SRI
Human Exposure
To Atmospheric Concentrations
Of Selected Chemicals
Office of Air Quality Planning and Standards
Environmental Protection Agency
-------
DC
450R790O4
ASSESSMENT OF HUMAN
EXPOSURES TO ATMOSPHERIC
ACRYLONITRILE
Final Report
August 1979
By: Benjamin E. Suta
Center for Resource and
Environmental Systems Studies
Prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
Task Officer Joseph D. Cirvello
Project Officer: Joseph D. Cirvello
Contract No. 68-02-2835, Task 20
SRI Project No. CRU-6780
CRESS Report No. 100
333 Ravenswood Ave. Menlo Park, California 94025
(415) 326-6200 Cable. SRI iNTL IvIPK TWX: 910-373-1246
-------
HOTICE
This report has been provided Co the U.S. Environmental
Protection Agency (EPA) by SRI International, Menlo
Park, California, in partial fulfillment of Contract
68-02-2835. The opinions, findings, and conclusions
expressed herein are those of the author and are not
necessarily those of EPA. Mention of company or pro-
duet names is not to be considered an endorsement by
EPA.
-------
CONTENTS
LIST OF TABLES iv
ACKNOWLEDGMENTS v
I INTRODUCTION 1
II SUMMARY 2
III CHEMICAL AND PHYSICAL PROPERTIES OF
ACRYLONITRILE AND ITS ENVIRONMENTAL BEHAVIOR 8
Introduction -. 8
Properties 8
Environmental Behavior .... 8
IV ACRYLONITRILE PRODUCTION AND USES 11
Production 11
Acrylonitrile Producers and Users 13
V POPULATION EXPOSURES FBOM ACRYLONITRILE PRODUCTION . . 16
Sources of Emissions 16
Atmospheric Emissions 16
Atmospheric Concentrations 16
Exposure Estimates 19
VI POPULATION EXPOSURES FROM PRODUCERS THAT USE
ACRYLONITRILE AS A FEEDSTOCK 23
Sources of Emissions 23
Atmospheric Emissions 23
Atmospheric Concentrations 26
Exposure Estimates 28
VII COMPARISONS OF MONITORING AND DISPERSION
MODELING CONCENTRATIONS 30
General 30
Atmospheric Monitoring 30
Comparisons 36
BIBLIOGRAPHY 37
ILL
-------
LIST OF TABLES
II-l Summary of Estimated Population Exposures to
Atmospheric Acrylonitrile From Specific Emission
Sources 4
II-2 Estimated Acrylonitrile Atmospheric Emissions 7
III-l Physical Propoerties of Acrylonitrile 9
IV-1 Acrylonitrile Production and Consumption 13
IV-2 AN Producers and Major Consumers 14
IV-3 1977 Use of Acrylonitrile Production Facilities .... 15
V-l Emission Factors for Acrylonitrile Production 17
V-2 Estimated Atmospheric Emissions of AN From
Monomer Production Facilities 18
V-3 Dispersion Modeling Estimates of Atmospheric
Concentrations of Acrylonitrile from Monomer
Production 20
V-4 Estimated Human Population Exposures to Atmospheric
Acrylonitrile Emitted by Producers 22
VI-1 Estimated Acrylonitrile Emission Rates for ABS/SAN
Resin Production 24
VI-2 Estimated Acrylonitrile Emission Rates for Acrylic
and Modacrylic Fiber Production 24
VI-3 Estimated Acrylonitrile Emission Rate for
Adiponitrile Production 25
VI-4 Estimated Acrylonitrile Emission Rate for
Nitrile Elastomer Production 25
VI-5 Dispersion Model Estimates of Omnidirectional Annual
Average Acrylonitrile Concentrations3 ( yg/m^)
for Various Products 27
VI-6 Estimates of Population Exposures to Atmospheric
Acrylonitrile Emitted by Plants that use it as a
Chemical Intermediate 29
VII-1 Atmospheric Monitoring Data for Acrylonitrile 32
VII-2 Comparison of Monitoring and Dispersion
Modeling Data 34
iv
-------
ACKNOWLEDGMENTS
It is a pleasure to acknowledge the cooperation and guidance given
by several individuals of the U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards. Joseph Cirvello,
Strategies and Air Standards Division, provided direction throughout the
study. David Mascone of the Emission Standards and Engineering Division
gave valuable assistance in regard to atmospheric emissions. George
Schewe and Philip Youngblood of NOAA provided input on atmospheric
dispersion modeling.
-------
I INTRODUCTION
This report is one in a series that SRI International is providing
for the U.S. Environmental Protection Agency (EPA) to estimate the
population-at-risk to selected pollutants. Primarily, this study has
sought to estimate environmental exposure of the U.S. population to
atmospheric acrylonitrile (AN) emissions. The principal atmospheric
sources we consider in this report are facilities that produce or that
use it as a chemical intermediate.
, 1
-------
II SUMMARY
The annual domestic AN production capacity is approximately 973,000
metric tons. During 1977, the industry operated at 76.62 of capacity,
producing 745,000 mt. Future growth of the market is expected to
average 6.0-7.52/yr through 1982, when the demand is estimated to be
850,000-920,000 tat/yr. AN is used primarily as a raw material in the
synthesis of other chemicals, in particular for acrylic and modacrylic
fiber (452),* acrylonitrile-butadiene-styrene (ABS) and
styrene-acrylonitrile (SAN) resins (192), nitrile elastomers (3%),
adiponitrile (102), acrylamide (32), nitrile barrier resins (12), and
minor miscellaneous uses (42). Exports account for approximately 152 of
domestic production. Miscellaneous uses include mixing AN with carbon
tetrachloride to produce a fumigant for stored tobacco and for equipment
used in flour milling and bakery food processing. Although AN is
registered as a pesticide in the United States and the EPA has
classified it for restricted use by certified applicators, manufacturers
have voluntarily withdrawn most pesticides containing AN (IARC, 1979).
AN is a clear, colorless, highly flammable liquid whose
characteristic odor is unpleasant and irritating. It boils at 77°C,
melts at -83.5°C, and has a specific gravity of 0.8. AN does not
react with water and is degraded slowly by both aerobic and anaerobic
bacteria. AN is photochemically reactive and has an estimated
atmospheric half-life of 9-10 hr. Because AN is a volatile compound,
considerable atmospheric discharges from land and water evaporation is
expected. This volatility, coupled with a sufficiently long atmospheric
half-life, results in aerial transport being a significant mechanism for
environmental distribution of AN.
*Numbers in parentheses are percents of total domestic production.
-------
Only a few environmental AN atmospheric monitoring data are
available from the vicinity of chemical plants. Going (1978) reported
on the sampling of atmospheric AN in the vicinity of 11 industrial
sites. They placed 4 to 8 monitoring stations at varying directions and
distances at each of the 11 sites. Duplicate samples were taken at some
stations for quality assurance. They collected samples for only one
24-hr period for 10 of the industrial sites. Two 24-hr samples were
collected at the eleventh site. AN was found in the atmosphere at all
monitoring stations. The average 24-hr levels for individual station
locations ranged from 0.1 to 249 pg/m , with the highest individual
24-hr integrated concentration at 325 ug/m .
Before January 1978, the U.S. Occupational Safety and Health
Administration's (OSHA) health standards for exposure to air
contaminants required that an employee's exposure to AN not exceed an
8-hr time-weighted average of 20 ppm (45 mg/m ) in the workplace air
in any 8-hr work shift of a 40-hr work week. In January 1978, OSHA
announced an emergency standard for AN that limits exposure to an 8-hr
time-weighted average of 2 ppm (4.5 mg/m ) AN in air. A ceiling
level of 10 ppm was also set for any 15-min period during the 8-hr shift
(LARC, 1979). The occupational standard was finalized at these limits
in October 1978.
Nonoccupational human population exposures to atmospheric AN have
been estimated for emissions resulting from its production and its use
as a chemical intermediate in the production of ABS/SAN resins, acrylic
and modacrylic fibers, nitrile elastomer, and adiponitrile. Because of
the lack of emission data, exposures were not estimated for acrylamide
production. According to Mascone (1979b) the AN emissions from the
production of acrylamide are expected to be negligible.
The population exposure estimates are given in Table II-l. The
total exposures for each chemical production are summarized both by the
total number of people exposed to annual averge concentrations exceeding
0.001 ug/m and by total risk. We define total risk as the sum of the
-------
Table II-l
SUMMARY OF ESTIMATED POPULATION EXPOSURES TO ATMOSPHERIC
ACRYLONITRILE FROM SPECIFIC EMISSION SOURCES
Annual Average
AN Concentration0
(ug/m3)
AN
Monomer
ABS/SAN
Resins
Acrylic and
Modacrylic
Fibers
Nitrile
Elastomer
Adiponitrile
15.0-19.9
10.0-14.9
5.0-9.9
1.0-4.9
0.50-0.99
0.10-0.49
0.050-0.099
0.010-0.049
0.005-0.009
0.001-0.004
Total
people exposed
Total risk
(person-pg/m3
2,700
1
64
140
,800
600
50
240
,000
,000
,000
,000°
n
?
Ob
850
73
79
680
1,200
1,400
510
790
,000
,000
,000
,000
,000
,000
,000
b
u
V
\J
4
52
70
370
190
260
,700
,000
,000
,000
,ooov
r»
,000°
or
0°
i,
22,
81,
650,
690,
2,700,,
5,
- 93,
,800
,000
,000
,000
,000,
' K
,000°
,100°
,000°
2,600,000 4,700,000 950,000 4,200,000
610,000
500,000 270,000
320,000
22,000
32,000
65,000.
120,000
9,400
aTo convert from ug/nP to ppb, multiply concentrations by 0.46.
"Exposures in these ranges are underestimated because calculations were only
made for exposures within 30 km of each plant.
-------
number of people exposed to a given concentration times the
concentration. Total risk relates directly to cancer mortalities if a
linear, no-threshold dose-response function is assumed. We found no
3
human exposures in excess of an annual average of 20 yg/m .
In estimating exposures, we first identified all plants that
produce the selected chemicals. EPA (Mascone, 1979a) provided estimates
of annual atmospheric AN emissions for each of the plants identified.
Dispersion modeling was used to estimate AN concentrations in the
vicinity of prototype plants. The emission estimates and source
characteristics served as input to atmospheric, dispersion models, from
which estimates of annual average atmospheric. AN concentrations in the
vicinity of each plant were obtained. Residential population was
estimated for 10 geographical rings about each plant. The ring radii
were taken as 0.3-0.5, 0.5-1, 1-2, 2-3, 3-4, 4-6, 6-10, 10-15, 15-20,
and 20-30 km. The SRI BESTPOP computer data system, which contains the
entire U.S. population on a network of 1-km grids, was used to estimate
ring populations. The exposure concentrations were taken as the
dispersion modeling estimated annual average concentration at the
midpoint of each geographical ring about each plant. Exposure estimates
were not made for annual average concentrations of less than 0.001
Ug/m . In addition, exposures were not estimated for people residing
beyond 30 km from the plants because the dispersion modeling results are
considered to be imprecise at the greater distances. Many locations
would have had people exposed, beyond 30 km from their plants, to annual
average AN concentrations ranging from 0.001 to 0.050 yg/m .
Consequently, the reported exposures in this range are underestimated.
Most AN exposures result from ABS/SAN resin and nitrile elastomer
production, with more than 4 million people exposed by each. AN monomer
production and ABS/SAN resin production result in the highest estimated
total risks considering exposed population and exposure
concentrationsmore than 500,000 person- yg/m for each.
-------
Table II-2 gives the estimated atmospheric AN emissons from AN
monomer production and from its use as a feedstock in other chemicals.
We estimate that emissions total about 10,822 mt/yr.
To arrive at these emission and exposure estimates has necessitated
reliance on very limited data. Because of the paucity of measured
atmospheric AN data, we had to approximate concentrations through the
use of dispersion modeling. Moreover, the resulting estimates are
subject to considerable uncertainty in regard to: (1) the quantity of
AN emissions, (2) AN production and consumption levels, (3)
meteorological conditions assumed, (4) control technologies employed,
(5) deterioration in control technologies over time, (6) physical
characteristics of AN sources (e.g., stack height), (7) details about
atmospheric dispersion and degradation, and (8) living patterns of the
exposed population. Given these complex variables, the accuracy of the
estimates cannot completely be assessed. Comparisons of limited 24-hr
monitoring data for 10 locations with dispersion modeling estimates of
annual average omnidirectional concentrations shows that, on the
average, the monitoring concentrations were about 30% less than the
dispersion modeling concentrations. This difference corresponds well
with the spiked sampling results, which indicate the average AN recovery
for AN monitoring is about 632.
-------
Table II-2
ESTIMATED ACRYLONITRIUE ATMOSPHERIC EMISSIONS
Emissions (mt/yr)
AN Production
Absorber vent 537
Flare stack 487
Product loading 136
Fugitive emissions 253
Storage tanks 788
Deep well pond 97
Subtotal 2,298
Production using AN as a feedstock
Acrylic and modacrylic
fibers 4,698
ABS and SAN resins 3,085
Nitrile elastomer 650
Adiponitrile 91
Acrylamide NE
Subtotal 8,524
Total 10,822
NE * Not estimated, but assumed negligible.
-------
Ill CHEMICAL AND PHYSICAL PROPERTIES OF ACRYLONITRILE
AND ITS ENVIRONMENTAL BEHAVIOR
Introduction
The-Chemical Abstracts service registry number for AN is 107-13-1,
AN has many synonyms and trade names, including: AN, 2-Propenitrile,
cyanoethylene, propenenitrile, VCN, vinyl cyanide, Acrylon, Carbacryl,
Pumigrain, and Ventox.
The composition and structure of AN is indicated by the molecular
formula, C H N, and the line diagram,
I
H
Properties
An is a clear, colorless liquid whose characteristic odor is
unpleasant and irritating. It is moderately soluble in watef, soluble
in acetone and benzene, and is miscible with ethanol and ether. Other
properties are given in Table III-l.
Environmental Behavior
Olefins as a class generally enhance atmospheric oxidation
reactions. AN, an olefin, might be expected to participate in these
reactions (Miller and Villaume, 1978). Joshi (1977) observed the
-------
Table III-l
PHYSICAL PROPERTIES OF ACRYLONITRILE
Molecular weight
Melting point, °C
Boiling point, °C
Vapor density
Specific gravity
Vapor pressure
Solubility
Exposive limits
Ignition temperature
Flash point
Conversion factor
53.06
-83.5
77-79
1.8 (air - 1.0)
0.8004 (25°/4°C)
83 mmHg at 20°C
110-115 mmHg at 25°C
7.3% by weight in water at
3.0-17.0Z by volume in air
at 25°C
481°C
-1°C (closed cup)
1 ppm vapor * 2.17 mg/m^
Sources: Patterson et al. (1976) and IARC (1979).
photochemical reactivity of AN through the use of a smog chamber. He
found acrylonitrile to be reactive with an estimated atmospheric
half-life of 9-10 hr. He hypothesized that the products of irradiation
of acrylonitrile/NO could be: hydrogen cyanide, carbon monoxide,
ozone, formaldehyde, and formic acid. Trace amounts of formyl cyanide
and nitric acid are also expected.
An atmospheric half-life of 9-10 hr is sufficiently long for aerial
transport to play a significant role in the neighborhood distribution of
AN. For example, with an average wind speed of 4 m/s, 86Z of the
emitted AN will survive at 30 km downwind from the source, and 78% will
survive at 50 km downwind.
Because AN is a volatile compound and is only moderately soluble in
water (72), it does not react with water. Consequently, it is labeled 0
(no hazard) in the NAS Hazard Rating System for reactivity with water.
Hydrogen cyanide is not an expected breakdown product because AN does
not dissociate appreciably in water (Miller and Villaume, 1978).
-------
Because the vapor pressure of AN is high, most atmospheric
emissions from its manufacture occur as vapor. Because of its only
moderate solubility in water, considerable evaporation from land and
water emissions into the atmosphere are expected to occur. This
volatility, coupled with a sufficiently long atmospheric half-life,
results in aerial transport serving as a significant mechanism for
environmental distribution of AN.
10
-------
IV ACRYLONITRILE PRODUCTION AND USES
Production
The annual AN production capacity is approximately 973,000 mt.
During 1977, the industry operated at 76.6% of capacity, producing
745,000 mt. Future growth of the market is expected to average
6.0-7.5%/yr through 1982 when the demand is estimated to be 850,000 to
920,000 mt/.yr (SRI estimate).
AN is used primarily as a raw material in the synthesis of other
chemicals, in particular for acrylic and modacrylic fiber, ABS and SAN
resins, nitrile elastomers, adiponitrile, acrylamide, nitrile barrier
resins, and other miscellaneous uses. Primary uses of these compounds
are as follows:
Acrylic and modacrylic
fibers
ABS resin
More than 60% of these fibers is
used in apparel. Carpeting is the
second largest use. Home furnishing
uses include blankets, draperies,
and upholstery. Industrial uses
include sandbags, filter cloths,
tents, and tarpaulins. The fibers
are also used in synthetic hair wigs.
Its major markets are pipes and pipe
fittings and automotive components.
Other important markets are large
appliances, housing for business
machines and telephones,
recreational vehicle components,
toys, sporting goods, and sheeting
material for luggage.
11
-------
o SAN resin
Its primary uses are for drinking
tumblers and other houseware items,
for automob ile ins trument pane 1s,
and instrument lenses.
o Nitrile elastomers
Its major uses are in rubber hose,
seals, gaskets, latex, adhesives,
polyvinyl chloride blending, paper
coatings, and pigment binders.
o Adiponitrile
It is hydrogenated to
hexamethylenediamine, which is used
to produce nylon.
o Acrylamide
Its largest use is in the production
of polyacrylamides for waste and
water treatment flocculants. Other
acrylamide products are used to aid
sewage dewatering, and for paper-
making strengtheners and retention
aids.
Nitrile barrier resins
It is used in the manufacture of
beverage containers.
Miscellaneous uses
These uses include the producton of
fatty amines and their derivatives,
cyanoethylation of various alcohols
and amines, fumigation of tobacco,
and as an absorbent.
Table VI-1 shows the quantities of AN consumed for these uses.
More than 50% of domestic consumption is used in the production of
acrylic and modacrylic fibers, and an additional 20% is used in the
production of ABS and SAN resins. Exports account for 15% of the total
AN production.
12
-------
Table IV-1
ACRYLONITRILE PRODUCTION AND CONSUMPTION (1977)
AN
(103 mt)
Projected Annual
Growth (Z)a
Distribution
U.S. production 745
Imports Negligible
Exports 109
Total domestic
consumption 636
Products
Acrylic and modacrylic
fibers 331
ABS and SAN resins 142b
Nitrile elastomers 24
Adiponitrile 73
Acrylamide 24
Nitrile barrier
resins 9
Other 33
NE
NE
NE
6.0-7.5
4.5-5.5
7.5-9.5
2.0-3.0
10.5-12.5
8.0-10.0
12.0
4.0-6.0
NE
Not estimated.
Projected annual average growth until 1982.
b!26,000 mt for ABS resin and 16,000 mt for SAN resin.
Source: SRI estimates.
AN Producers and Users
Table IV-2 lists the major AN producers and consumers, along with
their estimated installed production capacity as of January 1978. Most
of these producers use AN captively at other chemical plants. In fact,
during 1976 AN sales acounted for only about 40Z of production. The
numbers of plants producing AN and its major products follow:
13
-------
Producer
TOTAL
Table IV-2
AN PRODUCERS AND MAJOR CONSUMERS
(January 1, 1978, Production Capacities in Thouand* of Metric Tona of AN)
Location
ABTEC Co.a
American Cyanamide
American Cyanamide
American Cyanamide
Borg Warner
Borg Warner
Co polymer Rubber
Dow
Dow
Dow
Dow
Dow
Dow Badische
du Pont
du Pont
du Pont
du Pont
Eastman Kodak
B. F. Goodrich
B. F. Goodrich
Goodyear
Goodyear
Carl Corden
Mobil Chemical0
Honsanto
Monsanto1'
Monsanto
Monaanto
Honsanto
Monsanto
Nalco
Uniroyal
Uni royal
Vistron (SOHIO)B
Louiaville, KV
Linden, NJ
Pensacola, FL 49
West Hego, LA 120
Washington, WV
Ottawa, IL
Baton Rouge, LA
Allyn's Point, CT
Irontoun, OH
Midland, HI
Riverside, MO
Torrance, CA
Hi I lisas burg, VA 28
Canden, SC 65 3
Beaumont, TX IS9
Uayneaboro, VA 57
Memphis, TN 122
Kingsport, TN 7
Akron, OH
Louisville, KY
Akron, OH
Houston, TX
Worcester, MA
Joliet, IL
Alvin, TX 200
Texas City, TX 191
Decatur, At 122 9
Addyaton, OH
Muscatine, 10
Springfield, HA
Garyavilte, LA
Scotts Bluff, LA
Plainesville, OH
Lima, OH 181
Nitrite
AN Acrylic Hodacrylic ABS SAN Elaatomer
Monomer Fiber fiber Key in Re a in and late* Adiponitrile Acrylamidt
973
321
19
39
28
(b)
10
9
21
7
18
5
10
1
4
15
4
24*
196 26
87
28
87
(h)
SO
aJointly owned by Cosden Oil and Chemical and B. F. Goodrich.
bBorg Marner plan* to build a 55,000-mt ABS plant requiring IS,000 at of AN at capacity.
'Formerly owned by Dart Industries, Rexene Styrene Co.
dCapacity 11 to be increased by 95,000 mt in 1981.
eplans to increase capacity by late 1978 to require an additional 44,000 mt of AM.
f Plane to increase capacity by early 1979 to require an additional 23,000 at of AN.
^Includes 45,000 at older capacity that can be run or placed on standby.
"In 1976 Viatron closed a 7,000-mt plant at Lima, OH.
Source: SRI estimates.
-------
Number of
Product Plants
AN 6
Acrylic fiber 5
Modacrylic fiber 3
ABS resin 13
SAN resin 3
Nitrile elastomer 6
Adiponitrile 1
Acrylamide 4
Because many of these plants produce several AN products, 34 plants in
all manufacture them.
Chemical producers rarely operate at maximum production capacity
for a specific chemical. Table IV-3 shows the percent of production
capacities employed in 1977 to produce AN, as well as the major products
in which it is used as a feedstock. These percentages range from 48 to
97%.
Table IV-3
1977 USE OF ACRYLONITRILE PRODUCTION FACILITIES
(Thousands of Metric Tons)
Product AN Capacity3 AN Production13 Z Capacity Used
Acrylic and mod-
acrylic fibers 340 331 97.4
ABS resins 196 126 64.3
SAN resins 26 16 61.5
Nitrile elastomer 28 24 85.7
Adiponitrile 87 73 83.9
Acrylamide 50 24 48.0
Acrylonitrile 973 745 76.6
aSee Table IV-2. This is the amount of AN that would be used annually
if production were at 100% of capacity.
bSee Table IV-1.
15
-------
V- POPULATION EXPOSURES FROM ACRYLONITRILE PRODUCTION
Sources of Emissions
Table IV-2 lists the AN producers. Their annual plant capacities
total 973,000 tat. Table IV-3 indicates that approximately 77% of the
production capacity was employed during 1977.
Atmospheric Emissions
Six principal sources of atmospheric AN emissions have been
identified for AN production facilities: absorber vents, flare stacks,
product loadings, fugitive emissions, storage tanks, and deep well
ponds. Incinerators are also a source of emissions, but we did not
include them in this analysis because their emission rate is quite low.
Table V-l gives estimated AN emission factors and source descriptions
for the six principal emission sources. AN emissions total approxi-
mately 1.85 g/kg of product. For each manufacturing plant, Table V-2
gives the estimated emissions resulting from the emission factors. An
annual emission of 2,298 mt is estimated for all AN monomer production.
Atmospheric Concentrations
Because atmospheric monitoring data are insufficient (see Section
VII), we have employed atmospheric dispersion modeling to estimate
atmospheric AN concentrations and their neighboring population
exposures. In keeping with the generalized nature of this study,
approximate dispersion estimates were made using approximate
Gaussian-plume techniques. To facilitate the calculations, the
atmospheric dispersion models PTDIS and PTMAX were primarily used for
16
-------
Table V-l
EMISSION FACTORS FOR ACRYLONITRILE PRODUCTION
Source
Absorber vent
Flare stack
Product loading
Fugitive emissions
Storage tanks
Deep well pond
AN Emission Factor
(g/kg of product)
0.04a
0.50
0.14
0.26
0.81
0.10
Source Description
Elevated point source
Elevated point source
Point source at a
height of 5 m, no plume
rise
Uniform emissions at a
height of 10 m over a
100 m x 100 m area, no
plume rise
Uniform emissions at a
height of 10 m over a
100 m x 100 m area, no
plume rise
Assumed to be
well-controlled
a
The absorber vent emission factor is much larger for American
Cyanamide, West Wego (3.0 g/kg), and duPont, Memphis (1.2 g/kg).
We have assumed these to be at an effective stack height of 63 m.
Source: Mascone (1979a and 1979b).
17
-------
Table V-2
ESTIMATED ATMOSPHERIC EMISSIONS OF AN FROM MONOMER PRODUCTION FACILITIES
(Metric Tons per Year)
Emission Source
oo
Plant /Location
American Cyanamide, West We go, LA
du Pont, Beaumont, TX
du Pont, Memphis, TN
Monsanto, Alvin, TX
Monsanto, Texas City, TX
Vistron, Lima, OH
TOTAL
Absorber
360.6
6.4
146.4
8.0
7.6
7.2
536.4
Flare
60.0
79.5
61.0
100.0
95.5
90.5
486.5
Loading0
16.8
22.3
17.1
28.0
26.7
25.3
136.2
Fugitive
31.2
41.3
31.7
52.0
49.7
47.1
253.0
Storage8
97.3
128.8
98.8
162.0
154.7
146.6
788.1
Pond
12.0
15.9
12.2
20.0
19.1
18.1
97.3
Total
578.0
294.2
367.2
370.0
353.3
334.8
2,297.5
aLoading and storage emissions are assumed uncontrolled. Some producers may control up to 90%
of these emissions.
Source: Based on data supplied by Mascone (1979a and 1979b).
-------
point sources (absorber vent, flare stack, and product loading) with
backup by the single source model CRSTER; the dispersion model PAL
(Point-Area-Line) was used for the area sources (fugitive emissions and
storage tanks). Table V-l lists the source parameters used. The
meteorological conditions input to the models follow: neutral stability
(Pasquill-Gifford "D") and a wind speed of 4 m/s. EPA (Youngblood,
1977) conducted dispersion modeling to estimate the 1-hr downwind
concentrations at selected distances from a model plant. These 1-hr
average concentrations were adjusted to annual average omnidirectional
concentrations (for population exposure assessment) by first dividing
them by 20 for conversion to annual maximum values, then further
dividing them by 2.5 to smooth the maximum annual values with respect to
direction. These factors were derived by Youngblood (1978a) and are
based on empirical data and comparisons with more detailed dispersion
model results.
We used the dispersion modeling results for each of the six
emission sources and the related emission factors in Table V-2 to
estimate annual average concentration as a function of distance from a
plant that produces 139,700 mt/yr of AN. the estimated atmospheric
concentrations from each source type were summed at selected distances
from the plant to obtain an estimate of total atmospheric concentrations
due to all six emission sources (Table V-3). Because the American
Cyanamide plant at West Wego, Louisiana and the duPont plant at Memphis,
Tennessee have much larger absorber vent emissions than the other
plants, separate estimates for atmospheric AN concentrations resulting
from these emissions are given in Table V-3.
Exposure Estimates
The dispersion modeling in Table V-3 is for a model plant producing
139,700 mt/yr of AN, with emissions of 258 mt/yr. Atmospheric AN
concentrations around individual plants were estimated by multiplying
19
-------
Table V-3
DISPERSION MODELING ESTIMATES OF ATMOSPHERIC
CONCENTRATIONS OF ACRYLONITRILE FROM MONOMER PRODUCTION
Annual Average Omnidirectional AN Concentrations (ug/tn )
Absorber Vent Addition
Distance from American
Plant (tan) All Emission Sources3 Cyanamide duPont
0.4 11.0 0.12 0.05
0.75 5.5 0.59 0.23
1.5 2.4 1.36 0.54
2.5 1.2 1.09 0.43
3.5 0.75 0.80 0.32
5.0 0.46 0.52 0.21
8.0 0.23 0.28 0.11
12.5 0.13 0.16 0.06
17.5 0.08 0.10 0.04
25.0 0.05 0.06 0.03
aAssumes a plant producing 139,700 mt/yr of AN, with AN emissions of 258
mt/yr. Includes emissions from all six sources listed in Table V-l.
bThese concentrations are to be added to the estimated concentrations from
all other emissions sources. Only two plants are affected: American
Cyanamide at West Wego, Louisiana, and duPont at Memphis, Tennessee.
Source: Based on data supplied by Youngblood (1977).
20
-------
Che dispersion modeling concentrations for the model plant by the
plant's actual emissions (Table V-2), divided by 258 mt/yr- We then
added concentrations resulting from the absorber vent for American
Cyanamide, West Wego, and duPont, Memphis, to the estimated
concentrations from the other sources.
We estimated exposures for people residing within 10 concentric
geographic rings about each plant. The following radii were used:
0.3-0.5, 0.5-1,"1-2, 2-3, 3-4, 4-6, 6-10, 10-15, 15-20, and 20-30 km.
We assumed that no one resides within 0.3 km of any plant, since the
property boundaries extend at least this far. The estimated annual
average AN concentration at the midpoint of each radial ring was taken
to represent the exposure concentration for the entire ring.
SRI's computer system, BESTPOP (Suta, 1978) estimated the
population residing within the radial distances specified above. The
population file consists of a grid of 1-km square sections that span the
continental United States. This file was created by assigning the 1960
and 1970 populations to the grid network an by assuming uniform
distribution of population within each of 256,000 enumeration
districts. The computer software accesses the population file and
accumulates residential population within radial rings specified about
any given point. in addition, a rectangular map is printed for the area
around each specified point to show the population by square kilometer.
Table V-4 gives estimated population exposures to AN from monomer
production facilities. We estimate that approximately 2.6 million
people are exposed to annual average AN concentrations of 0.05 to 15
iug/m . No population exposure estimates were made for concentrations
of less than 0.05 Ug/m because the dispersion modeling was not
extrapolated beyond 30 km from any plant and concentration estimates
within 30 km were greater than 0.05 ug/m . All AN producers are
estimated to cause exposures of 0.001 to 0.05 ug/m at distances
beyond 30 km from their locations.
21
-------
Table V-4
ESTIMATED HUMAN POPULATION EXPOSURES TO
ATMOSPHERIC ACRYLONITRILE EMITTED BY PRODUCERS
Annual Average
Atmospheric AN , Number of
Concentration (ug/m ) People Exposed
10.0-14.9 50
5.0-9.99 240
1.0-4.99 64,000
0.500-0.999 140,000
0.100-0.499 1,800,000
0.050-0.099 600,000a
0.010-0.049 0*
Total 2,600,000
Estimated exposures at this concentration range are
underestimated. Correct estimates would have required
extrapolating the monitoring data beyond 30 km from each
production facility.
A measure of risk was defined as the product of the number of
people residing in a geographic ring around a plant times the average AN
concentration in that ring. Risk for each plant is the sum of the risks
in the 10 concentric rings evaluated. Risk, in units of person-ug/m ,
for all plants is the sum of the risks for each plant. This measure
relates directly to cancer mortalities if a no-threshold, linear dose
response relationship is assumed. Total risk for AN emissions from
monomer production is estimated to be 610,000 person-pg/m .
22
-------
VI POPULATION EXPOSURES FROM PRODUCERS THAT USE
ACRYLONITRILE AS A FEEDSTOCK
Sources of Emissions
We estimated human population exposures to atmospheric AN from
chemical production facilities that use AN as a chemical intermediate.
The products we considered are acrylic fibers, modacrylic fibers, ABS
resins, SAN resins, nitrile elastomer, adiponitrile, and acrylamide. In
some cases the same facility produces several of these chemicals.
Table IV-2 lists the facilities that use AN as an intermediate,
their locations, and their capacities. Table IV-3 indicates that,
depending on the chemical, from 48Z to 97Z of the production capacity
was used in 1977 for the chemicals noted above.
Atmospheric Emissions
Mascone (1978a) has estimated annual AN atmospheric emissions for
plants using AN monomer as a chemical intermediate. These emissions are
given in Tables VI-1 through VI-4 for ABS/SAN resin, acrylic and
modacrylic fibers, and adiponitrile production. According to Mascone,
data about AN emissions for acrylamide production are either unavailable
or confidential; nonetheless, he estimates these emissions to be
negligible. Consequently, we assume that the population exposures to AN
from acrylamide are negligible.
23
-------
Table VI-1
ESTIMATED ACRYLONITRILE EMISSION RATES FOR ABS/SAN RESIN PRODUCTION
Producer/Location AN Emissions (mt/yrl
ABTEC, Louisville, KY 125.2
Borg Warner, Washington, WV 1,769.0
Borg Warner, Ottawa, IL 387.4
Dow, Torrance, CA 9.1
Dow, Midland, MI 15.4
Dow, Riverside, MD 5.4
Dow, Allyn's Point, CT 8.2
Dow, Ironton, OH 8.2
Mobile, Joliet, IL 49.9
Monsanto, Addyston, OH 163.3
Monsanto, Muscatine, IA 362.9
Monsanto, Springfield, MA 27.2
Uniroyal, Scotts Bluff, LA 154.2
Total 3,085.0
Source: Mascone (1979a).
Table VI-2
ESTIMATED ACRYLONITRILE EMISSION RATES FOR
ACRYLIC AND MODACRYLIC FIBER PRODUCTION
Producer/Location AN Emissions (mt/yr)
American Cyanamide, Pensacola, FL 90.7
Dow Badische, Williamsburg, VA 725.7a
duPont, Camden, SC 479.9
duPont, Waynesboro, 7A 338.4
Monsanto, Decatur, AL 2,993.7a
Kodak, Kingsport, TN 69.4
Total 4,697.8
aThese plants are currently installing emission controls that
should reduce AN emissions by approximately 802 (Mascone, 1979a)
24
-------
Table VI-3
ESTIMATED ACRYLONITRILE EMISSION RATE
FOR ADIPONITRILE PRODUCTION
Producer/Location AN Emissions (mt/yr)
Monsanto, Decatur, AL 90.7
Source: Mascone (1979a).
Table VI-4
ESTIMATED ACRYLONITRILE EMISSION RATES
FOR NITRILE ELASTOMER PRODUCTION
Producer/Location AN Emissions (mt/yr)
Copolymer Rubber, Baton Rouge, LA 0.9
B. F. Goodrich, Akron, OH 123.4
B. F. Goodrich, Louisville, KY 303.0
Goodyear, Akron, OH 90.7
Goodyear, Houston, TX 78.9
Uniroyal, Plainsville, OH 52.6
Total 649.5
Source: Mascone (1979a).
25
-------
Atmospheric Concentrations
Because of insufficient atmospheric monitoring data in the
vicinities of production plants that use AN as a chemical intermediate
(see Section VII), dispersion modeling was used to estimate atmospheric
AN concentrations and their neighborhood population exposures. Due to
the relatively short stack heights and other source characteristics, the
dispersion modeling is primarily based on rough downwind calculations
under representative meteorological conditions of neutral (Pasquill "D")
atmospheric stability and moderate (4 m/s) wind speed. The 1-hr average
downwind concentrations were adjusted to annual average omnidirectional
concentrations by first dividing them by 20 for conversion to annual
values, then further dividing them by 2.5 to smooth the maximum annual
values with respect to direction. These factors were derived by
Youngblood (1978a) and are based on empirical data and comparisons with
more detailed dispersion modeling results. Table VI-5 gives the results
of atmospheric dispersion modeling for acrylic fibers, modacrylic
fibers, ABS/SAN resin, and nitrile elastomer production.
ABS/SAN resin plant emissions were assumed to emanate from a 100 m
x 100 m area at a height of approximately 17 m. Similarly, the
emissions from the modacrylic fiber plant were assumed to emanate from a
100 m x 100 m area at a height of 10 m. Calculations were made for both
industries using the PAL computerized dispersion model (Youngblood,
1978b).
The emissions from the nitrile elastomer plants were assumed to
emanate from six sources: spray driers, normal driers, liquid rubber,
concentrators, recovery, and fugitive emissions (Youngblood, 1978c).
The emissions from the acrylic fiber plant are assumed to emanate
from as many as 15 point sources (Schewe, 1978). The dispersion
modeling results are given in Table VI-5 for with and without Class "C"
stacks because some of the producers do not have these stacks.
