MgSO4 0.48
The other impurities were not determined.
Within recent years, manganous sulfate has been used on brine fields to obtain better and
coarser crys tals, as the result of high large clusters of crystals of sodium chloride are formed in the
crystallizing fields, yielding directly a crop of salt crystals containing as high as 95-96 per cent
NaCl with a fraction of one per cent of magnesium and calcium salts respectively left in the salt.
The proportion of manganous sulfate used is about 0.06 per cent on the weight of the saturated
brine, and may be reduced to one part per 50,000 by weight if the bittern is used over again with
the fresh brine repeatedly. No manganese salts are found in the salt crops obtained. The crystal are
large and hard and resemble weathered salt in dryness but they are in clusters. The yield of salt
from the brine is also increased by about 10 per cent. One crop gave the analysis shown in Table
22.
TABLE 22. Analysis of Crude Sea Salt Obtained from Sea Brine with Addition of
Manganous Sulfate.
Per Cent
NaCl 96.89
CaSO4 0.57
MgCl2 0.17
MgSO4 0.03
Insoluble Matter 0.33
Water (by. diff.) 2.01
The content tin magnesium salts is remarkably low (of ordinary sea salt analysis, Table 16).
The manganous sulfate is in the form of a crude bluish-gray powder, known in the trade as
“Mangan Crystallizer.” One commercial sample of the manganous sulfate used for this purpose (a
grayish dry powder) has the analysis shown in Table 23.
TABLE 23. Composition of Crude Manganous Sulfate Used as Crystallizing Assistant.
Per Cent
MnSO4 81.52
CaSO4 0.37
MgSO4 0.79
Fe2 O3 1.04
Al2 O3 0.02
SiO2 7.70
Silicates 1.20
Carbon 2.20
Moisture 5.03
Total 99.87
As shown above, sea salt contains as impurities considerable proportions of magnesium and
calcium salts. But the most economical method is not to work the brine for solid salt and then
dissolve it again to make saturated brine for use in the soda works, but to use the strong brine
directly from the evaporating fields, adding only what is necessary to bring it to full saturation. In
this case, however, all the impurities present in the seawater are retained in the saturated brine thus
obtained, and the difficulties of mud formation in the settling system, as described above, are
greatly enhanced. Hence pretreatment of the brine is essential. To guard against undersaturation of
brine direct from salt fields, means should be provided to bring the brine to full saturation by
dissolving in it solid salt in a rotary dissolver installed at the works.
Saturated brine has a specific gravity of 1.204 at 15 but impure brine containing large
amounts of dissolved salts has a little higher specific gravity reading. Theoretically, pure saturated
brine should contain 318 grams per liter of sodium chloride.
Table 24 gives the composition of tow grades of crude sea salt, representing the low and high
grades. The high grade is produced by washing the crystals with the saturated brine and draining,
as mentioned above.
TABLE 24. Composition of Crude Sea. Salt.
Low Grade High Grade
Per Cent Per Cent
Insoluble matter 0.58 Insoluble matter 0.24
NaCl 84.60 NaCl 89.83
CaSO4 1.02 CaSO4 0.61
MgCl2 2.56 MgCl2 21.49
MgSO4 0.81 MgSO4 0.40
Water and undetermined Water and undetermined
Substances by diff. 10.43 Substances by diff. 7.43
The data given in Table 24 represent the composition of fresh crops from crystallizing fields. If the
salt after harvesting is stored in mounds of thousands of tons for a year or more, by the natural
process of weathering, much of the impurities (especially MgCl2 ) is dissolved out and eliminated.
The composition would then be as shown by the analysis in Table 25.
T ABLE 25. Composition of Weathered Crude Salt after Storage for 12 Months.
Per Cent
Insoluble matter 0.64
NaCl 98.29
CaSO4 1.17
MgCl2 0.44
MgSO4 0.12
Water and undetermined
Substances by diff. 5.34
Table 26 gives the rates of solar evaporation at San Francisco Bay, U.S.A; and at Tangku,
China. The figures represent inches of water evaporated during the month.
TABLE 26. Evaporation in Inches.
Month *San Francisco Bay Cal., in 1911 Tangku, China in 1992
January …. 2.63
February …. 3.43
March …. 5.65
April 3.38 8.21
May 5.31 10.28
June 6.62 11.96
July 7.81 12.10
August 7.81 13.09
September 4.94 8.00
 October 2.94 8.06
November …. 4.38
December …. 2.77
Total 38.81 (for 7 months) 90.56
* Technology of Salt Making in U. S., W. C. Phalen, Bur. Mines Bull. 146, p.41.
The annual rate of evaporation at Searles Lake, Calif; is 72 inches.*
To give some idea of the world production of salt, it may not be out of place to give here
statistics compiled by the Bureau of Mines, the United States Department of the Interior, for
the year 1940.
* Chapman, L. W., Met. Eng., 24,687 (1921).
 Chapter V
Purification of Brine
Salt used in the ammonia soda industry is seldom pure and may have many constituents
besides sodium chloride, such as sodium sulfate, sodium iodide, sodium bromide, calcium
chloride and sulfate, magnesium chloride and sulfate, potassium chloride and sulfate, iron,
manganese, clay and sand. The amount of these impurities varies with the source of the salt supply.
For the ammonia soda industry, salt may come from one of the following sources:
1. Rock salt or rock-salt brine (artificial brine)
2. Natural brine
3. Sea salt
4. Sea brine (concentrated to saturation)
Rock salt is generally fairly pure, frequently containing over 98 per cent NaCl. As described
in the previous chapter, it occurs underground, and is obtained either by mining or by sending
water down a hole to dissolve the salt, and forcing saturated bring up by pumping, by air-lift, or by
the pressure of the entering water itself. Such brine is known as rock salt brine or artificial brine,
and is generally very pure. Gypsum is the chief impurity in rock salt, and magnesium is present
only in smaller percentages. Pretreatment of brine by addition of certain reagents to the water
before it is pumped into the wells, there by using the rock salt cavity as a setting reservoir, has
been suggested, but so far has not been practiced. However, some natural purification takes place
due to the alkalinity of the water used and to the insolubility of certain impurities and dirt which
settle at the bottom of the cavity, causing “blanketing”(See Chapter IV under Rock Salt.)
Natural brine occurs underground and exists either as a rock salt bittern or as brine formed by
the underground water flowing through salt strata. It is generally not so pure as rock salt or
rock-salt brine, especially when it exists as a natural bittern, such as the natural brine obtained at
Bay City or Midland, Michigan. Moreover, natural brine is seldom obtained in a completely
saturated condition. The best natural brines have only about 95-97 per cent saturation. Other
natural brines are very weak, some having only 6-12 per cent NaCl. Such brine must be
concentrated by means of steam evaporators or grainers, by solar heat, or by the fortification with
rock salt, before it can be used in the ammonia soda industry. Many ammonia soda plants use
natural brine, notably the I.C.I’s plant in Cheshire, England.
Sea salt is generally obtained by solar evaporation; besides sodium chloride the crystals
contain sodium sulfate, calcium sulfate and chloride, magnesium sulfate and chloride, etc. it may
be obtained in abundance and cheaply along the sea coast, where climatic and geographical
conditions are favorable. It contains more impurities than rock salt and more magnesium relative
to calcium in the impurities. It generally requires pretreatment for the ammonia sod manufacture.
Sea brine is a saturated brine obtained directly from salt fields where sea water is
concentrated b solar evaporation merely up to saturation, from which solid salt has not been
allowed to crystallize. The brine will contain all the impurities in the sea water and have far more
impurities than rock salt or sea salt. Pretreatment is necessary before it is used in the ammonia
soda process. This form of be, however, is very economical and does not require the roundabout
method of evaporating sea water to crystallize out solid salt and then re-dissolving the salt crystals
to make brine. It may be necessary to add some solid salt to it for fortification, to guard against
undersaturation. When sea brine is obtainable in abundance in this way, with a suitable means of
pretreatment it forms a very economical raw material for ammonia soda manufacture.
Pretreatment to brine for the ammonia soda process is now more generally practiced than
heretofore. This consists of treating the brine to eliminate calcium and magnesium before it is sent
to the different washers, and finally to the ammonia absorber. It eliminates the formation of scale
or sludge, consisting of calcium carbonate and magnesium carbonate and / or hydroxide in the
absorber system which causes obstruction in the passages, plugging in the pipe lines, clogging in
the setting vats, and many disagreeable operational difficulties. Further, sludge formation incurs
losses of ammonia and salt. Because of the necessity of frequent changeovers of the apparatus and
the cleaning thereof, it requires labor and time and causes loss of production. Also, some
impurities, notably magnesium, may fine their way to the finished product, giving cloudiness or
insoluble matter in the soda solution. Hence present practice generally resorts to some method of
purifying the brine before use.
There are several methods of treating brine, of which the more common ones may be
enumerated here:
(1) Treatment with ammonia and carbon dioxide gas.
(2) Treatment with lime, and ammonia and carbon dioxide gas.
(3) Treatment with lime and soda ash.
(4) Treatment with lime and sodium sulfate (followed by ammonia and carbon dioxide gas)
(1) Treatment with ammonia and carbon dioxide gas. This method of treatment is really
a part to the ammonia soda process and will be dealt with fully in Chapter VII,
“Admonition of Saturated Brine. “it eliminates calcium and magnesium by means of
NH3 and CO2 in hot solution. The difficulty lies in the removal of the sludge from the
settling vats and in the plugging of the piping and settling system. This difficulty is
aggravated by the fact that such treatment must be done in a closed system under a
partial vacuum to avoid loss of ammonia. Besides the necessity of guarding against the
loss of ammonia, the disposal of mud from the strong ammoniated brine by pumping it
to the distiller is accompanied by a considerable loss of salt. This method is
uneconomical when there are large amounts of impurities, such as in sea salt. It also calls
for a special provision in the absorber-vat system and in the arrangement of the washers
and absorbers. However, the treatment is carried out in the hot condition prevailing in
the ammonia absorber, and has a most favorable condition for settling. The reactions are
as follows:
Mg++ + 2NH4 OH Mg (OH) 2 + 2NH4 +
++
Ca +(NH4 )2 CO3 CaCO3 +2NH4 +
But it was found that in the presence to CO2 , magnesium is apt to come down as a carbonate or
basic carbonate instead of as hydroxide, and also as a double salt (NH4 )2 CO3 MgCO3 4H2 O, or
as a triple salt, NaCl MgCO3 Na2 CO3 , along with other salts, thus causing further loss of salt
(NaCl) as well as of (NH4 )2 CO3 and Na2 CO3 from the ammoniated brine. (See Chapter VII,
“Ammoniation of SATURATE brine.”) Moreover, in place of calcium and magnesium salts, the
corresponding ammonium salts are formed in the ammoniated brine, which may lower the
decomposition of NaCl in the columns. Hence this method is not economical when the raw brine
contains high calcium and magnesium impurities, besides the attending operational difficulties.
 A modified form of this treatment is the use of large settling vats with closed top and conical
bottom, each provided with a stirrer and treatment of brine in these vats with ammonia gas from
the distiller or from the weak liquor still. The brine is thus partially ammoniated and the impurities
of calcium and magnesium are allowed to settle off in these vats. The clarified brine is then sent to
the absorber. In this way, brine is pre-ammoniated and settled before it passes through the absorber.
This eliminates most of the impurities in the brine before it reaches the absorber and thus prolongs
the service life of the absorber system. Such a method is in common use in the alkali plants on the
Gulf Coast of the United States.
(2) Treatment with lime and ammonia and carbon dioxide gas. Advantage is taken of
the fact that Ca (OH) 2 even in cold brine very readily precipitates magnesium. Using a
theoretical quantity of lime, it is possible to eliminate 97-99 per cent magnesium from
the brine this way, and the precipitate may be settled off by decantation. The resulting
brine contains calcium equal to the amount originally present in the brine plus what was
added for the elimination of magnesium.
Mg++ + Ca (OH) 2 Mg (OH) 2 + Ca++
Since it is known that calcium can be readily precipitated as CaCO3 and settles more or less
readily, and since calcium carbonate mud does not form the kind of sticky, hard scale as does
magnesium mud, it can be drawn from the settling vats. Therefore, the idea underlying this
treatment is first to remove the troublesome and later to remove all the calcium by the regular
ammoniation process.
Ca++ +(NH4 )2 CO3 CaCO3 +2NH4 +
Although this does not form a double or triple salt in the mud from the ammoniated brine, it has
the same objection of leaving much fixed ammonia in the resulting ammoniated brine. The
treatment is particularly good for sea salt, which contains high magnesium.
(3) Treatment with lime and soda ash. As has been mentioned above, the removal of
calcium and magnesium from the raw brine by the ammonia and CO2 treatment results in the
formation of an equivalent amount of ammonium salts in place of calcium and magnesium salts in
the resulting brine. This substitution of ammonium salts for calcium and magnesium salts, when
present in large quantities, will interfere with the decomposition of salt by ammonium bicarbonate
in the columns. It is therefore preferable to introduce soda ash in place of ammonium carbonate
for the elimination of calcium after the precipitation of magnesium by lime.
Thus, sodium salts are formed instead of ammonium salts in amounts equivalent to the calcium
and magnesium salts originally present, and these sodium salts will form sodium bicarbonate in
the columns in the same way as sodium chloride, giving better decomposition. However, this
pretreatment must necessarily be done in the cold. Because the brine so treated has to be used in
various washers to scrub spent ammonia gases, on its way to the absorber. There will be certain
difficulty in settling the mud from the cold brine. When, however, the process is adequately
controlled, the magnesium and calcium precipitates form a reasonably quick settling
“floc”(probably also as a double salt), even in cold concentrated brine whose viscosity during
winter operations amounts to 3 centipoises. Therefore, given a properly designed settler, clear
brine with a highly concentrated sludge can be obtained. For this purpose, a Spaulding type of
“precipitator,” or reaction-thickener (see Fig.107, p.385), is most suitable. It is interesting to
observe that, as long as sufficient lime is present in the brine (but not too large an excess), the rate
of settling is quite rapid even in cold brine. A characteristic -settling curve is shown in Fig.14.Also,
when a sufficient quantity of lime (equal to, or in slight excess of, the equivalents of magnesium
present) is so employed, the removal of magnesium from the brine can be as high as 97-99 per
cent. Table 28 shows the composition of brine before and after treatment.
This method of purification requires a slight excess of lime and soda ash, and the finished brine
has a pH value of 9.0.The lime used so best in the form of milk of lime or hydrated power,
because quicklime often contains overburnt or underburnt pieces which do not dissolve readily in
brine .The slight dilution in the purified brine shown above is due to water added in the milk of
lime used. On the other hand, the gain in Na is due to the soda ash added. In rook-salt brine, the
magnesium present is sometimes negligible, so that lime may be omitted. Small amount of brine is
lost in purging the mud from the pretreatment system, but with proper thickener design, this loss is
seldom over 1 per cent. This is indeed negligible in view of the fact that the utilization of brine in
the process itself is not more than 72-73 per cent.
This method is advantageous because there is no ammonia loss involved and all calcium and
magnesium salts are converted to the corresponding sodium salts available for ammonia soda
production. However, it involves the consumption of soda ash.
(4)Treatment with lime and sodium sulfate. The underlying principle in this treatment is
first to precipitate magnesium with lime, as mentioned before, and then to precipitate most of the
calcium as CaSO4 by the addition of sodium sulfate. Thus, sodium sulfate is substituted for soda
ash, but it does not completely throw down all the calcium, especially in the cold condition. Using
a large excess of sodium sulfate-50 per cent or more of the theoretical quantity-we may eliminate
under good conditions 85-90 per cent of the total calcium present in solution after lime has been
added for the precipitation of magnesium in the raw brine. But the reaction is preferably carried
out in the warm condition, and at least two hours’ setting time must be allowed for coagulation.
The much larger quantity of sodium sulfate required may appear, at the first thought, to be
uneconomical; but if a crude sodium sulfate bittern with high salt content is available, this cheap
material may be utilized for this treatment .The use of excess sodium sulfate does no harm, as
sodium in the form of sulfate or chloride is equally available for the production of sodium
bicarbonate in the columns. On the other hand, no ammonium salts are formed, as would be the
case if ammonia and carbon dioxide gas were used .The last portion of calcium left in the treated
brine may be thrown out with a small amount of soda ash, or it may be taken care of in the settling
vats in the process of ammoniation.
Reference has already has already been made to the objection to the presence of ammonium
salts in ammoniated brine, especially when sea brine is used, when ammonia and carbon dioxide
gas (ammonium carbonate) are employed for the treatment .In the tower operation, the
decomposition of sodium bicarbonate yields ammonium chloride, as shown in the following
reaction:
NaCl+NH4 HCO3 NaHCO3 +NH4 Cl
It is to be expected that the presence of ammonium salt in the ammoniated brine (i.e., the presence
of fixed ammonia) will shift the equilibrium point to the left of the equation, suppressing the
reaction. Indeed, when fixed ammonia is present in excess of 35g (as NH4 Cl) per liter of the
ammoniated brine, the reaction is noticeably affected, and the decomposition of sodium chloride is
lowered. This corresponds to about 0.65 equivalent of the combined calcium and magnesium salts
per liter of the raw brine .In terms of sodium chloride concentration in the ammoniated brine, this
represents 14.5-15.0 per cent of Nazi. Normally, in rock-salt brine, the combined calcium and
magnesium salts are present to the extent of 0.045-0.050 equivalent per liter (about 70-80 per cent
Ca, 20-30 per cent Mg), and in sea salt that are 0.050-0.055 equivalent per liter (about 30-40 per
cent Ca, 60-70 per cent Mg). This indeed exceeds the limit mentioned above and would cut down
the decomposition of NaCl in the columns by 5-7 per cent. This difficulty is aggravated because of
the much large proportion of magnesium present in the impurities .The extent to which
decomposition is lowered depends upon the amount of fixed ammonia present. Hence, in the case
of sea brine containing impurities beyond 0.65 equivalent of the combined calcium and
magnesium salts per liter, it is preferable to use sodium carbonate instead of ammonium carbonate,
i.e., ammonia and carbon dioxide gas, for the precipitation of calcium. However, as discussed in
Chapter XXVII, when ammonium chloride is to be recovered from the mother liquor, which is to
be made up and sent back to the columns, the treatment with ammonia and carbon dioxide gas
entails no such disadvantages.
 Chapter VI

Burning of Limestone
GENERAL AND T HEORETICAL
Line burning is one of the oldest industries carried on by mankind. From ancient times, lime
has been used as a building material, although the chemistry underlying the process was not
understood until the time of Lavoisier. Lime burning consists in driving out carbon dioxide gas
from limestone by heat. This occurs at a comparatively high temperature, because the
decomposition absorbs heat. The amount of heat absorbed per gram mol of limestone decomposed
is 42900 calories at room temperature or 38500* cal. at the temperature of decomposition. The
former is the heat effect for the process starting and ending at 15 . and at constant pressure. The
thermo-chemical equation is:
CaCO3 Cao+CO2 -42,900 cal.
The minus sigh before 42 900 cal. shows that this amount of heat is absorbed per g-mol of
limestone during decomposition at room temperature. The heat is supplied by the burning of some
fuel, generally coke. Anthracite is sometimes used in the limekilns in the ammonia soda industry,
but there would be more impurities in the gas.
There is a definite vapor pressure of carbon dioxide gas from limestone at a given
temperature, just as there is a definite aqueous tension above water at a given temperature. Like
aqueous tension, too, this vapor pressure of carbon dioxide increases very rapidly with the
temperature. Like the system of water with its vapor, this system, CaCO3 CaO CO2 , is a
mono-variant one, there being two independent components, CaO and CO2 , and three phases, two
solid phases, CaCO3 and CaO, and one gaseous phase, CO2. Hence there is only one degree of
freedom.
The equilibrium of the reaction, CaCO3 CaO + CO2 ,gives
(solid) (solid) (solid)

where Cco2, Ccao, and Ccaco3 are moral concentrations of the respective components, but both
Ccao and Ccaco3 are constant, being present in solid form.
Cco2 =Kc , or Pco2 =Kp;
*J. Johnston, Thermal Dissociation of CaCO3 J.Am. Chem.Soc.,938 (1910)
where CCO2 is the molal concentration and P CO2 is the vapor pressure of CO2 gas.
Table 29 shows the vapor pressure of carbon dioxide from calcium carbonate at a given
temperature according to the investigation by Johnston. *
From Pco2 =Kp above, by the van’t Hoff equation, we have

where H is positive when the heat is absorbed. It is seen that Kp. increases, i.e., the partial
pressure of carbon dioxide in equilibrium with the limestone increases as the temperature is raised.
Since Kp is numerically equal to the partial pressure of carbon dioxide, with two sets of
temperatures and the corresponding partial pressures of carbon dioxide, it should be possible to
calculate the theoretical heat of decomposition of calcium carbonate. If we accept Johnston’s
figures, i.e., at 891 .PCO2=684mm., and at 894 . PCO2=716mm., we may calculate the heat of
decomposition as follows. Within the range of 3 . we may assume that H is constant, and the
expression after integration between the limits 891 and 894 . is shown below.

whence

Although theoretically a temperature of 898 . Is sufficient to decompose limestone with a

vapor pressure of CO2 of one atmosphere, the rate of dissociation is slow; thus, in practice, the
temperature in a lime kiln is considerably above 900 C.C. Furnas * in his study of the rate of
calcinations gave the relationship between the temperature at which calcinations takes place and
the duration of time required as shown in Fig.15.Calcination temperature is generally takes to be
940 . while the temperature at the decomposition zone is more nearly 1050 .under kiln
conditions. Sometimes the temperature rises as high as 1200 .When the lime is allowed to
remain for a few hours at this high temperature, it is changed to a denser mass, showing a dark,
grayish, or yellowish tings with minute checks. This dense lime presents fewer surfaces and
requires a much longer time for solution, and the particles suspended in the milk of lime react
more slowly with the ammonium chloride solution. Such lime is called overburnt or dead-burnt
lime. Besides giving difficulties as the result of its slow reaction in the prelimer and lime still, it
may cause difficulties by lowering the settling rate in the causticizing operation. Properly burnt
lime is practically cream-white and is readily hydrated to a loose, pure-white power. Hence the
name “fat lime.”
*Furnas, C. C., The Rate of Calcination of Limestone, Ind. Eng. Chem., 23, 534 (1931)

FIG 15

Relation between time

required for calcinations,

temperature, and thickness

of piece (Furnasilats).

Lime is a comparatively active base. It is sometimes called quicklime or caustic lime.

Hydrating consists in converting the oxide to hydroxide by adding just enough water to the lump
when the mass disintegrates and turns white, but does not become pasty or wet. Considerable heat
is involved during hydrating or slaking. the following thermochemical equation shows the heat
effect.
 Courtesy The Kritzer Co.

FIG 16 Kritzer continuous cylindrical hydrator.

CaO + H2 O Ca(OH) 2 + 15.540cal.(Thomsen)

(solid) (liquid) (solid)
The plus sign shows that the heat is evolved when quicklime absorbs water in slaking. During
hydrating the lump disintegrates to a fine powder leaving unburnt stone in a lump form. This is a
good way to separate lime from the unburnt core.
 FIG 17 The Schaffer continuous hydrator.

Hydrated lime has become a favored form in which lime is handled in commerce. Various
types of machines called hydrators have been designed for hydrating the lump lime. They consist
of either a vertical shelf construction, such as the Schaffer continuous hydrator, or of sections of
stationary cylinders, one on top of another, with a rotating screw ribbon in each cylinder, such as
the Kritzer continuous hydrator. Lumps of lime are charged into the top and a small excess of
water over the theoretical quantity necessary to form Ca (OH) 2 is sprinkled over them. The
stirring action helps to distribute the water and scraps the lime along. In the Kritzer type (Fig.16) a
number of cylinders one on top of another, work the lime sown in a zigzag way, taking the lime
from a small pre-hydrator on the top into which water is sprinkled and discharging the hydrated
lime from the button cylinder, much in the same way as the Hasenclever mechanical chamber for
the bleaching powder manufacture (see Chapter XX under “Manufacture of Bleaching Powder”)
The Sc haffer machine (Fig.17) consists of a number of shelves placed one on top of the other with
revolving rabbles arranged much in the same manner as in the Herreshoff roaster. The hydrated
lime is then screened to separate out unburned cores. Hydrated lime has many advantages over
lump lime .It is purer and more uniform, does not burn the bags id they should get wet during
transportation, does not readily become “air-slaked,” and has the advantage of being “aged.”
Quicklime (CaO), on exposure to air, absorbs carbon dioxide from the air and reverts to calcium
carbonate. Such lime is termed “air-slaked;” for the reaction is a reversible one:
CaCO3 CaO+CO2
At ordinary temperature, the partial vapor pressure of carbon dioxide in equilibrium with calcium
carbonate is so very small (only 1 mm. at 587 . And 23mm. at 701 ) that the small amount of
carbon dioxide gas in the air is sufficient to set up equilibrium and convert calcium oxide (lime)
back to calcium carbonate.
Lime ( CaO ) is very slightly soluble in water, its solubility product being 0.64×10-5 at 20 .
(Shipley and McHaffie). In grams per liter. its solubility is given by Guthrie in Table 30:

An interesting fact is that its solubility decreases as the temperature is raised.

Lime is so sparingly soluble that its solution would be too dilute for general purposes.
Therefore it is introduced usually in the form of a suspension called milk of lime .In milk of lime
there is an equilibrium between the solid lime and the lime that has gone into solution, small as to
is. And what is dissolve is again in equilibrium with the calcium and hydroxyl ions in solution.
Thus,
Ca(OH) 2 Ca(OH) 2 Ca+++2OH-
(solid) (diss.)

As the reaction goes on (i.e., say as the OH- ions are used up) more Ca (OH) 2 (diss) must
dissociate and more Ca (OH) 2 (solid) must dissolve. Hence all solid particles of Ca (OH) 2 in the
lime suspension are thus readily available for reaction. Table 31 shows that milk of lime of
different concentrations has the following specific gravity given by Lunge.

With lime containing as impurities considerable quantities of sand, clay, finely divided unburnt
stone (CaCO3 ), etc. the specific gravity of the milk of lime is always greater and its determination
furnishes only a very rough guide. Milk of lime from a poor grade of limestone may have a
specific gravity of 1.3 and yet contain not more than 200 grams of CaO per liter.
Limestone is frequently found to contain magnesium carbonate, varying from 45 per cent in
the case of dolomite, CaMg (CO3 ) 2, to but a fraction of 1 per cent in a high-grade limestone .A
large percentage of MgCO3 in the limestone is rather objectionable for its inertness in the
distillation of ammonia .It may be remarked here that the presence of magnesium in lime is, as a
rule, undesirable. One can think of the objection to magnesia in the cement industry, in the
bisulfate process in the paper industry, in the calcium carbide process in the air nitrogen industry,
in the manufacture of blearing powder, and in the lime mortar and lime plaster (as” lean lime”).
For furnishing CO2 gas in the ammonia soda industry, however, MgCO3 in limestone serves
equally well, and in fact, carbon dioxide gas from MgCO3 is driven off with greater ease; but for
ammonia distillation, magnesia in the lime is entirely unavailable .The thermochemical equation is
MgCO3 MgO +CO2 -28, 900 cal. (de Forcrand)
The minus sign again shows the heat absorbed during decomposition per mol MgCO3 . Table 32
gives data on the vapor pressure of carbon dioxide from MgCO3 (native magnesite) by Mitchell.

The heat of decomposition of MgCO3 is only 28,900 cal. while that for CaCO3 is 42,900 cal.
(room temperature) and the decomposition temperature for MgCO3 is 756 .as against 898 . For
CaCO3 The smaller quantity of molal heat absorbed and the lower calcinations temperature show
greater ease of decomposition of MgCO3 in comparison with CaCO3 .
The objection to MgO for ammonia distillation is not that it will not react with NH4 Cl and
(NH4 ) 2SO4 to free NH3 , but that its reaction is slow and must be carried out under special
conditions. Under certain conditions the following reaction does occur:
Mg (OH) 2 +2NH4 Cl MgCl 2 +2NH4 OH
As a matter of fact, go has been proposed for ammonia distillation in soda works, because it is
considered possible to recover by-products of MgCl2 from the distiller waste in the form of
chlorine, Mg2 OCl2 , HCl, etc. Under ordinary conditions, however, all the MgO passes through the
distiller unreacted upon, and only CaO is utilized .The solubility product of Mg (OH) 2 is only
1.2×10-11 at 18 . (Johnston) as against 0.64×10-5 for CaO at 20 .The heat of hydration of MgO,
too, is less.
MgO + H2 O Mg (OH) 2 +5,000 Cal. (Marignac)
(solid) (liquid) (solid)
The corresponding heat of hydration for CaO is 15,540 cal.

Furans (loc. it.), in his study of the calcinations of limestone, differentiate the rate of
calcinations and the rate of heat penetration (temperature wave). By the rate of calcinations is
meant the rate of advance of the line of calcinations from the outside toward the center of the
stone, which is dependent on the heat conductivity of the limestone; and by the rate of heat
penetration is meant the rate of advance of temperature wave. It is found that decomposition
(calcinations) in the center of the stone may not take place, even if the calcinations temperature
has been reached there .The center portion is then in a meta-stable state. According to Furnas, the
rate of calcinations (R) is a logarithmic function with respect to the temperature (t). Thus,

FIG 18 Rate of advance of line of calcinations vs. temperature of surroundings(Furnas data).

Unfortunately, Furans’ study covered only the size range between 2.5cm. (1”) and 8.5cm. (33 /8”),
but he stated that it might hold up to 15-20 cm. (6”-8”). It is thus seen that the length of time
required for complete calcinations is directly proportional to the size (diameter) of the stone (as
expected), other conditions being equal. In each individual piece of the stone, the calcinations
zone or boundary is very sharp and the temperature travels faster than line of calcinations, so that
even though the center portion might have attained the temperature of the outside surroundings,
decomposition lags behind. Decomposition of limestone, like the boiling of water, is marked by a
definite temperature which remain constant until the process is complete, after which the
temperature begins to rise to approach the outside temperature .The rate of advance of the lime of
calcinations may slightly decrease as calcinations progresses, because of the longer path through
the calcined zone, over which heat for calcinations must travel.
* Mitchell, A. E., J. Chem. Soc. (London), Trans., Part I, 123, 1055.
Based on the above, the rate of CO2 gas evolution may be estimated for a given piece of
high-quality limestone as follows:

The rate of the advance of the boundary line is R radially. The depth of calcinations boundary (of
line of calcinations from the surface) is R at any given instant. Consider a spherical form. The
diameter of the spherical boundary then is D-2R , and spherical surface of the boundary is .The
molecular weight of CaCO3 is 100 and the G.M.V. is 22.4 liters.

The rate of advance of calcinations inward depends upon the temperature only and is specific for a
given grade of limestone, i.e., R is independent of * as a first approximation. Integrating, we
obtain,

At 1052 ., which is near the average temperature of calcinations in kiln practice, R is very nearly
one cm.per hour on the average for a high-grade limestone (about 97 per cent CaCO3 and * per
cent MgCO3 ).
Therefore, V=0.61p (3D2 -6D +4 2 )
The above equation holds true only for a high-grade limestone, because it has not taken into
consideration CO2 that comes from the small quantity of MgCO3 present. Further, it is based on
the assumption that the rate of calcinations is constant and is approximately 1 cm. Per hour at a
temperature of calcinations of 1052 . This rate of calcinations would be considerably above 1cm.
When there is a higher percentage of magnesium carbonate present in the limestone, due to the
greater ease with which MgCO3 is calcined (decomposed) compared with CaCO3.

QUANRRYING OF LIMESTONE AND PREPARATION OF THE CHARGE

Limestone is generally quarried by open pit methods. Unless the over-burden is thick, mining
is seldom resorted to. Often the surface soil is just stripped off and open stop is employed. But
where the top soil is extensive and thick, or where the upper layers consist of a high silica or
alumina limestone (such as the cement rock frequently encountered), tunneling from the sides into
lower levels, where there is a high quality limestone, is a common practice. In such cases quarry
cars are usually run on rails into the tunnels at a slight grade .As a rule, pumping to keep the pits
dry is a very essential operation as in all mining stopes, if there is much underground water.
Details of quarrying depend upon the physical characteristics of the rock, its composition, its
bedding angle, and the amount of over-burden. Where the formation permits, deep faces from 30
to 125 feet are employed, but where the bedding is flat, the face may be any position, governed by
the convenience in the transportation of the stone .If the beds are steep and almost vertical, the
face should best be at right angles to the strike of the beds rather than parallel to it. Blasting of the
rock then should start from the bottom of the face, as it is easier to remove the stone from the
bottom. Dynamite is usually used for blasting. Dynamite sticks are 8 inches long, but may be 7 /8
1,11 /8 11 /4 etc, in up to 4 or 5 inches in diameter. The 11 /2 -inch size is the commonest. Each of
the 11 /4 by 8-inch sticks weighs about 1 /2 lb. and a large number of these sticks are generally placed
in one hole, one above the other. Fro a limestone quarry, it is generally figured that on the average
3 to 4 tons of stone are blasted per lb. of dynamite charge.
Blasting is done by drilling the rock with jackhammers, generally of 55-lb.weight with 1-inch
hexagonal hollow drill rods. These jackhammers require 80-90 lbs. of air pressure, and each
requires for operation 90-110 cubic feet of free air per minute, depending upon the air pressure.
Usually one man work on each jackhammer of this size. The drill bits may be detachable or
integral with the drill rods. Detachable bits may be re-ground three or four times before they are
worn and discarded. Integral bits may be clamped, shaped, formed and sharpened from the drill
rods in suitable die blocks and dillies and the center holes punched, using a drill sharpener (also
driven by compressed air), until the steel stock is used up. The bits may be 11 /2 13 /8 and 11 /4 or
13 /4 ,11 /2 and 11 /4 inches in diameter respectively. The hole are drilled to 12 to 16 feet, in steps,
using 1-inch hexagonal drill rods. The holes are drilled in rows at a distance of 5 to 6 feet between
centers. They may be drilled horizontally or vertically, depending on the position of the face or
large. Drilling may be done dry, wet, or dry with vacuum suction using a dust separator (such as in
the “drill-vac” outfit). Simple dry drilling is not to be recommended as it creates too much dust in
closed space for the health of the workmen. Where the rock is silicious, laws are quite strict in
prescribing the dust limits to protect wokmen against silicosis.
For large-scale quarrying operation, holes as 6 inches in diameter are drilled into the rock,
and may be as deep as 125 feet, or about 5 feet below the floor of the face. These holes are drilled
about 20 feet from the edge and are spaced at 20 foot centers in rows about 20 feet apart.
Dynamite blasting is done by special workers. These holes at a greater distance apart are used for
economy of labor.
When tunnelling is required, the tunnels may be from 35 to 50 feet wide by about 40 feet
high.” Pillars” are left between tunnels or “drifts” for roof support. The width (or diameter) of
these pillars varies from the width of the tunnel to * times the width, and the roof of the tunnel
may be flat or arced, depending upon the strength of the rock. As a rule, timbering is seldom
necessary. Shafts or “raises” are located at high places at the back of these tunnels for the purpose
of ventilation. Except in extreme cases, a couple of these raises, suitably located, suffice for
natural draft ventilation, and no mine fan or exhauster is necessary.
Blasting in tunnels is done by drilling holes first at the top horizontally into the face and
blasting off the top portion to form a bench. Then the holes are drilled vertically into the ledge at a
proper distance from the face. Generally these holes are drilled in rows and spaced 5-6 feet apart,
and about 8 feet from the face. Dynamite is loaded into the holes and fuses are put in with varying
lengths increasing by about 6 inches from the center to each of successive holes toward the side,
so that when electrical contact is made the powder is set off first in the center holes and then in the
other holes a few seconds apart from the center outward.
Frequently in large quarries a heavier drill mounted on a portable carriage and suspended on
a chain from a small mast is employed. Such drill is then operated by two men to each drill and
have maximum maneuverability .The carriage may be anchored in any position and the drill
adapted to different locations. Deep holes may be drilled this way.
The stone is removed by steam shovels and loaded directly into cars on a track or into trucks,
which haul it into the crushing and screening plant.
One source of serious trouble in quarry tunnels is the spelling or peeling off of the ceiling in
the tunnels, especially when the ceiling is a layer of shale. This is found to be caused by the humid
conditions in the tunnels (sweating during summer days), fluctuations of temperatures, and general
weathering. Falling more likely occurs near the entrances or at places exposed to outside
atmosphere. Such falling of pieces from the ceiling is a menace, endangers human lives, and
interferes with the quarry operation. Often painting the ceiling with a waterproof coal-tar paint
such as “EBONOL ” made by Sherwin Williams Company will prevent or minimize the trouble.
This may be applied by a spray gun. Sometimes, Gunite is a great help, and if properly applied,
will give an impervious coating and add strength to the ceiling .The Gunite should not be less than
an inch thick and may be applied in two layers, the first or bottom layer being only * inch thick
and the over * inch thick. Limestone surface is less liable to peel off this way. If the shale ceiling
the is thin with another layer of limestone just above it, it is best to take the whole shale roof down,
leaving the limestone surface exposed, which will be more permanent.
Ammonia soda manufacturers sometimes have quarry and haul limestone over a long
distance, either by rail or in river barges. Certain alkali plants are not so fortunately situated as to
have both limestone and salt close at hand .It is often a debatable question whether an ammonia
soda plant should be located near the source of limestone supply or near the source of salt supply,
if good limestone is found near the locality where salt is available. One of the plants on Lake Erie
in the United States has to ship limestone by water from the coast of Michigan .One of the
recently established alkali plants I the South has to transport limestone for a considerable distance.
An alkali plant in North China hauls its limestone to the plant site over a distance of from 56 to 90
miles by rail, and occasionally also by coastal boats over 200 miles across the sea. Limestone so
one of the two most important raw materials used in ammonia soda manufacture, and only a good
grade should be used for economical operation.
Limestone for ammonia soda operation is generally quarried under the supervision of the
soda ash plant management. Almost invariably some sort of over-burden must be removed and
main kiln supply must be blasted, crushed, screened, (washed if necessary), and stored away
preparatory to feeding the kiln .The most common procedure is to transport the stone from the
quarry to the plant after all crushing, screening and washing operations have been completed.
Occasionally, because of lack of water or other facilities at the quarry, the screening and washing
operations are conducted near the plant.
Often two strata of stone are worked at the same time, and these have different composition
as regards impurities and physical characteristics. This may necessitate proper mixing and
proportioning to obtain best results. If the variations in mixture do not involve any deep-seated
effect on kiln operations, accurate control in the blending is, of course, unnecessary.
Since the stone for any one kiln should be as uniform in size as possible, screening
operations are conducted at the quarry to give certain cuts or screen sizes of stone. The large sizes
in common use are about 6 inches in diameter. The range of sizes generally permits a ratio of
two-to –one or even three-to-one variation. If a quarry is operated for a plant having only one lime
kiln in operation, everything below the minimum allowable limit must be rejected and sold for
other purposes, such as railroad track ballast, road building, concrete aggregate, etc. If there are
several kilns in operation at the plant, one of them may be kept operating on the smaller sizes of
stone which result from the crushing and screening operations. This is quite feasible, provided the
latter are reasonably uniform in size. The smaller size simply requires somewhat greater blast
pressure for its burning. Vertical kilns very seldom use any pieces smaller than 2 inches, but the
small pieces of fines are occasionally brought to the plant and burned in rotary kilns.
The type of crushing equipment depends on the size of the quarry. The largest plants will
place the quarrying “shots” so as to produce relatively large pieces. Quarrying shovels, are, etc.
are then all proportionately large and all the quarrying product is put through a huge primary
crusher, sometimes capable of taking a piece 4-5 feet in the maximum dimension. In smaller
quarries this entire line of equipment is scaled down, requiring closer spacing of blast holes and a
larger number of “pot shots” for reducing the size of individual pieces shot down from the quarry
face. In some plants the preliminary crusher is of the gyratory type; in others the jaw type is used.
From the preliminary crusher the product is given a size separation, the over-size pieces being
returned to the crusher for father reduction. This is one example of closed-circuit operation. From
here on, secondary crushers and screens prepare the various grades of stone most suitable to the
particular operation.
Many quarries have an over-burned such that it may introduce an objectionable quantity of
clay or siliceous matter, which would obstruct the passages in the kiln by the formation of
low-melting clinkers. Washing operations may be necessary under these conditions, and more or
less standardized washing equipment is used.
The stone is transported from the quarry or washing plant to the kiln site by water, in
narrow-or standard-gage cars, on cableway systems, or occasionally by direct conveyors when the
two are close enough to each other .A stock pile of kiln size stone is kept at both ends of the
system to insure continuity of soda plant operation against breakdown or bad weather conditions.
Limestone is sedimentary in origin and its deposits often contain many impurit ies. It has
many different physical characteristics and chemical compositions. Natural products range from
the purest marble, calcite, aragonite, etc., to ordinary grades of limestone for industrial purposes.
There are argillaceous limestone containing considerable clay; erinaceous limestone containing
large amounts of silica; magnesia limestone containing 10 to 30 per cent magnesium carbonate;
and dolomite containing 45.65 per cent magnesium carbonate. These argillaceous and erinaceous
limestones, while very desirable for the manufacture of cement and hydraulic lime, cannot be used
in the ammonia soda industry on account of their silicious content. Dolomitic or magnesian
limestones are not desirable, or at least are uneconomical because of their magnesia content,
although small proportions (1 to 6 per cent MgCO3 ) can be tolerated. In general, the ammonia
soda industry demands a high –grade limestone in order to have successful and economical
operation. Especially is this true when the waste liquor from the distiller is to be worked for CaCI 2
or other by-products. Specifications are given in Table 33, although no strict line of demarcation
can be drawn:
Table 33. Speciflcations for Limestone.
Per Cent
CaCO2 90-99
SiO2 ,Al2 O4 ,Fe2 O3 0-3
MgCO2 0-6
 Since limestone is a natural product, its physical appearance (color, texture, hardness, etc) is
very much affected by impurities, although they may be present only in small percentages. Its
geological age of formation also has considerable bearing on its physical properties. Samples from
different deposits seldom look exactly alike. They range in color from deep blue, light gray, and
stripy white, to streaky brown. Their texture and hardness vary from a dense flinty appearance to a
loose crystalline cleavage. Both calcite and quartz are associated with limestone and appear white.
Dull color may indicate high magnesium, and crystalline luster ay be indicative of high silica
content. To an experienced eye, the general appearance of limestone will reveal much about its
composition, so that high-magnesia or high-silica limestone can be distinguished by inspection
with considerable accuracy.
In a quarry, close supervision is exercised in selecting the best face for working, the seam
that appears undesirable by inspection being left out or blasted down for some other work..

PRACTICAL OPERATION OF LIME KILN

Although lime burning is an important industry in itself, the ammonia soda industry
undoubtedly constitutes one of the biggest lime producers and consumers. Much of the
development of burning technique we owe to the alkali manufacturers.
Take for example the year 1937 when the quantity consumed by all chemical industries other
than the ammonia soda industry totaled 2151444 short tons, distributed as in Table 34
TABLE 34 Lime Consumed in Different Industries in United States (1937).
Glass Works 167,438 short tons
Metallurgy 694,817 ” ”
Paper Mills 447,728 ” ”
Sugar Refineries 21,211 ” ”
Tanneries 61,544 ” ”
Water Purification 212,213 ” ”
Other uses 546,496 ” ”
Total 2,151,444 short tons
If we include lime used in building and agricultural work, the total for the year 1937 was 3506439
short tons, thus:
*“Lime,” Minerals Year Book, Bureau of Mines, Dept of the Interior, Washington, D.C; 1938
Building 948,533 short tons
Agriculture 406,462 ” ”
Chemical Industries other
Than Ammonia Soda
(as above) 2,151,444 ” ”
Total 3,506,439 short tons
In the same year this country produced 2205006 short tons of ammonia soda ash and 488807 short
tons of caustic soda (by lime process).
Assuming an average figure of 1.25tons of limestone required for the manufacture of ton of
ammonia soda ash, or 0.75 ton of lime as burned per ton of soda ash, the total lime consumed in
the manufacture of ammonia soda ash is 0.75*2205006, or roughly 1654000 short tons. If we take
roughly 0.80 tons of lime per ton of caustic made, 488807 short tons of the caustic would require
approximately 391000 short tons of lime.
 Therefore, the total quantities of lime consumed by the ammonia soda and lime caustic
manufacturers for the year 1937 may be estimated at 2045000 short tons, and this quantity is
considerably more than one-half the total quantity of lime used in all other chemical industries, in
building construction, and for agricultural purposes; namely, a total of 3506439 short tons.
For this industry, both lime and carbon dioxide are needed, and lime kiln operation plays an
important role in the efficiency of the ammonia soda process. Hence it is imperative that kiln
operation should attain high efficiency in order to obtain rich CO2 gas. The scientific efforts of the
ammonia soda technologists have done much t develop the lime industry, although little of their
work has been published.
There are several different types of lime kilns. The earliest and crudest form is a small
mound built of bric ks of the stone itself with a pit inside, in which limestone is burned by means
of wood or cola (Fig 19). Natural draft, of course, is employed. A good product is turned out in
this type of kiln. The lime is obtained in very large lumps. However, the process is tedious and
there is no way to recover the gas; also the kiln has a very small capacity and low fuel efficiency.
In the ammonia soda industry, by far the commonest type is the vertical-shaft kiln, also known as
the Belgian kiln. The fuel used is generally coke; although anthracite may be use din certain cases.
In certain localities anthracite is considerably cheaper than coke. The gas obtained when anthracite
is used as fuel has a somewhat lower test and must be thoroughly scrubbed free from sulfur
dioxide and coal-tar products. The best proportion for a vertical-shaft kiln is to have the height a
bout +6-7 times the inside diameter. To economize the heat and consequently to make the richest
possible baste modern tendency is toward increasing the height. The lining of fire brick and
common brick should be about 24 inches thick for small kilns and 36 to 48 inches for large kilns,
and there should be from 3 to 6 inches or space between the steel casing outside (3 /8 inch plate) and
the brickwork, to be filled with sand or an insulating material in order to cut down convection and
radiation losses at casing outside. The modern standard vertical kiln is about 14 feet inside
diameter and 80 to 90 feet high. It has a capacity of about 300 to 350 tons of stone per 24 hours.
The thickness of the brick lining is as much as 48 inches. The kiln is provided with a skip hoist
either inclined or vertical, or an inclined belt conveyor. There is charging hopper on the top with
 FIG 19 Primitive lime kiln.
one or more cones for distributing the stone and coke evenly over the entire internal area of the
kiln. The entire assembly is supported ion cast-iron or steel columns surrounded by a steel hood or
air shield. Inside the air shield is the discharge mechanism on which the entire weigh to the charge
of limestone, coke and lime may be carried. The lime is drawn from the bottom by this automatic
discharge mechanism. A knife-edge scrapes the lime off the revolving table. This was originally
designed by Solve and may modifications have been introduced. They operate very successfully in
large units. Air is introduced through the center ad through four or eight tuyeres laced at
equidistant points on the circumference at the bottom of the kiln, much in the same way as in a
blast furnace. The valve at each tuyere can be set to admit more or less air, as desired.
The charging hopper at the top of the kiln is provided with a heavy 45 cast-iron cone
having an apron extension attached to the bottom of the hopper, consisting of four quadrants, one
deflecting a portion of the charge to center of the kiln, the next putting another portion little farther
out, the third still farther, and the fourth around the internal circumference of the kiln near the fire
brick lining. The apron revolves with the hopper a quarter or a fifth of revolution each time a
charge is dumped in, thus distributing the charge evenly over the whole area in the kiln. Stone and
coke charged into the kiln reach the top either by means of an electrically-driven skip hoist,
inclined belt conveyors, or cable ways. Proportioning of the coke with respect to stone charged
may be done either by automatic -weighing scales, automatic -weighing belt sections, or feed tables,
depending ton the means of conveyance employed.
 The discharge mechanism consists of a heavy, cast-iron spiral cone whose steps and slope are
so proportioned das to take out lime from the center and the sides proportionally. The spiral cone
works the lime from the center to the outside circumference. An annular table made of heavy plate
steel is attached to the base of the cone and revolves with it. Lime gradually drops onto the
annular table ad is scraped by a stationary scraper knife, as the cone is slowly revolved. The lime
scraped from the table drops onto a star feeder, which feeds it into an inclined bucket elevator,
carrying the lime into the bin above. The cone may be driven by a worm gear and chain sprockets.
It may be mounted on twelve steel rollers running on circular cast-steel treads, forming a large
roller bearing at the bottom. The lime is drawn intermittently, generally a few minutes in each
15-minute interval.
With the foregoing arrangement, both the charging of the stone and coke above, and the
discharging f lime below, are very uniform and kiln operation is easily controlled, the top (gas)
temperature can be maintained at as low as 40-60 and the bottom (lime) temperature at 40-80
,depending on the atmospheric temperature, the ate of draw, and the distance of the fire zone
from the bottom of the kiln. The kiln gas test can be maintained at an average of 42 per cent CO2
for 24 hours a day, with lime uniformly and properly burned. Little overburnt lime or unborn core
is found in the draw.
The top and bottom temperatures of the kiln vary considerably according to the height of the
kiln, the arrangement for lime draw, the blast pressure, and the rate of the kiln, the arrangement for
lime draw, the blast pressure, and the rate of operation. With an automatic discharge mechanism
(mechanical discharge), it is quite normal to obtain a top temperature not exceeding 60 in
summer and 30 in winter and a lime discharge temperature of 40-980 at the bottom. This
draw temperature is mainly dependent on the blast pressure and the rate of draw. When the wind
pressure at the bottom is low, or when, for some reason, the blast has been shut off, the fire stays
lows and the temperature lime drawn would be higher. On the other hand, when for some reason
lime has not been drawn out for some time, the fire zone will creep up and the lime drawn will be
colder.
Recently rotary kilns have been adapted to lime burning, but the size of the stone burned is
limited to1 /2 to 11 /2 inches, and the lime produced is fine. For making hydrated lime, this product is
not objectionable. The rotary kiln is especially suited for burning fines from the crushing plant, or
calcium carbonate mud from the causticizing process. The fuel used is oil, powdered coal or
producer gas. A waste heat boiler can be installed at the gas exit, end to recover the heat. Due to
the larger radiating surface and impracticability of extensive heat insulation or heat regeneration,
the fuel efficiency of the rotary kiln much lower than that of t vertical-shaft kiln. Because of this
drawback and certain mechanical difficulties in making the rotary joints perfectly air-tight, the
CO2 test in the gas obtained is generally lower. In the ammonia soda industry, therefore, its use is
confined to the burning of the fines or caustic mud, as mentioned above; the gas then is normally
not used.
The vertical-shaft kiln with mixed feed possesses many advantages. The burning zone, or
decomposition zone, is well insulated and is adapted to heat regeneration both in the exit gases at
the top and in the burnt stone at the bottom. The fuel economy is good and gas containing a
maximum percentage of CO2 can be maintained. The whole secret in securing capacity production
fro the kiln lies in the ability to use forced draft at the bottom of the kiln.
The lime as drawn from the bottom contains portions of undecomposed limestone and
sometime also particles of unburnt coke of glowing ember. The undecomposed limestone is
generally in the form of hard cores in the centers of lumps of lime. The presence of such a core in
a lump is often revealed by its abnormal weight. In may older plants where the kilns are relatively
short for the diameter, the production of these cores is a considerable part to the total lime. Such
cores are rejected by the slacker and are returned to the kiln as a regular fraction of the fresh stone
charged in to the kiln. Further, in short kilns, it is often extremely difficult to maintain the average
gas test above e40 percent CO2. The quantity of cores produce depends on the efficiency of
operation of the kiln, its design and constructions regards heat insulation, height in proportion to
the diameter, etc; and the rate at which it is operated. Even in a most efficient kiln the modern
height, a certain quantity of core will be produced above a certain operating rate.
The size of limestone for soda work varies from 2 to 6 inches; the coke is usually 2 to 3
inches. The ratio of fuel to stone depends largely upon the heat efficiency of the kiln, the quality
of the stone, and the calorific value of the fuel. With a tall kiln, a heavy brick lining, and good
coke, the ratio of coke to stone can be as low as 1 to 16 by weight with out leaving an excessive
amount of unburnt core in the lime and the gas obtained contains 41 to 43 per cent carbon dioxide.
This low full ratio and high-test have been attained and maintained continuously in soda plants in
kilns of sufficient height. At times, the gas test may jump to 48 or 54 per cent carbon dioxide, but
this occurs only momentarily when the blast is suddenly cut off while the kiln is in full operation.
For good gas production an excess of fuel should be avoided. The carbon dioxide produced
from the limestone approximates 100 per cent. Dilution is caused by nitrogen gas introduced in the
air required to burn the fuel charged with the limestone. To be able to use the smallest possible
must be highly efficient in the conservation of heat and must have the necessary height. It is often
a question of judgment in the operation of ammonia soda works whether one should maintain a
good gas test and stand a little more returned stone.
Fuel used to burn the stone should not contain too high ash, because in the mixed-feed kiln,
high ash not only contaminates the lime but also cause clinkering in the kiln. The acidic properties
of SiO 2 Al2 O3 and Fe2 O3 in the ash tend to form fused clinkers with lime, causing clinkering
between pieces of lime and adhesion of the clinker to the refractory walls of the kiln. Coke use
should contain not more than 6 to 10 per cent ash, but in no case should the ash content in the cold
exceed 15 per cent. Hard coke is preferred, because soft coke produces too much fines during
crushing and has a tendency to be crushed in the kiln, causing excessive resistance to the blast.
The kiln should be as far as possible adiabatically isolated from its surroundings. This means
heavy lining from bottom to top. There should be height enough to get a good regenerative effect.
Te exit gas at the top should be cooled by the incoming charge (limestone and fuel) so that its
temperature is low as it emerges. The lime at the bottom should be cooled by the incoming draft
and drawn out cold. In other words, both top and bottom of the kiln should be cold; the burning
zone is in the region from a fourth to a tired the way up from the bottom. The temperature at this
point is around 1050 ; though a temperature of 1100 is not uncommon. The height of the
burning zone depends upon how the kiln is force, i.e.; how much blast is sent through, and how
fast the lime is drawn. Under normal conditions the burring zone should be kept as low as is
consistent with a reasonably cooled lime for drawing.
For large kilns, that is, over 14 feet in diameter with a height of more than 80 feet, the draft
pressure drop is about 6 inches of water. In such kilns the distribution of air tends to be uneven.
The paths of least resistance are through the coarsest particles of stone and coke. These larger
pieces tend to segregate at certain places across the kiln cross-section, depending principally on
the method of charging. If the stone Fed is merely dumped into a single small hole in the middle
of the kiln top deck, the larger pieces roll toward the walls of the kilns where the gas resistance is
in any case low because of the relatively smooth kiln walls. This type of segregation is well
understood and affects many operations, such as feeding coal onto automatic stokers in wide
boilers,

FIG 20 .Lime kiln showing normal distribution of blast.

feeding coke to a gas producer, “stock-pile classification” such as is met with in mineral
operations, etc. Although the distribution of air up through the kiln can be slightly aided by proper
location of tuyeres around the sides and in the center, by far the most important factor is the
distribution of stone and coke uniformly over the entire kiln cross-section. The difficulty with a
single hole charging device is illustrated in Fig. 20.
In mixed feed charging, when the charge is dumped onto a slope or is allowed to fall into the
kiln from a 45 bell hopper, the large pieces of stone will roll forward the farthest, while the smaller
ones together with the larger part of the coke will merely drop down almost vertically off the edge
of the slope.
 FIG 21 Lime kiln showing tendency of different sizes of materials to segregate during Charging.
Consequently, the charge is not uniform inside the kiln; the area vertically below the edge of the
cone has more of the finer pieces of stone with more than its share of coke, while that farther away
from the edge has most of the larger pieces of stone and yet an insufficient quantity of coke to
burn them (Fig 21)* Thus, in the zone directly under the edge of the slope, the stone becomes
overburnt and considerable coke is left over, and the lime drawn is red how. Here little air can pass
through, because of the denser mass of the finer pieces of stone and coke. The result is that the fire
zone stays very low in the center of the kiln. On the other hand, in the zone farther away from the
edge of the cone or next to the refractory walls, air passes through more freely because of the
lower resistance between the charge and the wall surface. However, because of the larger sizes of
stone and of insufficient coke, the stone is only partially burned, a large core being left in each of
the larger lumps. Here the fire is carried high up in the kiln, and cold but superficially burned lime
is drawn at the bottom. This segregation is very serious, or under these circumstances not only is
the stone not properly burned, but the gas test is also low. In such cases, experience has shown that
there are not more than 30 to 32 per cent carbon dioxide in the gas obtained and 3 to 5 per cent
oxygen, while 0.5 to 1 per cent carbon monoxide is also present. Typical analyses are given in
Table 35.
*This may also cause excessive wear on the surface of the kiln lining at the top by the impact of stone against the
bricks.
Table 35. Typical Analyses of Kiln Gases under Abnormal Conditions.
I II III
 Slight Improvement Attained
Per Cent Per Cent Per Cent
By Vol. By Vol. By Vol.
CO2 30.2 CO2 31.6 CO2 34.6
O2 4.6 O2 3.6 O2 2.4
CO 0.8 CO 0.6 CO 0.4
Such a situation is aggravated if the limestone pieces are not of nearly uniform size. This,
however, puts an undue burden on quarrying operation and requires rejection of a very large part
of potential quarry output. When, however, there is more than one kiln in operation, the remedy is
simple. Each kiln may burn a different size of stone so that there will not be such a wide range of
sizes in each individual kiln. That kiln which burns a smaller size, of course, requires a little
higher wind pressure, which may be readily adjusted by means of different valves provided. But
proper mechanical designing of the charging mechanism is usually the best remedy, and a very
large number of ingenious devices have been introduced to accomplish this result. The simplest is
the use of several charging holes in the deck of the kiln, into each of which the same amount of
material is charged per unit of time. Or, an eccentric charging hopper may be used which can be
revolved over a certain angle after each charge, as in a large gas producer. Other devices consist of
mechanically rotating covers with the feed opening being continually moved by means of cams,
etc. One of the more complicated devices consists of feed cones with spiral edges so that the angle
of each element is different and is designed to produce that trajectory which gives uniform
distribution. As was described on page 87, such cones are turned a certain number of degrees for
each batch of feed charged over them by a skip hoist. In this way it is possible to maintain the fire
zone horizon tally at the same level in the kiln.
In addition to uniformity of size, it is important that the stone should be free from clay, dirt,
or fine dust; otherwise the air passage may be choked, and the fine, overburnt pieces may react
with clay or dirt to form clinkers. Suc h clinkers may attain the size of 24 inches and may
completely block the lime passage at the draw opening at the bottom..
The blast used in vertical limekilns varies widely, anything form 4 to 9 inches of water
pressure having been used. It depends on the rate of operation, the size of the stone and coke
employed, and the height of the kiln. With large kilns, the capacity of which is ample and the rate
of burning low, less air pressure is required for operation. Taller kilns burning a wide range of size
of stone, from which only the minus 2 in fraction has been removed, and kilns which are being
operated at high rates of production, sometimes require as much as 10 in. W. G. Smaller sizes of
stone necessitate a hither blast and require less time for burning. If large amounts of the smaller
sizes or fines are mixed with the larger pieces, the stone will not be evenly burned. The smaller
pieces become overburnt, while the larger lumps are incompletely burned. Also, smaller pieces, or
fines, tend to segregate themselves so as to choke the air passage. Consequently, when it is desired
to increase the rate of operation by increasing the blast to maintain a slight positive pressure on the
top of the kiln, air may blow in channels thought the larger lumps of the stone so that free oxygen
is found in the gas at the top of the kiln giving low gas tests as shown in Table 35.
From the foregoing it will be seen that to secure successful operation of a limekiln i.e., to get
a high-test gas and to secure uniformly and reasonably completely burned lime, too much
attention cannot be given to the following matters:
(a) Even distribution of the fuel in the stone
 (b) Absence of segregation of smaller pieces of stone from the larger lumps.
(c) Even distribution of air in the charge.
Any tendency to segregate the fuel from the stone, or the smaller pieces of the stone from the
larger ones, should be corrected at once; otherwise as has been noted above the gas would contain
(1) an excess of oxygen, (2) considerable carbon monoxide and (3) a low percentage of carbon
dioxide. Also, there may be large amounts of unburned limestone from one side of the kiln, and at
the same time overburnt lime from the other. Improvement may always be secured by having the
pieces of limestone and the pieces of fuel (coke) as nearly of uniform size as possible by adjusting
the wind dampers or butterfly valves to control the amount of air admitted at different points at the
bottom of the kiln, and by having air admitted at different points at the bottom of the kiln and by
having the charging mechanism at the top of the kiln so designed as to let the charge in more
evenly i.e. without any tendency toward segregation.
The gas from the 16 to 20-inch down-take located at the top of the kiln is draw off by carbon
dioxide compressors through a cast-iron scrubber, frequently 6 feet in diameter and about 20 feet
high, with a division plate and a mushroom at the bottom above the gas inlet. Above this
mushroom section, the whole space is filled either with Ranching rings or cylindrical tiles
(chemical tiles), 4 inches in inside diameter and 5 inches high, or packed with 6-inch coke,
through which water trickles down from the and gas passes up from the bottom counter currently.
Sometimes large pieces limestone itself are used for packing, and this has a tendency to neutralize
any and tarry matter but cools it to below 30 . on its way to the carbon dioxide compressor
intake .It is important to see that there is no excessive resistance due to the depth of the wash in
the passette; otherwise suction in the gas main at the depth of the wash in the scrubber should
barely touch the water surface; chemical tile serves the purpose better than coke or stone packing.
To adjust the amount of suction between the soda dryers and the lime kilns, if the two branches of
the gas mains are connected, the level of the water in the scrubber section can be raised or lowered
by means of the water in the scrubber section can be raised or lowered by means of an overflow
loop or a half-moon segment at the water overflow outlet, so that the carbon dioxide compressors
will draw the soda dryer gas before they draw the lean gas from the lime kilns Also, it is
frequently arranged to have a separate valve in each branch one on the kiln gas main and one on
the furnace gas main ,so that the differential pressures in both branches may be set as desired an d
mechanically controlled in relation to each other by a sensitive automatic mechanism.
The foregoing type of kiln gas scrubber, which is ordinarily used in the ammonia soda plants,
is generally incapable of removing the finest particles of dust in the gas, and ass a consequence,
the co2 gas compressor cylinders have high maintenance charges caused by frequent renewals of
discharge valves, piston rod, piston, or cylinder wall liner, The best type of scrubbing with
comparatively little loss of pressure through the scrubber is found to be a spiral whirlpool type of
washing, such as the gas cleaner developed by the Blaw-Knox Co., which may be installed before
the intake to the gas compressors so that only pure ,clean gases are handled by the compressors.
The gas from the limekiln should have the composition shown in Table 36. Excessive oxygen
in the gas indicates (1) uneven air distribution or channeling in the air passage as the result of
TABLE 36. Composition of Kiln Gas
Per Cent
CO2 41-43
O2 0.0-0.2
 CO 0.0-0.0
segregation of larger pieces of stone from smaller ones; (2) unequal heights of combustion zones
above the different air inlets, as a result either of channeling as stated above, or of obstruction at
certain air inlets; and (3) creeping of the fire too high up toward the top of the kiln as the result of
the excessive air blast employed and the low rate of lime drawing. Sometimes the presence of
oxygen in the gas is caused by a leakage of air at the top of the kiln when a partial vacuum exists
there. The presence of 1 per cent or more of carbon monoxide may indicate (1) channeling in the
air passage from the causes mentioned above ;or (2) too high a ratio of fuel, causing a reduction of
carbon dioxide by incandescent carbon just above the decomposition zone.
CO2 + C 2CO
In the latter case much of the lime, of course, would be overburnt. In the course of operation, as
mentioned above, it may be found that the lime drawn from one side of the kiln comes out red hot,
while at the other side it is stone cold. On the side of the kiln to which most of the air goes an in
which less fuel is present to burn the stone completely the burning quite cold and only partially
burned. The other side, in which the lime comes out red hot and overburnt, has a large excess of
fuel in proportion to the limestone; and because of denser mass little air is allo9wed through, with
the limestone; and because of denser mass little air is allowed through, with the result that the fire
stays low. The result of faulty air distribution may be illustrated by an incident: a 41 per cent
carbon dioxide test dropped to 32 when the air which was accidentally obstructed went to one side
of the kiln, This decrease occurred as soon as the obstruction took place and remained until the
obstruction was removed.
All the foregoing troubles lead to a low gas test.. The operation of a lime kiln in the ammonia
soda industry is a delicate matter, The ability to diagnose trouble comes only through keen
observation and long experience, The condition of the lime can be readily judged by its
appearance and in weight, Partially burned lime which has a raw core is heavy. Deadburnt or
overburnt lime is darker in color and denser, with the appearance of checks. For the control of the
process, an hourly test of the gas is made by means of a corset apparatus, and the amount of
returned stone (unburnt core) and that of rejected “sand” in the lime drawn are determined and
recorded.
The fused mass of lime adheres to the brick walls and may build “scaffolds” against the
brickwork. The descending stone then may tear off a portion of the firebrick from the face of the
lining. Nothing is more detrimental to the life of the kiln ids heard due to the falling of the stone
arched up over a void space. In the worst case, the “arching” and “scaffolding’ may become so
serious that the stone may actually be held up so that it cannot descend to make room for the
charge in the top of the kiln. The abrasive action of the descending stone on the face of the fire
brick is normally slight; but if the stone level inside the kiln should be allowed to fall low, the
pounding of the stones falling from the should be allowed to fall low, the pounding of the stones
falling from the top through some distance against the brickwork during charging may do
considerable damage to the fire brick lining. A kiln kept only partially filled cannot give good
results anyway, and means are generally provided to gauge the level of the charge at the top of the
kiln. If the kiln is kept full of the charge all the time and a fairly gook quality of limestone and
coke is used which have little tendency to cause serious clinker ing inside the kiln, to operate day
in and day out without repairs for a period of three to five years.
Lumps of clinker formed in the lime by the firebrick falling from the kiln lining give the
following analysis:
Table 37 Analysis of Lime-Firebrick Clinker.
Per Cent
CaO 61.00
MgO 1.96
SiO3 17.84
Al2 O3 16.10
Fe2 O3 2.90
High-grade limestone is essential since impurities in the limestone cause a great deal of
trouble in operation. Magnesium oxide in lime is an inert substance, although carbon dioxide gas
is liberated more readily from magnesium carbonate. The presence of more than 5 per cent of
silica in the limestone causes clinkering or semi-fusion, and under the worst conditions, arching or
“scaffolding” in the lime kiln, as mentioned above, due to its reaction with lime, forming a
calcium silicate:
xCaCO3 + ySiO2 → xCaO.ySiO 2 + xCO2
Pure lime fuses only at temperature of about 2550, which is beyond the temperature range in the
kiln. Alumina in the limestone has a more harmful effect than silica. Not only has it a greater
tendency to lower the fusion point of lime, but it also forms a sticky paste in the lime slaker. Like
silica, it enters into reaction with lime, giving a calcium aluminate (probably tricalcium aluminate)
at the temperature of the burning zone.
3CaCO3 + Al2 O3 → 3CaO Al2 O3 + 3CO2
It has a tendency to reduce the temperature at which lime clinker (of cement character) is formed.
Because of the quick-setting tendency of this aluminate, a rather sticky paste is formed in the
slaker, clogging the passage of the milk of lime in the slaker. This may seriously interfere with the
slaking operation and consequently the distiller operation with the result that normal strength of
milk of lime cannot be maintained and the lime pipes and pumps are clogged.
Table 38 shows the composition of the dark-colored, semi-fused clinker in the lime drawn
from the kiln.
Table 38 Composition of Lime Clinker.
(Air-dried Sample.)
Per Cent
CaO 75.55
CaCO3 3.70
MgO 2.29
SiO2 5.58
Al2 O3 9.24
Fe2 O3 0.64
Water by diff. 3.00 (due to exposure)

The composition of the limestone yielding such clinker can be calculated, with the results
given in table39
Table 39 Composition of Limestone Forming Clinker.
Per Cent
CaCO3 87.30
 MgCO2 3.52 3.02
Silica 0.40 9.73
Ferric oxide 5.81
Alumina
The sticky lime paste formed in the slaker which stops the passage of the milk of lime at the
discharge has the composition shown in Table 40.
Table 40. Composition of Sticky Lime Paste.
Per Cent
CaO 27.80
CaCO3 9.64
MgO 2.35
SiO2 4.47
Fe2 O3 0.24
Al2 O3 2.73
Water by diff. 52.73
Calculated to limestone, the composition of the limestone is given in Table 41.
Table 41. Composition of Limestone Calculated from the Sticky Lime Paste.
Per Cent
CaCO3 82.80
MgCO3 6.87
Silica 6.24
Ferric oxide 0.34 10.39
Alumina 3.81

It is obvious from the foregoing that any stone containing as much as 10 per cent silica
alumina and iron is substantially impossible to burn successfully in an ammonia soda plant.
Occasionally, high-silica (quartzite) and high-alumina (clay) limestone escapes attention through
the quarrying and washing operations and reaches the kiln, whit momentary results in the plant
operations as described above.
The down-take for gas at the kiln top and the gas inlet to the scrubber are frequently clogged
with lime dust, which comes over with the gas. At these points, instead of “elbows”, tees are
used with blind flanges which can be removed for cleaning when necessary. When the “tee ”at the
top of the kiln is choked by dust, the vacuum becomes excessive in the down take pipe below.
When, however, the gas inlet at bottom of the scrubber is choked, the vacuum at the top of
scrubber or inside the scrubber will be high. The point of constriction can be located by spotting
the vacuum readings at different points (at which water manometers should be provided). At times,
it will be found that the vacuum at the carbon dioxide compressor suction main is high, but the
kiln top, is still under considerable pressure. This is the time to look for choking at different points
in the kiln gas down-take, at the inlet to the scrubber, and at the bottom of the scrubber, caused by
the accumulation of dust from the gas .In this case, the vacuum in the scrubber above the bottom
section would be high, but the gas down-take leading to the inlet at the bottom of the scrubber
would be under pressure. On the other hand, when water used in the scrubber is low, or when for
some reason the water spray at the top of the scrubber is cut off, the seal at the bottom of the
scrubber may be lost and air may rush in through the water overflow opening at the scrubber
bottom in large volumes, so that considerable pressure would be felt at the soda furnace end .All
these things interfere with satisfactory operation, and demand close attention and prompt
correction.
Well-controlled soda plants have posted the pressure drop for the gas through every part of
the system when the part is reasonably clean. Shut-downs for cleaning operations are planned
when certain predator-mined increases are reached .The cleaning of the gas from lime kilns is an
expensive and never completed operation .The residual dusts find their way through the system
and, besides introducing high pressure drops in the various parts of the system, may cause high
maintenance costs in compressor operation.
In soda works the lime for distillation is slaked to thick milk, containing 200 to 250
grams per liter CaO. The slaking can be done in a large steel tank provided with a stirrer. Lime is
charged into a steel basket immersed in the water in the tank. The stirring action aids the
disintegration of the lumps and carries the fine particles in suspension. Unburnt core is caught in
the basket. The machine now most commonly used, however, is the rotary slaker, which has a
large capacity and is continuous in operation. It is of the general rotary type, having two support
points. The shell may be made of -inch steel plates, 5 by 6 feet inside diameter and 40 to 50 feet
long, having a lining. The revolving shell is provided with a number of short sections of angle iron
flights to carry the solids forward to the discharge end. It works on the countercurrent principle,
lime entering the discharge end and water the tail end, where it serves also to wash the “sand” and
unburnt core as these are carried out by a ribbon conveyor. Such a rotary slaker has a capacity of
200 to 250 tons of limestone per 24 hours. The unburnt stone is charged back to the kiln and the
“sand” which is rejected makes a good road filling material after it is allowed to age by exposure.
If the limestone is of low grade, containing, containing sandstone, clay, etc., the returned stone
will consist mostly of these impurities having little limestone distributed in the mass. Under these
conditions it is wise to reject these stones instead of charging them back to the kiln with the fresh
stone. Besides sand and other siliceous matter, the “sand” consists of large amounts of fine lime
that have been overbrunt and have passed through the slaker without being slaked. Generally it
will be found that this overburnt lime slowly disintegrates on exposure to the air (two or three
days), when it will pay to send it through the slaker again.
Many slackers, however, do not operate on a perfect counter-current principle. Both lime and
water (i.e., well over 90 per cent of the water) enter at the discharge end and travel through an
inner shell for about or of the length of the slaker. Then milk of lime is drawn from between the
inner and outer shells at the same end. The slowly digesting and the undigested matter is propelled
to the back, so that it has a very long time to give up all its suspended particles to the milk. At the
back end it is lifted up into a conical section and finally into a cylindrical screen section, where the
remaining small amount of water (i.e., about 10 per cent of the total) washes it free of the adhering
CaO. In most of these plants the quantity of cores and rejects from the slaker is very small, and
these are not returned to the kiln. They are occasionally ground up and added to the milk of lime,
when and if they consist mostly of overburnt particles. Grinding decidedly makes the overburnt
lime more available.
Steam is saved at the ammonia distilling operation if the milk of lime is used hot .In well
regulated plants, lime reaches the prelimer at a temperature above 80°C. At lower temperatures
not only is an excess of steam used, but also the prelimer reaction is slower, and there is a greater
tendency toward the formation of scales. The slaking operation is exothermic enough so that milk
of lime is automatically produced at almost 100°C. However, unless the slaker is well designed
and all pipe lines and storage tanks for milk are lime are thoroughly insulated, this temperature
falls rapidly because of air-cooling. The water used for slaking should be at a high temperature, to
facilitate the operation and increase the capacity of the slaking apparatus. Cold water and cold
lime often refuse to react when a slaker is just being started anew. There are several sources of hot
water for the slaking operation, the commonest being the condenser water from the distillation
operation. Theoretically, the water for the slaking operation would be heated to over 100°C. from
the heat of hydration, and the steam formed is removed via the slaker stack. The hot milk of lime
can be handled by means of plunger pumps or centrifugal pumps. The duty of the pumps is quite
severe because of the sus pended abrasive particles, or “sand” in the milk of lime. Large,
well-designed centrifugal pumps with open impellers and renewable sleeves, designed for
low-cost replacement, work successfully. To prevent the pumps from clogging, the milk of lime is
screen immediately after slaking. The over-sized particles retained on the screen are washed with
a small amount of water and then added to the rejects from the slaker before grinding for recovery,
or before sending to waste.
Milk of lime storage reservoirs generally consist of steel or concrete tanks, well insulated and
provided with mechanical agitators.
In the practical construction of the limekiln, the top cover with the charging hopper and cone
is never airtight. Consequently it is necessary to carry a positive pressure of about inch of water at
the top, so that leakage will be from the kiln to the air rather than from the air to the kiln with
resulting dilution of the gas. Also, an atmospheric vent pipe is provided so that, in case of any
excess of lean gas, the kiln gas can be let out to the atmosphere in order that the dryers may have
sufficient suction to pull out the gas from them. In the case of rotating kiln tops in large kilns,
however, the charging mechanism is sealed with water or mineral oil against leakage at a low
vacuum. During the periodic charging operation, the gas may be temporarily turned to the
atmosphere and kept entirely off the gas main.
Table 42 gives the approximate capacities of mixed-feed, vertical shaft limekilns. It is
understood that the capacity of a given set of dimensions depends upon many factors. It depends
upon the quality of the stone charged, the use of a forced draft, the use of a mechanical discharge
mechanism, the grading of the sizes of the stone and coke in the charge, etc.
TABLE42. Capacity of Mixed Feed, Vertical-shaft Lime Kilns.
(Working on good grade of limestone with forced draft and mechanical discharge.)
Inside Dimensions of Kiln Thickness of Lining
Diameter Height Bottom Top Average Tons of Limestone
Feet Inches Burning per 24 Hours
10 60-70 30 18 24 120-150
12 80-90 42 30 36 200-250
14 90-100 48 36 42 300-350
16 120 54 42 48 400-500
Large kilns are more economical in the utilization of heat, because heat losses by radiation
and convection are proportional to the surface exposed (the square of the linear dimensions),
whereas the capacity is more nearly proportional to the volume (the cube of the dimension), sine
the height of the kiln is more or less dependent on the diameter. On the other hand, in the case of
very large kilns, there is greater tendency for the segregation of the stone and coke and for uneven
distribution of the air current throughout the whole cross-section of the kiln. To get the fullest
advantages of very large kilns, therefore, requires the introduction of a refined mechanism for an
adequate feed distribution, automatic lime discharge, and blast distribution and regulation.

BURNING OF ANOTHER FORM OF LIMESTONE

Other forms of limestone have been known. In fact, lime in China was produced for centuries
by burning oyster shells; and this source of lime was know much earlier than the lime rock there.
However, the application to the ammonia soda industry of oyster shells, as a source of lime is a
recent one. On the Gulf Coast of North Americ a, a special form of oyster shells found in
abundance has been used in the ammonia soda industry since the new soda industry was
established there. At first, attempts were made to use the same shaft kiln, but these met with
failure. Rotary kilns are now used for the burning of such oyster shells there. However, the
operation is attended by same characteristics as in the burning of the fines from a limestone quarry
in any rotary type kiln. If the larger shells are screened out from the smaller ones, the larger shells
may be burned satisfactorily in the vertical shaft kiln, provided some limestone is mixed with
them for burning. Therefore it is often seen that when these oyster shells are used, both the rotary
kiln and the vertical shaft kiln are employed side by side.
Oyster shells are an important source of limestone along the American Gulf Coast, where
sedimentary deposits of limestone are deeply buried under later deposits and alluvial strata. Oyster
shells in the South have been used since Colonial times as a source of agricultural lime along
much of the American seacoast. The oyster breeds along shores where the water is not entirely
salty. Occasionally nature, through flood or excessive rainfall, makes the water too fresh for the
oyster to survive. Reefs of shells are found along substantially all seacoasts, particularly near the
mouths of streams or bays, into which fresh-water streams empty. Through metamorphic pressure
of superimposed strata, such reefs of oyster shells have become compressed and eventually turned
into limestone. It is, therefore, to be expected that chemical analysis of the oyster shells will be
similar to that of limestone. A sample of thoroughly washed oyster shells is hardly distinguishable
from a high-grade low-magnesium limestone by a chemical analysis of the principal inorganic
constituents. The following table gives an analysis of a sample representative of each on the dry
basis:
TABLE 43: Comparison of Chemical Analyses of Oyster Shells and Limestone.
Oyster Shells Limestone
Calcium carbonate, CaCO3 (Gulf Coast) (Amherstburg, Canada)
Magnesium carbonate, MgCO2 96.05% 96.70%
Calcium sulfate, CaSO4 1.22 2.11
Alumina oxide and iron oxide, R2 O3 0.36 0.05
Silica, SiO 2 0.33 0.57
1.28 0.52

The above limestone is, of course, “high-grade,” that is, it is from a deposit, which contains very
little earthy matter and had been thoroughly washed. Relatively few deposits are so low in
magnesium. Oyster shells as dredged generally contain appreciable quantities of organic matter.
Oyster shells present problems in the burning of lime and CO2 gas for calcium carbonate. The
shaft kiln, which is the standard equipment in the ammonia soda plant for burning limestone, is
not burning shells. The individual pieces of oyster shells are so small that there is an excessive
pressure drops in the arc -blast. The organic. The organic material, and probably also silica in the
shells, combine to cause the formation of large agree - gate of clinkers. Furthermore, since the
shells are thin and fragile, they are easily crushed, and so they are not suitable to the operation of
shaft kilns of modern height. From the standpoint of the ammonia soda industry, it would be
desirable to develop a shaft kiln suitable to oyster-shell burning. Agricultural oyster-shell lime has
been produced in short shaft kilns, but since the development of the rotary type of cement kiln,
shells have been burned on a large-scale commercial basis in that type of equipment.
Oyster shells require no special preparation before burning in a rotary kiln. Depending only
on the purity of the lime desired, they are washed in conventional types of washing
equipment .The process of collecting the shells when dredged by a dredger is in itself a washing
operation. When very high-grade white finishing lime is desired, the shells are given a thorough
supplementary washing and generally also a crushing operation, to dislodge dirt particles
deposited in the crevices of the shells .In the ammonia soda industry. The process can, the stand a
certain amount of impurity in the lime from the shells as dredged .The shells are therefore not
crushed and, except for the washing received on the dredge, are not washed .The shells, as fed to
the kiln, contain up to 8 per cent moisture, depending upon the length of time they have in storage
at the plant.
The kiln is a long rotating cylinder, set on a slope of 1 /2 inch per foot. The main equipment is
similar to that used in the burning of caustic mud in caustic soda manufacture (see Chapter XIX),
but more particularly in .the burning of raw mix in dry-process cement manufacture .The kiln
must be considerably longer than the usual apparatus of this type. Because higher fuel efficiency is
justified in order to obtain gases rich in CO2 .For the same reason, it is important to cool the
discharged lime for preheating the air required for combustion. This entails a separate lime cooling
apparatus, similar to the rotary clinker coolers used in the cement manufacture. However, as the
gas is also wanted, the kiln must be closed at both ends.
The shells are fed into the high end of the kiln by means of a constant-weight feeder and the
gases emerging from this end of the kiln are handled just like vertical kiln gases, except that a dust
separation in the hot and dry state precedes the usual scrubbing .The shells, in their slow progress
down the kiln, are first freed from surface moisture, then gradually are preheated to the claiming
temperature, and finally in direct contact with a pulverized coal, oil or gas flame, are calcined and
discharged to the cooler. As is blown and drawn through the shell lime in the cooler. A part of air
preheated in the cooler is drawn off and used as “primary air” for injecting the fuel through the
burner into the kiln .The rest of the air goes directly from the cooler the kiln hood, which is kept
under accurately balanced pressure by a delicate control which is kept under accurately balanced
pressure by a delicate control mechanism .The primary control of this operation is affected by the
speed of rotation of the kiln .The oyster-shell feeder is electrically or mechanically interlocked
with the kiln drive ,and speed at which fuel is injected is likewise accurately proportioned to the
kiln speed .Increase or decrease in fuel is based on a combination of factors which include:
(a) Analysis of the lime discharged,
(b) Temperature of the lime to the cooler.
(c) Analysis of the gases obtained.
If the discharged lime drops in CaO content (or increases in CO2 content), it is generally an
indication of insufficient fuel .It may be an indication of faulty fuel-air ratio, which can be
determined from an analysis of the gases .If CO has increased, there is insufficient air; whereas if
too much O2 is present, this is indicative of excess air .Any out-of-control condition results in a
lowered CO2 concentration in the gases. Hence this decrease in concentration is generally the first
indication of maladjustment of some operating factor.
The cooled shell lime discharged from the cooler is conveyed and elevated to the bins located
above the slaking apparatus. The slaking of oyster-shell lime is essentially the same as that of
stone lime .The physical properties of the milk of lime are similar to those any low-magnesium
limestone milk.
Partly because of the large amount of kiln surface from which heat is radiated relative to lime
burning rate, the unit fuel consumption of the rotary kiln is appreciably higher than that of the
shaft kiln. Consequently, the strength of the CO2 gas is lower, which requires more compressing
power for the carbonating towers. These effects can be held to a minimum only through the very
best design of equipment, which will permit extremely accurate control and very quick response to
control. Whereas in the very large and tall shaft kilns used in the ammonia soda industry, good
operating conditions are almost impossible to upset by a short-time maladjustment, the rotary kiln
is extremely sensitive to changes in blast pressure, suction pressure, rate of shell feed, fuel-to-shell
ratio, fuel –to-air ratio. Etc. The following table gives some estimated operating data for kilns
burning oyster shells with various kinds of fuel:

Because of the high calcium content of the shells, the consumption per ton of ammonia soda
is lower than that for plants using ordinary limestone. A total consumption of 1.15 tons of shell
per ton of soda ash represents good actual practice, and compares with 1.45 in a plant whose stone
contains 8 to 10 per cent magnesium carbonate, and with 1.25 in a plant which uses a high-grade,
low –magnesium stone.
Since oyster shells are a peculiarly thin from of limestone, heat is able to penetrate all the
way through with considerable rapidity; therefore the kiln required to burn the oyster shells
adequately (that is with very little CO2 left in the lime) need not be very long .The length of the
kiln should be sufficient to exchange properly the sensible heat of both the gases of combustion
and decomposition with the oyster shell feed on its way to the calcining zone. Only by good heat
regeneration can the fuel consumption be kept down to anywhere near vertical kiln practice.
Excess fuel requires oxygen to burn it, and this oxygen brings nitrogen from the air into the kiln,
thus diluting the gases .for the same reason, as much as possible of kiln cylinder itself should have
a good layer of insulating material (such as Sil -O- Cel) between the refractory lining and the steel
shell.
The lime kilns in ammonia soda plants must always operate to suit the demand of both lime
and CO2 gas in the distillation and carbonation processes. With vertical kilns, in which there is a
very large volume of preheated stone and burned lime, changes in the demand of either gas or lime
that are of ordinary duration generally only raise or lower the burning zone. Normally there is an
excess of CO2 gas in an alkali plant .In some of the kilns during a part of time of operation, the gas
is turned to the atmosphere, thus avoiding any dilution, which might take place during periods of
increased gas demand.
With the rotary kilns this type of control is impossible. There is only a small amount of
material in the kiln .A gas holder is therefore placed between the kiln and the CO2 compressor
suction .The kiln is then operated as required by the reservoir levels in the dry lime and milk and
milk of lime systems. This requires extremely precise control of kiln cylinder speeds. In this
respect oyster shell burning is very different from other kiln operation .Its rate of operation is
controlled for instantaneous needs rather than for predetermined optimum rates.

The kiln speed therefore establishes the primary impulse for all other operating controls .Gas
pressures throughout the system are under interlocking control and excess CO2 vents out
automatically. Momentary deficiencies of CO2 are handled by the gasholder. Table 45 above
shows a comparison of limes obtained from the oyster shells and from a good grade limestone.
From the above, it may be noted that the CO2 gas obtained from the burning of the oyster
shells in a rotary kiln is on the whole, much weaker than that obtainable from the burning of
limestone in a vertical shaft kiln of modern height. This is a handicap for an ammonia soda plant
depending on oyster shells for its source of lime and CO2 supply, although shell lime obtained is
of the first quality; because weak gas for the ammonia soda manufacture has several disadvantages,
as will be pointed out later on.

Kiln Calculations; Vertical Shaft, Mixed Feed Kilns

As remarked above, the kiln lining must be thick and the kiln must have a good regenerative
effect both at the top and the bottom. Following are cause for the loss of heat:
(a) Unburned fuel in the lime drawn out,
(b) Radiation and convection losses from the body of the kiln outside,
(c) Heat left in the lime drawn out,
(d) Heat left in the gases emerging from the top of the kiln
Good kiln operation requires cold conditions both at the top and the bottom. Items (c) and (d) can
be determined from the specific heats and the temperature ranges. Item (a) can be only roughly
estimated but generally the unburned fuel is so small that it can be assumed that all the fuel
charged is burned. Item (b) can only be established by an assumption.
The following is a typical calculation of the theoretical percentage of coke required and the
percentage of carbon dioxide in the gas obtained. Given (limestone and coke on moisture free
basis):

Under normal conditions no appreciable amount of carbon monoxide is found in the kiln exit
gases, so we can assume that all carbon burns to carbon dioxide, and we shall assume also that all
the carbonates in the limestone are completely decomposed.
(a) Unburned fuel in the lime: The unburned fuel present in the lime is an indefinite
quantity but generally it should be very small. If the kiln is not being drawn too fast, with a proper
fuel ratio, the lime should come out comparatively cold, and no unburned fuel, such as pieces of
unburned or glowing coke, should come out with the lime.
(b) Radiation and convection losses from the kiln surface: With a heavy lining, the outside
of the kiln is quite cold except for a section at the reaction zone, and the loss of heat is also small.
Estimates have been made which would put the losses at 10 per cent of the heat input of the kiln.
While the temperature inside and outside can be measured, the most uncertain factor is the thermal
conductivity of the combined path through the fire-brick and common –brick wall, the sand space
(or insulating material) and the casing, on account of their unknown physical safe to estimate that
10 per cent of the total heat generated should be considered as lost.
(c) Heat left in the lime drawn out: The red-hot lime should be cooled by the incoming air
before it is drawn out .It will be shown that there is more than enough air to cool the lime .Let us
consider that the lime comes from the burning zone at 1050. . and that it is to be cooled down to
50. .before it is drawn. Let us further consider that we use 150 lbs. of coke (85.25 per cent
fixed carbon) per 2000 lbs. of stone (89.90 per cent CaCO2 +5.94 per cent MgCO3 ). 100kg.of
limestone will yield
(100-89.9-5.94)+89.9 56 +5.9 4 40.3+7.5 .1475=58.4Kg
100 84.3
of “lime” including the ash from the coke .The specific heat is 0.19.Therefore heat to drawn from
the lime per 100kg.of limestone is
58.4 (1050-50) 0.19 = 11,100 kg. Cal
This 100kg. of limestone takes 7.5kg.of coke or 7.5X0.8525 fixed carbon which requires for
combustion
7.5 .8525 32 1=73.7 kg. of air.
12 0.231
The specific heat of air is 0.2389.If air is preheated from 15 . to 800 . it requires
73.7X0.2389X (800-15) . =13,750kg.Cal.So there is generally not enough heat in the lime to
preheat all the air if the heat interchange is good.
(d) Heat in the exit gases: On the other hand, there is not enough limestone to cool the exit
gases. Using the same assumption as above, we have of gases per 100kg. Of stone. The
specific heat of this mixed gas (CO2 +N2) can be taken as 0.23.Yherefore to cool the
gases from 1050. To 15.C. it would require to take away heat =1035 0.23
122.7=29,300kg.Cal.But there are 100kg.of limestone with 7.5kg. Of coke to be taken in.
Since the specific heat of coke =0.20 and the specific heat of limestone =0.22,if it could
be heated from 15.to 1050.C., it could only extract heat=1035 0.22 100+1035 0.2
7.5=24,350kg.Cal.So,granting good heat interchange and the maximum possible heat
extraction, not all of the available heat in the kiln exit gases can be utilized by the
incoming charge (stone and coke)-an unavoidable large item of heat loss in the kiln
operation.

FIG 22 Heat loss in kiln gases and lime drawn.

From the above it can be seen that the exit gases at the top must necessarily be at a much
higher temperature than the lime drawn at the bottom, if there is no loss of heat due to radiation
and convection at the top portion of the kiln. This is the reason for having a tall kiln with a storage
space for preheating the charge and explains the modern tendency to increase the height of the
kiln .It is also the basis for the theory (confirmed by actual practice in the kiln operation) that the
reaction zone must be kept as low in the kiln bottom as is consistent with a fairy cold lime draw.
There is so much surplus heat (as sensible heat) in the exit gases that probably not more than 75
per cent of it is recovered even in a tall kiln (height six diameters) with the reaction zone kept
about a fourth of the height from the bottom of the kiln.
 Fig.22 shows heat losses in the kiln gases and lime drawn respectively at any given
temperature. These losses are practical values, which will allow certain range of variation in the
conditions of operation. Based on the analyses of stone and coke as given before, the ratio of fuel
to stone may be determined as follows:
Let X =kg. Of coke required for every 100kg.of stone. Let the temperature of lime drawn be
50 .and that of the kiln exit gases 80 .The rest of the data is given in the foregoing example.
Let it further be assumed that 10 per cent of the heat be lost by radiation and convection in the kiln
body and top. Starting and ending at 15 ., we have

This is the theoretical percentage of carbon dioxide, assuming no excess of air and no carbon
burnt to carbon monoxide and that the kiln exits gas temperature is 80 .and the lime draw
temperature 50 . The above figures afford valuable guide in ascertaining what the kiln should do.
The limiting case is when there is no loss of heat by radiation or convection from the kiln, i. e,
when the brick lining of the kiln is so thick that the outside casing is cold and the heat losses are
negligible. Under such condition with the same grade of stone and coke and with the same set of
operating conditions in the kiln as above, Equation (1) becomes
With the ordinary grade of coke used in the ammonia soda industry, the term 6.65X is small
compared with 7094yX(being only about 01.0-0.12%). So we can write the equation (with an
error within 0.12%)
7094yX = 432.7C + 347.3M + 665
or 100yX = 6.10C + 4.90M + 9.38 (3)
Therefore for a given grade of limestone and a given set of kiln operating conditions, X varies
inversely as y. :In other words, given a grade of limestone, the amount of coke used varies
inversely as its fixed carbon content. The percentage of CO2 in the gas then (assuming 10% heat
loss)
Therefore the percentage of carbon dioxide in the gas obtained, assuming all the carbonates in the
stone are decomposed and no carbon monoxide is present in the gas, is a function only of the
percentages of CaCO3 and MgCO3 in the stone.
From Equation (3) we know % fixed carbon in coke % coke on the stone = 6.10C + 4.90M
+ 9.38. This expression is useful in calculating the amount of coke that should be required for the
kiln operation, knowing the analysis of the coke and the composition of the stone.

Expression (4) enables us to determine the percentage of carbon dioxide in the gas that
should be obtained from the lime kilns, knowing the composition of the limestone being burned,
i.e., % CaCO3 and % MgCO3 .
When the kiln lining is so extraordinarily thick that the outer casing is cold throughout, even
at the reaction zone, the heat losses by radiation and convection are negligible. Then equation (3)
becomes
100yX = 5.48C + 4.40M + 8.42 5
and Equation (4) becomes
% CO2 in the gas obtained = 457C+1.551M+0.701 100 6
3.185C+2.938M+3,357

From Equation (6) the theoretical maximum possible percentage of CO2 in the kiln gases
assuming no losses by radiation and convection from the kiln (with 80.C.exit gas temperature and
50.c. lime temperature) can be calculated whenever the composition of limestone (CaCO3 and
MgCO3 ) being burned is known. Also, it should be noticed that, other conditions being equal, a
high magnesium stone gives a better gas. From Equation (6), a 100 per cent CaCO3 stone gives a
theoretical maximum gas test of 45.6 per cent CO2 by volume; whereas a 100 per cent MgCO3
stone gives a theoretical maximum gas test of 52.4 per cent. Stones having various percentages of
CaCO3 and MgCO3 have theoretical maximum gas tests ranging between these limits.
Close perusal of the above mathematical relationships should make it apparent that the last 2
or 3 per in the percentage of carbon dioxide above 40 per cent calls for a tremendous effort in the
lime kiln operation. Further, since limekilns generally furnish a larger portion of carbon dioxide
for the column operation than the dryers, a drop of 2 or 3 per cent of carbon dioxide I the kiln gas
materially lower the percentage of carbon dioxide in the mixed gas obtained. Consequently, if only
a weak gas is obtainable from the lime kilns, the consumption of salt (from the poor
decomposition in the columns), of ammonia (from the ammonia losses in the cycle per ton of
soda made), of the limestone used and of coke to burn the limestone, of the power required in the
CO2 compressors to force the gas through the columns-in fact of most of the important items in
the manufacture of soda ash –will be high and the efficiency of the whole operation will drop off.
Hence limekiln operation occupies no minor position in the ammonia soda industry.
 Chapter VII

Ammoniation Of Saturated Brine

The saturated brine for soda production is sent to the system generally treated according to
one of the methods described in Chapter V. In the soda works the buildings are usually tall so that
the brine can be pumped to a head tank, on top of one of the buildings, were it flows down by
gravity to various units in the system. The fresh brine from this storage tank is employed first to
scrub small quantities of ammonia from the filter gas in the filter washer, from the absorber outlet
gas in the absorber weak washer, and from the tower outlet gases in the tower washers (Fig.23).
From this apparatus, the brine having picked up a small amount of ammonia and absorbed as
much carbon dioxide gas as can be retained by the alkalinity of the slightly ammoniacal solution,
flows to the absorber washer and thence to the absorber itself (Fig.24).

FIG 23 Tower washer FIG 24 Absorber

Fom the outlet of apparatus to the inlet of another a long “U” loop is usually provided in the
one end is working under pressure while the other is under partial vacuum, or if two units are
under different vacua. Theoretically speaking, each leg of these “U” loops need be just so long
that the static head of the column of brine will balance the differences of pressure above the
brine in the two legs of the loop. Where the brine is flowing down with some velocity, this
allowance is generally not sufficient and a longer must be allowed to take care of the velocity
head.
In the process, the brine is allowed to absorb the desired amount of ammonia prior to its
treatment with carbon dioxide in the carbonating towers, This order of treating the brine first with
ammonia and then with carbon dioxide is a natural one, for it would be impossible to reverse the
order, i.e. to treat the brine first with carbon dioxide and then with ammonia successfully, as
carbon dioxide is sparingly soluble in the neutral brine.
Formerly there was a process or distilling the filter liquor and condensing the ammonia
vapors to form an ammonia solution of suitable strength in which solid salt was then dissolved to
full saturation. Although this insured good saturation of sodium chloride in the resulting
ammoniacal brine, the actual operation of the process was attended by considerable mechanical
difficulties because of impurities in the salt and loss of ammonia. This procedure was therefore
abandoned.
The ammoniation of the brine serves a two –fold purpose. First, it is a means of introducing
the required amount of ammonia into the brine, and secondly, it is a means of eliminating the
impurities of calcium and magnesium necessary for the production of a high quality soda ash. For,
if calcium and magnesium are not eliminated at this stage, they will be precipitated with sodium
bicarbonate in the carbonating towers later. The gases that carry ammonia also carry carbon
dioxide, whether they come from the filters, from the absorber vacuum system, or from the
carbonating towers. Consequently, ammonia dissolved in the brine exists partly also in the form of
(NH4 ) 2 CO3. Hence calcium and magnesium are precipitated in the brine as CaCO3 , MgCO3,
MgCO3 .(NH4 )2 CO3 .4H2 O, MgCO3 .NaCl.Na2 CO3 , and to a small extent basic magnesium
carbonate or Mg (OH) 2 .

However, in the tower washers and weak washers these precipitates of calcium and magnesium
carbonates formed in the cold are in a milky suspension so that fortunately little is settled. Most of
the precipitate flows with the brine stream to the absorber washer, thence to the absorber and the
settling vats, where these impurities finally coagulate at a higher temperature.
Besides eliminating calcium and magnesium from the brine, the introduction of ammonia
into brine, or the ammoniation with ammonia gas from the distiller has the following effects:
(a) The decrease in solubility of sodium chloride in the resulting brine.
(b) The increase in the volume of the resulting ammoniated brine.
(c) The unavoidable dilution of the saturated brine by steam carried over with the
ammonia gas.
(d) The generation of a large quantity of heat of solution of ammonia in the resulting
ammoniated brine, heat of partial neutralization of ammonia by CO2 , and heat of
condensation from the steam carried over.
These individual factors will be treated more fully below
(a) The effect of the decrease in solubility of NaCl. It is a general rule that substances,
which dose not react with the solute, or form any solute when the substance is
introduced into the same solution. The decrease in the solubility of salt upon the
introduction of ammonia is one of the factors that reduced the quantity of sodium
bicarbonate which could be formed in the towers due to the decrease in the
 concentration of sodium chloride in the resulting ammoniated brine.
TABLE 46. Specific Gravity and Concentration of Pure Sodium Chloride
Solution at 15 .
(Based on Garlic’s figures.)
NaCl in Brine NaC1 Chlorine
Per Cent Sp. Gr. Grams per liter Titer
1 1.0073 10.073 3.447
2 1.0145 20.290 6.492
3 1.0217 30.652 10.488
4 1.0290 41.160 14.084
5 1.0362 51.812 17.729
6 1.0437 62.620 21.427
7 1.0511 73.576 25.174
8 1.0585 84.681 28.976
9 1.0659 95.934 32.826
10 1.0734 107.34 36.730
11 1.0810 118.91 40.688
12 1.0886 130.63 44.698
13 1.0962 142.51 48.763
14 1.1038 154.54 52.880
15 1.1115 166.72 57.048
16 1.1194 179.10 61.284
17 1.1273 191.64 65.575
18 1.1352 204.34 69.920
19 1.1432 217.20 74.320
20 1.1551 230.21 78.773
21 1.1593 243.45 83.302
22 1.1676 256.86 87.892
23 1.1758 270.43 92.536
24 1.1840 284.17 97.236
25 1.1923 298.07 101.99
26 1.2010 312.25 106.84
26.395 1.2043 317.88 108.77
*One titer is one-twentieth of one normal.
Table 46 gives the specific gravity and concentration of sodium chloride in C. P. brine, based
on Gerlach’s figures. With the introduction of ammonia into the brine, the solubility of sodium
chloride in the resulting ammoniacal solution is decreased according to the concentration of
ammonia in the brine.
 FIG 25 Curve showing solubility of sodium chloride in aqueous ammonia solutions at 20

Data on the solubility of sodium chloride in aqueous ammonia solutions seem to be still
somewhat in doubt. The original published data were given by Schreib in “vol. %,” i.e., in grams
NH3 per 100 cc. of the solution. Schreib admitted that the results were only approximate. Hempel
and Tedesco conducted the same investigation in 1911 at 30 . These results we desire to check.
We have made the study at 20 . approaching the equilibria from both directions, i.e., from the
solution of sodium chloride in aqueous ammonia solutions and also (for four readings) from the
ammoniation of the saturated brine, both at 20 . The results are given in Tables 47 and 48. For
comparison we have also plotted in Schreib’s and Hempel and Tedesco’s figures (Fig. 25)
Regarding the specific gravity of the resulting ammoniated brine, figures vary considerably.
The is due principally to the varying carbon dioxide contents in the ammonia solutions: the higher
the concentration of ammonium carbonate in the ammoniated brine, the higher the specific gravity
for the same NH3 and Cl-titers. High temperatures, of course, lower the specific gravity readings.
In our investigation we have taken C. P. ammonia solutions; the specific gravity readings of the
resulting solutions will therefore be lower than those of the strong liquors (containing carbon
dioxide) in actual plant operation at the same temperature.
A word may be said about the ammonia component in the ammoniated brine. The solubility
of ammonia in brine solutions is far greater than is required in operation. Given sufficient cooling,
the ammonia titer could easily be brought beyond 110, but it would be quite out of proportion to
the sodium chloride concentration, which tends to decrease, as the ammonia concentration gets
higher. When the ammonia concentration in the brine is made greater, the chlorine titer will
necessarily be lowered because of the decrease in solubility of sodium chloride and of the increase
in volume in the resulting ammoniated brine. This greatly decreases the concentration of sodium
chloride in the liquor. Hence for optimum operating conditions we ammoniate the brine only up to
or below 100 titer and do not attempt to push it to the limit of its solubility, because we wish at the
same time to maintain the concentration of sodium chloride (i.e., chlorine titer) in the resulting
ammonia brine as high as possible. (See Tables 47 and 48.)
*Z. angew. Chem., 1888,
+Ibid., 1911.
 TABLE 47. Solubility of Sodium Chloride in Aqueous Ammonia.
(From Saturation of Sodium Chloride in Ammonia Solutions at 20 )
Ammonia Solutions Same Solutions Saturated with C.P. Sodium Chloride
Sample Sp. Gr. NH3 NH3 NH3 Sp. Gr. NH3 Cl NH3 NaC1 NH3 NaCl
No. at 20 Titer g./l. Per Cent at 20 C Titer Titer g./l. g./l. Per Cent Per Cent
1 0.998 4.92 4.18 0.42 1.196 4.52 107.41 3.84 313.93 0.32 26.25
2 0.995 9.86 8.38 0.84 1.192 9.05 106.61 7.69 311.62 0.64 26.14
3 0.993 14.92 12.68 1.28 1.189 13.16 105.49 11.18 308.37 0.94 25.85
4 0.991 19.79 16.82 1.70 1.185 17.48 104.81 14.86 306.30 1.25 25.67
5 0.989 24.65 20.95 2.11 1.182 21.79 103.81 18.52 303.45 1.57 25.67
6 0.988 29.62 25.18 2.55 1.178 27.32 102.6 23.22 299.9 1.97 25.46
7 0.986 34.51 29.33 2.97 1.175 30.43 101.8 25.87 297.53 2.2 25.32
8 0.985 37.06 31.5 3.19 1.173 34.95 101.4 29.71 296.4 2.53 25.27
9 0.982 44.54 37.86 3.84 1.168 41.06 100.2 34.9 292.9 2.92 25.08
10 0.981 49.3 41.9 4.26 1.166 43.79 99.72 37.22 291.48 3.19 25.00
11 0.979 54.23 46.1 4.7 1.164 48.83 98.7 41.5 288.5 3.57 24.79
12 0.977 59.12 50.25 5.12 1.16 52.63 97.8 44.74 285.86 3.86 24.64
13 0.975 64.4 54.74 5.6 1.157 56.33 97.59 47.88 284.96 4.13 24.63
14 0.974 68.85 58.52 6 1.152 61.27 96.19 52.08 281.15 4.52 24.41
15 0.972 73.86 62.78 6.45 1.15 64.98 95.99 55.23 280.56 4.8 24.4
16 0.971 77.98 66.28 6.82 1.148 69.49 94.80 59.07 277.1 5.15 24.14
17 0.968 84.42 71.76 7.4 1.146 75.25 94.19 63.96 275.3 5.58 24.02
18 0.967 89.09 75.72 7.83 1.142 78.95 93.59 67.11 273.55 5.88 23.95
19 0.965 93.76 79.7 8.26 1.139 83.06 92.58 70.6 270.6 6.2 23.76
20 0.963 98.39 83.63 8.68 1.136 87.17 91.98 74.09 268.87 6.52 23.67
21 0.962 102.26 86.92 9.04 1.134 91.7 91.38 77.94 267.1 6.82 23.56
22 0.961 103.36 87.86 9.14 1.133 93.29 90.94 79.3 265.88 7 23.46
23 0.959 109.38 92.97 9.7 1.132 96.63 90.6 82.14 264.82 7.26 23.4
24 0.958 112.79 95.87 10.01 1.129 100.76 89.93 85.64 262.86 7.59 23.28

TABLE 48. Solubility of Sodium Chloride in Aqueous Ammonia.

(From Ammoniation of Saturated Brine.)
Sp. Gr. At NH Cl NH3 NaCl
20 . Titer Titer Grams per liter
1.186 18.09 103.82 15.38 303.46
1.175 35.36 100.6 30.06 294.05
1.147 73.6 93.4 62.56 273.01
1.124 109.38 88.13 92.97 257.6

The presence of sodium chloride and of ammonium carbonate tends to raise the specific
gravity of the ammonia solution, while ammonia tends to lower it.
If the process is started by introducing into the tower washer or filter washer cold saturated
brine, theoretically some sodium chloride should separate in the absorber during ammoniation. Yet
in practice the ammoniated brine from the absorber is generally undersaturated in sodium chloride
at the temperature in question due to the following reasons:
(1) That the brine has picked up some moisture during the scrubbing of the gasses saturated with
water vapor at a higher temperature; (2) that in absorbing ammonia gases from the distiller
considerable quantities of water are unavoidably introduced into the brine from the steam carried
over; and (3) that, to start with, the brine, especially from brine wells, is seldom absolutely
saturated. Consequently, the resulting ammoniated brine falls below full saturation with respect to
sodium chloride, despite the decrease in solubility of sodium chloride by ammonia. Formerly, an
effort was made to increase the concentration of sodium chloride by introducing solid salt into the
ammoniated brine after it had left the absorber. As the operation was attended with mechanical
difficulties, the practice has been generally discontinued. The increase of concentration of sodium
chloride would undoubtedly increase the production of sodium bicarbonate in the towers by its
mass action, although the gain may not warrant the trouble.
However, when ammonium chloride is to be recovered from the mother liquor from the
columns, it is worth while to fortify the ammoniated brine with pure solid salt to bring the
concentration of NaCl to theoretical saturation in the ammoniated brine at the top of the columns,
or before the ammoniated brine is sent into the columns for precipitation of NaHCO3 .
(b) The increase in the volume of brine. The ammonia gas dissolved in the brine
increases its volume just its volume just as it increases the volume of water. In the case of pure
water, the increase in volume of the aqueous solution of ammonia at 15 . as shown by the
decrease in the specific gravity is given by Lungs and Wiernik in Table 49.
TABLE 49. Specific Gravity of Ammonia Solutions.
Grams NH3 in Grams NH3 in
1000 cc. Sp. Gr. 1000cc Sp. Gr.
Solution at 15 15 /15 Solution at 15 15 /15
0 1.000 51.8 0.978
4.5 0.998 56.6 0.976
9.1 0.996 61.4 0.974
13.6 0.994 66.1 0.972
18.2 0.992 70.9 0.970
22.9 0.990 75.7 0.968
27.7 0.988 80.5 0.966
32.5 0.986 85.2 0.964
37.4 0.984 89.9 0.962
42.2 0.982 95.1 0.960
47.0 0.980 100.3 0.958

At a concentration of 80 grams of ammonia per liter, the increase of volume is a little less
than 12 per cent. In the case of brine, at the same concentration of ammonia, the volume increase
in the resulting ammoniated brine is slightly greater (about 13 per cent), as compared with plain
brine of the same sodium chloride concentration. Fortunately, the presence of carbon dioxide in
the ammonia gases forming some ammonium carbonate in the ammoniated brine tends to offset
this increase in volume.
(c) The dilution of the saturated brine by steam condensate. Theoretically it is impossible
to get perfectly dry NH3 gas from the distillers. Production of water vapor can be minimized only
by operation at a low temperature, taking advantage of the fact that aqueous tension of water is
much less at a high temperature (at 50 . it is only 12 per cent of what it would be at 100 .).
The question is: Can we work at as low a temperature as we desire? In practice, we cannot,
because when we cool below about 55 . the ammonia gases from the distiller containing carbon
dioxide, there is a tendency for the solid ammonium carbonate, ammonium bicarbonate and
ammonium bicarbonate crystals to deposit, blocking the gas passage. Equilibrium data from the
NH3 -CO2 -H2 O system indicate the possibility of the existence of the following compounds:
Ammonium Carbonate, (NH4 )2 CO3 . H2 O
Ammonium Carbonate, NH4 O. CO. NH2
Ammonium Bicarbonate, NH4 HCO3
and the double salts, NH4 HCO3 . NH4 O. CO. NH2
and 2NH4 HCO3 . (NH4 )2 CO3. H2 O
This limitation is a practical one. Following are temperatures at which such substances exist in the
solid state at atmospheric pressure.
Solid (NH4 )2 CO2 . H2 O exists below 58
Solid NH4 HCO3 exists below 60
Solid NH4 O. NH2 CO exists below 60
Solid NH4 HS exists at low (unknown) temperature
Solid (NH4 )2 S exists at low (unknown) temperature
The crystallizing temperature of the ammonia gases varies according to the relative concentration
of carbon dioxide the ammonia vapors carry.
Since the pipes are generally large, occasional periods of too low a temperature, if followed
promptly by periods of higher temperatures, will clear up the passage by sublimation of the
deposited crystals, or the blockade may be broken through by injecting exhaust steam. In such
cases of fluctuation, generally considerable steam will be carried over to the absorber with the
ammonia gas, diluting the ammoniated brine.
Theoretically at 55 . and 754 mm. Pressure (1/4”vacuum Hg) the partial vapor pressure of
steam is 118 mm. Hg., and that of ammonia hydrogen sulfide, carbon, and air, therefore, 636 mm.
Hg. Of these gases, ammonia and carbon dioxide exist in predominating quantities, and hydrogen
sulfide and air may be neglected in our calculations.
In the ammoniated brine obtained the ammonia “titer” (i. e., number of cc. Of N H2 SO4 per
20 cc. of solution) is about 95 (or 80.7 g. /l. solution), and CO2 , 500 cc. per 20 cc. (or 49.1 g. /l.
solution). From this, the ratio of the partial vapor pressure of ammonia to that of carbon dioxide is
as 80.7/17 is to 49.1/44 or as 4.75 is to 1.115. Therefore the partial vapor pressure of ammonia is
515mm. The density of the ammonia vapor at 515mm. Hg is 0.0395 lb. Per cu. Ft. and that of
water vapor at 118mm.Hg is 0.00650 lb. per cu ft. This gives 164.5 grams of steam per 1000
grams of the ammonia distilled over. In this calculation we have assumed that all ammonia and
carbon dioxide gases in the ammonia vapors coming over from the distiller have been completely
taken up by brine in the absorption system and that the temperature of the ammonia vapors
entering the absorber is held at 55 . Constantly. Actually some carbon dioxide gas is left
unabsorbed in the spent gas and the temperature of the ammonia vapors from the distiller
condensers does fluctuate and rise considerably above 55 .So in practice from twice to three
times the amount of water will go over to the absorber.
 In practical absorption operations the 95 ammonia “titer” and 89 chlorine “ titer” correspond
to 80.7 grams of ammonia and 260 grams of sodium chlorine per liter of the ammoniated
brine .The Cl- titer of the cold saturated brine was 108, but the brine actually obtained has
generally 104-105 Cl- titer .The combined increase in volume due to ammonia absorption and to
dilution by the steam condensate in the absorber is therefore 17.6 per cent, of which 4.1 per cent
has come from steam carried over by ammonia vapors from the distiller condensers and 13.5 per
cent is due to the increase in volume from ammonia absorption ,when the ammonia gas from the
distiller condensers is maintained at 55 .
(d) The generation of heat by the solution of ammonia vapor in the brine and partial
neutralization of ammonia by CO2 .The heat of solution of ammonia gas is 8430 calories per
g-mol-a very large quantity. Therefore, to make ammonia gas stay in the brine this quantity of heat
must be extracted, i. e., and the brine must be cooled during absorption. Water in the brine absorbs
ammonia gas with great avidity and there would be no difficulty in preparing an ammonia solution.
The whole trouble lies in the ability to cool. Cooling on a large scale is attended with considerable
difficulty because of incrustation of the cooling tubes. After ammonia absorption, the impurities I
the brine begin to separate in the form of calcium and magnesium carbonates, ferrous sulfide, etc.
On striking the cooling tubes these precipitates coagulate and settle on the tubes, in the apparatus,
and in the piping system. This crust formation seriously reduces the cooling capacity.
Occasionally the entire opening, both in the inlet and outlet of the apparatus and of the pipe, is
filled with s much grayish sludge or black scale that no brine can flow through. This may result in
the complete interruption of operations.
An approximate estimate of the quantities of heat liberated during absorption follows. This is
based on the assumption of a 95 titer of ammonia and 500 cc. of carbon dioxide per 20 cc. of
ammoniated brine and 400 kg. of water vapor per 1000kg. of ammonia .For every 1000kg. Of
ammonia in the absorber, there are 12.4 cubic meters of ammoniated brine made .The specific heat
of this ammoniated brine is 0.78 and the specific gravity is 1.175. (The specific gravity varies as
the amount of carbon dioxide dissolved in the ammoniated brine.)
Assuming that the temperature of the brine entering the absorber is 25 , the amount of heat
liberated in it (if not conducted away resulting ammoniated brine above its boiling point
(91+25=116 .). This means that heat is being generated in the absorber at the rate of one-half of
one million kilogram calories per ton of ash made, assuming the carbon dioxide titer in the
ammoniated brine is 500 cc. per 20cc. sample.
Consequently, inefficient cooling or an insufficient cooling surface may cause the whole
absorber to become hot to top – a phenomenon known as “ hot top.” In this condition ammonia
gas cannot remain dissolved in the brine in the absorber but passes through to the absorber washer
and even to the weak washer; and it may be sent with the exhaust gas to the air through the
vacuum pump or “exhauster.” This is a source of loss of ammonia which should be watched for. It
is the temperature that makes the operation of the absorber the reverse of that of the distiller.
After a preliminary absorption of ammonia gas in the absorber washer and in the upper part
of the absorber, the liquor is usually cooled in outside coil of the trombone type (called absorber
coolers) made of 8- inch pipes, with return bends over the top tier of which cooling water is
allowed to play and flow down the outside surface of the pipes in thin films. The liquor from the
upper part of the bottom of the absorber from the top tier for final absorption. Here the liquor is
finally ammoniated to the proper “titer” and flows to the settling vats at a temperature of 60 to
65C. This temperature range is found suitable for the settling of the precipitate. These settling vats
may be some 14 feet in diameter or larger by 25 feet overall height. They are provided with 45
conical bottoms they may be made of inch steel plates, welded or riveted together. Three of such
vats are connected in series so as to get the ammoniated brine from the last vat perfectly clear.
Each vat is provided with an agitator to keep the precipitate of “mud” loose from the sides and
bottoms. The agitators make from 1r.p.m. In the first vat to 1 revolution per hour in the last vat
(outlet vat). A set of three such vats in series has a capacity of from 100 to 250 tons of soda ash
every 24 hours, depending upon the quality of brine used. The mud is drawn out regularly from
the bottom of each vat by means of a small slow-moving plunger pump, and sent to the distiller to
recover ammonia. But before doing this it is best to send the mud to an intermediate settler, called
the mud settler, to concentrate the mud further so that as small a volume of mud settler, to possible
is sent to the distiller, and the clear ammoniated brine is returned to the system. This is to avoid
pumping too large a volume of good brine to the distiller where all the also if contains is lost.
Besides, it would tax the capacity of the distiller unnecessarily. Since the mud so filter liquors to
the mud pipe so that the mud pump may be flushed and the mud may be diluted by the filter liquor
through the pump to the distiller.
The large settling capacity of the vats also helps maintain an ammoniated brine of constant
strength for the vats also helps maintain an ammoniated brine of constant strength for the column
operation, making it less susceptible to changes in the concentration of the ammoniated brine
made in the absorber.
Normally the heat transfer between a gas and a liquid is not so effective as between a liquid
and a liquid. It would seem, therefore, to be more effective to cool the liquor rather than the gas in
the absorber by means of cooling of the liquor alone would not be adequate. In practice, the
method of cooling both the liquor and the gases by cooling tubes is employed in the absorber. The
liquor-cooling surface should have roughly 25 sq. ft. per ton of soda ash, while the gas -cooling
surface should have 15sq.ft. Per ton of ash, making a total cooling area of 40 sq. ft. per ton of soda
ash output. The estimate of cooling surface required, however, depends upon (1) the distribution
of cooling area in the apparatus, (2) the temperature of the cooling water available, and (3) the
scaling characteristics of the ammoniated brine (i.e., the amount of impurities in the brine). High
magnesium brine, such as sea brine, causes the formation of hard scale which may completely
counteract the cooling effect of the liquor cooling system; whereas a high carbon dioxide content
in the ammonia gases may cause ammonium carbonate (or bicarbonate) crystals to deposit on the
cooling tubes in the gas -cooling system if the temperature drops excessively low. Inasmuch as so
many factors are involved, the foregoing figures for cooling area per ton of ash merely afford a
rough guide in the design of the absorber.
The absorber and vat house operation is chemically simple but mechanically very
troublesome. The plugging of the piping system but the mud incrustation is a serious matter
especially when sea salt is used. Duplicate piping systems and cooler sets are employed to
facilitate cleaning without inter interrupting the operation. Tees and crosses are employed in
pipelines and overflow connections to facilitate opening up for cleaning. Tables 50 and 51 five
typical analyses of the “mud” from the sea brine pipe main.
The molecular ratio of MgCO3, NACI and NaCO3 are approximately 1:1:1. Hence the tripe salt in
question should be: MgcO3 Na2 Co3 NaCl. This would correspond to Northupite crystal in natural
deposits.
The reasons that tend to prove that the double salt of CaCO3 and Na2 CO3 (the Gaylussite
CaCO3 Na2 CO3 5H2 O formation) does not exist here given below:
(1) That CaCO3 is present in different scale sample in widely varying ratios with respect to
Na2 CO3 .
(2) That CaCO3 percentage in each case here is too low to form t Gaylussite (CaCO3 . Na2 CO3 .
5H2 O).
(3) That Water present is also far too low to form the hydrated crystals of Gaylussite.
Table 51 Analysis of Ammoniated Brine Scale, Sample B.
(Hard Scale from Brine Overflow Pipe Using Sea Brine.)
The molecular ratios of MgCO3 NaCl and Na2 CO3 are as 1:1:1 and the triple salt in question is
again MgCO3 Na2 CO3 NaCl.
These analyses show that the predomination constituents in the scale are MgCO3 Na2 CO3
NaCl and CaCO3 Several points are noteworthy:
(1) Magnesium here is not precipitated from ammonia solutions as Mg (OH) 2 but in the form
of MgCO3 which is normally more soluble than Mg (OH) 2 ;(2) NaCl is present in the
mud in excessive quantities, which have been precipitated from an unsaturated
ammoniated brine (unsaturated with respect to NaCl) and (3) Na2 CO3 is formed at this
stage instead of the far more insoluble NaHCO3 Further study of the composition of the
scale confirms the fact that there is a molecular relation-ship among the three main
constituents and that they are in the following molecular ratio:
MgCO3 : NaCl: Na2 CO3 = 1: 1: 1
so that NaCl and Na2 CO3 must have been formed from their respective ionic constituents with
MgCO3 in the form of tripe MgCO3. NaCl. Na2 CO3 in the mud.
The foregoing description applies to brine that has not been treated before ammoniation .If
the brine has been pre-treated, no such mud would be formed in large quantities. The loss of NaCl
and Na2 CO3 present in the mud would be avoided and the troublesome operation in the absorber
and vat system would be entirely eliminated.
Table 52 gives figures showing the concentrations of magnesium, chlorine, and total
ammonia. free ammonia, and fixed ammonia at different stages of ammoniation in the absorber
system. Certain irregularities, however, are to be noted.

