United States Office of Air Quality

EPA-450/4-84-007i
Environmental Protection Planning And Standards
Agency Research Triangle Park, NC 27711
September 1985

AIR

EPA

LOCATING AND ESTIMATING AIR

EMISSIONS FROM SOURCES OF
PHOSGENE

L& E
 EPA-450/4-84-007i
September 1985

Locating and Estimating Air Emissions

From Sources of Phosgene

U. S. ENVIRONMENTAL PROTECTION AGENCY

Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711

September 1985
This report has been reviewed by the Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
and has been approved for publication. Mention of trade names and commercial products does not constitute endorsement
or recommendation for use.

EPA-454/R-98-013
 CONTENTS

Page
Figures iv
Tables v

1. Purpose of Document 1

2. Overview of Document Contents 3

3. Background 5

Properties of phosgene 5
Overview of phosgene production and use 7
Miscellaneous phosgene sources 7
References For Section 3 11

4. Phosgene Emission Sources 12

Phosgene production 12
Isocyanate production 20
Polycarbonate production 25
Herbicides and pesticides production 29
References For Section 4 35

5. Source Test Procedures 37

References for Section 5 39

Appendix
Phosgene Emissions Data A-1
References For Appendix A-14

iii
 FIGURES

Number Page

1 Chemical Use Tree for Phosgene 9

2 Basic Operations in a Phosgene Production Process 14

3a Flow Diagram of a Phosgene Emission Control System for

Merchant Phosgene Operations 16

3b Flow Diagram of a Phosgene Emission Control System for

Phosgene Production and Onsite Consumption 17

4 Basic Operations Used in the Production of Diamino Toluenes 21

5 Basic Operations Used in the Production of Toluene

Diisocyanate 22

6 Flow Diagram of a Phosgene Emission Control System 24

7 Basic Operations Used in the Production of Polycarbonates 27

8 Control System for Polycarbonate Production 28

9 Basic Operations Used in the Production of Phenyl Ureas 31

10 Emission Control System for Phenyl Urea Production 32

11 Sampling Train for the Measurement of Phosgene 38

iv
 TABLES

Number Page

1 Some Physical Properties of Phosgene 6

2 Companies That Produce Phosgene 8

3 Estimated Phosgene Emissions From a Hypothetical Phosgene

Plant 20

4 Estimated Phosgene Emissions From a Hypothetical Toluene

Diisocyanate Plant Using Phosgene Produced on Site 25

5 Estimated Phosgene Emissions From a Hypothetical Polycar-

bonate Plant Using Phosgene Produced on Site 29

6 Estimated Phosgene Emissions From a Hypothetical Herbicide

and Pesticide Plant Using Phosgene Produced on Site 34

A-1 Summary of Estimated Phosgene Emissions from Hypothetical

Phosgene and Phosgene Derivative Production Facilities A-3

A-2 Estimated Fugitive Phosgene Emissions From a Hypothetical

Phosgene Plant Producing 200 Million Pounds of Phosgene
Per Year A-8

A-3 Estimated Fugitive Phosgene Emissions From a Hypothetical

Toluene Diisocyanate Production Facility A-9

A-4 Estimated Fugitive Phosgene Emissions From a Hypothetical

Polycarbonate Production Facility A-10

A-5 Estimated Fugitive Phosgene Emissions From a Hypothetical

Herbicide and Pesticide Production Facility A-11

A-6 Process Fugitive Emission Factors for Plants A-12

v
 SECTION 1
PURPOSE OF DOCUMENT

The U.S. Environmental Protection Agency (EPA), States, and

local air pollution control agencies are becoming increasingly aware
of the presence of substances in the ambient air that may be toxic at
certain concentrations. This awareness, in turn, has led to attempts
to identify source/receptor relationships for these substances and to
develop control programs to regulate emissions. Unfortunately, very
little information is available on the ambient air concentrations of
these substances or on the sources that may be discharging them to
the atmosphere.
To assist groups interested in inventorying air emissions of
various potentially toxic substances, EPA is preparing a series of
documents that compiles available information on the sources and
emissions of these substances. This document specifically deals with
phosgene. Its intended audience includes Federal, State, and local
air pollution personnel and others who are interested in locating
potential emitters of phosgene and making gross estimates of air
emissions therefrom.
Because of the limited amounts of data available on phosgene
emissions, and because the configuration of many sources will not be
the same as those described herein, this document is best used as a
primer to inform air pollution personnel about 1) the types of
sources that may emit phosgene, 2) process variations and release
points that may be expected within these sources, and 3) available
emissions information indicating the potential for phosgene to be
released into the air from each operation.
The reader is strongly cautioned against using the emissions
information contained in this document in any attempt to develop an
exact assessment of emissions from any particular facility. Because
of insufficient data, no estimate can be made of the error that could
result when these factors are used to calculate emissions from any
given facility. It is possible, in some extreme cases, that

1
orders-of-magnitude differences could result between actual and
calculated emissions, depending on differences in source
configurations, control equipment, and operating practices. Thus, in
situations where an accurate assessment of phosgene emissions is
necessary, source-specific information should be obtained to confirm
the existence of particular emitting operations, the types and
effectiveness of control measures, and the impact of operating
practices. A source test and/or material balance should be
considered as the best means to determine air emissions directly from
an operation.

2
 SECTION 2

OVERVIEW OF DOCUMENT CONTENTS

As noted in Section 1, the purpose of this document is to assist

Federal, State, and local air pollution agencies and others who are
interested in locating potential air emitters of phosgene and making
gross estimates of air emissions therefrom. Because of the limited'
background data available, the information summarized in this
document does not and should not be assumed to represent the source
configuration or emissions associated with any particular facility.
This section provides an overview of the contents of this
document. It briefly outlines the nature, extent, and format of the
material presented in the remaining sections of this report.
Section 3 of this document provides a summary of the physical
and chemical characteristics of phosgene and an overview of its
production and uses. A chemical use tree summarizes the quantities
of phosgene consumed in various end use categories in the United
States. This background section presents a general perspective on
the nature of the substance and where it is manufactured and
consumed.
Section 4 of this document focuses on major industrial source
categories that may discharge. phosgene air emissions. The
production of phosgene is discussed, along with the use of phosgene
as an intermediate in the production of isocyanates, polycarbonates,
carbamates, chloroformates, and other esters of carbonic acid.
Example process descriptions and flow diagrams are provided and
potential emission points are identified for each of the major
industrial source categories discussed. Where the limited data
allow, emission estimates are presented that show the potential for
phosgene emissions before and after industry-applied controls.
Individual companies reported to be involved with either the
production or use of phosgene are named.
Section 5 summarizes available procedures for source sampling
and analysis of phosgene. Details are not presented, and the EPA

3
neither gives nor implies any endorsement of these sampling and
analysis procedures. Because the EPA has not yet made a general
evaluation of these methods, this document merely provides an
overview of applicable source sampling procedures and references for
the use of those interested in conducting source tests.
Companies that produce or use phosgene, State air control
agencies, and other authorities were contacted in an effort to locate
data representing measured phosgene emissions. Only one known direct
measurement has been made of phosgene emissions from industries that
produce or use phosgene. Aside from this single measurement, the
only emission data found were company engineering estimates. These
estimated emission levels are included in this report even though the
companies provided no bases for them.
Other information was used to obtain phosgene emission
estimates. For example, health effects and air monitoring programs
are discussed, but only to the extent that they were used to estimate
phosgene emissions. References are cited and the methodology is
discussed in sufficient detail to allow the reader to assess the
probable limitations of' these estimates. Additional background
information is included in Appendix A to assist the reader in
understanding the basis for all of the estimates presented in the
report.
Comments on the contents or usefulness of this document are
welcomed, as is any information on process descriptions, operating
practices, control measures, and emission information that would
enable EPA to improve its contents. All comments should be sent to:

Chief, Source Analysis Section (MD-14)

Air Management Technology Branch
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711

4
 SECTION 3
BACKGROUND

3.1 PROPERTIES OF PHOSGENE

Phosgene is a highly toxic, colorless gas that condenses at 0°C

to a fuming liquid. Impurities can discolor liquid phosgene and
cause it to turn a pale yellow to green color.1 The human nose can
detect its characteristic odor only briefly at the time of initial
exposure. At a concentration of about 0.5 ppm in the air, this odor
has been described as similar to that of new-mown hay or cut green
corn. At higher concentrations, the odor may be strong, stifling,
and unpleasant. A common decomposition product of chlorinated
compounds, phosgene is noncombustible. Its molecular formula is
COCl2, and it has the following planar structure.