26
-------
Table VI-5
DISPERSION MODEL ESTIMATES OF OMNIDIRECTIONAL ANNUAL AVERAGE
ACRYLONITRILE CONCENTRATIONS* (ug/m3) FOR VARIOUS PRODUCTS
Acrylic Fibers
Distance
From Plant
(km)
0.4
0.75
1.5
2.5
3.5
5.0
8.0
12.5
17.5
25.0
Without
Class C
Stacksb
20.5
9.6
3.6
1.7
1.0
0.60
0.30
0.16
0.10
0.06
With
Class C
Stacksb
6.0
2.8
1.1
0.60
0.42
0.30
0.19
0.11
0.08
0.06
Modacrylic
Fibers
27.5
11.0
3.8
1.7
1.0
0.60
0.30
0.16
0.10
0.06
ABS and
SAN resins
11.5
7.5
3.3
1.6
0.96
0.59
0.29
0.15
0.10
0.06
Nitrile
Elastomer
7.4
5.7
2.8
1.4
0.89
0.55
0.29
0.15
0.09
0.06
All concentrations assume a continuous combined annual average AN
emission rate of 10 g/s from all sources.
See Schewe (1978) for a description of Class C stacks.
Sources: Based on data given by Youngblood (1978b, 1978c) and Schewe
(1978).
27
-------
Exposure Estimates
The dispersion modeling concentrations given in Table VI-5 are for
model plants emitting a total of 10 g/s (315.4 mt/yr) of AN to the
atmosphere from all sources. We estimate atmospheric AN concentrations
around individual plants by multiplying the appropriate dispersion
modeling concentrations for the model plant by the plant's actual
emissions (Tables VI-1 through VI-4) and dividing the total by 315.4
mt/yr. We assumed that the emission source characteristics (but not
emission rates) for adiponitrile production are similar to those for
modacrylic fiber production.
We estimated exposures for people1 residing within 10 concentric
geographic rings about each plant. The following radii were used:
0.3-0.5, 0.5-1, 1-2, 2-3, 3-4, 4-6, 6-10, 10-15, 15-20, and 20-30 km.
We assumed that no one resides within 0.3 km of any plant. The
estimated annual average AN concentration at the midpoint' of each ring
was taken to represent the exposure concentration for the entire ring.
As we did for AN monomer production plants, we used SRI's computer
system BESTPOP (Suta, 1978) to estimate the population residing within
the radial distances specified above.
Table VI-6 gives the population exposures to AN from the production
of ABS/SAN resins, acrylic and modacrylic fibers, nitrile elastomer, and
adiponitrile. Note that because exposures were estimated to only 30 km
distance from each plant, very few people were found in the lower
exposure concentrations (0.001-0.010 yg/m ). It is reasonable to
expect that a great number of people are exposed to these concentrations
beyond 30 km from the plants.
We defined the measure of risk as the product of the number of
people residing in a geographic ring around a plant times the average AN
concentration in that ring. Risk for each plant is the sum of the risks
in the 10 concentric rings evaluated. Risk for all plants is the sum of
the risks for each plant. The units on risk are person-yg/m . This
28
-------
measure relates directly to cancer mortalities if a no-threshold, linear
dose response function is assumed. Total risks are estimated as 500,000
3 3
person-pg/m for ABS/SAN resins, 270,000 person-yg/m for acrylic
2
and modacrylic fibers, 320,000 person-pg/m for nitrile elastomers,
and 9,400 person-yg/m for adiponitrile.
Table VI-6
ESTIMATES OF POPULATION EXPOSURES TO ATMOSPHERIC ACRYLONITRILE
EMITTED BY PLANTS THAT USE IT AS A CHEMICAL INTERMEDIATE
Annual Average
AN Concentration
(u g/m3 )
Chemical Product
Acrylic and
ABS/SAN Modacrylic Nitrile
Resin Fibers Elastomers
Adiponitrile
15.0-19.9
10.0-14.9
5.00-9.99
1.00-4.99
0.500-0.999
0.100-0.499
0.050-0.099
0.010-0.049
0.005-0.009
0.001-0.004
Total People
Total Risk
2,700
850
73,000
79,000
680,000
1,200,000
l,400,000a
510,000a
790,000a
4,700,000 .
500,000
4,700
52,000
70,000
370,000
190,000
260,000a
oa
oa
950,000
270,000
1,800
22,000
81,000
650,000
690,000
2 , 700 , 000
5,100a
93,000a
4,200,000
320,000
22,000
32,000
65,000a
oa
oa
120,000
9,400
aExposures in these ranges are underestimated because they were only
estimated to 30 km from each plant.
29
-------
VII COMPARISON OF MONITORING AND
DISPERSION MODELING CONCENTRATIONS
General
We calculated population exposures by using dispersion modeling to
estimate atmospheric AN concentrations. We would have preferred to base
the exposure estimates solely on monitoring data; however, few AN
monitoring data have been collected in the vicinity of the production
plants. It is, nonetheless, of value to compare the dispersion model
estimates of atmospheric concentrations with those limited monitoring
data to determine if the two are in reasonable agreement with one
another.
Atmospheric Monitoring
Going (1978) sampled atmospheric AN in the vicinity of 11
industrial sites. The selected sites included 2 producers of AN, 1
producer of AN and acrylamide, 1 producer of acrylamide, 3 producers of
acrylic and modacrylic fibers, 2 producers of ABS and SAN resins, 1
producer of ABS and SAN resins and nitrile elastomers, and 1 producer of
nitrile elastomers.
Four to eight monitoring stations were positioned at varying
directions and distances from each of the 11 sites. More than one air
sampler was deployed at some of the stations as part of the quality
assurance program. Samples were collected over one approximate 24-hr
period for all stations, except for Monsanto in Decatur, Alabama, where
two 24-hr samples were collected.
30
-------
Quality assurance of the AN sampling program was maintained in
three ways. First, blanks were taken to the site and treated as if they
were real samples. At least four blanks were used at each site. The AN
detected in each of the 16 blank samples analyzed was less than 0.3 yg,
which corresponds to the detection limit. Second, duplicate air samples
were collected at one or more stations per site, with at least one of
the stations downwind of the source. The average difference between the
13 sets of duplicate samples was 20%, with a standard deviation of 25%.
Finally, spiked sampling tubes were used at two or more stations per
site to establish the actual recovery of AN under real conditions. The
average recovery was 63% for the 53 spiked tubes analyzed.
Table VII-1 sets forth the monitoring results. Wa have averaged
the concentrations over duplicate quality assurance samples and over the
two 24-hr periods for Decatur. We have estimated the distance that each
monitoring station is from the AN producton facilities within a plant.
Because some of the plants are fairly large, monitoring stations placed
at their boundary can be 0.5 km or more from the AN production
facilities.
AN was found in the atmosphere at all monitoring stations at
average 24-hr levels for individual stations ranging from 0.1 to 249
3 3
Vig/m . The highest individual 24-hr concentration was 325 ug/m .
Because all sampling was conducted over one or two 24-hr periods, the
recorded concentrations depend greatly on the meteorological conditions
at the time of sampling, the placement of the monitoring stations, and
on the AN production within the plant at the time of sampling.
The elevated concentrations found near the American Cyanamide Plant
in Linden, New Jersey, indicate that significant environmental AN
emissions may result from the production of acrylamide. We have assumed
in the exposure assessments that acrylamide production has negligible AN
emission.
31
-------
Table VII-1
ATMOSPHERIC MONITORING DATA FOR ACRYLONITRILE
Plant/Location
American Cyanamide,
New Orleans, LA
American Cyanamide,
Linden, NJ
Monsanto
Texas City, TX
Monsanto
Decatur, AL
duPont-May
Lugoff, SC
duPont,
Waynesboro, VA
Station
Number
1
2
3
4
5
1
2
3
4
5
6
1
2
3
4
5
6
7
1
2
3
4
1
2
3
4
5
6
7
8
1
2
3
4
5
6
Distance*
(km)
0.7
1.8
1.2
0.6
0.6
1.8
1.4
1.3
1.0
1.8
0.5
1.5
2.6
1.9
1.6
0.8
2.2
0.7
5.0
1.4
1.8
2.2
1.3
0.7
2.1
1.9
1.5
2.2
1.6
1.3
0.8
0.9
0.3
0.4
0.7
0.5
0.5
Average AN
Concentration11
(ug/m3)
12.3 (13.6, 11.0)
0.4
0.6(0.6, 0.5)
15.9
1.6
<0.1
0.7
0.5
8.9
5.2
2.1
<0;3
2.4 (2.2, 2.6)
0.9
2.3
<0.2
<0.2)
1.8 (4.0, 2.5, <0.3)
2.3 (4.2, <0.3)
0.8 (1.6, <0.2, 1.3,
0.2)
0.6 ( 0.1, < 0.1, 1.1)
0.3
0.2
0.2
0.1
0.1
<0.1
0.3
1.1
<0.2
7.0
<0.2
<0.2
<0.2
<0.2
.2, <0.2)
32
-------
Table VII-1 (Concluded)
Plant/Location
Borg-Warner
Washington, WV
B. F. Goodrich
Louisville, KY
Station
Number
1
2
3
4
1
2
3
4
Distance
(km)
1.3
0.7
0.5
1.2
0.3
3.0
0.4
2.2
Average AN
Concentration
(yg/m )
0.3
86.1 (99.6, 72.5)
249.4 (173.7, 325.1)
<0.2
4.3
<0.2
<0.2
<0.2
0.2)
Monsanto,
Addysto-n, OH
Uniroyal,
Plainsville, OH
Vistron,
Lima, OH
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
0.7
0.3
1.0
0.7
0.8
0.9
0.3
0.7
0.4
0.4
0.3
0.4
0.2
0.3
0.6
<0.2
<0.2
1.1 (1.1, 1.0)
0.9
<0.2
0.9
3.1
0.4
0.7 (0.7, 0.7)
<0.2
14.5
16.1
141.0 (134, 148)
<0.2
Approximate distance from the AN production or use facilities within the
plant.
"Most concentrations are for one 24-hr period. Some are averaged over
duplicate samples taken at the same stations, and some are averaged over two
24-hr periods. Individual values for multiple sampling are shown in
parentheses.
33
-------
Comparisons
Table VII-2 compares the monitoring concentrations and the dispersion
modeling concentrations by 0.5-km radial increments around AN plants. The
dispersion modeling concentrations are estimated annual omnidirection averages
from all sources. They were calculated for the midpoint of each 0.5-km radial
increment with the procedures used to estimate population exposures. The
monitoring concentrations are the averages over all sampling locations within
an 0.5-km radial increment. Thus, each monitoring concentration is the
average of one or more sampling station over one or two 24-hr periods.
Because of the differences in the two methods (monitoring and modeling),
we cannot expect close agreement between their estimated concentrations. Of
the 29 comparisons in Table VII-2, the monitoring and modeling concentrations
were equal in one case, the modeling concentrations were larger in 22 cases,
and the monitoring concentrations were larger in 6 cases. Overall, the
monitoring concentrations were about 30% less than the dispersion modeling
concentrations. This difference corresponds well with the spiked sampling
results, which indicate that the average recovery for AN monitoring is 63%.
That is, the monitoring concentrations are expected to be about 37% less than
the actual concentrations due to incomplete AN recovery. Since the modeling
concentrations are 30% higher than the monitoring concentrations, estimates
based on modeling should be fairly close to actual concentrations.
34
-------
Table VII-2
COMPARISON OF MONITORING AND DISPERSION MODELING DATA
Plant/Location
American Cyanamide
New Orleans, LA
American Cyanamide
Linden,. NJ
Monsanto
Texas City, TX
Monsanto
Decatur, AL
duPont-May
Camden, NJ
duPont
Waynesboro, VA
Borg-Warner
Washington, WV
B. F. Goodrich
Louisville, KY
Distances3
(km)
0.50-0.99
1.00-1.49
1.50-1.99
0.50-0.99
1.00-1.49
1.50-1.99
0.50-0.99
1.00-1.49
1.50-1.99
2.00-2.49
2.50-2.99
1.00-1.49
1.50-1.99
2.00-2.50
2.50-4.99
5.00-5.49
0.50-0.99
1.00-1.49
1.50-1.99
2.00-1.50
0.30-0.49
0.50-0.99
0.50-0.99
1.00-1.49
0.30-0.49
0.50-1.99
2.00-2.49
2.50-2.99
3.00-3.49
AN Concentration (ug/m-3)
Monitoring^
4.3
0.1
0.1
0.5
6.0
0.7
2.4
nd
3.8
0.9
5.2
1.2
2.3
0.8
nd
0.2
0.7
0.3
0.1
0.2
3.6
0.2
157.6
0.3
2.3
nd
0.2
nd
0.2
Dispersion0
Modeling
4.3
2.5
1.9
d
d
d
6.5
2.5
1.8
1.3
21.0
9.3
7.2
3.1
7.4
2.3
1.3
1.1
5.2
3.4
42.1
30.4
7.2
1.7
1.0
35
-------
Table VII-2 (Concluded)
AN Concentration (ug/m^)
Plant /Location
Monsanto
Addyston, OH
Uniroyal
Plainsville,
Vistron
Lima, OH
Distances^
(km)
0.30-0.49
0.50-0.99
1.00-1.49 1.1
0.30-0.49
OH 0.50-0.99
0.30-0.49
0.50-0.99
Monitoring*3
0.2
0.4
2.8
1.3
0.7
43.4
0.2
Dispersion0
Modeling
5.9
3.9
1.2
0.9
10.6
5.2
aEstimated distance from the AN production within the plant.
bAverage of all monitoring stations within the indicated distances; all
concentrations reported as "less than" (i.e., < ) are assumed to be at their
upper limit.
cEstimated concentrations at the midpoint of the distances.
^Dispersion modeling estimates were not made for acrylamide plants.
nd « no data.
36
-------
BIBLIOGRAPHY
Going, J. E., "Environmmental Monitoring Near Industrial Sites:
Acrylonitrile," Midwest Research Institute (1978).
International Agency for Research on Cancer, "IARC Monographs on the
Evaluation of the Carcinogenic Risk of Chemicals to Humans," Volume
19 (February 1979).
Joshi, S. B., Northrop Services Corp., Letter to J. Cirvello, USEPA,
regarding photochemical reactivity of acrylonitrile (July 10, 1977).
Mascone, D. C., EPA, private correspondence (May 24, 1979a).
, EPA, personal conversation (May 31, 1979b).
Miller, L. M., and J. E. Villaume, "Investigations of Selected Potential
Environmental Contaminants: Acrylonitrile," The Franklin Institute
Research Laboratories, EPA 560/2-78-003 (1978).
Patterson, R. M., M. I. Bornstein, and E. Garshick, "Assessment of
Acrylonitrile as a Potential Air Pollution Problem," GCA
Corporation, GCA-TR-75-32-G(6) (1976).
Schewe, G. J., EPA, memo concerning "Rough Dispersion Estimates of
Ambient Acrylonitrile Concentrations from Acrylic Fiber Plants," to
J. O'Connor (November 7, 1978).
Suta, B. E., "BESTPOP: A Fine-Grained Computer System for the
Assessment of Residential Population," SRI International (1978)-
37
-------
Youngblood, P. L., EPA, memo concerning "Dispersion Modeling for
Determining Population Exposures to Benzene," to R. Johnson
(January 4, 1978a).
^ , EPA, memo concerning "Rough Estimates of Ambient
Concentrations of Acrylonitrile," to J. D. Cirvello (January 26,
1978b).
, EPA, memo concerning "Rough Dispersion Estimates for
Acrylonitrile from Nitrile Elastomer Plants," to J. O'Connor
(September 15, 1978c).
, EPA, memo concerning "Rough Estimates of Ambient
Concentrations of Acrylonitrile," to J. D. Cirvello (September 22,
1977).
38
-------
Final Report
July 1978
ATMOSPHERIC ETHYLENE OIBROMIDE:
A SOURCE-SPECIFIC ASSESSMENT
Prepared for:
Office of Toxic Substances
U.S. Environmental Protection Agency
Washington, D.C. 20460
Project Officer: Vincent J. DeCarlo
Task Officer: Richard J. Johnson
SRI Project 6839
Prepared by:
Susan J. Mara
Shonh S. Lee
Center for Resource and Environmental
Systems Studies Report \lo. 39
333 Ravenswood Ave. Menlo Park, California 94025
(415) 326-6200 - Cable: STANRES, Menlo Park «TWX: 910-373-1246
-------
NOTICE
This final report represents the results of work completed in
November 1977 and is provided to the U.S. Environmental Protection Agency
(EPA) by SRI International, Menlo Park, California, in fulfillment of
an agreement with the American Public Health Association. Comments on
the draft report (November 1977) were received from EPA in Spring 1978
and have been incorporated into this final report. The opinions, findings,
and conclusions expressed here are those of the authors, and not necessarily
those of EPA. Mention of company or product names is not to be considered an
endorsement by EPA.
-------
CONTENTS
NOTICE iii
LIST OF ILLUSTRATIONS vii
LIST OF TABLES ix
ACKNOWLEDGMENTS _ xi
I SUMMARY 1
II ETHYLENE DIBROMIDE IN THE ENVIRONMENT 5
A. Introduction 5
B. Chemical and Physical Properties 6
C. General Methodology 8
III CHEMICAL MANUFACTURING AND FORMULATING FACILITIES 11
A. Sources 11
B. Methodology 14
C. Exposures 20
IV GASOLINE SERVICE STATIONS 23
A. Sources 23
B. Methodology and Exposures 23
V PETROLEUM REFINERIES 39
A. Sources - 39
B. Methodology 39
C. Exposures 44
VI STORAGE AND DISTRIBUTION OF GASOLINE 47
A. Sources 47
B. Methodology and Exposures 47
VII URBAN EXPOSURES RELATED TO AUTOMOBILE EMISSIONS 55
A. Sources ,.
B. Methodology and Exposures
_>o
BIBLIOGRAPHY 65
APPENDIX - CAPACITIES AND EXPOSED POPULATION BY
PETROLEUM REFINERY AND STATE 71
-------
ILLUSTRATIONS
III-l Commercial Pathways of Ethylene Dibromide 12
III-2 Projected Dispersion Curve for Manufacturing
and Formulating Facilities 18
IV-1 Average Concentration of Ethylene Dibromide in
Air at 18 Sampling Stations Near a Retail Gasoline
Site in Phoenix, Arizona 34
IV-2 Projected Dispersion Curve for Annual Average
Concentrations in the Vicinity of Service Stations 35
V-l Projected Dispersion Curve for Petroleum Refineries 43
VI-1 The Gasoline Marketing Distribution System in
the United States 50
VI-2 Vapor and Liquid Flow in a Typical Bulk Terminal 52
VII-1 Isopleths of Mean Annual Wind Speed Through
the Morning Mix Layer . . . . , 59
-------
TABLES
1-1 Summary of Human Exposures to Atmospheric
Ethylene Dibromide from Emission Sources 3
II-l Physical Properties of Ethylene Dibromide 7
III-l Capacities of Manufacturing and Formulating Facilities . . 13
III-2 Results of Ethylene Dibromide Monitoring in the
Vicinity of Manufacturing Facilities 15
III-3 Estimated Population Exposures from Manufacturing
and Formulating Facilities 21
IV-1 Self-Service Operations 25
IV-2 Gasoline Market Share of Self-Service Stations
in Four AQCRs, Spring 1977 26
IV-3 Gasoline Market Share of Self-Service Stations
in 'Two Metropolitan Areas, 1976 27
IV-4 Sampling Data from Self-Service Gasoline Pumping 28
IV-5 Estimates of Ethylene Dibromide Exposures from
Self-Service Gasoline Pumping 29
IV-6 Service Station Density in Four Metropolitan AQCRs .... 31
IV-7 Summary of Population Exposed to Ethylene Dibromide
from Gasoline Service Stations 37
V-l Monitoring Data in the Vicinity of Petroleum Refineries . 40
V-2 Estimated Population Exposed to Ethylene Dibromide
from Petroleum Refineries 45
VII-1 Automotive Ethylene Dibromide Emission Factors
VII-2 Monitoring Data for Ethylene Dibromide in Urban Areas . . 56
VII-3 Estimates of Average Annual Ethylene Dibromide
Concentrations for Cities with Populations
Exceeding 1,000,000 61
VII-4 Estimates of Annual Average Ethylene Dibromide
Concentrations for Selected SMSAs 63
VII-5 Urban Population Exposures Related to Automotive
Emissions 64
-------
ACKNOWLEDGMENTS
It is a pleasure to acknowledge the cooperation and guidance given
by several individuals of the U.S. Environmental Protection Agency.
Dr. Vincent DeCarlo, Office of Toxic Substances, was project officer.
Mr. Richard J. Johnson, Office of Air Quality Planning and Standards,
provided vital direction throughout the study. Mr. Phillip L. Youngblood,
Office of Air Quality Planning and Standards, provided valuable suggestions
about the application of dispersion modeling results to the study of
atmospheric ethylene dibromide.
Messrs. Troy P. Miller and Benjamin E. Suta, SRI project supervisors,
gave vital support. Mr. Scott W. bailey edited the report superbly.
-------
I SUMMARY
This report.is one in a series that SRI International is preparing
on a quick-response basis for the U.S. Environmental Protection Agency
(EPA). Populations at risk to selected pollutants are being quantified
for input to other,-more inclusive studies. This study was undertaken
to quantify the environmental exposure of the population to atmospheric
emissions of ethylene dibromide (EDB).
Ethylene dibromide is used primarily as a lead scavenger in gasoline.
Additional uses include the production of vinyl bromide, soil and space
fumigant, production of other plastics, and solvent. There are no known
natural sources of the compound.
The five primary sources of atmospheric ethylene dibromide are:
chemical manufacturing and formulating facilities, gasoline service stations,
petroleum refineries, the storage and distribution of gasoline, and urban
exposures related to automobile emissions. Because ethylene dibromide is
combusted with gasoline, only evaporative emissions from automobiles have
been estimated.
Few quantitative data were available for this study. When data were
available, source locations were identified and ethylene dibromide emission
rates were estimated. Ambient concentrations of ethylene dibromide were
then estimated by applying approximate results of dispersion modeling
developed by EPA for an exposure study of benzene, and then extrapolating
the results to ethylene dibromide based on differences in vapor pressure
and in concentration. Population exposed to concentrations of 1.0 part
*
per trillion (ppt) and greater were estimated. When data were unavailable,
best estimates were developed to provide a reasonable basis for comparison.
*
The assumed limit of detection.
-------
All estimates presented In the report are subject to considerable
uncertainty concerning: (1) the quantity of ethylene dibromide emissions,
(2) production aad consumption levels of ethylene dibromide, (3) source
locations, (4) control technologies employed, (5) deterioration of control
technologies over time, and (6) the dispersion characteristics of ethylene
dibromide. The estimates, though not precise, provide an approximate
estimate of expected conditions.
Table 1-1 summarizes the results of the study. Exposures from gasoline
service stations and urban exposures related to automobile emissions con-
stitute the two largest sources. Petroleum refineries rank third, exposing
more than two million people. Chemical manufacturing and formulating
facilities are sources of exposure for more than one million people.
For an approximate comparison of different emission sources, we have
calculated exposures in similar units by multiplying the number of people
exposed by the annual average concentration of ethylene dibromide within
each range. Those values were then summed for each emission source
(Table 1-1). Thus, the units become "ppt-person-years." For exposures
at self-service gasoline stations, an exposure time of 1.5 hours per
person per year was assumed.
By far, the most highly-weighted exposures are to people who live in
the vicinity of gasoline service stations. Urban exposures related to
the evaporation of gasoline from automobiles have the next highest value.
Third are chemical manufacturing and formulating, followed by petroleum
refineries. These results differ because they are weighted by the number
of people exposed to a particular level of atmospheric ethylene dibromide.
Thus, they provide a useful basis for comparison and, assuming a linear
dose-response relationship, are directly related to human health.
As mentioned previously, the estimates given in this report are sub-
ject to considerable uncertainty. Further monitoring and sampling data
would be required to improve the accuracy of the analysis. Although the
estimated population exposed is substantial, significant reductions in
exposures can be expected by 1980 after leaded gasoline is phased out.
-------
Table 1-1
SUMMARY OF HUMAN EXPOSURES TO ATMOSPHERIC
ETHYLENE DIBROMIDE FROM EMISSION SOURCES
Population Exposed to Ethylene Dibromide Concentrations (ppt)a
8-Hour Worst Case:
Annual Average:
Source
Manufacturing and Formulating
Gasoline Service Stations
1. People Using Self-Service
2. People Living in the Vicinity
Petroleum Refineries
Storage and Distribution
Urban Exposures - Automotive
10.0 - 50.0
1.0 - 5.0
580,000
100,000,000
2,000,000
e
24,000,000
50.1 - 100.0
5.1 - 10.0
250.000
10.000,000
170,000
100.1 - 200.0
10.1 - 20.0
160,000
53,000
200.1 - 400.0
20.1 - 40.0
99.000
16,000
> 400.0
> 40.0
84,000
d
3,000
.Totalb'C
1,200,000
30,000,000
110,000,000
2,200,000
24,000,000
Comparison
Among Sourcesc
(10° ppt-person-years)
14.0
1.3
380.0
8.7
72.0
To convert Co ug/m , divide each exposure level by 130.
Population estimates are not additive vertically, because some double counting may exist.
Totals are roundud to two significant figures.
Estimated at 260 ppt for 1.5 hr/yr/person.
Estimated at « 1.0 ppt annual average.
-------
II ETHYLENE DIBROMIDE IN THE ENVIRONMENT
A. Introduction
The primary objective of this study was to quantify the environmental
atfmospheric exposure of the general population to emissions of ethylene
dibromide (EDB). As noted in the summary, this report is one in a series
of studies being conducted by SRI for the U.S. Environmental Protection
Agency (EPA) to quantify populations at risk to selected pollutants.
The studies are generally conducted on a quick-response basis to provide
input to other, more inclusive studies. In this project, we identified
sources of ethylene dibromide emissions, estimated the resultant atmospheric
environmental concentrations of ethylene dibromide, and estimated human
populations exposed to various levels of ethylene dibromide concentration.
This study has not considered the degree of biological sorption of
material. No attempt was required or has been made in this report to
assess potential health effects.
Ethylene dibromide is used primarily as an additive in leaded gasoline,
a use that accounts for more than 85% of domestic consumption. Ethylene
dibromide is used also in vinyl bromide, fumigants, solvent, and other
plastics. Although ethylene dibromide has been used in both soil fumigants
and commodity and space fumigants, one producer (Dow Chemical USA) has
recently ceased marketing its commodity and space fumigant mixture contain-
ing the chemical (SRI estimates). Because of the current phasing out of
leaded gasoline, production of ethylene dibromide can be expected to fall
toward 15% of its estimated current level (Mitre, 1976).
Concern over the possible carcinogenic properties of ethylene dibromide
has led the National Institute for Occupational Safety and Health (NIOSH)
3
to issue a new standard for the workplace of no more than 1.0 mg/m (0.13 ppm)
of the compound as determined by a 15-minu'te sampling period. Occupational
exposures in ethylene dibromide manufacturing facilities are often above
-------
3 3
1.0 mg/m and have been measured as high as 140 mg/m (18 ppm) (Joiner,
personal communication, 1977). Generally, the concentrations measured
at worksites at which ethylene dibromide is unloaded, transferred, or
stored, or where process vessels are being maintained, are consistently
higher than samples taken around the manufacturing or blending equipment
itself.
Although environmental concentrations of ethylene dibromide are
typically three orders of magnitude below occupational levels of exposure,
the widespread nature of the emission sources requires that serious atten-
tion be given to estimations of environmental exposures.
For this report, five sources of atmospheric ethylene dibromide were
considered: manufacturing and formulating; gasoline service stations;
petroleum refineries; storage and distribution of gasoline; and urban
exposures related to automobile emissions. Because ethylene dibromide
undergoes combustion with the gasoline, we have considered only evaporative
losses.
B. Chemical and Physical Properties
Ethylene dibromide (C-H.Br-; 1,2-dibromoethane) is a dense, colorless,
nonflammable liquid that resembles chloroform in odor. The principle
physical properties of the compound are shown in Table II-l.
Ethylene dibromide shows a reasonable tendency to degrade in both
atmospheric and aquatic receiving environments. The chemical shows
a half-life of five to 10 days toward hydrolysis under neutral conditions
at ambient temperatures in the aquatic environment (Mitre, 197"6). Ethylene
dibromide resists atmospheric oxidation by peroxides and ozone, typically
showing half-lives in excess of 100 days in those reactions (Mitre, 1976).
The compound is generally less reactive in the atmosphere than correspond-
ing alkanes or olefins.
Ethylene dibromide is generally inert at normal temperatures, although
slight decomposition may result from exposure to light. It is hydrolyzed
to ethylene glycol and bromoethanol at elevated temperature. When heated
to 340 to 370C, ethylene dibromide decomposes into vinyl bromide and
-------
Table II-l
PHYSICAL PROPERTIES OF ETHYLENE DIBROMIDE
Chemical Formula BrCH -CH Br
Molecular Weight 187.88
Boiling Point 131.6C
Melting Point 9.9C
Vapor Pressure, 25C 12 mm
Specific Gravity, liquid 2.17
Specific Gravity, vapor, 25C 6.5
Refractive Index 20C 1.5379
Solubility, water, 20C 4.3 g/1
Solubility, Octanol °°
Source: Kirk, 1968
-------
hydrobromic acid. The terminal halogen atoms are reactive, making the
compound a useful synthetic intermediate. It is the least expensive
organic bromine compound available (Mitre, 1976).
Because of a lack of relevant data, it is impossible to assess
environmental accumulation of the chemical. It is not known whether
rates of atmospheric degradation are sufficient to handle the environ-
mental burden adequately.
C. General Methodology
Few studies have addressed the dispersion of ethylene dibromide in
the atmosphere. Little information is available concerning emission factors
and precise locations of sources. This study represents a rough approxi-
mation of ambient concentrations of ethylene dibromide in the vicinity of
sources, and an approximation of exposures to the estimated population.
Much of the work was based on a previous SRI study (Mara and Lee, 1977)
that evaluated the population exposed to atmospheric benzene. In most
cases, it was necessary to rely on dispersion modeling or emission factors
developed for benzene sources and to extrapolate to the expected values
for ethylene dibromide.
A discussion of the general methodology is warranted. Because evapor-
ation accounts for the major loss of ethylene dibromide to the atmosphere,
it is necessary to determine the evaporation rate of the chemical with
respect to the evaporation of benzene. It is known that the evaporation
rate is proportional to the vapor pressure, solubility, and the molecular
weight. Thus, the following equation can be used to estimate the ethylene
dibromide emission factor (or emission rate) related to evaporation:
E P S
3 a a «Jma O 1 >
0< a \t-'±)
Eb Vb ^
where the subscript a refers to ethylene dibromide and the subscript b
refers to benzene; E is the emission rate (or emission factor); P is the
vapor pressure; S is the solubility; and m is the molecular weight.
K
This study has subsequently been revised, but the new findings have not
been incorporated into this report.
8
-------
For an estimation, (S) Gv/in) may be approximated by x» tne molar
fraction or concentration. Thus, Equation (2.1) can be written as
follows:
^a ^ Pa Xa (2.2)
p1 "P V
t> b ' b
Because the vapor pressures for ethylene dibromide (12 mm) and benzene
(100 mm) at room temperature (<«25C) are known, and the volume concentra-
tions of ethylene dibromide (0.05%) and benzene (2.0%) in gasoline are
also available, the emission factor (or emission rate) of ethylene dibromide
can be estimated by the following equation:
12 x 0.05 (2.3)
a 100 2.0 T>
E =
a 320
This ratio was applied to benzene emission factors related to gasoline
emissions in order to estimate comparable factors for ethylene dibromide.
When gasoline was not the source of emission(s) the differences in vapor
pressure were used alone (i.e., E = 0.12 K ).
We emphasize again that few quantitative data were available for
this study. All estimates given in the report are subject to a large
degree of uncertainty related to the quantity of ethylene dibromide emissions,
production and consumption levels of ethylene dibromide, source locations,
control technology employed, deterioration of control technology over time,
and dispersion modeling. Although the estimates are not precise, they do
provide a reasonable evaluation of expected conditions.
-------
Ill CHEMICAL MANUFACTURING AND FORMULATING FACILITIES
A. Sources
Ethylene dibromide is manufactured in locations that have commercial
bromine deposits. Approximately 75% of the production is clustered in the
Gulf Coast region of the United States. The capacity for producing ethylene
dibromide was approximately 365 x 10 Ib/yr (168 x 10 kg/yr) in 1975 (SRI
estimates)- Since that time, capacity is estimated to have been reduced
only slightly to 350 x 106 Ib/yr (159 x 106 kg/yr).
Approximately 340 x 10 Ib (154 x 10 kg) of ethylene dibromide were
sold in 1973 (SRI estimates).
As mentioned earlier, ethylene dibromide is used primarily as a lead
scavenger in gasoline. Domestic consumption for that use in 1973 was
estimated at 225 x 10 Ib (102 x 10 kg), which represents more than 85%
of the compound's consumption. The chemical has shown an annual growth
of 3 to 5 percent during the last decade, but as noted previously, consump-
tion is expected to fall off sharply because of the declining use of leaded
gasoline. The use of ethylene dibromide as a lead scavenger in gasoline
requires blending into an antiknock "motor mix," also known as tetraethyl
lead or TEL. The blending takes place at six locations scattered through-
out the United States. Figure III-l shows the commercial pathways of
ethylene dibromide in the economy.
Additional uses of ethylene dibromide include the production of vinyl
bromide, (5%), fumigant and solvent (10%), and in other plastics (minor
market). There are no known natural sources of the chemical. Locations
and information on approximate capacity for each manufacturing plant are
shown in Table III-l.
Limited monitoring data for two manufacturing facilities were collected
by Midwest Research Institute (MRI) in 1975. One air sampling site was
11
-------
MILLION LB/YR
270
IMPORTS
NEGLIGIBLE
MILLION LB/YR
EXPORTS
NEGLIGIBLE
MILLION LB/YR
TOTAL U.S.
PRODUCTION
(P)
315.6
MILLION LB/YR
TOTAL U.S.
CONSUMPTION
(0
315.5
MILLION LB/YR
LOSS OF
COMPOUND
4.7
MILLION LB/YR
LOSS Of
BY-PRODUCT
N.A.
MILLION LB/YR
DISPERSIVE USE
SCAVENGER FOR
LEAD IN GASOLINE
1.6
FUMIGANT
28.2
299.7 MILLION LB/YR
CAPTIVE USES
SOLVENT, OTHER
15.8 MILLION LB/YR
SYNTHETIC
INTERMEDIATE
Source: After Brown et al, 1975
FIGURE 111-1. COMMERCIAL PATHWAYS OF ETHYLENE DIBROMIDE
-------
Table III-l
CAPACITIES OF MANUFACTURING AND FORMULATING FACILITIES
Location
Arkansas
El Dorado
Great Lakes Chem-
ical Corp.
Magnolia
Dow Chemical Co.
Ethyl Corp.
(Bromet Co.)
Total
California
Antioch
E. I. du Pont de
Nemours & Co. Inc.
Louisiana
Baton Rouge
Ethyl Corp
Michigan
Midland
Dow Chemical Co.
New Jersey
Deepwater
E. I. du Pont de
Nemours & Co. Inc.
Texas
Beaumont
Houston Chemical Co.
(PPG Industries)
Freeport
Nalco Chemical Co.
Pasadena
Ethyl Corp.
EDB
Capacity
106kg
23
14
li
82
39
16
Production
of Vinyl
Bromide
Production
of
Fumigant
TEL
Capacity
106kg
79
79
79
54
18
79
Estimated
EDB use for
TEL (105kg)
14
14
14
10
3
14
Total
160
388
69
Source: SRI estimates.
13
-------
established within 100 m of each plant. The results are shown in Table
III-2. Because emission factors were not available, the monitoring data
were used to estimate the emission rate at manufacturing facilities.
Consequently, the exposure estimates are considered to be crude approxi-
mations.
B. Methodology
Each chemical manufacturing and formulating facility may have different
production rates, chemical processes, geographic locations, pollution con-
trol technology, and meteorological conditions. Thus, detailed dispersion
calculations are impractical, given the scope of this study. A simple
method of assessment was therefore developed to allow comparative analysis.
Variations in geographic locations and meteorological conditions were not
considered in the analysis. The results are not precise; rather, they
provide a reasonable order-of-magnitude estimate of atmospheric concentra-
tions of ethylene dibromide. A single dispersion curve was projected and
applied to all chemical manufacturing and formulating facilities based on
previous work done for benzene (Mara and Lee, 1977). The derivation of
the methodology is discussed below.
No information on emission factors or characterization of emissions
was available for ethylene dibromide manufacturing and formulating facil-
ities. The available monitoring data indicate that emissions are high in
comparison to those from other sources. The monitoring data were used as
the basis for extrapolation. Rough dispersion modeling of benzene by
Youngblood of EPA (1977b) was used and adjusted by monitoring data to
project a dispersion curve for ethylene dibromide. The shape of the
curve remains approximately the same for ethylene dibromide, but its
relative position changes, based on the difference in emission rates.
The general dispersion equation developed by regression analysis is
as follows:
-1 48
C = A EA D -1--40 (3.1)
where C is concentration at distance D; A is the system constant, and E
A.
is the estimated emission rate.