The preceding analyses made on the clear portion of each sample. There were 854-mg/l.
calcium in the fresh brine sample but only traces in all other samples, showing that calcium is very
readily precipitated as CaCO3 .
 FIG 26

Photomicrograph of artificial crystals

of ammonium-amgnesium carbonate

tetrahydrate [(NH4 )2 CO3 .MgCO2 .4H2 O]

found in ammoniated brine mud.

The question of impurities in the sea will center upon magnesium because there magnesium
exists in larger proportions and also because the chemistry of magnesium is far more complicated
than that of calcium. In the presence of CO2 in an ammoniacal solution, magnesium seems to have
the tendency to precipitate as the carbonate (MgCO3 ) or at least as a basic carbonate, rather than as
the hydroxide (Mg (OH) 2 ), as at stated above. Seldom, however, does the magnesium precipitate
exist as a simple salt: usually it comes down as a double or triple salt. Except where the ammonia
and Co2 gases are weak (such as the tail end of a washer, where such a simple salt as the
magnesium carbonate cryohydrate (MgCO3 .3H2 O) may separate out instead). As the gases
become stronger and the brine absorbs more ammonia and CO2 , the double salt of
magnesium-ammonium carbonate tetra hydrate (MgCO3 .NH4 ) 2 CO3 .4H2 O) and the triple salt of
magnesium- sodium carbonate and sodium chloride (MgCO3 .Na2 CO3 .NaCl) will occur in place of
the simple salt. These salts appear under the microscope at high magnification as in Figs.26 and 27.
These double and triple salts of magnesium separate most readily at a temperature above 50°C
Table 53 below reveals the crystal aggregate in the mud at various points starting from the
saturated brine to strong liquor at the carbonating tower end.

FIG 27 Photomicrograph of artificial Northupite crystals (Na2CO3 NaCl MgCO2) found in ammonated brine mud.

Most of the foregoing crystals are transparent with a wide range of sizes varying from what can be
seen at 60-diameter magnification to what cannot be clear discerned at 800-diameter
magnification. They are generally difficult to see at high magnification because of their
transparency, but can be seen best art 100-diameter magnification. From the above table it is
apparent that at the bottom of absorber washer, absorber, brine cooler and ammoniated brine
storage, the triple salt constitutes a large part of the mud formed.
The clear ammoniated brine is cooled to 30°C .in vats liquor coolers similar to the absorber
coolers previously described, since hot ammoniated brine, if fed to the top of the columns, would
cause excessive and brine, if fed to the to the top of the columns, would cause excessive
vaporization of ammonia by the carbon dioxide spent gases from the columns.
* Cf. Wilson, E.O. and Chiu, Y.C, “Brine Purification,” Ind. Eng. Chem. 26,1099(1934)

On cooling the liquor, any magnesium mud, if not completely settled in the vats, would separate in
the vat liquor coolers. This causes the formation of hard scale inside the cooling tubes, which
obstructs the liquor passage and renders the coolers ineffective. This occurs especially when the
temperature of the ammoniated brine if reduced below 30°C. Since the difference in temperature
between the ammoniated brine and the cooling water in the vat liquor coolers is smaller these in
the absorber coolers, much larger cooling area is required in these vat liquor coolers to get a cool
liquor for the column operation. This can be shown mathematically as follows. Suppose in the
absorber coolers the ammoniated brine enters the coolers at 60°C. and leave at 40°C., and in the
vat liquor coolers it enters the coolers at 50°C. and leaves at 30°C. Suppose also that the cooling
water available has a temperature of 18°C. and its temperature is raised to 25°C. at the outlet in
each case, the cooling system in each case being arranged counter-currently. Then the average
temperature difference between the liquor and the cooling water in each case is represented by the
logarithmic mean
so that, other conditions being equal, the cooling surface for the vat liquor coolers must be about
60 per cent larger than that for the absorber coolers. In practice 100 per cent larger area is
recommended for the vat liquor coolers.
It is quite important that the ammoniated brine for the column operation should be well
settled and free from mud. That good soda ash is made in the absorber and vat system should be
kept in mind. For without perfectly clear ammoniated brine, mud would be carried over to the
columns, be precipitated with the bicarbonate, and finally find its way to the soda ash, causing a
turbid appearance or, in the worst ease, a flocculent precipitate in the soda ash solution. This
affects the quality of the product of the plant. Furthermore, if mud is accumulated in the columns,
cleaning by the usual methods would not be effective and the columns might eventually have to be
shut down for a boiling operation and mechanical removal of the scale. The scale would consist
mostly of magnesium carbonate, which is difficultly soluble in hot water.
Magnesium mud is most difficult to settle out completely. If the brine used contains large
amounts of magnesium with other impurities, magnesium is the last to be removed and small
amounts of it will pass through with the brine stream. The amount of magnesium mud left in the
ammoniated brine depends upon the settling capacity of the vat system and is determined by the
rate of flow vs. the volume of the settling vats. With the usual proportion of the height to the
diameter (or cross-section area) of the vat construction, it is the volume, rather than the area, of the
vat that determines the settling capacity. Table 54 shows the relation of the rate of flew vs. the
amount of magnesium left in the ammoniated brine (sea bring) coming from a settling vessel.

Table 54. Rate of Flow vs. Magnesium Carried in the Effluent.*

(Settling Vessel has 231/2 cm. 231/2 cm. 35cm. Capacity.)
Mg. Carried in the Mg. Carried in the
Rate of Flow Ammoniated Brine Rate of Flow Ammoniated Brine
(cc. per min.) (gram per liter) (cc. Per min.) (gram per liter)
83.3 0.229 25.0 0.0961
 77.0 0.1878 20.0 0.0926
50.0 0.1214 14.3 0.0899
33.3 0.1057 11.1 0.0882
*Thesis by Dr. W. C. Hsieh with the suthor.
These experiments were carried out with a 5-gal. Tin can (23 1/2 23 1/2 35 cm. high), using
normal ammoniated brine from the absorber. Brine enters at the bottom and flows out from the top.
The rate of outflow is measured and the magnesium in the effluent determined by analysis.
Samples in each case were collected after then hours of constant flow.

FIG 28 Curve showing rate of settling vs. volume of flow.

These results are plotted in a curve (Fig.28). From this curve it will be seen that for a setting
volume of 23.5 23.5 35 cc., to reduce the magnesium content to within 0.1 milligram per
liter in the ammoniated brine, the rate of flow must be below 30 cc. Per minute. Granting that this
is the highest limit in the green liquor for the column operation, with the usual proportion of the
height to the diameter for each of the settling vats, the minimum volume necessary for settling
such brine per ton of soda ash made per 24 hrs. Would be 95 cu. Ft. for each settling vat.
23.5 23.5 35
6 cu. m. 35.3 cu. ft. = 95 cu. ft.
30 60 24
Therefore for a 100-ton plant employing such brine, a minimum of 9500 cu. ft. of settling capacity
is needed for the vats. Usually a capacity much in excess of 100 cu. ft. per ton of soda ash made
per 24 hrs. Must be provided, since allowance must be made for fluctuation of the flow. This
indeed calls for a huge settling capacity, and applies to un-pretreated sea salt.
The matter of having a proper ratio of ammonia to sodium chloride in the ammoniated
brine is important, for upon this depends the efficiency of the column operation. The Cl-titer
(which represents sodium chloride) should be kept as high as possible’; but because of dilatation
of the volume of brine by ammonia and dilution by the steam condensate as described in the
earlier part of the chapter, the sodium chloride concentration in the ammoniated brine is limited by
the absorber operating conditions to below 90 Cl- titer. Should the ammoniation in the absorber be
carried further, the concentration of sodium chloride (as shown by the Cl- titer) in the ammoniated
obtained would be greatly reduced. If the ammonia concentration in the brine were too high,
excessive amounts of ammonium bicarbonate would also be precipitated together with sodium
bicarbonate in the columns. If it is too low, the decomposition of sodium chloride in the columns
would be low and too much salt wasted, this affects the rate of salt consumption and the utilization
of ammonia and is vitally important to the financial success of the undertaking.
It is found best to have the ratio of 1.08 to 1.21 of NH3 titer to 1.00 of Cl- titer in the
ammoniated brine. Table 55 is based on Hempel and Tedesco’s results. It will be noticed that a
very high efficiency in the utilization of ammonia is necessarily accompanied by a low efficiency
in the utilization of salt, and vice very. The optimum point is reached

when the utilization of both of these principal raw materials reaches, of is slightly below, 75 per
cent, i.e., 99 t0 97 for NH3 titer, and 89 to 90 for Cl- titer with the ratio 1.10 of NH3 to 1.00 Cl.
These figures represent the best conditions obtainable in actual plan operation, although
theoretically such ammoniated brine is still slightly undersaturated with respect to sodium chloride
at this ammonia concentration.
Since in the ammoniating brine, there is a purification effect resulting in the precipitation of
CaCO3 , MgCO3 , Mg (OH) 2, etc., in the hot condition, a small amount of NH3 in the brine exists as
NH4 Cl or (NH4 ) 2 SO4 , i.e. as fixed ammonia as mentioned in Chapter V. The amount of the fixed
ammonia in the ammoniated brine from the absorber depends upon the purity of the brine used.
Generally, the fixed ammonia in the ammoniated brine is from 3 to 5 titer when rock salt brine is
used, but it may be as high as 8 to 9 titer when sea salt is used. For the effect of this fixed
ammonia on the column reaction, see Chapter V, on the purification of brine. Unless sea brine
(which is obtained from sea water directly without crystallization) is used, this would not
materially affect the decomposition of NaCl in the columns.
It should be remembered in designing the absorber that the absorption of ammonia requires
depth of wash (baronage). It is not unusual to provide a net depth of 18 to 20 inches in each
passette from the serrated edge of the mushroom to the lower edge of the overflow opening.
As has been previously pointed out, if the brine was pre-treated to eliminate calcium and
magnesium before it was sent to the system, ammoniation would be exceedingly simple. Most of
the difficulties above described would then be avoided. The capacity of the absorber and the vat
system would also be considerably increased, and the operating cost greatly lowered. The
ammoniation then form the final step of removing the last traces of impurities in the brine under
the best settling conditions, making it possible to produce the highest grade of soda ash.
 Chapter VIII

Carbonation of Ammoniated Brine

In early times solutions of salt and ammonium bicarbonate were used in the ammonia soda
process. A great step, indeed, was made when ammonia first employed in the form of a gas to be
dissolved in brine thus formed. This can be readily appreciated when one attempts to precipitate
sodium bicarbonate in the laboratory, using brine and ammonium carbonate or bicarbonate
solution, even though both these solutions are saturated to start with. From the saturated brine
mixed with an equal volume of saturated ammonium bicarbonate in the cold, no immediate
precipitate of sodium bicarbonate is obtained. Crystals separate out only after long standing.
Sodium bicarbonate in the cold seems to have a great tendency to form surpersaturated solutions.
On the other hand, granting that both solutions were saturated before they were mixed, the
resulting solution is only half saturated in respect to either NaCl or (NH4 ) 2 CO3 (or NH4 HCO3 ). A
big advance was made when a method was discovered for introducing ammonia and carbon
dioxide both in the form of a gas separately into brine without materially lowering the strength of
the brine. The present method of ammoniation with tolerably dry ammonia gas followed by
carbonation with carbon dioxide gas is indeed the key to the success of the ammonia soda process.
Furthermore, ammoniation is incidentally an important process for the purification of crude brine
when the brine has not been pre-treated, while the heat so developed is essential to the separation
of calcium and magnesium precipitates from the resulting ammoniated brine. Then again,
carbonation with CO2 (though necessary to be extracted by cooling water towards the bottom of
the column) is essential to the formation of good sodium bicarbonate crystals in the columns.
The ammoniated brine from the settling vats should be clear; other wise the suspended solids
would be precipitated with the bicarbonate in the carbonating towers, contaminating the soda ash
made. It should be further cooled to about 30 .in vat coolers, or brine coolers, before being sent
to the towers for carbonation; otherwise, because of the increased vapor pressure of ammonia at an
elevated temperature, an excessive quantity of ammonia gas that is otherwise available for
reaction would be distilled over by the bubbling of CO2 gases out of the tower into the tower
washer, there, on the other hand, more brine would be required to scrub the exit gases free of
ammonia before letting them emerge into the atmosphere. This can be done only within certain
limits, because the quantity of fresh brine that can be so employed we limited by the output of
soda ash in the plant, i. e., by the quantity of the ammoniated brine made in the absorber at affixed
rate of ammonia distillation. For this reason, some of the older carbonating towers provided the
admission of the fresh ammoniated brine, not at the top of the tower but near the middle. This was
working on the theory that the free ammonia in the liquor would be converted to ammonium
carbonate by CO2 gas before it could come to the top of the tower, and thus less ammonia would
be distilled over by virtue of the lower ammonia vapor-pressure over ammonium carbonate than
over free ammonia in the brine. In any case, there is some loss of ammonia in the towers due to
gas distillation. This is one reason why the ammonia titer in the ammoniated brine should be
slightly higher than the chlorine titer.
 FIG 29

Carbonating tower

FIG 30

Carbonating tower cooling

boxes.

The clear, cooled, ammoniated brine of proper concentration in sodium chloride and
ammonia is treated with CO2 gas in a carbonating apparatus. There are several types of this
apparatus, e.g., Solvay, Honigmann’s,
 FIG 31 Carbonating tower upper ring.

Boulouvard’s, etc. Among these, Solvay’s type is the most universally used (Fig.29). A tall
vertical structure consisting of a number of the mushrooms and division plates, or ”passettes,” one
above the other, is typical of Solvay’s design. The carbonating tower is an example. It consists of
cast iron cylindrical sections, or ” rings” of about 6feet or 2 meters inside diameter and 3 feet 4
inches high for each of the bottom cooling section (Fig.30) and 18 inches high for each of the
upper plain rings (Fig.31). The cooling system is of the Cogswell type, consisting of a large
5
number of horizontal cast-iron tubes of 2 inches outside diameter, 1 inches inside diameter,
16
and about 7 feet 3 inches long, laid in rectangular, cast-iron tube sheets.
 FIB 32 Carbonating tower or “column.”

The whole tower is about 75 feet high from the base to the top gas outlet. Carbon dioxide gas
enters at the bottom, bubbling up around the edges of an inverted cone opening (Fig.32). The draw
opening through which the magma or bicarbonate sludge is drawn out is located at the bottom
diametrically opposite the gas outlet.
 FIG 23 Another type of oarboatbag tower.

Instead of working the bicarbonate on the bottom floor, however, the bicarbonate sludge is
brought up through a 5-inch draw pipe to a height of 35 to 40 feet by the pressure in the tower
itself. It then flows down to the bicarbonate filters located some 35 feet above the bottom floor.
This elevation is necessary for vacuum filter working in order to provide a seal column under high
vacuum above the level in the storage tanks which are located on the bottom floor, vented to
atmosphere.
Within recent decades a departure in the size of the “standard” carbonating tower or
“column” has been made. The diameter has been increased from six feet to seven feet six inches
for substantially the same height with a large increase in capacity.
Another form of passette construction for the carbonating tower proposed by Mr. E. N.
Trump is shown in Fig. 33. This type of passette has a clear passage and is especially good for the
precipitating columns, in which fine crystals settle fast and have a tendency to block the passage.
The passettes are particularly suitable for the manufacture of refined bicarbonate of soda.
The composition of the normal ammoniated brine to the towers is given in Table 56.
 FIG 34 Approximate temperature gradient in eartonating tower.

In tower operation temperature regulation is very essential. The temperature gradient in the
towers should be so graduated by “bleeding” out of the cooling water that the bicarbonate crystals
will grow larger and larger toward the bottom as the liquor descends and is cooled .The
approximate temperature curve along the height of the tower is shown in Fig.34.
The lowest temperature is of course to be found at the draw ring at the bottom as the cooling
is arranged counter currently. Various statements appear in textbooks for the best drew liquor
temperature, such as 30 , 320 , 360 ,etc. Practice does not confirm these statements. With a
well-constructed apparatus and good cooling facilities, as low a draw temperature as 22 .or 20
.has been found advantageous. It would seem indeed that the temperature for the draw liquor is
primarily limited by the cooling capacity with the cooling water available, rather than by the
desirability of working at a certain temperature for any chemical reasons. Low draw temperature
does not necessarily mean poor crystals or pasty formation of sodium bicarbonate that is hard to
filter. The formation of good crystals in the columns depends upon:
(a) High concentrations of ammonia and sodium chloride in the cooled ammoniated brine with a
proper ratio between the two titers.
(b) Rich carbon dioxide gas properly cooled.
(c) High reaction temperature at a point about two-thirds of the height of the column above the
base.
(d) Gradual cooling from this point down, ending with thorough cooling with as much cools
water as possible as the draw point is reached.
This is consistent with the theory of crystallization. In a solution containing high
concentrations of the reacting constituents, sodium chloride, ammonia and carbon dioxide, the
ionic concentrations are such that the solubility product of the desired compound is quickly
exceeded and the sparingly soluble substance (in this case, sodium bicarbonate) is irreversibly
precipitated in a solid form. Then when the reaction occurs at a high temperature, i. e .,when the
precipitation, as of sodium bicarbonate, takes place in the hot condition, the mother liquor to a low
temperature for drawing helps to throw as much of the substance out of solution as possible. Thus,
good conversion ratio,” per cent decomposition” and good crystals for filtration and washing
operations are simultaneously effected. Furthermore, as the draw liquor is expose to the air during
filtration the loss of ammonia gas will cause personal inconvenience to the workmen in the room.
By following this procedure the draw temperature can be advantageously brought down as low as
can be secured with the cooling water available with no undesirable results.
To get as low a draw temperature as 20 to 22 . a very low cooling-water temperature is
required. It would be impossible to get this draw temperature if surface water or river water were
used for cooling during the summer. Under ground water, or well water, is best suited for this
1
purpose, as the temperature is always between 15 and16 ,summer and winter, at a depth of
2
400 to 450 feet below the surface. This uniform temperature of the cooling water all the year
round is desirable, as uniformity and regularity are required in the ammonia soda industry.
The main reaction in the tower is the double decomposition,
NH4 HCO3 +NaCl NH3 Cl +NaHCO3
The heat effect is, the tower, very small as shown below, and the reaction is a typical one of
equilibrium.
NH4 HCO3. Aq +NaCl. Aq NH3 Cl. Aq +NaHCO3. Aq +Q1
(-199,000) + (-96,600) = (-72,800) + (-222,700) + Q1
where Q1 = -100 cal .and is the heat of reaction at 15 .Because of limited accuracy of the heat
data available, these figures are evidently beyond the range of significant figures. It seems that the
reaction is thermochemically neutral. But when the sparing soluble sodium bicarbonate (which is
still less soluble in the presence of an excess of sodium chloride and ammonium bicarbonate),
separates out in a solid form, then there is a positive heat of reaction, which a moment’s thought
will tell us is nothing but the heat of solution of NaHCO3 crystals. The reaction is then exothermic.
NH4 HCO3 Aq +NaCl Aq NH3 Cl Aq +NaHCO3 Aq +Q2
(-199,000) + (-96,600) = (-72,800) + (-222,700) + Q2
Q2 = 4,900cal
Or per ton of ash, at an average of 72 per cent decomposition,260 grams per liter of sodium
chloride and 6 cubic meters of ammoniated brine,
Q2 = 260 0.72 4,900 cal 6,000 l =94,000,000 cal
58.5
The main heat effect, however, comes from the formation in the liquor of ammonium
carbonate from the neutralization of the dissolved ammonia by CO2 gas. It should be noted that
not all this neutralization of the ammonia in the ammoniated brine is effected in the towers, as the
ammoniated brine has picked up some carbon dioxide in the absorber system to the extent of about
500cc .per 20cc.Assuming 500cc of CO2 per 20cc.Aussuming 500cc. of CO2 per 20cc of sample,
this leaves 80.7-37.9 or 42.8 grams ammonia per liter yet to be neutralized to ammonium
carbonate in the towers. Finally all this (NH4 )2 CO3 is converted to acid carbonate,NH4 HCO3 by
further treatment with CO2 gas in the making towers. The system is unfortunately complicated by
the fact that before all the ammonia is neutralized to form normal carbonate, a part of the normal
carbonate,(NH4 )2 CO3, formed is immediately hydrolyzed to form NH4 HCO3 and NH4 OH,so that
(NH4 )2 CO3 and NH4 HCO3 are not formed in distinct stages as described above. Further, ammonia
is never 100 per cent bicarbonated to NH4 HCO3
NH4 HCO3 + H2 CO3 NH4 HCO3 +H2 O because of hydrolysis
The reaction takes place in this way because it is a salt of a weak acid reacting with a weak base.
2NH4 OH + H2 CO3 (NH4 )2 CO3 +2 H2 O
+
H2 O

NH4 HCO3 + NH4 OH

Hence the division of reaction between the cleaning and the making towers is not so distinct
as described above.
J. Thomsen gives the heat of neutralization,
NH3 Aq + CO2 .Aq (NH4 )2 CO3 Aq + 2 H2 O + Q2
Q3 =16,850 cal. per mol and the heat of solution of CO2 gas to CO2 .Aq is 58880 cal. per mol.
so that the neutralization of 42.8 grams NH3 per liter with CO2 gas to form (NH4 )CO3 .Aq would
cause the evolution of
42.8 (16,850 +5880) = 42.8 22,730 = 28,600 cal
34 34
or per ton of ash, 28,6000 cal. 6000 l. = 171,600,000 cal.
The second stage
(NH4 )2 CO3 .Aq + H2 CO3 .Aq 2NH4 HCO3 .Aq + Q4 ,
(if the reaction were so definitely marked ) gives further heat of reaction
(-221,600) + (-168,000) = 2(-199,000) +Q4
Q4 = 8400 cal. per mol.
The heat of reaction per mol of carbon dioxide due to this reaction then is
80.7 1 8400 = 19,950 cal
17 2
or per ton of ash, 19,950 cal. 6000 l. = 119,700,000 cal.
To counteract the heat evolved due to the foregoing reaction it is sometimes argued that heat
is absorbed because the carbon dioxide expands from atmospheres (34 to 36 lbs. gage pressure at
the inlet) to atmospheres (3 to 4 lbs. gage pressure at the top of the tow er).But this Joule-Thomson
heat effect is in reality very small here and need not be taken into account. Hence per ton of ash,
the total quantities of heat generated in the cleaning and making towers are:
(1) Heat from carbonation of NH3 to (NH4 )2 CO3 corresponding to 171,6000,000 cal.
the first stage on the tower side 119,700,000 cal.
(2) Heat from carbonation to (NH4 )HCO3 by CO2 corresponding 94,000,000 cal
to the second stage 385,300,000 cal.
(3) Heat from precipitation of NaHCO3 per ton of ash or
Total 385,300 kg. Cal.
per ton of ash
The entire reaction starting from the saturated brine and gaseous ammonia and gaseous carbon
dioxide is represented as follows:
NaCl Aq + NH3 (gas) +H2 O(liquid ) + CO2 (gas ) NaHCO3 (solid ) +NH4 Cl Aq + Q5
(-966000) +(-12000) +(-69000) +(-94500) =(-227700) +(-72800) + Q5
 Q5 =28400 cal per g-mol NaHCO3
This thermochemical equation represents the total heat effect in the absorbers and in the towers.
Apart from the heat of condensation of steam carried by the ammonia gases to the absorber. Of
this all the heat of solution of ammonia and a part of the heat of neutralization(2NH4 OH + CO2
(NH4 )2 CO3 + H2 O ) are liberated in the absorber, while the balance is liberated in the towers. The
total per ton of ash is
28,400 1,000,000 = 536,000,000 cal. or 536,000 kg Cal. per ton of ash.
53
with the heat of condensation of steam 226,000 6 =109,000
12.4
(see p.121), the total becomes 536,000+109,000=645,000kg.Cal.per ton of ash. This is the
theoretical minimum. Actually more ammonia and carbon dioxide are present in the liquor than
the quantities called for by the above equation. Hence assuming 75 per cent ammonia efficiency,
the minimum value may be put at.
645000 1.33= 0000 kg Cal per ton of ash
In the cleaning tower, the heat of carbonation starting with green liquor containing 500cc. of
carbon dioxide per 20cc. is
64-49.1 (16850+5880) 6,000,000 =46100 kg Cal per ton of ash
44 1000 1000
In the making tower, the heat liberated by the reaction is:
(a) Heat of carbonation(171,600 + 119,700 – 46,100) = 245,200 339,200kg. Cal
(b) Heat of crystallization(or solution )of NaHCO3 = 94,000 per ton of ash.
The heat generated in the cleaning tower is sufficient to raise its liquor temperature by 46,100

=8.5
(1.127+1.175) 0.78 6,000
(1.175=av. specific gravity of the green liquor, 1.127=av. specific gravity of the draw liquor,
0.78=mean specific heat), so that if the green liquor fed to the cleaning tower is at 30 ,the
temperature of the precarbonated liquor sent to the making towers will be about 38.5 .to 22
.(the draw temperature) which requires
(38.5-22) (1.127+1.175) 0.78 6,000 = 89,000 kg. Cal. per ton of ash.
2
Therefore this heat of cooling must be added to the heat of reaction in the making towers above.
The total heat whic h must be extracted by the cooling water in the making towers is then 339,200
+ 89,000 = 428,200 kg. Cal. Per ton of ash.
It is thus seen that the making towers need intensive cooling while the cleaning towers do not need
any cooling at all. But in normal operation the temperature of the liquor above the cooling section
is about 55 . and it is lowered by the cooling water to about 22 .at the bottom of the tower.
While the above quantity of heat does not represent any greater quantity of heat to be
conducted away than that in the absorber, because of the low temperature to which the draw liquor
is to be reduced, the mean temperature difference for the cooling water in the making towers is
even less than in the vat coolers. The cooling tubes here are likely to be coated with crystals of
NaHCO3 and NH4 HCO3 so that the coefficient of heat transfer is very small. Consequently, a
much larger cooling surface is required in the towers than in the absorber coolers. More than 75 sq.
ft of cooling surface are required in the carbonating tower per ton of soda ash per day, but much
depends upon the distribution of the cooling surface in the tower. As many of the cooling tubes as
are consistent with the passage of gas and bicarbonate should be concentrated at the bottom
sections of the tower, The number of these tubes per section gradually decreases with the height,
leaving a very few only on the top sections which are located about half way up in the tower.
It should be remembered that many of the foregoing heat effects were merely calculated and
have not been experimentally determined. The results are to be taken as a guide in the absence of
experimentally determined data.
In tower operation we are more concerned with the reaction
NaCl. Aq + NH3 . Aq + H2 O + CO2 (Gas)
NaHCO3 (Solid ) + NH4 Cl . Aq + Q6
Or, NaCl + NH4 HCO3 NaHCO3 + NH4 Cl
At equilibrium at any given temperature.
[NaHCO3 ][NH4 Cl]
= Constant = Kc
[NaCl][NH4 HCO3 ]
where the brackets show molal concentrations of the substances per liter.
For most effective decomposition. Kc should be as high as possible. And for a given Kc at a
given temperature the absolute value of the numerator is the larger, the larger the absolute value of
the denominator.
This means
(1) high molal concentration of NaCl in the ammoniated liquor.
(2) high molal concentration of NH3 in the ammoniated liquor.
(3) high molal concentration of CO2 in the tower liquor.
For given concentrations of these substances in solution, the value of the equilibrium
constant Kc with respect to the temperature is given by van’t Hoff’s equation,

where t is the temperature at which the reaction is carried out; T the absolute temperature = t +
273; R the perfect gas molal constant = 1.985 in cal; and Q the heat of reaction, which is positive
when the heat is evolved in the reaction and negative when absorbed. As Q is positive as shown
above (i.e., heat is evolved in the above exothermic reactions), this derivative is negative. That is,
Kc is increased, as the temperature is lowered, and vice versa. This is in accordance with the
LeChatelier principle. Hence,
(4) low draw temperature in the tower.
These four factors are discussed further.
(1) High concentration of NaCl. It is apparent from the absorber operation that owing to
the increase in volume of brine as the result of ammonia absorption and the decrease in
the concentration of sodium chloride as a result of ammoniation as well as the inevitable
dilution by steam condensate, the concentration of sodium chloride in the resulting
ammoniated brine can mot much exceed 260 grams per liter with an ammonia
concentration of about 80 grams per liter. Usually the concentration is between 250 and
260 grams of sodium chloride per liter. The practice of strengthening the liquor by
means of solid sodium chloride was attended with some difficulties and has been
generally abandoned in modern soda works. For the time being, at least, we must rest
 contented with a concentration of about 260 grams per liter of sodium chloride in the
ammoniated brine.
(2) High concentration of NH3 .With good cooling facilities, it is not difficult to get an
ammonia concentration higher than 80 grams per liter, while in poorly constructed
absorbers with poor cooling arrangements, a phenomenon of “hot top” may be the result.
With a limited cooling capacity in the absorber, loss of ammonia through the vacuum
pump exhaust is to be watched for, since the absorption of ammonia is simply a matter
of cooling. But because of the tendency of the NaCl concentration to decrease in the
ammoniated brine as ammoniation is carried further, it is inadvisable to employ an
ammonia concentration very much greater than 80 grams per liter. Theoretically, an
equivalent strength of ammonia with respect to sodium chloride is all that is required
(see equation above). Based on the mass action principle, however, a higher
concentration would of course be desirable, but it is found in practice that a very much
higher concentration of ammonia greatly reduces the sodium chloride concentration in
the resulting ammoniated brine and is not economical in the are of ammonia, because
large amounts of free ammonia would be left in the draw liquor. The general practice
now is to employ not more than 12 per cent excess of ammonia over the equivalent
amount of sodium chloride (i.e., not over 1.12 equivalent of ammonia for every
equivalent of sodium chloride, with about 260 grams of sodium chloride per liter in the
ammoniated brine). When a much greater ammonia concentration, say 150 per cent as
much as that of sodium chloride, is used, ammonium bicarbonate itself begins to
separate in the towers causing cooling difficulties. And considerable ammonium
bicarbonate will be formed with the sodium bicarbonate especially at a low draw
temperature, While this does not ultimately contaminate the soda ash, it represents just
so much ammonia to be recovered from the dryers, and to be circulated between the
towers and the dryers; so the loss of ammonia is proportionately greater. The plugging of
the furnace condensers by the formation of ammonium carbonate crystals would also be
more likely to occur.
(3) High concentration of CO2 gas. The amount of carbon dioxide dissolved is
proportional to the partial pressure of CO2 gas in contact with the solution. This is
Henry’s law. It means a high concentration of CO2 in the gaseous phase, i.e., “rich gas”
Theoretically, an equal quantity of CO2 is recovered from calcinations of the bicarbonate
in the furnace. As there is always unused CO2 in the tower waste gas, escaping in the
exhaust into the atmosphere and, as is sometimes done, the exhaust gas from the
absorber vacuum pump or “exhauster” is allowed to escape into the atmosphere instead
of being returned to the carbonating tower, there is a constant loss of CO2 . Consequently,
more kiln gas, or “lean gas,” must be drawn from the limekiln. The ratio of the weight of
CO2 in the kiln gas taken to that furnished by the furnace gas is 1.0-1.2 to 1, depending
on the rate of the dryer operation relative to the tower operation. Since 42 per cent CO2
in the kiln gas and 95 per cent CO2 in the returned gas represent a very good operating
result, the percentage of CO2 in the mixed gases from the carbonating towers cannot
average more than 58.2 per cent without pushing the dryer operation beyond the rate of
the column operation. When the furnace gas has only 90 per cent CO2 , and the kiln gas
41 per cent, the theoretical percentage of CO2 in the mixed gas into the column is only
 56.3 per cent, when the rate of furnace operation is in step with column operation, i.e.,
when the furnaces contribute the same amount of CO2 by weight to the columns as the
lime kilns. The volume of the kiln gas taken by the CO2 compressors is then 2.2times
that of the furnace gas. There is a limit to the CO2 concentration if the dryer gas is to be
mixed with the kiln gas. When the dryer gas is introduced separately as in the double
entry arrangement, then the percentage of CO2 at the bottom gas inlet to the making
towers could go up to as high s 90 per cent of CO2 by volume, if desired.
Fortunately, in Solvay’s carbonating towers the gas is forced up through a high column
of liquor under a pressure of 34 to 36 lbs. The partial pressure of CO2 is thus materially
increased, giving higher absorption of CO2 at the bottom of the tower and better
utilization of the CO2 gas. This is a distinct advantage in the Solvay type of carbonator
and the power spent in compressing the gas is well worthwhile. It is equivalent to
introducing richer CO2 gas for reaction.
The exhaust gas from the absorber system should contain rich CO2 gas when the
absorber system is reasonably airtight. It should be as high as 60 to 70 per cent CO2
when there is not much air leakage. It pays to keep the distiller, absorber and vat
systems reasonably air-tight to enable this gas to be returned to the towers, not only for
the recovery of CO2 but also of some hydrogen sulfide which is also present in the
absorber gas.
(4) Low draw temperature. Low temperature at the draw has been said to cause fine
crystals of bicarbonate that cannot be filtered dry. About the two-thirds point and a
proper temperature gradient from that point down, it is unquestionably better to draw the
liquor at 22 ,for example, than at 30 .The decomposition is improved and the loss of
ammonia gas by exposure is minimized. This demands a well-distributed and
proportioned cooling surface, ample cooling area, a good cooling water, and close
regulation of the cooling watercourse. The bicarbonate crystals will then give a good
settling test and can be filtered, washed and calcined without difficulty.
In the course of tower operation, the internal surface and the cooling tubes are
gradually coated with bicarbonate crusts, with the result that it is difficult to cool the
liquor to the desired draw temperature with a force the gas from the CO2 compressor is
materially increased because of the plugging tendency of the nest of cooling tubes by the
crystals at the bottom section. But often the pressure decreases due to the gas lift effect
through the restricted space in the towers. In this connection it may be remarked that
never for any length of time should the gases be cut off from any tower while making
bicarbonate, or the tower would be plugged solid and serious difficulties ensue. To
overcome such difficulties, if the case does happen, steam is injected into the bottom of
the tower and into the draw pipe to dissolve out the bicarbonate. When the bicarbonate
has set solid at the bottom of the tower, thawing with steam in the draw opening may
only open a channel through which the gas would blow through the draw opening and
out through the draw pipe. In serious cases the gas comes out thus at the draw opening
and the tower must be drawn empty, filled with water and cooked with steam. This
always results in off-colored bicarbonate and is at best unsatisfactory. In the normal
operation there should be no occasion to resort to such drastic measures. All that is
needed would be to cut off the cooling water, change to the lean gas, draw the tower low
 and thin of bicarbonate, throttle down the CO2 gas from the compressor, and fill the
tower with green liquor, All green liquor is then passed through this cleaning tower to
feed making towers. The cooling water is completely cut off to permit the temperature
of the liquor to rise to about 39 . The green liquor tends to dissolve the bicarbonate
crust and the higher temperature drives the reaction backward, unmaking the
bicarbonate:
NaHCO3 + NH4 Cl NaCl +NH4 HCO3 NaCl +NH3 +CO2 H2 O
This constitutes the cleaning tower operation and a schedule of about 4 days should be adopted for
cleaning the making towers rotation. The cleaning operation lasts from 18 to 24 hours, depending
upon the effectiveness of cleaning (i.e., the temperature of the ammoniated brine and the volume
of the green liquor put through). If a tower is over cleaned or” cleaned to the metal,” the
bicarbonate made may show a reddish tint and the ash obtained become slightly off-colored. The
liquor from the cleaning tower is sent to the making towers by means of CO2 lift or booster pump.
Generally, in cleaning, one tower can best supply carbonated liquor enough to feed 4 making
towers.
In the cleaning tower, it is generally considered that the following reaction takes place,
2NH3 Aq +CO2 (gas ) (NH4 )2 CO3 Aq
This would mean that with 80.7 grams per liter of ammonia, CO2 can be present in the carbonated
liquor up to grams CO2 per liter. In practice such a concentration cannot be reached; and before we
approach this figure, hydrolysis of ammonium carbonated set in so far that the concentration of
HCO3 - has attained such a magnitude that [HCO3 -] [Na+] exceeds the solubility product of
NaHCO3 and sodium bicarbonate separates out. Consequently, with the normal strength of
ammonia in the ammoniated brine, at no time at any point either in the cleaning or making tower ,
is there ever in the liquor such a high concentration of CO2. Generally, crystals begin to how in the
carbonated liquor containing more than 80 grams of CO2 per liter. In actual cleaning operation the
liquor is carbonated up to about 64 grams of CO2 per liter, a figure representing only 61 per cent
of the above concentration. All green liquor is cut off from the making towers and it is sent
through the cleaning tower s fast as possible to make the cleaning most effective. The
precarbonated liquor from this cleaning tower is then fed to all making towers in the group.
In the tower liquor there are present in solution the ionic constituents, Na+, HCO3 -, NH4 +, and
Cl-.There are thus four equilibrium equations, one for each of the four salts, NaHCO3 , NH4 Cl,
NaCl, and NH4 CO3 , expressing the equilibria between the ions of these salts and their
undissociated molecules .Two solubility products for the two least soluble ssalts,NaHCO3 and
NH4 HCO3 , are important. Unfortunately, the exact data on the degree of dissociation (or the
activity coefficient) for each of these four salts when present together at high concentrations in
solution are mot known at present. The solubility of each of these salts when present alone is
given in Table 57.
 These figures show that sodium bicarbonate is the least soluble of the four. Next come
ammonium bicarbonate and then sodium chloride and ammonium chloride at the temperature of
the draw. When, however, the four salts are present together, the salting-out effect of any three
salts in solution upon the fourth generally decreases the solubility of each. Then again the
common-ion effect materially reduces the solubility of the salts concerned, e. g., NaHCO3 by
NaCl and NH4 HCO3 and NH4 HCO3 , and NH4 HCO3 by NH4 Cl.Consequently, NaHCO3 and
NH4 HCO3 are much less soluble in the liquor than when present alone in plain water at the same
temperature. The solubilities of NaHCO3 in NaHCO3 in NaCl solution and in NH4 HCO3 solution
saturated with CO2 is given by Fedotieff as follows:
Table 58 Solubility of NaHCO3 in NaCl Solution Saturated with CO2 .
At 15 At 30
Solubility Solubility
NaCl NaHCO3 NaCl NaHCO3
Sp. Gr. Grams/100 cc. Water Sp. Gr. Sp. Gr. Grams/100 cc. Water Sp. Gr.
1.056 0.00 8.80 … … …
1.063 3.02 6.86 1.066 0.00 11.02
1.073 6.01 5.36 1.079 5.99 7.28
1.096 12.31 3.48 1.100 12.19 4.73
1.127 18.72 2.30 1.127 18.63 3.20
1.158 25.69 1.61 1.156 25.60 2.23
1.203 35.46 1.00 1.199 35.81 1.39
Table 59. Solubility of NaHCO3 in NH4HCO3 Solution Saturated with CO2

Table 59 shows that the molal decrease by NH4 HCO3 on the solubility of NaHCO3 is not so
much as the molal decrease by NaCl.
Fedotieff also gives the solubility of NH4 HCO3 in NH4 Cl solutions saturated with CO2 .
 The solubilities of sodium bicarbonate and ammonium bicarbonate in water in the range of
the tower operating temperatures(between 22 and 32 ) have been determined by the author.
The solubility in each case was determined by taking an excess of C.P. sodium bicarbonate or C.P.
ammonium carbonate in 600cc. water in a tall glass beaker of about one liter capacity, provided
with a glass stirrer. The beaker was immersed in an oil bath containing a refined petroleum oil
which was circulated by a

FIG 35 Curve showing solubility of sodium bicarbonate in water.

pump between the bath and an outside reservoir provided with both electric heating elements and
copper cooling water coils for the purposes of heating and cooling respectively. Carbon dioxide
gas bubbled continuously through the suspension in the beaker through small tubing introduced
through a hole in the cover, which loosely fitted the mouth of the tall beaker, so that the solution of
NaHCO3 or NH4 HCO3 was in constant contact with, and kept under the atmosphere of, the CO2
gas. After the temperature in the beaker had been maintained constant for about 30 minutes with
constant stirring, the suspension was allowed to settle for 15 minutes, after which a 20 cc. sample
of the clear supernatant liquid was drawn into a 250-cc. volumetric flask and the volume made up
to the mark. Ten cc. of the diluted sample was then taken for titration against
 FIG 36 Curve showing solubility of ammonium bicarbonate in water.

a standard N/10 hydrochloric acid solution using methyl orange as the indicator. The results,
expressed in grams per liter of the solution, are given in Table 61(see also Figs 35 and 36). In the
case of ammonium
Table 61. Solubilities of Sodium Bicarbonate and Ammonium Bicarbonate
in Aqueous Solutions.
NaHCO3 NaHCO3
Temp. Grams /1000 cc. Sol. Temp. Grams /1000 cc. Sol.
21.90 94.4 23.05 213.4
24.10 99.1 25.90 230.5
28.00 103.0 28.90 255.8
32.05 108.2 32.10 286.0
bicarbonate the salt in solution was constantly decomposing, especially at the higher temperatures.
So that fine bubbles were evolving from the solution.
In the normal tower liquor after the reaction, there are dissolved in the liquor the quantities of
the four salts shown in Table 62.
Just how much sodium bicarbonate remains dissolved depends upon the concentrations of the
other three salts in the liquor and upon the temperature of the draw. Calculations on the basis of
dissociation constants, solubility products, and general mass action laws do mot give accurate
results in these strong electrolytes at such high concentrations. Bradburn ( Z. angew. chem. .,p
82,1898)gave valuable experimental result results for the conditions of operation and the strength
of ammoniated brine used in his time .They are given in Table 63.

The free NH3 titer is lowered if the temperature of the draw liquor is lowered, i. e. ,if there is
sufficient cooling capacity in the tower ,and if good regulation is maintained .At a draw
temperature of 22 . the free titer in the draw liquor normally is about 20.The free titer in the
draw liquor and its temperature run parallel but will be dependent upon the initial ammonia
concentration in the green liquor employed .Low ammonia liquor may give only 12 to 16 free titer
in the draw liquor. Yet the percent decomposition is low .The amount of CO2 dissolved in the draw
liquor depends upon its free ammonia titer.
As had been remarked above and from a table giving the results obtained by Hempel and
Tedesco, page 130, it may be shown that a very high nH3 titer as compared to Cl- titer, say 157
NH3 titer: 76 Cl- titer, causes considerable ammonium bicarbonate to precipitate in the tower with
the sodium bicarbonate .The utilization of the available ammonia in the ammoniated brine then is
low, i, .e the ratio:
 Is low, and the handing of ammonia is greater per ton of ash made. Consequently ,the loss of
ammonia is necessarily greatly .On the other hand ,if the NH3 titer is low as compared to the CI-
titer, say the ratio of NH3 to is less than 1, the utilization of available sodium chloride in the brine
is low ,i. e the percentage decomposition is poor. Table 64, based on that given by P. P. Fedotieff
(see chapter XII), illustrates this point.
Table 64 Efficiencies of Na and NH3 in Varying Ratios of Cl- to NH3 Titers.

It will be seen from table 64 that there should have been solid sodium chloride in the
ammoniated brine in the forest five samples .The percentage utilization of sodium chloride is thus
decreased when that of ammonia is increased, and vice versa. Only samples Nos.6 and 7 will give
results of any value to actual plant operation.
Table 65 gives average values of the different substances in the normal draw liquor (clear
portion).

It is evident from the Cl- titer in the draw liquor that the liquor volume is here reduced by
about 10 per cent, as compared to the ammoniated brine used.
This draw titer (NH3 titer) depends largely upon the ratio of NH3 titer to Cl- titer in the green
liquor (i.e., the composition of ammoniated brine). When the NH3 titer is low relative to the Cl-
titer, a very low draw liter will be obtained. For example, when the NH3 titer is only 82 and the Cl-
titer 91,a draw of 16 may be obtained. When the NH3 titer in the green liquor is high, a
correspondingly high draw titer will result .For example, when the NH3 titer is 102 and the CI-
titer 89, a draw titer of 30 will show good decomposition .On the other hand, when the Cl- titer in
the green liquor is low (i.e., when brine is too dilute), a high draw titer always ensues. For
example, when the Cl- titer is 80 or 78 and the NH3 titer 90, a draw titer of 40 may be obtained,
and the loss of salt will be great.
Table 66 shows an experimental run in a miniature carbonating tower holding a gallon id
ammoniated brine, employing as high an ammonia concentration as is practically feasible in actual
plant operation .The ammoniated brine (green liquor) had a 107.7 NH3 titer but the Cl-titer
obtained was necessarily lowered to 87.0.As refined salt was used in this experiment, only a
negligible amount of fixed ammonia was present in the resulting ammoniated with practically pure
ammonia gas and little CO2 was present in the ammoniated brine. Hence it should be observed
that the specific graving of this ammoniated brine is lower than what it would be under actual
plant operating conditions. In the course of carbonation in the experimental tower, the specific
gravity increases, reaching a maximum of 1.202 and then decreases as the precipitation of sodium
bicarbonate sets on. The specific gravity readings were determined by a pycnometer and so are
more accurate than are actually needed .The CO2 content in the tower liquor also per liter in the
highly ammoniated green liquor, representing corresponding higher values than would obtain in
plant operation with normal ammoniated brine. This is also indicated by the high free titer in the
draw.
Unless otherwise specified, the figures in Table 66 represent “titer,” i.e., the number of
neither cc. of nor, mal solutions per 20-cc. sample.
After 6 hours of carbonation, equilibrium was reached and the “free titer “ in the draw
reached a constant value. The draw temperature was regulated by the amount of cooling water sent
through the cooling sections in this experimental tower. Hence it fluctuated considerably .The
average of CO2 gases used in the carbonation was 59.6 per cent by volume, representing good
plant practice, where mixed gases are used.
From the result in Table 66 it will be seen that with an ammonia titer as high as 1.24 times as
much as the chlorine titer, keeping the latter at its highest value (87.0 titer), the sodium efficiency
obtained (i. e percentage conversion) was increased only to 76.0 per cent, while the ammonia
efficiency was lowered to 68.7 per cent, and this with the mixed gas at about 60 per cent
concentration and a low draw temperature of 20 to 21 .
These results are plotted in Fig. 37, and the curves show that these items are straight-line
functions of the fixed ammonia titers in the liquor.
The S--content in the draw liquor shows that there is an excess of sulfide in the system. It is
necessary to keep this small excess of sulfide in the tower to ensure the production of white soda
 FIG 37 Curves showing condition in making tower using a highly ammoniated brine.

ash. The soluble sulfide in the liquor reacts with the exposed surface of the apparatus and pipes
forming a black coating of ferrous sulfide, which is not soluble in alkaline solution and so is not
attacked by either sodium chloride or CO2 . the sulfide is present naturally in the crude ammonia
liquor from the gas works, which is the usual source of ammonia supply for alkali plants. In the
case of a deficiency of sulfide, sodium sulfide, preferably in a fused form (60%), can be purchased
and a solution made from it fed to the still with the filter liquor to compensate for the loss of
sulfide from the system. In the towers, some sulfide is oxidized by the air in the CO2 gases, and
some is blown out with the waste gases through the tower washer into the atmosphere. Hence the
necessity of replenishing sulfide in the system to maintain a small excess.
In the cleaning tower, CO2 in the waste gases should be very small (a fraction of 1 per cent);
whereas in the making towers, it runs between 3 and 5 per cent. Sometimes a double gas inlet to
the tower, as mentioned above, is provided so that the rich gas fro the furnace can be admitted to
the bottom of the tower and the best effect of mass action can be obtained. The lean gas is then
admitted to the upper gas inlet a few rings above, where the partial pressure of CO2 in the gas falls
to the same value as in the lean gas, which is to be admitted there.
The rate of production in the tow ers and the success of tower operation depend largely upon
the richness of CO2 gases introduce into the towers. This in turn depends upon the percentage of
CO2 in the kiln gas and in the dryer gas. Efforts to secure as rich a CO2 gas from the limekilns as
possible are well repaid. For every ton of ash there should be delivered to the tower a volume of
800 to 900 cubic meters of the average gas under standard conditions of temperature and pressure.
The CO2 compressors must provide for a piston displacement of 1000 to 1200 cubic meters per
ton of ash:
Between the top of the carbonating towers and the tower washer, there is inserted in the gas
line a cyclone separator that separates the entrained liquor in the outlet gases and lets it drain back
to the top of the towers.
 The capacity of a tower depends largely upon the cooling surface. A 6-foot standard tower,
about 75 feet high, has a capacity to 50 to 60 metric tons of ash per 24 hours. For a 7 1- foot I.D.
2
tower, the capacity is increased to 100-120 tons of soda ash for 24 hours.
The CO2 compressor in the ammonia soda industry requires some special provisions in its
design. In principle it is nothing more or less than a gas pump. But the corrosive character of CO2
and the dust and ammonia, gases are likely to carry, make the duty severe. It should meet the
following specifications:
(1) The inside of the gas cylinder should have lining, which can be re-bored or renewed
when worn.
(2) It should have simple gas valves to insure action because of the corrosive property of
CO2 gas and the presence of some dust in the kiln.
(3) The part in contact with the gas should be all iron owing to the possible presence of some
ammonia in the furnace gas.
(4) Ordinary steel cup valves used in the air compressors for gas outlet are rapidly corroded
by CO2 and simple Stainless steel plate valves or heavy rubber disc valves may be used
instead. Mechanically operated, plunger-seated steel cup valves of special design,
however, have been found very satisfactory. But Stainless steel construction is preferable
and resists corrosion much better.
(5) Ordinary lubricating oil in the gas cylinders is affected by CO2 and becomes sticky. Soda
water of about 10-titer strength injected into the gas cylinders gives clean surfaces and
good lubrication.
(6) Besides jacket cooling, water may be injected into the gas outlet pipe. Soda water in
large volumes with provisions for circulating and cooling will serve for both lubrication
and cooling in the gas cylinders.
Because of low piston speeds (less than 500 f.p.m.) the gas cylinders can tolerate
considerable quantities of soda water without knocking; but it is best to locate the discharge valves
at the bottom of the gas cylinder, i.e.; with top gas intake and bottom discharge.
Between the CO2 compressors and the tower inlet, the gases pass through a separator so that
soda water may not be carried over to the towers. The gas passing from the CO2 compressors to
the columns should be cooled to about 30 . The column inlet pipe is provided with a tall loop
somewhat higher than the tower itself so that liquor from the column cannot run back through the
gas pipe, if for any reason the CO2 compressor is stopped or when the gas delivery fails. The CO2
compressor for the Solvay tower is the biggest power consumer in the soda plant outside of the
main power-generating units.
Lately, some ammonia soda works have installed turbo-compressors for compressing CO2
gases for carbonation. These compact machines have a very large capacity compared to their size,
15,000 to 25,000 cu. ft. of free gas per min at a speed of 10,000 to 12,000 r.p.m. It is clearly seen
that one such machine can do the work of 5 or 6 piston-type low-speed compressor units (each
unit of which may be in itself a duplex machine, having two gas cylinders each delivering gas to a
tower). The use of these turbo-compressors results in a great saving in floor space, and also a
saving in the first cost is not low, and the power consumption is even higher. Its drawback is that
the gas delivery is not positive, so that if one column is partially plugged and offers greater
resistance, little gas is delivered to that column. Also, because of its huge capacity, a large number
of columns must be grouped together and run in parallel. This makes the regulation of gases to
each individual column rather difficult. As for smaller units with a capacity comparable to that of
one unit of the piston-type CO2 compressor, these turbo-type compressors would be economical
neither in first cost nor in maintenance. For large-scale operation, however, these turbo-units offer
great possibilities of application.
In one instance, a four-stage turbo-compressor running at 12,500 r. p. m. at 40 lbs. Pressure
gage and delivering CO2 gas to a single making tower (1300 cu. ft. per min. piston displacement)
requires more than twice as much steam per I.H.P as would a piston-type compressor for the same
service, but the first cost is about the same. At twice this capacity (2600 cu. ft. per min piston
displacement for two making towers), the steam consumption per I.H.P is still about 80 per cent
higher, while the first cost is cut to a half. At a still higher capacity, however, where one
turbo-compressor serves eight making towers, the steam consumption begins to compare
favorably with that of the piston-type low-speed compressors, and the first cost is then reduced to
a small fraction. This applies to turbo-compressors driven by either single-stage non-condensing
turbines or multi-stage bleeder turbines.
At present, the Elliott Company, Ridgway, Pa; the Ingersoll-Rand Company, New York, N, Y;
and Allis-Chalmers Mfg. Co; Milwaukee, Wis., are leading in the manufacture of turbo-type
compressors in the United States.
The speed of the piston type CO2 compressors is taken at about 60 to 80 r. p. m ; because it is
believed that, at this speed, the pulsation effect of the gas delivered to the towers will help keep
them open, i.e.; the bicarbonate in the towers is not so likely to form incrustations tending to plug
them up. But this is just a belief, which has grown up in this industry, and lacks scientific
background.
 Chapter IX

Working of Carbonating Towers or “Columns”

The carbonating towers or “columns” of Solvay design which have about 35 gas-distributing
mushrooms, stacked one above the other, separated by plain rings or cooling tube sections, as the
case may be, are efficient in gas absorption, but they are very delicate pieces of apparatus when a
slurry is involved. In the first place, under no circumstances mush the flow of gas entering the
tower is interrupted while the tower is full of sodium bicarbonate so as to permit the suspended
bicarbonate to settle in the tower. The surging (or pulsation) effect of the carbon dioxide gas
delivered from the carbon dioxide compressors is believed to be beneficial in keeping the
bicarbonate in the columns in suspension. The flow of gas to the towers will be interrupted either
by the failure of the carbon dioxide compressor or, what is more likely to happen and what
constitutes a treacherous affair for even an experienced operator, by the failure of one or more of
the gas delivery valves to seat properly, so that little or no gas is being delivered to the tower, even
though the compressor may be running at its normal speed. Such a case may be diagnosed by the
surging up and down of the water levels in the U-shaped manometer located on the suction main.
If this condition is allowed to go on undetected for any length of time, the effect will soon begin to
show on the tower. Large quantities of sodium bicarbonate crystals will settle in the tower, of
which little can be drawn out through the draw line. The only remedy is to introduce steam into
the tower bottom. The steam can be introduced best through the steam connections at the sides and
through the draw opening in the bottom section. The steam dissolves the bicarbonate, reversing
the process according to the following reaction:
NaHCO2 +NH4 C1 NaC1+NH2 +CO2 +H2 O
and opens its way through the blockade. The introduction of steam must be carefully carried out,
or the bicarbonate will show a reddish tint. No more steam should be passed in than is necessary
to open up the tower, and the duration of steaming should be as short as possible.
In the second place, even when a tower has been operating smoothly, it needs frequent
cleaning. The cooling tubes become coated with precipitated bicarbonate after three or for days’
running (“making ”) and the tower is then said to be getting “dirty.” This “dirty” condition is
shown by the increase in the percentage of carbon dioxide in the exit gas and in the rise of the
draw temperature, since the cooling water is unable to absorb the heat generated in the tower by
the chemical reaction because of insulation of the cooling tubes by the bicarbonate scale. The
resistance to gas passage in the tower may have been increased, but in point of fact it is more often
attended by a fall, rather than a rise, in the gas pressure through the column. This can be
understood when we realize that the space in the tower is decreased by the deposition of large
amounts of solid bicarbonate in each section and that the net height of the static column of the
liquor is reduced. Such has actually been found to be the case. Only in case of obstruction of the
tower bottom by the settling of sodium bicarbonate does the gas pressure increase in the curse of
operation. The rise in draw temperature is, however, always an infallible indication that the tower
needs to be cleaned. In the worst case, the magma at the draw comes out , not in a smooth,
continuous stream, but in broken jets, and considerable carbon dioxide gas foam appears in the
draw liquor. This shows that the tower is badly plugged and that a channel has been cut in the
obstruction in the bottom ring where the draw opening is located, through which the gas finds its
way out with the magma. Such a case requires immediate attention.
In a carbonating tower possessing a large number of cooling tubes (hence a large cooling
capacity but a limited liquor space), if the exit gas rises excessively on account of a serous
shortage of green liquor (or ammoniated brine) or because of the decrease in the rate of draw
due to partial plugging of the pipe or of the draw opening, large amounts of bicarbonate
accumulate in the tower not only at the bottom but also in the upper cooling sections. If these
conditions are allowed to continue, bicarbonate may completely plug the column and the draw
stops by itself. When such conditions occur, the best remedy is to shoot in steam momentarily
through the draw pipe into the tower bottom. If after repeated application of steam, the tower
refuses to be opened up, the only safe way is to turn it to cleaning immediately. This is done by
closing off all cooling water, changing over to “lean gas” with gradual bleeding out of the gas, and
drawing out the bicarbonate with or without the help of steam at the bottom of the column. In one
case, because of shortage of green liquor, the exit gas in one making tower rose as high as 45 per
cent carbon dioxide, the inlet gas pressure fell from 38 to 30 lbs. Per sq. in; the upper tower
temperature rose from 55 to 67 , and the draw became so sluggish that it stopped by itself. The
draw liquor became foamy; and after each application of steam, the draw came out in a
discontinuous stream and gradually stopped of it self. The ultimate salvation was to turn the tower
to cleaning at once. Hence the simultaneous occurrence of an excessive rise in the percentage of
carbon dioxide in the exit gas, of a drop in the inlet gas pressure, of a rise of temperature in the
upper portion of the tower, and of the appearance of foamy and everdecreasing volumes of draw
liquor indicates that the tower is being badly plugged and that the condition demands immediate
attention.
The process of cleaning consists inputting the tower on the lean gas, cutting out the cooling
water entirely, bleeding out a portion of the lean gas, and at the same time drawing the tower hard,
until the bicarbonate in the magma is about 8 per cent or less by volume. The normal percentage
of bicarbonate by volume after 3 minutes’ settling is about 40 to 45 per cent. All the green liquor
then is fed to the tower, from the bottom of which this cleaning (precarbonated) liquor is sent to
feed the other (“making”) towers by means of a bocster pump, or better, by means of a carbon
dioxide gas lift. The cleaning tower is kept warm (approximately 40 ) by shutting off the cooling
water to help dissolve the deposited bicarbonate therein. By bleeding out the desired portion of the
lean gas, the liquor in the cleaning tower is carbonated to just below the precipitating point (65
grams of carbon dioxide per liter). The process of cleaning lasts from 18 to 24 hours. At the end of
the cleaning period, the normal process of “making” is resumed, rich gas being turned in and the
cooling water admitted about 15 to 20 minutes later. About 5 minutes after the cooling water has
been admitted, the tower can be drawn, at first lightly, then in gradually increasing amounts until a
normal stream is obtained. The following is an actual sequence of operation when a tower was
changed from cleaning to making:
At 2:40 the draw had been increased to the normal stream. Thus, it takes tow to three hors after
cooling water has been admitted for the draw to reach its normal temperature. At the start the
temperature above the cooling section in the upper portions of the column is 33 to 34 and the
exit gas contains only 0.8 to 1.0 per cent carbon dioxide. The tower being cold at the start, the
grain of the bicarbonate crystals from the tower is necessarily irregular, and the bicarbonate from
the filter is wet and difficult to dry. Such bicarbonate needs greater amounts of returned ash for
drying. But this does not affect dryer operation, when normal crystals from the other (making)
towers are mixed with such bicarbonate in drying. The crystals improve when the tower reaction
temperature increases to above 50 at about the tow-thirds point. When the operation is finally
normal, this temperature will be about 60 .
Experience has shown that a column must be put on cleaning after it has been making for
four days (96 hours). This means that with a group of five towers, one tower must be “cleaning”
every day, and the cleaning is done by rotation. With les than four towers, as in very small plants,
an equivalent length of time must be adopted for the cleaning period, but the cleaning is less
effective because of the small volume of green liquor that can be put through the cleaning tower to
feed the making towers. Under these circumstances, it may be necessary to boil the contents of the
tower with water and steam since every four to six months. The tower then is first emptied of the
bicarbonate and the liquor. It is then filled with water about 3 ./4 full, and steam is introduced at the
middle rings as well as the bottom ring of the tower at several points. A small quantity of lean gas
is admitted to help agitation, and the circulating pump or gas lift is used to circulate the water in
the tower. The temperature of the water in the tower should not be increased above 80 ; as a
higher temperature is neither desirable nor necessary. With tow or three changes of water in the
course of steaming, which lasts for 36 to 48 hours altogether, the entire content will have been
practically all dissolved out. All bicarbonates are dissolved, what is left being small pieces of iron
scale and the incrustation of magnesium carbonate. Scale in the lower portions of the tower
normally consists exclusively of sodium out through manholes or flushed down to the bottom of
the tower. it is column, they will form nuclei for scale formation in the column, blocking the draw
pipe. For this purpose, as well as for inspection, the minimum number of columns in a plane must
be three. This defines the smallest economic unit possible for an ammonia soda plant, than will
work efficiently must have a minimum daily output of 100 to 150 tons.
With a group of five columns as a working unit and with the green liquor reasonably free
from magnesium, the drastic be necessary and the usual method of cleaning by running through
warm produce on the average 200 tons of soda ash a day. Larger plants have more than one group
of five each.
Historically, in the Solvay Process plant, Syracuse, N. Y., the first carbonating towers were 2
meters in diameter and had flat plates with 9-inch holes in the center. These towers were showered
with water outside but had no internal cooling. They had a capacity of ten tons of soda ash each
per day. Mr. W. B. Cogswell suggested internal cooling, and nozzles were bolted on several of the
rings, through which passed 4-inch tubes carrying cooling water. The capacity was then increased
to 20 tons each.
When the second set of towers was built, the lower 8 rings in each tower were provided with
2-inch cast-iron tubes held in the tube sheets were made and the center hoses were increased from
9 inches to 20 inches in diameter. Those center holes were also provided with to subdivide and
distribute the gas bubbles in the column.
No steaming or heating of the towers to dissolve the scale from the running green liquor
through them, as mentioned above. These columns were always five in a group, one “cleaning”
and for “making” each that in this way each “making “column of 2-meter diameter averaged 60
short tons of soda ash a day.

The tower draw pipes are also subject to scale formation. Crusts of sodium bicarbonate are
formed around the internal walls of the pipes. By a gradual building up of these scales to a
thickness of 1 to 2 inches, the pipe opening is gradually reduced, restricting Goth passage of the
raw magma. Frequently these pipes have to e taken down and replaced by new sections of C.I.
pipes. Scale best leaned out by heating the pipes externally over a small fire or blowing steam
through them. Analyses of the scales in the upper and lower portions of a column, respectively, are
given it Table 68 and 69.
Thus it will be seen that the scale in the lower portions of the column, on the internal walls,
and around the cooling tubes is practically all sodium bicarbonate, but that that of the upper
portion consists also of magnesium carbonate and iron scale.
Magnesium in the tower comes from the mud in the ammoniated brine, which has not been
completely settled. It is deposited in the upper portions of the tower where the green liquor first
strikes the carbon dioxide gas. This is a serious matter. It may be the result of not having sufficient
capacity in the settling vats for the quantity of the brine fowl, or else it may be due to excessive
amounts of magnesium in the brine. As boiling-out is the most drastic means of cleaning a column,
magnesium scale, being difficultly soluble in wart, must be removed by hand after the boiling-at
operation. As not all rings are provided with ample manhole area. And not all manholes a re so
located as to reach every cooling tube or every corner of the column sections, blocking caused by
magnesium scale in the upper cooling sections is detrimental to successful operation of the column.
Magnesium mud (MgCO3 ) in the ammoniated brine is very much more difficult to settle out than
calcium mud (CaCO3 ) as has been mentioned many times elsewhere. The fact that considerable
magnesium and no calcium is found in the column scale bears out the
truth of this statement. With an effective depth of not less than 15 feet in the vats it is found that a
minimum of 6 sq ft settling area is necessary per gallon (U.S) of liquor flow per minute for an
unprotected sea brine in order to bring the total mud content in the clarified ammoniated brine to
below 0.1 gram per liter.
Following each boiling, an off-color ash is certain to be produced, or an ash yielding a cloudy
solution containing considerable insoluble matter. Hence boiling-out or steaming of a column is
not to be recommended unless absolutely necessary.
For a good temperature gradient in the tower, the cooling water coming out of the uppermost
cooling section of the tower should be quite warm and have a temperature of about 45 There
should not be a sudden chilling of the liquor upon striking the first bank of cooling tubes as it
descends from the top. Cooling water entering the bottom section should be as cold as possible but
it must be able to take up heat gradually and become warmed to about 45 at the top outlet ( I, e.,
from the uppermost cooling section of the tower ). The maximum temperature in the tower occurs
at about tow-tires of the eight from the bottom and is in the neighborhood of 60 , as stated
above.
Table 70 gives the temperature observations in a carbonating tower in operation.
In order that the outlet water leaving the tower may be warm (45 ), if we start with a very
low temperature of the cooling water (16 to 161 /2 ) at the bottom, it is evident that there must be
a large cooling surface and a long course of water flow. This is secured by having multiple
cooling-water passages in the tower cooling sections and by having a number of these sections,
one on top of the other, with a gradually decreasing number of cooling tubes toward the top, the
tower’s overall height Under these conditions, the lower the temperature of the cooling water, the
lower will be the draw temperature, and the lower the free ammonia titer in the draw.
Consequently, the higher will be the decomposition, with excellent bicarbonate crystals, provided
the gas and liquor concentrations are up to normal.
 Good bicarbonate crystals should feel gritty or granular between the thumb and the finger,
and the layer of bicarbonate on the vacuum drum filter should be about 1/2 inch thick at a vacuum
of 8 to 10 inches Hg. The free ammonia titer in the draw runs parallel to the draw temperature in
the normal rate of draw. For instance, with normal ammoniated brine, when the draw temperature
varies from 22 to 40 . (a very wide range of temperature change )the free titer in the draw
follows closely or may be one or two points behind the temperature. This shows how important
the draw temperature is and how essential it is to have an ample cooling surface and a cold
underground water supply for the cooling medium. Whit the above cooling arrangement
maintaining a high reaction temperature at the two-thirds point in the column, there can be no
lower point for the draw temperature, the limit of the draw temperature being the ability of the
cooling water to cool and the heat conductivity of the cooling tubes. Using well water having an
average temperature of 16 to 16.5 , it is impossible to cool the draw liquor below 20 .This at
present is the draw temperature limit, unless the cooling water is refrigerated.
The author has actually known of a case where, in the course of years of operation, free
sulfur was deposited in the top rings of a column, obstructing the gas passage. This deposit was
undoubtedly formed from the oxidation of hydrogen sulfide by the oxygen in the carbon dioxide
gas. Thus,
2H2 S + O2 → 2H2 O + S2
Such an excess of oxygen comes from the leakage of air into the furnace gas, which is to a certain
extent unavoidable. It has been occasionally observed in the gage glass at the top of the column
that the liquor contains suspended free sulfur in the form of a yellowish-white, fine emulsion. The
analysis of the free sulfur deposited at the top ring of the column is given in Table 71.
Table 71. Analysis of Air-Dried Sample of Deposited Sulfur at the Top of the Column.
Per Cent
Free sulfur 59.50
NaCl 13.51
NaHCO3 24.11
Na2 CO3 1.14
NH3 Trace
Moisture by diff. 1.74
Evidently considerable ammonia had disappeared here and the small amount of sodium
carbonate present came form the loss of carbon dioxide in the sodium bicarbonate on exposure to
the air.
The capacity of a carbonating tower, like the boiler rating, depends much upon how it is
operated. With a given amount of cooling surface (which determines the capacity to a large extent)
the quantity of production from a tower depends upon the richness of the CO2 gases pumped to the
tower (also its volume-rate) and the concentration of salt and ammonia in the ammoniated brine
fed to the tower, i.e., the Cl-and free NH3 titers. A large tower may have very low production when
working on a dilute mixed gas or on ammoniated brine too dilute in chlorine or ammonia.
Furthermore, the crystals obtained from columns under such conditions, i.e., with low chlorine
titer (weak brine) or with weak mixed gases, or both, are very fine and the soda ash obtained has
an abnormally low density (lighter than normal light ash). The one most important factor that
determines the capacity of a column is the temperature of the cooling water available.
 Chapter X

Filtration of Crude Sodium Bicarbonate (Ammonia Soda):

Composition of the Bicarbonate
The crude sodium bicarbonate drawn from the columns is filtered off from the mother liquor.
The older type of filter consisted of an open wooden tank with a false bottom, over which was laid
a filtering fabric. Mother liquor was sucked from the compartment below the false bottom by
means of a piston pump. This inefficient batch filter is now obsolete.
At present, the rotary open-drum filter and the centrifugal filter are the two standard
machines: The drum filter has a pope through the journal at one end, where suction is applied, and
a smaller pipe through the journal at the same end or at the other end, where compressed air or
waste carbon dioxide gas under pressure from the column exit is introduced. The filtering medium
is a woolen felt or blanket wrapped on the drum.
The vacuum helps dry the bicarbonate on the filter and sucks the ammonia gas inward so that
ammonia loss to the atmosphere is reduced.
The woolen fabric has a tendency to become plugged by crystals of the bicarbonate, and
frequent blowing out with compressed air or carbon dioxide gas is necessary to open the pores. A
self-blowing drum filter has one compartment that revolves just past the scraping knife or “doctor”;
this compartment is about to enter the liquor, blowing under pressure, while all other
compartments are under vacuum. An internal sliding pressure valve connects the opening to this
compartment with the pressure pipe, so that each compartment is blown once during each
revolution. In this way the drum can pick out bicarbonate continuously without stopping for
blowing or cleaning. Water in a thin film from a weir is allowed to play on the surface of
bicarbonate to wash the salt (NaC1) from it. The weir box is provided with a standpipe, at the
bottom of which is an orifice plate. The standpipe is provided with a gage glass to show the level
of water above the orifice plate, feeding the weir box below. Generally, a cold, weak ammonia
water from the cooling tower is used for this purpose. Between the weir box and the scraper, or
doctor, is provided a number of press rolls made of heavy cast iron (each having about 300 1bs.
pressure per foot of length) resting on the surface of the bicarbonate layer, the pressure on the
bicarbonate being adjustable. Each press roll serves to level the surface of, and fill any channels in,
the bicarbonate layer, and to squeeze out water by its pressure so as to give drier bicarbonate for
franking. Because of this crushing action, too, the bicarbonate leaves in a looser form, free from
solid cakes. Each press roll revolves at exactly the same peripheral speed as the peripheral speed
of the bicarbonate on the drum (making allowance for the thickness of the bic arbonate layer), It is
independently driven from the same drive of the filter-not merely caused to revolve by the rolling
contact with the face of the drum. These pressure rolls, however, have the disadvantage of causing
more rapid wearing out of the felt, especially when the edges of the slots on the cylindrical
segments of the cage below are not properly rounded off and smoothed. When the bicarbonate
crystals are good, the yield ( i. e., the weight of sodium carbonate plus sodium chloride obtained
per 100 grams of the filtered and washed sodium bicarbonate) should be about 54 grams (i.e., with
free moisture about 14 per cent).
The filter drum is a squirrel-cage affair, having slotted segments or cross bars with slots
between them. The ends of the drum are closed with covers. It should be all cast iron. Sometimes
the cage is covered with cylindrical segments of cast iron. Sometimes the cage is covered with
cylindrical segments of cast iron or steel plates having a large number of oblique slots. On the face
of the slotted drum is wrapped layer of coarse backing nickel wire cloth or screen, which supports
a fabric wrapped over it for the filtering medium. This fabric is in the form of a woolen felt or
blanket, about inch thick, and is of a loose but uniform texture. Neither the nickel wire cloth nor
the woolen blanket stands the corrosion of the mother liquor very well. The life of a woolen
blanket for this purpose is about three weeks of continuous service. Ammonia seems to have
considerable action on the animal fibers. Yet vegetable fibers like cotton or hemp, which stand the
alkaline action of ammonia much better, do not serve the purpose well, because they do not
possess the even porosity of woolen felt, and they swell up unevenly and become tight when
soaked in the liquor. Consequently, instead of a uniform layer of bicarbonate, there would be
uneven spots overt the larger pores in the cotton or hemp fabric. Through these pores considerable
bicarbonate may pass into the filter liquor and be lost. In place of the nickel wire cloth, finely
fabricated bamboo matting made of small bamboo sticks serves the purpose very well. It lasts
longer, and the costly nickel wire cloth can be successfully replaced by the inexpensive, inert
bamboo material.
It may not be out of place to dwell somewhat on the early history of the filters introduced
into the ammonia soda industry. In the early days of the Solvay Process plant at Syracuse, N. Y.,
long rectangular filter tanks were used; some of these were 60 feet long by 30 inches wide, and
were provided with woven flannel cloths supported on grates at the bottom. In the false bottom
below, mother liquor was drawn off by vacuum and the bicarbonate was shoveled out by hand into
bottom-dump cars standing alongside. The crystals were washed by means of a swinging
perforated pipe. This involved a great deal of labor.
Mr. Ernest Solvay, who was visiting the Syracuse works, suggested to the engineer, Mr.
Edward N. Trump, that he devise “a sort of a revolving perforated cylinder with vacuum inside,”
and the result was the first rotary vacuum filter built by Mr. Trump for the Syracuse plant. Six of
these rotary units replaced all the long, rectangular filter tanks. At first, the scraper knife tended to
“glaze” the surface of the blanket so that vacuum had to be taken off every 15 minutes and the
blanket washed. The machine was anything but continuous. To avoid “glazing” of the crystals on
the blanket, the scraper knife was then attached to the agitator shaft which gave it an oscillating
motion, so that knife would not bear on the same spot during each revolution and form a
corrugated surface on the bicarbonate. Later, compressed air was used for blowing back and back
washing. This was a decided improvement, but the filter still had to be stopped for back washing.
A big stride was made when it was found possible to do away with this intermittent operation in
back washing and introduce an automatic self-blowing device. A continuous rotary vacuum filter
was developed. The inside of the filter drum was divided into compartments around the
circumference, and compressed air was introduced by a valve closing the commutator leading to
that particular compartment having just passed over the scraper knife and blowing it when that
compartment just entered the liquor again, thus enabling it to pick up a fresh layer of bicarbonate.
Then it emerged from the liquor at the back where the cake was washed by a film of water over a
weir. The next development was the addition of heavy press-rolls, driven independently at a speed
somewhat faster than the peripheral speed of the filter drum itself, to level the bicarbonate layer
and squeeze the moisture out of it after it had been washed and before it was scraped off.
 The filter thus developed was soon adopted by European Solvay plants as one of the
standard pieces of equipment. From a filter one meter wide and two meters in diameter, it was
found possible to obtain enough ammonia soda (crude bicarbonate) to make 120 tons of soda ash a
day, using a 10-inch mercury vacuum. Filters were then used also for other service. For ammonia
soda service, flannel cloth or felt was used; but for filtering caustic mud, a close-woven
rectangular-mesh nickel wire cloth (20×120 mesh or finer) was used in place of the woolen
blanket. Narrow grooves supported by internal flanges, or circumferential grooves supported by
cross ribs, were milled in the cylinder, and the flannel cloth, or nickel wire cloth, as the case might
be, was held on the drum by winding it with nickel wire.
These continuous rotary vacuum filters could give bicarbonate cake containing 12 per cent
moisture when good crystals were obtained from the columns. Vertical-suspended,
bottom-discharge batch-type centrifugals had been tried in place of these filters, but it was found
that the capacity was low, the labor required high, and, above all, the loss of ammonia excessive.
A combination of the continuous filter and centrifuge had been successfully employed. While it
did not cause much ammonia loss and did yield a drier bic arbonate, the labor involved was
considerable, and the somewhat better yield in the ammonia soda obtained offered no particular
advantage in the rotary dryer operation as conducted in present practice.
Continuous rotary vacuum filters are now made in units as large as three meters in diameter
and two or more meters long.
The salt content can be reduced but it cannot be entirely eliminated. The “yield” of the
bicarbonate (which measures the dryness of the bicarbonate from the filters) is very essential to
the successful operation of the soda dryers. Table 72 shows a control analysis of the normal crude
bicarbonate from the filter.

TABLE 72 Result of Control Analysis of Filter Bicarbonate.

Per Cent

(a) Yield 53.52

(b) NaC1 0.36

(c) NH2 (total) 0.92

A complete laboratory analysis on a typical sample is given in Table 73.

TABLE 73. Analysis of Crude Bicarbonate from the Filters,

Per Cent

NaHCO3 75.60

Na2 CO3 6.94

NH4 HCO3 3.42

NaC1 0.39

CaCO3 Nil
 MgCO3 Trace

Water by diff 13.65

With the introduction of wash water, the mother liquor is diluted by about 8-10 per cent, i.e.,
from 6 cubic meters of the liquor per ton of ash the total volume becomes 6.5-6.7 cubic meters.

FIG 38 Multiple passette filter washer.

This is found by the determination of the chlorine concentrations, Cl titers, in the liquors entering
and leaving the filters. When bicarbonate crystals are poor, more water is consumed in washing
out the sodium chloride to the allowable limit, and then higher vacuum is required. The loss of
bicarbonate through solution by the wash water is estimated normally at 2-4 percent.
The filter wash water should be a good, soft water. Any hardness in the water will be thrown
down as calcium and magnesium carbonates or magnesium hydroxide, plugging the filter blanket.
This is a serious matter because it would soon throw the filter out commission. Generally, one
reserve filter is standing by, running empty in water to clean the blanket. Each filter in the battery
can be cleaned in this way by rotation. The bicarbonate should not appear dead white but should
have a slightly bluish tone, showing a slight excess of sulfide. It should be light and loose, feel
like sand between the fingers, and should not produce slimy juice when continually squeezed by
forced kneading in the palm of the hand. The bulk specific gravity of the bicarbonate varies with
the moisture content and with the crystalline character, i.e., whether it is grainy or pasty. Normally,
it should weigh 38 to 42 (average 40) pounds per cubic foot when loosely packed. The vacuum for
drum filter is 8 to 10 inches Hg at the filter. Coarser crystals of bicarbonate require fewer vacuums
than the pasty kind, and the filter takes on a heavier layer of bicarbonate on the blanket-as thick as
inch. The coarse crystals also wash better, yielding lower sodium chloride content and drier
sodium bicarbonate.
The gases from the filter are drawn through a small washer called the filter washer (Fig 38),
which is similar to the tower washer, but smaller in size and with fewer passettes. The ammonia
present in the gases is scrubbed in this filter washer with fresh brine and the brine sent to the tower
washer, thence to the absorber washer, and finally to the absorber for making ammoniated brine.
The gases are sucked through by means of a vacuum pump, or exhauster, and the exhaust goes to
the atmosphere. The volume displacement of the filter exhauster is 120-200 cubic meters per ton
of soda ash produced, depending on the crystalline character of the bicarbonate and the size of the
filter unit.
The drum filter should have a peripheral speed of 20 to 40 feet per minute. If it runs much
faster, the bicarbonate may not filter dry. Too slow a speed, of course, gives a slightly drier
product but cuts down the capacity. A self-blowing filter, one meter wide and two meters in
diameter, today has capacity of 150-200 tons of ash per 24 hours working on good bicarbonate. A
small filter, 30 inches wide and 4 feet in diameter, has a capacity of 800-100 tons of ash per 24
hours. The capacity of the filter is largely dependent upon the crystalline character of the
bicarbonate. A filter turning out 80 tons of ash on coarse-grained bicarbonate may work with great
difficulty on pasty bicarbonate at the rate of only 45 tons per 24 hours.
The litter operation is important because a difference of a difference of a few percent in the
moisture content, measured as the “yield” of the bicarbonate, makes a great difference in the
behavior of the bicarbonate in the drying operation.
The composition of crude sodium bicarbonate has been carefully studied and the results of
analyses are given in Table 74.
 From the foregoing results it is proved beyond doubt that the bicarbonate actually contains
normal sodium carbonate (Na2 CO3), which is present to the extent of 7 per cent. When the acidic
and basic constituents from the analysis are combined to form various salts, it is found that there is
a surplus of Na+ unaccounted for and at the same time also a corresponding surplus of CO3 --, after
all possible acidic radicals have been combined with Na+ and the small amounts of NH4 + and Mg++
present. Independent checks from the “yields” of the bicarbonate and from the losses on ignition
both give concordant results. So it can be concluded that the normal carbonate is present in the
sample of the crude bicarbonate from the filters. Study of the mother liquor from the towers shows
that this sodium carbonate is largely formed by the rapid loss of carbon dioxide on exposure to the
air (see Chapter XI). Further, if the fresh solution of the crude bicarbonate is immediately taken
from the filter without exposure to the air, it gives no coloration to phenolphthalein at first.
However, the red color develops gradually but increases rapidly on standing i.e., on exposure to
the air.
Some ammonium bicarbonate is also found to be present to the extent of about 5 per cent
NH4 HCO3 on the weight of dry NaHCO3 in the crude bicarbonate from the filters. The presence of
NH4 HCO3 is due to an entirely different source, however. No matter how thoroughly the crude
bicarbonate on the filter has been washed, so NH4 HCO3 is always found in fact in a definite
percentage of ammonium bicarbonate on the weight of dry sodium bicarbonate obtained. It would
thus be seen that NH4 HCO3 and NaHCO3 have separated in the columns in the form of a solid
solution with a definite ratio between them. This molar ratio is 1 mol NH4 HCO3 per 19 mols
NaHCO3 , which gives approximately 5.0 per cent NH4 HCO3 .
The size of the crystals of the bicarbonate formed in the towers has an important bearing on
the drying properties of the bicarbonate in the dryers. When the bicarbonate crystals are coarser,
they filter better, wash better, and come drier, i. e., the moisture content is lower and the “yield” in
Na2 HCO3 higher. Drier bicarbonate always means fewer tendencies to scale formation in the dryer,
less returned ash required, larger drying capacity for the dryer, and less coal consumption in the
furnace per ton of soda ash calcined. When the bicarbonate crystals are coarser, a thicker layer of
bicarbonate (as much as inch thick) can be picked up on the filter, even though a smaller vacuum
is required. Poor crystals, formed from cold towers, for instance, require a high vacuum for
filtering (as much as 12 to 18 inches Hg); yet only a thin layer can be picked up, yielding a wet
bicarbonate which is difficult to dry and which contains high sodium chloride. This cuts down the
capacity of the filter as well as that of the dryer. Bicarbonate that will not produce any slimy juice
on long squeezing and forced kneading in the palm of the hand will dry without difficulty. This is
a rough and ready test for the quality of the bicarbonate. Bicarbonate having a “yield” of more
than 47.5 per cent will not produce any slimy juice in this way, nor stick together to form a ball in
the hand. Good bicarbonate crystals should feel granular, like fine sand between fingers, while
poor crystals will feel sticky, like a starch paste. Good, coarse crystals will settle to a compact
sludge in a 100-cc. cylinder in less than three minutes, while poor ones will require a much longer
time (7 to 10 minutes), and they will not settle to a compact sludge but to a voluminous, mobile,
slimy paste. Too high a draw temperature may also produce this condition because of partial
re-solution of the bicarbonate crystals in the liquor. Under normal conditions, the volume of the
completely settled bicarbonate at the bottom of the glass cylinder should be 40 to 45 per cent of
the total volume of the tower draw sample. But when the crystals are poor, the precipitate is less
compact and may occupy a volume of 50-55 per cent. The percentage of suspended solids
(NaHCO3 crystals, etc.) in the slurry from the columns generally runs 20.5-21.0 per cent by
weight.
As mentioned above, filtration of crude bicarbonate by means of continuous, rotary vacuum
filters has been most generally employed in ammonia soda plants. This is essentially a
constant-pressure filtration, in which the pressure differential between the atmospheric pressure
outside the drum and the partial vacuum inside furnishes the driving force. In ammonia soda
plants, the rate of draw from the columns to the filters is maintained constant, and so is the output
from each filter. Due to the care exercised to maintain the crystal size, the thickness of the cake on
the filter and vacuum required are practically constant day in and day out. To understand better the
principles underlying filtration, we may pause for a few moments to dwell briefly on some
theoretical discussion. Not unlike a filter press, the cake on the filter has to be built up under a
pressure differential to a desired thickness from solids caught on the surface of the filtering
medium when the liquid is forced through. It is thus seen that the resistance through the cake
increases proportionally as the thickness of the cake that is being built up, other conditions being
equal, and so also increases proportionally as the volume of the filtrate that is being collected. At
the same time it is also easily seen that this resistance varies inversely as the filter area. The
filtering medium in the form of a woolen felt possesses numerous pores through which liquor
passes, closely approximating the condition of flow called for in Poiseuille’s law, i.e., the pressure
drop across the filtering medium is directly proportional to the viscosity of the liquid, the velocity
of the flow, and the length of the path; and it is inversely proportional to the cross-sectional area.
From the slowness of infiltration of the liquor through the pores of cake and of the blanket, it is
safe to assume that the flow within these pores would be “viscous” or “streamline”, and we may
expect Poiseuille’s law to hold. Under these conditions, we may expect the rate of filtration to be:

Where V=volume of the filtrate in cu. ft. passing through the filter;
è=tine of filtration in sec.;
A=area of filter in sq. ft.;
P=pressure differential in 1b. per sq. ft.;
ì= viscosity of the liquor in poises;
R=characteristic constant depending on resistance of the cake, etc.
For a given filter operating on a given liquor at a substantially constant-pressure differential, is a
constant (=K’), i.e.,

where íc is an integration constant having the significance of an initial volume of filtrate building
up an imaginary thickness of a cake having the same resistance as the filtering medium itself, and
èc a constant establishing the starting time from which this imaginary cake may be said to have
been formed. The cure showing the relationship between the volume of the filtrate passing through
(í) and the time of building up the cake (è) is thus a parabola, and is true of a constant-pressure
filtration operation, such as in the filtration of the crude bicarbonate from the columns, where the
solids are highly crystalline and the cake is substantially non-compressible under the pressure
differential employed.
 As mentioned above, a centrifugal filter is sometimes used as a final filter, which brings the
moisture content in the bicarbonate down still further to about 58 per cent of the yield, with the
moisture content less than 7 per cent. Sometimes the moisture content in the bicarbonate from the
centrifuge is claimed to be as low as 4 per cent, i.e., the yield as high as 60 per cent. Under such
conditions little returned ash would be required in drying the bicarbonate in the rotary dryers. The
centrifugal filter is, however, an intermittent machine operating in batches with a bottom discharge
for the vertical type, and side discharge for the horizontal type. It revolves at from 1000 to 1800 r.
p. m. A batch-type machine with a 48-inch basket will produce 30-35 tons of bicarbonate per 24
hours. The intermittency of the batch and the difficulty of keeping down the loss of ammonia
caused a serious a serious problem in the adaptation of the centrifugal to soda-ash manufacture.
However, the use of centrifugals for filtering is more general in connection with the manufacture
of refined sodium bicarbonate (see Chapter XVIII). A continuous centrifuge is now offered in the
market. It greatly cuts down the labor of handling and increases the output per unit with a saving
in floor space.

Fig 39. Continuous centrifuge. ( screen type).

*Ruth, B. F., Studies in Filtration, III. Derivation of General Filtration Equations, Ind. Eng. Chem.,
27,708(1935) ; Studies in Filtration, IV. Nature of Flow through Filter Septa and Its Importance in the
Filtration Equation, Ibid., 27, 806 (1935) .
Continuous centrifuges may be classified into two types. One is the automatic type, which is
continuous as regards the different operations in the cycle, but intermittent in feeding and
unloading. It is really an automatic intermittent machine. The other type is a strictly continuous
one, both taking the feed and discharging the cake continuously. To avoid confusion we shall refer
to the former as the “automatic” type and the latter as the “continuous” type. The automatic
machines are generally horizontal and have a speed of from 500 r. p.m. in the large machines to
1800 r. p .m. in the small ones (which a basket of small diameter). They may be single or double
(i.e. single basket, or double baskets placed back to back).
The continuous type, i., e., one having the slurry fed continuously at one end of the basket
and the dried crystals unloaded at the other end, is, generally speaking, still in the process of
development as far as the filtration of crude bicarbonate (ammonia soda) is concerned, and has so
far been confronted with several mechanical difficulties. Some of these may be mentioned here.
For very fine crystals (such as NaHCO3 ), this continuous type requires a very large number of
extremely fine slits in the basket (such as 0.004 to 0.006 inch wide by 4 to 1 inch long), which is
rather difficult and expensive to make. The screen is either cut from a solid plate or made up of a
large number of spec ial-shaped bars of uniform cross-section. Even then the smaller crystals (the
fines) are apt to pass through the screen and find their way to the filtrate. This is sometimes a very
serious matter if the filtrate is to be processed subsequently. Furthermore, the mechanism of
transferring (or conveying, or pushing) the crystals on the screen from the feed end to the
discharge end, is generally a difficult problem because of the high speed. This mechanism is now
either in the shape of a ribbon screw conveyor, running at a differential speed to the basket,
usually at a differential of from 1 /60 to 1 /80 of the basket speed, as in the Bird continuous machine

Fig 40 Continuous centrifuge (pusher type).

(Fig, 39), or in the form of a piston pusher madding 8 to 18 strokes per minute with the length of
stroke about 40 mm., as in the Escher Wyss continuous machine+ (Fig. 40). The push-type
mechanism runs very smoothly and has no tendency to crush the crystals in the basket. The
differential speed conveyor as developed by the Bird Machine Co. is now operating steadily and
stably, but the ribbon conveyor edge has a scraping action, grinding the crystals forward over the
surface layer of the cake in the basket. Both types at the present stage require the crystal size to be
no finer than 0.20mm., or 65 mesh, and are not adaptable to crystals finer than 0.10 mm., or 150
mesh; otherwise too much of the fines would pass through the screen. To meet this difficulty, the
Bird Machine Co. put out a solid bow1 machine having solid walls in the “basket” (I, e, doing
away with the slots and depending solely upon the centrifugal action to separate the solids from
the liquid in the bow1 (Fig.41). The machine is otherwise no different from the continuous basket
type. This solid bow1 machine is particularly suitable for hanging fine slimes, such as cement clay,
sewage sludge, lithopone pigment, etc. But for bicarbonate of soda crystals, it is not so satisfactory,
because it has a tendency to break up the crystals and knead them into a dough-like mass, so that
the dryness or “yield” of the bicarbonate obtained is not much better than the “yield” obtainable
from a well-designed modern rotary vacuum filter. Strange to say, however, bicarbonate of soda
crystals could be handle in the ordinary batch-type, vertical, suspended centrifuges, without such
difficulties, and these centrifuges are used extensively for refined bicarbonate of soda, as
mentioned above.
FIG 41 Bird solid bowl centrifugal Slter.
 FIG 42 Escher Wyse ( ter Meer ) automatic centrifuge.

At present, only certain of the automatic type centrifuges notably the Escher Why (ter Meer)
automatic machine (Fig. 42) have been used for the crude bicarbonate of soda service in place of
rotary filters. This ter Meer automatic will handle the fine crystals of bicarbonate of soda because
of the special feature in its basket construction, but more particularly because it also uses a woolen
felt as the filtering medium (Fig. 43) of the bicarbonate of soda crystals from the columns.

FIG 43

Eecher Wyse automstic centrifuge

using felt as filtering medium.

On the other hand, these automatic centrifuges are very well suited to the drying of coarse
crystals, e.g., NaCl, NH4 Cl, (NH4 )2SO4, Na2 SO4, CuSO4, or other substances that habitually
crystallize in large crystals. The better known ones are the tar Mere automatic of from 1700 to
2100 mm. Diameter, and the Sharples 63” automatic and 20” Super-D-Hydrator.*
As regards the cycle or the period of intermittency, these automatic machines vary widely. For
example, the Sharples Super-D-Hydrator with 20” basket has a cycle of only 45 seconds, while a
large ter Meer machine has a cycle of 18-20 minutes. Other automatic machines such as the
Haubold machine+ have a cycle of from 4 to 7 minutes. These cycles are, however, adjustable
within a certain range. Table 75 illustrates the distribution of time among the several events in the
cycle in some typical makes.

The present tendency is to go to smaller meters, but higher speeds (Sharples

Super-D-Hydrator, 20” diameter, and the Bird continuous, 18” diameter) with smaller amounts of
material in the baskets. According to the simple laws of physics,
If = acceleration in feet per sec. per sec.
v = peripheral velocity in feet per sec.
r = radius of basket in feet
v2
then =
r

This centrifugal force as developed commercially varies within a wide within a wide range of
from 200 to 1800 times gravity, the higher figure being for speeds as high as 1800 r. p .m., using a
40-inch basket in a batch type machine. Normally, a high-speed commercial centrifuge develops
about a thousand times gravity.
It is apparent that the greater the centrifugal force, the greater the drying power, or for the
same dryness, the shouter the time required for spinning. Now, to accomplish this, the speed, N, or
the radius of the basket, must be increased. But since the centrifugal force is increased as the
square of the speed, but only linearly as the diameter, it is advantageous to increase the speed
rather than the diameter. Also, from the consideration of mechanical strength, it is also desirable to
keep the diameter small. Hence the tendency of the design for a good factor of safety is rather to
have a smaller basket with a correspondingly smaller mass in it, working at a higher speed, than to
have a larger diameter, larger volume and consequently greater mass.
The following screen analyses on some of the typical crystals, such as NaCl, (NH4 )2 SO4 ,
NH4 HCO3 and NaHCO3 , are given to illustrate the range of crystal size ordinarily met with in
industrial plant(see Table 76) It can be seen that the refined bicarbonate of soda has the smallest
crystals among them ;next comes the crude bicarbonate of soda (ammonia soda).Crystals such as
NaCl(NH4 )2 SO4 ,NH4 Cl and the like ,which have over 80 per cent larger than 0.20 mm. Or coarser
than (retained on) 65 mesh, ate all suitable for use in the continuous centrifuges developed at the
present time. Crystals such as refined NaHCO3 , over 50 per cent of which will pass through 150
mesh, cannot be used on the continuous machine, using metal screens. The Bird Machine
Company’s solid bowl machine, which has no screen at all, is unique among all types of
continuous machines in that it can handle any fine, slimy solids down to even the cement size,
This solid bowl machine will handle sodium bicarbonate crystals (crude of refined), but the
moisture content in the cake is not less than that obtainable from a high-yield rotary vacuum filter.
Further, the bicarbonate cake obtained from the solid bowl machine becomes denser and glazed
because of the kneading action of the screw conveyor on the crystals.

From the above, it can be safely said, as regards continuous type centrifuges in general, that
they have so far not found satisfactory application in the ammonia-soda service. But because of
their strictly continuous character and because of the higher “yield” possible, there always exists a
temptation to try to adapt them to the ammonia-soda industry; indeed extensive efforts are being
made in these direction. There are, however, certain drawbacks to their use, among which are (1)
that the cost of the machine is far too great as compared to the capacity obtainable, and (2) that the
power required to drive such a machine is also greater, Compare this with a moderate sized,
continuous, rotary, vacuum filter of the latest high-yield type, which has a capacity of 200-300
tons of soda ash per day (i.e., 300-500 tons in terms of sodium bicarbonate per day),requiring only
5 H.P. motor to drive it, taking into consideration the fact that the difference in the “yield”
obtained from the filter vs. the continuous centrifugal may be only a trifle.
Also, with the advent of the continuous and automatic horizontal machines, it does not
necessarily follow that the batch-type, vertical, suspended centrifuges will be entirely superseded.
Nowadays, these bath type centrifuges (which, since they are most generally used in the sugar
industry, have become known as sugar centrifugals’) have also been brought to a high degree of
perfection .A 48-inch vertical-suspended machine is mow capable of operating at 1200 r.p.m.
while a 40-inch machine may now run at as high as 1800 r. p. m .being suspended and free to seek
its balance ,this batch-type vertical machine may be operated safely at a higher speed than a
two-bearing or overhung horizontal continuous machine of the same basher diameter. Further,
certain automatic mechanisms, such as the charging and unloading devises .may now also be fitted
to these bath-type machines, so that one person also can look after a number of them at the same
time.
As regard the crystalline forms of the four main salts in equilibrium in the liquor, the
crystals of sodium chloride are of course cubic (isometric); those of sodium bicarbonate
monoclinic; those of ammonium bicarbonate are also monoclinic; and those of ammonium
chloride isometric or tetragonal. Under normal operating conditions, the ammoniated brine
is under saturated with respect to sodium chloride, and so no crystals of sodium chloride
should deposit out. Because of the mode of formation in the towers, the cr4ystal of sodium
bicarbonate are necessarily small. Under a microscope even at a magnification of 400
diameters, these bicarbonate crystals appear merely as long rods, the individual shape of the
rod being still difficult to examine. This is because of the violent stirring effect due to the
bubbling of the carbon dioxide gases in there, with the result that the crystals formed are
very small, especially when the tower temperature has not been properly regulated or when
the carbon dioxide gas delivere d to the tower is great weakened, these bicarbonate crystals
have a “twinning “ habit and often form “daisies.” These are seldom seen as perfect, well
–developed, crystal individuals .As seen under a microscope, a few of these crystals as seen
under a microscope and estimated on a “counting chamber” is between 100 and 200 microns
(00.10-0.20 mm), while under poor operating conditions yielding a pasty sludge, the size may
be between 50 and 100 microns 90.05-0.10mm.) or smaller.
Under certain conditions the crystals of ammonium bicarbonate may be deposited with the
sodium bicarbonate. If this happens, it will be seen that they are coarser and are more plate –like
in formation, quite different from the rod formation of the sodium bicarbonate, though belonging
to the same system (monoclinic). The crystal s of ammonium chloride are not formed under these
conditions because it is extremely soluble and exist in the crystal of sodium chlorides, are merely
carried by the mother liquor which has not been completely washed room the crystal cake.
Good tower control should include a daily photo micrographic study of the bicarbonate
crystals from the towers. Not only must these crystals be maintained coarse and filterable but also
they should be of uniform size, free as far as possible from the secondary crystals.
Attempts have also been made to study the composition of the mother liquor from the
towers .It is of a much more complex character than the precipitated sodium bicarbonate with
which it is in equilibrium .the result of this investigation will be given in the following chapter.
 Chapter XI

Composition of mother Liquor from Carbonating Towers

Chemical reactions taking place in the carbonating towers are of a most complex
character. They are so-called ionic reactions with the soluble components in reversible
equilibria with the solid phase in the form of the precipitated sodium bicarbonate and
possibly also ammonium bicarbonate. These equilibrium points shift with the change of
temperature so that the draw liquor from the carbonating towers must be studied at the
temperature of the draw, and kept at this temperature. Not only does the solubility of the
sodium bicarbonate formed increase with the rise of the temperature in the draw, but also
the main reaction is reversed by the rise in temperature (i.e., from solid bicarbonate formed
in contact with ammonium chloride in solution, sodium chloride can be re-formed with the
rise of temperature). Equilibrium among the various salts in the solution phase, i. e., in the
mother liquor, also charge with the temperature. There are eight different kinds of ions in
solution (NH4 + ,H+ ,OH-,Cl-,HCO3 -,CO3 =,Na+ ,So4=)not including the small amount of sulfide
present; and from these ,at least nine different conventional combinations can exist. It is a
difficult matter to determine the exact amount of each of these constituents present in

solution .Any heating to which the solution of subjected in the cause of chemical analysis
tends to alter the equilibrium conditions; and where this is necessary ,only the total
quantities of the components in question (regardless of whatever combinations they enter
into ) can be determined .Hence most of the analytical results can be expressed only as total
quantities present .It has been found insufficie nt ,however ,to determine the quantities of all
the individual components from these total quantity determination without the help of the
physicochemical relationship ,as will be seen in the following discussion.
To illustrate the complexities of these salts in the mother liquor, the above physicochemical
equilibrium are given on page 182.The heavier arrow indicates the predominating tendency for the
components to exist in the combinations shown on one side of the equilibrate equation.
Before we proceed to give the method of determination and the calculation necessary to
obtain the desired result, we shall first give the analyses and data covering the ammoniated brine
fed to the towers and also the tower operating conditions at this particular period.
For convenience in calculation and comparison, the analytical results are expressed in the
number of cc. of normal solutions per 10cc.of the draw liquor. These then are converted into the
number of equivalents per liter by simply shifting the decimal point one place to the left. Finally
the concentration of each concentration thus found is given in grams per liter.
Data concerning ammoniated brine:
(1) Temperature of ammoniated brine as introduced to the towers- 49.3
(2) Specific gravity of ammoniated brine as introduced to the towers-1.170 at 20
(3) Total ammonia-50.988 cc. N. solution per 10 cc. Sample (or 101.976 titer)
Free ammonia- 48.396 cc. Per 10 cc. Sample (or 96.792 titer)
Fixed ammonia-2.592 cc. Per 10 cc. Sample (or 5.184 titer)
(4) Total chlorine- 44.570 cc. Per 10 cc. Sample (or 89.140 titer)
(5) SO2 as (NH4 )2 SO4 -0.634 cc. Per 10 cc. Sample (or 1.268 titer)
The operation conditions of the particular column from which the draw liquor was obtained were
as follows:
(1) Draw temperature……………………………………………………………………… . 30.0
(2) Temperature of the reaction zone about two-thirds of the height of the tower ………… 63.0
(3) Temperature of cooling water at exit from top of cooling suction ……………………… 29.0
th
(4) Temperature of cooling water at exit from bleeding point (at the 4 ring from bottom )..19.0
(5) CO2 in tower exit gas ………………………………………………………… .6.3% by volume
(6) Inlet mixed gas to towers…………………………………………………… . 54.0% by volume
(7) Volume of bicarbonate after 5 min. settling ………………………………….44.0% by volume
(8) Sp. gr. of draw liquor (clear portion) ……………………………………………1.126at 20.5
We shall give in some detail the analytical methods employed and the calculations made to
obtain the desired results.
(1) SO3 as (NH4 )2 SO4 . This is determined in the usual way by acidifying with HCI and
precipitating with BaCl2 in hot solution .As practically all sulfates come from gypsum
(CaSO4 ) in the case of rock salt or from MgSO4 and CaSO4 in the case of sea, salt, they
are converted to (NH4 )2 SO4 by ammoniation when CaCO3 , and MgCO3 or
(NH4 )2 CO3 .MgCO3 .4H2 O separate out in therefore of the “mud”.
MgSO4 + (NH4 )2 CO3 → (NH4 )SO4 +MgCO2
Ten cc. of the mother liquor contain (NH4 )2 SO4 =0.685cc.N solution.
(2) Total ammonia (NH4 +NH3 ). This is determined by distillation with an excess of NaOH I
the usual manner. For this purpose a 5cc.sample of the mother liquor is taken.
Ten cc. of the mother liquor contain total ammonia =45.402cc.n H2 SO4 .
(3) Total alkalinity. This is determined by taking 10cc.of the mother liquor, adding an
excess of a known amount of N/2 H2 SO4 and titrating back with N/2 NaOH using
methyl orange ad the indicator .ten cc. of the mother liquor =9.440cc.N H2 SO4 .
(4) Total chlorides, Cl-.this is determined by taking an aliquot portion containing 0.500 cc.
of the mother liquor, neutralizing, adding 0.2 gram CaCO3 , diluting it to about 100cc.
and titrating with N/10 AgNO3 , using 3 drops of 15% K2 CrO4 as indicator .The end
point is where the faintest red in the background of white AgCl precipitate is just
observable.
Ten cc. of the mother liquor =9.440cc. N H2 SO4
(5) Total sodium, Na+. Acidify an aliquot portion containing 4 cc .of the mother liquor with
dilute HCl, using a small excess. Heat to boiling and remove the sulfate with a slight
 excess of BaCl2 .Filter and wash with hot water. Make the filtrate distinctly alkaline with
NH4 OH, and add small excess of (NH4 )2 CO3 and (NH4 )2 C2 O4 .Allow to settle ,and in a
large casserole ,and ignite gently until 0 fumes are seen. Dissolve the residue with a
little water and transfer the solution to a small tarred evaporating dish, cautiously heat to
dryness and ignite over a low flame once more. Cool and weigh the NaCl residue for an
alternative method using uranyl magnesium acetates see Chapter XXXVIII,” Chemical
Analyses and Tests in Alkali Industry.”
Ten cc. of the mother liquor contain sodium =13.908 cc. N solution (by calculating from
the weight of NaCl).
(6) Total CO2 .The most accurate method for determining CO2 is by evolution and
absorption of CO2 in a purifying and absorption train gravimetrically, as described on p.
472, Chapter XXVIII. Unfortunately the evolution and absorption method is somewhat
long and tedious. The method of adding an excess of NaOH, heating, precipitating CO2
asBaCO3 by BaCl2 filtering washing and titrating the precipitated BaCO3 with an acid is
likely to give too high results as it is very difficult to wash out the excess of NaOH
completely from the fine precipitate of BaCO3 The result may be 4 to 5 per cent to high.
The method of liberating the CO2 from solution and measuring its volume before and
after absorption, on the other hand, gives too low results. The error is due partly to the
solubility of CO2 gas in water and partly to the inaccurate theoretical factor used from
the volume so obtained. The figures may be a half percent too low. The method of
driving off CO2 gas and weighing the loss in weight by means of an alkalimeter is likely
to give high results because of the loss of a small amount of water vapor being counted
as CO2 gas even if air bubbling (instead of heating) is employed to drive ort all the CO2 ,
The results, however only one -half of one per cent too high, when air bubbling (without
heating ) is employed to drive off the CO2 These results may mot be of value for
checking purposes inasmuch as the magnitude of error is sometimes much greater than
some of the quantities sought, if such quantities are to be found by difference.
Ten cc. of the mother liquor contain CO2 =14.245 cc. N solution by calculating from the
weight of CO2 obtained by the evolution and absorption method above.
(7) Ammonium chloride, NH4 Cl this is determined by alcoholic extraction. It is found that
(NH4 )2 SO4 (NH4 )2 CO3 and NH4 HCO3 are very sparingly soluble in a 97 per cent of
stronger alcohol. Ammonium hydroxide, which is soluble in the 97 per cent alcohol,
along with the ammonium chloride, cam be driven off by evaporating over a steam bath
without any material loss ofNH4Cl, The detailed procedure is as follows :
Pipette 5 cc. of the mother liquor into 200 cc . absolute alcohol. Stir allows settling and
filtering by decantation. Wash the residue with small portions of absolute alcohol, settle
and filter, combining the washing and the alcohol filtrate over a steam bath to about
one-half of the volume of until no alkalinity is detectable with litmus paper. Distill off
ammonia from these alcoholic solution of NH4 Cl with a small excess of NaOH into an
Erlenmeyer flask containing a known amount of H2 SO4 as in the “Fixed Ammonia”
determination. Evaporate off the excess alcohol from the distillate over steam bath to
about one-half of the volume Col, dilute with water and titrate the excess of H2 SO4 with
N/10 NaOH using methyl orange as the indicator. Loss of H2 SO4 in the distillate
represents fixed ammonia (NH4 Cl)
 Ten cc. of the mother liquor contain NH4Cl =36.062 cc. N H2 SO4
A summary of the analytical results is given in Table 77.

* Cf. K. W. Jurisch, Chem. Ztg., pp. 1091-1093 (1906)

From the analytical data in Table 77, we next proceed to carry out different calculation, but it
will be seen that the desired quantities cannot be completely determined without the help some
physicochemical relationship.
The mother liquor is alkaline to methyl orange, but barely turns phenolphthalein red; hence it
is practically neutral to phenolphthalein.
Let
 Then from the total CO2 we have
C2 = 1.425 = 2y + 2r +2u (1)
From total ammonia
C1 = 4.5402 = z + x (2)
From electrical neutrality or equivalent quantities of positive ions to negative ions,
C4 + z + t =C5 + C3 + s + y+ 2u
Since the solution is very closely neutral to phenolphthalein its pH value is not far from 8,i. e.,
in the order of magnitude of 10-8 mol of H+ per liter and s 10-6 mol of OH-. Hence without error
we can from the ionization t and s
C4 +z = C5 + C3 + y + 2u (3)
Also ,the ionization constant of water at 25 is
Kw = 10–14 = ts (4)
Now from the ionization equilibrium of ammonia, and from the first hydrogen and the
second hydrogen of the carbonic acid in the mother liquor, we have
 Kb = ionization constant of ammonia = 1.8 10-5 = 8z (5)
x

K1 = ionization constant of first hydrogen of carbonic acid = 3.5 10-7 =ty (6)
r

K2 = ionization constant of second hydrogen of carbonic acid from HCO3 -

tu
= 6.0 10-11 = (7)
y
We have here seven unknowns and seven equations, so that we are able, at least , to solve for these
unknowns. But solution by simultaneous equations with the elimination of one variable after
another becomes very complicated. Fortunately, the process can be shortened by solving them by
the method of successive approximations, noticing that u is only about 0.006 time as large as y,
and r about 0.03 time as large as y.

u 6.0 10-11 6.0 10-11

For from (7) y = = =0.006(approx.)
t 10-3

r t 10-3
For from (6) = =0.03(approx.)
y = 3.5 10-7 3.5 10-7

We can leave u and r out of consideration from (1) and (3) for the first approximation and then
come back to correct them afterwards, and so on. Solving by successive approximations in this
way we get after the third approximation the following values:
x = 0.2502 mol NH4 OH per liter
y = 0.6892 mol HCO3 - per liter
z = 4.290 mols NH4 + per liter
r = 0.0187 mol H2 CO3 per liter
s = 1.05 10-6 mols OH- per liter
t = 9.52 10-9 mols H+ per liter
u =0.00434 mol CO3 = per liter
Since from the independent determination above
NH4 Cl=3.6062 equivalents per liter (normal)
Therefore from the total chlorine,
NaCl =4.9144-3.6062 =1.3082 equivalents per liter (normal)
and from the total sodium,
NaHCO3 + Na2 CO3 =1.3908-1.3082=0.0826 equivalents per liter (normal)
Again from the total ammonia,

NH4 HCO3 and (NH4 )2 CO3 = c 1 -x-NH4 Cl-(NH4 )2 SO4 =

4.5402-0.2502-3.6062-0.0685 = 0.6153 equivalents per liter (normal)
The results are summarized in Table 78.
Table 78. Composition of Mother Liquor.
No. of Equivalents per
Constituents Liter (normality ) Grams per liter
 NH4 OH(x) 0.2502 4.25 (as NH3 )
NaCl 1.3082 76.5
NH4 Cl 3.6062 193
(NH4 )2 SO4 (c 5 ) 0.0685 4.52
NH4 HCO3 +(NH4 )2 CO3 0.6153 48.6 (all expressed as NH4 HCO3 )
NaHCO3 +Na2 CO3 0.0826 6.93 (all expressed as NaHCO3 )
H2 CO3 ( r) 0.0178 2 = 0.0374 1.16 (as H2 CO3 )
Total CO2 (diss.) 1.4245 31.3
Total “free ammonia” 0.8655 14.7 (as NH3 )

It will be seen that we have not used the data of total alkalinity. These can be used as an
independent check on the accuracy of the results Thus, total alkalinity from these results would be
NH4 OH + NH4 HCO3 +( NH4 ) HCO3 +NaHCO3 + Na2 CO3 =0.9481
From independent titration the value obtained above is 0.9940.
From the foregoing results we can obtain important information concerning the tower
operation.
In the first place, the shrinkage of volume of the ammoniated Brie due to fixation of free
ammonia in the towers is given by
 The amount of CO2 in the mother liquor from the towers, however, increases as the “free
titer” in the green liquor and consequently in the draw liquor gets higher.
From the above it will be seen that the clear portion of the mother liquor contains only an
inappreciable quantity of dissolved NaCO3 the It would thus be very doubtful if there could be
much sodium carbonate in the solid phase with the precipitated sodium bicarbonate in equilibrium
with the liquor in the towers Hence we must conclude that the normal sodium carbonate found in
the bicarbonate obtained from the filters owes its presence largely to the spontaneous
decomposition of sodium bicarbonate which occurs almost instantaneously on contract with the
air.
In the foregoing calculation we have made use of the electrolytic dissociation constants .We
are aware that these figures are not sufficiently accurate and that there is some disagreement in the
values obtained by several workers, especially in the case of the dissociation constant for the
second hydrogen of carbonic acid. Further, in using these data, we have neglected the factor of the
degree of ionization (or the activity coefficient) and have assumed complete dissociation for salts
and none at all for ammonium hydroxide and carbonic acid a condition not entirely justifiable in
such a complex and concentrated solution as the draw liquor from the towers. Also, it must be
mentioned that in the case of the un-ionized carbonic acid the total concentration really includes
carbon dioxide gas dissolved and carbonic acid undissociated, according to the equilibrium CO2
gas (dissolved)+H2 O (undissociated) H+ HCO3 -.However, it is believed that the results obtained
above were not materially affected by these assumptions, and that they are probably correct to the
third significant figure.
It must be pointed out that while sodium bicarbonate must be classified as a fairly soluble salt
having a solubility of the order of a hundred grams per liter at room temperature, it is tendered
very much less soluble in the presence of an excess of the reagents in solution (NaCl and
NH4 HCO3 ), which decreases its solubility to less than a tenth of its normal value when present
alone in water.
*See “The Theory and Use of Indicators” by E. B. R. Prideaux, New York, D. Van
Nostrand Co., pages 293-306.
 Chapter XII
Phase Rule in Tower Reactions: Graphical

Representation of Ammonia Soda Process

The reaction between sodium chloride in the ammoniated brine and ammonium bicarbonate
formed in the brine represents a system of four components. Unlike the four-component system
consisting of three salts with common ion in an aqueous solution, such s sodium potassium and
ammonium chlorides, which can be conveniently represented by the three sides of a regular
tetrahedron placing the fourth component at the vertex the system, which consists of two salts in
an aqueous solution with no common ions ,forming together with water also a four component
system, as best represented by the four sides of on-hale of a regular octahedron , In these case,
salts having common ion are represented on the adjacent side if a solid angle of the octahedron
The isothermal diagram in each case will be a plane orthogonal projection of the space figure onto
the triangular base in the case of the tetrahedron ,or on to the square mid-section in the case of the
octahedron (Fig.47). In the last case, with which we are here concerned, the isothermal diagram is
represented by four rectangular coordinate axes, each axis being separate and meeting at original
at right angles to the adjacent ones so that they do not form two continuous axes intersecting at the
origin. This is ethereal diagram is noting more or less than a plane projection from the space figure

1
onto the square base with the units measured along the four axes equal to of the units
2
measured along the four sides of the upper half of the octahedron starting from the vertex and
making an octahedron at a constant temperature.

FIG 44 For space figure.

 FIG 45 For space figure (Cf. Fig 44)

FIG 46 For plane isothermal projection.

By double decomposition the two salts with no common ions can be supposed to rearrange
themselves. Forming another pair of salts in solution in the following fashion:
 FIG 47 For plane isothermal projection (Cf. Fig 46)

It was from this consideration that the name Thus the reaction that the name “the reciprocal pairs
of salts” originated. In this way there would be formed four individual salts, which might be
thought to constitute, together with water, a five-component system. But the four salts are so
related that any three of them could determine the forth. In other rewords, only three of them are
independent, and the three salts, with water, form the for-component system. Thus, the tower
reaction
NaCl +NH4 HCO3 NaHCO3 + NH4 Cl
Constitutes a for-component system. Inactive, however, an ammoniacal brine is taken and
carbonated with carbon dioxide gas precipitation. of sodium bicarbonate progresses as the
carbonation proceeds , so that ammonium bicarbonate may be said to be formed in site. It is
evident then that the ammoniacal brine is always saturated with decagon dioxide egos at the
temperature in question. It may be thought that this carbon dioxide gas should be considered as
another component, but this is taken care of by the restriction impose that the partial pressure of
carbon dioxide, together with the partial pressures of ammonia and water, must add up to a given
total pressure above the solution. Since in this heterogeneous system we must have a vapor phase
and a solution, the maximum possible umber of solid phases at the invariant point would be four.
Degrees of Freedom + No. of Phases = No of Components +2 =6.
At a given temperature we have used up one degree of freedom, and the number of solid phases
possible is then only three. The number of combinations of three out of a total of four salts is
4 3 2 = 4 These four possible cases are enumerated below:
3
(1) Solid NaCl + solid NH4 HCO3 + solid NH4 Cl in co-existence with solution;
(2) Solid NaCl + solid NH4 HCO3 + solid NaHCO3 in co-existence with solution;
(3) Solid NaCl + solid NH4 Cl + solid NaHCO3 in co-existence with solution;
(4) Solid NaHCO3 + solid NH4 HCO3 + solid NH4 Cl in co-existence with solution.
 +- +- +- +-
Returning to the general expression above AB + CD = AD + CB, when the four solid salts are in
+- +-
equilibrium with the solution we have the product of molal concentrations of the salts AB and CD
+- +-
equal to that of the molal concentrations of AD and CB. This is a necessary condition for the two
+-
pairs of solid salts AB +-and AD
CD +- CB +- to be in stable equilibrium with each other and with
+- +-
the solution. If the product of solubilities of one pair of the solid salts AB and CD is greater than
+- +-
the corresponding product of the solid salts AD and CB at the same temperature, the firs pair are
said to be in a metastable equilibrium with the second pair and will disappear, yielding the second
pair as the solid phase in contact with the solution. Conversely, if the product of solubilities of the
+- +- +-
second pair of salts, AD and CB, is greater than the corresponding product of the firs pair, AB and
+-
CD, the second pair will disappear, leaving the first pair in a stable equilibrium with the solution.
In this reaction we have the solubility relationship at different temperatures as given in Table 79,
page 198.
Thus, we see that the product of solubilities of NaCl and NH4 JHCO3 at each temperature is
greater than the corresponding product of NaHCO3 and NH4 Cl, and the ratio is roughly 2 to 1 each
case.

