Pi

leita fot hts

Laboratory glass
LABORATORY GLASS-WORKING
FOR SCIENTISTS
 LABORATORY
GLASS-WORKING
EORSSCIENTEISTS

A. J. B. ROBERTSON M.A., Ph.D.

Reader in Physical Chemistry, University of London
Formerly Fellow of St. John’s College, Cambridge

D. J. FABIAN B.Sc., Ph.D.

Research Associate, Department of Chemistry
University of British Columbia, Vancouver, Canada

A. J. CROCKER B.Sc., Ph.D.

Department of Chemistry, King’s College, London

Je DEW
IN Ge BiSc.ePh.D:
Post-Doctorate Research Fellow, Division of Applied Chemistry
National Research Council, Ottawa, Canada

With a Foreword by
SIR ERIC K. RIDEAL F.R.S.

WMULUEHIA, De,
NEA p
|

y
\ 4 oe
ve —

B-2- 6
——$—-

ONS
4),
.%

- | (- > =
‘ ee %
\ ‘ ry ae cae *La ‘fif
/; = eek ON eal
‘
.0-. eAy o
vs Q ,
a fg £
ASS
* ,

LONDON

BUTTERWORTHS SCIENTIFIC PUBLICATIONS

1957
 BUTTERWORTHS PUBLICATIONS LTD.,
88 KINGSWAY, LONDON, W.C.2
AFRICA: BUTTERWORTH & CO. (AFRICA) Ltd.
DURBAN: 33/35 Beach Grove
AUSTRALIA: BUTTERWORTH & CO. (AUSTRALIA) Ltd.
SYDNEY: 8 O’Connell Street
MELBOURNE: 430 Bourke Street
BRISBANE: 240 Queen Street
CANADA: BUTTERWORTH & CO. (CANADA) Ltd.
TORONTO: 1367 Danforth Avenue
NEW ZEALAND: BUTTERWORTH & CO. (AUSTRALIA) Ltd.
WELLINGTON: 49/51 Ballance Street
AUCKLAND: 35 High Street

U.S.A. Edition published by

ACADEMIC PRESS INC., PUBLISHERS
111 RIFT AVENUE
NEW YORK 3, NEW YORK

in ee

i)MM
FTRI-MYSORE

Laboratory glass.

Set in Monotype Times Roman type

Printed in Great Britain by Robert MacLehose and
Company Limited
The University Press, Glasgow
 FOREWORD

OF RECENT years, with increasing vigour, the argument is advanced

that the day of the individualist in the experimental sciences is over.
Working parties, sections and units appear with increasing frequency
in our research centres and laboratories.
This is an unusual book and will be found to appeal most strongly
to those that believe that in England at any rate individualism in
science has by no means had its day, but is of the very ‘stuff? of our
continued existence as a serious contributor to the scientific advance-
ment of the world. ; :
Many in this country owe a great debt to Sanderson of Oundle who
believed that, in basic education, work with the head should always
be accompanied by work with the hand. Elementary training which
permits a man to achieve a scholarship at Oxford or Cambridge, or
to obtain a good post in a gas works, might be regarded by some as
too diffuse; but at any rate it serves as a corrective to the increasing
specialization which we note around us.
In experimental work it is frequently necessary to construct
apparatus. If standard investigations are under consideration,
standard apparatus may probably be purchased for the purpose; but,
in general, apparatus should be designed and constructed to suit the
purpose in view, otherwise ‘standard results’ only are obtainable and
the investigator is barred from a wide field of scientific enquiry, or,
what is worse, his mind becomes guided and contained by the
machines available and he becomes a captive.
There have been notable blowers of glass amongst our most
eminent scientists, especially in the United Kingdom and in the
United States of America, mainly chemists and physicists; and it is
in these branches of the experimental sciences where apparatus
fashioned in glass is essential for progress. Thanks to the progressive
attention of the manufacturers of scientific glassware, the tasks con-
fronting the scientific investigator are not nearly as complex as they
were in Faraday’s time; but whilst a large variety of standard fittings
is now available, modern apparatus has become more complex and
involves greater manipulative skill in its production. To all those
who are confronted with such problems this little book is heartily
commended.
LONDON 1956 ERIC K. RIDEAL
 CONTENTS

FOREWORD _... ee ws i - - Vv
ACKNOWLEDGEMENTS om . a 7 s Xi
pee ae Bs am a - ne *: Xiil
. INTRODUCTION .. e, l
2. PHYSICAL AND CHEMICAL “PROPERTIES OF
GLASS Re ; oa 7
The structure of glass me wh a os rs
General physical properties of glass os Me #5 9
General chemical properties of glass ote ae ip 12
Devitrification oa =i ‘i oe os - 18
Thermal strain in glass ee he a re a 19
Annealing of glass .. ee a aA 20
Some types of glass for general use we: of an 21
Some types of glass for sealing to metals .. v ve 22
Some types of glass for special purposes .. - vt 2
The ease of working different types of glass os a 25
Distinguishing between different types of glass .. ae 26
3. GLASS-WORKING EQUIPMENT ne Ke a 28
The glass-blower’s table... ms ny < - 28
Blowpipes .. oy a A oe cs x 28
Hand torches - Pe i as fe << 29
Cylinder heads and valves .. oO hi be 2h 30
Bellows v p s as ‘as ey ae 30
Air blowers .. ; 3 - < a s 30
Glass-working tools 2 te BE > 30
Wax for tools - a4 be Re Be 32
Carbons - A a + - ae 33
Gauges for measuring oe a ¥ - 4, 34
Rubber caps 9 aS a 7 = os 34
Rubber stoppers... i “f 7 ve 35
Blowing tubes 3 ee a8 35
The uses ofasbestos. . ” a ae a = 36
Glass-blowing spectacles .. ae By ay _ 37
Glass knives and diamonds ye ‘fe a - 37
Tube-cutting device
 CONTENTS

Clamps
Abrasives.
Glass holders
Large tube supports. .
Adjustable rollers
Glass-blower’s swivel
Polarized light strain-viewer
Glass-cutting wheels
Lapping wheel
Carborundum grinding wheel
Annealing oven
Treatment of slight burns
The amount of equipment required for different purposes
4. BASIC GLASS-WORKING OPERATIONS
The preliminary preparation of materials
The cleaning of glass tubes . .
Breaking glass tubes ~_
Knocking off small bits of tubing w
Holding and rotating tubes in the bench flame
Bends and spirals
Putting a handle on the end of a tube
Drawing tubes off to spindles
Round-bottomed closures of tubes
Fabrication of thin glass rod
Blowing holes in glass tubes
Joining tubes of similar sizes
Joining tubes of different sizes
T-joints
Y-joints
4- and 5--way junctions
Working capillary tubes
Blowing bulbs in the middle of a tube
Blowing bulbs at the end ofa tube
Multiple perforations in bulbs
Flanging, flaring or ee are:
Internal seals
Thin glass windows . .
Spinning out feet
Tapering glass tubes :
Sealing-in sintered glass discs
Mending cracks
Detection and removal of leaks and holes. .
Viil
 CONTENTS

Grinding glass
Releasing frozen glass surfaces
Polishing glass
Drilling holes in glass
Cementing glass
Silvering of glass :
Depositing copper on glass
Metallizing preparations for the mg on “of metalI ayers
on glass
oe CHE MANIPULATION OF LARGE TUBING
General observations
Bending big tubing ..
Drawing off large tubes
Closing large flasks ..
Joining big tubes
Big T-joints ..
Big internal seals
Mending cracks in large tubing
Mending cracks near complex and large seals
6. SOME OPERATIONS WITH A GLASS-WORKING
MACHINE
7. METAL-TO-GLASS SEALS
Matched and mismatched seals
Sealing tungsten into Pyrex
Sealing tungsten into special ag gee
Multiple wire seals .
Prepared copper-tungsten— nickel v wires
Sealing platinum into soda or lead glass
Sealing platinum into Pyrex
Special alloys for sealing to glass
Copper-to-glass seals
8. THE CONSTRUCTION OF SOME TYPICAL SINGLE
PIECES OF eessitt 7 -*
Ampoules
Sealed tubes for reactions oa
Break-tip seals
Distillation flasks .
Distillation splash heads
Dewar seals
Dewar vessels
Cold finger refrigerant traps
iX
 CONTENTS

Filter pumps
Phosphorus pentoxide traps
Condensers . -
Fractio nal distilla tion condens ers 15
Fractionating columns
A fractional distillation receiver
Double surface condensers
Mercury cut-offs
McLeod gauges
Traps to catch mercury displaced iinn apparatus
Valves Me
Soxhlet extraction apparatus
Mercury vapour pumps
Electrolytic gas generator ..
Thermostat regulators
Gas flow meters
Spectrum tubes
Discharge tube lamps ;
Apparatus containing many electrodes
Leaks for molecular flow of gases ..
Bourdon gauges and glass spiral gauges
Circulating pumps ..
Apparatus for semi-micro qualitative analysis
Gas analysis apparatus
3; THE ASSEMBLY OF COMPLEX APPARATUS
General observations
Joining closed systems
Making more than one joint at once
Removal of strain and clamping of apparatus
Annealing by flame .
Use ofa bent blowpipe and double tippingdevice .
Danger of flame cracks
10. THE MANIPULATION OF SILICA
The properties of fused silica and general remarks
General technique for working fused silica
Silica torsion fibres ..
AUTHOR INDEX
SUBJECT INDEX
 ACKNOWLEDGEMENTS

We thank the following Companies and individuals:

Aero Research Ltd., Duxford, Cambridge, for information on cements
for glass
The British Heat-Resisting Glass Company Ltd., Phoenix Works,
Bilston, Staffs., for information about Phoenix glass
The British Thomson-Houston Company Ltd., Rugby, for infor-
mation about glasses manufactured by them
Edwards High Vacuum Ltd., Manor Royal, Crawley, Sussex, for
information about their glass-working machine and for permission
to reproduce FiGurREs 34 and 35
A. Gallenkamp & Company Ltd., 17-29 Sun Street, Finsbury Square,
London, E.C.2., for information about the Davies double surface
condenser, and for permission to reproduce FIGURE 4
The General Electric Company Ltd., Osram Glass Works, East Lane,
Wembley, Middlesex, for information about glasses manufactured
by them
James A. Jobling & Company Ltd., Wear Glass Works, Sunderland,
for information about Pyrex glass
Johnson, Matthey & Company Ltd., 73-83 Hatton Garden, London,
E.C.1., for information about preparations manufactured by them
for the production of fired-on metallized layers on glass
John Moncrieff Ltd., North British Glass Works, Perth, Scotland,
for information about Monax glass
Plowden & Thompson Ltd., Dial Glass Works, Stourbridge, Worcs.,
for information about glasses manufactured by them
Stone-Chance Ltd., 28 St. James’s Square, London, S.W.1., for infor-
mation about glass-working burners manufactured by them
The Thermal Syndicate Ltd., Wallsend, Northumberland, for infor-
mation about Vitreosil
Wood Brothers Glass Company Ltd., Borough Flint Glass Works,
Barnsley, Yorkshire, for information about a glass
D. W. Bassett and J. A. Stone of King’s College, London, for reading
the original manuscript and making many suggestions
V. J. Clancey for information on the method used by him for making
fused silica fibres.

xi
 PREFACE

A NEED for working glass is frequently encountered in scientific

research and teaching, particularly in the field of physical chemistry.
In many laboratories this problem is solved by the employment of
professional glass-blowers, and the research worker requires little or
no skill in the manipulation of glass. Or it may be that a laboratory
has no glass-blower and the scientist has to rely on his own resources,
combined perhaps with the services of some outside specialist. In
other laboratories the scientist may find the available services to be
deficient in various respects, or overloaded, so that more satisfactory
progress is made when he himself becomes an amateur glass-worker.
This possibility was emphasized by W. A. Shenstone in 1889, when
he wrote that the amateur with practice can make almost all the
apparatus he needs for lecture or other experiments with a consider-
able saving in expense ‘and, which very often is more important,
without the delay that occurs when one depends upon the pro-
fessional glass-worker.’ This latter advantage is, he writes, for
himself a very weighty one.
We became interested in working glass ourselves when engaged in
various researches in the Chemical Laboratories at King’s College,
London, and over the last few years we have become increasingly
convinced that the scientific glass-worker can use methods which
differ from those of the professional. The scientist is primarily
interested in apparatus which gives him the results he seeks for, and
if this apparatus lacks elegance in appearance and is made by
methods which are looked upon unfavourably by the professional
glass-blower, it by no means follows that the apparatus is defective
from a scientific point of view.
We have therefore compiled this work in the hope that it will be
useful to scientists faced with problems of glass-working. By avoid-
ing the more difficult manipulations involved in professional glass-
working, it seems possible for a scientist to assemble quite complex
apparatus, including, for example, his own mercury diffusion pumps,
mass spectrometer tubes, molecular beam generators and silica
fibres; the preliminary practice required, which depends upon in-
trinsic ability, may take some hours a week for a period of two or
three months. We have also endeavoured to bring together data
Xiil
 PREFACE
scattered in the literature, and to give an account of some of the
more difficult methods of manipulating glass.

The researches which led to our interest in glass-working were

supported by grants from the Research Fund of the University of
London, from the Institute of Petroleum and from the Department
of Scientific and Industrial Research (maintenance grants to A. J. C
and J.D.).
a: R.
Bary.
December,: 1956 vo.

XiV
 Chapter |

INTRODUCTION

GLASS has been known to mankind for some thousands of years. The
manufacture and working of glass became a stable industry in Egypt
at the beginning of the 18th dynasty. W. M. Flinders Petrie (1909)
has described how, about 1370 B.c. in the time of Amenhotep IV, in
the works at Tell-el-Amarna, a lump of glass was patted into a
cylinder and then rolled into a rod which was drawn into cane about
g in. thick. This cane was wound on a mandrel to produce hollow
vases. Later the art of pressing glass into open moulds was dis-
covered. The blowpipe—an iron tube 4-5 feet long with a mouth-
piece at one end—was probably discovered about the beginning of
the Christian era, and, according to G. W. Morey (1938), caused an
industrial revolution. The use of tongs for manipulating glass was
known to the Romans in A.D. 300. By this time, therefore, several
of the basic methods which are now useful in constructing complex
scientific apparatus had been developed. The glass in those early
times was similar in many respects to a modern soda-—lime-silica
glass. Scientific method does not seem to have influenced glass pro-
duction very much before the present century, although of important
early investigations we may mention those of K. W. Scheele and
A. L. Lavoisier on the durability of glass exposed to water and weak
acids, and those of W. V. Harcourt and M. Faraday on the production
and properties of glass.
The value of glass as a laboratory material is very great. A
moderately skilled worker can fabricate complex glass equipment
using simple tools; and perhaps of even greater value in research is
the ease with which complex glass apparatus can be modified and
added to with little or no dismantling. Glass is sufficiently chemically
inert for most purposes, and vitreous silica may be used when extra-
ordinary inertness is required. The transparency of glass is often
valuable. Glass is a good electrical insulator, and metal electrodes in
a glass envelope can be raised to incandescence by eddy-currents
induced by a coil, outside the envelope, carrying high-frequency
current. In vacuum researches glass is valuable because of the ease
with which leaks are found with a Tesla coil, and on account of its
l
 INTRODUCTION

ready outgassing on baking. For nearly all practical purposes glass

is vacuum-tight. Glass—to—metal seals of various kinds may be made,
and different types of glass can be sealed together with appropriate
intermediate glasses.
The manipulation of glass is a craft and has been handed down
over the centuries largely by personal example and tradition. Venice
was the most important centre of glass-working for the four cen-
turies following its rapid development in the eleventh century to a
dominating position; in 1279 a system of apprenticeship needing
eight years was introduced there, and the closed and somewhat secre-
tive nature of the craft was reinforced by the employment of assassins
to remove glass-blowers who seemed likely to give away valuable
secrets to other centres of the industry. Perhaps the aura of mystery
which still to some extent surrounds scientific glass-working can be
traced back to these times.
The basic techniques developed for scientific glass-working involve
the manipulation of glass in the form of tubing or rod, using a flame
as a source of heat. The article to be fashioned is held by the operator
in both hands, and the part to be worked becomes a semi-molten
mass in the flame and is shaped principally by blowing. Generally it
is necessary to rotate the article in the flame, and this often requires
a very high degree of muscular co-ordination, which can be developed
by professional workers to a remarkable extent. This rotation opera-
tion is not easy; M. FARADAY (1842) stated the outstanding difficulty
in the following words: ‘But when the heat has brought the glass
into a soft state, it is by no means easy so exactly to turn the tube at
both ends alike, and so lightly yet equally to hold them, that the soft
part shall retain its cylindrical shape; being neither twisted, nor bent,
nor elongated, nor thrust up.’ R. E. THRELFALL (1946) considers that
one third of the art of glass-blowing consists essentially in being able
to move both hands about, rotating a tube with each finger and -
thumb, and keeping constant both the distance between the hands
and the speed of rotation. Considerable practice is necessary to gain
great mastery of this technique, which we can call the bench-flame
method of work, since the flame is in a fixed position on the bench.
The problem facing the scientist who has to engage in his own glass-
working is that of simplifying or modifying those operations which
depend on extremely good muscular co-ordination. We consider this
problem to be soluble when elegance can be sacrificed to utility.
The requirements of a scientist differ from those of a professional
glass-worker. The starting point of a new research is néw ideas, and
when these require subsequent experimental investigation the ap-
paratus used need be only good enough to give results which are
Zz
2
 INTRODUCTION
satisfactory from a scientific point of view. A glass apparatus must, in
fact, hold together and work, but it may have an unsightly appear-
ance. However, we must also note that experimental skill in itself is
of great value to the scientist in pursuing his ideas, for as Faraday
(1842) wrote: “The person who could devise only, without knowing
how to perform, would not be able to extend his knowledge far, or
make it useful; and where the doubts or questions that arise in the
mind are best answered by the results of an experiment, whatever
enables the philosopher to perform the experiment in the simplest,
quickest, and most correct manner cannot but be esteemed by him as
of the utmost value.’ In addition to these considerations, experi-
mental skill is of particular value in developing new techniques; and
these may lead to many unexpected developments. H. E. ARMSTRONG
(1924) tells how the introduction in 1865 of the Sprengel pump—a
fairly simple piece of glass-working—revolutionized the English
water supply. Again we may note the important practical conse-
quences following on the early work of H. L. CALLENDAR (1892) on
vitreous silica tubes, and that of C. V. Boys and R. Threlfall on
vitreous silica fibres.
In view of the difference in objective between the scientist and
professional glass-worker, we consider that glass-working for the
scientist should develop as its own craft, and indeed this has hap-
pened to a certain extent. An early systematic account is that of J. J.
BERZELIUS (1833), which describes some of the basic operations.
Faraday’s Chemical Manipulation is a masterly treatment, referred to
frequently in R. THRELFALL’S notable work On Laboratory Arts
(1898), which devotes 107 pages to glass-blowing and manipulation
of glass. Threlfall seems to have been one of the first to describe
what is now often called ‘in-place’ glass-blowing, in which the glass
pieces are kept stationary by clamps, and the flame is moved. Joints
are made with thin glass rod, now often called welding rod, by a
method very similar to that used in 1370 B.c. in the Tell-el-Amarna
factory. The difficulty of holding and rotating the glass is avoided,
and thus, as Threlfall says, the method is most useful to the experi-
menter who wants to get on to other things before sufficient skill is
acquired for the rotation method. The tradition of the craft of glass-
working for scientists is continued, we think, by the works of W. A,
SHENSTONE (1889), T. Boras (1898), B. D. Boras (1921), F. C.
Frary, C. S. TAyLor and J. D. Epwarps (1928), R. H. WRIGHT
(1943), and J. D. HELDMAN (1946). Amongst eminent scientists who
have carried on the glass-working craft with their own hands we may
mention R. W. Bunsen, whose skill at the oil-fed blowpipe and whose
exemplary patience when one of his pupils rapidly and several times
B ed
3
 INTRODUCTION

in succession broke the apparatus glass-blown by the master himself

are described by H. E. Roscog (1901); J. Dewar, whose vacuum-
jacketed glass vessels marked a new era in low temperature research ;
M. Bodenstein and W. Ramsay, both of whom are described by E.
K. Ret (1951) as uncannily skilled in the art of glass-blowing ;and
F. W. Aston, whose glass bulb discharge tubes and cooled charcoal—
in-glass pumping systems first gave those exact mass relations of
isotopic chemistry which contributed decisively to the opening of the
atomic age. In modern times the old tradition that the masters of
science should also be themselves masters of the practical craft of
science has failed to persist, and the more eminent scientists are now
usually not to be found at the working bench in the research labora-
tory. There is, we think, a consequent weakening of the craft basis
of practical scientific work, and some retreat from the view expressed
by I. Langmuir that work in the laboratory can be fun.
The increasing use in scientific research of borosilicate glasses over
the last thirty years has not, on the whole, been accompanied by
much departure from the traditional methods of glass-working.
Generally it is much easier to rotate the flame about a stationary
clamped article than to rotate the article in a stationary flame.
Quite complex apparatus can be constructed by working with a hand
torch—giving a movable flame held in the hand—on clamped appar-
atus. This method of work is specially suitable for the borosilicate
glasses of low thermal expansion, which can be worked into quite
knobby apparatus without there being much danger of cracking on
cooling—in contrast to soda—lime-silica glass. The joining of two
clamped tubes with a movable flame has been described by SHEN-
STONE (1889), THRELFALL (1898) and TRAVERS (1901). A .L. REIMANN
(1952) has described some further uses of this general method, but
the great number of complex operations which can be carried out
with facility using a hand torch in place of a bench torch does not
seem on the whole to have been realized ; indeed, HELDMAN writes:
‘End-seals with both tubes of approximately the same diameter and
T-seals are, with practically no exceptions, the whole repertory of
in-place glass blowing.’ We are by no means in agreement with this
statement. Many operations can be carried out with a hand torch
on completely fixed glass, or on a fixed piece of glass to which some
other part of glass can be joined by holding it in one hand whilst the
other hand manipulates the hand torch. The results are usually not
as elegant as those obtained by a skilled worker using rotation in a
bench torch, but less skill and practice are required for the hand torch
method. It has the further great advantage that the complexity of the
apparatus being constructed can be steadily increased without
4
 INTRODUCTION
greatly increasing the glass-working difficulty. Furthermore, the
scientist will probably only work glass from time to time, depending
upon the requirements of the research or other work, and the hand
torch method is far more suitable than the bench torch method for
the operator who does not want to spend time on preliminary
practice; it is also far more suitable for the scientist working in
awkward positions on complex apparatus. We find that very com-
plex apparatus may be assembled with a hand torch.
A clear account of professional methods of glass-working has been
given by W. E. BARR and V. J. ANHORN (1949). Valuable information
on some glass-working problems is also found in the works of J.
STRONG (1940), M. C. Nokes (1948), A. J. ANSLEY (1950), A. ELLIOTT
and J. Home Dickson (1951), and H. J. J. Brappick (1954).
Accounts of the simpler operations have been given by W. E.
PARK-WINDER (1947) and E. H. MorGAN (1953).

REFERENCES
ANSLEY, A. J., 1950, An Introduction to Laboratory Technique, 2nd
Edn; London, Macmillan.
ARMSTRONG, H. E., 1924, Chemistry in the Twentieth Century, edited
by E. F. Armstrong; London, Benn.
Barr, W. E. and ANHoRN, V. J., 1949, Scientific and Industrial Glass
Blowing and Laboratory Techniques; Pittsburgh, Instruments
Publishing Co.
BERZELIUS, J. J., 1833, Traité de Chimie, (Trans. Esslinger) Vol. 8;
Paris, Firmin Didot Freres.
Botas, B. D., 1921, A Handbook of Laboratory Glass-Blowing ;
London, Routledge.
BoLas, T., 1898, Glass Blowing and Working ;London, Dawbarn and
Ward.
Brappick, H. J. J., 1954, The Physics of Experimental Method;
London, Chapman & Hall.
CALLENDAR, H. L., 1892, J. Iron St. Inst., 1, 164.
Exuiotr, A. and Home Dickson, J., 1951, Laboratory Instruments ;
London, Chapman & Hall.
FARADAY, M., 1842, Chemical Manipulation, 3rd Edn; London,
Murray.
FLINDERS Petrie, W. M., 1909, The Arts and Crafts of Ancient Egypt;
Edinburgh and London, Foulis.
FRary, F. C., TAYLor, C. S. and Epwarps, J. D., 1928, Laboratory
Glass Blowing, 2nd Edn; New York, McGraw-Hill.
HELDMAN, J. D., 1946, Techniques of Glass Manipulation in Scientific
Research; New York, Prentice-Hall.
5
 INTRODUCTION
Morey, G. W., 1938, The Properties of Glass; New York, Reinhold.
MorGan, E. H., 1953, Newnes Practical Mechanics ; issues of Octo-
ber, November and December.
Nokes, M. C., 1948, Modern Glass- Working and Laboratory Tech-
nique, 3rd Edn; London, Heinemann.
PARK-WINDER, W. E., 1947, Simple Glass-blowing for Laboratories
and Schools; London, Crosby Lockwood.
REIMANN, A. L., 1952, Vacuum Technique; London, Chapman &
Hall.
RIDEAL, E. K., 1951, J. Chem. Soc., 1640.
Roscoe, H. E., 1901, Chemical Society Memorial Lectures 1893-—
1900; London, Gurney & Jackson.
SHENSTONE, W. A., 1889, The Methods of Glass Blowing, 2nd Edn;
London, Rivingtons.
STRONG, J., 1940, Modern Physical Laboratory Practice; London and
Glasgow, Blackie.
THRELFALL, R., 1898, On Laboratory Arts; London, Macmillan.
THRELFALL, R. E., 1946, Glass Tubing; London, British Association
of Chemists.
TRAVERS, M. W., 1901, The Experimental Study of Gases; London,
Macmillan.
WRIGHT, R. H., 1943, Manual of Laboratory Glass-Blowing ;Brook-
lyn, N.Y., Chemical Publishing Co.
 Chapter 2

PHYSICAL AND CHEMICAL

BRUCE R TIES. ORSGLASS

The Structure of Glass

A GLASS is a product of fusion which has cooled to a rigid condition
without crystallizing. This definition includes a large number of
organic glasses, and does not restrict the term ‘glass’ to inorganic
substances, which is a frequent practice in the U.S.A. This restriction
seems somewhat arbitrary, particularly when we consider how G.
Tammann established the general principles of the glass-like state by
research on organic glasses, and how the devitrification of technical
glasses is paralleled by that of organic glasses. Tammann concluded
that a glass could be regarded as a supercooled liquid in which the
rotational movements of the molecules had been frozen (see W. E.
GARNER, 1952). In fact, as R. Boyle described it about 1660, ‘the
particles of the glass agitated by the heat, were surpriz’d by the cold
before they could make an end of those motions which were requisite
to their disposing themselves into the most durable texture.’ In
modern terminology, a glass is thermodynamically unstable with
respect to the corresponding crystal.
Crystalline silica (quartz, tridymite or cristobalite) in its various
modifications is built up of SiO, tetrahedra linked together in a
regular manner so that every oxygen is between two silicons. The
tetrahedra therefore share corners. The arrangement in space of the
tetrahedra is different in the various crystalline forms, but is always
strictly regular. A silica glass, in contrast, again contains SiO, tetra-
hedra with every corner shared ; but by slight distortions of the valency
angles, as compared with the crystal, a continuous and irregular three-
dimensional network is built up. The orientation about the Si-O-Si
bond of one SiO, tetrahedron with respect to another can be practi-
. cally random. Thus a two-dimensional picture of a silica glass would
show a series ofirregular rings, with an average number of about six
tetrahedra in each ring, but with the number of tetrahedra in indi-
vidual rings varying from three to ten or more. The silica glass
satisfies the condition for glass formation proposed by W. H.
7
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

ZACHARIASEN (1932), namely that the substance can form extended

three-dimensional networks lacking periodicity, with an energy
content comparable with that of the corresponding crystal network.
A glass does not, therefore, produce a regular diffraction pattern
with x-rays; but a monochromatic x-ray beam incident on a glass is
scattered, and a radial distribution curve may be constructed. The
space average of the distribution of atoms round a given one can be
deduced (see J. T. RANDALL, 1938). Much work of this kind has been
carried out by B. E. Warren and his colleagues.
A soda-silica glass results from the fusion of Na,O with SiO,. The
number of oxygens is more than twice the number of silicons, and
some of the oxygens are bonded to only one silicon. A silicon bonded
to one of these oxygens is at the centre of a tetrahedron which shares
only three corners with other tetrahedra. With each singly bonded
oxygen there is associated one negative charge. The sodium ions are
found in the holes in the three-dimensional silicon—oxygen network.
On the average, each sodium is surrounded by about six oxygens, and
each silicon by four oxygens. In a soda—boric oxide glass of low soda
content the extra oxygen is bonded between two borons, and there
are no singly bonded oxygens. This can happen because in a boric
oxide glass the co-ordination of boron by oxygen is triangular, and in
the mixed glass some of the boron atoms become tetrahedrally co-
ordinated by oxygen. When there is more than about 13-16 per cent
of Na,O in the glass, the boron atoms cease to change their co-
ordination, and some singly bonded oxygens exist (B. E. WARREN,
1942).
Soda-silica glasses are not formed when the soda content exceeds
that given by the formula Na,SiO,. For this formula, if every silicon
atom is surrounded tetrahedrally by four oxygen atoms, then on the
average two oxygens round every silicon are singly bonded, and a
continuous network is just possible. With still more oxygen it is not
possible. In a soda-silica glass with much less soda, there are many
SiO, tetrahedra sharing every corner, and a number sharing only
three corners. The way in which these different tetrahedra are dis-
tributed is not yet quite clear. There may be small regions where all
the tetrahedra share four corners, and such regions are composed of
pure silica; they may alternate with regions of, for example,
Na,O 2SiO,. The composition may vary through the glass when
sufficiently small regions are considered.
The general picture of a glass as a negatively charged irregular
framework containing holes with positive ions in them enables a
distinction to be made between the network-forming ions, which
comprise the framework, and the network-modifying ions which go
8
 GENERAL PHYSICAL PROPERTIES OF GLASS
in the holes. Silicon, boron and phosphorus are important network-
forming ions. Sodium and potassium are important network-
modifying ions. Other ions can act in both capacities. This is
probably true of aluminium, beryllium, zinc, iron and titanium. Ina
lead-silica glass it seems that lead atoms can take part in the network
and link SiO, tetrahedra together. Cobalt ions in network-modifying
positions tend on heating to move into the network, and this can
use a colour change from pink to blue.
&Most commercial glasses are based on silicates or borosilicates. A
pical hard borosilicate glass for chemical work may contain 80 per
cent SiO,, 12 per cent B,O, and 4 per cent Na,O. A soft soda—lime-—
silica glass (usually referred to as soda glass) may contain 70 per
cent SiO,, 17 per cent Na,O and 5-4 per cent CaO. Lead glasses, used
for lamp and valve stems, may contain 30 per cent PbO, 57 per cent
SiO,, 5 per cent Na,O and 7 per cent K,O. These glasses have high
electrical resistance. Glasses with exceptionally high softening tem-
peratures contain 20-25 per cent of Al,O;. Borate glasses, sub-
stantially free from silica (8 per cent SiO.) are used for sodium
discharge lamps. )
General accounts of the structure of glass have been given by J. E.
STANWORTH (1950), B. E. WARREN (1940) and C. J. PHILLIPs (1948).

General Physical Properties of Glass

The physical properties of a given specimen of glass may depend
upon the previous history of the specimen. This is particularly the
case for the mechanical strength under tension, when the surface pre-
treatment of the specimen is of decisive importance. The thermal
expansion and viscosity of glass also depend to some extent on the
history of the specimen. The importance of this factor has been
emphasized by A. E. Date and J. E. STANWORTH (1945). R. vy
Douctas (1945) has given a valuable review of the physical pro-
perties of glass.

Mechanical Strength
The important property for the practical worker is the strength of
glass under tension. The surface of glass very probably contains
numerous extremely small cracks extending into the glass, and when
a tensile stress is applied there is a concentration of stress at the ends
-of these cracks, which causes them to grow further into the glass,
until at some crack breakage occurs and is propagated through the
specimen. Glass usually breaks in a direction at right angles to the
is
direction of maximum tensile stress. A newly-drawn glass fibre
than an old
free from these surface cracks and is much stronger
7
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS
fibre ; but touching a new fibre, even with the fingers, greatly weakens
it. An old fibre is actually strengthened by removing the surface
layer with hydrofluoric acid, even though the cross-section is reduced.
The strength of a glass under tension varies from one specimen to
another. A further complication is the variation of the tensile
strength with the time for which the stress is applied. A tensile stress
which does not cause fracture after a short time of application may
do so after a long time. There is in fact a delayed fracture of glass.

Jime

BO Oh AS 2 lk OP ee ee eee s
Breaking stress kg/mm

Ficure |. Nature of the relation between time of loading

and breaking stress for glass. This property is of a
statistical nature, and the particular curve shown can
only be taken as representative

The nature of the relation between time of loading and breaking

stress for a borosilicate or soda—lime-silica glass is shown in
FIGURE
1. The curve given must be taken as representative only.
A typical
figure for the safe tensile strength for prolonged loading
times is
0-7 kg/mm? (1000 Ib/in.*). Similar results for the relatio
n between
bending stress and time to fracture are found when a
tensile stress is
produced by bending a glass rod into an arc of
a circle. From
FIGURE 1 we note that an increase of stress by a
factor of 4 reduces
the time required for fracture by over 10° times.
If a certain load is
supported for one hour by a certain piece of glass,
one quarter of the
load should be supported for a million hours.
This can be made use
10
 GENERAL PHYSICAL PROPERTIES OF GLASS
of in testing the extent to which stress can be applied to a glass
apparatus. Four times the stress the glass must support can be
applied for a short time.
The delayed fracture of glass, shown in FIGURE 1, must be borne in
mind in assembling apparatus. If, for example, an apparatus is
clamped so that bending stress is introduced, the apparatus may
break after a long interval. Chemical reactions at the surface of the
glass may be partially responsible for delayed fracture: C. GURNEY
and S. PEARSON (1952) found a soda—lime-silica glass to be stronger
in vacuum, and to be weakened by carbon dioxide and water in the
surrounding atmosphere.

wee Expansion
The coefficient of linear thermal expansion is almost constant, for
most types of glass, for temperatures up to 400-600°C. )The actual
value depends on the chemical constitution of the glass. It then
increases rapidly above a certain temperature, often called the

+ B
S 20
Se
& 18
Fe 10
Q 5
us
0 100 200 300 400 500
Temperature °C
Ficure 2. A typical expansion curve for a hard borosilicate glass
(Phoenix)

‘transformation point’. This is not, however, a characteristic tem-

perature, since it depends on the thermal history of the specimen and
the rate of heating. At a higher temperature the glass softens and
ceases to expand. This is sometimes called the ‘softening tempera-
ture’ or the ‘Mg point’. Confusion may result from another defini-
tion of softening temperature, depending upon the rate of extension
of a fibre by viscous flow. This latter softening temperature, which
corresponds to a viscosity of 107*° poises, is very much higher than
the Mg point. A typical linear expansion curve for a borosilicate
glass (Phoenix) is shown in FiGure 2. A is the transformation point
and B the Mg point. The temperature corresponding to A is often
called the ‘lower annealing temperature’, and corresponds to a
viscosity of about 10" poises ;that corresponding to B is often called
the ‘upper annealing temperature’, and corresponds to a viscosity of
11
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

about 10! poises. Another definition in common use, especially in

the U.S.A., is to call the ‘annealing temperature’ that at which the
viscosity is 10!84 poises. This is then between A and B on the thermal
expansion curve. It is useful to bear in mind the confused state of
terminology and definitions when using tabulated data on the
thermal properties of glass. The viscosities mentioned above are,
perhaps, not established with certainty. At the lower annealing
temperature, annealing is actually extremely slow. This temperature
is not used for the practical annealing of laboratory apparatus.

Thermal Endurance
This measures the ability of the glass to stand sudden changes of
temperature without fracture. When a specimen of glass is suddenly
heated uniformly over all its surface, the heat penetrates slowly into
the interior. The outside layers are heated first, and being unable to
expand fully they become subject to a compressive stress, whilst the
inner layers become subject to a tensile stress. When the specimen at
a uniform temperature is suddenly cooled over all its surface, the
surface layers are subject to tensile stress. Since the mechanism of
fracture usually involves surface cracks, glass is more likely to break
on sudden cooling than on sudden heating. The magnitude of the
stress produced on sudden cooling depends on the modulus of
elasticity and the coefficient of linear thermal expansion, and, in a
way not important in practice, on Poisson’s ratio. Thermal endur-
ance is measured by somewhat empirical methods, and is again a
statistical quantity. A heat-resisting glass is one having a high
thermal endurance; a hard glass has a high softening temperature.
A I-mm thick beaker of a hard borosilicate glass, such as Pyrex,
Phoenix or Firmasil, will require a thermal shock, by sudden cooling,
of about 325°C to give appreciable probability of fracture. For a
soda—lime-silica beaker the corresponding figure is about 120°C.
Beakers of Monax glass stand a much greater thermal shock than
the soda—lime-silica beaker; the beakers of standard thickness can
usually survive a thermal shock of 240-250°C. Thick glass fractures
with less thermal shock than thin glass. .
The glass-worker subjects tubing to thermal shock by suddenly
placing itina hot flame. The inner surface ofthe glass tube is then not
heated directly, and is very quickly subjected to tensile stress. The hard
borosilicate glasses as tubes can usually be placed immediately in an
oxy-coal gas flame without fracture, but complex apparatus, espec-
ially when internal seals are present, requires more gentle heating.
Soda—lime-silica glass tubes need gentle warming at first, particularly
when the end of a tube which has not been fire-polished is put in the
J2
 GENERAL PHYSICAL PROPERTIES OF GLASS
flame. The end of a tube is fire-polished by fusing it in a flame, and
this process closes up surface cracks. Vitreous silica has very great
thermal endurance: small red-hot articles can be quenched in water
without cracking.
Thermal Conductivity
For the hard borosilicate glasses and the soda—lime-silica glasses this
is around 0-0025 cal °C-1cm- sec. For vitreous silica (Vitreosil) in
the transparent form it is 0-0025 up to 500°C, and 0-0035 from 500
to 1000°C;; for the translucent form it is 0-0033.
Viscosity and Softening Temperatures
These properties have already been mentioned in connection with
thermal expansion. The viscosity decreases rapidly with increasing
temperature. A linear relation is found between the logarithm of the
viscosity and the reciprocal of the absolute temperature. This is
convenient for extrapolation. When the viscosity has the value 107°
poises the glass is mobile enough to be drawn into threads, and the
temperature is sometimes called the softening temperature (see p. 11).
At temperatures between the lower and upper annealing tempera-
tures (A and B in FiGure 2) the viscosity can change with time—
when the glass is suddenly cooled the viscosity slowly increases to an
equilibrium value and when the glass is heated the viscosity slowly
decreases to an equilibrium value—in fact time is required for the
equilibrium viscosity values to be attained. Glass is often worked
when its viscosity is about 104 poises; for a hard borosilicate glass
this corresponds to a temperature of about 1200°C.
Elastic Properties
Pyrex Chemical Resistance Glass has a Young’s modulus of 6-1 x 10"
dynes/cm?, a modulus of rigidity of 2-510" dynes/cm® and a
Poisson’s ratio of 0:22. Similar values are found for other glasses.
The extension of an amorphous material under a tensile force can
be resolved into three parts; first, an immediate elastic extension,
which is immediately recoverable on removing the tensile force;
secondly, a delayed elastic extension which is recoverable slowly ;and
thirdly, a plastic extension, viscous flow, or creep, which cannot be
recovered. With glass at ordinary temperatures, this plastic exten-
sion is practically absent. A very slow delayed elastic extension
oceurs. This effect can be troublesome in work with torsion fibres.
The delayed elastic effect in vitreous silica fibres is 100 times less than
in other glass fibres, and viscous flow of silica is negligible below
800°C (N. J. TiGHe, 1956). For exact work vitreous silica torsion
fibres are therefore used.
13
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

Thermal Capacity
For Pyrex glass the thermal capacity (specific heat) between 0 and
250°C is given by C (calgm°C) =0:174 + 0-00036t where t is the
temperature in °C.
Electrical Resistance
The resistance of vitreous silica (Vitreosil) in the translucent form at
room temperature exceeds 2x 10'4 ohmcm. Glasses containing
metal ions in network-modifying positions are ionic conductors. In
a soda-lime-silica glass, for example, the current is carried by sodium
ions and the resistance at 150°C may be around 10° ohm cm.
Lemington W.1., a hard borosilicate glass, has a resistance of about
10/° ohm cm at 200°C. A typical lead glass, Wembley L.J., has a very
much greater resistance both at room temperature and normal lamp-
operating temperatures than a soda—lime-silica glass, and is therefore
valuable for lamp and valve pinches. The resistance of L.J. at 150°C
is 104? ohm cm. Generally the volume resistance due to ionic con-
duction decreases rapidly with temperature. The logarithm of the
conductivity is a linear function of the reciprocal of the absolute
temperature. The surface of most glasses is very hydrophilic, and
there is a surface conductivity which depends upon the relative
humidity. For Phoenix glass, for example, the volume resistance of a
centimetre cube at room temperature is about 3 x 10" ohm, but the
surface resistance at 60 per cent relative humidity is 7 x 10" ohm, and
at 81 per cent relative humidity it is 5-4 x 10° ohm. In very humid
atmospheres it is possible to have an electrical shock by touching the
surface of a soda glass apparatus containing electrodes at high
potential. The water layer on the glass becomes slightly alkaline
after a time by reaction with sodium from the glass; the apparatus
should be wiped from time to time with a cotton cloth. It is best in
these cases to use a borosilicate glass. The surface conductivity of
glass was discovered by M. FARADAY (1830).
Density
For the soda-lime-silica glasses this is about 2:5 gm/cm*; for the
borosilicate glasses it is very nearly 2:25 gm/cm3 and hardly changes
with slight variations in composition. Wembley L.1. lead glass has a
density of 3-08. A very dense lead glass has a density of 5-2.
Hardness
Generally glasses with a high silica content are more resistan
t to
abrasion than low silica content glasses. The hardness therefo
re
increases with increase of softening temperature. Lead glasses
can
be scratched quite easily.
14
 GENERAL PHYSICAL PROPERTIES OF GLASS
Transmission of Light
A 1-mm thick sheet of Phoenix glass will transmit 90 per cent or more
of the light incident on it, for wavelengths of 350 millimicrons to
almost 2 microns. In the infra-red region a strong absorption occurs
at 3 microns and little transmission beyond 4 microns. In the ultra-
violet region increasing absorption occurs as the wavelength falls
below 350 millimicrons and very little transmission occurs below 270
millimicrons. The transparent variety of vitreous silica (fused quartz)
has very superior optical properties, and is widely used in photo-
chemical and optical researches. In the ultra-violet region it trans-
mits at high efficiency down to 1850 Angstrom units (185 milli-
microns). ‘Quality O.H. Vitreosil’ of The Thermal Syndicate Ltd is
a special optical quality in which the absorption band at 2400
Angstroms has been eliminated. A special quality of fused quartz is
also available which transmits infra-red up to 3-5 microns approxi-
mately (‘I.R. quality Vitreosil’ of The Thermal Syndicate Ltd). In this
Vitreosil the absorption band at 2:7 microns has been much reduced.
The Stress-Optical Coefficient
It is not usual to take quantitative measurements of the strain in glass
apparatus made for research; when a strain-viewer is used (p. 43)
qualitative observations are normally made. Quantitative measure-
ments can be made when the stress—optical coefficient is known. The
theory involved requires a knowledge of the optical behaviour of
doubly refracting materials and depends on the fact that a ray of
plane-polarized light entering strained glass is broken into two rays—
the ‘ordinary ray’ and the ‘extraordinary ray’—vibrating at right
angles to each other. For glass subject to simple axial tension or
compression, the extraordinary ray vibrates in the plane which
includes the axis of the stress. The birefringence of strained glass is
proportional to the strain, and thus to the stress. The stress—optical
coefficient is the maximum double refraction or birefringence ob-
served in polarized sodium light for 1 cm path length when there is a
uniform stress of 1 kg/cm?. It is expressed either in wavelengths of
sodium light or in millimicrons. This coefficient varies from one
glass to another; it is around 3-5 millimicrons, or 0-006 wavelengths
of sodium light. A. JOHANNSEN (1918) has given an account of
methods for determining double refraction, and very valuable data
for practical work are given by J. H. PARTRIDGE (1949).
General Chemical Properties of Glass
Resistance to Chemical Actions
Vitreous silica is the most chemically inert glass for most purposes.
15
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

It is not affected by halogens or acids{ except for phosphoric and

hydrofluoric acids. Phosphoric acid attacks fused silica at tempera-
tures of 300-400°C, and hydrofluoric acid attacks it at room tempera-
ture,) forming silicon tetrafluoride and water. At high temperatures
silica reacts with caustic alkalis, certain metallic oxides, and some
basic salts, and cannot be used for incinerating these materials. Over
1600°C, fused silica is reduced to silicon by carbon. It can also be
reduced at high temperature by hydrogen. It is unaffected by water
under normal conditions but is attacked by strong solutions of
alkalis.
The hard borosilicate glasses are highly resistant to attack by
water; but just as the sodium ions in the glass are slightly mobile
under the influence of an electric field (p. 14), so also they can be
mobile by thermal agitation and escape from the glass into water in
contact with it and be replaced by hydrogen ions. This effect is
slight: for example, a Firmasil beaker in an autoclave containing
water at 150°C loses about 0-00015 gm of sodium per dm? in four
hours. A soda-—lime-silica glass loses sodium to water at a much
greater rate. The resistance of borosilicate glass to most acids is very
good, but strong aqueous alkalis produce visible attack. The net-
work of triangles and tetrahedra is attacked, so the glass tends to
dissolve as a whole. Soda-lime-silica glass usually has less chemical
resistance than a borosilicate glass. Alkaline attack, however, be-
comes much greater on glasses with high silica content. Alkalis can
also leach out boric oxide from a borosilicate glass. Hydrofluoric
acid dissolves glass, and glacial phosphoric acid attacks most kinds
of glass.