The physical properties of phosgene (also known as carbonyl chloride,

carbon oxychloride, carbonic acid dichloride, chloroformyl chloride,
and combat gas)2 are presented in Table 1.
Phosgene is soluble in aromatic and aliphatic hydrocarbons,
chlorinated hydrocarbons, carbon tetrachloride, organic acids, and
esters, and it is only slightly soluble in water.1 It is removed
easily from solvents by heating or air blowing. Because the density
of phosgene is more than three times that of air, concentrated
emission plumes tend to settle to the ground and collect in low
areas.3
Phosgene decomposes to hydrogen chloride and carbon dioxide if
contaminated with water. Hence, wet phosgene is very corrosive and
poses an additional hazard from pressure buildup in closed
containers.3

5
 TABLE 1. SOME PHYSICAL PROPERTIES OF PHOSGENE1
Properties and characteristics Value
Molecular weight 98.92
Melting point, °C -127.84
Boiling point (at 101.3 kPa = 1 atm), °C 7.48
Density at 20°C, g/cm3 1.387
Vapor pressure at 20°C, kPaa 161.68
Vapor density (air = 1.0) 3.4
Critical temperature, °C 182.0
Density at critical point, g/cm3 0.52
Critical pressure, MPab 5.68
Latent heat of vaporization (at 7.5°C), J/gc 243
Molar heat capacity of liquid (at 7.5°C), J/Kc 100.8
Molar heat of formation, kJ
from elements 218
from CO and Cl2 108
a
To convert kPa to psi, multiply by 0.145.
b
To convert MPa to psi, multiply by 145.
c
To convert J to cal, divide by 4.184.

Phosgene reacts with many inorganic and organic reagents.1 The

reaction of oxides and sulfides of metals with phosgene at elevated
temperatures yields very pure chlorides. Phosphates and silicates of
metals react with phosgene at elevated temperatures and yield metal
chloride, phosphorus oxychloride, or silicon dioxide. Anhydrous
aluminum chloride forms a variety of complexes with phosgene:
Al2Cl6 • 5COCl2 at low temperatures, Al2Cl6 • 3COCl2 at 30°C, and Al2Cl6
• COCl2 at above 55°C. Ammonia reacts vigorously with phosgene in
solution; the products are urea, biuret, ammelide (a polymer of
urea), cyanuric acid, and, sometimes, cyamelide (a polymer of cyanic
acid).
Phosgene also reacts with a multitude of nitrogen, oxygen,
sulfur, and carbon compounds.1 Reaction with primary alkyl and aryl
amines yields carbamoyl chlorides, which can be dehydrohalogenated
readily to isocyanate (an intermediate in the manufacture of
polyurethane resins). Secondary amines also form carbamyl chlorides
when reacted with phosgene. The reaction of phosgene with amino
acids has been used to isolate and purify chloroformate derivatives.
Hydrazine reacts with phosgene to yield carbohydrazine. The reaction

6
of phosgene with alcohols, which yields esters, is commercially
important because it serves as a basis of widely used polymer systems
(polycarbonates).
3.2 OVERVIEW OF PHOSGENE PRODUCTION AND USE
Phosgene is used as a chemical intermediate (i.e., feedstock) in
the production of various commercial products. Most commercially
produced phosgene is used captively at the production sites in the
manufacture of other chemicals. Less than 2 percent of the phosgene
produced reaches the marketplace.3 Phosgene is currently produced in
the United States by 14 companies at 17 manufacturing facilities
(Table 2). As of January 1983 the annual estimated production
capacity was about one million tons.4
The chemical use tree in Figure 1 shows the current uses of
phosgene. The manufacture of isocyanates consumes about 85 percent
of the world's phosgene production.3,5 The primary use of phosgene is
in the production of toluene diisocyanate (TDI),3 a precursor of the
polyurethane resins used to make foams, elastomers, and coatings. A
rapidly growing use of phosgene is in the manufacture of
polymethylene polyphenylisocyanate (PMPPI), which is used in the
production of rigid polyurethane foams.3 The polycarbonate resins used
in appliance and electrical tool housings, electronic parts, and
break-resistant glazing are also phosgene-based. About 6 percent of
the phosgene production is consumed in the polycarbonate industry.5
The remaining 7 to 9 percent is used in the manufacture of
herbicides, pesticides, dyes, pharmaceuticals, and other specialty
chemicals. The latter include acyl chlorides, chloroformate esters
(intermediates in the production of ore flotation agents and
perfumes), diethyl carbonate, and dimethyl carbamy1 chloride.5

3.3 MISCELLANEOUS PHOSGENE SOURCES

3.3.1 Atmospheric Photoxidation of Chlorinated Hydrocarbons
Under laboratory conditions, phosgene has been shown to form
when chloroform, methylene chloride, perchloroethylene, and
trichloroethylene are irradiated with ultraviolet light. Ambient
phosgene measured in urban and nonurban air samples in California
appears to confirm the possibility of photochemical phosgene
formation in the troposphere.6,7

7
 TABLE 2. COMPANIES THAT PRODUCE PHOSGENE4
Company Location End Product

BASF Wyandotte Corp. Geismar, La. Isocyanates

Dow Chemical Co. Freeport, Tex. Isocyanates
E.I. duPont de Nemours & Co., Isocyanates,
Inc. Deepwater Point, N.J. carbamates
Essex Chemical Co. Baltimore, Md. Pesticides
General Electric Co. Mount Vernon, Ind. Polycarbonate
ICI Americas Geismar, La. Isocyanates
Laurel Industries La Porte, Tex. Merchant phosgene,
chloroformates
Mobay Chemical Co. Cedar Bayou, Tex. Isocyanates
New Martinsville, W. Va. Isocyanates
Olin Corp. Lake Charles, La. Isocyanates
Moundsville, W. Va. Isocyanates
PPG Industries Barberton, Ohio Pesticides
Stauffer Chemical Co. Cold Creek, Ala. Pesticides
St. Gabriel, La. Pesticides
Upjohn Co. La Porte, Tex. Isocyanates
Union Carbide Corp. Institute, W. Va Isocyanates
Van De Mark Chemical Co., Inc. Lockport, N.Y. Merchant phosgene
Note: This listing is subject to change as market conditions change,
facility ownership changes, plants are closed down, etc. The reader
should verify the existence of particular facilities by consulting
current listings and/or the plants themselves. The level of phosgene
emissions from any given facility is a function of variables such as
capacity, throughput, and control measures. It should be determined
through direct contacts with plant personnel.

It is difficult, however, to assess the amount of phosgene

formed in the atmosphere. Although phosgene is evidently one of the
photolysis products of a number of high-volume chlorinated
hydrocarbon solvents, the role and significance of each solvent, the
half-life of phosgene in the air, and the atmospheric fate of
phosgene are not well understood. The quantities of phosgene
produced by photolysis, however, may be much higher than those
emitted by the chemical industry.7

8
3.3.2 Thermal and Ultraviolet Decomposition of Chlorinated
Hydrocarbons
Phosgene can be produced from the heating and resulting
decomposition of many chlorinated hydrocarbons, including methylene
chloride, monochlorobenzene, and dichlorobenzene (used as solvents in
polymerization reactions involving phosgene), carbon tetrachloride,
chloroform, ethyl chloride, polyvinyl chloride, and various
fluorocarbons (Freons).3,8 When heated, chlorinated hydrocarbon
vapors react with oxygen or water to form chlorine, hydrogen
chloride, phosgene, and other toxic substances; therefore,
incineration used for the control of volatile organic compound
emissions can become an inadvertent source of phosgene emissions. A
properly operated caustic scrubber can reduce phosgene emissions in
exhaust gases from the incineration of chlorocarbons.
The potential for phosgene generation by chlorocarbon
decomposition exists at chlorocarbon producing facilities,
metallurgical operations, drycleaning and degreasing facilities,
certain types of industrial fires, and wherever solvents contact heat
or ultraviolet light.

10
 REFERENCES FOR SECTION 3

1. Kirk-Othmer. Encyclopedia of Chemical Technology. 3rd Ed.,

Volume 17. Wiley Interscience Publication, New York. 1979. pp.
416-425.

2. U.S. Department of Health, Education, and Welfare. Occupational

Diseases - A Guide To Their Recognition. June 1977.

3. U.S. Environmental Protection Agency. Office of Pesticides and

Toxic Substances. Phosgene. Chemical Hazard Information Profile.
June 1977. pp. 226-236.

4. SRI International. 1983 Directory of Chemical Producers, USA.

1983.

5. Faith, Keyes, and Clark's Industrial Chemicals. Phosgene. 4th

Ed. John Wiley & Sons. November 1975. pp. 624-627.

6. Smith, A. J. Measurements of Some Potentially Hazardous Organic

Chemicals in Urban Environments. Atmospheric Environment,
15:601-612, 1981. Pergamon Press, Ltd., Great Britian.

7. Singh, H. B. Phosgene in the Ambient Air. Nature, 264:428-429,

December 2, 1976.