This means that at al these temperatures the coexistence eon solid NaCl and solid NH4 HCO3 in
contact with the solution is in a metastable equilibrium and these tow solid phases cannot exist
side by side. Consequently, of the four possible combinations of the three solid salts in equilibrium
with the solution enumerated above, only combination (3) to (4) need be discussed, because both
combination (1) and (2) would involve the coexistence of solid NaCl and solid NH4 HCO3 in
contact with the solution. In the actual study at a fixed temperature there exist only two invariant
points such as j and k illustrated by isothermal diagram for 15 . (Fig.48)*
From the phase rule we know that each area in this plane projection represents a region where
one particular salt will separate out as a solid phase from the solution of the composition given by
a point within the area, and that each internal line such as EJ, FJ, JK, KG, or KH represents the
locus where two salts represented by the two areas adjoining each other by such a line will
separate out as the coexisting solid on this line. These are known as univariant lines. The
intersections of these univariant curves give the invariant points such as J and K shown in Fig.48
These invariant points are located by plotting the compositions of the solutions in the following
manner: Set off the molal concentration of NaCl in the solution from the common point O along
the NaCl-axis and the difference between the molal concentrations of NH4 Cl and NaHCO3 from O
along that axis which has the greater concentration. The two coordinates will determine a point.
Notice that at either J or K the composition of the solution is determined by the molal
concentrations of the three salts, namely NaCl, NH4 Cl and NaHCO3 , the difference between the
last two being taken as one of the coordinates. But J and K represent the two invariant points at the
temperature in question where three solid phases can coexist with one another and the solution. At
J the three solid phases are NaCl NaHCO3 and NH4 Cl and at K they are NaHCO3 NH4 HCO3 and
NH4 Cl. It is thus seen that at K the three salts separating out from the solution do not correspond
exactly to those present in the solution. Hence K is called the “incongruent ” point, while J, which
marks the same three salts in the solid phases as in the solution from which these salts separate out,
is designated as the “congruent” point. It might be added here that by adding sodium chloride to
the incongruent system, solid ammonium bicarbonate can be gradually used up until, when solid
sodium chloride has been added in excess, solid sodium chloride remains in place of the solid
ammonium bicarbonate and the systems is again congruent.
* Cf. P. P. Fedotieff, “Der Ammoniaksodaprozess vom Standpunkte der Phasenlehre,” Z. physik. Chem.,
1904

FIG 48 Plane isothermal projection from octahedral space figure for system eonsisting of reciprocal pairs NaCl + NH4 HCO3

and NaHCO3 +NH4 Cl.

From the diagram (Fig.48) it is seen that at 15 .the NaCl-area does not come in contact at
any point with the NH4 HCO3 -area, and as the temperature is raised the two areas do not approach
each other. In other words, at all these temperatures solid sodium chloride cannot coexist with
solid NH4 HCO3 in contact with the solution. This just corroborates the above statement that two
combinations (1) and (2) are nonexistent and need not be considered.
As our aim is to precipitate sodium bicarbonate from the ammoniated brine we need only
consider those three border lines of the NaHCO3 -area, namely GK, KJ and JF. A moment’s
thought will tell us that in the diagram (Fig.48) points lying higher up above O represent higher
concentrations of NaCl in solution, and that points lying farther to the right of O represent greater
concentrations of fixed ammonia in the form of NH4 Cl in solution.To secure a high decomposition
of NaCl to NaHCO3 in the towers, therefore, we must get as close to the NH4 Cl-axis from above
this axis as possible, and as far to the right of O along the NH4 Cl-axis as possible. But for a high
conversion of NH4 HCO3 to NH4 Cl we must operate as far up from B and above the NH4 Cl-axis as
possible. Therefore it is to be expected generally that a high percentage conversion of NaCl to
NaHCO3 is likely to be accompanied by a low percentage conversion of NH4 HCO3 to NH4 Cl. As a
matter of fact, percentage conversion of NH4 HCO3 to NH4 Cl increases as we move from G to K
along GK, from K to J along KJ, and from J to F along JF, unit at F it reaches a maximum of 100
per cent. On the other hand, percentage conversion of NaCl to NaHCO3 increases as we move
down from F to J along FJ and from J to K along JK; but at K it reaches a maximum (not100 per
cent) and starts to fall off again as we move from K across the NH4 Cl-axis in the direction KG.
Along FJ, NaCl can exist in the solid phase as well as NaHCO3 . This does not represent the actual
fact, because under plant-operating conditions the NaCl concentration in the ammoniated brine
employed practically never even reaches its saturation point. Therefore the curve JF need not be
considered. Again, since the traction between NaCl and NH4 HCO3 is a reversible one, there will
always be some NaCl left in the solution. That is to say, the concentration of NaCl in solution can
only approach zero and is never equal to zero, and such points will lie above the NH4 Cl-axis. KG
intersects the NH4 Cl-axis at M. That section of KG lying below the NH4 Cl-axis therefore has no
meaning. By this elimination there are left only the curve KJ and a section KM. KJ is the curve
along which NH4 Cl reaches its saturation point and tends to separate out with NaHCO3 in the solid
phases. Further, it is the curve along which the percentage decomposing of NaCl from the reaction
decreases, the concentration of NaCl left in solution increasing as we move up from K to J. In the
plant operation any possibility for solid NH4 Cl to separate out with NaHCO3 is to be carefully
avoided. For it would lead to a high salt content in the soda ash after calcinations.
NH4 Cl + NaHCO3 NH3 + H2 O +CO2 +NaCl
This consideration, together with the facts that, because of mechanical perfection, the loss of
ammonia in the cycle in the present-day ammonia soda plants has been reduced to a very low
figure, and that the cost of ammonia is now a small item in the cost of manufacture. Has dictated a
condition of operation whereby a high percentage conversion of NaCl alone is aimed at. The
present tendency in ammonia soda plants is to go to a higher ammonia concentration in the
ammoniation operation (in the neighborhood of 100 titers of free NH3 ), while making every effort
to maintain as high a NaCl concentration in the ammoniated brine as possible — at best
considerably below saturation with respect to NaCl. The ratio of free NH3 to Cl- in the resulting
ammoniated brine now employed (expressed in molal concentrations ) is 1.08 or even 1.12 to 1.
This means that modern plants operate approximately within the narrow strip of area bordering on
the short section of the curve KM. As shown by the shaded area in the diagram (PQKM), where
there is a tendency for a small amount of solid NH4 HCO3 rather than solid NH4 Cl to separate out
with the precipitate of NaHCO3 . The presence of NH4 HCO3 in the crude NaHCO3 from the towers
is less objectionable and all of it will be decomposed and recovered in the dryers.
Another revision of the conclusion reached by P.P. Feedstuff in his classical research on the
ammonia soda process in 1904 is that it has not been found desirable to operate at high
temperatures such as 32 ., at which point he said that both ammonia efficiency and the sodium
Efficiency was 84 per cent and that there was “a greater yield of sodium bicarbonate” from the
reaction. If solid ammonium bicarbonate were used in the reaction instead of the ammonium
bicarbonate formed in solution by carbonating the ammonia in the brine with carbon dioxide gas,
this temperature might be desirable. But as we have to bubble rather dilute carbon dioxide gases
through ammoniated brine in the towers, and as the draw liquor has to be exposed to the air before
and during filtration, high temperatures of the liquor at the draw have been found to be
undesirable. For the sake of keeping down ammonia losses as well as getting a high percentage
sodium chloride decomposition in the brine, a draw temperature below 25 . And even as low as
20 .has been found beneficial. The point K, where the decomposition of NaCl reaches its
maximum, is unique, furnishing a good guide for plant operation. This point moves toward the
NH4 Cl-axis as the temperature rises above 15 , according to the results of P.P. Feedstuff, and
lies on the NH4 Cl – axis at about 32 . Consequently, as the temperature rises above 15 ., the
section of the curve KM dwindles in length. Therefore it might be correct to say that we are then
operating around the point K on the NaHCO3 side.
GRAPHICAL REPRESENTATION OF AMMONIA SODA PROCESS
We shall make use of triangular coordinates to represent the composition and solubility
relationship in a more complex system such as the liquor and slurry in the ammonia soda process.
This graphical representation has the advantage that it enables the reader to visualize the steps in
the process more clearly. But before we present these diagrams it may be well to review briefly the
properties of the triangular graphs. It is evident that triangular coordinates present to us plain
curves .By adding a vertical coordinate----a fourth variable----a space curve is readily obtained in
the body of a prism.
Let the triangle be an equilateral triangle ABC; each of the vertices represents a variable. In
our present case, each of the three vertices may represent a component in a solution or slurry. Thus,
we may let A be Cl, B be OH– and C be HCO3 -as in Fig. 49 (b). Each side of the equilateral
triangle represents molar percentage or fractions of a two-component system and its total length
adds up to 100 per cent (from 0 to 100 per cent of one component and conversely from 100 to 0
per cent of the second component). Thus, point R in (b) represents 15 per cent OH-and 85 per cent
Cl- on one side (AB) , and point P represents 15 per cent OH-, 55 per cent HCO3 - and 30 per cent
Cl-. The three coordinates drawn parallel to the respective sides of the equilateral triangle in
successive

FIG 49 Triartgular graphs.

order are R’R,Q’Q and R’P’,or R’R ,QQ’ and RP (=RQ). Therefore the following propositions
may be readily proved:
(1) The sum of the three coordinates of a point in a triangular graph is a constant and is
equal to the length of one side of the equilateral triangle, and each side may thus be
graduated to read from 0 to 100 per cent of each of the two components.
(2) If M and N, Fig. 49(c), represent the composition of two different solutions, the
composition of a mixture of these two solutions mixed in any proportion is represented by a
 point on a straight line connecting M and N, such as O, and by no other point. Further if O
is the composition of the resulting mixture of the two solutions whose compositions are
represented by M and N, MO then represents quantitatively (to scale) the quantity of the
solution N and NO the quantity of the solution M present in the mixture.
(3) If P represents the composition of a system of three components, Fig.49 (a), the
removal or addition of one of three components is represented by points on the straight line
connecting P and the vertex representing that component. Thus, in Fig.49 (a), connect P and
the vertex C, and produce CP to intersect the opposite side AB at T. Addition of component
C to the system is represented by a point moving from P toward C; while removal of
component C is represented by a point moving away from P along the line CP produced,
and at T, component C is entirely eliminated, and the system then becomes a
two-component one. AT then represents the percentage of B, and BT the percentage of A in
the resulting two-component system. Further, by taking points beyond T, such as T’, we can
conceive that T’ represents algebraically a negative coordinate C in the three-component
system. Suppose now C represents water in a solution, in which A and B are two solutes,
addition of water (dilution of solution) is represented by a point moving from P toward the
vertex C; while removal of water (concentration of solution) is represented by a point
moving away from P toward T on the line CP produced.
(4) If in Fig.49 (d) a represents the composition of a double salt in a system where C
represents H2 O and avow represents the solubility curve of the double salt at a certain
temperature t, a straight line connecting any point, S, on this solubility (saturation) curve
with a represents a mixture of the solid double salt and its mother liquor. If L represents
such a mixture, the ratio of the weight of the double salt to that of the mother liquor present
in the slurry is as SL a L. Further, if S’ l’ represents the compositions of another slurry at
different temperatures yielding a series of saturated solutions with the same double salt in
the solid phase, the intersection point a between the two straight lines SL and S’l’ then
determines the composition of the double salt a itself and thus locates it on the graph.
In general, as can be seen above, if a point P, representing the composition of a system, is to
vary with respect to one component only ,join P with the vertex of the triangle representing that
particular component .Then any point on the line or on this line produced ,represents the
composition resulting from the variation of this one component alone ,and consequently is a locus
of constant ratio between the two components represented by the two remaining vertices.
Graphical representation of the ammonia soda process by means of a triangular prism is
shown in Fig .50. This process diagram is based on a system of triangular coordinates for OH-,
HCO3 – and Cl- with a fourth (vertical) ordinate for NA+ ----NH4 + in the space relationship. Each
edge of the prism is graduated to read from 0 to 100 per cent of one component or from 100 to 0
percent of the other, and the six corners (or apices) represent respectively NaCl, NaHCO3 , NaOH,
NH4 Cl, NH4 HCO3 and NH4 OH . We have not considered H2 O, but represent these six components
on the dry basis.
Let us start with brine at the NaCl apex. Because of the presence in the salt of certain calcium
and magnesium salts which upon ammoniation yield equivalent amounts of fixed ammonia (3-5
per cent as NH4 Cl), this point is represented not by the NaCl apex but by A, along the edge
NaCl---- NH4 Cl. We ammoniate this brine in the absorber with ammonia gas from the distiller,
cooled through the distiller condenser, containing about 0.2 mol CO2 per mol of NH3. In other
words, this distiller ammonia gas is already about 20 per cent bicarbonate and is therefore

FIG 50 Ammonia Soda Process Diagram.

represented by B. Join A and B. As ammoniation proceeds, we arrive at C where the titer of Cl is

about 89 and that of ammonia 99, or when the ratio of ammonia to sodium chloride in the
ammoniated brine is as 1.11 to 1. This ammoniated brine, after settling and cooling, is pumped
first to the cleaning tower where a small amount of lean CO2 gas is passed in. The liquor is
allowed to be carbonated in the cleaning tower to less than 30 per cent bicarbonate, or until it
contains about 65 g CO2 per 1. This is represented by point E at the intersection of the carbonation
line KC and the line AD connecting A with the 30 per cent bicarbonate point D. This
precarbonated liquor is then sent into the making towers where it is finally carbonated to the
highest degree possible, The carbonation line, KC, is not exactly parallel to the ammonia plane or
NH4 HCO3 - NH4 OH edge, but is slightly inclined away from it on account of loss of ammonia gas
from the ammoniated brine due to the stripping effect of the CO2 gas in the towers.
Precipitation of NaHCO3 gradually occurs in the making towers; for above about 32 per cent
bicarbonate, the solubility product of sodium bicarbonate (Na+HCO3 -) begins to be exceeded, and
the precipitation commences long before the ammoniated brine reaches its maximum bicarbonate.
Because of hydrolysis, it is not possible to get 100 per cent bicarbonate. In fact, precipitation of
NaHCO3 in the making tower takes place considerably at the 40 per cent bicarbonate point G, and
continues as carbonation progresses until the maximum point K is reached at about 95 per cent
bicarbonate, which occurs at the bottom of the making tower, where the richest CO2 gas comes
into contact with the liquor, and the magma is about to be drawn out. But, strange to say, the
mother liquor in the magma in equilibrium with the bicarbonate crystals is still only about 82 per
cent bicarbonate. The highest degree of carbonation, K, is at the point of intersection of the
carbonation line KC and the line from A to H, the 95 per cent bicarbonate point.
The slurry at K is filtered, giving the filter liquor P and the crude bicarbonate (ammonia soda)
N. Because of the presence of NH4 HCO3 to the extent of 5 molal per cent in the crude bicarbonate,
the line KN does not pass through the NaHCO3 apex but some point N near it. Upon calcinations,
NH4 HCO3 is driven off so that Q is reached. NaHCO3 Q is then decomposed to light ash, R, which
CO
does not quite coincide with the middle point Na( 2 ) or Na2 CO3 of the NaHCO3 -NaOH edge, as
3

it should. Here again the calcinations line QR does not exactly follow the NaHCO3 - NaOH edge
on account of the presence of some NaCl as one of the impurities in the soda ash.
The filter liquor at P is first sent to the distiller heater and the “free ammonia” is decomposed,
yielding about 80 per cent bicarbonate ammonia gas M, and a heater liquor L, containing about 72
per cent fixed ammonia (as NH4 Cl). This heater liquor flows into the prelim and finally passes
through the lime still giving rise to CaCl2 in the distiller waste. The 80% bicarbonate heater gas is
mixed with the ammonia gas from the lime still, coming over as 20% bicarbonate distiller gas to
the absorber.
 Chapter X

Decomposition of Sodium Bicarbonate by

Calcination
The washed sodium bicarbonate from the filters drops onto a belt conveyor, which feeds the
furnaces in which the bicarbonate is decomposed by calcination. The bicarbonate is charged into
the furnaces through star feeders, or better still, by feed tables whereby a good air seal is secured
so that air may not be drawn into the furnaces with the bicarbonate feed nor the furnace gas blown
out.
The reaction in the furnace is the decomposition of sodium bicarbonate into soda ash, carbon
dioxide and water. The process looks very simple, and it is so in a laboratory where the product
can be stirred and no attempt is made to recover either ammonia or CO2 gas and this CO2 in a
concentrated form.
But in actual operation it is one of the most difficult parts of the whole process. Difficulties in
this operation have brought much trouble to the project, and plants have failed because this
particular part of the process has not been successful. The difficult nature of this operation can be
judged from the number of patents, which have been issued on the apparatus invented to perform
it. For it is not simply a drying operation. The whole difficulty arises from the following facts: (1)
Sodium bicarbonate has a tendency to cake into lumps or balls, especially when its moisture
content is high, these lumps having under composed bicarbonate cores which heat can hardly
reach from the outside. (2) The wet sodium bicarbonate forms a hard scale on the surface of the
steel shell through which heat cannot penetrate, and consequently the localized heat is likely to
burn out the metal. (3) The dried ash has a great avidity for the steam condensate and becomes
“set,” plugging such passages as the return ash chute, the extract barrel, etc., if the steam is not
conducted away fast enough, i.e., if the dryer happens to be under pressure. In addition, it is
necessary to get a bicarbonate-free soda ash, to lose little or no ammonia or CO2 gas, and finally to
maintain a high CO2 concentration in the returned gas.
In the early days, an enclosed dishpan, 10 to 14 feet in diameter, provided with a stirrer, and
an opening in the cover for gas outlet, was used. This, of course, did not give continuous operation.
In 1882, Striebeck introduced into the ammonia soda industry the ten pans in a somewhat
modified form, which had been used successfully in the LeBlanc soda industry. From that time on,
this type of pan superseded older types of calcining apparatus. Even today some are still in use in
Europe. It was also customary to divide the calcining operation into two stages, one a preliminary
drying operation and the second a finishing operation. The former involved driving out moisture,
ammonia, and a large part of the CO2 (about 75 per cent), while the latter drove off the rest of the
CO2 and completed the operation. The preliminary operation was carried out either in the dishpan
or in a lightly fired Thelen pan, and the partially calcined bicarbonate was finally finished off in
the finishing Thelen pan. It is claimed that a denser soda ash was obtained this way. One
construction of Thelen pan consisted of 6 cast-iron semi-circular sections, 7 feet in diameter, 11
inches thick and 5 feet 4 inches long, each pan being 32 feet long. The cover consisted of convex
wrought-iron plates. A series of cast-steel scrapers inside was actuated by a common shaft, which
oscillated through an angle of about 90 . Openings for bicarbonate charge and gas outlet were
provided in the cover at the front near the furnace end. The ash extract was located at the rear end
in the bottom of the shell below the pan. A Thelen pan, 7 feet in diameter and 34 feet long, has a
capacity of 20 tons of ash per 24 hours, and requires coal consumption of about 3 tons, and power
consumption of 10 to 15 hp. It is thus seen that fuel economy in the Thelen pan is excellent, and
the ash obtained generally somewhat denser. It can be arranged so that no returned ash is required
in the Thelen pan calcinations, thus obtaining good fuel economy.
The ash made by the ammonia process is very loose and light. It takes more packing space
and a larger bag per unit weight than the LeBlanc ash. Hence there is a heavier packing cost per
unit weight of ash. In a number of industries this light powder is not in favor, especially in the
glass and ultramarine industries. This is a distinct disadvantage for the ammonia process as against
the LeBlanc process. In earlier times, this was overcome by subjecting the ash again to an open
fire in a gas –fired reverberate furnace. The ash was thus brought to a state of incipient fusion and
became densified. If the calcining operation was carried out in two stages, the partially
decomposed bicarbonate, or “roaster ash”, was best utilized for this dense ash making. This was
done either in an evolving furnace like a black ash revolver, or in the Mactear mechanical furnace,
which had been in use in the LeBlanc industry since 1876. The latter is a type of reverberate
furnace having a circular, horizontal revolving firebrick bed as large as 20 feet in diameter. The
“roaster ash” was dumped onto this horizontal bed and the charge was turned over continuously
by a number of small revolving stirrers rotating about axes stationary with respect to the revolving
bed. When the operation was completed, the dense ash was scraped out through a central opening
at the bottom of the revolving bed. This is thus a batch process.
The apparent density of the ammonia soda ash varies according to the manner in which it is
determined. The apparent density of the normal ash, or light ash, when loosely packed, is about
0.5,i.e; almost 500 grams per liter. When the same ash is well shaken in a cylinder and tapped until
no more diminution in volume is observed, the apparent density becomes approximately 0.7.
Similar differences exist in the case of a dense ash. A dense ash of 0.9 to 1.0 apparent densities
when loosely filled, will give a value of 1.2 to 1.3 when thoroughly shaken down. By the
densifying process described above, the ash volume is contracted under heat and the density can
be made equal to that of the LeBlanc ash. From a poorer grade of bicarbonate crystals having a
lower yield, the ash obtained may be denser, and conversely.
The practice of dividing the calcining operation into two stages and of making dense ash by
the fire method described above is now obsolete. In the United States none of the equipment
described above was in use to any great extent in the ammonia soda industry except possibly some
Thelen pans in the Columbia Alkali Corporation.
In America in the early days, calcining of the wet crude bicarbonate (ammonia soda), which
has a slight amount of salt and ammonia, was at first done in shallow pans having revolving
knives supported from the cover. The moist bicarbonate was shoveled into a side door and, after
being heated by the fire below to a temperature of 180 ; a part was drawn out with a hoe, leaving
enough to mix with the wet charge to prevent scaling of the pan. These “secheurs” were not
efficient. Air leaked in and CO2 and ammonia gases escaped, so that the CO2 in the gas was not
much over 60 per cent and considerable ammonia was lost.
The next development was rotary cylinders some five feet in diameter and 60 feet long,
revolving inside a furnace on tires supported on wheels at the ends. These cylinders were later
changed to cylinders 6 feet in diameter, driven by gears, at first with steam engines and afterward
with motors. These are the prototype of the present rotary calciners so universally used in
American ammonia soda plants. Improved feed tables and feed screws were devised to mix the
returned ash with the charge in order to reduce the moisture to about 6 per cent to avoid scaling of
the cylinders. Improved furnace design and automatic stokers have made these rotary calciners
efficient machines in performing this difficult operation.
The rotary dryer has been developed to a high degree of perfection in America. It possesses
many advantages over the European soda ash roasters, including the Thelen pans. It has a larger
capacity and less power consumption per unit of output. It is capable also of giving a high-test
returned gas. Formerly the shell was riveted (butt-strap joints); now these shell joints are either
electrically welded or forged. The present standard machine is 6 feet in inside diameter and 62 feet
in over-all length. The body of the cylinder is made of7 /8 -inch steel plates (20×20 feet or 20×10
feet sheets). The joints in the body are best forged together, or else are V-ed at 75 or 60 both
outside and inside for electric welding, and the welded seams inside the shell ground smooth. In
electrically welded cylinders the girt-welded seams must be staggered to avoid continuous joints.
The two end sections are of cast steel or cast iron, conical in shape, 6 feet inside diameter at the
larger end, 5 feet at the smaller end, and 5 feet long each, riveted to the ends of the steel cylinder
body. The whole kiln is enclosed inside the brick setting with the exception of a part of the conic al
heads at both extreme ends. The kiln is supported at the extreme ends on two pairs of rollers. The
girt gear is attached to the extract end just outside of the brickwork. The tires at both ends and the
four rollers are made of cast steel. The dryer runs at about 4 r. p. m. The kiln is supported in a
horizontal position. Bicarbonate is charged in through a feed table and the dry ash from the
returned ash screw conveyor is fed to the feed barrel at the front end, mixed with the wet
bicarbonate by a feed screw, and conveyed into the furnace by a ribbon conveyor. The gas uptake
is located at the feed end just behind the feed inlet pipe and the ash is extracted by means of a
screw conveyor at the opposite end. The furnace is located at the feed end so that this arrangement
of heating is not counter-current. Throughout the length of the dryer, dragging along at the bottom
inside is a C.S. scraper chain with hardened steel blades attached to the links to keep the shell free
of soda scales. The chain is fastened to a fixed-feed support at the feed end and to the extract
barrel at the extract end. The ash from the extract barrel discharges to a screw conveyor called the
“collecting conveyor,” is cooled in an ash cooler, and finally goes to the packing bin through a
rotary screen of 4 to 8 mesh. The screened ash is packed in double gunny bags holding 200 lbs. or
90 kilograms per bag. A portion of the unscreened ash is diverted to the returned ash conveyor to
be fed back to the dryers as returned ash mentioned above. Tailings from the screen are sent back
to the dryers with the returned ash. When the bicarbonate crystals are normal, very few lumps are
formed in the soda ash from the dryers and the amount of such tailings is very small. The amount
of the returned ash employed depends entirely on the dryness, or the “yield,” of the bicarbonate
from the filters. This in turn is determined by the crystalline character of the bicarbonate obtained
from the columns. Good filters working on good bicarbonate crystals should give a yield of 53 to
55 per cent or 12 to 13 per cent H2 O as free moisture on the weight of the filter bicarbonate.
Sodium bicarbonate as obtained from the filters is too wet to be fed to the dryers alone (i.e.,
without a proper amount of returned ash) and would form hard scale and lumps in the dryers,
causing serious difficulties. The returned ash mixed in the wet bicarbonate brings down the
moisture percentage in the aggregate and consequently causes the bicarbonate to dry to a loose
powder, forming little or no scale on the shell of the dryer. The necessity of feeding back the dry
ash with the bicarbonate naturally cuts down the capacity of the dryer and also lowers the heat
efficiency. In practice, a limited quantity, and no more, of the returned ash are employed which
will just eliminate scaling in the dryers.
Rotary cooling conveyors with inside spirals deliver cooled ash to screens and packing bins,
from which the ash is either packed in bags by means of screw packer, or loaded by means of
pneumatic conveyor into enclosed railway cars and shipped in bulk, or conveyed into great
concrete silos from which it can be retrieved by conveyors, as the case may be.
The brickwork for the rotary furnace, as in steam boiler setting, requires considerable study
and experience. The space for the combustion chamber, the relative height and location of the
bridge wall, and the spaces allowed under and around the shell for flue passage are the results of
close study and observation. Other conditions being equal, i.e., with a given grade of bicarbonate
crystals and a given grade of coal, these considerations and the grate area of the furnace determine
the drying capacity of the dryer. A standard six-foot-diameter rotary dryer has a capacity of 50 to
60 short tons of soda ash per 24 hrs. The power required to drive the rotary dryer at 4to 5 r.p.m. is

FIG 51 Position of ares of rollers with respect to axis of cylinder.

15 to 20 hp. But the capacity of a dryer is largely dependent upon the crystalline character of the
bicarbonate. A 72-inch by 60-foot rotary dryer working on a coarsegrained dry bicarbonate will
turn out 55 tons of soda ash per 24 hours with ease, whereas the same machine working on a wet
bicarbonate with poor crystals operates with difficulty at the rate of 30 tons of soda ash a day. AT
present, there is a tendency to build larger rotary dryers, 8 feet or more in inside diameter and 80
feet long, using steel plates 11 inches thick, all of welded construction.
In this connection it may not be out of place to point out the effect of the central line position
of the shafts of the rollers on the behavior of the dryer cylinder. A rotary cylinder, be it a rotary
soda dryer, a rotary lime slaker, or a rotary lime kiln, supported at two points on two pairs of
rollers, one pair at each end, is frequently found to have a persistent tendency to creep forward or
to recede while it is running. This end thrust is sometimes so great that it causes the flanged tires
or guide rollers (whichever method is used to take care of the axial thrust) to wear out very fast,

with serious results. Experience has shown that this end thrust is caused by the axes of the
rollers (roller shafts) not being placed parallel to the axis of the cylinder shell (or square to the
surface of the cylinder tire). Careful reasoning will make it evident that in Case (Fig.51) the
cylinder will tend to creep forward and in Case the cylinder will tend to creep backward, if the
rotation in each case is clockwise as viewed from the front. Reversing the direction of rotation in
each case reverses the direction of the thrust. This occurs regardless of whether the cylinder
body is placed horizontally or inclined toward the back end or toward the front end. Conditions
shown in Case, where two rollers in one pair are set in the opposite direction to the other will
not produce any axial thrust in the cylinder; nor will conditions shown in case, where one roller
in a pair is set in the opposite direction to the other. Conditions shown in Case may cause the
cylinder body to vibrate while running. In any case where the axis of a roller is not placed
parallel to that of the
cylinder body, one end of the roller face will bear on the cylinder tire harder than the other, and
that end will wear down more rapidly. If all rollers are placed square to the tire with roller shafts
parallel to the axis of the cylinder body, the cylinder shell will run very steadily and no axial thrust
one way or the other will be thrown on the guide flanges or guide rollers. The behavior of a rotary
cylinder described above can be explained on the same principle as the motion of a boiler tube
expander shaft when a mechanic turns the shaft of the expander inside a boiler tube at the tube
sheet.
In the dryers, free moisture is first driven out; ammonium bicarbonate is then decomposed
and its constituents, ammonia, carbon dioxide and water, are also driven out. The sodium
bicarbonate is only slowly decomposed and the last portion is decomposed with some difficulty.
The following determinations in Tables 80-88 were made in a small, electrically heated rotary
oven on a commercial sample of crude sodium bicarbonate from the filters and also on a sample of
C.P. sodium bicarbonate for comparison. The gases were vented to the air and in the case of the
crude sodium bicarbonate the charge was also mechanically crushed to prevent caking.
 In the decomposition of the bicarbonate one difference is to be noted, and that is the factor of
the composition of the vapor phase in contact with the bicarbonate. If the gases are vented to the
air as in the laboratory experiments, the gas phase may contain different concentrations of CO2
and H2 O from those in the closed system in actual operation, and proper allowance should be
made regarding this point in the interpretation of the results. As can be seen from the
decomposition of C.P. NaHCO3 , complete decomposition of the bicarbonate is not only a function
of temperature but it is also a matter of time (duration). It will be seen that at or above 175 . the
order of the completion of decomposition is : First the free moisture in the bicarbonate, then the
ammonium bicarbonate, then the decomposition of the bulk of the sodium bicarbonate, and finally
 FIG 53 Curves showing program of calcinations of crude sodium bicarbonate from filters.

the last portion of sodium bicarbonate, the last trace of sodium bicarbonate in the ash being rather
difficult to decompose completely.
Curves showing the increase of sodium carbonate, the decrease of sodium bicarbonate, the
decrease of ammonium carbonate and the decrease of moisture with time during calcinations at
different temperatures are plotted in Figs. 52,53,54,55,56,and 57. These temperatures are selected
 FIG 54 Curves showing progress of calcinations of crude sodium bicarbonate from filters.

as representing the whole range of the ash extract temperatures in the operation of rotary dryers. It
is seen from these observations that at 160 the crude bicarbonate can be completely
decomposed, but it requires a longer time than at higher temperatures. This is borne out in practice,
for it is found that at 160 . extract temperature, the ash can be completely freed from bicarbonate
provided that it is allowed to remain in the furnace for a sufficient time. At a high rate of operation,
whereby the time is greatly reduced, a higher extract temperature (minimum 175 .) is maintained
to insure complete decomposition of the bicarbonate. The best working extract temperature may
be said to lie between 175 and 190 .
Chemically the decomposition reactions are as follows:
NH4 HCO3 NH3 +CO2 +H2 O
2NaHCO3 Na2 CO3 +CO2 +H2 O
Any ammonium chloride in the filtered bicarbonate will be converted to sodium chloride, a
reaction representing the reverse of the tower reaction.
NH4 Cl+NaCl NaCl+NH3 +H2 O+CO2
 FIG 55 Curves showing progress of calcinations of crude sodium bicarbonate from filters.

The sodium chloride formed this way, together with that carried in the mother liquor left in the wet
bicarbonate, constitutes the total sodium chloride in the finished ash. Any incompletely settled
mud in the ammoniated brine fed to the towers would show in the soda solution as a turbidity or
precipitate. The insoluble matter in the ash made from the ammonia process is generally
magnesium carbonate and ferric oxide (Fe2 O3). Any hardness in the wash water used on the filters
adds to the insoluble content of the ash (as CaCO3 and MgCO3 ). The bicarbonate carries some
iron
 FIG 57 Curves showing progress of calcinations of C.P. sodium bicarbonate.

rust from the towers and from the ammoniated brine fed to them. All iron will be converted to
Fe2 O3 in the furnace, although in some abnormal cases where there is too large an excess of
sulfide in the ammoniated brine, the iron may appear as the gray FeS in the ash extracted from the
dryers; but it gradually turns brown later on exposure to the air. Sometimes little brown particles
containing iron may be discovered in the ash. The yellowish-red appearance of the product is due
to Fe2 O3 present in the ash, and its color is very much intensified by having a soft paste made of
the ash with a little water. A simple and quick test for the quality of the product is to dissolve 15
grams of the turbidity or precipitate and for any reddish or yellowish coloration developed in the
solution. A specification of ammonia ash of salable quality is given in Table 89.
 The above specification applies to fresh ash only, which has not been kept in storage for any
length of time. Old ash may have absorbed moisture and CO2 from the air to such an extent that
the sodium bicarbonate content would be much greater and the water-absorbing power less. A
good grade of ammonia soda ash should have the composition shown in Table 90

Again, such results are obtained only in samples taken fresh from the furnace or from
samples that have been recalcined in the laboratory to remove moisture and CO2. It is frequently
reported that soda ash from the furnace contains caustic soda to the extent of 0.1 to 0.2 per cent.
This seems to be not incompatible, especially when the bicarbonate fed is wet and the temperature
in the furnace is abnormally high.
The whole secret of dryer operation lies in having a proper amount of returned ash. It is of
course impossible to ascertain just what the exact ratio must be between the weights of the dry ash
returned to the furnace and that of the bicarbonate fed at any moment. This could only be
ascertained by observing the character of the ash extrac ted, by the shell condition inside, and by
the temperature of the ash so extracted. If the returned ash is deficient, the shell would tend to
form scale, the extracted ash would contain balls or lumps, and, as a rule, the shell expansion
would be excessive. Under these conditions the shell is likely to become overheated, and in
serious cases the forged joints or the electrically welded seams may crack and the cylinder shell
fail. On the other hand, if the returned ash is sufficient for the rate of bicarbonate feed, the ash
would be practically free from balls and come out loose and dead like flour meal, only somewhat
coarser. The grains would be loose, but not light and “lowing”, as when the ash contains much
under composed bicarbonate. The temperature of the extract then can be maintained above a point
where all the bicarbonate is completely decomposed and the extraction of the ash from the dryer
can continue practically without interruption during the 24 hours. The shell would be clean and its
expansion reduced to a minimum for the quantity of the output. The quantity of the returned ash
varies as the wetness of the bicarbonate: the lower the yield, the more returned ash is required for
a given rate of the bicarbonate feed.
When the bicarbonate crystals are poor, the bicarbonate from the filters is wet and the drying
operation becomes exceedingly difficult. With a modern filter and fair bicarbonate crystals from
the tower (3 to 4 minutes test), a yield of 51 to 54 per cent is not a difficult matter. Anything above
a 50 per cent yield, however, is fair. Very wet bicarbonate (say 28 per cent moisture) could be
dried if a large excess of the dry ash is taken with it. This amount of returned ash, however,
decidedly outs down the capacity of the dryer and raises the fuel ratio required. Table 91 gives a
rough guide as to the amount of returned ash required on the weight of the crude bicarbonate
charged into the furnace.
 When the ash is to be packed as it is produced, the ash should be packed as hot as the bag can
stand to avoid absorbing moisture from exposure to the air; a temperature of about 80 is
favorable. Hot ash from the furnace can be air-cooled in sections of long screw conveyors
generally constructed across the building. Sometimes it is carried to the top of the building and
down to give a long path for air-cooling. If an ash cooler is desired, cooling can be carried out in a
rotary cooler consisting of a steel cylinder of about 5 feet inside diameter and 25 to 30 feet long
fitted at both ends with cast-iron conical sections, through one and of which a screw conveyor
carries in hot ash and from the other end another extracts cooled ash. The cooler is inserted in
conveyor lines in a horizontal position, using angle-iron sections riveted to the shell inside to carry
the ash forward. The discharge end of the cooler is provided with scoops, which pick up the ash
and feed it to the exit screw conveyor. The cooler is supported on rollers at the extreme ends, the
arrangement being much the same as in a rotary dryer. The shell revolves in a basin of cooling
water and the heat is carried away by evaporation in the same manner as in surface cooling. Soft
water should be used in the basin; otherwise the cooler surface will be completely coated with
hard scale, resulting in poor heat conduction. Frequently a cylindrical section of a rotating cooler
is air-cooled and forms a part of the conveyor system located outside the building.
The rotary dryer is best operated under neutral conditions, i.e.; without pressure or vacuum.
The gas main is provided with an automatic butterfly valve regulated by suction in the main, and
this can be set by a counterweight to give an exactly neutral condition in the dryers. Too much
vacuum would let air leak in, giving a low gas test. Too high a positive pressure makes the gases
and steam blow out through the feed and extract openings with losses of CO2 and ammonia, and
would cause steam to condense in the feed barrel and returned ash chute, giving rise to a
troublesome plugging. To avoid dilution of CO2 by air, the extract barrel and returned ash chute
should run full to secure a proper “seal”, so that no space is left for air to be drawn in. Another
place where air leakage may occur is in the bicarbonate feed opening. With a feed table of good
design, the leakage can be practically eliminated. Butterfly valve packings, dryer neck packing
ring, and baring holes provided for chiseling off furnace scale are sometimes also sources of air
leakage. To get a high gas test from the dryer requires great effort and continued vigilance on the
part of the operators. In the old days returned gas tested only 25 to 30 per cent CO2 . With the
introduction of the Thelen pan, 60 to 80 per cent CO2 test was considered very good. In rotary
dryers it is now possible to get a 90 per cent gas test with no great effort. A rich gas from the
furnace is worth a great deal to the works, for upon this depend the formation of good bicarbonate
crystals, the high decomposition of NaCl in the towers, and the capacity o f these towers (i.e.; the
output of the plant).
The outlet gases containing ammonia, CO2 and steam, together with the mechanically
entrained soda dust, are first drawn by the suction of CO2 compressors (which at the same time are
taking kiln gases through a second branch of piping to form the mixed gases for the tower
operation) through a scrubber called the furnace gas scrubber, to dissolve out the soda dust and a
part of the ammonia and incidentally to cool the hot gases; then through a set of surface
condensers with a large number of cast-iron cooling tubes, called the furnace gas condensers, to
condense the steam; and finally through a scrubbing tower filled with coke or tiles to scrub out the
ammonia in the gas and to cool the gas. The gas then enters the CO2 compressors free of ammonia,
and cool (and consequently also dry). The filter liquor is used in the furnace gas scrubbers to scrub
the soda dust, whence it goes to the distiller, incidentally having been preheated by the furnace
gases for the distiller operation. The condensate from the furnace gas condensers containing strong
ammonia and CO2 drains to the filter liquor main on its way to the distiller. Good, soft water is
used to scrub ammonia in the final scrubbing tower, and this wash water is then used on the filters
to wash sodium chloride from the bicarbonate.
Normally, light ash has a bulk density (by the loosely-packed method) of above 0.540
(density of water equals unity). But this property has a great bearing on the condition of the
column operation as regards the proper ammonia and chlorine teeters in the green liquor fed to the
columns and the richness of the CO2 gases passed to the columns. For instance, in one case where
the chlorine titer in the ammoniated brine fell from 89-90to 85-86 (ammonia titer normal, 96-97)
because of accidental leakage of water into the absorber at its cooling box, the density of the ash
obtained was reduced to 0.464. In another case where the gas to the columns was diluted by
accidental opening of the kiln main for cleaning, the density of the ash obtained was only 0.508.
With low ammonia or chlorine titer, or both, in the ammoniated brine, or with a weak CO2 gas in
the columns, bicarbonate crystals form in the columns would be finer and the bulk density
(apparent density) of the soda ash obtained would be also lowered.
The density of soda ash depends on the physical state of aggregation. It is found that soda ash
absorbs water and has a tendency to become “set” or conglomerated, i.e.; to from crystals of the
monohydrate, Na2 CO3 .H2 O. If then water is again driven out, the conglomerate structure remains.
Hence the present method of making dense ash is by adding about 16 per cent water to the ash,
mixing thoroughly in a screw mixer, and drying the mass again in a rotary dryer. The amount of
water added corresponds roughly to the formation of the monohydrate, Na2 CO3 .H2 O, and the
conversion of the light ash to the monohydrate crystals explains the fact that the light ash becomes
granular and dandified. It is found that, to obtain the desired density, one-half of the dandified ash
should be fed back with the light ash in the dense ash furnace in the same way as the returned ash
in the light ash dryers. The rotary type of dryer is very suitable for this work and has an even
greater capacity on dense ash making. The outlet gas (which contains steam and a small amount of
air) is drawn through a surface condenser by an exhauster, discharging it to the air. The steam
condensate is used over again in the light ash feed. The density of the ash is brought up to 0.95 to
1.0(loosely packed). This dense ash, therefore, is also called “water ash”. The dense ash,
especially the granular dense ash is favored by glass manufacturers, because it possess advantage
over the light variety in that (1) it gives larger charging capacity for the melting posts, (2) it gives
a longer time for reaction, and (3) it causes less dust loss.
In operating the dense ash furnace, the extract temperature should be maintained at 135 to
155 . It should not be permitted to rise above 160 ; or the density of the ash obtained may fall.
The screw conveyor system used for conveying the dense product must be arranged so that it is as
short as possible to avoid any unnecessary grinding action on the dense ash particles. The dense
product is then passed through a hummer screen to grade the different sizes for packing.
 It was noted above that in order to insure complete decomposition of sodium bicarbonate the
temperature of the ash extract must be maintained between 175 and 190 . The temperature in
the front of the dryer is of course much higher. A temperature of 20 is commonly given for the
complete decomposition of NaHCO3 . Gautier [ Ber. Deut. Chem.Ges; 9, 1434(1876)] found that at
100 to 110 decomposition of sodium bicarbonate could go to completion. But this, of cause,
is a matter of time as noted above, and is also dependent upon the rate of removal of CO2 and
water. In practice, under good working conditions of the dryer, the gas obtained from the dryer
main is over 90 per cent CO2 ; and the partial pressures of CO2 and water in contact with the solid
bicarbonate inside the dryer are probably around 0.25 and 0.65 atmosphere respectively. the flue
gases at the tail end of the furnace are usually at 350 to 450 ., depending upon the furnace
setting and the draft, which the furnace carries.
In the dryer reaction
2NaHCO3 (solid)=Na2 CO3 (solid)+CO2 (gas)+H2 O(gas)
NaHCO3 and Na2 CO3 being in the solid phases,
Pco2 × Ph2 o (expressed in am)=constant at a given temp.
=0.23 at 100 .
Theoretically, there are three phases (two solid phases and a vapor phase) and three
components (Na2 CO3 , CO2 ,and H2 O),and the system is therefore invariant. That is to say, the
equilibrium depends not only upon the temperature but also upon the composition of the vapor
phase (molal concentrations of CO2 and water vapor). With wet bicarbonate there may be also a
solution phase, and as long as the water (free moisture) is present in the bicarbonate this solution
phase remains. Then, according to the phase rule, this system would be invariant. Since the
decomposition can take place only when the product of the partial pressures of CO2 and of water
vapor in the vapor phase is less than a certain value(such as 0.23 at 100 ), it is easily seen that
when the bicarbonate from the filters is very wet the partial vapor pressure of water in the dryer
may become very high, ad that the product Pco2 × Ph2 o may then be nearly equal to the
equilibrium value, in which case the rate of decomposition
*“Chemical Principles” by A.A. Noyes and M.S. Sherrill, P.166. The Macmillan Company New York.
(of calcinations) would be very low unless the dryer shell is subject to a much higher temperature.
This is why it is essential to keep the moisture content of the bicarbonate as low as possible, to
facilitate the dryer operation and not “punish” the shell, to say nothing of the excessive fuel
consumption in the dryers.
The heat required to decompose 2 mol of NaHCO3 , giving 1 mol Na2 CO3 , 1 mol CO2 and 1
mol water vapor all at room temperature, is :
2NaHCO3 (solid) Na2 CO3 (solid)+ CO2 (gas)+ H2 O(gas)+Q
(-2×227,700) (-272,600) (-94,400) (-57,800)
Q=-30,600 Cal.
To estimate roughly the theoretical consumption of the coal required per 1000 kg of the
bicarbonate from the filter, given the following data:
Filter bicarbonate at 53% yield
Free moisture in bicarbonate …………………………………………………… .14%
NH4 HCO3 in bicarbonate ……………………………………………………… ...3.5%
Room temperature…………………………………… …………………………… 25
Temperature, ash extract ………………………………………………………… ..180
 Temperature, flue gases …………………………………………………………………
380
Temperature, gases from dryer …………………………………………………………
.200
Coal, bituminous slack ………………………………………………… ...7,00 kg. Cal. Per kg.
Heat of decomposition of NaHCO3 to CO2 and H2 O (gas) ……… 30,600 kg. Cal. Per kg. Mol
Sp. ht. Flue gases ……………………………………………………………………… .0.024
Sp. ht. Soda ash………………………………………………………………………… 0.256*
Sp. ht. CO2 gas ………………………………………………………………………… .0.21
Sp. ht. Steam ……………………………………………………………… …………… .0.47
Latent heat of evaporation of water at 25 to steam at 25 ………………… .582 kg. Cal.
Radiation loss from furnace assumed at 25 per cent of total heat supplied.
Flue gases from combustion at 80 per cent excess air (including water vapor) 15kg. Per kg. Of
coal+
That is, 0.15 ton of coal is required per ton of ash calcined. The fuel consumption, however,
depends also upon the quality of coal burned, the method of firing (hand firing, stoker firing or
pulverized coal firing), the construction of the furnace (brick-setting and the construction of flue
passages), and the manner and the rate of operation. The above calculation is a very rough one,
one biggest factor being the unknown radiation losses from the furnace, which are assumed to be
25 percent.
In practice, the fuel consumption is 20-25 per cent per ton of soda ash calcined, depending
upon the dryness of the bicarbonate and the quality of coal used. The efficiency is therefore even
lower. One American manufacture in the South has been successful in performing this operation
with high-pressure, steam-heated equipment and obtains a higher overall efficiency in this
operation (i.e.; in the combination of steam boiler and rotary calciner).
The foregoing is a description of the decomposition of sodium bicarbonate into normal
carbonate by calcinations. Very recently alkali manufacturers in the United States are installing
so-called “decomposers” for converting the crude sodium bicarbonate (ammonia soda) directly to
the carbonate in solution, furnishing soda solutions for the manufacture of both caustic and
bicarbonate o f soda. This we have called “Wet Calcination”(see Chapter XXI).
 Chapter XIV
Recovery of Ammonia-----Efficiency of Operation in

Ammonia Still
The price of soda ash is normally about $30 per ton but that of ammonium sulfate $30 to
$40.* Based on a 25 per cent ammonia content, the price of ammonia is $120 to $160 per ton, or
about 4 to 5 times that of soda ash. Evidently we cannot afford to lose much ammonia in the
manufacture of soda ash. Ammonia is a comparatively expensive commodity. One of the reasons
for the failure of the ammonia soda process earlier in the history was the lack of an efficient
method of recovering ammonia from the mother liquor and the inability to minimize ammonia
losses during operation. The heavy losses of ammonia in the cycle made it impossible for the
process to compete with the LeBlanc process. It is now a systematic practice in ammonia soda
plants to save all ammonia-bearing liquors and send them to the distiller for recovering NH3 . So
much attention is now devoted to saving ammonia that it has become almost a matter of habit on
the part of plant operators to consider the loss of sodium chloride as a very minor matter in
comparison with that of ammonia.
Naturally in the early history of the ammonia soda process the loss of ammonia was great.
Toward 1880, the loss of ammonia was still in the neighborhood of 4 to 8 per cent as (NH4 )2 SO4
on the weight of soda ash made. Now the conservation of ammonia in the soda ash plants has
received so much attention that the loss is only from 0.20 to 0.40 per cent as (NH4 )2 SO4 on the
soda ash produced.
The draw liquor contains from 65 to 74 titer of fixed ammonia and 22-26 titer of free
ammonia. In other words, there are approximately a little less than 3 parts of fixed ammonia to 1
part of free ammonia. But the proportion of free ammonia to fixed ammonia in the feed liquor is
increased by (1) the addition of the furnace condenser condensate which contains high ammonia
and ammonium carbonate;(2) the absorption free ammonia when a portion of the filter liquor is
employed in scrubbing furnace gasses (in the furnace gas scrubber);(3) the addition of crude liquor
which contains only the free ammonia (but little fixed ammonia) to replace the ammonia lost in
the system; and (4) the addition of sodium sulfide solution to make up for the deficiency of sulfide
in the ammoniated brine system. The sodium sulfide converts an equivalent quantity of fixed
ammonia to “free ammonia.”
Na2 S + 2NH4 Cl 2NaCl + (NH4 )2 S
* At present, the price of ammonia because of the production by the direct synthesis is rather low, but the
demand and supply will adjust themselves as times getnormal.
On the other hand, if crude liquor is not available, ammonium sulfate solution is generally made
from ammonium sulfate crystals and added to the filter liquor for distillation. This then would
increase the ratio of fixed ammonia to free ammonia in the feed liquor to the distiller. On the
whole, however, the proportion of free ammonia to fixed ammonia is considerably increased as
the filter liquor arrives at the distiller, and the distiller feed may have the ratio of free ammonia to
fixed ammonia as high as 3 to 7. At the same time, the liquor on its way to the distiller is diluted,
as far as Cl- is concerned, by :
(1) wash water on the filters;
(2) steam condensate from the furnace condenser drain line;
(3) steam condensate from gases in the furnace gas scrubber;
(4) water contained in the crude liquor, or in the ammonium sulfate solution, added to replace the
ammonia lost.
(5) water contained in the sodium sulfide solution added when the sulfide in the crude liquor is
insufficient.
The filter liquor is pumped up to feed the distiller through an orifice regulation.
The composition of one sample of crude ammonia liquor is given in Table 92

There are two types of modern ammonia column stills or distillers, one of the internal
overflow type and the other of the external overflow type, both consisting of a number of
compartments or rings with mushrooms, division plates, and overflow pipes. These rings are
placed one above another like a sort of column or tower (Fig. 58). The internal overflow type has a
pipe inside each division plate, and the upper end of the pipe controls the depth of liquor or
barbotage on the division plate, while the lower end dips well below the surface of the liquor in
the next compartment below to give the necessary seal. These overflow pipes alternate from one
division plate to another in a diametrically opposite position, so that the liquor is obliged to travel
across each plate and down in a zigzag manner (Fig. 59).
The external overflow type has these overflow connections outside of the distiller from one
ring to the next below, alternating in diametrically opposite positions as above. The edge of the
overflow opening in this case controls the height of the liquor in each section and the inlet opening
into the section below is located well below the level of the liquor line. These overflow pipes are
usually rectangular in cross-section, long horizontally and narrow vertically; and in order to
provide the most direct flow for the lime sludge which has a great tendency to plug the passage,
the horizontal part of the shoulder connecting the overflow pipe with the side of the distiller is
made as short as possible. In fact, one construction has the overflow trough directly attached to the
side of the distiller, making the outside wall of the distiller the bottom side of the rectangular
overflow trough.
The advantages of the internal overflow type are a more direct passage for the thick lime sludge to
travel from one section to another below, and no heat losses by radiation from these overflow
conduits. The advantages of the external overflow type, on the other hand, are that a clear space
inside the distiller is available for ammonia distillation and that the overflow passages are more
accessible. At present the external overflow type is common in soda plants (Fig. 60).
The size of the mushroom in each compartment is dependent on the shape of its sides or
wings. The diameter of the mushroom at the base of its serrated edge is generally from 55 to 70
per cent of the inside diameter of the compartment. For example, for a 2500-mm inside diameter,
the diameter of the mushroom may be from 1400 to 1700 mm. For a mushroom with a steep slope
at the sides, of the type shown in Fig.59, the diameter should be smaller; while for a mushroom
with widespread sides of the type shown in Fig.61, this diameter takes a somewhat higher ratio.
Often the bottom of the compartment (the division plate) is curved in the
manner shown in Fig.60, the idea being so to direct the flow o f steam that it may come into
intimate contact with the whole volume of the liquor and may spread itself in fine bubbles in the
liquor, encountering least resistance in its radial sweep. With a large number of mushrooms placed
one on top of another throughout the height of the distiller, the resistance may be considerable,
especially if the depth of the liquor in each compartment (the depth of wash or barbotage) is
considerable. For ammonia absorption, such as in an absorber where the gas is to be dissolved by
the brine, the depth of the wash should be greater; but for a distilling operation as in the distiller
discussed here, the quantity of the liquor carried in each compartment should be small.

FIG 61 “Bubble-Cap” type of heater sections.

If the diameter of the compartment is large, a multiple system of mushrooms is provided to

distribute the steam evenly over the entire area of the compartment; in this case, either several
long, narrow mushrooms or several small, circular mushrooms are placed at equidistant points on
the division plate.
One important point that should be observed in the design of the distiller is so to proportion
the mushroom dimensions and so to locate the overflow points that the liquor shall have no
tendency to be blocked up by the upward flow of steam. A condition in which the upper
compartments tend to run full and the liquor refuses to descend readily is known as the
“suspension” of the liquor (similar to the “gas-locking” or “gas -binding” in the absorber).
Normally, this is caused by the deposit of scale or sludge at the overflow points; but with faulty
design of the bubble-caps and division plates, this tendency is greatly increased, with the result
that steam enters the liquor passage and blocks its downward flow.
A column distiller for soda manufacture consists of three main parts: the upper part above the
lime inlet, called the “heater”; the lower part below the lime inlet, the lime still, or the distiller
proper; and the partial condenser located on the top of the heater. The heater is to vaporize “free
ammonia” from the filter liquor by steam, and the lime still is to decompose the fixed ammonia by
milk of lime and vaporize the free ammonia so liberated by steam. The partial condenser is to
provide the overflow to the top of the heater and to secure richer (drier) ammonia gases distilled
over. Under “free ammonia” are included (besides free NH3 ), NH4 HCO3 , (NH4 )2 CO3, (NH4 )2 S,
etc., i.e., all the ammonium compounds that are readily decomposable and volatilized by steam.
The fixed ammonia, which requires lime is mostly in the form of NH4 Cl, and in small quantities
also in the form of (NH4 )2 SO4. This column type of distiller was the invention of Dr. Ludwig
Mond, who designed and built it in 1883.
The older forms of lime still consisted of four large tanks in series. Steam entered the one
containing the weakest ammonia liquor, and the outlet from this went to the next, and so forth.
When all the ammonia in one unit had been boiled out, it was cut out and the content emptied.
Steam entered the next one in the series as a head tank. The empty kettle was filled with filter
liquor and the requisite quantity of milk of lime, and became the last one in the series. This is a
batch process. The handling required considerable labor. Although there might be a saving in lime
by being able to control the quantity of lime in each batch, the time and labor involved made it
unsuitable for any large-scale production.
The reactions in the heater are as follow:
(1) NH3 .Aq NH3 gas
(2) NH4 HCO3 NH3 +H2 O+CO2
(3) (NH4 )2 CO3 2NH3 +H2 O+CO2
(4) NaHCO3 +NH4 Cl NaCl+NH3 +H2 O+CO2
(5) Na2 CO3 +2NH4 Cl 2NaCl+2NH3 +H2 O+CO3
(6) (NH4 )2 S 2NH3 +H2 S
(7) NH4 HS NH3 +H2 S
(8) Na2 S+2NH4 Cl 2NaCl+2NH3 +H2 S
(9) Na2 S+CO2 +H2 O Na2 CO3 +H2 S
The reactions in the lime still or distiller proper are:
(10) Na2 CO3 +Ca(OH) 2 2NaOH+CaCO3
(11) CO2 +Ca(OH) 2 CaCO3 +H2 O
(12) 2NH4 Cl+Ca(OH) 2 CaCl2 +2NH3 +2H2 O
(13) (NH4 )2 S+Ca(OH) 2 CaSO4 +2NH3 +2H2 O
The following reaction:
(14) CaCO3 +2NH4 Cl(or(NH4 )2 SO4 ) CaCl2 (or CaSO4 )+2NH3 +CO2 +H2 O
requires prolonged cooking over a long period to get satisfactory completion. It takes place only
when finely divided calcium carbonate, such as the precipitated carbonate from the causticization
tanks, is used in excess and also when the ammonium chloride concentration is high. Under the
prevailing conditions of distiller operation, however, such a reaction does not occur to any extent.
For the same reasons, with the insoluble magnesium oxide in milk of lime, the reaction
(15) MgO+2NH4 Cl MgCl2 +2NH3 +H2 O
can hardly take place in the lime still under ordinary conditions.
The heater is either packed with coke, about 6 inches in size, or with earthenware tiles, the
so-called chemical tiles. The spiral tiles made of stoneware are best of 5 inches inside diameter by
6 inches outside diameter and 6 inches high, with a spiral ribbon inside (Fig.62). some heaters also
 FIG 62 Spiral chemical tile. Fig 63 “Spray dish ”type of heater sections.

have mushrooms and division plates (the “bubble-cap” type) as in the distiller proper (Fig.61),
while others have, instead of these, inverted spray dishes having serrated edges with serrated
overflow openings in the center of the division plates underneath (Fig.63). still others have
sections consisting of long, narrow mushrooms on each division plate (as in Fig.38,p. 170). A
good distribution method is to use a large number of long and narrow distributing troughs or trays,
packed one on top of the other and staggered over each other, in the manner of the packing in
some of the deaerating heaters for the liberation of oxygen from boiler feed-water (see Fig. 64).
But in a large majority of cases a combination of these “bubble” or “spray” devices and the tile or
coke packing is used. Coke packing offers more resistance to the gas passage than tile packing and
may build up excessive pressure. The coke-packed heaters while working properly when properly
packed, are likely to become partially plugged by dirt. When this happens, steam blows to one
side and channelling occurs, when a portion of the coke may be blown up forming a pit on the top
of the packing. The distiller proper (the lime still), on the other hand, always has the simple
mushrooms and the division plates (or bubble-caps and plates).
On top of the heater are sections of the coolers and condensers, usually consisting of a large
number of horizontal 2-inch cast iron tubes (Fig.65). These are divided into two main divisions,
the upper division of which uses water as the cooling medium, while the lower division uses the
filter liquor as the cooling medium so that the feed liquor is preheated on its way to the heater.
These

are partial condensers for the ammonia gases, and the condensate is allowed to drain back to the
heater. The feed liquor is thus heated to about 70 . before entering the heater. The gases from the
distiller consist of ammonia, carbon dioxide, steam, and some hydrogen sulfide and air. The steam,
of course, should be condensed out as far as possible, but with it also a portion of the ammonia
vapor. Hence the condensate returned to the distiller is necessarily a strong ammonia solution
carrying some CO2. The purpose of the distiller coolers and condensers is to condense out most of
the steam, thereby obtaining a sufficiently dry ammonia gas for the absorber operation. Hence the
top portion of the distiller, including the coolers and condensers and a part of the heater, may be
called a rectifier of dephlegmator. As was explained in detail in ChapterVII, excessive steam
carried over by the ammonia gas from the distiller would unnecessarily dilute the brine and
generate too much heat in the absorber. Sometimes the distiller coolers and condensers are
installed separately and located below the distiller top, but their proper place is on the top of the
distiller so that the condensate may be drained back directly to the heater where the strong
ammonia liquor properly belongs. When they are installed separately, they are placed high enough
so that the condensate may flow into the heater through a U-loop seal. This is done to avoid
pumping the hot, corrosive ammonia carbon dioxide condensate back to the top of the distiller.
The distiller coolers and condensers are subject to very severe chemical corrosion because of the
presence of ammonia, CO2 , H2 S, etc., carried by the ammonia gases at an elevated temperature.
The cooling tubes must be made of a food grade of cast iron (not steel) to withstand the corrosion.
Liquor from the bottom of the heater, which should be free from CO2 , flows to the top of the
lime still or the prelimer, where milk of lime is introduced to decompose the fixed ammonia; and
the free ammonia so liberated is distilled off by steam just as in the heater above, as the liquor
flows down from one passette to another in the lime still. In actual operation, in the arrangement
in which the heater is placed on top of the lime still, not all the free ammonia can be driven from
the heater liquor. On arriving at the bottom of the lime still, however, all ammonia should have
been distilled out, leaving the waste liquor with a small excess of lime to be run off to waste. Open
steam enters at the bottom compartment in the lime still diametrically opposite the blow-off
opening. In actual practice, exhaust steam from the engines, or better from the bleeder turbines, is
employed for this purpose. The engine cylinders are so designed as to give 8 to 10 pounds back
pressure in the exhaust. Generally, there is a tendency toward shortage, rather than surplus, of the
exhaust steam for distillation, and at times some live steam may have to be introduced into the
distiller. Live steam is not so suitable as the exhaust steam which, by virtue of its low pressure, has
a large volume, giving a large contact surface and more even bubbling action in the distiller. The
exhaust steam, however, must be dry and the condensate trapped out in order to avoid unnecessary
dilution in the distiller.
The rate of decomposition of the fixed ammonia by lime is wholly dependent upon the rate of
hydration and solution of CaO in the lime still. While the ionic reaction is instantaneous, the
solution of CaO is slow; which, therefore, controls the speed of its reaction with fixed ammonia.
CaO solid CaO diss Ca++ +2OH-
Ca++ +2OH- +2NH4 + Ca++ + 2NH4 OH
Lime has the peculiar property of becoming less readily soluble when burned at a high
temperature, i.e; when it is overburned or dead-burned it dissolves with great difficulty. In order to
make more lime available for ammonia liberation, the time of contact in the liquor should be
lengthened. This is accomplished by having a large reservoir outside where the heater liquor is
intimately mixed with the milk of lime and allowed much longer time of contact before entering
the lime still. This reservoir, called the prelimer or preliming tank, is provided with a stirrer to
keep the milk of lime in suspension, and the heater liquor in intimate contact. By gradual
displacement, the top portion of the liquor overflows to the lime still. Thus, a major part of the
reaction is caused to take place in the prelimer outside, returning virtually free ammonia liquor to
the lime still for distillation. There are several advantages in having the prelimer. First, it gives a
more thorough mixing of milk of lime and heater liquor; secondly, it allows more time for lime to
dissolve and react with the fixed ammonia; and thirdly, it affords easier control of the excess of
lime in the distiller, avoiding sharp fluctuation in the lime used for distillation.
The volume of filter liquor by the time it reaches the distiller is about 7 cubic meters per ton
of ash. Table 93 gives its approximate composition after the liquor has received various ammonia
condensates and liquors.

The heater liquor contains all the fixed ammonia and sodium chloride, and should contain no
CO2 or free ammonia. But in practice, as mentioned above, there still remains about 4 titer of free
NH3. The freedom from CO2 in the heater liquor is a test for the efficiency of the heater, and is
generally dependent upon (1) the temperature in the heater, (2) the intimate and thorough contact
between the liquor and the steam in the heater, and (3) the time that the liquor is exposed to the
action of steam, assuming a uniform gradient of the partial vapor pressure of ammonia throughout
the heater. The heater liquor should be practically free from CO2 before entering the prelimer, or
the calcium hydroxide would be converted to calcium carbonate, which is unavailable for
ammonia liberation in the lime still. For the following reaction, as mentioned above,
CaCO3 +2NH4 Cl CaCl3 +2NH3 + CO2 +H2 O
does not take place under normal operating conditions in the lime still.
The pressure at the bottom of the distiller should be from 8 to 10 lbs. gage (sometimes as
high as 12 lbs.); and there should be a slight vacuum at the top of the distiller, about 1 /4 inch Hg,
so that the pressure in the distiller decreases from 8 to 10 lbs. per sq. in. at the bottom to a slight
vacuum at the gas outlet at the top. The use of a slight vacuum at the top decidedly helps the
vaporization of ammonia in the liquor. The whole distiller and absorber system must be airtight; or
air will leak in and render the exhaust gas from the absorber system too weak in CO2 to be
returned to the carbonating towers. In any case a slight vacuum is to be preferred both from the
viewpoint of avoiding loss of ammonia gas to the atmosphere and of the more rapid removal of
ammonia vapors from the distiller. The top temperature is 82 to 86 .
Roughly speaking, as described above, the heater and the condenser may be said to
correspond to a rectifier and the lime still to the distiller proper. However, as it is actually
conducted in ammonia soda plants, the function of each portion of the distiller is not so simple.
When the heater is set on top of the lime still integrally, the rich ammonia gas from the lime still
below is not (and cannot be) in equilibrium with the liquor in the heater. The heater then cannot be
purely a rectifier or dephlegmator, because while receiving the overflow at the top from the partial
condenser, it also receives dilute feed liquor. Its function is primarily to distill “free ammonia”
from the feed liquor, and consequently cannot be simply one of rectifying or dephlegmation.
Again the heater liquor containing fixed ammonia is limed in an outside tank (preliming tank or
prelimer) and then returned to the top of the lime still, containing some excess lime. This is really
another feed to the distiller introduced at another point of the distiller body. Hence the whole
distiller is in reality a two-feed column (filter liquor at the top of the heater and a limed liquor at
the top of the lime still coming from the prelimer), entailing complexities inherent to a two-feed
system-nevertheless a simple two-component (ammonia-water)system where the composition in
the vapor phase from the top of the lime still is not in equilibrium with that in the liquid phase
from the bottom of the heater, and where the rich ammonia gas from the to of the heater is not in
equilibrium with the rather cold, dilute filter liquor introduced there.
As the bottom of the distiller is under a pressure of 8-10 lbs. gage, the liquor is therefore at a
temperature above the boiling point under atmospheric pressure. When, therefore, the liquor is
blown out and is released to the atmospheric pressure, it partially flashes into steam, which can be
utilized for distillation at a lower pressure. Indeed, in practice, the blow-off is first blown into a
flash tank near the bottom of the distiller and the steam from the uptake of the uptake of the flash
tank is utilized to operate a weak liquor distiller which is merely a small still about 1 m. in
diameter and 8 m. high, distilling floor liquor and other weak ammonia liquors. Such a weak
liquor distiller has been installed in most ammonia soda plants, utilizing the blow-off steam.

In the lime still or the distiller proper, hard scale is likely to form and the distiller needs to be
shut down and cleaned every so often. This happens once in 1 or 2 months, depending upon the
rate of operation and the quality of lime used. Frequent interruption in the distiller operation
would cause solids to settle in the lime still, necessitating more frequent cleaning. The scale
formed in the upper portion of the lime still contains much calcium carbonate; while calcium
sulfate predominates below, although lime is also present in considerable quantities.
Table 94 gives the results of analysis of a scale taken near the bottom of the distiller.
 FIG 66 Analyses of distiller liquor showing condition of operation.

This scale is very hard and attains a thickness of 1 to 2 inches. It has to be chiselled off by
hand or by a pneumatic scaling hammer. Before shutting down a distiller, it should be cooked for
several hours with steam to drive out all ammonia. The distiller is then cut out and allowed to cool.
The manhole covers are opened and the cleaning work is started.
Table 95 gives some idea about the ammonia concentrations in the different samples at
different points in a distiller in operation. Figure 66 shows the ammonia gradient in the liquor
phase in the lime still and the prelimer.

As to the composition of the liquors in different rings of the distiller, Bradburn’s * data in
Table 96 illustrate the conditions of operation in the distiller at his time.

*Ullmann, “Enzyklopadie der technischen Chemie,” Vol. 8, p.31, Berlin, Urban & Schwarzenberg , 1931.

The distiller waste, or run-off liquor, contains all the undecomposed salt in the filter liquor,
calcium chloride in a quantity equivalent to the
amount of sodium chloride decomposed, a small quantity of calcium hydroxide(2 to 4 grams per
liter) from the slight excess of lime used to insure the complete decomposition of fixed ammonia
in the lime still, and all the impurities of sand, unburned and overburned stone, etc. present in the
lime. The composition of the waste liquor varies greatly according to the quality of brine and
limestone used, the volume of the feed liquor to be distilled per ton of ash, the percentage of
decomposition in the towers, and the strength of the milk of lime employed. Normally, there are
about 10 to 12 cubic meters of the waste liquor per ton of ash made. (see Table 97.)

The composition of the distiller waster also varies greatly. Its approximate composition is
shown in Table 98.
Table 98. Composition of Distiller Waste.
Sp. gr. of waste liquor 1.10-1.13
CaCl2 85-95 grams per liter
NaCl 45-50 grams per liter
 CaCO3 46-15 grams per liter
CaSO4 3-5 grams per liter
Mg(OH) 2 3-10 grams per liter
CaO 2-4 grams per liter
Fe2 O3 and Al2 O3 1-3 grams per liter
SiO2 1-4 grams per liter
NH3 0.006-0.012 grams per liter

The composition of the distiller waste liquor and the relative concentrations of various
substances therein is a barometer indicating the efficiency of the whole ammonia soda process.
For, from the analysis of this liquor, it is possible to obtain many data bearing on the consumption
of raw materials employed and consequently the efficiency of the operation. For instance,
(1) Low specific gravity of the liquor and the proportionately low concentrations of various
substances present indicate a large volume of feed liquor to be distilled and consequently a
high consumption of steam and coal, and a greater ammonia loss per ton of ash.
(2) Low ratio of calcium chloride to sodium chloride indicates low decomposition in the
carbonating towers (or columns) and high wastage of sodium chloride.
(3) Presence of much ammonia indicates a heavy loss of ammonia through the distiller.
(4) High percentage of lime indicates a large excess of milk of lime and high consumption of
limestone.
(5) High percentage of calcium carbonate indicates inefficient operation in the heater (from CO2
in the heater liquor) and high consumption of lime and coke.
(6) Large amounts of insoluble solids in the waste liquor relative to calcium chloride, such as
magnesium oxide, silicon dioxid e, ferric oxide, and aluminum oxide (al of which must have
come from the milk of lime), indicates poor quality of limestone being used.
The distiller waste liquor carries considerable solids in suspension. The disposal of this waste
is a matter of no little concern to ammonia soda manufacturers. The solids accumulate very fast,
and the liquid portion containing free lime and calcium chloride, if allowed to run into a steam,
may kill the fish and contaminate the waster. Towns where the industry is located will pass laws
forbidding the discharge of this waste to the river in the community. The general practice is to
employ a large tract of land, 100 acres or larger, adjoining the factory site and pipe the waste in 8-
to 12- inch steel pipes first to one-half of this field and then to the other. Settling takes place and
the solids are left behind. Only the liquid portion is allowed to flow out. How fast these solids
accumulate can be appreciated from the fact that one plant producing 100 to 150 tons of ash daily
filled a piece of land 10 acres in area to a height of 20 feet in about 9 years. As the solids build up
in one half of the land, they are dug out and thrown on the sides, making an embankment
surrounding it. Meanwhile, the discharge goes to the other half. As the land is being filled up, the
elevation becomes too much for the gravity flow of the liquor from the bottom of the distiller. A
centrifugal booster pump is then employed to give the head necessary to discharge to these fields.
If limestone quarries lie near at hand, the waste is best sent to fill the old quarry pits.
The waste as shown above contains chiefly calcium chloride, sodium chloride, calcium
carbonate, silicious matter, and lime. Attempts have been made to work up this waster liquor for
anhydrous calcium chloride, refined salt, etc. In fact, several ammonia soda plants are doing this.
But the limited demands for calcium chloride in comparison with the quantity produced can utilize
hardly a fourth or a fifth of the total output. The balance has to be disposed of.
The blow-off liquor from the bottom of the distiller is at a temperature slightly above the
boiling point of plain water. Its sensible heat is therefore enormous because of the very large
volume involved. This will later be shown by the fact that it takes about 64 per cent of the total
heat required just to heat up the liquor. The heat in the blow-off liquor should therefore be utilized
for preheating, but unfortunately the liquor contains considerable suspended solids which may
settle and form scale in a preheated or heat exchanger, making the scheme impracticable. Some
heat, however, is utilized in evaporation when the clarified blow-off liquor is evaporated for the
recovery of calcium chloride and salt.
The distiller operation is rather delicate in its control, and good judgment must be exercised
at all times. The rate of feed, the varying composition and concentrations of the feed liquor, the
adjustment of the quantity of milk of lime to give the proper excess, the varying strength in the
milk of lime----all demand close attention on the part of the operators. If there is a deficiency of
steam, ammonia may not be completely driven out before the liquor reaches the bottom. This is a
phenomenon called the “Sour Bottom,” in which case the sour liquor at the distiller bottom must
be temporarily emptied out to a pit to be pumped back to the distiller later; otherwise much
ammonia would be lost in the waste. Also in such cases CO2 may not be completely driven out
from the heater liquor, and the sulfide from sodium sulfide added may not be completely
decomposed: some may remain in the distiller waste and be lost. If, on the other hand, too much
steam is used, it may be beyond the capacity of the condensers and coolers to cool, the gas outlet
temperature would be too high, and the gas would carry over too much steam to the absorber.
Normally the gas enters the absorber at a temperature of 55 to 60 . A temperature much lower
than 55 may, if allowed to remain for any length of time, cause crystallization of “ammonium
carbonates” both in the distiller cooler passage and in the gas mains leading to the bottom of the
absorber, thereby plugging up the system. On the other hand, a temperature above 60 would
bring too much steam condensate to the absorber, unnecessarily diluting the brine, with the result
that the Cl- titer in the ammoniated brine falls low. This, of course, causes low decomposition in
the column because of low concentration of Na+ in the green liquor. A serious case that sometimes
ensues is the “hot-top” condition in the absorber created by the heat thus carried to the absorber
from the condensation of much steam.
Further, if there is a deficiency in lime, there will be some undecomposed fixed ammonia at
the distiller bottom. The liquor, of course, would then smell “sweet,” but ammonia may have been
sent to waste unnoticed. As ammonia is an expensive material and normally the still is the largest
single source of ammonia loss, the liquor at the still bottom should be continually titrated for an
excess of lime and tested for “sweetness” by its odor as often as every few minutes.
Regarding the corrosion of the apparatus by ammonia, Pennock and Morton* found that if
the iron surface was clean and free from rust, ammonia not only would not corrode the iron but
would protect it. If, on the other hand, corrosion of iron had started, ammonia would accelerate it.
It is therefore necessary to have a ferrous sulfide coating on the exposed surface of the iron to
prevent corrosive action.
* J.Am. Chem. Soc., p. 377 (1902)
EFFICIENCY OF AMMONIA OPERATION
The distillation of ammonia from the filter liquor is an example of separating amore volatile
constituent from a less volatile one, when the two are soluble in each other in practically all
proportions, to form one liquor phase. It is a binary solution with one liquid and one gaseous
phase. The system is therefore divariant. The fixing of any two factors, or independent variables,
out of the three —temperature, pressure, and composition in the liquor —will thus determine the
whole system. These are generally the temperature in the still and the composition in the liquor,
the two being as far as possible maintained constant or uniform in the distiller operation.
Aqueous ammonia solutions behave quite normally. The ammonia concentration in the vapor
phase in equilibrium with the solution is consistently greater than its concentration in the solution
at all temperatures and in all concentrations, there being no maximum or minimum constant
boiling point for the solution. In a dilute solution its vapor pressure may be said to follow
Henry’s law.

That is, if c is the concentration of ammonia in solution, and C that in the vapor, C>c at all
concentrations in the solution (Fig.67).The partial vapor pressure of ammonia in equilibrium with
dilute solutions at 30°C, within the range of from 0 to 40 grams NH3 per liter, may be expressed as
follows:
p=0.99C
Where
P =partial vapor pressure of NH 3 in mm Hg
C=conc. of NH 3 grams per 1000 grams of H2 O
As described above, the modern ammonia distiller consists of a lime still and a heater set on
top of the lime still, receiving ammonia vapors arrangement, it is not done without some
sacrifice .For, while a countercurrent arrangement in the heat transfer is thus maintained, this is
not true of the ammonia vapor flow, and also frequently is not true of the introduction, of the feed
liquor.
In order that a distiller may operate efficiently the following conditions must be satisfied:
(1) There shall be a countercurrent heat supply to the distiller.
(2) The distiller body shall be insulated against radiation losses.
(3) The partial pressure of ammonia in the vapor rising from bottom to top shall not at any
point be greater than the ammonia vapor tension in the liquor with which the vapor comes
 in contact.
(4) The feed liquor shall be preheated and introduced at a point where its ammonia vapor
tension is greater than the partial pressure of ammonia at that point.
(5) The volume of filter liquor to be distilled and of the milk of lime employed per ton of
soda ash shall be as small as possible, i.e., both of these shall be as concentrated and hot as
possible.
(1) Countercurrent heat supply. Although ammonia gas is readily expelled from the
solution, the last trace of it is driven out with considerable difficulty. Hence, as ammonia in the
liquor gets dilute, more steam and a higher temperature are required to complete the work with a
given excess of lime. The supply of steam is so arranged that the weakest ammonia liquor meets
the fresh steam, and is subject to the highest temperature available, before the waste liquor is run
out. The supply of exhaust steam in the form of open steam is introduced at the bottom of the lime
still and the countercurrent feature is closely adhered to in practice.
(2) Lagging of distiller body against radiation. The vapors of ammonia and steam in the
distiller rise and gradually meet the stronger overflow liquors from the upper division plates. In
equilibrium with the increasing ammonia concentrations in the overflows above, the vapor in
contact with the liquors is gradually enriched in ammonia. When the vapor strikes the condensers
at the top of the distiller, a large portion of steam with a comparatively small portion of ammonia
is condensed and flows back to the still. The distiller condensers on top of the hear form what is
known as a partial condenser as mentioned before. In the ideal condition, the vapor from one
compartment below should be condensed on the division plates above, giving all its heat of
condensation for the vaporization of a richer liquor on that a poorer ammonia vapor disappears
and in its place appears a richer one. This process is repeated until the partial condenser on the top
is reached. Furthermore, it can be shown that in order to obtain a maximum separation between
ammonia and water, the amount of overflow from each division plate throughout the body of the
distiller should be the largest possible. That means that all condensation by cooling should be done
in the partial condenser at the top of the heater, and the overflow should start from this partial
condenser down. If there is a loss of heat by radiation from the body of the distiller, the vapor will
be condensed in the lower part of the distiller before it reaches the partial condenser at the top, and
this condensate gives rise to overflows arising from of vapor plates only. Hence, to utilize all the
heat coming in the form of vapor below for vaporizing a richer liquor above, and to secure the
maximum overflow possible from the division plates throughout the height of the distiller, it is
necessary to insulate the body of the distiller and to minimize the heat loss by radiation, so that all
the cooling shall be done in the partial condenser and cooler at the top of the heater, and not in any
part of the body of the distiller.
(3) Larger ammonia tension in liquor phase than partial ammonia pressure in gas
phase. In order to cause ammonia to vaporize from the liquor, its vapor tension in the liquor must
be greater than the partial pressure in the vapor. If the partial pressure is greater than that which
corresponds to the equilibrium value of the tension in the liquor, ammonia from the vapor will be
dissolved in the liquor. At equilibrium, the rate of vaporization of ammonia from the liquor equals
that of solution in the liquor. While this countercurrent principle is followed on the whole in actual
practice, there are departures at certain points in the distiller, notably at the junction between the
heater and the lime still and at the feed inlet at the top of the heater. In the distiller construction
prevalent in ammonia soda plants, the heater takes all the heat from the lime still, and all the rich
ammonia gas from the lime still below is allowed to come in contact with the heater liquor at the
bottom of the heater; which liquor contains practically all the ammonia as fixed ammonia. The
ammonia vapor tension from the fixed ammonia liquor is practically nil, and consequently at this
point some ammonia from the vapor phase will be redissolved in the liquor until equilibrium takes
place. This why the heater liquor leaving the bottom of the heater cannot be completely free from
free ammonia in this type of arrangement. This decreases the efficiency of ammonia separation in
the distiller and increases the burden on the lime still. The old arrangement of having a separate
heater and lime still taking their heat supply (steam) independently and delivering the ammonia
gases to a common condenser, also independently, did not have this objectionable feature.
(4) Feed liquor introduced at a point where its ammonia tension is not smaller than
partial pressure of ammonia in gas at that point. Another point is the feed inlet at the top of the
heater. In order not to break the gradient of ammonia concentrations in the overflow liquors, the
feed should be introduced at a point where its ammonia vapor tension is just large than that of the
overflow liquor at that point. At the top of the heater where the filter liquor is customarily
introduced, the hot condensate from the partial condenser is usually at a much higher ammonia
vapor tension than that in the feed liquor introduced there, especially when the latter is dilute and
comparatively cold; and so the gas is richer in ammonia than that which corresponds to the
equilibrium value in the feed liquor. As a result, ammonia vapor is actually rediscovered by the
feed liquor under such conditions. In the present-day ammonia soda plants effort has been made to
preheat the feed liquor by using it as a wash liquor in the furnace gas scrubber and further as a
cooling medium in one section of the partial condenser at the top of the heater. This not only
conserves the heat by treating the partial condenser as a heat interchanger, but also improves the
ammonia tension gradient in the heater. The condition is illustrated qualitatively in Figs. 68,69,
and 70.

FIG 68 Curve showing vapor tension of ammonia in liquor phase.

At x1 , Fig 69 ,re-solution of ammonia from the vapor always occurs, and at x2 re-solution of
ammonia may occur if the feed liquor is dilute and not properly preheated. Fig.70 shows the
resultant condition from the combination of the conditions in Figs .69 and 68 The same is
particularly true of the mud inlet .The mud from the settling vats, which contains rich free
ammonia and CO2 , should enter at the top of the heater.
(5) Minimum volume of filter liquor to be distilled and of milk of lime used. From the
above it is seen that one objection in the distiller operation is the weak feed liquor, i.e.,a large
volume of filter liquor to be distilled. In the matter of heat consumption in the distiller, a large
volume of filter liquor requires a correspondingly large amount of steam for distillation. This
means a wastage of steam and consequently of coal. In the matter of ammonia loss, a large volume
of filter liquor means a correspondingly large volume of the waste liquor to be disposed of . As
there is always a small amount of ammonia left in the distiller waste liquor, a large volume of this
waste liquor would result in a higher ammonia loss through the distiller,. Indeed, in normal
operation, the distiller is a source of the largest single item of ammonia loss. For the same reason
the milk of lime must be made as strong as can be handled, i.e.; containing about 250 grams per
liter, or 180 titer lime. For high efficiency it is therefore imperative to keep the volume of filter
liquor as small as possible and to employ as strong a milk of lime as can be handled through the
pumps and

FIG 69 Curve showing theoretical partial pressure of ammonia in gas phase.

piping lines, and also a hot milk of lime. The effect is cumulative: weak filter liquor and milk of
lime require more exhaust steam per ton of soda ash made, which would cause a still larger
volume in the distiller waste. When calcium chloride and sodium chloride are to be recovered

FIG 70 Curve showing resultant partial pressure of ammonia in gaseous phase.

from the distiller waste, it becomes a sine qua non that the distiller waste (and therefore the filter
liquor and the milk of lime entering the distiller) should be as concentrated as possible.
 To aid in the understanding of the distiller operation, we give a boiling-point table for various
concentrations of ammonia in aqueous solutions (Table 99). The values are plotted in Fig. 71.

(a) The velocity of the vapor passage through the distiller and the diameter of the
distiller.

(b) The surface contact between the vapor and the liquor and the agitation of the liquor by
steam.
(c ) The depth of the liquor on the division plate and the number of passettes.
(a) Velocity of vapor passage through distiller and diameter of distiller. Efficiency in
the separation of ammonia gas from the liquor is decreased if the liquor on one passette
(bubble-cap assembly) is carried up and mixed with the liquor in the other passette above. Such
entrainment disturbs the equilibrium relationship between the vapor and the liquid phases. It is
generally caused by spattering and splashing of the liquor by steam, and in the worst case the
liquor may be “suspended” due to the gas -lock effect of the vapor. Such phenomena are generally
caused by excessive velocity of the vapor passing up through the distiller, when the diameter of
the distiller is inadequate for the volume of vapor handled. Normally the average vapor velocity
from one passette to the other is about 3-6 feet per second, and so the spacing between two
adjacent passettes is 2 feet 3 inches on the top to 3 feet 4 inches toward the bottom, with one
deepest ring of about 5feet at the bottom of the lime still where low-pressure steam enters. This is
approximately the proportion for an 8-foot distiller. Thus, an 8-foot diameter distiller with
standard proportions will under normal conditions have a production capacity of 200-250 tons of
soda ash per unit per day, with out having too high vapor velocities.
(b) Surface of contact between vapor and liquor and agitation of liquor by steam. The
lower edge of each mushroom is serrated, so that steam in passing around the edge and bubbling
up through the liquor on the division plate may be subdivided and broken up into numerous small
bubbles, thus getting more intimate contact with the liquor. Also, the bottom of the division plate,
instead of being flat, is best curved to conform to shape of the path of the flowing steam under the
mushroom, so that there may not be any dead spot on the division plate where liquor does not
have a chance to come into direct contact with the steam. This helps break up the steam bubbles
and stir up liquor body so as to present fresh surfaces of contact between the vapor and the liquid,
and to destroy the inactive film layer existing at the liquid-vapor boundary. As the result, the molal
fraction of ammonia in the vapor phase approaches more closely the theoretical value in
equilibrium with that in the liquid phase on each plate. Further, the overflow outlet is located
diametrically opposite the inlet, and the liquor is compelled to travel across the full area of the
division plate before leaving the plate. All this results in high efficiency in the performance of
each passette.
(c) Depth of liquor on division plate and number of plates (passettes). To obtain good
contact and effective stirring action, it is necessary that not to large a volume of the liquor, nor to
deep a liquor body, be allowed to remain on each division plate. The depth of the liquor on each
plate is generally no more than 6 to 7 inches. Contrary to ammonia absorption in the absorber, the
depth of wash n the distiller is thus less than that in the corresponding passettes in the absorber.
Also the number of passettes in the distiller is larger. The depth of liquor on the plate is controlled
by the overflow outlet above the bottom of the plate, while the depth of wash is determined by the
submergence of the serrated edge below the liquor surface. The number of plates (passetes)
required may be determined by McCabe and Thiele’s graphical method. Take, for example, the
lime still. If we determine the molal fraction of NH3 (1) in the limed liquor fed from the prelimer,
(2) in the blow-off, and (3) in the overhead condensate overflowing back to the heater, we can
estimate graphically the number of plates required in the lime still, knowing the volume of
overflow. However, since the relationship between the lime still and the heater is complicated and
the function of the heater is not purely one of rectification, the results so obtained are misleading,
and so generally considerably more plates over and above what is theoretically required according
to the graphical solution will be needed.
To estimate the quantity of heat to be extracted by the distiller condenser and cooler per ton
of soda ash made and the cooling surface necessary, take for example
 Therefore partial pressure of ammonia alone at 82 is

and partial pressure of ammonia alone at 56 is(assuming same ratio of NH3 to CO2 )

Therefore the ratio of ammonia to steam in the vapor coming over from the distiller top at
82 is as the density of ammonia at 305 mm is to that of steam at 385 mm or as 0.0235
lb/ft.3 :0.0198 lb/ ft.3 = 1.186:1.Similarly the ratio of ammonia to steam in the vapor at 56 is as
the density of ammonia at 521 mm to that of steam at 123 mm as 0.0387 lb/ft.-3 :0.00683 lb/ ft. -3
=5.67:1. And the composition of the condensate is 16 per cent ammonia and 84 per cent water,
neglecting the carbon dioxide dissolved. At 56 per ton of ash there will be:
6.1 cu. m × 80.7kg NH3 /cu.m. =492kg NH3 going into the absorber
*McCabe, W.L; and Thiele, E.W; Graphical Design of Fractionating Columns, In. Eng. Chem;
17,605(1925); and Badger, W.L; and McCabe, W.L; “Elements of Chemical Engineering,” pp. 340-372,
McGraw-Hill Book Co; Inc; New York, Second Edition, 1936.
Among the various cooling duties to be performed in the soda ash plants, such as in the absorber
coolers, vat coolers, carbonating tower cooling boxes, furnace condensers, etc; the cooling duty of
the distiller condensers is one oft he most important and should require provision for one of the
largest cooling surfaces in the whole equipment. Fortunately, the mean temperature difference for
the flow of heat from the condensing vapor to the cooling water is generally great. Thus, suppose
the temperature of the ammonia gases is 82 and it is desired to cool the gases to 56 . Suppose
the cooling water has a temperature difference then is

The heat transfer from condensing steam to water should also be favorable. It is, however, very
difficult to calculate exactly the cooling surface required because the coefficient of heat transfer
depends largely upon the tube conditions both inside and outside. An insulating layer of
ammonium carbonate crystals, iron scale, etc; on the vapor side of the cooling tubes, and a hard
scale formed from the hardness or mud in the cooling water on the water side are liable to be
formed. These would greatly reduce the rate of heat flow from the condensing steam to the
cooling water.
The specific gravity readings (at room temperature) of different liquors are given in Table
100.

Between the distiller top outlet and absorber bottom inlet, the temperature of the ammonia
gases should drop from about 86 to 56 . In other words, the distiller condensers and coolers
should have sufficient capacity to cool the hot ammonia gases coming from the heater by about
30 . As a rough guide there should be 70 square feet of total cooling surface in the distiller
coolers and condensers combined per ton of soda ash made per day.
Following is an estimation of the exhaust steam required in the distillers. For the estimation
of steam required in the distiller per ton of soda ash made, not all the necessary data are available;
some assumptions must be based on practical observations. We shall assume, in the absence of the
specific figures available, that the heat effect of the neutralization of the first hydrogen of the
carbonic acid by ammonia is 11,000 cal per mol. Next, we shall place the figure for heat loss by
radiation from the well-lagged distiller at 10 per cent of the total heat required per ton of ash. We
shall take the following set of data:

Note that the heat of reaction for the 350 kg.NH3 in the following reaction is
negligible ,because it is an equilibrium reaction with ago components in solution:
2NH4 Cl.Aq+Ca(OH) 2 .Aq=CaCl2 .Aq+2NH4 OH.Aq+Q
(-72,800 2) (-244,200) (-214,000) (-88,000 2)
Q is negligible
That is, the beat effect in the above equilibrium reaction may be neglected.
From the above it will be seen that the greatest item of heat required is that required to heat
up the liquors. This takes 64 per cent of the total heat required. Therefore, for good efficiency, the
feed liquor and the milk of lime must be preheated and kept hot and be as concentrated out. Tin
practice, the smaller the volume of the filtered liquor and of the milk of lime, the smaller will be
the amount of exhaust steam required for distillation, and consequently the smaller the volume of
the distiller waste, and the smaller the losses of both ammonia and lima in the distiller waste sent
out.
 Chapter XV

By-products from Distiller Waste

The distiller run-off liquor consists of some 85 grams per liter of calcium chloride, 50 grams
per liter of sodium chloride, and some excess of lime, together with other solid matter such as
siliceous matter, calcium carbonate, magnesia, calcium sulfate, etc. The concentration of these
constituents varies widely in individual plants depending upon the operating conditions and upon
the efficiency of the plant .The liquor contains all the chlorine from the salt used; for, unlike the
LeBlanc process, the ammonia process makes no use of the chlorine. Because of the high cost of
the fuel required to concentrate such a weak liquor, and because of the limited demand for calcium
chloride and the cheapness of this product, and salt, comparatively little of this distiller waste has
been worked up, considering the very large volume of the waste turned out daily by the
ammonium soda industry.
The solid matter from the waste accumulates so fast that the disposal of this waste is a
matter of no small concern to ammonia soda manufacturers. Calcium chloride in solution
permeates the soil bed and contaminates sources of water supply. Municipal authorities permit
only a very limited portion of the waste (except the clear portion of the liquid) to be send to a river
or any public waterway, and that only after complete settling to eliminate the solids. Cases ate on
record where fish in the river were killed by the harmful constituents especially the free lime
carried in the liquor, so that ammonia soda manufacturers are obliged to do something to this
waste. Proposals have been made to utilize it in many different ways, and various products have
been claimed to have been prepared there from. Many of these proposals have no practical
value .The preparation of chlorine gas (from which bleaching powder is made) or the production
of hydrochloric acid from the calcium chloride in the waste by electrolysis can no longer be
considered practicable in view of the present much more economical methods for preparing these
products.
There is a considerable excess of unc onverted salt in the waste. It must be said that the
ammonia soda process is very wasteful in the use of salt, the average efficiency in conversion of
salt to sodium bicarbonate in the carbonating towers in actual operation being not more than 75
per cent; considering some additional losses in the form of “mud” pumped to the distiller, more
than 25 per cent of the total salt is wasted. As a consequence, the very existence of an ammonia
soda works presupposes a cheap and abundant supply of salt. Because of this fact it generally does
not pay to recover this salt from distiller waste. The salt so recovered, however, possesses a high
degree of purity and is comparable to the product obtained from brine purified by the soda-lime
treatment. This salt is recovered in conjunction with the manufacture of calcium chloride.
We shall enumerate a few of the principal products that have been prepared from distiller
waste on a commercial scale; namely, refined salt, calcium chloride, calcium, sulfate and
ammonium chloride.
Refined salt and calcium chloride. It is the aim of ammonia soda operators to keep the
volume of distiller waste as small as possible, especially when any of it is to be worked up. This is
done by keeping the volume of the filter liquor as small as possible, by securing the highest
decomposition possible in the columns, by using as little wash water on the filters as is necessary
to reduce the sodium chloride content in the ash within allowable limits, and by employing a
strong milk of lime containing not less than 250 grams of lime per liter (sp gr 1.21). In certain
cases even solid lime has been tried in the preliming tanks for distiller operation. This waste liquor,
which contains a small excess of free lime as mentioned above, is carbonated by passing some
waste CO2 gases through it. On settling, the precipitated calcium carbonate and all the solid matter,
such as magnesia, silica, calcium sulfate, etc. carried in the liquor separate out. The clear solution
containing calcium chloride, sodium chloride, and a very small proportion of dissolved calcium
sulfate, is concentrated in a double-effect evaporator, when sodium chloride together with the
small amount of calcium sulfate commences to crystallize out. The salt crystals are dried in
centrifuges, a little water being sprayed on them to wash any calcium chloride from the salt. From
the centrifuges the salt can be packed directly in bags of 200 pounds or in barrels of 400 pounds,
frequently without further drying. If drying is necessary, the salt from the centrifuges or drum
filters is charged to steam-heated, hot-air rotary dryers using exhaust steam.
The liquor entering the evaporator with a specific gravity of 1.11 is concentrated to 1.22 in
the first effect, and evaporation is carried on until the specific gravity of the liquor in the last effect
is 1.42.At this point practically all the salt can be settled out.
If necessary, some “chloride of lime ” may be added to oxidize the iron and the liquor is then
allowed to settle. The liquor shouldn’t be allowed to cool too much, or hexhydrate calcium
chloride crystals (CaCl2•6H2 o) would separate out. This clarified calcium chloride liquor
containing 42 to 45 per cent calcium chloride is known in commerce as “liquid calcium chloride. ”
It is shipped in tank cars like liquid caustic .It is used in refrigerating plants for road sprinkling
and for making liquid concrete accelerators.
If solid or flake calcium chloride is desired, the clarified liquor may be further concentrated
in cast-iron open pots until the molten mass, solidifying on cooling, gives what is known as 70-per
cent calcium chloride, corresponding to the formula CaCl2•2H2 O. Amore recent method is to use a
high-pressure, single-effect evaporator to concentrate this 1.42sp.gr.liquor directly to 75-per cent
calcium chloride, using a higher pressure steam instead of the exhaust. The beat economy in the
evaporators is far greater than in the cast-iron open pots. This product in the molten state is
pumped to steel drums, each holding about 650 lbs., and or the drums are sealed. The composition
of 75-cent solid calcium chloride is given in Table 101
Table 101 Composition of 75-per cent Calcium Chloride.
CaCl2 73.6
NaCl 1.4
Alkalinity as Ca(OH) 2 0.14
Insoluble matter 0.08
Water by difference 24.78
Flaked calcium chloride is made by sending the 75-per cent molten calcium chloride to
flaking machines provided with revolving water-cooled drums, as in flake caustic manufacture.
Anhydrous calcium chloride can be made by heating 75-per cent calcium chloride in a
reverberate or pudding furnace, fired by producer gas of oil. The molten mass is frequently raked,
when some calcium chloride is decomposed, giving lime and hydrogen chloride. The escape of
this hydrogen chloride gas leaves a porous mass of fused calcium chloride, which is known as
95-per cent CaCl2 . The cooled anhydrous mass, somewhat fluorescent, is packed in steel drums or
paper-lined wooden barrels, hermetically sealed, each holding from 350 to 400 pounds.
Fused calcium chloride, like fused caustic soda, has a density of about 2(sp.gr.of anhydrous
CaCl2 ,2.22); but unlike caustic soda the melting point is much higher (775 ). Hence molten
anhydrous calcium chloride could not be readily obtained in fusion pots, as is the case with fused
caustic soda.
There is a process for making powdered calcium chloride by drying the granular calcium
chloride in a rotary furnace by means of furnace gases entering at 400 to 475 .and leaving at
100 .The coarse particles drop out of the lower end of the rotary, while the fine dust suspended
in the current passes out of the upper end and is recovered in a collector containing about 73 per
cent anhydrous calcium chloride, which is non-caking.
Calcium chloride is used as a refrigerant in the manufacture of ice, as a dust palliative in
flake or granular form to minimize dusts on gravel roads, and as an anti-freezing solution (freezing
point-50℉) In the anhydrous form small quantities are used as a catalyst to promote the traction
between calcium carbide and nitrogen gas in the manufacture of calcium cyanamide, permitting a
lower temperature to be employed.
Calcium sulfate. The crystalline CaSO4•2H2 O is known as “Crown Filler” among the Solvey
products .It is a much purer product than natural gypsum .The process of manufacture consists in
setting the distiller waste and treating the settled liquor while still hot with salt cake. The excess of
free lime carried in the liquor will be neutralized by the slight excess of sulfuric acid in the salt
cake. If the resulting solution is acid some milk of lime may be added. The precipitated
CaSO4•2H2 O in finely divided form is filtered on a vacuum drum filter and further dried in a
centrifuge, water being sprinkled on the crystals to wash away salt. The crystals of gypsum are
dried at a reduced temperature in much the same way as sodium bicarbonate (see Chapter XVIII).
It is packed in wooden barrels of about 350 to 400 pounds and used as paper filler. A small
quantity is used as a water hardener in the fermentation industries.
When dried at or below 190 , this gives plaster of Paris (CaSO4•1 /2 H2 O), which is used for
making molds . When calcined at a temperature above 204 . it loses all its water ,care being
taken that none of the sulfate is reduced to sulfide. When traces of iron are present, and the
calcined mass is discolored and gives a yellow tint.
Calcined calcium sulfate is used in the manufacture of certain tooth pastes and in the
manufacture of wall and flooring plasters which give a smooth and hard finish.
The filtered liquor, after removal of the precipitated calcium sulfate, is very pure brine which
is sent to the soda works for making saturated brine for soda manufacture.
Ammonium chloride. In this connection may be mentioned sal ammoniac (ammonium
chloride), which can be obtained by concentrating the carbonated and settled waste liquor to
remove most of the salt (NaCl) and treating it with ammonium sulfate (NH4 )2 SO4 , whereby the
following reaction takes place.
CaCl2 + (NH4 )2 SO4 → CaSO4 + 2NH4 Cl
The mother liquor is further concentrated, allowed to settle, and cooled. Crystals of ammonium
chloride then separate out. This process was originally patented by Richardson.
Another process proposed for the manufacture of sal ammoniac that may involve important
changes in the ammonia soda industry is as follows: From the mother liquor (or filter liquor)
before distillation, it is proposed to separate ammonium chloride by cooling. This was done in
Germany during the World War by the Badische Anilin-und Soda-Fabrik using synthetic ammonia.
In Japan it was claimed that by freezing out ammonium chloride from the mother liquor and by
adding a fresh supply of salt, the resulting liquor can be sent back to the carbonating towers for
making sodium bicarbonate, thus doing away with distillation. It is claimed that in this way the
distiller can be dispensed with and the operation in the ammonia soda works simplified.
The detailed procedure consists of introducing more ammonia and salt to the mother liquor
after filtration and carbonating the resulting liquor again to throw out more sodium bicarbonate,
using a rather low temperature. The bicarbonate is filtered off as usual, and the mother liquor,
which is now very rich in ammonium chloride, is strongly cooled so that ammonium chloride
crystals separate out. The crystals are then centrifuged and washed in the usual manner.
Solvay makes a large amount of ammonium chloride-more 50 or 60 tons a day. It is made
from distiller waste liquor by first settling off the solids from the liquor and concentrating the
solution in evaporators .The small percentage of lime remaining dissolved in the solution up to
about 2 grams per liter makes calcium chloride liquor less corrosive on evaporators. The solution
is concentrated to 40 Âé, containing about 40 per cent CaCl2 ,when most of the salt separates
out. The slurry containing salt crystals is filtered off and the filter liquor ammoniated .The
ammoniacal calcium chloride solution is then carbonated in tanks at a temperature of 60 .with
stirrers; at this point very coarse calcium carbonate crystals are formed, which filet very beadily,
and the filtrate containing high concentration of ammonium chloride is cooled when the
ammonium chloride crystals begin to separate.
CaCl2 + 2NH3 + CO2 + H2 O→ CaCO2 + 2NH4 Cl
The liquor is then concentrated so that further crops of ammonium chloride crystals ate obtained.
Solvay at first used wooden tank filters with false bottoms for such purpose, although ordinary
filters work very successfully. Ammonium chloride crystals made in this way contain only 0.3 to
0.4 per cent sodium chloride.
The ammonium chloride crystals are dried with hot air at a low temperature. The same
equipment as is used for drying refined sodium bicarbonate is suitable for this purpose. (See
Chapter XVIII, Manufacture of Refined Sodium Bicarbonate.)
This process of making ammonium chloride runs exactly parallel to the gypsum process of
making ammonium sulfate used by I. G. Farbenindustrie in Ludwigshaven-Oppau, Germany, the
differences being that calcium chloride is used in place of calcium sulfate, and that ammonium
chloride is obtained instead of ammonium sulfate.
Ammonium chloride is used in the manufacture of dry cell batteries and in metal working for
soldering and galvanizing. For this purpose, a mimimum purity of 99.2-99.4 per cent NH4 Cl is
required; whereas for fertilizer, anything above 97 per cent NH4 Cl is acceptable, the impurities
being mostly NaCl.
Soil Lime. Solids carried in the waste liquor, consisting of waste sand, finely divided calcium
carbonate, excess lime, calcium sulfate, magnesia, etc, are allowed to separate by settling, and the
sludge is sometimes used as “fertilizer” for correcting acidic soil. However, the calcium chloride
content of the waste lime should be allowed to weather and be washed off by rain before it is
applied to farms. What fertilizing value it has is difficult to determine, but some farmers do use
this waste sludge on their farms with favorable results.
An ammonia soda plant is considered to be happily situated if it can send its distiller waste to
the neighboring lands freely. If a plant is situated among quarry hills where the distiller waste can
be sent to fill the quarry pits or the adjoining valleys, or if a plant can be situated near the sea or
on a large lake to which distiller waste can be discharged freely, the situation approaches an ideal
one. For the time is still far distant when such waste, which is produced daily ni enormous
quantities, can all be worked up into by-products. The great bulk of waste must be disposed of and
the space for its disposal must receive attention from ammonia soda manufacturers at the outset.
Distiller waste accumulated in the neighborhood of an ammonia soda plant leaves a rather
unsightly feature in our landscape. In the course of years, after weathering has set in, green
vegetation will grow on it .The muck may then become a good farm land again.
 Chapter XVI

Alkali Products of Ammonia Soda Industry

Besides the products naturally allied to soda ash, namely chemical caustic and refined
bicarbonate, ammonia soda manufacturers make many miscellaneous products from the
combination of soda ash, caustic soda, sodium bicarbonate, etc, to meet special needs of certain
industries. The manufacture of these products in most cases involves no special technical
operations, but it helps the manufacturers to sell their products, as they claim to be specially
prepared to meet the specific needs of certain classes of consumers.
In this chapter we shall not include those products known as by products from distiller waste,
such as calcium chloride (liquid, solid of flake), refined dolt, calcium sulfate, etc, which are dealt
with in Chapter XV, “By-products from Distiller Waste.” We shall consider only those special
names in the trade catalogs of the ammonia soda manufacturers.
Extra Lithe Ash, “Fluf,” or Fluffy Soda. This is a very loose and bulky form of lithe ash,
desired by a certain class of household suppliers who would like to have large bulk for their
packages. It has the advantage of dissolving more easily. This product is generally collected from
cyclone separators or dust collectors at the drying and finishing end of the process. The quantity
available is therefore not large. It has a specific gravity of about 23 pounds per cubic foot.
Light Ash. This is ordinary soda ash obtained directly from the calcining of the filtered
bicarbonate in the rotary dryers, or in Thelen pans. Because of the violent agitation in the
carbonating towers by the bubbling action of the CO2 gas, the crystals of NaHCO3 formed are
necessarily very fine; thus the soda ash obtained from the calcinations of these crystals is loose
and fine. It occupies a comparatively large volume per unit weight, and its specific gravity is only
about 35 pounds per cubic foot. The density of the light ash varies a free at deal, depending upon
the character of the bicarbonate crystals from the columns, the moisture content of the bicarbonate,
and the method of calcining it. Soda ash is now added to the clay mix in small quantities in the
manufacture of firebrick etc.
Modified Sodas. These are various mechanical mixtures of sodium bicarbonate (NaHCO3 )
and soda ash (Na2 CO3 ) in ratios varying from 25 to 75 per cent sodium bicarbonate by weight.
They are also called “neutral sodas,” and are used where a milder alkali is desired. Depending
upon the percentage of sodium bicarbonate admixed, they as sold under different names allegedly
to meet specific needs of certain industries. “tanners ‘ Soda,” “Textile Soda,” “Laundry Soda,” etc;
come under this group.
Sodium Sesquicarbonate. This is a special form of “neutral soda.” It is a double salt having
the formula Na2 CO3 NaHCO3 2H2 O, and hence has a definite composition and crystalline
structure. It consists of fine, lustrous, needle-like crystals. The product is known as “Concentrated
Crystal Soda ”in England, but as “Show Flake Crystals” among the Solvay products in America.
In its natural state it is known as “trine” or “urea. ”It is a stable compound, neither hydroscopic
like soda ash, nor effervescent like Sal soda; and it does not form lumps on storing. It is readily
soluble in water and dissolves without evolution of heat like soda ash (Na2 CO3) or monohydrated
(Na2 CO3 .H2 O), and without absorption of heat like Sal soda (Na2 CO3 .10 H2 O). This compound
exists extensively in natural soda deposits, as it is the ultimate stable form of sodium carbonate in
contact with the atmosphere.
The crystals can be manufactured by adding an excess of sodium bicarbonate to a hot soda
ash solution somewhat above the ratio of 84 parts NaHCO3 to 53 parts soda ash by weight. The
resulting strong solution is allowed to cool slowly, whereupon fine, shiny crystals are formed.
There is a process for making refined sodium bicarbonate from the crude bicarbonate by
dissolving the latter hot, distilling off the ammonia as in the distiller operation but without
addition of lime, settling the solution, cooling, and finally recarbonating the cooled, settled liquor
as in the ammonia soda operation. During distillation, a considerable portion of the bicarbonate is
decomposed to normal carbonate, and by close regulation, just enough sodium bicarbonate can be
decomposed to form normal carbonate so that the resulting solution is of the right composition to
form sodium sesquicarbonate crystals by cooling and reprisal libation.
Again, sequicarbonate may be made by dissolving soda ash, recarbonating with lean gas
(weak CO2 gas) in a wooden tower to a proper the sequicarbonate crystals. One of the natural soda
manufacturers in California partially carbonates soda brine to precipitate sodium sesquicarbonate
this way.
The crystals of sodium sesquic arbonate are dried at a reduced temperature in the same type of
equipment as the sodium bicarbonate dryer, and are packed in wooden barrels containing 280 lbs;
or in kegs containing 85 lbs each. The sesquicarbonae, being stable, is not affected by atmospheric
moisture or CO2 and is therefore free from lumps. J it is readily soluble and finds much favor in
wool scouring and laundering .
Granular Sodium Carbonate or Crystal Carbonate. This is the monohydrate of
soda(Na2 Co3 . H2 O) and is obtained by crystallization from solution like Sal soda, but at a very
much higher temperature. Theoretically, a temperature above 35 will yield these crystals: in
practice a boiling temperature is employed because of the ease of concentrating the solution. This
is a fairly stable compound to which Sal soda is first converted, on exposure, in the form of a
white outside crust. Natural Sal soda deposits yield this white powder on the surface layer by
efflorescence. In the natural state it is known as thermonatrite, Na2 CO3 .H2 O. by absorption of
CO2 and moisture from the air, crystal carbonate finally goes over to the sesquicarbonate
(Na2 CO3 .10 H2 O). hence the preponderating proportions of the latter in natural sod a deposits.
Crystal carbonate is first formed when soda is recovered from burkaite (Na2 CO3 .2 Na2 SO4 )
in the plant of the American Potash & Chemical Corp. at Trona on Sealers Lake. It is then
calcined to dense ash.
Sal Soda. Sal soda, washing soda, or soda crystals is the decahedra of sodium carbonate
(Na2 Co3 .10H2 O) occurring in nature as “matron.” The crystals are formed from solution at a
temperature below 32 . In the old days, when pure soda could not be obtained from the LeBlanc
process, people preferred to use soda crystals to avoid having caustic in the soda and to insure
getting a purer product. It was used in washing and in the scouring of wool. Nowadays a very pure
soda ash can be obtained from the ammonia soda process, and the trade in this article has greatly
declined on account of its high freight cost and the difficulty of keeping the crystals in hot weather.
These crystals contain more than 60 per cent of their weight of water and dissolve in their own
water of crystallization in tropical climates. Sal soda manufacture requires a large floor space.
When the solution prepared from the ammonia soda, a tank provided with a stirrer much like the
caustic zing tank is used. Water is heated, to which soda ash together with the desired amount of
sodium sulfate is added (see below) until its specific gravity shows 36 be at this temperature.
 This will take about 50 per cent soda ash on the weight of water. To aid solution, however,
the liquor is heated by closed steam coils to 60 to 70 , and while settling and clarifying must be
kept above 50 to prevent its crystallization and clogging of pipes, etc. for this reason, the
reservoirs, pipes, etc; should be lagged. After all has been dissolved, the stirrer is stopped, the
solution allowed to settle, and the clear portion drawn into shallow crystallizing pans. These are
10 feet wide by 10 feet long by 2 feet deep from the top to the pyramid bottom. On reaching the
pans, the solution should have been cooled to about 32 . The pans are set on brick or concrete
piers so that the mother liquor can be drawn from the bottom. They are set aside to crystallize for
4 to 10 days, depending upon the size of the pan, the weather conditions, etc. on hot summer days
this operation is attended with great difficulty. Across the pans are suspended galvanized iron wire
nettings with strings or wires barely submerged under the surface of the liquor. Crystals are
formed hanging from these wires immersed in the liquor. After crystallization, the mother liquor
shows a Be reading of only 20 to 22 when it is drawn off. The crystals formed on the sides and
bottom are chiseled off and charged to centrifuges to whiz out the enclosed mother liquor. A little
fresh water is sprayed on the centrifuge to wash out the remaining mother liquor. The crystals are
broken up in an impact crusher and packed in wooden barrels, each containing 280 lbs.
Instead of the dissolving tanks described above, a small rotary dissolver with angle-iron
flights riveted on the shell inside gives continuous operation and has a very large capacity, the
solution being settled in a separate tank outside.
The manufacture of Sal soda possesses some advantage in that discolored soda ash and floor
sweepings can be dissolved and utilized for its manufacture.. In this case the solution must be well
settled and decanted from the dirt. Thousands of tons of red soda produced when starting a new
ammonia soda plant have been utilized in this way or in the manufacture of caustic soda.
In order to obtain larger crystals from ammonia soda, anhydrous sodium sulfate is added with
the soda ash to the extent of 2.5 to 3.0 per cent by weight. This, of course, contaminates the
product but the users generally do not object to it.
Sal soda is made more extensively by natural soda manufacturers who purify the natural soda
deposits by ridiculing and recrystallisation, yielding Sal soda crystals (Na2 CO3 .10H2 O) or crystal
carbonate (Na2 CO3 .10H2 O).
Block Soda .In the orient, much soda is handled in block form. Large quantities of soda ash
are made into blocks and sold in this form, because of the old practice in which natural soda was
handled. Soda ash in a finely powdered from has a tendency to “set” in contact with water when
the mass is not disturbed by stirring, and further will unite with a considerable amount of water,
maintaining its firm, solid form when cooled. From fresh ammonia ash in a finely divided state
which has not absorbed any moisture by exposure, a suspension, or thin paste, formed by stirring
the fine powder into as much as 135 per cent of its weight of water, will “set” to a firm, solid form
on cooling. The amount of water the soda will take in this way depends upon (1) percentage of
Na2 CO3 (not as NaHCO3 ) in the ash, (2) the fineness of the crystals, and (3) the weather
conditions. The ability to absorb water and form a firm, solid block is materially impaired if soda
ash in part been already converted to NaHCO3, or if it has absorbed much moisture from the air
through long storage or after transportation over a long distance. To make these blocks, soda ash is
screened and the coarser particles or lumps are pulverized until all pass through at least a 20-mesh
screen. The powder is gradually stirred into 100 to 135 per cent of its weight of water at about 30
and the mixture continuously stirred. This stirring operation is very important and serves to
mix the ash thoroughly with water. As it is stirred, the suspension gradually becomes thick.
Stirring is continued until the suspension thickens to the consistency of a starch paste, whereupon
it is transferred to wooden molds, 34 inches long by 14 inches wide by 8 inches deep (inside
dimensions), and allowed to “set” for 24 to 36 hours, after which the blocks formed can be taken
out by loosening the sides of the molds.
Some sodium bicarbonate is added to the paste toward the end of the stirring just before
pouring into forms, especially in summer. This serves to hasten the “setting” time by “setting” and
to give better grain appearance in its cross-section. The amount of sodium bicarbonate required is
1
2 to 2 per cent in winter and 3 to 5 per cent in summer, based on the weight of soda ash taken.
2
The blocks will thus analyze less than 50 per cent Na2 CO3 A typical analysis is given in Table 102

In the East, soda in this form, or in the decahedra form, is used in making steamed bread,
noodles, cakes, frying, etc, and also in washing.
Causticized Ash or “Super Alkalies. ” These are various mechanical mixtures of caustic
soda and soda ash containing 10 to 50 per cent by weight of caustic soda. They are put up to meet
those requirements in certain industries where a stronger alkali is desired. Depending upon the
percentage of caustic soda admixed, they are sold under different names, such as “Tanners’
Aldali,,” “Bottlers’, Alkali,” “Metal Cleaning Alkali,” “Creamery Alkali, “ “Boiler Compounds,”
etc.
Dense Ash. The old Leblanc industry produced a heavy form of soda ash, more than twice as
dense as ordinary ammonia soda ash. Certain manufactures (such as glass and ultramarine makers )
objected to this loose form of product, and ammonia soda manufactures, in attempting to meet
their requirements, at first re-fired the ordinary ash in a reverberate (or gas-fired, down-flame)
furnace to incipient fusion to bring up its density. The present method, however, is to add a
proper percentage of water to the ordinary ash, mix it thoroughly, and re-dry it in a rotary dryer.
Hence also the name “water ash.” Its density can be increased by 80 to 95 per cent over the light
ash. This increase in density also materially reduces the cost of packing material per unit weight. It
is generally divided into two classes: the medium dense, weighing about 50 lbs per cu ft and the
heavy dense, weighing about 65 lbs or more per cu ft.
Granular Ash. This is a coarse-grained, heavy, dense ash made by employing a rather high
percentage of water and drying. The product is sized by screening and the coarse particles on the
screen give the “granular ash.” Its dustless property is much favored by the glass industry.
Fused Soda Ash. Soda ash has valuable properties as the alkali material required for
metallurgical processes, such as purification of molten iron and analogous operations. Light ash,
or other form of dry sodium carbonate, is fed into a refractor-lined furnace through suitable
charging openings in the roof. The furnaces are fired with either fuel oil or pulverized coal directly
over the hearth into which the soda ash is being fed. The soda ash melts (m.p.1564 F) and
maintains a level in the hearth of the furnace above the run-off spout. The stream of molten soda is
cast into iron molds carried on carts or a continuous chain. The castings of soda ash are dumped
out of the pans as soon as frozen and allowed to cool. They are shipped either in bulk or in cartons
of one kind or another. The actual temperature of casting is about 1600 to 1650 F. The low latent
heat of fusion and the low specific heat cause rapid freezing and cooling.
When the product is being made with a pulverized-fuel fire, the bloc is become contaminated
with the ash of the fuel. For many purposes to which this form of alkali is put, there is no
objection to such contamination. For other purposes the oil- or gas-fired furnace product is made.
The physical form, the weight of block and the composition are all varied to suit customers
and specific uses. In the iron foundry business three-pound blocks of uncompounded fused soda
ash have become more or less standard. Manufacturers of water-treatment chemicals require
blocks compounded with phosphorus, etc. Other special users request tinted blocks. The smallest
block on the market is about one-half pound and the largest is four pounds. Soda ash is also fused
with sulfur for feeding into a waste liquor claimer in craft paper manufacture.
A type of globular dense ash has been made by melting the light ash and atomizing it by
means of compressed air into an insulated chamber lined with refractory material. This gives small
globules of soda ash which is much denser-even denser than” water ash” –and is particularly
dustless for glass manufacture.
Ammonium Carbonate. This product is made by passing carbon dioxide gas through an
aqueous solution of ammonia in a still. The gases consisting of ammonia, CO2 , and water vapor
are distilled off and condensed in a water-cooled condenser and caused to solidify to a dense,
crystalline mass. The mass is then passed through steel circular saws in directions at right angles
to each other, forming inch cubes. These ammonium carbonate cubes are used as “smelling salts”
and for other medicinal purposes. They are among the common chemicals in chemical laboratories
and are also employed as baking powder. Such crystals of ammonium “carbonate” generally
contain also ammonium carbonate and ammonium bicarbonate.
Ammonium Bicarbonate. This product is obtained by carbonating the ammonia solution in
a tower much in the same way as in the precipitation of sodium bicarbonate from the ammoniated
brine. The fairly soluble salt of ammonium bicarbonate, which his analogous to sodium
bicarbonate, is rendered difficultly soluble in the presence of the excess reagents in solution,
separates out as fine crystals, an disinterred and centrifuged in the same way. The fine crystals are
dried by hot air at a temperature of about 50 on belt conveyors in a closed chamber heated by
steam coils. The dried crystals finally leave the belt as fine powder which Is then elevated,
screened, and packed in paper-lined wooden barrels, or, in small sizes, in paper boxes. These
ammonium bicarbonate crystals are used as a valuable baking powder which yields carbon dioxide
and ammonia gases copiously at which leaves no residue in the bread as do the other kinds of
baking powder. Ammonium bicarbonate may also be used as fertilizer.
In the Solvay Works not so much ammonium bicarbonate is made because of its limited
use — only 5 tons being made a day. What is made, however, is very pure and free from iron. For
this purpose a 26 be aqua ammonia containing about 29 per cent of NH3 is used. This is refined
aqua ammonia in which hydrogen sulfide has been removed by reinstallation. It is made by
running crude aqua ammonia through a distiller and passing the ammonia vapor through a water
scrubber or condenser in which hydrogen sulfide is washed down by the down-stream of water in
the partial condenser.
In the Solvay practice, this 26 Be fined aqua ammonia is run through a very small
tower —much smaller than the standard Solvay tower —and lean gas (or kiln gas)is passed through
the tower, with a certain amount of cooling, so that the slurry is kept at about 25 . The slurry
contains about 25 per cent crystals. Wooden tank filters with false bottoms are used to filter these
bicarbonate crystals so that no iron will be picked up from the filters. Art 25 ; these crystals,
though having high vapor pressure, lose ammonia only very slightly. But above 366 the vapor
pressure becomes excessive and decomposition of the ammonium bicarbonate crystals into
ammonia and CO2 is active.
In the ammonia soda plants, it is best to pre-carbonate the aqua in a separate tower and
precipitate the bicarbonate in the making tower. In this way the precarbonating tower containing
the normal ammonium carbonate solution is not subjected to corrosion, all the corrosion occurring
in the making tower. Only lean gas or kiln gas is used in these towers. Old discarded Solvay
towers can be used for such purpose; but if no iron can be tolerated in the ammonium bicarbonate,
wooden towers are constructed for this purpose.
Ammonium bicarbonate crystals are generally bigger than ammonia soda crystals and easy to
filter, but Solvay use exclusively open wooden tank filters because the amount made is small and
the loss of ammonia due to exposure is not excessive. Standard rotary filters or centrifuges also
work satisfactorily.
Because of high vapor tension, ammonium bicarbonate decomposes very readily into CO2 and
NH3 Hence only airtight containers can be used. Lately a very stable crystalline form of
ammonium bicarbonate has been prepared by the workers of the Gesellschaft fuer Kohlentechnik,
Germany, and it is claimed that such crystals of ammonium bicarbonate could be packed in gunny
bags and transported or stored in these bags for months without material loss by decomposition.
Ammonium Chloride. This can be obtained from the mother liquor from the columns by
dissolving nit more ammonia and pulverized sodium chloride and again carbonating the resulting
solution to precipitate out more sodium bicarbonate, working at a lower temperature. Sodium
bicarbonate is filtered off as usual and the mother liquor is strongly cooled to crystallize out
ammonium chloride by the addition of further quantities of pulverized salt, and so on. the crystals
are centrifuged, washed and dried at a low temperature.
The manufacture of ammonium chloride in conjunction with synthetic ammonia manufacture
is an important feature in the modified Solvay process and will be dealt with fully in Chapter
XXVII, “Modification and New Developments of Ammonia Soda Process.”
Liquid Caustic. This is a strong caustic liquor from the evaporators of 48 to 50 Be
containing 47 to 49 per cent NaOH. It is completely settled and clarified to separate out sodium
chloride, sodium sulfate, and sodium carbonate. Liquid caustic is a comparatively new commodity,
and is handled in rubber-lined railroad tank cars. Although the freight cost is higher (because of its
water content), large consumers find it advantageous to receive liquid caustic in tank cars, as this
minimizes the labor of handling. Liquid caustic is convenient to use and is cheaper than solid
caustic because it does away with labor and fuel in dehydrating it to solid caustic. As the caustic
freezes in winter, the tank cars must be provided with steam coils to thaw the content for
unloading.
Flake caustic. This is a solid caustic in thin pieces about 3 /8 inches by 1 /4 inch by 1 /32 inch
thick. It is made by sending molten caustic to the flaking machine, in which a water-cooled drum
rotates with the bottom segment immersed in the molten caustic. A thin film of caustic is picked
up, solidified on the surface of the drum by cooling, and scraped off as the drum passes under a
scraper. Flake caustic is favored by many consumers because of its ease of solution.
 Powdered or Ground Caustic. These are forms of solid caustic that have been pulverized or
ground in crushers to different degrees of fineness. The large surface area aids in solution.
Limestone and Limestone Products. At the quarries operated by most alkali plants, there
are generally a number of strata of limestone of varying composition. Depending on the nature of
these strata, the alkali plant operator markets limestone products as part of his business.
Occasionally the strata must be separated to get the high quality kiln stone in the most economical
manner. Such separation represents an appreciable part of the cost of quarry operation. Stone too
high in silica, alumina, etc. Is sold for cement manufacture, for road-building purposes, and as the
coarse aggregate for concrete, and is occasionally finely pulverized and bagged as agricultural
lime. Occasionally an alkali plant itself operates a cement plant from the waste mud of its caustic
operation. It then quarries certain clay rocks selectively for the use as cement raw materials.
Liquid CO2 and Dry Ice. As has been pointed out in a number of places in this book, an
ammonia soda plant has, as a general rule, an excess of carbon dioxide. This is over and above
that available as boiler plant flue gases. Boiler flue gases from a specially constructed furnace
which furnishes the raw material for most Dry Ice manufacture may run between 12 to 16 per cent
CO2 . The alkali plant has available varying quantities of gas ranging from a low of 23 per cent
CO2 (lime recovery rotary kiln gas) up to a maximum of 41 to 42 per cent (excess gas from a
vertical lime kiln).
The gases from vertical and rotary kilns contain soils particles or dusts, which must be
separated out completely in order to avoid contamination of the absorbent. In ammonia soda plant
practice this is commonly done with mechanical separators such as multilane, etc. Followed by a
thorough water scrubbing. For certain types of purification reagents, it is also important to
separate out organics, particularly utilizable impurities. This can be done with a potassium
permanganate scrubbing tower.
The absorption medium requires occasional purification as in all Dry Ice processes. The
peculiarities of the gas available in an ammonia plant are in no essential way different from those
of any other Dry Ice operation. Hence they are not discussed here.
The ammonia soda manufacturers’ principal reasons for producing solid and liquid CO2 are
twofold. First, they may have a large excess of relatively strong CO2 gas available, which gives
them an advantage over other types of manufacture. Secondly, they may locate the plants where
there is a fairly good market for either solid or liquid CO2 . Since the alkali manufacturers are very
large coal consumers, the ammonia soda plants are located close to coalmines and thus have a
potential market for CO2 in the manufacture of min explosives. CO2 for explosive use is produced
and shipped in liquid form. In small cylinders (up to 50 pounds net capacity) the liquid CO2 is
shipped at the atmospheric temperature, which corresponds to an equilibrium pressure of about
1000 pounds per square inch. In large vessels, such as tank motor trucks and tank railroad cars, the
common practice is to ship it at about 300 lbs. Gauge pressure which corresponds to an
equilibrium temperature of about 0 F. The tanks must therefore be very well insulated with
low-temperature lagging. Liquid CO2 explosive is particularly suitable for use in bituminous coal
mines, which require a slow but progressive action in blasting, to obtain a large percentage of
lumps.
Dry Ice is extensively used in soda fountains, ice-cream chests, and meat refrigerators. It is
also useful in fishing boats for preserving new hauls. It is, however, not so adaptable to the
preservation of vegetables or fruits, as these plant products tend to ripen or rot rapidly in the
atmosphere of CO2 gas.
Chlorine and Chlorine Products. As will be seen in the following chapter, ammonia soda
manufactures have established electrolytic caustic plants themselves, making caustic soda, liquid
chlorine, bleaching powder, and many chlorinated products. They are thus able to adjust the
production of caustic soda between the lime process (cauterization) and the electrolytic process in
accordance with market demands. Because of increased uses of chlorine, the electrolytic
production has made great strides in the last few years. Details of manufacture are given in
Chapter XX, “Manufacture of Electrolytic Caustic, Chlorine and chlorine Products.”
Sodium Nitrate. Ammonia soda manufactures often establish and operate a nitrogen fixation
plant for the manufacture of anhydrous ammonia by the direct synthesis of hydrogen and nitrogen
gases. A portion of the ammonia nay is oxidized to nitric acid, which is then used in converting
soda ash to sodium nitrate, selling this as nitrate of soda for fertilizer. It is in a much purer form
than Chile saltpeter and is a better fertilizer than ammonium sulfate when used alone, because the
latter leaves the soil acidic. Also the nitrogen in the nitrate form is more readily available to the
plants.
Recently sodium nitrate was made directly from salt and concentrated nitric acid in the
Solvay Process Company’s plant (Hopewell, Virginia), yielding chlorine and sodium nitrate. Thus,
chlorine is obtained without the joint product of caustic soda, or rather chlorine is obtained with its
joint product in the form of sodium nitrate instead of caustic soda. This in time may constitute an
important source for chlorine in conjunction with the nitrate manufacture from the oxidation of
synthetic ammonia, which can now be made so advantageously by ammonia soda manufacturers.
To complete the list may be mentioned by-products from distiller waste, such as calcium
chloride (liquid, flake, or granulated), calcium sulfate, refined salt, soil lime, etc. Sometimes alkali
manufacturers also make such products as sodium chromate and other sodium compounds, for
which soda ash is used as a raw material. It is evidently impossible to include all the products
manufactured by the ammonia soda manufacturers because of the many diversified products
covered by various ramifications of the industry.
 Chapter XVII

Position of Ammonia Soda Industry

The ammonia soda industry has played an important role in the fabric of the national
industries. Not only in war but also in peace does this industry form one of the most important
pillars of the national industrial structure. The production of soda ash, like that of sulfuric acid, is
an indicator of the industrial status of a country; however, unlike the sulfuric acid industry, the
soda ash industry is controlled by a comparatively small group of people who supply the needs of
the whole nation. This is also more or less true of other countries. Unlike the sulfuric acid industry,
too, the ammonia soda industry is so far, for the most part, a closed one. Soda ash ranks next in
importance to sulfuric acid. The acid and the alkali are the two most important basic chemicals a
nation possesses, for, directly or indirectly, they are used as the law materials for many other
industries.
Having displaced the LeBlanc process, which was the sole process for soda ash manufacture
for more than a century, the ammonia soda process has assumed a leading position. It is practically
the only process by which the bulk of soda ash now is made. Like its predecessor, the LeBlanc
soda industry, the ammonia soda industry forms a nucleus for many other industries. Besides the
three main alkali products, namely, soda ash, caustic soda and sodium bicarbonate, many other
products are made by ammonia soda manufacturers. The list of products has been described in
detail in Chapters XV and XVI.
Recently the electrolytic caustic industry has been added to the ammonia soda constellation.
It is a well known fact that chlorine, which is produced jointly with caustic soda in the electrolytic
cells in the approximate ratio of 20 tons of chlorine to every 23 tons of the caustic, frequently sets
a limit to the production of the electrolytic caustic industry. Formerly, the uses to which chlorine
might be put were few, and the demand for chlorine was limited. It was the disposal of chlorine
that formerly limited the output from the electrolytic industry; and the ammonia soda
manufacturers had to adjust themselves between production by the chemical method and that by
the electrolytic method. Today, the situation has entirely changed: the demand for chlorine has
greatly increased and has, within the last decade, spurred the electrolytic production to such an
extent that now the quantity of caustic soda produced by the electrolytic process as a joint product
bids fair to exceed that produced by the lime (chemical) process.
Since 1926, six of the ammonia soda manufacturers have gone into the manufacture of
electrolytic caustic. These are Solvay Process Co., Syracuse, N. Y.; Diamond Alkali Co., Fairport,
Ohio; Solvay Process Co., Baton Rouge, La.; Columbia Alkali Corp., Barberton, Ohio; Southern
Alkali Corp., Corpus Christi, Texas; and Michigan Alkali Co., Wyandotte, Michigan. There are
thus altogether seven such plants, including the Matheson Alkali Works, Inc., with its separate
plant at Niagara Falls, N. Y. Ammonia soda manufacturers are enviably situated for this enterprise
because of the availability of cheap and abundant sources of brine and fuel. As will be mentioned
in Chapter XX on the manufacture of electrolytic caustic, the two largest single items of cost
entering into the manufacture of electrolytic caustic and chlorine are (1) salt (brine) and (2) power;
the latter may now be generated very efficiently from coal by high-pressure turbo-generator plants.
Mercury-are rectifiers are now being used successfully in place of the dynamos or
motor-generator sets for converting into direct current, especially where the total voltage in the
direct current circuit is high-at least 440v or more.
Also, the operation of an electrolytic plant gives ammonia manufacturers an extended range
of products, such as liquid chlorine, bleaching powder, mono-chlorobenzene, Para- and or
the-dichlorobenzene, synthetic hydrochloric acid and numerous other chlorine products, in
addition to those discussed in Chapter XVI.
shows these electrolytic plants owned and operated by the alkali manufacturers in the United
States and their plant capacities as of 1940.