The Weathering of Glass

A reaction between sodium from the glass and atmospheric water and
carbon dioxide can lead to the formation of sodium carbonate, which
crystallizes in fine needles. A potash glass forms potassium car-
bonate, which is too deliquescent to crystallize out. A lead glass can
react with hydrogen sulphide, and to a smaller extent with carbon
dioxide, sulphur dioxide, and acid vapours.

Phenomena Arising from the Heating of Glass

A rapid evolution of adsorbed water first occurs on heating glass;
this is followed by a persistent evolution, due to gas (mostly water)
diffusing from the interior. Above 300°C the two processes are fairly
clearly separated. The adsorbed water is rapidly and completely
removed, and the quantity of gas evolved by the persistent evolution
16
 GENERAL CHEMICAL PROPERTIES OF GLASS
is proportional to the square root of time. The process has an
activation energy. For a soda—lime-silica glass over 98 per cent of
the evolved gas is water. B. J. Topp (1955) has studied these effects
in detail. The adsorbed water on glass can be troublesome in gaseous
manipulation, as R. W. Bunsen first appreciated.
At high temperatures glass loses its more volatile components. The
loss of silica, lime, magnesia and alumina is negligible, but boric
oxide, lead oxide, sodium oxide and potassium oxide can also be lost.
When the glass is heated in a flame, reaction may occur with some of
the flame gases; sulphur dioxide can react with soda glass and lead
glass to form sodium sulphate and lead sulphate respectively, and of
these only the former can be washed off. An account of these effects
is given by W. E. S. TURNER (1945). The loss of weight of vitreous
silica on ignition is negligible ;crucibles can be heated to 1050°C, and
precipitates can be ignited at 1000°C in crucibles with a porous base
of vitreous silica.

Diffusion through Glass

The mobility of the sodium ions in a soda—lime-silica glass at ele-
vated temperatures is fairly high; if an evacuated bulb of such a glass
is dipped into molten sodium nitrate and electrolysis is brought about
by bombarding the inside of the bulb with electrons, the circuit being
completed with an electrode in the sodium nitrate, then metallic
sodium appears in the bulb. By immersing the bulb in other molten
salts the sodium ions can be replaced by ions of silver, copper,
thallium and vanadium. These ions also diffuse into glass from their
molten salts in the absence of an electric field. When potassium is
distilled in a borosilicate glass vessel it becomes slightly contaminated
with sodium which diffuses from the glass and is replaced by potas-
sium (D. K. C. MACDONALD and J. E. STANWoRTH, 1950). Vitreous
silica allows helium, hydrogen, neon, nitrogen, oxygen and argon to
diffuse through it, with the permeability decreasing in the order given.
The permeability of silica becomes greater if the glass devitrifies. The
permeability to helium of soda—lime-silica glass is 10° (or more)
times less than that of vitreous silica. For practical vacuum purposes
soda and borosilicate glasses can be regarded as impermeable to
gases at ordinary temperatures, except in work at extremely low
pressures when the diffusion of atmospheric helium through the glass
may become significant.
The permeability of glass at high temperatures seems to have been
discovered by R. Boyle. In his collected works published in 1744
there is a paper in Volume III ‘A discovery of the previousness of
glass to ponderable parts of flame’ in which he writes *. . . it is plain
Ly
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

that igneous particles were trajected through the glass, which agrees
with the Epicureans...’

Some Surface Properties

The chemical properties of aieare largely determined by its surface,
because the metal ions present in most types of glass are in fact re-
active enough, but only those at the surface can react. The very slow
rates of diffusion in glass at ordinary temperatures limit its reactiv-
ity, so that, as M. FARADAY (1830) wrote ‘Glass may be considered
rather as a solution of different substances one in another, than as a
strong chemical compound; and it owes its power of resisting agents
generally to its perfectly compact state, and ‘the existence of an
insoluble and unchangeable film of silica or highly silicated matter
upon its surface.’ The surface composition of glass may be very
different from the bulk composition, for volatilization occurs during
the forming process, and weathering occurs subsequently; both
processes produce a surface resembling vitreous silica. It is possible
that the SiO, tetrahedra on the surface terminate in OH groups to
which adsorbed water is normally bound. The glass surface can be
made hydrophobic instead of hydrophilic by allowing adsorbed water
and surface hydroxyls to react with various monoalkyldichloro-
silanes (RHSiCI,), when hydrogen chloride is formed by elimination.
The surface properties of glass are of great importance in many
reaction kinetic studies, particularly those involving the termination
of reaction chains on the walls of the vessel. When a glass reaction
vessel is used in such cases, it is usually found that it must be
matured, by carrying out a number of reactions in it, before repro-
ducible results can be obtained. In many cases, reproducible results
are only obtained when the vessel is kept continuously at the reaction
temperature, and exposed only to the reaction mixtures; if the
apparatus is cooled, and air let in to make an alteration or repair,
different results may be obtained subsequently. This is particularly
the case in oxidation reactions.

Devitrification
This is the process of the crystallization of one or more of the con-
stituents of glass. Generally a glass is thermodynamically unstable
with respect to these crystals, but at ordinary temperatures the
crystallization rate is quite negligible. Crystallization may occur
when the glass is worked at high temperature. The crystals which
appear in a supercooled melt are not necessarily those of the stable
solid phase at the temperature concerned :for example, cristobalite can
appear at temperatures for which tridymite is the stable crystalline
18
 DEVITRIFICATION
form of silica. The crystals which are most likely to separate
from soda-lime-silica glass are those of calcium silicate, tridymite
and cristobalite. Calcium silicate occurs usually as the monoclinic
form wollastonite, but sometimes in a hexagonal form. The mono-
clinic form tends to appear as very long, thin crystals bunched to-
gether to give a brush-like appearance. The hexagonal form is not
acicular. Tridymite is hexagonal, and forms hexagonal stars, like
snow, by twinning at 60°. The cristobalite forms twins at 90°. Devitri-
fication on the surface of glass seems to depend upon loss of alkali
oxides, and may be assisted by dust particles.
Transparent vitreous silica (transparent Vitreosil) is liable to de-
vitrify if potassium or sodium compounds are present. The surface
of the material must be thoroughly cleaned, and the part to be
worked should then not be touched with the fingers.
When devitrification is observed in the soda and borosilicate types
of glass, the crystals may sometimes be removed by fusion of the
glass in the flame; alternatively the semi-molten devitrified portion
may be removed with tongs, and replaced by fresh glass added as thin
rod. If there is a large extent of devitrification the portion of glass
should be completely cut out and replaced. Old soda glass apparatus
is very liable to devitrify when repaired: F. C. FRARY, C. S. TAYLOR
and J. D. E>warps (1928) advise, in such cases, a preliminary wash-
ing of the glass with dilute hydrofluoric acid to remove the surface
layer.

Thermal Strain in Glass

When a block of glass is suddenly heated on all its faces, the outer
layers are under compression and the inner layers are under tension,
as explained on p. 12. This strain is temporary, however, since it
vanishes as soon as the temperature gradient vanishes. Temporary
strain is similarly produced when the surface of a block of rigid hot
glass is cooled. Above the upper annealing temperature (p. 11) a
stress can only exist in glass for a short time, because the glass flows
to relieve the stress. Stress is relieved only very slowly at the lower
annealing temperature. Thus over a certain temperature range glass
changes from a viscous to a rigid body. Consider a block of glass to
be rapidly cooled through this temperature range, so that there is
always a temperature gradient. In the viscous region the glass is
strain-free, and thus when it first becomes rigid it is also strain-free.
Thus there is a rigid block of glass containing a temperature gradient
but free from strain. When this temperature gradient is removed, the
inner layers of the glass are in tension and the outer layers in com-
pression. There is then a permanent strain in the glass. Clearly, when
c 19
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

a complex glass object is cooled rapidly there will finally be a complex

distribution of permanent strain, and this can be great enough to
cause fracture of the glass.
The drops of glass produced by Prince Rupert of Bavaria by
dropping molten glass into oil become rigid when there is a great
temperature gradient and the surface is consequently in strong com-
pression. This makes the drops very strong, but they disintegrate
violently as soon as the tail of the drop is cut off, when the internal
stresses are no longer balanced.

Annealing of Glass
The object of this process is to prevent permanent strain arising from
the cooling of glass. The glass must be cooled slowly through the
critical temperature range in which it becomes rigid and ceases to
relieve internal stresses by viscous flow. The rate at which these
stresses are relieved in the annealing range of temperatures (A to Bin
FIGURE 2) depends on temperature; when this is such that the
viscosity is 10!*-4 poises, the glass will become practically stress-free
in 15 minutes (A. E. DALE and J. E. STANWorRTH, 1945). Below the
lower annealing temperature the glass can be cooled quickly without
introduction of permanent strain, but the temporary strain could
become great enough to fracture the article.
Annealing is carried out most satisfactorily in an oven (p. 45).
Complex articles of Pyrex glass can be annealed at 560°C for 30
minutes, followed by slow cooling with the oven door shut. Articles
of Firmasil glass should be annealed at 575°C, but even at 475°C
strain is very slowly removed. For Phoenix, the upper annealing
temperature is 600°C and the lower annealing temperature is 520°C.
Annealing at 560—580°C is therefore satisfactory for this glass. The
annealing temperatures of these borosilicate glasses are not at all
critical. The article must not, of course, be madé too hot, or it will
deform. Annealing is of great importance for articles made of a
soda—lime-silica glass. Wembley X.8. soda glass should be annealed
in the range 520-400°C, and the General Electric Company, which
makes this glass, recommends annealing at a high temperature of
520°C for 5-10 minutes, followed by cooling to an intermediate
temperature of 460°C at a rate dependent upon the glass tubing
thickness. These rates are :—
3°C per minute for }mm wall thickness
2°C per minute for 1 mm wall thickness
1°C per minute for 3 mm wall thickness.
The glass should be cooled from the intermediate temperature of
460°C to a low temperature of 400°C at double the above rates. The
20
 ANNEALING OF GLASS
article can then be cooled to room temperature at any rate possible
without cracking it by temporary thermal strain. The corresponding
high, intermediate and low temperatures for Wembley L.1. lead glass
are 430°C, 390°C, and 340°C. The same annealing schedule can be
used. For Wembley M.6. ‘white neutral’ glass the temperature range
is 580-450°C. Again the same schedule should be used.
Complex apparatus assembled on the bench must be annealed by
flame, and this method must also be used when no oven is available
(see p. 170). In our experience, very complex apparatus of Pyrex glass
can be flame-annealed satisfactorily. Usually with Pyrex the appar-
atus either cracks in a day or two after making, or else not at all.
Complex apparatus of soda-—lime-silica glass can be annealed by
flame, but we do not find this satisfactory. For research apparatus it
is best to avoid this kind of glass. With a complex vacuum apparatus
of borosilicate glass a fracture can often lead to unfortunate conse-
quences, especially when there are many mercury cut-offs present;
and in such cases it is well, before evacuating, to wait for a few days
after a repair or alteration has been made in a position where flame
annealing is difficult.
Some Types of Glass for General Use
Many different kinds of glass are made. In this Section and the
following Sections we mention only a few of these which are useful
in the laboratory.
Wembley X.8.
This is a soda—lime-silica glass, containing magnesia and boric oxide
(less than 1 per cent), made by the General Electric Co. It is often
described as GEC X.8. or simply as X.8. The linear coefficient of
thermal expansion between 20 and 350°C is 9:65 + 0-10 x 10-®. This
glass is available as tubing and rod in a wide range of sizes.
BTH No. 94
Made by the British Thomson-Houston Co., this is a soda—lime-
silica glass. The linear expansion coefficient is 9-5 x 10~® (50-400°C).
The glass is available as machine-drawn tubing for bench working
etc.
Monax
This is a borosilicate glass, free from arsenic, antimony and lead,
made by John Moncrieff Ltd. The coefficient of linear expansion is
4-4 x 10-8/°C. The glass is available as tubing and rod of various
sizes; in addition many standard pieces of laboratory glass-ware
constructed in Monax are available.
21
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS
Pyrex
Pyrex is a borosilicate glass free from alkaline earth metals, zinc and
heavy metals, made by James A. Jobling & Co., and in the U.S.A.
at the Corning Glass Works, where Pyrex was developed. The linear
expansion coefficient (20-400°C) is 3:2x10-® Pyrex tubing is
supplied in a wide range of sizes, with standard wall, heavy wall, or
as extra heavy tubing. Capillary tubing and rod is supplied in a range
of sizes. Precision-bore tubing with very accurate inside dimensions
is also made. A very wide range of laboratory apparatus is manu-
factured in Pyrex glass.

Phoenix
Phoenix, also, is a borosilicate glass made by the British Heat
Resisting Glass Co. The linear expansion coefficient (50-450°C) is
3:24 x 10-®. Tubing and rod are supplied in a wide range of sizes.
The tubing is made with either standard wall or heavy wall.

Firmasil
Laboratory glassware made from this glass is available. It is a
borosilicate glass manufactured by Wood Bros. Glass Co. The
linear expansion coefficient (0Q-400°C) is 3-2 x 10-8,

Some Types of Glass for Sealing to Metals

Again we mention only a selection from the many sealing glasses
which are made.

Lemington W.1.
This is another borosilicate glass, and was developed to make strain-
free seals with tungsten. It is made by the General Electric Co., and
often described as GEC W.1. or simply as W.1. The linear expansion
coefficient (20-350°C) is 3-75 x 10-8,

BTH No. 9
A borosilicate glass for sealing to tungsten, and is made by the British
Thomson-Houston Co. It is often described as BTH C.9 or as oe
The linear expansion coefficient (50-400°C) is 3-65 x 10-8,

Dial 36
Made by Plowden & Thompson Ltd, Dial 36 is a borosilicate
glass
for sealing to tungsten. The linear expansion coefficient (20-300°C)
is 3°6 x 10-°,
22
 SOME TYPES OF GLASS FOR SEALING TO METALS
Corning No. 7720 (Nonex)
Corning No. 7720, known commercially as Nonex, is a borosilicate
glass containing lead and made at the Corning Glass Works. It is
used for sealing to tungsten metal. The coefficient of linear expan-
sion of Nonex is 3-6x10-§/°C. This is relatively low, although
slightly higher than that of Pyrex brand glass (Corning No. 7740),
and matches sufficiently with that of tungsten, 4-5 x 10-§/°C, to cause
the glass to make satisfactory seals to this metal.
Nonex begins to soften, and can be worked, at temperatures a
little below those required for Pyrex. Care must be taken when
working it to avoid reducing the lead and causing the glass to blacken.
The tip of an oxidizing flame must therefore be used. Nonex will seal
well to Pyrex, though sometimes uranium glass is used between the
two.

BTH No. 40
A borosilicate glass for sealing to alloys of the Kovar type (p. 111);
it also seals to some electrical porcelains. It is made by the British
Thomson-Houston Co., and is often referred to as BTH C.40 or as
C.40. The linear expansion coefficient is 4-85x10-® from 50 to
400°C.
Lemington FCN
Made by the General Electric Co. and often called GEC FCN or
FCN, this is a borosilicate glass for sealing to Kovar-type alloys. The
linear expansion coefficient (20-350°C) is 4-75 x 10~°.

Kodial
A Koyar-sealing borosilicate glass made by Plowden & Thompson
Ltd, this has a linear expansion coefficient (20-300°C) of 4-9 x 10~°.

Dial 444
This is a soda glass for sealing to platinum, and is made by Plowden
& Thompson Ltd. The linear expansion coefficient (20-300°C) is
8-8 x 10-®.

Some Types of Glass for Special Purposes

Here we mention only a few of the glasses which are made for special
purposes.

Wembley L.1. Lead

Made by the General Electric Co., this glass was developed primarily
for pinches and exhaust tubes in lamps and valves. It seals directly
23
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

to platinum and to copper-clad wire (p. 111). The coefficient of linear

expansion is 9-05 x 10-® from 20 to 320°C. A lead glass, such as L./.,
is useful as an intermediate glass for joining soda glasses of different
thermal expansions. ;
BTH No. 12
BTH No. 12 is a lead glass for lamp and valve manufacture, made by
the British Thomson-Houston Co. The linear expansion coefficient
(50—400°C) is 9-1 x 10-®.
Wembley M.6, ‘White Neutral’
This is made by the General Electric Co., and is used for medical
ampoules. It loses very little alkali to aqueous solutions. The linear
expansion coefticient is 7-3 x 10-® from 20 to 350°C. This is inter-
mediate between the soft and hard glasses, and M.6. is sometimes
used in graded seals.
Wembley ‘Amber Neutral’
This glass, made by the General Electric Co., is similar to white
neutral above, but is coloured amber by addition of iron and man-
ganese. The linear expansion coefficient is 7-5 x 10-® from 20 to
350°C. It is useful for ampoules when the contents need protection
from ultra-violet radiation.
Lemington H.26X.
Lemington H.26X., made by the General Electric Co., is a very hard
borosilicate glass of high softening temperature. The Mg point is
780°C. It is used in high pressure mercury vapour lamps. The linear
expansion coefficient is 4-6 x 10-® from 20 to 580°C. Sodium and
potassium are absent and alumina is present in quantity in this glass.
Sodium Resistant, NA.10
This type, made by the General Electric Co., is resistant to sodium
vapour and is used in sodium vapour discharge lamps. It has a very
high boric oxide content, a low softening temperature, and a low
electrical resistance. It is used as an internal layer in soda glass
tubing (X.8.).
Dial 43
A borosilicate glass with an expansion coefficient intermediate be-
tween that of the tungsten-sealing and Kovar-sealing glasses. It is a
useful sandwich glass for joining between these glasses. The linear
expansion coefficient (20-300°C) is 4-2x10-® It is made by
Plowden & Thompson Ltd.
24
 SOME TYPES OF GLASS FOR SPECIAL PURPOSES

Fused Silica
The Thermal Syndicate Ltd make fused silica (Vitreosil) tubes of
several kinds. Translucent and transparent Vitreosil are manu-
factured ; the former is supplied with ‘sand’, ‘satin’ or ‘glazed’ sur-
face, and the latter usually has a glazed surface. A glazed surface
should be used in vacuum work; the transparent tubing is best.
Vitreosil tubing is available with a wide range of sizes and with
several wall thicknesses. Rods, bars and capillary tubes are available.
Many items of laboratory equipment are manufactured in Vitreosil.
They are specially valuable when high temperatures and high thermal
endurance are needed. The linear expansion coefficient is 0-54 x 10-8,

Graded Seal Glasses

The General Electric Co. makes a range of sealing glasses. For
joining silica to hard glass, and hard glass to soft glass, the sealing
glasses given below can be used. The coefficients of linear expansion
are also given.

li glass type
Sealing Expansion
een Purpose

WOQ.31 1:0 x 10-*

WO.34 9-1 x 10-8 for silica
H.428 3-2 x 10-8 to hard glass

G.S.1 2 LOS
G.S.2 Ye ere
G.S.3 6:6 x 10~° for hard to
G.S.4 1 2elOe soft glass
G.S.5 7:8 x 10-§
G.S.6 8-4 x 10-8

The G.S.1 can be joined to a hard borosilicate glass using as inter-

mediates a Kovar-sealing glass, a sandwich glass such as Dial 43,
and a tungsten-sealing glass.

The Ease of Working Different Types of Glass

The working of fused silica is described in Chapter 10. Pyrex,
Phoenix and Firmasil require either a gas—oxygen flame or a gas—air—
oxygen flame. All the operations we describe for Pyrex can also be
carried out with Phoenix. These two glasses are very easy to work,
and very suitable for research work. They join together well. Monax
can be worked without oxygen, and is a most valuable glass for
jis
 PHYSICAL AND CHEMICAL PROPERTIES OF GLASS

laboratories, such as those in schools, where oxygen is not available.

The methods we describe for Pyrex can be used for Monax. When a
hand torch is used for working Monax, it should give a flame with a
definite inner blue cone, and it may be necessary to work the Monax
with the hand torch in small portions at a time, especially if the air
pressure is low and the tubing is large. Finished articles in Monax,
even when badly made, do not usually crack after a flame annealing.
Soda glass is very easy to work in a gas—air flame, but skill is neces-
sary to make apparatus which does not crack on cooling. Complex
apparatus of soda glass cannot be made satisfactorily with a hand
torch by many of the methods we describe for Pyrex ;the bench torch
methods are far better for soda glass, but they are far more difficult.
We do not recommend soda glass for research apparatus unless this
is of very simpie construction.
The working characteristics of the other glasses mentioned can
usually be deduced from their expansion coefficients. A glass of
lower expansion than Monax needs an oxygenated flame, and one
with higher expansion needs only a gas—air flame.

Distinguishing between Different Types of Glass

In a bunsen flame, soda glasses give a yellow flame and are softened,
borosilicate glasses keep their hard edges, and a lead glass is black-
ened. When a joint is made between two glasses of different com-
position and pulled out, the softer glass will pull out further. The
softer glass has the lower softening temperature.

REFERENCES
BoYLe, R., 1744, The Works of the Honourable Robert Boyle ; London,
Millar.
Date, A. E. and STANWoRTH, J. E., 1945, J. Soc. Glass Tech. aes Narae
DouGLas, R. W., 1945, J. sci. Instrian. PH
FARADAY, M., 1830, Phil. Trans., 120, ‘
FRARY, F. C., TAYLOR, CES; and Epwarps, J. D., 1928, Laboratory
Glass Blowing: 2nd Edn; New York, McGraw-Hill.
GARNER, W. E., 1952, J. chen Soc., 1961.
GURNEY, Cc. and PEARSON, S., 1952, Report No. 10. Selected
Govern-
ment Research Reports, Vol. 10, Ceramics and Glass; London
,
H.M. Stationery Office.
JOHANNSEN, A., 1918, Manual of Petrographic Metho
ds, 2nd Edn;
New York, McGraw- Hill.
MAcDOanak
oNnaLp,
Gn D. K. C. and STANWORTH, J. E., 1 950, Proc. ) Soc.
phys.

26
 REFERENCES
PARTRIDGE, J. H., 1949, Glass-To-Metal Seals; Sheffield, Society of
Glass Technology.
PHILLIPS, C. J., 1948, Glass: The Miracle Maker, 2nd Edn; London,
Pitman.
RANDALL, J. T., 1938, Annual Reports on the Progress of Chemistry
for 1937; London, Chemical Society.
STANWORTH, J. E., 1950, Physical Properties of Glass; Oxford,
Clarendon Press.
TiGHE, N. J., 1956, National Bureau of Standards Circular 569;
Washington, U.S. Government Printing Office.
Topp,B. J., 1955, J. appl. Phys., 26, 1238.
TURNER, W. E. S., 1945, The Elements of Glass Technology for
Scientific Glass Blowers (Lampworkers), 3rd Edn; Sheffield, The
Glass Delegacy of the University.
WARREN, B. E., 1940, Chem. Rey., 26, 237.
WARREN, B. E., 1942, J. appl. Phys., 13, 602.
ZACHARIASEN, W. H., 1932, J. Amer. chem. Soc., 54, 3841.

zi
 Chapter 3

GLASS-WORKING EQUIPMENT

The Glass-Blower’s Table

THE usual practice is to have a special table for the bench blowpipe.
This laboratory blowpipe table can be 2-3 feet square for most pur-
poses; but for the construction of large apparatus a good size is
6 feet long and 3 feet wide. This size will permit the use of rollers for
large tubes. The table should be placed so that long pieces of glass
can extend beyond the ends. The top of the table should be of heat-
resisting material. The blowpipe is placed near one edge, which
should be flat. The other three edges should have a raised rim to
prevent articles rolling off the table. Gas, oxygen and compressed air
must be available ; the latter is often obtained from bellows fixed under
the table. Drawers for tools are desirable. A rack of some sort should
be available in which hot glass objects can cool; a large block of wood
bored with holes of different sizes to hold the ends of the objects is
very useful for this purpose. A small flame should be present on
the table for relighting the blowpipe if it is extinguished during work.
The table should be placed so that direct sunlight does not fall on
the blowpipe flame. If this can happen, there should be a blind to cut
out the sunlight. It is difficult to work glass in a flame in sunlight
because the flame is then so difficult to see and the temperature of the
glass cannot be judged properly.
The height of the table and the accompanying stool should be such
that the glass-worker can rest his elbows on the table. A stool of
adjustable height (like a music stool) is very convenient, and was
advised by R. THRELFALL (1898). Both he and W. E. BARR and VY, J.
ANHORN (1949) advise a table of height 38 inches. A table of slightly
less height is also satisfactory.

Blowpipes
The term ‘blowpipe’ is used for both the flame generator employed in
glass-working and the iron tube used in blowing glass (p. 1). In the
glass trade the term ‘glass-blower’ describes a person who uses glass
melted in a pot, and the worker of tube or rod in a flame is a ‘lamp-
worker’, who uses a blowlamp; but in the laboratory this usage is not
28
 BLOWPIPES
customary. There a ‘glass-blower’ is a person who works tubing ina
blowpipe flame.
Satisfactory blowpipes or bench burners are available from labora-
tory furnishers. They usually have a range of jets to give different
flame sizes. The changing of jets during work is avoided in a burner
with a turret head. The burners normally burn coal gas with either
compressed air or oxygen, or air-oxygen mixtures. Crossfire burners
produce a number of small flames which heat both sides of a tube at
once, as shown in Ficure 3. They are very useful for large tubing.
Special burners producing a long thin flame (‘ribbon burners’) can be

FiGurRE 3. Two crossfire burners

used for glass tube bending; they are made giving lengths of flame of
50, 100, 200 and 300 mm. A cracking-off burner giving a thin line of
intense heat is also made. All these special burners can be obtained
from Stone-Chance Limited.
I. C. P. SmitrH (1947a) has given an account of the construction of
burners for the glass-worker. The adjustment of the flame size and
temperature with a given burner is best found by experience.

Hand Torches
A hand torch is a moderately light blowpipe designed to be held in
the hand and moved around the apparatus. Various hand torches
are available. We find the Flamemaster hand torch, made by Stone-
Chance Limited, to be very useful. This can be fitted with a number
of different jets. One gives a wide range of flame sizes with air—coal
gas mixtures. Three jets give different flame sizes of oxygen—coal gas
flames. With the largest of these, 4-cm diameter Pyrex tubing can be
worked with a single hand torch; with two hand torches as a cross-
fire, 6-cm bore tubing can be worked. A double-tipping device can
also be put on the torch in place of the usual nozzle; this gives two
small flames at an angle to each other of about 120°. This is useful
in ampoule sealing. Controls are provided on the hand torch for
Fi
 GLASS-WORKING EQUIPMENT
regulating the flow of gas and air or oxygen. The Flamemaster can
also be clamped to the bench and used as a bench torch.

Cylinder Heads and Valves

The working of glass such as Pyrex and Phoenix requires either an
oxygen-coal gas flame, or an oxygen—air—coal gas flame. The oxygen
is obtained from cylinders. For glass-working it is desirable to use a
cylinder head which has both a regulator for controlling the pressure
of the oxygen and keeping it at a steady value, and a fine adjustment
valve for regulating and turning off the gas flow. In operating a
blowpipe or hand torch, the oxygen is frequently turned off at the
blowpipe or hand torch. The regulator then automatically stops the
flow of oxygen from the cylinder. If only a valve is used the oxygen
continues to come out of the cylinder, and the pressure in the rubber
tubing to the blowpipe becomes so great that the tubing is forced off
the blowpipe or the cylinder head ; this can be very inconvenient when
one is working in an awkward position with a hand torch on a
complex apparatus.
Cylinder heads are liable to be damaged when they are attached to
cylinders left free to roll about on the floor, and to avoid this a
cylinder stand should be used.

Bellows
Foot-operated bellows fitted with a rubber disc enclosed in a string
net, to provide an air reservoir, give a steady pressure of about
20 oz/in.? They are suitable for air-gas flames of moderate size.
Bellows with a spring-controlled reservoir can also be used, but they
do not give such a steady pressure.

Air Blowers
A compressor driven by a } h.p. electric motor and giving a pressure
above atmospheric of 10 Ib/in.? is adequate for all normal glass-
working operations, including those with a small lathe (Chapter 6).
A filter should be fitted on the suction side of the compressor.
A filter pump discharging into an aspirator, fitted with an outlet
at the top for air and an outlet at the bottom for water, will give
enough air for small-scale operations of glass-working. W. A.
SHENSTONE (1889) described this device.

Glass-Working Tools
The tools necessary for most glass-working operations are simple
and
can best be described in relation to their uses. Probably the most
useful general-purpose tool is an iron or steel spike about 14 inches
30
 GLASS-WORKING TOOLS
long set in a wooden handle (FIGURE 4). This spike can be used
for
manipulating the hot glass into the required place and in
many
instances can also be used in the same operations as the
more
specialized tools described below, although it may not be so con-
venient. Throughout the book this tool will be referred to as a spike.
Tools designed for various operations are marketed by a number of
companies dealing in laboratory supplies. FIGURE 4 shows a selection
of such tools. Triangular flaring tools are intended for flaring out the

FIGuRE 4. Some tools used in the working of glass

ends of tubes. They are usually made of brass sheet or of thin carbon
plates, as are the other tools for shaping glass by means of pressure
of the flat surface of the tool applied to the glass as it is rotated. The
metal tools should be lubricated to prevent the glass sticking to the
tool. Carbon tools require no such lubrication but achieve the same
effect by wearing away in use. Hexagonal tapered reamers are used
for working tubes to standard tapers for the purpose of making
stopcocks or ground joints. These reamers may either be made of
carbon or of metal—usually aluminium alloy. Similar tools with a
much sharper taper and mounted on a handle may be found more
convenient than the normal flaring tool for opening out the ends of
tubes. We have found that the mounted needles from a set of dis-
secting instruments are useful in some cases with capillary tubing
since they are fine enough to be inserted into the bore of the tube.
31
 GLASS-WORKING EQUIPMENT
These needles have the disadvantage that, owing to their fineness,
they tend to oxidize away readily and may also be melted if held in a
hot flame for any appreciable length of time.
Forceps will be found useful in many circumstances. These are
usually made of steel and should be designed so that there is no
tendency for the soft glass to slip out easily. This will be avoided if
the points are bent inwards so that they meet at an angle of about
25°. The forceps should either be long enough not to get too hot in use
(about 10 inches long) or else they should be insulated in some way.
Flask clamps are made to fit a wide range of sizes of flask. These
are devices for holding flasks while the necks are being worked and
enable the whole flask to be rotated easily. They usually consist of a
handle with three or four sprung prongs which hold the body of the
flask. Some clamps are made so that they can be adjusted to fit a
range of flask sizes. °
Many operations require only a very modest set of tools, and too
great a stress can be laid on the variety of tools required for any
operation. As an example, forceps are useful for drawing off small
pieces of glass or for pulling out excessively thick parts of the wall of
a joint, but both these operations can be carried out by fusing a
length of glass rod to the part to be drawn off and using this as a
handle for the drawing operation. As in many other fields of crafts-
manship, a skilful operator can work wonders with inadequate and
makeshift tools, but his job would be much easier if he had all the
equipment he needed, although for much of the time he may only use
a very few of the tools at his disposal. There are times when the less
commonly used tools will be useful, but on the other hand the glass-
worker should not allow himself to be deterred by the fact that he -
does not have a vast array of tools.
J. C.’P. SmitH (1947b) has reviewed some uses of tools in lamp-
working.

Wax for Tools

(Tools can be used without lubrication, but care is necessary in
this
case to prevent the tool oxidizing and the oxide entering the glass,
which is then liable to crack on cooling. A lubricated tool
slides
more easily over the hot glass. Beeswax and paraffin wax have
been
used as lubricants} a thin film only is used. (A better
lubricant,
ere I.C. P. Smith (1947b){ is paraffin wax mixed with some
loidal raphite. The colloidal graphite is not rapidly burnt
col-
off the tool
and it retains its lubricating properties up to 600°C.
An over-heated
tool can also be dipped into distilled water (5 0z) containing
colloidal
graphite (4 oz of Aquadag). The tool picks up a
layer of graphite.
32
 WAX FOR TOOLS
When soft asbestos is used on the bench, waxed tools are likely to
pick up shreds of asbestos which can cause trouble. It is best to avoid
soft asbestos in this case.
When constructing high vacuum apparatus, waxed tools should
only be used when the apparatus under construction can be
thoroughly cleaned afterwards. Otherwise wax may be present in
the vacuum system, and will be removed by pumping only with
extreme oS)
Carbons
Carbon Plates
A few thin, rectangular plates of carbon can be very useful on the
glass-working bench. The faces of these plates are of use in flatten-
ing the ends of tubing, in blowing flat-bottomed bulbs, and in shaping

@®D

Ficure 5. A carbon plate and a

carbon rod. Both are attached to
a handle

pieces of molten glass as, for example, in making the paddle for a
glass stirrer. For shaping purposes it is sometimes preferable to have
a handle attached (FiGURE 5). Plates ranging from } to } inch in
thickness and of the order 3 by 6 inches are convenient sizes.
Carbon Rods
Carbon rods with one end filed to a conical shape make very good
33
 GLASS-WORKING EQUIPMENT
ng from
flaring tools. It is useful to have a few of these handy, rangi
1 to } inch diameter. Glass handles for these rods can easily be made
filed near
by shrinking the end of a piece of glass tubing into a groove
to one end of the carbon rod (FiGuRE 5). When working silica,
carbon rods become a necessity as flaring tools, etc., because no metal
will withstand the working temperature of this glass for long enough
for much tooling to be done.
Gauges for Measuring
Some apparatus must be made to a definite size. The external dia-
meter can be measured by an ordinary calliper gauge (FIGURE 6, J).
The internal diameter can be measured either by the prongs at the

Ficure 6. Gauges useful in measuring glass, and for working glass

to fixed dimensions

back of the calliper gauge, or, for greater distances inside a tube, by
an internal pair of callipers (FIGURE 6, //); the separation of the feet
of this gauge is measured after it is withdrawn from the tube. The
internal diameter at the end of a tube can be measured accurately
with a cone or taper gauge (FIGURE 6, JV), which also reveals devia-
tions from a circular cross-section. A pair of callipers of the type
shown in Ficure 6, J/I, is also useful in measuring the outside of a
tube. The gauges in FiGure 6, J, JJ and J// can be used on hot glass;
in working glass to fixed dimensions it is convenient to set a gauge at
the required dimension and use it to measure the glass when it is hot.

Rubber Caps
A very convenient way of closing the ends of glass tubes for blowing,
particularly with very small bore tubing into which it is difficult to fit
34
 RUBBER CAPS
a cork or stopper, is to employ rubber caps. These are made in a
variety of sizes to fit different diameter tubing, ranging up to about
30 mm, and when of good quality rubber can easily be slipped over
the open end. They create a compression rather than a tensile strain
in the glass tube and are therefore useful for closing tubing with
ragged ends when it is desirable to avoid fire polishing. These caps
are usually supplied made from rubber approximately 1 mm thick.

Thin Rubber Sheet

Pieces of very thin sheet rubber, of the order of 0-1 mm thick and of
very elastic quality, as, for example, pieces cut from toy rubber
balloons, can be very useful in closing ragged ends of large tubing.
They are of particular use with tubing larger than 20 mm diameter.
A small piece, of about the same diameter as the glass tube itself, can
be stretched over a ragged end of the tube and will be found readily
to stay in place by means of its own tension—provided excessive
pressure is not built up when ‘glass blowing. For detailed use of
rubber caps and pieces of sheet rubber see Chapter 9.

Rubber Stoppers
Rubber stoppers are used extensively for making a closed system for
blowing out. A complete overlapping range is a necessity to the glass-
worker. The rubber should be of good elastic quality since poor
quality stoppers often will not ‘give’ sufficiently to be squeezed easily
into the ends of glass tubing. The stress on the glass is less and an air-
tight seal is more easily obtained when good rubber is employed.
Two sets of stoppers are very useful, one set unbored and a second
set with a single hole bored in them to carry blowing-tube connectors.
A longer piece of glass tubing inserted into the hole will also form a
very adequate handle when working small pieces of apparatus.
An assortment of small corks for closing small bore tubing can, in
addition, be very useful. If the flame is to come near to a cork or a
rubber stopper it is best to wrap either in thin asbestos paper before
inserting into the tube. When pyrolysis of the cork or rubber
stopper seems probable it is best to use a cork and not a rubber
stopper, because the pyrolysis products from a cork can be cleaned
off more easily than those from rubber. Sometimes a plug made from
damp asbestos paper is adequate for closing a tube of small bore, and
this plug can be heated quite strongly.
Blowing Tubes
Rubber tubing is necessary for blowing when the piece being worked
cannot be brought to the mouth. A length of about 80 cm is convenient
D 35
 GLASS-WORKING EQUIPMENT

and the size most useful is that of normal condenser tubing, or

approximately 5 mm bore with 1 mm wall thickness. Lightweight
rubber is advisable since weighty tubing tends to drag. It is occa-
sionally helpful to have a finer blowing tube, and for this a length of
bicycle valve tubing, about 2 mm bore, is quite satisfactory.

Mouthpieces
Mouthpieces for these blowing tubes vary with individual inclination.
Some workers dispense with a mouthpiece and use the bare end of the
rubber tubing. By gripping this in the teeth the flow of air can be
controlled. Others desire a firmer mouth piece and use a short piece
of glass tubing which must, of course, be fire polished at the end. It is
perhaps also safer to strengthen the tube by thickening the end. A
short piece of rubber tube, about 2 cm long, attached to the glass
mouthpiece, is another modification desired by some. It combines
the features of a firm mouthpiece with one where the air flow can be
controlled with the teeth. A glass tube inside a rubber tube can also
be used as a mouthpiece.
An assortment of glass connectors for joining blowing tubes and
for connecting blowing tubes to rubber stoppers is desirable. For
blocking holes in stoppers short pieces of rod or closed tubing are
useful.
A holder for a blowing tube can be made out of wire. This fits
round the operator’s neck, and holds the mouthpiece near the mouth.
It is useful when engaged in complex operations which require the
use of both hands for a long time.

The Uses of Asbestos

Asbestos in one form or another finds many uses in laboratory glass-
working operations. Asbestos paper or tape is very useful for block-
ing the ends of tubes which are so placed that they get too hot to
allow the use of corks or rubber stoppers. A satisfactory blocker can
be made by winding asbestos tape round a length of glass rod until it
fits the tube. It may be found necessary to moisten the asbestos to
make it adhere to itself. Wound round a length of tubing in a
similar way, asbestos tape provides a convenient method of inserting
a blowing tube into a system. Asbestos paper may be pulped with
water to form a pulp which can be used for blocking irregularly
shaped holes while glass-blowing is carried out on some other part of
the system; but if it is used in this way care must be taken to prevent
shreds of asbestos falling inside the apparatus, and one must
not
leave any shreds adhering to the glass when it is subsequently worked.
When it is necessary to work glass under cramped conditions
and
36
 THE USES OF ASBESTOS
there is a likelihood of the flame being played on to other parts of the
apparatus, it is advisable to screen these parts with sheets of asbestos
board or by binding them with asbestos cord or paper. The bench is
also liable to suffer from the effects of hot glass and hot tools being
laid down unless it is protected by a sheet of asbestos board or similar
material. Hard asbestos board is better for this purpose than the soft
type as the latter has a tendency to contaminate the tools with loose
fibres which may be transferred to the glass being worked.
Asbestos gloves protect the hands both from glassware which
would otherwise be too hot to hold and from the effects of radiant
energy. But they suffer from the disadvantage that, owing to the
nature of the material, they tend to be clumsy and do not allow small
objects to be picked up easily. Unless large sizes of tubing are to be
worked there seems to be little advantage in the use of these gloves.
But with operations with a hand torch on large Pyrex tubing,
asbestos gloves are often essential—after a time the torch becomes
too hot to hold, because it is heated by radiation and convection from
the hot glass. The gauntlet type of asbestos glove is desirable for
these operations.
When new asbestos paper and tape is used on apparatus, it may be
desirable first to heat it in a gas—air flame to remove the binder.

Glass-Blowing Spectacles
It may be found that in operations in which it is necessary to observe
the glass deforming or flowing whilst it is in the flame, vision is
obscured by the sodium glare from the glass. Special glasses are
made to filter off this yellow light and also much of the infra-red
radiation. Spectacles or goggles made from such glasses are available
commercially. The glass itself resembles the cobalt glass used ‘in
qualitative analysis flame tests, and may also contain rare earth
oxides. Didymium goggles contain neodymium and praseodymium,
and selectively absorb light in the sodium D region. The effect of the
radiant heat from the glass is tiring to the eyes and may be reduced
considerably by wearing ordinary sun glasses made with Crookes
glass. These glasses reduce the light and heat intensity but do not
significantly cut out the sodium glare. If ordinary sun glasses are
used, care must be taken to see that the frames are not easily in-
flammable in view of the possibility of accidents.