14
-------
Table III-2
RESULTS OF ETHYLENE DIBROMIDE
MONITORING IN THE VICINITY OF
MANUFACTURING FACILITIES
Total
Number ^ Sampling , Concentration Standard
Company Location of Sites Time (hr) (ppb) t Deviation
Dow Magnolia, Ark. 1 4 13.2 0.9
Ethyl Corp. Magnolia, Ark. 1 8 3.1 4.3
A
All sites were within 100 m of the plant.
Samples were discontinuous; two were taken near the Dow facility and five near
the Ethyl Corporation facility.
+ 3
tTo convert to yg/m , divide by 0.13.
Source: Midwest Research Institute, 1975
15
-------
The monitoring data for the Dow facility show that the ethylene
3
dibromide concentration at 0.1 km is approximately 100 yg/m (13 ppb),
sampled at a downwind location. (Because the Ethyl Corporation data are
highly variable and the sampling site somewhat upwind from the facility,
those data were not used in this analysis.) With wind variability considered,
20 yg/m (3 ppb) is considered to be a reasonable value for the ambient
concentration at 0.1 km. Equation (3.1) can be rewritten as follows:
C = A' D-1'48 (3.2)
where A' is a value representing the product of the system constant and
the emission rate. The monitoring data for the Dow facility is then incor-
porated into Equation (3.2) to solve for A':
(20) = A' (O.I)'1'48 (3.3)
A1 = 0.717
Thus, Equation (3.2) becomes:
C = 0.717D'1'48 .(3.4)
Rearranging Equation (3.4), it becomes:
D = 0.798 (i) °'6757 (3.5)
L<
The projected dispersion curve for ethylene dibromide emissions from
manufacturing and formulating facilities based on Equation (3.4) and
(3.5) is shown in Figure III-2. Because Equation (3.5) and Figure III-2
are based on the production level at the Dow facility, the equation must
be normalized to account for varying production rates. The normalized
equation can be written as follows:
D.= 0.798 |f (I) °'6757 (3.6)
i
where C± is the concentration at distance D.; and Pn/Pd is the ratio of
the production rate at the facility of interest to the Dow production
rate (14 x 10 kg/yr).
For uniformity, we have established the ranges of ethylene dibromide
concentrations that follow and that apply to all sources:
16
-------
1.0 - 5.0 ppt
5.1 - 10.0 ppt
10.1 - 20.0 ppc
20.1 - 40.0 ppt
> 40.0 ppt.
For this analysis we used a computer program originally developed to
estimate population exposures to benzene (Mara and Lee, 1977). The dis-
tance (D.) at which the specified concentration (C.) was found was calcu-
lated from Equation (3.6\. Using the following equation, we then estimated
the population that resided within a circle of radius D..
P. = d T D^ (3.7)
i i
where d is the city or state population density, and P. is the population
exposed to concentration C. or greater.
The five main assumptions included in this analysis are:
The ethylene dibromide source is in the center of the city
(if the city has a population greater than 25,000).
The maximum radius considered is 20 km.
When a city has more than one plant, it is assumed that
these plants are co-located and their corresponding
emission rates are summed.
The population density is uniform over the exposed area.
If the city has a population of less than 25,000, state
density is used.
To accommodate those assumptions, the following steps were included in the
computer program. The radius of each city was determined by Equation (3.8)
1/2 (3.8)
where D is the estimated radius of the citv; P is the population of the
c - c
citv (1970 Bureau of Census data); and d is the average city density
c
(1970 Bureau of Census data available for cities of population greater
than 25,000) .
17
-------
500
100
DISTANCE FROM SOURCE-km
Band on a production rate of 14 x 10a kg/yr (Dow facility)
Souroa: SRI utimatai after Youngblood (1977b)
FIGURE 111-2. PROJECTED DISPERSION CURVE* FOR MANUFACTURING
AND FORMULATING FACILITIES
18
-------
When D calculated from Equation (3.6) is greater than D , or when
1 c
no city density is available, Equation (3.9) is substituted for Equation
(3.7) to calculate the exposed population on the basis of state density.
pi = PC + S - (D;; - D*) (3.9)
where d is average state population density; D. is the distance at which
s i
concentration C. is found; D is the radius of the city calculated in
ic J
Equation (3.8); and P. is the population exposed to concentration C. or
greater. P and D equal 0 when no city density is available.
The cumulative population totals that resulted were then automatically
subtracted, so that the total population within each range of concentra-
tions was printed out. For example, for the range 1.0 to 5.0 ppt, the
program subtracted P_ (a smaller number) from P (a larger number).
In other words, P, is the population exposed to concentrations of 1.0 ppt
or greater. P, is the total population exposed to concentrations of 5.0
ppt or.greater. By subtracting the two values, the total population exposed
to concentrations between 1.0 and 5.0 ppt is determined.
Equation (3.6) requires an estimate of the production or use of
ethylene dibromide. Data on capacity were used for ethylene dibromide
manufacturing facilities. For tetraethyl lead manufacturers, we multiplied
the TEL capacity by 18% (the percentage of ethylene dibromide in TEL) to
determine the ethylene dibromide capacity (see Table III-l). Because the
same manufacturers produce soil fumigant and vinyl bromide, it was not
necessary to make separate estimates for those processes.
The statistics on population were obtained from density data derived
from the 1970 census (U.S. Department of Commerce, Bureau of the Census,
1972 County and City Data Book). When the population density for a city
was unavailable, we used the average statewide population density. Although
population density in the vicinity of chemical manufacturing plants can
vary widely; this method provides a reasonable overall estimate of the
exposed population.
19
-------
C. Exposures
The estimated population exposed to atmospheric ethylene dibromide
in the vicinity of chemical manufacturing and formulating facilities is
shown in Table III-3. More than one million people are estimated to be
exposed to annual average ethylene dibromide concentrations of more than
1.0 ppt, with 30% exposed to annual average levels greater than 10 ppt.
The population exposed to the lowest levels may be underestimated because
we did not consider concentrations at distances greater than 20 km. At
each facility, ethylene dibromide concentrations were estimated to be
greater than 1.0 ppt at that distance.
20
-------
Table III-3
ESTIMATED POPULATION EXPOSURES FROM MANUFACTURING A.TO FORMULATING FACILITIES
Location
Arkansas
El Dorado
Great Lakes Chem-
cal Corp.
Magnolia
Dow Chemical Co.
Ethyl Corp.
(Bromet Co.)
Total
California
Antioch
E. I. da Pont de
Xemours & Co. Inc.
Louisiana
Baton Rouge
Ethyl Corp.
City State
Density Density
People/km^ People/tan2
617
1,481
1,604
15
15
51
32
City
Population _^
103 1.0-5.05.1-10.010.1-20.0267f-40.0
Population Exposed to EDB (pot) '
>40.0
25 10,000 3,000 1,000 -10,000 10,000
0 8-.000 7,000 3,000 2,000
28 60,000 5,000 2,000 10,000 10,000
166 30,000 70,000 60,000 20,000 20,000
Total
Exposed JF
PopulationT
34.000
20,000
37,000
200,000
Midland
Dow Chemical Co.
Xev Jersey
552
Deepvater
E. I. du Pont de
Neaours & Co. Inc.
Texas
Beauaont
Houston Chemical Co.
(PPG Industries)
Freeport
Nalco Chemical Co
Pasadena
Ethyl Corp.
Total*
641
985
61
376
18
18
18
35 40,000 20,000 10,000 20,000 20,000
400,000 40,000
116 20,000* 70,000
3,000* 300
89 20.000* 30,000
580,000 250,000
10,000
6,000
10,000
40
20,000
99,000
4,000
8,000
30
10.000
34,000
110,000
460,000
140,000
3,500
120,000
1,200,000
*A maximum radius of 20 km was considered in estimating exposed population. Consequently, these estimates may be low because
ambient EDB concentrations were above 1.0 ppt at a distance of 20 km.
TTo convert to ug/m3, divide by 130; to convert to 8-hour worst-case, multiply by 10; rounded to one significant figure.
1 Rounded to two significant figures.
Source: SRI estimates.
21
-------
IV GASOLINE SERVICE STATIONS
A. Sources
Leaded gasoline contains ethylene dibromide as a lead scavenger.
The amount of ethylene dibromide added depends upon the lead content of
the particular gasoline mix. The 1975 model year was the first in which
catalytic converters were required. In that year, 100% of Ford and GM
cars and 96% of Chrysler cars required unleaded gasoline. In 1976,
approximately 20% of the gasoline sold in the United States was unleaded
(Ethyl Corporation, 1976).
The antiknock "motor mix" added to gasoline is a combination of
ethylene dibromide, ethylene dichloride, lead, and other alkyl groups.
Enough ethylene dichloride is used to supply two atoms of chlorine and
enough ethylene dibromide to supply one atom of bromine for each atom of
lead. Motor gasoline antiknock mixes typically contain 18.8% ethylene
dichloride by weight and 17.9% ethylene dibromide by weight as scavengers.
Whereas the average lead content in all gasoline is 1.5 g/gal,
leaded gasoline contains approximately 2.5 g/gal. According to EPA's
phase-down schedule for lead, the average lead content for all gasoline
will be 0.5 g/gal by 1 October 1979 (Stolpman, personal communication,
1977). Because of the low lead content in unleaded gasoline (approximately
0.01 g/gal), ethylene dibromide is not required and is not added. There-
fore, our analysis of population exposures related to gasoline service
stations considers only leaded gasoline. The concentration of ethylene
dibromide in leaded gasoline is approximately 0.05 percent by liquid volume
(Mitre, 1976).
B. Methodology and Exposures
1. Self-Service Operations
Service stations are characterized by their services and business
operations; full-service stations, split-island stations, self-service
23
-------
stations, and convenience store operations. In full-service stations,
attendants offer all services, including gasoline pumping and other
mechanical check-ups. If fuel is obtained at any of the last three classes
of stations, the customers may fill up their tanks themselves. In split-
island stations, both self-service and full-service are offered. At the
two remaining types of stations, only self-service is available. While
pumping gasoline, an individual is exposed to ethylene dibromide released
A
as vapor from the gasoline tank. Although occupants in the car at both
self-service and full-service operations are exposed to some degree to
ethylene dibromide, the highest exposures are to the person pumping the
gas. Because it is difficult to estimate level and length of exposure
for occupants, only those who pump gasoline from self-service pumps are
considered. (It is not within the scope of this report to evaluate occu-
pational exposures.)
Self-service sale of gasoline is a relatively new marketing
method pioneered by independent operators on the West Coast and in the
southern United States. Today, it accounts for 30% of gasoline sold.
The national market share of the major gasoline producers has decreased
recently as independents and others specializing in high-volume, low-margin
sales capture a larger percentage. Of the approximately 184,000 conven-
tional service stations and tie-in gasoline operations in the United States,
service stations with some self-service operations account for 39%
(Arthur D. Little, 1977). Table IV-1 indicates the types of service
stations offering self-service gasoline.
«
Vapor recovery systems can reduce exposure levels significantly, if
properly working and operated. Such systems are required for service
stations in parts of California.
24
-------
Table IV-1
SELF-SERVICE OPERATIONS
Percent of
Outlets Offering Self-Service_ ' U.S. Total
Total self-service 9%
Split island with self-service 26%
Convenience stores 4%
Total Outlets with Self-Service 39%
Source: Arthur D. Little (1977).
A recent Arthur D. Little report (1977) revealed that 71,300
outlets offer self-service gasoline. Gasoline sold at service stations
9
for the year ending May 30, 1977, equals approximately 87.4 x 10 gal in
9
the United States. Of that amount, 27.0 x 10 gal (31%) is estimated to
have been dispensed at self-service pumps. The market share of self-
service stations was surveyed for four metropolitan Air Quality Control
Regions (AQCR): Boston, Dallas, Denver, and Los Angeles. The market
share held by self-service operations varied from 9% in Boston to 45% in
Denver (see Table IV-2). Another study by Applied Urbanetics, Inc. (1976)
surveyed Baltimore and Madison, Wisconsin. The results of that study are
shown in Table IV-3. It appears that self-service operations account for
about 40% of the market in urban areas.
To estimate the people exposed to ethylene dibromide from this
source, several assumptions were necessary. The gasoline pumped through
9
self-service outlets is estimated at 27.0 x 10 gal. The annual average
fuel consumption per vehicle is 736 gal (U.S. Department of Transportation,
1974). If it is assumed that on the average, a person who primarily uses
self-service gasoline makes one trip per week to the gasoline station, an
average fill-up amount of 14 gal is determined by dividing 736 gal/vehicle/yr
by 52 wk/yr. By dividing the average fill-up into the self-service gallons
9
pumped, we estimate trips per year to self-service operations at 1.9 x 10 .
When this number is divided by 52 trips per person per year, the people
exposed to pumping self-service gasoline is estimated at 37 x 10 . We
25
-------
Table IV-2
GASOLINE MARKET SHARE OF SELF-SERVICE STATIONS
IN FOUR AQCRs, SPRING 1977
Type of Operation
Boston AQCR
Full-service
Self-service (total)
Split island
Self-service
Convenience stores
Dallas AQCR
Full-service
Self-service (total)
Split island
Self-service
Convenience stores
Denver AQCR
Full-service
Self-service (total)
Split island
Self-service
Convenience stores
Los Angeles AQCR
Full-service
Self-service (total)
Split island
Self-service
Convenience stores
Numb er of
Outlets
2,253
100
8£
92
621
656
310*
226
120
2,518
4,780
Sales Market
Volume Sharing
(106 gal/yr) Percent
1,045.1
108.6
292.1
235.7
2,472.6
2,154.8
91%
9
2,094
1,124
480a
444
200
924.6
593.8
61
39
55
45
53
47
1,022
126
Split-island operations offering full service and self-serve
islands.
Of these 445 are split island operations that offer full service
and mini-serve (attendant-operated) islands.
Source: Arthur D. Little (1977).
26
-------
Table IV-3
GASOLINE MARKET SHARE OF SELF-SERVICE
STATIONS IN TWO METROPOLITAN AREAS, 1976
Type of Operation
Baltimore SMSA
Sales Volume
(10 gal/yr) Market Sharing Percent
Full-service
Self-service (total)
Split island
Self-service
90.5
25.5
65.0
55%
45%
Madison SMSA
Full-service
Self-service (total)
Split island
Self-service
56.0'
77.0
17.0
60.0
42%
58%
Includes the sales from mini-serve (attendant-operated)
stations and 50% of the sales from split islands.
Source: Applied Urbanetics, Inc. (1976).
27
-------
can further assume that only 80% of these people are pumping leaded gasoline
containing ethylene dibromide. Therefore, the number of people exposed from
this source is estimated to be 30 x 10 . This estimate of the population
exposed assumes that the individuals using self-service gasoline never
obtain gasoline at full-service stations.
Attendants at gasoline service stations were monitored by NIOSH.
No samples were above the 0.03 mg analytical level of detection. There-
fore, based on the length of sampling periods, concentrations were assumed
to be below 30 ppb (Hartle, 1977).
A rough estimate of ethylene dibromide exposures was made by
extrapolating the results of benzene monitoring by Battelle (1977). In
that study, three samples of ambient air were taken in the breathing zone
of persons filling their tanks at self-service gasoline stations. The
results, shown in Table IV-4, indicate a wide range in the benzene concen-
trations of the emissions. The variations seem to be related to the sub-
ject's position in relation to the tank opening and the wind direction.
i
Because all measurements were taken on the same day and at approximately
the same time, ambient temperature did not cause the variation. Basically,
if the subject was downwind of the tank opening, higher levels were
recorded. The corresponding ethylene dibromide exposures were estimated
based on these data and are presented in Table IV-5.
Table IV-4
SAMPLING DATA FROM SELF-SERVICE GASOLINE PUMPING
Customer
1
2
3
Sampling Rate
(mL/min)
31
31
31
Nozzle
Time (min)
2.5
1.1
1.6
Gallons
Pumped
14
8
9
Sample
Volume
(L)
78
34
50
Benzene
, 3
Ug/m
115
324
1740
Level
PPb
43
121
647
Source: Battelle (1977).
28
-------
Table IV-5
ESTIMATES OF ETHYLENE DIBROMIDE EXPOSURES
FROM SELF-SERVICE GASOLINE PUMPING
Estimated EDB Level
Customer
1
2
3
Time (mln)
2.5
1.1
1.6
Pumped
14
8
9
yg/m3
0.345
0.972
- 5.220
PPt
45
126
679
Average Nozzle Time =1.7 min
Time Weighted Average Exposure = 260 ppt
Source: SRI estimates based on Battelle monitoring data (1977).
The conversion is based on differences in vapor pressure
and concentration between benzene and ethylene dibromide
in gasoline (see Chapter II).
29
-------
The estimated exposure levels are based on the information con-
tained in Table IV-5. It is recognized that those data are limited and
highly variable. However, they do allow a reasonable estimate of expected
exposure levels from self-service gasoline pumping. In states where vapor
recovery systems are used, the estimated exposure level may be much lower.
It can be estimated that approximately 30 x 10 persons use self-service
stations. While filling their tanks once a week, they are exposed to an
estimated ethylene dibromide level of 260 ppt for 1.7 minutes. Their
annual exposure is estimated at 1.5 hr. (Table IV-7 summarizes this infor-
mation. )
2. Vicinity of Service Stations
People residing in the vicinity of service stations are exposed
to ethylene dibromide from gasoline evaporation. Ethylene dibromide
emissions result from gasoline pumping by attendants and customers, and
from gasoline loading by distribution trucks. The amount of ethylene
dibromide emitted depends on the ambient temperature, vapor recovery con-
trols, the ethylene dibromide content in gasoline, and the volume of leaded
gasoline pumped. Approximately 80% of the gasoline currently sold is
leaded. With approximately 184,000 service stations in the United
States, it is expected that many people are exposed to ethylene dibromide
from those sources. Because density of service stations in urban areas
is high and is expected to correlate well with urban population density,
only urban areas are considered in this analysis.
An emission factor of 0.00157 g/g/lead/gal from refueling losses
has been estimated based on testing at EPA's Mobile Source Air Pollution
Control Laboratory in Ann Arbor, Michigan (Kittredge, 1977). The factor
considers spilling, vapor displacement, entrained liquid gasoline losses
and volume of gasoline pumped. Assuming an average lead content in gaso-
line of 2.5 g/gal, the estimated emission factor for ethylene dibromide
is 0.00039 g/gal.
The number of service stations in urban areas can be estimated
based on service station density and total U.S. population in urban areas.
Service station density in urban areas can be extrapolated from the data
presented in Table IV-6. The service station density shown for four
30
-------
Table IV-6
SERVICE STATION DENSITY IN FOUR METROPOLITAN AQCRs
AQCR
Boston
Dallas
Denver
Los Angeles
*
Number of
Service Stations (1977)
2,353
3,218
1,277
7,298
A*
AQCR
Population
(1975)
4,039,800
2,970,900
1,389,000
14,072,400
Service Station
Density
(number/ 10 00
population)
0.6
1.1
0.9
0.5
Source:
*
ADL
U.S. Department of Commerce, Bureau of Economic Analysis, 1973.
SRI estimates.
31
-------
metropolitan AQCRs is somewhat variable, with no apparent regional pattern
evident. Based on these data, an average of 0.7 service station per 1000
population was estimated. This number can be applied generally to urban
areas throughout the United States. Urbanized areas provide the best
population base. The 1970 population residing in urbanized areas was
118,447,000 (Bureau of the Census, 1975). Thus, service stations in urbanized
areas are estimated at 83,000, or 45% of all stations.
An emission rate can be estimated as follows:
(1) 70.0 x 109 gallons of leaded gasoline are sold annually by
service stations.
(2) The average number of gallons pumped per service station^
is 3.8 x 105 gal. (There are approximately 184,000 service
stations in the United States.)
(3) Assume that all service stations have uniform pumping
volumes and that three service stations are co-located
in urban areas. Although there is great variability in
both the pumping volume and the number of stations
located in the same area, we believe that our assumptions
provide a reasonable estimate considering the limited
data available.
(4) The ethylene dibromide emission rate for three co-located
service stations is:
(Number of Stations) (Volume of Gasoline Pumped) (Emission Factor)
= Emission Rate;
- o
that is, (3) (3.80 x 10 gal/yr) (0.00039 g/gal) = 4.45 x 10 g/yr
= 1.41 x 10-5 g/s.
Monitoring data for ethylene dibromide were collected in the
vicinity of service stations in three cities (Phoenix, Los Angeles, and
Camden, New Jersey) by Midwest Research Institute (1976). There was no
information given on the amount of leaded gasoline pumped. The ambient
ethylene dibromide concentrations ranged from 10 to 60 ppt with sampling
periods of from 12 to 18 hours. Los Angeles showed the lowest concentra-
tions (averaging about 14 ppt), indicating that the vapor recovery system
on gasoline pumps effectively reduces ethylene dibromide releases to the
atmosphere.
Defined by the Bureau of Census as the central city or cities and surrounding
closely settled territories. All sparsely settled areas in large incorporated
cities are excluded by this definition. Densely populated suburban areas,
however, are included (U.S. Department of Commerce, Bureau of the Census,
1972 County and City Data Book).
32
-------
The most comprehensive data were collected in the vicinity of a
heavily used intersection in Phoenix and were used as the basis for estimat-
ing a dispersion curve. Seven service stations were located in the sampling
area. Measured ethylene dibromide concentrations ranged from 25 to 65 ppt.
Figure IV-1 shows the variations in concentration within the sampling area.
The samples were collected over an 18-hour period. The mean value within
the 500 m radius was 48 ppt with a standard deviation of 10 ppt. The esti-
mated concentration at 1 km is 30 ppt if no service stations are located
outside the perimeter. The traffic passing through the intersection was
reported to be 38,000 vehicles per day.
Assuming that 30 ppt at 1 km approximates the concentration
representative of a 24-hour sampling period, an 8-hour worst-case concen-
tration can be estimated at 40 ppt. Because that value is related to seven
service stations, the corresponding value for three service stations is
estimated as 17 ppt. The annual average exposure can then be estimated
as 2.0 ppt (0.013 yg/m ) at 1 km.
The dispersion modeling curve applied to exposures in the vicinity
of service stations was extrapolated from rough dispersion modeling for
an area source conducted by Youngblood of EPA (1977c) for benzene. The
shape of the curve remains approximately the same for ethylene dibromide,
but its relative position changes based on the difference in emission rates.
2
The dispersion curve for an area source of 0.25 km (500 m on a side) was
considered to be representative of three co-located service stations. Mara
and Lee (1977, p. 33) developed an equation by regression analysis to
characterize this curve for benzene. By incorporating the emission rates
for ethylene dibromide into that equation, a new equation is formulated,
as follows:
C = (0.013) D~°'91 (4.1)
where C is the concentration at distance D, and 0.013 is the product of
the emission rate and the system constant. The results of the calcula-
tions are shown in Figure IV-2. Note that concentrations can be expected
to drop below 1 ppt just beyond 1 km from the source area.
33
-------
RETAIL
GASOLINE STATIONS.
. & I « I IWI^t«9^
\j \L
Cactus Road
tooo
2) Sampling Station
(35) Average concentration in pot
Source: Midwest Ronarch Institute, 1976
FIGURE IV-1. AVERAGE CONCENTRATION OF ETHYLENE DIBROMIDE IN AIR
AT 18 SAMPLING STATIONS NEAR A RETAIL GASOLINE SITE
IN PHOENIX, ARIZONA
34
-------
100
90
80
70
60
50
40
P)
a
o
7
z
o
H
oc
O
O 30
ui
Q
O
DC
m
5
UJ
01
20
10
0.1
0.2
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
DISTANCE FROM SOURCE - km
2.0
Source: SRI estimates after Youngblood (1977c)
FIGURE IV-2. PROJECTED DISPERSION CURVE FOR ANNUAL AVERAGE CONCENTRATIONS IN THE
VICINITY OF SERVICE STATIONS
-------
The population exposed to annual average concentrations between
5.1 and 10.0 ppt is estimated as follows:
9 it 2
Exposed Population = IT (0.3 km) (27,633) (1318 people/km )
= 10,000,000
The population exposed to annual average concentrations between
1.0 and 5.0 ppt is estimated as follows:
Exposed Population = TT [ (.1 km)2 - (0.3 km)2] (27,633)* (1318 people/km2)
= 100rOOO,000
The summary results are presented in Table IV-7. It is recognized
that these estimates are only rough approximations, based on assumptions of
uniform distribution of three co-located service stations in urbanized areas,
uniform pumping volume, and average population density- In reality, more
service stations are located in commercial areas than in residential areas,
and pumping volumes vary substantially. In addition, although it is likely
that several service stations are located in the same general area, the
average number is not known. If these areas are considered to be commercial,
they may have either a higher than average population density within 1 km
(because of a high percentage of apartments nearby), or one much lower than
average (because of a high percentage of businesses and few residences of
any kind). People residing near areas with more than three co-located
service stations may be exposed to higher annual average benzene concentra-
tions than those estimated. It is likely from this analysis that population
exposed is overestimated, whereas the exposure levels themselves may be under-
estimated. Further study is warranted to determine a more accurate estimate
of exposure levels based on pumping volumes, co-location of service stations,
their distribution within an urban area, and emission rates.
*
Number of locations of service stations within urbanized areas
assuming three at each location.
36
-------
Table IV-7
SUMMARY OF POPULATION EXPOSED TO ETHYLENE DIBROMIDE
FROM GASOLINE SERVICE STATIONS
EXPOSURE TYPE
SELF-SERVICE
PUMPING
RESIDING IN
THE VICINITY
EXPOSURE
TIME
1.7 MIN.
24 HR.
ANNUAL
EXPOSURE
1.5 HR.
ANNUAL
AVERAGE"
POPULATION EXPOSED TO EDB CONCENTRATIONS (pptl*
1.0-6.0
-
100,000,000
5.1-10.0
-
10,000,000
260.0
30,000,000
-
TOTAL
30,000,000
110,000,000
To convert to f/g/m3, divide concentrations by 130.
"To convert annual average exposures to 8-hour worst case, multiply concentrations by 10.
Source: SRI estimates
-------
V PETROLEUM REFINERIES
A. Sources
Petroleum refineries are sources of atmospheric emissions of ethylene
dibromide. Ethylene dibromide is a major ingredient in the antiknock
"motor mix" blended into leaded gasoline at each refinery. Motor mix is
manufactured at five locations (see Chapter III) and shipped in tank cars
or trucks to petroleum refineries. It is then transferred to storage sites
and used as needed. The average content of ethylene dibromide in antiknock
mixes is 17.9% by weight.
A limited sampling study in the vicinity of two petroleum refineries
was conducted by Midwest Research Institute (1976). The results are shown
in Table V-l. Concentrations of ethylene dibromide ranged from 6 to 16 ppt
within 3 km of the refinery at Ponca City, Oklahoma, and from 10 to 26 ppt
within 2 km of the site 'at Paulsboro, New Jersey. MRI concluded that the
source of all ethylene dibromide emissions was the gasoline bulk loading
area. Although samples were collected in the vicinity of the tetraethyl
lead storage, the mixing facility, the leaded gasoline storage area, and
the gasoline pumping station, none of these samples suggested that those
areas were the source of substantial emissions.
In light of the monitoring data, we decided to limit our analysis
of petroleum refineries to an estimate of exposures related to bulk load-
ing of leaded gasoline. The next section describes the methodology and
the estimated population exposed to this source.
B. Methodology
The general methodology described in Chapter III was used as the
basis for determining exposure levels from petroleum refineries. The
dispersion of ethylene dibromide was approximated based on the rough
dispersion model developed by Youngblood of EPA (1977b) for benzene.
39
-------
Table V-l
MONITORING DATA IN THE VICINITY
OF PETROLEUM REFINERIES
Refinery
Conoco
Mobil
Location
Ponca City, OK
Paulsboro, NJ
Average
Sampling
Time (hr)
18
18
Average EDB
Concentration
(ppt)*
8.8
15.5
Standard
Deviation
2.4
5.0
* 3
To convert to yg/m , divide by 130.
Source: Midwest Research Institute, 1976.
40
-------
Mara and Lee (1977) present a complete discussion of the model and its
application to benzene emissions. A projected dispersion curve was developed
for ethylene dibromide and applied to the largest refineries by computer
program to estimate the exposed population.
To estimate the amount of leaded gasoline bulk loaded at each refinery,
a number of assumptions were necessary:
(1) Gasoline production can be estimated at approximately
45% of crude' capacity (Moore, personal communication,
1977) .
(2) Approximately 75% of gasoline is transported by pipeline
(Bureau of Mines, 1977); thus, 25% is bulk loaded.
(3) Approximately 80% of gasoline is leaded (Ethyl Corporation,
1976) .
Appendix A presents the results of these calculations.
An emission factor was estimated for bulk loading of leaded gasoline
at refineries based on earlier work conducted by PEDCo (1977) for benzene.
They had estimated an emission factor for bulk loading of gasoline. That
factor was extrapolated to approximate ethylene dibromide as follows:
[Benzene Emission\/ Vapor Pressure EDS \ /Cone. EDB in Gasoline _ \ =
V Factor / \Vapor Pressure Benzene/ \ Cone. Benzene in Gasoline/
EDB Emission Factor;
that is,
(1.1 x 10~4 kg/m3) () () = 3.3 x 10~7 kg/m3.
If more than one refinery was located in a particular city, we assumed
that the refineries were co-located, and we summed their emission rates.
Although several cities have three or more refineries, it is also true that
few people generally live near such complexes. Thus, with this method, the
exposed population is minimized, whereas the exposure level is maximized
for a particular city.
ft O
Locations having refineries with crude capacities exceeding 5 x 10 m /yr
were evaluated individually. The remainder was evaluated on a statewide
basis. Appendix A lists the location of and the estimated gasoline produc-
tion at each refinery.
41
-------
Emission rates were estimated for each location based on the estimated
volume of gasoline bulk-loaded and an 8-hour-a-day, 7-day week. Appendix A
lists the estimated emission rates for each location or state.
The dispersion curve of ethylene dibromide from petroleum refineries
was approximated from the curve representing ground-level point source
emissions (curve A) developed by Youngblood for benzene (1977b). As des-
cribed in Chapter III, the shape of the curve remains approximately the
same for ethylene dibromide, but its relative position changes based on
the difference in emission rates. Mara and Lee (1977, p. 24) developed an
equation by regression analysis to characterize the curve for benzene. By
incorporating the emission rates for ethylene dibromide and by adjusting
to the available monitoring data, a new equation can be formulated.
Monitoring data from the Ponca City site were used because the refinery
there represented an isolated location with few other ethylene dibromide
sources. The estimated emission rate for the refinery (see Table A-l in
Appendix A) is 2.07 x 10 g/s. The average measured ethylene dibromide
3
concentration at 2.5 km is approximately 0.08 yg/m (10 ppt) . Those data
are then substituted into the general equation to determine the system
constant (A) as follows:
C = A EAD~1>48 (5.1)
where C is 8-hour worst-case concentration at distance D, and E is the
A.
emission rate for the particular refinery.
(0.08) = A (2.07 x 10~5) (2.5)'1'48 (5.2)
A = 1.5 x 104 (5.3)
A is then extrapolated to annual average conditions by multiplying
by 0.1. Equation (5.1) is then rearranged as follows:
9 E .6757
D = (1.40 x 10Z) (-^) (5.4)
Ci
where C is the specified concentration (i.e., 1.0, 5.0, 10.0, and so on;
3
input data, however, are in yg/m ); and D is the distance at which the
specified concentration is found. The projected dispersion curve for
petroleum refineries
shown in Figure V-l.
petroleum refineries based on an emission factor of 10 x 10 g/s is
42
-------
n
I
O
<
X
z
m
O
O
O
S
I
O
oc
CD
Q
Z
Ul
X
111
10
-2
1 I I
10
DISTANCE FROM SOURCE-km
100
Bated on an emiwion rate of 10 x 10"8 g/s
Source: SRI estimates after Youngblood (1977b)
FIGURE V-1. PROJECTED DISPERSION CURVE FOR PETROLEUM REFINERIES
-------
The computer program developed by SRI (Mara and Lee, 1977) was used
to calculate the exposed population within ranges of concentrations at
each location from Equation (5.4). A complete discussion of the program
is found in Chapter III. Because the dispersion modeling results are
unverified at distances greater than 20 km from the source location, the
computer program automatically cut off calculations when a distance of
20 km was attained and calculated the concentration (C.) at 20 km. Distances
within 500 m of the point source were assumed to be within the plant perim-
eter and were not included in the estimate of exposed population.
C. Exposures
The population exposed to atmospheric ethylene dibromide from petro-
leum refineries by plant location and by state is shown in Appendix A.
Table V-2 presents the summary results. More than 99% of the exposed
population was found at locations having a total refining capacity exceed-
r Q
ing 5 x 10 m /yr. Of the more than 2 million people exposed to annual
average ethylene dibromide concentrations greater than 1.0 ppt, 90% are
estimated to be exposed to concentrations from 1.0 to 5.0 ppt. Approxi-
mately 0.1% are estimated to be exposed to ethylene dibromide levels
exceeding 40.0 ppt (400.0 ppt 8-hour worst case).
44
-------
Table V-2
ESTIMATED POPULATION EXPOSED TO ETHYLENE
DIBROMLDE FROM PETROLEUM REFINERIES
^ +
Annual Average Ethylene Dibromide Concentrations (ppt)
1.0-5.0 5.1-10.0 10.1-20.0 20.1-40.0 >40.0 Total
Exposed
Population 2,000,000 170,000 53,000 16,000 3,000 2,200,000
*
To convert to 8-hour worst case,
multiply concentrations by 10.
t 3
To convert to yg/m , divide by 130.
Source: SRI estimates.
45
-------
VI STORAGE AND DISTRIBUTION OF GASOLINE
A. Sources
Storage and distribution of gasoline represent potential sources of
atmospheric ethylene dibromide in the environment. Ethylene dibromide
can escape via two main emission pathways: (1) evaporation and spills
during loading and unloading of gasoline at bulk terminals and (2) spills
from collisions in transportation. Ethylene dibromide transfers normally
occur at manufacturing and formulating facilities and were considered in
Chapter III.
Gasoline transfers normally occur at petroleum refineries and at
numerous storage sites throughout the United States. Gasoline is usually
stored in closed containers in remote locations. Although loss from
storage tanks through evaporation has been observed, most of the ethylene
dibromide in the environment is believed to have been released during
the loading and unloading of the gasoline. Spills from collisions involv-
ing gasoline transfer vehicles account for negligible losses of ethylene
dibromide.
B. Methodology and Exposures
1. Storage
Storage facilities consist of closed storage vessels, including
pressure, fixed-roof, floating-roof, and conservation tanks. Ordinary
fixed-roof tanks store less volatile petroleum products, whereas floating-
roof tanks are most commonly used to store gasoline. Emissions of ethylene
dibromide from storage in a floating-roof tank occur primarily from stand-
ing and withdrawal (wetting) losses. Fixed-roof tanks have "breathing"
losses caused by expansion and contraction of the vapors because of diurnal
changes in atmospheric temperature. Because of the low volume of gasoline
stored in fixed-roof tanks, breathing losses are not a significant source
of ethylene dibromide.
47
-------
Standing emissions are caused by improper fit of the seal and
shoe to the vessel shell. Small losses also occur when vapor escapes be-
tween the flexible membrane seal and the roof. Withdrawal or wetting
losses are caused by evaporation from the tank walls as the roof descends
during emptying operations (PEDCo, 1977).
Emission factors of ethylene dibromide as a result of these
losses were-estimated by extrapolating from benzene emission factors developed
by PEDCo (1977, p. 4-65) as follows:
Benzene - A Estimated EDB
Storage Emission Factor (kg/m ) Emission Factor (kg/m )
Gasoline
Standing losses , 3.3 x 10
Withdrawal losses 2.6 x 10
Total: 5.9 x 10 5 1.77 x 10 7
PEDCo estimates (1977).
SRI estimates based on differences in vapor
pressure and concentration between benzene
and ethylene dibromide. See Chapter II for
a more detailed discussion.
Gasoline bulk storage terminals are generally near urban demand
centers, commonly in highly industrialized areas or on the city periphery
where population density is low.
Rough estimates of ambient ethylene dibromide concentrations for
the vicinity of storage sites can be based on the emission factors, assumed
storage volumes, and the results of the dispersion model discussed in
Chapter III. An average gasoline storage terminal is assumed to have the
following characteristics: average tank size, 8.7 x 103 m ; 30-day
retention time; 10 gasoline storage tanks of average size; facility size,
2
0.25 km . The emission rate is calculated as follows:
(emission factor) x (tank volume) x (number of tanks) = emission rate;
5.13 x 10~4 kg,
5.94 x 10~6 g/s.
that is, (1.77 x 10 kg/m3) (8.7 x 103 m3/30 days)(10) = 5.13 x 10~4 kg/day
48
-------
The ambient ethylene dibromide concentrations can be estimated
by extrapolating from the dispersion modeling calculations of Youngblood
(1977c) that assume uniform emissions througnout the terminal area. By
applying the estimated emission rate to the dispersion modeling results
presented in Table IV-4 (Mara and Lee, 1977) for the indicated terminal
2
area of 0.25 km , the following estimate can be made:
8-hour Worst-Case Exposure Levels at 300 m
(5.94 x 10~6 g/s 1 ,Qn/, ,3. -5 ,3
\ 100 g/s / yg = 5.35 x 10 ug/m
= 6.95 x 10~3 ppt.