8. Bjerre, A. Mathematical Modelling in the Hazard Assessment of

Substances Forming Toxic Decomposition Products. The Example of
Carbon Tetrachloride. Annals of Occupational Hygiene, 24(2)
175-183, 1981. Pergamon Press, Ltd., Great Britain.

11
 SECTION 4
PHOSGENE EMISSION SOURCES

This section describes industrial processes that are sources of

phosgene emissions, including direct phosgene production and the use
of phosgene as an intermediate in the production of isocyanates,
polycarbonates, carbamates, thiocarbamates, and phenyl ureas.
Included are process descriptions and emission estimates for
hypothetical facilities involved in the making or use of phosgene.
Because the production of chloroformates and chlorocarbonates
represents a minor end use of phosgene, this is not described. The
processes and the phosgene emissions and controls associated with the
production of these chemicals, however, are similar to those
described for polycarbonate production (Section 4.3).
Most phosgene is produced for onsite consumption, with merchant
phosgene accounting for less than 2 percent of total production.1
Hence, phosgene production operations will generally be found at
facilities engaged in the manufacture of isocyanates, polycarbonates,
carbonates, etc. The production of phosgene is discussed in
Section 4.1, and the use of phosgene as a chemical intermedate, in
the following sections. For economy, the discussion of phosgene
production is not repeated in each of the sections in which its
intermediate Uses are discussed. Instead, the reader should refer
back to Section 4.1.

4.1 PHOSGENE PRODUCTION

4.1.1 Process Description
Phosgene is produced by the reaction of carbon monoxide and
chlorine over a highly absorptive activated charcoal catalyst at
200°C and 14 to 28 kPa (2 to 4 psig):

CO + Cl2 COCl2 (1)

12
The reaction is rapid and exothermic. Because phosgene decomposes at
temperatures above 300°C, a water-cooled reactor is used to remove
the excess heat. Figure 2 presents a flow diagram of the production
of phosgene from carbon monoxide and chlorine.
Phosgene production is continuous and highly automated and
proceeds as follows:1
• Preparation and purification of carbon monoxide
• Preparation and purification of chlorine
• Metering and mixing of reactants
• Purification and condensation of phosgene
• Control of phosgene emissions to assure worker and
environmental safety

Carbon monoxide may be manufactured either by the reduction of

carbon dioxide over coal or carbon or by the controlled oxidation of
hydrocarbon fuels. Chlorine is usually purchased from manufacturers
who use the electrolysis of sodium chloride brines (caustic chlorine
process). These reactants must be pure. Objectionable impurities
include water (which can produce hydrogen chloride, hydrocarbons,
and hydrogen that may trigger a reaction between chlorine and steel
and destroy the equipment), sulfides (which can produce undesirable
sulfur chlorides), and other impurities (which could deactivate the
catalyst).
As shown in Figure 2, carbon monoxide (Stream 1) and chlorine
(Stream 2) are mixed either in equimolar proportions or with a small
excess of carbon monoxide to ensure complete conversion of the
chlorine. The product gases (Stream 3) are condensed, the liquid
phosgene (Stream 4) is sent to storage, and the remaining gases
(Stream 5) are scrubbed with a hydrocarbon solvent to remove residual
phosgene. Uncondensed phosgene and the solvent that is used in the
scrubber may be used for subsequent processing (e.g., in the
production of isocyanate).
The liquid phosgene is stored in pressurized steel tanks. A
typical precautionary measure is to store the material in two tanks,
neither of which is filled to more than half of its capacity.2 This
allows the transfer of the phosgene to either tank in case a leak
develops in one of the tanks or its piping system.

13
4.1.2 Emissions and Controls
Phosgene emissions fall into three categories process emissions
(including storage tank vents which are exhausted to the control
system), fugitive emissions, and emissions that occur during process
upsets. Each type is discussed, estimates are presented, and
controls are explained. The development of emission estimates is
discussed further in Appendix A.
Process Emissions--
All process emissions from phosgene production and utilization
are typically routed to a caustic scrubber. The caustic scrubber is
the control of choice because phosgene is rapidly and completely
destroyed by aqueous sodium hydroxide, as shown in the following
reaction:
COCl2 + 4NaOH 2NaCl + Na2CO3 + 2H2O (2)
The sodium hydroxide concentration should be maintained at
between 3 and 8 weight percent, and the sodium chloride and sodium
carbonate must not precipitate and clog the reactor. It should be
noted that the solubility of these components is appreciably lower in
caustic solution than in water.3 These requirements are met by
continually replacing the solution in the scrubber with fresh caustic
solution. Data generated by the U.S. Amy indicate that a two stage
scrubber can reduce phosgene emissions to below 0.5 ppm by volume.3
This study demonstrated that phosgene control is severely reduced if
1) the phosgene flow to the scrubber exceeds the design capacity of
the scrubber, or 2) the caustic concentration in the scrubber is not
maintained between 3 and 8 weight percent. The design of the
scrubber therefore must be such that it can accommodate any phosgene
surge. It is estimated that a phosgene plant producing 200 million
pounds per year would emit 300 pounds per year after scrubbing.
Figure 3a and Figure 3b present flow diagrams for phosgene
emission control systems. Figure 3a shows a possible control system
for a plant that produces phosgene for sale without any subsequent
onsite processing. Control can be achieved with a single caustic
scrubber. Figure 3b shows an emission control system for a plant
that produces phosgene and then processes it on site to produce other
products. These subsequent operations generate additional emissions
that must be controlled.

15
16
 In all commercial phosgene processes, the chlorine atoms react
with active hydrogen atoms to produce hydrogen chloride (HCl).
Hydrogen chloride is an acid gas, and like phosgene, it can be
controlled with a caustic scrubber; however, it is usually desirable
for a water scrubber to precede the caustic scrubber. This not only
permits HCl to be recovered as a byproduct, but also reduces the
loadings to the caustic scrubber. The production of toluene
diisocyanate includes a nitration step that generates acidic nitrogen
and sulfur oxide emissions, which would be routed to a caustic
scrubber. These reactions take place in an organic solvent medium,
and the solvent is a source of volatile organic compound (VOC)
emissions. Solvents include chlorinated compounds such as methylene
chloride, monochlorobenzene, and ortho-dichlorobenzene, as well as
pyridine, xylene, methanol, and aliphatic hydrocarbons. Whereas VOC
can be controlled by incineration, the incineration of chlorinated
VOC can produce hydrogen chloride, chlorine, and phosgene emissions.
Therefore, a second caustic scrubber is required in series with the
incinerator.
Fugitive Emissions--
Pumps and valves are the major sourcer of fugitive phosgene
emissions at facilities where phosgene is produced or used. No
compressors are used on phosgene process flows, and emissions from
flanges and drains are considered negligible. Because phosgene is
known to be very toxic, industry typically takes measures to minimize
fugitive emissions. These measures include:4
1. Welding pipe joints and monitoring of the quality of all
welds.

2. Enclosing the reactor and condenser in a negative-pressure

building, and venting the exhaust to the caustic scrubber.

3. Employing special construction materials and techniques

for all piping and valves handling hazardous or corrosive
substances. Installing plugs or caps on all open-ended
lines and plug valves to minimize stem leakage

4. Enclosing pump couplings and drivers on all pumps handling

phosgene. Using special mechanical seals on other pumps
(dual seals with barrier fluids) and closed purge sampling
systems.

5. Continuous area and individual monitoring and the

installation of phosgene-release alarms.

6. Establishing procedures and training for prompt response

to phosgene leaks and releases.

18
 7. Practicing intensive preventive maintenance during plant
shutdowns and turnarounds. (A turnaround is a planned
shutdown to allow equipment to be used to make a different
product.)

All plants that produce and use phosgene have individual and/or
area monitors to detect excessive phosgene levels in plant air.5-9
Based on the sensitivity of these alarm systems, the fugitive
phosgene emissions from a phosgene plant producing 200 million lb/yr
are estimated to be 120 lb/yr or 0.6 lb/million lb of phosgene.
Alternatively, this emission rate can be estimated by counting
the valves, pumps, and flanges at a typical plant and applying the
fugitive leak rates and control efficiencies developed by the
Environmental Protection Agency for the synthetic organic chemicals
manufacturing industry (SOCMI).10 This approach yields an estimated
fugitive emission rate of 6600 lb/yr, or 33 lb/million lb of phosgene
produced. Fugitive emission estimates by both the monitoring
approach and the equipment count/emission factor approach are
presented hereafter as a range for each plant. It is observed,
however, that fugitive controls used in phosgene production are
actually much more stringent than those reported in Reference 10.
In phosgene plants where the reactor and condenser are enclosed
in a negative-pressure building and the exhaust is vented through the
caustic scrubber, 99 percent of the fugitive phosgene emissions are
destroyed in the caustic scrubber, and phosgene emissions are further
reduced to an estimated 1 to 66 lb/yr (0.005 to
0.3 lb/million lb phosgene produced). Fugitive emission estimates
are derived in Appendix A.