Glass Knives and Diamonds

Glass knives are of two main kinds. The inexpensive ones are made
of hardened carbon steel, and the more expensive ones are made of a
very hard alloy. The latter kind keep their edge for a long time, but
37
 GLASS-WORKING EQUIPMENT
a
the former soon lose their edge. Both kinds can be sharpened on
type a fine hone is desirabl e ;
carborundum hone. For the expensive
for the other kind a coarse hone may be used and a scythe stone 1s
satisfactory.
A glazier’s diamand is very satisfactory for cutting sheet glass. It
is not satisfactory for cutting the outside of glass tubing; the inside of
glass tubing can be cut with a diamond by the device described below.
A triangular file can often be used instead of a glass knife.

Tube Cutting Device

A very useful device for cutting tubing too large to break by hand is
shown in FIGURE 7.* It can very easily be made in the machine shop.

Ficure 7. Tube cutting device

The bush 4 is hinged at B and carries a rod C whose position with

respect to B can be adjusted. At one end of Cis set a tiny diamond D.
The tubing to be cut is placed over C so that the diamond is in con-
tact with the inside of the glass wall and the tube is lowered on to the
rollers E. It is then rotated through 360° while upward pressure is
exerted at F causing the diamond to scratch the glass. A very clean
break is obtained either right away, or on local heating of the glass
(p. 50). The rollers are reduced in diameter over a short length at G
to take a tube whose end is flared, as, for example, a- test tube.
Notches in the end-plates H and J, in which the ends of the rollers sit
as shown, are cut at equal intervals on either side of the central point,
* We are grateful to MrA. J. Hawkins, glass-blower at the University of British
Columbia, whose design we have shown here.
38
 GLASS KNIVES AND DIAMONDS
sO that the distance between the rollers can be varied according to the
size of the tubing to be cut.

Clamps
Good clamps for holding glass apparatus in position are essential to
the scientist. Also, in many of the operations described in this book
the glass-working scientist will need to hold the glass tubes in
position with clamps.
The best form of clamp has both arms of the claw independently
movable with a screw to tighten up each (FiGurE 8, J). Many stan-
dard forms have only one tightening screw, and one movable arm is

I
FIGuRE 8. Two types of clamp

drawn towards the other whose position remains fixed (FIGURE 8, //).
These are less convenient. Two prongs in each arm of the claw are
desirable and lead to more stability than a two-and-one prong
arrangement. A claw lining of asbestos or cork is preferable to
rubber, and a machined boss makes for ease of manipulation and
uniformity of pressure.

Abrasives
Glass may be ground to shape by using one of five abrasives. These
are (/) grinding sand, (2) emery, (3) carborundum, (4) pure alumina
or corundum, and (5) diamond powder. Each abrasive has its own
particular property. Grinding sand is used for very rough grinding.
It is mixed with water and the glass is ground to shape against a
suitable surface. Grinding sand is very coarse and unsuitable for
refined work, for which emery, carborundum or corundum are used.
Emery is graded from coarsest to finest as 1, 0, 00, 000, 0000. The
a
 GLASS-WORKING EQUIPMENT

grading is done by shaking the emery with water and leaving the
suspension to stand. After a specified time the liquid suspension is
removed and dried, so the longer the standing time the finer the
emery. Emery can also be graded by particle size, by giving the size
of a mesh which will just retain the particles. A 100-mesh sieve, for
example, has 100 square holes per linear inch. (Often in a sieve only
about 25 per cent of the area consists of holes; the remainder is the
wire network. The wire size is equal to the aperture size when 25 per
cent of the sieve area is holes.) When grinding glass with increasingly
fine grades of emery it is essential to clean it so as to remove the
coarser emery before using the finer. The glass is ground with emery
in water either on a metal former or, when flat surfaces are required,
on another flat piece of glass, with the emery between the two glass
surfaces.
Carborundum is used mainly for fast work when much glass needs
to be removed, and for grinding quartz. It is cheaper and harder than
emery. It can be graded by particle size; the coarsest normally used
is 90 mesh. The medium size is about 180 mesh, and the fine about
300 mesh, with very fine of 600 and 900 mesh; this last is usually
unnecessary since better results can be obtained with jeweller’s rouge.
Carborundum is also often graded in F numbers: F consists of 240
mesh and finer; 2F of 280 and finer; and 3F of 320 mesh and finer.
Carborundum is used in the same way as emery, and has the advan-
tage that carborundum wheels and blocks can be obtained.
Corundum or alumina is again faster than emery, but it is much
more expensive. Finally, diamond powder can be used, and is very
fast but by far the most expensive. Once the glass has been ground it
can be polished with jeweller’s rouge or Cerirouge (see p. 83). With
these polishes an optical finish may be obtained.

Glass Holders
One often needs to manipulate a piece of tubing so short that the
normal method cannot be employed because the glass becomes too
hot to hold. Various types of holder can be used. A rubber bung,
covered with asbestos paper and bored with one hole containing a
glass rod or tube, can be inserted into the tube as in FIGURE A &
Asbestos paper can be wrapped round a small tube until it fits into
the tube to be worked (FiGure 9, JJ). The smaller tube may now be
manipulated with comfort and hence any desired operation can be
performed on the larger tube. A variation of these methods is
to
employ a glass holder of the type shown in FIGURE 9, IIT. The holder
is made byjoining a small tube to a large tube ;the two tubes
must be
coaxial. The large tubing is then cut about 2 inches from the
joint.
40
 GLASS HOLDERS
In use, the tube to be worked is wound with asbestos paper and fitted
into the large tube of the holder; a rubber bung can often be used
instead of the paper.
Holders for spherical glass bulbs are mentioned on p. 32. BARR
and ANHORN (1949) give many examples of the use of holders of the
type of FiGure 9, ///; in many laboratories, however, they are rarely
or never used.
- Asbestos 2

ee ae
IT

FiGureE 9. Some ways of holding glass tubes

Large Tube Supports

Large tubing is often too heavy to rotate as well as support with the
left hand, so a support is used to take some of the weight. The very
simple device shown in FiGurRE 10 consists of a piece of wood with a
‘V’ cut in the top, and a slit to let a thumb-screw slide up and down,

[i FiGureE 10. A support

for large tubing

—
can be
It is fixed to the stand by the thumb-screw so that the height
atom models )
varied. A few freely rotating wooden balls (e.g. small
A clamp
can be fixed along the ‘V’ to increase the ease of rotation.
4]
 GLASS-WORKING EQUIPMENT
and stand is often useful; the glass can be rested on the rod of the
clamp, between it and the boss.

Adjustable Rollers
The effort required to work long heavy pieces of tubing, even with
supports as described in the previous section, is tiring and may pre-
vent adequate rotation of the glass in the flame. This effort may be
reduced by the use of adjustable rollers (FIGURE 11), which may be
in the form of two pairs of rollers on one stand or two sets of single
rollers. This latter arrangement is rather more versatile, since they
can be placed far apart to support long tubes, whereas the two pairs

[- ai
Ficure 11. Adjustable rollers for rotating glass tubes

on one stand would have to be placed near the point of balance.

When the glass is placed on the rollers it can be rotated easily with the
palm of the left hand whilst the right hand is left free to work the
glass at the end of the tube. If glass-blowing is to be carried out on
the rollers, as distinct from flaring or reaming, the blowing tube must
be connected by a swivel connection. These rollers can be driven by
an electric motor, but for most purposes this is an unnecessarily
elaborate device.

Glass-Blower’s Swivel
For the amateur glass-worker it is very helpful if the rotating move-
ment of a piece of work held horizontally in the flame can be con-
tinued in a horizontal line, with the work lifted slightly out of the
flame for blowing, instead of having to lift one end to the mouth.
The glass-blower’s swivel enables him to do this.
This small device is shown in FiGure 12, /., It is a small union,
to
which a rubber blowing tube can be connected, and inside
which an
42
 GLASS-BLOWER’S SWIVEL
inner tube can rotate. It can be machined in metal or can be made
quite satisfactorily from glass (FiGuRE 12, IJ). The inner tube has
holes around its perimeter directly in line with the union T-piece. If
made from glass these holes can be formed by puncturing molten
glass with a tungsten needle, or by drawing out tiny spindles and
breaking these off. The inner tube must fit closely into the tube in
which it rotates, to minimize air leakage.

{To blowing tube tTo blowing tube

Vg

——S
———

To apparctus To apparatus .

FiGuRE 12. Two designs of glass-blower’s swivel. J is of metal and

IT is of glass

Polarized Light Strain- Viewer

When intricate pieces of apparatus have been constructed it can be
useful to observe the presence or absence of strain with a polarized
light viewer. A. L. REIMANN (1952) gives descriptions of various
systems and explains the optics of these systems.
The most sensitive form has a diffused white light shining through
a polaroid screen, then a mica tint plate and a second polaroid screen
crossed with respect to the first. The piece of glass apparatus under
examination is held between the first polaroid screen and the tint
plate. Unstrained glass appears coloured violet-red, and the presence
of strain is indicated by change of colour in the strained portions
through red or blue-green to yellow, depending on the nature and
amount of strain. The orientation of the work must be varied to give
maximum colour change, otherwise quite a large strain can be
missed. This can happen because glass subject to simple axial
tension or compression behaves as a uniaxial crystal with optic axis
along the direction of stress. Along this direction, therefore, no
double refraction occurs.
The tint plate produces a retardation of about 570 millimicrons, so
that the yellow components of white light vanish. The resulting
colour of violet-red (or ‘sensitive violet’) is changed when an extra
retardation is introduced by the strained glass. The eye is very
43
 GLASS-WORKING EQUIPMENT

sensitive to colour changes in this region. J. H. PARTRIDGE (1949) has

given a table of colours observed in a strain-viewer without and with
a tint plate.
Glass-Cutting Wheels
Large glass-working laboratories may be equipped with a glass-
cutting machine. This usually consists of a high-speed, power-
driven, fine abrasive wheel, or, better, a steel wheel in whose
perimeter are embedded fine diamond particles. A lubricant—water,
or water and cutting oil emulsion—is played on to the faces of the
wheel from jets on either side. The glass to be cut is held on a mov-
able steel table mounted on rollers.

Ficure 13. Cutting through

a glass tube with a high
speed wheel

For an oblique cut the glass can be clamped in position. When a

straight cut across the length of the tubing is required the tube is
carefully rotated during the last stages of the cut, in the direction
shown by the arrow in FiGuRE 13; this keeps the wall of the tubing,
at the point of cutting, perpendicular to the cutting blade and
prevents a portion of the wall from breaking away at the completion
of the cut. By varying the manipulations the ends of tubing can be
slotted and small holes or slots can be cut in the walls. A small hole,
for example, is produced when the cutting operation is stopped just
as the wheel breaks through the wall. )

Lapping Wheel
For many grinding operations a lapping wheel can be most useful. A
horizontal metal table rotating at about 30 r.p.m., sprinkled with
carborundum powder and moistened with water or cutting oil—
water emulsion, forms a very satisfactory grinding table. A raised
rim (see FIGURE 14) prevents the grinding paste from being thrown
from the circumference of the table while an inclined baffle, as shown,
44
 LAPPING WHEEL
made of wood with a lower strip of leather, keeps the paste from
piling up on the outside.
In most cases the glass being ground can be held by hand, but for
making accurately parallel faces, for instance, some form of adjust-
able mounting is useful. This can usually be built to suit the partic-
ular operation.

Ficure 14. A lapping wheel

Carborundum Grinding Wheel

A very fine, high-speed, carborundum wheel—of the sort employed
for grinding very hard metals—can be used for grinding glass. A
delicate pressure must be used, particularly with tubing, to avoid
shattering the glass. The wheel must be continuously moistened
either with a jet of water or by holding a wet cloth against a portion
of its face.
The method gives a very coarse finish but enables a large piece of
glass to be ground down quickly. The work must be finished with
carborundum powder by hand or using a lapping wheel.

Annealing Oven
Glass-working shops that are called upon to make large or intricate
apparatus should be equipped with an annealing oven. With some
ingenuity a home-made one can be built using Nichrome wire and
fire-brick.
The glassware should be transferred to the oven as soon as possible
and the temperature slowly raised to the annealing point of the glass
used. This temperature must be held for from five to ten minutes or
more and the oven then allowed to cool slowly. The maximum
cooling rate depends on the glass and wall thickness, and data can
usually be obtained from the manufacturer. For Pyrex brand glass,
Corning No. 7740, and wall thickness not greater than 2 mm, the
maximum cooling rate allowed should be 56°C/min. For a wall
thickness of 5 mm the cooling rate should not exceed 9°C/min.
Annealing schedules for some glasses are given on p. 20.
45
 GLASS-WORKING EQUIPMENT

Sometimes it is possible to transfer the article straight into a hot

oven. This depends upon the particular piece of apparatus. With a
diffusion pump, for instance, where most of the joints have been
allowed to cool completely before the final parts are assembled, the
whole article is placed in a cool oven which must slowly be raised to
the annealing temperature. On the other hand, it may be desirable
to transfer a complicated part, a multiple glass-to—metal pinch seal,
for example, directly from the glass-working flame to a hot oven
already at, or near to, the annealing temperature. This is the case
when cooling without careful annealing can set up large strains.

Treatment of Slight Burns

Trivial burns are sometimes produced by the accidental handling of
hot glass. These may be smeared with vaseline or skin cream and
covered with clean or preferably sterile linen. Blisters should not be
broken. Tannic acid jelly is not recommended.

Amount of Equipment required for Different Purposes

The various items of equipment we have described are all useful in a
laboratory where a great deal of difficult glass-working is carried out.
Much difficult work can be done without a tube-cutting device, glass
holders, large tube supports, adjustable rollers, a glass-blower’s
swivel, strain viewer, cutting wheel, lapping wheel, grinding wheel
and an annealing oven. These items, however, together with a glass-
working machine (Chapter 6), are desirable in a laboratory where all
kinds of difficult jobs are tackled. In school laboratories only simple
jobs will normally be done. The necessary equipment for such jobs
consists only of a hand torch, an air blower or bellows of some kind,
a glass knife, a spike, a flaring tool and forceps, and a set of rubber
stoppers with a blowing tube. Stands and clamps will be available as
ordinary laboratory equipment. An asbestos board is useful to
protect an ordinary bench used for glass-working.
In research work a lack of the proper facilities may lead to a con-
siderable loss of time because jobs are done far more slowly when the
proper equipment is not available. For example, much time is lost
when a complex apparatus cracks because an annealing oven is not
available, and when large tubing is manipulated without a glass-
working machine. We think that R. THRELFALL’s (1898) remarks are
still worthy of note. He wrote ‘Though, no doubt, a great deal can
be done with inferior appliances where great economy of money and
none of time is an object, the writer has long felt very strongly that
English physical laboratory practice has gone too far in the direction
of starving the workshop, and he does not wish, even indirectly, to
46
 AMOUNT OF EQUIPMENT REQUIRED
give any countenance to such a mistaken policy. Physical research
is too difficult in itself, and students’ time is too valuable for it to be
remunerative to work with insufficient appliances.’

REFERENCES
Barr, W. E. and ANHOoRN, V. J., 1949, Scientific and Industrial Glass
Blowing and Laboratory Techniques; Pittsburgh, Instruments
Publishing Co.
PARTRIDGE, J. H., 1949, Glass—to—Metal Seals; Sheffield, Society of
Glass Technology.
REIMANN, A. L., 1952, Vacuum Technique; London, Chapman &
Hall.
SHENSTONE, W. A., 1889, The Methods of Glass Blowing, 2nd Edn;
London, Rivingtons.
SmiTH, I. C. P., 1947a, Chem. Age, Lond., 165.
SmiTH, I. C. P., 1947b, Chem. Age, Lond., 571.
THRELFALL, R., 1898, On Laboratory Arts; Macmillan, London.

47
 Chapter4

BASIC GLASS-WORKING OPERATIONS

The Preliminary Preparation of Materials

BEFORE any glass-blowing is started it is highly desirable that the
tools and all pieces of glass rod, tubing, etc., should be collected
together and laid out systematically. At times this may seem un-
necessary and a waste of time, but it will be found to be worth-while
in the long run. When deciding what will be required, the complexity
of the job in hand must be borne in mind: the more complex the job
the more care required in preparing the materials and in working out
mentally the way in which it is to be done. If this is not done, sooner
or later it will be found that some tool is missing when it it required
and that the work cools and cracks while the tool is being sought. B.
D. Boras (1921) wrote ‘A large amount of glass-blowing is spoiled
through carelessness in arranging the work beforehand’.
The need for system and order in all practical work has been
stressed by M. FARADAY (1842), who wrote, in giving general rules
for young experimenters, “They all relate to method, that great source
of facility and readiness which is equally influential in the perform-
ance of the most common and the most difficult processes.’ His
remarks are as pertinent today as when they were made, particularly
as the tradition of the handing on of practical knowledge is not as
strongly maintained as it was in the latter part of the nineteenth
century. (R. W. Bunsen used to teach his students glass-blowing and
other practical arts and even went so far as to make the more difficult
pieces of apparatus himself.)
There seems to be a general psychological tendency in doing
practical work with glass (and other things) to try and take short cuts
and save time by using improper and hazardous methods, and in our
experience this tendency in the end leads to more waste than saving
of time. One must also avoid the temptation, in any given case, to
think that just for once a risk can be taken.
The Cleaning of Glass Tubes
Before any glass-working is started care should be taken to make sure
that the glass is thoroughly cleaned. This cleaning is necessary for
48
 THE CLEANING OF GLASS TUBES
two main reasons. One is that if inorganic material is left on the glass
it will not burn off in the flame and will become embedded in the
glass. This can, under some circumstances, lead to local strain in the
joint and may also prevent the two pieces of glass fusing together
properly, leaving a pinhole in the joint. The second reason why
thorough cleaning is important is that it is essential, if the apparatus
is to be used for high vacuum work, for the inner surfaces to be
grease-free. It is much easier to do this cleaning before a complex
piece of apparatus is assembled.
Normal chemical cleaning procedures are satisfactory for cleaning
glass tubing for glass-blowing. Much of the dirt can be removed by
the use of water, a detergent and a long brush suchas a burette brush. It
is desirable not to scratch the inside of the tube with the wire of the
brush. The use of a brush can be avoided by pushing or pulling a piece
of soft cloth through the tube. Chromic acid is satisfactory for cleaning
glass, after removal of the gross dirt, provided care is taken to see that
the glass is free from mercury. The usual laboratory cleaning solu-
tion, known colloquially as ‘chromic acid’ contains about 50 ml of
saturated aqueous sodium dichromate in a litre of concentrated sul-
phuric acid. If mercury is present a residue is precipitated which is
difficult to remove. Thorough washing with water is necessary after
cleaning in chromic acid. Organic solvents are adequate for remov-
ing grease from the inner surfaces of tubes where no other dirt is
present; while acetone or alcohol may be useful for drying glassware,
care should be taken that liquid or vapour from these solvents does
not remain inside the glass when glass-blowing is commenced, be-
cause of the risk of explosion. When glass which has been in contact
with mercury is worked, great care should be taken to clean the
mercury from all parts of the system likely to become hot, since mer-
cury vapour is toxic.
It is very difficult to remove the last traces of chromium from glass
cleaned with the usual chromic acid mixture. The remaining traces
may cause error in oxidation studies. R. A. WILKINSON (1955) found
serious errors from this cause in fat oxidation studies, and he cleaned
greasy pipettes in a boiling, } per cent aqueous solution of “Teepol”’
(sodium secondary alkyl sulphate). Glass can also be cleaned in
boiling aqua regia.

Breaking Glass Tubes

Tubing up to about 15 mm diameter can be broken by scratching the
surface in one place with a file or glass knife, making the scratch at
right angles to the axis of the tube and then placing the thumbs on the
side of the tubing opposite to the scratch; the thumbs are pressed
49
 BASIC GLASS-WORKING OPERATIONS

forward and the rest of the hands are slightly pulled apart (FIGURE
15). This is a gentle process and not a test of strength. The scratch
on the tubing can, with practice, be made with one stroke of the glass
knife ;it is not necessary to try and cut through the wall of the tubing,
and the knife is soon blunted when it is used like a saw. Most glass-
workers have the impression that wetting the initial cut makes the
glass break more easily. Soda glass is easier to scratch with the knife
than the borosilicate glasses.
We find that tubing in the range of 15-30 mm diameter (approxi-
mately) can be broken by a variation of the above method : The tub-
ing is scratched fairly strongly, and placed, with the scratch upper-
most, so that it rests on a copper wire, about 3 mm diameter, which is
opposite to the scratch and parallel to it. The scratch is moistened,

Pressure

Pressure Pressure

ark

Ficure 15. Breaking a small glass tube

and the ends of the glass tube are forced down strongly while the tube
is balanced on the copper wire: a square break is then usually
obtained. This method works best with moderately thin-walled
tubing.
Tubes exceeding 15 mm diameter (and smaller) can be broken by
the ‘hot-spot’ technique: A fairly deep scratch is made in the glass in
one place, and wetted. The point of a piece of glass rod is then heated
white hot, to give a hot spot a few mm in diameter, and this is placed
on the end of the scratch. The glass may crack right away, or it may
crack when the hot spot is removed and damp cotton wool is applied
to the hot glass. Usually, with borosilicate glass, the crack goes only
part of the way round the tube; a soda tube often cracks completely.
The crack can be continued by placing another hot-spot against its
end. This is continued until, with luck, the ends of the crack meet.
With borosilicate glass the crack usually wanders, and a square end is
not obtained. A cracking-off burner (p. 29) can be used in place of a
hot spot.
A square end is more easily obtained by the hot wire method. The
tubing is first scratched all the way round. This may be done by
50
 BREAKING GLASS TUBES
hand, a line being first drawn round the glass with a wax pencil—the
line can be checked to see that it is perpendicular to the tube axis—
and the scratch made along the pencil mark. Without the pencil
mark, the scratch may deviate from the perpendicular. A peripheral
outside scratch may also be made by rotating the tube against a
rotating steel wheel. An internal scratch may be made with a dia-
mond by the method described on p. 38.
After scratching, 26 s.w.g. (or thicker) Nichrome wire is wrapped
once round the scratch, which is wetted. The wire is heated to red-
ness electrically, and kept taut round the glass—for this purpose it is
convenient to have insulated handles on the wire. The tubing may
then crack smoothly; if not, it may be induced to crack by touching
the scratch with wetted asbestos when the wire has been hot for about
half a minute. An internally scratched Pyrex tube may not crack
until it is put in a flame.
Sometimes a large bottle is cut to make a heavy beaker or a bell
jar. Big bottles are generally of soda glass, and can be cracked with a
flame. The bottle is first fairly heavily scratched all the way round,
and then a very small flame is directed on to the scratch. A crack
should start, and it can be led round by the flame. A cracking-off
burner (p. 29) is specially useful here. This operation is quite difficult
for beginners, because the glass often just shatters; but with practice
it can be mastered. Increased safety is obtained by wrapping strips of
wet blotting paper or asbestos paper round both sides of the crack,
leaving a gap, about }-4 inch wide, containing the crack. An alter-
native method for large bottles, etc., is to spin them on a turn-table
and to have a fine pin-point flame playing at the level of a small initial
scratch; after a few minutes the flame is removed and a piece of wet
cotton wool pressed on the scratch.
I. C. P. Smit (1947) has described a simple method of grinding
through glass tubing with a hacksaw: the blade is replaced by soft
wire (such as iron, copper, or Nichrome) of 24 to 18 gauge, and this is
ground through the glass, using 180-mesh carborundum powder and
water and an ordinary sawing action. A glass-cutting wheel is
described on p. 44.
Knocking off Small Bits of Tubing
Sometimes there is not enough tubing left beyond the scratch to hold
in the hand, so that the normal method for small tubing (p. 49) can-
not be used. The easiest method to use in this case is to scratch the
tube and then place it on the top of a triangular file, with the scratch
opposite to and above the edge ofthe file (FIGURE 16). The end of the
tube is then knocked off by giving it a sharp blow with a heavy object.
1 51 | |
ow ae
es th oad
 BASIC GLASS-WORKING OPERATIONS

Often the end of a broken tube is not quite square. This can be
somewhat corrected, especially in larger tubes, by using a piece of
wire gauze: The gauze is stroked against the end of the tubing, when
small pieces of glass are broken off the protruding edges until the end
assumes a roughly square shape. Another method is to heat the end
of the tube strongly and cut off the protruding parts with scissors.

Scratch

ii
Triangular tile
Ficure 16. Knocking the end off a small piece of tubing

Holding and Rotating Tubes in the Bench Flame

Holding and rotating tubing in the flame is a most important part
of glass-blowing with small tubing. Usually the weight of the glass is
supported by the left hand, which is held over the tubing, and the
right hand, under the tubing, maintains only a light grip. Although
this may feel odd to the beginner, it will be found satisfactory because
the right hand is free to do the manipulations and to bring the tube to
the mouth for blowing. However, the right hand may also hold the
weight with an overhand grip while the left hand has the underhand
grip. In research work it is convenient to be able to use both
methods.
The more usual disposition of the hands is shown in FiGuRE 17.

Ficure 17. Holding and rotating a tube. ‘The

disposition of the hands is as seen by the operator

The tubing is rotated by the forefinger and thumb of the left hand,
and rotates in the groove between the thumb and forefinger of the
right hand, and can rest lightly on the middle finger of the right hand.
It is easy enough to get a smooth rotation of the glass when it is rigid,
but when it is soft both hands must rotate the glass without distorting
52
 HOLDING AND ROTATING TUBES IN BENCH FLAME
it, and this is difficult (p. 2). One must get the ‘feel’ of the glass. Its
softening is a gradual process, and its ‘feel’ during working is deter-
mined by viscosity, surface tension, thickness etc. Practice is neces-
sary here.
The glass should be rotated horizontally, unless it is desirable for it
to flow towards one end. Usually it is rotated with the upper side
moving away from the operator, but this seems a personal preference.
When the glass becomes very hot and plastic, it is essential for the
right hand to keep the tubing rotating at the same speed as the left-
hand piece, or streaks and ridges will appear. This operation is found
difficult by some people, but others master it with practice.
With soda glass the rotation of the article in the flame is of prime
importance, since if part of the work is allowed to cool, re-heating
may cause cracking and the destruction of the article; but with Pyrex
and similar glass it is often possible to heat only a portion of the
article at a time.
When the tubing is too heavy to hold and rotate, one end can be
rested on a support (p. 41), or rollers can be used (p. 42).
The beginner should not be disappointed if at first these operations
produce a hopeless mess ; with practice the facility is gained.

Bends and Spirals

To bend a piece of tubing, the left-hand end is first closed. The tube
is rotated in a fairly big flame with not very much air (for soda) or
oxygen (for Pyrex) init. A special flame from a ribbon burner (p. 29)
is best, but not essential; an ordinary bench blowpipe gives an ade-
quate flame. The tube is held as in FiGure 17. When soft over the
desired length it is brought, held horizontally, up to the chest. The
right hand then brings its end vertical, and gentle blowing into this
end brings the tube into shape with no kinks or flat portions. If a bad
bend is made it can be corrected by heating and blowing. Many
other methods of making a bend have been described.
Spirals
To make spirals of uniform size throughout is difficult; but useful
spirals can be made fairly easily by a beginner even though the pro-
duct may not look too elegant: Longer sections than usual must be
heated, and care must be taken not to flatten the tubing. It is easier to
make spirals with smaller tubing because the ratio of wall thickness to
diameter is highest.
Spirals can be made either free-hand or with a mandrel or former.
For the free-hand method a right-angle bend is made (FiGurRE 18, /),
with a length of tube (between A and B) of about 6 inches left to act
53
 BASIC GLASS-WORKING OPERATIONS
(FIGURE 18,
as a handle. The tube is next bent into a right angle at C
II) so that the length BC is equal to the desired radius of the spiral.
over a
It is then heated near C, both by rotation and lengthways,
e.
fairly large area—the major part of the heating being on the outsid
is held horizo ntally in
When the glass is soft enough, the length AB
first arc of
the left hand and the length CD is slowly bent to form the
the
the spiral (FIGURE 18, I//). The tube is next heated adjacent to
first arc, and, when sufficiently soft, is bent to continue the arc
previously made. The bending is continued until a complete turn is
made. The next step is to ensure that the tubing AB is axial. This is

Ficure 18. The production of a spiral

done by heating at B and bending AB into the correct position. All

the bending should be done outside the flame and it is advantageous
to look at the spiral from various angles to check the uniformity of
the coils. The spiral is completed by continuing as before.
To make an elegant-looking product it is necessary to pay attention
to the uniform diameter of the coils and to keep the spacing between
them constant. The final product (FIGURE 18, JV) usually has to have
the two ends aligned. To do this, the last turn is heated at a distance
from the end of the last arc approximately equal to the spiral radius
and a right-angle bend is made. The end of the last arc is then
heated, and this right angle is pushed inwards. Final adjustments can
be made after rotating the spiral to check alignment: The spiral
should appear as in FIGURE 18, V.
It is of ten easier to wind a spiral on a mandrel or former of carbon
metal tubing or glass tubing wrapped with asbestos paper. A right-
angle bend is made in the tube to give a short portion at the end which
54
 BENDS AND SPIRALS
is strapped to the mandrel. The tubing is heated in sections of about
1 inch and wrapped round the mandrel. The asbestos can be used as
a guide to ensure that even turns are made—one way is to wrap
Asbestos

Ficure 19. Production of a spiral with a

mandrel

asbestos tape round the mandrel in a spiral path (FiGuRE 19).

Finally the asbestos is softened in water and the spiral is removed
from the former.
A flat circular spiral can be made by the free-hand method: a
right-angle bend is made, and the glass near the bend is bent round
the short section used as a handle. By repeated heating and bending
a flat coil is produced.

Putting a Handle on the End of a Tube

When it is necessary to work comparatively near to the end of a piece
of tubing it usually becomes too hot to hold. It may be inconvenient
to support the whole weight of the glass in one hand, and the un-
supported end will almost certainly drop under its weight when the
tube is hot. A handle on the end of the tube is therefore useful.
First let us consider the case of a tube which is to be drawn off near
the end. In this case the end need not be airtight. The tube to have
the handle attached is held in one hand and the piece of rod or tubing
to be used as a handle in the other. The end of the tube is rotated in _
the flame until the glass is soft for a length about equal to its radius. ~
The handle should be heated until its end is molten and then attached
to a selected point on the end of the tube. The tube should be taken
out of the flame and the handle pulled across until the point at which
it is attached to the tube comes into contact with the far side of the
tube. This should be heated so that the two sides of the tube and the
handle are thoroughly fused together and the handle should then be
pulled approximately central on to the axis of the tube. The whole
mass of glass should be kept heated for a short while by rotating in
the flame to allow the strain in the glass to be reduced. Such a handle
should not crack off when the glass cools, and is very useful when the
tube has to be drawn off or bent near to the end. It does, however,
leave the end of the tube in such a state that it has to be either drawn
off or cut off and no blowing can be carried out as it is not air-tight.
a2
 BASIC GLASS-WORKING OPERATIONS

A handle can also be fused on to the closed end of a tube to support

it while glass-blowing operations are carried out: This is done by
heating the end of a piece of rod for the handle, heating the centre of
the closed end of the tube, and fusing them together. When the
handle is to be removed after the operations are finished the point
where it is fused on can be heated, the handle drawn off, and the end
of the tube blown out again to its original shape.
With big tube it may be convenient to prepare a handle from rod
shaped at its end like a Y, so that it sticks to two points along a dia-
meter at the end of the tube. It is then not necessary to push in the
ends of the glass tube. The Y-shaped handle is specially useful for
very big tubes.

Drawing Tubes off to Spindles

Drawing a tube off to a spindle is the simplest of the basic glass-
working operations. It is a convenient alternative to cutting glass
tubing when short lengths are required and it is the preliminary stage
to many other operations.
To draw off a tube using the flame of a bench torch, one places it
in the flame and rotates it, so that heating is uniform, until the tube
starts to collapse under the surface tension of the molten glass. The
glass should then be removed from the flame and the two parts drawn
steadily apart at a rate which gives a strong spindle as the glass cools.
The spindle can then be cut by melting it in the flame and pulling the
two bits apart. A more axial spindle is obtained when rotation is
continued as the glass parts are drawn apart.
It may be convenient (specially for big tubes) to shape the glass
somewhat in the flame before drawing the spindle. When the glass
starts to soften, the two parts of the tube can be drawn apart slightly
to maintain an approximately uniform thickness of glass as the tube
is drawn down. In this way a tube may be tapered or constricted. If
the two parts of the tube are drawn apart too rapidly, before the glass
has collapsed sufficiently, thin walls are obtained, and, if they are not
drawn apart enough, thick walls result. The rotation of the two parts
of the tube must be the same and the tube must be kept approxi-
mately straight during this process if a symmetrical taper or con-
striction is to be obtained. Continuation of the process greatly
reduces the size of the tube at the point of heating.
The spindle makes a very useful handle for holding the tube while
other glass-blowing is carried out, and owing to its much reduced
diameter it does not become hot when the tubing is worked near
to the end. The tip of the spindle can be cut off to attach a rubber
tube for blowing. If for any reason an accurately central spindle is
56
 DRAWING TUBES OFF TO SPINDLES
required it may be found convenient to put in a second spindle using
the first one as a handle, since short lengths of tubing are normally
more easily worked than long pieces which are unevenly balanced
about the point of working. The operations involved in putting a
central spindle near the end of a piece of tubing are shown in FIGURE
20. In many glass-blowing operations, a long spindle is extremely
useful. Spindles can be closed very quickly in a flame, or opened
again with a glass knife; we find it preferable to use them, when
possible, instead of rubber stoppers.
Experience of the feel of the molten glass as the tube is rotated and

+
a
——OLSS—
OO —
Curt

FiGure 20. Drawing tubes off to spindles

drawn down will enable a happy medium, between the extremes of

thin-walled spindles and spindles of glass rod, to be obtained.
The operation of putting a spindle in a tube is often described as
‘pulling a point’ or ‘pulling a spear’. This is a most important opera-
tion in bench flame work. The spindle must be central, otherwise the
rest of the work will not be central.

Round-Bottomed Closures of Tubes

When a good spindle is obtained and a rounded end is required the
glass should be heated at the end of the taper (as shown in FIGURE
21): the glass here should be collapsed down in a small flame and
pulled out comparatively slowly so as not to leave a blob of glass on
the end of the tube. The whole end of the tube should then be heated
by rotating it in a fairly large flame until the glass at the end of the
tube has started to collapse a little. The tube should be removed from
the flame at this stage and should be blown into gently. The glass
should be heated relatively strongly—most beginners tend to try to
ff
 BASIC GLASS-WORKING OPERATIONS
finish.
work the glass before it is hot enough which leads to an uneven
blow gently so that the end of the
It is advisable in the first place to
tube is stretched slowly. In this way the blowin g can be stoppe d at
any required stage, and this is desirable since if any part of the glass
is thinner than the rest this part will start to blow out first because of
its lower viscous drag but will cool faster than the thicker parts
which will continue to blow out after the thinner glass has become
too cool to extend much more. In this way a more even rounded end
will be produced than if the glass had been blown rapidly and then
allowed to cool.
The most usual fault in making a rounded end is to leave a lens of
thicker glass at the point where the spindle was drawn off. This may
be because either insufficient glass was drawn away with the spindle—
if this is the case more can be pulled off with a piece of rod—or else

ee
JctTe ae
re Tae)
FiGureE 21. Closing a tube with a
round end

the glass was not heated enough to allow the unevenness to even itself
out by flowing under surface tension. A lens of thicker glass left in a
round-bottomed closure can be removed by heating it with a small
flame and blowing it into a small bulb. The whole end of the tube
should be re-heated by rotation in a medium-sized flame until the
glass starts to flow. The end is then blown out to a hemispherical
shape and annealed.
Fabrication of Thin Glass Rod
In some of the following sections the use of thin ‘spills’ of glass rod
(welding rod) is described. They are most easily made from glass rod
by heating a portion and pushing the ends together until a reasonable
amount of molten glass is accumulated. This is drawn out at a rate
which gives a rod of 1-2 mm diameter.
If no rod is available the thin glass rod can be made from tubing:
thick-walled capillary tubing is the most convenient but normal 8
mm tubing will do quite well. A length of tubing, about 10-15 cm for
58
 FABRICATION OF THIN GLASS ROD
thick-walled capillary or 15-20 cm for 8 mm tubing, is rotated in the
flame until a part collapses down and fuses to a solid mass. This
process is continued, pushing the ends inwards slightly to build up a
mass of fused glass with no fine capillary through the centre. Such a
capillary is easily seen while the glass is hot. The molten glass should
be drawn down in the same way as for rod, taking care that the tube
next to the fused mass does not draw down to form a length of fine
capillary. This will be prevented if the initial stage of the drawing
down is carried out very slowly, Straight rod is obtained by rotating
during the drawing-out.

Blowing Holes in Glass Tubes

There are basically two sites for blowing holes in glass tubing, the
first where the hole is to be made in the end of a tube which has been
drawn off, and the second where the hole is required in the side of a
tube. The principle is the same in both cases: the glass to be blown
out is heated until it is soft enough to flow easily and is then blown
out into a very thin bubble. The bubble is then removed. If the glass
is too hot or it is blown too hard, the bubble will not be formed, but a
small irregular hole will be blown and a cloud of very fine glass flakes
produced. This is undesirable.
In the case in which the end of the tube is to be blown out care
should be taken to see that the tube is heated evenly over the whole
end, otherwise the hole produced in it will be skew. If the hole is to
be smaller than the bore of the tube only a small area of the rounded
end must be heated and the blowing must not be continued too long,
since once the bubble of glass is formed it is enlarged principally by
drawing more glass from around the edge of the hole in the original
tube, thus making the hole larger. This also applies to the blowing of
holes in the side ofa tube.
The second case, the blowing of a hole in the side of a tube, is a
necessary preliminary to making a T-joint, and therefore it is essen-
tial to be able to produce the right-sized hole in the correct place. The
procedure is very similar to that for blowing out the end of a tube,
except that a small area on the side of the tube is blown out. Gener-
ally speaking it is more convenient to heat a comparatively large area
on the side of the tube and blow a slight bulge, and then heat the
smaller area to be blown out more strongly with a rather smaller
flame. This should give a hole, if blown out correctly and not for too
long or too hard, which is surrounded by shoulders so that the rim of
the hole stands clear of the general line of the tube walls. In many
cases it is desirable to have these shoulders present (e.g. making
T-joints).
59
 BASIC GLASS-WORKING OPERATIONS
The removal of the thin glass bubble from the hole does not present
any problems. It can be easily knocked off, but this scatters flakes of
glass which may get inside the glass being worked. The glass bubble
may be crushed in a sheet of paper, which certainly causes less mess
and may be preferred in the interest of general tidiness when the work
is being carried out on fixed apparatus. A solution to the problem,
which does not cause flakes of glass and leaves a supply of glass for
further working about the hole, is to fuse the bubble down into a rim
around the hole. This usually happens as a matter of course when a
fairly cool flame is played on the thin glass, but if too hot a flame is
used the glass is boiled which does not improve its properties (see
Devitrification, p. 18).
To blow a very small hole the method described for multiple
perforations ii bulbs is useful (p. 72).

Joining Tubes of Similar Size

In this section detailed descriptions of the operations will be given for
the benefit of the complete novice. It is convenient to deal with the
subject in three sections, (/) the case in which both tubes can be
rotated in the flame of a bench torch, (2) the case of one fixed tube
and the other movable, and (3) the case of two fixed tubes.

(1) Joining Two Tubes in the Flame ofaBench Torch

The first step'in producing a satisfactory joint is the preparation of
the two ends to be joined. The most satisfactory method is to draw
the tubes off and to blow the ends out. If this is done correctly the
ends should be square. Cutting the tube with a file scratch leaves a
line of fine bubbles in the joint due to the remains of the original
scratch. This spoils the appearance of the joint when finished and
may also cause a weak spot at which a pinhole may develop when
testing for leaks with a Tesla coil. Jagged ends of tubing can be
heated until the glass is quite soft and trimmed with scissors.
The second step is to block the end of one of the tubes and heat the
ends of the two tubes in the flame until the glass is quite soft and
starts to collapse. Uniform rotation is necessary for this, as in all the
heating operations for this method. When this stage is reached the
glass is removed from the flame and the ends pressed lightly to-
gether; initial contact is best made at some point by holding the two
tubes at a slight angle (FIGURE 22). The ends are then brought
together by bringing the two pieces of the tube into line. A slight
pressure is then applied during the heating of the join so that the
glass is thickened around it. This glass is worked thoroughly by
alternate heating and blowing, care being taken to see that the two
60
 JOINING TUBES OF SIMILAR SIZE
parts of the tube are rotated at the same speed during the heating
process. When the glass has been worked sufficiently, so that the
original junction of the two joints cannot be distinguished, a slight
bulge is blown in the tube and it is drawn down until the walls are of
uniform thickness and the tube has a uniform bore. The whole
operation is sometimes described as ‘butt sealing’ or ‘butt welding’
and the joint as a ‘butt seal’.

Heat and
rotate

Slight pressure
EE

Heat and
VSR Ee

NN =— flow

Draw down
WN to uniform
bore

FiGure 22. Joining two tubes

It will probably be found that much practice is necessary before

good joints can be made in this way. A common difficulty is that the
two halves of the tube must be rotated at the same speed, and to avoid
this many novices tend to work round the joint in segments once the
two pieces of tube have been pressed together. This method, while
capable of producing vacuum-tight joints, leaves an irregular wall
thickness which is much more liable to crack from strains induced on
cooling, even though it may have been carefully annealed by cooling
slowly in the flame. However, it is satisfactory for Pyrex and similar
glass.

(2) Joining Two Tubes, one of which is Fixed

The necessity of joining a tube to another which is fixed arises
frequently in the construction of laboratory apparatus where the
61
 BASIC GLASS-WORKING OPERATIONS

main part of the apparatus is either fixed to a stand or too large and
unwieldy to allow the use of a bench torch. It is very often easier for
a novice to produce a good joint by clamping the glass in position and
working with a hand torch as described below, since it does not
require the same degree of manual dexterity.
The two ends to be joined should be prepared carefully. The best
method is by sealing off and blowing out, but in some cases this may
not be possible. If the end of one tube is jagged, resulting from a
break in the apparatus, and it is not practicable to close it and blow
it out, the end should be smoothed by heating and working to a flat
surface with a tool. This is done by removing protuberances with
forceps or scissors, and then pushing the surface flat with the tool.
The importance of working the ends of the tubes so that they are a
good-fit when pushed together cannot be overemphasized at this
stage, since a good fit eliminates much work in the later stages, and
overworking of glass leads to devitrification of the surface.
The working of the joint is carried out by first clamping the mov-
able tube loosely in position so that the ends of the two tubes to be
joined are nearly touching. The clamp should be left loose so that the
two tubes can be pushed together later, but it should be tight enough
to support the tube and not allow it to sag. The prepared ends of the
tubes are heated, gently at first, with a hand torch, taking care to heat
more or less evenly from all sides. Asbestos screens may be useful in
joins of this type for protection of other apparatus. When the ends of
the two tubes are hot enough to flow, the loosely clamped tube is
pushed gently into place so that the glass is in contact all round the
joint. The clamp is then tightened to prevent further movement. The
glass is worked by heating and blowing in much the same way as
described in the preceding section except that blowing is carried out
by a blowing tube attached to one or other of the two tubes. In
making fixed joints such as this it is undesirable to overheat the glass
so that it becomes very liquid, since it flows under gravity. This flow
should be kept to a minimum since it leads to uneven wall thickness.
The effects of this flow under gravity can be mitigated in a number
of ways. If the tubing being worked is vertical it may be drawn down
a little just below the join to reduce the wall thickness in this region;
then, as the glass flows, this thin section will thicken so that a uniform
wall thickness is achieved. Another method which can prove very
useful to the amateur glass-blower is to make use of a spike to push
the glass back to its original position. If the glass is heated until it is
fairly soft, the point ofthe spike can be used to lift the glass which has
dropped down. It is best to lift this glass above the position of the
joint as it has to be worked to remove the local thickening of the wall
62
 JOINING TUBES OF SIMILAR SIZE
produced by lifting, and during this working it will tend to flow down
again. This also applies to a horizontal joint where the glass tends
to collect at the bottom of the tube.
It is very often possible to produce a better finish to a joint if the
glass is actually blown while it is in the flame, but if this is done great
care must be exercised to blow very slowly and with short puffs rather
than a long steady pressure. It is also virtually impossible to do this
in a system which has a volume in excess of two litres since the pres-
sure response is subject to too great a time lag and when the blowing
is stopped the pressure inside the system may not drop for several
seconds, which may lead to a hole being blown in the glass.
An alternative procedure is to hold the movable tube in one hand
while the other hand operates the hand torch: the tubes are heated
and pushed together as before, and the movable tube is supported
and manipulated with one hand. With tubing over about 1 cm dia-
meter, the movable tube needs little or no support most of the time,
as the side not being worked on is usually rigid enough to hold it.
Hence it is possible to use welding rod for filling up any holes in the
joint. The blowing tube should be attached to the fixed tube, to
reduce the weight on the movable tube to a minimum. It is conveni-
ent to have the tubes vertical with the movable one held below. This
general method is usually preferable to the method involving clamp-
ing the movable tube, unless the latter is very heavy.
An elegant joint can be obtained by heating the joint strongly all
round, and as it thickens moving the lower tube up; then the whole is
blown out and the lower tube is moved down, so that a smooth joint
is made. This procedure resembles the joining of two tubes by
Method 1.