Therefore, annual average and 8-hour worst-case concentrations at 300 m
of the site are well below the detection level of 1.0 ppt. From this
analysis it appears that the number of people exposed to ambient ethylene
dibromide concentrations above the detectable limit in the vicinity of
gasoline storage terminals is negligible.
2. Distribution Systems
The gasoline distribution system involving transport from the
petroleum refineries to the consumer may also be a source of atmospheric
ethylene dibromide. The U.S. gasoline distribution system is illustrated
in Figure VI-1. Bulk terminals represent intermediate stations set up to
serve near-source regional markets. Gasoline at bulk terminals is trans-
ferred directly from the refinery by ships, rail tank cars, barges, and
pipelines. Bulk plants, on the other hand, are designed for local markets
and their supplies are distributed by tank trucks. Service stations
that fuel privately owned motor vehicles are supplied by tank trucks from
either bulk terminals or bulk plants. Privately owned commercial operations,
such as those providing fuel for vehicles of a company fleet, are generally
supplied by tank trucks from an intermediate bulk installation.
Most of the emissions take place during transfers of the gasoline
to tanks and tank trucks. Such losses occur at a rate directly proportional
to the amount of gasoline passing through 'the particular location. Because
many tank trucks are filled at one bulk terminal or plant, ethylene dibromide
emissions from that procedure are potentially much greater. As empty tank
trucks are filled, hydrocarbons in the vapor space are displaced to the
49
-------
SHIP, RAIL. BARGE
^SERVICE STATIONS
REFINERY STORAGE
BULK TERMINALS
TANK TRUCKS
AUTOMOBILES, TRUCKS
PIPELINE
BULK PLANTS
TRUCKS
COMMERCIAL,
RURAL USERS
SOURCE: PEDCa. 1977
FIGURE VI -1. THE GASOLINE MARKETING DISTRIBUTION SYSTEM
IN THE UNITED STATES'
50
-------
atmosphere unless vapor collection devices have been provided. The
quantity of hydrocarbons contained in the displaced vapors depends upon
the vapor pressure, temperature, method of tank filling, and conditions
under which the truck was previously loaded. Figure VI-2 is a schematic
drawing of liquid and vapor flow through a typical bulk terminal.
Gasoline is loaded from storage tanks to transport trucks (tank
cars) by two basic methods: top loading and bottom loading (PEDCo, 1977).
Top loading .can be done by splash fill or submerged fill. The former
method involves free fall of gasoline droplets and thus promotes evapora-
tion and.possibly liquid entrainment of those droplets in the expelled
vapors. In subsurface or submerged filling, the gasoline is introduced
below the surface of the tank. Bottom loading of gasoline is comparable
to submerged top loading.
Vapor recovery systems are designed to reduce the overall hydro-
carbon emission losses (including ethylene dibromide) for both loading and
unloading. For bottom loading, the vapor recovery system may achieve 100%
efficiency (PEDCo, 1977). Although it is difficult to quantify, vapor
collection for top loading is generally not so efficient as that for bottom
loading. An overall 95% efficiency of vapor recovery and containment can
be assumed for both loading and unloading (PEDCo, 1977, p. 4-60).
Rough estimates of ambient ethylene dibromide concentrations
related to gasoline distribution can be based on emission factors, assumed
transfer volumes, and the dispersion modeling results discussed in Chapter
III. Emission factors related to the loading and unloading of gasoline
were estimated for benzene by PEDCo (1977, p. 4-65), and have been extrapo-
lated to estimate ethylene dibromide concentrations. The estimates follow:
Benzene Emission Estimated EDB Emission
Distribution Factor (kg/m3)* Factor (kg/m3)t
Loading 1.1 x 10~ 3.3 x 10~?
Unloading 1.1 x 10~5 3.3 x 10-8
PEDCo estimates (1977).
SRI estimates based on differences in vapor pressure
and concentration between benzene and ethylene dibromide.
See Chapter II for a more detailed discussion.
51
-------
PIPELINE GASOLINE
TO STORAGE
STORAGE TANK
LOADING VAPORS
TO~ RECOVERY UNIT
TERMINAL
TRANSPORT
GASOLINE TO
LOADING RACK
VENT GAS
VAPOR
RECOVERY
UNIT
RECOVERED i
GASOLINE
SOURCE: PEDCo. 1977
FIGURE VI-2. VAPOR AND LIQUID FLOW IN A TYPICAL BULK TERMINAL {Floating-Roof Tank)
-------
A gasoline bulk storage terminal of the same characteristics as
described in the previous section is assumed. In addition, continuous
loading and unloading operations are assumed over a five-day work week,
eight hours a day. The emission rates are calculated as follows:
Loading
(Emission factor) x (Average Tank Size) x (// of Tanks) = Emission Rate
that is,
(3.3 x 10~7 kg/m3) (8.7 x 103 m3/30 days) (10) = 9.57 x 10~4 kg/day
= 3.32 x 10 5 g/s.
Unloading
(3.3 x 10 8 kg/m3) (8.7 x 103 m3/30 days) (10) = 9.57 x 10 5 kg/day
= 3.32 x 10~6 g/s.
Total emission rate = 3.65 x 10 g/s.
The ambient ethylene dibromide concentration can be estimated
from the dispersion modeling calculation of Youngblood (1977b) by assuming
ground-level point source emissions (Curve A). When the estimated emission
rate is applied to the results presented in Table III-4 (Mara and Lee, 1977),
the following estimate can be made:
8-Hour Worst-Case Exposure Levels at 300 m
>5 x 10"
100 g/s
3.65 x 10 5 g/s\ (14,000 yg/m3) = 5.11 x 10 3 pg/m3
/ = 0.7 ppt
Approximate annual average con- = 0.07 ppt.
centration
From this analysis, it appears that concentrations at 300 m
of the loading and unloading area are generally below 1.0 ppt. Concentra-
tions may be higher in some cases if a large volume of gasoline (larger
3 3
than the average value used in this analysis2.9 x 10 m /day loaded and
unloaded) is loaded and unloaded during one 8-hour period. Thus, although
occupational exposures may be high, exposures to the general public appear
to be minimal.
53
-------
VII URBAN EXPOSURES RELATED TO AUTOMOBILE EMISSIONS
A. Sources
Urban exposures to ethylene dibromide come from many sources, includ-
ing gasoline evaporation, gasoline service stations, losses through trans-
portation and storage of gasoline, and, in some cases, manufacturing and
formulating facilities. However, ethylene dibromide is not routinely
monitored in the ambient air, and few sampling data exist. Therefore, to
determine average urban exposures, it is necessary to restrict the analysis
to automobile emissions.
Although antiknock mixture for automobiles contains 17.9% by
weight ethylene dibromide, antiknock mixture for aviation fuel contains
35.7% by weight. It was beyond the scope of this study to evaluate air-
ports as a source of urban emissions. However, ambient ethylene dibromide
concentrations in the vicinity of airports servicing a high percentage of
piston-engine planes may be higher than ambient concentrations in the sur-
rounding area.
As previously discussed, the ethylene dibromide content in leaded
gasoline averages 0.05% and accounts for 80% of the gasoline sold. Tests
by EPA's Mobile Source Air Pollution Control Laboratory in Ann Arbor,
Michigan, have indicated that ethylene dibromide is destroyed in the com-
bustion process, producing various bromines in the exhaust gases (Kittredge,
personal communication, 1977). However, evaporation of ethylene dibromide
on the carburetor and from the fuel tank does occur. The testing results
for three categories of automobiles are shown in Table VII-1.
A few monitoring data have been collected in urban areas. An early
study conducted by Midwest Research Institute (MRI) (1975) measured
ethylene dibromide concentrations in four" locations in three cities
(Phoenix, Los Angeles, Seattle). As shown in Table VII-2, the concentra-
3
tions ranged from 8.3 to 13.0 ppt (0.69 to 0.11 ug/m ). In a second study,
55
-------
Table VII-1
AUTOMOTIVE ETHYLENE DIBROMIDE EMISSION
FACTORS (g/g lead/gal gasoline)
Vehicle Type Low High
Uncontrolled vehicle (pre-1972) 0.00144 0.00362
Pre-1978 controlled vehicle 0.00098 0.00250
Post-1978 controlled vehicle 0.00033 0.00085
Source: Kittredge, 1977
56
-------
Table VII-2
MONITORING DATA FOR ETHYLENE DIBROMIDE IN URBAN AREAS
Location
Phoenix
Los Angeles
Seattle
Phoenix
Los Angeles
A
Kansas City
Number
of Sites
1
1
2
10
9
1
Average
Sampling
Time (hr)
17
16
12
18
18
18
Average
EDB
Concentration
yg/m3
0.069
0.110
0.083
0.360
0.124
0.060
PPt
8
13
11
47
16
8
Reference
1
1
1
2
2
2
Suburban area
1. Midwest Research Institute (1975)
2. Midwest Research Institute (1976)
57
-------
MRI (1976) conducted much more extensive sampling near highly trafficked
sites in Phoenix and Los Angeles. While the two Los Angeles average con-
centrations were very close, the Phoenix measurements differed by a factor
of 5. No detailed meteorological information was available. An additional
sampling site in a suburban area of Kansas City provided results very close
to the Los Angeles measurements. Although no explanation was offered for
the Phoenix samples, MRI concluded that the effect of heavily trafficked
freeways on the ethylene dibromide levels in the two cities was not dis-
cernible. They further concluded that the ubiquitous nature of ethylene
dibromide is probably the result of widely dispersed sources of emission
in urban/industrial areas.
B. Methodology and Exposures
Only limited data are available concerning urban exposures from auto-
mobile emissions. Consequently, it is difficult to develop accurate
techniques to predict ethylene dibromide levels in urban areas. Uncer-
tainties include the ethylene dibromide content in gasoline, control tech-
nology, the deterioration of the control technology over time, and the
dispersion characteristics of ethylene dibromide under variable meteor-
ological conditions. A simplified model was employed to provide general
estimates of ambient concentrations and of exposed population.
The emission, factor is estimated from the information presented in
Table VII-1. The average emission factor for an uncontrolled vehicle is
0.00253 g/g lead/gal. Assuming 2.5 g lead per gallon of leaded gasoline,
the estimated emission factor is 0.0063 g/gal. This factor will provide
a slightly high estimate of ambient ethylene dibromide levels because it
assumes all automobiles using leaded gasoline have emissions comparable
to pre-1972 models. In fact,,the 1975 model year was the first in which
automobiles were required to run on unleaded gasoline, but controls to
reduce carburetor evaporation were introduced in the 1972 model year.
The Hanna-Gifford dispersion model (Gifford and Hanna, 1973) as
applied by Schewe (1977) for benzene is used for this analysis and
modified for ethylene dibromide. Mara and Lee (1977) contains a complete
discussion of this model and its application to benzene. Because ethylene
dibromide undergoes combustion, only evaporation is considered. The
58
-------
modified equation to estimate the emission rate for ethylene dibromide is
as follows :
_ fr. nn,-T / ..^ /annual travel miles per vehicle \ ,,, . , , _L,
Q evap = (0.0063 g/gal) - - d - I (# veh. regis.) ( - )
x * & ° \ average miles per gallon / 6 area
If 12,000 miles per year for each vehicle and 12 miles per gallon are
assumed (Department of Transportation, 1974) , the above equation becomes
^ ^o n -in"7 / \
Q evap = (2.0 x 10 g/s) x
Ven- Regis.
& -
3.1T63.
To calculate the annual average areawide ethylene dibromide concentra-
tion, the following equation can be used:
225 Q evap
XA ~ u
where u is wind speed (m/sec). The average annual wind speed, u, in the
area of study was obtained from Figure VII-1. Because wind speed (and thus
dispersion) increases in the afternoon, the morning values were used to
estimate higher concentrations. The number 225 is an empirical factor
derived from several studies that gave very good results for long-term
averages; it applies to light-duty vehicles such as passenger cars.
Source: EPA, 1971
FIGURE VII -1. ISOPLETHS (m/sec) OF MEAN ANNUAL WIND SPEED
THROUGH THE MORNING MIXING LAYER
59
-------
Because of the general unavailability of 1976 data for all urban
areas, 1973 data were used as much as possible in this estimate. Compari-
sons of 1973 with 1976 data indicated that the change was less than 3% and
had a negligible effect on the final results. The following data sources
were used:
1973 Standard Metropolitan Statistical Area (SMSA) and
county populationsU.S. Bureau of the Census, 1976,
Series P-25, No. 618.
1973 SMSA and county automobile registrationsU.S.
Department of Transportation, Federal Highway Administration,
1974, Table-MV-21.
Average annual wind speedEPA, 1972, Publication No. AP-101.
SMSA, county and city land areasBureau of the Census, 1972
County and City Data Book.
A detailed analysis was conducted for the six largest cities in the U.S.
(populations of more than 1 million). Table VII-3 presents the results.
Suburban areas are defined as those areas outside the central city but
within the SMSA. Because no VMT and registration data were available
at the city level, they were extrapolated either from SMSA data or county
data and were based on the fraction of the population residing in each
area. The results show that the annual average estimated ethylene dibro-
mide concentrations in city and suburban areas range from 0.5 to 2.2 ppt
and 0.1 to 1.0 ppt, respectively.
It is expected that people living in urban areas are exposed to
higher levels of ethylene dibromide from automobile emissions than those
living in rural areas. Consequently, our approach was designed to maximize
the urban population considered in the analysis. Although 43% of the total
urban population resides in central cities (as defined by the Bureau of the
Census), 83% of the total urban population resides in SMSAs. Thus, a greater
percentage of the urban population is captured by using SMSAs as study areas.
The six largest cities are in SMSAs with more than 2 million population.
60
-------
Table VII-3
ESTIMATES OF'AVERAGE ANNUAL
ETHYLENE DIBROMIDE CONCENTRATIONS
FOR CITIES WITH POPULATIONS EXCEEDING 1,000,000
City
Chicago
Detroit
Houston
Los Angeles
New York
Philadelphia
SMSA
Population
103
6,998.8
4,446.3
2,163.4
6,938.3
9,746.4
4,826.3
City
Population
10 3
3,173
1,387
1,320
2,747
7,647
1,862
City
Area
0.57
0.35
1.1
1.2
0.77
0.33
Automobile
Registration
1,324,171
675,065
701,766
1,490,483
1,707,891
944,660
Qt
10-10
g/s-m2
3.71
3.08
1.01
1.98
3.54
4.57
Wind
Speed
m/s
5
6
6
3
7
6
EDB Concentration
Central
10~3wg/m3
16
11
4
14
11
17
City
PP.t
2.1
1.4
0.5
1.8
1.4
2.2
Suburban
10~3pg/m3
6.6
7.7
0.74
3.0
1.8
1.0
PPt
0.9
1.0
0.1
0.4
0.2
0.1
Assume 80% of vehicles use leaded gasoline.
Source: SRI estimates based on Hanna-Gifford dispersion model as applied by Schewe (1977).
-------
To analyze the remaining SMSAs, the following population size categories
were employed (U.S. Bureau of the Census, 1976, Series P-25, No. 618):
SMSA Population Size Category Number of Areas
2,000,000 or more 15
1,000,000 - 2,000,000 20
500,000 - 1,000,000 37
250,000 - 500,000 63
fewer than 250,000 124
SMSA composite ethylene dibromide concentrations were estimated for
seven areas that represent four population size categories (see Table VII-4).
All calculations gave ethylene dibromide concentrations below 1 ppt.
The estimates of urban exposures from automobile emissions are approx-
imate estimates that are based on a simple dispersion model. In certain
locations and under certain meteorological conditions, ethylene dibromide
concentrations may be a factor of 10 higher than those listed. In addition,
central city areas (as shown in Table VII-3) may have consistently higher
levels than surrounding areas because of traffic density, frequency of inter-
sections, and street density. Because the model only includes automobile
emissions, areas with substantial commercial or bus transportation may have
higher levels than estimated. Also, the model is extremely sensitive to
area size, as Table VII-2 indicates. Thus, ethylene dibromide concentrations
in the composite SMSA provide the most reasonable estimate of the average
annual exposures for an urbanized area.
The total estimated urban population exposed to annual average ethylene
dibromide in concentrations greater than 1.0 ppt from automobile emissions
is shown in Table VII-5. The 1974 SMSA populations of Detroit and
Chicago were summed along with the central city population of Los Angeles,
New York, and Philadelphia to estimate the population exposed to average
annual ethylene dibromide concentrations of 1.0 to 5.0 ppt. The
results indicate that 24 million people, or 15% of the total SMSA
population, are exposed to average annual ethylene dibromide concentrations
greater than 1.0 ppt. The apparent differences between the monitoring data
can be explained in several ways. Our estimates concern annual average
62
-------
Table VII-4
ESTIMATES OF ANNUAL AVERAGE ETHYLENE
DIBROMIDE CONCENTRATIONS FOR SELECTED SMSAs
SMSA
SMSAs > 2, 000, 000
Pittsburg
San Francisco
SMSAs 1,000,000
Columbus
Milwaukee
SMSAs 500,000 -
Sacramento
Providence
Warwick
Pawtucket
SMSAs 250,000 -
Wichita
Harrisburg
Population
2,333,
3,135,
- 2^000
1,055,
1,423,
600
900
,000
900
200
Area
(109m2)
7
6
6
3
.8
.2
.2
.7
Automobile
Registration
2,358,600
688,
567,
642,
300
803
531
*
Qt
10-11
g/s-m2
4.8
1.
1.
2.
8
7
8
Wind
Speed
m/s
5
3
5
5
EDB
Concentration
10~3ug/m3 ppt
2.1 0.3
1.4 0.2
0.66 0.09
1.2 0.2
1,000,000
851,
854,
500,000
375,
425,
300
400
600
500
8
2
6
4
.7
.4
.2
.1
439,
869,
221,
198,
803
100
715
997
0.
5.
0.
0.
80
8
57
77
3
7
7
5
0.60 0.08
1.9 0.2
0.18 0.02
0.35 0.05
Assume 80% of vehicles use leaded gasoline.
Source: SRI estimates using Hanna-Gifford dispersion model as applied by Schewe (1977)
-------
ethylene dibromide exposures from gasoline evaporation from automobiles
within an entire urban area. The monitoring data, on the other hand,
were collected generally near highly trafficked areas with nearby gasoline
stations; such areas are treated as a separate source in this report.
Incomplete combustion is not considered in the model and may also contribute
to higher ambient levels. Uncertainties in the data used for the model
calculations and the limited nature of the monitoring data are additional
reasons for the differences. Given the widespread nature of ethylene
dibromide in the urban atmosphere, additional research is required to
improve these estimates of exposure.
' Table VII-5
URBAN POPULATION EXPOSURES RELATED TO AUTOMOTIVE EMISSIONS
Population Exposed to
Annual Average* EDB
Concentration of
Source 1.0 - 5.0
Automobile Emissions 24,000,000
*
To convert to 8-hour worst case, multiply by 10.
To convert to yg/m , divide concentrations by 130,
Source: SRI estimates.
64
-------
BIBLIOGRAPHY
Applied Urbanetics, Inc., "Market Share Study," FEA Contract No. CO-06-60435,
200 pp. (1976).
Arthur D. Little, Inc., "Self-Serve Market Shares in Four Metropolitan
Areas," memo to Richard J. Johnson, EPA, from E. Quakenbush and
P- E. Mawn, June (1977).
Battelle-Columbus Laboratories, letter to Richard J. Johnson (Office of
Air Quality Planning and Standards, EPA, Research Triangle Park),
from C. W. Townley, concerning "Results of Self-Service Exposure
Samples," May (1977).
Brown, S. L., F. Y. Chan, J. L. Jones et al., "Research Program on
Hazard Priority Ranking of Manufactured Chemicals, Phase II-Final
Report, Chemicals 1-20," SRI International, Menlo Park, California,
report to the National Science Foundation, April (1975).
Ethyl Corporation, "Yearly Report of Gasoline Sales by States," Houston,
Texas (1976).
Gifford, F. A. and S. R. Hanna, "Technical Note: Monitoring Urban Air
Pollution," in Atmospheric Environment, Pergammon Press, Vol. 7,
pp. 131-136 (1973).
Hartle, Richard, Industrial Hygiene Section, memo to Dr. Jerry Chandler,
Criteria Manager, National Institute for Occupational Safety and
Health concerning "Ethylene Dibromide Environmental Air Samples
Collected at Gasoline Service Stations," 4 October (1977).
Joiner, Ronald L., Manager Toxic Sciences Group, Center for Occupational
and Environmental Safety and Health, SRI International, personal
communication, October (1977).
Kirk, R. E., Encyclopedia of Chemical Technology, 2nd Edition, Vol. 3,
John Wiley and Sons, Inc., New York, p. 771 (1968).
Kittredge, George D., Senior Technical Advisor for Mobile Air Source Air
Pollution Control, Office of Air and Waste Management, U.S. Environ-
mental Protection Agency, memo to files concerning "Up-to-Date
Estimate of Automotive Emission Factors," 26 September (1977).
65
-------
Mara, S. J. and S. S. Lee, "Human Exposures to Atmospheric Benzene,"
SRI International, Menlo Park, California, prepared for U.S.
Environmental Protection Agency, Contract No. 68-01-4314, October
(1977).
Midwest Research Institute, "Sampling and Analysis of Selected Toxic
Substances, Task II-Ethylene Dibromide," EPA 560/6-75-001, U.S.
Environmental Protection Agency (1975) .
, "Sampling and Analysis of Selected Toxic Substances,
Task IV-Ethylene Dibromide," EPA 560/6-76-021, U.S. Environmental
Protection Agency.
Mitre Corporation, "Air Pollution Assessment of Ethylene Dibromide,"
prepared for the U.S. Environmental Protection Agency, Contract
No. 68-02-1495 (1976).
Moore, Michael A., Manager, Petroleum Refineries, Energy Center, SRI
International, personal communication, October (1977).
PEDCo Environmental, "Atmospheric Benzene Emissions," prepared for U.S.
Environmental Protection Agency, Research Triangle Park (1977).
Schewe, George J., Monitoring and Data Analysis Division, Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency,
memos to Richard J. Johnson, Strategies and Air Standards Division,
U.S. Environmental Protection Agency, concerning "Estimates of the
Impact of Benzene from Automotive Sources," of June 20, August 9,
August 12 (1977).
Stolpman, Paul, Office of Air and Waste Management, U.S. Environmental
Protection Agency, personal communication, October (1977).
U.S. Department of Commerce, Bureau of the Census, "Population Estimates
and Projections," Series P-25, No. 618, Washington, D.C. (1976).
, 1972 County and City Data Book-, Washington, D.C. (1973)
_, Statistical Abstract of the United States, Washington,
D.C. (1975).
U.S. Department of Commerce, Bureau of Economic Analysis, "Projections
of Economic Activity for Air Quality Control Regulations," NTIS
PB-259-870 (1973).
U.S. Department of Interior, Bureau of Mines, "Mineral Industries Surveys-
Crude Petroleum, Petroleum Products and Natural Gas Liquids,"
Washington, D.C., May (1976).
66
-------
U.S. Department of Transportation, Federal Highway Administration, "Annual
Miles of Automobile Travel," in Nationwide Personal Transportation
Study, Report No. 2, 32 p. (1972).
, "Highway Statistics," Washington, B.C. (1974).
, "Motor Vehicle Registrations by Standard Metropolitan
Statistical Areas," Table MV-21 (1974).
U.S. Environmental Protection Agency, "Mixing Heights, Wind Speeds, and
Potential for Urban Area Pollution Throughout the Contiguous United
States," in Publ. No. AP-101, Office of Air Programs, Research
Triangle Park (1972).
_, "Compilation of Air Pollution Emission Factors,"
2nd' Edition, Publ. No. AP-42, Research Triangle Park (1976).
Youngblood, Phillip L. , Monitoring and Data Analysis Division, U.S.
Environmental Protection Agency, concerning "Use of Dispersion
Calculations in Determining Population Exposures to Benzene From
Chemical Plants," September 20 (1977b).
, memo to Richard J. Johnson, Strategies and Air Standards
Division, U.S. Environmental Protection Agency, concerning "Population
Exposures to Benzene from Petroleum Refineries and Large Coking
Plants," September 21 (1977c).
67
-------
APPENDIX A
CAPACITIES AND EXPOSED POPULATION BY
PETROLEUM REFINERY AND STATE
-------
Table A-l
POPULATION' EXPOSURES FROM
PETROLEUM REFINERIES WITH CRUDE CAPACITIES
EXCEEDING 5.0 x 106m3/yr
Crude* Casolinet Emissiont
Capacity Bulk Loaded Rate
Population Exposed to EDB (ppt)^
Location
California
Bakersfield
Chevron USA Inc
Kern Co. Refinery Co.
Lion Oil Co. (TOSCO)
Mohawk Petroleum
Road Oil Sales
Sabre Refining Co.
Sunland Refining Co.
West Coast Oil Co.
Total
Benicia
Exxon Co.
Carson
Atlantic-Richfield
Fletcher Oil
Total
El Segundo
Chevron USA Inc.
Los Angeles
Union Oil Co. - Calif.
Martinez
Lion Oil Co. (TOSCO)
Shell Oil Co.
Total
Richmond
Shell Oil Co.
San Francisco
Union Oil Co. - Calif.
Santa Fe Springs
Gulf Oil Co.
Powerline Oil Co
Total
106m3 106ra3 10 5g/s 1.0-5.0 5.1-10.0 10.1-20.0 20.1-40.0 >40.0
1.51
0.92
2.21
1.28
0.09
0.20
0.81
0.87
7.89 0.71 2.24 21,131 1,647 645
5.12 0.46 1.45 578 45
10.16
1.11
11.27 1.01 3.18 52,301 4,076 1,598
23.51 2.12 6.68 4,554 355 139 55
6.27 0.56 1.76 34,468 2,686
7.31
5.80
13.11 1.18 3.72 2,065 161 63
21.20 1.91 6.02 69,854 5,751 2,254 883
6.44 0.58 1.83 244,090 19,024
2.99
2.56
5.55 0.50 1.58 649 51
71
-------
Table A-l (Continued)
Indiana
East Chicago
Energy Coop. Inc.
Whiting
Amoco Oil Co.
Crude* Casolinet Emissiont
Capacity Bulk Loaded Rate
Location
California (continued)
Tor ranee
Mobil Oil Corp.
Wilmington
Champ 1 in Petroleum Co.
Shell Oil Co.
Texaco Inc.
Total
Delaware
Delaware City
Getty Oil Co. Inc.
Georgia
Savannah
Amoco Oil Co.
Illinois
Joliet
Mobil Oil Corp.
Lemon t
Union Oil Co. - Calif.
Robinson
Marathon Oil Co.
Wood River
Amoco Oil Co.
Shell Oil Co.
Total
106m3 106m3 10"5g/s
7.17 0.65 2.05
1.78
5.22
4.35
11.35 1.02 3.21
8.13 0.73 2.30
8.71 0.78 2.46
10.45 0.94 2.96
8.76 0.79 2.49
11.32 1.02 3.21
5.51
16.43
21.94 1.97 6.21
7.31
21.19
0.66
1.91
2.08
6.02
741
6,230
Population Exposed to EPS (ppt)'
1.0-5.0 5.1-10.0 10.1-20.0 20.1-40.0 >40.0
45,977 3,583
1,692 132
2,367 185
58
52
23
54,874 4,277 1,676
1,812 141 55
2,554 199 78
486 190
27,199 2,120
4,422 345 135
75
53
72
-------
Table A-l (Continued)
Locations
Kansas
El Dorado
Cecty Oil Co.
Pester Refining Co.
Total
Kansas City
Phillips Petroleum Co.
Phillipsburg
CRA Inc.
Catlettsburg
Ashland Petroleum Co.
Louisiana
Baton Rouge
Exxon Co.
Belle Chasse
Gulf Oil Co., Alliance
Minnesota
Rosenount
Koch Refining Co.
Crude* Gasolinet Emission!
Capacity Bulk Loaded Rate
106m3 106m3 10"5g/s
4.57
1.31
5.88
5.22
15.32
7.88
29.60
0.53
0.47
1.38
0.71
-*
2.66
1.67
1.48
4.35
2.24
8.38
653
Population Exposed' to EDB
1.0-5.0 S_. 1-10.0 10.1-20.0 20.1-40.0 >40.0
137 11
17,172 1,338
500 39 15
51
Refinery
Conven t
Texaco
Garyville
Marathon Oil Co.
Lake Charles
Cities Service Oil Co.
Continental Oil Co.
Total
Meraux
Mucphy Oil Co.
Norco
Shell Oil Co.
11.
8.
11.
15.
4.
20.
5.
13.
40
13
61
?6
82
38
37
93
1
0
1
1
0
1
.03
.73
.04
.83
.48
.25
3.
2.
3.
5.
1.
3.
24
30
28
76
51
94
1,075
676
1,093
2,339
383
1,400
84
53
85
189
30
109
142,341 14,995 5,877 2,303
33
33
71
43
28
7.39
0.67
2.11
357
28
73
-------
Table A-l (Continued)
Location
Mississippi
Pascagoula
Chevron USA lac.
Crude* Gasolinet Emissionf
Capacity Bulk Loaded Rate
106m3 IQ^m3 lQ-5g/s
16.30
1.47
4.63
Population Exposed to EDB
_ _
1.0-5.0 5.1-10.0 10.1-20.0 20.1-40.0 >40.0
22,980 2,990
Missouri
Sugar Creek
Amoco Oil Co.
6.21
0.56 1.76
398 31
Montana
Billings
Continental Oil Co. 3.05
Exxon Co. 2.61
Total 5.66
New Jersey
Linden
Exxon Co. 16.54
Paulsboro
Mobil Oil Corp.
Perth Amboy
Chevron USA Inc.
Westville
Texaco, Inc.
Ohio
Lima
Standard Oil Co. - Ohio 9.75
Toledo
Gulf Oil Co. 2.92
Standard Oil Co. - Ohio 6.96
Sun Petroleum Prod. Co. 7,26
Total 17.14
Oklahoma
Ponca City
Continental Oil Co. 7.31
0.51
1.49
0.88
1.54
1.61
4.69
2.77
4.85
21,121 1,646
43,867 6,270
5.69
9.75
5.12
0.51
0.88
0.46
1.61
2.77
1.45
4,908
34,629
4,261
383
6,750
332
47,356
105,557
3,755
8,227
2,457
2,646
0.66 2.07
9,845 767
1,472
3,224
301
1,264
74
-------
Table A-l (Continued)
Locations
Oklahoma (continued)
Tulsa
Sun Petroleum Products Inc.
Pennsylvania
Marcus Hook
B P Oil Corp.
Sun Petroleum Products Co.
Total
Philadelphia
Atlantic-Richfield Co.
Gulf Oil Co.
Total
Texas
Bay town
Exxon Co.
Beaumont
Mobil Oil Corp.
Union Oil Calif.
Total
Borger
Phillips Petroleum Co.
Corpus Christi
Chanplin Petroleum Corp.
Coastal States Petrochem.
Hovell Corp.
Quintana Refining Co.
Saber Refining Co.
Southwestern Ref. Co.
Sun Petroleum Products Co.
Total
Deer Park
Shell Oil Co.
Houston
Atlantic-Rlchfield Co.
Charter Int. Oil Co.
Crown Central Petr. Co.
Crude*
Capacity
106m3
5.14
9.34
9.58
18.92
10.74
11.85
22.59
22.60
18.86
6.96
25.82
5.80
7.26
10.70
1.23
1.36
0.54
6.96
3.31
31.36
17.06
17.76
3.77
5.80
Casolinet
Bulk Loaded
106m3
0.46
1.70
Eraissiont
Rate
10-5g/s
1.45
5.36
Population Exposed to EDB
(ppt)'
2.03 6.41
1.0-5.0
8,397
I
6,764
5.1-10.0 10.1-20.0 20.1-40.0 >40.0
654
527
207
81
2.03 6.39 492,642 38,396 15,048 5,897
37,288 4,390 1,721
674
2.32 7.32 64,166 5.001 1,960 768 495
0.52 1.64 241 19
2.82 8.88 103,118 8,037 3,150 1,234 796
1.54 4.85 1,043 81 32 12
75
-------
Table A-l (Concluded)
Crude* Gasollnet Emissiont
Capacity Bulk Loaded Rate
Population Exposed to EDB
Location §
Texas (continued)
Houston (continued)
Eddy Refining Co.
Total
Port Arthur
American Petroflna Inc.
Gulf Oil Co.
Texaco Inc.
Total
Sweeney
Phillips Petroleum Co.
Texas City
Amoco Oil Co.
Marathon Oil Co.
Texas City Refining Inc.
Total
Washington
Anacortes
Shell Oil Co.
Texaco Inc.
Total
Ferndale
Atlantic-Richfield Co.
Mobil Oil Corp.
Total
Wyoming
Casper
Amoco Oil Co.
Little American Ref. Co.
Texaco Inc.
Total
**
Total Exposed Population
0.18
27.51
6.38
18.11
23.56
48.05
6.04
20.20
3.83
4.32
28.35
5.28
4.53
9.81
5.57
4.15
9.72
2.56
1.42
1.22
5.14
106m3 !Q-5g/a 1.0-5.0 5.1-10.0 10.1-20.0 20.1-40.0 >40.0
2.48
4.32
0.54
0.88
0.87
0.46
7.81
13.61
1.71
2.55 8.46
2.77
2.76
121,104
544
541
9,439
46,105 8,376
255 20
27,648 2,155
42
42
3,699
3,283
1,450 934
1,287 829
845
17
17
331 213
1.45 15,706 1,224
2,000,000 170,000 53.000 16,000 3,000
fOil and Gas Journal, May 28, 1977.
,SRI estimates.
JTO convert to ^/m3, divide by 130.
^^When more than one refinery is located in a city, it is assumed that they are co-located and emission levels are summed.
Rounded to two significant figures.
76
-------
Table A-2
POPULATION EXPOSURES IN STATES
HAVING REFINERIES WITH CRUDE CAPACITIES
LESS THAN 5 x 106m3/y
Population Exposed to
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Florida
Georgia
Hawaii
Illinois
Indiana
Kansas
Kentucky
Louisiana
Maryland
Michigan
Minnesota
Mississippi
Montana
Nebraska
New Hampshire
New Jersey
Number*
of
Locations
3
2
1
4
11
3
1
1
2
6
5
7
3
12
1
6
2
4
5
1
1
1
Total*
Crude
Capacity
106m3
2.86
4.24
0.23
3.51
13.75
3.73
0.33
0.28
5.74
16.11
4.81
13.71
1.66
17.83
1.65
8.71
5.19
2.81
3.42
0.29
0.75
0.35
To tall-
Gasoline
Bulk Loaded
106m3
0.26
0.38
0.02
0.32
1.24
0.33
0.03
0.03
0.52
1.45
0.43
1.23
0.15
1.60
0.15
0.78
0.47
0.25
0.31
0.03
0.07
0.03
Totalt
Emission
Rate
10-5g/s
0.81
1.20
0.06
1.01
3.91
1.04
0.09
0.09
1.64
4.57
1.35
3.87
0.47
5.04
0.47
2.46
1.48
0.79
0.88
0.09
0.22
0.09
Ratet
Per
Location
10-5g/s
0.27
0.60
0.06
0.25
0.36
0.35
0.09
0.09
0.82
0.76
0.27
0.55
0.16
0.42
0.47
0.41
0.74
0.20
0.18
0.09
0.22
0.09
EDB Concentrations
of 1.0 to 5.0 ppt
Population
per Location
32
3
0
16
88
15
15
0
268
364
67
31
0
68
396
125
151
15
0
0
31
100
State
Total
66
6
-
64
968
45
15
-
536
2184
335
217
-
816
396
750
300
60
-
-
31
100
-------
Table A-2 (Concluded)
Population Exposed to
oo
State
New Mexico
New York
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Tennessee
Texas
Utah
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total Exposed
Number*
of
Locations
7
2
3
3
10
1
7
1
25
5
1
2
3
i
8
Population'
Total*
Crude
Capacity
6.92
6.21
3.40
7.35
19.03
0.81
5.20
2.55
30.63
9.18
3.08
1.76
1.12
2.64
5.82
' Totalt
Gasoline
Bulk Loaded
106m
0.62
0.56
0.31
0.66
1.71
0.07
0.47
0.23
2.76
0.83
0.28
0.16
0.10
0.24
0.52
Totalt
Emission
Rate
10-5g/B
1.95
1.76
0.98
2.08
5.39
0.22
1.48
0.72
8.69
2.61
0.88
0.50
0.32
0.76
1.64
Ratet
Per
Location
10-5g/s
0.28
0.88
0.33
0.69
0.54
0.22
0.21
0.72
0.35
0.52
0.88
0.25
0.11
0.76
0.21
EDB Concentrations
of 1.0 to 5.0 PPt
Population
per Location
4
842
6
420
45
0
85
172
30
14
271
21
0
151
2
State
Total
28
1684
18
1260
450
-
595
172
750
70
271
42
-
151
16
12,000
*0il and Gas Journal, May 28, 1977.