Process Upsets--
Some phosgene emissions result from process upsets, e.g., pump
failures and inadvertent opening of the wrong valve. Based on 15
process upset reports during a recent 6-year period, three Texas
plants released a total of 900 pounds of phosgene. Phosgene releases
in the 15 episodes ranged from 1 to 220 pounds. The stated amount
released in each case usually represented an estimate, and it often
was not clear what part of the process was involved. For example, one
upset was reported as a ruptured line. Based on the size of these
three plants, the total phosgene release, and the number of releases
of phosgene (and assuming all releases were reported), it is
estimated that a plant producing 200 millions pounds per year of

19
phosgene will average one process upset per year during which 50
pounds of phosgene is released (0.0005 lb/ton phosgene produced).
Total Phosgene Emissions--
Total phosgene emissions for a plant that produces 200 million
pounds of phosgene a year is estimated to be 470 to 7000 pounds,
comprising 300 pounds from the process vents, 120 to 6600 pounds in
the form of fugitive emissions, and 50 pounds as a result of process
upsets. This is equivalent to 2.35 to 34.8 lb phosgene emissions per
million pounds of phosgene produced. Table 3 presents a summary of
phosgene emissions from phosgene production.

TABLE 3. ESTIMATED PHOSGENE EMISSIONS FROM A HYPOTHETICAL

PHOSGENE PLANTa(lb/yr)
Emission factor,
Process vent Fugitive Emissions due Total lb per million
emissionsb emissionsc to upsets emissions lb processed

300 120 to 6600 50 470 to 6950 2.35 to 34.8

a
based on facilities with a hypothetical rate of 200 million pounds of
phosgene production per year.
b
Incinerator and scrubber exhausts.
c
Estimated fugitive emissions would be reduced by a factor of 100 if
reactor and condenser are enclosed in a negative-pressure building and
vented through a caustic scrubber.

11,12
4.2 ISOCYANATE PRODUCTION
4.2.1 Process Descriptions
The commercial production of aromatic and aliphatic isocyanates
is accomplished through the reaction of amines and phosgene.
Aromatic isocyanates are more important commercially than aliphatic
isocyanates. In 1978, the estimated world production of the two
principal aromatic isocyanates was 635,000 metric tons of toluene
diisocyanate (TDI) and 454,000 metric tons of diphenyl methane-4,4'-
diisocyanate (MDI).
Aromatic Diisocyanate Production--
Figure 4 and Figure 5 show the reaction sequences for the
production of the major aromatic diisocyanates. Toluene (Stream 1)
is the starting material for the production of TDI and 3,3'-

20
dimethyldiphenylmethane 4,4'-diisocyanate. Toluene is converted to
dinitrotoluene with a mixture of solvent (Stream 2) and sulfuric and
nitric acids (Stream 3). (The sulfuric acid ties up water formed in
the reaction.) The proportions of dinitrotoluene isomers prepared
can differ depending upon operating conditions. Varying the mixture
of isomers allows flexibility in the properties of the diisocyanate
polymers:
! When dinitrotoluene is made without separating the
mononitrotoluene isomers (Stream 5), the resulting mixture
is 80 percent 2,4-isomer and 20 percent 2,6-isomer.

! The nitration reaction can be interrupted after the

formation of the mononitrotoluenes (Stream 4), and the
ortho- and paramononitrotoluenes (Stream 6 and Stream 7)
can be separated by distillation. Nitration of
paranitrotoluene yields 100 percent 2,4-dinitrotoluene.

! Nitration of orthonitrotoluene yields a mixture of 65

percent 2,4-dinitrotoluene and 35 percent 2,6-
dinitrotoluene (Stream 8). Alternatively,
orthonitrotoluene can be reduced to orthoaminotoluene (not
shown in Figure 4 and Figure 5) and form benzidine through
the benzidene rearrangement. Benzidene can then be
phosgenated to form 3,3'-tolidene 4,4'-diisocyanate.

! The nitrotoluenes (Stream 8) are reduced to the amines

with hydrogen (Stream 9). The aminotoluenes (Stream 10)
react with phosgene (Stream 11 and Stream 13) to form the
isocyanates in a two-step phosgenation process shown in
Figure 5. Phosgene is first added at a temperature range
of -20° to 60°C and again at 100° to 200°C.
Polymerization takes place immediately, but some monomers
remain. The product (Stream 14) is then distilled to
remove and recover solvent and unreacted monomer.

Aliphatic Diisocyanate Production--

Diphenylmethane-4',4'-diisocyanate, an aliphatic diisocyanate,
is produced by reacting two moles of analine with one mole of
formaldehyde, followed by phosgenation of the diamine and
polymerization of the resulting diisocyanate. The reaction and
process conditions are similar to those for the formation of TDI.
The phosgenation and polymerization reactions are carried out in
a solvent medium. Although the role of the solvent is unknown, the
choice of solvent influences the rate and extent of the reaction.12
The solvent must dissolve the amines, phosgene, isocyanate monomers,
and at least the lower molecular weight polymers. Typical solvents
are aromatic compounds such as xylene, monochlorobenzene, and o-
23
dichlorobenzene. Aliphatic solvents such as methanol or hydrocarbons
may be added to precipitate the polymer from solution.
Process economics require that the solvents be recovered and
recycled. All phosgene used in the process reacts with active
hydrogen atoms to form hydrogen chloride, which is recovered and
either sold or decomposed by electrolysis to yield chlorine (used in
the production of phosgene) and hydrogen (used to reduce nitro
compounds to amines).
4.2.2 Emissions and Controls
Potential process emissions from the production of isocyanate
include phosgene, hydrogen chloride, aromatic and aliphatic solvents,
aromatic amines, aromatic nitro compounds, isocyanates, nitrogen
oxides, and sulfur oxides. Because these emissions include a number
of toxic and corrosive chemicals, controls are necessary. A typical
control system (as shown in Figure 6) would include:
1. A water Scrubber to remove and recover hydrogen-chloride.

2. A caustic scrubber to provide removal of VOC and COCl2

from the water scrubber as well as to remove VOC from the
nitration and distillation processes.

3. An incinerator for volatile organic compounds:

4. A second caustic scrubber for treatment of the incinerator

exhaust to remove residues from the combustion of
chlorinated hydrocarbons.

24
 As shown in Table 4, total annual phosgene emissions, after
controls, are estimated to be 705 to 9760 pounds for a plant
producing 200 million pounds of phosgene and using it on site in the
production of TDI. This estimate includes emissions from phosgene
production, which were developed in the preceding section and are not
reported here. Almost always, phosgene is produced at the same plant
where phosgene derivatives, such as isocyanates, are produced.
Derivation of phosgene emissions from TDI production is documented in
Appendix A.

TABLE 4. ESTIMATED PHOSGENE EMISSIONS FROM A HYPOTHETICAL T0LUENE

DIISOCYANATE PLANT USING PHOSGENE PRODUCED ON SITEa
(lb/yr except as noted)
Emission
factor, lb
Emissions per million
Process ventb Fugitive due to Total lb phosgene
emissions emissionsc upsets emissions produced

Phosgene 300 120 to 6600 50 470 to 6950 2.35 to 34.8

production

Toluene 150 60 to 2640 25 235 to 2820 1.18 to 14.1

diisocyanate
production
Total plant 450 180 to 9240 75 705 to 9760 3.53 to 48.9
a
Based on facilities with a hypothetical rate of 200 million pounds of
phosgene production per year.
b
Incinerator and scrubber exhausts.
c
Estimated fugitive emissions would be reduced by a factor of 100 if reactor
and condenser are enclosed in a negative-pressure building and vented
through a caustic scrubber.

Derivation of phosgene emissions from TDI production is documented in the

appendix.