(3) Joining Two Tubes both of which are Fixed

This situation may arise in the repair of apparatus or in making a
closed system. If the joint is a good fit no great difficulty is to be
expected as the case is similar to that just described, except that the
ends cannot be pushed together unless they are in fairly close contact,
when they can usually be pushed together with a spike or some other
tool when the glass is hot.
If the tubes are so placed that there is a gap between them it is
possible to close this gap by various stratagems. One of the two
tubes may be drawn down slightly so that their ends are brought into
- closer contact. If this tube is vertical it is best to make the point of
drawing just below the proposed joint and to use forceps to hold the
short end of the tube just above the point of drawing. This will pro-
vide the constriction to take up the glass which flows down under
63
 BASIC GLASS-WORKING OPERATIONS
gravity. If the gap is not too great it may be filled in by building up a
network of glass in it, using thin rod. This process is rather similar
to darning the gap: a closely packed network of thin rod is built up
in the gap and is then heated fairly strongly to fuse the individual rods
together. The glass should not be heated too strongly while the
original network is being built up, as strong heating very often leads
to complete closure of the tubes. This method can be used for filling
in irregularly shaped holes and finds many applications, particularly
in the repair of apparatus which has been broken.
If the gap between the two pieces is too large for the above method
to be attempted it may be possible to fit a sleeve of wider tubing over
the two tubes and fuse its ends down on to them. This method in-
evitably leads to a very thick-walled joint, but the excess glass can be
pulled off with forceps or by fusing a glass rod on to the thick part
and pulling it out.
All joints, no matter how they are made, should be annealed with a
relatively cool flame (p. 170).
When joints are made by the above methods, pinholes are very
often left in them, particularly when they have been built up of thin
glass rod. These pinholes are usually fairly easily visible when the
glass is hot—at red heat they usually show up as fine orange lines
running through it. They are sealed by placing a small spot of glass
over their outer end using a thin glass rod for the purpose. The spot
of glass is then fused in thoroughly. It is obviously more convenient
to avoid such pinholes when the joint is made than to have to re-heat
it later. It is desirable to wear glass-blowing spectacles when looking
for pinholes.
Devitrified glass should be pulled-out of a joint with forceps or a
glass rod, and if this leaves a thin-walled joint extra glass can be built
on using thin rod.
The aim in all these operations is to produce a joint with the same
bore as the tube and with uniform wall thickness. Only practice will
allow this to be achieved. However, with Pyrex, very bad-looking
joints will hold together.

Joining Tubes of Different Sizes

There is very little basic difference in the procedures for joining
tubes
of different sizes and for joining tubes of the same
size. The main
difference lies in the preparation of the end of the larger
tube: the
end of this tube should be drawn down to a spindle
and the spindle
cut off in the flame. The closed end so produced
should be heated
locally so that when it is blown out the hole produc
ed will be of the
Same size as the bore of the smaller tube. But
it is usually better to
64
 JOINING TUBES OF DIFFERENT SIZES
aim to produce a hole rather smaller than this, since it is a simple
matter to open out the hole with a tool or to collapse the smaller tube
down slightly ; whereas it may be more difficult to close the hole in the
drawn-down portion of the larger tube if this has relatively thin walls
or uneven wall thickness. The above remarks apply equally to the
use of a bench torch and to the use of a hand torch with fixed tubes.
When the joint is to be made using a bench torch it will usually be
found easier to work so that the glass is blown through the smaller of
the two tubes, when the joint is more easily visible. If the larger tube
is used for blowing, the joint may be screened from view by it. When
the novice starts trying to make this type of joint with a bench torch,
difficulty may be experienced in rotating the two tubes of different

Thin wol/ =

RAGG0Gane Sie eee AON

ac se es

Thick wall >

Original mana
protile ———

Ficure 23. The effect of wall thickness

on the joint, when tubes of different size
are joined

bore at the same speed. The only remedy for this difficulty is prac-
tice, but it may be possible to avoid it by clamping the glass and using
a hand torch. If only one tube is clamped, it should be the larger one.
The wall thickness of the taper produced by drawing down the
larger tube can have an effect on the profile of the finished joint
(Ficure 23). If this taper has a wall thickness about the same as the
unworked tube it will tend to retain its profile in working, but if it is
thin-walled the joint will tend to have a rounded profile. This pro-
vides a method of controlling the type of joint produced. A rounded
profile can be produced by starting with a uniformly thick rounded end.
T-Joints
The making of T-joints is probably the most common laboratory
glass-working operation and therefore should be mastered thoroughly.
65
 BASIC GLASS-WORKING OPERATIONS
by the use
There are two basic methods of making these joints, (/)
of a bench torch and holding both tubes, (2) by the use of a hand
torch and clamping the tubes (or one of them).

(1) The Use of a Bench Torch

Once again the first step is the adequate preparation of the surfaces to
be joined. First the side arm should be prepared, as for butt-sealing
two tubes, by blowing out the end (p. 59). A hole is then blown in the
side of the tube at the position required (see p. 59 for details). Care
should be taken not to make this hole too large. Both ends of the
tube should then be blocked since it is best to blow through the side
arm so that the whole joint can be seen while blowing. Some workers,
however, prefer to blow through the cross-piece.

Ficure 24. T-joints

The side arm is taken in the right hand and the cross-piece in the
left and they are held in the flame so that the end of the side arm, and
the area just around the hole in the cross-piece, are heated. This is
done by directing the flame almost perpendicular to the side arm and
swinging both pieces of glass through an angle of about 200° (FIGURE
24). These motions should be synchronized in order to obtain uni-
form heating. A great deal of practice is necessary before this can be
done satisfactorily.
The two tubes are pressed lightly together when the glass is fused
and the ‘swinging’ of the glass in the flame continued until the region
of the join is thoroughly fused. The joint is then blown through the
side arm. This is repeated until the glass has been worked sufficiently
for a good joint to be obtained and it should then be annealed. ~
The above procedure can be simplified in two ways. First, it is not
necessary to swing the cross-piece in the flame during the preliminary
heating before joining the tubes, although this operation accustoms
66
 T-JOINTS
the two hands to synchronous working. The side arm only is rotated
in the flame, which is meanwhile directed on the hole in the cross-
piece. Secondly, a T-piece can often be made by working only a por-
tion of the joint at a time.
The procedure for this type of joint is the same for all sizes of
tubing. In some ways it is rather easier to carry out for relatively
large sizes (about 2 cm) than for smaller sizes (less than 8 mm). In
the latter case it is very easy to heat the cross-piece in such a way that
it bends at the joint. This may be due to several possible causes: too
large a flame; failure to hold the glass at the correct angle to the
flame; or overheating of the glass in the vicinity of the join, the heat
being transferred by conduction.
A special technique for making T-joints, which does not distort the
bore of the main tube if it is done correctly, is to make the joint by
sucking: the end of the side arm, and a localized area on the main
tube corresponding in size to the cross section of the side arm, are
heated strongly until the glass is very soft. The end of the side arm
is then placed on the heated area of the main tube and the heated area
sucked up strongly into the side arm which must have a plug of cotton
wool in it to catch any glass fragments. This produces a sound joint,
but it requires very thorough annealing because of its very local
heating.

(2) The Use of a Hand Torch

A casual study of any laboratory vacuum apparatus will show how
commonly T-joints must be made with either one or both of the tubes
clamped. Generally speaking this is a much easier technique than
that previously described, as it removes the necessity for synchronized
manipulation of the glass.
The glass is prepared in the same way as previously described, and
a blowing tube is connected to some part of the system. The tubes
are clamped in position and one of them left slightly loose so that they
can be pushed into close contact at a later stage. The glass is heated
in the vicinity of the join by working around it with the hand torch
and the two tubes are pushed together. The glass is then worked by
collapsing it down and blowing out until it has fused together
thoroughly. The joint is then blown out to the desired profile.
If difficulty is experienced with the glass flowing under gravity the
sarhe stratagem may be used as described in the section on butt-
joining two tubes (p. 62). If for any reason the tubes cannot be
pushed together the two pieces should be as good a fit as possible
before the joint is started and may be pushed into closer contact when
hot with the aid of a spike, or if this fails the gap can be filled in with
F 67
 BASIC GLASS-WORKING OPERATIONS

thin rod. This may leave pinholes in the joint and a careful inspec-
tion should be made while it is hot. When larger size tubing is used
it may be found advantageous to blow the holes in the tubes rather
smaller than their bore, since this provides a means of thickening the
wall of the joint as it is blown out to a smooth profile.
When making a T-joint with a hand torch it is often best to clamp
the cross-piece and fix a blowing tube to it, and to manipulate the
side arm with one hand. As with the joining of two tubes (p. 63), the
side arm will be supported by the rigidity of the glass on one side of
the joint most of the time, unless it is of small tubing. This method is
usually preferable to that involving clamping both tubes.
The construction of big T-pieces is described on p. 94.

Y-Joints
Joining two tubes in the form ofa Y is most easily effected by bending
one of the tubes to the desired shape and then making a normal
T-joint at the bend. If the T-joint is made first and then one of the
arms is bent, the resulting Y-piece will probably not be symmetrical
and the finish on the surface of the glass may not be good owing to
volatilization of some of its components during the making of the
joint. This causes the surface to wrinkle on bending the tube.
The most convenient order of carrying out the various steps is to
prepare the side arm first, bend the main tube and blow out the hole
for the T-joint, and then make the joint before the glass has time to
cool.

4 and 5 Way Junctions

Once the simple T-joint has been mastered the manufacture of 4 and
5 way junctions presents no difficulty: the operations described for
making T-joints are just repeated the required number of times. The
various side arms are all prepared as the first stage, and sufficient
corks, or other means of blocking the tubes as they are joined on,
must be at hand. The only other step which requires extra care is to
see that completed joints are not overheated in making others, since
if any of the side arms are unsupported they will bend under gravity.
Adequate annealing is, as in all other joints, essential.

Working Capillary Tubes

Working capillary tubing requires a slightly different technique from
the working of normal tubing. This is principally due to the small
bore of the tube and the very thick walls. Generally speaking it is
necessary to use a smaller, hotter flame for capillary than would be
68
 WORKING CAPILLARY TUBES
used for normal tubing. Three types of operation will be described
in this section, (/) joining capillary to wider bore tubing, (2) joining
capillary to capillary, (3) capillary T-joints.

(1) Joining Capillary to Wider Bore Tubing

The only difference between this type of joint and a normal butt joint
is in the preparation of the capillary tube: first the end of this is
sealed off and a small fairly thin-walled bulb blown equal in diameter
to the hole in the wide tube. The end of this bulb is then blown out
and the joint made by the normal method. The flame must be
directed so that the capillary is not collapsed. If the capillary has a
very fine bore some difficulty will be found in blowing through it.
This may be overcome by connecting it to a compressed air line while
the small bulb is blown; the end of the bulb can then be pulled out
with a rod and cut off to give the prepared end.
If for any reason it is found to be impossible to blow out the end of
the tube, the cut end of a capillary can very often be opened out by
using a mounted needle from a set of dissecting instruments as a tool;
but this is not recommended if the tube can be blown.

(2) Joining Two Capillary Tubes

Owing to the fact that capillary tubing is used in many systems to cut
down the gas space it is often desirable for the finished joint to have
the same bore as the tubes. Therefore it is undesirable to prepare the
ends by blowing out; it is much better to join the ends of tubes cut at
a file mark. They should be heated until the glass is soft, and the
tubes then held at an angle to one another. The two tubes are
brought into contact by touching their edges together, and then
brought into line ensuring that the bores of the tubes are coaxial.
The joint is then finished by heating as small a length of the tube as
possible, collapsing the bore down to about half its size and blowing
it out to about twice its size. This operation should be repeated until
a good joint is obtained. These joints require careful annealing
because of the heavy wall thickness.

(3) Capillary T-Joints

The side arm for the T-piece should be prepared by blowing the end
out of a piece of capillary as described above for joining capillary to
wider bore tubing. The cross-piece is prepared by heating a short
length of the capillary in the flame and blowing a slight symmetrical
bulb. This should not be more than twice the original bore of the
tube. A very small hot flame is used to heat one side of this small
bulb, which is then blown out as for a normal T-joint. If difficulty is
69
 BASIC GLASS-WORKING OPERATIONS
to
experienced here, it may be found advantageous to attach a glass rod
this out into a
the point heated on the side of the small bulb and pull
of
cone. The rod should next be cut away with the flame at the point
the cone, which can then be blown out much more easily. Lastly the
joint is completed in the normal way and annealed thoroughly.
Fine capillary tubes may become blocked with water when blown
into from the mouth; it may be desirable to put a drying tube in the
blowing tube or to blow with a rubber bulb.

Blowing Bulbs in the Middle of a Tube

To ensure that the centre of the bulb lies on the axis of the tube from
which it is made, the tubing, when in the flame, must be rotated
continuously as described on p. 52; otherwise the glass will sag to one

a iT

RASS Ee
POSS a eee ae

vie lV
Ficure 25. Blowing a bulb in the middle of a tube

side. Bulbs can be blown in the middle or at the end of a tube; the
former operation is probably easier at first. A fairly large and
relatively cool flame is used to heat a large portion of the tube, and as
the tube is rotated in the flame the ends are pushed in to form a con-
centration of glass (FIGURE 25, J). The tube is taken from the flame
and gently blown so that a slight bulge is formed (FIGURE 25, //).
The danger of distortion of the glass is much reduced when the
blowing is done with a glassblower’s swivel (p. 43): In this case the
blowing tube is attached to the end of the tube—usually the left—
which is held in an overhand grip. The process of accumulating glass
is continued after the bulge has been re-heated, and the bulge is re-
blown a number of times until enough glass has been accumulated
for a final sphere to be blown. This may be blown by a series of
puffs ;the thinner parts blow out first, but they also cool first, and by
blowing at the right moments a spherical bulb may be obtained. The
70
 BLOWING BULBS IN THE MIDDLE OF A TUBE
tubing should be rotated when it is being blown into; this is facili-
tated if the lips are moistened before blowing.
Instead of accumulating the glass in one bulge, an alternative
method is to make three bulges by pushing the glass together (FIGURE
25, III) and then reduce these to a single bulge by repeated blowing
and shrinking. This reduces the time required to collect enough glass
for the final bulb to be blown.
If the bulb to be blown is of radius much greater than that of the
tube, a large tube should be joined into the small tube, as in FIGURE
25, JV. The three portions of tube should be coaxial. The large one
is then shrunk and blown a number of times, and finally blown to the
required bulb. It should be longer than the required radius of the
bulb, as its length will decrease with heating and blowing.
To blow an elegant large bulb is very difficult. It is much simpler
for the scientist to join a round-bottomed flask into the tube.

Blowing Bulbs at the End of a Tube

Blowing bulbs on the end of tubing may be found difficult because
there is no portion of tubing to help to keep the bulb central whilst it
is being blown. A round end is first made on the tube, which is

FIGURE 26. Blowing a bulb at the end of a tube

rotated continuously so that the glass thickens (FIGURE 26, J). The
tube is taken from the flame, held near the vertical, and, with con-
tinual rotation, blown into until a slight bulb is formed (FIGURE 26,
II). This is returned to the flame and collapsed and blown a number
of times until enough glass has collected for the final bulb. When
blowing this it is essential to rotate the tubing all the time, and to
have a gradual air pressure, or a series of puffs, rather than a sudden
pressure. The bulb must be watched as it is blown; if it is becoming
off-centre it may be best to shrink it down and blow again. Beyond a
certain size, however, a bulb cannot be shrunk, specially if it is thin.
Another method of blowing a bulb at the end of a tube is first to
put a central spindle in it, and then accumulate glass just beyond the
shoulder of the spindle by pushing the glass together with the spindle,
and occasionally blowing through it. When enough glass has been
71
 BASIC GLASS-WORKING OPERATIONS
collected, the spindle is drawn off in a very small flame, and the bulb
is blown as described above.
To blow a big bulb a larger tube must be joined to the end of the
tube and shrunk to give the necessary accumulation of glass. The
beginner should be able to make useful small bulbs which will become
quite elegant with practice. For research, however, it is best to fuse a
round-bottomed flask to the tube.

Multiple Perforations in Bulbs

A bulb with a number of small holes in it is useful in bubbling a gas
through a liquid. These holes are easily made: a small bulb is
blown; a very small area of the bulb is heated strongly, and the glass
is pulled out by touching it with the fine point of a glass rod, and then

(o}
°o

a A I LT

FiGuRE 27. Making a number of holes in a bulb

drawing out the molten glass into a thin thread. This is repeated in a
number of places, to give the result shown in FiGuRE 27, I. The
threads are then broken off near the bulb, and each hole so made is
heated and rounded to give the finished perforated bulb shown in
FIGURE 27, JI.
An alternative method is to use a very small flame on the original
bulb, and to blow a thin-walled bubble, as in FiGuRE OMBee EE RON
number of these bubbles are blown, then they are all broken off and
each hole is smoothed off. If a bubble is broken accidentally, the
hole can be blocked with damp asbestos paper so that more bubbles
can be blown.

Flanging, Flaring or Bordering Tubing

A number of tools are useful in putting flanges of different kinds on
the end of a glass tube. The operations are purely manipulative and
require no blowing. Various tools which can be used in flanging
are
shown in FiGurE 4; perhaps the triangular-shaped tool is one
of the
most useful. Carbon rods are also useful (p. 33), and they have
the
advantage that they do not mark the glass; an overheated metal
tool
72
 FLANGING, FLARING OR BORDERING TUBING
sometimes leaves a small deposit of meta
l or oxide on the glass. The
general method used in flaring is to heat the
end of the glass tube over
a length of 2-3 times its diameter, and when
the glass has become soft
and thickened (FiGurE 28, I) the tool is inser
ted and the tube rotated
The glass should be rotated fairly quickly
backwards and forwards
by rolling it between the forefinger and thum
b of the left hand; or
normal rotation may be used (p. 52). The
glass should be rested on
the tool, and the action should be gentle. The
spinning, as this action
is called, Should be done with the flame impi
nging on the base of the
cone formed (FiGurE 28, //). For a standard
taper, a hexagonal

———- a

=> =
SSS _—

FiGuRE 28. Some methods used for flanging, flaring or

bordering tubing

reamer is used (p. 31). Lubrication is desirable to prevent the glass

becoming twisted round the reamer. Various ways of using tools are
shown in FIGuRE 28, I//, JV and V.

Internal Seals
Internal seals can be made by a variety of methods, which will be
described ;some indication will be given of situations in which they
might be useful. Internal seals are also called ring seals.

Method 1
In this method a single piece of tubing is sealed into the larger tube.
The first step is to form a slight ridge around the inner tube, at the
point where it is to be sealed, by rotating it in a small hot flame and
pushing its ends inwards slightly. If this is done correctly a sym-
metrical ridge should be formed.
The outer tube is then closed to a round end and a hole blown
73
 BASIC GLASS-WORKING OPERATIONS
centrally in this, large enough for the internal tube to be erage.
small enough to prevent the ridge on this tube passing through. e
outer tube should be longer than the inner so that the open end may
be closed by a cork or rubber stopper (FIGURE 29, I); the blowing is
carried out by blowing through the projecting inner tube. It may be
more convenient to close the other end of the outer tube with a
spindle. This gives a useful handle.

Vil ‘ VT
FiGure 29. Various methods for making an
internal seal

The glass is then heated by rotation in the

flame so that the outer
tube is fused to the ridge formed on the inner
, and finally the joint is
blown until a good-shaped joint is obtained.
To get a good-shaped
shoulder it is necessary to direct the flame
mostly against the smaller
tubing. The heating and blowing should
be repeated until good fus-
ion of the two tubes is achieved. Any
pinholes present can be sealed
as before by the application of a small
spot of glass from:a piece of
rod. These spots of glass should be fused
in thoroughly.
The unsupported internal tube will
sag under gravity during the
above operations, but if the rotation
of the work is sufficiently
uniform it can be centralized as the
glass cools by allowing it to drop
74
 INTERNAL SEALS
to its central position and keeping it there by rotation. Alternatively
it can be supported on a rod held by the cork blocking the end of the
outer tube. If this latter method is used care should be taken to see
that the flow of air from the inner to the outer tube is not hindered
since it is essential for the air pressure to be equal in the two tubes at
all times.
Internal seals require more thorough annealing than the more
simple joints, and if any pinholes require subsequent repair great care
should be taken in reheating the joint.
Method 2
In this method the internal tube is in two pieces and there is less
likelihood of making a joint with pinholes in it. The first step is the
preparation of the internal tube: a length of tubing is drawn off, the
end blown out and polished smooth in the flame. The projecting part
of the internal tube should be prepared in the same way. The outer
tube should be prepared by closing one end to a round end.
The joint can be made most easily if the internal tube is supported
almost in its final place by pads of asbestos paper or by a glass holder,
but whatever support is used it must allow the air pressure to be the
same on both sides of the internal tube. The inner tube is put in place
and the rounded end heated. When this is starting to collapse in-
wards the internal tube is pushed forward slightly so that its prepared
end comes into contact with the heated end of the outer tube. This
glass is worked thoroughly by heating and blowing until a good fus-
ion between the two parts is achieved (FIGURE 29, //). The end of the
tube is allowed to cool slightly and then, with a fairly small flame, the
area of glass covering the internal tube is heated and blown out. The
other end of the outer tube is then blocked with a stopper (FIGURE
29, III) and the remaining part of the inner tube joined on as in a
simple butt joint (FIGURE 29, JV). Once again it is important that the
air pressure should be the same in both inner and outer tube and that
the tube should be rotated uniformly. Finally the joint should be
annealed.
These joints can be made with the internal tube unsupported by
inserting the inner tube only after the outer tube is hot enough for it
to stick firmly and then proceeding as above. This is rather more
difficult and is seldom as satisfactory as wedging the inner tube,
although at times one is tempted to avoid this added step.
This method is not on the whole satisfactory: with soda glass, the
cold inner tube will probably shatter, and if a borosilicate glass is
used to avoid this, the difficulty of getting the inner tube central still
remains. A better method is the following: the outer tube is
aD
 BASIC GLASS-WORKING OPERATIONS

rounded off and pulled to a slight taper. The inner tube is flared
slightly so that it fits into the taper at the desired point and is thereby
centralized. The tubes are fitted together when cold, and the opera-
tion is continued as before.

Method 3
The third method of making internal seals is not so widely applicable
as the previous two, but it is very useful for certain joints. It is best
described in terms of one of its uses—in the manufacture of a Liebig
condenser.
Let us assume that the condenser is in the state shown in FIGURE
29, V: the first internal seal has been made by one of the two
methods described above, and a side arm has been put into the water
jacket. A connection should be made with rubber tubing, so that the
water jacket of the condenser is connected to the inner tube, and then
the outer tube is heated with a hot narrow flame to collapse it down
on to the inner tube for the internal seal. When the tube has col-
lapsed on to the second tube the heating is continued and both tubes
drawn off together (FIGURE 29, VJ). The system is then blown
slightly to give a good round profile to the end of the outer tube;
when this has been done the end of the inner tube is blown out and
the joint finished as in Method 2. Another method of performing this
operation is described on p. 130.

Method 4
This method involves the fusing of a flared-out narrow tube into a
wider one. This type of seal has its uses in pieces of apparatus such as
McLeod gauges. The first step is to flare out the end of the inner
tube, using a flaring tool, until it will only just pass inside the wider
tube. This is done by rotating the tube in a flame and using the
flaring tool as shown in FiGure 29, VII, or as described on pds ite
very large flare is required it will probably be necessary to fuse a
wider tube to the narrow one as the initial stage in flaring. When the
flare has been made to fit it is packed in place with asbestos paper or
held with a rod, and the outer tube is collapsed down on to it
(FIGURE 29, VIII). The outer tube is blown out again and the process
repeated until a good fusion of the two pieces is achieved. This type
of seal requires particularly careful annealing.

Use of the Hand Torch

Generally speaking it is best to make internal seals using a. bench
torch and rotating the glass in the flame. If this is not practicable
it is
possible to work with a hand torch with the glass clamped if Method
76
 INTERNAL SEALS
1 is used: in this case a rim of glass, left by the blowing of the hole in
the outer tube through which the other tube passes, can be heated
with a hand torch and pushed in against the walls of the inner tube
with a spike. This can be done also in cases where the formation of a
ridge of glass on the inner tube has been impracticable. Once an air-
tight joint has been made by pushing the glass into place, the hand
torch can be used to fuse the two tubes together by working round
the joint. As before it is essential to have the blowing tube connected
to both inner and outer tubes.
One great difficulty with this type of work is the force of gravity
acting on the inner tube: If this is unsupported it is very difficult to
obtain a good-looking joint, and even if it is supported the results are
not always satisfactory, owing to non-uniformity of heating. It is
best to clamp the tubes vertically. The inner tube may be held in one
hand. Difficulties may possibly arise if the outer tube has thin walls,
because any holes will enlarge as the glass is worked. In this case they
may be filled with rod as in the other cases described. Care should be
taken in inspecting these joints for pinholes and annealing the work
when completed.
Thin Glass Windows
In recent years the need has arisen for very thin glass windows cap-
able of withstanding a vacuum yet thin enough to transmit a and f
particles. S. ROSENBLUM and R. WALEN (1945) have described the
following method for putting very thin windows into capillary tubing
of up to2 mm bore. The process is very simple but may require some
practice before the precise conditions necessary for a satisfactory
window are achieved.
The end of the capillary tube is heated until the glass is soft, then
before it has time to cool it is touched on to the surface of a thin
bubble of glass and a slight suction applied. This forms the window
into a concave shape and draws it slightly down into the capillary,
whose ends then protect it from damage. The bubble of thin glass
should be thin enough to show interference colours. These windows
will stand a vacuum provided atmospheric pressure is on the concave
side of the window. If they are subjected to a pressure difference in
the other direction, failure occurs due to the reversal of curvature.
A method for making larger thin windows has been described by
J. T. Ltoyp (1949). For windows in the end ofa tube he advocates the
blowing of a thin-walled bulb at that end and then drawing it into the
tube to form the window by the following method: the thin-walled
bulb is blown at the end, the wall of the tube is then heated all round
at a short distance from the end until the glass is fused, and then the
77
 BASIC GLASS-WORKING OPERATIONS
bulb is immediately re-heated and sucked inwards. The bubble is
drawn into the tube and is fused to the walls in the region where the
tube was heated. This gives a relatively robust window capable of
withstanding a vacuum and protected by the projecting ends of the
tube. :
Lloyd also gives details of a similar method for the formation of
windows in the sides of tubes. The bulb is blown on the side of the
tube in the same way as it would be in the first stage of making a
T-joint; it is then heated and sucked in. Lloyd’s method requires
more practice than the other method before satisfactory results can
be obtained, but it has the merit of being more versatile.

Spinning Out Feet

The stands on measuring cylinders, wine glasses, etc., are called ‘feet’
by the glass-blower. They are made by an operation similar to
flanging: the glass must be thick enough for spinning out. A piece
of 3-cm diameter tubing about 10 cm long, with spindles (spears) at

=< =e
FiGure 30. The production of ‘feet’ for vessels

each end, is prepared. One end is then heated in a large, relativ

ely cool
(soft) flame until it thickens, and, by pushing with the spindl
e and by
blowing, the shape shown in Ficure 30, /is produced. The
spindle at
A is removed and the end closed. This closed end is then
joined to the
closed and rounded end of the tube for which the
feet are required.
The two ends are simply brought together (FiGur
e 30, 7), and, by
blowing through both D and E and slightly pulling
, a good joint is
made. The tubing is next cut at C; and the feet
are worked to shape
with a rounded tool (FIGURE 4, G), using
the same operations em-
ployed in flanging (p. 72). It is important, howeve
r, to get the glass
very hot and soft, so that the feet can be spun
out in one operation
and not to push too hard with the tool, to
avoid producing buckled
78
 SPINNING OUT FEET
feet. As the hot glass is rotated horizontally, the tool should be
slowly brought from the horizontal to the vertical. The spinning out
should be done along the whole width of the glass to produce the shape
in FiGure 30, J//; if only the outside is spun out the shape in FIGURE
30, JV is produced.

Tapering Glass Tubes

Glass tubing may be tapered by two methods, depending upon the
wall thickness required. For a thin-walled taper, a central spindle is
first put in the tube, and the shoulder is heated in a large soft flame.
The tubing is rotated and slightly collapsed, and then taken from the
flame and pulled out to the required taper, with constant rotation.
This will give a fairly thin-walled taper, conical in shape; but by
slightly pushing in the ends while heating the shoulder a thicker wall
can be obtained.
For a thick-walled taper the operations are the same except that
the tubing at the shoulder is shrunk down and blown out a number
of times to accumulate glass. When enough glass is present, the
shoulder is made hot, slightly blown out, and then drawn down out of
the flame.
To taper a tube from one size to another, two tubes of the desired
sizes are joined and worked to a taper at the joint.
When making tapers to fixed dimensions it is convenient to draw
the required taper on asbestos paper and work the glass to match the
drawing.

Sealing-in Sintered Glass Discs

The glass disc to be sealed into a tube must be made of glass having
the same thermal expansion coefficient as the tube. Discs of different
porosities can be obtained ; they consist of a large number of particles
of glass sintered together so that numerous channels remain. When

SR es
Tne Ly ns /)
SE
Ficure 31. Sealing a sintered disc into a tube

working with a disc it should be warmed up carefully. One end of the

tube into which the disc is to be sealed is closed with a rubber stopper
carrying a tube A (FIGURE 31). The disc is rested against this tube,
and should just fit into the outer tube. If that is too big, a constriction
must be put in.
79
 BASIC GLASS-WORKING OPERATIONS

The tube and disc are carefully warmed until the flame is yellow;
one side of the tube is then heated strongly and attached to the disc.
The disc is next made perpendicular to the tube by pushing with a rod
through the end B, and manipulation with A if necessary. The tube
is allowed to cool slightly so that the disc is firm, and this is then
attached again to the tube by strong local heating opposite the first
point of joining. The whole tube is then shrunk down on to the disc.
It will probably be necessary to blow the tubing out a few times,
blowing through both ends. The tubing can finally be pulled down
to a uniform diameter. The joint should be annealed with extreme
care, preferably in an oven.

Mending Cracks
Unfortunately, mending cracks in laboratory glassware is one of the
more common tasks. Cracks usually occur in glassware for one of
two reasons—either failure to remove intrinsic strain from the glass
by annealing, or else straining of the glass by careless handling of the
apparatus itself. Whatever the cause of the crack the procedure for
repairing the damage is the same:
The first stage in mending cracks and the one requiring most care is
the initial heating of the glass: If this is done carelessly the crack will
almost certainly spread much further. The glass should be warmed
up slowly with a small flame, very little oxygen or air being mixed
with the gas. This small flame should be played over the surface of
the glass in such a way that it does not remain directed at any one
part. The whole region surrounding the crack should be heated and
not just the immediate area, particular care being taken if the crack is
near an internal seal to see that this is heated evenly. As the glass
warms up the size of the flame and the proportion of air or oxygen in
it can safely be increased until the glass becomes soft and there is no
danger of further cracking. For very hazardous repairs it is desirable
to heat first with hot air, and then a flame with no air or oxygen until
the article is covered with soot, which probably assists the uniform
distribution of heat. When the soot starts to burn off, the most
dangerous part of the operation has been reached, in our experience.
When the glass is soft the flame should not be made any hotter—a
relatively cool flame is much more effective for sealing cracks. The
flame is then played on to the crack, starting at one end, and the glass
heated gently. It should be possible to see the crack sealing up under
these conditions and the flame should be moved along it as it seals.
If too hot a flame is used the crack will open as it is heated because
the glass at its edges will become sufficiently hot to draw back under
surface tension before the inner parts of the crack have had a chance
80
 MENDING CRACKS
to fuse together: this is then a progressive process and the crack can
only be sealed by filling in with glass rod or pushing the sides together
with a tool.
When the crack has sealed, a hotter flame can be used and the glass
worked by heating and blowing until the crack has been completely
sealed. It is necessary to keep the glass in the region of the original
crack, and all the surrounding parts, at a dull red heat for a short
while (for Pyrex) and then to cool it slowly to allow sufficient anneal-
ing time. Care must be taken with this general annealing of a fairly
large area of glass if further cracks are to be prevented.
Extremely unpleasant cracks can be sealed in Pyrex glass. With
soda glass, however, the sealing of cracks is a hazardous undertaking,
and the whole apparatus may be lost. Cracks can be sealed in Monax.
Detection and Removal of Leaks and Holes
Holes in awkward places are often difficult to find when one is
assembling complex apparatus. If the joint in question is put in a
luminous flame, or a flame put round the joint, and the apparatus
blown into, the air will rush out of the holes and blow the flame away
in a blue jet. The hole, when found, can then be repaired with thin
glass rod (p. 58); or, if it is very big, it can often be filled in with a
piece of glass. When a piece of glass falls out of an apparatus, it can
be fused to a rod, to provide a handle, and put back again with the
help of thin glass rod.
These methods work well with Pyrex glass, but not with soda glass,
which cracks too easily on re-heating. A hole which has been mended
with blobs of glass from a rod needs to be fused down and blown out
a few times, to make it more homogeneous. Fusing down and blow-
ing out glass often closes a small hole—this method works better
with soda than Pyrex glass.
Small holes and leaks can often be found by covering the outside of
the apparatus with soap solution and noting where bubbles appear
when it is blown into. Very small holes can be found with a high
frequency discharge: the apparatus is evacuated with a rotary pump
(or a good water-pump) and the electrode of a Tesla coil is brought
near it. A discharge appears inside the apparatus, and if the electrode
comes near a hole or fine capillary in the glass, the discharge passes
through it and a very bright spot of light appears at the leak. With
a water-pump vacuum the method is not so sensitive, because only a
poor discharge is obtained in the apparatus. When the pressure in
the apparatus is about 0-1 mm, very fine leaks can be found.
However, the Tesla discharge may blast a hole in the tubing, and
is not useful for glass-to—metal seals, because it goes straight to the
81
 BASIC GLASS-WORKING OPERATIONS

metal. If the seal is covered with cotton wool soaked in carbon tetra-
chloride or alcohol, this will pass through any leak, and should pro-
duce a greenish or bluish colour in the Tesla discharge. This method,
however, is not reliable.
The minute capillary leaks found with a Tesla coil should be
plugged with a blob of hot glass; the whole joint then needs heating
and annealing. When assembling complex apparatus, parts which
will become inaccessible to the flame must be tested for leaks before
they are put in their final positions.

Grinding Glass
Glass surfaces which must fit closely are usually ground to shape.
The most common example is the cone-and-socket joint: the sur-
faces are first worked to an approximate fit with fairly coarse car-
borundum (about 100 mesh) which is put between them and wetted.
The cone is rotated by hand or in a lathe or drill chuck, or by a
special machine, until the surfaces meet closely—the carborundum
must be kept moist all the time. The coarse abrasive is then washed
off, and the grinding continued a number of times with progressively
finer abrasive. Various possible abrasives are mentioned on p. 39.
Surfaces which require regrinding, such as those of stopcocks,
usually fit fairly well, and a fairly fine grade of carborundum can be
used for the grinding. Stopcocks must not be ground too much, or
the holes in the plug will no longer meet those in the barrel.
A flat surface can be put on the end of a tube by grinding. Such a
surface may be required when a flat plate has to be cemented on to a
tube. The flange is ground against a flat metal or glass plate using
wet carborundum as an abrasive. The flange must be kept steady and
not rocked, and is ground with a rotating motion of the whole tube
about a point outside it. Progressively finer grades of carborundum
are used. An operation of this kind is carried out very easily when a
lapping wheel is available (p. 44).

Releasing Frozen Glass Surfaces

Ground surfaces are often found to be jammed together. If the parts
are robust, a sharp tap will often release them. A glass stopper stuck
in a bottle is often released by tapping with another glass stopper, or
something solid. When the substance in the bottle has cemented the
stopper in position, the joint may be put in a suitable solvent. Frozen
surfaces are often more easily released after the joint has been im-
mersed in a detergent solution for some time.
Fragile joints which cannot be struck may often be released by
82
 RELEASING FROZEN GLASS SURFACES
quickly warming the outer member. This may be done with hot
water, or, less safely, with a flame. If the surfaces are made too hot
they are melted together and the apparatus must be discarded.

Polishing Glass
For many years, the standard method for polishing glass in the
laboratory has been by rubbing the glass over a sheet of brown paper
covered with a mixture of flour emery and a solution of camphor in
turpentine. C. V. Boys (1927) has stated that he can find no reason
other than tradition for the use of the solution of camphor in turpen-
tine rather than pure turpentine for lubricating grinding media.
When the best possible polish has been achieved with the above
mixture, rouge and water should be used for improving it. During
the polishing only a fairly light pressure should be applied and the
specimen should be moved fairly rapidly. If the polishing is being
done manually, pressure should be applied evenly over the whole
surface being polished. It is possible, however, to mount small
samples in pitch and use a rotating polishing surface.
Cerium oxide (Cerirouge) has been recommended as being superior
to rouge as a polishing material (J. sci. Instrum. 22, 1945, 224) and
special cloths are marketed impregnated with cerium oxide. These
are useful for polishing glass and other materials such as fluorite. For
details of the methods of polishing optical surfaces, works on that
subject should be consulted.

Drilling Holes in Glass

The problem of drilling holes in glass has occupied the attention of
scientists since the days of Faraday and many methods have been
described. Articles reviewing the subject have been published by C.
V. Boys (1927) and more recently by P. GRODZINSKI (1953).
The method which has been used most widely is probably that
involving a copper or brass tube in conjunction with a grinding mix-
ture of carborundum and water; a slotted copper rod is also effective
for small holes. To obtain the best results by this method a fairly
high drill speed should be used and the copper tube drill should be
raised from the cut periodically to allow a fresh supply of the grind-
ing medium to enter the hole. It is also desirable to have some slots
in the end of the tube to hold the grinding medium. The pressure on
the drill should be fairly light, as too great a thrust may splinter the
glass and may also slow up the cutting by preventing the grinding
medium from getting between the rotating drill and the glass. Grod-
zinski favours the use of diamond powder in place ofthe carborundum,
G 83
 BASIC GLASS-WORKING OPERATIONS

but in most laboratories this is not obtainable. It is easy with a tube

to obtain holes 3 cm in diameter.
Many variations of the shape of steel drills for glass have been
described—from Faraday who used the ground end of a triangular
file to C. V. Boys who designed a special drill of tool steel—while
other workers (B. BROWN, 1927; N. G. HEATLEY, 1938) advocate the
use of a normal twist drill. These drills are used in conjunction with
some lubricant—usually the traditional camphor in turpentine, al-
though pure turpentine or other mobile liquid seems to be as good,
while Heatley states that dilute sulphuric acid is effective. To avoid
splintering, the hole should be drilled from both sides.
Grodzinski, commenting on the use of sintered carbide drills,
states that high speed is necessary for their use and strongly favours
diamond drills. He gives a full account of their use and gives details
of the mounting of diamonds for this purpose. Probably the simplest
method of preparing a diamond drill for small diameter holes was
described by Heatley (1938) who broke off the end of the elongated
eye of a sewing needle and soldered a small diamond chip into this.
He was able to drill quite deep holes, of diameter less than 1 mm, in
glass with this type of drill.