TSRI estimates.
tRounded to two significant figures.
-------
ASSESSMENT OF HUMAN
EXPOSURES TO ATMOSPHERIC
ETHYLENE DICHLORIDE
Final Report
May 1979
By:
Benjamin E. Suta
Prepared for:
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
Task Officer: Jack K. Greer, Jr.
Project Officer: Joseph D. Cirvello
Contract No. 68-02-2835 Task 17
SRI Project CRU-6780
Center for Resource and Environmental Systems Studies
Report No. 82
-------
NOTICE
This report has been provided to the U.S. Environmental Protection
Agency (EPA) by SRI International, Menlo Park, California, in partial
fulfillment of Contract 68-02-2835. The opinions, findings, and
conclusions expressed herein are those of the authors and are not
necessarily those of EPA. Mention of company or product names is not to
be considered an endorsement by EPA.
-------
CONTENTS
LIST OF TABLES v
ACKNOWLEDGEMENTS vii
I INTRODUCTION 1
II SUMMARY 2
III CHEMICAL AND PHYSICAL PROPERTIES OF EDC
AND ITS ENVIRONMENTAL BEHAVIOR 8
Introduction 8
Physical Properties 8
Chemical Properties 9
Environmental Behavior 11
IV EDC PRODUCTION AND USES 14
Production 14
Uses 14
EDC Producers and Users 17
V POPULATION EXPOSURES FROM EDC PRODUCTION 21
General 21
Sources of Emissions 21
Emissions 21
Atmospheric Concentrations 23
Exposure Estimates 28
VI POPULATION EXPOSURES FROM PRODUCERS
THAT USE EDC AS A FEEDSTOCK 34
General 34
Sources of Emissions 34
Emissions 35
Atmospheric Concentrations 35
Exposure Estimates 35
ill
-------
VII POPULATION EXPOSURES FROM EDO IN AUTOMOBILE GASOLINE .... 40
General ........ 40
Exposures from Self-Service Operations 41
Exposures in the Vicinity of Service Stations 48
Urban Exposures Related to Automobile Emissions .... 54
Summary of Urban Exposures from Automobile Gasoline . . 59
VIII OTHER ATMOSPHERIC EXPOSURE ROUTES 60
General 60
Dispersive Uses 60
Transportation 61
Waste Disposal 64
BIBLIOGRAPHY 66
iv
-------
TABLES
II-l Summary of Estimated Population Exposures to
Atmospheric EDC from Specific Emissions Sources .... 4
II-2 Estimated Atmospheric Emissions of EDC for 1977 .... 5
III-l Physical Properties of EDC 10
IV-1 EDC Consumption 16
IV-2 EDC Producers and Major Consumers 18
IV-3 1977 EDC Production by Direct Chlorination and
Oxychlorination 19
IV-4 1977 Use of EDC Production Capacities 20
V-l EDC Oxychlorination Vent Emissions 22
V-2 Estimated Atmospheric Emissions from EDC Production
Facilities 24
V-3 Atmospheric EDC Monitoring Data for Calvert City,
Kentucky 25
V-4 Atmospheric EDC Monitoring Data for Lake Charles,
Louisiana 26
V-5 Atmospheric EDC Monitoring Data for New Orleans,
Louisiana - 27
V-6 Estimated One-Hour Average Downwind Atmospheric
Concentrations of EDC (yg/m3) 29
V-7 Estimated Human Population Exposures to
Atmospheric EDC Emitted by Producers 31
V-8 Comparison of EDC Monitoring and Modeling
Atmospheric Concentrations (ppb) 34
VI-1 Estimated EDC Atmospheric Emissions (g/s) for
Plants that Use EDC as a Feedstock 37
VI-2 Estimates of Population Exposures to Atmospheric EDC
Emitted by Plants that Use EDC as a Feedstock in
Various Products 39
VI-3 Estimates of Total Population Exposures to Atmospheric
EDC Emitted by Plants that Use EDC as a Feedstock ... 40
VII-1 Self-Service Operations 43
VII-2 Gasoline Market Share of Self-Service Stations in
Four AQCRs, Spring 1977 44
v
-------
VII-3 Gasoline Market Share of Self-Service Stations
in Two Metropolitan Areas, 1976 46
VII-4 Sampling Data from Self-Service Gasoline Pumping .... 47
VII-5 Estimates of EDC Exposures from Self-Service
Gasoline Pumping . 49
VII-6 Service Station Density in Four Metropolitan AQCRs ... 51
VII-7 Rough Dispersion Modeling Results for EDC Emissions
for Gasoline Service Stations 54
VII-8 Automotive EDC Emission Factors 56
VII-9 Distribution of Cities by 1970 Population 58
VII-10 Estimated U.S. City Exposures to EDC from the
Evaporation of Automobile Gasoline 59
VIII-1 Summary of Uncontrolled Emission Factor for the
Transfer of Benzene 63
VIII-2 Estimated 1977 EDC Emissions as Solid Waste and to
Water from EDC. Production 65
vi
-------
ACKNOWLEDGEMENTS
It is a pleasure to acknowledge the cooperation and guidance given
by several individuals of the U.S. Environmental Portection Agency,
Office of Air Quality Planning and Standards. Ken Greer, Strategies and
Air Standards Division, and Dr. George H. Wahl, Jr., EPA Consultant,
provided direction throughout the study. David Mascone of the Emission
Standards and Engineering Division gave valuable assistance in regard to
control technology and emission factors and George Schewe of NOAA
provided input on atmospheric dispersion modeling.
Mr. Casey Cogswell, SRI International, Chemical Industries Center,
generously provided information and guidance concerning the
manufacturing and uses of ethylene dichloride.
vii
-------
I INTRODUCTION
This report is one in a series that SRI International is providing
for the U.S. Environmental Protection Agency (EPA) to estimate
populations at risk to selected pollutants. Primarily, this study has
sought to estimate the environmental exposure of the U.S. populaton to
atmospheric ethylene dichloride (EDC) emissions. The principal
atmospheric sources we consider in this report are facilities at which
EDC is produced or used as a chemical intermediate and gasoline that
contains EDC as a lead scavenger. Possible exposures from
transportation of EDC, disposal of EDC wastes, and other minor uses are
also described.
-------
II SUMMARY
EDC is one of the highest volume chemicals used in the United
States, with approximately 5 million metric tons (mt) synthesized during
1977. More than 80% of the EDC produced is used in the synthesis of
vinyl chloride monomer (VCM). The majority of the remaining production
is used in the synthesis of 1,1,1-trichloroethane (1,1,1-TCE or methyl
chloroform), trichloroethylene (TCE), perchloroethylene (PCE),
vinylidine chloride (VDCM), and ethyleneamines (EA); EDC is also
employed directly as a lead scavenger for gasoline.
EDC is a colorless, oily liquid that has a sweet taste, a
chloroform-like odor, and a volatility similar to that of gasoline. It
boils at 83.5°C, melts at -35.4°C, and has a specific gravity of
1.2. It is relatively stable in water, but evaporates rapidly from
water to the atmosphere where it is destroyed by photooxidation.
Estimates of EDC'a half-life in the atmosphere range from weeks to
months, a period sufficiently long for aerial transport to play a major
role in its distribution but relatively short for it to accumulate in
the terrestrial and aquatic environment. It has a potential for
bioaccumulation; however, no firm evidence now exists to support
bioaccumulation in the marine environment or in other biota.
EDC does not occur naturally in the environment. Environmental
exposures occur mainly from EDC Lost during production, from EDC used as
a chemical intermediate in producing other chemicals, or in its use in
gasoline as JL lead scavenger. Minor environmental exposures may occur
through dispersive uses of EDC such as in grain fumigants, paints,
coatings, adhesives, cleaning, and in the preparation of polysulfide
compounds. Additional environmental exposures may occur from spills and
venting during EDC transportation and from evaporation resulting from
waste disposal.
-------
Monitoring data for occupational exposures to EDC have been report-
ed in the literature for more than four decades; however, the environ-
mental monitoring data that have been collected date only from around
1975 and they are limited. A number of environmental monitoring studies
have failed to find atmospheric EDC in the general U.S. environmental at
the ppt detection level (Grimsrud and Rasmussen, 1975; Singh et al.,
1977; and Hanst, 1978). PEDCo (1978) found EDC at the ppb level in the
atmosphere surrounding three EDC production facilities. (The maximum
integrated 24-hr value was 180 ppb.) Pellizzari (1978) reported the
detection of EDC concentrations of less than 55 ppb near a chemical dis-
posal site in New Jersey.
The current Occupational Safety and Health Administration (OSHA)
standard for occupational exposure to EDC is 50 ppm (8-hr time-weighted
average). In March 1976, the National Institute of Occupational Safety
and Health recommended an exposure limit of 5 ppm (time-weighted averge
for a 10-hr workday or less, a 40-hr workweek). These levels, however,
were designed to protect against toxic effects other than cancer and may
not provide adequate protection from potential carcinogenic effects
(NIOSH, 1978).
Human population exposures to atmospheric EDC have been estimated
for emissions resulting from its production, its use as a feedstock in
the production of other chemicals, and its use as a lead scavenger in
automobile gasoline. Other potential exposure routes such as emissions
resulting from transportation and emissions from other product uses have
also been described. The population exposure estimates given in Table
II-l are based on the calculated atmospheric emissions given in Table
II-2.
These emission and exposure estimates have necessitated reliance on
very limited data. Because of the paucity of measured atmospheric EDC
data, it was necessary to approximate concentrations through the use of
dispersion modeling. Moreover, the resulting estimates are subject to
considerable uncertainty in regard to: (1) the quantity of EDC emis-
sions, (2) EDC production and consumption levels, (3) certain source
3
-------
Table II-l
SUMMARY OF ESTIMATED POPULATION EXPOSURES TO ATMOSPHERIC EDC FROM SPECIFIC EMISSIONS SOURCES
Annual Average
EDC Concentra-
tion (ppb)a
10
6.00-10.00
3.00- 5.99
1.00- 2.99
0.60- 0.99
0.30- 0.59 1
0.10- 0.29 4
0.060-0.099 1
0.030-0.059 3
0.010-0.029
EDC
1,700
3,300
28,000
280,000
400,000
,500,000
,300,000£
,900,000*
,500,000f
550,000*
Total 12,500,000
Production Facilities'1 Gasoline
1^1,1- Lead Service Automobile Automobile
VCM TCE TCE PCE EA VDCM Scavenger Stations0 Emissions*1 Refueling6
1
( g )
1,300
360 70
30,060 1,700 390 80 17,000 270 1,900
42,000 16,000 10,000 500 8,000 3,400 3,400
260,000 83,000 47,000 17,000 43,000 34,000 25,000
940,000 170,000 140,000 250,000 37,000 90,000 350,000 1,000,000 13,000,000
1,300,000 260,000 200,000 270,000 110,000 130,000 380,000 1,000,000 13,000,000 30,000,000
To convert to /ig/m^, multiply each exposure level by 4.1.
b Production facilities that either produce EDC or use EDC as a feedstock in the production of another chemical.
c These are exposures to people who reside near gasoline service stations.
d These are exposure* from evaporative emissions from pre-1975 automobiles.
e These are exposures to people while refueling their automobiles at self-service gasoline stations.
* These are underestimates because the dispersion modeling results were not extrapolated beyond 30 km from each EDC production facility.
There are additional people Who ate exposed to EDC concentrations of 0.01-0.1 ppb at distances greater than 30 km from the larger producton
facilities.
8 Estimated as 30 million people exposed to an EDC concentration of 1.5 ppb for 2.2 hr/yr. The annual average time-weighted exposure is
0.0004 ppb.
-------
Table II-2
ESTIMATED ATMOSPHERIC EMISSONS OF EDC FOR 1977
Emissions (1,000 mt/yr)
EDC production
Fugitive
Storage
Direct chlorination
Oxych1orination
Subtotal
Production using EDC as
Feedstock
VCM
1,1,1-TCE
TCE
PCE
EA
VDCM
Lead scavenger
Subtotal
Automobile gasoline
Service stations
Auto emissions
5.2
14.5
6.3
17.9
43.9
1.1
0.4
0.2
0.3
0.3
0.2
0.2
2.5
0.1
1.2
Subtotal
Other
Dispersive uses
Transporation3
Waste disposal3
Total
5.0
"BTT
aNot included. Rough order estimates place these emissions as much
less than 2,400 mt/yr for transportation and much less than 29,100 mt/yr
for waste disposal.
-------
locations, (4) control technologies employed, (5) deterioration in con-
trol technologies over time, (6) physical characteristics of EDC
sources, (e.g., stack height), (7) details on atmospheric dispersion and
degradation, and (8) living patterns of the exposed population. Given
these complex and variable factors, the accuracy of the estimates could
not be assessed. Nevertheless, the estimates are believed to be a
reasonable approximation of actual conditions. Comparisons of atmo-
spheric monitoring and modeling data for EDC concentrations near 3 sites
used in this report shows agreement within 70% and averaging 25%.
During 1977, 18 facilities produced an estimated 5 million mt of
EDC. It is estimated that approximately 12.5 million people are exposed
to average annual EDC concentrations of 0.01 to more than 10 ppb from
this production. Estimates of exposures to concentrations of less than
0.1 ppb from production facilities are underestimates because the dis-
persion modeling results were not extrapolated beyond 30 km from the
plants. There are additional people who are expected to be exposed to
EDC concentrations of 0.01-0.1 ppb at distances of greater than 30 km
from the larger production facilities. However, it is generally assumed
that disperions modeling results are unreliable beyond 20 to 30 km from
the source.
Estimates are given for exposures to EDC used as a feedstock in the
production of VCM, 1,1,1-TCE, TCE, PCE, EA, VDCM, and gasoline lead
scavengers. Many of these chemicals are produced at the same facilities
that produce the EDC feedstock. Of the 28 producton plants involved, 18
also produce EDC. In 1977, approximately 5 million mt of EDC was re-
quired, with more than 80% used in producing VCM. More than 2 million
people are exposed to annual average EDC concentratons of 0.01 to 1.0
ppb from these operations.
Leaded gasoline additives contain EDC as a lead scavenger.
Although the EDC is expected to be destroyed during combustion, evapora-
tive emissions occur during refueling operations and from the gas tanks
and carburetors of automobiles. These emissions are expected to de-
crease as newer model automobiles replace the pre-1975 models. It has
6
-------
been estimated that approximately 30 million people are exposed to EDC
concentrations of 1.5 ppb for 2.2 hr/yr while refueling their auto-
mobiles at self-service stations. Similarly, approximately 1 million
people residing near gasoline service stations are exposed to average
annual EDC concentrations of 0.01 to 0.03 ppb from refueling losses.
Another 13 million are exposed to annual average EDC concentrations of
0.01 to 0.03 ppb from automobile evaporative emissions.
-------
Ill CHEMICAL AND PHYSICAL PROPERTIES OF EDC
AND ITS ENVIRONMENTAL BEHAVIOR*
Introduction
The Chemical Abstracts Service registry number of EDC is 000107062;
the NIOSH number is K005250. To minimize confusion between EDC
(C H Cl ) and cis and trans dichloroethylene (C2H2C12),
Drury and Hammons (1978) recommend that EDC be referred to as
1,2-dichlorethane in place of ethylene dichloride. Many synonyms and
trade names are also used: Brocide; Destrucol Borer-Sol;
Di-chlor-mulsion; sym-dichloroethane; alpha, beta-dichloroethane; di-
chloroethylene; Dutch liquid; EDC; ENT 1,656; ethane dichloride;
ethylene chloride; glycol dichloride; and oil of the Dutch chemists
(NOISH, 1977; Mitten et al., 1970).
The composition and structure of 1,2-dichloroethane (EDC) are in-
dicated by the molecular formula, CCL,, and the line diagram,
H H
Cl - C - C - Cl
H H
Physical Properties
EDC is a colorless, oily liquid that has a sweet taste and a
chloroform-like odor (Hawley, 1977). It is volatile and evaporates at a
rate 0.788 time that of carbon tetrachloride or gasoline (Whitney,
The discussion given here has been summarized from a draft report by
Drury and Hammons (1978).
-------
1961). Air saturated with EDC contains 350 g/m at 20°C and
*j
537g/m at 30 C. EDC is completely miscible with ethanol, chloro-
form, ethyl ether, and octanol (Windholtz, 1976; Johns, 1976). The par-
tition coefficient, log P, of EDC between octanol and water is 1.48
(Radding et al., 1977), reflecting preferential solubility in organic
media.
Vaporized EDC solvent is readily ignitedthe closed cup flash
point being only 13°C. The liquid is also flammable, burning with a
smoky flame, but the ignition temperature is high, 413°C. Under a
pressure of 1 atm, EDC steam distills at 71.9°C. The binary azeotrope
contains 19.5% water; 14 other binary azeotropes are known (Mitten et
al, 1970). A ternary azeotrope containing 78% 1,2-dichloroethane, 17%
ethanol, and 5% water boils at 66.7°C. Other properties are given in
Table III-l.
Chemical Properties
EDC is stable at ambient temperatures but slowly decomposes in the
presence of air, moisture, and light. During decomposition the liquid
EDC becomes darker in color and progressively acidic. It can corrode
iron or steel containers, but these deleterious reactions are completely
inhibited by small concentrations of alkyl amines (Hardie, 1964).
Both chlorine atoms in EDC are reactive and can be replaced by
other substituents. This bifunctional nature of EDC makes it useful in
the manufacture of condenstion polymers (Rothon, 1972). Hydrolysis of
EDC, with slightly acidic 160°C to 175°C water at 15 atm, or with
140°C to 250°C aqueous alkali at 40 atm, yields ethylene glycol; at
120°C, the addition of ammonia under pressure to EDC yields ethylene-
diamine. 1,1,2-TCE and other higher chloroethanes are fomred by chlor-
inating EDC at 50°C in light from a mercury vapor lamp. EDC reacts
with sodium polysulfide to form polyethylene tetrasulfide, and with
oleum to give 2-chloroethylsulfuryl chloride. With Friedel-Crafts ca-
talysis, both chlorine atoms in EDC can be replaced with aromatic ring
compound (Hardie, 1964).
-------
Table III-l
PHYSICAL PROPERTIES OF EDO
Molecular weight 98.96
Density, g/ml at 20°C 1.2351
Melting point, °C -35.36
Boiling point, °C 83.47
Index of refraction, 20°C 1.4448
Vapor pressure, torr, at °C
-44.5 1
-13.6 10
10.0 40
29.4 100
64.0 400
82.4 760
Solubility in water, ppm w/w at °C
20 8,690
30 9,200
Biochemical oxygen demand (5 days), % 0
Theoretical oxygen demand, mg/mg 0.97
Measured chemical oxygen demand, mg/mg 1.025
Vapor density (air =1) 3.42
Flash point, open cup, °C 13.0
Ignition temperature, °C 413.0
Explosive limit, % volume in air
Lower 6.2
Upper 15.9
Specific resistivity 9.0 x
Viscosity, cP, at 20°C 0.840
Dielectric constant, 10.45
Surface tension, dyne/cm 33.23
Coefficient of cubical expansion, 10°C-30°C 0.0016
Latent heat of fusion, cal/g 21.12
Latent heat of vaporization, cal/g, at boiling point 77.3
Specific heat, cal/g °C
Liquid at 20°C 0.308
Vapor, 1 atm at 97.1°C 0.255
Critical temperature, °C 288
Critical pressure, atm 53
Critical density, g/cm3 0.44
Thermal conductivity, Btu/hr-ft2 at 20°C 0.825
Heat of combustion, cP, kcal/g-mole 296.36
Dipole moment, ESU 1.57 x 10~18
Conversion factors, 25°C, 760 torr 1 mg/L* = 1 g/m3 = 247 ppm
1 ppm =4.05 gm/m3 =4.05 g/L
Source: Draft report by Drury and Hammons (1978),
10
-------
Environmental Behavior
Bioaccumulation and Biomagnification
EDC's physical and chemical properties exhibit opposing tendencies
with respect to bioaccumulation, but high vapor pressure and low latent
heat of vaporization argue that the compound is exhaled from the lungs
in the same condition in which it was inhaled. In fact, no firm evi-
dence exists for the bioaccumulation of EDC in food chains under envi-
ronmental conditions (Radding et al., 1977). Pearson and McConnell
(1975) in searching for simple aliphatic chlorocarbons in several tro-
phic levels of the marine environment near the industrialized area of
Liverpool, found no evidence of EDC. In laboratory studies on oysters
and fish using EDC labeled with carbon -14, Pearson and McConnell did
see rapid storage of the chlorinated hydrocarbon up to an asymptotic
level, but this accumulation was followed by loss of EDC on transfer of
the organisms to clean sea water. Parallel analyses by chromatographic
techniques showed reduced levels of EDC in the organisms, indicating
that metabolism of the compound occurred in the tissues of both fish and
oysters.
Biological Degradation
The conclusions of the few literature references to microbial de-
gradation of simple chlorinated hydrocarbon compounds conflict. Some
authors report these compounds are not metabolized either by aerobic or
anaerobic microorganisms (Pearson and McConnell, 1975). Other micro-
biologists believe biodegradation can occur via co-metabolic processes
(Horvath, 1972), but no evidence supporting biodegradation of EDC has
been found. There is general agreement, however, that mammals metabo-
lize these compounds, producing chlorinated acetic acids either directly
or via chloroethanols. All of the resulting chlorinated acetic acids
are susceptible to further degradation by microorganisms in sea water
(McConnell et al., 1975).
11
-------
Chemical Degradation
Photooxidative reactions involving atmospheric EDC probably result
in monochloroacetyl chloride, hydrogen chloride, and monochloroacetic
acid (Spense and Hanst, 1978). Alcohols, ketones, alkyl nitrates, and
cleavage produces arising from intermediate alkoxy radicals are also
possible products. Preliminary data indicate the half-life of EDC in
the atmosphere may be about 3 to 4 months (Pearson and McConnell, 1975;
EPA, 1975). Based on an average HO radical concentration of
0.8 x 10 M, Radding et al. (1977) estimated a combined
oxidativephotolysis half-life of 234 hr. The recent calculations of
Altshuller and the re- cent experiments of Snelson et al. (1978)
indicate tropospheric life- times of EDC of approximately 0.75 to 1
year. Although the half-life remains to be determined definitively,
available estimates make it clear that the lifetime of EDC in the
troposphere, although short in an abso- lute sense, is sufficiently long
for aerial transport to play a major role in its distribution.
EDC is resistant to hydrolysis. Radding et al. (1977) estimated a
hydrolysis half-life of approximately 50,000 yr. This estimte is much
longer than the 6- to 18-month half-lives observed for similar, but not
identical, compounds subjected to a combination of hydrolysis, oxida-
tion, and photolysis (Dilling et al., 1975); nevertheless, it appears
that hydrolysis of EDC is slow compared to other pertinent environmental
processes, such as volatilization or photolysis.
Dilling et al. (1975) and McConnell et al. (1975) studied the re-
moval of compounds similar to EDC from water by adsorption on several
common substrates. Dilling et al. observed little or no adsorption of
chlorinated hydrocarbons on clay, limestone, sand, and peat moss in lab-
oratory experiemtns that involved aqueous solutions containing 1 ppm
organic contaminant. McConnell et al. reached similar conclusions about
adsorption of chlorinated hydrocarbons from seawater by coarse gravels,
but they found relatively high adsorption by Liverpool Bay sediments
rich in organic detritus. The divergent conclusions of these studies
probably reflect different experiemtnal conditions: The hydrocarbon
12
-------
concentrations in the experiments of Billing et al. were well below the
solubility limits of the various compounds used, and adsorption under
these conditions is less likely than in Liverpool Bay, which receives
large volumes of industrial and domestic effluents.
The relatively high vapor pressure of EDC causes rapid volatil-
ization of the hydrocarbon from aqueous effluents. After 96 min at am-
bient temperature, about 90% of the EDC initially present in water at a
concentration of 1 ppm evaporated (Billing et al., 1975). this rate
corresponds to a vaporization half-life of 29 min. Comparison of these
data with those for other environmental removal processes indicate that
volatilization is the chief process for removal of EDC from water.
Persistence
EDC has a long hydrolysis half-life, a short vaporization half-life
from water, and a relatively short photooxidative half-life in the atmo-
sphere. It is unlikely to accumulate in the environment. Note, how-
ever, that one of EDC's photooxidative products is chloroacetyl
chloride, which may be sufficiently stable to reach the stratosphere and
interact destructively with the ozone layer.
Environmental Transport
Because the vapor pressure of EDC is moderately high, most emis-
sions from manufacturing operations occur as vapors that are vented
directly to the atmosphere. Even when initially present in wastewater
or solid waste products, EDC tends to transfer rapidly to the atmo-
sphere. This volatility, coupled with an atmospheric half-life suf-
ficiently long for aerial transport, results in most distribution of EDC
in the environment occurring by aerial transport (McConnell et al.,
1975; Pearson and McConnell, 1975). Some transfer of EDC from air to
water also occurs, particularly as a result of rainfall. This effect is
assumed to be minor when compared to aerial transport, but quantitative
data comparing these transport routes are lacking.
13
-------
IV EDC PRODUCTION AND USES
Production
The annual EDC production capacity for U.S. plants is approximately
7.3 million mt. From 1973 to 1977 the industry operated at about 60-70%
of capacity, producing 4.6, 4.7, 3.7, 5.0, and 5.2 million mt, respec-
tively. The overall production during these years may have been even
higher than indicated because captive production is not always adequate-
ly recorded in published data. Future growth of the market is expected
to average 4% to 5%/yr through 1981, at which time the demand for EDC is
expected to be 6.6 million mt. Five of the major producing companies
are currently expanding production facilities or are planning increased
production in the near future (Chemical Marketing Reporter, 1977).
Uses
EDC is used primarily as a raw material in the synthesis of other
chemicals, in particular for VCM, 1,1,1-TCE, TCE, PCE, VDCM, EA, and as
a lead scavenger for gasoline. Primary uses of these compounds are as
follows:
VCM Its major use is in the production of
PVC and its copolymer resins. Small
amounts are used in polyvinylidene
chloride and other copolymers.
1»1»1-TCE Its major use is for solvent clean-
ing. Minor uses include aerosol pro-
pellant, solvent in adhesives and
coating formulations, drain cleaner,
and fabric spotting fluid.
It is almost entirely used as a
metal-cleaning solvent.
14
-------
Its major uses are for metal cleaning
and dry cleaning.
EA Its major uses are as a chelating
agent and carbamate fungicide. Other
uses include detergents and softening
agents, specialty resins, epoxy hard-
eners, and corrosion inhibitors.
VDCM It is used mainly in the production
of polyvinylidene copolymers
Lead Scavenger It is used in gasoline antiknock mix-
tures.
The quantities of EDC consumed for these and other uses are shown in
Table IV-1. More than 80% of the EDC produced is used in the manufac-
ture of VCM. Each of the other compounds listed above requires 2% to 3%
of the total EDC produced. Exports account for about 3.4% of the EDC
produced, and other minor products require less than 0.2% of EDC produc-
tion.
Auerback Associates (1978) estimated EDC consumption for other
minor uses in 1977 at about 5,000 mt. Of this subtotal, about 28% was
used in the manufacture of paints, coatings, and adhesives. Extracting
oil from seeds, treating animal fats, and processing pharmaceutical pro-
ducts required 23% of the subtotal. An additional 19% was consumed in
cleaning textile products and polyvinyl chloride manufacturing equip-
ment. Nearly 11% was used in the preparation of polysulfide compounds.
Grain fumigation required about 10%. The remaining 9% was used as a
carrier for amines in leaching copper ores, in the manufacture of color
film, as a diluent for pesticides and herbicides, and for other
miscellaneous purposes.
EDC Producers and Users
Table IV-2 lists the major EDC producers and consumers along with
their estimated January 1979 installed production capacity. As the
15
-------
Table IV-1
EDC CONSUMPTION
(Thousands of metric* tons per year)
YEAR
Use
VCM
1,1,1-TCE
TCE
PCE
EA
VDCM
Lead scavenger
Other
Net exports
1973
3,645
184
141
104
128
83
106
b
1974
3,871
198
121
98
132
92
97
b
1975
3,015
155
92
89
123
83
80
b
1976
4,079
213
99
89
132
88
93
b
1977
4,300
215
83
87
136
97
89
b
19823
5,635-
6,140
260-280
85-110
87-95
113-119
125-135
39
b
167
(133)
(26)
(199)
177
180
Total
4,558
4,742
3,663
4,992
5,194
6,524-
7,098
Source: SRI estimate.
aProjected consumption.
"Other uses, which are not included in consumption, in 1974 were estimated
at 7,000 rat and at 5,000 mt in 1977.
16
-------
table indicates, most EDC producers have the capacity to use most of the
EDC they produce as feedstock for other products within their own
plants. In fact, in recent years only a small fraction (10% to 15%) of
the total production of EDC has been sold on the open market (US. Inter-
national Trade Commission, 1973-1977).
EDC is produced by the "balanced process." This process involves a
combination of direct chlorination of ethylene and oxychlorination of
ethylene using hydrogen chloride, which in turn is produced in the
cracking of EDC to VCM. The EDC manufactured by oxychlorination of
ethylene is generally used captively as an intermediate in VCM pro-
duction. Table IV-3 shows the percentages of EDC produced by direct
chlorination and by oxychlorination, by producer.
Chemical producers rarely operate at maximum production capacity
for a specific chemical. Table IV-4 shows the percentage of production
capacity employed in 1977 to produce EDC and the major chemicals in
which it is used as a feedstock.
17
-------
00
Table IV-2
EDC PRODUCERS AND MAJOR CONSUMERS
(January 1, 1979, production capacities in thousands of metric tons)
Producer
Location
Borden Chemical
Conoco Chemical
Diamond Shamrock
Diamond Shamrock
Dow Chemical
Dow Chemical
Dow Chemical
Dow Chemical
duPont
duPont
duPont
Ethyl Corp.
Ethyl Corp.
B. F. Goodrich
Houston Chemical
ICI America*
Honochem
Nalco Chemical
PPG Industries
PPG Industries
Shell Chemical
Shell Chemical
Stauffer Chemical
Stauffer Chemical
Union Carbide
Union Carbide
Vulcan Chemical
Vulcan Chemical
Geismar, LA
Lake Charles, LA
Deer Park, TX
La Porte, TX
Freeport, TX
Oyster Creek, TX
Pittsburg, CA
Plaquemine, LA
Antioch, CA
Corpus Chriati, TX
Deepwater, NJ
Baton Rouge, LA
Houston, TX
Calvert City, KY
Beaumont, TX
Baton Rouge, LA
Geismar, LA
Freeport, TX
Lake Charles, TX
Guayanilla, PR
Deer Park, TX
Norco, LA
Carson, CA
Louisville KY
Taft, LA
Texas City, TX
Geismar, LA
Wichita, KS
Total
Source: SRI estimates.
7,316
VCM
1.1,1-TCE
TCE
PCE
EA
VDCM
Scavenger
524
145
719
726
499
953
318
118
454
318
544
379
635
544
154
68
68
150
224
525
17 45
749
150 167 51 Ob 60 45C
525 Ob
Ob
936 112 45C
20C
Ob
20C
248 15 14 20C
20C
749
15C
224
Ob
5C
229 130 68 54 30C
375
629
525
130
Ob
70
60
Ob 41
Ob
6,218
a Plant was purchased from Allied Chemical in September 1978.
b Process does not use EDC as a feedstock.
c Rough order estimates.
409
151
154
190 120
100
-------
Table IV-3
1977 EDC PRODUCTION BY DIRECT CHLORINATION AND OXYCHLORINATION
Direct
Chlorination
Oxy-
chlorination
Producers
Conoco Chemical
Diamond Shamrock
Dow Chemical
Dow Chemical
Dow Chemical
Ethyl Corporation
Ethyl Corporation
B. F. Goodrich
ICI America3
PPG Industries
PPG Industries
Shell Chemical
Shell Chemical
Stauffer Chemical
Union Carbide
Union Carbide
Vulcan Chemical
Locations
Lake Charles, LA
Deer Park, TX
Freeport, TX
Oyster Creek, TX
Plaquemine, LA
Baton Rouge, LA
Houston, TX
Calvert City, KY
Baton Rouge, LA
Lake Charles, LA
Guayanilla, PR
Deer Park, TX
Norco, LA
Long Beach, CA
Taft, LA
Texas City, TX
Geismar, LA
49.2
35.8
57.1
(b)
51.7
52.7
100.0
33.3
66.7
77.2
(b)
66.2
(b)
69.2
100.0
100.0
0.0
50.8
64.2
42.9
(b)
48,
47.
0.0
66.7
33.3
22.8
(b)
33.8
(b)
30.8
0.0
0.0
100.0
Source: Draft report by Drury and Hammons (1978).
aPlant was purchased from Allied Chemical in September 1978
available.
19
-------
Table IV-4
1977 USE OF EDC PRODUCTION CAPACITIES
(Thousands of metric tons)
Product
EDC
VCM
1,1,1-TCE
TCE
PCE
EA
VDCM
Lead scavenger
EDC
Capacity3
7,316
6,218
409
151
154
190
120
100
EDC used
in 1977
Production^3
5,194
4,300
215
93
87
136
97
89
Percent
Capacity Used
71.0
69.2
52.6
61.6
56.5
71.6
80.8
89.0
aSee Table IV-2. This is the amount that would be used annually if the
product was produced at 100% of capacity.
bSee Table IV-1.
20
-------
V POPULATION EXPOSURES FROM EDO PRODUCTION
General
As was shown in Table IV-2, most of the EDC produced is used as
feedstock in the production of other chemicals, particularly VCM. the
majority of both the EDC produced and the chemicals that use EDC as
feedstock are made at the same facilities, thus, people residing near
these production facilities can be exposed to atmospheric EDC from
several types of production. Section VI sets forth the exposure from
chemical production facilities that use EDC as a feedstock.
Sources of Emission
EDC producers and their individual capacities are listed in Table
IV-2. The total annual capacity of the 18 plants listed is 7.3 million
mt. Table IV-4 indicates that approximately 71% of the production capa-
city was used during 1977. Because production data for each plant are
unavailable, we have assumed that each operates at 71% of capacity.
Emissions
Four principal sources of emissions have been identified: direct
chlorination vent stack, oxychlorination vent stack, fugitive emissions,
and emissions from tank storage.
As part of the Synthetic Organic Chemical Manufacturing Industry
study under way in EPA's Emission Standards and Engineering Division of
the Office of Air Quality Planning and Standards and as a result of the
detailed study of the VCM industry, a significant amount of engineering
data on the EDC industry are available. Table V-l summarizes the data
collected for oxychlorination vent emissions for 10 production plants.
We applied the averge vent emission factor of those plants (1.0%) to
plants for which no emission data are available. We estimated an emis-
sion factor of 0.22% for the direct chlorination process vent emissions,
with fugitive emissions estimated at 0.1% of plant
21
-------
Table V-l
EDC OXYCHLORINATION VENT EMISSIONS
Plant and Location
Oxychlorination
Production3
(g/s)
Conoco, Lake Charles, LA 5,987
Diamond, Deer Park, TX 2,871
Ethyl, Baton Rouge, LA 3,392
Goodrich, Calvert
City, KY 6,819
ICI America, Baton
Rouge, LA 2,366
PPG, Lake Charles, LA 2,800
Shell, Deer Park, TX 4,464
Stauffer, Long Beach, CA 1,071
Vulcan, Geismar, LA 3,408
Dow, Oyster Creek, TX 5,609
EDC
Emissions'
(g/s)
12 3
2.7
43.6
26.4
70.1
0.0
25.2
19.6
81.6
0.0
Emission Factor
(g emission/
g production)
0.0021
0.0009
0.0129
0.0039
0.0296
0.0000
0.0056
0.0183
0.0239
0.0000
Average
0.0097
aTotal capacity given in Table IV-2 times 71% use times percent
oxychlorination production given in Table IV-3.
bSource: EPA (1978).
cBased on an EPA engineering estimate.
22
-------
production. Emissions for storage tanks at production facilities were
estimated as 2.8 g/yr/kg of annual capacity (Mascone, 1978). Table V-2
gives the estimated emissions resulting from emission factors. Total
EDC atmospheric emissions are estimated as 43.9 thousand mt/yr (1,312
g/s) or about 0.8% of the amount produced. The storage and direct
chlorination emission rates are for uncontrolled plants. The industry
now has some controls on these two emission points but data are insuf-
ficient to estimate the present degree of control.