4.3 POLYCARBONATE PRODUCTION14,15

4.3.1 Process Descriptions

In general, polycarbonates are formed by the reaction of a diol
(a molecule with two alcohol groups) and a carbonic acid derivative
(phosgene is the chloride of carbonic acid). Because most commercial

25
polycarbonates are derived from the reaction of bisphenol A [2,2
bis(4-hydroxyphenyl) propane] and phosgene, this process is discussed
here. Polycarbonates can also be formed by the reaction of other
aromatic or aliphatic diols (dihydroxy alcohols) and phosgene.
The sequence of reactions for producing polycarbonate from
bisphenol A and phosgene is presented in Figure 7. The basic
reaction is:

Bisphenol A and 1 to 3 mode percent monofunctional phenol (to

control the molecular weight of the carbonate polymer) are dissolved
or slurried in aqueous sodium hydroxide (Stream 1). A solvent and a
tertiary amine catalyst (such as pyridine) are added, phosgene gas is
bubbled in (Stream 2), and the resulting mixture is vigorously
stirred. Additional caustic is added as needed to keep the mixture
basic. As the polymer is formed, it is filtrated in the solvent
layer. When the reaction is completed, the aqueous phase (Stream 3)
contains sodium chloride, sodium carbonate (formed by a side reaction
of phosgene and caustic), and possibly traces of phenols. The
organic phase (Stream 4) is a polymer solution containing
polycarbonate, residual catalyst, and solvent. This polymer solution
is washed with water, extracted with acid to remove residual catalyst
(Stream 5), and washed again (Stream 6) with water until neutral
(Stream 7). The solvent's then stripped from the polymer by
evaporation (Stream 8 and Stream 9). These reactions take place at
or about room temperature. The reaction may also be carried out in a
solvent medium in which a large quantity of pyrldine is used to tie
up the hydrogen chloride formed by the reaction of phosgene and
bisphenol A.
Possible solvents (Stream 1) include methylene chloride,
aromatic liquids, chlorinated aromatic liquids, and aliphatic
chlorohydrocarbons. Process economics require the recovery and
recycling of all organic solvents.
Most of the processing conditions (including reaction
conditions) are closely guarded secrets, particularly with regard to
26
processes for isolating the polymer from the solvents. Possible
procedures for separating polymers and solvent include nonsolvent
precipitation, spray-drying, multistep total solvent evaporation, and
partial or complete solvent removal in boiling water followed by
oven-drying. Total solvent evaporation is effected by the use of
wiped-film evaporators and multiport vacuum-vented extruders. The
total removal of a low-boiling chlorinated hydrocarbon (such as
methylene chloride) from a very high-viscosity, high-melting polymer
is complicated by two factors:
1) foam formation at low temperatures impedes heat and mass transfer,
and 2) the solvent can react with water or thermally decompose and
cause product contamination.
4.3.2 Emissions and Controls
Potential emissions from phosgene and polycarbonate production
include phosgene, hydrogen chloride, aromatic and aliphatic
hydrocarbons (some of which are chlorinated and could produce
phosgene on incineration), and phenols. Emission controls for the
reactors and solvent recovery systems include incinerators and
caustic scrubbers. These controls are similar to those used for
isocyanate production.

The wastewater from the reactor, acid wash, and water wash (see
Figure 7) is acidic and may contain small amounts of organic
compounds. These compounds would probably have high molecular
weights, have low water solubility, be nonvolatile, and thus would
not be a significant source of air emissions.

28
 As shown in Table 5, total annual phosgene emissions are
estimated to be 580 to 8190 pounds for a plant producing 200 million
pounds of phosgene and using it on site to produce polycarbonates.
Derivation of these emission estimates is documented in Appendix A.

TABLE 5. ESTIMATED PHOSGENE EMISSIONS FROM A HYPOTHETICAL POLYCARBONATE

PLANT USING PHOSGENE PRODUCED ON SITEa (lb/yr)
Emission
factor, lb
Emissions per million
Process vent Fugitive due to Total lb phosgene
emissionsb emissionsc upsets emissions produced

Phosgene 300 120 to 6600 50 470 to 6950 2.35 to 34.8

production

Polycarbonate 70 30 to 1160 10 110 to 1240 0.55 to 6.2

production
Total plant 370 150 to 7760 60 580 to 8190 2.90 to 41.0
a
Based on facilities with a hypothetical rate of 200 million pounds of
phosgene production per year.
b
Incinerator and scrubber exhausts.
c
Estimated fugitive emissions would be reduced by a factor of 100 if reactor
and condenser are enclosed in a negative-pressure building and vented
through a caustic scrubber.

4.4 HERBICIDES AND PESTICIDES PRODUCTION16

Phosgene is used in the synthesis of some pesticides and
herbicides. The active chlorine atoms of phosgene react with hydrogen
to produce hydrogen chloride, and the carbonyl group (C=O) is added
to the reacting molecule. The three general classes of chemicals
comprising herbicides and pesticides are phenyl ureas, carbamates,
and thiocarbamates. Section 4.4.1, Section 4.4.2, and Section 4.4.3
discuss the production of each of these classes of herbicides and
pesticides. The phosgene emissions from the production of each of
these are similar. (Phosgene emission estimates for herbicide and
pesticide production are presented in Section 4.4.4.)

29
4.4.1 Production of Substituted Phenyl Ureas
The herbicidal activity of substituted phenyl ureas was
discovered in the late 1940's. There are currently 20 to 25 phenyl
ureas on the commercial market. Although initially developed as
industrial herbicides, they also have been used in selective
agricultural applications.
A general reaction for the substituted phenyl ureas (e.g.,
monuron) can be written as follows:

For monuron, x is hydrogen, y is -OCl, and R and R' are CH3 . Other
substituted phenyl ureas have been prepared and are in use with
different substituents for x, y, R, and R'.
Figure 9 presents the basic operations used in substituted
phenal urea production, and Figure 10 presents a flow diagram of a
control system for such a process. The synthesis of monuron [3-(p-
chlorophenyl)-1,1-dimethyl urea] is typical of the general commercial
method used for the production of substituted phenyl ureas. For this
synthesis, the p-chloroaniline in dioxane or some other inert solvent
(Stream 1 and Stream 2) reacts with anhydrous hydrogen chloride and
phosgene at 70° to 75°C (Stream 3) to form p-chlorophenyl isocyanate
(Stream 4). This aromatic isocyanate further reacts with
dimethylamine at 25°C to give monuron (Stream 5), which is then
separated from the solvent by precipitation and evaporation. (see
Reaction 4.)

30
4.4.2 Carbamates Process
Carbamates are used as herbicides, insecticides, and medicinals,
and for the control of nematodes, mites, and mollusks. They are
obtained either by the reaction of a substituted analine with a
chloroformate ester or the reaction of an isocyanate with an alcohol.
The chloroformate ester is made by reacting phosgene with an alcohol
(ROH + COCl2 --> ClCOOR + HCl), and the isocyanate is made by
reacting phosgene with an amine (as described in Section 4.2). The
basic reactions are:

Different R groups and substitutions on the benzene ring yield a

variety of useful products.
4.4.3 Thiocarbamates Process
Thiocarbamates are used primarily as herbicides, but some have
value as fungicides. The phosgene reaction is the same as in the
other processes: the chlorine atoms react with hydrogen to form
hydrogen chloride, and the carbonyl group is added to the molecule.
Thiocarbamates are formed in a two-stage reaction. The first
stage is the reaction of phosgene and a secondary amine to yield a
carbamyl chloride:
R2 NH + COCl2 --> R2 NCOCl + HCl (7)
followed by reaction with a thiol to yield a carbamate:
R2 NCOCl + NaSR' --> R2 NCOSR' + NaCl (8)
Alternatively, the secondary amine can react with an alkyl
chlorothiol formate in the presence of a proton acceptor to tie up
the HCl formed in the reaction:

RSCOCl + NHR'R'' RSCONR'R'' + HCl. (9)

Varying the R, R', and R'' groups will produce different

thiocarbamates.

33
4.4.4 Emissions and Controls
Generalized flow diagrams for herbicide and pesticide production
and emission controls are shown in Figure 9 and Figure 10,
respectively. State-of-the-art controls include incineration (to
control VOC emissions) and caustic scrubbers (to control phosgene and
HCl produced either by reactions involving phosgene or by
incineration of the VOCs).
As shown in Table 6, total annual phosgene emissions are
estimated to be 580 to 8220 pounds for a plant producing 200 million
pounds of phosgene on site and using it to produce herbicides and
pesticides. Derivation of this emission estimate is documented in
Appendix A.

TABLE 6. ESTIMATED PHOSGENE EMISSIONS FROM A HYPOTHETICAL HERBICIDE

AND PESTICIDE PLANT USING PHOSGENE PRODUCED ON SITEa
(lb/yr)
Emission
factor, lb
Emissions per million
Process vent Fugitive due to Total lb phosgene
emissionsb emlssionsc upsets emissions produced

Phosgene 300 12O to 6600 50 470 to 6950 2.35 to 34.8

production

Herbicide and 70 30 to 1190 10 110 to 1270 0.55 to 6.4

pesticide
production
Total plant 370 150 to 7790 60 580 to 8220 2.90 to 41.1

a Based on facilities with a hypothetical rate of 200 million pounds of

phosgene production per year.

b Incinerator and scrubber exhausts.

c Estimated fugitive emissions would be reduced by a factor of 100 if reactor

and condenser are enclosed in a negative-pressure building and vented
through a caustic scrubber.