Because of the demand for a pure grade of caustic for the rayon industry and others, liquid
caustic of 50 or 70 per cent NaOH has been produced with a high degree of purity. As results,
there is a large increase in the production and shipment of liquid caustic by the ammonia soda
manufacturers and a corresponding decrease in the fused caustic. On account of difficulty caused
by freezing of the liquid caustic in winter, especially the 70 per cent grade, insulate and lined tank
cars have been specially built. These retard solidification of the liquid caustic in cold weather, and
also reduce contamination of the caustic by the container during shipment. For coastwise
transportation, even a tank boat of considerable capacity has been constructed for shipping a
high-grade liquid caustic in nickel-lined compartments somewhat similar to an oil tanker.
The ammonia soda manufacturers have long since identified themselves with the by-product
coke industry. They frequently supply coal gas to local cities. The incentive was evidently to
secure a cheap source of ammonia supply in the form of crude ammonia liquor containing all the
sulfides. Also, coke is required for burning limestone in soda manufacture. The gas, of course, can
be used for power generation in alkali plants. The coke plant supplies ammonia and coke at low
costs to the soda plants. Large alkali manufacturers generally run their own by-product coke ovens
and the by-product coke industry has grown considerably with them. Among them may be
mentioned such larger alkali manufacturers as the Solvay Process Co., Syracuse, N. Y.; the
Michigan Alkali Company, Wyandotte, Michigan; and the Diamond Alkali Company, Fairport,
Ohio. It is interesting to note that because of the need for ammonia in their process, the Solvay
Process Co. Was the first to introduce by-product coke ovens to American. The Semet-Solvay Co.
Had its origin in this and the plant was designed by Louis Semen of Solvay and Co. Of Brussels.
The situation is about the same with the European alkali manufacturers. For instance, the Imperial
Chemical Industries, Ltd. (the former Brunner, Mond & Co., Ltd.), also operates, or controls the
interest in, the by-product coke plants in England; and in fact a type of gas generator, called the
Mond Gas Producer, was originally developed by Dr. Ludwing Mond, one of the chief founders of
Brunner, Mond & Co., Ltd.
With the rise of the synthetic ammonia industry we find the ammonia soda manufactures
going into the manufacture of synthetic ammonia also, notably the Solvay Process Co., Syracuse,
N. Y. (Allied Chemical & Dre Corp., Hopewell, Va.), and the Mathieson Alkali Works, Inc.,
Niagara Falls, N. Y. The raw materials for ammonia synthesis are hydrogen and nitrogen. Nitrogen
gas in a concentrated form (95 per cent N2 or more) is a waste product from tower washers in the
ammonia soda process. Another good source of nitrogen and hydrogen is semi-water gas
producers which use coke. Since ammonia soda manufacturers already operate by-product coke
plants producing coke in quantities, they have an abundant supply of these raw materials. Further,
they can use for the manufacture of ammonia soda the carbon dioxide gas washed from the water
scrubber. Hydrogen gas may also be collected from cathode chambers of the electrolytic caustic
cells operated by them. Moreover, nitrogen fixation plants need soda ash as one of the important
raw materials for the production of synthetic sodium nitrate. For these reasons, the synthetic
ammonia industry may well associate itself with the ammonia soda industry. Indeed, the Solvay
Process Co. At Hopewell is the largest anhydrous ammonia producer of its kind in the United
States.
It may be noted that with the advent the synthetic ammonia industry there is a tendency to
modify the orthodox Solvay process to permit the use of synthetic ammonia in the process for the
simultaneous production of ammonium chloride as fertilizer or as a chemical for dry battery
manufacture. This undoubtedly has paved the way for the union of the ammonia soda industry
with the synthetic ammonia industry.
As already pointed out, the recent market relationship between chlorine and caustic has been
somewhat disturbed. Increased demand for chlorine has created a situation whereby increased
output in caustic soda can find no new outlet. This situation is exactly the reverse of what existed
in the early days of the alkali-chlorine industry. On this accounted another process of obtaining
chlorine without its counterpart in caustic soda production has attracted some attention. This was
undoubtedly one of the stimuli that caused the Solvay Process Co. At Hopewell, Va., to perfect a
process for obtaining sodium nitrate directly from salt. Prior to 1936, the plant at Hopewell had
used exclusively soda ash from the Syracuse soda plant and nitric acid from the oxidation of the
synthetic ammonia produced at the Hopewell plant, for the manufacture of sodium nitrate. When
the price of sodium nitrate fell to the low level, it became necessary to substitute a cheaper raw
material for soda ash for the manufacture of sodium nitrate. An old process therefore was revived
and improved for the production of sodium nitrate directly by treating salt with nitric acid,
obtaining liquid chlorine as a by-product. The reactions are

NaCl + HNO3 → HCl + NaNO2

3HCl + HNO3 → Cl2 + NOCl + 2H 2 O

and the combined result is

3NaCl + 4HNO3 → 3NaNO3 + Cl2 + NOCl + 2H 2 O

The nitrify chloride NOCl may be absorbed in sulfuric acid, forming nitrify sulfuric acid:
NOCl + H 2 SO4 → SO2 .OH .ONO + HCl

or it may be absorbed in a ferric sulfate solution containing strong sulfuric acid. The nitrify
sulfuric acid is then decomposed by steam and air into NO2 and H2 SO4 , the latter being recovered,
concentrated and used again. Absorption in sulfuric acid with the formation of nitrify sulfuric acid,
however, has been abandoned because of the large quantities of sulfuric acid required and the cost
of re-concentrating large volumes of dilute sulfuric acid. Briefly, the process may be described as
one in which salt (NaCl) is treated with dilute nitric acid (about 70 per cent HNO3 ) and the
mixture heated in a reaction vessel. After most of the gaseous products have been driven off,
boiling is continued and the steam evolved is led into a fresh reaction mixture that is being heated
to the boiling temperature. This may be carried out in a series of reaction vessels or in bubble-cap
columns. The gases coming out from the top consist of a mixture of NOCl, Cl2 and H2 O and are
treated with hot concentrated nitric acid (about 80 per cent HNO3 ) in an oxidizing column for the
oxidation of NOCl to Cl2 and NO2 .

2 NOCl + 4 HNO3 → 6 NO2 + Cl2 + 2 H 2 O

or are oxidized to NO2 and Cl2 by air in a reaction vessel using calcium aluminum silicate(elite)or
aluminum oxide as a catalyst at temperature of 300-400 . And under a pressure of about 8
atmospheres.* The resulting gas mixture is them separated into its components NO2 and Cl2 in a
rectifying column, where NO2(or N2 O4 ) is cooled and condensed, and the free chlorine gas leaving
at the top is liquefied to liquid chlorine in the usual way. The condensed NO2 is then heated and
oxidized to HNO3 in an absorber with air under a pressure of 4-8 atmospheres, and the HNO3
formed is absorbed in dilute nitric acid solution (40-60 per cent HNO3 ) obtained from the nitrate
evaporators, so as to enrich the acid to 70 per cent HNO3 to be used again in the reaction vessels.
From the bottom of the absorber, concentrated nitric acid (about 80 per cent HNO3 ) is obtained,
and this is used for the oxidation of NOCl in the oxidizing column described above.
The sodium nitrate solution from the reaction vessels contains excess of NaCl and some free
HNO3 which is neutralized with soda ash. The neutralized solution is at first concentrated under
reduced pressure at a low temperature to separate sodium nitrate crystals and then under
atmospheric pressure to remove NaCl. The crystals are filtered off and the mother liquor
containing NaCl and NaNO3 is returned to the process.
Some NO gas is always present in the gas as the decomposition product or combined with
N2 O4 as N2 O3 which must be eventually oxidized back to nitric acid:
2 NO + N 2 O4 → 2 N 2 O2
N 2 O3 + O2 + H 2 O → 2HNO3
The process is a complicated one because of many side reactions which yield various lower
oxides of nitrogen, and of the corrosive character of the large volumes of hot acids handled. The
oxidation of NOCl for the separation of NO2 from Cl2 and absorption of NO2 as strong HNO3 ,
leaving chlorine as liquid chlorine, are the essential steps involved in the process. The Allied
Chemical & Dye Corp. Have devoted several years’ time and heavy expenditure to perfect the
process. A pilot plant was first erected, giving results that fully justified the construction of a
commercial plant, producing 60 tons of sodium nitrate and 25 tons of chlorine per day. This plant
was put in operation in 1936. It now produces 120 tons of sodium nitrate and 50 tons of chlorine
per day.
Thus, a cheaper source of sodium has been used in place of soda ash for the manufacture of
sodium nitrate, chlorine being produced as a valuable by-product. Such a production of chlorine
without the accompaniment of caustic soda is bound to affect greatly the caustic -chlorine situation
and will have important economic significance, now that a cheaper raw material than soda ash is
used for the manufacture of the nitrate and a valuable by-product obtained in the form of liquid
chlorine.
*Beaus, H.A., U. S. Patents No. 2,138,016-7, Nov.(1938); No. 2,148,429, Feb.(1939); No. 2,150,669,
Mar.(1939).
With the synthetic ammonia industry taken up by the ammonia soda manufactures, there are
available considerable amounts of waste CO2 gas from the water scrubber for soda manufacture,
and certain quantities of waste ammonia gas purged off from the synthesis and purification system
of the ammonia plant for use in replenishing ammonia loss in the soda plant. Further, with the
abundant supply of free ammonia which must be fixed into some form of solid ammonium salt for
fertilizer purposes, the ammonia soda process is bound to undergo certain modifications whereby
ammonium chloride will be made more abundantly and cheaply for fertilizer than ever before.
This will ultimately do away with the whole distiller operation and possibly also with limekiln
operation, if the synthetic ammonia manufacture and the ammonium soda production are geared in
proper proportions.
A “synthetic sulfate of soda” plant was established by Matheson Alkali Works, Inc., at Lake
Charles, La., for the manufacture of a substitute for salt cake used in craft paper manufacture, by
sintering soda ash with elemental sulfur in approximately molecular proportions, giving an
equivalent of 96-97 per cent sodium sulfate. The manufacture is profitable and the product is taken
b local paper mills, but it cannot compete with burette (Na2 CO3 .2Na2SO4 ), a double salt of sodium
sulfate and carbonate, which is obtained as a by-product at the American Potash & Chemical
Corp.’s plant at Trine, Calif.) shipped to the Gulf Coast from Los Angeles, Calif. One sample of
the “Synthetic Sulfate of Soda” has the analysis given in Table 104

*Gillespie, W. F., “Operation of Sulfite Pulp Recovery Units on Soda Ash-sulfur Mixture,” Paper Trade Jour.,
Technical Section, P.36 (Dec. 26, 1940).
Thus it is seen that this is merely a fused block of soda ash and sulfur, having potential sulfate
content as in salt cake for craft paper manufacture. It is a granular mass of greenish yellow color
and is used as a make-up material for salt cake in the black liquor furnace, although the sulfur loss
in the flue gases is somewhat higher.
Soda ash is consumed in large quantities by the glass industry, which has developed
simultaneously with the ammonia soda industry. The need of soda ash for glass manufacture is so
great and so essential that it has been the practice with large glass manufacturers to operate their
own soda ash plants. The Columbia Alkali Corp., for instance, is a subsidiary of the Pittsburgh
Plate Glass Co., the Michigan Alkali Co. Was started by the J. B. Ford Glass people (Capt. John B.
Ford), and the Diamond Alkali Co. Is privately owned by Evans Glass interests. This vertical
combination between the raw material and the finished product manufacture as exemplified by the
relation between the soda ash and glass manufacturers is also common in other parts of the world.
For example, in Japan, the Asahi Glass Co. Owns and operates the Asahi Soda Co.’s plant located
at Tobata. Japan. It is reasonable to predict that there may be a similar relationship between the
caustic soda industry and the newly developed rayon industry. The historical relationship between
the soap industry and the soda industry and the soda industry is too well known to be mentioned.

*Chem. Met. Eng., Jan. 1951; Feb. 1941; Biennial Census of Manufactures, Washington, D. C.
In Table 105 and 106 are given the quantities of soda in one form or another consumed by
different industries, From these data the relationship between the ammonia soda industry and the
other industries will be evident.

Ammonia soda manufacturers have also become interested in cement manufacture ni an

attempt to utilize the lime sludge from the manufacture of chemical caustic and the abundant
supply of high silica and high alumina limestone unsuitable for soda manufacture. The lime sludge
from the cauterization of soda ash consists mainly of limestone (CaCO3 ) with fine “sand” and a
small excess of lime. If the caustic in the sludge is washed to below 2 per cent (NaOH+Na2 CO3 )
on the dry weight, it does no harm to the cement product. At least two ammonia soda
manufacturers in this country, the Michigan Alkali Co. And the Diamond Alkali Co., have
operated or are affiliated with cement plants for this purpose. The latter produces several hundred
barrels of cement a day in the same plant in which it manufactures its soda. The ammonia soda
industry forms a good combination with the cement industry; for not only is the waste sludge from
the manufacture of caustic utilized, but also the high-silica limestone, which is not suitable for
lime burning in the shaft kiln, makes excellent raw material for cement manufacture. Cement so
made helps carry the quarry costs.
In the ammonia soda industry, considerable “lean” carbon dioxide gas from the limekilns
cannot be used up in the manufacture of soda ash. This “lean” gas is a good deal richer in CO2
than are flue gases from combustion of fuels. The carbon dioxide from such surplus gases may be
absorbed in a series of two absorption towers using cooled potassium carbonate lye. It is then
boiled off in a lye boiler. The liberated carbon dioxide is cooled, condensed, and compressed in a
three-stage compressor to 1050 to 1070 lbs. Per sq. Inch. The compressed gas is again cooled in
coils until it liquefies. It then is caused to freeze by expansion and the “pop-corn” snow is
compressed into 10-inch cubical blocks weighing about 55 lbs. each. these compressed blocks are
the so-called “Dry Ice.” Often the gas is merely liquefied, placed in 50-lb. Steel cylinders as liquid
carbon dioxide, and sold to beverage and carbonated water dealers. Substantial amounts of liquid
CO2 have been used in coal mining. for this purpose, it is supplied in a cartridge with a heater
element, which by its combustion furnishes enough heat to vaporize CO2 . The coal vein is thus
loosened off by steady “push” and not by sharp explosion or detonation, with the result that a high
percentage of lumps are obtained. Liquid CO2 is also used in fire extinguishers.
T ABLE 107. Annual Production of Liquid CO2 and of Dry Ice
In the United States.*
Year Liquid CO2 (lbs.) Dry Ice (lbs.)
1931 153,574,997 84,954,018
1933 117,382,256 59,057,600
1935 87,657,446 165,123,912
1937 100,715,662 313,217,310
1939 102,208,118 356,893,516

*statistics of Bureau of Census, Washington, D.C.

Dry Ice is 141 F. Lower in temperature than water ices and has about 10 times its
refrigerating effect. Most of the Dry Ice production in 1940(about two-thirds of the total made)
was used in the ice-cream trade. Railroad refrigeration cars for transporting meat and frozen food,
and fishing boats also consumed considerable quantities. The annual production of liquid CO2 and
of solid CO2 (Dry Ice) in the United States is given in Table 107.
Ammonia soda manufacturers have taken up this Dry Ice industry very actively. The
Michigan Alkali plant at Wyandotte, Mich., has a daily capacity of some 200 tons of solid CO2 ,
while the Mathieson Alkali plant at Saltville Va., has about the same capacity. The fact that
ammonia soda manufacturers have gone into high-pressure work, such as ammonia synthesis and
liquid and solid CO2 manufacture, is significant: it means that the industry is extending its
ramifications, utilizing its available raw materials advantageously, and that it will assume a more
important position in the national industries than heretofore. The industry has a group of
chemically trained and mechanically experienced men well qualified for activities in new
chemical fields.
 Chapter XVIII

Manufacture of Refined Sodium Bicarbonate

Sodium bicarbonate, commercially known as baking soda or bicarbonate of soda, is a product
directly related to the ammonia soda industry. Sodium bicarbonate in the crude states is first made
in the carbonating towers or columns and filtered, and hence is also called “ammonia soda.” From
it soda ash is obtained by calcining. But this crude bicarbonate (NaHCO3 ) from the filters contains
as impurities sodium chloride (0.3 to 0.4 per cent NaCl on dry basis), ammonium bicarbonate (3.5
per cent NH4 HCO3 on dry basis), a small percentage of ammonium chloride, and sometimes also
traces of magnesium carbonate. Ammonium bicarbonate is decomposed in the dryers.
Ammonium chloride is converted to sodium chloride,

NH 4 HCO3 → NH3 + CO2 + H 2 O(steam)

NH 4 Cl + NaHCO3 → NaCl + NH 3 + CO2 + H 2 O( steam)

and the sodium chloride formed, together with sodium chloride present as such, is left in the soda
ash. Any precipitate of magnesium carbonate in the bicarbonate will find its way to the soda ash
which, when dissolved, shows turbidity in the solution. When the wet bicarbonate is dried by long
exposure at a low temperature to prevent decomposition of sodium bicarbonate, it is difficult to
decompose the ammonium bicarbonate completely and to 1 drive off all ammonia from
the sodium bicarbonate obtained. For , in a bicarbonate layer 2 as thin as inch, dried at a
temperature of 60 to 70 . Over steam coils for a period of 12 hours, enough ammonia remains in
the dried bicarbonate to give a distinct odor, while about 10 per cent of the bicarbonate has been
decomposed to normal carbonate. This is an unsatisfactory situation, to say nothing of the
impurities left in the product obtained. Therefore pure sodium bicarbonate cannot be made directly
this way.
At first thought it would seem natural to start with this crude bicarbonate from the filters and
refine it. There are, indeed, many patents dealing with the refining of this bicarbonate by
redissolving, recrystallization, etc. On account of the presence in it of 0.7 to 0.8 percent total
ammonia (as NH3 ), which must be recovered, such methods cannot be carried out simply, and a
step involving distillation of the bicarbonate solution is necessary if ammonia is to be saved. The
ammonia soda manufacturers generally prefer to make it from soda ash. This seems to be a
roundabout way, but after all it is simple and convenient from the manufacturing standpoint,
because:
(1) there are no ammonia losses to contend with;
(2) carbon dioxide gas is recovered in a rich form from the dryers, while lean gas from
lime kilns, which is usually in excess, may be used for recarbonating the sodium
monocarbonate to bicarbonate;
(3) almost the same apparatus can be used in settling, carbonating, filtering, etc., as in
soda ash manufacture. In fact, some of the antiquated apparatus from soda works is used
in the manufacture.
On the other hand, there is a process employed by several alkali plants in this country,
working from the crude bicarbonate by dissolving it hot, settling distilling off ammonia,
carbonating the partially decomposed bicarbonate and filtering the precipitated bicarbonate. This
process differs from the above only in the distillation operation, which would not be necessary
when starting from soda ash. This process has merits, because ammonia may be saved and the
CO2 gas from the partial decomposition of the bicarbonate is recovered in a rich form by passing
the exit gas through a condenser and cooler-to condense out steam-in much the same manner as in
the distiller condenser or as in a furnace condenser system for handling CO2 gases from the soda
ash dryers. The same method has been used for the preparation of soda liquor for causticization
from ammonia soda, except that the decomposition is carried somewhat further. (See Chapter XXI,
“Wet Calcination of Sodium Bicarbonate.”) Sesquicarbonate of soda (Na2 CO3. NaHCO3 . 2H2 O)
may be made from ammonia soda most advantageously in this way by decomposing (distilling)
the bicarbonate to a proper degree of bicarbonation, cooling and crystallizing.
The process of manufacture of sodium bicarbonate from soda ash consists in dissolving soda
ash to make a saturated solution, for which purpose a tank dissolver provided with a stirrer, or,
better still, a rotary dissolver is used. The solution is settled is settled in a series of vats (generally
two or three); after cooling it flows by gravity to a large storage tank from which it is pumped to
the bicarbonate tower, which is much the same as that used in soda ash manufacture but has less
cooling surface. Cooling water is introduced at about 18 feet from the bottom of the tower. The
gas employed is generally kiln gas and is introduced by carbon dioxide compressors as usual. The
exit gas from the bicarbonate tower contains as high as 10 to 15 per cent CO2. The draw
temperature is higher (about 39 .) than that in soda towers and the crystals are filtered and
washed, first on a rotary drum filter, as in soda ash manufacture, and then in centrifuges much the
same as those employed in the sugar refinery. Bicarbonate crystals from the filters here contain
about 12-14 per cent water but are dried further in these centrifuges to moisture content of about 8
per cent. Long cycle (as much as 8-10 min.) is used in these centrifuges in order to get dry
product.
These centrifuges reduce the moisture left in the bicarbonate so as to enable it to dry more
readily with less danger of decomposition. The filter liquor contains about 9 titer (i.e., 9 cc. N acid
per 20 cc. sample) and is used over again in making soda solution. Thus, impurities such as
sodium chloride, etc., are accumulated in the liquor. When this filter liquor gets too high in sodium
chloride content through repeated use, it is sent to the soda ash plant for making saturated brine.
Fresh water is then added for replenishment. When a bicarbonate tower gets “dirty” it must be
cleaned; this is done by simply filling the tower with water and heating with steam. The boiled
water is drawn out, and after settling is stored n feed liquor tanks for making soda ash solution. No
sulfide is introduced to the soda solution for carbonation in these bicarbonate towers.
In the study of equilibrium relationship of the Na2 CO3 -NaHCO3 -CO2 system in water, Harte,
Baker and Purcell* arrived at a useful empirical equation
x 2 C 1 .29
= 10
SP(1 − x)(185 − t )
Where C= concentration of total Na in solution in equivalents per 1
T= temperature of solution in .
S= solubility of CO2 gas in water at temp. T under 1 atm. Of CO2 partial pressure
P = partial pressure of CO2 gas above the solution
F= fraction of Na existing as NaHCO3
The relationship was established for a concentration (C) between 0.5 and 2.0 normal of Na,
for partial pressures of CO2 (P) within 1 tam. And for a temperature range from 20 to 70 .
While the investigation has not been extended to such a high partial pressure of CO2 as exists at
the bottom of the bicarbonate tower, this relationship undoubtedly holds good for a partial
pressure considerably beyond 1 atm. With the concentration and temperature of the solution and
the partial pressure of CO2 above the solution given, the degree of bicarbonate (x) can be
determined, because the solubility (s) of CO2 in water under 1 atm. Partial pressure is fixed by the
temperature (t) in question. Values of S are as follows:
T ABLE 108. Solubility of CO2 in water at 1 atm. Partial Pressure.
t( . ) S (Mol CO2 per liter)
15 0.0455
25 0.0336
35 0.0262
45 0.0215
55 0.0175
63 0.0151
75 0.0120
85 0.0090
100 0.0065
*Harte,C.r., Baker, E.M., and Purcell, H.H., “Absorption of Carbon Dioxide in Sodium
Carbonate-Bicarbonate Solutions,” Ind. Eng. Chem., 25,528 (1933).
This Na2 CO3 -NaHCO3 -CO2 is a three-component system. In the Dry-Ice operation where
absorption does not cause any sodium bicar-bonate crystals to separate out, it has only two
phases-solution and gas-and consequently it has three degrees of freedom. When, however, a solid
phase separates out, as in the precipitation of NaHCO3 in the bicarbonate tower, there remain only
two degrees of freedom. The concentration of the solution then corresponds to a point of
saturation which is a constant dependent upon the temperature only. This saturation point in the
bicarbonate tower liquor is approximately 1.5 normal of total sodium in solution.
In the bicarbonate tower operation where the kiln gas (lean gas) containing 41 per cent CO2 is
pumped through the towers at a pressure of approximately 38 lbs. Gage, and the draw temperature
is around 35 ., it is possible to estimate the degree of bicarbonation in the soda solution at the
tower draw. Here 38 lbs. Gage pressure and 41 per cent CO2 by volume correspond to 1.47 atm.
Partial pressure of CO2 gas. Then
x=98% of total Na in solution.

x 2 (1.5 )1 .29
= 10
(0.0262 )(1.47 )(1 − x )(185 − 35 )
 FIG 72 Bicarbonate belt conveyor dryer.

The solid phase is then pure sodium bicarbonate. That is, at such a high partial pressure of
CO2 in the bicarbonate tower, the conversion of soda ash to sodium bicarbonate is practically
complete. This is borne out in practice.
The bicarbonate from the centrifuges is dried, either on a continuous belt conveyor consisting
of a number of returned conveyor sections, all enclosed in a chamber with a steam stove at the
entrance end to heat the incoming air and an exhaust fan at the other end to pull the hot air through
and exhaust it to the atmosphere (Fig.72) or, better still, as described below, in a vertical tube,
about 50 feet high and 30 inches inside diameter (Fig.73). the bicarbonate from the centrifuge is
charged in through a feed table and is blown up by hot air, heated by steam stoves, from a
centrifugal blower at the bottom of the tube. The bicarbonate particles suspended in the hot blast
become dry as they reach the top, where two large cyclones in series are attached to the outlet of
the tube to receive the dried bicarbonate. the exhaust from the second, and generally much larger,
cyclone containing some extra-light particles, is led into a spacious dust-collecting chamber on its
way to the atmosphere.
 FIG 73

Bicarbonate drying tube.

This dust-collecting chamber is built of wooden frames, covered with cheesecloth. The
temperature of the hot air is maintained at 70-90 . And the blast from the centrifugal blower is at
a pressure of 10 inches of water. A 30-inch tunnel, 50 feet high, of such a dryer will produce over
50 tons of dried bicarbonate per 24 hours. From the bottom of each cyclone, the bicarbonate is
conveyed in a screw conveyor to packing bins, from which it is packed into wooden barrels
weighing 400 pounds each or into small wooden kegs weighing 112 pounds each.
A slightly different arrangement is illustrated by Trump’s Vertex Dryer (Fig.74). the
difference lies in placing the exhauster at the exit end of the second cyclone and in placing a larger
cyclone collector directly next to the outlet of the dryer as a bin. The air at the inlet to the dryer is
also heated to a higher temperature.
It might be interesting to follow the historical development of the manufacture of the pure
bicarbonate of soda in the Solvay Syracuse plant. The manufacture of the refined bicarbonate of
soda began in Syracuse in October, 1887. The plant started with five tons of bicarbonate a day and
gradually worked up to 300 tons a day. A plant has recently been added which may double this
output.
 FIG 74 Trump Vertex dryer.

For this purpose a solution of soda ash was made, settled, and recarbonated by sending lean
gas (kiln gas) into a making tower. The crystals were centrifuged, and at the beginning were dried
on trays using a low temperature. Then continuous belt conveyers were developed, consisting of
eleven belt sections one on top of another, each section being about 50 feet long and 6 feet wide
overall. These tiers of belt conveyors were housed in a casing with a fan blowing hot air at a
temperature of 95 . from one end to the other. The bicarbonate was fed onto the top belt by a
roller and dropped from one belt to the other below (similar to that shown in Fig.72) and was
discharged from the bottom belt into a hopper. These dryers were doing 25 tons a day.
When more capacity was needed, through the inventive genius of Mr. Trump, a conical tube
24 inches diameter at the bottom and 48 inches at the top and about 50 feet high, with a feed table
located near the bottom of the tube and an exhauster fan connected at the top of the cyclone, was
built (see Fig.74) and found to have a very large capacity. The air entered at the bottom and was
heated in passing through a steam heater. At first it was thought that the temperature was limited
to 95 . But later it was found to be possible to increase it to as high as 150 . in contact with the
wet blanket, without driving off much CO2. This was tested by putting an opening in the body the
tube, one at every five feet, and measuring the temperature of the bicarbonate at each point
throughout the height. It was found that the temperature of the bicarbonate never got above
40-50 . The exit air temperature is at 50 .and the air velocity in the cone is about 600 feet per
minute. The whole equipment extends to a height of more than 50 feet.
The volume of air required in this dryer varies from 8500 C.F.M. to 15,000 C.F.M. with a
drop in air temperature of 100 . (from 150 to 50 .) The moisture evaporated may be 500 to 900
Kg. Per hour. This gives a capacity of 150-270 tons of pure bicarbonate per day of 24 hours, when
the bicarbonate from the centrifuges fed to the dryer contains 8 per cent moisture. The finished
bicarbonate is fine and may be bolted for different meshes. A little dust may be saved by blowing
the exit gases into bag filters or dust chambers as referred to above.
This type of dryer is also useful for drying other fine crystals, especially when they must be
dried at a reduced temperature. Such materials may include ammonium bicarbonate, ammonium
chloride, sesquicarbonate of soda, monohydrate of soda, etc. The dryer has a thermal efficiency of
70-75 per cent.
The bicarbonate so manufactured has a remarkably high degree of purity, as is shown in
Table 109.
T ABLE 109. Composition of Bicarbonate.
Per cent
NaHCO3 99.80-99.92
NaCl 0.10-0.20
Fe2 O3 trace
This can be understood when we remember that it is made from ammonia soda which had, to
start with, an exceedingly high purity of 99+ per cent Na2 CO3 and 0.3-0.6 per cent NaCl. Students
of Analytical Chemistry will understand why textbooks recommend the use of sodium bicarbonate
as the starting point for preparing standard alkali and acid solutions for Quantitative Chemical
Analysis. The alkali manufacturers undertake to manufacture it to meet the U.S.P. specifications
directly.
Sodium bicarbonate is used for making baking powder, carbonated waters, beverages, in fire
extinguishers, and in the manufacture of leather, drugs, and chemicals. It is also used in small
quantities in stock food for the conditioning of hogs and cattle.
 Chapter XIX

Manufacture of Caustic Soda-Chemical Process

Caustic soda is another important product manufacture by ammonia soda manufacturers. In
many important industries, such as the soap, paper, mercerized cotton, viscose process for
artificial silk, explosive, dyestuff, etc., alkali is employed in the form of the caustic and not the
carbonate. The rayon industry is now on of the principal consumers of caustic. Ammonia soda
manufacturers possessed some advantages for undertaking the manufacture of caustic. First, in
conjunction with the manufacture of sodium bicarbonate which requires an excess of carbon
dioxide gas from the lime kiln, the corresponding excess in lime obtained is best utilized in
causticization; secondly, ammonia soda makes an extremely pure raw material for caustic
manufacture, yielding a product at least comparable in purity with electrolytic caustic; and thirdly,
unlike the electrolytic caustic industry, it has no joint product such as bleaching powder or
chlorine to dispose of.
Electrolytic caustic usually contains considerable amounts of sodium chloride, while the
chemical caustic made from ammonia soda and good lime will yield a product of the degree of
purity shown in Table 110.
T ABLE 110. Purity of Chemical Caustic from Ammonia Soda.
Per Cent
Na2 O 76.8
Na2 CO3 0.6
NaCl 0.2
Further, the lime process has the advantage that salt (NaCl) present in the caustic soda made can
be reduced to the amount corresponding to that present in the original soda ash used for
causticization, by so controlling the process that all of the steam condensate (and little of no raw
water) is used for the make-up (i.e., for washing the mud in the tail washer, for slaking the lime
with the wash water, etc.). Electrolytic caustic is normally saturated with salt (NaCl) at the
temperature of the strong liquor and at the concentration of NaOH in the strong liquor settlers. In
the electrolytic process, calculated on a dry basis, the strong liquor made generally contains about
2 per cent NaCl and 98 per cent NaOH unless a special treatment is further applied (2 per cent
NaCl being the amount corresponding to saturation in the strong liquor at the temperature and
concentration in question); whereas in the lime process, NaCl per cent can be maintained at as low
as 0.2-0.3. This figure is far below the solubility of NaCl in the 48-50 be liquor at the
temperature of 25-35 and merely corresponds to the small amount of NaCl originally present in
the ammonia soda used. For electrolytic caustic, therefore, good cooling and ample settling
capacity must be provided to eliminate as much salt as possible from the strong liquor.

*At present, demand for chlorine has been very much increased and the production of electrolytic
caustic has risen sharply.
The method of causticization is well known:
Na2 CO3 +Ca (OH) 2 2NaOH +CaCO3
The reaction depends upon the low solubility product of Ca++ and CO3 -- ions causing the removal
of solid CaCO3 from solution. As both sodium carbonate and sodium hydroxide are very soluble,
the reaction depends upon the relative solubility of calcium carbonate and calcium hydroxide in
the solution in question. As the reaction proceeds, the concentration of sodium hydroxide
increases. This increase, by common-ion effect, greatly decreases the solubility of calcium
hydroxide until ultimately it is no more soluble than calcium carbonate, when the following
equilibrium is set up:
Na2 CO3 +Ca (OH) 2 2NaOH +CaCO3
Thus conversion of Na2 CO3 to NaOH cannot proceed to completion. Consequently at the end of
the reaction in the resulting solution (lye), neither calcium hydroxide nor calcium carbonate can
remain in solution to any large extent. That is to say, in the causticization reaction the equilibrium
is reduced to that between sodium carbonate and sodium hydroxide in the resulting lye:
Na2 CO3 +Ca (OH) 2 (solid) 2NaOH +CaCO3 (solid)
Therefore,

This equilibrium constant, K, is significant. For, to get high conversion (i. e; to get OH-
concentration in the solution a high as possible and CO3 = concentration as low as possible), the
value of K must be as high as possible. Now in the resulting lye, since the concentrations of
calcium hydroxide and calcium carbonate are small, we can assume complete ionization. So the
solubility product of Ca (OH) 2 (K1 )will be (Ca ++)× (2OH-)2 = 4(OH-)3 , and that of CaCO3 (K2 )will
be (Ca++)× (CO3 = ) =(CO3 = ) 2 Therefore,

Therefore, K is a function of the solubilities of calcium hydroxide and calcium carbonate. From
the mathematical expression we know that the smaller the constant K2 and the larger the constant
K1 , the large K will be. In other words, the more soluble the alkaline earth hydroxide, and the less
soluble its carbonate, the more suitable it will be as a causticizing agent. This is a general principle
governing the choice of a causticizing agent for the conversion of sodium carbonate to sodium
hydroxide or of potassium carbonate to potassium hydroxide.
From the above expression

Where K is a constant at a given temperature, by dividing both sides by (CO3 =) and extracting the
square root, we obtain

Where K’ is the square root of K. Examining the mathematical expression of equilibrium, it may
be seen that the more dilute the soda solution (i.e., the smaller the concentration of CO3 =), the
higher will this conversion percentage be (i.e., the ratio OH-/CO3 = increases as
decreases), with 10 per cent soda solution, using a 1.13 specific gravity liquor
containing 12 per cent sodium carbonate, the conversion ratio is not over 92 to 93 pre cent.
 FIG 75 The causticization equilibrium.

In practice it does not average much over 90 per cent conversion. Too dilute liquor requires an
excessive amount of fuel for concentration; hence a close study must be made between the degree
of conversion obtainable and the cost of fuel for the concentration of the caustic lye. Fig. 75 shows
the theoretical conversion curve giving the percentage of conversion from different concentrations
of soda ash solution on the basis of using dry slaked lime.
Further, in the two reciprocal pairs: Na2 CO3 +Ca(OH) 2 and CaCo3 +2NaOH, there are four
components-any three of the above four compounds with water as the fourth component. There
are therefore two degrees of freedom (temperature and pressure).At a given temperature and under
a given total pressure, there are all together only four phases -two solid phases, a solution, and a
vapor phase. Normally Ca(OH) 2 and CaCO3 in the sludge are the two solid phases in question, but
nothing can prevent the formation of a double salt in place of one of the two solid phases. In fact
such a double salt as Personate (Na2 CO3 . CaCO3 . 2H2 O) or Gaylussite (Ns 2 CO3 . CaCO3 . 5H2 O)
has been found in the sludge from the more concentrated soda solutions, especially at low
temperatures. This would lead to greater losses of soda when too great a soda concentration is
employed.
It may not be out of point out here that the temperature does not influence the degree of
conversion to any great extent. This means that the temperature coefficient in van’t Hoff’s

equation of chemical equilibrium: is small; i.e., -Q, the heat of causticization, is

small. In practice, so long as the temperature is high enough to bring abort the reaction quickly
and to get a rapid settling for the sludge, the temperature of causticization is rather immaterial.
This may not be the case if a causticizing agent other than lime is used .For instance, if strontium
oxide is used instead of lime, the heat of causticization is positive and large, and a higher
temperature would yield a lower conversion.
The following description applies to a small caustic plant .The soda 1.13(17 Bé.) is
causticized in large tanks, 10 to 14 feet in diameter, 8 to 9 feet deep with a flat bottom. Steam Is
used to heat the liquor to 185 to 195 F and a compressed air line may be provided for agitating.
The tanks are each provided with a mechanical stirrer having a set of bevel gears driven by a
pulley from a counter-shait. Line in lumps of 6-inch size or smaller is added to a steel basket
attached to the side and partially submerged in the liquor, in an approximate ratio of 60 pounds
lime to every 100 pounds of sodium carbonate in solution. The exact amount of lime depends
upon the quality of the lime and the amount of excess desired. The bottom and sides of the basket
are perforated with -inch holes. When closer control is desired, instead of solid lumps, milk of
lime having a concentration not less than 250 grams of CaO per liter is used. Several of such tanks
can be used in rotation, one being in operation while the others are settling and washing.
Decantation of the supernatant caustic solution is performed through a swing pipe wit outlet
nozzle attached to the side near the bottom of the tank.
After sufficient lime has been added and the desired conversion percentage obtained as
shown by titration with “double indicators,” the liquor is allowed to settle for about 2 hours. The
clear portion is drawn off through the swing pipe. This gives lye containing about 10 per cent
sodium hydroxide. The lime sludge, with whatever lye remains, is washed with the third liquor
from a previous batch, the stirrer is started, and the liquor heated with steam. After settling, this
gives the second liquor of 7 to 8 be that may be added to the lye tank. This slime is again washed
with the fourth liquor in a similar way; this gives the third liquor, which is used for the second
washing and also for making soda solution for causticization. Finally, the slime is further washed
with fresh water, giving the fourth liquor. This washed slime containing the precipitate of calcium
carbonate with a little alkali that remains, is washed out and sent to waste. A batch of
causticization can be made in every to 3 hours; but this, coupled with the time for washing,
settling and decanting, makes the cycle for each tank about 16 hours. This washing process is a
very important one, although slow and tedious; considerable soda could be lost through careless
washing. Sodium carbonate seems to be held quite tenaciously by the slime, supposedly as a
double salt formation with calcium carbonate in solid from.
A more modern practice that gives more thorough washing with a smaller quantity of water
and a stronger caustic is to use the Dorr system of continuous causticizing agitators and
thickeners.
The following is a description of a large causticizing plant using the Dorr agitators,
thickeners, lime slaker, classifier rotary filters, diaphragm pumps, and rotary limekiln.
Strong soda solution of about 18-20 per cent sodium carbonate is made in a rotary dissolver
using the weak wash liquor for solution. Lime is slaked to a dry hydrate in a rotary lime slaked
using also the weak wash liquor. The hydrated lime coming from the slaker is then suspended in
larger quantities of the weak wash liquor to form a strong milk of lime containing about 250
grams of CaO per liter. The milk of lime passes through a mechanical “classifier” in which the
unburnt stone, the finely divided over burnt lime, and sand are separated, yielding a rather pure
lime suspension. The soda solution is first causticized with a large excess of lime in the primary
causticizing agitators, generally three in a series, from the sat of which the she suspensionis then
settled in large continuous thickener. The overflow clear lye from the first thickener furnishes the
strong caustic for evaporation. The sludge from the bottom of the first thickener is pumped out by
means of a diaphragm pump to a rotary drum filter (generally of the Oliver type), from which the
cake containing an excess of free lime is further treated with about 10 per cent excess of weak
soda solution in a secondary causticizing agitator. The filter liquor from this filter is returned to
the first thickener. From the secondary agitator the lye suspension I settled in the second thickener,
the overflow from which furnishes the weak wash liquor for making soda solution in the dissolver
and for hydrating the lime and making milk of lime in the lime slaker and “classifier.” The sludge
from the second thickener is pumped to the third thickener, in which fresh water is introduced for
the final washing and from which the sludge is pumped to another rotary drum filter. Fresh water
also is used on this filter to wash the cake and the filter water is returned to the third thickener. In
this way, the recovery of alkali is as high as 99.7 per cent. The upkeep is slow, the power
consumption quite reasonable, and the floor space required comparatively small.
Instead of using milk of lime and soda solution, it is also feasible to introduce solid soda ash
(ammonia ash) into milk of lime in a mixing trough together with some wash liquor. Soda ash is
introduced by means of an adjustable pusher driven by a worm gear and is dissolved in the milk of
lime in the mixing trough. About 90 per cent of the causticization work then takes place in this
mixing trough, from which the suspension overflows to the bottom of the causticizer in which the
liquor is heated by steam coils. From the top of the causticizer, the suspension in which the
chemical reaction is practically completed overflows to the settlers or thickeners.

FIG 76 A small caustic plant with lime recovery.

The mud from the last filter is either burnt in a rotary kiln or sent to waste. Rotary limekilns
are suitable for burning the fines like the caustic mud. Too much caustic left in the mud may cause
attack of the refractory lining of the rotary kiln, necessitating frequent renewal and interruption of
operation. This caustic mud is nowadays sometimes utilized in cement plants operated in
connection with the ammonia soda industry. If the alkali content is kept below 2 per cent on the
weight of the dried cake, it does no harm for cement manufacture. If a good quality of limestone is
available, and if the milk of lime has been purified in the “classifiers,” the caustic mud obtained in
this way can be used for the manufacture of white
A very compact arrangement representing a practice different form the above is illustrated in
a small modern caustic (15-20 ton capacity) using 23º Be. Soda solution (18 per cent Na2 COa) for
causticization and causticizing it with concentrated milk of lime (250 g CaO/1)to obtain 12 per
cent caustic liquor . This has the arrangement represented diagrammatically in Fig 76.
In this plant the caustic sludge is reburnt in a rotary kiln (11) and the reburnt lime is stored in
a bin (1) next to a rotary slaked (4) by means of a mechanical feeder (apron type conveyor) and
slaked with water containing weak liquors form wash water. Then the milk of lime goes to a Dorr
classifier (5) for removing unburnt stone or dead burnt lime. This concentrated milk of lime
together with some weak liquor is sent to causticizers (6)----two in series in which a turbine type
agitator is used----and the effluent goes to a Dorr washing tray thickener (7) in which there are
three compartments for counter current washing. The washed sludge is pumped by means of a
diaphragm pump (9) and filtered on a rotary vacuum filter (10), on which the cake is further
washed with fresh water is utilized for slaking lime in the rotary slaker and stored in storage (3),
but the stronger wash liquor may be added to bathe milk of lime for causticization. The clear
caustic liquor (12 per cent Na OH) overflowing form the washing tray thickener is sent to double
effect evaporators for thickener is used instead of a series of three individual t5hickeners, as
mentioned above, to save floor space. Lime recovery is effected by burning the sludge in a notary
kiln. A small amount of fresh lime is required for the make—up.
As has been noted above, the washing of the caustic sludge is a most tedious job. The success
of this chemical process of caustic manufacture, therefore, depends largely on the efficiency of the
washing operation both to cut down the loss of alkali and to keep the volume of the liquor as small
and the caustic as concentrated as possible. Without an efficient means of washing, both of these
requirements would be impossible of fulfillment. It can be readily understood that the rate of
settling of the suspended lime sludge in the caustic liquor has a great influence on the washing of
the sludge by the sludge by the process of decantation either in settling tanks or in the Dorr
continuous thickeners. AS great deal of study has been made of the conditions and ways and
means where by a high rate of settling i. e., of separation of the solids form the liquid phase, can
de secured. It has been found that the coarser the solid particles held in suspension, the higher the
temperature of the solution during causticization (within certain limits), and the more dilute the
lye, the better is the settling rate for the solids form the liquid phase and the more complete is the
separation. In washing the caustic sludge, the purpose is to separate the soluble alkali as much as
possible from the solid lime sludge held in suspension; and the rate of settling of the suspended
solids form the wash liquor has a direct bearing on the ease of washing. In the first place, lime, if
first slaked in the weak caustic liquor to form a dry hydrate and later re-pulped in a desired
quantity of that liquor to form the milk of lime, aids the rate of settling.* In the second place, long
and violent agitation in the causticizing agitators, or high peripheral speed at which the stirrer is
driven, impedes the rate of settling. It has been pointed out that a high concentration of the soda
solution results in a low conversion of the caustic. By using a strong soda solution in the so-called
primary causticizing agitators and a large excess of lime, a strong liquor coupled with a high
conversion is thus secured, the excess lime in the sludge being later treated with more soda
solution in a secondary causticizing agitator. To aid the separation of the liquid from the solids, a
rotary filter is employed, so that the caustic solution may be more completely separated from the
solid sludge. As it is theoretically impossible with a finite number of washings to wash out all the
alkali from the sludge, considerable alkali remains in the sludge in the form of a liquid retained in
the solid particles. The moisture in the sludge from free settling in the thickeners is generally as
much as three times the weight of the dry solid particles. If the moisture content in the mud can be
cut down it is evident that the loss of alkali in the washed mud will be greatly reduced. Hence to
minimize the loss of alkali in the caustic mud, the sludge leaving the third thickener is pumped to
another rotary filter and the mud cake washed on the filter with fresh water. Thus, the loss of the
alkali is cut down to 0.3per cent, giving a recovery o f99.7 percent. The efficiency of the Dorr
system lies in (1) the efficient separation of the solids from the liquid phase leaving the solids in a
compact form of sludge, and (2) in the strictly countercurrent arrangement whereby the sludge is
washed as it passes through each stage. A filter inserted between the primary and secondary
causticizing agitators serves to remove the strong caustic lye from the sludge containing excess
lime, to be further treated with fresh but rather dilute soda solution in the secondary causticizing
agitator. The final filter at the end of the system which handles mud from the third thickener
greatly reduces the amount of the alkali-bearing water in the mud, thus minimizing the loss of
alkali from the system. It was found also that with a moderate rate of stirring during casticization
and employing motto long duration for stirring, so long as the temperature is maintained above 85
(or 185 F) a favorable rate of settling could be secured. A good grade of lime (i.e.; high CaO
lime) also gives better rate of settling than a poor grade.
*W.E.Piper, “Recent Advances in Causticizing Theory and Practice,” Trans. Am. Inst. Chem. Eng;1923
The lye from the strong liquor storage tanks is concentrated in multiple-effect vacuum evaporators,
having generally two of these effects. The commonest types of these evaporators are the Zaremba,
the Scott, the Swenson, and the Kestner, using exhaust steam. Because of the elevation of boiling
points as the caustic liquors become more concentrated, the temperature difference between the
steam side and the liquor side in each effect is limited, and it is generally impractical to use more
than three effects in the evaporator system using exhaust steam. These evaporator tubes are made
of steel or copper (21/2 inches outside diameter and about 6 feet long) but for manufacturing caustic
for the rayon industry, the use of copper is objectionable. Steel rubes are almost as good as nickel
up to 50 per cent, but above this concentration pure nickel tubes are superior. Where steel tubes
are used, the effect of embitterment of the thin steel tubing by the action of strong caustic makes it
rather unsafe to use steam of higher pressure than exhaust steam. The use of a high-pressure steam
(say 125 pounds per square inch) is, however, made possible by the introduction of pure nickel
tubes. Ordinarily, the caustic liquor from the last effect is concentrated up to 46 to 49 Be.,
corresponding to a point of minimum freezing temperature in the freezing point curve for caustic
solutions within this range. This hot strong caustic is drawn out into tall, cast-iron tanks in which
salt (NaCl) and soda (Na2 CO3 ) separate out at the bottom, these two salts being only sparingly
soluble in the caustic solution at this concentration. Separation of these two salts from the caustic
is aided by cooling ; but if it is allowed to cool too far, the caustic becomes so viscous that the
rate of settling may be retarded and considerable sodium hydroxide crystals at the bottom of the
settling tanks. One analysis of this concentrated caustic hydrate sediment (from liquor of 1.50 sp.
gr. at 15 .) is given in Table 111.

The composition of , caustic is approximately as follows:

 It is necessary to bring the caustic liquor in the multiple-effect evaporators to 47 to 49
Be. So that on settling and cooling both sodium carbonate and sodium chloride may settle out ,
yielding a high-grade caustic for fusion; for these two salts become sparingly soluble at this
concentration. Even then, if these two salts are present in quantities up to their respective
saturation points (NaCl in the electrolytic method and Na2 CO3 in the lime process), there could be
approximately 1.0 per cent NaCl and 0.5 per cent Na2 CO3 remaining completely dissolved in a 50
per cent NaOH liquor at 30 . Therefore, the best way is to keep out these tow salts as far as
possible. If, for some reason, these two salts have not been eliminate from the caustic liquor, it
would be impossible to expect a high-grade caustic from the pots. For good caustic can come only
from a well settled, properly concentrated liquor. A sample of solid caustic obtained from a well
settled, properly concentrated liquor. A sample of solid causticobtained from a will-settled but
insufficiently concentrated caustic liquor(38 Be) gave the following analysis:

It will be seen that if the caustic liquor from the evaporators were not up to normal strength
for settling, many of these impurities would remain dissolved and would not settle out.
The clarified 46 to 49 are. Solution can be drawn to a number of cast iron, open-fusion
pots, generally direct-fired with natural gas, oil, or coal. Sometimes hydrogen gas from the cell
room is burnt under these pots, in which case a pilot gas pipe must be provided burning natural
gas or coal gas all the time to prevent accident of explosion caused by the hydrogen flame being
extinguished, leaving an explosive mixture in the furnace. As these heavy, cast-iron, direct-fired
pots are very inefficient in the utilization of heat, attempts have been made to substitute
evaporators for the pots at least during the earlier part of the pot fusion work. With a single-effect,
nickel-tube evaporator, using 125 pounds per square inch of steam, it is feasible to concentrate
further the caustic by means of evaporators from 48 Be to as high as 75 per cent NaOH
concentration using forced circulation. A few alkali plants have already adopted this procedure.
The cast-iron fusion pot above referred to is approximately 10 feet inside diameter and 6 feet 3
inches deep, the shell of the pot being about 3 inches thick at the bottom. Each pot of this
dimension will hold 19 tons of finished solid caustic. Some caustic pots in use are much larger
than these, but these are considered to be the standard size now. (on the other hand, some pots are
only 6 feet in diameter by 4 feet 6 inches deep, holding 10 to 11 tons of finished caustic per batch.)
These pots are frequently arranged in two rows. The back pots, placed 12 inches higher than the
front ones, are for preliminary concentration, while the front one are use for the final
concentration. Strong caustic liquor enters the back pots and receives preliminary concentration
before it is transferred to the front pots, where more intense heating occurs. This arrangement
gives a countercurrent gradation in firing and avoids the inconvenience experienced when adding
weaker caustic liquor to a pot containing nearly completely dehydrated molten caustic. It also
shortens the working cycle for the final pots and prolongs their life. The pots are made of dense
gray iron, preferably with some nickel added (1 to 11 /2 per cent nickel) to make the composition
more resistant. The greatest skill is required in casting them. In their settling, these pots should be
supported at the bottom rather than suspended from the top flanges (Fig.77). Local overheating is
very serious and should be guarded against, especially near the bottom of the pot toward the fire.
The flame usually comes in near one side of the furnace wall, is caused to surround the pot, and
leaves by the other side of the furnace on its way to the stack; a narrow partition wall divides the
front into two halves. Recommended For this reason it is Recommended that these pots be turned
around 60 or 90 after about every 15 heats. With proper care in maintenance a fusion pot
should last for about 200 heats.

FIG 77 Caustic pot as installed.

When the 48 Bé. Caustic is charged to back pots, the liquor at first boils violently, but soon
the ebullition subsides although the temperature steadily rises. Pungent odors are noticed. When it
has been concentrated in these back pots for from 24 to 48 hours depending upon the size of the
charge and the method of firing, it is transferred to the front pots continuously and raised to a
strong heat. Finally small bubbles come up flickering through the quiet molten mass. The
temperature here will have reached 500 . ( as shown by the incipient fusion of anhydrous cupric
chloride contained in a thin iron tube immersed in the molten caustic.) * After an hour or more,
the fire should be gradually withdrawn and the pots allowed settling until a temperature of about
350 . is reached.
In settling, most of the sodium carbonate in the molten mass separates out, but little sodium
chloride is thrown down at this final stage. Practically all the iron oxide, however, should have
settled to the bottom. The end point of fusion can easily be judged by the behavior of the molten
mass in the final pots and by examining the texture of a sample of solidified caustic obtained. The
time required for cooling and settling is from 16 to 24 hours, making a complete cycle of
operation from 3 to 4 days. The color of the caustic obtained sometimes slightly reddish. The
former condition is caused by the incomplete oxidation of iron and other impurities, due to
insufficient settling. Proper fusion followed by a quiet settling is a sine qua non for a colorless,
high-grade product. The presence of manganese in the form of Na2 MnO4 causes a dirty green
color, in which case a little flowers of sulfur should be thrown over the surface as it cools, but
before it is cooled to 400 . Too much sulfur in the caustic imparts a yellow-red tinge due to the
formation of fused sodium sulfide. Normally 5 to 15 lbs. of sulfur are added to each pot. If fusion
is carried out properly and the molten mass given proper settling, nothing need be added to obtain
a good color and a proper product, but generally the sulfur treatment is desirable. When the
content of a final pot has been allowed to settle and cool to about 350 ., the molten mass is
ladled, or better, pumped out by an air-driven vertical centrifugal nickel pump, into drums made
from No.22-gage steel plates. Each drum contains 700 pounds net of solid caustic. The mass
solidifies on cooling with a conical depression at the center and the drum is immediately sealed to
prevent absorption of moisture. This is the commercial 76 percent (Na2 O) caustic containing over
98 percent sodium hydroxide. The specific gravity of the solid caustic varies somewhat but is
almost 2.0. The chemical analysis of a good “76” caustic is given in Table 114.
*A recording pyrometer is now regularly installed to register the temperature cycle in the pot.

When the strong caustic liquor for some reasons contains large amounts of iron, a dark scum
appears on the surface of the caustic at the beginning of concentration in the back spots. The scum
is found to have the composition given in Table 115.

Unlike the LeBlanc the black liquor, there is no sulfide or thiosulfate in the ammonia soda
and no niter need be added. But good settling is very essential. That portion of molten caustic left
at the bottom after settling, known as the caustic “bottoms”, is of reddish-brown color and
contains considerable ferric oxide and sodium carbonate and some sodium chloride. One analysis
of such a sample is given in Table 116.

When there is a ready market, the caustic bottoms are sold as such; otherwise it is dissolved
to make 48 liquor and the iron then separated by decantation.
In Table 117 is a temperature record for an experimental (11 /4 ton) fusion pot whose pot
whose cycle of operation was about 42 hours.
 It will be noticed that at the beginning the temperature of the caustic rises only slowly, but
toward the end when the specific heat of the more or less completely dehydrated caustic decreases,
it rises rapidly. The trend of the curve in Fig. 78 shows this.
In Table 118 is given a temperature record of a 19-ton coal-fired caustic pot whose cycle of
operation was about 72 hours. Chemical (Lime) caustic of 50 be. Was concentrated in these pots
and the temperatures were recorded by an alumel-chromel pyrometer with a compensated cold end.
A long C. I. tube is used for the protection of the pyrometer bulb. The pot was fired rather slowly
and the highest temperature of the molten mass did not attain quite usual normal peak.

Fig. 78 Temperature of caustic in pot vs. time in hours,.

Again we note that except for a short duration at the start when dilute caustic is being boiled,
the temperature of the caustic in the pot rises only slowly, but toward the end of the fusion
operation when the specific heat of the more or less completely dehydrated caustic decreases, it
rises rapidly.
Solid caustic is also made in a flake form or in a stick form, the former being frequently
preferred by customers. It is made by scraping off the solid film from a rotating water-cooled
cylinder immersed in molten caustic (Fig. 80), while the latter is made by casting it in molds.
Powdered caustic is made by grinding the solid caustic At present much of the caustic is sold in
liquid form containing about 50 per cent, or sometimes even 70 per cent NaOH.

This method of concentrating the caustic liquor to a molten state is rather wasteful of fuel. It
is estimated that about 0.5 ton of coal is required under the pot to obtain one ton of soled caustic
starting from the 50 per cent solution. Even when the strong liquor has been concentrated in the
evaporators to 70 percent NaOH, the fuel required in the pots still amounts to approximately 400
lbs. of coal per ton of solid caustic made.
Attempts have been made to do away with these fusion pots and substitute evaporators of a
special design, or stills, using a liquid not miscible with NaOH. To obtain the required temperature
for evaporation, the use for the heat vehicle of some working fluid other than steam, whose vapor
has a high temperature at a comparatively low pressure is very essential, and biphenyl* offers
some possibilities. Recently a process was developed for the concentration of 50 or 70 percent
liquid caustic by partial pressure evaporation in a steel evaporator, using kerosene of an average
boiling point of 431 F. The process may be carried out in a nickel-clad still pot provided with an
agitator and a closed-coil heater using superheated steam or a high-boiling liquid. The ratio of
kerosene mixed with liquid caustic to caustic is 2:1, and the kerosene in the vapor phase is
separated from the water in a condenser. A bubble-cap or packed column may be provided on top
of the pot still as a rectifier. The weight ratio of kerosene and water-Crystals of caustic separate
out in a very fine form of the size of 70-100 mesh in the kerosene slurry drawn from the bottom of
the still and may be separated from the kerosene by a centrifuge. Using a long cycle of 22 minutes
in the centrifuge, it is possible to reduce the kerosene content in the caustic crystals to 1.8 per cent
by weight or less. Caustic crystals from the centrifuge can further be washed with solvent naphtha
until the residual kerosene in the caustic is reduced to 0.1 per cent or less. The washed crystals are
then dried at about 170℉. The naphtha kerosene mixture is then separated in a fractionating
column. Anhydrous caustic obtained in this way is fine, crystalline, and free flowing, but may
contain small amounts of kerosene and become discolored. The average temperature in the pot
still is 423o F, and the heat required for dehydration is only about one-third of that required in
ordinary caustic pot fusion. Compared with solid caustic in a large mass obtained by pot fusion,
and packed in steel drums, these finely divided crystals have many advantages in industrial
applications.

FIG 79 Curve showing temperature of caustic in fusion pot.

* Badgen, W. L., Monrad, C. C. and Diamond, H. W. “Evaporation of Caustic Soda to High Concentrations by
Means of Diphenyl Vapors,” Ind. E ng. Chem., July , 1930.
+”Anhydrous Sodium Hydroxide” by D. F. Othmer and J.J. Jacobs, Ind. Eng. Chem., Feb., 1940, pp. 154-160.
 FIG 80 Caustic flaking drum.

However, a complete elimination of the last trace of kerosene from the caustic obtained is rather
difficult and there may be objection in its use in certain industries where traces of kerosene are not
tolerated. Also, the process requires a temperature around 900oF and is wistful of heat and severe
on the pots, the distillation method requires a temperature of only 430o F.and the heat efficiency is
much better. Further, the operation by this partial pressure method may be made strictly
continuous and carried out under a partial vacuum.
As has been mentioned before, caustic soda used for rayon manufacture must be of the
highest grade and free from many impurities. For this, liquid caustic is generally supplied. Rayon
and Cellophane manufacture by the viscose process require the digestion of some form of natural
cellulose with caustic soda. The very rapid growth of the rayon industry has appreciably increased
the market for caustic soda. Presumably, all impurities have a harmful effect in the viscose process,
but any discussion of the production problems of rayon is beyond the scope of this book. There is
an inherent and almost insurmountable difficulty in preparing the anhydrous material with
negligible impurity content. Solutions up to 50 per cent NaOH by weight are much less corrosive,
and a 45-50 per cent caustic solution has also a lower freezing temperature in this range as seen
from the freezing point curve of the caustic soda solution (see Appendix). Consequently a 50 per
cent solution is made and delivered to the rayon plant in a high degree of purity. For this reason
and with the incentive of favorable freight raves established for tank-car lots, the growth of the
rayon industry has brought with it a rapid increase in the quantity of caustic soda shipped by tank
-cars as a 50 per cent solution, It also contributed to the commercial standardization of the 50 per
cent strength.
The following analysis represents an acceptable liquid caustic.
TABLE 119.Chemical Analysis of Rayon Caustic ("50%"Solution).
Na2 O 37.4%
NaOH 47.3
 Na2 CO2 0.18
Al2 O3 0.02
Fe2 O3 0.02
NaCl 0.30
Na2 SO4 0.15
SiO2 0.34
CaO 0.01
In 1924 the rayon market took only a small fraction of the output from the chemical caustic
plants, but in 1929 the rayon trade represented an appreciable portion. The effects of impurities
were not yet understood but were commencing to be appreciated. By 1934 the demand for higher
purity had increased in intensity. Quite remarkable decreases in the content of alumna, iron, and
silica had already been achieved. The effort to reduce further the impurities of calcium,
magnesium, sulfates, and heavy metals is still proceeding. Iron and silica have been the cause of
the most frequent complaints. Lately, sulfates and alkaline earths, as affecting the clarity of the
solution at low temperatures, have been receiving attention. Occasional complaints regarding
impurities not even included in the analysis above have been made.
New laboratory technique and analytical methods have been introduced and are still being
perfected in connection with this requirement of a high degree of purity. Even the spectrograph
and accompanying light densitometry are now equipment for routine for routine laboratory
procedures.
Alkali manufacturers use two general schemes to produce such high-grade liquor. The first is
to use the purest of raw material and avoid all possible contamination of impurities (such as by the
use of all-nickel evaporators, nickel-clad vessels, etc.) The second is to use various treatments or
purifying processes on the finished caustic just before shipment. The method of manufacture gives
rise to the presence of certain characteristic impurities in the liquid caustic obtained. For example,
the electrolytic process, when using the diaphragm cell, introduces a unique problem with the
NaCl. The mercury cell avoids this. The chemical (lime) process from ammonia soda ash
introduces unique problems with sulfates. The heavy metals as impurities are common to all
processes because of contamination from the concentrating apparatus. Competition on purity us
extremely keen in America. The alkali manufacturers jealously guard their methods of producing
pure caustic.

FIG 81 Tank car for liquid eaustic.

 This very pure product cannot be shipped in ordinary iron tank-cars without acquiring very
appreciable iron contamination during transit. A number of coatings which are principally
rubber-base paint have been in use for several years for lining the insides of caustic tank-cars. A
typical construction and arrangement of a caustic tank-car is shown in Fig.81.It is equipped with
steam coils for thawing the frozen content for unloading at sub-zero winter temperatures.
For the shipment of liquid caustic by water, a tank boat similar to an oil tanker has been used.
Since attention was turned to the manufacture and use of liquid caustic in the industries,
much investigation has been conducted to determine the thermo chemical properties of caustic
soda solutions, such as heat content (enthalpy) and specific heat (heat capacity) in relation to the
concentration, temperature, and vapor pressure of caustic soda solutions. Merkel in Germany first
studied these properties and in 1928-1929 published diagrams and tables. Later McCabe and
Beretta, at the University of Michigan, extended this study for concentrations from zero to 50 per
cent NaOH and below 200o F. ; and McCabe and Wilson further extended it to concentrations
between 52 and 78 per cent NaOH in 1934-1936.These results may now be shown in Fig.82.

FIG 82 Heats of dilution for caustic soda solutions (enthalpy chart)

Although certain data beyond the range of 50 per cent NaOH and above 200o F.are not yet
final, these workers have given us a set of valuable data in a tabular form as a chart, that serves
much the same purposes for the caustic soda manufactures and users as the steam tables do for
power plant engineers. This so-called Enthalpy-Concentration Chart (Fig.82) has been a great help
to all calculations in problems dealing with caustic soda concentration, evaporation, mixing and
dilution, and in the thermal effects of a caustic soda solution used as a medium in thermo
compressor work. Details of physicochemical research, its method of attack, and its theoretical
derivation cannot be entered into here, but the readers are referred to these original articles.* A
Boiling Point-Concentration curve for solutions of sodium hydroxide at atmospheric pressure is
given in Fig.83. A specific heat table for caustic soda solutions of various concentrations is found
in the Appendix.

FIG 83 Boiling point of caustic soda solutions.

*Merkel, Z. Ver. Deut. Ing., 72,109 (1928).

Merkel, Z. ges. Kalte-Ing., 35,130 (1928)
Merkel, Arch. Warmewirt., 10,13 (1929)
Merkel and Bosnjakovic, “Diagramme und Tabellen zur Berechnung der Absorptions-Kaltemaschinen,”
Julius Springer, Berlin (1929).
McCabe, W. L., “The Enthalpy-Concentration Chart-A Useful Device for Chemical Engineering
Calculations,” Trans. Am. Inst. Chem. Engrs., 31,129 (Mar. 1939 for Nov. 1934).
Haltenberger, W., Jr., “Enthalpy-Concentration Charts from Vapor Pressure Data,” Ind. Eng. Chem., 31,783
(1939).
Bertetti, J. W., and McCabe, W. L., “Sodium Hydroxide Solutions,” Ind. Eng. Chem., 25,247-8 (1936).
Bertetti, J. W., and McCabe, W. L., “Specific Heat of Sodium Hydroxide Solutions,” Ind. Eng. Chem.,
28,375-8 (1936).

TABLE 120.Production of Caustic Soda in the United States.*

Chemical Process Electrolytic Process Total
Year (short tons) (short tons) (short tons)
 1921 163,004 75,547 238,591
1923 314,195 122,424 436,619
1925 355,783 141,478 497,261
1927 387,235 186,182 573,417
1929 524,985 236,827 761,792
1931 455,823 203,657 658,889
1933 439,363 247,620 686,983
1935 436,980 322,401 759,381
1937 488,807 479,919 968,726
1939 530,907 494,104 1,025,011
1940(Estimated) 500,000 595,000 1,095,000
*Chem.Met.Eng.,48,92(1941);U. S. Census of Manufactures.

It must not be supposed that since the advent of the electrolytic process the quantity of the
caustic soda made by the chemical(lime) process has been diminished. On the other hand, as Table
120 shows, chemical caustic production in the United States has been holding its ground, although
the electrolytic caustic production has of late increased very sharply because of the demand for
chlorine.

Fig.84 shows the caustic production by the lime and electrolytic processes during the last two
decades. From this it is clearly seen that the electrolytic production has now overtaken the
chemical (lime) production.
Also, because of increased competition from the electrolytic caustic, the ammonia soda
manufacturers themselves today have gone into the manufacture of electrolytic caustic. Since the
advent of the alkaline-chlorine industry in the beginning of the present century, electrolytic caustic

FIG 84 Curves showing production of lime and electrolytic caustic.

production in this country has come to the fore very rapidly, growing from a very insignificant
production in 1910 to a point where its volume of production is now greater than of chemical
caustic. For this reason a chapter will be devoted exclusively to the electrolytic production of
caustic soda and chlorine. (See Chapter XX.)
 RECOVERY O F LIME FROM CAUSTIC M UD
In Fig.76 we have shown a rotary kiln as a part of the equipment in the chemical caustic plant
for returning the caustic sludge. We shall now devote some space to the description of details
concerning the recovery of lime from caustic mud.
The calcium carbonate precipitate resulting from the causticizing reaction between milk of
lime and sodium carbonate solution is handled in one of three different ways.
(1) "Returned" in long rotary kilns to recover lime for additional reaction.
(2) Sent to a by-product plant as raw material of cement manufacture.
(3) Sent to waste.
Of the total chemical caustic produced, probably at present more is made in plants
employing the third method of disposal of the mud than either of the other two. This is explained
by the fact that soda manufacture is most generally conducted where lime is very cheap. The
tendency, however, is in the direction of installing more equipment for returning. Cement is made
only where there is a rather rarely occurring fortuitous combination of cement market, alkali
industry, and the presence of other suitable raw materials for cement manufacture.
Regardless of the method by which the NaOH content in the precipitated mud is minimized,
subsequent preparation of the material for feed to a returning kiln is to get its moisture content as
low as is economical. The physical character of these mud is such that a high degree of dryness is
generally uneconomical, though not impossible, to maintain continuously. The calcium carbonate
in the mud is present in a variety of forms. There are also present (in small concentrations) a
number of double or triple compounds of calcium, magnesium, aluminum, iron and silicon,
present as oxides, hydroxides, carbonates, chlorides, sulfates or silicates. They are generally in the
forms of calcium or magnesium-aluminum silicates, carbonates and sulfates. Variations in the
composition undoubtedly account for the large variations in moisture-retaining properties of the
sludge.
In spite of the fact that the water in the sludge must be converted to vapor at the expense of
additional fuel over and above that required in the simple decomposition of the carbonate to the
oxide, it is probable that a high degree of dryness would bring about materially lower fuel
consumption only if expensive heat recuperation equipment has to be added to the installation.
Therefore, as in the cement industry where the “wet process” is, in actual practice, substantially as
economic al as the “dry process,” the alkali operator strives for a consistency of 35 to 40 per cent
moisture. He is generally much more concerned with a low alkali content in his preparation of the
moods for returning, to avoid excessive attack on the refractory lining.
The kilns for returning are essentially similar to those used in wet-slurry cement practice. A
steel cylinder,6 feet or more in diameter by 150 feet or more long, is set on a slope of a out 1/2
inch per foot. It is lined with refractory materials throughout its length, part of which may be
backed by insulating materials. It is mounted for slow, controllable-speed rotation. The lower end
is fitted with a hood carrying an adequate fuel burner with primary and secondary air inlets. This
end is also arranged so that the lime discharged is removed continuously and automatically either
by drag chain conveyors, or via a lime cooler and conveyors to storage.(See Fig.85 for one type of
rotary kiln.)
The high end of the kiln is arranged to discharge the gaseous products and to receive the feed
of lime sludge. The gases are not recovered for soda manufacture. The upper half or third of the
kiln length is fitted with "chains" in the usual fashion of a wet-process cement kiln. Such chains
are interlaced lengths of common chain, the links generally being about 2-1 /2 inches long and of
3/8-inch rod diameter. One end of a length is fastened to a point in the shell from the top in a
position of 60' to the left of the vertical line. The opposite end of that chain-length is then fastened
to a point in the top of the shell about 12 to 16 feet away axially and at 60'-to-the-left and
60'-to-the-right position each loop of chain is long enough so that the bottom of the centenary is
about two feet above the bottom of the kiln. At one axial location, from 12 to 24 chain lengths
originate and generally two sets of lengths are staggered. As the kiln rotates these lengths dip into
the mud which is slowly flowing along the bottom of the kiln and they emerge thickly coated with
the mud, thereby forming a much extended so face between the mud and gases. This promotes
drying and preheating of the calcium carbonate mud and at the same time the cooling of the
gaseous reaction and combustion products. As the chains rub over each other, embryo lumps of
lime are formed.
Another type of the chain section is made by fastening sections of the chain in planes
perpendicular to the axis of the kiln at equidistant points on a number of angle-iron guide spirals
attached to the shell of the kiln. The sagging of the chains between points of support causes them
to dip into the mud and lift it in contact with the hot gases as the cylinder rolls over.
The chain section extends as far down the kiln as the conditions of operation will permit. If
the chain sec tion is carried too far down toward the fire, the chains will be operating at too high a
temperature. They would not only be heated to destruction at that point, but also are objectionable
in creating dust loss if the material is too dry.
If the chain does not extend far enough toward the fire end, there is no good heat interchange
between incoming wet mud and combustion products,

FIG 85 Sludge recovery rotary lime kila.

with consequent high fuel consumption a more serious difficulty, brought about by insufficient
number of chains, is the formation of "rings." At some intermediate state of dryness the physical
characteristics of the mud in the kiln are such that the material agglomerates or clots into masses.
These will at times adhere to sides of the kiln where there are no chains to break them up and a
ring of smile-baked material builds up and acts as a dam restricting the flow of slurry from above
and introducing a high pressure drop from below.
Ring formation is not alone due to incorrect proportioning of the chain section. Variations in
concentration of the impurities such as iron, alumna, silica, sulfate, etc; and particularly, alkali
content also have a bearing on ring formation. Under good operating conditions the mass emerging
from the chain section is more or less granular. It consists of lumps, varying from 1/2 inch to 3/2
inches in diameter just perceptibly damp and with very little dust and no large lumps.
The best fire is in the nature of a long plumy flame. Pulverized coal or any powder fuel
mixture with sufficient volatile matter to maintain a steady ignition state, is the most suitable.
Natural gas, producer gas, or any kind of oil can be used if the proper burner design and burning
conditions are provided. The choice of fuel depends almost entirely on local cost. Another factor
occasionally affecting the choice of fuel has to do with the strength of CO2 in the kiln gas. If Dry
Ice or liquid CO2 being manufactured from these gases, the fuel, which is richest in carbon, has the
advantage.
The gaseous products of calcination and combustion are cooled by the incoming wet mud,
particularly in the chain section of the kiln. Depending on the coat of fuel, additional kiln length
may be justified to cool these gases to low temperatures if an induced draft fan is provided. In
usual practice they leave the high end of the kiln from about 230-350℉. At very low rates of
operation in long kilns they may occasionally fall to the dew -point. The gas may contain from 18
to 30 per cent CO2 on the dry basis, the strength depending mainly on the moisture content of the
mud feed, the carbon content of the fuel, the excess air required, and the efficiency of the burner
and kiln. It is a common practice to carry a slight excess of oxygen so that there will be
substantially no CO.
The gases precipitate a portion of their dust burden in a chamber at the high end of the kiln
and are then, by means of induced draft fans, exhausted through a suitable chimney. When a liquid
or solid CO2 operation is included, the gas is taken from the chamber just in front of the high end
of the kiln.
Control of this operation is essentially similar to that of burning oyster shells in rotary kilns,
which is described in Chapter VI. Capacity or rate of output, is controlled by kiln speed, and the
firs is controlled by the combination of lime and gas analyses. The "wet kiln" operator has the
additional problem of relatively large fluctuations in the content of moisture in the feed to his kiln,
but on the other hand, he is less concerned with the quality of gas from his apparatus.
The rate of operation of a returning kiln must keep reasonably well in pace with the rate of
operation in the caustic plant which it serves. The only latitude in out-of-step conditions is that
which is provided for in storage reservoirs for:
(1) Washed caustic sludge
(2) Returned dry lime
(3) Milk of lime at returning plant
(4) Milk of lime at caustic plant
All these kinds of storage are relatively expensive. The dry lime storage requires elevators
and conveyors for handing a seriously hot and abrasive material. Both the milk of lime and the
unburned sludge storage require power-consuming agitation to maintain the material in suspension
suitable for pumping.
Storage tanks must be adequate in size to take care of the usual operating difficulties. In case
a breakdown in the recovery plant prevents continuation of the returning, there might be a loss of
sludge, were there no reservoir space to receive it from the caustic plant where it continues to
settle out of solutions even for a long time after the reacting constituents are no longer fed.
Likewise interruption of operation of the recovered lime slaking operation would require
expensive slow-down or shut-down of the returning kiln itself, if there were no place to store
additional freshly burned lime. On the other hand, a breakdown of moderate duration at the caustic
plant should not require an immediate corresponding adjustment of the lime recovery operation.
The most frequent "operation" of the storages in such a system is to take care of variations in
the settling rates of the causticizing precipitate. Probably for the same reasons as were advanced to
explain the variations in moisture-retention characteristics, the rates of settling vary between wide
limits. Only the most skillful operating engineers know the fine points of mud settling.
Consequently, that portion of the mud, which is at any one time held in the reaction vessels of the
caustic plant, also varies between wide limits. The handing of storages is one of the most difficult
operations involved in the returning operation. Expert design is needed to provide adequate
storages, which will not cause excessive power or maintenance costs.
The feed of slurry to the kiln is generally controlled with a "Ferris wheel "feeder similar to
those used in wet-process cement practice. This consists of cups or buckets mounted on the rim of
a wheel. The wheel is driven from the kiln by a synchronous tie, which is connected, to the kiln
motor. The sludge is thus delivered to the kiln in an exactly controllable relationship to kiln speed.
Since there is a small and continuous loss of lime from the system, a lime “make-up” must be
provided (see Fig.76, p. 289). This can be lime from the kilns of the adjacent ammonia soda plant,
or it can be fines of limestone from crushers to be fed into the kiln together with the caustic sludge.
The kiln operator strives to maintain the loss at a minimum. Most of the losses occur at the kiln
end of the system, substantially none at the caustic end. The two essential sources of loss are in the
dust carried off with the chimney gases, and in the undigested lumps rejected from the screen
while the returned material is slaked to milk of lime. The first loss is generally a function of rate of
operation. At normal rates, it will be quite low, possibly 1/2 to 2 per cent of the feed. When the
apparatus must be forced to highest production, this loss assumes rapidly increasing magnitude.
The loss as slaker rejects has little to do with rate of operation. It is a function of impurity
content and also of burning conditions. Certain kinds of silicates are formed in a material of this
nature, particularly when the feed is insufficiently freed of alkali. Some of these silicates are very
high in lime, but they digest in the slaker very slowly (if at all).
Occasionally a portion of the lime made from the returning of the mud is purposely fed into
the prelimers of the adjacent ammonia soda plant and fresh lime from the burning of quarried
stone in vertical kilns from the ammonia soda plant is used as "make-up" in the caustic plant. By
such an expedient, impurities such as sulfates, iron, alumina and silica can be purged from the
system.
The burning of the lime is controlled to give the lowest cost of production for caustic. If there
were nothing but fuel economy in the kiln to consider, a kiln could undoubtedly be designed to
burn the material to a negligible percentage of residual CO2 in the lime. This is very seldom done.
It requires a higher temperature and protracted holding of the materials at the higher temperature.
Under such conditions, the lime "shrinks" and becomes what is known as "dead burned." Such
lime may have very poor reaction characteristics in the causticizing operation. It may react slowly
and may produce a precipitate, which settles very slowly and is difficult to wash free of alkali.
A "soft-burned "or fluffy lime generally contains from 4 to 8 per cent CO2 or more. It
frequently forms a "fat" milk of lime, that is, one having a very high relative viscosity. Such
material is generally difficult to pump and frequently causes trouble in pipelines when a high
strength of milk of lime is desired.
Minimum CO2 in the lime requires high investment in the kiln. Incompletely returned mud
does not generally involv e large cores such as in vertical shaft kiln operation. Nor does it
necessarily involve a high fuel loss, because the moisture evaporated per ton of caustic made is not
strictly proportional to the moisture evaporated per ton of caustic made is not strictly proportional
to the moisture per ton of CaO burnt. A'' quick-settling" lime of good filtering and washing
properties may have contained 10 per cent CO2 leaving the kiln, and may return to the kiln with
lower moisture content than would a slimy mud obtained from lime burned to say 3 per cent CO2.
A skilful lime recovery operator learns by long experience the peculiarities of the materials,
which he is handling. The successful and economical operation of the caustic plant depends on his
skill, care, and vigilance.
 Chapter XX
Manufacture of Electrolytic Caustic,

Chlorine and Chlorine Products

As early as 1800 the electrolytic method of caustic soda manufacture was known, but it was
not until about 1885 that successful commercial application was accomplished. Undoubtedly, the
perfection of the direct current generator, or dynamo, contributed much to the success of the
process. Now for more than four decades it has been an important method for the manufacture of
caustic soda, chlorine, and chlorine products. Caustic soda produced by this process has become a
product of standard purity in the chemical trade. Wherever electric current can be economically
generated, either because of the cheapness of coal or rather because of the presence of abundant
waterpower, the electrolytic method has found wide application. This is notably true of Niagara
Falls. Cells of the diaphragm type are most commonly used for this purpose in the United States;
the commonest ones are the Voice, Allen-Moore, and Hooker S cells, although the mercury cells
of the Caster type are fast coming into use in one modified form or another.
In Germany the horizontal Billiter-Siemens type cell is used very extensively. This type of
cell has a horizontal diaphragm at the bottom, he caustic liquor produced at the cathode flowing
away from the diaphragm under static pressure so that there is no chance for the caustic to come
into contact with the anode above. This separation is aided by the increase in the sp. gr. of the
liquor during electrolysis. Unlike the vertical diaphragm cells, the static head over the whole area
of the diaphragm here is uniform. The diaphragm is made of sheet asbestos covered with
precipitated barium sulfate and fine asbestos fiber. The anode disc is made of graphite, and the
cathode consists of a lattice of iron or nickel wires. The cell liquor is maintained at 70 or 80C.by
means of stoneware heating coils.
Electrolytic cells using brine can be divided into three main classes:
(a) The diaphragm type, in which the cathode liquor is separated from the anode by means of
an asbestos diaphragm. This class has many representatives and is by far the most commonly used
in the United States. Some of these cells have already been mentioned above.
(b) The bell type which has no diaphragm; but the brine is fed into the bell and flows out by
gravity, thus keeping the cathode liquor from coming into contact with the anode. This class of cell
is not in general use in the United States.
(c) The mercury cell, in which mercury is used as the cathode and an amalgam is formed with
the metallic sodium. The amalgam is decomposed in the central compartment or outside chamber
by water, forming almost pure sodium hydroxide solution in a rather concentrated form (sp. gr.
about 1.3). A well-known representative of this type is the Castner-Kellner cell installed at Niagara
Falls. One of the ammonia soda manufacturers (at Wyandotte, Michigan) also uses this type of cell.
The Krebs cell is an outgrowth of this type.
It is not within the province of this work to describe each type of individual cell in detail.
Suffice it to discuss a few of the typical cells, which are most generally installed in this country,
and their operation in detail.
In what follows, we shall describe in considerable detail the application of the process by
means of the Allen-Moore cells of the vertical-diaphragm, rectangular type, which are widely used
in paper and pulp mills throughout the United States; the Voice cylindrical cells, which are now
perhaps more extensively used than any other diaphragm type cells in the Unite States; and the
Hooker S cells, which have been developed in large units with high electrical efficiencies and large
output.
Brine. The brine used is a saturated salt solution made from rock salt To prepare the brine,
rock salt is charged to the top of a saturator, a tall square concrete tank about 6 feet square (inside
dimensions) by 20 to 25 feet high, or in cypress tanks having a diameter of 30 to 40 feet and a
height of 15 to 20 feet. The tank is kept full of salt. Water is introduced from the bottom of the
tank; and as it forces its way to the top, it becomes saturated and flows from the outlet at the top as
saturated brine. The amount of water passed in is so regulated that the overflow is saturated. Salt is
constantly added to keep the tank full, and the mud that accumulates in the saturator is flushed out
at intervals through the outlet opening at the bottom. Salt suitable for this purpose must be rather
free from calcium and magnesium compounds and sulfates, and the brine made must be purified
by chemical treatment before use; otherwise the diaphragms are likely to be clogged by the
precipitated calcium and magnesium hydroxides in the cathode chamber. Of the two main
impurities in the brine (calcium and magnesium), magnesium is the more objectionable, because
magnesium hydroxide, unlike calcium hydroxide, cannot be washed out of the asbestos
diaphragms by flushing with warm water. Renewals for the diaphragms must therefore be more
frequent. This would shorten the life of the cells. It is clear that for this purpose rock salt is
superior to sea-salt obtained from solar evaporation, for the sea-salt usually contains more
magnesium salts. Good rock salt should have the analysis give in Table 121.

The brine overflowing from the saturator should contain about 315 grams of sodium chloride
per liter at a specific gravity of 1.20 at 25 . Good, soft water (warm if it is available) is best for
brine saturation in the saturator.
If rock salt brine is available from brine wells by pumping down water to dissolve the rock
salt to saturation (or nearly so), the rate of pumping must be so regulated that such a concentration
of brine shall be maintained. Natural brine in this country is frequently not saturated, and must be
brought to saturation at the plant. Where such unsaturated brine is impure, it is generally
evaporated in multiple-effect evaporators to crystallize out the salt in order to eliminate the
impurities, and then the salt crystals are dissolved to make saturated brine.
The brine from the saturator or from the wells flows to storage tanks from which it is pumped,
or flows by gravity, to the purifying tanks. These are usually cypress tanks 25 to 35 feet in
diameter and 15 to 20 feet high, to which soda solution (10per cent strength) is added in quantities
sufficient to precipitate the calcium and magnesium compounds as carbonates, the solution being
agitated y compressed air. The wash water containing caustic soda and salt (see below) is used to
make the soda solution. Saturated brine made from precipitated salt (which carries some caustic)
from the salt catchers at the bottom of the evaporators is best admixed to the raw brine in sufficient
quantities to precipitate magnesium compounds as Mg (OH) 2 .Soda ash alone does not precipitate
all the magnesium unless a large excess is employed, and then the precipitate formed does not
settle well. To aid settling, the brine can be heated to 140 F.by steam coils. This treatment removes
the calcium and magnesium, but not the sulfate ions. To test whether sufficient soda ash and
caustic have been added, take about a 100-cc. sample of the brine, filter it into a 100-cc.graduate
containing 10 cc. Na3 PO4 (10per cent solution). Stir and observe any cloudy appearance by
looking down vertically through the solution. After a sufficient quantity of soda. solution has been
added as determined by this test, the brine is filtered through a sand filter constructed of wood and
having a false bottom which is covered with burlap, over which is placed a 4-inch layer of coarse
cinders, then a 2-inch layer of fine cinders, and finally about a 6-inch layer of clean, coarse sand.
Mud collects on top of the sand and the clear brine flows to the acid-proof storage tanks. To
prevent clogging of the filter the sand should be back-washed at intervals. The schedule for
cleaning is to be determined by the individual plants. To save labor, the treated brine may be run
through filter presses, such those made by the Sweet land or other companies.
To correct the alkalinity in the treated brine, commercial hydrochloric acid is added to the
acid-proof tanks in slight excess (about 7 /2 pounds of the commercial acid to every 1000 gallons of
brine), and the solution is agitated by compressed air. This acid reaction is necessary to prevent
any hypochlorite formation in the anode compartment. From these tanks the acidified brine flows
to a constant-level head tank and then to the individual cells through the float-feed control system.
This treated and acidified brine has the composition shown in Table 122.
TABLE 122 Analysis of Treated and Acidified Brine.
SP.GR. 1.196 at 20
NaCl 315 grams per liter
HCl 0.15 grams per liter
NA2 SO4 2.00 grams per liter
CaO 0.67 grams per liter
MgO nil
Fe2 O3 nil
Allen-Moore Cells (KML type). The detail of construction of these cells is shown in the
accompanying photograph (Fig.86) and cross-section view (Fig.87).
 FIG 86 The Allen-Moore cell.

To illustrate one typical installation, the following arrangement may be described. A340-KW
dynamo driven by a steam engine, Diesel engine, or A.C. motor, having a terminal voltage of 230
v. feeds the direct current through 64 such cells connected in series and arranged in two rows along
each side of the room. The bus bars connecting the anodes or graphite electrodes are copper bars
of varying cross-section, which are connected to each graphite electrode by means of a copper
wire. Here are 24 of such graphite anodes arranged in two rows with 12 on each side. These
anodes are placed close to the perforated iron sheet cathode and are separated by asbestos sheets
which serve as the diaphragm supported by the cathode. The gap between the cathode and the
anode surface is about 1 /2 inch, and the lower ends of the graphite anodes are separated from the
cathodes, and from each other, by means of porcelain spacers. From the anodes, the current travels
through the brine, and through the asbestos diaphragm supported by the cathode, which in the
KML type is W-shaped. From the cathode the current is collected by another bus bar which
delivers it to the graphite anodes of the cell next in series, and so on Enclosing the perforated iron
sheet and the graphite anodes is a space, or cathode chamber, in which the electrolyzed brine
collects and drains out of the cell through a U-loop seal at the bottom. The cathode chamber is a
steel tank enclosing the W-shaped perforated iron cathode, and carries near the top of the tank a
vent type for the exit of the hydrogen gas.(See cross-section view,Fig.87.)
 FIG 87 Allen-Moore type KML alkali-chlorie cell.

The anode chamber, which is set on the top of the steel tank, is constructed of carefully
deposited consisting of 1 part cement to 3 parts clean stone or quartz crushed to about 1 /2 -inch size.
The inside surface of the concrete is painted with a coal-tar protective paint. On the top of the
concrete anode chamber is located the chlorine outlet. The anodes consist of individual Acheson
graphite boards 5 /4 " thickx30" long x6"-8" wide. Joints, if any, are to be carefully machined and
fitted together to minimize voltage drops. These graphite anodes last from 12 to 18 months, and
the asbestos diaphragms last from 12 to 18 weeks.
The cell effluent has the composition shown in Table 123.

TABLE 123. Composition of Cell Effluent.

Sp.gr 1.21-1.22 at 25
NaOH 100-110 grams per liter (8.5-9.0%)
NaCl 180-190 grams per liter (15.0-15.5%)
Na2 SO4 10-12 grams per liter (0.8-1.0%)

The reactions involved are as follows:

NaCl Na+ + Cl-
2Na+ + 2 faradays + 2H2 O 2NaOH+H2
2Cl- + 2 faradays Cl2
So that, apart from secondary reactions, hydrogen gas is evolved at the cathode where NaOH
is formed, while chlorine gas is given off at the anode. The brine passes through the asbestos
diaphragm is electrolyzed at the iron cathode and flows off, carrying with it a large percentage of
undecomposed salt so that NaOH may not furnish OH-ions carrying the electric current to anode.
 NaOH Na + + OH-
2Na+ + 2 faradays + 2H2 O 2NaOH + H2
2OH- + 2 faradays H2O + 1 /2 O2
For if the last reaction were allowed to take place, it would cause loss of current as far as
chlorine and caustic soda production is concerned, and so decrease the current, efficiency. Besides,
the oxygen thus formed would attack graphite, resulting in a short life for the graphite anodes and
also in the contamination of the chlorine gas obtained, This is also true when much sulfate is
present, as the SO4 --ions would tend to carry current to the anodes liberating oxygen at the anodes.
It is to prevent these excessive losses that electrolysis is stopped when less than one-half of the
total sodium chloride in the brine has been converted to sodium hydroxide. As chlorine gas is
fairly soluble in brine, if it for some reason comes into contact with sodium hydroxide, either
hypochlorite or chlorate will be formed, according to the temperature conditions, as follows:
2NaOH+Cl2 NaClO+NaCl+H2 O in the cold
but
6NaOH + 3Cl2 NaClO 3 + 5NaCl + 3H2 O in the hot
Either HClO or HClO3 will attack graphite anodes, producing CO2 ; therefore a small amount
of sodium carbonate is always found with the caustic in the cell effluent.
CO2 + 2NaOH Na 2 CO3 + H2 O
The theoretical voltage required can be roughly calculated from the following
thermochemical equation, although the result is not strictly accurate thermodynamically.
2NaCl + 2H2 O 2NaOH + H2 + Cl2 + Q; Q = -H
Heat of formation of 2NaCl in aqueous solution = 2x 96,600 cal .
Heat of formation of 2NaOH in aqueous solution= 2x 112,450 cal .
Heat of formation of 2H2O in liquid state = 2x 69,000 cal .
Q = 2× -53,150 cal

2 × 53,150 × 4.187
E= = 2.3volts.
2 × 96,540
Strictly speaking, we have to use the Gibbs-Helmholtz equation, expressing the relationship
between the voltage and the heat of reaction, plus a temperature-coefficient term. The
Gibbs-Helmholtz equation is
−H dE
E= +T
NF dT
Where E=decomposition potential in volts.
H=heat of reaction in joules for N gram equivalents of the substance=4.185xNo. of cal.
per N equivalents. H is positive when heat is absorbed and negative when it is
evolved.
F=96,540 coulombs per grain equivalent.
T=absolute temperature in degrees Kelvin at which the current is passed through

Separating variables we get

NF dT
dE =
NFE + H T
Integrating between limits we obtain,

E 2 E1 − H 1 1
− =  − 
T2 T1 NF  T2 T1 
 This assumes that the heat of reaction H. between T1 and T2 within not too wide a range of
temperature is substantially constant .N and F being constants, if H is known and assumed
constant the decomposition voltage at any other temperature can be found from a known value at a
given temperature .It will be seen that the above voltage of 2.3 (=E) is obtained by neglecting the
temperature-coefficient term and calling

−H
E= = 2.3
NF
This is the minimum voltage. The actual voltage required is much higher, because of the cell
resistance and of polarization at the electrodes. In practice, 3.5 volts is close to the average figure
for this type of cell. Some energy is dissipated as heat by the resistance of the electrolyte. To
minimize such losses, i.e., to secure highest energy efficiency, the resistance through the
electrolyte is reduced by placing the electrodes as near to each other as possible (about 1/2 -in.
clearance); by having the electrode surface as possible (about 25 sq. ft. to each cell); and by
running the electrolyte rather hot (about 45 ). It is understood that the electrolyte should be
saturated to start with.
To cut out a cell, the current is short-circuited by clamping a copper bar onto the carbon
anode bus bar on the cell out and connecting it to the anode bus bar of the next cell. The brine
level in the cell is maintained constant by means of a glass float valve. Brine is fed from a
cast-iron header from a head tank to the anode compartment, the header being provided with a tee
for each cell, connected by a nipple next to the rubber valve which carries a flexible rubber tubing
for delivering brine to each cell. The weak liquor, or cell effluent, from each cell is drained
continuously into an open concrete trough and then to weak liquor storage tanks. Hydrogen gas is
let out into the room through a vent pipe on top of the cathode chamber, or else it is collected and
compressed for use. In some plants, hydrogen gas produced this way has been utilized for the
hydrogenation of oils. Hydrogen is often burned with chlorine to make synthetic ammonia process
by direct union between hydrogen of the synthetic ammonia process by direct union between
hydrogen and nitrogen gases, hydrogen gas from this source is now an important raw material.
Chlorine gas is collected through a 6-inch stoneware pipe located above the anode compartment,
and is drawn by means of a stoneware exhauster or a lead-coated steel fan through the absorption
towers or bleaching powder chambers. The top of the anode chamber is covered with
cement-asbestos bricks and the whole is looted over with a coal-tar cement or rubber -wax
composition to prevent leaking of air into, or of chlorine gas out of, the anode chamber,.
The voltage across each cell is taken once a day and the cell effluent tested for strength of
caustic and the presence of hypochlorite. Current efficiency may be determined from the gas test
with an or sat. Systematic records are kept of each cell. The asbestos diaphragms gradually
become clogged by solid deposits from the electrolyte, and the resistance to the flow is thus
increased. To maintain the constant flow, therefore, it is necessary to raise the float and run a
higher static head of electrolyte on the diaphragms. With new diaphragms in a cell newly cut in,
the rate of flow may be as high as 35 liters per hour, but in the course of time it gradually
diminishes to 20 liters per hour. Efficiency tests show that with a total current of 1500 amperes
passing through the cells, the volume of effluent should be from 20 to 22 liters per hour .At the
end of about 16 weeks, the asbestos diaphragms may become so badly choked that proper flow
cannot be maintained with the brine level at the highest point. It is then necessary to cut the cell
out, remove the asbestos, and clean the iron plates. A new diaphragm is made with three layers of
asbestos Paper for the bottom half and two layers for the top half. After a cell has been cleaned
and assembled with new diaphragms, the brine is fed in. When the electrolyte has reached the top
of the cathode plate and normal flow has been reached, then and only then, should the cell be cut
in. At first, when the cell is cold, the resistance is high, and the voltage drop across the cell is
temporarily excessive. But as the cell is warmed up, this voltage will drop to 3.3 volts and then
rise up gradually to about 3.45volts.On the other hand, the new asbestos diaphragms being open,
the flow for the first few days will be high, causing excessive amounts of unconverted salt in the
cell effluent for a short time .As the diaphragm becomes clogged, however, the flow is diminished.
After three or four months of continuous running it may fall below 16 liters per hour. This is about
the time to change the asbestos diaphragms, as the reduced rate of flow would cause the formation
of considerable hypochlorite or chlorate in the anodic liquor and reduce the current efficiency,
Oftentimes, the addition of hydrochloric acid to the feed helps prevent the formation of
hypochlorite.
It is often inadvisable to maintain a very high decomposition of brine in the cell effluent. For,
in the first place, a high decomposition may be brought about by increasing the thickness of the
asbestos diaphragm, thus decreasing the volume of the effluent; in which case. The resistance to
the passage of the current would be increased, with con-sequent loss of energy. In the second place,
a high decomposition may be brought about by forcing a larger current through the cells; in which
case, the cell terminal voltage would rise, causing a lower power efficiency and more vigorous
oxidation of the graphite anodes by the oxygen liberated. Low percent decomposition of the brine,
on the other hand, increases the cost of evaporation, especially if the fuel in the locality is
expensive. Therefore, there is an optimum percent decomposition based on the cost of the electric
power vs. the cost of the fuel. With a high power cost and low fuel charges, the decomposition can
be advantageously kept low; whereas with a low power cost and high evaporation charges, the
decomposition should be kept high.
The current density employed in these cells is about 0.4 ampere per square inch .The current
efficiency is 94 per cent, and the power efficiency is about 62 per cent .The decomposition
efficiency of NaCl is less than 50 per cent. Some of the operating results of these cells are
tabulated below:

In analyzing the losses of voltage in the cell, it may be stated that the theoretical
decomposition voltage requires 2.3 volts, the over-voltage necessary (due to polarization, etc)
takes about 0.5 volt, and the ohmic resistance of the cell (0.005,ohm at 1500 amperes) makes an
IR drop of 0.7 volt, giving a total of 3.5 volts for the terminal voltage across such a cell.
 FIG 88 Cross-section of Vorce cell with flow-sheet diagram.

Recently the Vorce cells have been making great headway and are now in certain cases
replacing the rectangular type of cell. The Vorce cell is of circular type (Figs. 88, 89 and 90). but it
differs in no essential details either in principle or in materials of construction from the other
vertical diaphragm cells . It consists of a circular steel tank (made of 16/3-inch steel plate), a
circular steel screen as the cathode, and a circular row of 24 two-inch square graphite sticks 36
long as anodes, with a vertical asbestos diaphragm on the inside of the steel cathode screen These
anodes are suspended from the cover ring. The cover and bottom of the cell are made of cement,
sand, and mixed asbestos fiber, mixed in proper proportions with water and formed in wooden
moulds. Alter these cement rings have taken an initial set, they are dipped in tar solution for
waterproofing. The inside dimensions of each cell are 22 inches diameter by 34 inches high, and
the overall outside dimensions are 26 inches diameter by 42inches high. Its chief advantages over
the rectangular type are (1) the saving of the floor space (the saving being from a third to a half for
 FIG 89

End cells in series showing

feeder connections to anode

and cathode.

the same capacity);(2)the larger anode surface and the integral construction of the anodes resulting
in a smaller anode current dens ity and lower voltage; (3) the simplicity and strength of
construction resulting in the lightness of weight per unit of the cell output and in the saving of the
massive concrete otherwise required;(4 ) the ease with which repairs and renewals may be effected
by lifting off (it necessary) the whole cell bodily for working, each cell empty of the brine
weighing only a little more than 500 lbs;(5) the high purity of chlorine gas obtained (97-98per
cent),which is advantageous when liquid chlorine is to be make from the gas. Consequently, Vorce
cells are now probably more extensively used in the United States than any other diaphragm cells.
What was

FIG 90 Cell room containing many Vorce cells.

said above about the rectangular type of cell, however, applies equally to the Vorce cells, so that
no detailed description of their operation and maintenance will be given here, except that the
anodes in the Vorce may last some-what longer (18 to 20 months )and the life of their diaphragms
is from three to four months between successive renewals. The life of the anodes may be increased
considerably by treating the graphite with linseed oil. The conversion of sodium chloride in the
Vorce cells is also the same as in the rectangular type of cell (105 to 115 grams sodium hydroxide
per liter) The Westvaco Chlorine Products, Inc, South Charleston, W. Va, which controls the sale
of the Vorce cells, gives the data in Table 125 for the monthly average on the cells installed in
their plant.

FIG 91 Hooker S type cell.

A new development in the construction of the Vorce cells consists in the addition of an extra
cathode inside the row of anodes, which adds about 60 per cent more cathode surface and which
also reduces the gap between the anode and cathode surfaces. This further reduces the terminal
voltage per cell. It is possible that a working average voltage per cell of about 2.9 volts may be
maintained with the same current capacity, thus increasing the power efficiency to 70 per cent.
The capacity of the cell may also be increased up to 2500 amperes and the throughput doubled,
without exceeding the present voltage of a standard cell. How-ever, the current efficiency is
lowered because of higher current density and a much higher current concentration (current
concentration is defined as current throughput in amperes per liter of the volume of cell liquor)
From the old Townsend cell (Hooker Type F cell) have been developed the Hooker Type E
cell (1913) and later (1929) the Hooker Type S* cell which has been installed in the Hooker
Electrochemical Co. Of Niagara Falls, N. Y, and works very successfully it is almost a square type
using concrete and steel construction (concrete bottom and top, steel middle section, see Figs.91,
92 ,93) The cell* is insulated to maintain a high temperature in the cell brine and is provide with a
double-walled coil for heat exchange between the effluent and the incoming brine feed.

FIG 92 Hooker S type cell with auodes in place ready for cathodes.

Hence the electrical resistance through the electrolyte is low and a higher concentration of caustic
is possible in the electrolyzed brine-as much as 135g per l. The distinctive features of this cell are
the deposited asbestos diaphragms and the compactly arranged alternate anodes and cathodes
filling the space inside the cell, so that the total anode or cathode area is more closely proportional
to the cube, rather than the square, of the linear dimension of the cell Further, the anodes are not
supported from the top and there is no anode or cathode portion exposed to the gas space above
the liquor. The anode graphite blades are embedded in a lead slab at the bottom of the cell, while
the cathode iron “fingers” are supported horizontally from the steel side frame. There are 28 of the
cathodes and 30 of the anodes, arranged symmetrically on each side of the cell, leaving a central
lane at right angles to the plane of electrodes for the brine passage. The anodes are thus submerged
in the cell brine, leaving the top cover entirely free, so that it may be conveniently lifted off. Thus
the assembly of the cell can be inspected and spacing between each adjacent anode and cathode
checked, before the cover is put on. The cell outside dimensions is 4 ft. 6 in. Wide, 5ft. Long, and
3 ft, 8 in. High; they occupy a floor apace of about 70 aq ft. each, including all surrounding
working spaces. Each cell is able to take 6000-7000 amps. as working average and has a large
*Stuart, Lyster, and Muray, “The Story of the Hooker Cell,” Chem. Met. Eng., 45,358 (1938)
Muray, R. L., “Growth of Electrolytic Alkali and Chlorine Industry in the U. S. Development and Importance
of Deposited Diaphragm Cell,” Trans. Am. Inst. Chem. Engrs., 36, 445 (1940)

output and high electrical efficiencies. There Hooker S type cells are coming into more and more
extensive uses so that more chlorine will soon be produced by this type of cell than by any other in
the country. Variable terminal voltage is used depending upon the age of the cell. Table 126 gives
the average operating results.
We shall briefly describe the mercury cell which is not a diaphragm type but which is coming
into use more and more extensively. The Michigan Alkali Company, Wyandotte, Mich; and the
Mathieson Alkali Works, Inc, Niagara Falls, N. Y, both have this type of cell in some modified
form. In the original Castner cell, the body was made of slate or concrete, having the general
dimensions: 4 ft. By4ft. -6 in high. The anodes were the graphite blocks placed at the end
compartments, while in the central compartment iron grids were used as the cathode. At the
bottom of the cell, mercury was used as a sort of “intermediate electrode.” In the end
compartments, brine was electrolyzed and chlorine was liberated at the anode, and sodium was
amalmagated with mercury at the bottom. At intervals, the cell was rocked by a cam mechanism
and the amalgam from each end compartment was caused to flow to the central compartment
where sodium in the amalgam came into contact with water, forming sodium hydroxide and
liberating hydrogen gas at the iron grid cathode.
2Na +2H2 O 2NaOH +H2
Table 126 Typical Operation Results of Hooker S Cells.
Current per cell 6000-7000 amp.
Voltage per cell 3.35v.
Temperature of cell liquor (preheated brine) 90
Current efficiency 95%
Voltage efficiency 69%
Power efficiency 65%
Active cathode surface per cell 130sq.ft
Active cathode surface per cell 110 sq.ft
Caustic soda in effluent 135g NaOH per l.
Lb. NaOH per cell per day 452-530
Lb. Chlorine per cell per day 400- 470
Life of anodes 450days
Life of diaphragm 200days
Chlorate formation per 1000 lb. NaOH2 1b.

The only criticism about this cell was the rather small capacity per cell at that time and the
expensive material (mercury) used. At present several modifications have been developed, some

of them having a very large output per unit .One of them is the Sorensen cell which is not much
different from the original Castner cell. It is constructed of a cast-iron frame lined with concrete.
The liquor contains 20 per cent NaOH and the caustic made is very pure. The terminal
voltage for each cell is 3.9 –4.3 volts, averaging 4 volts. Another modification is the Krebs type
cell, which uses ebonite and iron frame for its construction. It takes a very large current –as much
as 15,000 amp. per cell-and has thus a tremendous output in these later modifications, mercury is
used directly as the cathode and the amalgam is taken to a denuding chamber for decomposition.
The advantages of these cells are hat the caustic made is practically free from salt (NaCL) and has
a higher concentration, thus reducing the cost of evaporation. A pure product is thus obtainable for
rayon manufacture.
 FIG 95 Curves showing power and current efficiencies.