Atmospheric Concentrations
Atmospheric monitoring data have been collected from three loca-
tions that have EDC production facilities (PEDCo, 1978). These three
locations are (1) near the B. F. Goodrich plant in Calvert City, Ken-
tucky, (2) near the Conoco plant in Lake Charles, Louisiana, and (3)
near the Shell plant at Norco, Louisiana and the Union Carbine plant at
Taft, Louisiana. The Goodrich, Conoco, and Shell plants each have an-
nual EDC production capacities of approximately 500,000 mt. The Union
Carbide plant has an annual capacity of approximately 70,000 mt.
Twelve monitoring stations were positioned around each location.
Data were recorded for 10 days in New Orleans, 12 days in Lake Charles,
and 13 days in Calvert City. The preliminary results of the monitoring
data are summarized in Tables V-3 through V-5. Average 12- to 13-day
atmospheric concentrations ranged from 0 to 5 ppb for the Calvert City
stations, 1 to 43 ppb for the Lake Charles stations, and 0.1 to 12.1 ppb
for the New Orleans stations. Individual 24-hr concentrations were much
higher. Generally, the concentrations for locations near the plants in
the Lake Charles area were almost 10 times those for the Calvert City
and New Orleans areas. The differences may be attributable to meteor-
ological conditions, plant production at sampling times, emission con-
trols, positioning of the monitoring stations with respect to plant
location and wind direction, or other EDC sources in the areas.
The monitoring data show that elevated EDC concentrations exist in
the vicinity of at least 3 EDC production facilities; however, the data
23
-------
Table V-2
ESTIMATED ATMOSPHERIC EMISSIONS FROM EDC PRODUCTION FACTILITIES
NJ
Plant
Conoco
Diamond
Diamond
Dow
Dow
Dow
Ethyl
Ethyl
Goodrich
1C I America
PPG
PPG
Shell
Shell
Stauffer
Union Carbide
Union Carbide
Vulcan
Location
Lake Charles, LA
Deer Park TX
La Porte, TX
Freeport, TX
Oyster Creek, TX
Plaquemine, LA
Baton Rouge, LA
Houston, TX
Calvert City, KY
Baton Rouge, LA
Lake Charles, LA
Guayanilla, PR
Deer Park, TX
Norco, LA
Carson, CA
Taft, LA
Texas City, XX
Geiamar, LA
Total
Production3
103 mt/yr
372
103
510
515
354
678
226
84
322
226
386
269
451
386
109
48
48
107
(g/s)
11.800
3,265
16,190
16,345
11,235
21,455
7,160
2,660
10,220
7,160
12,250
8,533
14,295
12,250
3,470
1,530
1,530
3,380
Fugitive
11.8
3.3
16.2
16.3
11.2
21.5
7.2
2.7
10.2
7.2.
12.3
8.5
14.3
12.3
3.5
1.5
1.5
3.4
Storage
33.0
9.1
45.3
45.7
31.4
60.1
20.1
7.5
28.6
20.1
34.3
23.9
40.0
34.3
9.7
4.3
4.3
9.5
Emissions (g/s)
Direct
12.8
2.6
17.8
20.5
12.4
24.4
8.3
5.9
7.5
10.5
20.8
9.4
20.8
13.5
5.3
3.4
3.4
0.0
Oxychlorination
12.6
1.9
15.8
67.3
0.0
99.5
43.7
0.0
26.6
70.6
0.0
41.4
27.1
59.4
19.6
0.0
0.0
80.8
Total
70.2
16.9
95.1
149.8
55.0
205.5
79.3
16.1
72.9
108.4
67.4
83.2
102.2
119.5
38.1
9.2
9.2
93.7
5,194
164.9
461.2
199.3
566.3
1,391.7
Source: SRI estimates.
Assumed to be 71Z of production capacity.
-------
Table V-3
ATMOSPHERIC EDC MONITORING DATA3 FOR CALVERT CITY, KENTUCKY
Site
No.
1
2
3
4
5
6
7
8
9
10
11
12
Relation to
Goodrich Plant
0.8 km SE
1.8 km SW
1.7 km SSW
2.0 km SSE
3.3 km SE
2.9 km E
2.5 km ENE
3.4 km NE
2.3 km NE
2.8 km N
2.3 km NNW
3.0 km NW
Average
(ppb)
2.0
2.3
0.
0.
0.2
0.0
1.2
1.5
5.1
3.6
2.3
0.6
Average
(pg/m3)
8.0
9.3
0.5
2.8
0.6
0.1
4.8
6.2
20.6
14.6
9.4
2.3
Rangec
(pg/nr)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
- 37.7
- 72.2
- 4.1
- 18.0
- 3.2
- 0.5
-36.3
- 22.4
- 67.8
- 59.9
-55.0
-28.7
Source: Based on draft data supplied by PEDCo (1978).
Observations are for thirteen 24-hr periods between August 27, 1978, and
September 18,' 1978.
^When duplicate quality control samples were taken at one site, the average
of the two samples has been used.
25
-------
Table V-4
ATMOSPHERIC EDC MONITORING DATA* FOR LAKE CHARLES, LOUISIANA
Relation to
Conoco Plant
1.0 km S
0.7
1.2
0.7
km WNW
km WNW
km W
0.9 km SW
1.3 km WSW
3.0 km NW
2.8 km NNW
2.0 km NNW
1.5 km NNW
0.7 km NE
1.8 km ESE
Average
(ppb)
26.4
61.3
5.0
35.4
40,
11,
1.1
1.7
20.1
12.3
Average
(yg/m )
106.
248,
20,
143.4
162.7
45.4
4.5
4.0
6.5
6.7
81.4
49.9
Range
(ug/m )
1.4 - 269
6.0
0.0
651
67
1.8
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.5
- 744.8
- 383.3
- 171.6
- 27.3
- 32.8
- 30.2
- 36.2
- 581.6
- 497.8
Source: Based on draft data supplied by PEDCo (1978).
Observations are for twelve 24-hr periods between September 24, 1978, and
October 5, 1978.
"When duplicate quality control samples were taken at one site, the average
of the two samples has been used.
26
-------
Table V-5
ATMOSPHERIC EDC MONITORING DATA3 FOR NEW ORLEANS, LOUISIANA
Site
No.
1
2
3
4
5
6
7
8
9
10
11
12
Relation to
Shell
4.0 km NNW
4.0 km NW
3.0 km WNW
0.4 km SW
1.0 km NE
6.0 km WSW
3.0 km SW
2.0 km SW
1.5 km S
2.0 km SE
3.0 km SSE
14.0 km S
Union Carbide
6.0 km NNE
4.0 km NNW
3.0 km NNW
2.0 km NE
4.0 km NE
4.0 km NW
0.8 km NNW
1.5 km NE
2.0 km WNW
3.0 km WNW
3.0 km WSW
12.0 km S
Average
(ppb)
0.1
0.4
0.4
12.0
0.5
0.9
1.5
2.3
1.4
0.8
0.5
0.6
Average
(Mg/m3)
0.4
1.7
1.7
48.5
2.0
3.8
5.9
9.4
5.6
3.1
1.9
2.5
Range'3
(yg/m3)
0.0- 1.2
0.0- 6.2
0.0- 5.8
0.6-169.0
0.5- 9.1
0.0- 20.7
0.5- 24.3
0.5- 29.1
0.0- 17.6
0.0- 13.1
0.0- 8.7
0.0- 6.4
Source: Based on draft data supplied by PEDCo (1978).
Observations are for ten 24-hr periods between October 10, 1978, and
October 17, 1978.
^When duplicate quality control samples were taken at one site, the average
of the two samples has been used.
27
-------
available are insufficient for estimating population exposure for all
EDC producers. It is, therefore, necessary to use dispersion modeling
to estimate neighboring population exposures. In keeping with the
generalized nature of this study, approximate dispersion estimates were
made using rough-cut Gaussian-plume techniques. Centerline, ground
level, one-hour concentrations were calculated assuming a wind speed of
4 m/s and neutral ("D") stability. Based on engineering data
characteristic of the production facilities, a typical stack height of
25 m was used to assess point source emissions (oxychlorination process
vent and direct chlorination process vent). Fugitive and storage
emissions were treated as area sources. The production area was assumed
2
to be 0.01 km , and the area of storage tanks at production facilities
were taken from the following equation supplied by Mascone (1978):
Storage tank area (ft ) = ( /no. tanks - 1)60
total production capacity in 10 Ib/yr
where: No. tanks = c --rf ^ J
1.66 x 3
Table V-6 lists the results of the dispersion modeling, giving the
average 1-hr downwind concentrations. These 1-hr average concentrations
were adjusted to annual average omnidirectional concentrations by first
dividing them by 20 for conversion to maximum annual values, then
further dividing them by 2.5 to smooth the maximum annual values with
respect to direction. These factors were derived by Youngblood (1978)
and are based on empirical data from studies of industrial sources
similar to those modeled here. The atmospheric concentrations shown in
o
Table V-6 are in yg/m . These concentrations can be converted to
parts per billion (ppb) by multiplying by 0.244.
Exposure Estimates
We used the emission factors for the four sources of EDC emissions
(fugitive, storage, direct chlorination vent, and oxychlorination vent)
in Table V-2 to scale the generalized dispersion curves in Table V-6.
In this way, we estimated atmospheric EDC concentrations as a function
of distance from each plant, the atmospheric concentrations from the
28
-------
Table V-6
ESTIMATED ONE-HOUR AVERAGE DOWNWIND
ATMOSPHERIC CONCENTRATIONS OF EDC* (yg/m3)
Downwind Point Source Emitter with Emitter with
Distance (km) Emitterb 0.0625-km2 Areac 0.01-km2 Areac
0.30 3,400 4,000 10,000
0.45 4,800 3,400 7,700
0.60 4,400 2,900 5,700
0.75 3,700 2,500 4,300
1.00 2,700 2,000 2,900
1.25 2,100 1,600 2,200
1.60 1,500 1,200 1,600
2.50 800 720 810
4.00 410 380 410
6.00 230 220 220
9.00 120 120 120
14.00 66 64 66
20.00 39 39 39
aAssumes an emission rate of 100 g/s for each source, neutral ("D")
stability atmospheric conditions with a wind speed of 4 m/s.
bSingle stack 25 m high.
cEffective emission height of 10 m.
Source: Modeling data provided by P. Youngblood (EPA, 1978).
29
-------
four emission sources were summed at each downwind distance to give
total concentration. The point source emission concentrations of Table
V-6 were used for oxychlorination and direct chlorination sources; the
9
0.01 km emissions area concentrations were used for fugitive emis-
2
sions; and linear interpolation between the 0.01 and 0.0625 km area
concentrations were used for storage emissions, depending on the com-
puted storage area. The total annual average EDC concentration esti-
mates as a function of distance from each plant were used to determine
the radii at which the specified annual average concentrations (i.e.,
0.01, 0.3, 0.6, 1.0, 3.0, 6.0, and 10.0 ppb) are attained in the vicin-
ity of each plant.
The population residing within the radial distances to the concen-
trations specified above was estimated by SRI's computer system, BESTPOP
(Suta, 1978). The population file consists of a grid of 1-km square
sections that span the continental United States. This file was created
by assigning the 1960 and 1970 populations to the grid network and by
assuming uniform distribution of population within each of 256,000 enum-
eration districts. The computer software accesses the population file
and accumulates residential population within radial rings specified
about any given point. In addition, a rectangular map that is printed
out for an area around each specified point shows the population by
square kilometer.
We determined the latitude and longitude for each facility by con-
tacting the company directly, from regional planning groups, or from
other studies completed for EPA.
Table V-7 gives estimated population exposures to EDC from produc-
tion facilities. The number of people exposed to concentrations of less
than 0.1 ppb is underestimated because the dispersion modeling was not
extrapolated beyond 30 km from any plant. The larger EDC producers are
estimated to cause exposures of 0.01 to 0.1 ppb at distances beyond 30
km from their locations.
30
-------
Table V-7
ESTIMATED HUMAN POPULATION EXPOSURES
TO ATMOSPHERIC EDC EMITTED BY PRODUCERS
Annual Average
Atmospheric EDC Number of People
Concentration (ppb) Exposed
10.0 1,700
6.00 -10.00 3,300
3.00-5.99 28,000
1.00 - 2.99 280,000
0.60 - 0.99 400,000
0.30 - 0.59 1,500,000
0.10 - 0.29 4,300,000
0.060- 0.099 l,900,000a
0.030- 0.059 3,500,000a
0.010- 0.029 550,000a
Total 12,500,000
aThese are underestimates because the dispersion modeling results were not
extrapolated beyond 30 km from each EDC production facility.
31
-------
It is estimated that 12.5 million people are exposed to annual EDC
concentrations greater than 0.01 ppb from EDC producers. Approximately
6.5 million of these are exposed to concentrations greater than 0.1 ppb.
Comparison of Monitoring and Modeling Concentrations
Monitoring data were available for three locations having EDC pro-
duction plants (Calvert City, Lake Charles, and New Orleans). The moni-
toring data for each location are given in Tables V-3 through V-5. As
has been previously describved, the monitoring data were recorded as
24-hr samples taken for 12-13 days at each plant in the period of August
to October, 1978. Table V-8 averages these data for various distances
from each plant and also presents the average concentrations over the
three locations. The corresponding annual average dispersion modeling
concentrations are also given in Table V-8. Therefore, when comparing
the monitoring and modeling data, it is necessary to remember that the
two are not expected to agree precisely since the monitoring data are
site-specific and were recorded over a relatively short period of time
while the modeling data are based on general assumptions and are intend-
ed to represent annual average conditions. Thus, the degree to which
the annual and average monitoring concentrations are comparable depends
in part on the local meteorological conditions during sampling, the
placement of the monitoring stations, plant production during monitor-
ing, averaging times, and assumed source configurations. It is believed
that the monitoring data typify the year's average conditions.
Both monitoring and modeling data indicate that elevated concentra-
tions of EDC occur at distances of at least 14 km from the plants. Com-
parisons of the monitoring and modeling data at various distances from
the plants show that:
o The monitoring concentrations are approximately 20% higher
than the modeling ones for distances of less than 1 km.
o The modeling concentrations are 30-70% higher than those moni-
tored for distances of 1-4 km.
32
-------
o For distances of 4-14 km, both monitoring and modeling results
appear to be of about the same magnitude.
Thus, there is sufficient agreement between the monitoring and mod-
eling concentrations to conclude that the generalized modeling analysis
gives a reasonable first-cut estimate of ambient EDC concentrations near
production facilities allowing estimates of potential population expo-
sures.
33
-------
Table V-8
COMPARISON OF EDC MONITORING AND
MODELING ATMOSPHERIC CONCENTRATIONS (ppb)
Distance
(km)
0.7-1.0
1.1-1.5
1.6-2.0
2.1-3.0
3.1-4.0
4.1-5.0
5.1-6.0
14.0
Monitoring Average Concentrations
Calvert Lake New 3-Location
Cityk Charlesb Orleansb Average
2.0
c
1.0
2.1
0.9
c
c
c
36.7
6.0
7.0
1.1
c
c
c
c
6.3
1.4
1.6
0.7
0.3
c
0.9
0.6
15.0
3.7
3.1
1.3
0.6
c
0.9
0.6
3-Location
Modeling
Average3
13.5
9.1
6.5
4.3
2.1
1.6
1.2
0.4
aData are the average 24-hr concentrations over 10 to 13 days for monitoring
and estimated annual averages for modeling.
bThe EDC emissions have been estimated in Table V-2 as 72.9 g/s for B. F.
Goodrich, Calvert City, KY; 70.2 g/s for Conoco, Lake Charles, LA; and 119.5
g/s for Shell, New Orleans, LA.
clndicates that no monitoring data were collected.
34
-------
VI POPULATION EXPOSURES FROM PRODUCERS
THAT USE EDC AS A FEEDSTOCK
General
In estimating human population exposures to atmospheric EDC from
chemical production facilities that used EDC as a feedstock, products
considered include VCM, 1,1,1-TCE (or methyl chloroform), TCE, PCE, EA,
VDCM, and gasoline lead scavenger. Many of these chemicals are produced
at the same facilities that produce the EDC feedstock. Exposure
estimates are given for each product considered separately and also for
all products combined.
Sources of Emissions
Table IV-2 lists producers that use EDC feedstock and their
cpacities. The number of producers of each chemical that uses EDC as a
feedstock is as follows:
Chemical Producers Using EDC
VCM 14
1,1,1-TCE 3
TCE 4
PCE 4
EA 3
VDCM 3
Lead scavenger 6
See Table IV-2.
Table IV-4 indicates that, depending on the chemical, from 52 to 89% of
the production capacity was used in 1977 for the preceding chemicals.
35
-------
Because actual production data for each chemical at each plant are
unavailable, we have assumed that each operates at the percent of capac-
ities shown in Table IV-4.
Emissions
The VCM production process is well controlled to limit emissions.
These controls also reduce emissions from the EDC used as feedstock. It
is estimated that the EDC emission factor for VCM is 0.025% of EDC
input, the EDC emission factor for other processes that use EDC as a
feedstock is estimated as 0.2% of the EDC input (Mascone, 1978). The
estimated EDC emissions are given in Table VI-1. It is assumed that
these emissions are of a low level fugitive type resulting from leaks in
valves and other processing equipment and from storage-tank evaporation.
Atmospheric Concentrations
Because so few atmospheric monitoring data exist for the vicinities
of production plants that use EDC as a feedstock, it has been necessary
to use dispersion modeling to estimate neighborhood population
exposures. (Dispersion modeling is described in Section V.) The
2
dispersion estimates for a 0.01-km area source emitter (Table V-6)
was used for assessing exposure.
Exposure Estimates
The EDC emissions given in Table VI-1 were used to scale the
dispersion curve to estimate atmospheric EDC concentrations as a
function of distance from each plant for each product. Concentrations
were similarly estimated about each plant for emissions from all
products. The tables showing annual average atmospheric EDC
concentrations as a function of distance from each plant were used to
determine the radii at which specified annual average concentrations
(i.e., 1.0, 0.6, 0.3, 0.1, 0.06, 0.03, and 0.01 ppb) are attained. The
population residing within the distances to the concentrations specified
above was estimated by SRI's computer system, BESTPOP (Suta, 1978). We
determined the latitudes and longitudes for each facility by contacting
the company directly, by using information from regional planning
36
-------
Table VI-1
ESTIMATED EDC ATMOSPHERIC EMISSIONS (g/s) FOR
PLANTS THAT USE EDC AS A FEEDSTOCK
Plant"
Borden
Conoco
Diamond
Diamond
Dow
Dow
Dow
duPont
duPont
Ethyl
Ethyl
Goodrich
Houston
1C I America
Hal co
PPG
PPG
Shell
Shell
Stauffer
Union Carbide
Union Carbide
Vulcan
Location
Geismar, LA
Lake Charles, LA
Deer Park, TX
La Porte, TX
Freeport, TX
Oyster Creek, TX
Plaquemine, LA
Antioch, CA
Deepwater, NJ
Baton Rouge, LA
Houston, TX
Calvert City, KY
Beaumont, TX
Baton Rouge, LA
Freeport, TX
Lake Char lea, LA
Guayanilla, PR
Deer Park, TX
Norco, LA
Carson, CA
Taft, LA.
Texas City, TX
Geismar, LA
1,1,1-
VCM TCE TCE PCE EA
1.2
2.9
0.7 1.6
4.1
0.8 5.6 2.0 2.7
2.9
5.1 3.7
1.4 0.6 0.5
4.1
1.2
1.3 4.3 2.7 1.9
2.1
3.4
2.9
0.7
3.2
2.7
1.5
Lead
VDCM Scavenger Total
1.2
2.9
2.3
4.1
2.3 13.4
2.9
8.8
2.3 1.1 3.4
1.1 1.1
1.1 3.6
1.1 1.1
4.1
0.9 0.9
1.2
0.3 0.3
1.5 11.7
2.1
3.4
2.9
0.7
3.2
2.7
1.5
Total
34.1
13.6
6.0
5.5
8.6
6.1
5.6
79.5
Blanks indicate the chemical is not manufactured at the plant in question or that the plant has no EDC emissions.
Source: SRI estimates.
-------
groups, or from other studies completed for EPA. The population
exposures to EDC for individual products that require EDC as a feedstock
are given in Table VI-2.
VCM production results in the largest number of exposures about
1.3 million people about half of all the exposures for products using
EDC as a feedstock.
Table VI-3 gives the total exposures from all EDC feedstock
producers for all products. As shown in Table IV-2, many facilities
produce several products requiring EDC. Thus, two alternative estimates
of total exposures are given in Table VI-3: Alternative A combines all
the emissions reported in Table VI-1 and then uses these combined
emissions to estimate total exposures about each plant. Alternative B
is a summation of the individual product exposures shown in Table VI-2.
Because Alternative B counts some people twice (or more times), it
results in an overestimate of total exposures; however, the exposure
concentrations are lower than for Alternative A.
38
-------
Table VI-2
ESTIMATES OF POPULATION EXPOSURES TO ATMOSPHERIC EDC EMITTED
BY PLANTS THAT USE EDC AS A FEEDSTOCK IN VARIOUS PRODUCTS
Annual Average
Atmospheric EDC
Concentration (ppb)
0.600-0.999
0.300-0.599
0.100-0.299
0.060-0.099
0.030-0.059
0.010-0.029
Product
VCM
1,300
360
30 , 000
42,000
260,000
940,000
1,1,1-TCE
1,700
16,000
83,000
170,000
TCE
390
10,000
47 , 000
140,000
PCE
80
500
17,000
250,000
EA
70
17,000
8,000
43,000
37,000
VDCM
270
3,400
34,000
90,000
Lead
Scavenger
1,900
3,400
25,000
350,000
Total
1,300,000
260,000
200,000 270,000
110,000
130,000
380,000
-------
Table VI-3
ESTIMATES OF TOTAL POPULATION EXPOSURES TO ATMOSPHERIC EDC
EMITTED BY PLANTS THAT USE EDC AS A FEEDSTOCK
Annual Average Alternative A Alternative B
Atmospheric EDC Sum of Sum of
Concentration (ppb) Emissions3 Exposures"
0.600-0.999 1,300 1,300
0.300-0.599 2,100 430
0.100-0.299 110,000 51,000
0.060-0.099 210,000 83,000
0.030-0.059 520,000 510,000
0.010-0.029 1,500,000 2,000,000
Total 2,300,000 2,600,000
aExposures are based on the total feedstock emissions for each plant given
in Table V-l.
Exposures are based on the sum of the exposures for each product given in
Table V-2.
40
-------
VII POPULATION EXPOSURES FROM EDC IN AUTOMOBILE GASOLINE
General
Leaded gasoline contains EDC as a lead scavenger, with the amount
of EDC depending on lead content. Catalytic converters were first
required for the 1975 model year. In that year, 100% of Ford and GM
cars and 96% of Chrysler cars required unleaded gasoline. In 1976,
approximately 20% of the gasoline sold in the United Satats was unleaded
(Ethyl Corporation, 1976).
The antiknock "motor mix" added to gasoline is a combination of
ethylene dibromide (EDB), EDC, lead, and other alkyl groups. Quantities
of EDC and EDB are used sufficient to supply two atoms of chlorine and
one atom of bromine for each atom of lead. Major gasoline antiknock
mixes (which are added in small quantities to leaded gasoline) typically
contain 18.8% EDC by weight and 17.9% EDB by weight (SRI estimates).
*
Whereas the current average lead content in all gasoline is 1.5 g/gal,
leaded gasoline contains approximately 2.5 g/gal. According to EPA's
phase-down schedule for lead, the average lead content for all gasoline
is expected to be 0.5 g/gal by 1 October 1979 (Stolpman, personal
communication, 1977). Because of the low lead content in unleaded
gasoline (approximately 0.01 g/gal), EDC is not required and is not
added. Therefore, our analysis of population exposures related to
gasoline use considers only leaded gasoline. If we assume that gasoline
contains 0.425 units of EDC per unit of lead (by weight), this results
in an estimated 1.1 g of EDC per gallon of leaded gasoline, or 0.02% EDC
by volume.
We have used gallons rather than liters to represent gasoline volume
since these are the units commonly used in the United States for
gasoline sales.
41
-------
We evaluate nonoccupational population exposures to atmospheric
emissions of EDC from leaded gasoline for three sources:
o Exposures to people who refuel their automobiles at
self-service stations
o Exposures to people who reside in the vicinity of service
stations
o General urban population exposures from the evaporation of EDC
from automobiles.
Exposures from Self-Service Operations
Sources of Emissions
Service station types are characterized by the services they offer
and their business operations; they include (1) full-service stations,
(2) split-island stations, (3) self-service stations, and (4)
convenience store operations. In full-service stations (1), attendants
offer all services, including gasoline pumping and other mechanical
check-ups. If fuel is obtained at any class of stations (2) through
(4), the customers themselves may fill their tanks. In split-island
stations (2), both self-service and full-service are offered. At
stations (3) and (4), only self-service is available.
While pumping gasoline, an individual is exposed to EDC released as
vapor from the gasoline tank. Although occupants in the car at both
self-service and full-service operations are exposed to some EDC, the
highest exposures are to the person pumping the gas. Because it is
difficult to estimate level and length of exposure for car occupants,
only those who pump gasoline from self-service pumps are considered
Vapor recovery systems affixed to the gasoline nozzle can reduce
exposure levels significantly if they are working properly and are
operated correctly. Such systems are required for service stations in
parts of California.
42
-------
here. (Note it is not within the scope of this report to evaluate
occupational exposures.)
Self-service sale of gasoline is a relatively new marketing method
pioneered by independent operators in the West Coast and in the southern
United States. Today, it accounts for 30% of gasoline sold. The
national market share of the major gasoline producers has decreased
recently as independents and others specializing in high-volume,
low-margin sales capture a larger percentage. Of the approximately
184,000 conventional service stations with some self-service operations
account for 39% (Arthur D. Little, 1977). Table VII-1 indicates the
types of service stations offering self-service gasoline.
Table VII-1
SELF-SERVICE OPERATIONS
Outlets Offering Self-Service % of U.S. Total
Total self-service 9
Split island with self-service 26
Convenience stores 4
Total outlets with self-service 39
Source: Arthur D. Little (1977)
An Arthur D. Little report (1977) revealed that 71,300 outlets
offer self-service gasoline. Gasoline sold at U.S. service stations for
9
the year ending May 30, 1977 equalled approximately 87.4 x 10 gal.
Q
Of that amount, 27.0 x 10 gal (31%) is estimated to have been
dispensed at self-service pumps. The market share of self-service
stations was surveyed for four metropolitan Air Quality Control Regions
(AQCR): Boston, Dallas, Denver, and Los Angeles. The market share held
by self-service operations varied from 9% in Boston to 45% in Denver
(see Table VII-2). Another study by Applied Urbanetics, Inc. (1976)
surveyed Baltimore and Madison, Wisconsin. The results of that study
43
-------
Table VI1-2
GASOLINE MARKET SHARE OF SELF-SERVICE STATIONS
IN FOUR AQCRs SPRING 1977
Number of
Type of Operation Outlets
Boston AQCR
Full-service 2,253
Self-service (total) 100
Split island 8a
Self-service 92
Convenience stores
Sales Volume
(106 gal/yr)
1,045.1
108.6
Market
Sharing
Percent
91.0
9.0
Dallas AQCR
Full-service 2,094
Self-service (total) 1,124
Split island 480a
Self-service 444
Convenience stores 200
Deuver AQCR
Full-service 621b
Self-service (total) 656
Split island 310a
Self-service 226
Convenience stores 120
Los Angeles AQCR
Full-service 2,518
Self-service (total) 4,780
Split island 3,632a
Self-service 1,022
Convenience stores 126
924.6
593.8
292.1
235.7
2,472.6
2,154.8
61.0
39.0
55.0
45.0
53.0
47.0
aSplit-island operations offering full service and self-serve islands.
bOf these, 445 are split-island operations that offer full service and
mini-serve (attnedant-operated) islands.
Source: Arthur D. Little (1977).
44
-------
are shown in Table VII-3. It appears that self-service operations
account for about 40% of the market in urban areas.
Emissions
To estimate the people exposed to EDC from service stations,
several assumptions were necessary. The gasoline pumped through
self-service outlets is estimated at 27.0 x 10 gal/yr. The annual
average fuel consumption per vehicle is 736 gal (DOT, 1974a). If it is
assumed that on the average, a person who primarily uses self-service
gasoline makes one trip per week to the gasoline station, an average
fill-up amount of 14 gal is determined by dividing 736 gal/vehicle/yr by
52 wk/yr. By dividing the average fill-up into the self-service gallons
pumped, we estimate trips per year to self-service operations at 1.9 x
9
10 . When this number is divided by 52 trips per person per year, the
people exposed to pumping self-service gasoline is estimated at 37 x
10 . We can further assume that only 80% of these people are pumping
leaded gasoline containing EDC. Therefore, the people exposed from this
source is estimated to be 30 x 10 . For this estimate of the
population exposed, we assume that the individuals using self-service
gasoline obtain all of their gasoline at self-service stations.
Atmospheric Concentrations
A rough estimte of EDC exposures was made by extrapolating the
results of the Battelle (1977) benzene monitoring. In that study, three
samples of ambient air were taken in the breathing zone of persons
filling their tanks at self-service gasoline stations. The results,
shown in Table VII-4, indicate a wide range in the benzene
concentrations of the emissions. The variations seem to be related to
the subject's position in relation to the tank opening and the wind
direction. Because all measurements were taken on the same day and at
approximately the same time, ambient temperature did not cause the
variation. Basically, if the subject was downwind of the tank opening,
higher levels were recorded.
45
-------
Table VII-3
GASOLINE MARKET SHARE OF SELF-SERVICE STATIONS
IN TWO METROPOLITAN AREAS, 1976
Type of Operation
Baltimore SMSA
Full-service
Self-service (total)
Split island
Self-service
Madison SMSA
Full-service
Self-service (total)
Split island
Self-service
Sales Volume
(106 gal/yr)
90.5
25.5
65.0
56.03
77.0
17.0
60.0
Market
Sharing
Percent
55.0
45.0
42.0
58.0
alncludes the sales from mini-serve (attendant-operated) stat'ions and 50% of
the sales from split islands.
Source: Applied Urbanetics, Inc. (1976).
46
-------
Table VII-4
SAMPLING DATA FROM SELF-SERVICE GASOLINE PUMPING
Customer
1
2
3
Sampling
Rate
(mL/min)
31
31
31
Nozzle
Time
(min)
2.5
1.1
1.6
Gallons
Pumped
14
8
9
Sample
Volume
(L)
78
34
50
Benzene
Level
115
324
1,740
43
121
647
Source: Battelle (1977).
No EDC monitoring data obtained in the vicinity of gasoline
stations are available, therefore, by determining the evaporation rate
of EDC with respect to benzene, benzene monitoring data can be used to
provide a rough estimate of EDC exposures. It is known that the
evaporation rate is proportional to the vapor pressure, solubility, and
the square root of the molecular weight. Thus, the following equation
can be used to estimate the EDC emission factor (or emission rate)
related to evaporation:
PS /m~
e e e
where the subscript e refers to EDC and the subscript b refers to
benzene; E is the emission rate (or emission factor); P is the vapor
pressure; S is the solubility; and m is the molecular weight.
For an estimation, (s) (\Tm~) may be approximated by Xj the molar
fraction or concentration, thus, Equation (7.1) can be written as
follows:
47
-------
(7.2)
The Battelle benzene monitoring data were taken when the temperatures
was about 20°C. Because the vapor pressures for EDC (70 mm) and
benzene (80 mm) at 20°C are known, and the volume concentrations of
EDC (0.02%) and benzene (2.0%) in gasoline are also available, the
emission factor (or emission rate) of EDC can be estimated by the
following equations:
E = 0.009 E, . (7.4)
e b
This factor can be used to scale benzene atmospheric concentrations
o
( g/m ) to corresponding EDC concentrations becuse it is assumed that
atmospheric concentrations are proportional. the corresponding EDC
exposures were estimated, based on these data and are given in Table
VII-5.
Table VII-5
ESTIMATES OF EDC EXPOSURES FROM SELF-SERVICE GASOLINE PUMPING
Nozzle Gallons Estimated EDC Level
Customer Time (min) Pumped g/m^ppb
1 2.5 14 1.04 0.27
2 1.1 8 2.91 0.71
3 1.6 9 15.66 3.83
Average nozzle time =1.7 min
Time weighted average exposure = 1.45 ppb
Source: SRI estimates based on Battelle monitoring data (1977). the
conversion is based on Equation (7.4).
48
-------
Exposure Estimates
The estimated exposure levels are based on the information con-
tained in Table VII-5. It is recognized that these data are limited and
highly variable. However, they do allow a reasonable estimate of ex-
pected exposure levels from self-service gasoline pumping. In states
where vapor recovery systems are used, the estimated exposure level may
be much lower. Approximately 30 x 10 persons use self-service sta-
tions. While filling their tanks once a week, they are exposed to an
estimated EDC level of 1.5 ppb for 2.5 min (time required to pump 14
gal). Their annual exposure is estimated at 2.2 hr. This equates to an
annual averge time-weighted exposure of 0.0004 ppb.
Exposures in the Vicinity of Service Stations
Sources of Emissions
People residing in the vicinity of service stations may be exposed
to EDC from the evaporation of gasoline pumped by attendants and cus-
tomers, and from gasoline loaded by distribution trucks. These expo-
sures are in addition to those assessed previously for persons using
self-service gasoline stations. The amount of EDC emitted depends on
the ambient temperature, vapor recovery controls, the EDC content in
gasoline, and the volume of leaded-gasoline pumped. Approximately 80%
of the gasoline currently sold is leaded. With approximately 184,000
service stations in the United States, it is expected that many people
are exposed to EDC from those sources. Because most service stations
are located in urban areas and because their location is expected to be
highly correlated with urban population density, only urban areas are
considered in this analysis.
Emissions
An EDB emission factor of 0.00157 g of EDB per gram of lead per
gallon from refueling losses has been estimated, based on testing at
EPA1s Mobile Source Air Pollution Control Laboratory in Ann Arbor,
Michigan (Kittredge, 1977). The factor considers spilling, vapor dis-
placement, entrained liquid gasoline losses, and volume of gasoline
pumped. Assuming an average lead content in gasoline of 2.5 g/gal, the
49
-------
estimated emission factor for EDB is 0.00039 g/gal. The EDB emission
factor can be used to estimate the EDC emission factor through the use
of Equation (7.2) (and by substitution of EDB factors for benzene fac-
tors in the equation). We have assumed that the EDB vapor pressure is
12 mm at 25°C and that it constitutes 0.05% of the gasoline (by vol-
ume). Hence, we estimte that the EDC emission factor for automotive
refueling losses (E ) is:
e
E " T? x7T7?t x 0.00039 = 0.001 g/gal. (7.5)
e \-2. U. ID
The number of service stations in urban areas can be estimated,
based on urban service station density and total U.S. urban population.
Service station density in urban areas can be extrapolated from the data
presented in Table VII-6. The service station density shown for four
metropolitan AQCRs varies, with no regional pattern evident. Based on
these data, we estimate an average of-0.7 service stations per 1,000
population. This number can be applied generally to urban areas
throughout the United States. Urbanized areas provide the best popu-
lation base. The 1970 population residing in urbanized areas was
118,447,000 (Bureau of the Census, 1975). Thus, service stations in
urbanized areas are estimated at 83,000, or 45% of all stations.
An emission rate can be estimated by employing the following
assumptions:
(1) 70.0 x 10' gal of leaded gasoline are sold annually by
service stations.
(2) The average number of gallons pumped per service station is
3.8 x 105 gal. (The United States has approximately 184,000
service stations.)
. _.
Defined by the Bureau of Census as the central city or cities and
surrounding closely settled territories. All sparsely settled areas in
large incorporated cities are excluded by this definition. Densely
populated suburban areas, however, are included (U.S. Department of
Commerce, Bureau of the Census, 1972 County and City Data Book).
50
-------
Table VII-6
SERVICE STATION DENSITY IN FOUR METROPOLITAN AQCRs
Number ofa AQCR^ Service
Service Stations Population Stations0 per
AQCR (1977) (1975) 1,000 Population
Boston 2,353 4,039,800
Dallas 3,218 2,970,900
Denver 1,277 1,389,000
Los Angeles 7,298 14,072,400
0.6
1.1
0.9
0.5
Sources:
*A. D. Little (1977).
^Bureau of Economic Analysis (1973)
CSRI estimates.
51
-------
(3) All service stations have uniform pumping volumes.
(4) The EDC emission rate for a service station is:
(Vol. of gasoline pumped)(emission factor) = emission rate;
that is, (3.80 x 105 gal/yr)(0.001 g/gal) = 380 g/yr
= 1.2 x 10~5 g/s.
Atmospheric Concentrations
Dispersion modeling of benzene emissions from gasoline service
stations (Youngblood, 1977) employed the Single Source (CRSTER) Model
(USEPA, 1977). Meteorological data for Denver, Colorado were used to
represent a reasonable worst-case location. EDC emissions from gasoline
service stations are thought to be from sources similar to those for
benzene. The dispersion modeling assumed that the sources of emissions
are dispersed over a 50 sq ft area. The benzene concentrations of the
previous study were modified to reflect EDC emissions by first
multiplying the benzene concentrations (ppb) by 3.2 to convert them to
,g/m3
EDC.