34
4.5 REFERENCES

1. Kirk-Othmer. Encyclopedia of Chemical Technology.3rd Ed., Volume 17.

Wiley Interscience Publication, New York. 1979. pp. 416-425.

2. Personal communication with R. L. Matherne, Ethyl Corporation, Baton

Rouge, Louisiana, February 7, 1984. Based on Mr. Matherne's
experience at BASF Wyandott.

3. Kistner, S., et. al. A Caustic Scrubber System For The Control Of
Phosgene Emissions Design, Testing, and Performance. Journal of the
Air Pollution Control Association, 28(7):673-676,1978.

4. Enviro Control, Inc. Assessment of Engineering Control Monitoring

Equipment. Volume I. Prepared for the National Institute for
Occupational Safety and Health, Cincinnati, Ohio, PB83-15269.
Contract No. 210-79-0011, June 1981.

5. Personal communication with Marshall Anderson, General Electric

Corporation, Mount Vernon, Indiana, February 8, 1984.

6. Personal communication with George Dinser, Mobay Chemical Company-

Cedar Bayou, Texas, February 8, 1984.

7. Personal communication with Mark Kenne, ICI Americas, Rubicon

Chemical Division, Geismar, Louisiana, February 14, 1984.

8. Personal communication with Jerry Neal, PPG Industries; LaPorte,

Texas, February 14, 1984.

9. Personal communication with George Flores, Dow Chemical Co.,

Freeport, Texas, February 15, 1984.

10. U.S. Environmental Protection Agency. Fugitive Emission Sources of

Organic Compounds - Additional Information on Emissions, Emission
Reductions, and Costs. EPA-450/3-83-010. Research Triangle Park,
North Carolina. April 1982.

35
11. Personal communication with R. A. Campbell, Plant Manager, Olin
Chemicals Group, Moundville, West Virginia, May 3, 1984.

12. Kirk-Othmer. Encyclopedia of Chemical Technology. 3rd Ed., Volume 13.

Wiley Interscience Publication, New York. 1979. pp. 789-808.

13. Encyclopedia of Polymer Science and Technology. Volume 11. Wiley

Interscience Publication, New York. 1979. pp. 507-525.

14. Kirk-Othmer. Encyclopedia of Chemical Technology. 3rd Ed., Volume 18.

Wiley Interscience Publication, New York. 1979. pp. 479-492.

15. Encyclopedia of Polymer Science and Technology. Volume 10. Wiley

Interscience Publication, New York. 1979. pp. 71O-764.

16. Kirk-Othmer. Encyclopedia of Chemical Technology. 3rd Ed., Volume 12.

Wiley Interscience Publication, New York. 1979. pp. 319-326.

36
 SECTION 5
SOURCE TEST PROCEDURES

No EPA Reference Method has been established for measuring phosgene;

however, the NIOSH Manual of Analytical Methods contains a proposed method
for the collection and analysis of phosgene in air.1 This method involves
the reaction of phosgene with a solution of 4,4'-nitrobenzyl pyridine in
diethyl phthalate to produce a red color. The color reaction is measured
in a photometer.
In the NIOSH method, exhaust or air containing phosgene is passed
through midget impingers (Figure 11) containing a color reagent made up of
2.5 g 4,4'-nitrobenzyl pyridine, 5 g N-phenylbenzylamine, and 992.5 g
diethyl phthalate. Fifty liters are drawn through the impingers if the
phosgene level is in the range of 0.04 to 1 ppm, whereas a volume of 25
liters is drawn for phosgene levels above 1 ppm. The phosgene reacts with
the reagent to form a red color. The N-phenylbenzylamine in the solution
stabilizes the color and increases the sensitivity. The resulting red
color should be measured with a photometer within 9 hours of sampling.
Sampling efficiency is 99 percent or better.
Interfering compounds are acid chlorides, alkyl and aryl derivatives
(which are substituted by active halogen atoms), and sulfate esters. If
necessary, most of these interfering compounds can be removed in a
prescrubber containing an inert solvent, such as Freon-113, that has been
cooled by an ice bath. This method has not been validated by EPA.2

37
5.1 REFERENCES

1. National Institute for Occupational Safety and Health. NIOSH Manual

of Analytical Methods. Part 1 - NIOSH Monitoring Methods, Volume 1.
U.S. Department of Health, Education, and Welfare, Cincinnati, Ohio.
April 1977.

2. Knoll, J. Intra-agency memorandum to T. Lahre, U.S. Environmental

Protection Agency, Air Management Technology Branch, Research
Triangle Park, North Carolina, December 4, 1984.

39
 APPENDIX

PHOSGENE EMISSIONS DATA

About 89 percent of the phosgene production capacity in the United

States is located in West Virginia, Louisiana, and Texas.1 In an effort to
obtain phosgene emissions data, several plants and their respective air
control agencies in these States were contacted. Plants in two other states
were also contacted. The Texas Air Control Board (TACB) files were also
reviewed for information on the Texas plants. Only one direct measurement of
phosgene emissions was found, but the companies had made conservative
calculations of phosgene emissions. In one case, emission calculations were
based on the sensitivity of in-place monitors; no phosgene was actually
detected.
Engineering estimates submitted to state air control agencies by eight
phosgene producers range from 0 to 7.0 tons of phosgene per year. These
estimates contained no breakout of emissions due to phosgene production,
storage, use, etc., and the manufacturers did not indicate any basis for the
estimates.
Compounds other than phosgene that must be controlled include process
solvents, reactants, intermediate products, chlorine, carbon monoxide, and
hydrogen chloride (the latter is included unless phosgenation is carried out
in an alkaline medium). Most of these compounds are subject to Occupational
Safety and Health Administration (OSHA) regulations. A typical plant's
emission control system will include the following-

1. A water scrubber to remove and recover hydrogen chloride for sale or

reuse.

2. A caustic scrubber to control acidic gases, hydrogen chloride, and

phosgene. A backup scrubber, installed as a spare, is used if the
primary scrubber malfunctions.

3. An incinerator to control volatile organic compounds (some may

contain chlorine) and carbon monoxide. If both caustic scrubbers
malfunction, the phosgene will be routed directly to the incinerator
for destruction.

A-1
 4. An additional caustic scrubber that treats the incinerator exhaust
to remove residues from the combustion of chlorinated hydrocarbons.
In some plants where only one caustic scrubber is used, it is
located downstream of the incinerator.

5. For fugitive emissions, plugs or caps on open-ended lines, closed

purge sample systems, dual seals with barrier fluids on pumps, and
vent systems and rupture disks on safety relief valves. Phosgene
plants are also typically enclosed in negative-pressure buildings,
which may be vented to caustic scrubbers.

The following three types of emissions are found within a phosgene plant:
1. Process vent emissions--These are emissions from reactors, other
processing equipment (including storage tanks), and emission control
equipment, including incinerator and scrubber exhausts.

2. Process upsets emissions--These emissions represent inadvertent

releases due to equipment failures and human error.

3. Fugitive emissions--These emissions represent releases-due to leaks

in pumps, valves, and other phosgene handling equipment.

Phosgene Emission Estimates

Estimates of total phosgene emissions from plants producing 200 million

pounds of phosgene per year (a capacity chosen to represent production from a
large plant) are as follows:

1. From a plant producing phosgene for sale, 470 to 6950 lb/yr.

2. From a plant producing phosgene and converting it to toluene

diisocyanates, 705 to 9760 lb/yr.

3. From a plant producing phosgene and converting it to polycarbonates,

580 to 8190 lb/yr.

4. From a plant producing phosgene and converting it to herbicides, 580

to 8220 lb/yr.

The emission factors on which these estimates are based (in terms of pounds
of phosgene emitted per million pounds of phosgene processed) are presented
in Table A-1.
Phosgene emissions from processes that consume phosgene (TDI,
polycarbonate, herbicide, and pesticide production) have been estimated by
comparing the processes used to produce these chemicals with the process for
phosgene

A-2
 TABLE A-1. SUMMARY OF ESTIMATED PHOSGENE EMISSIONS FROM
HYPOTHETICAL PHOSGENE AND PHOSGENE DERIVATIVE PRODUCTION FACILITIESa
Emissions,
Process Fugitive Total lb/million
vents'b Upsets, emissions, emissions, lb phosgene
lb/yr lb/yr lb/yr lb/yr produced
• Phosgene production 300 50 120 to 6600 470 to 6950 2.35 to 34.8
at a merchant phos-
gene plant

• Toluene diisocya- 150 25 60 to 2640 235 to 2820 1.18 to 14.1

nate production

• Polycarbonate pro- 70 10 30 to 1160 110 to 1240 0.55 to 6.2

duction

• Herbicide and pes- 70 10 30 to 1190 110 to 1270 0.55 to 6.4

ticide production

• Total for TDI 450 75 180 to 9240 705 to 9760 3.53 to 48.8
plant (1 + 2)d

• Total for polycar- 370 60 150 to 7760 580 to 8190 2.90 to 41.0
bonate plant(1+3)d

• Total for herbi- 370 60 150 to 7790 580 to 8220 2.90 to 41.1
cide plant (1+4)d

a
Based on facilities with a production rate of 200 million pounds of phosgene
per year.
b
Incinerator and scrubber exhaust.
c
Fugitive emissions would be reduced by a factor of 100 if process reactor and
condenser are enclosed in negative-pressure buildings and vented through the
caustic scrubber.
d
Emissions from intermediate production are added to those from the phosgene
production operations to estimate total plant emissions.