Considerable advance has been made in the design and operation of the alkaline chlorine
cells of the terminal voltage and in the increase of the power as well as current efficiencies. Indeed,
electrical power and salt are the two largest items of cost entering into this electrolytic production
of caustic and chlorine amounting to fully two-thirds of the total cost of production, so that it is
imperative that power consumption (and consequently voltage per cell) be kept low, and the
utilization of salt (and consequently the current efficiency) be kept high. Today, the industry
occupies a very important position, as it concerns the production of not only caustic, chlorine, and
bleaching powder, but also hydrochloric acid, hydrogen, sodium chlorate, and other chlorine
products. Fig.94 shows the trend of the reduction in the terminal voltage for the past three decades;
Fig.95
shows the improvement the power and current efficiencies obtained during the same period.
Table127 shows the total daily capacity of the alkali-chlorine plants in the United States as of
November, 1940. Other products besides caustic soda and chlorine are also listed. This gives
clearly the possibilities of the caustic soda and chlorine production annually in the United States
by the electrolytic process.
Manufacture of Bleach. Chlorine gas from the anode compartments is conducted away in a
6-inch stoneware main and sent through the absorbing system (bleach towers) by means of a
stoneware blower. The joints in the mains are luted with tar cement, which does not crack on
being subjected to atmospheric changes. The main is provided with a U-tube manometer showing
about -inch water vacuum.
Chemically, as far as the solid bleaching powder is concerned, the bleach is not to be
considered as a mixture of Ca(ClO)2 and CaCl2 , but rather as a compound having the formula Ca
H2 O with a small excess of free CaO. There is evidence supporting this which cannot be discussed
here. Pure Ca (OH) 2, would absorb chlorine to the extent of 43.5 per cent . Theoretically,
1 OCl
according to the formula H2 O, pure calcium hypochlorite should contain 49 per cent
16 Cl and decomposition of the hypochlorite,
chlorine But because of incomplete chlorination of the lime
percentage of available chlorine is lower. he commercial basis ,there fore, is 35 per cent bleach
meaning 35per cent of its weight as available chlorine, although frequently the powder contains as
high as 39 to 40 per cent available chlorine .Even the bleach liquor, which is rather dilute and is
extensively used in ClO
paper and pulp mills, is referred to this 35 per cent basis, The lime used in
Cl
making bleach must be high in CaO, and must contain less than 5 per cent CaCO3 and 1 per cent
of magnesia and silicious matter, and not more than traces of iron or manganese oxides.
For making bleach liquor the lime is slaked either by hand in slaking tanks, or in a continuous
rotary slaked. The milk of lime must be allowed to cool before use.
CaO + Aq Ca(OH) 2 +Aq

The strength of the milk of lime depends upon the strength of the bleach required. For the
bleaching of pulp in paper and pulp mills. where the strength of 60 grams of 35 per cent bleach per
liter is generally required milk of lime containing about22 grams CaO per liter is employed. The
lime solution I solaced in a bleach tank provided with a stirrer. From the bottom of the tank a
pump sends the milk of lime to the top of the bleach towers whence it sprays down and runs over
internal overflows, meeting chlorine gas entering from the bottom of the tower countercurrently.
The tower is made of concrete, and has flat division plates and overflows arranged in a zigzag way.
This provides a long course for the flow of milk of lime and an intimate contact with the chlorine
gas sent in by a stoneware fan. The milk of lime flows from the tower through a gas seal back to
the bleach tank, from which it is again circulated by the pump through the towers until almost all
the CaO is used up (leaving only 2 to 4 grams CaO per liter).
ClO
Ca(OH) 2 + Cl2 Ca .H2 O
Cl

Milk of lime containin35 pounds CaO will make a bleach liquor containing 100 pounds of 35 per
cent bleach.
Usually two towers are used in series to insure complete absorption, and the spent gas is
exhausted to the atmosphere. When one bleach tank is finished the agitation is stopped, and a new
tank containing fresh milk of lime is cut in and circulated in the same way. The bleach liquor in
the tank then is allowed to settle and the clear liquor drawn out through a syphon for use in the
bleaching room. After all the clear liquor has been drawn out, the slime or “grout,” containing an
excess of lime, calcium carbonate, and malicious matter, is washed and strengthend with fresh
milk of lime to the proper strength for a neo batch. If lime is exhausted and the circulation of the
bleach continued, there is a danger of losing the bleach.
H2 O +Cl2 HClO+ H+ +Cl-
2HclO 2H+ +2Cl- +O2
CaClO Cl + H+ +Cl- Ca++ +2Cl- +HCl0
And
2HOCl 2H+ +2Cl- +O2
The liquor must be kept below 35°,or decomposition of the beach and conversion into
chlorate is likely to occur.
If much CaCO3 is present in the milk of lime, the acid is neutralized, and the bleach made is
unstable, the hypochlorous acid formed being soon decomposed.
Cl2 + H2 O HOCl + H+ +Cl-
2H+ + 2Cl- + CaCO3 Ca++ + 2Cl- + CO2 + H2 O
2HClO 2H+ +2Cl- + O2
a decrease in Ph value, i.e. the addition of a mineral acid, would push the reactions from right to
left; whereas an increase in pH by the addition of an alkali would cause them to proceed from lift
to right In the latter case the following equilibrium is established:
ClO- + H2 O = OH- +HClO
The presence of magnesia in the lime will cause the bleach to be decomposed rather quickly.
The maximum amount of magnesia in the lime should not exceed 2per cent.
The presence of iron or manganese oxides n the bleach liquor causes rapid decomposition of
the bleach, so that the lime used should contain very little of these oxides.
2CaClO.Cl 2Ca+++4Cl- +O2 in presence of iron or manganese oxides.
For a strong bleach solution containing 250 grams or more of 35 per cent bleach per liter,
normal strength of milk of lime is used in the beginning, but more lime can be added as absorption
proceeds. The process of chlorination thins down the suspension so that the addition of more lime
later does not affect the flow through the towers. Care must be taken to keep the solution
cold(about 20 ); otherwise the bleach will not settle well and may be chlorated to an excessive
extent:
6CaOCl 5Ca+++10CL- +Ca(ClO 3 )2
Sometimes, the cell effluent itself is used to absorb chlorine in making bleach liquor, but this
is very wasteful.
Manufacture of Bleaching Powder. For making bleaching powder, an absorption apparatus
known as the Hasenclevr’s mechanical chambers is used, consisting of 6 to 8 sections of
lead-lined cast-iron screw mixers, 20 inches in diameter and about 12 feet long, placed one on top
of another and driven by a system of cog-wheels so that the slaked lime travels down from the top
in a zig-zag way until it reaches the bottom tier, where it is discharged out as bleaching powder.
Dried chlorine gas diluted with air is blown in from the bottom and the spent gas let out from the
top countercurrently. The quicklime for this purpose must contain high CaO and very small
quantities of iron or manganese oxides, as mentioned before. Further, it must be slaked with
enough water so that there is present about 4 per cent free moisture in the Ca (OH) 2 formed, or
about 28 per cent total water on the weight of the hydrated lime. The screw consists of a shaft with
a number of spiral spuds attached to it, the spuds scraping the lime as well as pushing it along.
Any tow adjacent sections one above the other have the spiral spud shafts turned in opposite
directions from the motion of the spur gear wheels in direct gearing. Slaked lime sifted through a
screen enters from a hopper on the top. Considerable heat is generated during chlorination, and it
is necessary to keep the temperature below 40 and to run intermittently. On this account, dilute
chlorine gas is preferable. It is a good plan to run a slightly higher vacuum on the chlorine main,
as much as inch of water. This serves to dilute the gas and at the same time keep the loss to a
minimum. Excess of free CaO in bleaching powder runs about 10 per cent. The finished product is
packed in steel drums weighing from 300 to 400 pounds net.
A newer type of mechanical chlorinator is based on the same principle as the Herreshoff or
Wedge burner. There are about ten shelves with a 5- to 6-inch C.l. pipe shaft carrying scraping
arms which are provided with malleable cast-iron rabbles to scrape the lime from one shelf down
to another, first through the central opening and then the side opening in the periphery of the shelf
in a zig-zag way. Chlorine gas is led in through the bottom and the lime is charged from the top.
The shelves are made of concrete with cooling water pipes embedded therein, both for cooling and
for reinforcement. The circular body of the chlorinator is also made of reinforced concrete.
Formerly, large stationary bleach chambers were used, four of thesebeing connected in series.
The chambers were lead-lined concrete strue-tures with asphalt-covered flooring, and were cooled
by cooling coils located under the floor surface. Chlorine gas was passed through these chambers
in series and absorbed by lime on the floor. The lime was introduced from hoppers located on the
roof and spread over the floor by hand. When the desired strength was reached, the remaining
chlorine gas was exhausted from the head chamber to a secondary lime chamber. Men went in to
rake the bleaching powder and clear it out. A new batch of lime was then put in and this chamber
made the tail end lf the series. The conditions were rather hard on the workmen. A gas mask
should be provided for each man working inside.
Manufacture of Liquid Chlorine. Instead of making bleaching powder, large quantities of
chlorine are now liquefied as liquid chlorine and sold to the textile manufacturers. For this purpose,
the gas is kept as rich as possible by avoiding air leakages and excessive vacuum in the main. The
gas is first dried in two stoneware towers in series using 93 per cent sulfuric acid as the drying
medium. The gas is pulled through these towers by means of Nash “Hytor” pumps using 93 per
cent acid as seal in the easing. The pumps, which are made of cast iron, are not attacked by the dry
chlorine gas or the strong sulfuric aced. Each pump can develop a pressure of 25 to 50 pounds and
thus keep the gas under this pressure in the refrigerating system, using brine at a temperature of
about 10 F. Sometimes two such pumps are used in series. Chlorine under this pressure and
temperature is liquefied and the dried liquid chlorine is safely handled in steel cylinders of 100to
150 pounds or of one ton net weight, or in tank cars containing 16 tons net. Liquid chlorine is now
handled in the trade in place of bleaching powder. From it the bleach liquor can be conveniently
made in the consumer’s own plant by bubbling chlorine gas through a soda ash solution of suitable
strength of form sodium hypochlorite bleach liquors. It is thus in a concentrated form ready for
use and may be prepared whenever it is needed.
Manufacture of Synthetic Hydrochloric Acid. At present large amounts of chlorine are also
used in the manufacture of concentrated hydrochloric acid. For this purpose, the hydrogen gas and
the chlorine gas from the cell room are brought in two secarate pipes in suitable proportions under
slight pressure into a vertical brick-lined cylinder or horizontal brick furnace, where they are
burned.
H2 (gas)+CL2 (gas) 2HCl (gas ) +44,000. Q=22,000 cal./mol HCL
Considerable heat is developed but it is usually not utilized. The hydrogen chloride gas must be
cooled in silica and stoneware pipes and absorbed in water to form concentrated hydrochloric acid
in Cellarius tourills and acid –proof towers.
With the production of chlorine, the variety of products in the alkali plants is further
multiplied. Alkali manufacturers are now manufacturing many chlorine derivatives such as
mono-chlorobenzene (from the reaction of chlorine gas on benzene), para- and
ortho-dichlorobenzene, potassium chlorate, sulfur monochloride (from the direct union of chlorine
gas and sulfur), etc. These products should be added to the list enumerated in Chapters XV and
XVI.
Evaporation of Caustic Liquor. The cell effluent is collected in weak liquor storage tanks,
from which it is pumped to evaporators for concentration. Zaremba, Buffalo Foundry, or Swenson
evaporators of the submerged vertical tube type with salt catchers attached or with continuous salt
separators, are among the ones most generally used. Two effects (and sometimes three) are
generally arranged. According to the practice of one plant, each evaporator is 80 inches in
diameter and 12 feet high, holding about 1000 gallons of the liquor. In the first effect, exhaust
steam, or live steam throttled to 25 lbs. Pressure, is employed, and the pressure above the liquor in
the first effect varies from 5 lbs. Pressure to 3 lbs. Vacuum. The liquor is sucked over to the
second effect by vacuum and so to the third (if any). The vacuum above the liquor in the last effect
is from 26 to 27 inches Hg. Condensed steam from both effects goes to the boiler. The condensate
from the first effect is forced by its own pressure to the boiler feed tank, but that from the second
(and the third) requires a pump. On the top of the second effect (or last effect) is located a cyclone
separator, to the outlet of which a barometric condenser is attached. This may be a countercurrent
type with steam from the last effect coming into the bottom of the condenser and the water
entering at the top, flowing over zig-zag baffle plates, the vacuum may be created by means of a
steam ejector (using 60 lb. Or higher steam pressure), or by means of a dry-and-wet pump. This
vacuum connection is located on the crown of the condenser. In the second (or last) effect the
caustic liquor is concentrated to about 47o Be. And the concentrated liquor is drawn off and
allowed to settle in tall tanks where most of the salt separates out. The analysis of the strong
caustic from the evaporator is given in Table 128.

It will be seen that considerable sodium chloride and sulfate will have separated out from the
strong caustic inside the evaporators. The strong caustic after complete settling is now a
commodity on the market known as liquid caustic.
Caustic made this way will contain considerable NaCl as the main impurity which may be
objectionable in certain industries, such as the rayon industry. One method of manufacturing
NaCl-free caustic is to bring the caustic liquor to 47o be. and settle off sodium chloride as much as
possible, as described above. The clear liquor then is diluted down to about 37o be. And cooled to
about 35℉. When crystals of NaOH. 3 1 H2 O or NaOH. 4H2 O separate out, leaving NaCl in the
2
mother liquor. These crystals almost free from NaCl are filtered off and melted and the solution is
concentrated back as usual.
As the 47o Be. Caustic concentrated from the cell liquor is saturated with respect to NaCl, it
may still contain 2.2g NaCl and 0.08g Na2 SO4 per 100g Na2 O, after complete settling. Even the
fused caustic obtained from this would contain 0.7g NaCl per 100g Na2 O. It is therefore desirable
to get rid of the excess NaCl in the electrolytic caustic. D.A. pritchard* proposed to treat the 47o
be. Electrolytic caustic liquor by adding to the liquor finely pulverized anhydrous sodium sulfate
in an amount 3-4 times the weight of NaCl present, heating and stirring so as to form a triple salt
of NaCl, Na2 SO4 and NaOH, which will separate on cooling and settling. It is claimed that in this
way the resulting caustic contains as low as 0.2g NaCl per 100g Na2 O.
If solid caustic is desired, this clarified 47o Be. Caustic is decanted and further concentrated
in cast-iron open pots by direct heat for 3 to 4 days. The molten caustic is then pumped by means
of a centrifugal pump made of pure nickel or ladled to steel drums, as explained in Chapter XIX.
*B.P. 299,995 (1928)
The salt from the catchers is washed until it contains only 0.1 per cent sodium hydroxide,
when it is dissolved with warm water and the solution used for treating raw brine in conjunction
with the soda ash treatment mentioned at the beginning of the chapter. The wash water is sent back
to the evaporators with the weak liquor. When these sodium chloride and sodium sulfate crystals
are dissolved and sent back to the raw brine-treating tanks it is easily seen that sodium sulfate will
accumulate in the brine, causing low efficiency in the cells and much oxygen gas generated at the
anodes. To get rid of sodium sulfate, barium chloride solution may be employed when necessary.
As the evaporators are frequently made of steel, any hypochlorite carried in the weak liquor
(or cell effluent) would attack the steel material very rapidly, causing corrosion in the steam chest
and leaks in tubes. It cannot be too strong emphasized that every effort must be made to keep the
cell effluent free from hypochlorite, or the upkeep on the evaporator will be heavy. If proper
attention is given to the operation of the cells, these tubes should last for more than two years.
Nowadays these evaparator tubes are frequently made of pure nickel, giving a product of the
purity required by the rayon manufacturers. Such nickel-tube evaporators will take higher steam
pressures and with forced circulation can concentrate the liquor to 74 per cent NaOH or higher.
Nickel-tube evaporators permit the use of higher steam pressure, with consequent higher
concentration of the caustic; for then there is no danger of caustic embrittlement, as would be the
case in the steel tube evaporators when a high steam pressure is used.
*B.P. 299,995 (1928)
One manufacturer in the United States uses caustic evaporators made entirely of nickel-body,
tubes and all. This nickel equipment, coupled with vigilant care in selecting high-grade raw
materials and in keeping all impurities, both metallic or of alkaline-earth origin, away from
contact with the weak liquor in the course of manufacture both prior and subsequent to this
evaporation process, insures a pure product that will meet any requirements demanded of the
liquid caustic. In this way, recrystallization of the caustic as above mentioned is no longer
necessary.
Manufacture of Sodium Chlorate. The manufacture of sodium chlorate is done by
electrolytic methods using cylindrical cells somewhat like Vorce cells, or rectangular type cells.
The cells use graphite electrodes and iron cathodes, as in the alkali-chlorine cells, but without
asbestos diaphragm. The brine used, too, is saturated and must be free from calcium, magnesium,
sulfates, and iron. Here similarity ends, for to the brine in the cells must be added hydrochloric
acid to maintain an acidity of pH 6.4-6.6 so as to bring about the oxidation of NaClO 3 by the
hypochlorous acid in the cell; also sodium dichromate must be added to protect the iron cathode
from the action of hypochlorous and hydrochloric acids. Calcium and magnesium must be absent
from the cell liquor; otherwise these would be deposited on the cathode, causing high voltage drop
at the cathode surface. Sulfate must be kept within a very low limit because SO4 = ions would carry
current to the anode, liberating oxygen, oxidizing the anode and causing losses of graphite, iron
must be absent in order that hypochlorous acid and sodium hypochlorite may be stable, cathodic
reduction reduced, and current efficiency improved.
A cylindrical cell* 33 inches in diameter and 4 feet high will take 1300 to 1500 amp. Per cell
at a voltage per cell from 2.8 volts to 3.6 volts, depending upon the current density, when using 2”
by 2” (or 2” round) by 36” long (overall) graphite anodes, although the theoretical voltage is only
1.5 volts. The current density is about 04-06 amp. Per sq. cm. and the current concentration (i.e.,
amp. Per liter of cell liquor) is 4.0-5.0. cooling of the cell liquor is necessary; this may be
accomplished by internal cooling using either cooling coils as a part of the cathode in the cell, or
outside cooling by circulating cell liquor through outside coils made of iron pipes cathodically
protected. The cell liquor temperature is approximately 25-30 ., depending upon the cooling
surface provided and cooling water temperature available. High cell liquor temperature would
cause excessive graphite losses. A cell operating at 42 ., for instance, will have anode losses
about double that at 32 ., and a cell operating at 32 . Again double that at 25 .
The reactions are as follows:
(1) 6NaCl+6H2 O+6 faradays 6NaOH+3H2 +3Cl2
(2) 2NaOH+Cl2 NaOCl+NaCl+H2 O
(3) 2Cl2 +2H2 O 2HOCl+2HCl
(4) NaOCl+2HOCl NaClO2 +2HCl
(5) 4NaOH+4HCl 4NaCl+4H2 O
Adding these five equations we obtain
(6) NaCl+3H2 O+6 faradays NaClO 3 +3H2
and from the thermochemical equation
NaCl + 3H2 O NaClO 3 +3H2 + Q

(-96,400) (-3×68,370) (-81,100)

we estimate roughly

220,400 × 4.187
E= = 1.59volts
6 × 96,540
*Cf. Groggins, Pitman, McLaren and Davis, “Electrochemical Production of Sodium Chlorate,” Chem. Met.
Eng., June, 1937; Pitman, McLaren, Davis and Groggins, “Sodium Chlorate Production,” ibid., December,
1938; also Groggins, Pitman, and Davis, “sodium Chlorate Cell Design,” ibid., July, 1940.
If we take account of the temperature-coefficient term, the theoretical decomposition voltage
is 1.50. But actually with the space of 3 /4 inch between the graphite anodes and the iron cathode
and with a current density of 0.3 amp. Per sq. In., the terminal voltages is 3.3 volts. Thus, the
overvoltage due to polarization, etc. Is very high. The current efficiency is about 83 per cent
because of many secondary reactions, while the power efficiency is below 40 per cent. The anode
consumption is a big item of expense. Denser graphite of sp. Gr. 1.7 or higher lasts longer.
The brine entering the cells contains about 350 g/l NaCl and 2 g/l Na2 CrO4 and the cell
effluent contains 550 g/l NaClO 3 and 60-80 g/l unconverted salt. It is possible to operate four to
eight of such cells in series (chemically), transferring the liquor by syphon or gravity from the first
cell in the series to the next. In this way, it is possible to get low unconverted NaCl and high
NaClO 3 concentration and operate the cells continuously, feeding the brine at the head cell of the
series and drawing the effluent from the last cell in the series. To avoid grounding of the current, a
device is used to feed brine into the cell in drops or in a discontinuous stream (see Fig.96).
Hydrochloric acid may be added to each cell in the series (to the immediate cells as well as the
head cells ) and the pH value of each cell carefully regulated.
A rod-cathode cell*construction is found to possess a much higher efficiency than the
ordinary plate-cathode cell above described. Such a rod-cathode assembly (Fig 97) with a single
2-inch diameter central graphite anode surrounded by ten inch diameter steel cathode rods, equally
spaced around a circle of inches diameter, at a distance of about inch from the center of the rods to
the surface of the anode, has a much higher current efficiency and a lower terminal voltage. This
type of cell has less resistance because of shorter distance between the cathode and anode surfaces
(inch) and hence a lower terminal voltage (2.8-3.2 volts), and has a more efficient stirring effect so
that a current efficiency of 93 per cent is obtainable with higher current density (0.5-0.6 amp per
sq cm) and current concentration. The energy required is about 2.4 kw -hr. per lb of sodium
chlorate made. A single unit of this rod-cathode assembly, 2 feet 6 inches long, will carry 100-150
amp of current so that a number of such units may be conveniently placed in a cell and connected
in parallel for a large capacity. The capacity of each cell is then directly proportional to the
number of such units it contains. Thus, a very compact arrangement is obtained for a al rge
capacity cell. A number of such cells are then connected in series for operation as usual.
*Cf. a very recent article by James McLaren, etc. entitled “The Efficiency of a Sodium Chlorate Cell with Rod
Cathodes,” Trans. Elec. Chem. Soc., 79, for April meeting, 1941.

The rod-cathode cell tank may be made of earthenware, concrete, or even steel plate. If made
of concrete, it should be coated inside with a coal-tar protective paint or with Chinese wood oil or
linseed oil to preserve the concrete surface from crumbling. The upper part of each graphite
through the porous graphite. The top of each anode, where electrical contact is made, is best
copper-plated to secure good electrical contact. Care should be taken not to copper-plate lower
portion of the anode that may come into contact with the cell liquor; otherwise the anode would
deteriorate rapidly due to local cell action. The cell cover may be made of 1-inch asbestos
“Ebony” or “Chemstone” board of Johns-Manville make. Joints around the edge of the cover may
be sealed with water glass or rubber cement, or with any other inert plastics. One such
composition is prepared by melting together

Rosin 75 pts by wt.

 Gum rubber 15 “ “ “
Beeswax or other wax 10 “ “ “

and adding to the melt a small quantity of fine asbestos fiber. The proportion of rosin to wax may
be varied according to the consistency desired.
Also, it is found that in the rod-cathode cell if the sulfate ions (SO4 --) are kept out of the cell
liquor, the cell will operate with almost no chromate. Then the anode lift is very much
lengthened, while the cathode reduction is not increased. It seems that while sodium chromate
protects the cathode surface, it has a tendency decidedly to accelerate oxidation of the graphite
anodes.
Electrometric measurement of the pH values must be employed if accurate results are desired,
because the pH Comparator is rendered useless on account of the presence of strongly oxidizing
hydrochlorous acid, which would destroy the indicator in the solution, and because the solution is
highly colored. The color range of CrO4 = Cr2 O7 =, however, may be used as a rough guide in the
field for comparison against a series may be used as a rough guide in the field for comparison
against a series of known standards. The cell effluent is collected in a glass-lined settling tank
where the graphite slime is allowed to set off and time is allowed to complete the reaction between
sodium hypochlorite and hypochlorousacid.
NaClO + 2HClO NaClO 3 + 2HClO
2HClO + NaClO NaClO 3 + 2HCl
The following reaction
HClO + 2Cl2 +2H2 O HClO3 + 4H+ + 4Cl-
Increases the acidity and causes the lowering of PH in the liquor as the reaction proceeds.
This conversion of HClO to HClO 3 is arrested if the PH of the solution falls below 2 (for then
most chlorine would exist as Cl2 ); and again, if the PH increases beyond 9 (for then chlorine
would exist mostly as ClO - ion). Between the PH values of 2 and 9, this conversion proceeds
favorably, the optimum being at PH = 5, for then the concentration of HClO is at a maximum.
Table 129 shows the ratios of HClO to ClO - and Cl2 to HClO at different PH values at 25°C.
When the reaction is complete, the liquor is neutralized with caustic soda or soda ash and
passed through a sand filter to remove the residual.

*Rue. John D., “The Chemistry of Bleaching Chemical Wood Pulps,” Trans. Electro chem.
Soc., Paper Presented to Savannah Meeting, May (1938). Figures taken from “Die Bleiche des
Zellstoffs,” I,p.77 (1935).
graphite slime. Neutralization is performed to destroy the acidity, so that the liquor may be
concentrated in evaporators for the recovery of NaClO 3 crystals. This separation of NaClO 3
crystals from the liquor may be accomplished by (a) evaporating the liquor in vacuum evaporators
and cooling the concentrated liquor so as to throw out NaClO 3 , or (b) by salting out the NaClO 3
crystals in the liquor with solid salt in finely pulverized form to obtain saturation with respect to
NaCl and cooling the liquor to a low temperature, say 15 ., to crystallize out NaClO 3 . NaClO 3
crystals which appear yellowish because of the presence of a small amount of sodium dichromate,
are dried in a centrifuge and washed with a small amount of water on the centrifuge to obtain
white color. The mother liquor is reconcentrated and eventually returned to the cells for the
preparation of brine for electrolysis. The sodium chlorate crystals are dried in a hot-air shelf dryer
at 140-160 ℉., heated by low-pressure steam (10-15 lbs. Gauge) coils, and then packed in paper
lined, hermetically-sealed, tinned containers holding 100 lbs. each.
In the operation of the chlorate cells, it is very important to regulate the PH values in the cells
and also in the settling tanks where the oxidation of sodium hypochlorite to chlorate is allowed to
be completed, while the graphite slimes are being separated. After completion of the reaction, the
liquor is heated by steam jacket and agitated by compressed air to destroy excess hypochlorous
acid. Decomposition of the hypochlorous acid is aided by heat and aeration. Neutralization is then
effected by the use of soda ash or caustic soda, so that this neutralized liquor devoid of
hypochlorous acid may be concentrated in cast-iron of alloyed cast-iron (austenitic cast iron
containing small amounts of nickel or chromium) evaporators at an elevated temperature without
excessive corrosion,. It is, however, more advisable to use stainless steel or copper evaporators for
this purpose.
The liquor foams badly in the evaporators, but this trouble may be minimized by making the
liquor distinctly acid before heating and agitating to destroy the hypochlorite in the settling tank
mentioned above.
The mother liquor from the crystallizer and the centrifuge may be returned to the evaporators
for further concentration and a portion of it may be sent to the makeup tank for feeding into the
cells, thus returning the sodium chromate to the cell liquor. Care must be taken to avoid
accumulation of metallic impurities-iron oxide particularly in the liquor. The effect of iron rust in
the cells is to cause excessive losses by reduction, thereby lowering the current efficiency.
Consequently, the concentration of iron in the cell liquor must be carefully watched, and the iron
oxide formed in the cells must be allowed to settle quietly to the bottom of the cell and not to be
circulated with the liquor stream.
Sodium chlorate is used for killing the weeds in the farms and may be mixed with TNT or
nitroglycerine for the manufacture of potassium the cells, as might be inferred from sodium
chlorate manufacture, because potassium chlorate is difficultly soluble. Instead, sodium chloride is
used to make sodium chlorate liquor as described above. The cell effluent is then taken and treated
with finely pulverized potassium chloride, thereby causing potassium chlorate to be formed by
metathesis:
NaClO 3 + KCl KClO3 + NaCl
Potassium chlorate, being less soluble, separates from the liquor. The mother liquor is then
returned to the cell for the preparation of the brine for electrolysis. Thus, the process does not call
for concentrating the cell liquor by evaporation.
Potassium chlorate is used in the manufacture of matches and in the munitions industry.
 Chapter XXI
Wet Calcination of Sodium Bicarbonate
Decomposition of crude sodium bicarbonate, or ammonia soda, into soda ash is normally
done in a dryer with external firing. “Wet calcination” is a term applied to the operation of
decomposing sodium bicarbonate into soda ash in the wet way in contradistinction to dry
calcination. Wet calcination is advantageous when soda ash is to be used in the form of a solution
for the subsequent manufacturing processes, i.e., when it as the source of heat instead of direct
firing, wet calcination has the advantage over dry calcination by means of a dryer, because it has a
better over-all heat efficiency and because the operation is simpler and does not entail such
mechanical difficulties as are ordinarily encountered in an externally fired furnace, especially
when the crude bicarbonate obtained is wet and the crystals are poor and in the form of a fine
sludge. Further, it enables caustic soda to be manufactured directly from sodium bicarbonate
instead of from soda ash, into which sodium bicarbonate would have be first dried and calcined,
and from which a solution then would have to be made. And it is even more feasible to recover
CO2 gas from decomposition by wet calcination in a very rich form – in fact approaching 100 per
cent CO2 .
Decomposition of sodium bicarbonate by either dry or wet calcination is endothermic, and
the heat in the wet calcination operation comes from steam. In wet calcination it takes heat to
dissolve sodium bicarbonate to heat up the solution and to drive off CO2 from the solution.
Theoretically, the quantity of heat required per kg. Mol of NaHCO3 may be estimated from the
following thermochemical equation:
2NaHCO3 (diss.) = Na2 CO3 (diss.) + H2 O(liq.) + CO2 (sat.)+ Q
(-2 225,000) = (-278,200) + (-68,370) + (-103,680) + Q
Q equals –250 kg. Cal. For 2 kg. mols NaHCO3 , a very small figure of practically nil; i.e. ,the
above reaction, starting from sodium bicarbonate in soluti8on and ending in soda ash and CO2 in
solution, has practically zero heat effect. Therefore, not only does wet calcination provide better
heat transfer, but the quantity of heat required is actually smaller, because Na2 CO3 is obtained in
solution and H2 O in the liquid form. Heat, however, is required to dissolve solid sodium
bicarbonate (heat of solution), to heat up the solution, and to drive CO2 out of solution Moreover,
the steam used serves a twofold purpose: (1) to furnish the heat necessary for decomposition, and
(2) to act as a distilling medium lowering the partial pressure of CO2 above the solution. In this
way, the partial pressure of CO2 at the bottom of the decomposer is reduced to a very low figure,
less than 0.01 atm. Generally speaking, the bicarbonate slurry contains about 4.5 normals total
sodium, and the slurry goes into solution completely toward the end of decomposition. From the
equilibrium relationship derived by Harte, Baker and Purcell (see p. 278), taking the boiling
temperature of only 100 .and the partial pressure of CO2 at 0.01 atm.
 That is, the bicarbonate under such conditions could be decomposed to 91.5 per cent soda ash.
But in practice, because of the length of time it takes to reach the equilibrium and the excessive
steam required, decomposition generally stops at around 85-87 per cent at the exit from the
decomposer, and sometimes less.
A decomposer is very similar to a distiller in its operation. It is a CO2 stripping column with
rectification for CO2 gas, all arranged counter-currently. While a distiller distills off ammonia, a
decomposer distills off CO2 gas. The stream used in the decomposer is generally low-pressure
steam of 6-10 lb. Per sq. In. Gauge, but may be 60-lb. Pressure steam. One type of decomposer is
shown in Fig.98. the bicarbonate slurry containing about 4.5 normals of sodium is fed to the
bubble-cap portion on the top, trickling through packed sections to the bottom, where steam is
introduced countercurrently as in a distiller. CO2 gas from the decomposition, saturated with steam,
leaves the top of the decomposer, passes through a condenser to condense out as much steam as
possible, and is finally cooled on its way to the compressor intake main. This cooled gas will
contain 95 per cent CO2 by volume or more, if the system is reasonably tight. The slurry is made
up of various returned liquors, weak wash liquor or condensate, to which crude sodium
bicarbonate or any rejected soda is added, and the suspension is kept at about 50 .The crude
sodium bicarbonate contains, besides NaHCO3, small quantities of NH4 HCO3 , NaCl, Na2 CO3 and
NH4 Cl. The reactions in the decomposer are:
 (1) NH4 HCO3 NH3 + CO2 + H2 O
(2) 2NaHCO3 Na2 CO3 + CO2 +H2 O
(3) NH4 Cl+NaHCO3 NaCl +NH3 + CO2 +H2 O
NH3 and CO2 are distilled off, but all the NaCl remains in the decomposed liquor. Indeed, these4
reactions are very similar to those occurring in the ammonia distiller, except that no lime is
introduced. But whereas in the distiller operation, exhaust steam of 6-10 lb gauge is generally used,
in the decomposers the steam pressure may be as high as 60-lb gauge, to give more rapid and
complete decomposition, especially if the decomposed liquor is to be causticized for caustic soda
manufacture.
From the equation referred to above

Where õ represents the degree of bicarbonation (fraction as NaHCO3 ), it will be seen that a
higher steam pressure (consequently a higher steam temperature) has several advantages, because
(a) solubility of CO2 gas in the liquor is decreased with the increase in temperature;
(b) increase in temperature gives a smaller value of õ;
(c) the rate of decomposition is increased as the temperature is increased, it being doubled
for every 22-24 temperature rise;*
*Whitman, W. G., and Davis, G. H. B., “A Comparison of Gas Absorption and Rectification,” Ind. Eng. Chem.,
18,264 (1926)
Harte, C. R., and Baker, E. M., “Absorption of Carbon Dioxide in Aqueous Sodium
Carbonate-Bicarbonate Solutions,” ibid., 25,1128 (1933).
(d) increase in steam pressure and temperature causes less condensation and gives greater
steam-distillation effect in the decomposer.
The CO2 gas from the decomposer, after passing through a condenser and cooler, is usually
sent to the rich CO2 main at the intake to CO2 compressors to be delivered to the “making”
columns for ammonia soda manufacture; but if 60-lb steam is used, it is also possible to conduct
the rich gas from the decomposer directly to the bottom of the “making” columns by means of
constant-volume distribution control without going through the CO2 compressors.
Wet calcinations may be used to advantage in the following processes where crude
bicarbonate is used directly----all in the form of a solution:
(a) the manufacture of sodium sesquicarbonate;
(b) the manufacture of refined sodium bicarbonate;
(c) the manufacture of caustic soda.
Practically the same design of decomposer may be used for all three purposes, but the control
may be different as regards the slurry concentration, the degree of decomposition, and the
optimum steam pressure for the decomposer, all of which may vary in accordance with the
purpose for which the decomposed liquor is to be used. Frequently, certain alkali rejects are used
together with the bicarbonate to make the slurry for their alkali value. These may be soda dust
from dust collectors, fines from dense ash screening operation, off-color ash from the starting-up
of a new apparatus, floor sweepings, etc which are first dissolved and then settled and decanted
before they are added to the bicarbonate.
Unlike the distiller operation, the decomposer operation is rather simple, there being less
tendency for the formation of scales or the deposition of insoluble sludge. Hence, as pointed out
above, a higher steam pressure is more commonly used in the decomposer than in the ammonia
distillers. In general, also, plain-steel plates are frequently used for the construction of the
decomposer instead of the cast-iron rings, which are more generally employed for the distiller,
especially for the upper sections above the lime inlet.
The Manufacture of Sesquicarbonate. In the manufacture of the sesquicarbonate, the slurry
used consists of about 520 parts of crude sodium bicarbonate to 1000 parts by weight of return
water. The decomposition is carried up to about 78 per cent, i.e.; when 78 per cent of the
bicarbonate has been converted to soda ash. The blow-off is then taken to a crystallizing tank and
slowly cooled to allow needle-shaped (monoclinic), lustrous crystals of sodium sesquicarbonate,
Na2 CO3 NaHCO3 2H2 O to form . Attention is given to controlling the operation so that al
ammonia is driven out of the solution. At the end of the decomposition all bicarbonate should
have gone into solution. The decomposed liquor is so concentrated that it may freeze in the pipes
unless they are lagged. The following represents a typical condition of operation:

The Manufacture of Refined Sodium Bicarbonate. The crude bicarbonate (ammonia soda)
is suspended in the return liquor so that the slurry contains 480 kg of the bicarbonate per 1000 kg
of water. The dec omposition is carried to about 76 per cent using a moderate-pressure steam. At
the end of decomposition, all the bicarbonate should have gone into solution at the temperature of
the decomposer, and ammonia should have been completely distilled off. The liquor, however, is
so concentrated that all pipes and storage tanks should be carefully lagged or traced with steam
coils to prevent freezing. Decomposition is carried to the above extent in order to decompose all
ammonium bicarbonate and drive out ammonia; but further decomposition is unwise, because the
liquor has to be re-carbonated in the bicarbonate towers for the precipitation of sodium
bicarbonate. For this purpose, lean gas (kiln gas) is used, which is pumped into the bicarbonate
towers by means of the CO2 compressors in the same way as in ammonia soda manufacture. The
decomposed liquor is rapidly carbonated and cooled in the bicarbonate tower and the slurry is
drawn out at the bottom and centrifuged. For details of the operation and the drying of the refined
sodium bicarbonate, the readers are referred to Chapter XVIII. The following represent a typical
condition of operation:
 The Manufacture of Caustic Soda. In the manufacture of caustic soda, the decomposition is
carried as far as practicable (usually 85-88 per cent) to minimize the waste of lime for
causticization.
2NaHCO3 +Ca(OH) 2 CaCO3 +Na2 CO3 +2H2 O
NaCO3 +Ca(OH) 2 CaCO3 +2NaOH
Also, the slurry is made up in such a concentration that the decomposed liquor is of just the right
strength for causticization with milk of lime containing about 250g of CaO per l. In present
practice, such a liquor contains an equivalent of 18 per cent Na2 CO3 as it is delivered to the
causticizing tanks. For such purpose, the slurry is made up with the “return liquors” containing
about 400kg. Ammonia soda per 1000kg. Of return liquors. As the salt will be accumulated in the
decomposed liquor by using the return liquors over and over again, a portion is constantly beld out
and sent to the ammonia distiller. The decomposed liquor is kept in the hot condition in a storage
properly insulated against heat loss, and sent to the causticizers or reaction agitators, into which a
calculated quantity of milk of lime is continuously added for causticization. Details of the
operation are given in Chapter XIX to which the reader is referred. This operation avoids
calcination of the sodium bicarbonate to soda ash, and then dissolves the ash to make the liquor
for causticization. Also, the heat in the liquor from the decomposer is directly utilized in
causticization. As far as the CO2 recovery is concerned, it can be more readily obtained in a rich
form by wet calcination than by dry calcination, such as in the rotary dryers, where rotating parts
are difficult to make perfectly air-tight. The following represent a typical condition of operation:

As we are concerned here with a high degree of decomposition, we must make our best effort
to reduce the bicarbonate in the liquor as far as possible. It will be noticed that the slurry used here
contains less bicarbonate. One reason why we do not use a more concentrated slurry here is that a
much higher concentration of the slurry would lead to a region where sesquicarbonate would be
formed, from which CO2 is liberated only with difficulty. (See Fig. 99.) then again in caustic soda
manufacture, we do not need a more concentrated decomposed liquor for causticization.
For each of the foregoing three processes, to which wet calcination is applied, a diagram may
be drawn to represent each condition. In the diagram (Fig.99) the ordinate is the number of mols
of H2 O per mol of Na, while the abscissa is the degree of decomposition. The degree of
decomposition in sodium bicarbonate is represented by zero, while that in soda ash is represented
by 100 per cent. Axis-OY represents the composition of the slurry in mols of H2O per mol of Na,
 FIG 99 Process diagram for wet calcinations.

while Axis-OX represent progress of decomposition. Lines ab, cd, ef represent wet calcination for
the manufacture of sodium sesquicarbonate, refined sodium bicarbonate and caustic soda,
respectively. The end of the line b, d, or f represents percent decomposition at the outlet from the
decomposer. The inclination of these lines upward indicates dilution of the liquor during
decomposition , due to steam condensate in the decomposer and water formed from the
decomposition. The curves MN, NP and PQ are solubility curves for pure sodium bicarbonate,
sodium sequicarbonate, and sodium carbonate monohydrate around 60 . Espectively.
The heat required for wet calcination in caustic soda manufacture may be esstimated as
follows:
Given:
Slurry contains per 1000 kg. Water, 400kg. NaHCo3 on dry basis
Sp. gr. of slurry 1.21
Sp. Heat 0.84
Temperature of slurry feed 40 .
Exhaust steam 6 lbs. Gauge
Pressure at top of decomposer 0.6 lb/ ” gauge
 Steam pressure at 85 . 8.4 lb/ ” abs.
Bicarbonate required on dry basis
Per 1000kg. Soda ash 1580kg
Slurry required per 1000kg. Of
Soda ash 1400×1580/400 5500kg.
Heat Required for Wet Calcination:
1. Heat of solution of 1580 kg. NaHCO3 at 4300 kg. Cal per mol
1580×4000/84 80,900 kg. cal
2. Heat required to heat bicarbonate solution from 40 to 100 .
5500×0.84×(100-40) 277,000 kg.cal
3. Heat required to drive out CO2 gas from solution
1580/84×1/2×86%×5880 47,600 kg.cal
4. Heat carried of in steam at 85 . From decomposer top
1580/84×1/2×86%×8.4/(15.3-8.4)×18×548 97,100 kg. Cal
Net total heat required 502,600 kg. Cal
Heat loss through radiation, etc. (say 20%) 100,500 kg . cal
Total 803,100 kg . cal
Heat to be supplied from low-pressure steam:
Latent heat available in exhaust steam at 6 lb. gauge=532 kg. Cal
Wt of exhaust steam at 6 lbs. gauge
=603,100/532=1130 kg. Steam
This is a rough estimate. In plant practice, the figure is nearly 1.2 metric ton of steam per
metric ton of soda ash. This considerably less than the quantity of exhaust steam required for
ammonia distillation, namely about 2 metric tons of exhaust steam per metric ton of soda ash
output.
 Chapter XXII
Generation of Power for Ammonia Soda Plants
The requirements in the generation of power for ammonia soda manufacture are not in any
way fundamentally different from those in any other chemical industry, except that as the exhaust
steam is used for ammonia distillation, the plant generally runs non-condensing or uses bleeder
turbines. As the buildings in ammonia soda plants are, as a rule, tall and the machinery is
scattered, individual drive by electric motors is undoubtedly the best arrangement. Among the
largest units that require steam are the electric alternators, or turbo-generators, which are used for
power as well as for lighting, the power required being about 10 per cent of the total electric power
generated, although this figure varies with the layout of the plant. The next largest units requiring
steam are the CO2 compressors. Then come the cooling water pumps, the air compressors, the
filter vacuum pumps or "exhausters," the absorber "exhausters," etc., all of which may be
steam-driven to produce exhaust steam enough for ammonia distillation and to secure flexibility in
speed regulation to meet the different demands of changing conditions.
For low-speed drive, the horizontal, duplex, steam-driven type is best suited and should be
recommended for the construction of all these machines. In certain cases, however, the
single-cylinder, horizontal type may be used to advantage, as in the case of the CO2 compressors,
since each compressor can be better regulated to suit the individual column-operating conditions.
The vertical type is not recommended as it does not possess high stability, nor can it withstand
heavy duty. Corliss valves ni different form are commonly used. Standard Corliss valves are
recommended for low speeds, non-releasing Corliss valves for higher speeds, and half Corliss for
the inlet valves of CO2 and air cylinders. Recently "Unaflow" steam engines have been extensively
used, and have given better economy than countercurrent type engines. Exhaust steam is piped in
16-to 24-inch mains to the distillers. The exhaust from the feed-water pumps for the boilers and
other auxiliaries should also be introduced into the exhaust main. As there is normally a scarcity of
exhaust steam for ammonia distillation, every bit should be conserved and not allowed to escape
into the atmosphere. If, however, the steam valves and piston rings in the engine cylinders leak
because they are worn out by use, there would be an excessive quantity of exhaust steam. Under
these conditions there would be a tremendous demand for steam at the boilers and an excessive
amount of coal would be consumed. Leakages of this kind are very serious in Corliss valve
cylinders, because steam would leak around the admission valves and pass out directly to the
exhaust main around the exhaust valves at the same end, regardless of the events in the piston
operation. The back pressure then becomes excessive and the engines consume abnormal
quantities of steam in order to develop the required horsepower. A bak-pressure valve should.
therefore, always be provided in the exhaust steam system for atmospheric relief in case the
pressure should go up to an undesirable figure. Further, as mentioned in Chapter XIV, the volume
of the filter liquor for ammonia distillation should be kept as small as possible so as to require the
mini-mum amount of steam for distillation.
The equipment used outside of the boiler and engine rooms, such as the centrifugal type of
pumps used for brine and strong liquors; the plunger type of pumps used for mud and milk of lime;
stirrers for milk of lime, for prelimers, and for settling vats; conveyors and elevators of all
kinds--in fact all machinery outside of the power plant--should be individually electrically driven.
They are operated at more or less constant speeds, so that the cheap and rugged squirrel-cage type
of motor is best adapted for the work. This type of motor, of course, calls for A. C. power which is
preferable to D.C. The best voltages considering all phases would be 440 volts for power and 110
volts for lighting, unless the motor units are very large, in which case higher voltage, say 2200
volts, may be used. For all classes of work, except in driving direct-connected centrifugal pumps,
higher synchronous speeds than 1200 r.p.m. are undesirable because of the smaller starting torue in
these high-speed motors and of the expensive speed reduction mechanism necessary .Synchronous
speeds of 720, 900 and 1200 r.p.m. are most generally used. All motors from 50 hp. up should be
started with compensators, although this is largely a matter of electric power plant capacity. Small
motors may be started with across-the-line, full-voltage, magnetic starters with no-voltage release
and overload protection, but motors smaller than 2 or 3 hp. are seldom desirable in plant operation.
In America the preferred electrical characteristics are three-phase,60-cycle,440 volts,550
volts, or any other multiple of 110 volts, while in Europe three-phase,50-cycle,and 380/220 volt
star-delta connections are common. The generator and distribution voltage depends on the distance
of the power plant from centers of load. In American alkali practice, 2300 volts is reasonably well
standardized. In oil refineries, for example, where loads are less concentrated, higher voltages such
as 6600 volts or higher, are prevalent. The electric power distribution system of an alkali plant has
a peculiarity in that interruptions to service must be so nearly non-existent that the service shall be
absolutely continuous. it is therefore important that service should be given to each building
through at least two different feeders. A common system is to have a feeder for each building or
load center, and in addition ,a common circuit or spare feeder which can feed any one of several
load groups. Due to the corrosive nature of the atmosphere around any chemical plant, electrical
distribution equipment is subjected to rather severe service. It is therefore customary to protect
feeders, sub-station switch gear, switchboard apparatus, motor starting equipment, etc. with dust
and fume tight enclosures. Standard, open squirrel-cage motors are commonly adopted, although
drip-proof motors in certain wet places are desirable; but no explosion-proof type of electrical
installation is required. Slip-ring motors are seldom used except in special cases where a high
starting duty or regulation of speeds is desirable. Synchronous motors in units above 250 hp. may
be installed to advantage for power factor correction, especially in cases where a low speed drive
is required. Electric cables are best kept in iron conduits with vapor-tight fittings. The most
satisfactory main distribution is to use lead-covered or waterproof, rubber-sheathed cables with
power ducts underground which make the plant substantially immune to lightning.
The coal consumption in a soda plant varies greatly with the size of the plant (the output),the
efficiency of operation, the equipment in the engine and boiler rooms, and supervision of the
equipment. It varies from 0.2 ton of coal per ton of soda ash made in large, well-operated plants to
as much as one ton of coal per ton of soda ash made in very small plants.
The question arises as to whether the exhaust steam will be just sufficient to meet the
distillation demand, whether live steam will be just sufficient to meet the distillation demand,
whether live steam will have to be employed in part to meet the distillation demand, or whether
too much exhaust steam will be left on hand so that a good portion of it has to be allowed to
escape into the atmosphere. To answer this question, many phases have to be considered which
concern the economy of the whole plant operation. Any one of the three cases may exist.
Normally there is a tendency toward a deficiency of exhaust steam. It is largely a matter of
heat conservation (by means of preheating arrangements for heat regeneration) and the balancing
of operations between the engine room and the distillation plant. If the CO2 gases are dilute,
considerable power is consumed by the CO2 compressors per ton of soda output, and consequently
there is a corresponding increase in the exhaust steam produced from the compressors. If the
engine valves and piston rings leak because of poor upkeep of the engines, there is a
corresponding increase in the exhaust steam produced. Thus, there may be a large surplus of
exhaust steam on hand under these conditions.
If, on the other hand, there is an excessive dilution in the filter liquor and the milk of lime
employed is weak, a large amount of exhaust steam is needed for distillation. If the milk of lime is
cold and the filter liquor has not been properly preheated before entering the heater, steam
consumption in the distillers is greater. Thus there will be a deficiency of exhaust steam to meet
the distiller demands under these conditions.
A large volume of filter liquor and other weak liquors to be distilled causes a high static
pressure at the bottom of the still and consequently throws a high back pressure on the exhaust
main, causing excessive back pressure and high steam consumption in the various engines and
bleeder turbines. This creates a greater demand for high-pressure steam from the boilers.
For good economy of plant operation, therefore:
(1) The lime kilns must produce high-test CO2 gas, and the furnace (dryer) gas must be
kept concentrated by guarding against leakage of air into the dryers and the furnace gas system.
(2) The column (tower) decomposition must be kept high so that a smaller volume of mother
of mother liquor is produced per ton of soda ash made.
(3) Good bicarbonate crystals must be maintained in the columns so that a smaller amount
of wash water is required on the filters to keep the salt content in the ash below a maximum,
thereby preventing unnecessary dilution of the filter liquor, and so that the bicarbonate crystals
will filter dry, thereby cutting down coal consumption in the soda dryers. The wash water must be
accurately maintained at the optimum temperature and properly distributed across the filter surface
to effect efficient washing.
(4) The filter liquor must be kept concentrated and no more waste or weak liquors should be
sent to it than necessary.
(5) The filter liquor must be preheated distiller condensers to a high temperature before
entering the heater.
(6) The milk of lime must be used hot, and hot water from the exit of the condensers or
coolers should be used for making the milk of lime in the slaker
(7) The milk of lime must have a high concentration of lime so that as little water is
introduced into the milk of lime as possible (ef. attempt to use dry lime in prelimers.)
(8) All valves and piston rings in the steam cylinders must be kept tight and in good
condition, and the setting of these valves must be frequently checked by indicator cards or
otherwise. All power equipment, such as the compressor end of CO2 compressors, steam engines,
or turbines driving them, and all mechanical valves, must be kept as efficient as possible (low
efficiency revealing itself in hot bearings, high oil consumption or rapid wearing out of these
moving parts).
(9) All steam piping, distillers, prelimers, and other hot vessels should be adequately lagged
with sufficient thickness of heat-insulating materials, and all the condensate contained in the
exhaust steam should be trapped off at the distiller end as well, as in the exhaust steam main so
that only dry, saturated steam enters the distiller.
(10) Plants must be provided with an adequate number of meters and indicators which will
also record the periods and amounts of exhaust steam passing through the reducing valves or
blowing to the atmosphere. Frequently, under, maladroit control, a plant operating neat the balance
point has steam going through the reducing valve and at the same time blowing to the atmosphere.
(11) The equipment units, or trains of connected apparatus, should not be too large for the
plant. They should be small enough to permit certain units to be taken entirely out of operation
during a slowdown. This, of course, must be done consistently with the requirements for low
direct operating labor costs.
(12) Apparatus on the dry side of the plant should be arranged for minimum rehandling of
products.
(13) At the boiler end, the boiler feed water made up of the exit hot water from the different
condensers and coolers must be preheated to a high temperature in economizers heated by flue
gases, and further in heaters heated by exhaust steam, before entering the boilers. Economizers
under certain conditions may be installed at the back of the soda ash dryers to recover the heat
from the high-temperature flue gases there. Air preheaters and extraction steam feed-water
preheaters may be considered if the plant fuel costs warrant them .
All these factors tend to reduce coal consumption for the whole plant, so that, give a good
boiler outfit, with the usual necessary accessories for a modern steam-generating plant, a good coal
figure per ton of soda ash made will be shown.
Boiler feed water can (more or less successively) receive waste heat from:
(1) Distiller condensers
(2) Furnace gas cooler
(3) Steam flashed from distiller blow-off
(4) Rotary dryer-economizer
(5) Extraction steam
(6) Boiler plant "steaming" economizer
Steam flashed from distiller blow-off is better used in "weak liquor" distillation. This is the
distillation of ammonia from weak ammonia-bearing liquors, from the condensate resulting from
the cooling of the furnace gases and from the calcium and magnesium sludges settling out of the
ammoniated brines. Soda dryer economizers are expensive and, where the feed water is still quite
cold, often have high maintenance cost because of corrosive condensate on the flue-gas side
during a slow-down. Extraction steam preheating is justified only with fairly expensive fuel. The
boiler plant economizer is today built as a "steaming" economizer and as such is most generally
justifiable.
A modern steam generating plant minimizes the waste of heat, and recovers the sensible heat
from the flue gases as far as possible. The furnace is usually provided with an induced-draft fan
and a forced-draft blower, yielding a balanced draft which causes the least possible leakage of air
into the furnaces. Air for combustion is preheated by means of the flue gases in the air-preheater,
and the feed water on its way to the boiler is first preheated in the economizers heated by the flue
gases, and then in the feed-water heater and deaerator heated by the exhaust or extraction steam.
Often, the intake air is through ducts in the furnace walls to cool the fire-bricks an also to preheat
the air. Frequently water-cooled walls are provided, especially when pulverized coal is fired, and a
number of water tubes leading from the steam drums above are located around the combustion
chamber to absorb the radiant heat from the rolling flame as well as to protect the fire-bricks from
exposure to this intense heat .In this way ,a high rating of the boiler is secured .People nowadays
seldom figure on the normal rating of one B.H.P. per ten square feet of heating surface. From
three to five hundred per cent of this normal rating is often not difficult to obtain.
As mentioned above, the generator turbines should preferably be of the bleeder type with
condensing outfits, but having automatic bleeding points to bleed out extraction steam for
pressures varying from 8 to 10 lbs. pe sq. in. depending upon the requirements for ammonia
distillation. To take care of any unbalanced condition between the amount of exhaust steam
required and the amount of exhaust steam at hand, steam may be bled out from this unit in any
desired quantities to make up the deficit, while the remainder goes to the condenser. Thus, just
sufficient exhaust steam to meet the distiller requirements may be maintained without wastage.
Another good feature is that mixed steam pressures can be used. The present tendency is
toward high working pressures. If old, low-pressure boilers and engines are still in use in the plant
and it is considered that they are still too good to be scrapped, then mixed steam pressures must be
employed. New, high-pressure unite will be working together with old, low-pressure units. The
high-pressure steam can be employed to advantage in the bleeder type turbo-generators running
condensing, and the extraction steam from the bleeder at a lower pressure can be utilized to drive
the low-pressure engines or low-pressure turbines, so that the amount of exhaust steam from the
low-pressure cylinders can be adjusted to meet the demand of the distiller. Furthermore, if there is ,
for some reason, a shortage in the high-pressure steam, the low-pressure steam can actually be
introduced backwards into the turbo-generator through the bleeder connection to meet the
generator load. If the old equipment both at the boiler and engine ends operates at 200 or 180
pounds steam pressure and it is desired to install large units working at 900 or 600 pounds steam
pressure, a high-pressure bleeder type turbo-generator taking steam at 900 or 600 pounds pressure
can be installed to advantage, and the extraction steam at 200 or 180 pounds pressure taken out to
run the low-pressure engines or turbines.
For the same reasons at the electric end, when many induction motors of comparatively low
speeds and small capacities have been installed throughout the plant for a number of years, the
power factor of the plant is likely to be low and the power losses high. To correct the power factor
in such a plant, several synchronous motors in large units may be purchased when certain large
units of constant-speed machinery are to be installed, or else a capacitator of sufficient size may be
installed on the main feeders.
A still more flexible arrangement si to make an agreement with the city or local power
company, whereby electric current from the city line can be drawn to meet a part of the demand in
the plant whenever a surplus of exhaust steam is being accumulated, i.e .,during periods of low
production. The current will be returned to the city line when the plant is in full operation
producing all the exhaust steam required, i.e., during the time when large quantities of exhaust
steam ate required for distillation because of the increased output.
During recent decades, the rapid improvement in high-pressure and high-temperature steam
and power -generating equipment has permitted the ammonia soda industry to produce a large
number of surplus kilowatthours without increased (and at times with decreased) coal consumption.
Practically every American plant has taken advantage of this development in connection with the
installation of adjacent electrolytic caustic operations. Only three of the ten plants operating in the
United States are, as of 1939,without electrolytic caustic operations. As late as 1926 a steam
pressure of 150-200 pounds gauge and a temperature of 500℉ .was among the highest in the
industry, and was representative of fairly uniform practice in America.Steam conditions in certain
alkali plants in 1939 have reached as high as 2500 lbs.per sq.in.and 950℉ .but 450-,600-and
900-pound pressure plants are in general operation.
The power requirements of an ammonia soda plant revolve around the two major
consumptions which are,(1) low-pressure steam for ammonia distillation, and (2) power for
conpressing gases and elevating the brines. The first figure is generally of the order of 4000
pounds of steam per ton of soda, and the second is of the order of 75kilowatt hours per ton of soda.
As stated throughout this volume, figures of this character vary widely from plant to plant.
Variations in the above correspond to the variations in coal consumption given in Table 148.In
some cases high coal consumption may be due to the profit resulting from burning a local, but
inferior grade of coal, or from some other condition not indicative of low efficiency. Other
requirements for steam and power are space heating in the winter time, the various pumps, the
operation of all agitators, conveyors, elevators, rotary calciners, filters, power plant auxiliaries, etc.,
as stated above.
All mechanical power requirements in an ammonia soda plant can be

TABLE 133.Types of Drive for Machinery in Alkali Plant.

Load Driver
Electric generators Steam turbines, bleeder or straight
condensing
Carbonating tower gas compressors Horizontal reciprocating steam engines,
or steam turbine in centrifugal
compressors
Pumps for main supply of cooling water Steam turbines or electric motors
Pumps for handling brine and inter-mediate liquors Electric motors
Elevators, conveyors, agitators, etc. Electric motors
Filter vacuum pumps and absorber Steam engines, steam turbines or
vacuum pumps electrictic motors

arranged for electric motors, reciprocating steam engines, steam turbines and even water turbines
or internal combustion drives. In present practice, those given in Table 133 are usual.
Because of the convenience of wiring in the tall buildings and scattered machinery necessary
for the most economical arrangement in the process, all drives smaller than about 75 or 100
horsepower are usually electric motors. The combined efficiency of a large turbo-generator,
electric transmission wire and induction motor is superior to any other under such conditions.
From the maintenance cost standpoint such combination is also decidedly superior.
As one requirement for the ammonia soda power plant is to run the plant non-condensing or
with bleeder turbines with a back pressure of about 10 pounds gauge, any efficiency in the
utilization of heat must come from the use of higher steam pressure and superheated T 1 − T 2
T1
steam .For the higher steam pressure and the presence of high superheat mean a higher
initial temperature for the working fluid. According to the Second Law of Thermodynamics, the
maximum possible efficiency in utilizing the heat (i.e., for a reversible process) is given by the
fraction
where T1 is the initial absolute temperature and T2 the final absolute temperature of the working
fluid .As we cannot run the engines entirely condensing ,little can be done with T2 ,and the only
thing left to do is to boost T1, i.e., to employ a higher steam pressure and greater superheat in the
steam, in order to make the value of this fraction as large as possible. The modern tendency in
soda ash plants, as in the case of central power stations, is to use high working pressure and high
superheat.
The heats of formation (from the element) for the raw materials and final products of the
overall chemical process of the ammonia soda industry very nearly cancel each other .Hence
consumption of coal in a simple ammonia soda plant ,on the basis of the First Law of
Thermodynamics, would be almost zero .On the contrary ,as seen from the data given on page 516
and in Table 148,coal and coke together head the list of expenses of production .Direct operating
cost and at times total payroll are generally smaller than the fuel bill, though not invariably so. The
overall net reaction in the ammonia soda process:
2NaCl+CaCO3 Na2 CO3 +CaCl2 ± small heat effect
can be compared to the overall performance of an ice-manufacturing operation, the equation for
which might ,for purposes of his analogy, be stated:
Warm water +cold brine=warm brine +ice±"0" Calories
Therefore, the heat efficiency of an ammonia soda operation must be scrutinized on the basis
of the Second Law of Thermodynamics. Its overall operation is analogous to that of any operation
whose ideal is the Carnot cycle.
We may write the equation connecting heat and work in the Second Law of Thermodynamics
as

dw T − T " T"
= =1−
dQ T T'

wherein dQ represents the differential amount of heat to be extracted from any source of
heat(steam coke fire in kilns, fuel in rotary calciners)at the necessary temperature T, and dW
represents the reversible work that can be obtained from this heat on transferring it down to the
practical cold reservoir temperature T" of the plant in question.
If state 1 is taken as the initial condition and state 2 as the final condition of equilibrium at
cold reservoir temperature T'. then the above equation integrates to
W1 -W2 =Q1 -Q2 +T’(S1 -S2 )
dQ
(S=entropy or
T
Alkali plants should be compared by the above equation when studying possible heat or fuel
improvements. In usual good practice, alkali plants have a Second-law efficiency in the order of
40 to 50 per cent. That is, when operating neat the most profitable condition for the local costs of
fuel and fixed charges for equipment, "good" plants are burning about twice the amount of fuel
that is theoretically required. With present technical knowledge and available materials of
construction, this efficiency could be increased, but in all probability not much beyond 60 per cent.
Still higher efficiency would require untried and uneconomically expensive procedures, such as
binary fluid systems, or the decomposition of limestone, the generation of steam power, and the
calcination of bicarbonate in series from a single fire, etc.
As to the boilers, nothing but the water-tube type should be considered, The use of a shell
type or a fire-tube type in the form of a horizontal return tubular boiler is entirely out of date.
Water-tube boilers have many advantages. Chief among them are:
(a) Large grate area (for use with poorer grade coal); large heating surface per rated b.hp.,
and large steam-generating capacity.
(b) Greater safety and higher mechanical strength ni the construction; accidents, if any,
occurring in the form of tube fissure rather than shell explosion.
(c) Better circulation of water in the boiler, equalized temperature and uniform expansion in
all parts of the boiler, and consequently greater freedom from strains due to unequal expansions.
(d) Quick steaming.
(e) Accessibility for repairs and good facility for cleaning; all scale formed inside the thbes
being easy to get at and may be cleaned by means of turbine cleaners.
(f) Higher efficiency and economy in heat utilization.
From the above, it will be seen that a water-tube boiler is suitable from every point of view.
In large units a three-drum bent-tube type, or Stirling type, of water-tube boiler is especially well
adapted. Although it appears simple to clean the tubes of a straight-tube water-tube boiler, the
tubes are bent so as to take care of their own natural expansion. Therefore, in a bent-tube boiler,
tube trouble seldom occurs and removal is rarely necessary. In a straighr-tube boiler, especially
when it is subject to frequent shut-downs for cleaning, etc., the straight tubes may be bent or
twisted to such an extent that the tube ends shrink into the tube sheet at either end. This will
necessitate frequent renewal of the boiler tubes.
There are two general types of water-tube boilers available: those `where the tubes enter
common drums and those where the tubes are brought into relatively small headers. The sharp
distinction between these two types has disappeared to some extent in recent years because of the
rapidly spreading use of various kinds of "water-wall" furnaces. The latent heat of evaporation of
water decreases rapidly with increasing pressure in the water-steam system. Hence ,at higher
pressures the amount of heat which must be transferred through boiler heating surface for the
actual conversion of water into steam is appreciably smaller the at lower pressures. At the same
time, the amount of heat which must be transferred through boiler and economizer surface to bring
the temperature of the water up to the boiling point is reatively greater. Likewise, at higher
pressures the correspondingly high steam temperature requires extended superheater surface.
Therefore modern, high-pressure Stirling boilers with water-cooled furnaces often have more tubes
entering manifolds than drums. Those parts of a boiler in which the water is below the boiling
point are generally referred to now as "economizer sections." these parts may be either actual
economizers (operated above boiler pressure)or they may be "steaming" economizers. In the latter
case, they are really a part of the "convection" surface of the boiler. In either case, being
constructed of a multiplicity of tubes, they should be such that external and internal surfaces are
readily accessible for cleaning. The external surfaces, that is, those presented to the fire or
products of combustion, should be provided with properly located soot blowers or automatic
cleaning devices. The water treatment should be selected so that internal cleaning is very seldom
required, This suggestion is made more from the standpoint of operating safety than from that of
the reduction of maintenance costs.
In modern boiler equipment there should be a minimum of refractory surface, The highest
rates of heat transfer are obtained where absorption of radiant heat is possible. Consequently the
"all water-cooled furnace" permits very large steam generating capacities per unit of floor space
which, of course, provides a saving in initial cost; and the alkali plant, because of its high fuel
consumption, should take advantage of such developments.
The size of the boiler units is best kept relatively small from the standpoint of the shut-downs
required for routine inspections. On the other hand, units that are too small unduly increase the
initial investment and result in larger direct operating labor cost and in low efficiency.
For a soda plant which requires 1000 b.hp., or more, the boilers should be stoker-fired rather
than hand-fired, not only because of limitations of human physical strength, but also because
uniformity of operation and economy in utilizing the heat in the coal consumed. If the calorific
value of the coal is very low because of high content of ash or foreign matter, special effort should
be made to secure extra-large grate area. If the coal has only ten thousand or less B.t.u. per
pound,1 square foot of grate area to 4 to 5 rated b.hp. is none too much .A draft gauge should be
provided in every furnace to determine the amount of draft the furnace is carrying. If a CO2
recorder is not provided, the flue gases should be frequently tested with an Orsat apparaus. The
draft that gives the highest CO2 test in the flue gases should be considered as the optimum for the
furnace, and the damper position adjusted accordingly. For large soda plants, especially where
larger units of boilers are required and low-grade coal is to be used, pulverized coal with a unit
pulverizer system and a high-pressure boiler with water-cooled walls give very good effic iency
and economy in the utilization of coal. Pulverized coal does away with stoker firing and simplifies
the ash-handling problem materially.
Since fuel is generally, if not always, the largest single item among the operating expenses in
alkali plants, it is worth while to consider every possible "heat-trap". Justification of each heat
recuperation or the extent thereof depends on the local cost of fuel. If the fuel is coal, as it is in a
great majority of alkali plants, a decision must be made as to the method of firing. That is, whether
hand firing, automatic stoker or pulverized-coal firing is the most profitable. There is still an open
question whether pulverized fuel firing or mechanical stoker firing gives better results all around.
In America, pulverized fuel has come to the fore, especially for very large steam-generating units.
For steam generating capacity below 35,000 lbs. per hour, however, stoker firing is, as a rule
recommended. In Europe, the stokers still find great favor. With mechanical stokers, the operation
is simpler and the first cost is slightly less Mechanical stokers in Europe have been developed to
such a perfection that it is claimed that as good efficiency can be obtained with stokers as with the
pulverized coal method, particularly in small units. On the other hand, pulverized fuel firing is
adaptable to almost any grade of coal and yields good efficiency in the case of high-ash, low-grade
coal. although the power consumed in pulverizing high-ash coal is considerable. The maintenance
charge is higher with the pulverized coal system. In localities where the air is humid and there is
humid and there is much rain, the amount of moisture in the coal generally causes trouble in
pulverizing, and necessitates a separate coal drying plant when it much exceeds8-10 per cent.
Chemical treatment or conditioning of the feed water is carried on somewhere in the
preheating cycle and its various methods are discussed in detail in Chapter XXIII. In plants
making no caustic, substantially none of the condensate from the steam generated is available for
boiler feed water. In plants where the steam requirement of the concentration of caustic is a
considerable fraction of the total steam, most of that fraction is available for boiler feed water in
the form of the condensate, but it is frequently more desirable to use it for mud washing, etc.
Boiler feed water treatment is therefore an important item of expense in connection with the
operation of steam plants for soda plant operation. Not only is the water treatment equipment very
large per unit of boiler capacity, but the condensates, drips or such other sources of relatively pure
water from the processes which are occasionally used, may be contaminated with alkali. This
introduces danger of boiler embrittlement. Until the physical chemistry of the cause and
prevention of embrittlement is better understood and methods for its control made more precise,
the sulfate-to-carbonate ratios recommended by the ASME are probably the safest procedure for
the alkali plant operator, except where the drums are fusion-welded.
If the feed water contains considerable temporary hardness and calcium sulfate, an open-type
heater operating on the exhaust steam should be used in order to precipitate most of the calcium
acid-carbonate and sulfate from the water before it enters the boiler. Where exhaust steam from
cylinders is used for boiler feed, the oil must be carefully separated out in an efficient oil separator.
Formerly a closed-type heater was used, operating on the exhaust steam. The latter type is now
seldom used because leakages are likely to occur at the tube joints and the cleaning of the tubes is
rather difficult. Open-type heaters are now made with efficient of the oil. Many heaters are now
made in the form of deaerating heaters.
The entire operation of a soda plant depends for its economy on the maintenance of continuity,
with uniformly maintained accurate control. Hence boilers produce steam at a much more constant
rate than in daylight operating industries or in public utilities. There are no daily or diurnal
peak-load periods. Likewise there are no noon-hour "valleys" or other slack spells during which
some routine periodic operation like blow-down, soot blowing, fire slicing, etc., can be done better
than at another time. For this reason a number of special features are valuable to the boilers and
auxiliaries, such as special distribution of soot-blower elements, continuous blow-down equipment
and control, several emergency sources of fuel, design for moderate ratings, etc.
Since there is very little difference in seasonal load on an alkali power plant, the equipment is
designed to be also designed for a very high "availability factor".
The boiler plant itself, since it is the place where the most costly of a single item of material
is consumed, should, in a well operated alkali plant, be kept up-to-date and in first-class operating
condition. The local manager of an alkali plant is most frequently a graduate from the ranks of
operators of the chemical equipment. It is, therefore, not unusual to find the power plant a
somewhat neglected area insofar as application of the latest improvements is concerned. The boiler
house is more often considered a regrettable spot in connection with the operation of an alkali
plant than a place where ingenious and careful operation could effect large-scale economies.
On the other hand, all the large pieces of equipment in an alkali boiler plant are the ones
usually found in other industries. The boiler plant thus receives a degree of attention thrust upon it
from outside sources, which is not accorded to other parts of the plant. That is, sellers of modern
improvement devices for power plant equipment, in continually trying to familiarize their potential
customers with the advantages to be gained in utilizing such developments, appear froquently at
the plant. Unfortunately, many of these bring about a profit to the user in only a portion of the
possible case where they can be applied. Consequently, the alkali power plant operator and the
purchasing agent have adopted an attitude of wilful reluctance which occasionally shuts out a
useful improvement.
The boilers themselves, as pointed out above, differ from other power plant boilers in that
they must be designed for long-period, uniform-load, continuous operation. The happenings which
prevent unlimited continuous operation can be divided into two categories:
First, the accumulation of insolubles precipitating from water being evaporated inside the
boiler.
 Secondly, the accumulation of solid products from the fuel outside.
The boilers must, of course, be primarily designed and operated with due consideration for
the particular feed water which is available and its suitable treatment, as will be discussed in
Chapter XXIII. They will then come as close as is profitable to being always in condition to have
the accumulations of solids removed as solutions or suspensions by blow-down procedures. Such
operation will require a boiler to be opened only for quick statutory and insurance inspection,
insofar as internal surfaces are concerned.
On the fire side, the accumulation are either ash or soot. Adequate design requires soot
blower location to be such that the regular operation thereof leaves no accumulations to be
removed during outage. Ash removal must likewise proceed without interference with routine
combustion of fuel. Whereas in a normal industrial power plant, clinker masses can perhaps be
sliced up by hand during the night shifts, such procedure will surely penalize the night shift's
output in an alkali plant.
The steam required or for the distillation of ammonia could theoretically be used at highest
pressures, but a high-pressure distiller for materials having the properties of ammonia liquors
would be an expensive and, more particularly, an unwieldy unit. Furthermore, at the high
temperature involved in a pressure still, it is likely that difficult scale-forming conditions would be
set up in that portion of the apparatus which handles the limed solutions. Again, a distiller can be
designed which, theoretically at least, can operate at a very low steam pressure in the bottom. Such
a still would require mechanical design features to permit low pressure drop throughout the
distiller-absorber system. Consequently, its capacity per unit of cross-sectional area would be low.
Some operating experience with low-pressure ammonia distillation indicates that scale deposition
even more difficult than the present would have to be solved by expensive trial and error methods.
 FIG 100 Diagrammatic connection of power equipment in simple ammonia soda plant.

It is, therefore, universal practice to operate distillers at moderate steam pressures in the
bottom, that is, from about 2 pounds below, to a maximum of 10 or 12 pounds above the
atmospheric pressure. Since the consumption of steam is large compared to any other unit in the
plant (that is, any plant not including chemical caustic or electrolytic caustic operations), it
controls the balance of the power plant situation. The prime-movers are collectively arranged to
exhaust the required distillation steam, either entirely from electric generators (in a 100 per cent
motor-driven plant) or from these and the other steam-drive prime-movers. In the average (without
electrolytic) case, this involves a boiler plant in which steam at a medium pressure and
temperature is generated, a portion being delivered to electric generators and the rest to gas
compressors, vacuum pumps and turbine-driven water pumps. These all exhaust into a common
exhaust main which conducts the steam for the distillation operation and building heating
operations. A part of the exhaust steam is also frequently used for the final preheating of the boiler
feed water on its way to economizers, located in power plant chimney ducts.
 FIG 101 Diagrammatic connection of power equipment in an ammonia plant making custic soda also.

From the foregoing it will be apparent that all mechanical drives, even including the largest,
can be electric -motor driven. Then steam at boiler pressure is fed only to the electric generator
turbines, which exhaust directly to the distillers, etc. This would make a very clean arrangement,
but as far as is known it has never proved ecconomical.Fig.100 shows a diagrammatic connection
of power generating and consuming equipment in a simple ammonia soda plant.Fig.101 shows the
same arrangement as applied to an ammonia soda plant which also has chemical caustic soda and
electrolytic caustic soda operations served by the same power plant and under the same
administration. This latter illustrates the most frequent American practice.
The pressure required at the boiler plant for any one combination of ammonia soda, chemical
caustic and electrolytic caustic (and other byproducts) can be determined, and curves similar to
Fig.102 can be drawn. These curves are shown with boiler pressure as ordinate, independent of
steam temperature. Steam temperature is definitely limited by "creep" of metals and does not
influence the total price of boiler equipment up to 750F.,and then causes an increas e of price in one
step anywhere up to 950F.(the maximum temperature available today). Therefore, in this
elementary discussion it has been eliminated from consideration. The choice of boiler pressure is
then determined by a balance between the increasing cost of boiler plant with increasing boiler
pressure and the decreasing cost of fuel resulting from the corresponding lower total steam
consumption.
 The choice of boiler pressure is then determined by a balance between the increasing cost of
boiler plant with increasing boiler pressure and the decreasing cost of fuel resulting from the
corresponding lower total steamconsumption.
A second and even more important study must be made to determine the best boiler pressure.
This has to do with the fluctuation in the rate of production of the various products. That
fluctuation is generally not in the control of the plant operating personnel. Best results under
circumstances of fluctuating production rates require a different boiler pressure from that for an
absolutely steady production.
Fig.103 shows a Cartesian coordinate, wherein the number of kilowatts is the ordinate. The
abscissa is a non-rigorous function of the plant-operating rate. In a simple ammonia soda plant this
can be rigorous, i.e., the number of tons of soda ash produced per unit of time .In a plant making
both soda ash and chemical caustic, good coordination can be found when the number of tons of
soda ash and the number of tons of caustic soda produced are merely added together into a
"combined" product. Reasonable coordination is also obtained when electrolytic caustic is added
to the other products to form a "combined" total for plant production. It is immediately apparent
that wide variations in the production rate of one product relative to the others render such a chart
quite inconclusive. However, since the combination of circumstances which underlie wide
fluctuations in the rate of plant production generally affects all of the products alike, ratios are
reasonably well maintained regardless of the production rate. This type of chart is found to be very
useful in determining the most profitable boiler pressure, either when designing and entirely new
plant, when replacing obsolete power-generating equipment, or when considering the extension of
power generating equipment (topping plant) needed in connection with expansion of production or
addition of new products.
 FIG 103 Relationship of steam pressures to output of products manufactured.

As indicated by the crossed curves on Fig.103,it is characteristic of the modern equipment of

an ammonia soda plant to have a deficiency of electric energy available at low rates of operation
and excess at high rates of operation. This is because the efficiency of mechanical drives goes
down sharply at low loads, whereas the efficiency of steam consumers dose not .For example ,the
electric motor which drives the agitator in the prelimer tank will take exactly as much power,
whether that particular distiller assembly is operating at 100 tons per day or at 200 tons per day.
Where several distillers and prelimers are being operated in parallel at high rates and one or more
are shut down entirely at lower outputs of ash, the electric power drops proportionately; but such
control is seldom available. In either case, to a greater or less degree, the steam consumption drops
in almost exact proportion to the production, whereas electric power consumption does not. Many
of the other electric loads, like screw conveyors, are analogous to the agitator load used for the
illustration above. With a large number of the motors (such as drives for bucket conveyors or
centrifugal pumps), when the out put of the plant is reduced, the load falls somewhat, but not in
direct proportion.InFig.103, the two curves with the steeper slopes represent the electric power
which can be generated by sending the demanded process steam, at either of two boiler pressures,
through turbo-generators of average efficiency. The single flatter curve represents the demand
for electric power corresponding to the output of products, It is apparent from the chart that a
"balance point" exists for each boiler pressure in a given plant. This is the optimum production
rate at which there is neither deficiency nor excess of electric energy. It is an invariable practice to
bypass steam from boiler pressure through a pressure-reducing valve at production rates higher
than the balance point. At production rates lower than the balance point, either the excess exhaust
steam from power generation is blown to the atmosphere or the deficiency of electricity is made up
in condensing units. This deficiency in electric power may also be purchased from a public utility
or adjacent outside concern in the rare cases where such procedure is possible. In large plants
where, through the years that they have grown, a variety of types of boiler and electric generating
equipment are in operation or available for operation, the plant can usually be kept" in balance" by
an interchange of load between high-pressure and low-pressure stations, or between condensing
and non-condensing units. Smaller, simple ammonia soda plants get the balance point down to the
lowest output not requiring excessively expensive boilers, and then depend on the simplicity of the
reducing valve for giving a good year-in, year-out efficiency.
In many of the soda plants, all the large engines and mechanical drives, the gas compressors,
water pumps, air compressors, etc., are considered to be within the province of the "generation of
power." In the following, a brief discussion of the merits of various forms of such equipment is
given. An effort is made to show why local condition affect choice of the type of equipment.
The gas compressors for compressing, limekiln, rotary dryer or mixed gases into carbonating
towers (columns) are pieces of equipment most expensive to buy and to maintain in the power end
of the ammonia soda business. Regardless of the degree to which limekiln gases and rotary dryer
gases are scrubbed, an amount of dust finds its way into the compressors. The industry has
gradually evolved suitable poppet, plate, and rotary types of valves for the gas conpressors. These
valves are generally of such physical dimensions that slow-speed reciprocating engines are almost
mandatory .Low piston speeds from 450 to 500 feet per minute are considered desirable. Recently
the introduction of stainless steel, suitably resistant to the corrosive action of the gas, has permitted
some progress with "Feather" valves. This requires a special lubricant and lubricating procedure.
Inlet valves on the gas end of the compressors are generally of the half Corliss type without the
trip-mechanism. Stainless steel piston rods and metallic packings in gas cylinders as well as in
steam cylinders have in recent years reduced maintenance costs considerably. Central lubricating
systems for steam end and mechanical parts, with dilute soda solution system for gas ends also
tend toward lower maintenance cost. The heavy, mechanically operated gas delivery valves require
engines of very rigid frame and bearing design. These mechanically operated valves have done
away with much valve trouble. These compressors are generally more expensive per cubic foot of
gas compressed on account of low piston speeds than are ordinary air compressors.
Some rotary cycloidal compressors have been used in small capacities; and these units,
although quite inexpensive, and of relatively low maintenance cost in larger capacities, are low in
efficiency except at some one particular rate of operation or discharge pressure. Recently the
industry has shown interest in the centrifugal (or turbo) compressor because of its considerably
lower first costs. These units are rarely profitable for a small plant, but show a decided advantage
for a very large plant. The approximate dividing point between the "large" and "small "designation
depends on a number of factors peculiar to local conditions. In plants using the double entry
system at the making columns, the pure limekiln gas will be profitably compressed into entry by
centrifugal compressors, in capacities as low as 4000 cubic feet per minute. The bottom entry
requires one or two more stages in the centrifugal compressor, but units smaller than 5000 cubic
feet per minute would usually favor the reciprocating machines. The centrifugal machines are
generally direct-driven by turbines, taking steam at boiler pressure. When they are motor-driven,
economy of materials usually dictates a step-up gear set. The centrifugal compressor has a decided
advantage in low maintenance and first cost, but generally a decided disadvantage in thermal
efficiency. Further disadvantages of the centrifugal compressors are (1) that they are not positive
enough in forcing the gas into the columns like the piston-type compressors; (2) that their
effic iencies are low, especially in small units; and (3) that their output is so large that several
columns have to be served in parallel by one machine, causing difficulty in adjusting the volume
of the gas delivered to each column from a common main. The detailed requirements of the
compressor end as to internal design and materials of construction are not yet standardized to the
same degree as in reciprocating compressor construction. Another disadvantage, or rather point of
objection to the high-capacity centrifugal compressor, particularly where plants are large enough
to justify its adoption, is that the single unit requires installation of regulating devices for
distributing gas from a common discharge manifold to the several carbonating towers. In the cas e
of the smaller capacity reciprocating machines, one machine, or one cylinder of a duplex machine,
or even one end of a cylinder, is separately piped to each carbonating tower. Hence no volume
regulation beyond the speed of the compressor is required.
The vacuum pump for pulling air through the filter cake of crude bicarbonate is another piece
of mechanical equipment peculiar to the industry .The air is generally scrubbed (on the suction
side of these pumps) with cold brine to remove ammonia. The gas contains some CO2 flashed off
from the mother liquor under the mother liquor under the action of reduced pressure inside the
filter, and also contains a minute of brine spray and perhaps also a little mother liquor. The most
complete practical separation of these sprays is of great importance to the low maintenance costs
of the vacuum engines. Reciprocating vacuum pumps, with valves similar to those described for
the gas compressors, are used extensively. The rotary type of positive displacement vacuum pump
may be used in relatively small capacity units. The rotary pump of slide vane type, such as made
by Fuller Company, Catasauqua, Pa., may also be used. The gas handled by these machines is
from 3 to 8 per cent CO2 The development of superior corrosion-resisting materials of construction
has permitted operation of such units with the scrubber on the discharge side. This reduces the
vacuum piping cost materially and also reduces the size of both the vacuum pump and the
scrubber.
The exhausters for removing non-condensing gas from the ammonia absorber train may be
similar to the filter vacuum pumps described above. In the older ammonia soda plants, the
absorber vacuum pump was about 1 /4 to 1 /2 the size of the filter vacuum pump. In modern plants,
with the efficient types of absorbers made possible by purification of brine, these vacuum pumps
become extremely small, since CO2 is absorbed much more readily by ammoniacal brine in a
ceramic -ring packed than in a bubble-cap washer. Consequently, as there is practically no other
"non-condensable" gas in the distiller-absorber system except the residual air, the absorber vacuum
pump becomes a small control mechanism, rather than a high volume exhauster, provided there is
no leakage of air in the vacuum system.
Air compressors for the air requirements of the ammonia soda plant do not differ from those
of any other industrial plant. Compressed air is an expensive commodity, often too lightly regarded.
There is generally no justification for the use of compressed air for tank agitation, inducing
chminey drafts, cooling hot bearings, etc. As an aid to extensive maintenance work, it is almost
indispensable, but its indiscriminate use for other purposes should be carefully guarded against in
order to avoid waste of power and consequent high fuel consumption per ton of soda ash made.
The cooling water pumps in ammonia soda plants are generally of a very normal design. The
height of the water outlet in the carbonating (making) towers and the distiller condensers generally
determines the total head to be carried. Water pumping plant heads of about 130to 150 feet are
therefore normal, provided the pump capacity is adequate. In plants having various pieces of other
water-cooled equipment requiring different heads for example, caustic liquor evaporator
condensers it will often pay to have a two-head system. The materials of construction for the
pumps will depend principally on the nature of the impurities in the cooling water supply. In
general, a cast-iron pump with bronze fittings is suitable, but for the sake of interchangeability, an
all-iron construction is preferable.
Electric generators required for the ammonia soda industry are tailor-made to fit the balance,
which exists. The turbine generator manufacturers have highly developed designs, which can be
arranged to be applicable to any condition met in the industry. In general, the equipment should be
governed for constant speeds with a voltage regulator and a constant and adjustable extraction
steam pressure. Two-pole machines directly driven by extraction-type steam turbines are the
modern preference in America. Sizes today range from a minimum of 1000.KW up to perhaps
10,000KW. in single units.
As regards feed water pumps for the boilers, it is good practice to have at least two
independent unit motor-driven and the other turbine-driven, and in addition one steam ejector for
the emergency second reserve. The steam ejector is simple, convenient and economical, and may
be connected to a cold water supply for certainty of operation. The ejector will produce a pressure
of from 50 to 80 1bs. Per sq. in. greater than the boiler pressure and can tied over any period of
emergency. But it is not to be used, unless both the motor-driven and the steam-driven pumps are
out of commission; which indeed very rarely happens.
 Chapter XXIII
Boiler Operation and Conditioning of Boiler Feed

and Cooling Water in Ammonia Soda Plants

In an ammonia soda plant, where high steam pressures are used, feed water becomes a matter
of increased importance. The problem in an ammonia soda works is aggravated by the fact that, as
little returned condensate is available for the boiler feed, almost 100 per cent make-up is required.
This situation presents a problem common to many chemical industries where exhaust steam is
utilized in the process. Hence in what follows, while attention applies equally well to industrial
chemical plants in general.
We shall begin with a general treatment of the subject. Modern steam generators require
rather pure water for steam generation, not only because of the high working pressure maintained,
but also because of high steam generating rate, particularly when pulverized fuel is used with
water-cooled walls. Today, the boiler rating is no longer 10 sq.ft. of heating surface per boiler
horsepower(B.H.P.), but generally a figure of from 200 to 400 per cent of this rating. This, of
course, calls for a very high rate of heat transmission through the metal walls and requires good,
pure water to keep the internal surface clean.
No natural water (not even rain water on account of dissolved oxygen) may be used directly
in such boilers without some form of treatment. The nearest approach to a pure water would be
distilled water or the condensate from the steam system; but even in these cases the water may not
be completely free from oil or dissolved oxygen and CO2, nor can it be of proper alkalinity
(suitable pH) to prevent corrosion in the boilers. All natural waters contain more or less dissolved
solids and suspended matter, and the feed water as received in the boiler room may contain,
besides these mineral matters, oil, grease, organic matter, dissolved oxygen and CO2 , etc. All these
constituents have specific effects on boiler operation.
To summarize, the boiler feed water, if not properly cared for, may contain the following:
(a) Solids: (1) Soluble: mineral matters, alkali salts, alkaline-earth salts, (2) Suspended:
sand , clay, organic ,matter.
(b) Liquids: oil, grease, etc. (mostly from the lubricating medium).
(c) Gases: oxygen (from dissolved air), CO2 (from alkaline bicarbonates and carbonate
hardness).
Raw make-up water cones from two main sources: (1) surface water and (2) artesian Water.
Surface water may be lake water, river water, surface stream or any body of water collected from
the drain over the earth surface. Such water can be quite “soft,” but sometimes also “hard.”
Mountains stream and falls generally give soft water. Artesian well water comes from
underground layers of some depth below the surface (200feet or more). This is water that flows
through underground gravel layers or sand-bearing strata. It is generally a hard water, but
sometimes it may contain high alkaline bicarbonates with practically mo hardness. The source,
therefore, does not disclose the character of the water.
According to the character of the water we may therefore classify water for boiler purposes
as
Follows:

1. Hard: Temporary hardness

Permanent hardness
2. Soft .
3. Acid
4. Alkaline
Hard water is water containing dissolved sulfates, carbonates, chlorides, etc. Of calcium,
magnesium, iron, etc. If these mineral carbonates are in the form of bicarbonates they can be
decomposed by heating, and the resulting normal carbonates of calcium, magnesium, iron, etc.
Precipitated. Hence water containing these bicarbonates is said to possess “temporary hardness” in
contradistinction to the sulfates, chlorides, etc. Of calcium, magnesium, etc. which cannot be
precipitated by heating and are therefore said to constitute “permanent hardness.” “Temporary
hardness” is undoubtedly due to the solution of the alkaline-earth carbonates by CO2 dissolved in
water. When the water flows through limestone deposits.
Soft water is water that contains little or no alkaline-earth salts. It may, however, contain
considerable quantities of alkali salts, such as NaHHO 3 , Na2 CO3 , NaCl, and Na2 SO4, which do not
form are scale in the boilers. From the consideration of boiler operation, however. Water
containing such soluble salts in large quantities is not viewed with favor. No chemical treatment in
the ordinary sense can remove these sodium salts. One possible exception is the
“De-Mineralizing” process or the hydrogen exchange “elite” treatment (see below). Besides, if
these soluble salts exist in the form of sodium carbonate or bicarbonate, they may lead to caustic
difficulties in the boiler.
Acid water is water containing free acid (H2 SO4 ) or acidic salts [Al2 (SO4 )3 , FeSO4, MgCl2 ,
CaCl2 , and alkaline-earth bicarbonates]. It may have come from the neighborhood of coalmines or
peat regions where water has come into contact with sulfur-bearing minerals, such as the pyrites,
which through oxidation form sulfuric acid or sulfates of the metals. Such water is generally also
very hard. Treatment of line and soda followed by a elite treatment will render it usable for steam
generation.
Alkaline water is water containing alkali bicarbonates and carbonates. Generally such water
contains high bicarbonate of soda but hardness is practically absent. Absent. It is very similar to an
over-treated water having negative hardness, or alkaline hardness, from the soda-lime treatment.
The high alkalinity probably owes its origin to the decomposition of feldspars by the
dissolvedCO2 , yielding kaolin and potassium (or sodium) carbonate which remains dissolved in
the water,

Al2 O3 .k2 (or Na2 )0.6SiO 2 +H2 CO3 +H2 O Al2 O2 .2SiO 2 . 2H2 O+K3 CO3 (or Na2 CO3 )+4SiO 2
K3 CO3 (or Na2 CO3 ) + H2 CO3 2KHCO3 (2NaHCO3 )

or else to the reaction of calcium and magnesium bicarbonates in contact with natural zeolite bed
causing a base exchange, depriving the eater of all its hardness,
Al2 O3 .Na2 O.4SiO 2 .6H2 O+Ca(HCO2 ) Ai2 O3 .CaO.4SiO 2 .6H2 O+2NaHCO3
In this way, deep wells sometimes yield water which is not hard, but contains a high
concentration of sodium bicarbonate.
Hardness of water is expressed in various ways in different countries. It is, however,
generally expressed in terms of CaCO3 except in the German calculation where CaO is used. The
following table gives the comparison. Based on ppm. as CaCO3, the American Association of

Railway Chemists grading may be expressed as follows, although no strict line can be drawn
between different classifications, and opinions nay vary considerably with individual water
experts.
Table 135 Grading of Boiler Feed Water.
Hardness as CaCO3
(ppm.) Grading
140 or less Excellent
140 to 260 Good
260 to 350 Fair
350 to 520 Poor
520 to 700 Bad
700 up to 4000 Very bad
Just to show wide variation in hardness of water from different parts of the country, the
following table lists a few of such samples
Table 136 Hardness of Water in Different Localities.
Water Sample Hardness (ppm.)
Great Lakes water 120
Missouri River water at Omaha, Neb. (April) 215
Chicago Heights City Wells 600
Arkansas River at Little Rock, Ark. 189
Maumee River at Toledo, O. 257
Jacksonville City Wells (Fla.) 274
Red River at Shreveport, La. (January) 86
Greenwood City Wells (Miss.) 35
Detroit River (Detroit, Mich.) 103
Toledo City Wells O. 1540
Oklahoma Private Wells 4080
Hardness in the boiler water leads to scale formation or incrustation in the tubes and drums.
Of the various constituents present in boiler water, the scale formed by calcium sulfate, silica and
calcium hydroxide is the hardest and considered serious. The scale is a good insulator of heat and
its formation is side the tubes and drums is a matter of grave concern to boiler operators,
especially when the boiler is operating at a high rating. On account of excessive local heating of
the metal, the tubes may blister or even bulge under pressure, causing damage and serious
shutdowns. In bad cases, boiler explosions have occurred causing losses of life and property. In
general, as the incrustation is formed, the rate of heat absorption by the metal drops, and the flue
gas temperature rises. Following is an estimate of the percentage of heat losses through scale
formation in a boiler, as determined at the University of Illinois:

The above table gives some idea of the magnitude of heat losses, but the figures vary
according to the individual boiler setting. Nowadays chemical treatment has reached such a state
of perfection that the presence of any scale in the boilers of an ammonia soda plant is no longer
considered excusable.
To prevent scale formation in the boilers, chemical treatment is resorted to. The treatment
was formerly done inside the boilers (external treatment). Internal treatment was used to a very
limited extent in small soda plants, which did not warrant investment for a separate water-treating
plant and which employed only low-pressure steam. In such cases, chemicals suc h as trisodium or
disodium phosphate, sodium aluminate, soda ash, caustic soda, sodium silicate, vegetable (annins
or any combination of these, have been added to the feed water. These ate mostly alkaline
chemicals and the object is to convert CaSO4 , CaCl2 , etc.(which form a hard scale)to
Ca3 (PO4 )2,CaCo3 ,erc.which form a loose sludge .Introduction of such strongly alkaline chemicals
and organic matter into the boiler, however, is always objectionable and cannot be justified except
in very small soda plants with a low working steam pressure (180-160 lbs.) In all cases, the
addition of these substances to the feed water increases the soluble solids and alkalinity in the
boiler saline, and the phenomena of foaming and priming and even tube ruptures may occur. For
 FIG 104 Curves showing carbonate-to-sulfate ratio in boiler for prevention of hard smile

steam pressures above 200 lbs. Such internal treatment is not recommended.* Fig. 104 gives
empirical graphs showing the carbonate-to-sulfate ratio that must be maintained to prevent
formation of hard scale on the metal walls in the boilers.
These lines show the ratio of carbonate-to-sulfate that must be present in the boiler at a given
working pressure to secure loose carbonate scale. Points to the left or above the respective lines
mean deficiency in sodium carbonate in the boiler. In using this chart, however, one must always .
* At present, some water experts advocate internal treatment to limited extent for high-pressure boilers by
introducing a portion of raw water into the boilers together with treated water, in conjunction with sludge.
Removal by means of the so-called sludge deconcentrator.
+ Boiler Feed and Boiler Water Softening by H. K. Blanning and A. D. Rich, p. 83, Nickerson and Collins Co.,
Chicago.
remember that high alkalinity caused by the addition of an excessive amount of soda ash or an
alkali in the boiler should receive first consideration in this treatment.
From the practical side of the boiler operation, difficulties encountered in the boiler room
may be one or a combination of the following:
(1) Scale Formation (in tubes and drums) frequently causing bursting of tubes or boiler
explosions.
(2) Priming and Foaming (yielding wet steam or causing unsteady water level in the steam
drum, or trouble of deposit in the super heater or even on turbine blades).
(3) Corrosion and Pitting (either in the tubes of the economizer, or in the tube heads in the
drums or on the drum surface below water level, or at rivet heads of the mud drums).
(4) Caustic Embrittlement (in the plate along riveted joints, or causing bursting of tubes).
A word may be said about the chemical analysis of water and the analytical report on its
constituents. Practice varies: some chemists express the constituents in oxide form, such as CaO,
MgO, Na2 O, SO3 , etc., while others prefer the ionic form, Ca++, Mg++, Na+, SO4 -, Cl-, etc. It is
now pretty generally established that the constituents should be expressed in ionic form in p.p.m
(parts per million), or gr.p.g. (grains per gallon), and then in “equivalents per million” (e.p.m.).
This last is obtained by dividing grams of each constituent per million grams of water by its
combining weight, and is thus equal to million grams of water by its combining weight, and is
thus equal to milligram equivalents per liter. This has the advantage in that the relationship among
various constituents present may be seen at a glance. Take, for example, the composition of a river
water, such as water from the Yangtze, the chemical analysis of which, in the winter season, is
expressed as follows:

If the cum of the e.p.m.’s of Ca and Mg is greater than the e.p.m.’s of HCO3 -, there is
permanent hardness in the water and a hard scale may be formed in the boilers, if the water is not
properly treated. On the other hand, if it is substantially equal to the e.p.m. of HCO3 -, the water
contains only temporary hardness and a soft carbonate scale may be formed. If it is less than the e.
p.m. of HCO3 - which is sometimes the case with certain well water-the water in question is an
alkaline water and we must watch out for possible caustic embrittlement. From the point of view
of determining the necessary chemical treatment for the water in question, the advantage in using
the e. p.m. system is obvious. In the case of the Yangtze water in the winter season, it can be seen
from the above e. p.m.’ that the water contains 85 per cent of its total hardness (e. p.m. Ca+. p .m.
Mg=3. 007) as temporary hardness (e. p.m. HCO3 -=2.566), and the remaining 15 per cent as
permanent hardness, but the amounts are not large and therefore it is a rather soft water for the
source of raw feed water supply.
Attempts are often made to express the different constituents dissolved in the water as
conventional combinations, which would likely exist, if thes e constituents were present in solid
salt form. These are only hypothetical compounds and the results may be different depending on
the manner in which these constituents are combined. A general rule however, may be stated as
follows:
(a) Combine SO4 --with Ca++ as CaSO4
(b) Then combine balance of Ca++ with HCO3 -as Ca (HCO3 )2 or balance of SO4 -with Mg++ as
MgSO4
(c) Next combine balance of HCO3 -with Mg+ + as Mg (HCO3 )2 or balance of Ca++ with
Cl-asCaCl2 or balance of Mg++ with Cl- as MgCl2 or balance of SO4 -with Na+ as Na2 SO4
(d) Finally combine Na+(or balance of Na+) with CL-as NaCl
The total of the basic constituents should closely check with the total of the acid constituents,
if the analysis was correctly made, unless other acid radicals are present SiO 2 Fe2 O3 , and Al2 O3
are generally not combines. The above represents a general practice but the sequence is often
modified in special cases where certain other constituents are found to be present.
SCALE FORMATION
As has been mentioned above, scale formation is caused by hardness in the water,
particularly calcium sulfate, sili9ca, and calcium hydroxide. While temporary hardness may be
removed by heating and thrown down as the carbonate sludge in the preheated feed water,
permanent hardness must be removed by chemical treatment. This can be done by means of (a)
soda and lime,(b) lime and sodium aluminate, (c) caustic soda and sodium phosphate, or any
combination of these, followed by (d) a zeolite treatment, with subsequent addition of a proper
amount of trisodium or disodium phosphate or ‘Akon’ to the treated water entering the boilers.
Soda-lime Treatment (Modification of the old Scotch Professor, Thomas Clarke’s process)
By treating with lime:
Temporary hardness:
Ca (H3 CO2 ) 2 +Ca (OH) 2 2CaCO3 +2H2 O
Mg (HCO3 ) 2 +2Ca(OH) 2 MgOH2 +2CaCO3 +2H2 O
Fe (HCO3 ) 2 + Ca (OH) 2 FeCO3 +CO2 +H2 O
Or, on heating,
Ca (H3 CO2 ) 2 CaCO3 +CO2 +H2 O
Mg (HCO3 ) 2 MgCO3 +CO2 +H2 O
Fe (HCO3 ) 2 FeCO3 +CO2 +H2 O
The decomposition of the temporary hardness by heating is rather slow. Hence treatment with lime
is often preferred.
Permanent hardness:
CaSO4 +Na2 CO3 CaCO3 +Na2 SO4

CaSO4 + Na 2 CO3 → CaCO3 + Na 2 SO4


 MgSO4 + Ca (OH )2 → Mg (OH )2 + CaSO4
 MgCl2 + Ca(OH ) 2 → Mg ( OH ) 2 + CaCl2 
 
CaCl2 + Na2 CO3 → CaCo3 + 2 NaCl 
CaCl2 +Na2 CO3 CaCO3 +2NaCl
The soda-lime treatment is particularly suitable when water contains larger quantities of
non-carbonate calcium and magnesium hardness. Generally speaking, lime need be employed in
theoretical quantities, but soda ash added should allow foe an excess of about 25 p.p.m..
Lime-Sodium Aluminate Treatment. Sodium aluminate is hydrolyzed to caustic soda and
aluminum hydroxide:
NaAlO2 +2H2 O NaOH+ Al (OH) 3
So that the reactions are the sane except that aluminum hydroxide sepaates out as a voluminous,
gelatinous precipitate that serves as a coagulant foe the other precipitates. This treatment is
economical when water contains high temporary hardness, but watch out foe residual aluminum in
the treated water.
Caustic Soda-sodium Phosphate Treatment. The reactions are essentially the same, sodium
hydroxide taking the place of calcium hydroxide. Trisodium or disodium phosphate is used to
precipitate calcium, magnesium and iron. It is, however, important to observe that when caustic
soda is used, no large excess should be taken for treatment. Because of the expensive character of
the chemicals, they are often used only for the removal of the last trace of hardness.
3Ca++(or 3Mg++)+2PO4 -- Ca3 (PO4 )2 [or Mg2 (PO4 )2 ]
Zeolite Treatment. This treatment removes calcium and magnesium hardness almost
completely and can also be used independently when temporary hardness is low. Ordinary zeolite
treatment does not remove the carbonates, (CO3 --), does not lower the methyl orange alkalinity,
and does not reduce total solids in the boiler feed water so treated. Hence the zeolite treatment is
often used as an after treatment for one of the above preliminary treatments, (a), (b),and(c),and is
itself followed by an acid treatment and degasification to remove CO2 and to reduce the alkalinity
(or pH), as will be seen below. The soda-lime or lime-sodium aluminate treatment does not
remove all the hardness, especially in the cold; but it will remove most of the bicarbonates and
carbonates. Hence it forms a good combination with the zeolite treatment which forms a
subsequent treatment to remove the last portion of the hardness. This combination then gives
excellent feed water with practically zero hardness, but with low carbonates and hence low
alkalinity.
The zeolite treatment consists in treating the water by passing it through a thick zeolite bed,
generally 31/2-4 thick. The zeolite mass varies from 20 mesh to 40 mesh in size, the “effective
size” being about 35 mesh. The zeolites have approximately the composition:
Al2 O3 .Na2 O.4SiO 2 .6H2 O.Both natural and artificial zeolites are employed. When water is passed
through a zeolite bed, calcium and magnesium salts are converted into corresponding sodium salts;
hence this reaction is known as the Base Exchange Process.
Natural zeolite or Glauconite has a vase exchange capacity of 28003200 grains of equivalent
CaCO3 per cu. Ft., and artificial zeolite may have a capacity as much ah 7500 grains or more
Natural zeolite in the form of treated “Greensand” pf New Kerseu. Although it has lower base
exchange capacity than the artificial zeolite and consequently requires larger bed area, has,
however, several advantages: (1) natural zeolite is more durable and is not so likely to
disintegrate .as artificial zeolite; (2) natural zeolite can stand some turbidity in the water and the
bed is not so likely to become plugged by fine silt; and (3) the first cost of equipment, even
considering the larger units required. is generally in favor of the natural zeolite, It therefore has a
wide industrial application for boiler feed water ,whereas artificial zeolite has its field mostly in
domestic softening.
It is to be observed that water, after, passing through the zeolite bed, contains, besides sodium
sulfate and sodium chloride, sodium bicarbonate and carbonate, as mentioned above, in place of
the original calcium and magnesium bicarbonates and carbonates. Hence water containing high
bicarbonates and carbonates is often not suitable for zeolite treatment directly. For, if zeolite
treatment is used directly for water containing high bicarbonates and carbonates, the water may
become too alkaline, and sometimes boiler trouble is caused by this, if no correction is applied.
For this reason, in the case of very hard water, or water having high temporary hardness, it is often
preferred to give the water a preliminary treatment with soda and lime of lime and sodium
aluminate, before the zeolite treatment as mentioned above. This has the advantage of neutralizing
the acid carbonate in the temporary hardness with lime and throwing the calcium and magnesium
down as normal carbonates. For instance,

Ca (HCO3 ) 2 +Ca (OH) 2 2CaCO3 +2H2 O

Mg (HCO3 )+Ca (OH) 2 MgCO3 +CaCO3 +2H3 O
Thus, not only are the calcium and magnesium precipitated, but also the carbonate radical (CO3 --)
is eliminated by this treatment. A direct zeolite treatment would simply convert the alkaline-earth
bicarbonates to sodium bicarbonate, which remains dissolved in the resulting water, giving high
pH and requiring large quantities of acid for neutralization. This always results in high soluble
solids.
After the water has received a pretreatment with lime and soda, or lime and sodium aluminate,
it is settled off and filtered. After filtration the alkalinity in the water is neutralized with sulfuric
acid through. a feed proportioner with automatic control to keep the pH value between 7.4and 7.2
for the final zeolite treatment. The water is degasified in wooden or steel tanks; t is then sent
through zeolite softeners, and finally to the deaerator and thence to the boilers. This sulfuric acid
treatment before the zeolite treatment serves also to adjust the sulfate-to carbonate ratio so that no
treatment after zeolite softening is required, This combination keeps the soluble solids in the water
down to a minimum and at the same time removes the objection of leaving high sodium carbonate
and bicarbonate in the water after zeolite softening. A small amount of disodium phosphate is then
added to the boiler feed to maintain a suitable phosphate content (PO4 --) in the boiler saline.
For the removal of iron, a new zeolite is brought out by the National Aluminate Corporation.
This form of zeolite will remove iron and uses salt as the regenerating medium as usual The
mineral is black and of synthetic type It has a capacity of about 5830 gains per cu .ft, requiring
41 /2 lbs. of salt per cu ft. It weighs about 40 lbs. Per cu. ft It has little softening action, but will
remove iron, either in soluble iron salt form or as suspended iron. However, it will not remove
manganese.
To a limited extent, iron is also removed by the zeolite treatment, but to remove iron
completely, such as in the case where the water is used for caustic manufacture for the rayon
industry, a specially treated zeolite such as “Ferrosand” may be employed. “Ferrosand” is made by
treating the same natural Greensand with potassium permanganate forming a coating of MnO2 or
Mn3 O4 on the crystals. These manganese oxides oxidize iron in the water to the ferric state and
the Fe2 O3 by a thorough backwashing at 7-9 gals. Per sq. Ft. Per .min, treat the washed mass with
dilute KmnO4 from the bed. In actual operation, this is done in two stages by treating the water
first with Ferrosand to remove iron, then with the ordinary Greensand softeners for softening.
If the zeolite mass for some reason begins t disintegrate, a treatment with a dilute solution of
sodium silicate will garden the grain and stop disintegration. After the silicate treatment, the
zeolite should be thoroughly washed to get rid of the residual causticity. The base exchange
capacity then is temporarily greatly enhanced right after this treatment but will soon become
normal. The reaction underlying the silicate treatment is not clearly under stood but it seems that
Na2 O rather than SiO2 is absorbed by the zeolite mass, increasing the activity.
Lime-soda treatment will only reduce the hardness to a minimum of 2-21 /2 gt,per gal in the
cold, or 1-1 1 /2 gr. Per gal. in the hot process, with-out using a very large excess of the reagents in
each case; while caustic -soda and phosphate treatment may reduce it to 1 /5 gr .per gal .With the
zeolite -treatment ,the total hardness is reduced to 0.2 or less gr. Per gal, although zero hardness is
often spoken of in the trade parlance. For this reason, as seen above, the zeolite treatment is used
after the lime-soda treatment or lime-sodium aluminate treatment to remove the remaining
hardness.
After treatment according to one of the above methods, the water is checked for alkalinity
(pH value) For such purpose, a convenient type of the Hellige, LaMotte, or Taylor Comparator
may bi used, but electro-metric methods with the Leeds and Northrup pH potentiometer give more
reliable results. If the water is too alkaline (pH value is too high), sulfuric acid is cautiously added
to lower the pH value and also to bring up the sulfate content. For this purpose, a sulfuric acid
feed proportioner having an electric control, such as made by the American water Softener Co, or
the Permutit Co. or other company may be used. If, on the other hand, the water is not alkaline
enough (or pH value still too low), proper amount of caustic soda or soda ash is added to prevent
corrosion in pipe lines, economizers and boilers. If the sulfate-to carbonate ratio after pH
adjustment is still not sufficient, sodium sulfate or sodium bisulfite may be added to the treated
water in calculated quantities. Lastly, a small amount of trisodium or disodium phosphate is
generally introduced to take care of whatever residual hardness there is left in the feed water to the
boilers.
To prevent silica scale in the boilers, a certain phosphate concentration (30-50 p. p. m as PO4
--
) may bi maintained in the boiler saline. This phosphate constituent has a great inhibitive effect
against cars-tic embrttlement; but as it is reacted upon by the residual hardness in the water, it is
quickly depleted and must be replenished continually. For this reason, in considering inhibitants
for caustic embrittlement, only sodium sulfate is considered in figuring sulfate-t0carbonate ratio
according to the A.S.M.E. regulations, no allowance being made for the soluble phosphate
remaining in the boiler saline.
Silica present in the boiler feed water beyond a half grain per gallon, is considered
objectionable, because it forms a glassy, hard scale in the boiler which is particularly objectionable
in a high-pressure plant. The silica scale is a heat insulator and would cause the front bank of
tubes t o be burned out or to blister. Silica in water is a troublesome constituent to remove, as it is
rather inert and cannot be reacted upon readily by chemicals, producing insoluble sludge. One
common remove, as mentioned above, is to introduce a sufficient amount of disodium phosphate
into the boiler feed water .The phosphate then combines with Ca and Mg in the residual hardness,
thus avoiding the formation of calcium and magnesium silicates. If a special treatment for silica is
desired, the water may be given a pre-treatment with ferric sulfate solution and sulfuric acid to
give a pH reading somewhat below 4.0. Then lime is added to bring the pH to around 9.0. This
trivalent Fe+++ starting from a rather acid condition of pH=4.0 will bring down silica effectively
when the water is again made alkaline by lime with an alkalinity around 9.0.* But the simplest
way is to use the Akon treatment , or treatment with some form of activated magnesium oxide.
To a certain extent, silica in the water si removed in the hot lime-soda process. This is
probably due to the adsorption of silica by the magnesium hydroxide from magnesia introduced
with the lime. Magnesia is known to possess the greatest adsorption properties-more marked than
ferric or aluminum sulfate-and a method of the removal of silica from the boiler feed water has
 FIG 105 Relationship between removal of SiO2 and Sio2 in water.

been devised based on the use of magnesium.** Indeed, for hot removal of silica, Allis-Chalmers
Mfg. Co. also put out “Silimite” which is active lime especially burned from dolomitic limestone,
and W. H. L. D. Betz of Philadelphia, Pa., devised an adsorption process with their
“Remosil.” this treatment for silica with MgO is best done in the hot solution, say at 90 to 95 .
It can be combined to advantage with the hot treatment of lime and soda ash, so that hardness and
silica may be removed in one operation. Loose and light grade of MGO removes silica more
effectively. Also, a voluminous sludge and prolonged retention provide an intimate contact
resulting in a more complete removal.
* Schwartz, M. C., “The Removal of Silica from Water for Boiler Feed Purposes,” J. A. W. W. A., 30,659(1938).
Powell, Carpenter, Settar and Coates, “Recent Trends in Water Treatment, Chem. Met. Eng., 46,483 (1939).
** Tray, S. W., and Pankey, T. L., “Removal of Silica from Industrial Boiler Water Supplies ,” Combustion,
12,No. 5, p. 39 (1940)
+Betz, Noll and Maguire, “Removal of Silica from Water by Hot Process,” Ind. Eng. Chem., 32,1323 (1940)
Figs. 105 and 106 show results of treating silica with magnesia in the field tests made by Betz
Laboratories. The relationship shown in the two straight lines in Fig.105evidently indicates that
the action is one of adsorption* as it conforms to the Freundlich adsorption isotherm of the form
Y= KXn ; where Y=Removal of SiO 2 per unit wt. Of adsorbing substance.
X=Residual SiO 2 in water
K and n are constants.
Or, log Y=n log X + Const, an equation of a straight line on a logarithmic scale (see Fig.105)
Fig. 106 shows the decrease of silica in solution with the increase of magnesia by sludge
 FIG 106

Silica removal by MgO

with sludge recireulstion.

Cowtay Bets Laboration.

re-circulation.

In the operation of high pressure boilers (above 450 lb per sq in gauge) aluminum in the feed
water is objectionable due to the possibility of the formation of Analcite scale with silica on those
parts where there is high heat transfer, such as on the water wall tubes. Analicite scale is a double
silicate of sodium and aluminum (Na2 O Al2 O3 4SiO 2 2H2 O) and is formed even when there is no
hardness in the boiler, and when a residual phosphate of 40 p. p. m. is maintained. It is a tin, dense
scale, possessing high heat insulating properties. If the silica can be kept below 2 per cent of the
total solids in the boiler water and the alumina held to about 2 p .p .m., there is usually no trouble
experienced. But it is best to keep alumina out of feed water, as much as possible, for high steam
pressure operation.
* Whether it is by adsorption or chemical combination, the sludge formed contains silicates of calcium and
magnesium.
There are cases where alumina cannot be entirely kept out and where silica will run higher
than 2 per cent of the total solids. In such cases, the use of Akon (an electric are dispersion of
metallic iron in anhydrous ferrous oxide, which is put out by the Allis-Chalmers Mfg. Co.) will
generally remove the silica by adsorption and carry it down as a sludge. The silica can then be
kept within proper limits. With the use of Akon the residual phosphate required may be reduced to
about 5-10 p .p. m. Akon is in a liquid form and can be fed into the feed water at the intake to the
boiler feed pump. The usual dosage is about 3-5 p.p. m. Akon will also reduce foaming and
priming (which see) and small amount of oxygen in the feed water, thus helping prevent corrosion
in the economizer and boiler.
Residual aluminum in the water from the use of sodium aluminate or aluminum sulfate as a
coagulant is objectionable in the rayon-quality caustic manufacture and should be carefully tested
and removed. For this reason, coagulation by means of alum should not be carried out in an
alkaline condition (i.e., in high pH range).
Raw water often carries considerable mud, such as the muddy water from some rivers. If
such water is passed directly through the zeolite bed, the bed will soon be covered with a layer of
fine silt and eventually plugged. in such cases, a preliminary treatment of the water by coagulation,
sedimentation, and filtration is necessary before zeolite treatment, and this opportunity is afforded
by the soda-lime or lime-sodium aluminate pretreatment. However, the zeolie treatment will not
permit a hot pretreatment. The temperature of water for zeolite treatment should not exceed 100℉.
For natural zeolite and 90℉. For artiflcial zeolite; otherwise disintegration of the zeolite would
result. A brand of specially treated natural zeolite may, however, stand a temperature up to 150℉.
For coagulating the muddy water, aluminum sulfate, sodium aluminate, ferrous sulfate, or
lime may be used. Aluminum sulfate may be used in the cold but it reacts much more quickly in
hot water. In general, a hot treatment requires less reagents, gives more complete reaction, and
takes less time for settling; but the hot process is not suitable for the zeolite treatment as
mentioned above. Aluminum sulfate hydrolyzes in water and yields a flocculent precipitate of
aluminum hydroxide carrying the fine particles of mud down with it. This reduces the alkalinity
and leaves the water acid; neutralization with an alkali is necessary before the zeolite treatment.
Sodium aluminate has the advantage over aluminum sulfate because it hydrolyzes to sodium and
aluminum hydroxides, leaving the water alkaline. In conjunction with lime for removing
temporary hardness, sodium alumiante has been also extensively employed. Ferrous sulfate, upon
oxidation to the ferric state, acts like aluminum sulfate and has the advantage of being able to
remove dissolved oxygen and silica in the water. Ferric salts act like aluminum salts. Lime has
been found to possess some coagulating properties; but its action is much milder, and much large
quantities would be required. For cold treatment, it may take as much as four times the weight of
lime to coagulate the water as the alum. Lime, however, is cheap and is still the best reagent for
removing hardness where very hard water has to be treated. Whichever coagulant is used, the
water must then be settled and filtered.
Often the coagulating operation and the lime-soda treatment may be done in one step by
means of a large sedimentation basin. Excess lime is used in the treatment, together with other
regents, and the water is allowed to settle, recarbonated with CO2 gas and finally filtered. The
recarbonation is done to remove the excess of lime or to convert insoluble CaCO3 to soluble
Ca(HCO3 )2 so that it may not deposit itself in the pipelines or filters later on . This is called
“stabilizing” the water, in water works parlance. Very recently, Mr. C H. Spaulding* of Springfield,
I11., introduced a double-cone “precipitator” (Fig.107) based on the principle of “hindered

FIG 107 Spaulding precipitator.

settling,” using the sludge as its own filtering medium in a conical ring space, the vertical
cross-section of which is triangular, with the base line upward. The sludge particles carried up
by the up flow of water will distribute themselves evenly at a depth determined by the rate of up
flow. The mixing is done inside the central cone, so that water turns upwards around the base of
the central cone and flows upward in the sludge filtering space outside. This permits intimate
contact of the sludge with the stream of water, and allows time to complete the reaction so that
only a very small excess of lime is needed and no recarbonation is necessary. The sloping sides
of the sludge filtering space give a large and larger area on the top, which results in a
decreasing rate of upflow (velocity2” or so per min.), so as to give a maximum settling effect, as
the water reaches the surface and decants ino a collecting trough. This produces a “stable,”
clear effluent from the “precipitator” within a minimum detention period of only 45 minutes.
* Some New Practices in Water Softening, by Chas. H. Spaulding, Water Works and Sewage, 85 (March,
1938).
Filtering is generally done in sand filters. For the boiler feed in the ammonia soda plants,
many types of filters are on the market, notably the Cochrane filter, international Filter, etc. Such
filters may be divided into two main classes (1) Gravity (or open) filters, and (2) pressure (or
closed) filters. In point of construction they may be either vertical or horizontal, cylindrical or

FIG 108 Vertical cylindrical pressure filter.

rectangular vessels. Pressure filters are generally used in connection with a hot process of
softening, and are more desirable where water coming out of the filters has to be lifted has to be
lifted to a certain level. This eliminates another boosting pump. The rate of filtration varies from
gal per min per sp. Ft. of filter bed, depending upon the uniformity of the sand and gravel particles,
the condition of the filtering medium, and the thickness of bed. It pays to select the coarsest sand
available (mm. Size of sand is best) for the filtering medium; and then it is possible to filter at the
rate of 4 gal per min. per Sp ft filter area. One construction of a vertical cylindrical pressure filter
and one construction of square concrete open filter are shown in Figs.108 and 109. Open (gravity)
filters are, however, often preferred by some users because of simplicity in backwashing and in
general maintenance. As can be seen, the construction of the open filter is exactly the same as that
of the closed filter, minus cover over the top for gravity flow. Both pressure and gravity filters
must be washed frequently (once in every 8 or 12 hours) by backwashing with water at the rate of
about 9 gal per min per sp ft of filter area. If the sand is very coarse, it is possible and also
advisable to backwash at the rate as high as 20 gal .per min. The washing of gravity or open filters
may be aided by blowing compressed air into the filter or spraying the surface layer of sand by
means of high-pressure water sprays to break up the surface crust or “mud balls.” For the filtering

FIG 109 Open concrete filter.

medium, many substances have been used: calcite, magnetite, activated carbon,
anthracite(“Anthrafilt”), gravel, sand, etc; but in the hot softening process where hot water carries
considerable alkalinity, a silicious filtering mediuon, such as sand, should be avoided because of
the tendency of the alkaline hot water to attack the silicious material. This may increase silica
content in the water. In such cases, magnetite may be substituted.
Activated carbon has recently also been used as a filtering and purifying medium.
Incidentally, this also removes color and taste from the water. Its use for the filtration of water
after alum or aluminum sulfate coagulation is found also to reduce the residual content of
aluminum in the resulting water, which is most desirable for the manufacture of rayon-quality
caustic soda (see Chapter XIX).
The rate of percolation of water through the zeolite bed varies from 2 to 3 gal. Per min. Per
sq. ft. of bed area. The pressure drop through the bed 31/2’-4’ thick amounts to 2-5 lbs., depending
upon the condition of the bed and the volume of flow. The time of contact of water with the zeolite
mass is estimated at from 4 min. (for a rather soft water) to 10min. (for a hard water). When some
residual hardness in the effluent begins to show, it is time to regenerate the zeolite bed by means
of strong brine containing low calcium or magnesium. This is done generally once in 24 hours of
operation. For ordinary ziolite softeners using down-flow of water, an upward stream of washing
should precede the brine regeneration. An upflow of wash water of about 7 gal. Per sq. ft. of area
is generally sufficient. Care should be exercised nor to allow the upward stream of washing to
attain so high a velocity as to cause loss of zeolite carried over by the water. The whole
regeneration operation takes 30 to 45 min. Depending upon the physical characteristics of zeolite,
porous zeolite taking a somewhat longer time for regeneration. Regeneration takes 1 to 2 lbs. of
good salt (containing low Ca and Mg) per cu. ft. of zeolite, depending upon the base exchange
capacity of the zeolite in question. This corresponds to one-half pound of salt per 1000 grains
(kilograins) of hardness as CaCO3. The time of actual contact of the brine with the zeolite mass
varies from 10 to 20 minutes, depending upon whether the zeolite in question reacts on the surface
only or is porous and active throughout the mass. An average of about 15 min. Of contact is
sufficient to effect the reaction. As a rule, artificial zeolite is more porous than natural zeolite
(Greensand or Glauconite), and hence has a higher exchange capacity and takes a somewhat
longer time to regenerate. The bulk density of artificial (synthetic) zeolite is 50 to 55 lbs. Per cu ft.,
whereas that of the natural zeolite is 90 to 95 lbs. The heavier zeolite has less tendency to be
washed away during back-washing. Synthetic zeolite is made from sodium silicate and an
aluminum compound into a gel form, called zeolite gel, which is then dried to a hard granular
material. The ordinary zeolite softeners are arranged for downflow, i.e., the water to be softened
enters at the top, nows through the zeolite bed, and passes out at the bottom of the softener, the
gravel bed being located at the bottom to support the zeolite bed. Sometimes, where a high rate of
flow through the softener is desired, in order to avoid excessive pressure drop through the zeolite
mass, an upflow softener is employed in which the water is passed in through a manifold at the
bottom and flows out from the top of the softener, the construction of the softener remaining
essentially the same. This incidentally does away with the necessity of back washing of the zeolite
mass before each regeneration. The danger to be observed in this arrangement is the possible loss
of zeolite particles floating off with the upflow of water from the top of the softener, if the upflow
velocity through the softener should become excessive. To overcome such an objection, a
two-flow pair of softeners is installed, and water is passed first into the up flow unit, flows into the
top of the down-flow unit and out at the bottom. Thus, any zeolite particles floating out from the
up flow unit, Fig.110. Here the two units are connected in series, but the pressure drop through the
two units in series will by no means be small. Such are the details of the arrangement and design
of the zeolite softeners, and will not be further dealt with.
Some recent installations have used zeolite equipment operated entirely automatically by an
electric control mechanism. Some saving in salt and labor is claimed. A simple device has also
been worked out whereby spent brine from regeneration may be partially recovered and
strengthened for use.
 FIG 110 “Two-Flo” softener.