Cementing Glass
Standard cones and sockets of metal and glass can be joined with
Picien or W wax; such joints are vacuum-tight. Windows can be put
on metal apparatus with these waxes. Picien is a black wax of low
vapour pressure, and flows freely at about 80°C. Three useful
Apiezon waxes are W, W100 and W40, which are applied at tempera-
tures of 100°C, 80°C and 40-50°C respectively. When a large glass
socket is waxed to a métal cone, the strain set up on cooling can
easily crack W wax, but W40 is satisfactory for large joints. These
waxes can also be used to seal two glass surfaces together, and a
vacuum-tight joint can be obtained without ground surfaces.
Aero Research Ltd make two resin cements, Araldite type I and
Araldite 103 and Hardener 951. They are epoxy resin cements, the
former being thermal setting and the latter cold setting, and make
transparent seals. To use Araldite type 1, the surfaces are cleaned,
degreased and heated to 100°C, and the cement in solid form is
applied. It melts and runs smoothly over the surface. The surfaces
are pushed together and held together by a clamp while the whole is
heated for a curing time dependent upon temperature, varying from
10 minutes at 240°C to 7 hours at 140°C. These are minimum
times
and should be increased for stro nger joints.
84
 CEMENTING GLASS
To use Araldite 103, 8-10 parts of Hardener 951 are added to 100
parts of Araldite—they are both liquids. The mixture is stirred and
allowed to stand before use for 20-30 minutes at room temperature.
The pot life of the mixture is about 3 hours at 20°C, so it is made up as
required. The glass is degreased and the cement applied with a
spatula or brush as a thin, even coating on both surfaces. The sur-
faces are immediately joined and clamped lightly. The cement cures
at a rate depending on temperatures. 16-24 hours are required at
20°C and 10-20 minutes at 100°C. It is claimed that Araldite can be
used for vacuum-tight, glassto—glass seals.
Mica windows can be fused to glass using a special powder of a
lead borosilicate glass softening at 450°C (J. S. DoNnaL, 1942). A
suitable glass for solder seals of this kind is BTH No. 93.
Silvering of Glass
Glass surfaces may be silvered by chemical precipitation of silver on
to them. This is most conveniently done by reduction of a silver salt;
the speed of precipitation depends on the reducing agent used. The
glass must first be cleaned: it is degreased with benzene, washed with
acetone and water, and left in “chromic acid’ (see p. 49) for an hour
or more at room temperature. It is then washed with tap water,
followed by distilled water. Some workers prefer to then clean again
in concentrated nitric acid, followed by the water washings. The
glass must on no account be allowed to dry after these washings.
The speed of precipitation of silver required depends on the surface
to be silvered. If the rate of precipitation is too great, the silvering
agent may be spent before it can be removed from the surface, and a
‘bloom’ will appear because the silver is in contact with the spent
solution. This is a discolouration of the silver. Thus if a rapid silver-
ing agent is used for Dewar flasks, it may be difficult to avoid spoiling
the silvering.
The silvering agent usually consists of two solutions, sometimes
three, which are mixed immediately before use and poured on to the
object to be silvered. Various recipes have been proposed (I. C.
GARDNER and F. A. Case, 1931). The following are useful for slow
silvering (R. B. Scott, J. W. Cook and F. G. BRICKWEDDE, 1931):
Solution A—SO gm silver nitrate in 2 1. distilled water
Solution B—90 gm potassium hydroxide in 2 |. distilled water
Solution C—(80 gm cane sugar in 800 ml. distilled water) plus
(100 ml. ethanol and 3-5 ml. nitric acid sp. gr. 1-42).
Solution A should be kept in the dark, and C improves with age.
The solutions are used in the proportion A: B: C=16:8:1 by
volume. To use, concentrated ammonia is added to A drop by drop,
85
 BASIC GLASS-WORKING OPERATIONS

with stirring, until the precipitate first formed is redissolved. The

requisite amount of B is next added, and a dark brown or black pre-
cipitate forms. (But if enough ammonia has not originally been added
to A, a yellowish-green precipitate will be formed; the solution
should be discarded in this case). Ammonia is then added until the
brown-black precipitate is almost redissolved ;and to the resulting
solution the correct amount of C is added to give the final, prepared
solution. The intermediate solution prepared from A plus B can be
used immediately, or kept up to an hour, after which it loses its
activity. If allowed to stand for about half an hour it will become
practically clear.
To silver a mirror a bath should be used just larger than the article
to be silvered. The glass should be cleaned and under no circum-
stances allowed to become dry. The prepared solution is poured over
the glass and the bath is rocked steadily. The silvering must be
stopped at the right time or the mirror will be tarnished. The end-
point is when a flocculent precipitate is formed. The time for this to
happen depends on a number of factors: old silver nitrate solution
deposits silver more slowly ; temperature alters the rate, and a test run
to find the approximate time may be desirable. To see how the silver-
ing is proceeding one may take the mirror out of the solution, but it
should never be exposed to the air for more than 1 or 2 seconds. The
rocking during the silvering prevents precipitate adhering to the
surface and spoiling the mirror.
The mirror can be silvered a second time if a thicker coating is
required. Brighter and thicker coats are obtained by using a more
dilute solution twice instead of a more concentrated solution once.
Front-silvered mirrors can be polished with optical rouge on a pad
of chamois leather over cotton wool. Care is necessary because it is
easy to polish completely through the silver. Back-silvered mirrors
can be protected by first varnishing the silver and then painting the
varnish. Front-silvered mirrors can be protected with a thin coat of
lacquer, such as a dilute solution of cellulose nitrate in amyl acetate.
The lacquer should not be too thin, or interference colours may be
produced.
In order to silver Dewar flasks the jacket is cleaned and filled about
two thirds full with the prepared solution. The Dewar should be
rolled or shaken during the silvering. The end-point is determined by
occasionally pouring out a little of the solution, and when a floccu-
lent precipitate is formed the solution should be removed from the
flask as quickly as possible to prevent blooming. The flask is then
cleaned and dried. If the glass is first etched with a dilute solution
(about 0-7 molar) of hydrofluoric acid before its final cleaning, the
86
 SILVERING OF GLASS
stability of the silvering with respect to peel-off is increased, and spots
are less likely.
A simple method for filling and emptying Dewar vessels is men-
tioned on p. 123. F. E. J. OCKENDEN (1938) has given some hints on
silvering glass, and E. A. BAKER (1931) has described a method for
electrolytically removing silver from one face of a small mirror. P.
B. G. UpTon and E. F. G. HERINGTON (1950) have described a
method for depositing a heavy silver film by reduction of a silver
ammonionitrate solution by hydrazine sulphate. The deposition of
silver other than on the desired surface is prevented by pyridine. The
effect of the pyridine is due to the presence of surface-active impurity,
and the process becomes controllable by adding cetyl pyridinium
bromide to ‘inactive’ pyridine.

Depositing Copper on Glass

Copper can be electro-deposited on to a silvered article: the plating
bath consists of 10 gm of copper sulphate in 100 gm water. The
article is taken from the silvering bath, washed, and put wet in the
plating bath—on to a copper cathode so that electrical contact is
made to the silver. A copper anode is placed 1-2 cm above the
article. The electrodes should be about the same size as the article.
The current density should be about 0:05 amp/cm?, and in 24
minutes a thin copper film should be made. If the current density is
too high a granular deposit will be formed which will rub off, and if
it is too low a non-uniform deposit will be formed.
Copper can be chemically deposited in the same way as silver,
using phenylhydrazine as the reducing agent. F. D. CHATTAWAY
(1908) found this to give good results with blown glass but not with
glass that had been polished. For plane surfaces E. A. H. FRENCH
(1924) obtained good results with hydrazine sulphate.
Many metals can be deposited on glass in a vacuum, by vaporizing
the metal and condensing it on the glass. J. YARWooD (1955) has
given a general account of this method.

Metallizing Preparations for the firing on of Metal Layers on Glass

Preparations of silver, platinum, gold and palladium are available
which consist of a suspension or solution of the metal, or a com-
pound, in an organic material; adhesion-promoting materials are
also present. The cleaned glass is coated with the preparation by
brushing, spraying or dipping. It is then heated fairly slowly to a
high temperature (the firing temperature), when the organic mater-
ials are oxidized or volatilized and the metal becomes strongly bound
87
 BASIC GLASS-WORKING OPERATIONS

to the surface. The preparations mentioned here are all made by

Johnson, Matthey & Co.; they also supply thinners for them.
A glass surface to be silvered by a firing-on process should be
cleaned with soap and water or detergents. For metallizing with
platinum, gold and palladium, improved results are sometimes
obtained with glass which has been stored when, after cleaning, it is
immersed in 5 per cent acetic acid (cold) for six hours. The special
preparation is then applied. If necessary it is mixed before use with a
suitable thinner. The preparation is allowed to dry; this may be
accelerated by warm air. Dust should not be allowed to settle on the
surface before firing.
After drying, the coated glass is put in a furnace below 100°C and
then slowly raised to the firing temperature. The exact firing tem-
perature is best found by experiment, but data provided by Johnson,
Matthey & Co. give good guidance. A temperature 20°C below the
softening point (Mg point) of the glass is often used. The furnace is
kept at the firing temperature for about 20 minutes, and well venti-
lated so that the organic constituents burn away. With palladium,
however, oxidation of the metal occurs at 400°C and the supply of air
to the furnace should be restricted during the final stages of the firing
operation. The furnace can be cooled at any rate which does not
introduce permanent strain into the glass (p. 19). The cooling to the
lower annealing temperature of the glass should therefore be at the
same rate as that for annealing the glass, or at a lower rate.
Some further details concerning the particular metals are given
below.
Silver
The silver pastes give a thickness of silver of about 25-30 microns
after a single application and firing. Successive applications and
firings can be made up to a certain limit. Silver Paste Y.40 gives a
thick and strongly adherent silver film. One application of Silver
Paste X.351 followed by one of X¥.353 gives a thick, matt, silver
deposit. Several other silver pastes are made by Johnson, Matthey
& Co. Silver Paste O.182 is for mounting quartz crystals. Their
liquid burnish silvers are much less viscous than the pastes, and with
one application give a silver film about 3 microns thick.
The fired-on silver films on glass can be soft-soldered. They should
be ‘tinned’ with a solder containing some silver; ordinary soft solder
can bring about complete dissolution of the silver film. Matthey
Ceramic Solder No. 1 is a suitable solder for the films. A resin-type
flux should be used. Tinning can be carried out by dipping into the
molten solder.
88
 THE FIRING ON OF METAL LAYERS ON GLASS

Platinum, Gold and Palladium

The films of these metals are much thinner than those of silver. A
single application of Liguid Bright Gold (12 per cent), followed by
firing, gives a gold layer of about 0-12 microns thick ;one application
of Liquid Bright Platinum gives a layer of about 0-08 microns; one
application of Liquid Matt Platinum a layer of about 0-35 microns;
and one application of Platinum Paste gives a layer of about 2
microns. These are too thin for soldering, except with considerable
difficulty. Thicker films can be built up by successive applications
and firings. The Liguid Bright Platinum preparations give brilliant
mirror-like surfaces on glass; the films deposited by the Liquid
Bright Golds are useful as optical reflectors and filters. Films of gold
on glass can be prepared that reflect nearly all the infra-red radiation
and transmit most ofthe visible spectrum.

REFERENCES
BAKER, E. A., 1931, J. sci. Instrum., 8, 392.
Boas, B. D., 1921, A Handbook of Laboratory Glass-Blowing;
London, Routledge.
Boys, C. V., 1927, J. sci. Instrum., 4, 299.
Brown, B., 1927, J. sci. Instrum., 4, 205.
CHATTAWAY, F. D., 1908, Proc. roy. Soc., A80, 88.
DoNnAL, J. S., 1942, Rey. sci. Instrum., 13, 266.
FARADAY, M., 1842, Chemical Manipulation, 3rd Edn; London,
Murray.
FRENCH, E. A. H., 1924, Trans. opt. Soc., Lond., 25, 229.
GARDNER, I. C. and Case, F. A., 1931, Bureau of Standards Circular
389; Washington, U.S. Government Printing Office.
GRODZINSKI, P., 1953, Laboratory Practice, 2, 356.
HEATLEY, N. G., 1938, J. sci. Instrum., 15, 340.
Lioyp, J. T., 1949, J. sci. Instrum., 26, 126.
OCKENDEN, F. E. J., 1938, J. sci. Instrum., 15, 206.
ROSENBLUM, S. and WALEN, R., 1945, J. sci. Instrum., 22, 196.
Scott, R. B., Cook, J. W. and BRICKWEDDE, F. G., 1931, Bur. Stand.
J. Res., Wash, Te 930.
Seairn, 1. C..P., 1947, Chem. Age, Lond., 298.
Upton, P. B. G. and HERINGTON, E. F. G., 1950, Research, Lond., 3,
289.
WILKINSON, R. A., 1955, Chem. & Ind. (Rev.), 1426.
YaRwoop, J., 1955, High Vacuum Technique ;London, Chapman &
Hall.

89
 Chapter 5

THE MANIPULATION OF LARGE

TUBING

General Observations
WE apply the term large tubing to tubing with an external diameter of
about 35 mm and greater. The problems encountered in working this
size of tubing differ from those connected with the working of smaller
tubes. These problems usually arise from either the greater weight of
the tubing or from the greater heat losses from the site of working by
conduction along the glass and by radiation from its surface. The
weight poses no problem that practice in manipulation and the use of
supports or rollers cannot overcome, but the increased heat losses
cause the tube to become too hot to hold for a considerable distance
along its length, and the radiation from the large mass ofhot glass may
prevent the tubing from being manipulated with unprotected hands.
The obvious solution to the above problems is to protect the hands
with asbestos gloves, but even this has its disadvantages. Asbestos
gloves prevent the same degree of manipulative skill as can be
attained without them, which increases the difficulty of working the
heavy tubing, and if the radiation is great the air inside the glove may
become unbearably hot. For the amateur glass-blower it is frequently
more convenient to clamp the tubing and carry out the work with
hand torches.
Many operations with large tubing can be performed easily with a
glass-working machine (Chapter 6). The following pages describe
some methods for hand-working which are applicable to hard boro-
silicate glass tubes, e.g. Pyrex and Phoenix.

Bending Big Tubing

Method for the More Skilled Worker—Shrinking down a Short Portion
Bends cannot be made in tubing of greater diameter than 12 mm
without blowing. Skilled workers can bend tubing up to about
25-30 mm diameter by heating it with rotation over a length equal to
about twice the diameter, shrinking it over this length to about half
90
 BENDING BIG TUBING
its diameter until the glass is beginning to sag, removing it from the
flame, bending and blowing it out with one puff. The unskilled glass-
blower has, however, to rely on heating and bending a short portion
at a time.

Making Sharp Bends and Bending a Portion at a Time

A good method of producing the desired result is that of sealing two
separate pieces of glass held at the required angle, or, in effect, mak-
ing a T-seal at the end of a piece of tubing. This, in fact, is the only
way, for the unskilled worker, of producing good sharp bends in
tubing greater than 15—20 mm diameter.
The method of heating and bending a portion at a time will give a
satisfactory bend whose radius, for the outside wall of the bend, is
between five and six times that of the tube. The method has been
used satisfactorily on tubes up to 50 mm diameter. The tube, closed
at one end and with a blowing tube attached at the other, is clamped
horizontally near one end, and heated, where the bend is required to
start, over a length equal to about twice the tube diameter, with a
large but not too hot flame from the hand torch. As the glass heats
up it will begin to bend under its own weight, and the free end is
supported until the glass reaches a dull red heat when it is allowed to
fall a short way but prevented from sagging too far. The glass on the
outside of the bend tends to flatten and it is heated fairly strongly
with a hotter flame and blown out evenly. Kinks on the inside of the
bend are heated even more strongly, first with a pin-point flame until
they shrink in, and then with a larger flame which enables them to be
blown out smoothly. The clamp is then rotated in its boss until the
free end of the tube is again horizontal, and the procedure is repeated
on the neighbouring portion and so on until the required bend is
obtained.
With tubing up to about 30 mm in diameter the whole bend can be
annealed by flame at the finish, but above this diameter it is advisable
to anneal each portion for a few minutes after it is bent and shaped,
and before one moves on to the next portion. With extra large tubing
it is better to have two workers, one on either side of the tube, each
with a hand torch.
Drawing Off Large Tubes
The need for a strong spindle is avoided if the end of the tube is to be
‘blown out to make ajoint or if it is to be drawn off by way of cutting
the tube. In this case the tube may be drawn off under less well
controlled conditions than if a strong spindle is required. It is
possible to clamp the tube, preferably in a vertical position, and work
9]
 THE MANIPULATION OF LARGE TUBING
the glass with hand torches. It may be necessary for two persons to
carry out this operation so that the lower part of the tube can be
supported to prevent too rapid drawing off as it falls under gravity.
The lower part of the tube will remain much cooler than if the glass
were worked in a horizontal position.
Drawing large diameter tubing off to spindles can be accomplished
by a simple extension of the methods previously described. It is
almost essential to support the tube on a stand or set of rollers since
the great weight of a relatively long length of tubing will prevent
accurate rotation of the two pieces. It may also be found that the
normal bench torch does not produce a hot enough flame for the
purpose. If this is the case it is possible to arrange two hand torches
in laboratory clamps so that the tubing is rotated between them.
Then, if the tubing is not drawn down too rapidly, a strong spindle
will be produced. Asbestos gloves may be necessary to protect the
hands from the hot glass and the radiation if the tube is to be kept in
the flame for long enough to collapse down sufficiently to form a
strong spindle.
Cutting a length of tubing can be a useful initial stage in putting-in
a strong spindle. This reduces an unwieldy length of heavy tubing to
a more manageable size which can be held in one hand while a
handle is put in preparatory to drawing off the end to a strong
spindle. If an uneven end has been produced by faulty cutting and it
is not intended to draw the tube off, the end may be tidied up con-
siderably by local working with a hand torch and the use of forceps
and scissors for removing excess glass. Any glass so worked should
be annealed carefully to prevent cracking on cooling.

Closing Large Flasks

The necks of large flasks (about 5 litres capacity) are usually of thick
glass, and they must be heated slowly. Before the top is closed a side
arm must be put on the flask, and the neck is usually the best place
for it. The mouth of the flask is closed with a rubber stopper wrapped
in asbestos paper and carrying a blowing tube. A considerable area of
the neck of the flask is slowly heated, using initially a large smoky flame
and finally a hot flame. When the flame is tinged yellow a hole is
blown in the side of the neck, of the right size to fit the side arm. The
end of the side arm tube should previously have been thickened by
heating, flanging and running down, otherwise a bad joint will be
made because the thin, side arm glass runs so much more easily than
the thick glass of the flask neck. The side arm is joined to the neck by
the normal T-piece procedure, and the neck ali round the joint is
annealed carefully. The top of the neck is then heated and pushed in
a2
 CLOSING LARGE FLASKS
with a spike until a handle can be attached as in FIGURE a2 ed ee
neck is pulled off near the top, giving a spindle which will probably be
very thin (FIGURE 32, JJ). This is removed and the end rounded by
running down and blowing out a number of times—to blow out all at
once requires a very large flame, and the end must be blown out piece
by piece with all ordinary flames.

[ in
I II
FiGuRE 32. Closing the neck of
a large flask

Joining Big Tubes

Once again the methods previously described for joining two tubes .
are applicable provided the limitations imposed by the size of the
tubing are borne in mind. Joining two large diameter (say 4 cm)
tubes in the flame of a bench torch follows the same course as joining
smaller tubes by the same method. Care must be taken to see that the
tubes are long enough not to become too hot to hold. Tubing of
4 cm diameter may become too hot to hold for about ten inches on
either side of the site of working, since not only is more heat con-
ducted along the glass but also a much larger flame is necessary.
Working large diameter tubing clamped is very often more con-
venient than attempting to work it with a bench torch. Joining two
large diameter tubes by this method follows the same pattern as the
joining of smaller tubes in similar circumstances: first an adequate
supply of tools and thin glass rod to meet any emergency is collected.
Then the two tubes should be prepared by blowing out the ends. It is
best to blow holes of only about three quarters of the diameter of the
tube (FIGURE 33) as the ends of the drawn-off tubes are liable to have
‘rather thin walls and this allows them to be thickened as the joint is
made. The tubes are heated and pushed together as before, or, if this
is not possible, the flanges of glass left by blowing out the ends are
heated and pushed together with a spike, or the gap is filled with glass
93
 THE MANIPULATION OF LARGE TUBING
rod. While this is being done it is necessary to heat the joint more or
less evenly and not just on one side. If this is difficult two operators,
each with a hand torch, may be found an advantage.
All other operations—the working of the glass by alternate collaps-
ing and blowing, sealing pinholes, and finally blowing the joint to an
acceptable shape—are the same as previously described. The flow of
glass under gravity will be more noticeable when large tubes are
joined, principally because the time required to make the joint is

ma
greater. This flow can be combated as before.

Hh
Ficure 33. Joining large tubing

The final annealing of the joint and adjacent tubing is of the utmost
importance with large tubes. If this is neglected there is a high
probability that the joint will crack as a result of the strains induced
by working and cooling the glass.

Big T-Joints
Large sized T-joints can be made using a bench torch and manipulat-
ing the glass in precisely the same way as previously described. This
is difficult since the whole weight of the glass must be carried by the
hands, because the way in which the tubing must be turned in
the
flame precludes the use of supports. The simplest way of making
these joints is to clamp both tubes in such a way that they
can be
pushed together at the required time.
The two parts should be prepared by blowing a hole in the
end of
the side arm smaller than the bore of the tube and then
blowing a
similar sized hole in the cross-piece. These parts should
next be
clamped so that the tubes are nearly touching and can
readily be
94
 BIG T-JOINTS
pushed together. The two tubes are then heated as uniformly as
possible around the join, and when the glass has become soft the two
tubes are pushed together. If they do not make contact at any point
it may be possible to push the glass in this region together with a
spike or else to bridge the gap with thin rod. The joint is then
worked, by the usual method, to fuse the two tubes together. When
the joint is blown out to its final form it is advisable to heat a fairly
large segment of it at a time ;and blowing the glass carefully while it is
in the flame very often leads to a better finish. Mending any pinholes
in the joint is much more easily accomplished at this stage and there-
fore a careful inspection should be made before the final annealing,
which should be carried out thoroughly. For work on very large
diameter tubing it may be necessary to use two hand torches for the
above operations.
Very often it is found more convenient to replace a right-angle
bend by a T-joint made right at the end of the drawn-off cross-piece.
This makes a very neat right-angle join in tubing which has too large
a diameter to bend easily.

Big Internal Seals ”

Big internal seals can be made by any of the methods described in the
previous chapter although by far the easiest method is to work with
the glass clamped and to keep the internal tube in one piece.
Preparation of the materials is one of the most important phases of
the operation, since once the glass-blowing is started it is virtually
impossible to leave the work until it is finished and annealed thor-
oughly. The inner tube should be prepared as before with a slight
ridge round the tube at the position of the joint. The outer tube
should then be closed off to a round end and the hole to take the
internal tube blown. This should be left with a slight rim of glass and
not flame-polished. Before this can cool the internal tube should be
inserted and clamped in position and work started on the joint. If
the glass is not in good contact initially it may be heated and pushed
on to the internal tube with a spike. Care should be taken in so doing
to see that the flame is not concentrated too much on one side of the
partly made joint, allowing the other side to cool and crack—one of
the dangers of pushing the glass in with a tool. The attention is con-
centrated on the localized area being worked and the rest of the joint
is allowed to cool, thus straining the glass excessively. The localized
. working itself tends to strain the joint, but if care is taken to see that
the whole joint is kept hot the strain is dissipated by a continuous
annealing process.
Once the two tubes have been pushed into close contact and the
bs
 THE MANIPULATION OF LARGE TUBING

joint is air-tight it can be finished as previously described. These

internal seals require very thorough annealing and it is best to heat
them to a dull red heat for a short while before cooling slowly in the
flame. The use of two torches facilitates uniform heating during the
annealing process, and since no other operations are necessary one
can be held in each hand if they are both prepared before the com-
mencement of the glass-blowing. This use of two flames is probably
unnecessary except for very large tubing.

Mending Cracks in Large Tubing

When large tubing has been worked and has been annealed in a flame
it is usually strained to a considerable extent, which sometimes causes
it to crack. Mending these cracks without spreading them is an
operation requiring a considerable amount of patience. It is one of
the cases in which it is necessary ‘to make haste slowly’. The use of
two flames is an advantage as they allow the heating up of both ends
of the crack at the same time. The glass should initially be warmed
uniformly and slowly with small cool flames. When it has started
to heat up, the rate of heating is increased, concentrating the flame
slightly on the ends of the crack so that these parts warm up just a
little faster than the rest of the glass. Once the glass has heated up
sufficiently to flow slowly the crack is allowed to seal itself, using
only a cool flame as previously described. Finally, when a complete
repair has been effected, the glass is annealed.

Mending Cracks near Complex and Large Seals

The above remarks apply equally here, except that extreme care is
needed to prevent the crack spreading into the complex region. If the
crack spreads to an internal seal and the inner tube cracks, it will be
virtually irreparable. This spreading can be prevented if the whole of
the complex joint is heated at the same time and at the same rate as
the glass around the crack. The heating must be slow to allow time
for the inner parts of the joint to be heated adequately by conduction.

96
 Chapter 6

SOME OPERATIONS WITH A GLASS-

WORKING MACHINE

WE have mentioned (p. 2) the difficulty of holding and rotating a

piece of glass tube in each hand and operating on a semi-molten mass
between the two pieces. A very satisfactory way of resolving this
problem is to use a machine—the Edwards Model G 3 Glass-Working
Machine (made by Edwards High Vacuum Ltd) is of great value. The
glass to be worked is held in two chucks which grip the tubing and
rotate it synchronously at a speed which is variable over a considerable
range (120-500 rev/min). One of the chucks can be moved towards or
away from the other. The glass may also be held in only one chuck.
The glass can be worked both by blowing into it as it rotates and
by the use of tools which can be mounted in various ways or held in
the hand. Thus operations involving the production of articles with
circular symmetry can be carried out with great ease, and a skilled
operator is by no means necessary. The advantages of using this
machine, when possible, for work involving tubing exceeding 3-4 cm
in diameter are, we think, particularly manifest. The operation of
putting a spindle in a large tube, for example, is done by hand only
with difficulty, whereas with the machine it can be done rapidly and
effectively. In this Chapter we attempt to give merely a general idea
of how such a machine can be used; the detailed manipulations in-
volyed soon become clear after a little practice. The machine is
valuable either for the production of special parts in research and
development work or for continuous operation on repetition work by
unskilled personnel.

General Description
Two views of the glass-working machine are shown in FIGURES 34
and 35. It consists of a substantial cast base fitted with a fixed head-
.stock and a movable tailstock. This can be moved towards or away
from the headstock by operating the tailstock traverse control.
Rotating mandrels in the headstock and tailstock are synchronously
driven; the tailstock mandrel drive can be disconnected and locked
97
 WITH A GLASS-WORKING MACHINE
SOME OPERATIONS
can then be
in position so that the mandrel does not rotate, and it
lengths of
employed as a tool holder for tools used to work single
ds
tubing held in the headstock chuck. The tool can be moved towar
or away from the glass using the tailstock traver se as usual.
In the front of the machine there is a burner mounting bar which
or
normally carries either a cannon burner, operating with gas and air
re burner, operat ing with gas—
gas—air—oxygen mixtures, or a crossfi
oxygen mixtures. Both burners can be used together and their
flames directed on the same region. The burner mounting bar has a

Too/ mounting
bar
Flot forming
foo/

Cannon
Blowing burner

fixture

pulley
Connon burner Gx
Jets eS
Hand fool
hand/e

7 Air Universa/
Supplies 4 Gas bed shield clamp
Oxygen

supply points
Variable
speed contro/
Ficure 34. A glass-working machine

-rack-and-pinion mechanism so that the burner or burners can be

traversed a limited distance along the glass and be made to follow up
the work. |
At the rear of the machine there is a tool mounting bar. A flat
forming tool can be fixed to this, as shown in FIGURE 34; and by
using the operating lever (FIGURE 35), the tool is brought into contact
with the lower side of the rotating tubing, which, softened in a flame
is pushed into shape. A socket for a cone-and-socket joint is easily
made by setting up the flat forming tool at the proper angle. The
tool mounting bar.is fitted with an adjustable mechanical stop which
98
 GENERAL DESCRIPTION
is also an electrical contact for giving visual indication on a lamp
when pre-set limits are reached. This arrangement is useful for repeti-
tion work. Generally, when constructing specialized apparatus for
research work, operations with tools are performed by holding them
in the tailstock chuck (not rotating), or in the hand tool holder
(FIGURE 34), which is held in the hand.
The mandrel bore size is 32 mm, so that tubing of this size will pass
right through the headstock and tailstock. The maximum jaw open-
ing is 63 mm with the jaws in their normal position; but they are

Headstock
sree Mandre/
flange Flanging tool
Standard
chuck
Tailstock Q
Operating lever

Mechanical
stop and contacts

Burner bar traverse Tailstock lock

contro!

Tailstock Burner mounting bar

clamp Tailstock
traverse Clutch
contro!
Ficure 35. Another view of the glass-working machine

reversible, and by reversing them tubing up to 115 mm diameter may

be held.
There are various ways of blowing into the rotating work. If the
tube, or a portion of it, is long enough and less than 32 mm diameter
it may be passed right through the headstock and blown into using a
glass-blower’s swivel (p. 42). Otherwise spindles or rubber stoppers
carrying glass tubes may be used in conjunction with a swivel. The
machine itself can be used to put a short spindle in a tube, using, in
principle, the method described on p. 56. A blowing fixture can also
be fitted to the headstock or tailstock mandrel (see FIGURE 34); this
H 99
 SOME OPERATIONS WITH A GLASS-WORKING MACHINE

remains stationary while the work rotates, and so it is only necessary

to prepare a sufficiently air-tight joint between the tubing to be used
and the mandrel with the help of tape or rubber stoppers, or stoppers
and tape together.
Sketch of Some Possible Operations
A tube held in the headstock chuck can be flared by any of the
methods previously described (p. 72). Flares can be made very easily
with a cylindrical carbon rod attached to a special holder fitting into
the tailstock chuck, which is not rotated during the operation. The
flare is made by pushing the carbon rod, held at the desired angle,
into the end of the rotating and heated glass tube, for which purpose
the tailstock traverse is used. With this method the end of the tube
can be opened out until it resembles a funnel.
To put a spindle in a tube (p. 57), both ends are held in the
rotating chucks. It may be desirable to straighten the tube, by heat-
ing it in a gas—air flame as it rotates, if it is so bent that it is not held
well in both chucks. A spindle is put into the tube by the method
described on p. 56, except that the operator’s hands are replaced by
the two chucks. The length of the spindle which can be obtained is
limited by the length of the tailstock traverse.
A tube with a spindle in it can be worked to a tube with a round
end by removing the spindle just beyond the shoulder (see p. 57) with
a small flame, and then shrinking down and blowing out the end.
The blowing can be done while the glass is in the flame. A hole can
be blown in the end of the closed tube by heating the extreme end in a
small flame. The hole can then be opened out with a tool, using an
ordinary flaring method. Two tubes with holes of the same size in
their ends can be joined very easily by bringing their hot ends to-
gether as they are both rotated, one tube being in each chuck. The
joint is finished as described on p. 61. To join big tubes it may be
convenient to grind the two ends flat; no holes need then be blown,
since the tubes will meet well enough.
To blow a bulb in the middle of a tube, it is held in both chucks;
glass is accumulated by pushing the tailstock to the headstock as the
glass rotates, and the tube is blown into gently at the same time. The
bulb is blown as the glass accumulates; an elegant bulb can be
obtained very easily.
An internal seal can be made by Method | of p. 73. The hole in the
outer tube and the ridge or bulge on the inner tube are both put in on
the machine. The ridge can be put in by heating a small portion of
the tube as it rotates in both chucks, and quickly pushing one end
towards the other with the tailstock traverse. Alternatively, a small
100
 SKETCH OF SOME POSSIBLE OPERATIONS
bulge can be blown until it is the required size. This is conveniently
done with the help of a calliper gauge. When making the final inter-
nal seal, the inner tube of the seal can be supported in the outer tube
with corrugated cardboard. An internal seal can also be made by
Method 4 of p. 76. The correct size flange or flare is put on the inner
tube with a tool, and a calliper gauge is again used to check the size
of the hot glass.
A Dewar seal (p. 122) can be made without blowing by flaring the
smaller tube outwards until it will nicely meet the larger tube (as in
FiGure 50, //I). The smaller tube is mounted in the larger tube with
the help of corrugated cardboard, and the final joint is made by
pushing the flare of the inner tube on to the outer tube with a tool as
the glass rotates in the flame. If the tubes are first ground flat at the
ends to be sealed, an elegant joint can be obtained very easily.
Graded seals must be used for joining together glasses of such
different thermal expansions that a direct seal cannot be made
(p. 25). The best way of making a graded seal with the machine is to
use the various intermediate glasses as rod. The first piece of glass
tubing is held in the headstock chuck, and the first intermediate glass
is held in the tailstock chuck ; both are rotated. The end of the tube
is closed with the rod, which is then drawn off just beyond the closure.
The closed tube remaining then has its end worked to a rounded
shape. The next glass is attached as a globule on the extremity of the
hemispherical end and is worked to another hemispherical end; the
process is repeated with the next intermediate glass, and so on. The
final tube is joined as a tube.
A graded seal can also be made by fusing together rings of the
intermediate glasses, or by fusing on each intermediate glass in the
form of tubing, then drawing a spindle, making a round end, and
blowing a hole.
The making of glass bellows is an operation which is very greatly
facilitated by the machine. The glass is rotated in a small hot flame
and blown out to give a thin protruding portion of glass. More of
these thin portions are then introduced. Practice is necessary to
make bellows with the required degree of flexibility. The flexibility of
course increases with the number of thin portions of glass; the
fragility likewise increases.
Tapering can be carried out by the methods described on p. 79.
When a precise angle is required the flat forming tool on the tool
mounting bar (FIGURE 34) can be used.

101
 Chapter 7

METAL-TO-GLASS SEALS

ALTHOUGH the sealing of noble metals through glass has been known
for a great many years, even centuries, the extensive use of metal—to—
glass seals in laboratories has developed only in recent years. This
recent development has, however, been extremely rapid and a great
deal of detailed literature now exists. An excellent collation of much
of the data has been made by J. H. PARTRIDGE (1949) in his mono-
graph devoted entirely to glass-to—metal seals. We, in this Chapter,
attempt to supply the reader with information and detail sufficient
to be useful to the average laboratory experimentalist.
Matched and Mismatched Seals
Where a direct fusion of glass-to—metal or metal oxide is obtained
(sometimes termed the “wetting” of metal by glass) two types of seal
can result—one matched, the other mismatched, depending on how
the relative coefficients of thermal expansion of the glass and the
metal compare.
In a matched seal the stress is kept within safe limits by close
correspondence of the two thermal expansions throughout the
temperature range over which the seal has to be worked during its
manufacture.
A mismatched seal is one in which the glass and metal have differ-
ing coefficients of expansion. Here the resultant stress is reduced to
within the limits required to avoid fracture by special shaping of the
metal at the junction. In some cases relief to a small extent is
obtained by the use of a ductile metal which gives slightly, so reduc-
ing the stress in the glass.
When it is required to seal a metal into a glass with which its co-
efficient of thermal expansion is greatly mismatched, graded seals are
often employed from the given glass, through one or more inter-
mediate glasses, to a glass which will form a matched seal with the
given metal. Kovar-to—Pyrex seals (see below) are a good example
of this type.
A further combined characteristic that a glass and a metal must have
before they will form a seal, over and above that of the provision
102
 MATCHED AND MISMATCHED SEALS
by some property or other for the relief of stress in the glass, is that
the glass must wet and adhere to the metal or its oxide.
Another important consideration that must be given when em-
ploying a matched seal in the laboratory is the temperature range to
be encountered during use of the seal. The absence of stress relies on
the matching of thermal expansions throughout the temperature
range encountered when the seal is worked, and this matching must
continue over the temperature range in which the seal is to be used, else
fracture may result. This factor must be watched when, for instance,
the metal—to—glass seal is to be immersed in, or is near to, refrigerants.
Sealing Tungsten into Pyrex
Pyrex (Corning No. 7740) adheres well to tungsten oxides and will
give a vacuum-tight seal with this metal provided the joint is made
carefully. The seal is mismatched and an upper limit to the diameter
of tungsten rod that will give a satisfactory joint is 0-5-1 mm. The
exact figure varies ; the most important factor is probably the severity
of the service requirements.
For vacuum work tungsten rod is superior to wire because its
method of manufacture leaves the metal with less surface cavities and
less fine channels running within it. Centreless-ground tungsten rod
is advised. The rod is first beaded, and for this a small length of well-
fitting, thin-walled, Pyrex capillary is drawn; a piece about 1 cm long
and approximately 0-5 mm wall thickness is required. The portion of
the tungsten rod to be beaded is cleaned by heating to dull red heat
and bringing it immediately into contact with crystalline potassium
or sodium nitrite, either in the form of a stick or as a small heap of
powder balanced on a spatula. A strongly exothermic reaction
occurs during which the metal continues to glow red hot and is
rapidly dissolved. The reaction is allowed to proceed for a few
seconds, then the rod is washed free of the nitrite and wiped dry. The
surface of the metal is left clean. It is then reoxidized slightly in a
just non-luminous flame until it appears coloured a second order
green or blue. .
The piece of glass capillary is immediately slipped over the pre-
pared portion of the rod (FiGurE 36, 7) and heated progressivel y from
one end to the other. The rod must not be allowed to cool before
this is done or the oxide layer may break away from the surface of the
metal due to differing thermal expansion coefficients. As the glass
shrinks on to the metal and adheres to it the air is slowly driven out
ahead of the shrinking portion (FIGURE 36, //). The metal, when in
contact with the glass, appears larger in diameter and assumes a
reddish colour so that the progress of the beading is easily observed.
103
 METAL-TO-GLASS SEALS

The finished bead, shown in Ficure 36, JI/, will, when made cor-
rectly, be coloured something between a straw yellow and a reddish
tan. If desired, the bead may be enlarged at one end or in the middle
to assist in fusing it into a small bore tube as indicated in FIGURE 36,

Pyrex capillary
——— .

(eS ae

Prepared portion of rod ere

ff IT

“MWY OSS

IT IV
FIGURE 36. Tungsten—to—Pyrex seals

IV. A hole is made at the end of the tube, just large enough for the
bead to fit closely, and the end of the tube is fused on to the bead with
a hot pin-point flame. The resultant seal should be annealed well and
is usually a shade darker in colour than the initial bead. If the
tungsten wire or rod has to be cut to length, this is best done by grinding
or dissolving in molten potassium or sodium nitrite. Any attempt to
cut tungsten results in the metal being torn and in the opening of fine
channels which may cause leaks.
Wire Holders
The tungsten rod during the operations described above can, when of
sufficient length, be held in the fingers; but with short pieces of rod

FIGURE 37. Wire holders

some method of gripping is necessary. An ordinary pair of workshop

pliers will suffice, but if the metal-to—glass seals are made frequently
it is convenient to have at hand some form of small wire holder. A
useful tool is seen in FiGure 37, J. A small metal collet head, for
104
 SEALING’ TUNGSTEN INTO PYREX
gripping the wire, is set into a hollow handle—preferably of wood.
Another type of holder, simpler to make, is shown in FIGURE 37, IT.
With wire holders such as these a very even rotation of the wire
during beading is possible.
Sealing Tungsten into Special Glasses
With tungsten rods of diameter 0-5-1 mm it is desirable to use Special
sealing glasses for vacuum work, while with rods larger than 1 mm
diameter this becomes a necessity. Nonex (Corning 7720), uranium
glass (Corning 3320), G.E.C. W.1, B.T.H. C9, Dial 36 and B.T.H. C14,
all form a more matched seal with tungsten than does Pyrex (Corn-
ing 7740). Phoenix glass, which has an expansion coefficient almost
the same as that of Pyrex to which it fuses excellently, seals more
Specia/ Pyrex Special Pyfex
Uf

FiGuRE 38. Tungsten-to—Pyrex seals with sealing glasses. Phoenix can

always replace Pyrex in these seals

satisfactorily to tungsten than does Pyrex. It is therefore very useful

as an intermediate glass for Pyrex—to—tungsten seals with rod up to
about 1-25 mm diameter, and also in making pinch seals.
Of the above special glasses we find that W.J, C9 and Phoenix are
easily worked and give very satisfactory seals. Very great care is
necessary when using Nonex to prevent the lead in the glass from
reducing with resultant blackening; the tip of an oxidizing flame
must be used. The characteristics of sealing glasses are given in
Chapter 2. Those mentioned above seal well to Pyrex, though with
Nonex uranium glass is sometimes used as an intermediate.
When making a seal the tungsten rod is first beaded as with Pyrex.
The bead of the special glass may either be fused directly into Pyrex
(FIGURE 38, /) or first fused into a tube of the same glass and this tube
cut near to the seal and fused to a Pyrex tube (FIGURE 38, IZ).

Multiple Wire Seals

‘When a number of mutually insulated wires have to be carried into a
piece of apparatus through one glass tube, a multiple wire seal is
necessary. There are several methods for achieving this end, and the
one to use depends upon the particular purpose and circumstances.
105
 METAL-TO-GLASS SEALS
Multiple Wire Seals when Special Sealing Glasses are not available
One way of making a multiple seal is indicated in FIGURE 39, JJ. This
method is extremely useful when it is required to seal several tun gsten
rods or wires through Pyrex, and special sealing glasses are not avail-
able. Each of the wires is first beaded with Pyrex as described above.
One end of the tube, into which it is desired to seal the wires, is
rounded off and the smaller tubes are added one at a time. Tubing of
3-5 mm bore is used for these separate sheaths. Using a hand torch,
and employing a fine pin-point flame, it will be found with practice
that making a number of small joints, very close together in this
manner, is quite straightforward. The central sheath should be added

7k vi
FiGuRE 39. Multiple wire seals

first and the remainder positioned around it. As each tube is sealed
in position it can be truncated to the required length as shown in
Ficure 39, Jand closed with a small expendable cork.
When the multiple sheath is completed each cork is removed in
turn and replaced by a beaded rod which is fused into place
after
careful scraping away of any burnt cork. The finished work should
be well annealed. We find the method described here to
be very
serviceable and useful even when beads of a special sealing
glass can
be employed.
Pinch Seals
A higher degree of stress exists in pinch seals than
when the wires are
given separate sheaths. Direct Pyrex—to—tungsten
pinch seals, there-
fore, are not satisfactory. However, serviceabl
e multiple pinch seals
can be made with most of the special glasses
mentioned above. When
using Phoenix glass a special sealing glass for
beading the wires is not
necessary.
The end of the Pyrex tube that is to carry
the wires is heated and
shaped with tongs as shown in FIGURE 40,
Jor II, The desired number
106
 MULTIPLE WIRE SEALS
of wires, beaded with one or other sealing glass, is held in a clamp
of the form shown in Ficure 40, J//, and the latter is supported so
that the wires are introduced to the prepared tube with the beads
enclosed in the pinched section. The whole of the prepared portion of
the tube is then heated strongly and evenly and the glass pinched hard
on to the beads with a pair of tongs. The initial beads should be pre-
pared large enough for neighbouring ones to be touching when the
wires are clamped in position ; this ensures that sufficient of the special
sealing glass is present. During the working of the pinch seal all
trapped air must be carefully squeezed out. With most glasses flame
annealing is adequate.

Le

Ficure 40. Pinch seals

Finished multiple seals of this type are shown in FiGure 40, JV and
V. The outer tube may be sealed directly to the apparatus for which
it is required, or a Dewar seal may first be formed as shown in
FiGure 50, VII, making the multiple seal re-entrant. The re-entrant
form is very useful when, for example, the wires are directly attached
to, and are supporting, electrodes (see p. 157).

Prepared Copper—Tungsten—Nickel Wires

Long lengths of tungsten rod beyond a metal—to-glass seal are often
inconvenient because the metal is so brittle and will not give. Any
lateral strain at the end of a long length of rod or wire can easily
fracture the metal-to—glass seal. For this reason it is often desirable
to use prepared copper-—tungsten-nickel wires. These are obtainable,
or can be made by butt-welding braided copper at one end of a short
piece of tungsten and nickel at the other. Nickel also spot-welds well
‘to tungsten.
These prepared wires are very useful for sealing through Pyrex and
Phoenix. Care must be taken to see that potassium or sodium nitrite
does not come into contact with the metal junctions or the tungsten
107
 METAL-TO-GLASS SEALS

will not be left clean. When beading the tungsten the glass capillary
must be kept short, so that the glass comes only into contact with the
tungsten portion and does not overlap either junction, otherwise
fracture will result.

Nickel Tipping of Tungsten

In the prepared wires described above, the copper and nickel butt-

cam
welds at either end of the tungsten serve to seal up any fine channels
that might be present in this metal and to prevent leaks. If the tung-
sten rod or wire is to be used alone it is sometimes desirable to seal
the ends by welding or fusing a coating of nickel over the tip. A fine

. FiGure 41. Sealing platinum into soda

channel is much more likely in tungsten wire than tungsten rod.

With rod the nickel tipping is generally not required.

Sealing Platinum into Soda or Lead Glass

Platinum has a thermal expansion coefficient of 9 x 10-8/°C, and
therefore seals easily into a number of soft glasses. Fine platinum wire,
of diameter less than about 0-25 mm, can be fused directly into soda
glass to give a vacuum-tight seal. If wire above that diameter is to be
used and the apparatus is required to hold a vacuum, then an inter-
mediate lead glass that will fuse to soft glass should be used; alter-
natively Dial 444 may be used as a sealing glass. Platinum oxidizes
only slightly and thus no great care is needed when fiusing glass to it.
When a wire is to be fused into the wall of a tube without employ-
ing an intermediate glass a small tail is first pulled out (FiGure 41, J).
This is done by heating with a pin-point flame, touching the tiny hot
portion with the tip ofaglass rod (or with the platinum wire itself)
and
drawing out a fine spindle. The spindle is broken off about 2mm
108
 SEALING PLATINUM INTO SODA OR LEAD GLASS
from the wall and should at this point be of diameter just sufficient to
accommodate the wire (FIGURE 41, //). The wire is inserted, the fine
tail is collapsed and fused on to the platinum to form a small bead in
situ, which is heated strongly so that it shrinks back into line with the
walls of the tube (FiGuRE 41, //7 and JV). In this manner the seal is
made without any appreciable thickening of the glass around the
platinum. When the seal is finished the whole circumference of the
tube should be heated at this point and well annealed.