A-3
production and then using engineering judgment to estimate the emissions
relative to those for phosgene production. The processes were compared
with respect to operations where phosgene might still be present and
potential emission sources. For example, emissions due to TDI
production were estimated to be one-half of those due to phosgene
production, and those for polycarbonate production and herbicide and
pesticide production were estimated to be about .25 percent of those due
to phosgene production. The TDI estimates are higher because two
phosgenation stages are required (Section 4.2) as compared with one
phosgene reaction step for polycarbonate (Section 4.3) and herbicide and
pesticide production (Section 4.4).
Phosgene emissions from phosgene consuming processes would be
expected to be significantly lower than those from the phosgqne
production process. Phosgene is only used early in the process and is
almost completely consumed. No provisions for phosgene recovery and
storage after production are needed. Therefore, it is reasonable to
expect phosgene production to be the major source of phosgene emissions.
Basis For Emission Estimates
The emission estimates in Table A-1 are approximations based on
limited information, derived as shown below:
Process Emissions--
Process emission estimates were based on the following:
1. An assumed phosgene concentration of 0.5 ppm in the caustic
scrubber effluent, based on a U.S. Amy study.2 This study
indicates that actual phosgene concentrations in scrubber
exhausts could be either significantly higher or lower,
depending on whether the scrubber is properly designed and
operated. The range was 0.015 to 10.3ppm. A scrubber would
need to have the capacity to handle phosgene surges.

2. Stack flow rates calculated from stack and exhaust velocity

data submitted to the Texas Air Control Board by the four Texas
phosgene producers.

3. Assumed plant operations of 24 hours a day, 330 days a year (90

percent availability).

A-4
 Based on this information, phosgene emissions of 19 to 600 pounds per
year were calculated for each of these plants. Thus, phosgene emissions
from a plant producing 200 million pounds of phosgene per year are
estimated to be 300 pounds per year.
Fugitive Emissions (Based on Ambient Exposures)--
Individual and area monitors of phosgene exposures yield the best
available information about actual phosgene levels in plants. As
standard equipment at plants producing and handling phosgene, these
monitors allow Upper limit estimations of ambient phosgene
concentrations and emissions.
Phosgene exposures are measured in ppm-minutes (the product of the
phosgene concentration in parts per million and the time of exposure in
minutes). The maximum phosgene concentration that OSHA allows is 0.1
ppm for an 8-hour day, or 48 ppm-minutes. Area monitors can respond to
phosgene levels as low as 0.05 ppm. Film badges are worn by all
employees in Ohosgene areas. An exposure of 5 ppm-minutes can be
detected by a visible color change, and as low as 2 ppm-minutes can be
detected photoelectrically. Based on reported data, photoelectrically
detected color changes were infrequent--only five occurrences over an
18-ffionth period in a single plant.3 Other plants contacted indicated
that positive monitor responses were also infrequent. Because the
@badges would detect exposures of 2 ppm-minutes, a steady-state phosgene
concentration of 1.2 ppm-minutes for an 8-hour shift was assumed.
Other assumptions included an indoor work area measuring 200 feet
by 100 feet by 30 feet, 40 air changes per hour,4 and operated 24 hours
a day, 330 days a year.

1.2 ppm - min x 10-6 / ppm

Process fugitive emissions =
8 h x 60 min / h

x 330 days/yr x 24 h/day x 40 changes/h x 6xl05 ft3 /change x 98.92 lb/lb - mole
380 ft3 / lb − mole

= 120 lb/yr
This calculation yields a controlled process fugitive phosgene emission
of 120 pounds per year. This is a conservative estimate because the

A-5
steadystate phosgene exposure is assumed to be constantly high
throughout the building for every working day during the year.
Fugitive Emissions (Based on Equipment Counts and Emission Factors)--
An alternative method can be used to estimate fugitive emissions;
one which applies emission factors to each valve, pump, etc., based on
the phosgene content of each process stream. The following steps were
followed in this approach:
1. Develop a process flow diagram (see Figures 2, 5, 7, and.9 in
text).

2. Identify all process streams containing phosgene.

3. Determine the phosgene content of each stream and whether

phosgene is present as liquid or vapor.

4. Identify and estimate the total number of fugitive emission

points (valves, pumps, and relief devices).

5. Estimate phosgene emission rates based on the probable degree

of control, assuming very stringent inspection and maintenance
programs and typical phosgene plant control measures for
valves, open-ended pipes, pump seals, and vents.5

Fugitive emissions of phosgene and other volatile organics result

from leaks in process valves, pumps, compressors, and pressure-relief
devices. For the four processes discussed (phosgene production,
isocyanate production, polycarbonate production, and herbicides and
pesticides production) the phosgene emission rates are based on process
flow diagrams, process operation data, fugitive source inventories for
hypothetical plants, and emission factors for process fugitive sources.
The first step in estimating fugitive emissions of phosgene
entailed listing the process streams in the hypothetical plants and then
estimating their compositions. For a reactor product stream, the
estimated composition was based on reaction completion data for the
reactor and on the plant product mix. For a stream from a distillation
column or other separator, the estimated composition was based on the
composition of the input stream to the unit, the unit description, and
the general description of the stream of interest (i.e., overheads,
bottoms, or sidedraw).
After the process streams were characterized, the number of valves
per stream was estimated (based on the type of process). Pumps were
A-6
assigned to each liquid process stream, and relief devices were assumed
on all reactors, columns, and other separators. No compressors are used
on phosgene process flows.
Emissions were then calculated for pumps, valves in liquid and gas
line service, and relief devices. Welded pipe joints are used in lieu
of flanges, and emissions from pipe joints are negligible.4 Fugitive
emissions from a particular source were assumed to have the same
composition as the process fluid to which the source is exposed. For
example, phosgene emissions from valves in liquid service were
determined by taking the product of 1) the total number of liquid valves
in phoigene service, 2) the average phosgene content of the streams
passing through these valves, and 3) the average fugitive emission rate
per valve per unit time. Emissions from valves in gas service and pumps
were calculated in the same manner. For relief devices, the composition
of fugitive emissions was assumed to be the same as that of the overhead
stream from the reactor or column served by the relief device.
Emissions from the various fugitive types of sources were summed to
obtain total process fugitive emissions of phosgene.
Emissions from process fugitive sources depend on the number of
sources rather than their size; therefore, plant capacity does not
affect total process fugitive emissions. For this reason, overall
emissions are expressed in terms of kilograms per hour of operation.
The estimates of fugitive phosgene emissions are presented in
Tables A-2 through A-5. The emission factors used in these estimates
are summarized in Table A-6. At a hypothetical phosgene plant producing
200 million pounds of phosgene per year, the process fugitive emission
rate is 0.38 kg/h or 3000 kg/yr (3.3 tons/yr), assuming the plant
operates 24 hours per day, 330 days per year. For toluene diisocyanate
production, estimated process fugitive emissions are 0.15 kg/h or 1,200
kg/yr (1.3 tons/yr); for polycarbonate production, 0.067 kg/h or 530
kg/yr (0.58 tons/ yr); and for herbicide and pesticide production, 0.068
kg/h or 540 kg/yr (0.59 tons/yr), in addition to the emission rate
(determined above) associated with captive phosgene production, if
carried out at these facilities.
Reference 5, the basis for the emission factors and control
efficiencies in this analysis, does not consider fugitive emission
A-7
 TABLE A-2. ESTIMATED FUGITIVE PHOSGENE
EMISSIONS FROM A HYPOTHETICAL PHOSGENE PLANT
PRODUCING 200 MILLION POUNDS OF PHOSGENE PER YEARa
Uncon-
trolled
emissionb Controlb,c
factor, efficiency, Avg. COC12 Emissions,
Emission source Number kg/h % content, % kg/h
Valves
Liquid 150 0.0071/ 59 68 0.30
valve

Gas 100 0.0056/ 73 50 0.08

valve

Pumps 2 0.0494/ 100 68 0

pump

Relief valves on:

COC12 reactor 2 0.104 100 65 0
Condenser 2 0.104 100 100 0
Adsorption column 2 0.104 100 35 0
Storage tanks 4 0.104 100 100 0
All sources 0.38

a
Process streams and their composition at hypothetical plant:

Process streams Phase % phosgene

Reactor to condenser Gas 65
Condenser to storage Liquid 100
Condenser to absorber Gas 35
Liquid phosgene to plant or shipment Liquid 100
Phosgene solution to plant Liquid 5

b
Reference 5.
c
The control efficiencies are based on the use of plugs or caps on open-ended
lines, double seals with barrier fluids on pumps, vent systems and rupture disks on
safety relief valves, and a monitoring interval of at least monthly for valves.