As a recapitulation, we may say that the zeolite treatment, while excellent in removing
practically all the hardness in the water for boiler feed, does not reduce the total solids and often
aggravates the alkalinity situation in the feed water. It is often not suitable or direct treatment
when water contains high sodium bicarbonate or when water is high in temporary hardness or
carbonate hardness, without lime pre-treatment to get rid of most of the temporary hardness. Then
also the zeolite treatment must be preceded by an acid treatment (usually sulfuric acid) to
neutralize excess lime and excess alkalinity, and followed by neutralization and aeration. But then
total solids in the feed water are increased considerably, especially when it is necessary to
maintain the sulfate-to-carbonate ratio according to the A.S.M.E regulation.
A rather recent development is the use of hydrogen zeolite, such as “Zeo-Karb H” of the
Permutit Company. This hydrogen zeolite is not really a zeolite in the chemical sense, but on
account of its similarity in the manner in which it is used, it is referred to as a “zeolite.” It is made
by sulfonating and granulating bituminous coal or other carbonaceous material in manner
somewhat similar to the preparation of activated carbon. Hence it is classified as carbonaceous
“zeolite.” It can be regenerated either with a mineral acid (sulfuric acid) or with salt; in the latter
case it performs exactly the same as the synthetic or natural zeolites. The Permutit Company refers
to the former process as “Zeo-Karb H” and the latter as “Zeo-Karb Na”. For generality, we shall
call the former the hydrogen exchanger and the latter the sodium exchanger as being more
descriptive of their properties. This carbonaceous granule is coarser than natural zeolite and
therefore can handle high rates of flow of water through the softener (6-7 gal per min per sq ft of
bed area with the same pressure drop). One remarkable property is that it can replace sodium,
when used as the sodium exchanger, its base exchange capacity is 5000-8000 grains of CaCO3 per
cu ft; which I almost as high as the capacity of a synthetic zeolite. It will take 2-4 lbs of salt per cu
ft of the volume of sodium exchanger. As the hydrogen exchanger, it will replace al metallic
cations, Ca++, Mg++, Fe++, Na+, etc by hydrogen, has a capacity of 8000-9000 grains per cu ft; and
will take lbs of 66 sulfuric acid per cu ft. If we let R represent the anion, M the cation and HZ
the hydrogen exchanger,
MR+HZ MZ+HR
When R=SO4 ,Cl , CO3 HCO3 or NO3 -, we have, after treatment, H2 SO4 , HCl H2 SO3 , or
-- - --, -

HNO3 respectively, in the water. For boiler feed water or for any industrial purposes, these free
mineral acids must be neutralized back by an alkali, when sodium ion (Na=) will be restored, and
there seems to be no gain in this respect over the ordinary zeolite treatment. But the advantage
consists in converting the bicarbonates and carbonates into free carbonic acid which may be easily
liberated from water by degasification with out any acid treatment,
H2 CO3 CO2 +H2 O
Whereas in the ordinary zeolite treatment, all calcium and magnesium bicarbonates and
carbonates would be converted into the corresponding sodium bicarbonate and carbonate which
can only be decomposed by adding a mineral acid such as H2 SO4 . In this way, in the ordinary
zeolite-treated water, sodium bicarbonate and carbonate would be converted by the acid into
sodium sulfate, remaining as a soluble solid in the feed water, and consequently the total solids are
increased.
*Applebaum, S. B., and Riley, R., “Zeo-Karb H-A New Method of Conditioning Water to Remove Sodium
Bicarbonate Chemically Instead of by Distillation,” Ind. Eng. Chem., 30,80 (1938), Also, Applebaum, S. B.,
“Applications of Carbonaceous Zeolites to Water Softening,” J. A. W. A., 30,947 (1938).
+Tiger, H. l., “Carbonaceous Zeolites-An Advance in Boiler Feed-Water Conditioning,” Trans. A.S. M. E.,
60,315 (1938).
In actual practice, these free mineral acids may be neutralized after the hydrogen exchange
by mixing the effluent from the hydrogen exchanger unit with the effluent from the ordinary
zeolite softeners in a proper proportion to remove the acidity or to obtain any desired alkalinity in
the mixture. Thus, total solids in the feed water are reduced and the alkalinity can be regulated to
any desired degree by the combination of the hydrogen exchange unit with a sodium exchanger
unit or with an ordinary zeolite softer. The hydrogen exchanger is especially good for certain
artesian well water containing high alkaline carbonates (sodium bicarbonate mostly) having low
permanent hardness; in which case a calculated amount of the raw water may be by-passed from
the hydrogen exchanger unit and mixed with the effluent from the unit to neutralize its free acidity.
Also, the hydrogen exchanger is very good for water containing high temporary hardness because
the carbonic acid formed in the resulting water may be driven out by degasification, with the result
that no alkali neutralization is needed. This carbonaceous mass, not being a silicious compound,
offers absolutely no danger of increasing the silica content of the water treated with it. However,
the fact that the hydrogen exchanger will replace sodium ions as well thus converting sodium
sulfate, sodium chloride or sodium nitrate in the water into corresponding sulfuric acid,
hydrochloric acid and nitric acid, has no significance, because such free mineral acids in the
treated water would have to be converted into sodium salts again before the water is fit for use in
the boilers.
Still another good combination is to use cold lime treatment, employing a moderate excess of
lime and soda ash, and then follow it with the hydrogen exchanger treatment, by-passing a proper
portion of the water for neutralizing the free acidity in the effluent from the hydrogen exchanger
unit. This gives low total solids and permissible amount of the remaining hardness for the boiler
feed, while the alkalinity can be regulated to any desired pH reading. This combination is
applicable to any kind of water where the non-carbonate hardness is high.
 Before the advent of the hydrogen exchange process, the reduction of the excess of the
bicarbonate (and carbonate) in the feed water was effected wholly by lime treatment before zeolite
treatment. Then, in order to bring the sulfate-to-carbonate ratio to the A. S. M .E. specification
after the zeolite treatment, sulfuric acid or sodium sulfate is added to the zeolite-treated water in
predetermined quantities. now with the advent of the hydrogen exchange process, it is possible to
maintain this A.S.M.E sulfate-to-carbonate ratio by eliminating the bicarbonates (and carbonates)
in the water instead of adding sodium sulfate or sulfuric acid, thus keeping the soluble salts in the
feed water at a minimum.
To summarize, the advantages of the hydrogen exchange or Zeo-Karb H process may be
stated as follows:
A. Removal of Na as well as Ca, Mg, Fe, etc.
B. Decomposition of sodium bicarbonate or carbonate y this treatment and subsequent
degasifying.
C. Reduction of total solids to a minimum.
D. Freedon from silica contamination when low-silica water is desired.
E. High exchange capacity (5-8 kilograins per cu ft) and high rate of water flow through the
bed at 6-7 gals per sq ft area.
Its application is of particular interest in boiler plants,
a. where distilled water would be required for the make-up, as in the case of
high-pressure and high-temperature steam generators.
b. Where water available is the artesian well water containing high alkaline bicarbonate
and carbonate.
c. Where combination of other process or bypassing of the raw water can be utilized
for the neutralization of the free mineral acids in the treated water.
d. Where the feed water available contains very high carbonate hardness or very high
temporary hardness.
Recently synthetic resins have been developed for application to water softening and are
found to possess good base-exchange (cation-exchange) properties. These base-exchange resins
are prepared from phenolic compounds or tannins, by condensation with an aldehyde
(formaldehyde, acetaldehyde, etc.) in the presence of concentrated hydrochloric acid, and are used
in the same way as the zeolites. They are then regenerated with salt (sodium chloride) or
hydrochloric acid. By the use of selective resins, practically al metal ions present in the water-not
only calcium, magnesium, iron, manganese, etc; but also sodium, potassium, ammonium,
aluminum, lithium, etc-may be removed and replaced with sodium or hydrogen ion, so that either
their sodium salts or their corresponding free acids are formed in the water.
Most interesting is the behavior of the commercial sulfited tannic extracts (such as quebracho
extract, catechu extract, chestnut extract, or hemlock cellulose sulfite extract) which, when
insolubilized by heating to 80 with 4-5 per cent of concentrated sulfuric acid, or with petroleum
acid sludge, and then washed and dried, yield an insoluble mass possessing even higher
base-exchange properties than the phenolic or tannin –aldehyde resins, and much higher
base-exchange capacity thanthe Greensand. Further, the base-exchange capacity is not lowered by
low pH (or acid ) condition, nor is the mass attacked by dilute mineral acids, in the same way as
the Greensand or the gel zeolite is attacked.
*Burrell, H., “Organolites (Organic Base-Exchange Materials),” Ind. Eng. Chem., 30,358-363 (1938).
 On analogy to the zeloite, Carleton Ellis termed those resinified organic bodies.
“organolites.” As water-softening agents, they possess many advantages over the ordinary zeolites
in that (1) they have a high base-exchange capacity, (2) they are less affected by acid or alkaline
conditions in the water, (3) they may be regenerated either with sodium chloride or with
hydrochloric acid, if desired, so that either the sodium chloride or with hydrochloric acid, if
desired, so that either the sodium salts or corresponding acids are formed in the water so treated,
and (4) they may stand hot water almost up to the boiling point with out disintegration. As
mentioned in connection with Zeo-Karb Na and Zeo-Karb H, a combination of salt-regenerated
resin treatment with acid-regenerated resin treatment may be employed; or a suitable portion of
raw water may be by-passed and mixed with a stream treated by the acid-regenerated resin for
bicarbonate removal and alkalinity control. This method is especially adapted to the treatment of
certain alkaline waters which contain excessive sodium bicarbonate and carbonate, or to a kind of
hard water whose hardness consists of essentially calcium or magnesium bicarbonate (practically
all temporary hardness). The bicarbonate or CO2 in the resulting water is simply removed by
aeration, as has been pointed out before.
On the other hand, certain aromatic amines (such as aniline, phenylene-diamines,
toluidines,etc) may be similarly resinified by an aldehyed and fount to possess aced-exchange
properties. They will remove acid radicals, such as SO4 --, Cl-, and NO3- etc; from water, and may
be regenerated by a solution of ammonia, sodium hydroxide, sodium carbonate or sodium
bicarbonate. By first removing the metal from the water with acie-regenerated resin as mentioned
above, and there moving the acid radicals in this way, it is possible, by repeated treatment, to
reduce the total soluble salt content in the resulting water to 1 ppm. The cost of getting pratically
distilled water from sea water by this method would be considerably below that of distillation. The
study was made in the Chemical Research Laboratory, Teddington, England, in 1934-6 and in
subsequent years, and has aroused considerable interest in the field of water purification.
It would be beyond the scope of this book to go into the many interesting details regarding
these synthetic resins except mentioning that the base-exchange property of the phenolic or
tannin-abldehyde resins or of tannn-acid sludge-insolubilized bodies seems to be due to the
presence of the hydroxyl group or sulfonic group. The size of particles of these organolites is
about the same as that of the ordinary zeolite (through 20, on 40 mesh)but they are lighter than
Greensand, the bulk density being about 45-50lbs. per cu ft; which is of the order of the bulk
\density of the synthetic zeolites (alumino-silicantes).
It may be further remarked that all the above non-siliceous “zeolites,” even when used only
as a base exchanger using salt regeneration may sometimes be preferable to Greensand or gel
because they do not cause increase in silica content in the water so treated. Silica, however, is not
removed by these resins an dmust be first removed by other means.
*Goudey, R. F., “Removal of Salts from Water, ” Journal A. W. W. A., 32,435(1940).
+Adams, B. A., and Holmesm E. L., “Absorption Properties of Synthetic Resins,” J. Soc. Chem. Ind. 54,1-6
T(1935).
EXCESSIVE CONCENTRATION OF SO L U B L E SOLIDS
The first effect of a high concentration of soluble matter on the boiler water is foaming and
priming; these may also be caused by high suspended matter as well as high soluble solids, and
sometimes also by the presence of oil in the water. Accumulation of soluble matter in the boilers is
brought about by high soluble matter carried in the feed water and by the high rate of evaporation
in the boilers, when the rate of blow down becomes inadequate. When the engines and turbines
run condensing, the situation is not so serious, but, when something like 100 per cent make-up
water is required daily as in the ammonia soda works, the remedy may be costly, if no other source
of water is available. For only through distillation by means of an evaporator could these soluble
solids of alkali salts be eliminated. One remedy is to install a continuous blow-down for the
boilers. This provides heat recovery for heating the feed water, using a flash tank and heat
exchanger for this purpose. This continuous blow-down system works very smoothly and keeps
steady load and pressure conditions in the boilers.
As mentioned above, there is a kind of artesian well which contains high soluble salts but
very little hardness in the form of Ca or Mg salts fortunately, this kind of water frequently contains
high alkaline carbonates (mostly as NaHCO3 ) and the modern hydrogen exchanger (see above) I
sable to eliminate these soluble alkaline carbonates with out neutralizing the water with an acid.
When the suspended matter accumulates so fast that blow-down becomes in adequate to
prevent accumulation, especially when some internal treatment is carried on, it is best to install a
sludge “De-concentrator” so that the boiler saline may be passed through the “De-concentrator”
the clear water returned to the boiler, and the concentrated sludge drained out at intervals. A
notable type of such equipment is that manufactured by the Elgin Softener Corporation of Elgin,
Ill; specially recommended where a certain amount of raw water has been by-passed and fed to the
boiler along with the treated water; thus some internal treatment is effected in the boiler. For this
purpose, a predetermined quantity of the boiler saline is circulated continuously through the
“De-concentrator” installed outside the boiler, where the mud and sludge are separated and the
clear water is returned to the boiler. Circulation is continuous and automatic and is done without
loss of heat. The quantity of water circulated is proportioned according to the amount of
suspended matter in the saline. Here this equipment will perform more advantageously than the
continuous blow-down.
Foaming and priming depend upon many factors. While the chief cause may be excessive
concentration of suspended and soluble solids, the phenomena also depend upon(1) the
construction of the steam drum, the method of circulation of the water in the boiler, and the
amount of steam space in the steam drum; (2) the presence of oily matter, organic matter or high
alkalinity; and (3) the non-uniformity(or irregularity)of steam drawing and furnace firing. The
American Railroad Chemists have rated the foaming qualities of water on the basis of
concentrations of soluble salts, as follows (expressed in parts per million):
Table 139. Foaming Grading of Raw Waters for Boiler Feed.
Soluble Salts
(ppm.) Grading
120 or less Excellent
120 to 260 Good
260 to 430 Fair
430 to 700 Bad
700 and up Very bad
This is merely an attempt of classify water as regards its foaming tendency, and is , of course,
subject to difference in individual experience and opinion. In general, with the ordinary water-tube
boilers operating at low or moderate pressures, when a concentration of 4000 ppm of total soluble
solids in the boiler saline is reached, foaming difficulties begin to show. If the concentration is
increased beyond this figure, water level appears unsteady and priming out of the stop-valve is
generally observed. If for some reasons the concentration of total solids reaches 9000 ppm
because of increase in the viscosity of the water in the boiler, serious local heating may occur and
tube rupture may result. For modern high-pressure boilers, this situation is greatly aggravated and
the tolerable limit of total solids in the boiler saline is much lower.
CORROSION AND PITTING
Corrosion and pitting in the boiler tubes and on the drum plate or rivet heads are caused by
dissolved oxygen in the feed water or by dissolved CO2 from NaHCO3 present. If the water has a
low pH value or if it contains some acidic salts, such as MgCl2 , CaCl2 , etc; serious corrosion will
occur, if the acidity is not removed. Consequently, modern steam boilers call for a feed water
carrying a distinct excess of alkalinity to the extent of 10-15 ppm with a pH value of 8.5-9.0 on
leaving the feed-water heater. Corrosion may be found in the steam space in the economizer, or
around the tube edges, or in the lower end of the tube opening, or on the rivet heads above the
water line. But sometimes it occurs on the drum plates under the water line, and may be observed
also at the lower half of the mud drum as light brown spots with numerous depressions on the
surface.
A zeolite-treated water is often found to possess marked corrosion properties. This is due to
its low or zero hardness and the absence of any coating formed inside the boiler. The zeolite
treatment being made in the cold, the water may contain oxygen to the extent of 7-8 cc per liter
and much of the dissolved CO2 in the form of sodium bicarbonate. Unless the feed water is
acidified, degasified, heated, and deaerated, in the neighborhood of 212 f; much of the oxygen
and CO2 will find their way to the economizer and to the boilers. Fortunately, nowadays a
combination.of feed water heater and deaerator is a standard equipment, and practically all the
oxygen may be driven out (to less than 0.03 cc O2 per liter) in the deaerator if operation is
carefully controlled. However, if corrosion still persists, it is necessary to add some sodium sulfite
(Na2 SO3 ) to the feed water after deaerating, and maintain its pH value at 8.6-9.0. the sodium
sulfite will reduce oxygen and form sodium sulfate. The Na2 SO3 solution may be added by means
of a small plunger pump into the suction of the feed water pump. it is very essential to add the
solution uniformly to the stream of the feed water on its way to the boilers. Fig.111 shows the
 FIG 111 Relationship between temperatures and oxygen dissolved in water.

relationship between the water temperatures in the deaerating heater and the oxygen content in
the boiler feed water under various pressure and vacuum conditions in the heater. On the same
figure are shown quantities of anhydrous sodium sulfite to be added to the water for the removal
of xoxygen (if desired) corresponding to the oxygen content in the water.
Sometimes corrosion in the boiler is attributable to the presence of fatty acids from the
compounded cylinder oil in the exhaust steam of the engine-driven CO2 compressors, air
compressors, pumps, etc. oil in general should not be present in the feed water, no matter whether
it is an animal oil or mineral oil. If it is present in the exhaust steam intended for boiler feed, the
steam should be led through an efficient oil separator and filter. If it is in the condensate, treatment
with aluminum sulfate and soda ash followed by decantation may be resorted to, to remove the oil
before the condensate is returned to the boiler. If steam turbines are used to drive generators, water
pumps, gas compressors, etc in an ammonia soda plant, no oil will be present in the exhaust steam.
It should be observed that most of the CO2 comes from sodium bicarbonate and sodium
carbonate present in the feed water which has not been properly acidified and degasified. In the
boilers all the sodium bicarbonate is decomposed to normal sodium carbonate:
2NaHCO3 Na2 CO3 +H2 O +CO2 at about 212 f
and, further, the normal carbonate then is decomposed or hydrolyzed to caustic soda.
Na2 CO3 +H2 O 2NaOH +CO 2
The latter reaction, or hydrolysis, occurs to the extent of about 30 per cent of the total sodium
carbonate present in low-pressure boilers, but may be as high as 80 per cent in higher-pressure
 FIG 112 Curve showing pH values vs. ratio of methyl orange alkalinity in feed water.

boilers. CO2 gas so liberated comes out with the steam and finds its way to the exhaust steam and
condensate. Such CO2 gas present in the steam causes corrosion also in the steam main and steam
piping. The steam condensate which is to be retuned to the boilers may then contain considerable
free CO2 and become corrosive. In ammonia soda plants where exhaust steam is used for heating,
distillation or evaporation, low CO2 in the exhaust steam an din the condensate for the boiler feed
is desirable. If the feed water to the boilers carries low alkalinity and has been properly degasified,
there will be little CO2 generated in the boilers, and hence low CO2 in the exhaust steam or in the
condensate. The presence of free CO2 in the condensate. may be determined by its pH value as
shown in Fig.112, which shows the relationship of pH readings to the ratios of the methyl orange
alkalinity (calculated as CaCO3 ) to free CO2 in the condensate.
CAUSTIC EMBRITTLEMENT
For more than three decades boiler failures have been recorded when feed water from some
deep wells carrying high alkalinity, or when zeolite treated water, or feed water overtreated with
lime and soda, was used in the boilers. Gracks occurred along riveted seams; upon examination
these were found to be inter-crystalline. It was found that in all boilers that failed this way, the
water in the boilers carried high caustic alkalinity and the concentration of sodium sulfate in the
boiler saline was low; and that, on the other hand, where a certain amount of sulfate existed, no
such trouble occurred. Prof. S. W. Parr and F.G. Straub of the University of Illinois conducted
many investigations and found that besides the sulfate (SO4 --) other acid radicals such as PO4 -- ,
CrO4 -- , and even CO3 — also had such inhibitive properties, the soluble PO4 --- being most effective
many times as effective as the SO4 -- . The A.S.M.E. recommends that the ratio of sodium sulfate to
the total alkalinity calculated as Na2CO3 be maintained in the boiler saline as follows:
Table 140 A.S.M.E. Recommendations of Sulfate-to-Carbonate Ratio.
Sodium Sulfate
Boiler Working Pressure Total Alkalinity as Na2CO3
Up to 150 lbs. Gauge 1
150 to 250 “ ” 2
250 to 600 “ ” 3
Above 600 lbs. Only distilled water from an evaporator should be used for the small make-up
required. Ordinarily, sodium sulfate is the inhibiting agent used, although nowadays phosphate in
 Na2 SO4 + Na2 CO3
≥ 2.5
NaOH
the form of tri- or disodium phosphate has also been used extensively. At or above 600 lbs.
Pressure, the addition of such an amount of sodium sulfate may not be tolerated, but tri- or
disodium phosphate can be used in much smaller quantities (40 ppm.) for the same purpose. The
phosphate, however, is likely to be depleted by uniting with the residual hardness in the water and
precipitated as insoluble phosphates of calcium, magnesium, or iron in the boiler. Prof. F. G.
Straub of the University of Illinois gives the following curves in terms of the ratios of
Na2CO3/NaOH as ordinates and Na2So4/NaOH as abscissas. Points lying to the right and above
the curves are considered safe for the respective pressures in question. If we considered sodium
carbonate left under composed in the water as also contributing the inhibiting properties, the ratio
of the sum of sodium sulfate and sodium carbonate to sodium hydroxide equivalent to 5/2:1 or
higher pressures by line MN, Fig. 113) would seem to be safe for all ordinary pressures, especially
when phosphate is also present in the boiler saline. Thus,

*“Treatment of Boiler Feedwater by Zeo-Karb Process” by J. D. Yoder, Combustion, May, 1939.

+ Embrittlement in Boilers, University of Illinois Bulletin No. 216 by Frederick G. Straub, p. 77 (modified
slightly from original).
Of course, if the boilers are electrically welded, as is at present carried out under Class I
Welding with X-raying and annealing, the danger of caustic embrittlement a priori will be
practically eliminated.
The following figures are given as a guide to the boiler operation under working pressures
between 300 and 600 lbs. Per sq. In. Gauge in ammonia soda plants:
TABLE 141.Conditions for Boiler Feed Water.
1.Alkalinity: 15 ppm as Na2 CO2
2.Total solids: less than 100 ppm
3.Dissolved oxygen: less than .03 cc. per liter
4.Sulfate-to-Carbonate ratio: 3:1(unless the drums are all welded)
5.pH from deaerator to boiler: 8.7(or 9.3 if an economizer is used)
6.Total hardness: less than 1.0 grain per gal.
7. Total silica: less than 6 ppm as SiO 2

8.Oil: nil

TABLE 142. Conditions of Boiler Saline (Internal Water in Boiler).

1.Causticity: less than 200 ppm as NaOH
2.Total solids: less than 1500 ppm
3.Sulfate-to-Carbonate ratio: 3:1(unless the drums are all welded)
4.Phosphate excess in solution (in boiler saline): 35-45 ppm as PO4 ---
5.Oil: nil
6.CO2 present in steam generated: less than 10 ppm
7.Total silica: less than 30 ppm as SiO 2

The pH value in the raw water is 7.0. That of the water coming from the zeolite tank may be
7.2-7.4. That in the feed water heater or deaerator is usually increased as the result of the loss of
CO2 by the decomposition of the bicarbonate and may be from 8.6 to 9.0 ,while that of the water
inside the boilers taken from the blow-off (boiler saline ) may be as high as 11.0 with
sulfate-to-carbonate ratio of 3 to 1 or higher.
TREATMENT OF COOLING WATER
Cooling water in ammonia soda plants, as a rule, does not receive any chemical treatment,
except in a very special case where the only source of cooling water available is too hard or too
corrosive for the apparatus and piping. Ordinarily, cooling water should be cold and clean or
clear .If it is turbid or muddy, it should be settled or filtered. Cooling water for ammonia soda
plants usually comes from deep artesian wells because of its low temperature and of its constancy
in temperature all the year around, summer and winter. The requirements for cooling water for
ammonia soda plants are:
(a) Low temperature
(b) Clearness and cleanness (freedom from excessive suspended matter or mud )
(c) Freedom from scale formation
(d) Freedom from corrosive properties
In addition to the low temperature, especially during summer, and freedom from the
excessive suspended matter or mud, the water must not form scale coating in the cooling pipes or
in the coolers or condensers; otherwise the rate of heat transfer would soon be greatly lowered.
Generally, if the cooling water contains much hardness or possesses scale-forming properties,
scale or incrustation will build up in the pipe sections at the exit end or on the cooling surface in
the hot portion of the apparatus. This not only cuts down effective cooling but may also constrict
passage of water through the cooler. If the hardness is mostly temporary, the trouble may be
avoided by treatment with lime followed by settling. If the cooling water contains high permanent
hardness, such that it forms hard scale in the pipes and apparatus, lime-soda treatment, as
described under boiler feed-water treatment, may be employed; but this is seldom found necessary.
If the cooling water contains much acidic salts and is therefore corrosive, lime treatment with or
without soda ash will remove the acidity and corrosive properties. If the cooling water in question
causes much algae growth that interferes with effective heat transfer, treatment with copper
sulphate , liquid chlorine, or liquid chlorine and ammonia (chloramination) will stop the algae
growth on the cooling surface. * Chloramination, i.e., chlorination after adding ammonia to the
water, causes chlorine to react with ammonia to form mono-or dichloramine according to the pH
value. The chloramine is more effective and stable than chlorine alone, although the action is
slower.
Sometimes without going through a formal treatment, formation of scale may be avoided by
adding to the water a small amount of sodium metaphosphate or sodium pyro-phosphate. The
phosphate seems to prevent effectively formation of the scale or incrustation, although the
quantity added is far short in proportion to the amount of calcium and magnesium present. A case
may be cited where cooling water containing 12-15 gr. Of hardness (mostly temporary) as
Ca(HCO3 )2 which formed a bad scale in the cooling pipes, was successfully corrected by adding to
the water only 5 ppm of sodium metaphosphate. The role the phosphate plays is therefore
interesting.
More remarkable is the effect of the so-called “hexametaphosphate” on the behavior of
calcium and magnesium bicarbonates in the water (temporary hardness). The
“hexametaphosphate” seems to act as a deflocculant or peptizing agent to prevent deposition of
calcium carbonate and magnesium carbonate precipitate. A very small quantity is sufficient to
withhold precipitation of CaCO3 and MgCO3 . Coagulation comes only after the water so treated
has been heated, or after long standing. The addition to the water of such a small quantity of
“hexametaphosphate” brings about a metastable state of supersaturation of calcium and
magnesium carbonates in the water, and is known as a “threshold treatment” for the water. It
appears that the function of the “hexametaphosphate” is to bring about a condition of
supersatuation and inhibit the deposition of calcium and magnesium carbonate precipitate by
adsorption on their nuclei of “hexametaphosphate,” which thus prevents their growth beyond
colloidal dimensions. For example, “hexametaphosphate,” when added in a quantity of only 2 ppm
to water containing temporary hardness as high as 600 ppm of Ca(HCO3 )2 , withholds separation
of calcium carbonate until the water is heated to below the boiling point.Hence this treatment is
very useful for water which has a tendency to form hard scale on cooling tubes or in cooling pipe
lines for multi-tubular coolers, condensers or heat exchangers in an ammonia soda plant, where
water is the cooling medium. It appears, however, that the effect of “hexametaphosphate” is more
on the calcium carbonate or magnesium carbonate, and not so much on magnesium hydroxide.
Further, the use of hexametaphosphate-treated water tends also to loosen and re-dissolve gradually
the scale already formed in the pipes. The “hexametaphosphate” seems to be adsorbed on the
surface of steel, copper or brass, so that its effect is carried over. Therefore, by virtue of the
hysteresis in its effect, this action persists even after the addition of the “hexametaphosphate” has
been temporarily discontinued, so that the result is effectively and uniformly maintained,
unaffected by any momentary fluctuations in the amounts of “hexametaphosphate” fed to the
water. These properties have made “hexametaphosphate” a valuable reagent for the treatment of
the scale-forming cooling water in an alkali plant, where the temperature of the exit cooling water
seldom exceeds 80 . This sesquestration of calcium and magnesium and other polyvalent ions
from cooling water (though more or less temporary) is sufficient to free the cooling water system
from scale formation and is very economical in its application because the quantity of the
“hexametaphosphate” needed is so small. But because of the metastable and temporary character,
this treatment would not be suitable for the treatment of hard water for boiler feed, except of
course only for the prevention of the scale formation in pipe lines leading to the boilers.
* Nason, H. K., “Chemical Methods in Slime and Algae Control,” J. A. .W. W. A., 30,449 (1938).
Hatch, G. B., and Rice, O., “Surface-active Properties of Hexametaphosphate,” Ind. Eng. Chem., 31,51 (1939).
Beneficial effects have also been obtained from the use of “hexametaphosphate” in the
prevention of the formation of “mud balls” or in the removal of crust deposits on sand filters, and
in the prevention of after precipitation in the pipe line after the lime-soda treatment.
 Chapter XXIV
Special Requirements of Ammonia Soda Industry
The ammonia soda process has been perfected by methods of gradual evolution. It has been
developed by the highest skill and keenest observation. The pioneers in the industry incurred great
expenses in working out the different steps by the introduction of many special features in the
arrangement and construction of the apparatus to accomplish certain purposes. Experiments have
been costly and experience has been painfully gained. Such little details as the U-loop seal, gas
venting, overflow loop provision, orifice restriction, gas -lift effect, syphon effect, liquor
suspension, gas locking, etc., are of the utmost importance in dealing with liquor and gas spaces in
the practical arrangement of the apparatus, especially in a closed system under partial vacuum.
Plants have undergone many changes as the operating conditions and the principles underlying
each step of the operation have become better understood. Very few of the present ammonia soda
plants remain as they were originally designed. These changes give some idea of the
improvements that have been constantly effected in ammonia soda plants, though little
information has become public. The difficulty of operation can be appreciated by realizing the
single fact that the process tolerates no interruption, whether caused by the stoppage of the
continuous flow of liquors or gas, or by the breaking down of any unit of the machinery. Although
there is always a spare unit for every important unit of the equipment in the plant, absolute
continuity of operation is not secured without the greatest effort. Even the liquor and gas piping
sections must be carried in duplicate to enable one branch to be cleaned while the other branch is
being used to carry on the process.
The rate of operation must be kept as uniform as possible. It is almost impossible to provide
sufficient storage capacity for either the strong liquor the filter liquor, or the crude bicarbonate, to
tide over any length of period of stoppage in any one of the several divisions. Such matters, after
all, can be provided only to a limited extent. Much has to depend upon continuous operation,
drawing from the discharge of one unit and feeding into the next, and so on. Any trouble in the
operation of one division is bound to be felt by the other divisions, and its effect in most cases is
cumulative.
A study of the historical development of the ammonia soda industry shows that there were
two stages in the operation of a plant. The first stage was to maintain the operation of the plant
without interruption, so that a steady operation from the plant day after day for 365 days in the
year resulted. In the second stage an effort was made to run the plant on the most economical basis,
getting the maximum efficiency and capacity possible out of each piece of apparatus and reducing
the losses and waste of raw materials to a minimum. Then, and only then, could an ammonia
soda plant be said to have been placed on a firm foundation. Most of the work of modifying and
altering the plant output as mentioned above took place during the first stage.
The material used for the construction of the apparatus is limited. The presence of ammonia
in the gases and liquors unfortunately has eliminated the use of any copper, brass or bronze
materials, which can be very easily worked and which have sufficient tensile strength for most
purposes. They form, furthermore, valuable materials in combination with steel or cast iron in
general commercial design. For copper and copper alloys are as a rule readily attacked by
ammonia, showing a deep blue coloration.
Steel and wrought iron, although they have excellent tensile strength, lead very short lives in
contact with ammonia gases or brine liquors containing ammonia and carbon dioxide, especially at
an elevated temperature. The life of a steel tube in contact with hot ammonia and carbon dioxide
gases or with hot ammoniated brine is form six months to a year, and that with cold ammoniated
brine is from one to two years. Much trouble has been caused by the failure of cooling tubes made
of standard boiler steel tubing. Only cast iron will satisfactorily withstand the corrosion.
Unmachined cast-iron surfaces having a hard crust stand the corrosion better than machined faces.
Special grades of cast iron, e.g., “Meehanite,” can be prepared which withstand corrosion better
than the ordinary grades. It is imperative that only cast iron construction be used, wherever
possible; an all-iron construction is a familiar specification for ammonia soda equipment.
Lead has excellent corrosion-resisting properties, but its softness limits its use to such matters
as bell and spigot joint calking, gaskets for unmachined faces, cast iron tubing expanding, and
some low-pressure work.
Corrosiron or Duriron (especially Durichlor) is resistant to the action of ammoniated brine or
ammonium chloride liquors, but is very brittle and not machinable. It can be ground only, and is
used in a limited number of places where the metal is not subject to mechanical or thermal strain.
High-chrome steels or chrome-nickel Stainless steels resist corrosion eminently well,
although toward hot ammoniated brine, they are not so good as the high-silicon iron (Durichlor).
The Stainless steels, such as 18-8 (meaning 18 per cent Cr and 8 per cent Ni on the average), 19-9,
and scores of similar kinds under various trade names, are resistant to many chemicals. Among
these, Type 304, containing low carbon, (.05-.06 per cent) are especially recommended, although
special grades such as KA2Smo, containing also 3-4 per cent Mo (Type 317), are more
corrosionresistant and cost considerably more. It is essential that these Stainless steels be
heat-treated (“annealed”) after all heating operations, such as electric welding, forging, touching
with a torch, excessive cold working, etc., in order to restore their austenitic state. For only in the
austenitic state does the alloy develop its best corrosion-resistant properties, any crystallization in
its grain structure or any segregation of carbon in the grain boundaries inviting attack by the action
of chemicals. The so-called low carbon Stainless steels may not have carbon content low enough
to be used without annealing after welding. Only the stabilized Stainless steels (such as Type 347
Cb-stabilized and Type 321 Ti-stabilized) may be welded without subsequent heat treatment. It is
advisable to avoid direct impact of ordinary steel heavy tools (such as sledge hammers) on the
Stainless steel plate or sheets during fabrication. The heat treatment consists in heating the alloy to
above its critical temperature (generally from 1850 to 2050î F) and then quickly quenching it in
water (water quenching) or cooling it rapidly in air (air cooling), in order to preserve its austenitic
state and allow no chance for crystals to separate. Whenever practicable, it is always
recommended to “anneal” a Stainless steel piece after its complete fabrication and before it leaves
 FIG 114 Huey tests on type 304 steel tube. Average rate of penetration in 48 hours over a period of 240 hours.

the shop. With extremely low carbon content, however, such as in Type 304, annealing may be
omitted where such annealing is impracticable for mechanical reasons. These Stainless steels,
while quite resistant to the action of nitric acid and fairly resistant to the action of cold, dilute
sulfuric acid, are not resistant to hydrochloric acid or hot ammonium chloride liquors. For plain
nitric acid, straight chrome steels containing 14-18 per cent Cr (Type 430) or higher, are
serviceable and cost less than the 18-8, but these high-chrome steels are usually harder to work
with, and are not to be welded. To improve machining properties, sometimes .02.04 percent Se or
S is added to these Stainless steels. For high-temperature work, high chrome-nickel Stainless
steels such as 25-20 (Type 310) are generally required because of their high heat-resistant
properties. For electric resistance and high temperature, a high chrome-nickel alloy, “Nichrome”
(Cr 15, Ni 60. Fe 25), has proved serviceable.
To help maintain the austenitic state of a Stainless steel working in the range of 800-1400î F.,
a “stabilizer” such as Cb or Ti is often added in small quantities (Cb 10 time carbon and Ti at least
4 times carbon). (See Types 347 and 321.)
Table 143 gives the composition of some of the commoner grades of Stainless steel (Type
304) at different temperatures on the corrosion resistance, as determined by the Huey Test.
 In using Stainless steel in combination with other metals immersed in a chemical liquor, it
should be borne in mind that generally speaking a Stainless steel behaves as a more
electro-negative metal (is corroded faster) than silver, copper, brass, lead, nickel, or graphite, and
as a more electro-positive metal (is less attacked) than ordinary iron, steel, aluminum, zinc, or
cadmium, when local battery action sets in. In the ammonia soda industry, Stainless steel is
commonly specified for the construction of the CO2 compressor valves, valve seats, piston rods,
etc. and for the construction of centrifugal pump rods, steam turbine blades, etc.
In all hot-working operations dealing with Stainless steels at a certain range of temperatures,
care must be taken to look out for grain growth, air hardening, intergranular corrosion, cold
brithleness, and other metallo-graphical changes due to temperature effect.
For hydrochloric acid or strong ammonium chloride liquors, Hastelloy B, among
nickel-molybdenum alloys, has stood corrosion well. Illium alloys are resistant to the action of hot
ammoniated brine or hot ammonium chloride liquors. They are resistant to all chemicals (acids
and alkalis alike) with the exception of hot hydrochloric acid or substances tending to liberate
chlorine, such as FeCl3 . These alloys are machinable and are suitable for the construction of small
parts, such as valves for strong liquor lines, although their cost limits their use.
Recently tantalum metal has been produced in quantities sufficiently large for commercial
application. It is very inert metal like platinum and is not attacked even by aqua regia, chlorine, or
chlorine products. It is an industrially valuable metal, combining mechanical strength and
chemical resistance. Like tungsten, it has also a high melting point. It must, however, be treated as
a semi-precious metal and its cost is far too high for general applications. It would be useful for
making lining and clad vessels that resist corrosion of all chemicals except hydrofluoric acid and
molten caustic soda (or hot strong caustic).
It is unfortunately true that solutions of chlorides, such as brine and ammonium chloride, hat
are met with in the alkali industry, are generally more corrosive than solutions of the
corresponding sulfates and nitrates.
Aluminum is slowly attacked by strong ammoniated brine and its use is confined only to
plain brine or weak sodium carbonate solutions.
Even cast iron is attacked by hot ammoniated brine or hot gases containing ammonia,
carbon dioxide, and steam, but its corrosion-rate is slower than that of steel. Actually there is no
metal or alloy of metals that is absolutely unattacked by this com bination of materials at
elevated temperatures; it is only a matter of the relative lengths of life and the cost of the
materials that determine their fitness. Except in the gas system, where, if there are any very
serviceable. They give tight service against corrosive liquors and their plugs are easier to turn.
Several makes, such as the "Merco Nordstrom," "Barco," etc., are available, but the ammonia
soda manufacturers usually make them themselves. The industry is unique in that cocks instead
of valves are extensively used throughout the plant. For simple hand Operation, cocks armade
up to 6 inches in size. From 8 inches up, such cocks require worm gear mounting for their
operation. To enable all cocks to be opened or closed easily, whenever desired, their plugs are
turned and lubricated regularly once a day. There is a special man in each division whose duty
it is to look after these cocks and valves.
Nickel tubes and indeed all-nickel construction have been used in the construction of caustic
evaporators and tanks for the manufacture of rayon-quality caustic.
A long list of different ferrous and non-ferrous alloys containing various percentages of
chromium nickel, molybdenum, cobalt, silicon, etc. that have been placed on the market under
various tradenames, could be given; But as their usein the ammonia soda industry is greatly
restricted by their cost, such a list would have only an academic value.
A very complete catalog covering eorrosion-resistant properties of metals and their alloys
with their compositions as fax as known, is pub-

TABLE 144 Metals Suitable for Different Liquors.

Liquor Excellent Fair Unsuitable
Strong brine and (1) Lead (1) Cast iron
calcium chloride (2) Nickel (2) Copper
solutions (3) Monel (cast) (3) Aluminum
(4) Steel
Sodium carbonate (1) Chroane-nickel (1) Cast iron
 solutions Stainless steels (2) Steel
(2) Nickel (3) Copper
(3) Monel (cast) (4) Aluminum
(4) Illium
Caustic liquors (1) Chrome-nickel- (1) Cast iron
(strong or weak) molybdenum (2) Cast steel
Stainless steels (3) Copper
(2) Nickel (4) Steel (low pressure work only)
low-pressure (3) Monel (cast)
(4) Illium
Molten caustic and (1) Nickel (1) Dense gray east
fused calcium iron
ohloride (2) Nickel cast iron

Ammoniated brine (1) Chrome-nickel- (1) Cast iron (1) Copper, brass
molybdenum and bronze
Stainless steels (2) Wrought iron
(2) Lead and steel tubes
(3) Illium (3) Aluminum
(4) Nickel

Mother liquor and (1) Chrome-nickel- (1) Cast iron (1) Copper, brass
filter liquor molybdenum and bronze
Stainless steels (2) Wroughtiron
(2) Lead and steel tubes
(3) Monel (cast)
(4) Nickel
(5) Illium

lished by Chemical and Metallurgical Engineering, entitled "Data Sheets on Materials of

Construction."
Table 144 gives some suggestions concerning materials in be employed for parts in contact
with different liquors in the soda industry. "Excellent" means "rather durable"; "fair" means
"commercially applicable"; and "unsuitable" means "chemically attacked."
The contents of Table 144 can be compared with an article by Mr. J. L. Ever hart, entitled
"Materials That Resist Alkali Corrosions." *
From Table 144 it will be seen that a good grade of cast iron or nickel cast iron is by far the
most commonly used material. It is used because of its good resisting quality and its cheapness.
For caustic liquors at high temperatures and pressures, steel vessels and wrought-iron tubes are
subject to caustic embrittlement. For ammoniated brine which contains magnesium salts that form
 FIG 115

Slip blank.

scale or coating on the internal walls, steel vessels and steel pipes are protected against corrosion
by the coating formed. Often a metal surface that has been completely submerged in the liquor is
less attacked than the portion in the gas space or partially exposed to the air, i.e., at the junction
between the liquor and gas phases.
Impellers of centrifugal pumps made of east iron are frequently pitted or eaten up by the hot
ammoniated brine or hot filter liquor. So also are the cast-iron cocks. As a rule, the cocks for this
service will leak and will not hold the liquor if proper maintenance is not effected. When no
leakage may be tolerated, as is the case when one portign of piping, is taken down for cleaning or
for a changeover, a "slip blank" must be used. This slip blank is made of 1/16-inch or 3/32-inch
steel plate cut with the diameter, which just fits the inside circle of the belt holes in the flange. It is
inserted between the flange joints with a rubber gasket on each side. These slip blanks are
provided with handles (Fig. 115) and are very useful for the gas system as well as for the liquor
system. Sometimes no valves are provided at all, the cutting-out of one branch being done by
means of these slip blanks. The changeover from one section of piping to another frequently
involves removing these lip blanks from one set of joints and inserting them in another. Piping
systems of various sizes arc usually flexible enough to enable the flange joints to be pried open
with a flat chisel and these slip blanks to be inserted between the flange faces.
* Chem. Met. Eng., 39,88 (1932).
All over the plant, wherever liquids, such as raw brine, milk of lime, filter liquor, etc, are
introduced into any apparatus, they are measured through orifice plates similar to the above slip
blanks but with a round hole of the proper diameter in the center, regulating the flow by the head
of the liquid above the orifice plate. The orifice diameter is to be marked on the handle. Where a
constant rate of feed is desired, the constant orifice head is further regulated by means of a float
control Probably in no other industry is the regulation of rates of different liquor feeds carried out
more systematically and with more minute precision than in the ammonia soda industry.
Ammoniated brine and other hot ammonia liquors are also very corrosive in their effect on
the packing material and pipe joints. Drippings from packing glands and pipe joints are annoying
and make the plant intolerable. Packings for centrifugal andplunger pumps must be of a
lead-graphite-asbestos composition or of a very soft metallic packing.To avoid leaking, these
stuffing boxes must be extra deep. Flange gaskets are cut from 1/10- to 3/32-inch pure rubber
sheets.
 Leakage is something that cannot, be tolerated in ammonia soda plants not only from the
viewpoint of keeping the apparatus and floors clean, but also of protecting the life of the
equipment. Any leakage of liquor due to a slight ooze from the pipe flanges or threaded jointsr or
from the overflow cover, manhole cover, or tapped hole through the cast iron wall will soon cause
the Wetted metal surface, even of cast iron, to become corroded and gradually eaten through. It is
not an unusual thing to see steel bolts in a leaky flange joint "frozen” in their holes or else rusted
so badly that threads have been destroyed, because ammonia brine liquors are especially corrosive
in contact with air. Leakages at the top flanges of cast-iron cocks, though a rather common sight,
are subject to the same objection. All leaks should be stopped at the very beginning of their
appearance: the white salt crystals appearing on the surface of the apparatus or hanging from pipe
flanges should receive prompt attention from the men in charge. If these leaks are not stopped at
their first appearance, they will be more difficult to remedy later. If cast-iron cocks leak at the
paoking glands or do not hold tight, the packing gland bolts should be immediately tightened. If
such leakage is allowed to go on for several days, the contact surface of the plug will have been
corroded by the liquor and the plug will no longer fit tightly, however firmly the bolts are
tightened later. Cock plugs should always be kept down. Even if the gland bolts have to be
loosened for turning the plug, they must be tightened back immediately after the plug has been
turned to position.
As it is necessary for frequent change-overs to be effected very quickly, there will be
occasions for opening up pipe flanges, manhole covers, etc. Since the atmosphere is very
corrosive in the works, all bolts used in pipes, images, manhole covers, etc., must be of exact
lengths flush with the top of the nut. It should be a practice with the men in the plant never to use
longer bolts with their threads projecting much beyond the nut. Inside the apparatus in contact
with the liquor, nothing but through bolts preferably in cored slots is allowed, as stud bolts or cap
screws in contact with liquor are unsuitable where opening or overhauling for cleaning is
necessary.
The ammonia soda industry is also unique in that large volumes of liquids and gases are
handled throughout the entire process. Fortunately, generally speaking, they are all under low
pressures or partial valuta, and it is interesting to note that wooden plugs requently suffice to stop
various holes provided for rodding, for observation, or for other purposes.
Because of various streams of fluid flow under low pressure or underpartial vacuum, the
plugging tendency of different passages must be constantly guarded against by the operators. This
plugging maybe caused by the mechanical deposition of solids carried in the liquor, by the
formation of hard crust around the internal walls of liquor pipes, by the crystallization of ammonia
compounds from ammonia and carbon dioxide gases at a low temperature, or by the chemical
reaction of the constituents in solution upon the internal walls of the pipes. The difficulty arising
from this cause will not be fully appreciated by those unfamiliar with the process. Although it is
impossible to enumerate all points where plugging is likely to occur, several salient places can be
mentioned here. In the kiln gas scrubber, the gas down-take at the entrance to the scrubber is likely
to be choked by dust collected from the kiln gases. In the distiller condensers, the furnace
condensers, and other gas coolers for ammonia and carbon dioxide, the gas passages may be
choked by the deposition of ammonium carbonate and ammonium carbamate crystals when the
temperature is reduced below a certain point. In the liquor system of the ammonia absorber and
settling vats the overflow passages and piping connections tend to become plugged by the soft
"mud" accumulated in these places and by the bard scale formed in contact with the ammoniated
brine. The presence of "mud" in the liquor, coupled with the fact that the flow is under a partial
vacuum, makes the flow of the brine very sluggish and the "suspension" of liquor due to gas
locking and mud restriction is likely to occur in the absorber system, if such matters as gas venting,
inclined gravity flow, freedom from downward loop (mud trapping), freedom from upward loop
(gas pocketing), etc., have net been carefully observed.
It may be remarked that, in general, a simple and direct piping should be arranged for the
liquor flow; that these liquor pipes should be inclined wherever possible toward the direction of
the flow to avoid any upward or h0rizontal sections, where "mud" may be accumulated, and to aid
the venting of gases out of the liquor stream toward a higher point; that the ammoniated brine
pipes should have ample cross section for the brine to flow, allowing for the ebnstriction by scale
formation; that the overflow opemngs should be sufficiently wide to provide ample width forthe
liquor and the "mud" to flow; and that allover flow shoulders or horizontal edges should be as
short as possible. Even in the piping system, which contains clarified ammoniated brine, the flow
may often be restricted by the formation of hard scale in the pipes.
That this apparently clear ammoniated brine gives rise to scale formation in the pipes is
proved by the fact that an 8-inch strong liquor pipeleft only a 4-inch opening after having been in
use for about 6 or 8 months. This scale consists mostly of magnesium carbonate, for it is very
difficult for magnesium mud to settle completely and traces may remain in colloidal suspension in
the "settled" ammoniated brine. The mother liquor or filter liquor, however, is not so likely to
form scales in the cocks and pipes as ammoniated brine (or strong liquor) and it is also not so
corrosive. In the distiller, the heavy liquor in the lime still tends to deposit, solids on the division
plates, under the "mushrooms," and at the overflow edges and overflow trough bottoms, causing
plugging. In the course of time a hard scale will be formed as thick as 2 to 3 inches. The period
of cleaning may vary from a couple of weeks to a couple of months depending upon the character
of the lime and upon the manner of operation. In the blow-off line, where the distiller waste (or
run-off-liquor) assent out by the natural distiller pressure without the aid of a booster pump, the
discharge may become very sluggish because of restriction in the pipe lines by the formation of
hard scale or by the deposition of solids. In the furnace gas main at the uptake, the soda dust that
iscarried in the gas frequently collects in the main and chokes the gas passage. Provision for a
water pipe for flushing with water and an opening for rod ding the pipe is quite necessary. In the
carbonating towers, sods sodium bicarbonate coating formed on the cooling tubes and mushrooms
may choke the gas and liquor passages. The difficulty is overcome normally by the so-called
"cleaning" with green liquor at a comparatively high temperature every 4 or 5 days. If the green
liquor carries any suspended magnesium mud, magnesium carbonate will be deposited in the
columns, especially at the upper part, forming a solid scale that cannot be taken care of by the
ordinary method of cleaning, and the columns will eventually be put out of commission. If, for
any reason and for any length of time, the carbon dioxide gas from the compressors delivered to
the towers is interrupted, a complete settling and plugging of the column by the bicarbonate
certainly results. This is one of the most serious cases of plugging that can occur in any apparatus.
When such an accident happens nothing but shutting down the tower in question and cooking the
content with water by means of steam can remedy the trouble. This does not or should not usually
happen. This cooking may cause leaks in tube sheets or in cooling box covers, and will result in a
low-grade and off-color ash when the column is started again. In ordinary eases, however,
shooting in steam through the draw opening and around the sides of the bottom ring for a few
minutes suffices to open up the column.
In the suction and delivery lines of the lime pump handling milk of lime for the distiller and
of the mud pump pumping mud from the settling vats-for both of these services only a
plunger-type pump fitted with ball valves is satisfactory-unless the work is attended with
intelligence and care, plugging of the pipes by solids settled therein is a source of trouble and
delay. Provisions in such lines for flushing with water, rodding out the solids, and cleaning tile
sections concerned, are made by means of tees and crosses with blind flanges. In the centrifugal
pumps for ammoniated liquors--which are most extensively used for pumping liquors in the works
the impeller ports are frequently restricted or even plugged by a hard scale formation and then the
pump fails to deliver the quantity of liquor required. Unlike the pumps for filtered liquors, these
strong liquor pumps have a very severe duty. To give an extreme case where plugging is least
expected, it has been the writer's actual experience that the 8-inch vacuum main at the gas' outlet
from an ammonia washer using fresh brine was plugged by the gradual deposition of solid sodium
chloride entrained in the gas bubbling out of the top compartment over a tall gooseneck.

It is good practice to use crosses and tees in such pipe lines provided with blind flanges so
that, when cleaning is necessary, these blind flanges can be opened (Figs. 116, 117, 118, and 119).
The cocks are set next to the crosses or tees (Fig. 118) to enable them to be ridded in case of
plugging or restriction in the opening. Often, a hand pusher (Fig. 120) is provided at the end of a
 FIG 119

Special inlet nossle.

pipe or opposite an opening to scrape off scale in the opening of a cock or of an overflow.
In the cases of gases containing ammonia and carbon dioxide there is always danger of
crystallization when the temperature gets too low for any length of time. This is due to the
formation of solids which consist of carbonates and carbamate of ammonia. The composition of
these crystals and the critical crystallization temperature depend upon the composition of the gas

FIG 120 Scale pusher for cock opening.

and the relative concentration of ammonia and carbon dioxide in it. At the tail end of the absorber
system where the gas has been more or less scrubbed by brine, crystals are sometimes formed in
the main when the temperature gets very low, and these crystals are found to have the composition
given in Table 145.
TABLE 145. Composition of Crystals Deposited at Tail End of
Absorber System.
Per Cent
Ammonium carbonate, (NH4 )2 CO3 47.61
Ammonium carbamate, NH4 . NH2 CO2 42.23
Ammonium bicarbonate, NH4 HCO3 8.80
Ammonium chloride, NH4 Cl 1.34
In the distiller condensers where the gas is rich in ammonia but not so rich in carbon dioxide,
the critic al crystallization temperature is generally around 55~ C., as has been mentioned in
connection with the absorber operation (see Chapter VII).
In the case of furnace condensers, however, since the gas is richer in carbon dioxide than in
ammonia, the crystals formed at a low temperature consist almost entirely of ammonium
bicarbonate. The analysis of crystals deposited in the furnace gas main is given in Table 146.
TABLe 146. Composition of Crystals in Furnace Gas Main.
Per Cent
Free ammonia NH3 20.81
Total CO2 53.57
Total Cl2 Trace
+
Fixed ammonia (NH4 ) Nil

Thus these crystals contain 96.70 per cent NH4 HCO3 .

In the event of plugging or constriction of the gas passage in a cooler or condenser due to
crystallization of solids when the temperature for some reason falls too low, a sure remedy is to
introduce some exhaust steam to warm up the apparatus, thus decomposing the solids and opening
the passage.
In the operation of ammonia soda plants there are the so-called cycles. The well-known ones
are the ammonia cycle and the carbon dioxide cycle. Ammonia gas is generated in the distiller,
absorbed by brine in the absorber, passed in the form of ammoniated brine through the
carbonating towers, drawn out in the form of mother liquor with 2/3 to 3/4 of the free ammonia
converted to fixed ammonia, and finally sent back as filter liquor to the distillers for the
regeneration of ammonia. This is the ammonia cycle. As for carbon dioxide, half of the carbon
dioxide in the bicarbonate drawn from the tower is recovered from calcination in the soda dryers
and returned to the towers either separately or mixed with the kiln gas. This is one branch of the
carbon dioxide cycle. Another branch is represented by Carbon dioxide dissolved in the mother
liquor as ammonium carbonate, sent to the distiller in the form of filter liquor, recovered with
ammonia from the distiller, absorbed by brine in the absorber as ammonium carbonate, and sent
back to the towers in the form of the partially carbonated ammoniated brine (containing 45 to 55
grams of carbon dioxide per liter). That part of the carbon dioxide existing in the gaseous phase in
the absorber system under partial vacuum is pumped through the exhanster as spent gases
containing as high as 60 to 70 per cent carbon dioxide by volume (the balance air) and is sent back
to the columns with the dryer gas through the carbon dioxide compressor intake.
Then again there is the balancing of operations in the different divisions of the works. If the
operation in different houses is not properly balanced, the whole plant cannot operate to the fullest
capacity, and also high operating efficiency or economy cannot be secured. A balance of
operations exists in each of the following divisions:
(1) Between the Tower House and the Dryer Room. The rate of furnace operation must
keep step with tower operation. A changed condition, in the furnace operation causes a change in
the tower operating conditions. If furnace operation is interrupted for any reason, the volume of
the dryer gas is diminished or the concentration of the gas is decreased, the decomposition and the
output from the towers are correspondingly decreased, and further the crystalline character of the
precipitated blear bonate is altered. This means that less bicarbonate is produced to feed the dryers
and also that poorer bicarbonate crystals are formed for the dryers. The result is a still poorer gas
returned from the dryers to the columns. Thus, the cause and the effect are interrelated and persist
in a cycle only to become worse as the cycle is repeated. The solution lies in keeping a certain
quantity of bicarbonate on the floor to meet these emergency conditions. An attempt should, of
course, be made to avoid any occurrence of such conditions.
 (2) Between the Tower House and the Distiller House. The rate of distillation must keep
step with the tower operation. If the towers are drawn faster, more ammoniated brine is required,
and more filter liquor is produced; then the distillation operations must necessarily be speeded up
to take care of the increased volume of the filter liquor and to meet the increased demand for
ammoniated brine. If the filter liquor is not sufficient to maintain the distiller operation for the
production of the ammoniated brine required, then crude liquor or ammonium sulfate solution
must be added to replenish the deficiency of ammonia in the system. The two houses must be kept
in balance and the best policy is to secure absolute unifomity in the rate of operation on each side.
(3) Between the Engine Room and the Distiller House. The power consumption in the
engine room must be so adjusted that the exhaust steam produced is just sufficient for distillation
purposes. A higher power consumption, and consequently steam consumption, not only puts an
extra load on the boilers and increases the coal bill, but causes a surplus of exhaust steam over
what can be used in the distillers. This excess steam may have to be let out to the air. It represents
a waste of coal, unless it can be condensed back as in the bleeder turbine arrangement. This
problem has been discussed in the chapter on "Generation of Power for Ammonia Soda Plants,"
and the reader is referred to that chapter.
(4) Between the Lime Kilns and the Distiller and Tower Houses. The production of lime
in the kiln for the distiller operation and of CO2 gases for the column operation must be so
adjusted that the rate of operation in this triangular relationship may be kept in good harmony.
Usually there is an excess of the kiln gas to be allowed to escape into the atmosphere. The
determining factor is the rate of lime consumption in the distillers. Where there are simultaneous
demands for lime for caustic soda manufacture and for kiln gas for refined sodium bicarbonate
manufacture, the operation of the limekiln then may not be under exclusive control of the distiller
operation.
Consideration of the various phases shows that an absolutely uniform and constant rate of
operation in each and every division of the works is a sine qua non for ease of control, for
regularity of output, and for high efficiency and economy in manufacture.
To maintain continuous operation and to provide for failure in power transmission and
lighting lines, it is generally recommended to have more than one feeder circuit over which power
may be obtained. Since several of the drives in an ammonia soda plant require the utmost
continuity of service, a number of ways are adopted for obtaining it. A failure of power in the
electric motors at the rotary calciners (for instance) might cause an extremely expensive shutdown
due to the high temperature at which these shells are operated. Occasionally a gasoline engine is
installed and serviced for continual stand-by so that the shells of rotary dryers can be operated
almost immediately, should all sources of electric power supply fail. A source of purchased
electricity available for stand-by or breakdown service is often justified. Generally, however, to
avoid damage to the dryer, hand-operated mechanism is provided (in case of emergency) for such
purposes.
As the operation in the ammonia soda plant is strictly continuous and any accidental loss due
to personal inattention on the part of the operators may run to large figures, it is always a good
policy to provide recording and integrating meters at various points in the process. Such
installations will soon pay for themselves. These include recording thermometers or pyrometers,
recording pressure or vacuum gauges, recording flow meters for steam and water, recording CO2
meters in the flue gases, and such other recorders as the Esterline-Angus recorder for ammonia at
the bottom of the ammonia distiller, etc. The electrically operated type is often to be preferred.
As in all industrial plants, an iron cement, of which many commercial brands are available, is
very useful and handy in making repairs due to cracking or corrosion in cast iron or steel body,
and in making tight joints that do net have to be taken apart in the future.
Consideration Of special requirements will not be complete without a
discussion of surface protection for materials of construction in an ammo-
nia soda plant. The paint materials generally have the following bases:
(1) Oxidizing vegetable oils (linseed oil base)
(2) Phenol formaldehyde resinoid (Bakelite base)
(3) Asphaltic material
(4) Chlorinated rubber
and contain the following principal pigments:
(1) Iron oxide
(2) Basic lead oxide
(3) Titanium oxide
(4) Aluminum and aluminum alloy metal powders
(5) Graphite
A strong alkali is the active ingredient of one important class of "paint removers." It is
therefore to be expected that the surface protection problem in an alkali plant is difficult and not
entirely solved.
The service in the ammonia soda plant can best be discussed under separate headings of
"dry" and "wet." The dry side includes the interiors of those divisions for calcining the crude
bicarbonate, etc., and all other places where there are seldom occasions for steamy conditions or
condensations of vapor to form alkali solutions on the wall surface. The wet side covers those
divisions dealing with ammonlated brine, and CO2 and ammonia fumes, and has either frequent
condensations or occasional spillages and sprays of alkaline or ammoniasal liquors. In both
services, the air contains alkali dusts, CO2 gas and ammonlacal fumes.
Reinforced concrete columns and beams deteriorate in the presence of these fumes, if the
concreting job was not done properly and the concrete mass is porous. The steel bars inside the
reinforced concrete columns and beams have been known to corrode under these circumstances.
On structural steel around a high ambient temperature and near gas vent from an apparatus
which operates at a temperature of 60 to 90 C., a reasonably good and moderately permanent
protection against alkali dust and brine spray is obtained by spraying the surface with heavy fuel
oil containing a measure of rosin oil and Portland cement. This is an especially inexpensive
coating. On cooler surfaces it remains objectionably sticky, whereas on the warmer surfaces it
dries eventually to a sort of varnish-like surface.
In the dry parts of the plant the linseed oil-lead pigment paints have an appreciably shorter
life than the resinoid-titanium pigment paints, although the latter have a greater chalking tendency
than the former. The resinoid-aluminum pigment paints have preferred applications on warmer
surfaces (up to 250~ .). They also do reasonably well on surfaces occasionally subjected to
ammonlacal fumes. The iron oxide paints in linseed oil have the shortest life of those listed.
Where soda ash dust becomes damp and wet through humidity changes, the life of any paint is
short, the most severe service is near the outlets of the brine-scrubbed carbonating tower.
It appears that in the crude bicarbonate and soda ash divisions the asphalted base paints with
essentially graphitic constituents have the longest life. Their black color, however. makes them
undesirable in many locations and applications. One good grade is the so-called "Ebonol" made
by Sherwin Williams Company, and is applied by means of a spray gun. The air pressure on the
gun is 35 to 40 lbs. gauge and that Used on the tank is 15 lbs. Fifty feet or more of a 7/16-ineh
rubber hose is used for the air line to the tank; and 25 feet of a 5/16-inch hose are used for air and
25 feet of a 1/2-inch hose for paint, from the tank to the spray gun. When a light color is desired, a
thin coat of aluminum paint may be applied on the surface of "Ebonol."
In the caustic soda plant, corrosion is not as severe as in the soda plant. Unpainted structural
steel rusts more slowly, although wood and timber may suffer severe deterioration. For resistance
against the attack of caustic soda solutions in concentrations up to 50 or even 70 per cent, the
rubber base paints have been found to be the most suitable. The Hercules Powder Company's
chlorinated rubber paint, manufactured under one of the German patents and bearing the trade
name "Tornesit," and the Goodyear series of "Pliolites" are in use. Lately, for same service,
ethylcellulose has been patented by the Pittsburgh Plate Glass Company. These material]have been
used for lining stationary tanks as well as tank-cars, and also for painting structural steel, such as
columns, girders, beams, etc. In the caustic plant itself, neither linseed-oil base nor bakelite base
paints with lead or titanium pigments have been found satisfactory because of the caustic
atmospbere. The asphaltic paints de better and are cheaper.
In electrolytic and chlorine products plants, the corrosion is generally more severe than in
others. The same generalities apply, but life of coatings will be found to be shorter. Chlorine
attacks metal surfaces. Brick or concrete construction is preferable. In the caustic fusion room,
no paint will resist attack by caustic fumes and last for any length of time.
In the electrolytic sodium chlorate plant, no wooden structure is tolerated because of the
inflammable or explosive character of wood or any combustible matter when soaked and dried
with sodium chlorate liquor. Men around the plant must wear rubber garments and rubber boots
toavoid fire hazard caused by sodium chlorate.
The most logical attack on the alkali plant paint problem lies in the removal or mitigation of
the source of trouble. Exhaust vent stacks can be relocated or extended. Overflow ducts, splash
guards, ete., can be studied critically and provided for in order to eliminate causes of accidental
spillage, formation of spray, sources of fumes, etc.
"Guniting" has been found very useful for protecting the surfaces with cement where a rather
heavy coating is desired. For this purpose three bags of coarse sand are taken and mixed with one
bag of cement, using three gallons of water. Sand for "Gunite' should not be mixed with cement
more than an hour before application. The mixture must be of right consistency: if too wet, the
coating would be thin and porous, and if too dry, it would develop cracks and voids. The nozzle
of the gun should be held 3 or 4 feet from the surface and the air pressure should not be excessive.
Two coats are generally applied, the first coating being about 1/4 inch thick, while the second
coating may be as much as 3/4 inch thick. Workmen should be provided with goggles to protect
their eyes and with respirators to guard against inhaling silica dust.
 Chapter XXV

Control in Ammonia Soda Process

Chemical control in a process differs from chemical analysis in that, owing to lack of time,
only such tests or analyses as can be rapidly and readily carried out by the operatives in the field
are made and the results are awaited to guide the operation immediately. Less accurate but more
rapid methods must often be adopted, and tests, which are indications though not strictly accurate,
are used because of their convenience in exertion. More accurate and complete analyses are
usually relegated to the chemical laboratory to be accomplished by regular analysts. The control
tests should be made at least hourly and, where necessary, as for the excess lime at the distiller
bottom, and for NH3 and Cl- titers in the absorber bottom, the tests are made at intervals of a few
minutes. Unless explicitly mentioned, such tests are made hourly. Results, however, are always
entered hourly in the works log sheets.
In addition to chemical tests it is essential to regulate the temperature and pressure (and
vacuum) conditions in the apparatus and at various points in the system. Where orifice regulation
is used, the diameter of the orifice and the head of the liquor above the orifice must be properly
adjusted.
Brine department. It is first necessary to ascertain if the brine is saturated. Strictly speaking,
we are interested in the concentration of sodium chloride or other sodium salts present in the brine,
i.e., of sodium only, but as no rapid method is available for the determination of sodium, the
chlorine determination by Mohr's method is taken as representing sodium because of its rapidity.
This is not accurate inasmuch as not all Na is present as NaCl and not all Cl is combined with Na.
If the composition of brine runs approximately constant, however, a factor can be found to
determine the concentration of Na from that of Cl found. Usually the specific gravity of the brine
is taken. But if the brine contains many impurities, the specific gravity of the brine runs higher
than that which corresponds to the amount of sodium chloride present. Saturated sea brine gives a
specific gravity of 1.205 to 1.215 at 15~ ., which is considerably higher than that of the pure
saturated brine at this temperature. It is necessary to have the brine fully saturated. Its Cl- titer
should run 107 to 108.
When raw brine is to undergo chemical treatment before it is sent into the system, close
control must be exercised as to the regularity and quantities of the reagents added. The
temperature and the pH value of the brine during treatment must be carefully adjusted and
recorded The purging of the mud from the settling system, or mud pumping, should be done
regularly and scrupulously so that excessive quantities of brine may not be wasted.
Lime department. In the limekiln the kiln gas should be sampled at least hourly and tests for
the percentage of carbon dioxide, oxygen and carbon monoxide made with an Orsat. There
should be practically no. carbon monoxide, but there may be a small fraction of a cc. of oxygen.
For efficient operation the temperatures of the kiln exit gas from the top and that of the lime drawn
from the bottom should both be low. Thermometers should be provided in the top gas down-take
and in the lime draw. The temperature of the exit gas should be around 80~ . and that of the lime
50~ . The milk of lime going to the distiller should be kept hot (around 80~ .) and should be as
concentrated as the pipelines and pumps can handle it. The strength of the milk of lime should be
carefully watched. Ten cc. should be pipetted and diluted to 100 cc. and 10 cc. taken for analysis
by titrating with N HCl using phenolphthalein as an indicator. The CaO titer should be around
180. which corresponds to about 250 grams CaO per liter. Too weak (dilute) a milk of lime
unnecessarily increases the volume of the liquor to be distilled and causes high consumption of
steam for distillation and loss of ammonia in the distiller waste. Unless tile limestone in very poor
and the lime improperly burned, such strength of milk of lime (250 grams of CaO per liter) should
not give much trouble. When the limestone is of a very poor quality, such a high concentration of
milk of lime may be difficult to handle without plugging up the pipelines. To ascertain the strength
of the milk of lime the specific gravity can be taken with a hydrometer or, better, by weighing the
milk of lime in a bottle of a 100 cc. volume and known tare. The hot lime liquor of good strength
should have a specific gravity of about 1.20. Attention should be called to the fact that the
specific gravity reading gives only a very rough indication of the strength of the milk of lime, as
the reading is affected by the amount of solid impurities in the lime. The pressure of the blast
entering the bottom of the klin is taken by a water manometer and the blast adjusted to the rate of
operation. For a kiln say 70-80 feet high, the blast varies from 4 to 8 inches of water, depending
upon the rate of burning and the size of the stone charged. A water manometer is provided in the
gas outlet pipe to measure the pressure condition in the kiln top. Where the kiln cover and casing
are reasonably airtight, the kiln should be operated under a very slight vacuum to aid the
decomposition of limestone. Usually, however, for fear Of leakage of air inward, a slight positive
pressure instead of vacuum is employed. The ratio of the stone to coke is important. Both the
stone and the coke in the charge should be weighed and recorded. Too much coke gives a low gas
test and causes overburning of lime, while too little coke yields excessive underburned lime, i.e.,
lime with a stone core. The different sizes of the stone are best sorted and burned separately. The
smaller size of stone takes less coke than the larger size and less time to burn. The difference of a
single pound of coke in a carload of charge (1 ton of stone) makes a noticeable difference in the
results of the kiln operation. Sizes from 6 to 2 inches, however, may be burned together. The
proportioning of coke to stone may be done in a multiple bopper feedtable.
In the kiln gas main from the kiln down-take, through the scrubber and the cleaner, and
finally, all the way to the CO2 compressor intake, at a half dozen or more points the line must be
provided with U-tube manometers, each to indic ate the pressure or suction conditions at the
respective places. At the kiln end, the down-take and the inlet the scrubber are likely to be plugged
by the fine dust in the gas carried over from the kiln; while between the scrubber and cleaner
outlet and the compresser intake, the pipe is sometimes flooded at low points by the condensate or
water entrained in the gas from the scrubber and cleaner. This restriction to the gas passage causes
a high vacuum ahead of the obstruction and affects the distribution ratio between the kiln gases on
the one hand and the furnace gas on the other, when the compressor is to send mixed gases to the
columns from the two sources. The manometer indications help locate the place of restriction.
Generally normal vacuum (or pressure) readings are posted at the respective spots so that when
the manometer reading at any particular spot exceeds the reading given, it indicates to the
operators that the section beyond that spot needs cleaning or perhaps the condensate there needs
more thorough draining. A systematic charting of the suction or pressure conditions throughout
the whole system is worked out for the guidance of the operators so that internal obstruction at any
point of the system can be readily spotted out and the necessary cleaning or draining applied
immediately. All this should be attended to before the trouble begins to be felt in plant production.
Absorber and vat house. Tower washers and filter washers are placed trader the charge of
the absorber rather than the tower men, as the brine used for scrubbing is all lobe sent to the
absorber for final amrnoniation. All brine inlets to this apparatus should have orifice and float
regulations in order to adjust the rate of flow accurately. The head above the orifice plate for a
given orifice diameter should be recorded. In these washers, the temperature at the bottom
compartment should be takes as well as that of the fresh brine at the top inlet, as the absorption of
ammonia by brine is largely dependent upon the temperature of the brine. Sampiee of brine from
the top compartments of these washers should show no ammonia tiller. All the scrubbing brine
ultimately flows by gravity through the absorber washer to the absorber. From the bottom
compartment of the absorber, samples of ammoniated brine should be taken every 5 minutes or
oftener to determine the NH3 and Cl- titers. These two titrations can be profitably done in one
sample by first titrating for free ammonia with N H2 SO4,using methyl orange as the indicator and
then adding a little calcium carbonate powder after diluting the solution, titrating by Mohr's
method for chlorine with N AgNO3 , using polonium chromate as the indicator. The flow of brine
to the absorber is adjusted according to the rate of distillation. The Cl- tiler must be maintained
as high as possible but cannot be mush above 260 grams of sodium chlotide per liter. The NH3
titer is maintained at a fixed ratio to the Cl-(about 1.10:1).
Thermometers should be provided at the absorber bottom comportment, in the liquor line
entering the absorber cooler, in the liquor line returning from the absorber cooler, and finally at the
gas outlet from the absorber top. These temperatures are import an as ammonia absorption is a
function of temperature. When the body of the absorber is hot to the top, known as "hot top,"
because of either insufficient cooling surface in the absorber cooler, excessive steam carried over
by too hot ammonia gas to the absorber, or too little or too warm brine being used in the absorber
ammonia will not be readily absorbed in the liquor but will pass out from the top and may even be
drawn finally to the vacuum pump. The temperature at the absorber system, therefore, is very
important and a recording thermometer should be installed to show the condition of the absorber
exit during the 24 hours. This top temperature of the absorber washer is normally 35 to 40~ .
The absorber liquor outlet to the vats should be provided with a thermometer and this temperature
is from 60 to 65~ ., a temperature quite suitable to give good settling of the ammoniated brine in
the vats. Through these settling vats, by atmaospheric cooling, the temperature of the ammoniated
brine is reduced somewhat, until at the inlet to the brine coolers or vat coolers it is about 50~ .
The brine coolers (or vat coolers) should bring the temperature down to about 30~ . At these
places thermometers are provided to observe the temperatures mentioned. The temperature of the
cooling water at the inlet and at the outlet of these coolers should also be observed.
From the bottoms of the settling vats the mud should be pumped by a plunger type pump in
rotation. The number of minutes for mud pumping for any particular vat to reduce the mud content
therein to, say 25 per cent (by volume) should be noted. This establishes a time schedule for mud
pumping. In the absorber system mercury manometers are provided to register the vacuum
conditions at different points and to detect any gas -locking (liquor suspension) or plugging
tendency in any part of the system. The following points in the vacuum system should be provided
with manometers:--Absocher exhauster, top of absorber washer, fop of absorber, outlet of
&tiller condenser, and top of distiller. These vacuum conditions vary as the depths of wash
(harborage) in the different parts of the apparatus. If the apparatus and pipe lines are clean and
show no tendency to plug, the total vaenm for the whole system registered at the exhauster end
should be about 8 to 12 inches of Hg.
 When the brine has been pretreated, operation in the distiller and vat house becomes very
simple. The plugging of the apparatus and piping by mud will have conmpletely disappeared and
there is very little need for changing over to the standby unit for cleaning. Operation is less
likely to be interrupted, and the capacity of the system in terms of soda ash output can be greatly
increased.
Tower house. The main units in the tower house are the carbonating columns and the filters,
and the control tests as made in the field are outlined below. In the green liquor line to the columns,
a thermometer pocket should be provided to receive a thermometer for measuring the temperature
of the green liquor to the towers. This temperature should be not more than 30~ , a higher
temperature eansing excessive valatilization of ammonia from the top of the towers and putting
heavier duty on the tower washers. The green liquor is tested for NH3 and Cl- titers. Their
respective concentrations and relative titers should have the values given in Chapter VII, i.e., 98 to
99 titers for NH3 and 89 to 90 titers for Cl-. These titers should not differ by more than 1 or 2 units
from day to day, as with a large capacity in the vats there should not be any difficulty in
maintaining a constant strength of the ammoniated brine. The percentage of carbon dioxide in the
mixed gases to the towers should be analyzed with an Orsat. According to the efficiency of the
plant, this carbon dioxide percentage in the mixed gases will vary from 53 to 60 per cent.
Similarly the waste gases from the top of the towers should be tested with an Orsat. For the
making towers, the percentage Of carbon dioxide should be from 3 to 5 per cent by volume, while
for a cleaning tower it should be about 1 per cent. it is evident that the lower the percentage of
carbon dioxide in the waste gases, the better the tower absorption efficiency for carbon dioxide,
and the smaller the loss of carbon dioxide. For a cleaning tower the outlet liquor is tested for
carbon dioxide content by means of a simple evolution method (Fig. 124, Chapter XXVIII). There
should be 61 to 66 cc. per 2 cc. of the liquor corresponding to 60 to 65 grams of carbon dioxide
per liter.
Temperature regulation in the columns is vital to their successful operation. The point about
two-thirds of the height of the tower should be provided with a thermometer pocket to ascertain
the reaction temperature in the towers. This reaction point happens to be near the tower working
(or draw) floor or one floor above so that the reading can be readily taken there. It is above the
cooling section of the tower and should have a temperature of 55~ to 60~ . But unless there is
full decomposition reaction in the towers, such a temperature cannot be developed. And unless the
temperature is high here, good crystals of bicarbonate cannot be obtained at the bottom.
Consequently, while it is necessary to get full cooling effect at the bottom of the tower, a part of
the cooling water must be "bled out" from the middle of the coaling sections to permit the top
portion of the cooling cretins of the column to remain, at a comparatively high temperature. A
continuous temperature gradient in the towers is thus maintained, whereupon the bicarbonate
crystals may build up as the liquor is cooled. In a column having an ample cooling surface and a
long path of travel for the cooling water, the top cooling water outlet temperature should be
around 45~ . Consequently, the cooling water outlet temperatures as well as the temperatures at
the two-thirds point of the tower and at the draw must be closely observed. One thermometer each
is provided in the top and in the middle (bleeding) cooling water outlet points to observe the
temperature of the cooling water and adjust the amount of bleeding.
The tower draw temperature should be frequently tested at the draw funnel (every 5 to 10
minutes) and a portion is taken in a 100 cc. measuring cylinder to ascertain the quantity
(percentage by volume) of the bicarbonate in the draw liquor and to observe the number of
minutes the bicarbonate takes to settle out with a sharp boundary between the crystals and the
supernatant liquor. With good crystals it should not take more than 3 minutes to settle. Poor
bicarbonate crystals may take considerably longer time and give a milky precipitate with an
ill-defined and mobile boundary between the precipitate and the clear liquor. A 5-cc.portion of this
clear liquor is pipetted off and titrated for "free ammonia" with NH2 SO4 , using methyl orange as
the indicator. The same portion after "free ammonia" determination can be titrated for Cl- by
Mohr's method with the addition of a small amount of calcium carbonate powder to neutralize the
acidity. It should be noted that the "free ammonia" determination is in reality a total alkali titratio
and the result, strictly speaking, does not represent free ammonia alone, but the error is on the safe
side. With the ammonia and chlorine titers known in the green liquor and with the "free ammonia"
determined in the draw liquor, the tower decomposition, i.e., per cent conversion, can be roughly
estimated as follows:
% Decomposition=NH3 titer in green liquor -free NH3 titer in draw liquor
100
Cl-titer in green liquor

This is only a rough indication of more or less relative value to the operators for the reasons
mentioned and also because it does net take into account the change in volume between the green
liquor and the draw liquor and the loss of ammonia from the towers to the tower washers.
At the filters the head of wash water above the orifice should be regulated. No attempt is
made to make any chemical determination on the bicarbonate in the field. The filter liquor in the
filter liquor drop leg fromthe filter separator to the main is sampled for a Cl- determination. From
the Cl- tilter in the draw liquor and that in the filter liquor, the amountof dilution by wash water on
each filter can be readily ascertained. Another points in the filter liquor main where other solutions
are introduced similar Cl- determinatious are made to determine the amount of dilution in the filter
liquor at these respective points. The furnace gas, after passing through the condensers on its way
to the compressors, is scrubbed and cooled by soft water in a scrubbing tower and the wash water
is circulated in this scrubber by a centrifugal pump, until a titer Of 8 to 10 is reached or until the
water gets too warm. This wash water carryingammonia may be used for filter wash. The wash
water in the scrubbingtower is fixated for ammonia content in order that neither too much fresh
water may be added nor too high an ammonia concentration in the wash water may be reached.
The amount of cooling water through the furnace condenser is adjusted, so that ammonium
carbonate crystals may net be formed as the result of the temperature in the condenser getting too
low.With the insertion of a furnace gas, scrubber before the condenser, however, there is little
occasion for trouble arising from this source.
Distiller house. The milk of lime for the distiller should be titrated for total lime. It is
important to make as concentrated a milk of lime as can be handled in the piping system. Too
weak a milk of lime will cause unnecessary dilution of the liquor to be distilled, resulting in the
waste of steam, loss of ammonia and loss of lime as excess lime in the large volume of distiller
waste produced.
For the same reason, it is even more important to keep the filter liquor as concentrated as
possible. Waste liquor should be kept small in volume and floor wash water should not find access
to the filter liquor system. A dilute feed liquor puts a very heavy load on the distiller, making it
very difficult to make the required quantity of ammoniated brine in the absorber. This is a serious
matter, because so large a volume of dilute feed liquor would have to be put through the distiller
that the several compartments at the bottom of the distiller would tend to be filled up, making the
steam pressure at the bottom compartment excessive and the top temperature of the distiller fall
off. ThCs always results in heavy loss of ammonia in the distiller waste or run-off liquor.
Consequently in a large alkali works a special weak ammonia distiller is provided for distilling off
the various weak ammonia-bearing liquors.
On the top of the distiller, there should be a recording thermometer to register the
temperature of the exit ammonia gas. There a manometer should also be provided to adjust the
pressure condition (vacuum) in the distiller. The distiller top should have a vacuum of about 1/2
cm. Hg, both to prevent any possible leakage of ammonia to the room and to aid the distillation of
ammonia gas from the distiller. The temperature should be about 82~ to 86~ ．, but this
temperature is dependent upon the individual construction of the distiller and the location of the
distiller condenser in the system. Similarly the distiller condenser outlet should be provided with a
thermometer and a mercury manometer. The temperature of the ammonia gas at the outlet of the
condenser to the absorber should be above 55~ ． and the vacuum about 1.5 cm. Hg. If the
temperature of the outlet gas from the distiller condenser is much above 55~ ., excessive steam
would pass on to the absorber, causing dilution of the ammoniated brine (low Cl- titer), and also
poor ammonia absorption in the absorber (high temperature in absorber). This frequently leads to
the "hot top" phenomenon in the absorber and is detrimental to the absorber operation. On the
other hand, if the distiller condenser outlet gas temperature is below 55~ ．, ammonium
carbonate and ammonium carbonate crystals tend to form in the gas passage, clogging the
condensers. The heater liquor, i.e., liquor from the bottom of the heater before entering the
prelimer or lime still, should be tested for freeammonia and carbon dioxide. The carbon dioxide,
as in the cleaning tower liquor, is determined by the simple evolution method, and the volume
similarly recorded. There should be little or no carbon dioxide, but there will be some free
ammonia in the heater liquor in the type of arrangement in which the heater is
pleased directly over the lime still with strong ammonia gases from the lime still below passing in
contact with the liquor from the heater. The heater temperature should be closely observed, too
low a temperature will cause incomplete removal of carbon dioxide gas from the heater, and too
high a temperature will add an unnecessary burden to the condensers above. The distiller waste
should be tested for the presence of excess lime by titrating with hydrochlorie acid. The complete
removal of ammonia from the distiller waste can be ascertained by the sense of smell, which is
delicate enough, to guide the operation provided a sufficient excess of lime has been maintained.
A Nessler reagent is kept at hand to test ammonia whenever desired. The milk of lime and filter
liquor is both fed to the distiller through orifices whose diameters and liquor heads above the
respective orifice plates should be recorded. The amount of milk of lime used should be in
proportion to the quantity and the fixed NH3 titer in the filter liquor feed. With the strength of the
milk of lime and the composition of the filter liquor to the distiller running uniform Day and night,
this relation will soon be established, In normal operation , there is to be added a volume of milk
of lime equal to about a third of The volume of filter liquor fed. The prelimer liquor should be
tested for free ammonia and excess lime. To study the dilution of mother liquor by wash water,
furnace condensate, etc, samples of filter liquor are taken at different point and the CI -titers
determined. The volume is inversely proportional to the concentration or titer of chlorine. The
exhaust steam main to the distiller should be provided with a recording pressure gauge and the
distiller bottom compartment with an ammonia proof pressure gauge. The pressure at the bottom
compartment of the distiller varies From 8 to 10 pounds, depending upon individual construction
of the lime still and the heater, and the rate of operation.When the distiller shows plugging or
when the volume of the liquor in the distil for some reasons is excessive, the pressure may rise to
12 to 15 pounds. The distiller bottom compartment should be provided with an ammonia meter
(such as an Esterline-Angus recorder) to indicate the presence of ammonia in the distiller waste,
by means of conductivity measurement ,using a recording ohm-meter. The meter may be
calibrated in dollars of ammonia loss per ton of soda ash to impress the distiller men.
Furnace room.The furnace gas up-take should be provided with a water manometer to
indicate the pressure (vacuum) condition inside the furnace. The nearer the furnace approaches the
natural condition, the better it is for the furnace operation and furnace gas obtained. Too high a
vacuum causes much leakage of air inward and consequently a weak returned gas. Too high a
positive pressure would cause steam to condense in the returned ash chute and in the extract barrel,
clogging these passages. It may even cause blowing through the extract barrel .The flue gas up
take or down take as the case may be, at the extract end of the furnace should be provided with a
base metal then couple with a recording meter to show the temperature of the flue gases. This flue
gas temperature may be from 350~ to 450~ . The extract ash. is tested for temperature every
few minutes. The temperature should be between 175~ to 190~ ．
These extract and flue gas temperatures, especially the former, give indications whether the
bicarbonate in the ash is all out. These are not always absolutely reliable. They are reliable when
the furnace is operating uniformly, i.e., when it is charged and discharged regularly and
continuonsly. When, however, the discharge is interrupted for some reason, the ash that first
comes out may be completely free from bicarbonate, and yet has a lower temperature than it
should; on the other hand, when the furnace is pushed too hard, the fire being pulled
forwardtoward the extract end by the stronger draft employed, with the ash pouring out through
the extract, net all the bicarbonate will have been decomposed in the extract, though the extract
temperature does not appear partienlarly low. With some experience, the operator can judge
unmistakably whether the ash is free from bicarbonate. The ash that bas been completely calcined
looks powdery and "dead." The ash that contains under composed bicarbonate looks mobile and
flows like a fluid. The furnace condenser gas outlet should be provided with a thermometer and
the cooling water regulated according to this outlet temperature. The tem premature of the gas
here should not be below 30~ . and the use of a fur nice gas scrubber greatly lessens the duty of
the furnace gas condenser, and consequently the amount of cooling water required for this
condenser would not be large. In the furnace gas system, as in the lime kiln gas system or in the
absorber vacuum system, manometers are provided at various points to follow the vacuum
conditions and to observe any plug going tendency caused by soda dust at any point in the system.
This trouble is very likely to occur in the gas up-take main between the fur nice gas outlet and the
furnace gas scrubber, because of soda dust deposition carried over in the furnace gas. Butterfly
valves are provided to entree the vacuum conditions in the furnaces. In the main, beyond the gas
scrubber an automatic vacuum regulator should be provided operating on the geometer principle
for maintaining a constant pressure in the dryer. Chemical examinations, either on the bicarbonate
fed to the furnace or on the extract ash from the furnace, is net attempted by the operators in the
field, but the temperature of the extract is closely observed to insure a completely calcined ash.
The amount of returned ash is judged by the wetness of the bear bounty and by the behavior
of the furnace, i.e., sealing in the furnace. The bic arbonate from the filter should not cozen out
moisture on continued kneading in the hand. The returned gas should be analyzed for the
percentage of carbon dioxide with a Corset. The ability to get a large output from the towers, other
conditions being equal, will depend. upon the richness of the gas returned from the furnaces. Not
only the capacity of the columns but also the crystalline character of the bicarbonate obtained-- in
fact the whole successful operation of the towered--will be largely contingent upon the richness of
the gas from the furnaces. The ash coming from the furnace is cooled in an ash cooler to 80 or
70~ . and should be packed hot so that the ash may not absorb excessive moisture from their.
The ash as packed is sorted according in the color and grade, and classified into a number of
grades in accordance with market conditions. Normally there is only one grade, however.
Power plant. In the boiler room, the coal burned in the furnace should be weighed and also
the ash coming from the furnace. The pressure of the steam in the boilers should be recorded. It is
desirable to have a meter inserted in the feed line to measure the water fed to the boilers. The
temperatures of the feed water entering and leaving the feed water heater should be recorded to
determine what the heater is doing. Tests for oxygen content in the feed water from the
desecrating heater to the boilers should be regularly made and recorded. A draft gauge should be
installed at each furnace to regulate the draft to be carried by the boiler. A carbon dioxide recorder
will pay for its own installation. On the steam main taking steam from a battery of several boilers
there should be a steam flow meter to measure the quantity of steam generated per hour, and a
recording pressure gauge to register the average pressure of steam during the 24 hours. In the
engine room, there should be a complete record of the voltage, current, kilowatt-hours, etc. of the
generators, and the current in each feeder supplying power to each house in the plant. A complete
set of control instruments, such as frequency meter, power factor meter, voltage regulator,
synchroscope, and standard power plant accessories, should also be provided. On the carbon
dioxide compressors, the vacuum exhausters, the air compressors, and the main cooling water
pumps (all these large units being steam driven to get the exhaust steam required) the r.p.m.'s, the
pressures (or vacuum), and the temperatures at each gas inlet and outlet should be observed and
recorded. The fewer the number of r.p.m.'s in the carbon dioxide compressers for a given plant
output, the better would be the power economy, as this means smaller consumption in coal. The
titer of soda water for lubricating CO2 gas valves and cylinders should be determined at least twice
a week, The sodium carbonate content in the soda water used should have about 10-12 titer.
All samples, which the operatives cannot test in the field, are taken care of by the chemical
laboratory. Samples of the liquors or solid products (as the case may be) must be taken at
regular intervals as directed by the chief chemist and a composite for each during the shift sent
to the laboratory dally for analysis. This applies to all samples whether the operatives are
required to test in the field or not.
The smoothness of operation depends upon delicate control and absolute regularity of
operation. The process permits of no interruption in any part of the process for 24 hours s day
and 365 days a year. If, in a certain apparatus, e.g., the towers, the operation of the carbon dioxide
compressors is suddenly interrupted for some reason, it means that the columns will be plugged
solid, which may put them out of commission for days. Another requirement is that the operation
in all divisions be balanced, as the process depends upon the interlocking of the several divisions.
Trouble in any one division will soon be reflected in the others and the whole plant may be
thereby crippled.
 Table 147 gives an estimate of the cooling surface necessary in each apparatus per ton of
soda ash expected from the said apparatus. No hard and fast rule, however, can be laid down.

TABLE 147. Cooling Surface Required

(For average conditions.)
Cooling Surface Required per
Apparatus Ton of Soda Ash Made
Vat coolers 80-120 sq. ft.
Carbonating towers 75-90 sq. ft. in each tower
Distillers
Coolers on top of the distiller with filter liquor 40 Total
Condensers for NH3 gases with water 30 70 sq. ft.

Absorber coolers 25 sq. ft. for liquor Total

15 sq. ft. for gases 40 sq. ft.
Furnace condensers 30 sq. ft.

With good cooling water, the mount of cooling surface provided for each unit is ample.
Nevertheless, specific conditions in each individual plant may vary so that considerable
departure from the cooling surface giver is possible.
 Chapter XXVI

Losses and Consumption of Raw Materials in

Ammonia Soda Process
While the ammonia soda process is capable of the highest scientific control and is being
conducted exceedingly well, the efficiency of the main operation, the conversion of sodium
chloride into sodium bicarbonate, by the nature of the reaction, is unfortunately comparatively low.
The average figure for the decomposition of sodium chloride to sodium bicarbonate is below 75
per cent. Under the best conditions even for a short time, the percentage of conversion of sodium
chloride falls short of 80 per cent. This is quite different from the usual effic iency of 95 or even 98
per cent in most of the other chemical industries.This low efficiency has been so far tolerated
because salt is cheap and is a small iteming the cost of soda manufacture .In certain plants salt is
recovered from the distiller waste as refined salt. Operators in the plant are too accustomed to look
upon the losses of salt with little concern, although their vigilance in safeguarding the process
from loss of ammonia has been most particular. As a general rule in the plant an effort is made
only to recover ammonia other constituents such as salt etc.being discarded as distiller waste.
In the manufacture of ammonia soda the direct raw material are salt and limestone Ammonia
may be compared to a catalyst, but its loss in the cycle of operation is virtually chargeable to its
consumpition by the process. Coal and coke maybe regarded as indirect raw materials .We shall
now enumerate the sources of losses of different raw materials in the process and give normal
figures wherever possible, foot the rate of consumption of each according to the best modern
practice.

Carbon dioxide.

All carbon dioxide that goes to make soda ash is derived from limetone. Lime, the joint
product from the burning of limestone is used in the recovery of ammonia .The rate of limestone is
governed by the rate of consumption of lime and carbon dioxide gas, Generally however it is the
line, rather than the carbon dioxide gas, that determines the consumption of limestone in the plant.
But the loss of carbon dioxide ,especially the river carbon dioxide gases(either from the waste -age
of the rich gas or from the dilution of the gas by air ,thereby rendering the carbon dioxide less
available)has an important bearing on column operation and consequently cause a higher rate of
consumption of all material- salt ammonia, coal ,limestone and coke .The sources of loss of
carbon dioxide are as follows:
(a) Leakage on the top of the limekiln. The top of the limekiln is usually not gas-tight, and
any attempt to nm the kiln under vacuum to aid the decomposition of limestone generally results
in a low test gas caused by air leakage from without. Consequently, we are compelled to on the
kill with a slight positive pressure at the top. Hence there is a considerable loss of carbon dioxide
from the kiln to the atmosphere. Fortunately this is a "lean" gas and its lees is.net seriously felt.
Generally there is an excess of this "lean" gas, which is allowed to escape to the atmosphere
through the vent pipe provided on top of the kiln.
b Formation of calcium carbonate with milk of Lima in the distiller. If the heater is not
working efficiently, not all the carbon dioxide in theliquor will be driven out. Even in a distiller of
good design, considerable carbon dioxide at times may be founding the liquor so the bottom of
theheater. The loss, however, should be very small and should net be toleratsd when found in
eunsiderablc quantities, as it means a corresponding loss in available lime.
( c) Formation of calcium carbonate and magnesium carbonate in the and vats. Carbon
dioxide gas coming eve, with ammonia from the distiller is dissolved by brine in the absorber,
forming ammonium carbonate, etc. If tile brine contains impurities such as calcium sulfate, as if
rock salt brine, or calcium and magnesium sulfates and chlorides, as in sea brine, corresponding
calcium and magnesium carbonates are formed. constitoting a large part of the "mud" in the vats.
This gives rise to the presence of fixed ammonia in the ammaninted brine. The extent to
whichfixed ammonia occurs depends upon the amount of these impurities in the brine used.
Generally, it would be about 3 to 5 titer andmay be as high as 8 titer for sea brine. This generally
amounts to 2 to 3 per cent for the loss of carbon dioxide on the weight of earbob dioxide handled.
(See Chapter V, "Purification of Brine.").
(d) Exhaust from absorber vacuum pump. Just as ammonia is recovered from the mother
liquor from the columns by distillation, so also the carbon dioxide dissolved in the mother liquor
is recovered in the distiller. Normally with strong ammonia gas, carbon dioxide ought to be
absorbed by the resulting ammaniated brine. But as the temperature in the absorber bottom and in
vats is about 65~ C. the partial pressure of carbon dioxide in the hot ammeniated brine is
considerable. Considerable carbon dioxide will escape from the ammoniated brine. In the absorber
washer and weak washer, the outlet gas from the absorber and from thevat vent lined containing
ammonia and carbon dioxide is deprived of ail its ammonia content, leaving most of the earbob
dioxide as spent gas to be drawn through to the vacuum pump or exhanster. Except for the small
quantity of air leaking into the system, this exhaust gas should be essentially carbon dioxide. If the
absorber system is airtight, the percent age of carbon dioxide in the exhaust gas should be
considerably higher than that of the kilo gas. It may test around 70 per cent carbon dioxide. The
gas should be sent back to the tower gas main and made available for tower operation. If the
absorber system permits leakage of air through the pipe joints, manometer connections, valve
pickings, cover flanges, etc., the exhaust gas may contain leas nan 20 per cent carbon dioxide, and
the gas is thus absolutely worthless and has to be discharged to the air. The loss here is 10 to 15
per cent of the rich gus available. Such a loss cannot be tolerated if the soda plant is to be run
efficiently. Some sulfide is also recovered along with this absorber gas.
(e) Exhaust from the filter vacuum pump. Because of the high vacuum employed there is
a tendency to suck out carbon dioxide and ammonia from the mother liquor in the filters. As a
large quantity of air is sucked in through the pores in the filtering medium, carbon choice in the
exhaust of the filter, exhauster is very weak, and has to be discharged to the air. Fortunately,
because of the lower temperature of the draw liquor, the loss of carbon dioxide is not serious and
should be very small.
(f) Exit gases through the column washers. The column Washer is designed to scrub
ammonia from the spent gas from the carbonating columns. All the carbon dioxide is lost. The
amount of loss is determined by the efficiency of the carbon dioxide absorption in column
operation. Normally it should be less than 5 per cent of the total gas pumped to the columns, a
general mathematical formula in terms of the percentage of carbon dioxide in the mixed gases into
the columns and the percentage of carbon dioxide in the column exit is given below:
Let pm =% CO2 by violin the mixed gases to the columns
per =% CO2 by violin the column exit gas
L =% loss of CO2
Then L =Pe/100-Pe/Pm/100-Pm x 100=Per(100-Pm)/Pm(100-Pe) x 100

If the mixed gases in the columns are approximately 50 per cent, lO0-Pm is nearly equal to
Pm. As P6 is small, generally less than 5 per cent, 100-P, can be cancelled by 100. In such eases
Per can be taken roughly equal to L, the percentage loss of carbon dioxide from the total carbon
dioxide admitted to the columns. Good Solvay columns should have a carbon dioxide absorption
efficiency of more than 95 per cent.
(g) Carbon dioxide blown from soda furnaces. Normally the calcining furnace is run
under a very small vacuum or, better still, under exactly neutral conditions of pressrun. There
should not be much loss of carbon dioxide gas. But due to pressure fluctuations in the furnace and
the unavoidable opening of the barring hole for chiselling the scale, etc., sometimes the furnace
gases are blown or puffed out. When the furnace is not running smoothly because of scale
formation, poor bicarbonate crystals, or excessive moisture content in the bienrbonate, there is
sometimes a momentary pressure inside the furnace causing a lees of the seal in the barrel extract
or feed passages, and the gases containing ammonia, oarben dioxide and soda dust may be blown
out through these openings. In normal running, however, su0h cases seldom oeuvre and the loss of
earbun dioxide should not be large.
Any loss in carbon dioxide will have to be made up by increasing the rate of burning of the
limestone in the limekiln. As the carbon dioxide obtained in this way is in a comparatively dilute
form, the increase in the volume of the kiln gas relative to the volume of the rich gas from the
dryers and from the absorber system reduces the average percentage of carbon dioxide in the
mixed gases. This, as has been shown in Chapter VIII, is every serious drawback in column
operation. Let us examine this point further. Let
Vk =Vol. of kiln gas and Pk its CO2 concentration expressed as a fraction by vol.
Vf = Vol. of furnace gas and Pf its CO2 concentration expressed as a fraction by vol.
P m= CO2 concentration in the mixed gases expressed as a fraction by vol.
Then

or
or

That is, the ratio of the

volume of the kiln gas to the volume of the furnace gas is as the difference in carbon dioxide
concentrations between the furnace gas and the mixed gases is to the difference in carbon
dioxide concentrations between the mixed gases and the kiln gas. Again from
.'. Given Pk, Pm, Pf, let x = the fraction of CO2 (by weight) contributed by the lime kilns in the
mixed gas to the columns

Therefore if we know the concentrations Pk, Pm, Pf, we can readily find out the amount of CO2 (by
weight) furnished by the lime kilns or by the dryers.
An interesting case presents itself when the kilns and the furnaces furnish equal amounts
(weights) of CO2 for the tower operation, in which case

This relation can also be obtained by putting x = 1/2 in (3) above. That is, in the ideal case
where there is no loss of carbon dioxide gas and when the furnace operation is in step with the
column operation, so that the lime kilns and the furnaces furnish equal amounts of carbon
dioxide to the columns, the percentage of carbon dioxide in the mixed gases forms a harmonic
mean between the percentage of carbon dioxide in the dryer gas and the percentage of carbon
dioxide in the kiln gas. This relation affords a good guide in operation. When the furnace gas
contains 90 per cent carbon dioxide by volume and the kiln gas 41 per cent, and when both
contribute equal amounts of carbon dioxide to the column operation, then

But as is often the case, the furnaces contribute, fly 45 per cent of the total carbon dioxide
required for the column operation and the limekilns 55 per cent, then from (3)

This gives normal figures for the column mixed gases. But when the furnaces are furnishing
55 per cent and the lime kilns 45 per cent, the mixed gases will test higher and their value then
becomes 58.5 per cent.
If, under close control, the furnace gas tests 95 per cent CO2 by volume and the kiln gas 42
per cent CO2, if the two sources contribute equal amounts of carbon dioxide for column operation,
then the percentage of CO2 in the mixed gases will be

If, as may happen, the furnaces contribute 60 per cent of the total CO2 to the columns, then
In practice, because of other minor sources, of comparatively rich CO2 gas, such as the absorber
exhauster discharge, etc., the mixed gases to the columns may frequently average more than 60
per cent CO2 . Momentary readings may go as high as 67 per cent, but they are not the average. On
the whole, percent CO2 by volume in the mixed gases to the columns cannot exceed 60 per cent by
volume under normal working conditions. This is the argument in favor of the "double entry"
arrangement, whereby the original 90-95 per cent richness in the furnace gas can be utilized.
In the final analysis, all losses of CO2 must be made up from the burning of limestone in the
kiln.
It will be seen from the foregoing discussion that the loss of carbon dioxide (if the absorber
exhanster gas is returned to the columns), should be from 15 to 20 per cent of the total carbon
dioxide handled. Theoretically, to make enough sodium bicarbonate for I metric ten of ash there is
required a volume of 422 enable meters oil pure carbon dioxide under standard conditions.
Assuming a total loss of 17 per cent, the volume of carbon dioxide under standard conditions
would be 509 cubic meters per ton of ash. At an average of 57.5 per cent carbon dioxide in the
mixed gases in the columns, at a temperature of 30 , and at a vacuum of 2 inches H2 O, assuming
a volumetric efficiency of 80 per cent in the carbon dioxide compressor cylinder, there will be
required 1230 cubic meters of piston displacement in the gas cylinders per ton of ash.

Ammonia.

One of the greatest achievements made in the ammonia soda industry today is the blight
efficiency of the ammonia cycle. In contrast to a lees of ammonia of 50 to 80 kg. of anunonium
sulfate per 1000 kg. of soda in 1880, today's figure of 3 to 4 kg. of ammonium sulfate per 1000 kg.
of soda ash for small plants and still less for large planet marks a good step of advancement. Of
course, the modem soda plants have an advantage in that they work on a very much larger scale
and the loss of ammonia per unit of production is necessarily smaller. Following are possible
sources of loss of ammonia:
(a) Ammonia stills. Considerable effort is now directed to keeping the volumes of the filter
liquor and milk of lime for the distiller at a minimum. There are various reasons why the volume
of the liquors fed to the distiller should be kept as small as possible. One reason is to reduce the
loss of ammonia through the distiller. In the distiller waste there is always a small amount of
ammonia that has net been completely distilled out. The removal of the last portion of ammonia
remaining in solution requires a long time. Generally, not less than 0.005 gram of ammonia per
liter is present in the distiller waste. At the rate of 10 Cubic meters of distiller waste per ton of
ash, there is a loss of ammonia of at least 0.2 kg. ammonium sulfate per ton of ash. This does not
include losses caused by an accidental "sour bottom," or by a momentary deficiency in lime
causing fixed ammonia in the distiller waste, or by other accidents. Under the best conditions, the
distillers are one of the largest sources of loss of ammonia in the whole soda ash plant.
(b) Dryers. Considerable ammonia may also be long in the form of gases leaking from the
soda dryers, The dryers are never gas-tight at rotating joints, etc., and frequently there is need to
open a hole for rodding or chiseling the scale. The loss is considerable when au attempt is made to
run the furnace under a slight pressure to maintain a good gas test. Hence a very slight vacuum is
desirable, consistent with a good gas test.
(e) Tower washers. The function of the tower washer is to scrub ammonia from the tower
exit gases with saturated brine which is to be later sent to the absorber for ammonia absorption.
Under normal running conditions with an ammonia titer not over 12 in the bottom compartment of
the tower washer, using cold, saturated brine, the loss of ammonia through the tower washer is
email. When the mixed gases to the columns are low in carbon dioxide, when the carbon dioxide
absorption in the carbonating columns is poor, when the column decomposition is low, when the
green liquor fed to the top of the carbonating column is at an excessive temperature, and when the
saturated brine used for scrubbing cannot be kept cold, as is the case during the hottest summer
days, then the volume of brine is not sufficient to scrub the ammonia in thecolumn exit gases and
the loss of ammonia may be considerable. The body of the column washer will then be hot and
ammonia cannot be readily kept is solution. As the quantity of the saturated brine that cane put
through the column washers is limited by the rate of production, i.e., by the rate of ammonia
distillation, it would not be feasible to let down a larger stream of brine to the top of the column
washer than can be used in the system. Losses under these conditions cannot be ascertained except
by observing the total losses over a period of time by means of a gas meter measuring a small
stream of the column washer exit gas and determining the quantity of ammonia by an absorption
bottle.
(d) Absorber and filter exhausters. The duty of the filter washer is generally light. Unless
the design of the washer is faulty,, or the brine used for scrubbing is at an undesirably high
temperature, there is little loss of ammonia in the filter exhaust, in the absorber exhauster, however,
there is danger of losing ammonia during the "hot top" time. When, however, the exhaust gas is
sent back to the columns, as should be the case, such a danger does not exist. This is one of the
reasons why the exhaust gas from the absorber system should not be discharged to the atmosphere
and every effort should be made to stop air leakages in the absorber vacuum system to enable the
gas to be returned to the column gas main.
(e) Draw funnel and bicarbonate filters and exposed bicarbonate on the feed floor in
the furnace room. It has not been found successful to have the ammonia liquors going throng the
process in a closed system throughout, i.e., without exposing any portion of the liquor at any stage
to the open air. Under the present method of working the sherry, the magma is drawn from the
carbonating columns and exposed to the air at the receiving funnel and at the filters. At a low
draw temperature and with good decomposition, little odor of ammonia is noticed in the room and
the loss of ammonia is negligible. When, however, the decomposition of the sodium chloride in
the columns is low, when the free ammonia tiller is high, and when there is not enough cooling in
the columns to operate them at a low draw temperature, considerable ammonia is lost to the air
from the warm liquor, when the liquor is drawn out to the draw funnel and exposed at the filters.
Not only does this mean a loss of ammonia but it also cause inconvenience to the workmen in the
room, a condition that spells failure for the plant.
(f) Piping connections, gasket flanges, pump packing glands, valve and cock par. kings,
manometer connections, sampling pipe nozzles, etc. The minor losses of ammonia liquors
caused by leaks in the centrifugal and plunger pump packing glands, leaky valves and cocks,
manometer connections, sampling lines, etc., though much to be regretted, are to a certain extent
unavoidable, as the corrosion and wear on the materials by ammoniated brine liquors at an
elevated temperature create a problem which still confronts ammonia soda manufacturers. In a
poorly managed plant, the loss of ammonia from leaks everywhere and from let-outs may far
exceed the normal loss in the distillers. The total loss may be so great that it is hardly possible to
account for the depletion of ammonia in the system. A search for the sources of loss must be made
and the conditions immediately remedied. On the floors of each of the rooms where such pumps
are located or where such liquors are likely to be spilled, there should be provided a sump to
collect these ammonia-bearing waste liquors which are to be immediately returned to the weak
liquor distiller for distillation. In plants where the volume of the liquor handled is large, these
minor losses are insignificant. In a small plant having a small output such losses may amount to
quite an item.
Theoretically, according to the reaction
NH4 HCO2 + NaCl NaHCO 3 + NH4 Cl

1/2 (Na2 CO3 + H2 O + CO2 )

to make 1 ton of ash there would be required 0.32 ton of ammonia to be distilled. In practice,
however, owing to the presence of considerable free ammonia in the draw, at least 0.45 ton of
ammonia must be circulated, i.e., the ammonia efficiency in the towers is below 75 per cent on the
average. Every handling of ammoniais accompanied by some loss: therefore, for a given output in
soda ash, the higher the ammonia efficiency in the columns, the smaller the quantity of ammonia
circulated, the less is the ammonia loss, other conditions being equal.
(g) Tube leaks. Cooling-tube leaks at the tube sheets in the strong liquor coolers or gas
condensers are likely to occur even if cast-iron cooling tubes are used. Cast-iron tubes in the
distiller condensers are known to have been eaten through by hot, ammonia gases after five years
in service. To check these leaks, the exit cooling water is occasionally tested for ammonia. To
follow any loss, sewer samples taken around the plant are frequently tested for ammonia.
In Order to maintain the absolute regularity of operation which is required in the working of
the ammonia soda process, the rate of operation in the various divisions must be held constant.
Since the settling and storing capacity in the vats and storage tanks is practically constant, the soda
ash output of the plant depends upon how fast the ammoniatedbrine can be circulated, i.e., from
the absorber, through the towers, through the distillers, and back to the absorber. This, in turn,
depends upon how many hours the ammoniated brine takes for complete settling in the vats. For
the speed with which the ammoniated brine can be sentthrough the system is mainly governed by
how fast it can be put through the settling vats to get a complete settling. The settling capacity of
the vats and the storage capacity of strong liquor storage tanks has an important bearing on the
ultimate rate of production. Generally, with a high quality brine containing small amounts of
impurities, there should be capacity enough in the settling vats and in the storage tasks to hold the
volume of ammoniated brine in the form of stock and working liquors of normal strength
equivalent to at least as many tons of ammonia (calculated as NH4)2SO4), as the daily tonnage
output of soda ash in the plant. This figure varies greatly according to the purity of the brine
available. With the present day brine pretreatment, the ammoniated brine settling and storage
capacity can be greatly reduced.

Lime.

Limestone is one of the important raw materials in the ammonia soda industry, as it furnishes
all the carbon dioxide required for soda ash manufacture and all the time for the recovery of the
comparatively expensive material, ammonia, and also lime for eanstieisatiun. Theoretically, one
molecular weight of limestone is required for every molecular weight of soda ash viewed from the
standpoint either of the calcium oxide or of the carbon dioxide entering into the reaction. But
because of various losses, come of which are unavoidable in practical operation, the consumption
of lime aim consequently of limestone is greater.
The sources of loss of lime or limestone are as follows.
(a) Waste in breaking. In breaking and crus hing the limestone to the required size of 5 to 6
inches, there are always some fine pieces smaller than 2 inches. These have to be screened out and
discarded, although the fines may be valuable for concrete construction work, road pavement, etc.,
and high silica stone may be useful in the cement manufacture.
(b) Underburned or unburned limestone. In order to get a high-test gas, no more coke
than necessary is charged with the stone in the kiln. Sometimes, intentionally, slightly tests coke is
used than is necessary to burn the stone completely, and there will be some underburned core in
the lime drawn. The amount of unburned stone in the lime is 5 to 10 per cont. Although unburned
stones of the sizes above 2 inches are returnable to the kiln as returned stone, file smaller sizes of
unburned stone are rejected as "sand" and the fines which are suspended in the milk of lime
simply pass through the distiller and out in the distiller waste. The total loss is estimated to be
about 2 to 3 per cent. It is apparent that the unburned stone generally comes from that portion of
the stone that is of the poorer grade, which is less readily decomposed, unless the fuel and air
distributions in the limekiln are faulty.
(c) Overturned or dead-burned lime. There is frequently a small amount of overturned
lime in the draw, even when the fuel ratio is not excessive. This overburden or dead-burned lime
takes a long time, sometimes days, to become hydrated or slaked. In this form the lime is not
available for liberating fixed ammonia during the abort time that the liquor comes into contact
with lime in the distiller. Frequently this inactive lime in the milk of lime may amount to 5 per
cent, the active lime being defined as that portion of lime determined by the sucrose method. A
large amount of overburden lime sometimes exists in the rejected "sand" from the lime Maker.
Such overburden lime, when finely ground and added to the milk of lime, will become to a large
extent available for reaction in the distiller.
(d) Carbonated lime in the distiller. When the temperature of the heater or the surface of
contact between the steam and the liquor in the heater is insufficient, there is a considerable
amount of carbon dioxide left in the heater liquor flowing into the prelim and into the lime still.
Even when the heater is working properly, a small amount of calcium carbonate is formed from
this source. Lime that is carbonated is not available for liberating fixed ammonia under the
conditions of distiller operation. True, a portion of the calcium carbonate found in the distiller
waste comes from the unburned stone or air-slaked lime that is suspended in the milk of lime, but
at least half of the quantity so found is due to carbon ton by the carbon dioxide in the liquor. This
loss may amount to fully 2 per cent of the total limestone under good operating conditions. It must
be emphasized that the mud from the settling vats, which contains high concentrations of carbon
dioxide, should be sent up to the very top of the heater so that all the carbon dioxide may be
liberated before the liquor comes into contact with the milk of lime.
(e) Excess lime in the distiller waste. An excess of lime in the distiller waste is simply to
insure complete liberation of fixed ammonia. The time of contact between the lime and the liquor
in passing through the distiller is a matter of minutes, so that some excess of lime is necessary to
facilitate the liberation of ammonia and its removal from solution. Any excess, however, more
than 3 titer (or 4 grams of lime per liter) is unnecessary and represents just so much more available
lime lost. At an excess of 4 grams Ca0 per liter with a volume of the distiller waste at 10 cubic
meters per (on of ash, the loss is 6-8 per cent of the total good available lime, which is equivalent
to about 10 per cent of the limestone charged in.
The total loss of available calcium carbonate in the limestone is therefore about 20 per cent.
With a good grade of limestone containing 96 per cent calcium carbonate or better, the rate of
consumption of raw limestone per ton of ash is calculated at 1.22 tons, which is confirmed in
practice.

Salt.