Use of an Intermediate Glass

If an intermediate lead or sealing glass is used then the wire is beaded
separately and a slightly larger tail is drawn out—sufficient for the
whole bead to be inserted. The tail is prepared first so that the hot
bead can be inserted and fused into place without cooling. The seal is
otherwise finished as above; care being taken not to form a lump in
the wall of the tube around the wire. When platinum wire of dia-
meter greater than about 0-5 mm is used it is preferable to add to the
larger tube a small T-piece as a sheath for the beaded wire (FIGURE
41, V), as described above for tungsten—-to—Pyrex seals.

Sealing Platinum into Pyrex

No satisfactory, single, intermediate glass for sealing platinum metal
to Pyrex glass is readily available. Very fine platinum wire and very
thin-walled platinum tubing can, however, be satisfactorily sealed
directly to Pyrex, but such seals are not usually serviceable for
vacuum requirements.

Light Platinum Tube

With thin-walled platinum tubing, of wall thickness up to 0-15 mm,
the glass is fused around the outside of the tube for a length of
5-15 mm; slight blowing may be necessary. The metal in this form
will give sufficiently to take up the stress caused by differing thermal
expansions. If used for current carrying, the capacity of the platinum
can be increased by inserting a suitable wire and soldering it at one
end.

Housekeeper Seal
As described by W. G. HOUSEKEEPER (1923), platinum wire of 0-05
mm diameter can be successfully sealed through Pyrex if the wire is
‘first flattened and the edges filed to a sharp V-shape as shown in
FiGure 42, J. The flattened portion of the wire is fused into a tube
whose end has first been collapsed down until it will just take the
broad part of the wire. Alternatively the wire may be beaded first,
109
 METAL-TO-GLASS SEALS

but in either case a strong heat is required and the glass is pressed into
good contact with the metal with tweezers. The entire section of the
apparatus around the seal must finally be well annealed. The finished
seal is shown in FIGURE 42, LI.

Use of Mercury Cup

If the platinum wire is to be used as an electrode it may be preferable
not to flatten and shape it as in the Housekeeper method, but to use a
mercury cup on the outside of the apparatus (FIGURE 42, J//). The
fine wire is fused into a constricted part of a narrow bore tube and the

4 Th Lt

Ficure 42. Sealing platinum into Pyrex

exposed end, on the outside, is covered with mercury which serves to

seal up any leaks that might develop by the glass drawing away from
the metal. Another wire dipping into the mercury cup can serve for
electrical contact. Clearly this is of no avail for vacuum work, but is
useful in making electrolytic cells.

Use of Platinized Glass Rod

A method for effecting a platinum-—to—Pyrex seal by platinizing thin
glass rod has been reported by G. F. TAYLOR (1929) and described in
detail by C. C. CorFIN (1940). Pyrex rod of 1-2 mm diameter is
coated with metallic platinum by dipping in a platinizing solution,
allowing the excess to drain off, heating uniformly in a smoky flame
until a shiny black film appears and then slowly heating to redness to
burn the film into the glass (see p. 87). Three to four coats are re-
quired and then the platinized rod can be sealed into a Pyrex tube
directly. The platinum layer gives electrical conductivity. Coffin
reports that 10-15 applications will give a layer sufficient to carry
several amps.

Special Alloys for Sealing to Glass

Tungsten and molybdenum, which have been used extensively for
metal-to~glass seals, suffer from the disadvantages of being expensive
110
 SPECIAL ALLOYS FOR SEALING TO GLASS
and difficult to fabricate into rod and of being non-ductile. Platinum
is not so difficult to fabricate, but it is very expensive. An attempt
has been made, therefore, to replace these metals with metal alloys,
and extensive research in recent years has been aimed at find-
ing alloys whose thermal expansion characteristics match those of
glasses in the range from room temperature to the upper annealing
temperature of the glass The expansion coefficient of glass increases
rapidly in the range between the lower and upper annealing tempera-
tures and it has been found that magnetic Kovar-type alloys of nickel,
iron and cobalt, sometimes with a little chromium, have thermal
expansions which behave in this manner. With alloys of only nickel
and iron the increased expansion rate sets in at too low a tempera-
ture; however, such wires coated with copper, forming a Dumet
alloy, are used extensively even to the present day in the lamp indus-
try for sealing to soft glasses. A detailed survey of all the characier-
istics of the various alloys used for sealing to glass has been given by
PARTRIDGE (1949).

Copper-Clad Nickel-Iron Wires for Sealing to Soda Glass

Seals for carrying small currents through soft glass can be made
using Dumet alloy wires. These are drawn from rod which consists of
a core of 43 per cent nickel-iron alloy on which about 30 per cent of
its weight of copper has been electrolytically deposited. Wires up to
0-5 mm can be satisfactorily sealed into soft glasses. The seals are
made in the normal manner but are left not fully annealed so that the
glass remains in a state of compression in the longitudinal direction.
The deposit of copper is sufficient to raise the total thermal expansion
coefficient of the rod, in the radial direction, to the order of that of
soda glass, but full annealing would result in a state of tension in the
glass along the length of the seal.
If the seals are to be used for vacuum work it is advisable to tip the
ends of the drawn wire with nickel in order to seal up any fine
channels that may open between the copper layer and the central
alloy core.

Nickel-Iron—Cobalt Alloys for Sealing with Intermediate Glasses to

Pyrex
A large number of Kovar-type alloys of iron, nickel and cobalt of
slightly varying compositions are obtainable and are marketed under
. different manufacturers’ names. Common examples are Fernico I
(54 per cent Fe, 28 per cent Ni, 18 per cent Co), Fernico II (54 per
cent Fe, 31 per cent Ni, 15 per cent Co), Nicosel, Telcoseal and Nilo.
The last three all have a composition of the order 54 per cent Fe,
111
 METAL-TO-GLASS SEALS

29 per cent Ni, 17 per cent Co and expansion coefficient of

AS5axc10-FeC,
All the above alloys seal well to several special sealing glasses of
which B.T.H. C40, G.E.C. Lemington FCN, Kodial, Corning 7040,
Corning 7052 and Corning 7060 are common. The choice of alloy
and sealing glass depends (other than upon availability) upon the
service requirements. All give very satisfactory seals for normal pur-
poses, the residual strain being slight and negligible; while, for
instance, Fernico IJ and Nicosel (or Telcoseal) give practically com-
pletely strain-free joints with Corning 7060 and G.E.C. FCN respec-
tively.
Ha sealing to Pyrex glass a further intermediate glass is required
in order that expansion coefficients become sufficiently matched. For
example, when sealing B.7.H. C40 or G.E.C. FCN to Pyrex, C9 or
W.1. can be used as intermediates. Another intermediate glass, such
as Dial 43, may be desirable as a sandwich glass between the Kovar-
sealing glasses and the tungsten-sealing glasses. A Kovar to 7052 or
7040 tubular seal can be joined to Pyrex (7740) via Nonex (7720) if
the tubing is less than 1 cm diameter; while for larger tubing 7052
(or 7040) to 7050 to 7070 to 7740 is necessary.
Metal-to-Glass Seals using Kovar Tubing
Metal-to-glass seals with Kovar tubing up to about 8 mm diameter
can be made fairly comfortably by hand, but with tubing larger than
this it becomes increasingly desirable to use a glass-working lathe so
that uniform seals are obtaired. Firstly, square edges on the metal
must be rounded off, and for this machining or filing is advised rather
than grinding, since particles of abrasive are difficult to remove from
the metal and great care must be taken. The metal is degreased and
then cleaned by dipping in a solution of 10 per cent hydrochloric acid
and 10 per cent nitric acid at 70°C, followed by thorough washing and
drying. Touching of the cleaned portion of the metal, before the seal
is made, should be avoided.
The metal should be outgassed by heating in a stream of hydrogen
for 3-4 hours at 900°C, otherwise bubbles will be formed in the glass
and an unsatisfactory seal will result. With small size tubing, less
than 8 mm diameter, this cleaning in hydrogen is not absolutely
necessary but the resulting seal will be less strong hermetically and
leaks may develop. A useful makeshift hydrogen furnace consists of
a silica tube along which hydrogen is passed and which is kept glow-
ing white hot with a bench torch. The hot hydrogen must not flow
into the atmosphere and produce an explosive mixture.
Tubing of the glass to be used should be chosen of diameter
such
112
 SPECIAL ALLOYS FOR SEALING TO GLASS
that the Kovar tubing just slides inside. A short length of cleaned
metal tubing is mounted on a glass rod handle with asbestos tape and
oxidized by heating to dull red heat. It is allowed to cool sli ghtly and
inserted inside the glass tube to a depth of 3 mm. The glass is then
fused directly on to the metal starting from one end so that occluded
air is driven out. The flame should be played on the glass and not on
the metal. The glass tube is finally pressed lightly towards the metal
tube so that it builds up slightly over the inside edge of this tube, the
flame being directed at this portion during the operation (see FIGURE
43, I). Slight blowing may be necessary. The resulting seal is grey in
colour and is annealed in the same manner as a glass-to-glass seal.
With larger tubing a small 3-mm long ring of glass can be fused on to
the metal tube first (FiGuRE 43, //) and the bulged end of the glass
tube sealed to this (FiGuRE 43, I//). Much more satisfactory seals can
be made employing a glass-working lathe, for then the glass can be

Napap Special glass

if

Ficure 43. Sealing Kovar to a special sealing glass

tooled into good contact with the metal and all air carefully squeezed
out. Several types of lathe are available; the machine described in
Chapter 6 is satisfactory for these seals.

Copper—to—Glass Seals
Most glasses adhere strongly to copper, but the coefficient of expan-
sion of the metal (17 x 10~§/°C) is much higher than that of all known
technical glasses, resulting in a greatly mismatched seal. Copper is so
ductile, however, that when specially shaped the metal will yield,
relieving much of the stress and making the formation of satisfactory
copper-to—glass seals possible. The strong adhesion at the glass—
metal interface results in a preference for the metal to give rather than
to separate from the glass. A large contact surface between the glass
-and the metal is therefore provided. W. G. HOUSEKEEPER (1923) has
reported in detail on the methods for sealing ductile metals to glass
with special shaping of the metal at junctions. Oxygen-free high
conductivity copper is the best type of copper for seals.
Copper to soft glass seals are most easily made since the difference
in expansion rates is least here, but satisfactory seals of copper to
Pyrex can be formed with a little more care. A further factor with
Hh
 METAL-TO-GLASS SEALS
hard glasses is that a higher working temperature is required and
therefore the copper is subject to excess oxidation and the seal to a
larger cooling range during which the stresses are set up.
The method for sealing copper wire through any glass is the same
as described for platinum through Pyrex. However, with copper the
metal is very brittle after being worked and the seals are not fre-
quently used. Copper wire can be sealed through glass if it is
flattened to 0:05 mm thickness and the edges feathered. The flattened
portion is cleaned, pre-oxidized by heating to red heat for a few
seconds, allowed to cool, and then given a thin coating of borax.
This is effected by dipping the prepared metal in borax solution and
drying in the hot air above a flame. Borating of the copper prevents
excess oxidation when the seal is worked and is therefore of much
greater importance when employing hard glasses. With soft glasses
it can be dispensed with but this is not advised if the seal is to be used
in vacuum apparatus.
The prepared flattened section is inserted into a glass sleeve shaped
to fit, and the glass is fused on to the metal, pinched with tweezers
and strongly heated. The seal must be carefully annealed. With soft
glasses the copper glass interface is coloured red, while with harder
glasses the colour is more of a yellowish red.
Copper—to—Glass Tubular Seals
Tubular seals can be made by several methods reported by Houss-
KEEPER, Of which three are described here.
In the first the copper tube is machined so that a cone-shaped
flange is formed—30° to the axis of the tube—at the end to be sealed
(FicurE 44, 7). The metal flange is oxidized and while still hot the
molten end of the glass tube is pressed inside the cone and fused into
contact. Care must be taken not to let any glass overlap the end of
the copper tube, so that a seal is finally formed as shown in FIGURE
44, II. Gentle blowing helps to shape the join. That the junction is
bulky is the only disadvantage ofthis type of seal.
In the other two kinds of seal the glass is allowed to overlap the
edge of the copper tube, but this edge must be first machined to a fine
5° taper. After pre-oxidizing this feathered edge by heating to red
heat for a few seconds a well fitting glass tube can be fused to the
outside or to the inside without allowing the oxidized metal to cool.
The glass is bulged slightly over the feathered edge thus affording it
protection. The seal to use (FIGURE 44, JJJ and IV) depends upon
whether the seal is to undergo further heating or cooling in service,
consideration being given to the fact that the copper will attempt to
expand or contract much more than the glass.
114
 COPPER-TO-GLASS SEALS
Since tubular seals can be made in this manner to both soft glass
and hard glass, an intermediate copper tube can afford a very useful
method for making a soda-to—Pyrex seal. If Nonex is available a
stronger join is obtained by using it between the copper and the
Pyrex, because a lower working temperature is required for Nonex
and thus, on cooling the worked seal, the copper has a shorter
temperature range in which to contract at its greater rate.

pe ol
Copper, Glass

LT IV
FiGure 44, Copper-to-glass seals

Disc Seals
Copper disc seals can be used for carrying a wire conductor into a
glass tube: two pieces of glass tubing are flared, each at one end,
and a copper disc is sealed between the two flared ends. A suitable
thickness of disc for tubes of diameter 2 cm is about 0-3 mm.
To make the seal one of the tubes is clamped vertically and the pre-
oxidized disc is placed centrally upon the flare of this tube. Working
round the tube with a hand torch the glass is softened and the disc
fused into place. With the seal still hot, the tube is removed from the
clamp and, rotating by hand in a bench flame, the second glass flare

y.
FiGure 45. Copper-—to-—glass disc seals

is fused to the disc, centrally and directly opposite the first (FIGURE 45,
I)—if desired, the lower tube can be left clamped. Great care must
be taken during these operations to see that no glass runs over and
bridges the edge of the copper disc, otherwise fracture will result. After
annealing, one of the glass tubes may be cut off near to the disc and a
conductor wire may be soldered through the disc (FiGuRE 45, J/).
I 115
 METAL-TO-GLASS SEALS
Glass tubing must be sealed on either side of the disc in this man-
ner so that stress is taken up by the disc contracting and not by it
bending, which would fracture the seal. HOUSEKEEPER reports that a
disc up to 2:5 mm thick has been used but the resulting seal would not
stand heating or cooling.
The copper disc, if thin, will distort fairly easily if there is a circular
groove in it. This stratagem may be found useful if difficulty is
experienced in making a disc seal.

REFERENCES
CoFFIN, C. C., 1940, Canad. J. Res., 18B, 318.
HOUSEKEEPER, W. G., 1923, J. Amer. Inst. elect. Engrs, 42, 954.
PARTRIDGE, J. H., 1949, SLE SE IGE Seals ;Sheffield, Society of
Glass Technology.
TAYLOR, G. F., 1929, J. opt. Soc. Amer., 18,138.

116
 Chapter 8

THE CONSTRUCTION OF SOME

TYPICAL SINGLE PIECES OF
EQUIPMENT
IN this Chapter we describe the construction of various pieces of
apparatus by the methods of glass-working already described. All
the apparatus mentioned below can be made by quite simple methods
and without the use of the expensive items of glass-working equip-
ment such as lathes, ovens, various wheels and so on. Many pieces
of complex-looking apparatus are made merely by a succession of
simple operations.
Ampoules
Ampoules (FIGURE 46, J) are used when a material, such as a bio-
logical fluid, has to be stored in a sealed-off glass vessel. To make an

vf f/f
Ficure 46. Ampoules

ampoule of normal size, a tube about 4 cm long and 1 cm diameter is

prepared with a spindle at each end: a uniform taper is made at one
end (FiGuRE 46, J/); the tube is then heated at A and either rotated
against a metal object or else sucked-in so that a shoulder is formed
as in FiGuRE 46, /—this shoulder is not necessary, however, and the
operation may be omitted. The spindle at the other end of the tube is
117
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

taken off and a test-tube end (i.e. a rounded end) is made. This is run
down by heating in a flame, and either pressed against a carbon plate
or blown out against a carbon plate to form a flat bottom on which
the ampoule can stand.
Special glasses for ampoules are mentioned on p. 24; Monax is also
a good glass for ampoules.
Sealed Tubes for Reactions
Tubes for ‘sealed tube’ reactions, such as the Carius determination of
halogens and sulphur, can be made from Pyrex, Monax or soda glass.
The mechanical strength of the glasses is about the same, but a soda
Carius tube is much more likely to crack as a result of thermal strain
than a Pyrex or Monax one. The Carius tubes are usually made from
tubing of approximately 20-25 mm diameter and 3 mm wall thick-
ness—Pyrex extra heavy tubing of external diameter 22 mm has a
wall thickness of 2-5-4 mm and can be used up to 600°C.
To prepare a Carius tube, a length of tubing is cut off slightly
longer than the furnace into which it is to go. A handle is attached
to one end, which is then drawn off and rounded so that the round
end has about the same thickness as the wall of the tube. The tube
is cleaned, and the necessary chemical substances are introduced—
for a halogen determination silver nitrate is added, using a paper
cylinder so that none of the salt adheres to the walls except at the end
of the tube. Nitric acid is added with a long thistle funnel, and finally
the organic specimen in a small test tube is introduced. The open
end of the tube has to be drawn off to a thick spindle and sealed while
these substances are in the tube, so that the normal method of hold-
ing and rotating a tube horizontally cannot be used.
Two methods for sealing the tube can be used: in one, the tube is
held in one hand at about 45° to the vertical, a glass rod handle is
attached to the open end, and a thick-walled spindle or capillary is
put in, using a bench torch, in such a position that the final sealed
capillary will project about 2 cm from the iron tube which will con-
tain the Carius tube in the furnace. Following this operation, the
capillary is sealed at the proper place before the tube has cooled.
This ensures that the tube is cooling at the moment when it is finally
sealed, so that as the gas inside contracts this final sealing is assisted
by the atmospheric pressure which pushes the glass in. In the other
method the tube is clamped vertically and a handle is attached to the
top, open end. A thick-walled spindle is then put in the tube, using a
hand torch flame played round the glass; if at first the spindle is too
wide and too thin it is slowly thickened by shrinking in a flame which
is not too hot. The end of the capillary is sealed off as before.
118
 SEALED TUBES FOR REACTIONS
The external diameter of the tube and its wall thickness determine
the pressure which it will stand. Failure of a tube is due to the tension
at the surface becoming too great. For a cylinder of internal radius r
and external radius R, and with an internal pressure in the cylinder of
P, the maximum circumferential (or hoop) tension S (see A. MORLEY,
1940) on the inner surface of the cylinder is
R?+r
eae (ea)
Thus, if the maximum safe tension is 1000 lb/in.? (see p. 10), a Carius
tube with 3-mm wall thickness and 2-cm external diameter is suitable
for reactions at 340 lb/in.”. It is necessary to remember, of course,
that the strength of glass under tension is a statistical quantity.

Break-Tip Seals
These seals separate two parts of an apparatus, and they can be
broken by a slight mechanical blow. The simplest type is illustrated
in Ficure 47, //—the tip is broken off by a piece of iron, usually
enclosed in glass, dropping under gravity. This weight is moved and
lifted with a magnet.

Flame

Ficure 47. Break-tip seals

The seal is made by cutting 1-cm diameter tubing to give one piece
with a square end. Near the end of the other piece a spindle is drawn.
The shoulder of the spindle is heated in a small hot flame so that it
thickens, and then, with a fairly cool flame, the spindle beyond the
shoulder is drawn out to a fine capillary about 1 mm diameter. This
-is heated in a very cool flame and bent first one way and then the
other, as in FiGurE 47, IJ. The end is finally sealed off. The tip thus
prepared is inserted into the square end of the other piece of tubing,
as in FiGure 47, J, and a joint is made by directing a small flame at
119
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT
piece.
the piece of tubing with the tip, and not at the square-ended
and then
The joint is run down and blown out a number of times
It is
drawn out so that the outer tube has a constant diameter.
essential to direct the flame only at the tube with the seal on the end
of it, because if the other tube is heated strongly it will collaps e on to
the fine capillary. ©

Distillation Flasks
A round-bottomed flask can be made into a distillation flask simply
by the addition of a side arm, which is added just as a side arm is
added to a big flask (p. 92). The angle of the side arm to the neck is
usually about 75° (FiGurRE 48, J).

75.

fe i

Ficure 48. Distillation flasks and

Claisen flasks

A Claisen flask of the type shown in FiGure 48, JJ, has a wide side
arm sealed into its neck at an angle which, for convenience, is about
45°. Any liquid in the side arm will then readily run back into the
flask. A 45° joint of this kind can be made quite easily with a hand
torch: a round end is put on the side arm, using a bench torch, and
a hole is blown in this round end so that the plane in which the hole
lies is at 45° to the tube axis. The wide side arm is joined to the flask
neck, and the small side arm is then joined on.
The small side arms are very often broken off distillation flasks,
especially in teaching laboratories ;the most common place of break-
age is near or at the actual T-joint, and the neck of the flask is often
cracked. The crack can be repaired easily if the flask is of a boro-
silicate glass, as described on p. 80, and with great care a similar
repair is possible for soda flasks. The remaining stumpy portion of
the side arm is then removed with forceps and scissors, the hole is
filled in with thin rod, and a new hole is blown for a new side arm.
If the remaining part of the old side arm is long enough (over 1 cm)
120
 DISTILLATION FLASKS
it is only necessary to put a spindle in it and then blow a hole so tha
a new portion of tubing can be added.

Distillation Splash Heads

Splash heads are used in distillation to prevent the liquid in the dis-
tillation flask from splashing over into the condenser. Several types
are used, but probably the simplest is the pear-shaped one shown in
FiGureE 49, VJ. A bulb is blown in the middle of a tube of about 1 cm
diameter (FIGURE 49, J) and heated at B in a large flame, with rota-
tion, so that, by blowing and pulling, the pear-shaped bulb of
FiGure 49, IJ, is produced. The tube at A is removed and the top of
the bulb is rotated in the flame—this, with gentle blowing, should

‘ a

v/a I iF VI

Ficure 49. The construction of a distillation splash head

produce a fairly flat-topped bulb (Ficure 49, I/I). The tube to form
the top part of the splash head is heated in a large flame and tapered.
The top of the taper is next heated strongly with a small flame, and
gentle pressure applied at both ends of the tube so that a ridge is
formed (FiGurE 49, JV). The taper is then heated at D in a soft
flame and bent round. The curved part is cut off (most easily in the
flame) about 1 cm beyond the ridge, and the end is fire-polished (i.e.
heated in the flame to smooth it). In the centre of the flat top of the
bulb a hole is blown, just smaller than the size of the ridge; into this
the curved part is inserted, and an internal seal is made either with a
bench torch (Method 1, p. 73), or with a hand torch (p. 76), the
former method is preferable. The final splash head is shown in
Ficure 49, VI.

Dewar Seals
The Dewar seal is important and useful, although a ring seal can
, in
sometimes be substituted in simple apparatus—as, for example
shown in FIGURE 52, 1. In the making
the cold finger liquid air trap
121
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

of a Dewar seal the preparation of the parts is of extreme importance.

Basically, the inside tube is flared at one end, and the end of the out-
side tube is turned in so that with the two tubes coaxial the inward
and outward flares meet as closely as possible.
FicurE 50, J, shows the ideal starting point for obtaining a good
Dewar seal. The flare in the inside tube should be just large enough
for this tube to be held up by its flared end when sitting inside the
clamped outer tube. With small diameter tubing the flange on the
inside tube can be made by simply blowing out the end with the glass

Ks FE VIT VII
FiGure 50. Dewar seals

just slightly less hot than normal. To make the Dewar seal using the
hand torch, the whole joint is warmed and then a portion at a time
worked with a largish flame and enough oxygen to make it just non-
luminous. The glass of the inside flare is pushed down with a spike or
similar tool on to that of the outside tube, gradually working round
the seal, and when blowing becomes possible a pin-point flame is
used for blowing out and shaping a portion at a time.
Welding rod is employed when necessary, but if the parts are
prepared satisfactorily the flares should provide sufficient glass to
alleviate the need for rod. With care it will also be unnecessary to
hold the inside tube in position since only part of the joint is heated to
122
 DEWAR SEALS
softening point at a time and, in the final annealing, if the seal is
made well the inside tube will tend to hang centrally. However, the
inside tube may easily be kept central with asbestos paper or corru-
gated cardboard or with copper wire that can be dissolved out with
nitric acid. In making the seal care must be taken not to collapse the
outer tube on to the inner just below the join; otherwise little difficulty
should be encountered. The finished seal is shown in Ficure 50, II.
If the starting parts are shaped as in FiGure 50, J// or V, the result-
ing seals will be bulged outwards or inwards respectively (FIGURE 50,
IV and VI). These may satisfactorily serve their purpose but require
extra careful annealing. If the flare of the inside tube is turned right
back on itself (FiGuRE 50, V//), then the final joint is most easily
made.
Dewar Vessels
The work which led to the modern vacuum flask has been described
by J. DEWAR (1896). Vacuum jacketed flasks are used extensively in
scientific laboratories and can be constructed relatively easily: tub-
ing is chosen to give a flask of the required size and, after the inside
tube has been rounded off at one end, a Dewar seal to the outer tube
is made at the other end (Ficure 51, J). Ifa flask with a narrowed
neck is required the parts must be prepared as in FiGure 51, I/, before
the Dewar seal is made.
The open end of the outer tube is then drawn off, a 1-cm tube B is
attached, and a 2-mm internal diameter constriction is formed at A as
shown in FiGureE 51, ///. This is best done by rotation in a flame,
but, if the tubing is too large for the operator to manipulate, it can be
performed with the apparatus clamped as described in Chapter 5.
For efficient thermal insulation the inside walls of the jacket should be
silvered, and the silvering is performed at this stage of the construc-
tion by the methods described on p. 85 and in some detail by W. E
Barr and V. J. ANHORN (1949). FiGurE 51, JV, shows a simple
stratagem for filling and emptying the jacket with the cleaning and
silvering solutions respectively. After cleaning and drying, the
silvered vessel is ready for evacuation and sealing off.

Evacuation of Dewar Vessels

(Glass under vacuum slowly releases gas and vapours (p. 16) and
therefore, in order that a good vacuum will be retained in the closed
_ jacket, it is necessary to outgas the evacuated jacket before sealing
off. As shown by I. LANGMurR (1918) and described by Barr and
Anhorn, the most effective outgassing is obtained when the evacua-
tion is carried out in two or more stages of decreasing temperature.
123
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT
The prepared flask is supported so that the tube B can be sealed
to a vacuum system. If a suitable furnace is available the flask is
surrounded by this, with the constriction A still within reach of the
hand torch flame; otherwise the heating must be done by flame. A
liquid air trap is included between the flask and diffusion pump.
With a vacuum of below 10~-§ mm of mercury the temperature of the
flask is raised to near the softening point—the heating may be rapid

BS

L wh IT FIGuRE 51.
Construction of
Dewar flasks

To vacuum pump
ne od

with unsilvered vessels, but must be very slow when the walls have
been silvered. With Pyrex glass (Corning No. 7740) a temperature of
550°C is held for about one hour; the temperature is then reduced to
about 400°C and this held for a further thirty minutes with continu-
ous pumping. With the flask still at about 400°C the glass is slowly
collapsed at the constriction 4, allowing four or five minutes for this
process to enable the heated glass in this region to outgas thoroughly.
The tip on the bottom of the flask is rounded and annealed, and the
flask then allowed to cool as quickly as possible without setting up
Strain.
Both commercial and home made Dewar flasks used in the labora-
tory should be wound along the whole of their length with adhesiv
e
124
 DEWAR VESSELS
tape for safety. This prevents them exploding and shattering in the
event of fracture which sometimes results from strain set up during
pouring refrigerants, particularly liquefied gases, into or out of the flask.
Cold Finger Refrigerant Traps
Cold finger refrigerant traps are often used in vacuum lines as a
substitute for the more efficient total immersion traps which, how-
ever, tend to cut down the pumping speed of the system. In con-

Mti
structing a cold finger trap a Dewar seal is first made, then, before

se & Ficure 52. Construction of cold finger refrigerant

traps

annealing, a small hole is blown out near to the seal (FIGURE 52, I)
and a prepared side arm sealed in position. Care must be taken to
keep the far side of the Dewar seal tube hot while the side arm is
being added, to prevent cracking; finally the whole is annealed. The
finished trap is shown in FiGure 52, //.
Ficure 52, I/I, shows how a quite satisfactory cold finger trap can
be made with the Dewar seal replaced by a straightforward internal
seal. As before, the side arm must be added immediately the internal
seal is completed and the whole annealed together.
For high vacuum work, where it is desirable to shorten the pump-
‘ing path to a minimum, a useful built-in cold trap is shown in
Ficure 52, V. The outer jacket is prepared from a round-bottomed
flask to which the side arms are attached, as shown in FIGURE ave
The inner tube is shaped with a flare at its upper end and its heated
125
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

lower end is thickened and collapsed as for making a bulb (p. 71).
The lump of white hot molten glass is lowered through the top side
arm of the outer jacket and the bulb blown steadily and firmly to the
required size. Contrary to expectation it is found that the hot molten
glass does not adhere to the cold outer tube when it comes into con-
tact with it, provided the lowering operation is performed quickly.
The trap is finished by making a Dewar or ring seal at the upper end.
An unskilled worker will find, with practice, that it is possible to
make small, bulbular, cold finger traps of this type for pumping lines
up to 25-30 mm diameter. Anything larger requires a skilled hand.
With all these cold finger traps the rate of loss of refrigerant is
much greater than from the Dewars whose inner surfaces are silvered
to cut down radiation losses.
Filter Pumps
It is advisable before making a piece of apparatus such as a filter
pump to prepare all the separate components, since the work, if left
whilst a component is made, may crack on re-heating—in fact if soda
glass is used it probably will do so. Hence for a filter pump of the
type shown in Ficure 53, V, the jet, the cup, the side arm and the
tube for the jacket are first made separately.
The jet is made from tubing of about 1 cm diameter: the tubing
is strongly heated and a ridge put in. Another ridge is made about
1-5 cm below the first, and a taper is made 4-5 cm long with the
shoulder about 1 cm below this last ridge. A ridge is then made
at the shoulder of the taper, and the end of the taper is cut off and
fire-polished (FiGuRE 53, /).
The side arm is made from tubing of about 0:5 cm diameter by
drawing a spindle and making a ridge just above the shoulder. The
glass is cut off 1-2 cm from this ridge (FIGURE 53, IJ uh
The jet is sealed into the jacket with an ordinary internal seal
(Method 1, p. 73), and the side arm is attached at the shoulder; the
jet must be perfectly central (FicurE 53, IJ).
The cup is then made: a piece of tubing of the same diameter as
the j*zet (about 1-5 cm) is joined to a tube of the same size as the jet
tubing. A slight taper is made in the smaller tubing, and a ridge made
at the shoulder of the taper. A bulb is blown at the end of the taper.
This is not very easy since the glass is usually very thin at first—th
e
bulb must be blown and shrunk a number of times until enough glass
has been accumulated to blow a strong bulb. It should be 3-5
mm
diameter (FIGURE 53, JV). The end of the bulb is blown
out and
spun (p. 73) to give a thistle funnel shape (B, FiGurer 53,
V). The hole
should be just bigger than the end of the jet,
126
 FILTER PUMPS
The end of the jacket is rounded off so that when the cup is in-
serted the jet will be just inside the cup. The cup is next sealed in
with an internal seal. The larger tube is then rounded off at C
(FiGuRE 53, V), and a side arm is put on at the bottom and bent
round. Lastly, the two side arms are cut off and fire-polished to give
the final pump seen in FiGuRE 53, V.

viv! Fic. 53. The construction of a filter pump

Phosphorus Pentoxide Traps

Water vapour and some organic vapours are readily removed froma
gas stream by phosphorus pentoxide traps. A simple trap is shown
in Ficure’54, J; A is the P,O; boat.
In most cases a tube of 2 cm diameter will suffice for housing the
boat, and to one end of such a piece of tubing is sealed a ground glass
Cc
- socket B. The other end is rounded off at the required length. At
r and
and D two T-junctions are formed to tubing 8-10 mm diamete
is made
these side tubes are bent to the desired shape. A stopper E
Vacuum grease or
from the appropriate standard cone as shown.
127
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

wax is used for sealing the stopper in place. In this way the P,O; boat
can be easily recharged. When building the trap into an apparatus,
the side arms at C and D should point upwards so that any phos-
phoric acid syrup spilt from the boat will not get into the system.
Phosphorus Pentoxide Boats
A satisfactory P,O; boat can be readily made from glass tubing: a
short length of suitable diameter is cut along its length (FIGURE 54,
IT) either on a cutting wheel or by leading two cracks along the
length of the tube on opposite sides—this latter operation is fairly
easy with soda glass. Using one half, each end in turn is heated and

7)
cos
sna
eeee) — 5
8
r€ Hf§
FiGure 54. Phosphorus pentoxide trap

bent upwards, then finally shaped with tweezers to form a closed end
as shown. A pin-point flame is used for sealing the seams. Finally
the long edges are fire-polished and the boat annealed carefully.
A less elegant P,O,; boat can be made much more quickly by the
following method: the glass tube to be used is closed at each end
with a spindle so that the proper length is obtained. A hole is blown
somewhere in the side of the tube, and the glass is strongly heated at
one side of this hole and pulled away with forceps. This process is
continued until the glass has been removed from one whole side of
the tube. The ends are then worked with forceps and tweezers to
give the final boat. This method avoids both the use of a cutting
wheel, and the cracking of a tube along its length, an operation which
does not always go well because the crack may wander in the wrong
direction.
Condensers
The construction of a water-cooled condenser is a good example of a
case where all the parts must be prepared, and at hand, before the
final assembly is begun. There are several methods commonly in
Bh for setting about the construction ;FIGURE 55 shows one of
ese.
Lengths of tubing are chosen of suitable diameter for the inner tube
and the jacket. An internal seal (ring seal) is formed at one end and a
side arm, bulged to take rubber condenser tubing, is immediately
128
 CONDENSERS
added near to the ring seal while the latter is still hot. Thus, after
annealing, the stage shown in FiGure 55, JJ, is reached. Then, at a
point giving the required length of the condenser, the outer tube is
collapsed on to the inner tube and the two are pulled off together

Oe
Ficure 55. A method of making a condenser

(see p. 76) to a spindle (FiGuRE 55, III). Excess glass is pulled away
so that a rounded end is formed. Blowing through the side arm at the
far end prevents the outer tube shrinking on to the inner behind the
ring seal; while blowing down the central tube allows the end of this

M i}
FiGure 56. A method of making
a condenser

to be thinned and finally blown out as shown. A prepared piece is

then immediately added to the central tube (FIGURE 55, JV) followed
by the side arm attached to the water jacket. Careful annealing of
this whole end is required to relieve as much strain as possible from
the apparatus. The finished condenser is shown in FiGureE 55, V. If
129
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT
the outer tube has subsequently to be heated for any further glass
blowing, very careful warming is required because, at either ring seal,
the outer tube heats more quickly than the inner and therefore tends
to draw away from the latter.
A second method of construction is to form the first ring seal and
side arm from the prepared parts shown in FIGuRE 56, J. The inner
tube is slightly flared at the far end from the ring seal. In the second
stage of the construction the outer tube is drawn off so that the
shoulder is very close to this flared end (FIGURE 56, IJ). The shoulder

Se :
FiGure 57. A method of FiGure 58. A condenser with a
making a condenser bulbular inner tube

is shrunk on to the inside flare and, again with blowing through the
far side arm, a ring seal is formed. The condenser is then finished as
before.
In the third method the outside tube is rounded off at one end and
the inside tube prepared with a large flare and a small flare at oppo-
site ends as seen in FiGurE 57, J. A butt ring seal is first formed at the
closed end of the outer tube. This is blown out as shown in FIGURE
55, IV, a piece added to the central tube, and a side arm added to give
the stage represented in FiGure 57, IJ. In the last stage the outer tube
is shrunk on to the inner flare forming a large ring seal and the
second side arm added. This method gives a finished condenser as
shown in FiGure 57, IJ.
If the internal tube is required to be of some special shape, bulbula
r
or in the form of a spiral for example, then the best method
of
approach is probably the first method described above: the
inner
tube is first shaped as required and the outer jacket made to
fit as in
FiGurE 58, J. The ring seal can be formed at one end and
the side
130
 CONDENSERS
arm added to give the stage shown in FiGure 58, JI. Finally both tubes
are drawn off at the far end to give a ring seal and the condenser is
finished as in FiGuRE 55, JV and V.

Fractional Distillation Condensers

A condenser of the type shown in FiGure 59 is often called a frac-
tional distillation condenser and was advocated by H. Jackson,
Professor of Chemistry at King’s College, London—hence the con-
denser is sometimes called a King’s College condenser. It can be used
as an ordinary condenser and as a reflux condenser. To make it, a

Ficure 59. A fractional

distillation condenser

piece of 3-cm diameter tubing is joined to 8-mm diameter tubing by a

taper (D). 1-5-cm diameter tubing, for the water circulation, is
drawn out to a taper at its end, and closed. 8-mm diameter tubing is
prepared for the side arm A and 5 mm tubing for the water inlet and
outlet tubes B and C. The 3 cm tubing is rounded off and a hole
blown just large enough to let the 1-5 cm tube slide into it. A normal
internal seal is next made, but in this case the joint must be blown
from both ends. The side arm A is then added, followed by the side
arm C. The larger tubing to which C is joined is then rounded off,
blowing through C, and the central tube B is inserted in the usual
way, but the internal seal is somewhat more difficult to make since
‘the side arm C is already present. The whole apparatus requires
careful annealing.
Alternatively, the whole water-circulating part can be made separ-
ately and internally sealed into the wide tube, using a hand torch, and
K 131
 TYPICAL SINGLE PIECES OF EQUIPMENT
SOME
seal as in
blowing from both ends. Another design employs a Dewar
a rubber
Ficure 52, //, and the water-carrying tubes are inserted with
stopper.
Fractionating Columns
These are devices for establishing contact between ascending vapour
be
and descending liquid in a distillation. A variety of columns can
made. Perhaps the simplest (FIGURE 60, //) consists of a support ina
tube on which a packing rests, made up of short glass tubes, glass
rings or pieces of metal gauze. The support for this packing is made

wilh Ff i
FiGurE 60. Two types of fractionating column

by heating the tube with a small hot flame and pushing the wall in
with a spike almost to the centre of the tube; this is done a number of
times to give the support shown in Ficure 60, /. A thorough anneal-
ing should be given.
Another fractionating column consists of a number of pear-shaped
bulbs (FIGURE 60, V). A series of bulbs is blown (FiGure 60, ///), and
their ends are heated and blown and drawn so that pear-shaped bulbs
are formed (FIGURE 60, /V). The ends of these bulbs are heated (at A
etc.) and they are gently pushed in. The column should then be
thoroughly annealed. The whole sequence of operations can be
carried out very easily with a glass-working machine.

A Fractional Distillation Receiver

A receiver for low pressure distillations designed by G. A. R. Kon
(1930) is quite easy to make: the joints D and C (FiGurE 61) are
both standard B.19 joints ; the two taps at A and Bare standard designs
of three-way taps. For the part G a tube of about 3 cm diameter
132
 A FRACTIONAL DISTILLATION RECEIVER
is joined at both ends to tube of about 8 mm diameter. One of the
small tubes is joined to the tap B. To prevent distortion of the
barrel of this tap, the joint should be made 1-2 cm away from the
barrel. An internal seal is made into the cone D and a side arm
attached. The top of the internally sealed tube is then joined to the
tap B. The two tubes E and Fare joined to the tap 4. The socket C is
then joined to the larger tube.

J
C

a)
y
Ficure 61. A fractional distillation receiver

An advantage of this receiver is that receiving vessels connected at

D can be taken off and put on without altering the low pressure in the
main apparatus. When tap Bis shut during the operation of changing
a receiver the distillate collects in G.