A-8
 TABLE A-3. ESTIMATED FUGITIVE PHOSGENE EMISSIONS
FROM A HYPOTHETICAL TOLUENE DIISOCYANATE PRODUCTION FACILITYa
Uncon-
trolled
emissionb Controlb,c
factor, efficiency, Avg. COC12 Emissions,
Emission source Number kg/h % content, % kg/h

Valves
Liquid 100 0.0071/ 59 5 0.01
valve
Gas 200 0.0056/ 73 45 0.14
valve

Pumps 2 0.0494/ 100 5 0

pump

Relief valves on:

Phosgene line to
first-stage
phosgenator 2 0.104 100 85 0
Phosgene line to
second-stage
phosgenator 2 0.104 100 85 0
Unreacted phos-
gene to recycle
line 2 0.104 100 5 0
First-stage-
phosgenation 2 0.104 100 5 0
Second-stage
phosgenation 2 0.104 100 1 0
Distillation
column 2 0.104 100 1 0
All sources 0.15
a
Process streams and their composition at the hypothetical plant:

Process streams Phase % phosgene

Phosgene to first-stage phosgenation Gas 85
Phosgene to second-stage phosgenation Gas 85
Unreacted phosgene to recycle Gas 5
Monoisocyanote to second-stage
phosgenation Liquid 5
Diisocyanate to distillation Liquid 5
Waste phosgene to scrubber Gas 5
b
Reference 5.
c
The control efficiencies are based an the use of plugs or caps on open-ended
lines, double seals with barrier fluids on pumps, vent systems and rupture disks on
safety relief valves, and a monitoring interval of at least monthly for valves.

A-9
 TABLE A-4. ESTIMATED FUGITIVE PHOSGENE EMISSIONS
FROM A HYPOTHETICAL POLYCARBONATE PRODUCTION FACILITYa

Uncon-
trolled
emissionb Controlb,c
factor, efficiency, Avg. COC12 Emissions,
Emission source Number kg/h % content, % kg/h
Valves
Gas 100 0.0056/ 73 44 0.067
valve
Pumps 0 0 0 0 0

Relief valves on-

Phosgene line to
reactor 2 0.104 100 85 0
Unreacted phos-
gene to inciner-
ator 2 0.104 100 85 0
All sources 0.067

a
Process streams and their composition at the hypothetical plant:

Process streams Phase % phosgene

Phosgene to reactor Gas 85
Unreacted phosgene to incinerator Gas 2
b
Reference 5.
C
The control efficiencies are based on the use of plugs or caps on open-ended
lines, double seals with barrier fluids on pumps, vent systems and rupture disks on
safety relief valves, and a monitoring interval of at least monthly for valves.

A-10
 TABLE A-5. ESTIMATED FUGITIVE PHOSGENE EMISSIONS FROM A
HYPOTHETICAL HERBICIDE AND PESTICIDE PRODUCTION FACILITYA
Uncon-
trolled
emissionb Controlb,c
factor, efficiency, Avg. COC12 Emissions,
Emission source Number kg/h % content, % kg/h
Valves
Gas 100 0.0056/ 73 44 0.067
valve
Liquid 50 0.0071/ 59 1 0.001
valve

Pumps 1 0.0494/ 100 1 0

pump

Relief valves on:

Phosgene line to
reactor 2 0.104 100 85 0
Unreacted phos-
gene to scrubber, 2 0.104 100 2 0
All sources 0.068

a
Process streams and their composition at the hypothetical plant:

Process streams Phase % phosgene

Phosgene to reactor Gas 85
Unreacted phosgene to incinerator Gas 2
Aromatic isocyanate to second-
stage reactor Liquid 1
b
Reference 5.
c
The control efficiencies are based on the use of plugs or caps on open-ended
lines, double seals with barrier fluids on pumps, vent systems and rupture disks on
safety relief valves, and a monitoring interval of at least monthly for valves.

A-11
 TABLE A-6. PROCESS FUGITIVE
EMISSION FACTORS FOR PLANTS
Emission factor,
Facility Emission source kg/h
Marketable phosgene Phosgene production 0.38
producer

Toluene diisocyanate Phosgene production 0.38

producer TDI production 0.15
0.53

Polycarbonate Phosgene production 0.38

producer Polycarbonate production 0.067
0.45

Herbicide and pesti- Phosgene production 0.38

cide production Herbicide and pesticide 0.0068
production
0.39
For example, the most stringent level level of control for valves cited
in this report only involves monthly inspection and maintenance. This
program provides 73 percent leak control for valves handling gases, and
59 percent leak control for valves handling light liquids. A control
efficiency of 100 percent, however, was estimated for plugs and caps on
open-ended lines, dual seals with barrier fluids on pumps, and vent
systems and rupture disks on safety relief valves. These efficiencies
were used to estimate fugitive phosgene emissions from valves, pumps,
and relief valves.
The emission factors from Reference 5 are based on "leaking" and
"nonleaking" sources. Leaking is defined as "screening at or above
10,000 ppm with a portable VOC monitor.” Nonleaking is defined as
"screening below 10,000 ppm.” These data do not allow extrapolation to
the actual level of control most likely in phosgene plants. A 10,000
ppm phosgene concentration could not be tolerated because the average
lethal exposure is 400 to 500 ppm-minutes (concentration in ppm
multiplied by exposure in minutes). The current OSHA standard is 0.1
ppm. Phosgene plant monitoring equipment typically responds to
concentrations of 0.02 to 0.2 ppm. 6-8 Film badges typically respond to
exposures of 2 ppm-min.
The estimated fugitive phosgene emission rate of 120 lb/yr based
on the monitoring approach from a facility with a capacity of 200

A-12
million pounds corresponds to 99.3 percent control of valve emissions.
Considering that phosgene concentrations of 0.02 to 0.2 ppm6-8 produce
immediate responses from monitoring equipment, this level of control
seems achievable. Both estimates of fugitive phosgene emissions,
however, are presented in the report to provide a range of fugitive
phosgene emissions.
In plants where the reactor, condenser, and associated valves,
pumps, etc. are enclosed in negative-pressure buildings and the exhausts
are vented through a caustic scrubber, phosgene fugitive emissions will
be reduced by a factor of 100 from the above estimates.
Process Upsets--
As discussed earlier, a search of the Texas Air Control Board
files yielded process upset reports for three producers of phosgene.
Based on the 15 process upset reports during a recent 6-year period, the
three plants had a total phosgene release of 900 pounds. Phosgene
releases in the 15 episodes ranged from 1 to 220 pounds. The stated
amount rel.eased in each case usually represented an estimate. All
releases may not have been reported. One letter in the file responding
to a Notice of Violation stated that the company was not obligated to
report the release because none of the released material (not phosgene)
had left the company property. One of the 15 releases led to a
fatality, and two other releases were responsible for lost-time
accidents. Based on the size of the plants, the total phosgene release,
and the number of releases of phosgene, it is estimated that a plant
producing 200 million pounds of phosgene per year will have one process
upset per year during which 50 pounds of phosgene is released.

A-13
 REFERENCES FOR APPENDIX

1. SRI International. 1963 Directory of Chemical Producers, USA.

1983.

2. Kistner, S., et. al. A Caustic Scrubber System for the Control
of Phosgene Emissions Design, Testing, and Performance.
Journal Air Pollution Control Association, 28 (7): 673-676,
1978.

3. Personal Communication with R. L. Matherne, Ethyl Corporation,

Baton Rouge, Louisiana, February 7, 1984, concerning Mr.
Matherne's experiences at BASF-Wyandotte.

4. Pollution Engineering Practice Handbook. P. N. Cheremisinof

and R. A. Young, eds. Ann Arbor Science, Ann Arbor, Mighigan.
P. 212.

5. U.S. Environmental Protection Agency. Fugitive Emission

Sources of Organic CompoUnds--Additional Information on
Emissions, Emission Reductions, and Costs. EPA-450/3-82-010,
1982.

6. Personal communication with Marshall Anderson, General Electric

Corporation, Mount Vernon, Indiana, February 8, 1984.

7. Personal communication with Jerry Neal, PPG Industries,

LaPorte, Texas, February 14, 1984.

8. Personal communication with George Flores, Dow Chemical

Company, February 14, 1984.

A-14