Because brine is cheap the loss has not been a matter of grave concern to the manufacturer.
Of course, any loss in handling is overshadowed by the magnitude of loss caused by the low
average decomposition efficiency in the columns. By the very nature of the reaction, the loss of
sodium chloride is more than 25 per cent, or the average decomposition is less than 75 per cent
under good operating conditions. Hence a first requisite for the establishment of an ammonia soda
plant is a cheap source of sodium chloride in the form of saturated brine. There are three main
sources of less of sodium chloride in ammonia soda plants:
(a) Incomplete decomposition in the columns. Under good operating conditions the
decomposition of sodium chloride into sodium bicarbonate can average only 73 per cent, air
Hough as high efficiency as 82 per cent for the reaction
NaCl + NH4 HCO3 NaHCO3 + NH4 Cl
was attained, according to some laboratory investigations. If the mixed gases in the columns test
low, if the concentrations of ammonia and sodium chloride in the ammoniated brine fall below
normal figures, if the cooling surface in the columns is insufficient to get the desired draw
temperature, the rate of decomposition falls still lower, and as much as 40 per cent of the total
available sodium chloride in the ammoniated brine may be unconverted and discarded from the
distillers. In addition, some sodium bicarbonate is dissolved by the wash water on the filters (2-3
per cent), so that the actual decomposition of sodium chloride into soda ash is still leas.
(b) Salt in the mud sent to the distiller. As ammoniation of the brine is accompanied by
purification, a considerable amount of semi-colloidal mud is formed by the treatment of the
ammonia and carbon dioxide present. The precipitate should be completely settled in the vats, and
the mud from the bottom of the vats disposed of by pumping to the distiller to recover the
ammonia. The entire volume of the mud is saturated was sodium chloride and all the salt
contained therein is lost. In order to minimize the loss of salt the mud must be pumped in as
concentrated a form as possible. In order to prevent plugging of the vats by mud a slow moving
scraper must be provided to keep the bottom of the vats open. Normally for sea salt the loss of salt
through mud pumping is from 5 to 10 per cent when the sea salt brine has not been pretreated and
purified. A further step in the economic use of salt is to pump the mud from the vats first into
another settler, called the mud settler, where the mud is further concentrated and pumped from it
to the distinct, it is, of course, even more necessary to install a scraper or an agitator in this mud
settler to prevent settling of the thick mud.
With a mud settler where all the mud from different settling vats is re-settled (and thus
concentrated), it is possible to pump the mud continuously from the settler to the distiller at a
concentration of not less than 95 per cent mud by volume. This is the volume of fine mud in a
slimy form, measured in a 100-cc. measuring cylinder after 10 minutes of standing. The saving of
good brine thus affected is indeed very considerable, especially where much mud is to be handled
from the use of untreated sea salt in the absorber and vat system.
Besides the main losses mentioned above, there are many minor losses from the drippings of
the pump packing glands, the leakage at the pipe joints, and so on, which are, however,
insignificant in comparison to the others. The rate of consumption of salt per ton of soda ash made
is thus 1.6-1.7 tons, which is the figure obtained in practice with a good grade of salt (brine).
The day is approaching when the utilization of salt in the process may be above 90 per cent.
To this end an effort has been made to recirculate the mother liquor from the columns for the
production of ammonium chloride crystals in conjunction with the manufacture of synthetic
ammonia. A new modification of the ammonia soda process has been worked out, wherein the
utilization of salt is 90-95 per cent (see Chapter XXVII, "Modifications and New Developments of
Ammonia Soda Process")

Coal.

Coal in the soda plant may be used to generate electric current both for power and lighting, to
calcine ammonia soda, and to generate steam for driving the carbon dioxide compressors, the filter
and absorber exhausters, the air compressors, etc., the exhaust steam from which is utilized for
ammonia distillation. For simplicity, the statement is based on coal, although oil or natural gas
when available can be used to advantage in place of coal. If these larger units of machinery
(generally located within the main engine room) are steam driven, the exhaust steam should be
sufficient for the normal running of the distiller. If the filter liquor is excessively diluted and the
volume to be distilled large, live steam may have to be added unless a bleeder turbine is installed.
The sources of waste of steam leading to high consumption of coal are:
(a) Excessive power consumed by the carbon dioxide compressors. Because of weak kiln
gas or weak furnace gas, or both, the piston displacement of carbon dioxide per ton of ash, i.e., the
r.p.m. of the compressors, may be excessive with a corresponding loss of steam and consequently
coal. Even with the most simple gas delivery valves (and therefore low volumetric efficiency in
the gas cylinders) there should not be more than 1400 cubic meters of piston displacement in the
carbon dioxide compressor cylinders per ton of ash, unless the gas is abnormally weak in CO2 .
(b) Excessive power consumed by the absorber exhauster. If the absorber system is net
air-tight, considerable air will leak in, not only giving an exhaust gas too low in carbon dioxide to
be returned to the columns but also increasing the power required to draw the gas through the
absorber exhauster. The piston displacement of the absorber exhauster per ton of ash should not be
over 350 cubic meters at a vacuum of about 10 cm. Hg.
(c) Live steam for distiller operation. A greater loss, and a more serious one, in so far as
regularity of operation and efficiency of the process are concerned, is the live steam required to
supplement the exhaust steam in the distillers. The distiller foreman should avoid using live
steam unless absolutely necessary, as, apart from the waste incurred from the use of additional
steam, live steam is not as suitable for distillation as exhaust steam.
(d) Steam leaks around the valves and piston rings. The leakage of steam around Corliss
valves is a serious matter. It causes high steam consumption and produces large quantities of
exhaust steam. In worst cases such engines cannot run up to the rated speed due to high back
pressure without having the throttle valve wide open with the governor set for the latest cut-off.
The loss from this source depends upon the care and attention given at the repairs and
maintenance of the steam cylinders.
(e) Low fuel efficiency in the calcining furnace. With poor bicarcarbonate crystals that
cannot be filtered dry and that form scale in the furnace, furnace operation will be greatly
handicapped and its capacity reduced. Consequently, the fuel required per ton of ash is increased.
With the bicarbonate giving a 3-minute settling test, with efficient filters, and with a well-designed
furnace running at full capacity, the fuel consumption should be about 1 ton of good coal for every
5 to 6 tons of ash calcined. The efficiency of the furnace is dependent upon its construction and
the brick setting, but its capacity is also dependent upon the character of the bicarbonate crystals
to be dried, i.e., its crystalline character and moisture content. Hence efficiency in the tower
operation is tied up with the fuel consumption in the furnace operation.
In general, in small plants the coal consumption is about 1 ton of coal per ton of ash. In large
plants, considerably less than this figure (as low se 0A to 0.5 ton of coal per ton of ash) is obtained.
This includes coal used for the generation of power and in the dryers. The low fuel and power
consumption is another noticeable advancement made by the modern ammonia soda industry.

Coke.

As all the coke is used in burning limestone, the smaller the consumption of lime, the less the
amount of coke required. But it also depends upon heat insulation sad heat regeneration in the
limekilns and upon the grade of limestone burned, for a given rate of lime consumption. With a
properly proportioned lime kiln and heavy fire-brick lining, i.e., with a tan kiln having a thick
lining, the coke required should net be more than 1 ton to every 14 tons of normal stone. With a
good distribution of coke in the charge and of air in the kiln, and with a good regenerating effect
for heat both In the gas exit above and in the lime draw below, as low as 1 ton of coke to every 16
tons of normal stone has been Obtained. This gives coke consumption at the rate of less than 0.10
ton per ton of soda ash made.

Sulfide.

A soluble sulfide, or S= ion, is a necessary constituent in the composition of ammoniated

brine. This sulfide in contact with the cast-iron surface in alkaline solution forms a firm coating of
ferrous sulfide on the internal walls of the apparatus and pipes, so that the sodium bicarbonate, and
consequently the soda ash made, does not contain iron rust. When etude ammonia liquor is used,
a large portion of the necessary sulfide is furnished by the crude liquor in the form of (NH4 )2 S or
NH4 HS When the crude liquor does not contain enough sulfide, some Na2 S must be dissolved and
added to the filter liquor for distillation. This is especially true when rock salt brine is used in
which the magnesium content in the brine is low. Fused sodium sulfide (60 per cent, technically
pure) can be used for this purpose. The amount of sulfide consumed per ton of ash depends
largely upon the output of soda ash, upon the efficiency of the works, and upon the magnesium
content in the brine. When ammonium sulfate is the only source of ammonia supply, the entire
sulfide required must be obtained from the sodium sulfide solution added to the feed liquor for

TABLE 148. Consumption of Raw Materials Per Ton of Soda Ash

Under Good Under Average
 Conditions Conditions
Salt .......... .......................... 1.50 tons 1.70-1.80 tons
Limestone ............................... 1.20 tons 1.30-1.50 tons
Coke at 71/2% of the stone ................ 0.095 ton 0.10-0.1l ton
Coal for boilers ......................... 0.25 ton 0.50-0.80 ton
Coal for dryers ...... .................0.16 ton 0.20-0.25 ton
Ammonia loss as (NH4 )2 SO4 .........2-4 kg. 5.00-7.50 kg.
CO2 gases (mixed gases to towers) piston
displacement ........................ 1000 cu. m. 1100-1200 cu. m.
Ammoniated brine........................ 5.7 cu. m. 6 cu. m.
Distiller waste ......................... 8.6 cu. m. 10 cu. m.
the distillers. A certain amount of sodium sulfide then must be dissolved and added to the filter
liquor daily to replenish what is used in the system. The consum ption of sulfide calculated to
sodium sulfide per ton of soda ash is 0.001 to 0.002 ton when no crude liquor is available.
When sea brine containing considerable magnesium salts is used, the sulfide consumption is
considerably less than 0.1 ton per 100 tons of soda ash made.
A summary of the raw materials consumed is given here. Too much dependence cannot be
placed on these figures, since a plant, owing to certain favorable conditions, may use one raw
material more economically than another. Therefore a range is given in each ease, but even this
range may be too narrow. One single factor that anuses greater variation than say other is the else
of the plant, i.e., the dally output of soda ash. A small, plant will have a less favorable showing in
the consumption of raw materials, no matter what effort is made to run it on an economical basis
The figures in Table 148 apply for the most part to large ammonia soda plants.
The whole ammonia sods process may be summed up by the following equation:
2NaCl + CaCO2 Na2 CO3 + CaCl2
Of course, this reaction does not take place under ordinary conditions. Hence we have to
introduce ammonia to bring about the ultimate result. The part that ammonia plays may be likened
to a catalyst, but it is not strictly a catalytic action. A catalyst does not cause a change but simply
hastens the resection rate to the equilibrium point, whereas in this equation the reaction does not
go and cannot be said to be hastened by a catalyst. From this equation, however, it is readily seen
what relationship exists between the salt and the limestone required on the one hand and the soda
ash and the calcium chloride produced on the other. The low conversion ratio, however, makes
these figures run much higher, as seen from the foregoing discussion.
 Chapter XXVII
Modifications and New Developments of

Ammonia Soda Process

The ammonia soda process has proved its value for the past eighty years since the days of
Solvay Brothers. It. has reached a high state of perfection, but it must be said to have left
something to be desired. In the first place, one of the essential raw materials--salt--used in the
process is utilized only up to 73 per cent, and in small works the percentage conversion is still less.
This leaves about 30 per cent overall of salt taken into the process unconverted into sodium
bicarbonate, and all this salt is practically wasted. This is caused by the nature of the reaction; for,
unless the waste liquor containing the unconverted salt can be used over again in the process, or
unless effort is made to recover the salt from the liquor, little can be done to improve the situation.
In certain plants such salt has been recovered and turned into refined salt, but the amount is very
small. As a consequence, the existence of the ammonia soda industry presupposes a cheap source
of salt and plenty of it, preferably in the tore of saturated brine. No country which has an.
expensive salt supply, or whose salt bears a high tax under the government control, can afford to
use the salt for the ammonia soda industry without certain special concessions from the
government for tax exemption.
In the second place, ammonium chloride formed in solution after the reaction cannot be
readily utilized. The common practice is to subject the liquor to distillation with lime in order to
free the ammonia and use it over again in the process. The ammonium chloride in the liquor,
although in a useful form, cannot be available for commercial purposes without a special
treatment. Many processes have been proposed to obtain ammonium chloride as crystals, and
numerous patents have been taken out in regard to the recovery of ammonium chloride in a solid
crystalline form from the liquor. Among these may be mentioned those proposing to concentrate
the liquor and cause ammonium chloride to crystallize at a low temperature by dissolving in the
liquor, salt, sodium nitrate or other sodium salt. The operation is costly and the ammonium
chloride liquor is difficult to handle on account of its corrosive properties when it is hot and
concentrated. Such a process has from time to time been the subject of patents by the I. G.
Farbenindustrie, A. G. in Germany, the Brunner, Mond & Co., in England (now Imperial
Chemical Industries, Ltd.) and the Solvay Process Co., in America (now in Allied Chemical and
Dye Corp.). We shall not attempt to dwell on this phase of development at any length, but prefer to
refer those interested to the patent literature, because such processes have not been used
extensively.
One important modification which is being developed into a new process, and which was not
originally intended to be a part of the ammonia soda process, was due to the pioneering work of
Prof. W. Gluud and Dr. B. Loepmann when they were working in the Gesellschaft four
Kohlenteehnik Laboratories at Drafted near Dortmund, Germany, as early as 1924. Much of the
process is not new, however, for the use of solid ammonium bicarbonate for the precipitation of
sodium bicarbonates from brine was mentioned by H. Sehreib 1 as early as the beginning of this
century. The important feature in this process is the introduction into the solution of a soluble salt
called the "Intermediate Salt" having Na + as action and as anion any radical other than Cl-. The
regulation of the concentrations of various ions in solution for the precipitation of ammonium
chloride and sodium bicarbonate crystals from the liquor, sad the manipulation of the operations
are rather ingenious, and deserve discussion at some length. The Institute at Dorstfeld was
engaged in researches on coal products, carbon dioxide, ammonia and the by-products of the coal
and coke industry. Heretofore, ammonia was all converted to ammonium sulfate as fertilizer-by
means of free sulfuric acid, whic h after all has no value as an ingredient in the fertilizer. Besides,
the useful Sulfuric acid is rendered unavailable, and in certain cases the sulfate part of time
fertilizer may have some harmful effect on the soil. Consequently, to save this valuable sulfuric
acid, the I. G. Farbenindustrie, A, G. developed a process of fixing ammonia by means of finely
pulverized gypsum (calcium sulfate) suspended in ammonia solution, by carbonation with carbon
dioxide gas. This is the Gypsum Process. This successful process marks a distinct improvement in
the manufacture of ammonium Sulfate as fertilizer dispensing with sulfuric acid. These German
investigations in the Kohlentechnik Institute were attempting to fix ammonia in the form of some
other salt than the sulfate, possibly as the chloride, as ammonium chloride had been reported by
Prof. Bosch, Prof. Wagner and others to possess as favorable fertilizer properties as ammonium
sulfate for many crops. A natural course for these investigators to follow would be first to convert
the tree ammonia obtained from coke manufacture to ammonium bicarbonate by the waste CO2
gas, and then to convert the bicarbonate to the chloride is displacing the bicarbonate ion. It would
also be natural to utilize the chlorine from common salt for this purpose by removing the
bicarbonate ion with sodium as sodium bicarbonate, as in the Solvay Process. Now, it was an easy
matter to precipitate the bicarbonate us sodium bicarbonate, but how was it possible to precipitate
the highly soluble ammonium chloride by a less soluble sodium chloride? Here, an ingenious
device was worked out by the use of an Intermediate Salt to obtain a condition of equilibrium
relative to the solubility products involving Na+, NH4 +, Cl- and other ions
1
Schreiber, "Die Fabrication der Soda nacho dem Ammoniak-Verfahren."
in the liquor (see below) and to introduce NaCl in a finely divided state. This caused separation of
ammonium chloride from solution, thereby fixing it as chloride. This ingenious device was looked
upon as a successful attempt to fix ammonia as ammonium chloride in place of ammonium sulfate.
But in order to achieve this, the bicarbonate must first be removed as sodium bicarbonate, a rather
soluble acid salt of sodium, but very much less soluble in the presence of high concentrations of
other salts in solution. For example:
In water at 30 ., solubility of NaHCO3 is .......... 11.02 g/l00 cc.
In 11% brine solubility of NaHCO3 is about ......... 4.7 g/100 cc.
In 20% brine solubility of NaHCO3 is ................. 2.2 g/100 cc.
and in saturated brine solubility of NaHCO3 is ....... 1.3 g/100 cc.
Soon it was pointed out that this by-product of sodium bicarbonate should have a great
commercial value as a source of soda. Attention was then directed to working it up in the form of
caustic soda or soda ash. With the successful attempts, especially in converting the sodium
bicarbonate to caustic, soda directly, the situation has changed considerably; and it was gradually
pointed out that this was another form of the soda process, a modification of the Solvay Process,
whereby so1id ammonium chloride is produced in conjunction with sodium bicarbonate, making
ammonium chloride and soda ash or outside soda joint products. Great significance, however, is
attached to the fact that in fixing the free ammonia as ammonium chloride, it brings the two
branches of fundamental heavy chemical industry--the Solvay Process and the synthetic ammonia
industry--more closely together; in fact the two may now exist side by side in one group: long. For
the fixed nitrogen industry requires some means of fixing free ammonia as a nitrogenous fertilizer,
and this is one way of fixing it for fertilizer purposes, and for other uses as well. Further, in the
fixed nitrogen industry, in obtaining hydrogen from semi-water gas or coal gas, there is s great
excess of carbon dioxide gas, which is released to the atmosphere. This waste CO2 gas now may
be used for the production of ammonium bicarbonate as an intermediate step in the manufacture of
soda and ammonium chloride by this process.
The operations involved in the production to ammonium bicarbonate crystals are to pass
carbon dioxide gas into ammonia solution containing approximately 20 per cent NH3 by weight,
first in saturating and then in precipitating towers, in the same way as in the Solvay Process. The
crystals of ammonium bicarbonate are filtered off from the slurry in centrifugals or on rotary
vacuum filters, as for ammonia soda. The filter liquor is sent through the absorber for the
absorption of further ammonia, and the solution is cooled and pumped to the carbonating towers
arranged in much the same way as in the Solvay process. Ammonium bicarbonate crystals so
obtained are used to precipitate sodium bicarbonate crystals in a number of precipitation tanks,
each provided with a high-speed stirrer.
In fact, many Solvay works have been making ammonium bicarbonate this way for years; but
because of the limited field for the use of ammonium bicarbonate in quantities, it has not been
made on a scale comparable with that of sodium bicarbonate. Now, with the use of ammonium
bicarbonate as the intermediate product in fixing ammonia as ammonium chloride, this field will
be greatly extended as never before.
In this connection, by way of digression, it may be mentioned that a process has been worked
not for obtaining crystals of ammonium bicarbonate that have such a low vapor pressure the t they
may be packed in gunny bags and kept in storage for months without any material leas. This is
accomplished by coating or cementing the crystals with mineral oil, sugar or other inert
materials.* This will remove many objections to the Use of ammonium bicarbonate commercially
because of its unstable character and the necessity of special air-tight containers.
So far only the hatch process for the precipitation of sodium bicarbonate has been in use, but
it is not altogether impossible to develop this step further into a continuous process, such as in the
other steps of the operation. The development at the present stage is limited to a comparatively
small capacity on account of the batch character at this particular stage.
Quite different from the Solvay process is the fact that this process starts with a solution
virtually saturated with respect to both NaCI and NH4 Cl. For some time in the beginning of the
development, the difficulty had come from the fact that ammonium chloride and sodium
bicarbonate would precipitate together when ammonium bicarbonate was introduced into such a
concentrated solution of sodium chloride and ammonium chloride. Many attempts had been made
in the beginning to separate ammonium chloride from sodium bicarbonate in the precipitate, by
resorting to the use of strong ammonia solution (40 per cent NH3 ) for dissolving out ammonium
chloride is the form of a highly soluble ammonia complex such as NH4 Cl. 4NH3 . This involved
the use of a very strong ammonia solution, the handling of which caused considerable difficulties,
and the loss of ammonia placed a serious handicap on its development.
Finally, the us of strong ammonia solution was dispensed with, and the difficulty was solved
by dividing the operation into two steps: first the precipitation of sodium bicarbonate, and then the
precipitation of ammonium chloride in separate vessels at different temperatures by carefully
regulating the concentration of various ions in the solution so that only one precipitates at a time.
*D.R.P. -622,8876; D.R.P. -485,054; etc.
The principle underlying this process is to obtain a saturated solution containing NH4 +, Na+,
Cl-, and another ion (A-). This we may express as NH4 Cl, NaCl and NaA in solution, the NaA
being the intermediate salt present in quantities substantially equivalent to the amount of NaHCO3
to be precipitated from the solution. The balance of Na+ may be considered to exist as Na+ Cl-. At
the same time, NH4 Cl may be said also to be present to saturation under these conditions. When to
this solution are added ammonium bicarbonate crystals, NH4 HCO3 dissolves and gradually
displaces sodium bicarbonate as precipitate. Thus,
Na+A- + NH4 +HCO3 NaHCO3 + NH4 +A-
However, the concentration of the NH4 + ions in solution has been so regulated that, when
NaHCO3 is thrown down by NH4HCO3 , the additional quantity of NH4 + ions formed, together with
NH4 + ions already present in solution, does not exceed the solubility product of NH4 Cl with the
consultation of Cl ions present in solution at the temperature in question (e.g., 35 .), so that no
NH4 Cl separates out with NaHCO3 .But as much of Na+A- as is equivalent to the NH4 HCO3 added
has been converted to NH4 A in solution. This increases the NH4 + concentration in solution by the
amount equivalent to the ammonium bicarbonate added. When, therefore, the solution after
filtration is cooled to say 10 to 15~ .below the temperature of the sodium bicarbonate
precipitation (s.g, 20-25 .), and common salt is introduced, the salt dissolves and in its solution
furnishes additional Cl- ions which, together with the Cl- ions already present and with the
increased concentration of NH4 + mentioned above, now cause the solubility product of NH4 Cl to
be exceeded, and NH4 Cl begins to precipitate. This takes place in a separate precipitation tank.
Let CNH4 + = concentration of NH4 + ions originally present as NH4 Cl in soul-
ton.
Ccl = concentration of total Cl- ions originally present as NH4 Cl and
-

NaCl in solution.
CNa = concentration of total Na+ ions originally present as NaCl and
+

NaA (intermediate salt) in solution.

CHCO3 = concentration of HCO3 - ions.
-

C'NH4 += increase in concentration of NH4 + ions. by the addition of

NH4HCO3 .
C'cl = increase in concentration of Cl- ions by the addition of NaCl.
-

SNH4 Cl= solubility product of NH4 Cl.

SNaHCO3 = solubility product of NaHCO3 .
Then (CNH4 + CNH4 +) x Ccl- < SNH4 Cl at 35 .;
While CNa+ x CHCO3 - > SNaHCO3 at 35 .
However, when the solution is cooled and salt introduced, additional Cl- ions are formed, NaCl
Na+ + Cl-.
Then (CNH4 + + C'NH4 +) (Ccl- + C'cl-) > SNH4 Cl at 15 .
Then and only then will ammonium chloride precipitate. Further, it can be seen that by the
displacement of ammonium chloride from solution with an amount of NaCl equivalent to the
amount of ammonium bicarbonate previously added to the solution for the precipitation of
NaHCO3 , the original concentrations of the various ions present in the solution prior to the sodium
bicarbonate precipitation are restored, so that the treatment of the liquor (after filtration) with
ammonium bicarbonate may be repeated and the cycle of operations started anew, and so on
indefinitely. Also, the volume of the filtrate is substantially the same as that of the original
solution, unless too much wash water has been added or unless too much mother liquor has been
removed with the sodium bicarbonate or ammonium chloride crystals filtered off. This
increment of Cl- ion concentration, coupled with cooling, is all that is necessary for the
precipitation of NH4 Cl. But it is apparent that the presence of too much of the so-called
intermediate salt (NaA) is undesirable, bec ause of the depletion of Cl- ions, causing difficulties in
the precipitation of NH4 Cl and because of the presence of high intermediate salt and sodium
chloride in the ammonium chloride crystals formed, although it makes little difference as far as the
sodium bicarbonate precipitation is concerned. On the other hand, if too little intermediate salt
(NaA) is used, assuming that the solution is still saturated with respect to heth NaCl and NH4 4Cl
under these conditions, the preponderating amounts of NH4 Cl will suppress the equilibrium of the
following reaction, causing the reaction to be less complete:
NH4 HCO3 + NaCl NaHCO3 + NH4 Cl
or ultimately causing NH4 CI to separate together with NaHCO3 , contaminating the sodium
bicarbonate obtained. Hence there is an optimum range in which the intermediate salt should be
present for the reaction to take place in the presence of both NaCl and NH4 Cl at high
concentrations, if pure products are to be obtained. This concentration of the intermediate salt is
generally from 6 to 10 per cent by weight, depending on the character of the negative ion and its
equivalent weight. Precipitation of sodium bicarbonate or ammonium chloride may be said to be
due to this quantity of the intermediate salt present, which first assumes the form of an ammonium
salt and then of the corresponding sodium salt. The whole situation may be summarized trus: In
sodium bicarbonate precipitation the reaction takes place virtually as follows:
NaA + NH4 HCO3 NaHCO3 + NH4 A
whereas in ammonium chloride precipitation ,essentially the following reaction occurs:
NaA + NaCl NHCI + NaA
thus the condition is restored and the cycle may be repeated indefinitely. In this way, the
intermediate salt acts as a buffer for the exchange ,oscillating between the ammonium and sodium
salt as the cycle is repeated. Such is the role that the intermediate salt plays in this ingenious
combination .It can be seen that the anion A-may be anything other than the CI-ion ,,provided that
its sodium and ammonium salts are sufficiently soluble ,and provided that the anion of the
intermediate salt does not form an insoluble salt with calcium ,magnesium barium iron etc. which
may be present as impurities in the crude salt used. In practice sodium nitrate (NaNO3 ),sodium
sulfocyanate (Incans),sodium sulfate (Na2 SO4),etc or their corresponding ammonium salts have
been used; any one of these may serve the purpose equally well, apart from certain practical
difficulties attending the operation, and from their economic considerations. Sodium or
ammonium phosphate cannot be used because it would be precipitated by calcium, magnesium,
etc. in the crude salt used, thereby causing loss of the phosphate; and sulfate cannot be used when
the crude salt contains much barium chloride as impurity. Other conditions being equal, the lower
the equivalent weight of the intermediate salt, the better. If the anion of the intermediate salt is Cl-,
then the processes is nothing more nor less than the normal Solvay Process, in which case, with
the simultaneous presence of large NH4 Cl and NaCl in the liquor, decomposition of NaCl by
NH4 HCO3 would be greatly impeded,
as mentioned above.
For the process to work successfully, it is necessary also to have as large a difference (or
interval) as possible between the temperatures at which sodium bicarbonate is precipitated and
that at which ammonium chloride is precipitated. Too small a temperature interval would cause
difficulties in getting the two products, ammonium chloride and sodium bicarbonate, separated out
individually, thus contaminating sodium bicarbonate with ammonium chloride, and vice versa. On
the other hand, there are practical difficulties in trying to make the temperature interval as large as
might be desired. For, if the temperature of the NaHCO3 precipitation is raised too high, say above
40 ., decomposition of NaCl would be less complete because of the increased solubility of
NaHCO3 at the higher temperature. Moreover, the decomposition of NH4 HCO3 to CO2 and NH3
would be excessive, and the loss of NH3 in the reaction, as well as that due to exposure of the
warm liquor to the air, would be greatly increased. On the contrary, if we set the precipitation
temperature of NH4 Cl too low, say much below 20 ., we are confronted with difficulties in
cooling the liquor with the temperature of the cooling water generally available, unless
refrigeration is employed. As a rule, the crystal formation at a low temperature is poorer,
although the solubility of the crystals is much less, the efficiency of conversion better, and the less
less. In plant operation, the temperatures for the precipitation of sodium bicarbonate and
ammonium chloride are 40 and 25 ., or 35 and 20 ., or 30 and 20~ ., respectively. For
example, for laboratory test, take 5 liters of 8 per cent Na2SO4 solution and saturate it with table
salt at room temperature by stirring into it about 1600 g of NaCl. Bring the temperature to 40 .,
and filter off any undisclosed solids. Add to the solution 125 g of NH4 HCO3 (air-dried crystals)
per liter of the solution in small portions very slowly; keep stirring and filter off the precipitate of
NaHCO3 . Thoroughly wash the sodium bicarbonate crystals on the filter. Cool the filtrate to 20 ,
and introduce finely pulverized table salt (-40 mesh or finer) to the filtrate in the ratio of 85 g of
salt per liter of solution, and keep the agitator constantly working. The precipitate of NH4 Cl* is
centrifuged off and washed with a small amount of water. This filtrate is now heated to 40 .
again for sodium bicarbonate precipitation, as before. The cycle may be repeated an indefinite
number of times, it being essential to filter the cake dry, returning all mother liquor to the filtrate;
to regulate the concentrations of various ions in the liquor; and to keep the volume of the mother
liquor at each operation constant replenishing the loss of mother liquor with s little strong wash
liquor each time. A little intermediate salt is also added from time to time to replenish its loss in
the cycle.
*In the first few cycles, little ammonium chloride would separate because the solution is not yet
saturated with this salt.
In practice, there are a great many "kinks" attending the operations, which can be appreciated
and guarded against only by those of long experience in the working of the process. For instance,
if one does not regulate the relative concentrations of different ions in the liquor, one may get very
impure or unsolvable products. If one has net used proper quantities of ammonium bicarbonate
or salt per cu. m. of liquor, one may soon find the re]stave concentrations in the resulting liquor
out of the range, thereby permanently impairing the solution for subsequent operations; and this
could not be restored without special treatment. If for some reason the crystals obtained are too
fine, and cannot be filtered dry, much liquor would be entrained in the cake, and the inclusion of
much mother liquor in the cake may contaminate the product with many impurities from the
highly concentrated liquor, or may affect the relative concentrations of certain ions in the resulting
liquor because of loss of volume of the liquor. Consequently, either the concentration of the liquor
could not be maintained, or the volume thereof could not be kept constant, or the wash water
would be left on hand in disproportionate quantities, or the loss of the intermediate salt, or
ammoma, or both, would be excessive. Such conditions may spell failure for the undertaking, just
as much as in the early days in the history of the ammonia soda process before Solvay's time.
Again, the presence of NaHCO3 in NH4 Cl (or NaCl in NH4 Cl) may be caused by low Na+ ion
concentration during the previous precipitation of NaHCO3 , or by complete absence of
magnesium salts in the crude salt used; this may be avoided by having proper concentration of free
ammonia in the bicarbonate mother liquor before the addition of NaCl for the precipitation of
NH4 Cl (see below). With good experience, it is possible to follow the progress of the reaction by
inspecting the color change of the solution or by observing the change m the specific gravity of
the liquor. The pH values change somewhat, but this is not a good index because of very strung
buffer action and salt effect in the liquor.
*This free ammonia concentration concentration should have a certain definite value for proper precipitation of
ammonium chloride, while the extent of bicarbonation by passing CO2 gas into the liquor during sodium
bicarbonate precipitation varies according to whether the precipitation is carried out in open vessels or under
considerable pressure.
The key to success lies in so regulating the temperature, concentration and operating
conditions that reasonably eared crystals are obtained in order that the resulting sodium
bicarbonate may be filtered dry and free from impurities. This returns the mother liquor to the
system and keeps its concentration within the orbit for the subsequent cycles. It is found also that
coarse ammonium bicarbonate crystals are conducive to the formation of coarser sodium
bicarbonate crystals because of crystallographic isomorphism between the two. Also, slower rate
of addition of ammonium bicarbonate crystals to the solution causes the formation of coarser
sodium bicarbonate crystals, presumably because of the smaller number of nuclei on which
crystals may grow. For, upon the addition of a small portion of ammonium bicarbonate crystals, it
may be observed that at first they are dissolved and the solution remains clear. Here the
temperature decreases to a definite point, P (Fig. 121). Gradually, however, upon the addition of

FIG 121 Change of temperature during the addition of NH4 HCO2 to NaHCO3 prcipitataor.

further quantities, the solution becomes cloudy, the crystals appear, and the temperature starts to
rise to a maximum. When further quantities of ammonium bicarbonate are added, the crystals
grow coarser and settle more rapidly, and the temperature falls slowly and gradually until the last
portion of ammonium bicarbonate has been added. Under such conditions, the sodium bicarbonate
crystals are more easily filtered and are filtered drier. All such details cannot be entered into here,
but suffice it to Say that they have important bearing on the purity of the products obtained and on
the successful operation of the
process.
Under average operating conditions the sodium, bicarbonate obtained
should have [on dry basis) the composition as shown in Table 149.

Table 149. Composition of Sodium Bicarbonate.

1. NaCl 1 per cent or less
2. NH4 HCO3 3-4 per cent (which is recovered upon calcination)
$. Intermediate salt a small fraction of 1 per cent
4. NaHCO3 96 per cent
5. Free moisture 15-18 per cent H2 O when rotary vacuum filter is used
8. Decomposition of NaCl 91 per cent on the basis of NH4 HCO3 used;
DEVELOPMENTS OF AMMONIA SODA PROCESS
and the ammonium chloride obtained should have (on dry basis) the
composition as shown in Table 150.

TABLE 150. Composition d Ammonium Chloride.

1. NaCL 1-2 per cent
2. NaHCO3 A small fraction of 1 per cent
3. lntermediate salt a small fraction of 1 per cent
4. NH4 Cl 96-98 per cent
5. Free moisture 3-5 per cent when centrifuge is used
6. Yield of NH4 Cl 93-94 per cent on the basis of NaCl used.
The purity of NH4 Cl depends greatly upon the purity of salt used. Practically all the
impurities in the salt, such as Ca++, Mg++, etc., precipitate as CaCO3 , MgCO3 , etc., together with
NH4 Cl. In order to avoid contamination of NH4 Cl, a good grade of salt should be used, or
purification of brine for the winning of a high-grade table salt, is recommended. As far as the
operation is concerned, it is only possible to minimize the contamination of NH4 Cl from such a
source. Ammoniation (as mentioned above) of the liquor before the addition of salt is very
useful. This converts NaHCO3 held in solution at a higher temperature into normal sodium
carbonate, which will remain in solution upon cooling the liquor for ammonium chloride
precipitation. Further, this free ammonia present to the extent of only 6.3 per cent NH3 or less is
found to have also the tendency to hold in solution calcium and magnesium impurities in the salt,
so that to a large extent the impurities will not separate out with file ammonium chloride
precipitate. This may be explained by assuming the formation of some double salt of calcium
and magnesium carbonates which separates out only slowly, and which will not separate in time
with NH4 Cl crystals, or, if separated at all, will exist in small particles which may pass through the
screen when the NH4 Cl slurry is centrifuged. Thus, a purer ammonium chloride is obtained from
salt containing a small amount of Mg salts as impurities (.3 per cent as MgCO3 ), than otherwise
from salt, which contains no magnesium, salts. For it was found that without this small quantity of
Mg salts in the salt, the ammomum chloride obtained may contain high sodium bicarbonate,
sometimes as high as 7 per cent NaHCO3 .* The double salt formation then slowly separates out in
the settler provided for this purpose, thus eliminating from the ammonium chloride crystals most
of the impurities in the salt used. On the other hand, if the ammonia concentration is too high,
calcium and magnesium compounds will separate rapidly from the solution together with the
ammonium chloride crystals, thus causing all the calcium and magnesium in the halt to separate
with the ammonium chloride.
* See Loepmann, “Berichte der Gesellschaft fuer Kohlentechnik,” pp. 78-81, Band IV, Heft 1, 1931.
The presence in the liquor of free ammonia or ammonium hydroxide requires that it be again
neutralized with CO2 . This is done by introducing CO2 into the sodium bicarbonate precipitation
vessel (precipitator) after the addition of the last portion of NH4 HCO3 . The pressure of CO2 gas of
about 1 /2 atmosphere (gauge) above the liquor is sufficient for this purpose. This is in line with the
Solvay operation, but CO2 is introduced here for three purposes:
(1) The bubbling of C02 gas into the solution converts more of the CO3 -- ions back into
HCO3 - ions and further furnishes CO2 to push the equilibrium of reaction toward completion in the
conversion of NaCl to NaHCO3 by the mass action principle;
(2) This helps maintain the temperature of reaction by the heat liberated by neutralization of
ammonia by CO2 in order to secure good crystal formation, and also serves to stir the solution; and
(3) As the liquor has a tendency, in the course of the operation, to contain more and more free
ammonia (shown by its rising alkalinity) as the result of decomposition of NH4 HCO3 and
liberation of CO2 gas from the liquor, bubbling of CO2 gas here will counteract such a tendency.
In practice, about 10 per cent of CO2 required is passed in this way. The heat of neutralization
utilized in this way is desirable, inasmuch as the heat effect in the precipitation of NaHCO3 by
NH4 HCO3 is negative, i.e., the reaction is endothermic. This can be understood from the fact
that the heat of solution of NH4 HCO3 is -6.69 kg. Cal. per mol in a large quantity of water, while
the heat of solution of NaHCOa is only -4.30 kg. Cal. per mol. In fact, during the precipitation of
NaHCO3 the solution would be cooled about 3~ . but for the heat generated by neutralization.
In this connection it may be added here that the heat effect in the precipitation of NH,C1 by
NaCl is just the reverse, i.e., the reaction is exothermic, and there is a rise in the temperature of the
liquor when salt is being added; e.g., the heat of solution of NaCl per mol in 200 molls of water is
1.28 kg. Cal. as against that of NH4 Cl of 3.89 kg. Cal. per mol in the same amount of water. In
practice, a rise of temperature of more than 2 . is observed. Therefore, to keep the temperature
down it is necessary to pass cooling water at a good rate during ammonium chloride precipitation.
Losses hand consumption of raw materials in this process is comparable with those in the
Solvay process. The process is handicapped in that the precipitation operations of both NaHCO3
and NH4 Cl are not yet continuous, and larger planet beyond one hundred tons of soda ash per day
have not been practicable without multiplying the number of units greatly. Table 151 shows
losses and consumption of raw materials in this process under-average operating conditions.
TABLE 151. Consumption of Raw Materials.
1. Salt 1.23 tons per ton of soda ash made
2. Limestone 1.25 " " " " " " "
3. Intermediate salt 10 kg. " " " " " "
4. Loss of ammonia 3-4 kg. as (NH4 )2SO4 per ton of soda ash made.

The other items of consumption, such as steam, electric power, etc., are also not essentially
different although the ammonia loss may be somewhat higher because of smaller scale of
Operation. The quantities of cooling water required may be greater because of relatively
insufficient cooling in the NH4 Cl precipitators. Nowadays, the plant working on this process is
also conducted in a closed system throughout, and various ammonia gas washers and scrubbers
are in use.
Centrifuges of different types have been used to a great extent because more complete
separation of mother liquor from the crystals is essential (see below).
The consumption of salt given above is indeed very favorable, and shows considerable saving
as compared to 70-73 per cent conversion with 1.70-1.80 tons of salt consumed per ton of soda in
the Solvay process. In fact, it is possible to obtain average conversion efficiency in this process as
high as 95 per cent. This is an advantageous feature, and is of great value to a country where salt
supply is particularly expensive. Against. this must be mentioned that, whereas in the Solvay
process natural brine is used, in this process only solid salt can be used, which adds considerable
cost to the working up of brine for the dry salt crystals. Further, salt used in this process must be
pulverized to minus 40 meshes or thereabouts, and the power required is also an item in its cost.
But the fact that its conversion efficiency is high, coupled with the feat that soda here is obtained
as a by-product--if we view this process as a means of fixing free ammonia as ammonium chloride
for fertilizer--makes this process very interesting. It may also be pointed out that, whereas this
process requires as raw material ammonium bicarbonate crystals which necessitate additional
equipment and cost in their manufacture, the practice 'in the Solvay process is just to pass CO2 gas
into ammoniaeal brine, forming NH4 HCO3 in site, thus greatly simplifying the operation. However,
the sulfuric acid saved, which otherwise would have to be employed in fixing the free ammonia as
ammonium sulfate, of which the sulfate part. is only a sort of vehicle and has no value whatsoever
as a fertilizer, will considerably outweigh all such considerations and justify its adoption in
conjunction with the fixed nitrogen industry (synthetic ammonia industry), where ammonium
chloride can be sold as fertilizer. Here the happy combination exists, and here the nitrogen
industry with its large supply of waste. CO2 gas, or the gas and coke industry with its by-product
crude ammoma liquor, can join hands with the soda industry, partic ularly where salt is too
expensive for the establishment of a regular Solvay plant. The sodium bicarbonate here is
generally converted to caustic soda by wet "calcination" with steam, so that calcining in rotary
soda dryers to soda ash is eliminated, and furthermore caustic and mud from the eausticization
process may be utilized in the preparation of "ammonia lime"(see below).
Again, from the comparison of this process with the Solvay process, it is an undeniable fact
that the sodium bicarbonate eristic are likely not to be so large as those from the Solvay process,
because of lack of countercurrent character and lack of temperature gradient in such precipitators;
but with good regulation it is possible to obtain fairly satisfactory crystals that may be filtered
reasonably dry, washed well and calcined in rotary furnaces without difficulty, where soda ash is
desired. Under good operating conditions, the size of She crystals of sodium bicarbonate from the
Solvay process may be as large as 0.10-020 mm., whereas the size of the crystals of sodium
bicarbonate by this process is generally from 0.05 to 0.10 mm. Also, as the solution is so
concentrated (specific gravity 1.21-1.23, depending upon the quantity and nature of intermediate
salt used), i.e., as the liquor is so "heavily loaded" with soluble salts, the product from such a
liquor is likely to be contaminated with salts from the mother liquor included in the crystals. Here,
centrifuges serve well; either the automatic type, or the strictly continuous type, with robbercoated
baskets and screen, or made of Stainless steel material, such as KA2 SMo, has been found very
advantageous for large-scale production, as these centrifuges retain the minimum amount of
mother liquor in the crystals, But where rotary vacuum filters are used, as in the filtration of
sodium bicarbonate, the best arrangement is to have "double filtration," i.e., two filters installed in
series, the cake from the first filter being so. ended in a wash liquor; using a screw mixer. The
suspension is then registered on a second rotary filter, while fresh water is used for washing the
cake on the second filter, as shown diagrammatically in Fig. 122. These filters are so designed as
to be able to separate the wash water from the filtrate (mother liquor). In this way, sodium
bicarbonate containing only 0.10-.015 percent NaCl has been obtained.
One particular application of this process (as was mentioned above) is in connection with
working the sodium bicarbonate so obtained directly into caustic soda. This is done by directly
converting the bicarbonate into carbonate liquor used directly for cauterization. Here, smaller
crystals of sodium bicarbonate are of no consequence. For this purpose, crude sodium bicarbonate
from the filters is added to hot water or condensate to form a suspension, and is pumped into the
top of a column somewhat like a small distiller, while steam at moderate pressure (about 20-50 lbs.
gauge) is passed in at the bottom. This decomposes sodium bicarbonate suspension into a perfect
solution of soda ash and some sesquicarbonate, and the hot liquor is directly sent to the cauterize.
This is an application of "Wet Calcinations" (see Chapter XXI). A well-insulated tank must be
used for the storage of the hot liquor to prevent crys tallization of Na2 CO3 , or freezing. (For further
details, see also Chapter XIX on the manufacture of caustic soda by the lime process.)
2NaHCO3 + heat Na2 CO3 + CO2 + H2 O
The saving in the heat (steam) here is considerable, while the CO2 gas, after passing through the
condenser, is available for carbonation in a very rich form. It is likely that a little more lime is
required, because of incomplete decomposition of sodium bicarbonate. The causticized liquor may

FIG 122 Double filtration method for filtrate containing high soluble salts that would cause considerable impurities in filter cake.

then be concentrated for the manufacture of liquid or solid caustic, as usual.

 A word may be said of ammonium chloride as a fertilizer. It has been definitely established
that its fertilizer. properties are in many cases as effective as those of ammonium sulfate, barring
only certain plants, such as tobacco, citrus fruits and certain starch-forming plants which cannot
tolerate much chlorine in the soil .While ,in general ,excess chlorine is harmful to many
plants .small amounts seem to be beneficial in protecting them from excessive drying in drought
periods. The relationship of the chlorine content in tobacco leaf to its fire- holding capacity has
been established, a high chlorine tobacco leaf possessing fire-holding ability.
For crops in general ammonium chlorine has been found to have good fertilizing properties.
Like ammonium sulfate, it renders the soil acid, but when ammonium chlorine is used , acidity in
the soil may be washed away more readily than is the case ammonium sulfate. It may be mixed
advantageously with an alkaline fertilizer such as sodium nitrate, calcium nitrate or calcium
Cyanamid, or with pulverized limestone or dry caustic mud .Its relation to ammonium sulfate is
the same that of potassium chlorine to potassium sulfate, and so it is generally used as an
ingredient in mixed fertilizers. In the mixed fertilizer, such as 4-8-8 or 6-8-4,ammonium chloride
serves as the source of nitrogen just as well as ammonium sulfate, the only difference being when
ammonium chlorine is used for this purpose as the source of potash, potassium sulfate rather than
potassium chloride should be used. It is seldom that ammonium chloride is due alone,i.e. unmixed
with the other ingredient ,but for many crops and soils ,NH4 Cl may be freely substituted for
(NH4 )2 SO4 as a nitrogenous fertilizer, the reason that it has so for not been so used extensively is
due to the present high cost of this material.
Ammonium chloride is more concentrated and contains a higher percentage of nitrogen than
ammonium sulfate; consequently limestone or precipitin calcium carbonate sludge from the
caustic soda manufacture is frequently mixed with it so that it will contain about 16 per cent NH3
and 22 per cent CO2 This is sold under the name of “Ammonium lime”(Ammonkalk), about 50
per cent lime sludge to 50 per cent NH4 Cl being used.
For its manufacture, take the ammonium chloride crystals from the centrifuges and mix with
them about an equal weight of caustic mud (from the caustic filters) washed free from caustic. The
mixture, when still moist, is compressed to from noodles and chopped into short sticks to make
granular ammonium chloride for the convenience of drilling on the farm. It is then dried in a small
rotary furnace at a temperature on the farm. It is then dried in a small rotary furnace at a
temperature below 300 . The product is then sacked a sold to the farmers for fertilizer, such a
combination of CaCO3 with NH4 Cl ensures a neutral mixture and does not render the soil acid.
Often, however, farmers also mix some ammonium sulfate with it.
 Capter XXVIII

Chemical Analyses and Tests in Alkali Industry

I. Soda Ash Manufacture
This chapter deals only with the analyses of the typical raw materials and finished products
of the ammonia soda industry. The analyses of materials common to other industries, e.g., coal,
coke or water for boiler feed, will not receive consideration here, for they are discussed in
standard books on Technical Analysis.
We shall give in each case (1) routine laboratory analyses, and (2) rapid tests for control work.
The former are based on standard quantitative procedure emphasis being laid on accuracy and
reliability, while the latter, which only serve as a guide to aid operatives in the field, are generally
more rapid simpler but less accurate method .In the laboratory analyses, efforts have been made to
determine as many substances as possible in one porting of the sample in order to save time. In
places where two methods are equally well adapted, one possessing certain advantages over the
other, both will be given be and their metros or objections pointed out. In the control analyses, to
save calculation it is recommended that reagent solutions of suitable strength be employed
wherever possible so that ,with a predetermined size of sample the number of scoff the reagent
consume represents directly the percentage of the constituent sought.
IA.BRINE LABORATORY METHOD
Brine is almost the universal form in which salt is secured for use in the ammonium soda
industry, Where rock salt is the source of supply, it is always obtained in the form of brine,
whether it comes from mines or wells as natural brine, as in Chins England or by pumping water
down a bored hole to dissolve the salt below as in Syracuse N.Y., Barberton O., Hutchinson, Kans,
etc., in the United States. Where sea water is the source of supply, often solid crystals are first
obtained by solar evaporation to eliminate many of the impurities in sea brine, and the crude salt
so obtained is dissolved to make saturated brine .Sometimes concentrated sea brine is used.
In the routine laboratory analysis of salt the following determinations are made:
(a) Specific gravity and temperature
(b) Suspended matter
(c) Ferric oxide and alumna
(d) Cao
(e) Mgo
(f) SO3 in one portion
(g) Na2 Oand K2 O
(h) Total Cl-
(a) specific gravity and temperature. Specific gravity is usually measured by a
hydrometer(1.000-1.400 range )in a 500-cc.cylinder and the temperature is determined with a
thermometer at the same time .Specific gravity reading must always be accompanied by
temperature readings.
(b) Suspended matter. 500cc.of the natural brine or 100 cc. Of artificial sea brine before
settling is filtered through a previously prepared and weighed Gooch crucible or porcelain
filtering crucible and washed and dried at a temperature of 103 to 105 . to a constant
weight ,the residue on the filter is the suspended matter.
(c) Fe 2 O3 and Al2O3 .pipette 25cc.of the filtered sample into a 250-cc.volumetric flasks and
make up to the mark. Pipette 25cc.into a 350cc beaker and dilute to about 100cc.Heat to
boiling .Add 2 grams NH4 Cl and make the solution slightly alkaline with NH4 OH.Filter off
the precipitate by decantation wash ignite and weigh .The residue is the combined oxides of
Fe2 O3 and AI 2 O3.
(d) CaO. Concentrate the filtrate to about 10cc.cool and add 15cc 6N(NH4 )2 CO3 and 15cc.95
percent alcohol. stir vigorously and allow to settle in cold water for 30 minutes .filter off the
precipitate and wash with the (NH4 )2 CO3 reagent. Evaporate the alcoholic filtrate on a water
bath for SO3 and Na2 O determinations (see below).Dissolve the precipitate with a little dilute
hot HCI, pouring the acid repeatedly through the filter. Wash with hot water. make the filtrate
alkaline with NH4 OH, add 20cc.N/2 (NH4 )2 C2 O4 and boil. Set aside for 30 minutes. filter
ignites strongly and rapidly weigh the caO.
Or, if preferred , the CaC2 O4 precipitate after washing can be dissolved in dilute H2 SO4
solution .The solution ,after being heated almost to boiling ,can then be nitrated with N/10
KMnO4 .
(e) Mgo. Concentrate the filter from the CaO determination to less than 75cc.if the volume is
larger .cool and add 10cc.N (NH4 )2 HPO4 and one-third of the total volume of 6N
NH4 OH.Stir vigorously and allow to settle in a cold place for four hours or ,better,
overnight .filter ignite in a porcelain crucible and weigh the Mg2 P2 O7 .
Or, if preferred magnesium may be determined volumetrically by bringing the filtrate
from the CaO determination to 70 .,adding 10cc.8-hydroxyquinoline solution ,(25 g
8-hydroxyquinoline dissolved in 60 cc. Glacial acetic acid and diluted to 21.) for each per
cent MgO estimated to be present and then 4cc.conc.ammonia solution (sp.gr.0.90),and
stirring continually with a mechanical stirrer for 15 minutes. settle, filter and wash the pale
yellow precipitate with hot 3 per cent ammonia solution. After washing, dissolve the
precipitate in about 50cc.hot 10 per cent HCI solution, and when solution is complete, dilute
to 100 cc. And add 15 cc. conc. HCl (sp. gr. 1.20). Cool the solution to 25 . and add from a
burette 10 cc. KBrO3 -KBr solution (containing 1/6 equivalent of KBrO3 and KBr each) for
each percent of Mgo estimated to be present. Stir the solution for 30 seconds to insure
complete oxidation. Finally and 10 cc. 25 per cent KI solution, titrating the iodine with a
standard sodium thiosulfate solution and using starch as indicator. The difference in the
iodine values between the original KBrO3 - KBr solution and the back titration represents
magnesium.
(f) SO3 After alcohol is all evaporated from the (NH4 )2 CO3 filtrate in (d), take up with water,
acidify with HCl, adding 5 cc. dilute HCl in excess. Make the volume about 100 cc. and boil.
Add 5 cc. N BaCl2 in small drops with constant stirring. Digest on a steam bath for 1/2 hour.
Filter, ignite with moderate heat, and weigh the BaSO4 .
(g) Na2 O and K2 O Make the filtrate alkaline with NH4 OH Boil and precipitate Ba with a slight
excess of (NH4 )2 CO3. Filter by decan-tation and wash thoroughly. Evaporate the filtrate to
dryness in a casserole and gently ignite off all ammonium chloride fumes over a small flame.
Redissolve the residue with a little water and transfer to a tared platinum dish and ignite
again. Weigh the residue as NaCl and KCl combined.
Dissolve the combined chlorides and make the volume of the solution 100 cc. in graduated
flask. Take 10 cc. and dilute to 150 cc. in a 250-cc. beaker. Determine Cl in these combined
 chlorides according to Mohr’s method desoribed in (h) below. From the weight of combined
chlorides and the quantity of total chlorine present, Na2 O and K2 O can be calculated.
(h) Total Cl.. Pipette 10 cc. of the solution made in (c) for the Fe2 O3 and Al2 O3 determination and
make up to 150 cc. in a volumetric flask. Pipette 10 cc. and dilute to about 100 cc. Titrate
with N/10 AgNO3 solution to a very faint red coloration, using 3 drops of 15 per cent K2 CrO4
solution as indicator.
Other impurities, such as bromine, iedine, etc., that may be present in small quantities in
sea salt are not determined.
For control work in the brine department the specific gravity reading by mean of a
hydrometer is generally sufficient. A saturated C.P brine at room temperature has a specific
gravity of 1197. With impure brine such as sea brine the specific gravity may run as high as
1.205 to 1.215. Mohr’s method is frequently employed in determining the concentration of
brine used in operation. Operators often judge the degree of saturation of the brine by its
Cl-content, as the sodium determination cannot be carried out by operatives in the field.
Note.-Mohr’s method is sufficiently accurate with the size of the sample used if the volume of the solution
to be titrated is diluted to about 100 cc. And if the very faintest red coloration is taken as the end point. The
Volhard’s method, however, using excess of AgNO3 solution and titrating back with KCNS solution using ferric
alum as indicator is generally considered more accurate.
An alternative method for the determination of Cl - using dichlorofluouescein is given toward the end of
this chapter (see under “Alternative Laboratory Methods”). This method is accurate; especially when the
quantity of the chlorides present is small.
IB. Solid Salt. Laboratory Method
The same methods are used in the analysis of solid salt except that moisture and combined
water, or water of crystallization, are determined instead of specific gravity, and the insoluble
matter instead of suspended matter.
(a) Moisture and water of crystallization. Weigh quickly to the mearest milligram a 5-gram
portion of the crude salt sample. Dry to a constant weight, holding the temperature at 300 . In a
tared porcelain crucible until a constant weight is obtained. The loss of weight represents the
moisture and all water of crystallization.
(b) Insoluble matter. Weigh 10 grams of the crude salt and dissolve it in about 200 cc. of water .
Filter in a tared Gooch crucible or porcelain filtering crucible , wash thoroughly and dry at 103 to
105 . The residue is the insoluble matter.
For the other determinations, the same methods of procedure as given under brine analysis can be
followed, but notice the size of the portions taken for analysis.
For calculations of the individual salts, the basic and acidic constituents are combined as follows:
Combine all Na+ with Cl-; if Cl- is in excess, combine it with Mg+; if Mg+ is in excess, combine it
with SO4 -if SO4 - is excess, combine it with Ca++. If Na++ is in excess over Cl-, combine the excess
of Na++ with SO4 - . Then if SO3 is in excess, combine it with Mg++ and so on. Such combinations
are purely hypothetical, guided by what is thought to be probably present when the various salts
exist in solid form. Many combinations are possible with a given set of ionic constituents as
determined by chemical analysis. This calculation, however, serves as a check on the accuracy of
the work. If all constituents have been detemined except those which, even if present, can only
occur in very small quantities, the number of equivalents of the basic constituents should check
with that of the acidic constituents, so that after all possible combinations have been made in the
above calculation, no excess of either the acidic or the basic constituents should be unaccounted
for. The discrepancy should not be greater than 0.10 per cent in any case, and the total should
differ from 100 per cent by less than 0.15 per cent with the ordinary grade of salt.
Ammoniated Brine and Pre-carbonated Liquor
For laboratory analysis the following items are required:
(a) Specific gravity and temperature (f)Cl
(b) Suspended matter (g)SO3
(c) CaO and MgO (h)Total soluble sulfides
(d) Free NH3 (i)CO2
(e) Fixed NH3
Ammoniated brine is the brine in which ammonia gas carrying some steam CO2,H2 S, ect.,
has been dissolved to the proper strength for the manufacture of NaHCO3 and from which,
consequently, calcium has been precipitated as CaCO3 and magnesium as MgCO3 . This is
responsible for the presence of fixed ammonia in the ammoniated brine. Excess of CO2 is
dissolved in it in the form of (NH4 )2 CO3 . H2 S is dissolved as such, or as (NH4 )2 S or NH4HS from
combination with NH3 . Any ferric salts or iron scale are reduced to FeS and thrown out as
greenish-black particles. Ammoniated brine usually has a faint yellowish or greenish-yellow tint
due to a trace of FeS present in colloidal formation. Well-settled ammoniated brine should carry
no suspended matter, no calcium salts, a very small magnesium precipitate, and only traces of FeS.
There-fore these determinations, with the possible exception of Mg, are generally not necessary,
unless specifically required.
(a) Specific gravity and temperature. Determine the specific gravity of the ammoniated brine with
a hydrometer and take the temperature simultaneously.
(b) Suspended matter. If the suspended matter determination is requid, filter 500 cc. of the
ammoniated brine through a large, tared Gooch crucible or a porcelain filtering crucible, wash
thoroughly, dry at 103 to 105 . and weigh.
(c) CaO and MgO If calcium and magnesium determinations are required, pipette 25cc. of
the ammoniated brine sample and make the volume about 100cc. Bring to a boil and 10 cc.
N/2(NH4 )2 C2 O4 with constant stirring. Set aside for 30 minutes. Filter, ignite strongly and
weigh the CaO quickly.
For MgO determination, follow exactly procedure (e) under Brine Analysis.
(d) Free NH3 . Pipette 25 cc. of the original sample and make up to 500 cc. in a volumetric flask
for analyses (d) to (i). Pipette 25 cc. of this solution into a 250-cc. distilling flask provided
with a condenser, and add 50 cc. distilled water without adding any NaOH. Distill into a
250-cc. Erlenmeyer flask containing 100 cc. N/10 HCl until the volume in the distilling flask
is about 30cc. Titrate the excess of HCl in the Erlenmeyer flask with N/10 NaOH.
(e) Fixed NH3 . Introduce 25 cc. N/2 NaOH into the contents of the distilling flask after the free
ammonla determination and distill similarly into another Erlenmeyer flask containing 50 N/10
HCl. Titrate the excess of HCl similarly.
(f) Cl. Pipette 10 cc. of the above dilute solution. Neutralize the alkalinity with HNO3 using one
drop of 0.1per cent methyl orange as indicator. Introduce about 0.2 gram of C.P. CaCO3
powder and stir vigorously Dilute to about 100 cc. and titrate the CaCO3 suspension with
N/10 AgNO3 until a faint red coloration is just visible, using 3 drops of 15 per cent K2 CrO4 as
indicator.
(g) SO3 .Pipette 10 cc. of the above solution .Acidify with dilute HCI, using 5 cc. In excess.Bring
to a boil and add 5 cc. N BaCl2 with constant stirring.Filter the BaSO4 ,ignite with moderate
heat and weigh.
(h) Total soluble sulfides.Pipette 25cc.of the solution .Neutralize with dilute HCI in the cold and
immediately titrate with N/10 I2 solution ,using starch as indicator.**
(i) CO2 .Pipette 15cc. Of the solution ,and determine CO2 gas in the apparatus illustrated
(Fig.123), taking the temperature of the jacket water.Read the barometric pressure. One
cc.CO2 gas under standard conditions weighs 0.0019763 gram.

FIG 123 Apparatus for carbon dioxide determination by expulsion and abeorption.

Procedure : Place 15cc. Of the above solution in a 100-cc. Generating flask A. Run into this
flask 15cc. Of 6N H2 SO4 from funnel B. Heat to boiling ,and collect CO2 gas through the spiral
condenser C into the burette D provided with water jacket F; meanwhile cock I remains closed.
When all the gas has been driven out, fill the flask A and the condenser C with water through
funnel B so that water rises to cock H, Now close the cock H. With the leveling flask E, read the
volume of CO2 in the burette D. Record the thermometer reading T (thermometer K), and the
barometric reading P1.Open the cock I and bring the gas into the absorption flask containing
NaOH solution .Take the volume of the remaining gas in burette d again. Loss in volume V1
represents CO2 .Let a=aqueous tension at T .in mm.Hg.
For control work for the carbonated and green liquors in the tower house ,the ammoniated
liquor in the absorber house , or the filter liquor in the distiller house, the following determinations
are required:
(1) Free NH3
(1) Total Cl
(2) H2 S
(3) CO2
 The tests may be made in the field by men having no special chemical training .All they
require is rough information to guide their operations in the field, and that information must be at
hand quickly when desired. Simplicity and rapidity are of paramount importance. Frequently the
figures do not represent exact results ,but they give comparative values, which are sufficient to
enable the operators to fix and duplicate the operating conditions. All the results are calculated to
“titer” or the number of cc. Of any normal solution per 20 cc.sample.
Pipette 10 cc. of the strong liquor, mother liquor, or any other liquor, in a 100-cc. graduated flask
and make up to the mark.
(1)Free NH3 Pipette 10 cc. of the diluted solution and titrate against NH2 SO4 using methyl
orange as the indicator. The result multiplied 20 times gives the “free NH3 titer”
(2)Total Cl. After the free NH3 determination, drop in 0.2gram C. P. CaCO3 powder, dilute
the same portion to about 100 cc. Add 3 drops of 15 per cent K2 CrO4 solution and titrate against a
N AgNO3 solution until a faint red coloration is seen. The result multiplied by 20 gives the
“chlorine titer”
(3)H2 S Pipette 20 cc. of the original sample. neutralize with H2 SO4 until neutral to methyl
orange indicator. Titrate immediately against N/10 I2 solution, using starch as indicator. The result
divided by 10 gives the “sulfide titer”
(4)CO2 Determine the volume of CO2 dissolved in the liquor as follows (Fig.124).
Procedure: Pipette 10 cc. of the diluted solution into E and carefully introduce 10 cc. 6N H2 SO4
into the space outside E on the paraffin bottom. Place the stopper tightly. Level A and B. Record
reading on A Invert flask D and shake thoroughly. Immerse flask D in a vessel of water at room
temperature for a few minutes. Level A and B again and take reading on A. The increase in
volume represents CO2 .The result multiplied by 20 gives the volume titer of CO2 at room
temperature.
HIIA. Limestone. Laboratory Method
(b) Moisture (g)K2 O
(c) Silicious matter (h)Na2 O
(d) Total silica(i)CO2
(e) Fe2 O3 Al2 O3 (TiO 2 and P2 O5 ) (j)Cl
(f) CaO(k)SO3
(g) MgO

Preparation of Sample:
Samples of limestone are taken from each carload in about the same size, taking in different
varirties. The stones are then crushed in a small jaw crusher to about 1 inch in size and quarted
successively until about a 10-1b.lot so obtained. Instead of quartering ,a mechanical sampler, if
availed ,affords a more rapid and better means of sampling The sample id crushed to 4 mesh in
roller crushers and then pulverized to –20 mesh in a disc pulverize. Of this a 0.5-lb.portion is
taken and sent to the laboratory.
If the above equipment is lacking ,the samples of limestone from the carloads are broken by
hand, quartered as before, and the final portion crushed in a small jaw crusher operated by hand.
This crushed sample is mow pulverized in a porcelain mortar to 20 meshes.
(a) Moisture .Heat a 0.5-gram sample prepared as above in a tarred porcelain crucible at 105c. for
1.5 hours or to constant weight .The loss in weight represents moisture.
(b) Siliceous matter. Transfer the dried sample from (a) to a 250-cc.wide –mouthed flask. Add 30
cc.water and 25 cc.N HCl in small portions ,covering the mouth of the flask with a small watch
glass. When effervescence has subsided, trandfer to a 200-cc.casserels and evaporate just to
dryness .Dehyrate at 125c.for I hours. Take up with 10cc. 6N HCl,heat to dryness and dehydrate
once more .Take up with 10cc.6N HCl and 50cc.water,heat to boiling and filter. Wash the residue
with hot water until free from chlorides, collecting the washing with the main filter for the
determination of combined oxides see (d) below.
( c) Total silica. Fuse the siliceous matter with 4 times its weight of Na2 CO3 and 4 times its weight
of K2 CO3 in a platinum crucible. While it is still molten, tilt the crucible to spread the melt up the
sides, so that when cold the mass can be readily loosened. Transfer the mass to a 250-cc.casserloe,
washing the contents of the crucible into the casserole. Acidify very carefully with 6NHCI .When
effervescence subsides add 5cc.more.Evaporate to dryness and determine the silica by double
dehydration as in (b) above.
( d) Fe2 O3 and Ai2 o3 (TiO 2 +P2 O5 ).Combine the filtrates from (b) and (c) .Add 2 grams NH4 Cl and
make the solution very slightly alkaline with 6NNH4 OH.Heat to boiling .Filter by decantation
wash the precipitate and ignite. The residue is the combined oxides of iron and aluminum with
P2 O5 and TiO 3 if phosphorus and titanium are also present.
( e) CaO Heat the filter from the above and add 20cc.N/2 (NH4 )2 C2 o4 with stirring .Set the beaker
over a hot plate for 30 minutes. filter ,test the filtrate with a small amount of (NH4 )2 C2 O4 and
ignite first gently and finally strongly over a Meeker burner. Cool in a desiccators and weigh the
CaO quickly.
( f) MgO Evaporate the filtrate to not more than 75 cc.While hot, precipitate Mg(Nh4 )PO4 with
10cc.N(Na2)HPO4 and * of the volume of 1.12 NH4 OH.Stir vigorously and aloe to settle in ive
water for 4 hours or preferably in cold water overnight .Filter by decantation, wash and ignite the
precipitate in a tred porcelain crucible. The residue is Mg2 P2 O7 (g,h) K2 O and Na2 O Weight
another 0.5 gram portion and mix it thoroughly with an equal amount of nh4ci on a glassed paper.
Then mix with 8 times its weight c.p.CaCO3 powder, reserving a small portion o f CaCO3 for
“washing “ Place the intimate mixture in a platinum crucible and cover the crucible. Heat very
gently for 15 minutes so that no NH4 Cl fumes appear but the ammonia odor is noticeable. Then
heat with a very short flame for minutes so that only the lower third portion of the crucible
becomes red. Then cool and disintegrate the mass in 200 cc. Of water in a 500 cc. Casserole. Heat
to boiling and filter by decantation. Repeat the extraction and wash the residue until free from
chlorides. Acidify with 6N HCl and add 5 cc. In excess. Bring to boiling, add N BaCl2 solution
drop by drop until no further precipitate is formed, avoiding large excess. Make the solution
alkaline with ammonia and ad a slight excess of 6N (NH4 )2 CO3 . Concentrate the solution to about
100 cc., filter and wash with hot water. Add N/2 (NH4 )C2 O4 drop by drop and see if any
precipitate occurs. If so, add a small excess, boil and filter. Evaporate the filtrate to dryness and
ignite off NH4Cl in a casserole with a moderate heat. Take up with 10 cc. 6N HCl and re-ignite.
Dissolve with a little water, transfer and wash into a tared porcelain dish. Evaporate to dryness
over a steam bath, protection the dish from foreign matter. Finally dry in an oven at 105o C. The
residue is the combined chloride of sodium and potassium.
If sodium and potassium are to be determined separately, dissolve the combined chlorides
making the volume 150 cc. determine the chloride by Mohr’s method and calculate Na2 O and K2 O
from the two results.
(i) CO2 . CO2 may be determined in four ways, (1) by decomposition with an acid and
titration, (2) by evolution and measuring the volume of the gas as in (I) under Ammoniated Brine
analysis, (3) by expulsion of CO2 from a weighed flask and determining the weight of the CO2 lost
by difference, and (4) by evolution and weighing the absorbed gas. Of these the first is the least
accurate even when a non-volatile acid such as H2 SO4 is used for decomposition, because in
titrating back with an alkali, when the end point is registered even by methyl orange, a certain
amount of acid from hydrolyzable salts, e.g., Mg++, Fe+++, Al +++, Ti++++, etc., salts, which form
insoluble hydroxides with an alkali, will have inevitably been included. But its rapidity has much
to commend it for quick , tough tests.
The second method by the measurement of the volume of CO2 together with absorption in an
Orsat pipette is often used for determining CO2 in solution such as in ammoniacal liquors (see (I)
under Ammoniated Brine analysis). The result calculated from the volume of CO2 obtained is less
than 5 in 1000 or 0.5 per cent, too low. In the case of limestone, however, the very large volume of
CO2 liberated makes it difficult to handle, and for this reason the method is seldom used.
The third method by the expulsion of CO2 and weighing the loss in weight in various forms
of alkalimeters such as Mohr’s, may be used for limestone or alkali carbonates. This method gives
fairly accurate results (accurate to 0.1 per cent), provided that the gas, after liberation, is driven
out not by heating but by air bubbling in the cold.* the arrangement is shown in Fig.125.
Procedure: Weigh the thimble B and into it weigh about a 0.5-gram sample. Disconnect the
Erlenmeyer flask at G together with the rubber tubing and pinch cock, close the cock G and weigh
the Erlenmeyer flask with the acid, CaCl2 , and tube F, etc., complete. Place the thimble with the
sample into the solution in the flask. Tip the flask so as to overturn the thimble to allow the acid to
react with sample. Set aside for a few minutes. When the reaction has subsided, connect C, D and
E. Open the cock G. Apply suction for about 15 minutes. Disconnect at G. Weigh the complete set
of apparatus again. The loss in weight represents CO2 .

This method has drawbacks in that any other volatile gas set free by the acid will be included
in the CO2 determination.
The fourth method by weighing the CO2 absorbed is the most accurate (Fig,126). It measures
CO2 alone in the liberated gas, but unfortunately the method is long and tedious. It is used only
when accurate results are desired.
Arrangement of apparatus by the fourth method is as follows:

FIG 126 Abeorption train for carbon dioxide determination.

Procedure: Weigh a 0.5-gram sample into the generating flask and cover it with 30 cc. Water.
Connect the apparatus as in the diagram except that G, H is disconnected for weighing and that the
2N HCl is not put in the dropping funnel. While G, H are being weighed, apply suction to this
system to exhaust any CO2 from the air in the system. After weighing insert G and H in the train,
fill the dropping funnel with 20 cc. 2N HCl. Without opening the drop cock in the dropping funnel
apply suction to test if the system is air-tight. Then drop in 2N HCl slowly so that the gas bubbles
through the Gomberg bulb at a rate not faster than 2 bubbles per second. When all the acid has
been dropped in and no more CO2 is apparently liberated from the generation flask, heat the
generating flask very gently just to boiling and continue to draw air through the system to exhaust
all CO2 . Disconnect G and H, wipe clean and weigh. Remove the flame. The increase in weight
represents CO2 . One cc. CO2 under standard conditions weighs 1.9763 mg.
(j) Cl. Weigh another 0.5gram sample in a wide-mouth, 250-cc. Erlenmeyer flask, add 20 cc.
Water and 20 cc. 2N H2 SO4 slowly. When effervescence ceases, heat to boiling for 5 minutes.
While hot, make the solution slightly alkaline with 6N NH4 OH with constant stirring. Filter first
by decantation into a 250-cc. volumetric flask, and wash the precipitate in the Erlenmeyer flask
several times with hot water. Cool the filtrate and make the volume to the 250-cc. Mark. Pipette
100-cc. into a 250-cc. Beaker, and neutralize the solution carefully with a small amount if H2SO4.
Add about 0.2 gram pure CaCO3 . Determine chlorine by Mohr’s method on the usual way.
(k) SO3 . Weigh another 0.5-gran sample and fuse it with 4 times its weight of Na2 CO3 each.
Dissolve the melt by boiling in about 100-cc. Water in a 250-cc. Casserole with an excess of 6N
HCl added in small portions, avoiding loss by effervescence. Evaporate the solution to dryness
and bake at 130°C. For 1 hour; moisten with 5 cc. Concentrated HCl and bake again for 0.5 hour.
Add 100 cc. Of water and bring to boiling. Filter and wash the residue. Neutralize the filtrate with
6N HCl, add 5 cc. In excess and heat to boiling. Add 5 c. N BaCl2 in drops, stirring constantly.
Allow to settle on a hot plate for about 1/2 hour, filter and ignite slowly so as not to allow the filter
paper to burn in flames. Cool and add a drop or two of 6N H2 SO4 . Ignite for weighing.
In order to avoid unreliability in sampling, the mild of lime made from the limestone burned
is analyzed completely by the methods described. The amount of CaCo3 and MgCO3 calculated
from these results better represents the average, although the lime contains a small amount of ash
(1-1.5 per cent) from the coke introduced in the kiln.
IIIB. Rapid Analysis
When rapid analyses are required as in the case of exploring limestone deposits when 50 or
more field samples are to be tested and only comparative results are desired, the following
procedure is followed:*
Weigh a 0.5-gram dried pulverized sample in a wide-mouth 250-cc Erlenmeyer flask, wash
down with a little water and introduce 25 cc. N/2 HCl. Stopper the flask with a one-hole rubber
stopper carrying a glass tubing about 24 inches long. Heat very gently to boiling. Cool in water
and wash down the flask with a little water. When the solution has cooled, add a drop of methyl
orange and titrate the excess acid with N/5 NaOH, taking the first change from a pink to a
yellowish color or colorless as the end point. The loss in acidity represents CO2 .
To the solution add 10 cc. 2N NH4 Cl and a small amount of ammonia to give a distinct odor.
Heat to coagulate the precipitate. Filter through a tared porcelain filtering crucible, and wash the
residue and precipitate. Dry at 105o C. To a constant weight and weight and weigh. The difference
represents siliceous matter and iron, aluminum, titanium, phosphate, etc. The assumption is made
that the sum of CaCO3 , MgCO3 , the insoluble matter, and ammonia precipitate equals 100 per cent
in the dried sample.
LetC = %CaCO3
M = % MgCO3
R=% Residue and ammonia precipitate

0.025
Total CO2 as CaCO3= cc. N /2× × 100
0.5
= cc. N/2 HCl×5=A
 100M
Then C = A−
84.32

M+C+R=100%
M=5.38(A+R-100)
and C=A-1.19M
*Base on a paper by S.D.Averitt, J.Eng.Ind.Chem.,14,1139(1922)
The suitability of a new supply of limestone for ammonia soda manufacture is determined by
the content of CaCo3 , MgCO3 and siliceous matter in it. Excessive quantities of siliceous matter
tend to form fusible clinkers blocking the limekiln or attacking the kiln lining. The rapid analysis,
while not strictly accurate, is sufficient to determine the suitability of limestone for the ammonia
soda industry.
The physical appearance of limestone as regards color, luster, striations, structure, etc., varies
a great deal, depending upon the locality. Probably the age, manner of formation, etc., have much
to do with its physical characteristics.
In the lime department the operators may judge the strength of the milk of lime by means of a
specific gravity bottle, which is a tared bottle with 100 cc. Marked on it. The net weight in grams
divided by 100 represents the specific gravity. With the strength of milk of Lima as high as 250
grams CaO per liter, the suspension is so thick that the use of a hydrometer is not feasible. The
specific gravity determination is a very rough guide because the specific gravity does not represent
accurately the strength of milk of lime unless the limestone burned is of uniform quality, and of
uniform grade, which is seldom true even within the same quarry. In all cases the operators are
required to determine the actual lime content in order to check the specific gravity reading. For
this purpose, a 5-cc. Portion of milk of lime from a well-shaken sample is taken with a pipette
having the capillary tip cut off and a new mark made by recalibration. Titrate the milk of lime with
N H2SO4 using phenolphthalein as indicator. The first appearance of red coloration in the solution
is taken as the end point in the practical work..
Milk of lime is tested by the operators in the lime department, in the distiller house, and in
the causticizing room when milk of lime is used for causticization.

IV. Kiln Gas and Returned Gas from Dryers

The CO2 gas is analyzed for

(1) CO2
(2) O2
(3) CO
(4) N2 (by diff.), in the order named.
For this purpose, a 3-pipette Orsat apparatus is sufficient. Sodium hydroxide solution is used
for the CO2 determination, an alkaline solution of sodium pyrogallate for the O2 determination,
and acid or ammoniacal cuprous chloride solution for the CO determination. the methods of
preparing these reagents are as follows:
(1) sodium hydroxide: Dissolve 350 grams of NaOH stick in a liter of water.
(2) Alkaline sodium pyrogallate: Dissolve 350 grams of NaOH stick in a liter of water and then 50
 grams of solid pyrogallic acid.
CHEMICAL ANALYSES AND TESTS
(3) cuprous chloride:
(a) acid solution.* -- place a layer of copper oxide about 3/8 inch thick in a 2-liter bottle and a
bundle of 1/640 inch (or NO.20) copper wires about 8 inches ling each. Fill the bottle with
1.10 HCl so as to leave no air space when the stopper is put in place. Invert the bottle and set
aside with frequent shaking until the solution is almost colorless.
(b) Ammoniacal solution.----the acid cuprous chloride is made slightly alkaline with 0.90
ammonia and a bundle of copper wires are put in. The bottle is set aside as before.
The manipulation of the Orsat apparatus is simple and will not be described here. Where
hourly analyses are made, the above solutions must be changed not less than once a week. The
cuprous chloride and pyrogallate solutions must be tested more frequently to determine whether
they are still good.
At present, cuprous-ammonio format solution, especially in the cold, is considered to be a
very good absorbing agent for carbon monoxide.
When CO2 gas to be analyzed carries NH3 , such as in the furnace gas or tower exit gas, the
gas must be washed through a washing bottle containing H2 SO4 before entering the Orsat
apparatus.
V. Crude Ammonia Liquor
This is probably the cheapest source of ammonia for the ammonia soda industry. Ammonia in
the liquor is mostly in the form of (NH4 )2 S or NH4HS. The percentage of sulfide present should
be kept as high as possible . Concentration of ammonia in the liquor is from 16 to 25 per cent NH3
by weight. For the ammonia soda industry we are interested in the total ammonia content and the
sulfide (or total reducing sulfur). Calculate the sulfide arbitrarily to 60 per cent Na2 S and ammonia
to 25 per cent ammonium sulfate (25 per cent NH3 ).
(a) Total ammonia. Pipette 25 cc. Of the crude liquor into a 500-cc. Volumetric flask and make it
up to the mark. Place 10 cc. In a 150-cc. Distilling flask and distill with the addition of a small
amount of NaOH solution, absorbing the gas in 100 cc. 0.10N H2 SO4 in a 250-cc. Erlenmeyer
flask. Titrate the excess of acid with 0.10N NaOH, using methyl orange as the indicator.
(b) Total reducing sulfur. Pipette 25 cc. Of the above diluted solution in a 250-cc. Erlenmeyer
flask, acidify with 6N HCl and immediately titrate with N/10 I2, using starch solution as the
indicator. This represents sulfide, thiosulfate, etc.
(c) Sulfide sulfur. Pipette 25 cc. Of the diluted solution into a 150-cc. Erlenmeyer flask and dilute
to about 50 cc. Titrate with 0.10N ZnSO4 .
*The absorption power of acid solution is lower than that of the ammoniacal solution. See
“Absorption of Carbon Monoxide by Cuprous Ammonium Salts” by W. Gimp and I. Ernst, Ind.
Eng. Chem., Vol. 22, Apr. 1930, p. 382.testing the formation of the white precipitate by a drop of
ZnSO4 solution. Titrate until a drop from the clear solution on the stirring rod gives no immediate
brown spot on a piece of lead acetate paper(filter paper impregnated with 10 per cent PbAc2
solution).
(d)CO2 See (i) under Limestone analysis, or (i) under Ammoniated Brine analysis.
For control work, only the total ammonia and total reducing sulfur determinations are made.
VI. Sodium Sulfide
The function of a soluble sulfide in the ammoniated brunet is to give a coating of FeS in all
iron apparatus, which protects the surface of the iron from being attacked by the ammoniated brine,
because FeS is quite inert in a alkaline solutions. Sometimes the amount of sulfide in the crude
liquor introduced daily into the system is not sufficient. Some sulfide (usually a solution made of
60 per cent fused Na2 S) must then be added to the distiller feed liquor in order to maintain a good
color in the soda ash. Too much sulfide(from the use of too much Na2 S), on the other hand, is
decidedly harmful because excessive sulfide iron from the apparatus. Ammoniated brine having
high sulfide content will have a dark green coloration or residue from contact with the iron surface
in the apparatus, for instance, in the settling vats.
In the laboratory the percent Na2 S is tested. For this purpose some fragments of fused Na2 S
of about 5 grams weight are weighed and dissolved to make 500-cc.of solutions in a volumetric
sack. A 15-cc sample of this solutions placed in a 250-cc flask,10cc N NH4 OH and 5 cc N NH4 Cl
are added, and the solution I diluted to about 50 cc. proceed as in ( c)under Crude Ammonia
Liquor analysis above.
For control work in the field, the operators test the strength of the Na2S solution by titrating a
sample for total reducing sulfur as in ( b) under Crude Ammonia Liquor analysis. This, of course,
includes sodium thiosulfate, sodium sulfite, etc; besides sodium sulfide, if the former are present
in the fused sulfide.

VII Tower Draw Liquor(Mother Liquor)

This is the clear portion of the liquor from shish sodium bicarbonate has completely
separated out at the temperature of the draw. It is from the analysis of this liquor that we can
determine the percentage conversion of NaCl into NaHCO3 in the towers (“percentage
decomposition”). Unfortunately, the nature of the substances present in the mother liquor is
exceedingly complex and there are at least nine different combinations present in solution in
equilibrium with one another. The study of these compounds or combinations in their exact
relationship has formed the subject matter of a special chapter, and the exact methods of
laboratory analysis will not be discussed here (see Chapter XI). The following represents the
partial and approximate analysis commonly required in ammonia soda works.
(a) Specific gravity and temperature (f)SO3
(b) Free NH3 (or total alkalinity ) (g)Total Cl
(c )Fixed NH3 (h)Unconverted salt (NaCl)
(d) CO2 (i) percentage decomposition
(e) Soluble sulfide
(a) Specific gravity and temperature. These are determined with a 1000-1400 hydrometer and a
thermometer simultaneously.
(b) Free NH3 This, as usually carried out in the laboratory or in the field, comprises not only
NH4 OH, but also (NH4 )2 CO3 , NH4 HCO3 , NaHCO3 ,etc.
For analysis, pipette 25cc of liquor an make it up to 500 cc in a volumetric flask. Twenty-five
cc of this diluted sample are taken and the procedure given in (d) under Ammoniated Brine
analysis can be followed.

The two methods are really identical and the titration method is the shorter of the two.
(c ) Fixed NH3 Procedure (e) under Ammoniated Brine analysis is followed.
(d) Procedure (i) under Ammoniated Brine analysis is followed.
(e) Soluble sulfide (h) under Ammoniated Brine analysis is followed.
(f)SO3 procedure (g) under Ammoniated Brine analysis is followed.
(g) Total Cl. Procedure (f) under Ammoniated Brine analysis is followed.
(h) Unconverted salt (NaCl). This, as is usually carried out in the laboratory, comprises not only
NaCl as such but also an equivalent quantity of NaHCO3 from its reaction with NH4 Cl present in
the mother liquor.
NH4 Cl + NaHCO3 Nacl +NH3 +CO2 +H2 O

For analysis, pipette 25 cc of the diluted sample and evaporate it in a tared platinum dish.
Gently ignite the residue until no more ammonium chloride fumes are seen. Cool and weigh. This,
less the equivalent quantity of Na2 SO4 from (f) above, may be taken as NaCl, assuming
magnesium salts as entirely absent and all (NH4 )2 SO4 completely converted to Na2 SO4 by ignition
in the presence of NaHCO3
(i) Percentage decomposition. This represents the percentage of NaCl that has been
converted to sodium bicarbonate in the towers and is roughly calculated as follows:

All quantities must of course be expressed in the number of equivalent weights per liter or in titers.
It is at best an approximation but represents the method usually carried out in the laboratory of an
ammonia soda works.
For the filter liquor, the analysis follows the same scheme as for tower draw liquor and in the
draw liquor will show the extent of dilution by the filter wash water; hence the amount of water
used on the filters. This filter liquor can be sampled and analyzed at different points on its way to
the distillers to follow up the extent of dilution by the various condensates or liquors introduced at
these points.

VIIA. Tower Draw Liquor (Alternative Method)

Piper 10 cc of the clear mother liquor into a 150-cc evaporating dish, evaporate just to
dryness over a steam bath, and dry to a constant weight in an electric oven at 105 . Weigh the
contents of the dish. Ignite the contents over a low gentle flame until all ammonium chloride
fumes disappear. Cool and weigh again. the contents of the first weighing can be taken as NaCl +
NH4 Cl and that of the second weighing as NaCl, so that the difference is NH4 Cl and

VIII. Crude Bicarbonate from Filters and Refined Bicarbonate of Soda

The crude sodium bicarbonate from the carbonating towers, while a very pure product from
the commercial standpoint, contains a small amount of NH4 HCO3 , NaCl, NH4 Cl and sometimes
also MgCO3 . magnesium carbonate will be present if the brine contains an excessive amount of
magnesium salts, as sea brine does, inasmuch as it is very difficult to get rid of the last trace of
magnesium precipitate in the ammoniated brine. Calcium, however, is very seldom present,
because CaCO3 settles out quite readily in the settling vats. The small amount of NH4 Cl from the
mother liquor contained in the bicarbonate behaves in the ordinary analysis as if it were NaHCO3
during drying or distillation, or whenever the solution is heated.
NH4 Cl + NaHCO3 Nacl +NH4 HCO3 Nacl +NH3 +CO2 +H2 O
Therefore, in order to determine NH4 Cl, absolute alcohol is used to extract NH4 Cl from the
crude bicarbonate, the carbonates and bicarbonates of sodium and ammonium being very
sparingly soluble in absolute alcohol. From the alcoholic extraction excess alcohol and free
ammonia (if any ) are driven off by evaporation over a steam bath, leaving behind NH4 Cl. By far
the greatest part of the free ammonia “ in the mother liquor, or in the bicarbonate separating out
from the mother liquor, exists as NH4 HCO3. the evidence is that the mother liquor gives hardly any
red “coloration to phenolphthalein except on dilution with wart or on long standing. The fresh
solution of the crude bicarbonate filtered in the atmosphere of CO2 and carefully protected from
exposure to the air shows hardly any red coloration with phenolphthalein at first, but red
coloration gradually and rapidly develops on standing in contact in this solution, the red coloration
in the solution again disappears. This shows the absence of any large quantity of normal carbonate,
Na2 CO3 either in the mother liquor or in the precipitate of bicarbonate in the towers. much of the
Na2 CO3 found in the crude filter bicarbonate is formed during the short exposure to the air during
or after filtrating.
The complete analysis of the bicarbonate comprises the determination of :
(a) Total Cl (g)Bicarbonate CO2
(b) Total NH3 (h)Na2 O
(c) Fixed Nh3 (i)Moisture
(d) Total alkalinity (j)Insoluble matter
(e) SO3 (k)Fe2 O3
(f) Total CO2 (l)MgO
For analysis take 10 grams of the crude bicarbonate and dissolve it in cold water to make a
volume of 250 cc in a volumetric flask. Stopper the flask and shake until all has been dissolved
except possibly a very small dark residue.

(a) Total Cl pipette 15 cc o this solution .Exactly neutralize the solution with 6N HNO3 and
introduce 0.2 gram CaCO3 (C.P.power)
Make the volume about 100 cc and proceed as in (f)under Ammoniated Brine analysis.
(b) Total NH3 . Pipette 25 cc. Of the solution and determine the total ammonia by distillation as in
e for fixed NH3 under Ammoniated Brine analysis.
(c) Fixed NH3 . Weigh 3 grams of the crude bicarbonate and suspend it in 100 cc. Absolute
alcohol. Stir and finally filter by decantation. Repeat once more and wash the residue in the
beaker with a little absolute alcohol, filter off and wash again. Evaporate the alcoholic filtrate
over a steam bath until about one – third of the original volume is left, or until litmus paper
shows no trace of alkalinity. Distally into 50cc N/10 H2 SO4 contained in a 150 cc. Erlenmeyer
flask wit the addition of caustic sod absolution as in (e) under Ammoniated Brine analysis.
The H2 SO4 solution from the Erlenmeyer flask is evaporated over a steam bath to drive off
most of the alcohol until less than half of the volume is left. Cool and dilute t about 100 cc
and citrate the excess of H2 SO4 in this solution wit N/10 NaOH using methyl orange as
 indictor. Loss in acidity represents fixed ammonia from NH4 Cl.
Note:--Most of the alcohol should be driven off and the solution then diluted with water before
titration; otherwise the methyl orange indicator does not give a sharp endpoint in the strong
alcoholic solution.
(d) Total alkalinity. Pipette 10 cc of the solution prepared for analysis and citrate wit N//10 HCl,
using methyl l orange as the indicator.
(e) SO3 . Piipette 50 cc o ft solution, acidify carefully wit dilute HCl an dad 5 cc in excess. Bring
to a boil and determine SO3 with BaCl3 as in (g) under Ammoniated Brine analysis.
(f) Total CO2 Pipette 10 cc o the solution and determine CO2 as in (i) under Limestone analysis or
s in (i) under Ammoniated Brine analysis.
(g) Bicarbonate CO2 . Pipette 10 cc of the solution and add 50 cc. N/10 NaOH. Add a small
excess f neutral BaCl2 solution. Titrate the excess of the NaOH with N/10 HCl in the cold,
using phenolphthalein as the indicator. Loss N/10 NaOH represents bicarbonate, the CO2 from
which can be deducted from the total CO2 to give carbonate CO2.
(h) Na2 O Pipette 15 cc of the clear portion of the solution into a tared porcelain dish. Acidify
carefully with 6NHCl, adding a slight excess .heat to boiling and precipitate the sulfate with a
slight excess of BaCl2. Filter, make alkaline with dilute ammonia, and add a slight excess of
(NH4 )2 CO3 solution. Heat to boiling and filter. Evaporate the filtrate to dryness and ignite with
a low heat until no NH4 Cl fumes are visible. Take up with a little water, transfer to a tared
porcelain dish and carefully re-ignite. Cool and weigh the NaCl; whence Na2 O
For an alternative method, see the triacetate method, p 505
(i) Moisture Weigh a 5-gram sample of the crude bicarbonate, spread it in a stared porcelain dish
and heat gently, using a heavy sheet of asbestos with a hole cut in the center to project one
–third of the dish below the asbestos board. Finally heat at 300 to a constant weight with
frequent stirring. Cool and weigh rapidly. The loss in weigh less NH4 HCO3 and NH4 Cl from
(b)and (c) CO2 and H2 O from the bicarbonate determination in (g) represents moisture.
(j) Insoluble matter the dried sample from the moisture determination is dissolved in about 100
cc of water. The insoluble residue is filtered off and washed in a tared porcelain filtering
crucible. The crucible is dried at 105 and the residue represents insoluble matter.
(k) Fe2 O3 The filtrate from the insoluble matter determination is acidified with 6NHCl using
slight excess and oxidized with a little bromine. Add 2 grams NH4 Cl. bring the solution to
boiling, and make it slightly alkaline with NH4 OH. Filter off the precipitated Fe2 (OH) 3 , ignite
and weigh as Fe2 O3. or, use standard KCNS solution and compare the color with known
standard.
(l) MgO. The filtrate from the Fe2 O3 determination is concentrated to about 75 cc. Determine
MgO as in (e) under Brine analysis.
From the above calculate the percentages in the crude bicarbonate of NH4 HCO3 ,
NH4 CL,NaHCO3 , Na2 CO3 , Fe2 CO3 , MgCO3 , NaCl, Na2 SO4 , free moisture, and insoluble matter.
For control work, it is only required to know the “yield” of the bicarbonate and percentage of
NaCl, for which purpose the following method is usually followed in the reworks laboratory.
Take a 2-gram sample and dry it in a tared porcelain dish by heating to 300 with a gentle
flame and frequent stirring until the weight is constant. The represents the weight of soda ash from
the dryers obtains able from the bicarbonate and includes the impurities.
Dissolve the residue in about 100 cc of water and carefully neutralize with H2 SO4
Determine the chlorine by Mohr’s method according to (f) under Ammoniated Brine analysis.
Calculate to NaCl.
The weight of the yield less NaCl can be taken as Na2 CO3 from which NaHCO3 can be
calculated. Subtracting the percent NaHCO3 and per cat NaCl from 100 gives roughly the percent
free moisture. It will be seen that NH4 HCO3 , NH4 CL, Na2 SO4 , and insoluble matter (Fe2 CO3 ,
MgCO3 etc;)have been neglected in this control analysis.
The following volumetric method is a more convenient alternative for control work and is no
less accurate than the method just described: Weigh a 2-gram sample of the crude bicarbonate as
before and dissolve it in about 100 cc water. Heat the solution in a 250-cc. Erlenmeyer flash and
boil until no bubbles of CO2 are seen in ebullition, or until about half of the volume is left.(The
boiling operation is very smooth and requires little attention) Cool and citrate with N H2 SO4 , using
methyl l orange as the indicator. The alkalinity calculated as Na2 CO3 may be taken as the “yield.”
After the H2 SO4 titration, the solution is diluted to about 100 cc. Determine the chlorine as
before. Calculate to NaCl. The percent free moisture in the bicarbonate may be taken as roughly
equal to 100 less the percent NaHCO3 and percent NaCl as before.
In both of these control methods, NH4 CL is converted back to NaCl by NaHCO3 and left as
NaCl in the ash or in the solution
NH4 Cl + NaHCO3 Nacl +NH4 HCO3
NH4 HCO3 NH3 +CO2 +H2 O
Sampling. Sampling is a very important phase of chemical analysis. The analyses may be
made accurately and the results may be good to the fourth significant figure; yet sampling is
oftentimes an unknown factor. Frequently, the value of an analysis is impaired because of a
question as to correct sampling or as to whether or not the sample is really representative. In the
routine analyses on Soda Ash, Caustic Soda, Liquid Caustic, etc; many points need be observed. It
is evidently not possible to cover all possible cases in a chapter like this. Suffice it to mention a
few observations for some more important products.
A shipment of soda ash, whether in bulk or in gunny bags, should be carefully sampled by
taking samples from the outside layer as well as from the interior of the mass. All portions on the
surface or in the center, also both in lumps and in loose powder, should be included, and then an
average portion be secured by a mechanical sampler. soda ash is rather susceptible to the action of
atmospheric moisture and CO2 , and so care should be taken to avoid unnecessary exposure and to
place the sample in a glass-stoppered bottle . As an illustration of a difference in composition
between portions of soda ash in the center and on the outer layer in a bag, the reader is referred to
p 510 in the follwing chapter.
In the case of solid caustic shipped in drums, the mass of the caustic inside an individual
drum is generally not homogeneous. In the course of solidif ication of the molten caustic in the
drum, a partial segregation of some of the constituents usually takes place in the center
(conedepression) and at the bottom. It is necessary that suitable portions of the caustic from the
cone, from the sides, and from the bottom of the drum be included. Caustic soda is more likely to
take up Co2 and moisture from the air and it is exceedingly important that minimum exposure to
the air be permitted.
In the case of liquid caustic shipped in tank cars, transportation by rail over territories having
sub-zero temperatures usually causes separation from the liquid mass of solid crystals with various
degrees of hydration, and consequently a change in composition in the remaining liquid. Samples
should be taken only upon complete thawing with steam coils provided in the car, and after
complete mixing. Exposure to air causes very rapid absorption of the atmospheric CO2 and
moisture. Hence closed vessels are recommended even during weighing.
In the case of rofined sodium bicarbonate, shipment is generally made in paper-lined wooden
barrels, or in kegs similarly protected. Because of its high purity and its use in making drinks,
baking powder and pharmaceuticals, shipment in closed containers only is allowed. The fine,
loose crystals have a tendency to lose CO2 upon exposure to the air, and may also become lumpy
when in contact with damp atmosphere during storage or transit. Hence sampling must include
correct proportions of the outer and inside portions of the mass.
IX Soda Ash
The use of soda ash as a chemical has long been known. Historically, it is the father of many
industries. The test standards for purity are different in different countries. It is sometimes
expressed as the per cent Na2 O based on the incorrect atomic weight of sodium of 24 instead of
23, as in the Newcastle test in England. In America the ratio of the incorrect atomic weight of
sodium to the correct atomic weight was employed and a system known as the New York and
Liverpool test came into use. In Germany and Russia, the tests are reported as per cent Na2 CO3 . in
France and Belgium, the degree of purity is figured on the weight of H2 SO4 necessary to
neutralize 100 parts of soda ash (the Descroizilles degree). The available Na2 O based on the
correct molecular weigh of Na2 O(the Gay-Lussac degree) is becoming more and more widely
adopted. It is now customary to regard 99.16 per cent Na2 CO3 as 58 per cent ash, meaning 58 per
cent Na2 O on the correct atomic other alkaline products are expressed in these figures. Just how
such erroneous calculations based on wrong atomic weight came into use is a matter of historical
interest. We shall not discuss it here, except to give the relationship among the various degrees in
use in different countries.(See Table 152). Details are given in Table 163.

Soda ash absorbs moisture and CO2 rapidly. Different samples can only be compared on
water-free and bicarbonate-free bases.(See Chapter XXIX). This is the correct way to judge the
purity of a product. Samples received in the laboratory are first calcined and the analysis is then
made on the dried ash. Incidentally, this gives the free moisture and the bicarbonate — CO2 and
H2 O in the sample. The complete chemical analysis of soda ash includes:
Chemical Analysis Physical Tests
(a) Moisture (k) Density
(b) SiO2 (l) Degree of Fineness
( c) Fe2 O3 (m) Clearness of Solution
(d) CaO (n) Color
(e) MgO (o) Water Absorbing Power
 (f) SO3
(g) Na2 O
(h) Bicarbonate
(i) Carbonate
(j) Cl
Chemical Analysis. Laboratory Method
(a) Moisture and bicarbonate -CO2 and H2 O. Weigh a 10-gram sample and spread it in a tared
porcelain dish. Heat to 300 to a constant weight as in (i)under Crude Bicarbonate analysis.
The loss in weight less bicarbonate- CO2 and H2 O represents the free moisture.
(b) Si2 O. There is practically no silica in the ammonia soda ash. If it is necessary to determine Si2
O, dissolve the fried sample from the moisture determination and follow procedure (b) under
Limestone analysis.
(c) Fe2 O3 . Dissolve the dried sample inhabits 100 cc of water and carefully acidify with 6N HCl.
Or, take the filtrate from the Sao determination(b). heat to oiling, add 8 grams NHCl and
determine Fe2 O3 with 6N as in (k) under Crude Bicarbonate analysis.
Note-Avery accurate method of determining small amounts of ironies as follows: Take a 10 g.sample and dissolve it by

carefully adding a small excess of dill. HCl Heat the solution to boiling, reduce with SnCl2 solution drop by drop until just barely
colorless. Immediately cool in water completely and add an excess of HgCl2 when a silky fibrous precipitate is seen. Add 10 cc.
Cone. Phosphoric acid and citrate with N/10 KmnO 4 to a faint purple coloration. This is a modified Zimmermann-Reinhardt
method.

Instead of using KmnO4 solution as in the above mentioned Zimmer-mann-Reinhaet’s method,

the solution after the addition of the HgCl2 and of phosphoric acid may be titrated with 0.10N
K2 Cr 2 O7 solution, using 3 drops of 0.01 molar diphenylamine sodium suffocate solution as
indicator, until a drop of the K2 Cr 2 O7 solution turns the solution redish violet.
(d) CaO and (e) MgO. Follow procedures (e) and (f) under Limestone analysis.
(e) weight 10 grams of the original sample and dissolve in freshly boiled distilled water to make a
volume of 250 cc on a volumetric flask. Pipette 25 cc into a 250-cc beaker. Acidify slowly wit
6N HCl, adding 5 cc in excess and make the volume about 100 cc. Bring to boiling and
precipitate the SO3 with a slight excess of N BaCl2 . filter and ignite the precipitate as in (f)
under Brine analysis.
(f) Na2 O. The filtrate from the So3 determination (f) is concentrated to 15 cc and neutralized with
6NNh4OH. To the cold solution, add 155 cc 5N and 30 cc 95 per cent alcohols. Stir and set
aside for half an hour with frequent stirring. Filter into a 150-cc. Casserole and wash with a
little (NH4 )2 CO3. Evaporate off the alcohol from the filtrate over the steam bath and ignite off
(NH4 )2 CO3 and NH4 Cl. Dissolve in a little water and transfer to a tared porcelain dish.
Evaporate and ignite again, avoiding intense heat. Cool and weigh the NaCl; whence NaO
(g) Bicarbonate and (i) Carbonate. Pipette into a 100-cc beaker 10 cc of the clear portion of the
solution made for SO3 determination (f) dilute to 50 cc; and 5 grams C.P. NsCl and cool in ice
water. Add 2 or 3 drops of phenolphthalein and titrate the solution with N/10 HCl, keeping the
tip of the burette under the solution and moving the glass rod very gently. The end point is
reached when the red color just disappears. Add a drop of methyl orange, lift the burette tip
out of the solution , and continue the titration until the slight yellow color changes to pink.
(h) Bicarbonate* and (i) Carbonate. Pipette into a 100-cc. Beaker 10cc. Of the clear portion of the
solution made for SO3 determination(f), dilute to 50 cc. And 5 grams C.P.NaCl and cool in ice
 water. Add 2 or 3 drops of phenolphthalein and titrate the solution with N/10HCl ,keeping the
tip of the burette under the solution and moving the glass rod very gently. The end point is
reached when the red color just disappears. Add a drop of methyl orange, lift the breeder tip
out of the solution , and continue the titration until the slight yellow color changes to pink.
Note. This double indicator method does not give a very sharp endpoint with phenolphthalein. the red color of which disappears only

gradually in the presence of a large amount of NaHCO 3 in solution. With a little practice.Sundstrom’s method for determining bicarbonate i n

the presence of a large amount of carbonate (such as in soda ash )is both convenient and accurate.

Bicarbonate CO2 (alternative method, Sandstorm). Pipette 10 cc of the solution and

dilute to about 50cc. Titrate it with N/10 NaOH, using 10 per cent AgN O3 is seen. The end
point is registered by observing the slight dark color in the white precipitate formed
immediately and not on log standing. This value, when deducted from the total alkalinity
determined by titration with an acid using methyl orange as the indicator, gives the carbonate.
(J) Cl Pipette 10 cc of the solution made for SO3 date, oration (f) and neutralize carefully with 6N
add 0.2gram C.P. Stir and dilute to abbot 100cc. Titrate the chloride with N/10 AgNO3 , using
15 per cent solution as the indicator as in (f) under Ammoniated Brine analysis.
(k) Density. This is the apparent density i.e; the bulk weight of soda ash contained in a given unit
of volume. It is expressed in grams per liter or pounds per cubic foot. It expresses the
bulkiness of soda ash. Then again density depends upon how closely the ah is packed. In
practice the two extremes are taken — one is the weight when tightly packed or tapped until no
further shrinkage in volume is observed, and the other is the weight when loosely packed,
avoiding any exertion of pressure. The difference between the two is as follows: the light ash,
when tightly packed until no further shrinkage is observed, occupies a volume only 63 per
cent as great as when loosely packed. The density figures thus depend greatly upon how the
density is taken. The following table shows average figures for densities of various grades of
soda ash by the tow methods. The second method, however, is the one that is now generally
adopted, as it shows more nearly the space required by a consignment of soda ash during
transpiration or storage. (See Table 153)

In the elaborator the method of determination depends, of course, upon the system chosen. If
the loosely packed system is adopted, the procedure is to fill a known volume of a tared container
(say a brass cylinder of exactly 500 cc capacity), with a sample of soda ash, level the top off
without exerting any pressure, and weigh again. The number of grams of contents times two is the
density sought. The apparatus is arranged as in Fig.127.
Procedure: Fill the funnel A with the ash. Tip the flap cover B to allow the ash to flow into
the cylinder C below. With a large steel spatula, level off the top of the cylinder and weigh the
content. The weight less the tare times two is the density of the ash in grams per liter or kilos per
cubic meter.
(l)Fineness. Weigh 100 grams of the sample and put it through an 8-mesh screen. Weight the
portion that passes through the screen and the portion left on the screen. From the portion that
passes through a 12-mesh screen and that which is left on the 12-mesh screen. Repeat with
16-mesh screen, etc. the ash that is fresh from the dryers and that has not been weathered will
all pass through an 8-mesh screen. If it has been kept in storage or has taken up moisture
during transportation, large lumps may be formed. The screen analysis is either “differential”
or “cumulative” according to whether the result is expressed as the percentage retained by
each individual screen or as the total percentage retained by all screens up to a particular
mesh.
(m) Clearness of solution. This registers the amount of insoluble matter or “mud” in the ash in a
qualitative way and is generally due to the presence of MgCO3 . for determination, take 15
grams of the sample and dissolve it in 100 cc of water. Compare the turbidity of the solution
with known standards. Good ash should contain very little insoluble matter and the solution
should be practically clear and free from cloudy precipitate. At the bottom of the solution
there should not be present any red or dark particles. Magnesium that has not been completely
separated out as mud from ammoniated brine will cause cloudiness or even precipitate in the
solution.
(n) Color. Good as should be perfectly white. The presence of iron will cause red color in the ash.
Frequently iron is present as isolated red particles in the ash, which are detectable even by the
naked eye. Good-colored ash should not appear yellowish at all when solutions made as in
(m). the most severe test is to take 50 grams of the each and macerate it with just sufficient
water to form a thick paste. Any color in the ash will show in an exaggerated degree in the
 paste form. Light ash carries not more than 0.003 per cent Fe2 CO3 . Dense ash may carry a
little higher (0.004 to 0.005 per cent)
(o) Water-absorbing power. This test is sometimes made when the ash is to be made into blocks.
The water absorbing power is the maximum quantity of water that the ash will take before the
paste fails to set to a firm solid block on cooling. For testing, 100 grams of the ash is taken,
placed in a mortar and ground with the desired amount of water. During the curse of
continued grinding, the thin solution gradually thickens; and when the thickened paste is
poured into a mold, it should set or solidify to a firm block on cooling. The water-absorbing
power depends on (1) the purity of the ash (i.e.; per cent Na2 CO3 not includingNaHCO3 ), (2)
the freshness of the product, i.e.; how far it has been weathered or how much moisture and
CO2 it has taken from the atmosphere, and (3) the weather conditions, i.e.; in winter or in
summer seasons. In winter, good fresh ash will take as much as 135 per cent water on the
weight of the ash, but in summer 110-120 per cent water may be the maximum.
For routine analysis in ammonia soda plants only the following determinations are required:
(a) Moisture
(b) Insoluble matter
(c) Available Na2 O
(d) NaHCO3
(e) NaCl
(f) Na2 SO4
(g) Fe2 O3
(h) Density
In the works, moisture is to determine because fresh ash immediately after it comes from the
dryers contains no water.
(c) Insoluble matter. Weigh 10 grams of the sample and dissolve in 200 cc of water in a 350-cc
beaker. Bring the solution to a boil and filter through a tared filtering crucible and receive the
filtrate in 350-cc beaker. Wash thoroughly with hot water. Dry the crucible at 105 . The
residue is insoluble matte.
(d) Available Na2 O. cool the filtrate from above and make the volume 50 cc. Pipette 10 cc and
titrate with N/10HCl, using a drop pos methyl orange as indicator.
(e) NaHCO3 (Sundstrom’s method). Pipette 10 cc of the solution. Dilute it to 100 cc. Titrate with
N/10NaOH, using g10 per cent AgNO3 solution on a spot plate as an outside indicator. A drop
of the solution from the glass rod giving an immediate brownish cast in the bluish-white
precipitate with AgNO3 solution on the spot plate registers the endpoint.
(f) NaCl. Pipette 10 cc; neutralize with dilute H2 SO4 and dilute to 100 cc. Add 0.2 gram
C.P.CaCO. Determine chlorine by Mohr’s method as in (f) under Ammoniated Brine analysis.
(g) Na2 SO4 Pipette 100 cc and neutralize carefully with 6N HCl, adding 5 cc in excess as in (g)
under Ammoniated Brine analysis. Calculate as Na2 SO4.
(h) Fe2 O3 Weigh a 2-gram sample and dissolve it in about 50 cc of water. Add a slight excess of
6N HCl. Bring to a boil, filter if there is any residue, make the solution slightly alkaline with
6N NH4 OH, and determine iron as in (k) under Grade Bicarbonate analysis. When the
quantity of iron in the each is very small, a better way is to make the volume of the acidified
solution 100 cc and place it in a Kessler tube. Add 2 cc 10 per cent KCNS solution and
determine the percent Fe2 O3 by comparing with known standards made in the same way.
(i) Density. This is determined by one of the two ways described above, according to which
system is adopted.
For control work in the field, when the per cent Na2 CO3 and per cent NaCl are to be
determined, take 2.65 grams of the sample by placating this fixed weight on the scale pan,
dissolve the soda ash in about 25 cc water and titerate the solution with N H2 SO4. solution, using
methyl orange as the indicator. Twice the number of cc of the N H2 SO4. used is the per cent
Na2 CO3 in the ash. The same solution then is titrated. with 0.453 N AgNO3 solution as in (f) under
Ammoniated Brine analysis. The encumber of cc of the AgNO3 use is the percent NaCl in the soda
ash.
X Distiller Waste Liquor
The distiller run-liquor affords an index to efficiency of the whole process. From the
composition of the liquor may be calculated (1) the utilization of NaCl in the brine, (2) the
utilization o flame in the distiller operation,(3) the loss of NH3 in the distiller, (4) the volume of
brine, of distiller waste, etc; per ton of sod abash made, etc. the concentration of distiller waste
and consequently its specific gravity vary a great deal, depending upon the deficiency of operation
and the dilution of the mother liquor by filter wash water, steam condensate, and various
ammonia-bearing liquors at different stages. For complete analysis the following determinations
are required:
(a)Specific gravity and temperature
(b)Total NH3
(c )Excess CaO
(d)Total CO2
(e)Total Cl
(f)SiO2 and siliceous matter
(g)Fe2 O3 Al2 O3 etc.
(h)CaO
(i)MgO In one portion.
(j)Na2 O
(k)SO3

(a) Specific gravity and temperature. Determine the specific gravity of the cold sample with a
hydromever by stirring the suspension with a thermometer. Note the temperature at the same
time.
(b) Total NH3. Pipette 50 cc of the sample into a 250-cc distilling flask, using a pipette having a
large bore in the tip. Dilute to 75 cc. Determine ammonia by distillation in the usual way,
adding NaOH if necessary.
(c) Excess CaO .Pipette 100 cc of the sample and dilute to 500 cc un a volumetric flask. Pipette
10 cc of the suspension with a pipette having a large bore tip. Boil off any ammonia and then
dilute to about 50 cc. Titrate with N/10 Hcl, using phenolphthalein as the indicator. Take the
end point when the red color first disappears.
An alternative method for the determination of the excess CaO approximating the condition in
the distiller is as follows: Pipette off 10 cc of the well-shaken suspensions before, boil off
ammonia and place it in a 150-cc distilling flask, adding a small excess of neutral (NH4 )2 SO4
solution. Distill NH3 into the H2 SO4 solution, proceeding exactly in the same way as in the
 Fixed Ammonia determination.
(d) Total CO2 . Pipette 10 cc of the suspension and determine CO2 as in (i) under Ammoniated
Brine analysis.
(E) Total CL. Pipette 10cc. of the suspension made in © and neutralize with dilute HNO3 free
from chloride. Add 0.2 gram C.P. CaCO3 and dilute to about 100cc. Titrate with n/10 abno3
solution ,using 15 per k2 cro3 as the indicator in the usual way.
(f) Sio2 and siliceous matter. Pipette 10cc.of the original sample .Acidify with 6n hci and
determine sio 2 as in(b) under limestone analysis
(g) Fe 2 O3 Al2O3 etc, Determine the combined oxides in the filtrate from (f) above as in (d)under
Limestone analysis.
(h) CaO. Determine CaO in the filtrate from the Combined Oxide determination (g) as in
(e)under Limestone analysis.
(i) MgO and (j)Na 2 O .The filtrate from the CaO determination is acidified with an excess of 6N
H2 SO4 and ignited in a casserole until no dense fumes of SO3 come off. It is cooled
completely taken up with a little water, and transferred to a tared porcelain dish. It is then
evaporated to dryness and ignited as before. The residue is weighed as Na2 SO4 and MgSO4
The combined sulfates are dissolved in 75 cc/ pf water in a 250-cc. Beaker . MgO is
determined with Na2 HPO4 in the usual way.
The weight of combined sulfates less Mg SO4 from Mg2 P2 O7 gives Na2 SO4,from which Na2 O
is calculated.
(k) SO3 Pipette 25 cc. Of the diluted suspension made in (c) above, acidify with 6N HCl. adding 5
cc. In excess. Hilute to about 100 cc. Bring to a boil and filter. Determine SO3 in the filtrate with
BaCl2 in the usual manner.
For routine analysis in the plant, only the following are determined:
(a) Specific gravity and temperature
(b) Excess CaO
(c) Total NH3
(d) Total Cl
This last item total chloride is determined in order to estimate the degree of dilution and the
volume of distiller waste per ton of soda ash made.
XI.Heater Liquor
This is the sample of the liquor from the bottom of the heater, which enters the preliming
tank or lime still before adding lime. It should be analyzed for co2 in accordance with I under
ammoniatred brine analysis. This analysis should be made in the field as well as in the laboratory
to determine the work of the heater soaps to guide the distiller operation .The excessive amount of
CO2 left in the heater liquor cause a loss of availed lime in the lime still or in the prelimer.
The distiller feed liquor is essentially the tower liquor and can be analyzed in the same way
as the latter. Comparison of chloride concentrations in the distiller feed liquor and in the heater
liquor will show dilution by steam condensate in the heater. Comparison of chloride
concentrations in the mother liquor and in the liquor from “mud” pumped into the top of the
heater ,will show the volume of the distiller waste per ton of soda ash .The volume of milk of lime
required per tonic soda ash can be estimated from the fixed nh3 present in the disabler fed liquor
and the volume of the heater liquor ,of milk of lime mind of the lime still can be estimated.
 To check up possible leakage in the cooling cubes in all coolers ,condensers,
presenters ,etc.the exit cooling waster and sewer waste should be occasionally tested with nester’s
reagent .and if ammonia is found present ,the leakage should be traced and immediately attended
to .Wherever possible, arrangement should be made so that the cooling waster at the exit is under
a slight pressure so that ,should any leakage occur, it is the water that goes into the apparatus.
II.bicarbonate manufacture
For methods of analysis of solutions iced in the bicarbonate manufacture the procedures
given under the soda ash process are applicable .they will not be reposted here .The soda liquor
can be analyzed like any other naco3 solution with tests for the specific gravity and temperature.
the refined bicarbonate in the bicarbonate process can be analysis in the same way as the crude
bicarbonate from the filters in the soda ash process with the omission of the ammonia
determinations. The yield of bicarbonate from the filers or from the centrifuges is determined also
in the same way .however ,the product here is of higher surety than that from the soda ash process.
III.Chemical caustic manufacture
Caustic soda in the ammonia soda works is produced by the socalled chemical process i.e. by
cauterization with lime .This can produce a caustic of equal or higher purity than the electrolytic
has a high purity. This lime process furnishes roughly one half of the total caustic soda made in
the Unite state the remaining half being supplied by the electrolytic process.
XII.Caustic Liquors
These are liquors either from the cauterization tankers or at different stages of
concentrating .they are called weak or strong liquor according to their relative strengths. They are
analyzed in the laboratory for:
(a) Specific gravity and temperature
(b) Total alkalinity
(c) Na2 CO3
(d) NaCl
(e) Na2 SO4
(a) Specific gravity and temperature. These are taken with a hydrometer and a thermometer
as described previously.
(b) Total alkalinity
(c) Na2 CO3
Volume 250cc. In a flask . Take 10cc. Of the diluted solution and citrate with N/10 H2 SO4
using phenolphthalein as indicator. Note the reading Add a single drop of methyl orange and
continue the titration until the slight yellowish color changes to pink. The methyl orange
endpoint gives the total alkalinity. Calculate as per cent Na2 O. Twice the difference between
the methyl orange reading and the phenolphthalein reading represents Na2 CO3 .
(d) Nail .The solution from methyl orange titration can be further analyzed for chlorine by
adding 0.2 gram CaCO3 an distracting with N/10 AgNo3 with 15 per cent K2 CrO4
indicator in the usual way. Calculate to NaCl
(e) NaSO4 . Pipette 25 cc. of the solution made in (b) and (c)above, acidifies with 6N HCl and
determines SO4 with BaCl2 in the usual way. Calculate to Na2 SO4
In the caustic zing room for control work ,the specific gravity of the liquor and the percent
conversion only are required . The latter is determined by the double indicator titration as above
and the percent conversion is obtained by reference to a chart or table prepared for this purpose.
XIIA.Causitc liquor rapid method
This is a sample of the liquor being causticized and is taken from the causticizing tanks. it is
analysis for the excess of lime and percent conversion .The following is a rapid method for control
works in the field.
(a) Per cent NaOH
(b) Per cent Na2 CO3 In one sample
(c) Per cent Conversion
(d) Per cent Excess lime
(a) NaOH and 9b) Na2 CO3 .Pipette 5cc.of the suspension, using a pipette with a large bore
at the tip, and deliver onto a small filter paper on a porcelain suction fitter. Apply suction
and wash the residue rapidly with a little distilled water. Titrate the filtrate in the suction
flask with N/2HCl first using phenolphthalein and then methyl orange. The
phenolphthalein reading glass the difference between the methyl orange reading and the
phenolphthalein reading represents NaOH. Twice the difference between the
phenolphthalein reading and the methyl orange reading represents Na2 CO3
( c ) Per cent conversion
(d)Excess lime. Suspend the residue from the filter above in 100 cc. of distilled water in a
250-cc. beaker, taking with it the filter paper. Titrate the suspension with N/2 HCl, using
phenolphthalein as indicator. This represents excess lime.
XIII.Cautic mud
Cautic mud is the sludge from the causticization tanlscontaining precipitated calcium
carbonate, excess lime soluble alkalis and all the impurities in the lime. this is analyzed in the
laboratory for
(a) Total alkali
(b) Caustic lime, CaO
(c) CaCO3
(d) Moisture
(a) Total alkali and (b) Caustic lime. Take a 5-gram sample and suspend it in about 150cc. of
water with constant stirring . Titrate the suspension with N/2 HCl solution, using
phenolphthalein as the indicator. This gives the total alkali and Ca(OH) 2 Filter and wash
the residue by decantation Determine calcium in the combined filtrate by precipitating it
with (nh4)2 C2 O4 and subtract the CaO so found from the figure obtained by the above
titration. This gives the total alkali, and the calcium determination gives caustic lime
(c )Ca CO3 Suspend the residue from the above HCl titration in about 100 cc. of water .
Titrate the suspension with N/2 HCl solution using methyl orange as the indicator. Or,
suspend the residue in about 100cc. of water and add small excess of a known amount of N/2
H2 SO4 Boil off CO2 and titrate back with N/10 NaOH solution using methyl orange as the
indicator. This gives CaCO3
Two alternative methods for the total alkali determination are given below:
(a) Total alkali Leach a 25-gram sample of the sludge with about 50cc, of hot water allow
to settle and decant the clear solution. Repeat with another 50cc. of hot water . Finally
add about 75cc. of water heat carefully filter and wash. To the combined filtrates, pass in
CO2 gas for 10 minutes. Allow settling and filtering by decantation. Titrate the filtrate
with N/10 HCl ,using methyl orange as the indicator.
The other method is as follows:
(b) Total alkali Weight a 5-gram sample of the encaustic sludge into a 250cc. casserole.
Cover with about 25cc.of water and stir. Add 10cc. 5N(NH4 )2 CO3 solution stir and
evaporate just to dryness over a steam bath Break up the residue in the casserole and
repeat by adding another 10cc. 5N(NH4 )2 CO3 solution Add 75 cc. of water and heat to
boiling filter by decantation, washing the residue thoroughly . Titrate the Na2 CO3 in the
fitrate with N/10HCl solution using methyl orange as the indicator.
(c) Moiture. Weigh 5 g sample of the mud and dry it in an electric oven at 105 to constant
weight the loss in weight gives approximately water in the mud, neglecting the small
amount of CO2 that may be picked up by NaOH during drying.
In the causticizing room the amount of alkalis left in the mud is regulated by the extent
and efficiency of washing. This amount should be very small when mechanical washing
equipment is employed, such as a Dorr thickener.
XIV Caustic Salts (Fished Salts)
These are crystals of NaCl, Na2 CO3 or Na2 SO4 together with hydrates of NaOH
separating out from strong caustic liquor during or after evaporation on cooling they are
analyzed for:
(A) NaOH
(B) Na2 CO3
(C) NaCl
(D) Na2 SO4
For analysis, weigh 20 grams of the sample and the volume of the solution 500cc. in a
volumetric flask.
(a) NaOH and (h ) Na2 CO3 These are determined by the double indicator method in the usual
way, using gN/10 H2 SO4
(b) NaCl To the solution after total alkali titration and 0.2 gram C.P.CaCO3 and titrate with
N/10 AgNO3 using 3 drops of 15 per cent K2 CrO4 as indicator.
(c) Na2 So4 Pipette 25 cc. and acidify with 6N HCl adding 5cc. in excess. Heat to boiling and
determine SO3 with BaCl2 in the usual way. Calculate to Na2 SO4
XV. Solid Caustic and Liquid Caustic.
Liquid caustic is getting to be one of the standard products in comer

ce. Samples of solid caustic from caustic pots (76 per cent Na2 O)or liquid caustic of 48 to 50
be .From evaporators are analyzed for:
(a) Total alkali (d)Na2 Co3
(b) NaCl (e)Na2 So4
(c) NaOH (f)H2 O
Weigh quickly a 10-gram sample of the e solid caustic and make the volume of the
solution 500cc. in a volumetric flask.
(a) Total alkali. Pipette 10cc. of the solution and titrate with N/10 H2 SO4 using methyl orange
as the indicator . Calculate to Na2 O.
(b) NaCl the solution from the total alkali determination is titrated with N/10 AgNo3 for
chlorine as in I(c )under caustic salts (Fished Salts) analysis. Calculate to NaCl
(c) NaOH Pipette 10cc. of the solution , add 10cc N Bacl2 and titrate with N/10 HCl , using
phenolphthalein as the indicator.
(d) Na2 Co3 The difference between the total alkali and NaOH gives Na2 CO3
(e) Na2 SO4 Pipette 50cc. of the solution a determine Na2 SO4 as in (d)under Fished Salts
analysis .
(f) Water is usually formed by deducting the percentages of NaOH, Na2 CO3 NaCl and
Na2 So4 from 100.
XVA Solid Caustic or Liquid Caustic Laboratory Method
If SiO 2 Fe2 O3 Man and Cu in the solid or liquid caustic are to be determined, weigh a 10
g. sample of the solid caustic or 20 g. sample of liquid caustic and make the volume of the
solution 500 cc. in volumetric flask.
(a) SiO2 . Pipette 25cc. of the solution into a 200cc. casserole, acidify with 6N HCl and
determine SiO 2 by double dehydration as in (b) under IIIA Limestone analysis.
(b) Fe2 O3 Pipette 50cc. of the solution made above in a 250cc. beaker acidify with 6N HCl
and determine iron as in (c) under Chemical Analysis for Soda Ash, Laboratory
Method(p.484)
(c) Mn. Weigh a 5 g sample of solid caustic air 10 g sample of 50 per cent liquid caustic and
dissolve it in 50cc. water in a 200 cc. casserole. Acidify by adding 20cc.O4 and 5cc. of 70
per cent HclO 4 solution Evaporate until copious fumes of SO3 are evolved, continually
whirling the solution over a small flame. Cool completely and add cautiously 30cc.
distilled water . Filter off SiO 2 through a No. 44 Whitman filter paper, washing the
residue with hot water. Make the volume of the filtrate about 90cc. and 2cc. 85 per cent
H3 PO4 and 0.5 g KIO4 Boil gently for 15min . Cool the solution , transfer it to a clean
Kessler tube and make up to the e100cc. mark with 2N H2 SO4 which has previously been
boiled with a little KIO4
Prepare KmnO solution standards in a series of Nessleer tubes as in the test sample using
known strengths of KmnO4 and the same concentrations of H2 SO4 and H3 PO4 Compare the
sample solution with these standards and determine percent Mn in the caustic soda sample by
matching the color of the solution.
(d) Cu (Modified Low’s Method). Weigh the same size sample ad in ©and acidify with 6N
H2 SO4 Heat the solution to boiling , add 20cc. 10per cent sodium thiosulfate solution (see
Note below), and continue boiling for about hour . Filter on suction filter and wash
several times with hot water. Ignite the residue on a porcelain crucible, take up with a
little conc. HNO3 and ignite gently until nitric acid is all expelled and copper oxide
remains Dissolve the copper oxide with a small amount of 6N H2 SO4 with heating , and
transfer the solution to a 250 cc. Erlenmeyer flask . Neutralize with dilute Na2 CO3
solution until basic cupric carbonate precipitate persists on shaking Acidify with dilute
acetic acid cool and add 10 cc. 25 per cent KI solution . Titrate the iodine liberated with
0.1N sodium thiosulfate solution using starch as the indicator in the usual way (Note: If
the caustic contains much Fe2 O3 Na2 MnO4 or NaClO 3 sufficient sodium thiosulfate must
be added to reduce these oxidizing impurities before copper is precipitated.)
XVI Caustic “Bottoms”
They are analyzed in the same way as above but in addition Fe2 O3 etc. Determinations
must be made.
XVII Fused Calcium Chloride (Anhydrous 95 per cent Calcium Chloride)and Liquid
Calcium Chloride
These are products from the distiller waste and are analyzed for:
(a) Total chlorine , Cl (d) CaCl2
(b) Ca(OH) 2 in fused (95%) CaCl3 (e)NaCl
(c) CaSO4 (f)Water or moisture