Double Surface Condensers

A simple form of double surface condenser that can be made without
too great difficulty by an unskilled worker is shown in FIGURE 62, IV.
The parts seen in FIGURE 62, /, must first be prepared.
The ring seal at A is made in the early stages and is partially
annealed. This part of the assembly is then inserted into the prepared
tube B which is held clamped, and, when seated well, is itself clamped
at C (FiGuRE 62, JJ). Using a hand torch and exerting care to warm
133
 SINGLE PIECES OF EQUIPMENT
SOME TYPICAL

up the whole ares and uniformly the ring seal at D is now

area slowlyly an 5 flare of tube B on to the
made by tooling, a little at a time, the glas
C

Va \
Nh [= Bees te
——— G er re

| |
aN Water :

v4 iT Th

y =

| SY

is oh a Water
FiGure 62, A simple form of double surface condenser

bulge of the inner tube. Care must be taken to keep the whole of both
ring seals hot while this is being done; a large flame should therefore
be used and only reduced to a sharp hot flame for the final blowing
out. The water outlet tube FE is added while the whole area is hot.
134
 DOUBLE SURFACE CONDENSERS
This tube can be held by hand and a pin-point flame used to seal it in
position. Every few seconds the flame size should be increased and the
far side of the assembly re-warmed before continuing the sealing of Ein
position. Finally the whole section must be very carefully annealed.
The assembly may now be clamped ready for the final large ring
seal as in FIGURE 62, J//. For this operation, and to a lesser extent for
that described in the previous paragraph, it is advisable to have a
second worker, also with a hand torch, who can keep the far side of
the apparatus hot while one side is being worked. Extra careful
warming up is required at this stage before the large final ring seal can
be made. The warming should be commenced with hot air for about
10 minutes, then with the tips of large luminous flames and finally
with slowly increasing oxygen content. It will be found with’ such a
procedure that the initial double ring seal will survive without crack-
ing. The large ring seal is then worked in the same manner as the seal
D, and, with the entire section kept hot, the side arms F and G are
added. This region with three ring seals needs extremely careful anneal-
ing. We have found, however, that flame-annealing is adequate.
The completion of the outer water jacket, with ring seal at H and
water inlet tube J, is relatively straightforward. The tubes G and C
may be joined with a short piece of condenser tubing, though a more
elegant finish to the condenser is achieved if they are bent and sealed
to give a closed circuit as in FIGURE 62, JV. If this is done the upper
ring seals must be very carefully protected with asbestos paper, but
otherwise the procedure is again straightforward.
The design of this double surface condenser can be changed as
desired ; for example, FIGURE 62, V, shows a double surface reflux con-
denser which may be made by modifying the above general procedure.
The type of double surface condenser most often found in labora-
tories is the ‘Davies improved double surface’ condenser’ which
appeared in 1905, and was designed by J. Davies of A. Gallenkamp &
Co. as a direct development from Thresh’s modification of the Bidet
condenser. All rubber bungs and tubing were eliminated in the
Davies condenser, which has retained its popularity for 50 years. To
make a condenser of this kind the inner water jacket is connected to
the outer one by two internal seals which are made by a slight
modification of Method 2 described on p. 75. The rest of the
assembly is similar to that for an ordinary Liebig condenser.
Mercury Cut-offs
Mercury cut-offs are used in vacuum apparatus instead of taps when
tap grease is undesirable. A frequently used type is shown in FIGURE
63. A two-way tap is attached at the top of a reservoir for mercury.
135
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

The internal seal at A can be made with capillary tube extending right
down into the reservoir, but this necessitates very careful annealing.
It is easier to first join the capillary to normal tubing of equal external
diameter, and then to seal this normal tube into the larger tube. The
top part of the cut-off consists of an internal seal at B and a side arm.
This part is connected to the reservoir by about 70 cm of capillary

A Ee Vacuum
(aan
4)
c——
ee 4/705, phere

Ficure 63. A mercury cut-off

tube, which gives strength and reduces the volume of mercury re-
quired. To use the cut-off the mercury level is raised and lowered
between C and D.
W. E. BARR and V. J. ANHORN (1949) describe a mercury cut-off
with mercury return lines which prevent mercury surging into a
vacuum system. |
McLeod Gauges
The McLeod gauge is one of the oldest instruments for the measure-
ment of vacua and it has remained virtually unchanged since its
introduction in 1874 (H. McLeop, 1874).
A simple bench type McLeod gauge is shown in FiGure 64, JI//.
Before starting to make such a gauge an estimate should be made of
136
 MCLEOD GAUGES
the volume required to give pressure readings in the requir
ed range:
a gauge with a volume of 300 ml and using 0-75-mm bore
capillary
may be used to read pressures of 10-° mm of mercury with reason
able
accuracy, but if relatively high pressures are to be measured it
will be
more satisfactory to use wider bore capillary, say 2mm. This
avoids
difficulties encountered with mercury sticking in the capillary. The
bore of the capillary should be determined accurately by direct
measurement along its length to ensure uniform bore—unless

Ww vacuum

sphe re
\ To atmo

FiGure 64. A McLeod gauge for bench use

precision-bore tubing is used. This precision-bore tubing may be

contaminated by the lubricant used to prevent it sticking to the metal
mandrel used in its manufacture, and consequently great care should
be taken in cleaning it. Stripping the surface with hydrofluoric acid
gives a uniform clean surface if normal cleaning methods fail.
The simplest way of making a McLeod gauge is to make it in two
basic sections which are finally joined together. Since the capillary
’ on both parts of the gauge should be identical in bore and in surface
properties, it is best to make use of a single length of tubing so that
both capillaries are subjected to identical cleaning procedures. The
capillary should be carefully inspected for particles of dirt which may
137
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

not have been removed by the cleaning. FIGURE 64 shows a gauge

is
fitted to the mercury reservoir by a standard ground joint; this
usually the easiest method of assembl y.
The part in Ficure 64, //, can be made either by blowing a bulb of
suitable volume or by using a round-bottomed flask. If the latter is
used a slight bulge should be blown centrally in the bottom so that a
sharp shoulder is avoided at the capillary seal. A 15-20 cm length of
capillary is sealed to this bulge and its upper end left open. This
capillary can then be used as a handle and blowing tube for the subse-
quent work on the neck of the flask. The neck of the flask should then
be drawn down, a smaller tube joined to it and cut off short to form
the inner tube defining the cut-off volume. Alternatively, the cut-off
volume can be defined by sealing a flanged tube into the neck of the
flask, which then does not have to be drawn down. The standard
cone is then joined to the mercury supply tube by a ring seal to the
flared out end of the tube. This part is then joined to the end of the
flask as shown, allowing sufficient room between the two seals for the
attachment of the side arm to the vacuum system. The two side arms
should be inserted while the glass is still hot from working the internal
seals and the whole section annealed thoroughly.
The next step is the calibration of the cut-off volume: this can be
done by filling with water and weighing the volume contained in the
capillary and bulb up to the cut-off point ;for small volumes mercury
is a better medium but becomes unmanageable for larger volumes.
Once a satisfactory volume calibration has been obtained the end of
the capillary is closed to a square end. This can be easily achieved by
drawing down a piece of glass rod until it just fits into the bore of the
capillary and cutting a short length of this rod to plug the end of the
capillary before fusing it. If this is done carefully the capillary will be
closed with a square end and the walls of the tube will not be dis-
torted at the end of the gas column, which introduces uncertainties in
reading this position. It is important to avoid lens effects here if
accurate readings are to be obtained.
The part of the gauge in FiGureE 64, J, is simply constructed by
joining lengths of normal tubing to either end of the length of
capillary, bending these to a suitable angle and blowing out the ends.
One end of this composite tube is then blocked with a stopper, and
the other end joined to the main tube at an angle by a normal T-joint.
A hole is then blown in the main tube for the second T-joint and the
shape of the bends in the tubing is adjusted by gentle heating and
bending until the end is brought into close contact with the hole. This
joint is then completed. The parts Jand // (FiGuRE 64) are then joined
so that the two capillaries are parallel and a few millimetres apart.
138
 MCLEOD GAUGES
A useful safety precaution is to draw the end of the tube dipping
into the mercury reservoir down to a capillary to retard the rate of
flow of mercury into the gauge if excessive air pressure is applied
accidentally. The drawing down of the air inlet on the two-way stop-
cock makes the gauge much easier to operate. Various other methods
have been used for raising and lowering the mercury in the gauge but
the most convenient and most widely used is that shown—the use of
atmospheric pressure and a low vacuum which can be supplied by a
normal water pump if necessary. If this method is used it is advisable
to have the high vacuum connection at more than atmospheric height
above the mercury reservoir.

GF
FiGuRE 65. A small rotating
type of McLeod gauge

The range of a McLeod gauge may be extended by using various

sized tubes in conjunction with a small diameter capillary and similar
sized compensating tubes (L. DUNOYER, 1926). In this way compara-
tively high pressures may be determined on the same gauge as may
low pressures. The calibration technique is more complex as the
volume is not a simple fraction of the length of the compressed gas
column (see FIGURE 64, JV).
Small McLeod gauges have been developed for rapid pressure
measurements: they avoid the tedious raising and lowering of a large
volume of mercury by the gauge rotating about a ground glass joint
in the high vacuum connection. A gauge of this type is shown in
FIGURE 65. The volumes used are relatively small and can be made
‘conveniently from short lengths of tubing; the reservoir bulb should
be large enough to permit free passage of gas into the gauge when all
the mercury is in the reservoir. The volume of mercury required is
determined by the volume of the bulb and the volume of the tubing
139
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT
—the mercury level in the compensation capillary should be level
with the top of the closed capillary when the gauge is in a vertical
position and under vacuum. These gauges are best calibrated against
a normal type of gauge.

Traps to catch Mercury displaced in Apparatus

When parts of an apparatus contain mercury, as in cut-offs and
McLeod gauges, it very often happens that air rushing in or out of the

a
system will displace mercury in an undesirable manner. With cut-offs
across a differential pressure, bubbling of gas through mercury is
unavoidable, and in tubing of internal diameter less than 3-44 mm
bubbling will not occur but the gas will raise a column of mercury in
front of it. It is therefore desirable to insert anti-splash traps wher-

FiGure 66. Trap for catching displaced mercury

ever bubbling is likely to occur. A simple form of trap is shown in

FIGuRE 66, J//.
In the construction a bulb of the required size is first blown and
then a bulge is formed at A (FIGURE 66, J): the tip of the bulge is
blown out resulting in a hole which will just take the size of tubing
being used. A short piece of tubing is bulged and bent to shape
(FiGuRE 66, JJ) and a ring seal is formed at A. Care must be taken to
see that the lower end of the inside tube finally slopes downwards so
that drops of mercury will not lodge in it.

Valves
Valves are of two kinds: those meant to close a system as far
as
possible, and those designed to operate as check valves which
reduce
the flow through a system.
A simple and effective check valve which impedes the flow
of mer-
cury 1s seen in FiGure 67, J. It consists of a small plunger
floating on
mercury, which is pushed into a seating when the mercu
ry level rises
The plunger P is made by blowing an elongated bulb at
the end of a
140
 VALVES
tube and then drawing out a neck. The bulb is about half filled with
mercury and the neck sealed off. The tube C is joined to the tube 4
and the edge of the join at B is heated in places and pushed in to form
a few spikes to hold the plunger (FiGuRE 67, //). The plunger is then
inserted in C; the end of C is rounded off and D is joined. This must
be done with the tubing held vertically, or nearly so, to prevent the
plunger coming in contact with the hot glass; the joint can be made
very easily with a hand torch.
Valves that are to close a system must have two surfaces which
make intimate contact ; this is readily effected by grinding the plunger

Vo! D

S B
B

IT IV VI
FiGure 67. Some kinds of valves

into its seating. A simple design is illustrated in FIGURE 67, III: the
plunger is made by blowing a bulb on the end of a small tube, heating
the end of the bulb, touching it with a piece of hot glass, and pulling
out into a conical shape. The plunger must be able to pass down the
tubing to be used as C. The tubing A is joined to B by a taper with an
angle equal to that of the head of the plunger, and the plunger iS
ground into this taper with fine carborundum until a continuous
ground band 1-2 mm wide is formed on the glass. The cone 1s then
washed and dried, and B is joined to C. The end of the plunger 1s
tapered as shown, mercury is added, and the plunger is sealed off and
a
put into B through C. The junction of C and B is then pushed in at
spikes just allow enough freedom of
few places so that the glass
movement ofthe plunger.
Instead of a glass plunger a ball bearing may be used, as in FIGURE
to
67, IV and V. In both of these examples a larger tube is joined
. The
a smaller tube and the joint is ground to fit the ball bearing
141
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

grinding can be carried out with the ball bearing or with a glass grind-
ing tool of diameter equal to that of the ball bearing. This tool is a
glass rod with a sphere made on the end by rotating the end in a hot
flame. The glass sphere is ground into the joint, and since both surfaces
are ground away a slightly tapered seating is produced which is better
than that produced by direct grinding with the ball bearing. The ball
bearing can float on the mercury as in FiGuRE 67, V, or be kept in
place by a small indentation in the glass (FIGURE 67, JV) made by
pushing it in with a spike. The former valve is closed by the mercury
rising, and the latter may be closed by mercury also, or by a great
enough flow of gas or liquid pushing the ball bearing into place. In
this case the bulb shown in FiGuRE 67, JV, may be replaced by a tube
just bigger than the ball bearing.
Another form of valve which we find useful is shown in FIGURE 67,
VI. To make the plunger a sphere is made on the end of a glass rod,
and to this sphere a thin glass rod is attached, using a very small
flame so that the sphere is not distorted. The sphere is then removed
from the larger glass rod with a small and very hot flame, and it is
rotated with the thin rod until it is spherical. The glass sphere is next
ground into the end of the tube to be used as B; this can be done
easily by rotating the tubing with a lathe chuck and keeping the
sphere stationery. The tube B is then internally sealed into C, the
plunger is put in and finally C is rounded off and D is added. The
tail of the plunger should be long enough to prevent the plunger
coming out from the tube B. A valve of this kind is useful for liquids
and gases; it is opened by pressure of the fluid in B, but flow cannot
occur from D to B.
Soxhlet Extraction Apparatus
The apparently complex Soxhlet apparatus (FIGURE 68, III) is made
by three simple operations: first the large tube A is joined to the lower
main tube H, then the vapour bypass is fitted, and finally the liquid
siphon is put on. The main tube is usually made from 3-cm diameter
tubing about 15 cm long, and this is joined to 1-5-cm diameter tubing
H. There is not a continuous tube when these are joined, and hence
the joint must be blown from both ends. The joint is made by touch-
ing the small tube against the rounded end of the large tube, and then
directing the flame against the smaller tube, which is shrunk and
blown a few times, and then straightened out. Directing the flame
against the smaller tube prevents the larger tube becoming distorted.
A hand torch can be used for this joint.
. To make the vapour bypass, 1-cm diameter tubing is bent with two
right angles, and the ends are cut off so that the tube just fits at C and
142
 SOXHLET EXTRACTION APPARATUS
D and is parallel to the larger tube. Holes are then blown at C and
D. One end of the vapour tube is closed with a stopper and the joint
D is made; then, after removing the stopper, the upper right angle
bend of the vapour tube is heated and the tube pushed until the lower
end meets the hole at C. The end of the vapour tube is then heated
and pushed with a spike on to H until there are only a few small
holes; the joint is finished in the usual way. The top angle is then
made smooth and the whole annealed (FIGURE 68, J). It is advan-
tageous to have a bunsen burner handy with which the top joint is
kept hot while the lower joint is made.

vs
Ficure 68. A Soxhlet extraction apparatus

For the siphon tube some 2-3-mm diameter tubing has a circular
bend made at one end through about 180°, and then about 8 cm from
this bend the tube is bent sharply again through 180° (FiGuRE 68, J/).
The straight end of the tube is bent about 3 cm below the end of the
circular bend, into a right angle in line with the circle. The end is cut
off at Fso that E and F will meet the larger tubes as shown. Another
piece of the small tube is bent into a right an gle to make G. This piece
is internally sealed into H, and a fine hot flame is directed against the
centre of the join, which is blown out to a bubble. G is made central,
and the siphon tube F/E is joined to G by the appropriate internal
seal procedure: E is closed with a stopper. This is removed, a hole
is blown in the large tube A (for joining on £) and, with F heated, F is
a
manoeuvred into position. The joint is then made with the help of
spike as usual. Fis then made smooth and the whole annealed.
143
 SINGLE PIECES OF EQUIPMENT
SOME TYPICAL
of the tube to
aking the second joint £ it is helpful if the ends
d and sligh tly flanged
nes ‘6a hole ath a spike are thickene
the case for small
before the first joint is made. This is especially
bypass and the siphon
diameter tubing. The holes blown for both the
parallel with the main
should be so placed that both side tubes are
tube.
Mercury Vapour Pumps
tory equip-
Mercury diffusion pumps are standard pieces of labora
be constructed
ment and, for most cases, quite adequate ones can
is necessary
from glass. In general, Pyrex glass (or similar glass)
strain 1s
because hot mercury vapour circulates and differential
cy
a High vacuum
Backing pump

Ficure 69. A simple diffusion pump

rapidly set up at various, sometimes complicated, joints. Below,

however, is described a very simple glass diffusion pump which is
quite satisfactory for many purposes, and which can easily be con-
structed from soda or other glasses.
Very Simple Glass Mercury Diffusion Pump
This pump (FIGURE 69) is due to H. P. WARAN (1923) and is designed
such that it may be constructed by the most average of experimental-
ists. Soda glass is satisfactory because all joints are subject to a
minimum, if not a zero, temperature gradient. Pyrex glass would, of
course, be an advantage but is not a necessity.
The boiler A can be made from an ordinary laboratory conical
flask: a side arm is added and then the top of the flask is drawn off
and rounded—f a laboratory filter suction flask is available, and has
sufficient thermal endurance for a boiler, the side arm already exists
and need only be lengthened.
144
 MERCURY VAPOUR PUMPS
The side arm is next bent slightly downwards to form the nozzle B
which can be tapered as shown if desired. The mouth of the nozzle
should be of the order of 5—10 mm diameter, though the size is not
critical provided a good rotary, oil, backing pump is to be used. The
larger the nozzle the faster the pumping that can be obtained, while
the higher the backing pressure employed the finer must be the nozzle
and the higher the boiler pressure of mercury vapour.
The ring seal at C is made not too far from the flask so that the hot
mercury vapour has to travel as short a path as possible. The top of
the boiler and the exposed portion of the side arm may be lagged
with asbestos rope or paper to reduce loss of heat. The remainder of
the pump is easily constructed. The condenser F can be any reason-
able length, for the length does not affect the pump performance. D
is a 3-5 mm bore tube for returning mercury to the boiler, and must
lie below the level of the mercury in the boiler. The condenser jacket
is made separately and is fixed to the inner tube with rubber stoppers
or corks which are waxed to give a watertight seal. The jacket must
be positioned before the T-junction at F is made. If Pyrex glass is
employed then the condenser jacket may, if desired, be joined to the
inner tube with ring seals at either end in the normal manner.
The boiler can be heated with an ordinary bunsen burner but its base
should be coated with a 3’; in. layer of asbestos paste—this is abso-
lutely necessary with soda glass. It may be desirable to employ an
asbestos board G for deflecting radiant heat from the boiler, thereby
protecting the ring seal and the waxed cork of the water jacket.
Waran reports that an average pump of this design, working
against a backing pressure of less than 1 mm of mercury, will pump a
1 litre volume at atmospheric pressure down to 10-? mm of mercury
in 2-3 minutes. When a rotary backing pump is not available, a
diffusion pump with a 5 mm nozzle can be operated, with less effi-
ciency, against a backing pressure of 7 mm which can be obtained
with an ordinary water vapour pump. Sufficient heat must be sup-
plied to the boiler, in the latter case, to produce a mercury vapour
pressure of 1-2 cm.

Glass Mercury Diffusion Pump suitable for High Vacuum Work

A useful glass diffusion pump, due to J. Pollard, is shown in FIGURE
70, VIII. This pump is quite satisfactory for high vacuum work, and
our measurements on pumps made by hand torch methods show that
‘a pumping speed for air of about 10 litres/second at 10-° mm of
mercury can be obtained at the mouth of a pump of this design.
FIGURE 71 shows some results for the speed of this pump communi-
cated to us by J. Pollard.
145
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

This length not

critical Zemin |

:
J ' a
|
. eee t ; ra)
! : De | ,- 7 et ee 3 m,
Kcm 25cm
o aoe [2 Grind # | ee
gmm sharp eqge , 3mm
A
I

X\\ALA Pees meres

=
i

IT IV

f¥es
ae =

About
s0cm

FiGure 70. A mercury diffusion pump due to J. Pollard

146
 MERCURY VAPOUR PUMPS
Two 3-mm internal diameter constrictions at A and B form the
pump ‘jets’. These tapered jets are made in the first stages of the con-
struction. The small jet A is drawn down from 1 cm tubing by careful
rotation and shrinking in a not too hot bench flame. Care is neces-
sary to avoid thickening the glass overmuch. The angle of the taper
1s not critical but should be approximately that shown in FIGURE
70, J. The open end of this jet is ground to a sharp edge.

Curve A Boiler pressure 35 millibars

s}-— Curve B Boiler pressure 20 millibars | ,

millibars/sec.
litre.

Throughput

Curve A 73 millibars
a4 Curve 8 0-9 millibars

10 10 10° 10" 10'

Pressure at mouth of pump millibars

FiGuRE 71. The pumping speed of a Pollard type pump

The larger jet B is more easily made in two parts. The larger,
25-mm diameter tube is drawn down to the required taper to give a
cone-shaped end. As small a hole as possible is blown out at the apex
of the cone with a pin-point flame. A 1 cm tube is similarly drawn
down and its end blown out to a tiny hole. For joining these two
tapers, to give the jet B, the tubes are best clamped, since extremely
skilful handling is required to avoid buckling if they are rotated by hand.
A relatively cool flame will be found adequate for this joint. The joint
must be tested for leaks before the pump is assembled (see p. 82).
The larger tubing of the jet B is drawn off at the required length, the
end enlarged by blowing a slight bulb, FiGure 70, //, and the small
jet A inserted into place by forming a ring seal as shown. The side arm
is added while the seal is still hot and this portion is then annealed.
The ground end of A should be 35 mm from the constriction in B.
L 147
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

Next the water jacket is prepared. A piece of tubing of internal

diameter about 40 mm is drawn off to a spindle at both ends to give
an intermediate piece of the required length. The two side tubes for
in-flow and out-flow of water are added at this stage. The in-flow
tube takes the form of a jet C, FiGure 70, ///, and is inserted by
means of a ring seal. At the lower end of the water jacket is blown a
small hole to take the 1 cm tube. A rubber stopper carrying glass
tube will form a handle at this end for blowing the larger hole needed
at the upper end of the jacket. The two jet assembly in FiGure 70, J/,
is then bulged in the regions X and Y so that the water jacket fits as
shown in FiGurE 70, JV. The top upper ring seal can be sealed with
plasticine, or wrapped with a long strip of thin rubber sheet, so that °
blowing is possible for making the lower ring seal. If desired, the
upper end of the jacket can be permanently plugged with asbestos
paper and sealed with rubber strip or a high melting-point wax.
However, by clamping the apparatus, and with careful initial warm-
ing, we have found it relatively straightforward to make the large
upper ring seal of the water jacket at this stage. We advise a proper
seal here because a waxed joint is liable to spring a leak.
A more skilled worker may prefer to do the operations in the order
indicated by FiGure 70, V and VJ. Annealing is then only necessary
at two stages: first, when the work of inserting the smaller jet, ring-
sealing the upper end of the water jacket and adding the large side
arm is completed ;and secondly, when the work at the lower end of
the water jacket is completed.
When the water jacket is in place, the tubing below the jacket is
bulged and the backing pump side arm D is added—far enough from
the ring seal to cause no strain in the latter. The return drain tube E
for mercury can now be attached.
The boiler is made separately. The inner tube is chosen to fit the
heater to be used, and the outer tube is of such size that the annular
space has a volume of 30-40 c.c. to just above the top of the inner
thimble. A boiler holding more mercury can also be used.
A piece of outside tubing is first drawn off and attached to a piece
of 1-8 cm tube; then the other end of the large tube can be drawn off
and blown out at the required length. The inner thimble is shaped
separately so that a Dewar seal can be made as shown, FIGURE 70,
Vi. The annular space is usually quite small and care must be taken
with this Dewar seal to keep the open end of the thimble from shrink-
ing with the result that the heater cannot be inserted. The side tube
for returning mercury is added while the Dewar seal is still hot. The
completion of the pump after this stage is straightforward.
The heater can be made from Nichrome wire wound in the form of
148
 MERCURY VAPOUR PUMPS
a fine spiral which is itself spiralled round a cylinder of fire-clay. A
thin piece of sheet mica serves to separate the glowing wire from the
inner glass of the thimble. A 150 watt heater is adequate. The tube
from the boiler to the first jet should be lagged along the whole of its
length with asbestos rope.
If desired, a cold finger refrigerant trap can be built in, as shown in
FiGure 70, /X, with a draining tube for returning condensed mercury
to the boiler. The point of return must be below the mercury level in
the boiler so that the high vacuum is separated from the backing
pressure by mercury in the draining tube. It has been found desirable
to have some form of water-cooled condenser, preferably a double
surface condenser, in the backing line D. This condenser can be
positioned conveniently alongside the pump water jacket, and in this
way the tube D bends upwards so that mercury condensing in the
backing line runs back into the drain tube E.
Mercury used in diffusion pumps should be thoroughly cleaned by
vacuum distilling. The glassware should be flushed with chromic
acid, then rinsed and thoroughly dried before filling with mercury.
Any grease left inside the glassware will have a most deleterious effect
on the performance of the pump. If rubber tubes are used in the
backing line these should be freed of sulphur by boiling in concen-
trated sodium hydroxide solution.

Electrolytic Gas Generator

A very convenient method of preparing small quantities of pure
hydrogen is by the electrolysis of a 30 per cent solution of caustic
soda (A. FARKAS and H. W. MELVILLE, 1939). A cell for this purpose
(FIGURE 72) can be made simply. The cell itself is H-shaped with the
top part of the ‘H’ elongated. The side piece should be of tubing
approximately 2-5 cm bore and about 30 cm long, and the cross piece
should be of the same size tubing and about 10 cm long. A narrower
tube, 1 cm bore, for filling with the electrolyte extends from the mid-
point of the cross-piece, and is closed with a rubber stopper when the
cell is in use.
Nickel gauze or thin sheet make satisfactory electrodes, which are
suspended from tungsten seals in the top ofthe sides of the ‘H’. The gas
outlet tubes should be of fairly small diameter (8 or 10 mm), and should
include bulbs of about 20 ml volume to allow the gas to bubble through
any electrolyte which is displaced into the outlet tubes by uneven gas
‘pressure. Both sides of the gas generator should be connected to an
overflow (see FiGURE 72) which allows excess hydrogen and the un-
wanted oxygen to escape by bubbling through liquid (dibutyl
phthalate isconvenient). The gas pressures then remain approximately
149
 TYPICAL SINGLE PIECES OF EQUIPMENT
SOME
that this type of
the same on both sides of the cell. We have found
scale.
generator is not satisfactory when used on a very small
d out in the order con-
The glass-blowing of the cell may be carrie
we have found
sidered most convenient by the person making it, but
T-joint in a
it most convenient to make the ‘H’ first by making one
the glass
bench torch and then making the second T-joint with
sides of the ‘H’ the
clamped, since this makes it easy to get the two
the bench.
same length. This means that the cell will stand firmly on

Y,
LL
Li LE
LLL
LLL
ZZ LL
ites WY
Vin
WfLLL ZL Li

/p MEL
LLL

Z BIT
BE
AE
ENG
OEE
OEE
PLE
PEE
A
EO
I
LEDGE
LE
LE
98TAPE

Ficure 72. Anelectrolytic gas generator

The filler tube is next attached, followed by the gas outlet tube. The
metal—to—glass seals with the electrodes are put in last.

Thermostat Regulators
The regulating device for the heating current of a thermostat may
depend on the bending of a bi-metallic strip or on the expansion of a
liquid. The latter is more sensitive, and we describe some of these
regulators here. An apparatus represented diagrammatically in
FIGURE 73, I, constitutes the ‘make and break’ device. When the
liquid in the regulator expands, the mercury is pushed up and com-
pletes a circuit between FE, and £,, and a relay then operates to cut off
the heating current. When the circuit between E, and E, is broken
the heating current is turned on again. ;
150
 THERMOSTAT REGULATORS
For a sensitive regulator the surface area to volume ratio must be
high, so that thermal equilibrium is rapidly approached. Toluene
(or other convenient liquid) is used as the regulator liquid, with a
mercury column for the electrical contact between FE, and E,. Various
designs giving large surface area have been proposed (see J. REILLY
and W. N. Rag, 1954). A bulky design is shown in FiGure 73, JJ; the
tap T facilitates filling and the adjustment of the regulator for the
required temperature. The apparatus is made from 15-mm diameter
tubing. The construction is simple: perhaps it is easiest to make four
parts separately and join them between A and B, Cand D, and Eand F.

Mercury Toluene Mercury

Z

FiGuRE 73. Some thermostat regulators

The regulating head is an important part of the apparatus; only a

small current should flow, otherwise arcing will occur and the mer-
cury will oxidize. This oxidation is also reduced by keeping the
surface area of mercury exposed to the air a minimum by using
capillary tube. A useful form of head is shown in FIGURE 73, i173 the
capillary is about 1 mm diameter, the upper bulb about 5 mm dia-
meter, and the lower one about 1 cm diameter. An internal seal is
made round the upper bulb; this avoids the thick accumulation of
glass which would result from a direct internal seal to the capillary
tube. The tungsten rod or wire E, of 1 mm diameter can be sealed
in without a sealing glass (for a Pyrex apparatus). The electrode E£, is
held by a rubber stopper.
A convenient thermostat control for a beaker is simply a small bulb
d
filled with toluene to which a U-tube full of mercury is attache
(Ficure 73, IV). This can be made with 8 mm and 15 mm tubing.
The bulb may be filled by successive heating and cooling, with the
head in toluene.
151
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT
To keep to a minimum the current flowing through the mercury it
is best to employ a valve circuit with a relay, so that the regulator
current only controls the grid potential of a valve.

Gas Flow Meters

The usual laboratory gas flow meter is essentially a manometer which
measures the pressure difference across a capillary tube through
which the gas flows, as in FiGure 74, J, showing a design described by
W. BARKAS (1939). This is intended to prevent a sudden change in
pressure from forcing the liquid in the manometer into the rest of the

| 7o manomerer

v/A
FiGure 74. Two gas flow meters

apparatus. The bottom tube of the manometer is slightly larger than

the rest of the manometer tube, and it contains a fairly closely-fitting
ball bearing which is displaced by any sudden flow of liquid and thus
blocks the narrower tube. The construction of this apparatus is
straightforward. When the capillary tube is joined to the wide tube
the join should be as sharp as possible (i.e. there should be a sudden
change of internal diameter) if the flow rate is to be related to the
pressure difference using an equation for viscous flow of the gas
in
question; but a direct experimental calibration can also be
made,
using, for example, a large Berzelius gas holder.
A convenient flow meter is shown in FIGURE 74, II. The constru
c-
tion is simple. Splash heads or bulbs should be incorporated in
both
manometer limbs (as in FIGURE 74, 7). The capillary tube
at 4 may
152
 GAS FLOW METERS
be drawn out as required, in which case calibration is necessary, or a
precision-bore tube may be sealed on. With a drawn-out capillary,
the correct adjustment is found by breaking off portions of capillary
until the proper flow rate is reached; with the ground glass joint
arrangement shown this is very simple. A broken thermometer often
provides a useful capillary for small gas flows.

Spectrum Tubes
The spectrum excited by an electrical discharge through gas at low
pressure can be studied with two simple types of spectrum tube seen
in FIGURE 75. These tubes also act as weak sources of certain radia-
tions, such as monochromatic light for an optical spectrometer. The

‘4
FIGURE 75. Two types of spectrum tube

two tubes shown can be made of Pyrex with fused quartz windows
sealed on with W wax or a cement (p. 84); but for many purposes
these windows are not required. In both J and J// the electrodes are
metal cylinders (e.g. aluminium or stainless steel) joined to tungsten
rods sealed into the glass. Stainless steel cylinders | cm diameter and
3 cm long are suitable—the stainless steel is not so prone to splutter-
ing as most metals. The cylinders are hard-soldered to the tungsten,
or bolted on with a suitable connector. If the tungsten rod exceeds
1 mm diameter it should be sealed in with a sealing glass. 1 mm bore
capillary is suitable for the tubes.
To fix on the window as shown in FiGureE 75, J, a bulb is blown in
the capillary, the top of the bulb is blown out to a bubble, knocked
off, and the end remaining is fire-polished and finally ground flat.
The tube can be filled with gas through a side arm joined to one of the
electrode chambers; the spectrum tube may be sealed off, or pro-
vided with a stopcock. The capillary tube in Ficure 75, IJ, is sealed
to the glass at one end only of the capillary. To do this, the capillary
is joined to a wider tube, which is cut off and flared out to make the
153
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

necessary flare at the end of the capillary. The remaining operations

are simple.

Discharge Tube Lamps

Discharge tube lamps for laboratory work usually consist of three
basic parts: the tube to contain the discharge, the electrodes, and a
water cooling system. Many different types of lamp have been
designed, and while a complete review is outside the scope of this
book a few ofthe more simple ones may be described.
High voltage hydrogen discharge lamps which give continuous
radiation in the ultra-violet can be constructed quite simply and
elaborate designs are unnecessary for many purposes. A very simple

Ficure 76. A hydrogen discharge lamp

design which should give satisfactory results is shown in FIGURE 76.

The lamp consists of a straight tube with a fused quartz or fluorite
window waxed to the end and side arms carrying the electrodes and
connections for pumping out and filling with purified hydrogen. The
lamp is immersed in a trough of water with only the window project-
ing through a watertight seal in the wall. The electrodes should be
of aluminium or stainless steel as these metals are particularly resis-
tant to spluttering under the conditions of the discharge. The tube
containing the discharge should be lightly platinized (see p. 87) to
facilitate recombination of the atomic hydrogen produced in the
discharge.
The chief factors to be borne in mind in the design of these dis-
charge tube lamps are that the electrode material will splutter in the
discharge and, unless it is geometrically impossible, will coat the
window with a metal film, and that a large amount of energy is
consumed by the lamp, much of which is dissipated as heat.
This
154
 DISCHARGE TUBE LAMPS
demands an efficient cooling system, particular
ly in the region of the
waxed joint sealing on the fused quartz or fluor
ite window.
W. West (1949) gives a survey of some disch
arge lamps of different
types in which he provides details of the opera
ting conditions for such
lamps .

FIGURE 77.
A high pressure
mercury lamp

12cm

Air or water cooled mercury discharge lamps find many uses, one
of the more obvious of which is the study of photochemical reactions.
These lamps are usually made of vitreous silica because of its low
‘ thermal expansion, high melting point and its transparency to ultra-
violet radiation. Their operating pressure has a profound effect on
the spectral distribution of the radiation produced and ache it is
important to consider the requirements in the design of such lamps.
155
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

E. W. R. STEActE and N. W. F. PHILips (1938) describe in detail

an ingenious design of a low pressure mercury lamp with a rare gas
carrier. This lamp was designed for the study of mercury-sensitized
photochemical reactions sensitized by the 2537A mercury line. The
lamp itself in this design is of vitreous silica and Pyrex, joined by
graded seals, and mounted in the centre of the reaction vessel, thus
giving the maximum amount of reaction possible. Practical con-
structional details are given by BARR and ANHORN (1949). These
authors also give details of the construction of rare gas discharge
tubes and display signs.
A simple, high pressure, capillary, mercury arc has been described
by F. DAnieLs and L. J. Het (1932), which, although it has only
about twenty four hours useful life, may be useful on account of its
very simple construction. The lamp (FIGURE 77) is constructed from
silica tubing of 1-2 mm bore and 0-7-1 mm wall thickness. The bulbs
A and B confine the discharge and their spacing governs the operating
voltage. The third bulb constitutes the expansion volume and is not
filled completely with mercury. The tungsten electrodes are sealed
in with de Khotinsky cement (shellac and pine-tar), although it is
probable that other stable cements would serve equally well. The
lamp is started by heating the column of mercury between the bulbs
A and B with a small bunsen burner flame until the mercury is
vaporized. A rapid stream of cooling water is then turned on to
prevent the silica tube melting as a result of the discharge. These
lamps require a ballasting resistance since they have effectively zero
resistance before starting, and 20-500 2 resistance is generated when
operating, depending on the operating pressure and other factors.
W. WEsT (1949) also describes the operating conditions for this type
of lamp.

Apparatus containing many Electrodes

Hand made radio valves and other vacuum tubes containing several
electrodes, as for example, mass spectrometer tubes, present special
difficulties. In many cases, electrodes, plates and grids can be sup-
ported by spot-welding to nickel wires which in turn are spot-welded
to tungsten rods that form metal-to-glass seals with the vacuum
envelope. Various ways for making the metal-to—glass seals are
described in Chapter 7. A pinch seal is often employed, as shown in
FicurE 78, /. Careful planning of the position of the final joints, at
A and B for instance, allows for the electrodes to be bent first to any
required shape and position. Clearly an infinite number of particular
assembly arrangements is possible. Prepared copper—tungsten—nickel
wires are often very useful in preparing such electrode assemblies.
156
 APPARATUS CONTAINING MANY ELECTRODES
The above method is not generally suitable when precise spacing of
electrodes is necessary. In electron and ion guns, for example, some
form of rigid mounting with accurately ground spacers between
electrodes is used. The whole mounting is supported in some way in
its glass envelope and fine conducting wires make electrical contact
between the various plates and their respective tungsten—to—glass
seals. FIGURE 78, //, shows diagrammatically such an arrangement.
Copper connecting wires are generally used and must be connected
to the electrodes before the mounted assembly is introduced into its
envelope. The connecting wires can then be drawn one at a time out

rs
FiGure 78. Apparatus with many electrodes

through the tube A. We have found that fine wires are best (provided
low currents are to be passed) because these tend to work-harden,
with resultant fracture, less easily than thicker wires. Pieces of thin-
walled glass capillary, drawn from | cm tubing, make very good
insulating sleeves where these are necessary. The wires can be con-
nected with small connectors to the tungsten rods, or welded with a
pin-point flame to intermediate nickel wires. This is done as near as
is practicable to where the final joint B is to be made.
In high vacuum work all metal parts to be mounted inside the glass
envelope should be thoroughly cleaned by immersion in appropriate
cleaning and passivating dips. Noble metals can be polished with
3/0 or 4/0 (000 or 0000) emery paper. Thorough cleaning of the
metals lessens outgassing effects. Any fluxes used in hard or soft
_ soldering must be thoroughly washed away.
To prevent oxidation of electrodes in an envelope which is being
glass-worked, a current of carbon dioxide can be passed through the
apparatus. This is conveniently obtained from a filter flask containing
157
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

solid carbon dioxide. A side arm is provided to allow the gas to

escape when the final joint has been made in the glass envelope. The
carbon dioxide can also be used to blow the final joint; to do this
a rubber tube fixed to the side arm is squeezed when blowing 1s
desired.

Leaks for Molecular Flow of Gases

Molecular flow occurs when a gas passes through a tube or a hole
with cross-sectional dimensions which are smaller than the mean free
path of the gas molecules. The molecules then flow through inde-
pendently of each other. A leak consisting of a hole in a very thin
diaphragm is often useful. For most purposes this can be made of
metal foil, but sometimes a glass diaphragm with a hole in it is
useful—for example, when free radicals are to be passed through the
hole and the diaphragm material must be chemically inactive. We
find that microscope cover glasses can be used to make a suitable
leak. The cover glasses, about 0-1 mm thick, are first individually
measured with a micrometer and sorted out so that about twenty are
obtained with very nearly equal thickness. These are covered with
paraffin wax on one side, and on each one an area of about 1 mm? of
wax is scraped away. They are put in a basin, and the exposed 1 mm?
area is etched by placing on it a drop of dilute hydrofluoric acid (20
per cent). All are etched simultaneously and the etching is started as
nearly as possible at the same moment for each one. After a time,
two or three of the cover glasses will be completely penetrated by the
acid, the drop of which will be seen to pass suddenly through the
glass. The cover glasses are then all swamped in water, and those
which have not been penetrated will have very thin portions of glass
about 1 mm square; a thin part can be punctured by a spark so that
a smooth circular hole of diameter around 0:01-0:05 mm is formed.
A single spark is obtained from an induction coil by giving one light
tap to the trembler. Pointed electrodes are used one of which is placed
very near the thin part of glass. The cover glass is held in a slot cut in
a solid piece of rubber. A spark length of 1-5 cm is suitable, but this
can be varied depending on the size of hole required.

Precautions to be taken when using Hydrofluoric Acid

Precautions are necessary when using hydrofluoric acid. It must not
come into contact with the skin; the acid can cause severe burns
which heal extremely slowly. A large beaker of water should always
be handy so that any acid on the skin can be washed off immediately.
Running water should also be available. Rubber gloves can be used,
158
 LEAKS FOR MOLECULAR FLOW OF GASES
and they must be free from pin-holes. Barrier creams of the anti-acid
type can be used to coat the hands. A paste of magnesium oxide in
glycerine is an antidote for hydrofluoric burns.

Bourdon Gauges and Glass Spiral Gauges

Among the methods for measuring gas pressures down to 10-! or 10-2
mm of mercury, one of the most convenient is the use of a gauge
which depends upon the distortion of a thin glass envelope under
varying pressures. These gauges can be used above room tempera-
ture, and for corrosive gases which would attack metals.
The simplest form is shown in FiGurRE 79, J: T is a thin-walled
glass bulb with one side collapsed inwards. The shape alters with
pressure difference across the glass, and the pointer P moves. The
gauge needs calibration.
The Bourdon tube (as this bulb-shaped part is called) is made by
putting a spindle in tubing about 10 mm diameter and blowing a thin-
walled bulb just below the shoulder. The bulb is made slightly
elliptical by pulling whilst being blown into. The thinner the bulb the
more sensitive is the gauge. One side of the bulb is heated by a soft
flame until it collapses; the radius of the collapsed side should be
slightly less than that of the other side. On gently blowing into the
tube the pointer should move several millimetres. The spindle can be
left on as a pointer, or be replaced by a thin glass rod. The tip of the
pointer should be drawn out to a fine point. The lower internal seal
of the gauge can be made with a hand torch, the internal tube being
first bulged. A bench torch can also be used. The glass must be
blown into from both ends, since the end of the pointer P is closed.
If the method used for this internal seal involves first rounding off the
wide tube and then dropping in the Bourdon tube (Method 2, p. 75),
the final hole must be made by pulling away the very hot molten glass
with a glass rod, since it is not possible to blow out a bubble of glass.
The movement of the pointer may be observed with respect to a
reference pointer which is sealed into the jacket from above. The
pointer P can also be made to move a glass lever system which
magnifies the movement (S. G. FoorD, 1934).
Another type of gauge depends upon the movement of a thin-
walled glass spiral—S. G. YorKE (1945, 1948) has given construc-
tional details. The spiral is wound on a carbon mandrel (FIGURE 79,
II), which is tapered and threaded, and has a hole in it at the top of
the taper. The flame is directed against a metal shield just below the
mandrel; this gives uniform heating. The thin tubing collapses if
heated directly. The thin tubing for the coil is made by drawing out,
with blowing, 6-mm diameter tubing to 1-25 mm ina ribbon flame.
159
 SINGLE PIECES OF EQUIPMENT
SOME TYPICAL
el, and 7 or 8 turns
The thin tube is put through the hole in the mandr
the hole in the man-
are wound. The spiral is cut where it goes into
ontal (FIGURE 7,
drel. The gauge can be used with the spiral horiz
al form shown a mirror
IIT) or vertical (FIGURE 79, 1V ). In the vertic
lary tube C is
M is fixed to the pointer with Picien. A piece of capil
st excessive
also sealed into the jacket to protect the spiral again
nted in. A
movement. An optical window W is sealed in or ceme
n in FiGu re 79, II.
gauge of this type is more sensitive than that show
a compe nsa-
Both the Bourdon and spiral gauge can be used with
of the
tion method or with direct observation of the movement

7cm

‘6 IT IV
FiGureE 79. Some glass gauges for pressure measurement

pointer. In the compensation method a pressure is applied to the

outside of the bulb or spiral until the pointer deflexion becomes zero.
For maximum sensitivity it is best to control the gauge temperature
by adding an outer jacket through whicha liquid at constant
temperature is passed.
Circulating Pumps
Simple pumps for circulating gas or liquid can be made using two
valves with a chamber between them. One ingenious pump uses a
heat engine (A. R. PEARSON and J. S. G. THomas, 1925). The hard
glass or silica bulb A (FIGURE 80, 7) contains air initially at atmos-
pheric pressure. This is heated with tap T closed; and when the
160
 CIRCULATING PUMPS
pressure is 5-8 cm above atmospheric, T is quickly opened
and
closed. The pressure in the bulb suddenly drops and the mercur
y in
the U-shaped tube is set into oscillation and continues to oscillat
e as
long as the bulb is heated. To ensure regular motion the dimens
ions
of the bulb must be properly chosen. Another way of startin
g the
pump is to heat the bulb (to about incipient redness), then to
open T
and to fix on to the outlet from T a rubber bulb. By squeezing
and

y Mercury

Ficure 80. Circulating pumps

releasing this bulb the mercury is set into oscillation. The tap T is
then closed, when the oscillations should continue. The valves V,
and V, can be of the type shown in FiGurE 67, VJ. V, opens and V,
closes as the mercury rises. Another kind of valve is shown in
FiGureE 80, //. The U-tube can be made of 5 mm tubing and 4 can
be a 250 ml flask. A liquid or gas can be pumped.
The pump in FiGure 80, /, pumps at every other stroke, because
one stroke is required to fill the chamber between the valves. A. H.
CocketTT (1955) describes the pump shown in FiGure 80, J//, which
pumps at every stroke. The piston P consists of a magnet encased in
polythene—this avoids trouble which can arise from a glass piston in
a glass tube forming powdered glass so that the piston sticks. It is
161
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

made by pouring molten polythene into a mould containing the

magnet—the mould can be a piece of glass tube similar to that of the
piston chamber. The piston is finally rubbed with emery paper till a
good fit is obtained; it should not be too tight, nor too loose. The
piston is lifted with a solenoid of 1000 turns of 22 gauge wire (about
7 Q resistance). When the current is cut off the piston falls—the
magnet M prevents it from breaking the apparatus. As the piston
moves up, chamber A is emptied and B is filled; and as it moves

/
down, Bis emptied and 4 filled, so there is continuous pumping.
Apparatus for Semi-Micro Qualitative Analysis
Most workers develop their own particular methods of performing
semi-micro qualitative analyses, and the number of special appara-

lV

e-—
\ H Hl

FiGure 81. Some devices for semi-micro qualitative analysis

tuses is considerable. Usually, however, certain simple items such as

stirrers, spatulas, droppers etc. are used for most of the work. These
can be made with great ease. Stirrers and spatulas are made from
glass rod: thin rod about 10 cm long and 1-5 mm diameter is drawn
out from ordinary rod, or from tubing (p. 58). For a stirrer, about
5 mm of the rod is bent through 180° at the end (FiGure 81, J). Some
thick rod (or tube) can be left on as a handle, as shown. A spatula is
made by drawing a spindle in rod, putting a blob of glass on the end
and flattening the blob with tweezers. The edges can be bent up with
tweezers to form a spoon if desired.
The dropper shown in FiGure 81, //, has a small bulb so that for
maximum uptake liquid does not enter the rubber teat; reagent
solutions are then-less likely to be contaminated. The capillary tube
162
APPARATUS FOR SEMI-MICRO QUALITATIVE ANALYSIS
can be drawn after the bulb is blown, and it should not have too thin
walls. Some glass should be accumulated before the tube is drawn
down.
For detecting gases a tapered tube is made which fits snugly into
the top of a small test tube, as in FiGure 81, I//. The hole A should
be about 2 mm diameter, and the top about 5 mm diameter, so that
the reagent can be put in with a dropper. This apparatus is not
reliable because the gas usually escapes up the side of the tube instead
of through the reagent. The two parts can be ground together to stop
this; care is then necessary, if the test tube is heated and cooled, to
stop the reagent entering it. A thin tube can be prepared, as in
Ficure 81, /V, for bubbling gas through a reagent: this is again
tapered to fit the test tube, and is preferably ground in; alternatively
a rubber ring can be used as a gasket. The volume of the gas delivery
tube should be small, so the glass tube is first worked till the walls
are quite thick by pushing it together while it is rotated in the flame.