3 3
/ig/m and then by multiplying the /Ug/m by 0.244 to convert to ppb of
Table VII-7 presents the results of the benzene modeling modified
for EDC; an EDC emission of 0.01 g/s is assumed. Two conditions are
given: (1) the 8-hr worst-case concentrations for a service station
that operates only during daytime, and (2) the annual average
concentrations for a service station that operates 24 hr/day.
Exposure Estimates
Population exposures to EDC emissions from gasoline service
stations have been estimated by assuming that the population in
urbanized areas is uniformly distributed with a density of 1,318/km2
(based on the 1970 census). We have also assumed that no one resides
within 50 m of a gasoline service station. To calculate atmospheric
concentrations as a function of distance, the annual average dispersion
modeling data in Table VII-7 were scaled by the estimated EDC emissions
52
-------
Table VII-7
ROUGH DISPERSION MODELING RESULTS FOR EDO EMISSIONS
FOR GASOLINE SERVICE STATIONS3
Distance (m) 8-hr Worst Case (ppb)k Annual Average (ppb)c
50 12 1.0
100 6 0.5
150 3 0.3
200 2 0.2
300 1 0.1
aAssumes an EDC emission of 0.01 g/s during operation.
"Assumes continuous operation from 8 a.m. to 4 p.m., 6 days per week.
cAssumes continuous operation 24 hr per day, 7 days per week.
Source: Modified from Youngblood (1977) by adjusting ppb of benzene to ppb of
EDC.
53
-------
(ratio of estimated emissions to emissions on which the dispersion curve
is based. Two cases were considered: (1) the 83,000 urban service
stations were distributed by assuming that no two are closer than 300 m
to each other and that no people reside closer than 50 m, and (2) the
urban service stations were geographically distributed so that three are
always located in close proximity. Thus, for the 27,633 triplets we
have assumed that no people reside closer than 100 m. The EDC emission
rate for the single service station case is taken as 1.2 x 10 g/s
_ c
and for triplet service station case as 3.6 x 10 g/s. The actual
geographic distribution of urban service stations is assumed to be
somewhere in between these two cases; therefore, the exposures estimated
for these two cases are expected to bound actual exposures.
Because of the population exclusion radii (50 and 100 m) and
assumed emissions, no exposures are estimated to occur for EDC annual
average concentrations greater than 0.03 ppb. In some cases, people may
reside closer to service stations than these exclusion radii permit. In
these cases, some would be exposed to atmospheric concentrations in
excess of 0.03 ppb. It is estimated that 600,000 people are exposed in
the single service station case and that 1.4 million are exposed in the
triplet service station case. All of these exposures are estimated to
be in the 0.01 to 0.03 ppb annual average concentration range.
These estimates, which are only rough approximations, are based on
assumptions of uniform distribution of service stations in urbanized
areas, uniform pumping volumes, average populaton density, and on
dispersion modeling. In reality, more service stations are located in
commercial areas than in residential areas, and pumping volumes vary
substantially.
Urban Exposures Related to Automobile Emissions
Sources of Emissions
Urban exposures to EDC come from many sources, including gasoline
evaporation, gasoline service stations, losses through transportation
and storage of gasoline, and emissions from production facilities. Most
54
-------
of these sources have been treated as point sources and their emissions
are evaluated elsewhere in this report. This section presents analysis
of exposures due to emissions of EDC from automobiles.
Emissions
As previously discussed, the EDC content in leaded gasoline
averages 0.02% by volume, and leaded gasoline accounts for 80% of all
gasoline sold. Tests by EPA's Mobile Source Air Pollution Control
Laboratory in Ann Arbor, Michigan, have indicated that EDB is destroyed
in the combustion process (Kittredge, personal communications, 1977),
and it has been calculated that EDC is similarly destroyed (Mabey,
1978). However, evaporation from the carbureator and from the fuel tank
does occur. We have been unable to locate data on EDC emissions from
this type of evaporation; however, data are available for tests
measuring EDB (Table VII-8).
Table VII-8
AUTOMOTIVE EDB EMISSION FACTORS
(G/G OF LEAD PER GALLON OF GASOLINE)
Vehicle Type Low High
Uncontrolled vehicle (pre-1972) 0.00144 0.00362
Pre-1978 controlled vehicle 0.00098 0.00250
Post-1978 controlled vehicle 0.00033 0.00085
Source: Kittredge, 1977
The average EDB emission factor for uncontrolled vehicles, based on
Table VII-8, is 0.00253 g of EDB per gram of lead per gallon. Assuming
2.5 g of lead per gallon of leaded gasoline, the estimated EDB emission
factor is 0.0063 g of EDB per gallon. The EDB emission factor can be
used to estimate the EDC emission factor through the use of Equation
(7.2). By substitution in this equation, the EDC emission factor (E )
for uncontrolled automobile evaporation becomes:
55
-------
on n no
Ee = 12 X 0"1J5 X °-0063 = °-017 § of EDC/gal (7.6)
This factor will provide a slightly high estimate of ambient EDC levels
because it assumes all automobiles using leaded gasoline have emissions
comparable to pre-1972 models. In fact, the 1975 model year was the
first in which automobiles were required to run on unleaded gasoline,
but controls to reduce carbureator evaporation were introduced in the
1972 model year.
Dispersion Modeling
The Hanna-Gifford area-wide dispersion model (Gifford and Hanna,
1973) as applied by Schewe (1977) for benzene is used for this analysis
and modified for EDC. Mara and Lee (1978) contains a discussion of this
model and its application to benzene. Because EDC is destroyed during
combustion (Mabey, 1978), only evaporation is considered. The modified
equation to estimate the emission rate for EDC is as follows:
0 = (0 017 e/eaT) / annual travel miles/vehicle \
eva? gS } \ average miles/gal )
(vehicles registered) - r-j -j\
area \' ')
If 12,000 mi/yr for each vehicle and 12 mi/gal are assumed (DOT,
1974b), the above equation becomes
QevaP - (5.4 x lO"7 g/s) x ( vehicles^eglstered J (?
-------
where u is wind speed (m/s) and \ is the atmospheric concentration in
3
g/m . The average annual wind speed, u, in the area of study was
obtained from AP101 (EPA, 1972). Because wind speed (and thus
dispersion) increases in the afternoon, the morning values were used to
estimate higher concentrations. the number 225 is an empirical factor
derived from several studies that give very good results for long-term
averages for low-level emission sources (Gifford and Hanna, 1973).
Estimates of Exposures
Cities whose vehicular densities are higher should have the higher
EDC concentration from automobile evaporation. Because 916 cities in
the United States have populations greater than 25,000, we used a
statistical sampling approach to evaluate EDC exposures. Table VII-9
shows the distribution of cities by size. Because of the expected
higher EDC concentratons in the larger cities, it was decided to
evaluate the exposures for the 26 cities with populations greater than
500,000 and to do a fractional sample of the cities in the smaller
groups. However, because vehicle registration data were unavailable for
Boston and New Orleans, only 24 of the largest 26 cities were
evaluated. Of the cities in the 250,000-500,000 size range, 25% were
Table VII-9
DISTRIBUTION OF CITIES BY 1970 POPULATION
Number of Combined
Population Size Cities Population
1,000,000 6 18,769,000
500,000-1,000,000 20 12,967,000
250,000- 500,000 30 10,442,000
100,000- 250,000 100 14,286,000
50,000- 100,000 240 16,724,000
25,000- 50,000 520 17,848,000
Source: U.S. Bureau of the Census, Statistical Abstracts of the United
States-1974.
57
-------
sampled, as were 15% of the cities in the 100,000 to 250,000 size
range. No EDC exposures greater than 0.01 ppb were estimated for cities
with population less than 100,000. Exposures were estimated for people
within the sample, and these -sampling results were then projected to all
cities, based on the ratio of total population to sample population.
Table VII-10 sets forth the exposure estimates for the cities
sampled. No exposures were found to exceed 0.03 ppb. Table VII-11
gives the calculations used to project the sampling data to the total
population. Based on this projection we estimate that approximately 13
million people are exposed to annual average EDC concentrations of 0.01
to 0.03 ppb from the evaporation of gasoline from automobiles.
Summary of Urban Exposures from Automobile Gasoline
Exposures to EDC from leaded gasoline have been estimated for
people refueling their automobiles at self-service stations, for those
residing near service stations, and for those exposed to EDC evaporation
from automobiles. We estimate that approximately 30 million people are
exposed to an EDC concentraton of 1.5 ppb for 2.2 hr/yr while refueling
their automobiles. Approximately 600,000 to 1,400,000 (average of 1
million) people residing near gasoline service stations are exposed to
annual average EDC concentraions of 0.01 to 0.03 ppb from refueling
losses. Another 13 million people are exposed to annual average EDC
concentrations of 0.01 to 0.03 ppb from automobile evaporation.
58
-------
Table VII-10
ESTIMATED U.S. CITY EXPOSURES TO EDC FROM THE EVAPORATION OF AUTOMOBILE GASOLINE
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
17
18
20
21
22
23
24
25
26
27
32
37
42
47
52
57
62
72
82
87
92
97
112
117
122
127
132
137
142
151
157
162
Name
New York, NY
Chicago, IL
Los Angeles, CA
Philadelphia, PA
Detroit, MI
Houston, TX
Baltimore, MD
Dallas, TX
Washington, DC
Cleveland, OH
Indianapolis, IN
Milwaukee, WI
San Francisco, CA
San Diego, CA
San Antonio, TX
Memphis, TN
St. Louis, MO
Phoenix , AZ
Columbus, OH
Seattle, WA
Jacksonville, FL
Pittsburgh, PA
Denver, CO
Kansas City, KA
Atlanta, GA
Minneapolis, MN
Oklahoma City, OK
Miami, FL
Norfolk, VA
Akron, OH
Richmond, VA
Corpus Christi, TX
Ft. Wayne, IN
Fresno, CA
Santa Ana, CA
Lubbock, TX
Riverside, CA
Peoria, IL
Ma con, GA
S avannah , GA
Columbia, SC
Alexandria, VA
Allentown, PA
Hollywood, FL
Duluth, MH
Pueblo, CO
Sunnyvale, CA
Population"
(1,000)
7,895
3,363
2,816
1,949
1,511
1,232
906
844
757
751
746
717
716
697
654
624
622
582
539
531
529
520
515
507
497
434
367
335
308
275
250
204
178
166
157
149
140
127
122
118
114
111
110
107
101
97
95
Automobiles'*
(1,000)
1,707
1,476
1,515
954
796
692
412
741
391
392
237
328
367
388
333
306
295
357
334
273
355
252
331
264
354
237
238
221
144
153
132
102
112
86
98
80
73
66
72
63
63
57
58
82
45
50
55
Qevapc
(10-10 g/s-n^)
11.9
13.8
6.8
15.5
12.0
3.3
11.0
5.8
13.3
10.8
1.9
7.2
16.9
2.6
3.8
2.9
10.1
3.0
5.2
6.8
1.0
9.5
7.3
1.7
5.6
9.0
0.8
13.4
5.7
5.9
4.6
2.1
4.5
4.3
7.6
2.2
2.1
3.7
3.1
4.9
1.2
8.1
6.8
6.8
1.4
4.6
5.4
Wind Speed
(m/s)
7
5
3
6
6
6
6
6
5
5
5
5
3
3
6
5
6
4
5
5
6
5
4
6
5
6
6
6
7
6
5
6
5
3
3
6
3
5
5
5
5
5
3
5
6
4
3
Concentration
(ppb)
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
a 1970 census city population.
Registered automobiles by SMSA as given in DOT (1974b) have been assigned to cities, based on the ratio of city
population to SMSA population.
c See Equation (7.8).
59
-------
Table VII-11
CALCULATIONS OF NATIONAL EXPOSURES* TO EDC
FROM AUTOMOBILE EVAPORATION
City Population
1,000,000
500,000-1,000,000
250,000- 500,000
100,000- 250,000
Total U.S.
Population
18,769,000
12,967,000
10,442,000
14,286,000
Projected
Fraction Sampled Population
Population Exposed to Exposures
Sampled 0.01 ppb 0.01 ppb
18,769,000
11,733,000
2,670,000
1,892,000
0.43
0.19
0.13
0.08
8,130,000
2,460,000
1,310,000
1,140,000
Total
13,040,000
All exposures are in the 0.01 to 0.03 ppb range.
60
-------
VIII OTHER ATMOSPHERIC EXPOSURE ROUTES
General
Environmental exposures to EDC may occur through any of its
dispersive uses, including grain fumigants, paints, coating, adhesives,
cleaning, and the prepartion of polysulfide compounds. Additional
environmental exposures may occur from spills and venting during
transportation and from evaporation at waste disposal operations. Few
data are available on the amount of EDC used or lost to the environment
from these potential exposure routes; consequently, only qualitative
descriptions of exposures can be given.
Dispersive Uses
Auerback Associates, Incorporated (1978) estimated consumption of
EDC for minor uses in 1977 at about 5,000 mt. Of this total, about 28%
(1,400 mt) was used in the manufacture of paints, coatings, and
adhesives. Extracting oil from seeds, treating animal fats, and
processing pharmaceutical products required 23% of the total (1,150
mt). An additional 19% (950 mt)-was consumed cleaning textile products
and PVC manufacturing equipment. Nearly 11% (550 mt) was used in the
preparation of polysulfide compounds. Grain fumigation required about
10% (500 mt). The remaining 9% (450 mt) was used as a carrier for
amines in leaching copper ores, in the manufacture of color film, as a
diluent for pesticides and herbicides, and for other miscellaneous
purposes. It is generally assumed that all of this material is
eventually released to the atmosphere (Drury and Mammons, 1978).
Atmospheric exposures to EDC from these dispersive uses occur as
point source losses from the industrial sites where these products are
manufactured and from the use of end products.
61
-------
Atmospheric dispersion modeling was used to provide crude
estimates of magnitude of exposures that might be attained from the
manufacture of these other products. Use of the dispersion modeling has
shown that a plant would have to use more than 90 mt/yr of EDC to have
concentrations in excess of 0.01 ppb at the assumed plant boundary 0.5
km from the plant center. A plant that uses 1,000 mt/yr of EDC would
have concentrations of 0.1 ppb at the plant boundary (0.5 km) and 0.01
ppb at 2.5 km from the emissions. Production data are not available for
these dispersive uses. Since a total of 5,000 mt/yr are involved in all
dispersive uses, it is highly unlikely that any one production plant
would use as much as 1,000 mt/yr of EDC. Based on these preliminary
calculations, it appears that the EDC exposures to people residing near
these other manufacturing plants are minimal.
\
Nonoccupational exposures from end product use would occur
primarily from the use of paints, coatings, adhesives, and solvents, and
to people who inadvertently enter fumigated areas. All of these
exposures would be intermittent and are extremely difficult to estimate.
Transportation
EDC may be emitted to the atmosphere during transportation from
inadvertent spills and from venting. The amount of EDC transported each
year is not well known because companies transport the chemical between
their plants as well as to other companies or to places for export to
other countries. We estimate that at least 672,000 mt of EDC were
transported during 1977 (approximately 13% of production). This
includes 5,000 mt required for minor dispersive uses, 177,000 mt for
exportation, and an estimated 490,000 mt transported within the United
States for use in the major products shown in Table IV-2. The estimated
*
For the modeling, it has been assumed that EDC emissions are 1% of
input and that the emissions occur over a 0.01 km2 area (see Table
V-6).
62
-------
490,000 mt transported for major product use was obtained as the sum of
shortages between EDC production and EDC required at individual plants.
We assumed that each plants' production, for each product, was at the
total capacities given in Table IV-2 times the percent capacities used
during 1977 (Table IV-4). However, a chemical plant is flexible in
regard to the products it actually makes, and this factor could result
in considerably more EDC actually being transported than estimated.
The major transportation emissions would probably result from
venting and spillage during loading and unloading transportation
containers. Thus, these emissions would occur at the processing plants
and would add to the EDC emissions from other plant activities. Other
emissions could result from venting while in transport and from
accidental rupture of a container.
No data are available on EDC emissions during loading and
unloading; however, Mara and Lee (1978) give uncontrolled emission
factors for the transfer of benzene (Table VIII-1) by inland barge, tank
truck, and rail car. The benzene emission factors can be adjusted to
rough order EDC emission factors by adjusting for the differences in
vapor pressure through the use of Equation (7.2). This adjustment gives
an uncontrolled EDC emission factor of approximately 10 g/gal or 0.18%
by weight. This must be regarded as an upper limit because some of the
Table VIII-1
SUMMARY OF UNCONTROLLED EMISSION FACTORS
FOR THE TRANSFER OF BENZENE
Benzene Emission
Operation Factor (g/gal)
Inland barge 0.76
Tank truck 1.8
Rail car 1.8
Average 1.45
Source: Compiled by Mara and Lee (1978)
63
-------
transfer areas have controls. If we assume that 672,000 mt of EDC are
transported each year, that emissions might occur during loading and
unloading, and that the emission factor is less than 0.18%, an extreme
upper emission estimate of 2,400 mt/yr results.
Waste Disposal
EDC wastes may be generated in any process in which the chemical is
involved. The largest quantities of EDC wastes occur during the
synthesis of the compound and in the production of VCM, processes that
involve all of the EDC production and most of its consumption. The
treatment of wastes is of concern in estimating atmospheric emissions
because of EDC's moderately high volatility (see Section III).
Liquid wastes result from scrubbing vented gases or crude EDC with
water or caustic solutions. The treatment of these wastes varies from
plant to plant, but usually this wastewater is used for pH control in
other processing areas or is sent off site for final processing (EPA,
1974).
Solid wastes are usually disposed of by burial in a landfill or by
incineration (Patterson, 1975). Solid wastes also occur in EDC
manufacturing plants that use fluidized bed, rather than fixed bed,
reactors (EPA, 1974). These solid wastes are periodically removed from
the rejected water settling ponds and transported to landfills.
Monsanto Research Corporation (1975) has estimated EDC emission
factors emitted to solid waste and water for EDC formulation. We have
used Monsanto's conclusions, shown in Table VIII-2, to estimate that
during 1977, 10,600 mt were discharged to solid wastes and that 18,500
mt were discharged to water. Additional solid waste and water
discharges occur as a result of production of chemicals that used EDC as
a feedstock.
64
-------
Table VIII-2
ESTIMATED 1977 EDC EMISSIONS AS SOLID WASTE
AND TO WATER FROM EDC PRODUCTION
Solid Waste Water
Emission Factor3 (kg/mt)
Direct chlorination 1.5 2.9
Oxychlorination 2.8 4.6
Emissions** (1,000 mt/yr)
Direct chlorination 4.5 8.5
Oxychlorination 6.1 10.0
Total Emissions 10.6 18.5
aBased on Monsanto (1975).
^Assumes 58% direct chlorination and 42% Oxychlorination (Patterson,
1976) and an EDC production of 5,194,000 mt/yr.
Estimating atmospheric exposures from the solid waste and water EDC
emissions would require:
(1) An estimate of the rate of return to the atmosphere
(2) Identification of the contaminated sites, the amount deposited
at each site, and the site's location in respect to population
(3) A method for transforming emission estimates to estimates of
atmospheric exposure.
Currently, the data available on items (1) and (2) are insufficient for
estimating exposures. Table II-2 indicates that an estimated 52,000 mt
of EDC is emitted directly to the atmosphere annually from sources other
than transportation and waste disposal. If all the EDC solid waste and
water emissons evaporate to the atmosphere, an additional 29,100 mt/yr
of atmospheric emissions would result. This estimate, however, is
uncertain since the rate of evaporation from landfills and water
emissions cannot be adequately assessed.
65
-------
BIBLIOGRAPHY
Altshuller, A. P., "Lifetime of Organic Molecules in the Troposphere and
Lower Stratosphere," Environ. Sci. Tech. (to be published).
Applied Urbanetics, Inc., "Market Share Study," FEA Contract No.
CO-06-60435 (1976).
Arthur D. Little, Inc., "Self-Serve Market Shares in Four Metropolitan
Areas," memo to Richard J. Johnson, EPA, from E. Quakenbush and P. E.
Mawn (June 1977).
Auerbach Associates, "Miscellaneous and Small-Volume Consumption of
Ethylene Bichloride," unpublished report prepared for EPA under
Contract EPA-68-01-3899 and Auerbach Associates, Inc.,
AA1-2431-104-TN-1 (1978).
Battelie-Columbus Laboratories, letter to Richard J. Johnson, EPA, from
C. W. Townley concerning "Results of Self-Service Exposure Samples"
(May 1977).
Bureau of the Census, Statistical Abstract of the United States (1975).
Bureau of Economic Analysis, "Projections of Economic Activity for Air
Quality Control Regulations," NTIS PB-259-870 (1973).
Chemical Marketing Reporter, "Profile: Ethylene Dichloride," 212(3):9
(1977).
66
-------
Billing, W., N. Tefertiller, and G. Kallos, "Evaporaton Rates and
Reactivities of Methylene Chloride, Chloroform,
1,1,1-Trichloroethane, Trichloroethylene, Tetrachloroethylene, and
Other Chlorinated Compounds in Dilute Aqueous Solutions," Environ.
Sci. Technol. 9;833-837 (1975).
Drury, J. S., and A. S. Hammons, "Investigations of Selected
Environmental Pollutants: 1,2-Dichloroethane," Oak Ridge National
Laboratory, Draft Report prepared for the Office of Toxic Substances,
EPA (1978).
Ethyl Corporation, "Yearly Report of Gasoline Sales by States" (1976).
Federal Register, "Consolidation of Hazardous Material Regulations,"
41(74):15972-15990 (1976).
Gifford, F. A., and S. R. Hanna, "Technical Note: Monitoring Urban Air
Pollution," in Atmospheric Environment, Vol. 7, Pergamon Press (1973).
Grimsrud, E., and R. Rasmussen, "Survey and Analysis of Halocarbons in
the Atmosphere by Gas Chromatography - Mass Spectrometry," Atmos.
Environ. (England), ^:1014-1017 (1975).
Hanst, P., "Noxious Trace Gases in the Air," Chemistry 51(2):6-12 (1978),
Hardie, D., "Chlorocarbons and Chlorohydrocarbons," in Kirk-Othmer
Encyclopedia of Chemical Technology, 2nd ed., Vol. 5, pp. 171-178,
Interscience, New York (1964).
Hawley, G.G. (ed.), The Condensed Chemical Dictionary, 9th ed., Van
Nostrand Reinhold Company, New York (1977).
Horvath, R., "Microbial Co-metabolism and the Degradation of Organic
Compounds in Nature," Bacteriol. Rev. 36(2);146-155 (1972).
67
-------
Johns, R., "Air Pollution Assessment of Ethylene Bichloride," MTR-7164,
The Mitre Corporation (1976).
Kittredge, G. D., memo to files concerning "Up-to-date Estimate of
Automotive Emission Factors," (26 September 1977).
Mabey, W. R., Physical Organic Chemistry Laboratory, SRI International,
personal communications (December 1978).
Mara, S. J., and S. S. Lee, "Assessment of Human Exposures to
Atmospheric Benzene," SRI International (1978).
Mascone, D., EPA, personnel communications (16 November 1978).
McConnell, G., D. Ferguson, and C. Pearson, "Chlorinated Hyrdocarbons
and the Environment," Endeavor 34;13-18 (1975).
Mitten, M., K. Dress, W. Krochta, F. Ewald, and D. DeWitt,
"Chlorocarbons," in Encyclopedia of Industrial Chemical Analysis, Vol.
9, F. D. Snell and L. S. Ettre, eds., pp. 437-510, Interscience, New
York (1970).
Monsanto Research Corporaton, "Potential Pollutants from Petrochemical
Processes," as cited in Drury and Hammons (q.v.).
National Institute for Occupational Safety and Health, "Criteria for a
Recommended Standard Occupational Exposure to Ethylene Dichloride,"
NIOSH-76-139 (1976).
, Registry of Toxic Effects of Chemical Substances, Vol. II,
p.388 (1977).
_, "Current Intelligence Bulletin #25: Ethylene Dichloride
(1,2-dichloroethane)" (19 April 1978).
68
-------
Patterson, R. M., M. I. Bernstein, and E. Garshick, "Assessment of
Ethylene Dichloride as a Potential Air Pollution Problem," GCA
Corporation (1976).
Pearson, C., and G. McConnell, "Chlorinated C and C~ Hydrocarbons
in the Marine Environment," Proc. R. Soc., London, Ser. B,
189:305-332 (1975).
PEDCo, "Draft of Preliminary Monitoring Data," collected for EPA (1978).
Pellizzari, E. D., "Electron Capture Detection in Gas Chromatography,"
J. Chromat. 99;3-12 (1974).
Radding, S., D. Liu, H. Johnson, and T. Mill, "Review of the
Environmental Fate of Selected Chemicals, SRI International,
EPA-560/5-77-003, Office of Toxic Substances, EPA (1977).
Rothon, R. N., "Petroleum and Organic Chemicals," in Chemical
Technology; An Encyclopedic Treatment, Vol. 4, Barnes and Noble, New
York (1972).
Schewe, G. J., EPA, memos concerning "Estimates of the Impact of Benzene
from Automotive Sources" to R. J. Johnson (20 June, 9 August, 12
August 1977).
Singh, H., L. Salas, and L. Cavanagh, "Distribution, Sources, and Sinks
of Atmospheric Halogenated Compounds," J. Air Pollut. Control Assoc.
27(4);332-336 (1977).
Snelson, A., R. Butler, and F. Jarke, "Study of Removal Processes for
Halogenated Air Pollutants," U.S. Environmental Protection Agency,
EPA-600/3-78-058 (1978).
Spense, J., and P. Hanst, "Oxidation of Chlorinated Ethanes," J. Air
Pollut. Control Assoc. 28(3);250-253 (1978).
69
-------
Stolpman, P., EPA, personal communication (October 1977).
Storck, W., "Big Chemical Producers Post Moderate Growth," Chem Eng.
News 56(18);31-37 (1978).
Suta, B. E., "BESTPOP: A Fine-Grained Computer System for the Assessment
of Residential Population," SRI International (1978).
Toxic Materials News, "NCI Finds Ethylene Bichloride to be Carcinogenic"
(27 September 1978).
U.S. Environmental Protection Agency, "Mixing Heights, Wind Speeds, and
Potential for Urban Air Pollution Throughout the Contiguous United
States," in Publication No. AP-101 (1972).
, "Engineering and Cost Study of Air Polluton Control for the
Petrochemical Industry," Vol. 3, "Ethylene Bichloride Manufactured by
Oxychlorination," EPA-450/3-73-006-C (1974).
, "Report on the Problem of Halogenated Air Pollutants and
Stratospheric Ozone," EPA-600/9-75-008 (1975).
. , "Users' Manual for Single-Source (CRSTER) Model,"
EPA-450/2-77-013 (1977).
draft material relating to human exposure to atmospheric
ethylene dichloride near production facilities (1978).
u
U.S. Bepartment of Transportation, Federal Highway Administration,
"Highway Statistics" (1974a).
_, "Motor Vehicle Registrations by Standard Metropolitan
Statistical Areas," Table MV-21 (1974b).
70
-------
U.S. International Trade Commission, "Synthetic Organic Chemicals, U.S.
Production and Sales" (1973-1977).
Whitney, W., "Fumigation Hazards as Related to the Physical, Chemical,
and Biological Properties of Fumigants," Pest Control 29(7):16-21
(1961).
Windholtz, M., The Merck Index, 9th ed., Merck and Company, Rahway, New
Jersey (1976).
Youngblood, P. L., EPA, memo concerning "Use of Dispersion Calculations
in Determining Population Exposures to Benzene from Chemical Plants"
to R. Johnson (20 September 1977).
Youngblood, P. L., "Dispersion Modeling for Determining Population
Exposure to Benzene," EPA memo to R. Johnson (4 January 1978).
71
-------
ASSESSMENT OF HUMAN
EXPOSURES TO ATMOSPHERIC
PERCHLOROETHYLENE
Final Report
January 1 979
By:
Susan J. Mara
Benjamin E. Suta
Shonh S. Lae
Prepared for:
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Researcn Triangle Park, North Carolina 27711
Task Officer: JacK K. Greer, Jr.
Project Officer: Joseph D. Cirveilo
Contract No. 63-02-2835
SRI Project CRU-S7SO
Center for Resource and Environmental Systems Studies
Report No. 73
-------
NOTICE
This report has been provided Co the U.S. Environmental Protection
Agency (EPA) by SRI International, Menlo Park, California, in
fulfillment of Contract 68-02-2835.- The opinions, findings, and
conclusions expressed herein are those of the authors and are not
necessarily those of EPA. Mention of company or product- names is not to
be considered an endorsement by EPA.
-------
CONTENTS
LIST OF ILLUSTRATIONS t IV
LIST OF TABLES V
ACKNOWLEDGEMENTS 1
I SUMMARY _2
II PERCHLOROETHYLENE IN THE ENVIRONMENT 5
A. Introduction 5
3. Chemical and Physical Properties of Perchloroethylene ... 6
C. Sources of Perchloroethylene 8
III PRODUCTION FACILITIES 15
A. Source 15
B. Methodology 19
C. Exposures 25
IV DRY CLEANING OPERATIONS 28
A. Sources 28
1. Perc Consumption 28
2. Process Description 30
3. Emission Controls 31
4. Monitoring Data 33
B. Methodology 35
C. Exposures ^'
V METAL CLEANING OPERATIONS 46
A. Sources 4°
1. General 46
2. Service/Maintenance Industry Degreasing 50
3. Manufacturing Industry Degreasing 51
B. Methodology 53
1. Emissions - Cold Cleaners 53
2. Emissions - Open-Top Vapor Degreasers 53
3. Emissions - Conveyorized Degreasing "
4. Perchloroethylene Emissions ^
5. Exposure Estimates ^
C. Exposures
BIBLIOGRAPHY 64
APPENDIX A Number of Dry Cleaners in Urban Areas and Exposed
Population from Dry Cleaners, by State A-l
1M
-------
LIST OF TABLES (Continued)
20. Population Exposed to Perc From Dry Cleaners 44
21. Projected Growth in Solvent Metal Cleaning
Industry, 1974 - 1985 47
22. National Degreasing Solvent Consumption (1974) 48
23. U.S. Halogenated Solvent Consumption by Type
of Degreasing Operation (1974 and 1975) 49
24. Categories of Manufacturers 52
25. Estimated Perchloroethylene Used for Degreasing
in the Manufacturing Industry 54
26. Facilities of Selected SIC Codes and Sizes
in the United States 57
27. Estimated Perc Emissions From Metal Cleaning
in Manufacturing Plants of Various Sizes 59
28. Estimated Annual Average Atmospheric Concentrations of
Perc as a Function of Distance From Plants Using Perc
as a Degreaser . j60
29. Estimated Population Exposures to Atmospheric Perc
Emissions From Industrial Degreasing "62
A-l Estimated Urban Commercial Dry Cleaners That
Use Perc, by State A-2
A-2 Estimated Urban CoinJOperated Dry Cleaners That
Use Perc, by State A-4
A-3 Estimated Urban Industrial Dry Cleaners That
Use Perc, by State A-6
A-4 Estimated Population Exposed to Perc From Commercial
Dry Cleaners in Urban Areas A-8
A-5 Estimated Population Exposed to Perc From Coin-Operated
Dry Cleaners in Urban Areas . A-10
A-6 Estimated Population Exposed to Perc From Industrial
Dry Cleaners in Urban Areas A-12
V1
-------
LIST OF TABLES
1. Summary of Estimated Population Exposures
to Atmospheric Perc .4
2. Physical Properties of Perchloroethylene 7
3. Estimated Consumption of Perc by Type of Use, 1978 9
4. Summary of Ambie.nt Monitoring Data for Perc in Various \
Locations 13
5. Ambient Concentrations of Perc Measured by Rutgers
University, 1973-74 14
6. Locations of and Production Figures for Perc Facilities ... 16
7. Summary of Waste Disposal Practices at Perc Production
Facilities 18
8. Results of Monitoring for Perc at Facilities Producing
Trichloroethylene and Methyl Chloroform 20
9. Estimated Emissions From Perc Facilities 21
10. Locations of Each Perc Production Facility by Latitude
and Longitude 26
11. Population Exposed to Perc From Production Facilities.. ... 27
12. Perc Consumption by Type of Dry Cleaning Operation, 1978. . . 29
13. Perc Losses From Dry,Cleaning Processes and Equipment .... 32
14. Perc Losses From Dry Cleaning Plants With Vapor Adsorbers . . 34
15. Variation in Perc Emissions Based on Type of Load
for Dry Cleaning.
16. Perc Exposure Data for Employees in Commercial Dry
Cleaning Plants
by Size of Operation.
19. Estimated Emission Rates for Each Type of Dry Cleaning
Operation
36
17. Density of Dry Cleaners in Selected Cities .......... 3.
18. Estimated Number of Urban Dry Cleaners Using Perc
42
-------
LIST OF ILLUSTRATIONS
1. Market Distribution of Perc, 1978
2. General Dispersion Curve for Perc Based on an Emission,
Rate of 1.0 Gram per Second ................ 23
1-V
-------
ACKNOWLEDGEMENTS
It is a pleasure Co acknowledge the cooperation and guidance given
by several individuals of the U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards. Ken Greer, Strategies and
Air Standards Division, and Dr. George H. Wahl, EPA Consultant, provided
direction throughout the study. Valuable assistance concerning control
technology and emission factors related to sources of perchloroethylene
was provided by Charles Kleeberg, David Mascone, and Jeff Shumaker of
the Emission Standards and Engineering Division. Joseph D. Cirvello was
the Project Officer.
Mr. Casey Cogswell, SRI International, Chemical Industries Center,
generously provided information and guidance concerning manufacuring and
distribution of perchloroethylene. Ms. Lynn Manfield edited the report.
-------
I SUMMARY
This report is one in a series that SRI International is providing
for the U.S. Environmental Protection Agency (EPA) to estimate
populations-at-risk to selected pollutants. The primary objective of
this study was to estimate the environmental exposure of the U.S.
population to atmospheric perchloroethylene (perc) emissions. The three
principal sources of atmospheric perc considered in this report are
facilities in which perc is produced or used as a chemical intermediate,
dry cleaning operations, and metal cleaning operations.
The method of approach for estimating population exposures varied
with the type of data available. Production facilities were located by
latitude-longitude, and the perc emission rates for these facilities
were calculated. Average annual concentrations of perc in the vicinity
of these facilities were then estimated by applying approximate
dispersion modeling results. The population exposed at average annual
concentrations greater than 0.01 ppb was estimated.
The number of urban dry cleaning and metal cleaning operations that
use perc, and the emission rates associated with them, were estimated on
a state-by-state basis. Average annual concentrations were calculated
through dispersion modeling based on the assumption that plants are
uniformly distributed throughout urban areas and that each one will act
as a point source. The number of people in each area exposed at
*
concentrations greater than 0.05 ppb was estimated.
The resulting estimates are subject to considerable uncertainty in
regard to: (1) perc emissions from various facilities; (2) locations of
sources; (3) control technologies employed; (4) deterioration of control
The dispersion modeling technique was not considered valid below 0.05
ppb, because the closeness of the sources resulted in double counting
of the exposed population.
-------
technologies over time; (5) physical characteristics of perc sources;(6)
the number of metal cleaning and dry cleaning operations that use perc;
(7)
population density in the vicinity of sources in urban areas; and (8)
living patterns of the urban population. Given these complex and
variable factors, the accuracy of the estimates could not be assessed
quantitatively. Despite the uncertainties, however, the estimates
approximate expected conditions.
Table 1 summarizes the results of the assessment. The two largest
sources of 'exposures are commercial dry cleaners and metal cleaning
operations, each of which affect 30 million people. Metal cleaning
operations are also the source of the largest number of exposures at
concentrations greater than 1.0 ppb.
To facilitate approximate comparisons of different emission
sources, weighted exposures have been calculated in similar units by
multiplying the number of people exposed by the annual average perc
concentration, in each range. These values were then summed for each
emission source to produce, estimates expressed in ppb-person-years (see
Table 1). The results show that the highest weighted human exposures
resulted from metal cleaning operations, followed closely by commercial
dry cleaners. Because these results are weighted by the number of
people exposed to a particular level of atmospheric perc, they provide a
useful basis for comparison. Assuming a linear dose response
relationship, these weighted exposures are directly related to the
expected risk to human health.
These preliminary estimates indicate that the number of people
exposed is substantial. Further monitoring and sampling data are
required for a more thorough and accurate assessment. Potential health
effects of exposure to perc at environmental concentrations will be
addressed in a separate report being prepared by the EPA Cancer
Assessment Group.