Weigh rapidly 10 grams of 95 per cent CacL2 in a covered weighing tube and dissolve it
in water , making the volume 250cc. in a volumetric flask. In the cease of liquid calcium
chloride, weigh or pipette an equivalent amount and make up to this volume.
(a) Total Chlorine Cl Pipette 10cc. of the solution neutralize if necessary with a little dilute
HNO3 add a little C.P.CaCO3 powder and titrate with N/10 AgNO3 using 3drops of 15 per
cent K2 CrO4 indicator in the usual way.
(b) Ca(OH) 2 In the case of fused (95 per cent )CaCl2 pipette 20 cc. of the above solution and
titrate with N/10 HCL using phenolphthalein as the indicator Calculate to Ca(oh).
(c) CaSo4 Pipette 50 cc. of the above solution , ad about 100cc. of water acidify with dilute
HCl brings to a boil and precipitate BaSO4 in the usual way . Ignite with moderate heat
and calculate to CaSO4
(d) CaCl Pipette 20 cc. of the dilute solution ad d100cc. of water and 5 cc. dilute ammonia ,
heat to boiling and precipitate CaC2 O4 with 9nh4 )2 C2 O4 in the usual way . Deduct from
the total calcium found the calcium determined in (b) and (c) , and calculate the balance
to CaCl2 .
(e) NaCl Deduct from the total chlorine found in (a)the chlorine corresponding to CaCl2 in
(d)and calculate the balance to NaCl
(f) H2 O This can be found by difference.
IV Electrolytic Caustic and Bleach Manufacture
XVIII Cell Effluent or Weak Liquor
This liquor is analyzed for (a) NaOHand (b) NaCl if other elements are to be determined,
follow the procedure under Brine analysis.
(a) NaOH +Na2 CO3 Take 10cc. of the sample and make up to 250cc. in a volumetric flask .
Pipette 10cc. of the dilute solution into a 250cc. beaker and add 10cc. neutral H2 O2 to
decompose the hypochlorite. HNO3 solution using methyl orange as the indicator.
(b) NaCl After the NaOH determination, the same solution may be used for the NaCl
determination by diluting it to about 75cc. and adding about 0.2g C.P.CaCO3 power . Stir,
and determine chlorine by Mohr’s method in the usual manner.
An alternative method of determining chlorine capable of somewhat greater accuracy is
as follows :Pipette 10 cc. of the dilute solution and neutralize with 6N HNO3 Add an excess of
N/10 AgNO3 and titrate with N/10 KCNS solution using 2 to 3 drops of saturated ferric
ammonium sulfate solution as the indicator.
(c) CO2 Pipette 10cc. of the diluted solution and add 5cc. neutral N BaCl2 solution Stir and
titrate the NaOH, using phenolphthalein as the indicator. Subtracting this result from (a)
NaCH + Na2 CO3 gives Na2 CO3 , whence CO2 in the cell effluent may be calculated.
Table 154 is useful for reference work in the electrolytic caustic plant laboratory.
XIX Bleaching Powder and Bleach Liquor
 The bleaching power is sampled from drums or barrels by means of a sample tube 3 feet
long made of inch brass pipe with the end beveled to form a scoop . The pipe is driver into the
mass at different places and the composite sample quartered down to laboratory size. For
analysis, weigh 7.092 grams and suspend them in a liter of water in a liter volumetric. Flask .
Shake and pipette 25cc. of the suspension . Follow procedure(b)below. With this size of
sample and N/10 standard solutions , twice the number of cc. of the reducing solution used
represents the percentage of available chlorine in the bleaching powder.
In the case of bleach liquor; pipette 25cc. of this suspension for analysis.
Each of the above samples may be analyzed for (a) free lime and (b)available chlorine.
(a) Free Lime. Pipette 25cc .of the dilute suspension Add neutral H2 O2 solution to
decompose the bleach until the bubbling ceases. Titrate with N/10 HNO3 using
phenolphthalein as the indicator.
(b) Available Chorine (Pontius’ Method). Pipette 25 cc. of the diluted suspension into a
250-cc. Erlenmeyer flask and dilute to about 75cc. Add 3 g. NaHCO3 and titrate with
0.10/N KI solutions adding a little starch as the indicator. The end point is reached when
one drop of the KI solution turns the solution blue. Twice the number of cc. of KI solution
used represents percent available chlorine.
(1) 3CaOCl2 + 6NaHCO2 + KI 3 CaCO2 + 6NaCl + KIO3 + 3H2 O +3CO2
(2) KIO3 + 5KI + 6CO2 + 3H2O 3 I2 + 6KHCO3
Note that in the above titration 0.10/N KI solutions contain only 1/60 mol KI per liter.
XIXA.Bleaching powder or liquor bleach laboratory method
“Chloride of lime “ is represented by the formula CaCl , Cl ,H2 O, in which the theoretical
percentage of chloride is 49 .Commercial “chloride of lime “never reaches this figure, the usual
percentage of available chloride being 36-39 per cent Cl .This is because of incomplete chloride of
slaked lime in practice and also because of the decomposition of the calcium hychloride
formed .The analysis consist of determination of the following :
(a) Free lime (in “Chloride of Lime”)
(b) NaOH and Na2 CO3 (in soda bleach)
(c) Calcium or sodium hypochlorite
(d) Calcium or sodium chlorate
(e) Calcium or sodium chloride
For these analysis ,take 10g.sample of “chloride of lime “or liquor bleach and suspend it in
water, making the volume 1000 cc. in a liter volumetric flask.
(A) free time .pipette 25 cc. of the “chloride of lime “suspension made as above and determine
free as in a under XIX.
(B) NaOH and Na2 CO3 Pipette 25cc . of the liquid bleach solution made as above undetermined
sodium hydroxide as in (a ) and (c ) under XVIII.
(C) Hypochlorite. Pipette25 cc. of “Chloride of Lime” suspension or liquid bleach solution into a
250cc. Erlenmeyer and an excess of 0.10N standard sodium arsenate solution (see Note below).
Neutralize the solution with dilute HCl , stir constantly for 30 seconds and itirate back with 0.10N
I2 solution using starch as the indicator. This titration takes care of the hypochlorus acid only and
is not affected by HclO 3 . (Note :The standard sodium arsenite solution is made by dissolving 4.95
g . pure sublimed As 2 O3 in 150 cc. distilled water , to which have been added 15 g. Na2 Co3 . When
solution is complete., add 25 cc. N HCl and make the volume 1000cc. )
(d)Chlorate. Pipette 25 cc. of the “Chloride of Lime” suspension or liquid bleach solution into a
250-cc. Erlenmeyer flask and dilute to about 50cc. Add an excess of 0.10Nstandard ferrous sulfate
solution and immediately acidify with dilute H2 SO4 solution 10cc. in excess. Cover the flask ,
shake and titrate excess of ferrous sulfate with 0.10N K2 Cr O7 solution , using 3 drips of 0.10
molar diphenylamine sodium sultans the solution reddish violet . This titration gives the chlorate..
(e) Chloride . Pipette 25 cc. of the “Chloride of Lime” suspension or liquid bleach solution
into a 250cc. beaker and dilute to about 50 cc. Add H2 O2 solution to decompose the
bleach until the bubbing ceases. Heat just sufficient ferrous sulfate solution to reduce the
remaining hypochlorite and chlorate . Neutralize with dilute ammonia solution , heat to
boiling and filter off Fe(OH) 3 Cool and dilute the filtrate to abut 75 cc. and add 0.2g
C.P.CaCO3 powder. Titrate with 0.10N AgNO3 solution ,using 3 drops of 15 per cent
chlorine. Subtracting chlorine from (c )and (d)from this figure gives chloride-chlorine.
Calculate to CaCl2 or NaCl as the case may be .
XX. Electrolytic Cell Gases
The gases either from the alkali-chlorine cells or from chlorate cell sere to be tested to
determine their purity . Oftentimes, this is a very feasible and accurate way of
ascertaining the current efficiencies of the cells , e.g., the current efficiencies of a chlorate
cell as shown below.
XXA. Cathodic and Anodic Gases from Alkali-Chlorine Cells
The cathodic gas from an alkali-chlorine cell is essentially hydrogen, although there may
be a small amount of air. Hydrogen in the gas may be determined by combustion with a
known volume of oxygen(air)by passing the gas over gently ignited palladium asbestos at
temperature from 300 to 400°C. The gas mixture is passed back and forth three times
very slowly so that for the third time the gas is passed, the palladium will hardly glow.
Contraction of the volume is caused by the union of hydrogen and oxygen to from water
and two-thirds of the contraction in volume represent hydrogen.
The anodic gas is essentially chlorine, but there will always be some CO2 and perhaps also
oxygen. The gas being soluble in water, an Orsat apparatus using mercury or glycerin for
displacement is employed. 100 cc. of the gas are usually taken, and chlorine and CO2 are
absorbed in the first pipette containing 25 per cent NaOH solution. Oxygen is next absorbed in the
alkaline pyrogallol (sodium pyrogallate) in the second pipette. The balanoce of the gas may be
assumed to be nitrogen.
To determine chlorine, a separate sample of 100 cc. Is taken and shaken over 15cc. Saturated
KI solution placed on top of mercury in a glass bulb, provided with glass stop cocks. When the
cocks are closed and the bulb is shaken chlorine displaces I2 from KI and the iodine reacts with
mercury in the presence of Kl so that the red color of liberated iodine gradually disappears. The
remaining volume is again measured. The contraction in volume represents chlorine. Deducting
this from the Cl2 and CO2 determination by the Orsat analysis gives CO2 .
The gas volume is reduced to N.T.P. In all measurements, the barometric pressute must be
recorded. Where the gas is not dried, the temperature of the gas in contact with water or with a
solution must be Observed so that the pressure may be corrected for aqueous tension.
The rate of flow of the gas is measured by means of a gas meter. With the ordinary wet gas
meter, the anodic gas must be first scrubbed In NaOH and then washed in water, before it is led
through the meter. Thus chlorine and CO2 are removed and the volume measured is that of
Oxygen, hydrogen and air.
From the cell gas analysis and the rate of gas generation, it is possible to calculate the
calculate current efficiencies respectively, when the cell amperage is known. In the alkali-chlorine
cells, some chlorine and CO2 are dissolved in the electrolyte and it id therefore necessary to
determine these in the cell effluent. For the determination of dissolved chlorine and CO2 in the cell
effluent, the procedures given under XIX(b) and XVIII(a) and (c) may be followed. When the rate
of flow of the cell effluent has been determined, it is possible to find the quantities of chlorine and
carbon dioxide so dissolved per hour. From these the cathodie and anodic current losses, and
consequently current efficiencies, can be calculated, as will be shown in what follows.
XXB. Mixed Gas from Chlorate Cells
The sum of chlorine and CO2 in the gases, and then separately chlorine gas alone, are
determined as above. The rate of the gas generation is measured by a gas meter after the remove
of chlorine and CO2 by NaOH and washing so that only H2 ,O2 and possibly some air are
measoured. Determine O2 and then H2 in the gas respectively as in above.
Let Vg=volume of gases after Cl3 and CO2 have scrubbed, in liters per hour at N.T.P.
P=per cent oxygen by volume in Vg as measured.
Pn=per cent H2 by volume in Vg as measured
Pm=per cent Cl2 and CO2 by volume in the original gases
Pe=per cent Cl2 by volume in the remaining gas after chlorine has been removed
Call Vo=volume of original gases reduced to N.T.P. in liters per hour
F=per cent O2 by volume in the original gases
Fh=per cent H2 by volume in the original gases
Fc=per cent CO2 by volume in the original gases
Fh=per cent Cl2 by volume in the original gases
A=amperes of current passing through the cell
AlterNative Laboratory Method
Determination of Chlorine in Brine, Ammoniated Brine, filter liquor, Or other chloride
solution using dichlorofluorescein as adsorption indicator. Pipette an 25 cc. Of the sample and
dilute to 1000 cc. in a volumetric flask. Pipette an aliquot portion of 10 cc. into a 250-cc.
Erlenmeyer flask for analysis. Add a drop of phenolphthalein and make the solution slightly acid
with dilute nitric acid. Make the volume about 100 cc. Add 6-7 drops of dichlorofluoredcein
indicator prepared as in Note l Below, and triturate with N/10 silver nitrate solution with constant
stirring To coagulate the silver chloride formed until the silver chloride precipitate Turns pink and
the suspension of AgCl appears just faintly red.
Notes. 1. The indicator is prepared by dissolving 1 g. in 600cc. Of 95 per cent alcohol to
Which are added 25cc. N/10 NaOH. The volume is then made to 1 liter.
2. This method is accurate even when a very small amount of chloride is present (15-20mg.
Cl.
Per liter).* its endpoint is sufficiently sharp and with some practice may be readily recognized.
Compared with Mohr’s method, this method is generally more accurate and gives results with
0.55cc. of N/10 AgNO2 lassie. Moth’s method, this method is generally takes about 0.15cc. Of
n/10 AgNO2 in excess to Reach its endpoint. The solution to be triturated should not be more
concentrated than 0.01 N in Cl.
3. Unlike Mohr’s method, the solution need not be made exactly neutral. A slightly acid
Solution may be tolerated. This is particularly advantageous when triturating a chloride solution
that would hydrolyze to a certain extent.

Volumetric Determination of Sulfate (Hinman’s Method)

Preparation of samples. For crude salt weigh 10.g of sample and Dissolve them to make
250cc. Of solution in a volumetric flask. Pipette 25cc. For analysis. For brine or ammoniated
brine or filter liquor, pipette 50cc. Of sample and dilute to 250cc. In a volumetric flask. Pipette
25 cc. for analysis. For soda ash or crude sodium bicarbonate from filters, weigh 5g. of sample,
and dissolve in 100cc.of water for analysis.
Preparation of barium chromate reagent. Prepare the barium chro Mate reagent by
precipitating dilute barium chloride solution with potas Sium chromate solution at the boiling
temperature. Filter and wash the Precipitate with hot, dilute acetic acid, and then with warm water
until Free from chromates. Dissolve 3-4g. of dry BaCrO4 precipitate in 1000 cc. of N HCl in a
liter volumetric flask.
Procedure: dilute the portion taken for analysis to about 100cc. In A 250-cc.besaker if the
volume is not already 100cc. Make the solution Distinctly acid with dilute HCl. Heat to boiling
and slowly add a small Excess of the barium chromate reagent. Boil for 1 minute (or 5 minutes If
any carbonate is present). Digest on a hot plate for 15 minutes. To The hot solution cautiously
add CaCO3 (C.P. powder) in small portions Until present in slight excess. (use ammonia for this
neutralization if Much iron is present. Then boil off excess ammonia. See Note 3 below.) Cool to
room temperature and dilute to 250cc. In a volumetric flask. Filter through a dry filter paper,
rejecting the first 20-30cc. Of filtrate. Then take 100cc. Of the filtrate into a 250-cc. Erlenmeyer
flask. Add 2g. Kl and 15cc. Of 12N HCl. Stopper the flask loosely, shake well,
*Kolthoff, I.M., Lauer, W.M., and Sunde, C.J.,J.Am.Chem.Soc.,51,3273-7 (1929).
and allow to stand for 15 minutes. Titrate slowly with 02N sodiumThiosulfate, using the starch
indicator in the usual way.
Follow the same procedure, using 100cc. Of distilled water as a blank for correction.

Notes. 1. The principle is based upon the fact that although both barium sulfate and barium
Chromate are sparingly soluble in cold water, barium chromate is readily soluble in a dilute HCl
Solution. 1000cc. Of cold water dissolve 2.3mg. Of BaSO4 and 3.8mg. Of BaCrO4 but BaCrO4
Unlike BaSO4 may be held in solutions by HCl which displaces HcrO4 ﹣as it does H3 CO3 because
HcrO4 ﹣as an acid is comparable to H2 CO3 . If, then, a solution of BaCrO4 in dilute HCl is added
In slight excess to a solution containing SO4 ﹣﹣ions, BnSO4 is precipitated but BaCrO4 is not. If
The solution is then neutralized, the excess BaCrO4 is also precipitated leaving an amount of CrO4
In solution equivalent to the SO4 ﹣﹣originally present. After filtering off the precipitates.CrO4 In
the filtrate can be determined idiomatically.
2. this method is capable of giving accurate results if the solution taken for analyisis does not
contain more than 5mg.Of SO3 per 100cc. Of solution, provided certain conditions as herein men-
tioned ate observed. The result will be high if the barium chromate reagent contains other soluble
chromates. Since the solubility of BaCrO4 in hot water is appreciable, the result will also be high
if the solution is filtered hot. On the other hand, the result will be low if other chromate is pre-
cipitated from the solution with barium sulfate when the solution is neutralized (such as basic
ferric chromate). If the solution is not acid enough during the titration of CrO4 idiomatically
the reduction of chromate may not be complete, and hence the result is low. Then the return of
the blue iodo-starch color within a few seconds after the end point has apparently been reached
will be experienced. The reduction of chromic acid by hydroiodie acid is greatly accelerated by
high
hydrogen ion concentration, thus:
CrO4-+ 3HI + 5H+ > Cr ++ + 3I + 4H2 O
3. When iron nickel, or sine is present, the solutions cannot be neutralized with calcium car-
borate, because if any of these metallic salts is present, insoluble basic chromate may be formed
when the solution is boiled with calcium carbonate. Then too little CrO4 ﹣would remain in the
filtrate causing low results. In such a case. The solution may be neutralized with ammonia, using
a small excess, and then the solution is boiled until excess ammonia is expelled, and then filtered.
Alternative methods for the determination of (1) total ammonia,(2)CO2 and (3)sodium id
ammoniated brine, mother liquor, filter liquor and other brine liquors follow:
preparation of sample. Pipette 10cc. Of the strong liquor into a 25.-cc. Volumetric flask and
make up to the mark. Ten cc. If this dilute Solutions are pipette for the following analyses.
Total ammonia. Pipette 10cc. Of the dilute solution into a 250-cc.Erlenmeyer flask. Dilute
to 40cc., add a single drop of 0.1 per cent. Methyl orange indicator and neutralize with N/10 H2 SO
solution, adding 10cc. In excess. Bring the solution in the Erlenmeyer flask to a boil and
boil for 15 minutes gently to drive off CO2 . cool in cold water and neutralize with N/10 NaOH.
Add to the solution of 15cc. Of 37 per cent formaldehyde solution carefully neutralized with
N/10 NaOH, using 3 drops of 1per cent phenolphthalein solution as the indicator. Stir and triturate
with N/10 NaOH to a red coloration without further addition of phenolphthalein. The number of
cc. N/10 NaOH in this formaldehydes titration represents total ammonia.
Notes. 1. As the solution contains NH4 Cl it must be made acid with a considerable excess of
H2 SO4 before boiling off CO2 in order to avoid any possible loss of NH4 through hydrolyses of
NH4 Cl at the boiling temperature. Also the solution must be made neutral to methyl orange rather
Then phenolphthalein, in neutralizing back after boiling. If phenolphthalein id used here, the result
For total ammonia would be low. In standardizing N/10 NaOH with an acid, however, phenyl-
Ptyalin should be used as indicator.
2. The formaldehyde solution must be first made neutral to phenolphthalein, for methyl
orange color is destroyed by formaldehyde. There will be some uncertainty about the end point
when neutralizing it with NaOH. The best way is to take out 15cc. Of the 37 per cent solution and
dilute to about 50cc. Adding 3 drops of phenolphthalein indicator. Neutralize with N/10 NaOH to
a faint red coloration. Add this neutralized formaldehyde solution to the solution. The neutralized
formaldehyde solution contains enough phenolphthalein so that no more need be added.
3. Care must be taken to boil off CO2 completely from this strongly acid solution, otherwise
the result for total ammonia would be erratic. For this purpose the solution is boiled for 15 minutes
in the presence of an excess of H2 SO 4 . to avoid any accidental loss during boiling, the erleneyer
flask is fitted as follows(fig.128.).
Carbon dioxide. Pipette 10cc. Of the dilute solution into a 250-cc.Erlenmeyer flask and dilute
to 40cc. Add 30cc. N/10 NaOH, a blank Of which must be carried to determine the CO2 content of
the NaOH Reagent. Bring the content of the flask to a boil and gently boil off Ammonia for 20
minutes using the same arrangement as in the total Ammonia determination (see Fig. 128). The
volume left after boil Ingo is about 40cc. Rinse back through the glass tubing with a little Water,
and while the solution is hot, add 10cc. Of N BaCl2 . immediately Stopper the Erlenmeyer, shake,
and cool in cold water.

FIG 128

Flask for boiling off CO2 or NH3

To the cold solution, add 2 drops if phenolphthalein and exactly Neutralize with N/10 HCl.
Hen add a drop of methyl orange and triturate With N/10 HCl to the permanent end point of
methyl orange. The unmark of cc. N/10 HCl in this methyl orange titration represents CO2 .
Note. With an excess of NaOH CO2 will not be lost while ammonia is being boiled off. Any
Bicarbonate will be converted will not yield a voluminous precipitate of BaCO3 which cannot
readily By BaCl2 in a hot solution will mot yield a voluminous percolate of BaCO3 that cannot
readily Settle. Naturalization to phenolphthalein indicator in the presence of BaCO3 precipitate,
must be one in a perfectly cold solution and the endpoint must mot be overtopped. The methyl
orange Titration then gives a sharp end point and the final end point is taken when the red color of
the Methyl orange will not didapper after 20 seconds of continuous shaking.
Total sodium, Na2 O(Kahane’s Method). Pipette 10cc. Of the dilute Solution into a 150-cc.
Beaker. Add I drop of 0.1 per cent methyl red Indicator and exactly neutralize the solution with N
HCl solution, string to drive off most of CO2 . The volume will then be about 15cc. Or less Add
to the beaker 100cc. Of the unary-magnesium acetate reagent (see Note 3 below), stirring
continuously for 5 minutes. Cool in water and Lay aside for an hour or more, keeping the
temperature at 20°C. the Precipitate will be fine, heavy lemon-yellow crystals. Decant off as Much
as possible of the supernatant viscous solution without disturbing the precipitate. Add 25cc. Of
denatured alcohol (saturated with the Urany-magnesium-sodium triple acetate), stir and allow
settling again Decant off this solution similarly taking care not to lose any precipitate. After the
second decantation, add about 15cc. Of the denatured alcohol and filter the precipitate through a
tarred porcelain filtering crucible under Suction (as in the Gooch crucible arrangement). Wash out
the precipitate from the beaker and wash the precipitate again in the crucible in an electric oven at
108°C.for an hour. The content multiplied by 0.02022 represents Na2 O.
Notes. 1. The presence of ammonia, magnesium and chalkier, and also go the carbonates in
the Ammonia cal brine does not interfere if these directions are followed.
2. The unary-magnesium acetate is present as follows: take 32 grams (UO 2 ) Ao2. H2 O and 100
grime anhydrous MgAo 2 (or 150 grams MgAo 2 .4H2 O). Dissolve them in 500cc. (90 to 95 per cent)
denatured alcohol, adding 20cc. Glacial acetic acid and 400cc. Water. Warm to aid solution and
make up to 1000cc. In alter volumetric flask, when cooled. Let the solution stand overnight. Some
precipitate of the urinal-magnesium-solution triacetate will have separated out from the sodium
carried in the regents. Filter off the solution into a darkened bottle and keep it in a dark place for
use.
The precipitate renaming in the liter volumetric flask is covered with about 500cc. Fresh
Denatured alcohol. It is shaken and allowed to settle. The alcohol is filtered off and reserved for
Washing.
3. The supernatant unary-magnesium acetate reagent should be decanted off instead of
tempting to filter it through the curable, if it is run through the curable ,the filter becomes so badly
clogged that a long time is reducible. For the filtration operation. If these directions are followed,
the filtration rate of 50 to 100 drops a minute can be secured, and the whole operation is
accomplished in about 15 minutes. This determination for sodium is very accurate even if the case
of strong brine. The method is especially suited when sodium alone in the brine id to be
determined.
Oxidation-Reduction Methods Involving the Use of Cleric Ammonium Sulfate*
Many oxidation-reduction reactions may be utilized in determining Certain constituents in
the chemical analyses conducted in the ammonia Soda industry. The idiomatic method is a special
type of the oxidation Reeducation reaction. The other cases where oxidation-reduction reactions
Are involved ate the volumetric determination of iron in soda ash or in caustic “bottoms,” the
determination of sodium chlorate in the electrolytic caustic soda, etc. it has been found that cleric
ammonium sulfate as an excising agent possesses many advantages, among which may be
mentioned its keeping quality (less subject to decomposition on standing), its stability toward
elevated temperatures where the reaction must be carried out in hot solution, its high oxidation
potential, and its freedom from
*Willard, H.H., and Young, P.,J.Am.Chem.Soc.,55,3260 (1933).
Interference by chlorine in the titration of ferrous iron in a cold chloride Solution (in
contradistinction to potassium permanganate). However, Toward sodium arsenate the reaction is
rather slow, and so the addition of a small amount of osmium peroxide, OsO4, as a catalyst is
recommended. For this purpose, orthophenanthroline-ferrous sulfate is best used as the Indicator,
facing red coloration in the reducing solution, changing to pale blue coloration as soon as an
excess oxidizing agent is present.
The foregoing reactions are available as alternative laboratory methods and may be illustrated
as follows:
In the case of the calcium determination, CaC2 O4 precipitate is filtered, washed and
re-dissolved in a hot dilute H2 SO4 solution. The solution is heated to about 70°C.and excess if
0.10N ceria ammonium Sulfate solution is added with stirring for 30 seconds. A few drops of
Molar OsO4 solution and a drop if 0.025 molar orthophenanthroline ferrous indicators are added,
and the solution is nitrated with 0.10N standard resinous oxide solution until one drop of the
resinous oxide solution turns the solution pink.
In the case of he iron determination, after the reduction of the ferric To ferrous iron by SnCl2
and the addition of HgCl2 and then H3 PO4 after Cooling, the solution may be directly nitrated with
0.10N cerci ammonium Sulfate solution is the cold, using one drop of 0.025 molar orthophenan
Throline- ferrous sulfate solution as the indicator, until one drop of theCeric ammonium sulfate
solution discharges the pale-pin coloration.
In the case of the NaClO 3 determination in the electrolytic caustic Soda, to the sample portion
is added and access of standard resinous oxide solution. A citify with dilute H2 SO4 , adding a
slight excess. Add a few drops of osmium oxide solution as before, then heat to about 55 .
Titrate the excess of resinous oxide with cerci ammonium sulfate solution, using the
orthophenanthroline-ferrous sulfate as the indicator, until one drop of the cerci ammonium sulfate
solution discharges the pale-pink color of the solution.
U. S. GOVERNMENT SPECIFICATIONS
Soda Ash
Soda ash shall be of the following types:
TypeI. 58 per cent ordinary or height
Type . 58 per cent extra-light
Type III. 58 per cent dense
Soda ash shall be a high-grade anhydrous sodium carbonate in powdered form and shall be of
the type specified by the purchaser.
Soda ash shall conform to the following detail requirements for each type as Indicated:
(a) Chemical composition. Soda ash, types I,IIand III, after drying for hour at 150to
155°C.shall conform to the following requirements as to chemical Composition:
Total alkalinity, calculated as Na2 CO3 , maximum………………………99.2
Sodium bicarbonate (NaHCO2 ) maximum……………………………….. .5
Matter insoluble in water, maximum……………………………………. .25

(b) Loss in weight at 150 to 155°C. the loss in weight on heating about 2g.accurately weighed,
of the material (types I, IIor III) as received for 1 hour at 150to 155°C. shall not exceed
the following limits:
(a) If sample is taken at manufacturer’s works, maximum…………….. 1
(b) If a sample is taken elsewhere, maximum………………………….. 4
(c) Volume. Thirty grams of the soda ash as received shall have an apparent
Volume in milliliters for each type indicated, as follows:
Minimum maximum
Type I, light …………………………50 65
Type , extra-light……………………75 95
Type III, dense…………………………25 35
Methods of sampling, inspection and tests
When packed in cans or cartons-one can or carton shall be taken at random From not less
than 1 per cent of the candors’ shipping conditioners, provided such containers contain not less
than 50 pounds each. In the case if smaller containers a can or carton shall be taken at random
from each lot of containers totaling not to exceed 5000 pounds. The total sample shall in all cases
consist of not less than three cans or cartons taken at random from separate containers. With very
large Lots, where the sample drawn as above will amount to more than 20 pounds, the Percentage
of packages sampled shall be reduced, so that the amount drawn shall not exceed 20 pounds. Wrap
the individual cans or cartons tightly in par affined Paper at once and seal by rubbing the edges
with a heated iron. The inspector Should accurately weigh each wrapped can or carton, record its
weight and the date Of weighing on the wrapper, place the wrapped cans cartons in an airtight
container, which should be nearly filled, seal, mark, an send to the laboratory for Test. Samples
should be kept cool until tested. The seller shall have the option of being represented at the time if
sampling, and when he so requests shall be furnished with a duplicate sample.
When in bulk. A grab sample of not less than one-half pound shall be taken At random from
not lass than 1per cent of the vendor’s shipping containers, provided such containers contain not
less than 100 pounds each. In case of smaller containers a grab sample of not less than one-half
pound shall be taken at random from each lot of containers totaling not to exceed 10000 pounds.
The total sample shall in all cases consist of not less than three grab portions taken at random from
separate containers. with very large lots, where the sample drawn as above will amount to more
than 20 pounds, the percentage if packages sampled shall be reduced, so that the amount drawn
shall not exceed 20 pounds. The inspector should rapidly mix the sample, place in an airtight
container which shall be filled seal, mark, accurately weigh, record its weight, date if weighing
and place if sampling in the package, and sang to the laboratory for test. Samples should be kept
cool until tested. The seller shall have the option of being represented at the time of sampling and
when be so requests shall be furnished with a duplicate sample.
Preparation of sample. Rapidly mix the sample: if desired quarter down to about 1pound and
weigh out all portions for analysis at once. Unused portions of the sample used for analysis shall
be preserved in an airtight container in a cool Place. Note the condition of the sample. Analyses
for carbonate, bicarbonate and Matter insoluble in water shall be referred to theory basis or carried
out on Samples previously dried for one hour at 150 to 155 °C .
Total alkalinity as Na2 CO3 . weigh 5.30g of the sample and transfer to a 500-ml. Erlenmeyer
flask. Dissolve the sample in about 100ml, of distilled water add exactly 100ml if 1.0N H2 SO4
from a burette, place a small funnel in the neck of the flask and boil the solution gently for
5minutes to expel CO2 . cool rinse off the funnel and remove it, and then rinse down the sides of
the flask. Add 4 drops of methyl red indicator, and citrate back with 0.1N NaOH. Calculate the
total alkalinity as Na2 CO3 as follows:
0.1N NaOH
Na2 CO3 percent = ml. Of 1.0N acid –m of
10

Sodium bicarbonate weighs 4.20b of the sample and transfer to a 250-ml Beaker. Dissolve
the sample in about 100ml. Of distilled water at room temperature, and citrate with 0.5N NaOH
until a drop of the solution added to a drop of Freshly prepared AgNO2 solution (10 per cent) on a
spot plate gives a dark color instantly. Milliliters of 0.5N NaOH requited=percentage of NaHCO3 .
 Matter insoluble in water. Transfer 10g of the dried sample to a 400-ml Beaker, adds about
200ml of freshly boiled distilled water, and boil the solution for 10 minutes. Filter on a weighed
grouch crucible, wash thoroughly with hot water dry the crucible and residue at 105 to 110°C for
one hour, cool, and weigh. Calculate the percentages if matter insoluble in water.
Loss at 150 to 155°C. place about 2g accurately weighed, of the sample, without previous
drying, in a tarred wide –mouth short weighing tube provided with a glass stopper. Heat with
stopper, cool and weigh. Calculate the loss in weight to percentage.
Volume. Transfer 30g of the sample, without previous drying to a clean, dry 100ml.
Graduated glass cylinder, avoiding any packing. Rotate the graduate until the sample flows freely
and then, taking care to avoid jarring, level the surface of the sample and read the volume in
milliners occupied by the sample.

Packaging Packing and Marking for Shipment

Packaging. Unless otherwise specified, commercial packages are acceptable under this
specification.
Packing. Unless otherwise specified, the subject commodity shall be delivered in standard
commercial containers, so constructed as to insure acceptance by common or other carriers, for
safe transportation, at the lowest rate, to the point of Delivery.
Marking. Unless otherwise specified, shipping containers shall be marked with The name of
the material and the quantity contained therein, as defined by the Contract of order under which
the shipment is made the name of the contractor, And the number of the contract or order.

Bases of Purchase
The material should be purchased by net weight, provided the loss is weight At 150 to
155°C does not exceed 1per cent. Deliveries based on samples taken at The manufacturer’s works
and which yield more than 1per cent loss in weight at 150 to 155°C should be rejected. Deliveries
based on samples taken at any place other than the manufacturer’s works and which yield more
than 4per cent loss in weight at 150 t o 155°C, should be rejected. With deliveries based on
samples taken at any place other than the manufacturer’s works and which yield not over 4 per
cent loss, settlement should be made on the basis of 1per cent loss in weight; that is, 0.99 pound of
nonvolatile matter should be considered 1 pound of soda ash. For example: loss in weight at 150
to 155 °C=3 per cent, then:
Net weight as received (100-3)
Net weight of material to be paid for=
99

Notes.
This specification covers material for various washing, cleaning and scouring processes with
or without soap, as conditions may require, and where a moderately strong alkaline material is
desired.
Soda ash is hygroscopic and it may also absorb some carbon dioxide depending atmospheric
or storage conditions or both or on packaging. Therefore. .Care should be exercised in sampling in
other to obtain a representation sample. T he inspector should mark the sample to show the place
at which it was taken.
This specification is not intended to apply to soda ash used in glass making, but dose not
preclude the pure base of soda ash suitable for glass making.
It is believed that this specification adequately describes the characteristics necessary to
secure The desired material and that normally no samples will be necessary prior to award to
determine Compliance with this specification. If, for any particular purpose. Samples with bids are
necessary
They should be specifically asked for in the invitation for bids and the particular purpose to
be Served by the bid samples should be definitely stated the specification to apply in all other
respects.
This specification covers only the types class’s grades sizes etc. of the commodity as
generally Purchased by the Federal Government and is not intended to include all of the type etc.
which Ate commercially available.
 Chapter XXIX

Behavior of Soda Ash in Storage

Soda ash, or anhydrous sodium carbonate, is hygroscopic. It absorbs Moisture from the
atmosphere during storage or transit. A freshly made Ash is powdery loose and free from lumps. It
has the tendency to cake or “set” in contact with moisture and consequently becomes lumpy. This
Often causes a consignment of soda ash to fall below the guaranteed test when the ash has been
kept in storage for a long time or shipped a considerable distance. For example one shipment of
soda ash had the analysis given in table 155.
Table 155. Composition of Soda Ash After Six Months’ Storage.
Per cent
Na2 CO3 99.17
NaHCO3 Nil
NaCl 71
Water Nil
After six months in storage it had the composition given in table 156.
TABLE 156. Composition of Soda Ash After Six Months’ Storage
Per cent
Na2 CO3 82.81
Na2 HCO3 4.73
NaCl 0.61
Water 11.74
Complaints have frequently come from the customers, stating that the Ash has fallen in purity
and that a considerable amount of water has been found in the product. If the ash is sold to the
block makers, very likely they will find that it wills mot “set” to a hard block with the normal
Amount of water employed. On close examination it is found that the Weight per bag or per
barrel has increased correspondingly, so that while the alkali content has decreased, the bulk
weight has increased. No loss in soda has really taken place. The sample showing considerable
water and bicarbonate, the same test as when it left the works.
Moreover, samples from different parts of the same bag may yield quite different results. One
dense ash consignment consisting of 200-lb Bags of dense ash each gave two analyses shown in
Tables Nos. 156 and 157.
From the tables it will be seen that the outer portion of the bag has absorbed more moisture
and more CO2 from the air than the inner portion, the density of the outer portion is also higher.
Hence sampling is a very important factor..
TABLE 157. Analysis of a Dense Ash Sample Taken from
Outside Portion of a Bag.
Per cent
Na2 CO3 89.26
Na2 HCO3 2.84
NaCl 0.50
Na2 SO4 0.15
Fe2 O3 0.001
 Insoluble matter 0.04
Water 7.19
++
Ca nil
++
Mg nil
Density (by loosely-packed method) 0.998
TABLE 158. Analysis of a Dense Ash Sample Taken from Center of the Same Bag.
Per cent
Na4 CO3 92.00
NaHCO3 1.99
NaCl 0.53
Na2 SO4 0.15
Insoluble matter 0.04
Fe2 O3 0.001
Water 5.29
++
Ca nil
++
Mg nil
Density (by loosely-packed method) 0.987

The amount of water and carbon dioxide that the ash will absorb Depends upon the climatic
conditions (i.e., the humidity and the temperature of the air), the extent of exposure to the
atmosphere and the Length of time. Light ash is more readily affected by atmospheric moisture
and CO2 than dense ash. A kind of light ash that had been kept for 9 years (from 1921-1930) in a
stopper bottle gave the co imposition shown in Table 159.
TABLE 159. Composition of Light Ash Nine Years’ Storage.
Analytical results calculated results
Per cent per cent
Insoluble matter 0.03 Insoluble matter 0.03
Na2 CO3 51.25 Na2 CO3 .H2 O 16.23
Na2 HCO3 29.62 Na2 CO3 .Na2 HCO3 . 2 H2 O 79.69
NaCl 0.95 NaCl 0.95
Fe2 O3 0.002 Fe2 O3 0.002
Water 18.11 Free moisture 3.06
And a sample of granular dense ash kept for the same length of time
And under the same condition gave the result shown in Table 160.
 For many purposes, the presence of moisture and bicarbonate in the Ash is not objectionable,
but for certain purposes, such as for soda block Making, it may cause the ash to lose much of its
“setting” property.
The change which soda ash undergoes in contact with the atmosphere has been the subject of
some study. It not only reveals the natural soda Change the ash undergoes on exposure to he
atmosphere, but it also Throws much light on the formation and composition of natural soda
Deposits. A bag of freshly made soda ash is sent to the laboratory and Kept in a dry place. A
sample is taken from it once every two weeks and the necessary analyses made thereon. In this
way the change in the ash is observed by means of a bi-weekly record during a period of a whole
year. The results are given in table 161.
The results in table 161 show that in the course of 365 days soda ash Has picked up more
than 14 per cent of water but only 5per cent of CO2 It absorbs moisture from the air more readily
than it does CO2 , and takes Up moisture very rapidly in the beginning before it takes any material
amount of CO2.moisture seems to have been taken in to form first the Monohydrate
(Na2 CO3 .H2 O), which then in turn is gradually transformed into the sesquicarbonate with the
absorption of CO2 and a further amount of moisture. This is show by the fact that the monohydrate,
which is formed in comparatively large quantities at the beginning, rises rapidly to a
maximum; then it decreases and at the same time the quantity of the sesquicarbonate formed
steadily increases (see Curves I and II.Fig.129). Had we continued to keep record of the change
for A number of years it would have been found that all the soda ash (anhydrous carbonate) would
hake disappeared and that all the monohydrate Would have been converted in time to the
sesquicarbonate from. In other Words. All the soda ash would eventually be changed to the
sesquicarbonate as the final stable form in contact with the atmosphere. The difference between
the amount of water absorbed and that of CO2 absorbed Also points in this direction, as the value
has increased to over 9 per cent And was slowly increasing, approaching the value of 10.17 per
cent, which Is the theoretical difference in the percentages of water and CO2 in the

Pure sesquicarbonate from the pure anhydrous sodium carbonate (curves IV andV,Fig.130). thus,
3Na2 CO3 +5H2 O +CO2 à 2[Na2 CO3 . NaHCO3 .2H2 O]
318 90 44 452
 90 44
﹙ - ﹚×100=10.17%=Difference between %H2O and
452 452
%CO2 absorbed in the sodium sesquicarbonate.
Again, the ratio of water absorbed to that of CO2 first increases to a Maximum of 11, then
decreases to less than 3; and the trend of the curve Shows that it would continue to decrease
approaching the theoretical Value of 2.04,which is the theoretical ratio of water to CO2 absorbed
by Pure anhydrous sodium carbonate to form pure sesquicarbonate as the Ultimate product.
In spite of the continued increase in the water absorbed in the soda Ash, the percentage of
Na2 CO3 .H2 O is going over to the Na2 CO3 . NaHCO3 .2H2 O form faster than is formed from the
soda ash. It is not Unlikely that at the end there will come a time when all the Na2 CO3 .H2 O Will
have disappeared and the exposed mass then will consist of all Na2 CO3 .NaOH3 .2H2 O if the
process is allowed to proceed indefinitely. But as the course of the change was followed for only
365 days, it is Difficult to judge from the curves obtained when that time may come, i.e., how long
it will take to reach that condition, starting with anhydrous soda ash, because the curve flattens out,
asymptotically approaching the theoretical value. Nevertheless, it gives us a good hint as to the
manner of formation of the natural trine or urea, and a theory as to why the sesquicarbonate is the
most favored form of deposit in nature, although the sesquicarbonate with the favored form of
deposit in nature, although the sesquicarbonate with the formula Na2 CO3 .NaHCO3 .2H2 O
repreclosely approach in composition.

FIG 131 Curve showing ratio of water to carbon dioxide absorbed by soda ash.

The manner of formation of trine or urea in the earth crust is Explained thus: water carrying
soda in solution is collected in a depressed Region and is concentrated by natural evaporation due
to the dryness of The climate, the scarcity of rain, and the presence of more or less constant Winds
in these open places. Concentration of soda in the water becomes so great that during cold weather
sale soda begins to crystallize out on the Surface of the marshy ground. The upper layer of the sale
soda deposit Gradually effloresces in contact with the atmosphere and loses much of its water of
crystallization. Forming the white powdery monohydrate Na2 CO3 .H2 O. This in turn is gradually
converted by atmospheric H2 O And CO2 to the sesquicarbonate as the ultimate stable form. We
have Thus:
Na2 CO3 (diss.) à Na2 CO3 .10H2 O(cryst.)à
Na2 CO3 .H2 O(powder)àNa2 CO3 .NaHCO3 .2H2 O
 Coming back to the question of the weathered soda ash during storage Or transit, how are we
to distinguish it from an underfinished or wet ash? That is, how are we to determine, upon receipt
of a sample of soda ash Containing moisture and NaHCO3 , whether it was a good ash but merely
Weathered by long storage, or whether it was originally a kind of underfinished soda ash as it left
the dryers? A weathered ash had, as we saw above, a high percentage of water with a
comparatively low percentage Of CO2 and the percentage of total water absorbed in the ash is
always greater than twice that of CO2 , approaching 2.04 as the excess of CO2 in the form of
NaHCO3 . in calcining crude filter bicarbonate, the last traces of ammonia and water are driven off
far more readily than the remaining portion of NaHCO3 , and an underfinished ash may contain as
much as 3 to 5 per cent NaHCO3 and yet a very small fraction of one per cent of water .
consequently, from the proportion of water and excess CO2 in the ash, it is not difficult to
dirrerentiate a weathered ash from an underfinished ash , but when an underfinished ash has
further been excessively weathered, there is no way to tell whether or not this weathered ash was
originally a well-finished ash. Its past history then cannot be determined. Table 162. gives the
composition of a sample of underfinished ash.
TABLE 162. Composition of Unerfinished soda ash
Percent
Involution matter 0.06
Na2 CO3 79.55
NaHCO3 16.80
NaCl 1.54
Fe2 O3 0.005
Water 1.89
With the NaHCO3 as high as 16.8 per cent the amount of water Present is only 1.89 per cent.
If this were a weathered ash to correspond to this high percentage of NaHCO3 the water absorbed
would be considerably more than 10 per cent.
 Chapter XXX

Layout, Design and Location of Ammonia Soda Plant

Ammonia soda plants require as raw materials the following commodities arranged in the
order of quantities required:
(1) Salt (rock salt brine or sea brine)
(2) Limestone
(3) Coal
(4) Coke
(5) Ammonia [crude liquor or (NH4 )2 SO4
(6) Sulfide of soda, or 60 per cent Na2 S when (NH4 )2 SO4 alone
Is used
Arranged in the order of the expenses involved per ton of soda ash Made (quantity and cost)
these raw materials stand as follows:
(1) Coal (4) Coke
(2) Limestone (5) Ammonia
(3) Salt (6) Sulfide of soda
The latter arrangement may vary somewhat for different plants Because the comparative
costs of coal, salt, and limestone entering the Manufacture depend upon the relative location of the
plants in regard to The sources of supply of these materials, their quality, and their prevailing
prices in a particular locality.
The prices of the first three raw materials when they are plentiful And their source is in close
proximity to the works may be as low as:
Salt in the form of brine 0.580 per ton, delivered at plant
Limestone 0.60-- 0.75 per ton, delivered at plant
Coal 2.00-- 3.00 per ton, delivered at plant
But it is seldom that a site for an ammonia soda plant can be so fortunately located as to have
all three raw materials in close proximity, at The minimum prices, and in large quantities. With
soda ash at its normal Price, plants obtaining these raw materials at prices varying from 100 to 300
per cent higher than the above figures operate at a good profit, especially when they have a
moderately large output.
The foregoing is a direct cost dealing with cost of raw materials only. In the following we
shall include also the indirect costs ,thus approaching More nearly the cost statistics of a plant
operation. These may be arranged as follows:

(1) Fuel (5) Ammonia

(2) Labor (6) some form of sulfur
(3) Limestone (7) Factory supplies
(4) Salt
The positions of the first two items might be reversed in locations where fuel is exceptionally
cheap and where the proper arrangement of Equipment for reasonably low attendance labor does
not exist.
In only the most exceptional circumstances could limestone be more Expensive than either of
the first two items, although occasionally salt May cost mote than limestone. The last three items
are frequently in other relationships to one another since they depend on many factors. They have
been included because they have some bearing on the selection Of the plant site, and their effect
will be discussed briefly in this chapter “Depreciation and obsolescence” at the customary rate
of 10 per cent on, say, a 15000 investment per ton of daily capacity in soda ash and a 70 per
cent annual load factor introduce a cost per ton which leads the above list by an appreciable
margin. it is thus obvious that the lowest possible initial cost of plant is important to low-cost
production. A layout which results in low initial cost and yet permits the above seven items to
have a low total is obtained only from capable, highly experienced design engineers. This chapter
does not attempt to describe all the factors and variables entering into the problem of obtaining the
best layout and design of an ammonia soda plant it merely calls attention to the effects of some of
the more important variables.
Favorable unit prices for the four large items given in the foregoing List follow. These are
delivered prices and represent a fairly plentiful Supply close to the sites under consideration.
occasionally some one of the materials may be even higher. Well-designed plants can be operated
At a profit where raw material prices are cons iderably higher, since then the soda price is
generally also higher.
(1) Fuel
(a) Coal, 3 per ton, or 11 cents per million but.
(b) Coke, 6.50 per ton, or 23 cents per million but.
(c) Bunker “C” fuel oil, 0.85 per barrel or 13.5 cents per
Per million but.
(d) Natural gas, 15 cents per thousand cubic feet or 15 cents
Per million bru.
(2) Attendance and Maintenance Labor (weighted mean of all classes
Needed)
0.65 per man-hour
3 Limestone
0.60-- 0.75 per ton stone, as already given
4 Salt, as saturated raw brine
0.50 per ton NaCl, as already given

There is almost nowhere a peculiarly favorable site where all the raw Materials and
intelligent labor required at the lowest unit prices can be found. Furthermore, it is the alkali
manufacturer’s problem to deliver His product to his customer’s doorstep at the least possible cost.
In selecting a site he must therefore consider not only the purchase and Transportation of raw
materials to his plant, but also the transportation of his soda products to the plant of his customers
(or potential Customers).
The layout of an ammonia soda plant is a matter largely of local conditions and individual
preference. However, there are some general rules on which those qualified to make a design will
agree. For instance, the Boiler room should be located adjacent to the engine room furnishing
Steam to the carbon dioxide compressors, electric generators, exhausters, Cooling water pumps,
air compressors and other steam-driven machinery And the lime department should be adjacent to
the distiller house so that The concentrated milk of lime can be sent to the distillers in short pipe
Lines. Next to the distiller house should come the ammonia absorber House, with an adjoining
space for the location of settling vats. Next comes the tower house taking the ammoniated brine
from the vats in the Absorber system. The furnace room is located reasonably near the tower
Hours so that crude bicarbonate can be taken from under the filters in Belt conveyors to the
furnaces and the returned gas from the furnaces Sent back to the columns through a set of
condensers. The engine room, Where all the steam-driven machinery, the carbon dioxide
compressors Est. Are located, should be sufficiently close to both the furnace room and The tower
house to bring the gas from the furnaces through scrubbers and Coolers to the columns and at the
same time take the kiln gas from the Lime kilns through a kiln gas scrubber. The carbon dioxide
compressors, The generators, the exhausters, the cooling eater pimps and the air compressors may
be advantageously steam driven, and the exhaust steam Utilized for the distillation of ammonia in
the distiller house. Outside Of the engine room in the various houses, the machinery is
individually driven by electric motors.
The buildings in an ammonia soda works are usually high. The distiller house must be high
enough to accommodate the tall distillers Together with the condensers. It is general practice to
put the heater on Top of the lime still, and the condensers on top of the heater, making the Overall
height of the whole apparatus over 100 feet. The distiller house is then at least 125feet high and
frequently much higher.
In the absorber house, the elevation of the bottom of the absorber must be above the top of
the settling vats, so that the ammoniated brine can flow to the vats but gravity. The vats are located
on the main floor on pedestals high enough to provide a working height to pump “mud” from the
bottom. On the top of the absorber can be located an ammonia gas cooler, and on top of this the
absorber washer, necessitating a height about 80 feet for the absorber house. In between the
bottom compartments of the absorber and the middle cassette, the partially ammoniated Brine is
cooled in outside surface coolers (absorber coolers) before final Ammoniation at the absorber
bottom.
In the tower house are located the columns, or carbonating towers, Form 75 to 80 feet high.
The exit gas containing ammonia passes through A cyclone separator located near the top of the
tower and is scrubbed in Washers called tower washers, whose bottoms are below the top of the
towes, but must be above the inlet point in the inlet compartment of the Absorber washer to avoid
the use of a centrifugal pump. The brine from the bottom of the tower washers, containing some
ammonia, flows to the Middle compartment of the absorber washer through a long u-loop seal.
The brine from the bottom of the filter washer enters the tower washer at a few sections below the
top compartment. The filter gas washer Must be located on a higher level (not less than 10 feet
above the top Of the tower washer) so that the brine from the bottom of the filter Washer can flow
by gravity with a u-loop seal to the inlet compartment In the absorber washer from a higher
vacuum on the filter washer side Or to the inlet compartment at the tower washers against pressure
on the Tower washer side. This arrangement avoids the necessity of having Centrifugal booster
pumps to send brine from the bottom of the filter Washer to the inlet compartment of the absorber
washer or of the tower Washer. The total height of the tower hours including the washer house
Portion is over 120 feet. On top of the roof my be situated a large Brine storage tank feeding fresh
brine by gravity to the top of the tower Washers the filter washers and the absorber washers. The
draw liquor being cold, very little brine is required to scrub ammonia from the gas in the filter
washer. The absorber washer and the tower washer are the two places where fresh brine is
required in proper quantities. The Bicarbonate filters are generally located on a floor at least 30
feet above Ground so that the filter liquor can flow down to the filter liquor main under high
vacuum with sufficient liquor leg for a seal. From the filmier Trap. Or separator, the gas is piped
up to the bottom of the filter washer Located on approximately 100-foot level.
In the lime department, the lime drawn from the bottom of the kiln May be elevated to a bin
in bucket elevators. From the bin the lime is Fed to a rotary slaker by means of a conveyor feeder.
The discharge end of this horizontal slaker by means of a conveyor feeder. The discharge end to
be hauled out in cars from the main floor and the milk of lime to flow by gravity to a lime well
or storage tank situated on the main floor, where the lime pump sends the milk of lime to the
distiller. With the Lime kiln standing above the building the total height of the house should be
about 50 feet. On top of the limekiln, there should be a working Platform to which access may be
gained conveniently by means of a Stairway or elevator.
The furnace room contains the rotary furnaces and the conveyor System, with a clear floor
space on the top for storing and feeding bicarbonate to the furnaces below. The building need not
be over 45 feet high it is one of the lowest buildings in the whole plant. The soda conveyors can
go outside of the building or between buildings. A good overhead Space, however, must be
provided so that the place may not be too hot or oppressive to the workmen. Ventilation points
should be freely provided in the roof construction; otherwise the hot atmosphere in a low Room
coupled with the soda dust and ammonia gases (to a certain Extent unavoid able) may make the
place quite uncomfortable to the Workmen. This applies to the space required over the conveyors
as Well as over the dryers.
The boiler room is normally just high enough to accommodate the Boilers and their
accessories and its height depends upon the type of Boilers installed, but a full-height basement
must be provided. For Hauling out the ash. The engine room is a low building, generally one Story
high, to house various machinery; but it must be high enough to Allow an overhead space under
the trusses for installing a traveling Crane. It must also have a full-height basement to house
various steam Lines exhaust lines, etc.
The packing room and the rest of the buildings are low and generally Not more than two
stories high, but high enough to house various soda Ash elevators and conveyors. But the main or
ground floor of all the Houses in the works should be located 10-20 feet form the normal ground
Level with a full-height basement on the ground level. This basement Space may be utilized for
ash hauling (in the boiler room), for locating Steam mains and exhaust mains (in the engine room),
and for locating various gas liquor and water pipes and also waste liquor sumps (in the Tower and
distiller houses). This arrangement sets the working floor At 10-20 feet above ground lever with
various pipes introduced below the Floor and yet allowing a clear space overhead for working in
the basement. It is both a neat and convenient arrangement.
The packing and loading facilities for the dry products should be immediately adjacent to
loading tracks. And docks. There should also be proper approaches for motor trucks. The docks
for handling products onto both trucks and railroad cars should be designed for minimum labor.
The raw material storage yar d and the finished product storage should Be located near the
shipping points, in order to facilitate moving the Raw materials in and sending the finished
products out.
The packing and shipping buildings should be reasonably close to the Rotary furnace
building so that conveyors for the dry products may be as short as possible. This minimizes power
consumption, maintenance Cost and breakdown hazards. Prior to the filter operation, substantially
All the materials are in the liquid state so that the arrangement of Buildings need sacrifice
relatively little of the requirements of short Pipelines. Pipelines are very cheap compared to dry
product conveyors. Form the filters on, the handling of the solids by conveyors starts. The Furnace
room needs to be within good conveying distance of the crude Bicarbonate filters.
There must be a storage reservoir for cooling water and another for Brines. If these reservoirs
can be placed as high as needed to feed into all apparatus by gravity, an advantage is gained in that
on pump is Required and the hazard of pump breakdown is thereby eliminated
The buildings are of either steel or reinforced concrete structure (prêtearly the former), and
must be heavy to take care of the heavy, tall, Cast-iron apparatus. The foundation must be made of
heavy reinforced Concrete (made waterproof). With concrete or timber piles driven under each
piece of heavy apparatus, such as the ammonia distillers, the Carbonating towers, etc. where the
whole load is concentrated on a Small base. Overhead inter-building pipes should be supported
form the Trusses or beams. While those underground pipes between buildings Should be located
in a tunnel or trench, but not buried in the soil, so That settling of the buildings at either end will
not throw strain on the Pipes, and also so that repairs may be made conveniently. Where a pipe
Passes through a brick wall, clearance above the pipe in the wall should be allowed so that
settling of the wall will not bear on the pipe. This Applies especially to cast-iron pipes. Which are
used extensively in the Works. Sewers and fire hydrant lines should be located below ground
around the buildings. Cooling water from the exit of different units of the apparatus is run off to
the sewers, but the distiller waste liquor must be sent out with or without a booster pump to an
open field for Settling or to a large body of water through pipes or in a specially constructed
channel or canal.
The materials used in the construction of buildings and equipment Depend to some extent
upon local supplies .as pointed out in earlier Chapters, the apparatus for handling ammoniated
brine and filter Liquor must be of cast iron; high-copper alloys are entirely unsuitable. Building
frames constructed of mild steel will suffer from corrosion, and factors of safety in the original
design must take this into account. Timber buildings are reasonably resistant, but are exposed to
fire hazards Concrete made of ordinary silica sand and a coarse aggregate of good Rock does
reasonably well for floors if they are amply pitched to Minimize the effect of puddles of corrosive
liquor.
In the caustic plant, concrete floors should be very steeply pitched and the design must be
carefully prepared to avoid seepage of caustic Liquor under important weight bearing footings.
Although concrete is Not very resistant to caustic there is no other suitable material having an
economical advantage over the usual local concrete. The dust in the Air near the caustic operations
is corrosive enough to deserve careful Consideration in the design.
The tallest buildings in an alkali plant afford protection to portions of apparatus, which rarely
have men near them. it is sometimes advisable to consider the climate and the cost of a simpler
type of protection, Suc h as heat insulation ,painting weatherproofing , etc of the apparatus and
piping rather than a housing over them. .
The following summary gives a broad consideration regarding general Requirements in the
layout and the design of an ammonia soda plant:
(1) It must fit the contours of the land and take best advantages of High and low points.
(2) It must bear a proper relationship to existing or potential railroads motor highways, docks
and waterfront.
(3) It must be arranged for lowest handing costs.
(4) It must be arranged to permit extension of the plant and leave Room for additional units or
for possible by-product operations.
(5) Interconnecting piping and material conveyors should be as short as possible.
(6) Apparatus should be arranged so that directs attendance labor can control a maximum of the
operations.
(7) It should leave room for adequate storage. Of raw materials and finished products.
(8) It should bring about minimum maintenance cost.
Since the contours of the actual site and the variations in the eleven ton of the ground
affect the building arrangement more than and other Factor. No hard and fast rule can be laid
down for the best arrangement A number of expedients are possible for reducing installation
costs by Compact, and yet not cramped, layout. Further improvement can be attained by
ingenious interrelation of equipment so that a single operator can adequately attend a maximum
of it. Not only does this save money in direct operating labor cost, but also a more adequate
control is possible
Where a single individual depends least on the cooperation of others. Unfortunately, no
direct statement as to how this is achieved can be Mate because, depending on the-best location
for railroad and docks, The various products to be manufactured, etc , certain operations will Be
adjacent to each other in one plant and not in another.
SELECTION OF PLANT SITE
For the location of an ammonia soda plant the site must fulfill the following main
requirements:

(a) Good transportation facilities preferably

Both by boat and by rail.

(1) General
Requirements (b) Sufficient nearness to the market or to
The center of distribution

(2) Particular
Requirements (c) close proximity to the sources of salt,
Limestone, and coal (and consequently
Also coke and crude liquor).
(d) Plenty of underground cold water avail-
Able for cooling.
(e) Ample land or waterway available for
Waste disposal.
 If all these requirements cannot be met, expert judgment is necessary to decide, by a detailed
study of the conditions, which of the above items are most essential. For example. If both salt and
limestone are not available close together, it may be better to locate the plant near the source of
supply of limestone and transport the salt, than vice versa, Since salt can be transported in the
form of brine at little cost.
Items (a) and (b) above apply to most industries and are therefore the general requirements
for any given industry. They will not be discussed in detail here, as this is quite evident.
As salt (in the form of rock salt brine or sea brine), limestone, and Coal are the three main
raw materials in soda ash manufacture, it is Necessary that these materials be plentiful and cheap,
and their continuous supply assured. The first two, salt and limestone, further the manufacturers of
soda ash must produce more. The soda Works must own and operate salt wells (in the case of rock
salt brine) or salt fields (in the case of sea brine), and quarry their own stone. As these are the two
raw materials used in the largest quantities, the Economical development of the works depends
upon a cheap and Inexhaustible supply of both. They must also be of good quality, especially the
limestone. A poorer grade of limestone is very uneconomical to use. Coal can be purchased on the
open market but it must be sufficiently cheap. Where bituminous coal is abundant, and coalmines
are not far away, crude ammonia liquor and coke will be available form the gas works. Coke is a
standard commodity than can be bought in the market, but the crude liquor must be secured from
the gas works by Special arrangement. Ammonia and coke, though used in comparatively Small
quantities, constitute important materials in the ammonia soda Process because of their
comparatively high cost, so much so that the Manufacturers frequently operate their own
by-product coke ovens. The crude liquor is desirable not only because it is a cheap source of
Ammonia but also it contains the sulfide required for the system .at Present some synthetic
ammonia is being used, and sodium sulfide is all The more necessary in the process. All these
enter into the consideration of a plant site.
The question of water supply is important. Not only is water used in large quantities in the
boilers for power generation, but also a much larger Quantity is needed for various cooling
purposes. The requirements for the two purposes may not be the same. The surface water, e.g.
river Water, while it may be very good for the boilers, is not suitable for Cooling, as the
temperature may be to high in summer .on the other Hand, well water may be too hard for boilers
but it is excellent for cooling Purposes. A good source of supply of cooling water is from artesian
wells ,as it is practically at a constant temperature (16 to 18.5 . In Summer and winter at 400 to
500 feet below the surface level). Good Water is comparatively difficult to obtain for plants, which
are located Near the sea-coast, where sea brine may be used as the source of the salt Supply and
where probably only well water will be available for the Boilers. For small works, the quantity of
cooling water required is in the neighborhood of 2,000 to 2,500 cubic feet per ton of ash. For large
Works it is less per ton of the output; but much depends upon the cooling Surface available, the
design of the apparatus, and the arrangement of different cooling units. Therefore cooling water is
a factor in the Choice of the site.
The distiller waste is always a source of trouble. The quantity is Large (about 350 cubic feet
of liquor per ton of ash) and its disposal in Public waterways have been in most cases prohibited
by law. Calcium Chloride in solution permeates the ground and finds its way to the underground
water. It causes excessive chlorine in the soil. Solids separating Out from the waste do not readily
become hard and firm. If the waste Liquor is sent directly into a river, its lime and calcium
chloride content May be objectionable. Besides, the solids carried in it may silt up the River. Its
satisfactory disposal has become a problem unless the works is fortunate enough to be located
near the sea, near a big lake, or among the hills (in this last place, the waste may be sent to fill old
quarry pits Or a nearby valley). Not all the waste produced can be converted into By-products; a
large quantity must still be disposed of. If the works is Located near the sea, on a lake, or near a
very large body of water it is a comparatively simple matter, but inland this waste can only be sent
to a pond to settle, from which only the clear portion is allowed to run out to the river. The lime
content in the waste liquor is particularly harm Full to the fish in the river. This waste disposal
must receive consider Ton in the selection of the site.
The foremost problem is to find an area where potential market and the present delivery
costs of alkali are attractive and at the same time reasonably near sources of the necessary raw
materials and fuels. Facilities for adequate and rapid service to customers eventually must be
provided. This requires that a number of good transportation facilities be available (actually or
potentially) in the area, for ocean, rail and motor highway movements of products. Within such an
area, the few specific locations where the foregoing is favorable must be rated quantitatively on
the basis of many factors. The following detailed consideration may govern the choice of the site
and are given here as a resume:
(1) Delivered cost, quality and potential continuity of main Supply
(2) Delivered cost, quality and assurance of an alternative or Emergency fuel supply.
(3) Delivered cost, quality and reserves of lime stones available.
(4) Delivered cost, quality and reserves of salt supply.
(5) Total freight cost of all products to present and potential Consumers.
(6) Temperature, quantity, and chemical properties of cooling Water supply.
(7) Treatment cost and treatment results of best and second best Boiler feed water supplies.
(8)Cost of solid waste disposal in near and distant future.
(9) Cost of liquid waste disposal procedure.
(10) Soil-bearing capacity and character and cost of land.
(11) Availability, intelligence and character of labor.
(12) Adequacy of housing, schools and other living facilities for Personnel and workmen.
(13) Cost, adequacy and assurance of ammonia supply.
(14) Cost of coke and availability of crude ammonia liquor supply.
(15) Flood, hurricane, earthquake, etc. hazards.
(16) Cost of space for expansion or growth. .
(17) Potentialities or incentive for location and growth of alkaliconsuming industries.
(18) Attitude of adjacent communities toward large-scale industry.
(19) Adequacy of mill-supply and mechanical services.
In what follows we shall elaborate these points in greater detail. It Is necessary that expert
engineering judgment be exercised in weighting Factors such as listed above relative to one
another, and in computing or Estimating the rating of each of the prospective sites. Since fuel,
limestone and salt are the three main requirements, the selection of a site will be governed by the
quality and mining cost of those which are available within the territory considered. Some of the
factors, which require Consideration are further elaborated in the following:
 The quantity and kind of fuel required have been discussed in Chapters XXII and XXVI. In
this chapter it need be pointed out that Ammonia soda plants can burn any kind of fuel so long as
it is cheap. Fuel equivalent to half a ton of high-grade coal per ton of soda is near The minimum
consumption for all purposes in the direct manufacture Of soda by the ammonia process .a small
amount of the fuel will be Burned in preparing the lime. Almost invariably this must be
metallurgical coke. Which is a standard commodity with a standardized market, but can
conceivably be very expensive. Such is quite unlikely to be the Case in any territory having a high
alkali market, but is true with certain Alkali plants on the gulf coast. A somewhat larger amount of
the fuel will be required in soda dryers, and this may be bituminous coal. Since Most of the fuel is
used in the generation of steam, the fuel problems in selecting a site are much the same as for any
other industry. These were discussed in chapter XXII, and hence will not be dealt with here.
Continuity of operations, important as it is to economical production And high quality
product, requires either very large stocks of fuel and Raw materials, or assured emergency or
“second-string” supplies. A Combination of the two is usual. This is more important with fuel than
With limestone or salt, for the latter are usually under the control of Local management. The
second-string fuel supply can be more costly if it is dependable.
It is decidedly advantageous that the production of salt and limestone is under the control of
the soda ash manufacturer. As is generally done. That is, one department of the works should be in
charge of the Brine field operation and another in charge of the quarrying and preparation of the
stone .Not only must sources of these raw materials give assurance of loving-time inexpensive
supply, but their quality must also fulfill certain requirements. this is particularly important with
limestone, as is discussed in chapter VI. Likewise an impure brine or high magnesium sea-salt
introduces added costs, as discussed in chapters IV and V. the adequacy of each on the basis of the
principles mentioned must be exhaustively considered in selecting an ideal site.
More than eight tons of brine, stone and fusel are handled into the works per ton of soda ash
shipped out; and more than 12 tons, per ton of chemical caustic shipped out. On a strict ton-mile
basis, therefore, the Tion of the stone. Not only must sources of these raw materials give
Assurance of long-time inexpensive supply, but their quality must also fulfill certain requirements.
This is particularly important brine or limestone, as is discussed in chapter VI. Likewise an impure
brine or high magnesium sea-salt introduces added costs, as discussed in Chapters IV And V. the
adequacy of each on the basis of the principles mentioned Must be exhaustively considered in
selecting an ideal site.
More than eight tons of brine, stone and fuel are handled into the Works per ton of soda ash
shipped out; and more than 12 tons, per ton Of chemical caustic shipped out. On a strict ton-mile
bias, therefore, the Plant should occupy a position considerably nearer to its raw materials than to
its market. However, about 60 per cent of the raw materials are Very cheaply transported by pipe
line, and freight rates on stone and coal Are cheaper than on finished products like soda ash and
caustic soda. Furthermore, there are local discontinuities in freight rates on stone and coal Setting
up a “formula” for finding the site of minimum transportation Costs. A finite (and small) number
of potentially good sites can be determined by a rapid qualitative consideration of the most
Important Factors. For these and the few suitable raw material and fuel sources. An experienced
engineer should compute actual annual transportation Costs based on previously prepared market
estimates.
The question of cooling water supply is of considerable importance if during the months of
warm water, on reasonably cool (ten to fifteen Degrees below wet-bulb temperature) water is
available, and the production of the plant will suffer. Furthermore, the power required for
circulating
Excessively large volumes of water is considerable. If a site affords a large volume of water,
which gets quite warm (approaches wet-bulb temperature) on the summer time, it is necessary to
investigate the possibility of a supply of artesian well water for summer operation. Equipment for
refrigeration of water will, under certain condition, compete with Deep-well water if the latter id
scarce. Exhaust steam is suitable as the Source of power for such refrigeration and is thus a sort of
summer complement to the winter space-heating load. Only a small portion of each Unit of
cooling apparatus needs to be supplied with refrigerated water Besides the temperature of the
cooling water, it is necessary to investigate Its behavior in cooling equipment. Certain water s
contains temporary Hardness or other solids, which deposit difficulty removable crusts on cooling
surfaces. Many sources of water are conducive to algae-growth in warm weather when the plant
can least afford any impeded heat Transfer. Other has properties of causing corrosion of metals of
Construction in the cooling apparatus. So it may be necessary to give the cooling water a suitable
chemical treatment. This has been dealt with at great length in Chapter XXIII.
An additional supply of water may be desirable for steam generation. As mentioned before,
frequently a source of excellent cooling water is a Poor source for boiler feed water. Organic
contamination in this course of Water is not so harmful as in the cooling water, but scale-forming
solids in solution would bring about high operating expense. Since the supply of boiler feed water
should (for best fuel economy) be given as much Preheating as possible from sources of
low-pressure waste heat, it must also be considered as a cooling medium. The quality of the
contemplated Source of feed water, both from the standpoint of behavior in cooling Equipment
and from the standpoint of adequate treatment to inhibit Scale forming and boiler corrosion, must
be given very careful consideration. Alkali engineers generally employ water specialists and
finished
The difference between the quantities of raw materials and finished Products mentioned
above must be disposed of as solid liquid and Gaseous wastes. There should be a large area
adjacent to the plant for the disposal of the solid wastes, as mentioned above. There come from the
distiller blow-off, the lime caustic mud, the boiler ashes, and the slaker Rejects, etc. This is one
of the most serious problems connected with an alkali plant. The physical character of most of
these solids is such That they ate very poor material for land fills. The caustic mud contains excess
lime. The distiller waste does not dry or harden over long Periods of time. There is sufficient free
lime and chloride in it so that Vegetation will not grow on it for many years. Consequently most
people consider the deposits unsightly. It cannot be disposed of in the Usual Rivers or public
waterways. A prier site for an alkali plant Should have available a large area of swamp or marsh
land or valleys, Which can be cheaply disked or dammed to make room for the waste Disposal for
a number of years. The boiler ashes and the slaker rejects May be utilized in road making for a
country community.
All alkali plants have to dispose of large quantities of calcium Chloride. For plants located
on the seacoast, once the distiller blow-off has and the solids settled from it, this presents on
problem. Inland Plants, however, are faced with a difficulty caused by the introduction of The
chlorine into the irrigation waters of the district, an alkali plant Must therefore be located within
reach of a reasonably large stream So that the chloride to be disposed of does not add appreciably
to the Salinity of the river water.
As explained above, most of the buildings in an alkali plant are unusually tall.
Consequently unusually heavy foundation loads are encountered under the buildings as well as
under the taller equipment if an alkali plant is not located on good ground, the foundation costs
May be exorbitant. Furthermore, large foundation masses inhibit pipe Tunnel, elector pit, and
similar features of design, which add to the Cost of the entire plant. There should be an adequate,
reasonably level should avoid regions of congestion and high land values. Modern ammonia soda
plants use relatively little labor, compared to Most other industries, measured either per dollar of
investment or per Ton of the output. The usual complement of killed maintenance artisans should
avoid regions of congestion and high land caules. Modern ammonia soda plants use relatively
little labor, compared to Most other industries, measured either per dollar of investment or per Ton
of the output. The usual complement of killed maintenance artisans Employed on apparatus and
building repaired, as in most other Plants. The labor for handling and loading the finished product
is likewise the same in quantity and quality as in other industries, having a comparable division
between bulk and package shipments. In direct production, however, one painstakingly trained and
intelligent operator can, because of the continuity and automatic control in the operation, attend an
exceptionally large amount of apparatus. Abnormalities of operation occur so seldom that it takes
a long time to gain direct experience. Consequently ”steady” workers are demanded. Pay rates
must be such as to avoid high turnover of labor. The incentive of a clear and substantial route of
promotion for operating personnel must be “built” into the plant organization. Certain
communities afford a larger market for the steady type of workers, and these must be kept in view
in the selection of the site, as labor problems should not be allowed to interfere with the continuity
of operations.
If housing, school and living facilities are not available in the vicin ity of a favored plant site,
such site must be penalized to the extent of adding the cost of such facilities to the first cost of the
venture. In America, the operation of such facilities introduces additional management burdens.
However, it behooves the management to provide housing facilities on a cooperative basis.
Ammonia in the form of crude coke-oven liquor must generally be contracted for, since it is
hardly a commodity with standardized market availability. Since synthetic ammonia, as anhydrous
liquefied gas in tank-car lots, has become cheaper, the avidity of ammonia from coke-ovens is no
longer so important to a prospective alkali lent as it was 15 years ago. It was a widespread practice
before the first World War (1914-1918)for alkali manufacturers either to operate these own
by-product coke ovens or to locate near a by-product coke plant. Today, as a matter of fast, the
combination of the alkali industry with the synthetic ammonia industry has been worked out very
advantageously to the ammonia sods industry .
When anhydrous ammonia is being used, the consumption of sulfur will be much higher.
Although the quantity of sulfur in any case is small, the unit cost can be quite high, as was pointed
out in Chapter XXVL. Sine the quality of the product is quite seriously dependent on the sulfur
control, this item should receive due consideration. However , where good coking bituminous coal
is available in abundance, the problem of fuel, of coke , of ammonia, and of sulfur required will be
once and for all solved together, when a suitable site has been selected.
Flood, hurricane, and such force majored characteristics of sites should be compared on the
basis of insurance rates and probability of lost production based on known history. The elevation
of the ground and the design of the building should take into consideration the highest flood level
and the maximum wind pressure are noticeable in the construction of certain alkali plants on the
Gulf Coast.
The alkali plant site should leave room for growth. Not only can the main products be
expected to increase at a reasonable rate, but also it is reasonable to expect to expect that
by-products or end products will be developed to the point of finding profitable production
immediately adjacent to the main works. That land is preferred which permits a layout of the
original plant for a methodical growth without introducing undue expense in the piping and
handling of materials. Ultimate plans layout should be visualized also from the operating and
maintenance standpoint.
An alkali plant serves a number of major industries. The consumer products, for which alkali
is a large-scale raw material, are glass, soap, rayon, paper, petroleum oil, etc. The choice of a site
for a new alkali plant must consider the proximity of the other raw materials required for the
alkali-consuming industries .For example, if two sites are very nearly equal in all other respects,
but one of them is reasonably close to sources of glass sand and the other is not, the former would
certainly be the favored location, since it would double the incentive for the location of a glass
plant adjacent to it.
An alkali plant causes no severe industrial hazards such as the fire hazard connected with a
petroleum refinery or the explosion hazard of a powder works. It is not particularly malodorous,
dusty or smoky. Nevertheless, it has chimneys and machinery noises to which some communities
object. It has accumulation of wastes. Inordinate local taxation or other forms of burdensome
legislation may be found in a community unsympathetic to this large-scale activity.
The most carefully planned and operated plant supplies stored occasionally run out or fails to
stock sufficient maintenance materials of great importance. The store stock also represents a large
chare of working capital. Some sited , therefore, have considerable monetary advantage in being
chosen to a favorable market for factory and mill supplies. Likewise ready availability of services
of specialists like gas compressor experts or large foundry and machine shop facilities is a decided
asset in the assurance of lower operating costs.
As was explained in the preface and has been reportedly emphasized in the text, this book is
not intended to replace the services of a consulting engineer or an expert alkali plant designer.
Before the consideration of a new alkali project has advanced beyond the preliminary market and
raw material studies, capable engineering talent should be engaged to criticize such studies and
later to secure the best possible plant site, plant layout and design.
In the foregoing discussion one factor has not received consideration. This may have a
momentous bearing and may exert a great influence on the decision to be made, as far as plant
location and the method of construction are concerned. It may be that this single factor will
outweigh all other. This single factor is that of air-raid protection. When air-raid precaution is to
be considered, a great many of the orthodox provisions for an ammonia soda plant layout may
have to be discarded As long as aerial warfare remains one of the vital means of military attack,
no country can afford to ignore its potential menace. The Principe in voted in the grouping of the
various divisions of the works, in the sequence of operation, in the interrelation among the various
steps on the process, in the convenience of hacking close proximity of one produce to the next step
where it is to be further processed, in the relationship between the raw materials and finished
products-all these may, in part, have to be re-examined and modified. for instance, instead of
locating the buildings close to one another or in rows parallel to one another in a regular formation,
they may have to be scattered far apart and placed irregularly. The buildings them selves may have
to be arranged and constructed just in a certain way. this will be appreciated when we bear in mind
that an ammonia soda plant has many a tall building which might readily become a target from the
air, and that the ammonia soda industry is strictly one of the defense industries of a nation.
However, it has not been considered wise to include this in the above discussion except to mention
that air-raid protection will and must receive due consideration in the location, design and
construction of the ammonia soda plants of tomorrow.
 Chapter XXXI

Centrifugal PUMPS applied to Ammonia Soda Industry

The ammonia soda industry is unique in that large volumes of liquor Have to be handled in
the process: one of its most important raw materials-salt-is used in the form of brine,
approximately 200 cu. ft. of which are required per ton of soda ash made. Besides brine, other
liquors are handled in the distiller and column operations in even large quantities. Further, very
large amounts of cooling water are also required and Distributed throughout the plant. It is no
surprise to find that centrifugal pumps are used extensively everywhere. in an ammonia soda
plant. Centrifugal pumps today have been brought to such a stage of perfection that they are
applicable to a great multitude of services from handing suspensions, such as milk of lime,
sludge liquor,etc.to from pumping feed-water to boilers against a high pressure to supplying
cooling and scrubbing water at moderate or low heada. Indeed, centrifugal pumps are applicable
to all services in an ammonia soda plant except where a comparatively small volume of liquor is
delivered against a very high head, where probably a plunger type pump may be more
economical. Today large units of centrifugal pumps may have 86-88%of mechanical efficiency.
centrifugal pumps are of two main types-the turbine type and the volute type. The turbine type is
one in which diffuser vanes are incorporated in the casing to guide the flow of the liquid as it
emerges from the impeller, so as to convert as much velocity head as possible back to pressure
head. It is the reverse of a water turbine, although, as commonly constructed today, the canes in
the runner or impeller are curved in the opposite direction. The clouted type has no vanes in the
casing but depends upon a spiral-shaped chamber to guide the flow and to convert the velocity
to pressure. today, very few turbine pumps are made except in multi-stage, high-head pumps,
and by far the majority of the centrifugal pumps used in the ammonia soda plant are of the
volute type because of their simplicity of design and low first cost. In what follows, we shall
therefore confine our attention to the clouted Type pumps. (See Fig.132.)
 FIG 132 Centrifugal pump, volute type.

Centrifugal pumps may be classified as single-stage or multi-stage, According as one or

more impellers are used on the same shaft. They may be single-suction or double suction,
according as the liquid enters the Impeller from one side only or from both sides. They may be
horizontal or vertical, according as the impeller shaft is in a horizontal or vertical Position. The
vertical pump is also known as the vertical submerged Pump, with the pump casing totally
submerged in the liquor and the Motor or turbine vertically above it. Such vertical pumps are
frequently Used for pumping sump tanks, or pits, or underground storage tanks One special type
of vertical pump is the deep-wall pump capable of Pumping liquid (underground bring or oil) up
to 500 ft. depth below Surface with a large number of impellers. But these are generally turbine
type centrifugal pumps with impellers of comparatively small Diameter in the casings (bowls)
and with diffuser vanes turned axially upward to reach the suction of the next bowl above.
Because of limitation of the well bore, the impeller diameter must necessarily be kept Small. To
obtain the head required, as many as 30 stages are not infrequently employed.
The impellers may be classified as enclosed or open, depending on Whether or not the
impeller vanes are shrouded on both sides. Impellers May be further classified as
forward-discharge, backward-discharge, or Radial-discharge, according to whether the vane tips
toward or Away from the direction of rotation, or are straight radically (Fig.133). As may be
seen below, a forward-discharge impeller has the tendency to overload the pump or increase its
head, as the discharge is increased and so is not commonly used. A radial-discharge impeller has
no such Tendency and generally maintains a constant head, regardless of the discharge. This
type of impeller is sometimes used in high head pumped. A backward-discharge impeller, on the
other hand, generally causes a Drop of head as the discharge increases, and is the one most
generally.
used. These statements hold only generally, because there are other factors influencing the
ultimate characteristics of pump.
An open-impeller centrifugal pump has generally only one suction on one end of the shaft
opposite to the drive, the liquid entering axially from that end. Open-impeller pumps are used to
a certain extent in Soda plants for sludge-laden or scale-forming liquors for the following
Reasons:
(1) greater passage space for liquor through the impeller;
(2) simplicity and low cost of liquor through the impeller;
(3) ability to handle suspended solids, crystals or mud;
(4) ease in cleaning when the impeller is clogged or has scale built on it;
(5) one stuffing box or one source of leakage only.
Open-impeller pumps with a single suction t the end are now built integral with the mortar
frame with the impeller mounted on the motor extended shaft, such as “motor pump” made by
the Ingersoll-Rand Co. and“Close-Cupld” pump made by Gould’s pumps, Inc. This construction
is stung, compact, and rigid with no possibility for the pump and motor to get out of alignment.
Further, it is possible to turn the volute casing so as to set the discharge nozzle at any position
desired.
Recently, in order to avoid leakage of corros ive liquors though the packing gland around
the pump shaft, an open –impeller, single-suction centrifugal pump without nag stuffing box has
been perfected and placed on the market, such as that made by A. R. Wiley & Sons, Denver,
Colorado, and a vertical pump node by Allen-Sherman-Hoff. This is specially adapted to the
handing of corrosive acids, chemical liquors, slurries, slimes, sand sledges, etc. In the case of
the Wiley type pump, when the pump is running, an “expeller” by its centrifugal action prevents
leakage of liquor around the shaft, When the pump is not running a governor attached to the
motor draws the pump shaft back by its spring action, closes the clearance around the shaft by a
conical seal ring, and stops nag leakage.
On the other hand, open-impeller pumps have lower head and lower hydraulic efficiency
than enclosed impeller pumps. In all cases, where liquid in question is clean and clear, and is to
be headed in large volumes at a considerable head, a double-suction, enclosed-impeller (instead
of open impeller) pump is preferred, because here power saving will be an item. An
open-impeller pump with a single smuggle at the end is not balanced, the axial thrust being
toward the suction end. An enclosed impeller with double suction may be looked upon as tow
angle open impellers placed back to back with two inlets on opposite sides, and is therefore
essentially balanced. Besides, the shroud serves to guide the flow through the impeller more
accurately. Hence an enclosed-impeller pump has a higher head and greater efficiency then an
open-impeller one.
Compared with reciprocating pumps, centrifugal pumps have no valves, have only one
moving part, and are light in construction. They are self-balanced, so that only a light
foundation is necessary. They have a large capacity per unit weight and consequently a low first
cost. They may handle considerable solids in the liquid with immunity, and will not cause any
accident if the delivery valve happens be completely shut. Unlike a piston pump, a centrifugal
pump cannot stand much leakage of air at the intake or around the packing giants, and must be
primed to start the operation. Hence it is general practice to place the pump below the liquor
level wherever possible, I. e. , and flooded suction.
A single-stage enclosed-impeller pump can operate efficiently at a head from 160to 180 ft.
running at 1800 or 3600 r.p.m. with not too large an Impeller. The overall efficiency may be
from 80-83 per cent. For a high-head service, better efficiency is obtained with a small impeller
diameter and a high speed (r.p.m.), than with a large impeller diameter and a lower speed. The
reason is that the rotation loss, due to spinning of the impeller in the liquid, varies as the fifth
power of the diameter of
The number of the impeller vanes or blade is usually 6, and may very from 4 in small ones
to 8, 10 or more in large ones. Too few blades or vanes in the impeller cause insufficient
guidance to the flow of the liquid through the impeller, while too many blades may restrict the
liquid passage and cause excessive fiction losses to the flow. In very large impellers, sometimes
half vanes are provided near the periphery between two adjacent full bladed. This virtually
doubles the number of the impeller vanes, and improves the hydraulic efficiency, but without
much increase in friction losses.
Centrifugal pump has the follow in five items to be considered:
(1) Head
(2) Speed
(3) Capacity
(4) Brake Horsepower
(5) Efficiency
A head-capacity curve is shown in Fig. 134.
The highest head corresponding to no discharge is the so-called” shutoff” head. As the
discharge increases from zero, the head generally falls. Sometimes, as the discharge increases
from zero, the headfirst rises and then falls. Sometimes also the head at first remains practically
constant and then falls sharply as discharge is further increased. A pump is said to have a rising,
flat, or falling characteristic according as the head rises, remains practically constant, or falls, as
the delivery calve opens from the shut-off position. A falling characteristic cur very calve opens
from the shut-off position. A falling characteristic curve is again referred to as “steep” or “flat”
depending on whether, when the head decreases, the capacity increases slowly or rapidly.
Services where more or less constant discharge is required favor the steep characteristic type of
pump. It is generally true that a forward-discharge impeller has a rising characteristics, a
radial-discharge impeller a flat characteristic, and a backward-discharge impeller a falling
characteristic, although sometimes backward-discharge impellers may have a distinct rushing
characteristic (see below).
A centrifugal pump is essentially a high-speed relative machine adapted to be direct
connected to an electric motor or a steam turbine, although a belt drive, such as V-belt drive, is
often used. If the pump is driven by a steam turbine, the speed may be change as desired: but if
it is direct connected to a squirrel-cage motor, it is essentially a constant-speed machine. The
brake horsepower (i.e.) generally increases as the head decreases and the discharge increases.

FIG 134 Performance curve of a entrifugal pump.

The rated capacity is generally chose at a point of maximum efficiency on the

head-discharge curve, or at a little beyond that, and is known as the rated, or full-load, capacity.
At zero discharge (or shut-off position) the horsepower required may be only 20-40 per cent of
the full-load brake horsepower, At a head lower than the rated, the discharge greatly increases,
and the pump may require a much greater power for drive-so much so that a constant-speed
motor driving the pump may run dangerously hot at a low head and high discharge, especially it
the pump happens to have
rising characteristics. It is generally desirable so to design the impeller that the b.h.p.-capacity
curve is almost flat or slightly increasing to a maximum, then decreasing rapidly. Fig 134 shows
the head-capacity, b.h.p.-capacity, and efficiency-capacity curves of a backward-discharge,
single-stage, and volute pump.
The b.h.p. is determined by a dynamometer , and the discharge is generally measured by a
nozzle or an orifice between pipe flanges. A mercury manometer and a pressure gauge measure
the suction and delivery pressures respectively. A pyrometer gauge attached to the nozzle
determines the volume of discharge for a standard nozzle.
When an orifice is used, a differential manometer is required. The output
of the pump is :
HQg HQg
Liquid h.p.= *8.33=
33.000 3960
where H=head in ft
Q=discharge in gal.per min.
s=sp.gr.of the liquor
144PQ*8.33 PQ
or Liquid h.p.= =
62.4*33.000 1715
where P= pressure in lbs.per sq. in (ga.)
The second formula does not need the specific gravity of the liquor being
 Pumped. This is the pump output. The input is measured by a
dynamometer giving
WR 2nN WRN
b.h.p. = * =
12 33.000 63.000
where W=lbe. on the dynamometer scale
R=length of knife-edge arm in class
N=r.p.m.
This is the input which may be further checked by the voltage and
amperage of the dynamometer motor when its efficiency is known :
volts*amp*effciency
b.h.p.=
746
The efficiency of the pump is
liquid h.p.
e= *100%
b.h.p.

The efficiency of a pump of moderate size nowadays is as high as 80-84 Per cent or more.
his efficiency increases as the discharge increases Form zero, because at a low discharge, the
output is small and the mechanical and hydraulic losses constitute a big item ion the input. The
efficiency then gradually rises to a maximum as the pump is brought to the rated discharge. but
beyond this point ,the efficiency drops again, because at a high rate of discharge beyond a
certain limit, high velocity of the liquid at the entrance to the impeller causes excessive drop in
pressure, resulting in a high vacuum and consequently in the vaporization of the liquid,
rarefaction, or cavitations. This causes high consumption of power without commensurate
increase in the discharge. Therefore the eminency finally falls sharply, anf discharge reaches a
figure, which it cannot exceed even if the head drops to zero.
As mentioned above, at the rated capacity, the b.h.p.may not be at Its maximum, but may
continue to rise in spite of the fact that the head Is rapidly falling. When the pump is
direct-connected to a constant Speed motorist is advisable to use a motor with its rated h.p.not
merely Large enough to drive it at the rated capacity, but large enough also to Take care of any
increased power required when the pump happens to be operated at a lower head then the rated
and consequently at an increased discharge; otherwise the motor then might be dangerously
overloaded. Especially is this so in an ammonia soda plant where most of the liquors Form scale
in the pump and on the impeller, and where the liquors pumped may contain a certain porting of
crystals. This demands an increase in power in the pump operation and so generally a 40’C. rise
motor is preferable because of its 15 per cent service factor.
For the benefit of the operatives in the ammonia soda industry, we shall now discuss in
more or less detail the relationship between the speed and size of impeller, head, capacity, brake
horsepower, and efficiency of a centrifugal pump. These statements will apply mostly to the
backward-discharge, single-stage, and volute type, horizontal centrifugal pumps.
A. Effect of change of speed. If the r.p.m. of a centrifugal pump are increased say by a
change of the motor or steam turbine speed), the peripheral speed u2 of the
u22
impeller is increased proportionately. As the shut-off head h0= , the head
2g
increases as the square of the r.p.m.
To maintain the same efficiency, the relative velocity of the liquid, v2, through the impeller must
be proportionally increased. Since the discharge Q is directly proportional to this relative
velocity, the cross-section of the impeller passage remaining constant, the discharge increases
directly as the r.p.m. Since power is the product of the head into discharge per unit time, power
required increases as the cube of the r.p.m.
B. Effect of change in impeller Diameter, other dimensions remaining constant. The
increase in the impeller diameter, D, proportionally increases the peripheral speed of the
impeller for a given r.p.m. Therefore the relationship given above applies here. The general
practice with centrifugal pump manufacturers is to adopt, within a limited range, a certain size
of volute casing capable of being fitted with any of the several sizes of impellers of the same
pattern, but differing in the outside diameters. This leaves different space in the volute chamber
for free vortex. Also, a change in the diameter of the impeller sets up a new condition in the
passage for the liquid through the impeller. Thus, the relationship
h0&D2
Q&D
b.h.p.&h0Q&D2
will hold approximately. Therefore, to meet the requirements in the head desired when the
discharge required is very small, it is sometimes necessary to go into the extreme impeller
diameter, if the r.p.m. can no longer be increased. Here the friction losses due to disc rotation
increase rapidly with increase in the impeller diameter, and so high-head pumps that require
excessively large impeller diameter for a limited volume of discharge necessarily have lower
efficiencies.
D. Effect of Change of Impeller to Another in Homologous Series. In a homologous series
we have a family of impellers each having the outside diameter, the diameter of the: eye:, the
width of impeller, the cross-section of liquid passage, the space in the volute chamber, the
curvature of the vanes---all in geometrically similar proportions. Let
D=impeller diameter in inches, N=r.p.m, Q=discharge in gals.
per minute. Then

u22 (d/12*#N/60)2 DN
h0= = =( )2 approx, and
2g 2g 1840
Q=cross-section of impeller pam ge*v2
&D2*D
&D2
Therefore the shut-off head increases as the square of the diameter of the impeller for a
given N, and the discharge increases as the cube of the diameter. Consequently the power
required increases as the fifth power of the impeller diameter at a haven r.p.m.
 FIG 135 Disk friction as a function of viscosity.

HQs
Effect of Change of specific Gravity of Liquor. The liquid h.p. as given above equals .
3960
Therefore, other conditions being equal, the power required increases directly as the specific
gravity. Contrary to the common belief, the head of a centrifugal plump is in dependent of the
specific gravity of the liquid, provided the change in spacing gravity is not accompanied by a
change in viscosity. A centrifugal pomp rated at 100 feet head on water ill develop the same
head when it is pumping brine(specific gravity=1.20) or strong sulfuric acid (specific gravity
1.84), making suitable allowance for the increase in viscosity, provided there is sufficient power
in the motor to develop the same r.p.m. And this demands a greater power. Further, the power
required is actually Greater than corresponds to the specified gravity because of greater Friction
losses due to more viscous liquid. Also, the discharge would not be affected by a change in
specific gravity, provided the viscosity remains Substantially.
E. Effect of change of Viscosity of liquid .The increase in
Viscosity greatly increases friction losses in the suction pipe, the impeller the volute
chamber, and the delivery pipe. Teffect, however, is rather complicated and cannot be simply
formulated. An insight may be gained from the manner that this viscosity factory n is involved
Dus
in the Reynolds Number (where D=diameter of conduit in feet, v=relative velocity in feet
u
per sec,8=specific gravity, ands=absolute viscosity in poises).Friction factor increases as the
Reynolds Number decreases and so the larger u is, the smaller is the Reynolds Number, and the
greater is the friction. In general, increase in the viscosity of the liquid causes decrease in the
head, in the capacity, and in the efficiency. and increase in the disc friction(see Fig.135)and in
the power required. However, Prof. Daugherty *found that for small increase in viscosity,
Varying from 31 sec. (Say bolt Universal Viscosimeter) for water to 70sec. For light crude
oil-absolute viscosity varying from 1 to 10 centipoises-the characteristic head-capacity curve of
the pump was not materially affected. However, a larger increase in viscosity would cause a
steep drop in the head-capacity curve of a pump, and a considerable reduction of the maximum
efficiency point on the efficiency-capacity curves; and this effect is more marked for small
pumps than for large ones. For the same reason, increases in viscosity cause a larger percentage
decrease in efficiency, capacity, etc. in a small pump than in a large pump. Further, the increase
in viscosity affects the behavior of the liquid in the impeller passage and in the volute casing so
fundamentally that the same pump designed originally for maximum efficiency on water may
not operate most efficiently on viscous liquids such as heavy oils. It seems that for heavier oils
whose viscosities are much higher, an Impeller having a somewhat larger vane angle will
operate more efficiently.
F. Effect of change of Suction Lift, there conditions remaining unchanged. Increase in
partial vacuum or suction at the pump intake causes marked decrease in discharge at low heads
where very high discharge is normally expected. On the contrary, at high heads or with the
delivery valve throttled to reduce the volume of discharge, normal head.(or identical
characteristic head-capacity curve)thesis obtained How small this discharge should be throttled
to, in order to maintain the “Investigations of the Performance of centrifugal Pumps when
pumping Oils,” by. L. Daugherty, Gould’s Pumps, Inc, Bulletin 126. “A Further Investigation of
the Performance of Centrifugal Pumps When Pumping Oils,” by R.L.Daugherty. goulds Pumps,
Inc, Bbulletin 130. head, depends upon the suction conditions at the intake. At 25’Hg. suction,
the discharge must be reduced to less than one-half of the rated capacity before the normal head
capacity relationship can be maintained The b.h.p.required is greatly reduced at lo heads as the
vacuum at the intake increases, evidently because of the decreased volume of discharge. For the
same reason, efficiency of the pump generally decreases with a high vacuum at the intake. Also,
high vacuum tends to draw in air at the intake or around the packing gland. Consequently, the
pump will have difficulty in priming or, may even lose priming during operation. To prevent
leakage of air through the packing glands, pump manufacturers provide deep stuffing boxes so
as to receive not less than six rings of packing material with a land tern ring in the middle of the
stuffing box, into which liquid from the volute chamber may be led by connecting a small pipe,
one to each side of the pump.
F. Effect of Gases or Volatile Components in Liquid. When the liquor pumped contains
any volatile gases, such as CO2 or NHs in the warm filter liquor, centrifugal pumps
must not be operated at suction. As a rule, in a soda plant it is wise to locate, wherever
possible, all centrifugal pumps below the liquor level so that suction may be flooded.
When a centrifugal-type feed-water pump is installed in the boiler plant, handling hot
feed water from the desecrator to the boiler, static head of the hot water column above
the feed-pump suction sufficient to develop a pressure in excess of the vapor pressure
of water at the temperature at which the desecrator is operated, must be provided. this
may be as much as 20 feet above pump sucton. The presence of the gases in the
centrifugal pump causes decrease in the discharge and in the head attained. When much
gas is present, the pump fails to deliver any liquid,i.e., “loses priming.” Due to the
deceased volume of discharge, The power required is somewhat less, while the
efficiency drops greatly nested with that from a high suction that the same remedy is
resorted to, namely by throttling down the volume of discharge pending the removal of
the cause of trouble. bib-cock is provided at the highest point or “crown ” of the volute
chamber to vent gases, after bringing the pump to rest.
Two similar centrifugal pump a may be operated in parallel when they have about the same
head. Then the discharge will be the sum of the two, but the head will be the same as that of
each individual pump. This is like a parallel operation of two alternators with the same electrical
characteristics. The current is increased but the voltage remains the same. When tow such
pumps are connected in series, the discharge of one going to the intake of the other, the capacity
of the two so connected is the same as that of each, but the head is the sum of the two. These are
just like a two-stage pumo, each stage being a separate unit. But seldom are centyeifugal pumps
cannected this way.
In the trade, the size of a centrifugal pump is usually designated by the diameter of the
delivery nozzle. As mentioned above, the head and capacity of the pump are chosen at the point
of maximum efficiency of the pump when operating at a given speed (r,p.m.).The maximum
head developed by a pump is understood to be its shut-off head when the pump just fails to
deliver.A typical data sheet showing all characteristic curves of a particular centrifugal pump is
shown in Fig.136.
Centrifugal pump are designed and fitted specially for a given speed at a maximum
efficiency for a certain head and capacity. Changes in conditions of operation in the field may
cause the pump to fail to attain that efficiency.

Centrifugal pumps were known more than a century ago_in fact Euler in 1755 propounded
the theory-but real application did not begin until about 1849,when Appold in England invented
the pump and introduced it into practical use. Appold was the first to use curved vanes, but the
introduction of the diffuser vanes in the casing was generally credited to Osborne Reynolds of
Reynolds Number fame, who built a pump with diffuser vanes in 1875.Up to this, construction
was still imperfect and efficiency very low, due to fact that high tentative machines were at that
time unavailable. Not until about the close of the nineteenth century, when high-speed turbines
were perfected, was efficiency improved and the pump considered commercially applicable. In
1896,Sulzer of Winterthur, Switzerland, and about the sometime, Byron Jackson of San
Francisco, Cai, began to manufacture centrifugal pumps on a large scale.
As the name implies, centrifugal pumps work on the principle of centrifugal force. The
impeller, or runner, by its high relative speed, imparts centrifugal motion to the liquid. The
velocity attained by the liquid as it leaves the impeller is converted to pressure as much as
passable. While pressure of a liquid may be converted into velocity readily , the reverse of
transforming velocity back to pressure cannot be accomplished completely. Consider the
impeller shown in Fig.137 rotating in the direction of the arrow at N r,p.m. Let v1,v2 be the
tangential velocities at the circumference of the “eye” and at the impeller rim respectively.
V1,v2 be the relative velocities of

the liquid with respect to the impeller at the entrance and outlet of the impeller respectively.
V1,v2 be the absolute velocities of the liquid at the entrance and at the outlet of the impeller
A1,a2 be the angles, which these absolute velocities make with these tangential velocities
respectively.
B1,b2 be the angles, which the relative velocities of the liquid in the impeller make with these
tangential velocities respectively, produced in the mega tie direction. D is the impeller diameter
in feet. W is the weight in lbs. of the liquid passing through the impeller per sec and g gravity.
since W lbs. Of the liquid passing through the impeller per secant the impeller imparts to the
liquid an absolute volute velocity Vested impeller is imparting to the liquid a tangential
w
momentum *Vs coos &2per sec. By the principles of mechanics, change of angular
g
momentum (or moment of momentum) with respect to time is equal to the torque

w d
V cos &3 * =Torque
g 2 2
Also.
Torque×Angular Velocity=power
w
.V2 cos &2 ×Angular Velocity=Pwer=Wh2 (where h2is head in ft.
g
induced by the rotation
of the impeller).
Since
2#N
Angular Velocity= per sec.
60
 w D 2#N
.V2 cos &2× × =Wh2
g 2 60
But
#DN
=U2
60
u2V2cos &1
h2=
g
In the same way, we might derive the head induced by the rotation of
The impeller eye for the liquid at the entrance. Thus,
u1V1cos&1
H2=
g
And the net head developed is
u3V2cos&2-u1V1cos &1
H2-h2=
g
Since in the actual construction of the centrifugal pumps, the liquid
Enters the “eye”{of the impeller at right angles to the plane of the impeller
(or axially)and even with a whirling motion as it enters the impeller, &1
is practically 90· and cos &1 approaches zero.
u2V2cos&2
Theoretical head developed=h‘=h2-h1=
g
Now from plane geometry, in the parallelogram of forces we have
The following relationship:
V2 2 =u2 2 V2 2 —2u2 V2 cos&2
solving for u2 V2 cos&2
Solving for u2V2cos &2 we have
U@2+V22-v22
U2V2cos&2=
2
u22 v22 V22
H’= - +
2g 2g 2g
in the actual operation of pump ,the liquid passing through the impeller with a relative velocity
v22
V2 sustains some friction losses which we may put as K1 (K1 being some constant).Also,
2g
as mentioned above, it is not possible to convert all the velocity head into pressure head. Call
K2the fraction so converted . Therefore, in the actual pump, the actual head developed is
u22 v22 V2
h= -(1+K1) +K2
2g 2g 2g
Now when the pump is operation with the delivery completely cut off, or when the pump is
operating against a head at a point where the flow of the liquid (or discharge)just ceases, then
there is no definite flow of the liquid through the impeller, i.e, the relative velocity v2=0. Under
such conditions, the liquid just spends itself with violent eddies in the volute casing, and very
little of its velocity head is available for Building up pressure head. thus, the head is all due to
the :whirling
Head” of the liquid in the pump. This means that the fraction K2 is also practically zero.
This leaves on the means that the fraction K2 is also practically zero. This leaves on the
right-hand side of equation
u22
(3) only the first term, which we may call the shut-off head h0,of the pump and so we may
2g
set the actual head;
v22 v22
h=h0-(1+K1) +K2
2g 2g
u22
(4) and the shut-off head h0= . But in an actual pump, this is not exactly the case. for we
2g
u22
find that h0may be forum 0.85 to 1,10 Now when a pump is operating, as delivery valve
2g
is opened and Discharge increases form zero, the head may either increase, diminish, or Remain
practically constant, as it approaches its normal capacity, According as the second term is
smaller then, larger then, or equal to, The third term on the right-hand side of Equation(4),or as
u22 V22
(1+K) K2
2g 2g
This then determines qualitatively the nature of the characteristic headCapacity curve of the
pump in question. Again, from the parallogram of forces above,
V2 cos &2=PQ=u2-u2 cos B2
u2V2cos &2 u2(u2-u2 cos B2)
h’= = -form(1)
g g
u22 u2v2 cosB2
= -
g g
Here B2 is the vane angle of the impeller at its periphery.
For a given speed (N r.p.m.) of the pump, v2 is a constant. Therefore From Equation (5) the
smaller the angle B2 the induced. Consequently, And the smaller is the head, h’ that can be
induced. Consequently, for a high head pump, the designer often resorts to using a larger Vance
Angle B2. but a larger vane angle B2 given magnitude of the relative Velocity V2 leaving the
impeller for a given magnitude of the relative Velocity V2, which can be shown in Fig.138. thus,
a larger B2 results in a higher absolute velocity, V2 of the liquid as it emerges from the Impeller.
But this is disadvantageous, because the conversion of the Velocity head back into pressure head
 FIG 138 Abeolute velocities of liquld resulting form varying vane angles.

is at best inefficient. The usual Practice is to keep V2 low. Therefore, the design as regards the
vane Angle represents a compromise. In practice, the angle B2 is usually Between 20 and 30.
hence the impeller vane looks much like a long Scroll. To go much lower than 20 is not practical,
because this would Seriously. Restrict the passage of the liquid through the impeller. Also, for a
given head it would require a much higher peripheral speed of the Impeller (greater r.p.m.) for
the pump to develop the head required, thus incurring greater disc loss. Now let us again
consider equation (5) in which the second term on the right-hand side appears in the negative
sense. The vector quantities V2,V2 and g are positive. The value of h’ will increase, diminish,
or Remain the same, according to whether cosB2is negative positive, or zero. In a forward-
discharge impeller, B2 is greater than 90 and cosB2 is Negative. Fig.133(3). As the discharge
increases from zero, approaching the normal capacity, V2 increases. Consequently, the
induced h’ Increases. Hence a forward-discharge impeller has rising characteristics, and tends to
overload the motor as the discharge increases. On the Other hand, in a backward-discharge
impeller, B2 less than 90 Generally very much less-and so cosB2 is positive, and the head h’
Tends to diminish as the discharge increases from zero. This shows how a backward-discharge
pump has generally a falling characteristic. If Now an impeller has radial a-a radial-discharge
impeller-then B2=90, cosB2=0, and the second term on the right-hand side of Equation (5)drops
out. Consequently, in a radial-discharge centrifugal Pump, the head of the pump is independent
of the volume of discharge Hence by controlling the vane exit angle B2 it is possible to obtain a
Sharply drooping characteristic for the head-capacity curve, which will Not overload the motor
at low head and high discharge. Generally speaking, the smaller the angle B2, the less tendency
it has to overload. Also By changing the inlet angle B1 (Fig. 137) it is possible to obtain
different efficiencies from pump. A higher-speed pump generally requires a larger inlet angle B1
to secure a greater efficiency. Thus it is seen that by changing there vane angles B2 and B1, the
performance of the pumps May be altered as desired. And this is then definitely checked against
Test results in the laboratory.
However, these are only general and qualitative statements subject To the influence of
many other factors. For instance, as mentioned above, Cases are known where a
backward-discharge impeller pump with a vane Angle of 30 of less had a distinct
rising-characteristic head-capacity
Curve and tended greatly to overload the motor at a low head and high Discharge.
To understand these factors more Cleary, we shall discuss more in Detail the efficiencies
and the various sources of loss in a centrifugal Pump. Power is supplied to rotate the impeller at
a definite speed so That the pump may deliver a certain quantity of liquid in a unit time against a
definite head. This power is called the input, and the work in lifting the liquid to a certain height
against a certain pressure constitutes the output. This useful output is also known as liquid h.p.
On the input side, power is supplied sufficient not only to raise a certain Quantity of the liquid
in a unit time to a certain height, but also to Overcome (a) the bearing friction of the rotating
shaft, and b the Disc friction of the rotating impeller in the liquid. Bearing friction is
Generally a small item of loss, but the disc friction loss is often a considerable item, especially
when the impeller diameter is large and is running at a high speed, such as in a high-head pump.
This relative Disc friction consumes from 2 to 6 per cent of the total power input, while bearing
and gland friction amounts to only a fraction of 1 per cent for a moderate- size pump. The disc
friction for an open impeller is Greater than for an enclosed impeller of the same diameter, and
this is also one reason why an open-impeller pump is less efficient.
On the output side, we have (c) hydraulic friction losses, such as Friction in the passage of
the liquid through the impeller caused by the Rough surfaces underneath the impeller vanes.
There are also (d) eddy along the curvature of the impeller vanes. There are also (d) eddy
Current losses within the impeller passage, due to the fact that the Velocity of the liquid over the
whole cross-section of the impeller passage Is not uniform (the velocity of the liquid behind the
concave side of the Vane in a backward-discharge impeller being greater than of the Liquid in
front of the convex side); as a result, local eddies are set up Whirling from behind the concave
side of the vane to the convex side of The following vane in the passage, there eddy losses
within the passage In the impeller will be better appreciated when we consider that it is Mostly
by this sort of action that the liquid keeps itself circulating when The pump is operating at its
full speed with the delivery valve completely Shut off. The total hydraulic losses for ( c) and (d)
above may amount to from 2 to 5 per cent. Next as the liquid emerges from the impeller we
have (e) losses due to sudden change in the direction of the resultant Velocity (or “shock”
losses), which the liquid is obliged to make as it leaves the impeller. Such losses caused by a
change in direction are also called diffuser losses. “also , there are friction losses in the
whirlpool Chamber of the casing . of all the individual items of losses both on The input and
output sides, this single item of the diffuser loss is probably The largest, this may account for the
loss of from 7 to 15 per cent of The total power input depending upon the design of the impeller
relative To the casing and the viscosity and

FIG 139 Turbine pump with cirecular casing.

specific gravity of the liquid in Question. It is often said that while a pressure head of a liquid
can be Readily changed to a velocity head, the converse is not so; and by the Nature of the usual
construction of the pump casing with respect to the Location of its outlet nozzle, such losses will
be present and may be Minimized but cannot be entirely eliminated. On the other hand with A
poorly designed casing these “shock” losses may assume still greater Proportions it is to
minimize these “shock” losses that diffuser vanes Are provided in the casing, in which case the
angle of each of the diffuser Vanes is made equal to a2 (see Fig 139) for this purpose the
Pomona Pump company pompons California have developed a type of “bulbous” Vane end
shape for the diffuser vanes to give the streamline effect (see Fig 140) it is claimed that by the
use of this bulbous shape to guide The flow of the casing the efficiency of the centrifugal pump
May be increased by as much as 10 per cent
 FIG 140 Diffuser vane and shapes.

Finally, because of mechanical difficulties it is impossible to isolate Completely the suction

port of the pump from the delivery chamber and So a certain portion of the liquid under the
differential pressure between The delivery and suction will find its way back to the suction
around the So-called wearing ring into which the “eye” of the impeller fits, this Means that a
certain amount of the liquid from the delivery side will Return to t he suction and the leakage
represents just so much waste of Effort. Such leakage loss (f) in a large pump constitutes only a
very Small item but in a high-head pump, having a small volume of discharge, this leakage loss
represents a comparatively large item of power loss. This is another reason why a high-head
pump with a small capacity is not so efficient.
From the above it can be seen that efficiency of a centrifugal pump, defined as
Will be low, in view of the many sources of loss. Nevertheless, centrifugal pump
manufacturers today have so perfected the design with A view to minimizing all these losses
that in a pump of good size an Efficiency not lower than 84 per cent as mentioned above is
easily attainable, even for a volute type of pump without any diffuser vanes. This May be
estimated from the magnitude of losses in items (a) to (f) enumerated above.
As has been mentioned conditions in the centrifugal pumps are very complicated. There are
also opposing tendencies. Actual condition is seldom exactly as represented in the above
theoretical discussion. for Instance, prof. Daugherty of the California Institute of Technology
found that the angle the relative velocity V2 makes with the tangent was not B2 but 4-9 less
than B2 . Again conditions inside the liquid passage Space in the enclosed impeller. Bounded by
the two impeller shrouded and The adjacent scrolled of vanes, are very uncertain, and so the
behavior Of the liquid therein may be influenced by many unknown, and so the behavior
Sequent, exact mathematical relationship is difficult to establish and Much still depends upon
the test data.
 Appendix

Some Important Tables and Curves for Reference

in Soda Works

The following is a collection of some important data and curves not given in the foregoing
text. They should be in the possession of men engaged in the alkali industry as they refer to the
manufacture of soda ash, bicarbonate of soda, caustic soda, calcium chloride, chlorine, etc.
FIG 141 Freesing point curve for sodium carbonate solutions.
 FIG 142 Specific gravity of caustic soda solution at 60℉

FIG 143 Relation between degrees Baume of caustie soda solutions and per cent NaOH at 60℉
FIG 144 Freesing point diagram of pure NaOH and H2 O.*
FIG 145 Heat content-concentration chart of caustic soda solutions.
FIG 147 Viscosity of caustic soda solutions.
Wcathered soda ash, 515 Wyandotte, Mich, 48,56
Wedge furnace,10 Wyoming trona deposit, 21
Weldon, Walter, 8
Weldon process, 1,8,13 Y
Well water, 138,372,522,523 Yangtse river water, composition of , 376
West End Chemical Co., 19,21,24,25 “Yield” of sodium bicarbonate from filters,173
Westvaco Chlorine Products, Inc., 324, 329 Yoder, T.D.,398
Wet calcinations,342 Young, P. (with Willard, H.H.),505
,calculation of,349 Yugoslavis ,41,44
,factors in,344 Yugli Chemical Industries, Ltd., 44
,operating data, 349
,process diagram,348 Z
,steam consumption,349 Zahn & Co., G.m.b.H., 37
Whiechapel, England, 31 Zaremba evaporator, 291
Whitman, W.G., and Davis, G. H. B., Zeo-Karb,Na,390,392
344 ,uses of 392
Widnes, England, 4,8,31 exchange capacity, 390
Wilfiey, A.R.and Sons, 534 Zeo-Karb, Na, 390
Wilfiey, H.H. and Young, P., 505 ,uses of 390,391
Winnington, England, 4 exchange capacity, 390
Wooden plugs, 411 Zeolite, artificial, 379
Workable unit of soda plant 161 , natural, 379
Working of carbonating tower 158 et seq. , regeneration of ,381,388
World output of soda ash, 41 ,treatment of ,378 et seq.
World production of soda, 41 Zeolite softener, 388
World soda plants, 43,44,45 , operation of ,388
World Trade Notes, 45 , ‘”Two-flo”,389
World War I,5,6,20 Zeolites, non-siliceous, 390,391,392,393
Zimmermann-Reinhardt’s method, 484