Gas Analysis Apparatus

The art of analyzing gases by chemical means has made vast progress
since the early researches of R. W. Bunsen. Most gas analysis
apparatus, however, still contains a reservoir, a combustion chamber,
and an absorption vessel, and formidable-looking apparatus is found
on inspection to be capable of assembly by a series of moderately

Ficure 82. A simple form of gas analysis apparatus

M 163
 SOME TYPICAL SINGLE PIECES OF EQUIPMENT

straightforward glass-working operations. A simple form of appar-

atus is shown in FIGURE 82; B is the combustion chamber. The
apparatus can be used in different ways: for example, both C and D
can be used as reservoirs or absorption vessels—C is usually the
reservoir and D the absorption vessel.
The best size for the capillary tubing is 1-5 mm bore; the rest of the
tubing in the apparatus is 8 mm diameter. In making the part B, two
tungsten rods W, and W, (FiGure 82, //) are spot-welded or hand-
soldered to a platinum wire coil—if spot-welding is used it is much
easier if an intermediate piece of nickel is introduced. The tungsten
is beaded, with a sealing glass if desired ;two holes are blown in B for
the tungsten, which is put in, with the platinum, through the lower
tube (FIGURE 82, JJ). It is convenient if the beaded tungsten just
passes through the holes. The volume of the reservoir C from the
tap 7, to the mark M, must be measured (if it is needed) before this part
is joined to the apparatus. Similarly, the volume between the tap T,
and the mark M, should be found before assembly. A rigid frame is
needed for the apparatus because of the great weight of mercury in it.
The mercury reservoirs A,, A, and A, are standard equipment;
they can be made by joining at each end about 8 cm of 4-cm diameter
tubing to 8 mm tubing. The large portion of tubing is then blown to
a bulb of about 7 cm diameter, which is worked by pulling and blow-
ing to a pear-shape. The 8 mm tubing attached to the pointed end of
the bulb is corrugated to take rubber tube, and at the other end the
bulb is blown out and flanged.

REFERENCES
BARKAS, W., 1939, J. sci. Instrum., 16, 162.
Barr, W. E. and ANHorRN, V. J., 1949, Scientific and Industrial Glass
Blowing and Laboratory Techniques; Pittsburgh, Instruments
Publishing Co.
CockeTT, A. H., 1955, Nature, Lond., 175, 768.
DANIELS, F. and He1nt, L. J., 1932, J. Amer. chem. Soc., 54, 2381.
Dewar, J., 1896, Proc. Royal Instn G.B., 14, 1.
Dunoyer, L., 1926, Vacuum Practice; trans. J. H. Smith; London,
Bell.
FARKAS, A. and MELVILLE, H. W., 1939, Experimental Methods in Gas
Reactions ;London, Macmillan.
Foor, S. G., 1934, J. sci. Instrum., 11, 126.
Kon, G. A. R., 1930, J. chem. Soc., 182.
LANGMUIR, I., 1918, U.S. Patent 1, 273, 629.
McLeop, H., 1874, Phil. Mag., 48, 110.
164
 REFERENCES
Mor.ey, A., 1940, Strength of Materials, 9th Edn; London,
Longmans.
PEARSON, A. R. and THOMAS, J. S. G., 1925, J. chem. Soc., 127, 2450.
REILLY, J. and Rag, W. N., 1954, Physico-Chemical Methods, 5th
Edn. Vol1I; London, Methuen.
STEACIE, E. W. R. and PHILLIps, N. W. F., 1938, Canad. J. Res., 16B,
219.
WaARAN, H. P., 1923, J. sci. Instrum., 1, 51.
WEsT, W., 1949, Technique of Organic Chemistry, 2nd Edn. Vol. 1,
Part 2, Ed. by A. Weissberger ; New York, Interscience.
YorRKE, S. G., 1945, J. sci. Instrum., 22, 196.
YORKE, S. G., 1948, J. sci. Instrum., 25, 16.

165
 Chapter 9

THE ASSEMBLY OF COMPLEX

APPARATUS

General Observations
THE assembly of complex glass apparatus is comparatively easy pro-
vided it is done in a systematic manner. It is most important to work
out the general arrangement of the various parts of the system before
starting to assemble them, and the sequence of operations should be
planned in detail.
The arrangement of apparatus is usually dictated to some extent by
the laboratory space available; but within this limitation careful
preliminary planning can make the construction, use and mainten-
ance of the apparatus easier. The apparatus should be designed so that
all parts are easily accessible and so that its operation does not impose
too great a physical strain on the worker due to awkwardly placed
components. An apparatus, badly designed in this respect, will prob-
ably be broken much more often than a more carefully planned one.
Assuming that the apparatus has been carefully planned and that it
is to be a vacuum apparatus, it is generally more convenient to start
building it at the pump end of the vacuum line. This allows the joints
to be tested before the complete system is finished, which is an advan-
tage if some of the joints are screened by other parts of the apparatus
constructed later. Probably the most convenient method of working
is to join on each section with its far end finishing in a sealed-off
spindle which can easily be blown out preparatory to the joining of
the next section. This technique also excludes unnecessary dust and
contaminants from the system during building. If some other method
of rendering the system air-tight for blowing is used, care should be
taken to see that it does not contaminate the glass. Care should also
be exercised to avoid introducing flakes of glass into the apparatus
when holes are blown in the glass. If such flakes are carried in the air
stream during evacuation and lodge in the diffusion pump, its
efficiency may be seriously impaired.
The cleanliness of the inside of a vacuum system is of the utmost
importance, and therefore care should be taken at all stages to see
166
 GENERAL OBSERVATIONS
that the glass is thoroughly cleaned before assembly and that no un-
necessary dirt is introduced during assembly. The introduction of a
single length of improperly cleaned tubing can render useless all the
effort spent on cleaning the remainder of the apparatus. Open tubes
on the apparatus should be blocked in some way, even if not required
to be air-tight, to prevent dust from entering them. Sometimes it may be
desirable to prevent excessive amounts of moisture from entering the
apparatus. For this purpose a tube of drying agent can be included
in the blowing tube, although very often large amounts of water
vapour enter the apparatus as combustion products of the flame.
When glass-blowing is carried out on a system containing greased
stopcocks or ground joints they should not be overheated ; if they are,
grease is spread over the inside of the system. Silicone grease on
ignition forms a solid mass which cannot be burnt off.
Whatever the problems and however complex the apparatus to be
constructed, the sequence of operations should be carried out men-
tally before any practical work is started. The time spent in planning
any piece of practical work is well spent and will probably save time
eventually.
Joining Closed Systems
The need for joining closed systems arises quite frequently, and
various methods may be used to make the final joint. Difficulty
usually arises out of the necessity of blowing two holes in the system
in close proximity to one another.
One method of doing this has been described already in the section
on McLeod gauges; it involves blocking the first hole with a stopper,
blowing the second hole and then bending the glass until the holes are
in a suitable position for joining. This method may be varied so that
it is unnecessary to allow the glass to cool after blowing the first
hole. If care is taken in blowing the first hole, a bubble of glass,
sufficiently thin to be knocked off but strong enough to withstand the
blowing pressure, can be produced. It is then possible to blow the
second hole and remove the thin glass bubbles.
The most difficult part of joining a closed system is making the two
prepared tubes meet correctly. This can usually be accomplished by
bending either or both of the tubes, and perhaps in addition drawing
the tubes out until the desired fit is obtained. Once the joint has been
made by pushing the glass together or by filling the gap with rod, the
irregularities produced by bending and pulling can be evened out by
judicious heating and blowing, although for many purposes they
will not have any effect on the efficiency of the apparatus.
Another method which may be useful in joining closed systems is
167
 THE ASSEMBLY OF COMPLEX APPARATUS

described more fully in the next section. This involves making the
two final joints simultaneously, at either end of a reasonably long
length of tubing, inserted between prepared sites.
Making more than one Joint at once
When several tubes in a complicated piece of apparatus have been
cut in order to remove a large or intricate portion for repair or alter-
ation, difficulty is encountered, during reassembly, in obtaining a
closed system for blowing. It becomes necessary to make several
joints at once and we have found various techniques useful in these
circumstances.

Closing Tubes with Ragged Ends— Use of Rubber Caps

The cutting operation when the tubes are not easily accessible usually
results in slightly ragged edges. However, the two halves will fit
perfectly, and it would be convenient not to have to fire-polish either
edge, so that when the reassembly is undertaken the two will still be a
perfect fit. If a rubber stopper is used for blocking a cut tube, fire-
polishing is necessary in order to make an air-tight seal and in order
to reduce the somewhat large danger of splitting the ragged end when
the stopper is forced in. Rubber caps are therefore very useful here
and, with large tubing, thin pieces of sheet rubber (see p. 35).
Glass-Blowing Two Joints Simultaneously
It is occasionally convenient, both when dealing with complex
apparatus and when replacing a short section of a fixed glass tube, to
be able to glass-blow two joints at once. This can successfully be
undertaken when the two joints are reasonably close together and
accessible, and when the tubes involved are not larger than 15 mm
diameter.
The interposing portion of glass is first shaped and its ends worked
until it fits as closely as possible. It is then clamped in position. By
using the hand torch for a few seconds at a time on each joint they
can be worked simultaneously, care being taken to heat each evenly
all round so that one does not crack in the few seconds that the other
is being worked. By using the spike and welding rod where neces-
sary, the two joints can be first sealed and then blown to satisfactory
shape. Each joint must be separately and carefully annealed (unless
they are so close in the same tube that they can be heated together to
the annealing point).
Glass-Blowing more than Two Joints at once
The following method also applies to two joints that cannot be
treated by the procedure just described: Each joint is worked in turn
168
 MAKING MORE THAN ONE JOINT AT ONCE
without blowing and sealed up, using the hand torch, welding rod and
the spike. Blowing is not possible until the last joint is reached, and
so care must be taken not to collapse the tubes too far nor to apply
too much rod and produce large blobs which will sag. Too hot a
flame should be avoided so that the portions to be sealed do not
shrink away from each other. Pin-holes will not matter in this part of
the operation as long as a pressure sufficient for blowing can be built
up. After each joint has been worked in this way, it must be very
carefully annealed since the half-finished seal is normally very uneven
in thickness and resulting strains can easily cause cracking while the
next joint is worked. The joint that will give the most difficulty,
either because the tubing is large or because it is inaccessible, is left
to last since blowing will then be possible. Finally each joint is care-
fully warmed up, blown to shape, finished and annealed.
Clearly, when a portion of apparatus has been removed by cutting
several tubes, each of the tubes can be drawn off before reassembly
and, in the reconstruction, short pieces of new tubing can be used for
joining the various pairs of cut tubes in turn. However, when the
removed portion can be re-clamped in position with the cut tubes
meeting well, it has been found that the above procedure is quicker
and more satisfactory, particularly with tubing larger than 15 mm
diameter.

Use of Thin Rubber Sheet in making Several Joints at once

With extra large tubing, greater than 35 or 40 mm diameter, a joint
sealed with welding rod but not blown out to even thickness is un-
likely not to crack on cooling, however well annealed. The procedure
in the preceding paragraphs cannot therefore be used when more than
one such joint is involved. Thin rubber sheet has been found very
useful on such occasions. Small pieces, used to close the ragged end
of a tube (see p. 35), allow the two halves of a joint, that fit perfectly,
to be brought together with the rubber still between them. Then as
each joint is made, the thin sheets of rubber can be removed one by
one and the minute gaps can easily be closed with the welding rod.
Removal of Strain and Clamping of Apparatus
piece of
Strain in the glass, created during the construction of some
r, further strain
apparatus, is usually removed by annealing ; howeve
d to a rigid
is sometimes set up when a piece of apparatus 1s clampe
may not remove
framework at more than one point. Annealing alone
ed so that it will
strain produced in this way—the glass must be soften
rable to heat any-
flow. With some pieces of equipment it is undesi
for example, with a
where after the whole has been annealed—as,
169
 THE ASSEMBLY OF COMPLEX APPARATUS

diffusion pump—and if clamping at more than one point is unavoid-

able then this clamping must be performed with extreme care. The
claws of the clamp should be well padded with cork, asbestos ropeor
layers of asbestos paper. If rubber is used it should be dusted with
french chalk to prevent sticking which can cause fracture when un-
clamping at some later date. See p. 39 for remarks concerning types
of clamps.
Wherever possible, when a glass tube has been clamped rigidly at
two points, the tube between these points should be heated to the
softening temperature—so that the glass will flow and take up any
inaccuracy in the clamping—and then annealed. In this way an
almost completely strain-free system can be obtained. Clearly a rigid
framework is necessary if this strain-free condition is to be main-
tained.

Annealing by Flame
When an annealing oven is not available, even quite large pieces of
equipment can be annealed by flame. Apparatus assembled on the
bench can often only be annealed by a flame. As a general principle
each joint should be annealed as it is completed, unless a further join
is to be made, or another operation is to be performed immediately
in the same region or so near to it that the whole of the first joint can
be kept hot—then the whole is annealed together.
As large a flame as possible is used, with the oxygen adjusted so
that the inner blue cone is long and a very pale transparent blue (for
Pyrex and similar glass). The glass is warmed slowly—with large
equipment, made with tubing greater than 40 mm diameter, warming
should be begun with a luminous flame or even with hot air—until
the whole has been raised to a dull red heat and the glass is distilling
sodium (colouring the flame yellow) freely. As much of the glass as
possible is kept at this temperature for a few minutes, and then, by
slowly reducing the oxygen content and, finally, the size of the flame,
the glass is allowed gradually to cool. The flaming is continued until
an even, black deposit of carbon is formed.

Use of a Bent Blowpipe and Double Tipping Device

For getting at tubes placed awkwardly in complicated apparatus a
bent tubular blowpipe is very convenient. A useful design is shown
in FIGURE 83, J.
We think that a long thin blowpipe, as in FiGure 83, J, but bent at
the end through a greater angle—about 150°—would be very useful
for working on complex apparatus. It would enable the back of
tubing to be worked from the front.
170
 USE OF BENT BLOWPIPE AND DOUBLE TIPPING DEVICE
A double tipping device, which forms two fine flames directed
towards one another from nozzles about 2 cm apart, is shown in
FIGURE 83, LI (see p. 29). This blowpipe head is very handy for work-
ing small tubing in awkward positions. When joins have to be made
in such tubes it is desirable to arrange things so that the ends can be
pushed together whenever possible. The use of welding rod at long
range, or when the operator has to assume a tedious position, requires
a very steady hand.

Loy
fom

e v/h
Ficure 83. Specially shaped blowpipe
heads

Danger of Flame Cracks

Recognition of Flame Cracks
Care must be taken, when working on complicated apparatus, to
avoid producing flame cracks. These are caused by momentary
playing of a hot flame on cool glass. Cracks formed in this way may
not appear for several hours, and in some cases even days, after being
caused. They generally appear in the surface or bulk of the glass in
one of the two forms shown in FiGurE 84. The first is a small curling
crack, usually between 5 and 10 mm diameter if circular, and is very
characteristically a flame crack. The second general type has a
starred appearance with small cracks radiating from a tiny curl or
point. The exact mode of formation of these patterns is uncertain,
but it is clear that some form of differential strain is the cause.
Sudden heating of the surface layer makes it strain-free, and, on
cooling, tensile stress develops in the surface layer because the inner
layers of the glass are rigid. Glass in tension can undergo a delayed
fracture (p. 10).
Precautions for avoiding Flame Cracks
With intricate apparatus, particularly when a hand torch is being
171
 THE ASSEMBLY OF COMPLEX APPARATUS

used, it is sometimes impossible to avoid playing the flame on to some

portion of the assembly that is not actually being worked. Such por-
tions must be protected by wrapping with asbestos paper; this is
particularly important when a joint such as an internal seal is in
danger. When a hot flame has been inadvertently played on a piece
of cool glass, the latter should be annealed, after the work has been
completed, to avoid subsequent formation of aflame crack.

Em, CO

Oy
eae) <>
7s vhs
FiGureE 84. Flame cracks

Removal of Flame Cracks

Flame cracks are easily removed by warming the glass, gently at first
to avoid spreading the crack, and raising it to the annealing tempera-
ture. When joints are in the vicinity care must be taken to heat every-
thing uniformly and to anneal as a whole.

172
 Chapter 10 a

THE MANIPULATION OF SILICA

The Properties of Fused Silica and General Remarks

VITREOUS silica, commonly known as Vitreosil, is pure homogeneous
fused silica containing over 99-8 per cent SiO,. This glass has many
extremely useful properties—such as chemical resistance, extremely
low thermal expansion with resultant constancy of volume and resist-
ance to thermal shock, resistance to heat and transparency to ultra-
violet and infra-red radiations—making it a valuable material for
many scientific purposes. These properties are dealt with in varying
detail in Chapter 2. Resistance to heat is often of considerable
importance since Vitreosil has an extremely high softening range, in the
region of 1500°C, and can, under normal working conditions, be safely
kept at temperatures of 1100°C and as high as 1300°C for short periods.
In vacuum work, for evacuated furnaces for instance, fused silica can
be used at temperatures up to 1100°C; and provided no devitrifi-
cation has occurred the glass remains practically impermeable to gases
at these high temperatures, with the possible exception of helium.
Fused silica is manufactured in both translucent and clear, trans-
parent forms; the particular requirements govern the form to be used.
The translucent variety can be obtained with a sandy, satin or glazed
finish. An interesting historical account of the early work on the
practical manipulation of fused silica has been given by W. A.
SHENSTONE (1902).
Devitrification is a factor requiring particular attention. Impur-
ities on the surface of Vitreosil will readily cause crystallization of the
surface glass when it is worked. Salts of sodium, potassium and
lithium, salts of vanadic acids, fluorides, phosphates and radioactive
substances, are especially active in this respect. It is therefore very
important, particularly when it is required that the silica remains
clear and transparent, that the glass be washed carefully, inside and
out, before working. If it has been necessary, during preparation, to
grind the fused silica, then all traces of abrasive must be removed by
dipping in dilute hydrofluoric acid followed by careful rinsing. This
treatment, with subsequent heating to just below the softening point
173
 THE MANIPULATION OF SILICA

with a flame, is also effective in removing any condensed surface

deposits of silica due to volatilization during working. Precautions
advisable when using hydrofluoric acid are mentioned on p. 158.
Any cleaned portion of the Vitreosil must not be touched with the
fingers before being worked. When using a bench flame, if the part
being handled is subsequently to be worked, the piece of equipment
must be re-cleaned before this further work is carried out. For this,
cotton-wool dipped in alcohol is very useful—after the piece has
been allowed to cool.

General Technique for Working Fused Silica

Cutting Tubing
Fused silica is slightly more brittle than borosilicate glasses and is
therefore more easily fractured. Tubes up to just over 1 cm in dia-
meter can be snapped with the fingers after scratching with a file or
glass knife. Any ragged end so formed can be cleaned up with an
abrasive wheel.
With larger tubing a cutting wheel is necessary since the low
thermal expansion renders local heating entirely ineffective in starting
a crack. A silicon carbide cutting wheel, 345 to # inch thick and
running with a peripheral speed of 6000 feet per minute, is very suit-
able and quite adequate. A copious supply of water directed on the
cutting edge is necessary and should be directed from either side of
the cutting wheel. Copper or phosphor-bronze, bonded, diamond,
cutting wheels can also be used but are much more expensive items.

Fusing and Working Operations

As a medium for glass-working, fused silica has the considerable
advantage over the softer glasses that careful and uniform heating
and annealing are unnecessary. Slightly offsetting this advantage is
the necessity for careful cleaning, as described above, and for
stronger heating.
An ordinary oxy-coal-gas torch is only adequate for small size
equipment; when working with tubing of greater diameter than 15
mm, oxy-hydrogen or other very hot flames must be used. Special
burners are obtainable however—as, for example, the No. 8 Thermal
Syndicate torch—with which oxy-coal-gas, or, better, Oxy-propane
mixtures, will give a sufficiently hot flame for this larger tubing, so
avoiding oxy-hydrogen mixtures which are less safe in operation
unless anti-burn-back valves are employed.
In all fusing operations with fused silica, protective spectacles must
be worn—a glass whch absorbs at least 96 per cent of the incident
174
 GENERAL TECHNIQUE FOR WORKING FUSED SILICA
light is necessary. The recommended glass for best protection is
green Protex grade C, § inch thick (Chance Bros. Ltd).
For hand work with a bench torch, tubing is rotated in the flame
to secure even heating and hence even shaping; while when clamping
of the various tubes is employed and the hand torch is used, all the
dangers of cracking, which are present when borosilicate or soft
glasses are being worked, are negligible. The Vitreosil must be blown
while still in the flame; and therefore, when a bench torch is used and
the apparatus is manipulated by hand, a blowing tube must be
attached at one end. It is preferable then to use a glass-blower’s
swivel (see p. 42).
For tooling of molten Vitreosil 35 inch Nichrome wire may be
employed ; though if much tooling is required, then carbon rods or
plates are necessary. In general, most of the operations described in
earlier chapters can be modified for working with Vitreosil and, if a
sufficiently hot flame is available, less manipulative skill is usually
necessary because the material stays fairly viscous.
When operating with tubes it is well to take care to prevent the
flame licking the insides ; the vapours can cause contamination which
results in devitrification when more work is done.

Large Diameter and Thick-Walled Tubing

With very large silica tubing. over 4 cm bore, the walls are usually
very thick. Special methods are required when the walls are 5 mm
thick or more. The wall edges, at the ends to be joined, are cut or
ground to sharp knife-edges, so that a V-shaped notch is left all round
when the ends are butted together. The ends are fused together,
using a hand torch and 3—5 mm welding rod to fill in the V-shaped
notch. With very large tubes the flame must be manipulated to keep
the tubes hot all round, maintaining an even heat, and the apparatus
must be annealed—preferably in a furnace for one hour at 1050°C.
Translucent and transparent Vitreosil rods of the proper size for
the above operations are obtainable.
Silica Torsion Fibres
These fibres can be made by drawing by hand or by simple mechani-
cal means, and by blowing them out with a flame. Here we describe
only simple methods for making fibres by drawing them out with a
weight falling under gravity. With these methods fibres with a range
of sizes can be made quite quickly. T. C. KEELEY (1924) fixed a small
weight to a silica rod and heated this at a point until the silica
softened sufficiently for the weight to fall and draw out a fibre. V. J.
CLANcEY (1950), during an investigation of the evaporation and
Lio
 THE MANIPULATION OF SILICA
te torsion
surface mobility of naphthalene molecules, used a delica
d. He
apparatus, the fibre for which was made by the gravity metho
ter silica rod. The rod was
fused a silica hook to a 1-mm diame
on the hook.
clamped vertically and a weight of about 10 gm hung
as the
The rod was heated with a flame just above the hook, and just
fell on
weight started to fall the flame was whisked away. The wei ght
r sheet, to stop it bounc ing. By
to a soft material, such as a rubbe
varying the weight and the moment the flame is taken away, fibres
with different sizes can be made. Some preliminary practice 1s re-

Ficure 85. The production of a silica fibre

by drawing out a fused rod with a weight
falling under gravity

quired. If the flame is kept on too long the fibre will fuse and break,
and if the flame is kept on for too short a time the weight will not fall
the full length.
We find it more convenient to draw out fibres by the gravity
method with a weight of about 500 gm which is released from a pair
of electromagnets. The weight consists of an iron bar about 25 cm
long, to the centre of which the silica rod of 2-5 mm.diameter is
attached with a socket and bolt arrangement. The bar is held at each
end by a small magnet, and the two magnets are connected in series.
A portion of the rod is heated with a hand torch and the top of the
rod is held in one hand (FiGuReE 85). A semi-molten portion of
176
 SILICA TORSION FIBRES
desired thickness is obtained, the magnet current is turned off with a
foot-operated switch, the flame is whisked away, and the weight falls
into a box of sand. If all goes well a long fibre is produced. The
semi-molten part of the silica rod, when initially thin, gives a thin
fibre, and when thick, a thick fibre; the fibre thickness is also deter-
mined by the precise moment the flame is whisked away—for a thin
fibre this should be just after the weight starts to fall. Fibres down
to 5-10 microns diameter can be made in this way.
N. J. TIiGHE (1956) has given a comprehensive survey of the pro-
duction, properties and applications of silica fibres; G. A. Downs-
BROUGH (1937) has discussed the damping of fibres. With a rubber
catapult method, introduced by R. THRELFALL (1898), fibres in the
range of 2-30 microns diameter can be made. Smaller fibres can be
made by the drag of the flame gases in an oxy-hydrogen flame on a
tiny globule of molten silica, a method which Threlfall describes.
This author also gives an account of the bow-and-arrow method of
C. V. Boys.
Small fibres should be manipulated against a dark—preferably
black—background, and small camel’s-hair brushes moistened with
water are very convenient implements. Fibres, when examined under
a microscope, become more easily visible if they are mounted under
a cover slip and a drop of glycerine, coloured by a dye, is allowed to
run in.
P. L. Kirk and R. CraiG (1948) have discussed the construction of
fibre devices; L. WALDEN (1937) has given a practical account of
several uses of fibres as instrument suspensions ;D. R. BARBER (1930)
has described a device for mounting fibres with a degree of tension;
and H. V. NEHER (1940) has given a very practical account of the
manipulation of silica fibres.
REFERENCES
BARBER, D. R., 1930, J. sci. Instrum., 7, 105.
CLANCEY, V. J., 1950, Nature, Lond., 166, 275.
DOWNSBROUGH, G. A., 1937, Phys. Rev., 51, 877.
KEELEY, T. C., 1924, J. sci. Instrum., 1, 369.
Kirk, P. L. and Crala, R., 1948, Rey. sci. Instrum., 19, 777.
Neuer, H. V., 1940, in Modern Physical Laboratory Practice by John
Strong, Chap. 5; London & Glasgow, Blackie.
SHENSTONE, W. A., 1902, Proc. Royal Instn G. B., 16, 525.
_ THRELFALL, R., 1898, On Laboratory Arts; London, Macmillan.
Ticue, N. J., 1956, National Bureau of Standards C ircular 569;
Washington, U.S. Government Printing Office.
WALDEN, L., 1937, J. sci. Instrum., 14, 257.
Lay,
 AUTHOR INDEX

Anhorn, V. J., 5, 28, 41, 47, 123, 136, Gardner, I. C., 85, 89
156, 164 Garner, W. E., 7, 26
Ansley, A. J., 5 Grodzinski, P., 84, 89
Armstrong, H. E., 3, 5 Gurney, C., 11, 26
Aston, F. W., 4
Harcourt, W. V., |
Baker, E. A., 87, 89 Heatley, N. G., 84. 89
Barber, D. R., 177 Heidt, E. J:, 156, 164
Barkas, W., 152, 164 Heldman, J. D.. 3, 4, 5
pee E., 5, 28, 41,,47, 123, 136, 156, Herington, E. F. G., 87, 89
1 Home Dickson, J., 5
Berzelius, J. J., 3,5 ra par al WG 10991135) 114 >

Bodenstein, M., 4
Bolas, B. D., 3, 5, 48, 89
Bolas, T., 3, 5 Jackson, H., 131
Boyle, R., 7, 17, 26 Johannsen, A., 15, 26
Boys, C. V., 3, 83, 84, 89, 177
IBEAUCICKA EA ew isto Keeley sie Glial i
Brickwedde, F. G., 85, 89 Ritko Pee Le
Brown, B., 84, 89 Kon, G. A. R., 132, 164
Bunsen, R. W., 3, 17, 48, 163
Langmuir, I., 4, 123, 164
Callendar, Ho 1. 3,5 Lavoisier, A. L., 1
@ase;.F.A.,.85,:89 Lloyd, J. T., 77, 78, 89
Chattaway, F. D., 87, 89
Clancey, V.-J., 175, 177 MacDonald, D. K. C., 17, 26
Cockett, A. H., 161, 164 McLeod, H., 136, 164
Cofm C C., 110, 116 Melville, H. W., 149, 164
Cook, J. W., 85, 89 Morey, G. W., 1, 6
Grape ke. 177 Morgan, E. H., 5, 6
Morley, A., 119, 165
Dale, A. E., 9, 20, 26
Daniels, F., 156, 164 Neher, H. V., 177
Davies, J., 135 Nokes, M. C., 5, 6
Dewar, J., 4, 123, 164
Donal, J. S., 85, 89 Ockenden, F. E. J., 87, 89
Douglas, R. W., 9, 26
Downsbrough, G. A., 177 Park-Winder, W. E., 5, 6
Dunoyer, L., 139, 164 Partridge, J. H., 15, 27, 44, 47, 102,
111,116
Edwards, J. D., 3, 5, 19, 26 Pearson, A. R., 160, 165
Elliott, A., 5 Pearson, S., 11, 26
Phillips, C. J., 9, 27
Faraday, M., 1, 2, 3, 5, 14, 18, 26, 48, Phillips, N. W. F., 156, 165
83, 84, 89 Pollard, J., 145, 146
Farkas, A., 149, 164
Flinders Petrie, W. M., 1, 5 Rae, W. N., 151, 165
Foord, S. G., 159, 164 Ramsay, W., 4
Frary, F. C., 3, 5, 19, 26 Randall,J.T., 8, 27
Reilly,J., 151, 165
French, E. A. H., 87, 89
N 179
 AUTHOR INDEX
Reimann, A. L., 4, 6, 43, 47 Tighe, N. J:, 13, 27, 177
Rideal, E. K., 4,6 ToddeB alee lies,
Roscoe, H. E., 4, 6 Travers, M. W., 4, 6
Rosenblum, S., 77, 89 Turner, W. ES. 17527

Scheele, K. W., Upton, P. B. G., 87, 89

Scott, R.B. zas, 89
Shenstone, W. Ar 3, 4, 6, 30, 47, 173,
Walden, L., 177
LZ
Smith; Gs besa aAg sho Walen, R., 77, 89
Stanworth, J. E., 9, 17, 20, 26, 27 Waran, H. P., 144, 145, 165
Steacie, E. W. R., 156, 165 Warren, B. E., 8, 9, 27
Strong, J., 5, 6 West, W., 155, 156, 165
Wilkinson, R. A., 49, 89
7 Wright, R. H., 3, 6
Tammann, G.
5; 19,26.
Taylor, C. S., 35.
Taylor, G. F.F 110, 16 Yarwood, J., 87, 89
Thomas,J. S. G. 160, 165 Yorke, S. G., 159, 165
Threlfall, R., 3, - 28, 46,
gnc> 47,177
Threlfall, R.2,
E a Zachariasen, W. H., 8, 27

180
 SUBJECT INDEX
Abrasives, 39 Cerirouge, 40, 83
Alkaline attack on glass, 16 Chemical properties of glass, 15
Alumina as abrasive, 39 ‘Chromic acid’ for cleaning, 49
Ampoule sealing, 29 Claisen flask, 120
Ampoules, 117 ae apparatus, delayed fracture
Annealing of glass, 20, 170 ol,
Annealing oven, 45 pn ie apparatus, working on, 4,
Annealing temperature, 12
Asbestos gloves, 37, 90, 92 Clamps, 39, 170
Asbestos, uses of, 36 Cleaning glass tubes, 48
Closed systems, 167
Bellows for blowing, 30 Closing a large flask, 92
Bellows, glass, 101 Closing tubes with round ends, 57
Bench flame method of work, 2 aoe tubes with rubber caps, 34,
Bends, 53 1
Bends in big tubing, 90 Closing tubes with stoppers, 35
Blowers for air, 30 Cold finger traps, 125
Blowing tubes, 35, 56 Complex apparatus, 166
Blowpipes, 28, 170, 174 Condenser, double surface, 133
Boats, glass, 128 Condenser, fractional distillation, 131
Borate glasses, 9 Condenser, Liebig, 76, 129
Bordering, 72 Condensers, 128
Boron atoms, coordination of, 8 Copper, deposition on glass, 87
Borosilicate glass composition, 9 Copper-to-glass seals, 113
Bottles, cutting big ones, 51 Copper-tungsten-nickel wires, 107
Bourdon gauge, 159 Corks, 35
Breaking glass tubes, 49 Corning No. 3320 glass, 105
Breaking of glass under tension, 9 Corning Nos. 7040, 7052 and 7060
Breaking of glass when heated, 12 glasses, 112
Break-tip seals, 119 Corning No. 7720 glass, 23, 105
BTH No. 9 glass, 22, 105, 112 Corning No. 7740 glass, 105
BTH No. 12 glass, 24 Corundum, 39
BTH No. 14 glass, 105 Cracking-off burner, 29
BTH No. 40 glass, 23, 112 Cracks, flame, 171
BTH No. 93 glass, 85 Cracks, mending, 80, 96
BTH No. 94 glass, 21 Crossfire burners, 29
Bubbles, glass-blowing of, 59 Crucibles, vitreous silica, 17
Bulbs at the end of a tube, 71 Cut-offs, mercury, 135
Bulbs, holes in, 72 Cutting a tube internally, 38
Bulbs in a tube, 70, 100 Cutting a tube with a flame, 56
Burns, treatment when slight, 46 Cutting a tube with a wheel, 44
Cylinder heads, 30
C. 9 glass, 22, 105, 112
C. 40 glass, 23, 112 Delayed fracture ofglass, 10, 11
Capillary tube working, 68 Density of glass, 14
Carbon tools, 33, 72, 175 Devitrification ofglass, 18
Carborundum, 39, 44, 82, 83 Devitrification ofsilica, 173
Carborundum grinding wheel, 45 Dewar seal, 101, 107, 121
Carius tubes, 118 Dewar vessels, 86, 123
Cements for glass, 84 Dial 36 glass, 22, 105

181
 SUBJECT INDEX
Dial 43 glass, 24, 112 Graded seal, 101
Dial 444 glass, 23, 108 Graded seal glasses, 25
Diamond powder as abrasive, 39, 83 Grinding glass, 40, 45, 82, 141
Diamonds, use of, 38, 84 Grinding wheel, 45
Diffusion through glass, 17
Disc seals, copper, 115 Hand torch, 29, 171
Discharge tube lamps, 154 Hand torch, value of, 4
Distillation flasks, 120 Handle for tubing, 55
Distillation splash head, 121 Hardness of glass, 14
Distinguishing types of glass, 26 Heating of glass, effects of, 16
Double tipping device, 171 Helium, diffusion through glass, 17
Drilling holes in glass, 83 Holders for tubes, 40
Dumet alloy wire, 111 Holes, blowing of, 59, 72
Holes, drilling of, 83
Elastic properties of glass, 13 Holes, filling of with rod, 64
Electrical resistance of glass, 14 Holes, finding of, 81
Electrodes in apparatus, 156 Hot spot for breaking tubes, 50
Electrolytic gas generator, 149 Housekeeper seal, 109, 113
Emery, 39 Hydrofluoric acid, for removing sur-
Equipment required for different pur- face layer, 19
poses, 46 Hydrofluoric acid, precautions with,
Extension of glass under tension, 13 158
Extraction apparatus, Soxhlet, 142 Hydrogen discharge lamps, 154
Hydrophobic glass surfaces, 18
FCN glass, 23, 112
Feet for vessels, 78 ‘In-place’ glass-blowing, 3
Fernico, 111 Internal seals, 73, 100
Fibres, delayed elastic effects in, 13 Internal seals, large, 95
Fibres, making silica ones, 175
File, triangular, 38 Jeweller’s rouge, 40
Filter pumps, 126 Joints between two big tubes, 93, 100
Fire-polishing, 13 Joints between two similar tubes, 60,
Firmasil glass, 22 100
Flame annealing, 170 Joints between tubes of different size,
Flame cracks, 171 64, 100
Flame gases, reaction with glass, 17 Joints, borosilicate glass to vitreous
Flamemaster hand torch, 29 silica. 25
Flanging, 72 Joints, borosilicate to soda glass, 25
Flaring, 72, 76, 100 Joints in closed systems, 168
Flaring tools, 31 Joints, Pyrex to Phoenix, 25
Flask clamps, 32 Joints, tungsten-sealing to Kovar-seal-
Flow meters for gas, 152 ing glasses, 24
Forceps, 32 Joints, Vitreosil, 175
Fractional distillation condenser, 131 Joints with capillary tube, 69
Fractional distillation receiver, 132 Junctions, 4 and 5 way, 68
Fractionating column, 132
Frozen surfaces, 82 King’s College condenser, 131
Knives, glass-cutting, 37
Gas analysis apparatus, 163 Knocking off bits of tubing, 51
Gas generator, electrolytic, 149 Kodial glass, 23, 112
Gas flow meters, 152 Kovar alloys, 111
Gauge, Bourdon, 159 Kovar-to—Pyrex seals, 102, 112
Gauge, glass spiral, 159
Gauge, McLeod, 76, 136 L.1 glass, 23
Gauges for measuring glass, 34 Lamps, discharge, 154
Glass, defined, 7 Lampworker, 28
Glass holders, 40 Lapping wheel, 44
Gloves, asbestos, 37 Large tube supports, 41
Gold, firing on to glass, 87 Lathe, glass-working, use of, 97, 113
182
 SUBJECT INDEX
Lead glass, 23, 108 Resistance, chemical, of glass, 15
Lead glass composition, 9 Resistance, electrical, of glass, 14
Leaks, finding of, 81 Ribbon burners, 29
Leaks, molecular flow, 158 Ring seals, 73
Lemington FCN glass, 23, 112 Rod, thin glass, 58, 64
Lemington H.26X. glass, 24 Rollers, 42
Lemington W.1. glass, 22,105, 112 Rotation of tubes in the flame, 52
Light transmission through glass, 15 Rubber caps, 34, 168
Liquid air traps, 125 Rubber stoppers, 35
Lower annealing temperature, 11
Lubrication of tools, 32 Sand for grinding, 39
Scissors for cutting hot glass, 52, 92
M.6. glass, 24 Sealed tubes for reactions, 118
Machine for glass-working, 97 Sealing glasses, 22
McLeod gauge, 76, 136 Sealing glasses for tungsten, 105
Mercury lamps, 155 Seals, multiple wire, 105
Mercury, traps for, 140 Semi-micro analysis apparatus, 162
Mercury vapour pumps, 144 Silica fibres, 13, 175
Mesh size for particles, 40 Silica, properties of, 25, 173
Metal layers on glass, 87 Silica, working of, 174
Metal-to-—glass seals, 102, 156 Silvering glass, 85, 87
Meters for gas flow, 152 Sintered discs, sealing in, 79
Mg point, 11 Soda-—boric oxide glass, 8
Mica windows fused to glass, 85 Soda—lime-silica glass composition, 9
Monax glass, 21, 118 Soda-silica glass, 8
Monax glass, ease of working, 26 Soda-to—Pyrex seal with copper tube,
Mouthpieces for blowing tubes, 36 115
Sodium Resistant NA.10 glass, 24
Network-forming ions, 8, 9 Softening temperatures, 11, 13
Network-modifying ions, 8, 9 Soldering glass after silvering, 88
Nicosel, 111 Soxhlet extraction apparatus, 142
Nilo, 111 Spectacles for glass-working, 37, 174
Nonex glass, 23, 105, 112 Spectrum tubes, 153
Spike, 31
Optical properties of glass, 15 Spindles, 56, 91, 99, 100
Spinning, 73, 78
Palladium, firing on to glass, 87 Spiral gauge, 159
Perforations in bulbs, 72 Spirals, 53
Phoenix glass, 22, 105, 106 Splash head for distillation, 121
Phosphorus pentoxide traps, 127 Strain in clamped apparatus, 169
Physical properties of glass, 9 Strain in glass, thermal, 19
Pinch seals, 106 Strain-viewer, 15, 43
Pinholes, 64, 74 77 Strength ofglass, 9, 10
Platinizing glass, 87, 110 Stress-optical coefficient, 15
Platinum, sealing into glass, 108 Structure of glass, 7
Polishing glass, 83 Supports for large tubes, 41, 42
Potassium, contamination of in glass, Surface properties of glass, 18
17 Swivel connection, 42, 70, 99
Prince Rupert’s drops, 20
Pulling a point, 57 T-joints, 65
Pulling a spear, 57 T-joints, capillary, 69
Pump, filter, 126 T-joints, large, 94
Pumps, circulating, 160 Table for glass-blowing, 28
Pumps, mercury vapour, 144 Tapers in tubes, 56, 65, 79, 101
Pyrex glass, 22, 105, 118 Telcoseal, 111
Tensile strength of glass, 10
Reaction vessels of glass, maturing of, Tetrahedra of SiO,, linking of, 7
18 Thermal capacity of glass, 14
Reamers, 31 Thermal conductivity of glass, 13

183
 SUBJECT INDEX
Thermal endurance of glass, 12 W.1 glass, 22, 105, 112
Thermal expansion of glass, 11 Water, evolution from glass, 16
Thermal strain in glass, 19 Wax for tools, 32
Thermostat regulators, 150 Waxes for joining glass, 84
Tools for glass-working, 30, 33, 73 Weathering of glass, 16, 18
Tools with glass-working machine,98 — Welding rod, 58, 64
Torsion fibres, silica, 175 Wembley ‘Amber Neutral’ glass, 24
Transformation point, 11 Wembley L.1. Lead glass, 23
Traps, liquid air, 125 Wembley M.6. ‘White Neutral’ glass, 24
Traps, phosphorus pentoxide, 127 Wembley X.8. glass, 21
Traps to catch mercury, 140 Wheel for cutting glass, 44, 51
Tubes, holding, 41 Wheel for grinding, 45
Tungsten, nickel tipping of, 108 Wheel for lapping, 44
Tungsten, sealing to Pyrex, 103 Windows, mica, 85
Windows, thin glass, 77
Upper annealing temperature, | 1 Wire holders, 104
Uranium glass, 105 Working various types of glass, 105

Vacuum flasks, 123 X.8. glass, 21

Valves, 140 X-rays, scattered by glass, 8
Viscosity of glass, 11, 13
Vitreosil, 25, 173 Y-joints, 68

184
7 a * = ee
 as
4

te

-
oy we -
ag AR doleyNat os
Oh" pane ae _—

We C700 -RARY,* OF

6 Yi4)
Pe {o008

RIFIED
Os."

CFT AL-MY SORE

1 TIAMAT
432?
Laboratory glass