-------
Table 1
SUMMARY OF ESTIMATED POPULATION
EXPOSURES TO ATMOSPHERIC PERC
Number of People Exposed to Fere Concentrations (ppb)
Comparison
Source
Production Facilities
Dry Cleaning:
Coumerc ial
i Coin-operated
Industrial
8-Hour Worst-Case 0.25 - 1.3 1.4 - 2.5
Annual Average 0.01 - 0.05 0.06 - 0.10
ties 300,000 20,000
J 20,000,000
f 2,600,000
f 2,900,000
2.6 - 25.0 25.1 - 100.0
0.11 - 1.00 1.01 - 4.00
5,000
11,000,000 41,000
790,000
1,800,000 4,000
TOTAL
300,000
30,000,000
3,000,000
5,000,000
Among Sources^
(10 ppb-persons-yrs)
0.01
7.8
0.6
1.2
Metal Cleaning
20,000,000
11,000,000 120,000 30,000,000
8.0
* To convert to jug/nr, multiply each exposure level by 6.7. A dash ( - ) signifies that no expoaed population
was estimated by our method for the annual average concentrations listed. There may be some people exposed Co
those concentrations for shorter periods of time. In addition, the analysis for dry cleaning and metal cleaning
operations assumed that no people lived within certain distances of the plant.
* Totals are rounded to one significant figure. Population estimates are not additive vertically because double-counting exists,
due to overlap of sources.
T Not estimated, because the closeness of the sources resulted in double-counting of the exposed population.
Source: SRI estimates.
-------
II PERCHLORETHYLENE IN THE ENVIRONMENT
A. Introduction
Perchloroethylene (C^l^) is one of a number of chlorinated
hydrocarbon solvents that have come under investigation by the federal
government. The National Cancer Institute (NCI) has recently determined
that perc causes cancer" in mice when it is administered orally. The
National Institute of Occupational Safety and Health (NIOSH) has urged
industry to consider perc a human carcinogen as well. Currently, the
Occupational Safety and Health Administration (OSHA) has set an exposure
limit of 100 parts per million (ppm) for perc concentrations in the
workplace. This limit may be revised downward as a result of the recent
findings of NCI. The Consumer Product Safety Commission has
provisionally proposed to classify perc as a Class A carcinogen, a
procedure that is the first step in restricting the use of perc in
consumer products. The outcome of the various proposed actions
affecting perc could have a significant effect on its future use as a
solvent.
The primary objective of this study has been to estimate the
atmospheric exposure of the U.S. population to perc from each major
source. This exposure information together with the health effects
document being prepared by the U.S. Environmental Protection Agency
Office of Research and Development will provide the necessary data for
EPA's Cancer Assessment Group to make a determination of risk.
Because few quantitative data were available, all estimates given
here are subject to considerable uncertainty. This uncertainty is
related to: quantity of perc emissions, locations of emission sources,
estimates of perc production and consumption, control technology
employed, deterioration of control technology over time, and dispersion
modeling. Insufficient atmospheric, monitoring data are available to
-------
assess the accuracy of Che modeling results. Comparisons of short-term
ambient concentration:} established by monitoring with annual average
ambient concentrations estimated from dispersion modeling are tenuous.
In addition, meteorological conditions and operating characteristics of
facilities may be significantly different from the average. Despite
these uncertainties, a comparison of monitoring data with dispersion
modeling results suggests that the agreement between the two is
sufficient to support the accuracy of the modeled concentrations.
B. Chemical and Physical Properties of Perchloroethylene
Perchloroethylene is a colorless, extremely stable, and
nonflammable liquid with an ether-like odor. It is insoluble in water
.but miscible with alcohol, ether, and oils in all proportions. Perc has
a relatively high density and a moderate boiling point and heat of
vaporization. Pertinent physical properties are listed in Table 2.
Few reactivity studies of perc have been conducted. Huybrechts et
al. (1967) studied the photochiorination and oxidation of perc under
specialized conditions at 80 and 100°C. Analysis of the products in
the completed experiments showed that 85 + 52 of the oxidized perc
appears as trichloroacetyl chloride and 15 _+ 51 as phosgene. Trace
quantities of carbon tetrachloride and tetrachloroethylene oxide were
also detected. Studies to determine the rate constant for
photochlorination of perc have been made by Horowitz et al. (1968),
Goldfinger et al. (1961), and Dusoleil et al. (1961).
In 1976, Rutgers University (Appleby, 1976) completed an evaluation
of the atmospheric fates of perc in a polluted atmosphere. In each
system investigated, perc was chemically reactive under the influence of
the light source used. The time needed to achieve a given level of
decomposition varied significantly. The major decomposition products of
perc included: phosgene, carbon tetrachloride, trichoracetyl chloride,
and trichloroacetaldehyde. On the average, perc photodecomposition was
found to cause the formation by weight of about 82 carbon tetrachloride
and 70 - 85Z phosgene in 7 days. Phosgene is a highly toxic material
-------
Table 2
PHYSICAL PROPERTIES OF PERCHLOROETHYLENE
Properties Value
Density, Ib/gal* 13.55
Boiling point, °C** 121.0
Heat of vaporization, Btu/lb* 90.0
Flashpoint* none
Freezing point, °C** -22.4
Specific gravity @ 20°C** 1.625
Refractive index @ 25°C** 1.5029
Color* water-white
*7isher, 1977
**Hawley, 1977
-------
with a threshold limit value (T.L.V.) of 0.05 ppm. A half-life of 2
days for photodecomposition of unstabilized perc has been reported in
the literature (Lapp et al., 1977).
Fuller (1976) reported that perc reacts rapidly with hydroxyl ions
(half-life of 8 days) and slowly with alkyl peroxy radicals and ozone
(half-lives of 220 da/s and 11 years, respectively). Fuller found that
perc decomposed to a mixture of phosgene and trichloroacetyl chloride in
the presence of ozone.
Perc is corrosive to metals in the absence of chemical stabilizers
such as organic amines. Below 140 C, stabilized perc is inert to air,
water, light, and common construction metals (Franklin Institute
Research Laboratories, 1975). In the absence of moisture, oxygen, and
catalysts, the compound"is stable to about 500°C. At 700°C, it
decomposes on contact with active carbon to yield hexachloroethane and
hexachlorobenzene.
The decomposition rates for perc in aerated water both in the
presence of sunlight and in darkness were determined by Dillings et al.
(1975). The half-life for the reaction in darkness was 8.8 months,
whereas the half-life for the reaction in sunlight was only 6 months.
The researchers believed that degradation was probably caused by
oxidation and was probably free radical in nature. Dillings et al. also
found that unstabilized perc decomposes to trichloroacetic and
hydrochloric acids after contact with water for long periods at elevated
temperature, and that perc is relatively resistant to hydrolysis.
Because of its high vapor pressure and low solubility, perc will be
rapidly removed by volatilization from agitated natural bodies of water
(Dillings et al., 1975).
C. Sources of Perchloroethylene
The dry cleaning industry is the major user of perc, as Table 3
shows. The compound is also used in metal cleaning as a chemical
intermediate for the production of C fluorocarbons, in textile
8'
-------
Table 3
ESTIMATED CONSUMPTION OF PERC
BY TYPE OF USE, 1978
Type of Use
Dry cleaning
Metal cleaning
Chemical intermediate
Textile processing
^^t
Miscellaneous
Consumption*
(106 Kg)
160
50
40
20
30
Percent of
Total
53
17
13
7
10
Totals
300
100
*Rounded to the nearest ten.
u^U
Includes product going into inventory.
Source: SRI estimates
-------
processing, and for miscellaneous purposes (e.g., as a solvent for silicones
and as a constituent in aerosol specialty products for the laundry
pretreatment market). The market distribution of perc is presented in
Figure 1. Note that imports and exports of perc nearly balance and that
U.S. production accounts for more than 90% of U.S. consumption.
Domestic consumption of perc in 1977 was about the same as it was in
L972, but it was 7% less than the peak reached in 1974. Growth to 1982 is
expected to be smallprobably less than 1.0% annually (SRI estimates).
In this report, three sources of perc emissions are evaluated:
production facilities, dry cleaning operations, and metal cleaning
operations. The analysis of production facilities includes those in which
perc is used as a chemical intermediate. Consequently, our analysis covers
all perc production (290 x 10 kg per year) and 83% of perc consumption
(250 x 10 kg per year).
Exposures from textile processing facilities are not considered in our
analysis. Although textile processing accounts for 7% of perc consumption,
no information is available that would permit us to determine which
facilities actually use perc. A quick survey of applicable Standard
Industrial Classification (SIC) codes from the 1970 Census of Business
showed that more than 2,500 facilities are engaged in textile processing
operations that might involve the use of perc. The size of the group means
that each facility probably uses only a small amount of perc and would emit
a similarly small amount of perc to the atmosphere.
Perc is a manmade substance that is not released from any natural
sources. It is not sampled regularly in any continuing air quality
monitoring program. A few ambient measurements of perc have been made, and
suggest that typical ambient concentrations are generally below 1.0 ppb. A
report on atmospheric halogenated compounds prepared by Rutgers University
(Appleby, 1976) summarized the measurements of perc available in the
10
-------
IMPORTS
30
U.3. PRODUCTION
290
DRY CLEANING
160
METAL CLEANING
50
CHEMICAL
INTERMEDIATE
40
TEXTILE PROCESSING
20
MISCELLANEOUS
30
EXPORTS
20
Soum: SRI «stinr»te»
FIGURE 1. MARKET DfSTRIBUTION OF PERC, 1978 (106kg)
n
-------
literature, as shown in Table 4. All measurements were 1.0 ppb or below.
Rutgers also conducted a monitoring program in eight eastern U.S. cities.
Their results are presented in Table 5. Note that New York City and
Bayonne, New Jerseyboth densely populated regionshad perc concentrations
greater than 1.0 ppb and maximum concentrations near 10.0 ppb. On the other
hand, Baltimore, Maryland, and Wilmington, Delawarealso densely populated
regionshad mean perc concentrations less than 0.5 ppb.
The EPA Environmental Monitoring and Support Laboratory (draft report,
1979) has recently collected ambient measurements in New York City, Houston,
and Detroit. These cities were selected because they were suspected to have
a large number of perc sources in relation to population density. New York
City has a large number of dry cleaning facilities, Houston has a perc
production facility in the vicinity, and Detroit has a large number of metal
cleaning operations. The 24-hour average concentrations of perc jfound at
each sample station varied as follows: 0.16 - 10.61 ppb in New York City;
<0.1 - 4.52 ppb in Houston, and <0.1 - 2.16 ppb in Detroit. The high
concentrations found in New York City probably result from emissions from
dry cleaners and metal cleaning facilities.
* The detection level is 0.1 ppb for sampling techniques with easily
replicable results.
12
-------
Source of Data
Kuriyang
Williams
Reid et al.
Simmonds et al.
Lillian and Singh
Table 4
SUMMARY OF AMBIENT MONITORING DATA FOR
PERC IN VARIOUS LOCATIONS
Location
Unknown
Unknown
Unknown
Los Angeles area
New Brunswick
Perc Concentration (ppb)
1.0
1.0
1.0
0.125
0.112
Source: Appleby, 1976
13
-------
Table 5
AMBIENT CONCENTRATIONS OF PERC MEASURED
BY RUTGERS UNIVERSITY, 1973-74
Sampling Concentration (ppb)
Location
Baltimore, MD
B ay onne, NJ
New York, NY
Sandy Hook, NJ
Seagirt, NJ
White Face Mountain, NY
Wilmington, DE
Wilmington, OH
Period (days)*
2
12
2
4
2
Y 4
3
' 4
Mean
0.18
1.63
4.5
0.39
0.32
0.07
0.24
0.15
Maximum
0.29
8.2
9.75
1.4
0.88
0.19
0.51
0.69
Minimum
0.02
0.30
1.0
0.15
0.10
0.02
0.02
0.02
* Discontinuous sampling for a few hours at a time.
Source: Appleby, 1976
14
-------
Ill PRODUCTION FACILITIES
A. Source
Perc is produced at 10 facilities in 5 states: California, Kansas,
Kentucky, Louisiana, and Texas. The company name and the location of each
facility are shown in Table 6. The 1978 capacity for all facilities
combined was 515 x 10 kg.'
estimated U.S. production.
combined was 515 x 10 kg.* Actual production is estimated at 56% of the
Perc is produced by three major processes. The most common process and
one of the most economical is chlorination of C, to C- hydrocarbons or
their partially chlorinated derivatives at high temperatures (500-540°C)
with or without a catalyst. Propane is a common feedstock. Perc and carbon
tetrachloride are both produced, and it is possible to vary the product
distribution to produce more than 90Z of either material. Large quantities
of hydrochloric acid are produced by this process. The six production
facilities that use this process account for 52% of the total estimated perc
production.
Ethylene dichloride is used as the feedstock for two of the perc
manufacturing processes. Perc is produced by high-temperature
(400-450 C), noncatalytic chlorination of ethylene dichloride.
Adjustments of the chlorine/ethylene dichloride ratio favor production of
perc over trichloroethylene, which is also produced by this process. Large
quantities of hydrochloric acid are produced. The three facilities that use
this process account for 31% of the total estimated perc production.
*This fizure does not include the capacity at the Hooker Chemical
Corporation facility in Taft, Louisiana, which reportedly ceased
production in March 1978.
15
-------
Table 6
LOCATIONS OF AND PRODUCTION FIGURES FOR PERC FACILITIES
Location
Company*
Manufacturing
Process**
1978
Capacity
(106 kg) +
1978
Estimated
Production
(106 kg)*
California
Pittsburg
Kansas
Wichita
Kentucky
Louisville
Louisiana
Baton Rouge
Geismar
Lake Charles
Plaquemine
Texas
Corpus Christi
Deer Park
Freeport
Total
Dow Chemical USA Cl-HC
Vulcan Materials Co. Cl-HC
Stauffer Chem. Co. Cl-HC
Ethyl Corporation Cl-EDC
Vulcan Materials Co. Cl-HC
PPG Industries, Inc. Ox-EDC
Dow Chemical USA Cl-HC
DuPont 5 Cl-HC
Diamond Shamrock Cl-EDC
Dow Chemical USA Cl-EDC
20
20
30
20
70
90
50
70
75
70
515
10
10
20
10
40
50
30
40
40
40
290
*An August 1978 article in Chemical Marketing Reporter stated that the Hooker
Chemical Corporation facility at Taft, Louisiana stopped producing perc in March
1978.
r Key to symbols: Ci-HC - Chlorination of C^ to Cj hydrocarbons or their
partially chlorinated derivatives; Cl-EDC chlorination of ethylene dichloride;
Ox-EDC - oxychlorination of ethylene dichloride.
* SRI estimates.
§ Captive use only.
-------
Oxychlorination of ethylene dichloride is used to produce perc ac one
facility. In this process, which accounts for 17Z of estimated production,
perc and trichloroethylene. may be produced either separately or as a mixture
in varying proportions by reacting ethylene dichloride (or other C?
chlorinated hydrocarbons) with a mixture of oxygen and chlorine or
hydrochloric acid at about 430 C over a suitable catalyst. This process
avoids net production of hydrochloric acid and provides an outlet for
unwanted hydrochloric acid from other processes.
Approximately 13Z of the perc produced is consumed as a chemical
intermediate in the formulation of a series of C? fluorocarbons: F-113,
trichlorotrifluoroethane; F-114, dichlorotetrafluoroethane; F-115,
chloropentafluoroethane; and'F-116, hexafluoroethane. All the perc produced
at the DuPont facility at Corpus Christi is used captively in the production
of fluorocarbons from perc primarily F-113. Of the four fluorocarbons,
F-113, which is used primarily as a premium solvent, is produced in the
(55
largest volume. F-113 is the basis for DuPont's Valclene ^ dry cleaning
agent and is used as a cleaning solvent for special applications such as
electronics and aerospace equipment. F-113 is also used to make
chlorotrifluoroethylene, a monomer for fluorocarbon resins. Despite the
high cost of F-113, use of it is expected to increase because of its low
toxicity and its usefulness in special applications. F-114 is the second
most important fluorocarbon derived from perc, but the volume of F-114
combined with that of the smaller members of the group is no more than half
that of F-113.
Midwest Research Institute (MRI) conducted a survey in 1977 of the
waste disposal practices of producers of chlorinated solvents for the EPA
Office of Toxic Substances. The survey results are discussed in a draft
final report (Lapp et al.t 1977) that will soon be released in final form.
The principal wastes from perc production are tarry residues ("hex" wastes)
that contain hexachlorobenzene, hexachlorobutadiene, hexachloroethane, and
other chlorinated compounds. Industry practices for disposal of these
wastes vary, but generally include incineration and landfill. Table 7
summarizes the relevant information derived from the MRI survey.
17
-------
Table 7
SUMMARY OF WASTE DISPOSAL PRACTICES
AT PERC PRODUCTION FACILITIES
Company
Diamond Shamrock
Ethyl Corporation
PPG Industries
Vulcan Materials
Location
Deer Park, TX
Baton Rouge, LA
Lake Charles, LA
Wichita, KN
Geismar, LA
Disposal of Wastes
from Perc Production
Shipping in sealed con-
tainers to private com-
pany for incineration
Deep-well disposal
Incineration that oper-
ates at 1,371° C and
uses aqueous scrubbers
for HC1 recovery and
alkaline scrubbers for
chlorine destruction.
Incineration
Impounded in a landfill
Source: Lapp et al., 1977
-------
Only one report en measurements of perc in the vicinity of production
facilities has been found. 'Battelle (1977) sampled ambient concentrations at
several facilities manufacturing trichloroethyelene and methyl chloroform.
Most of these facilities also produce perc, and therefore perc measurements
were made in conjunction with the other sampling. The pertinent results are
shown in Table 8. All measurements were taken within 4 km of the facility,
most were taken within 2 km.
B. Methodology
Because of the scope of this study detailed dispersion calculations that
incorporate, individual site meteorology are impractical. A simple method of
assessment was therefore developed to permit comparative analysis. Variations
in pollution control technology, physical surroundings, and meteorological
conditions were not considered in the analysis. Because of these assumptions,
the results are not precise but do provide a reasonable estimate of
atmospheric perc concentrations and associated exposed population.
To assess the ambient perc concentrations in the vicinity of production
facilities, two factors must be estimated: perc emission rates at each
location, and atmospheric dispersion in the vicinity of the facility.
Emission rates are estimated by applying an emission factor to estimated total
production at the facility- One of two rough emission factors was used,
depending on the particular manufacturing process used at the facility:
Direct chlorination process 0.002 kg emitted/kg produced
Oxychlorination process - 0.005 kg emitted/kg produced.
These factors were determined by EPA (Greer, personal communication, 1978) on
the basis of several very limited studies of perc losses during production.
They do, however, provide a reasonable estimate of emissions from each
facility. The emission rate for each production facility is shown in Table
9. Twenty-four hour operation over 365 days was assumed.
19
-------
Table 8
RESULTS OF MONITORING FOR PERC
AT FACILITIES PRODUCING TRICHLOROETHYLENE
AND METHYL CHLOROFORM
Company
Dow Chemical
Ethyl Corp.
PPG Industries
Vulcan Materials
Location
Freeport, TX
Freeport, TX
Baton Rouge ,
Lake Charles
Geismar, LA
Type of
Production*
MC
TCE
LA TCE
, LA MC,TCE
MC
Number
of
Samples
36
51
49
47
66
Perc
Max.
ND**
3.4
37.0
5.0
23.0
Concentration
(ppb)f
Min.
ND
0.3
0.3
0.3
0.3
Mean
ND
0.5
1.6
0.7
2.2
*I.e., Che chemical produced at Che particular facility sampled. Perc is
also produced at each location, but Che facilities may be separate. Infor-
mation on the perc production facilities at each location was not available
from chis study. MC * methyl chloroform; TCE a trichloroethylene.
Limit of detection was reported Co be 0.3 ppb. To convert to M
multiply by 6.7.
**
ND 3 not determined.
Source: Battelle, 1977
20
-------
Takie $
EHESSHSS FROM 7ES£ FAC
Location
California
Pitts burg
Kansas
Wichita
Kentucky
Louisville
Louisiana
Baton Rouge
Geismar
Lake Charles
Plaquemine
Texas
Corpus ChriseL
Deer Park
Freeport
Efesr OhesrcsL OS&
Cou
SCauffar Chemical Co.
Ethyl Corp-oratian
7uican Materials Co.
IntfaiaCriaB, Inc-
ChesicaL
HuEonC*
OiaiaoTid Shamrock
Dow Chemical USA.
do kg)
40
20
220
60
ao
ao
so
a.63?
L.3Q
Q..6J
Z.ST-
7.9Q
Z.5Q
Z«53
2.50
* Captive use for C^ fluoracarfacna; emiaaioa factor is onehanged.
Source: SRL eacimacea
-------
The ambient concentrations of perc present in the vicinity of production
facilities are estimated based on.an earlier application by Youngblood (1977;
described in Mara and Lee, 1978). Dispersion curves developed in a study of
benzene exposures were modified for application to perc emissions (refer to
Mara and Lee, 1978). A single dispersion curve was developed by regression
analysis to reasonably represent dispersion from three source categories:
ground-level point source (effective stack height, 0 m); building source
(effective stack height, 10 m); and elevated point source (effective stack
height, 20 m)- The meteorological conditions assumed were: wind speed, 4
m/s; stability class, neutral (Pasquill Gifford "D"). Exhaust gas
temperature, which is important in determining near-source concentrations, was
not considered. The general dispersion curve for perc is shown in Figure 2.
This curve was used to estimate ambient perc concentrations for all source
categories: production facilities, dry cleaning operations, and metal
cleaning operations.
Equation (3.1) is used to estimate 8-hour worst-case concentrations:
C = 16.48 E_ D'1'48. (3.1)
a
C is the 8-hour worst-case perc concentration in Mg/m , E is the emission
rate for the location of interest in g/s, and D is the distance from the
source in km.
Annual average concentrations are estimated by including a multiplier of
0.04 in the equation. The equation, thus becomes:
C - 0.659 Ea D"1'48. (3.2)
This 0.04 conversion is based on results from previous studies. It
represents the ratio of 8-hour concentrations to directionally-averaged annual
concentrations. This conversion gives reasonable concentration estimates in
keeping with the general nature of this study.
22
-------
GENERAL EQUATIONS:
ANNUAL AVERAGE - C
8-HOUR WORST CASE - C
0.01
673
DISTANCE - km
Source: SRI estimates
FIGURE 2. GENERAL DISPERSION CURVE FOR PERC BASED ON AN
EMISSION RATE OF 1.0 GRAM PER SECOND
23
-------
In this study, all population exposures are estimated on the basis of
annual average concentrations. This approach is consistent with the
requirements of a linear dose response model which assumes that exposures to
higher concentrations for short periods of time have no higher risk associated
with them.
For the sake of uniformity, the following ranges of perc concentrations
were used in this study:
0.01 - 0.05 ppb
0.06 - 0.10 ppb
0.11 - 1.00 ppb
1.01 - 4.00 ppb.
All exposure estimates were confined to these ranges. Note that 0.10 ppb is
the present limit for detection of perc that produces easily replicable
results.
For each facility, the distance at which the specified concentrations
are found is determined by rearranging equation (3.2) as follows:
(3.3)
D. is the distance in km at which the specified concentration is found; Ea
is the emission rate at that location in g/s; and C. is the specified
annual average concentration (i.e., 0.01, 0.05, 0.10, 1.0, and 4.0 ppb;
2
input data, however, are in ^g/m ). Emissions from production facilities
were assumed to be the only contributors of perc to the atmosphere in the
vicinity of the facility.
The population residing within a circle of radius D. was estimated by
SRI's computer system, BESTPOP (Suta, 1978). The population file consists
of a grid of 1-km-square sections that span the continental United States.
This file was created by assigning the 1970 populations to the grid network
and by assuming uniform distribution of population within the 256,000
enumeration districts in the 1970 census. The computer software accesses
-------
the population file and accumulates residential population within specified
radial rings about any given point. In addition, a rectangular map for and
area around each specified point is printed out, and it shows the population
by square kilometer.
The latitude and longitude of each facility was determined according to
information provided by the company itself or from other studies completed
for EPA. These data are shown in Table 10.
C. Exposures
Estimates of population exposed to perc emissions from production
facilities are shown in Table 11. Approximately 5,000 people are exposed to
perc at annual average concentrations greater than 0.1 ppb. Louisiana and
Texas have the largest exposed populations. Exposures are relatively low
level for two reasons: (1) perc emission rates are low; (2) facilities are
usually located in sparsely populated areas. Consequently, few people are
estimated to be exposed to perc from production facilities at concentrations
higher than the detectable limit of 0.1 ppb.
-------
Table 10
LOCATIONS OF EACH PERC PRODUCTION FACILITY
BY LATITUDE AND LONGITUDE
(Degrees, Minutes, Seconds)
Source of
Company
Diamond Shamrock
Dow Chemical
DuPont
Ethyl Corporation
PPG Industries
Stauffer Chemical
Vulcan Materials
Location
Deer Park, TX
Pittsburg, CA
Plaquemine, LA
Freeport, TX
Corpus Chris ti
Baton Rouge, LA
Lake Charles, LA
Louisville, KY
Wichita, KA
Geismar, LA
Information*
1
2
1
1
3
1
1
3
3
1
Latitude
29.43.34
38.01.35
30.19.15
29.01.59
27.52.30
30.29.20
30.12.36
38.12.30
37.34.50
30.10.23
Longitude
95.07.16
121.51.22
91.14.56
95.13.16
97.15.00
91.11.13
93.17.06
85.52.30
97.25.16
90.57.58
*1 - EPA
2 » Association of Bay Area Governments, Berkeley, CA
3 a Company spokesmen
26
-------
TABLE 11
Location
Callfornia
Pitcsburg
Kansas
Wichita
Kentucky
Louisville
Lou isiana
Baton Route
Geismar
Lake Charles
Plaquemine
Company
Dow Chemical Co.
Vulcan Materials Co.
Stauffer Chemical Co.
Ethyl Corporation
Vulcan Materials Co.
PPG Industries Inc.
Dow Chemical USA
POPULATION EXPOSED TO PERC
FROM PRODUCTION FACILITIES
Population Exposed* to Perc (ppb)**
0.01 - 0.05 0.06 - 0.10 0.11 - 1.00
Texas
Corpus Christ! DuPont
Deer Park Diamond Shamrock
Freeport Dow Chemical USA
Total Exposed Population*:
* Rounded to one significant figure.
20,000
600
40,000
300,000
500
30,000
10,000
100,000
20,000
20,000
90,000
1,000
500
-
20,000
500
_
100
-
20,000
***
5,000
60
5,000
Total Exposed
Population*
20,000
600
40,000
30,000
10,000
100,000
20,000
20,000
90,000
1,000
300,000
** Annual average concentrations; to convert to 8-hour worst-case estimates, multiply by 25;
to convert to ^g/m3, multiply by 6.7; a dash (-) signifies that our method estimated that no
people were exposed at the annual average concentrations listed, although some people may be
exposed at those concentrations for short periods of time.
*** Less than 10 people.
Source: SRI estimates.
-------
IV. DRY CLEANING OPERATIONS
A. Sources
1. Perc Consumption
Two types of solvents are used extensively in the dry cleaning
industry: petroleum solvents (called "Stoddard") and perchloroethylene.
Perc has been used since about 1950, and dry cleaning plants currently
account for 53% of perc consumption in the United States. Although
available statistical data indicate that the number of dry cleaning
establishments has decreased in the past 20 years, consumption of perc in
the industry has remained fairly constant and is currently estimated at 160
'x 10 kg per year (see Table 12).
The more than 40,000 U.S. dry cleaning establishments are
classified by SIC code as commercial cleaners (7216), coin-operated cleaners
(7215), and industrial cleaners (7218)*. Data from the International
Fabricare Institute (IFI) indicate that 82% of the dry cl'eaning plants in
the United States use perc (Fisher, 1977). As Table 12 shows, more than 75%
of the perc consumed by the industry is used in commercial cleaning
es tab1ishment s.
The overall number of facilities in the three SIC-code categories
decreased by 4% from 1970 to 1974 (Bureau of Census, 1970 and 1974 County
* SIC 7215, Coin-Operated Laundries and Dry Cleaning, is defined as
establishments primarily engaged in the operation of coin-operated or
similar self-service laundry and dry cleaning equipment for use on the
premises, or in apartments, dormitories, and similar locations} SIC 7216,
Dry Cleaning Plants, Except Rug Cleaning, is defined as plants primarily
engaged in dry cleaning or dying apparel and household fabrics other than
rugs; SIC 7218, Industrial Launderers, is defined as establishments
primarily engaged in supplying laundered or dry cleaned work uniforms,
laundered wiping towels, safety equipment, dust control items, and other
selected items to industrial or commercial users. Establishments included
in this industry may or may not operate their own laundry or dry cleaning
facilities.
23
-------
Table 12
PERC CONSUMPTION BY TYPE
OF DRY CLEANING OPERATION, 1978
Type of
Operation
Number
of
Plants*
Percent
Using
**
Perc
Number
Using
Perc
Estimated
Perc
Consumption
do6 kg)*
Consumption
as Percent
of Total
Commercial 22,000
74
16,000
123
77
Coin-
operated 11,000
Industrial
700
97;
50
11,000
350
23
14
14
9
Total 33,700
27,350
160
100
**.
*The number of coin-operated and industrial plants with dry cleaning
equipment was estimated based on data from the Bureau of Census, 1974
County Business Patterns4 The number of commercial cry cleaners was
taken directly from that publication.
'Fisher, 1977.
' SRI estimates
* Kleeberg, 1978.
32 use Fluorocarbon-113.
29
-------
Business Patterns). However, only commercial dry cleaners have shown an
actual decrease (down 15%) while coin-operated plants have increased by
17% and industrial plants have increased by 45%. (Note that some plants
in the latter two categories are laundries with no dry cleaning
equipment.)
2. Process Description
Dry cleaning is a three stage process. In the wash cycle,
perc is passed through a rotating cylinder to remove soil from
garments. The perc is continuously recirculated and usually passes
through a diatomite or a cartridge filter for clarification. After the
wash cycle, an extraction or centrifuging process is employed to remove
as much perc as possible from the garments before they are dried.
Extraction is usually carried out in the same cylinder that is used tor
washing. Perc that remains in the garments after extraction is removed
in the drying cycle. Air is heated in a rotating cylinder and passed
through the garments. The perc-laden air leaving the cylinder is passed
over a water-cooled condenser to recover the perc. The same air is then
reheated and recirculated in the recovery chamber. After an appropriate
drying period, fresh air is drawn into the chamber and then exhausted to
deodorize the garments. All three processes may be carried out in a
single machine called a "hot unit", but it is more common to use a
separate unit, called a recovery tumbler, for drying.
Filtration is not the only method used for clarifying perc.
Periodic distillation at atmospheric pressure in either a still or a
combination cooker/still with water separators is often used to recover
perc. Because perc and water are immiscible, separators are usually
simply small metal containers with baffles to facilitate the separation
process.
The largest single loss of perc from dry cleaning plants that
do not use carbon adsorption results from direct venting of solvent
vapors to the atmosphere. An IFI survey (1975) of California dry
30
-------
cleaners estimated average losses at 20 gallons per plant per week.*
Smaller amounts of p«rc also escape from solid residues, such as filter
powder cookers and cartridge filters, and are lost in leaks and spills
that occur during handling or at pipe joints. The average loss from
solid waste has been estimated by IFI at 0.1 to 6 gallons per week per
plant. Emissions associated with various dry cleaning processes and
equipment are summarized in Table 13. The most common filtration system
in commercial plants is a regenerative diatomite filter connected to a
muck cooker (Fisher, 1977)
Most commercial dry cleaning establishments use separate
dryers and transfer clothers manually. Clothes are sometimes
transferred automatically in industrial plants, which decreases perc
losses. Coin-operated cleaners always wash and dry in the same unit.
According to IFI, the relationship between the weight of the
clothes cleaned and the capacity is the most significant factor
affecting perc emissions. Fisher (1977) used that factor in estimating
perc losses in California. His estimates, in pounds lost per California
resident per ^year, are given in the following tabulation:
Total losses
By reported perc consumption 1.99
By calculated perc consumption 2.26
Atmospheric losses
By reported perc consumption 1.75
By calculated perc consumption 1.99
3. Emission Controls
Adequate control equipment can substantially reduce perc
emissions. At present, the primary emission control techniques in dry
cleaning plants are adsorption, filtration, distillation, condensation,
and housekeeping. Vapor adsorbers (carbon sniffers) are the most
sophisticated devices and are intended to remove all of the perc from
the air before it is exhausted to the atmosphere outside the plant.
*No plant size was given in the report.
31
-------
Table 13
PERC LOSSES FROM DRY CLEANING
PROCESSES AND EQUIPMENT
Average
Perc Losses
Process (kg lost/100 kg used)
Washer 0.5-1
Dryer 3-6
Filter X
- Diatomite (no cooker) 14
- Diatomite (with cooker) 1-1.6
- Cartridge 0.5 - 1.8
Still residue 1.6
Miscellaneous 1.2
Average total loss 12
- Range without vapor adsorber 8-21
- Range with vapor adsorber 3-18
Source: Fisher, 1977
32
-------
Data from a California survey (IFI, 1975) show that when washers and dryers
are ducted to adsorbers emissions of perc from these two sources are
effectively eliminated. However, exhaust air from about half of the cookers
or stills is either vented into the processing room or piped outside the
plant. Very high concentrations of perc are likely to occur at the outlets
of such vents or pipes. Floor-level pick-ups for the adsorbers are often
used to capture solvent vapors in the processing area (IFI, 1975).
Vapor adsorpers- are estimated to be used in about 35% of commercial
and industrial dry cleaning operations that use perc (Kleeberg, 1978).
Adsorbers are rarely found in coin-operated dry cleaners. A recent survey
of San Diego County (Evatovich, 1978) showed that 100% of the dry cleaners
in the county have vapor condensers and 45% have vapor adsorbers. Data on
adsorber efficiency and measurements of vapor concentrations in adsorber
exhaust are given in Table 14. The San Diego survey showed that about
two-thirds of the absorbers in dry cleaning plants were exhausting up to 40
times more perc vapor than manufacturer's specifications claimed they would
(Evatovich, 1978).
4. Monitoring Data
Three studies of perc emissions from two types of dry cleaning
operations have been conducted, and the results were summarized by Lapp et
al. (1977). Midwest Research Institue (1976a) sampled a large industrial
facility in San Antonio, Texas. The plant employs a 250-lb transfer system
equipped with a single carbon adsorption bed. The vapor adsorber is used
for only a part of each cleaning cycle; during other segments of the
operation, the system is either sealed or vented directly to the
atmosphere. Perc concentrations measured in the outlet air vent averaged 3
ppm for the first 2 days and 1 ppm for the third day of testing.
Midwest Research Institute (1976b) conducted a second study of a
small, neighborhood commercial dry cleaning operation in Kalamazoo,
Michigan. The facility used a 40-Ib, dry-to-dry system with dual carbon
adsorption beds. Average concentrations in samples from the outlet air vent
33
-------
CO
Table 14
PERC LOSSES FROM DRY CLEANING PLANTS WITH VAPOR ADSORBERS
Plant
No.
1
5
10
12
13
14
20
22
24
27
29
30
31
34
35
43
44
45
48
Hours Used
Per Day
6
8
8
10
8
8
8
Every
8
8
8
8
8
Only
when
cleaning
8
8
Only
when
cleaning
8
8
Frequency of
Stripping Vapor
Adsorber Bed
Daily
1 Bed/day
Every 3 days
Weekly
3 1/2 days
1 Bed/day
Each load
Daily
Daily
Daily
Daily
1 Bed/load
Daily
Daily
Daily
Daily
1 Bed/day
1 Bed/day
No. Loads
Per Week
75
109
65
50
211
(70 per
machine)
85
60
35
270
30
80
35
40
45
110
40
145
75
Base Reading
(ppm)
500
200
1,000
900
500
20
NA
240
70
10
10
10
700
30
NA
220
700
NA
Maximum
(ppm)
900
1,000
1,000
NA
No peak
No peak
No peak
No peak
No peak
No peak
No peak
NA
No peak
No peak
NA
Perc Losses
(Gal/Week)
26.8
13.5
20.7
19.8
37.5
1.0
1.3
12.2
3.4
0.5
0.5
0.5
15.8
1.7
150.8
4.0
38.1
Source: IFI, 1975
-------
ranged from 73 Co 138 ppm. Table 15 gives an example of how emissions
varied with the type oE the load.
A third study sampled a fairly large commercial operation in
Hershey, Pennsylvania (Scott, 1976). This facility used a 110-lb washer and
two dryers with a dual carbon adsorption bed. Hourly average concentrations
of perc in the outlet air vent pipe were measured for a 3-day period. The
average perc concentration was 22.8 ppm and ranged from 5 to 72 ppm.
Results of occupational exposure monitoring indicate that machine
operators are exposed at the highest average concentrations (see Table 16).
The lowest exposure concentrations are experienced by the employees who work
at the counter.
B. Methodology
A simple method of assessment similar to that described in Chapter III
was developed for dry cleaning operations that allows a comparative analysis
of exposures from various sources. Variations in pollution control
technology, physical surroundings, meteorological conditions, and weight of
the garments cleaned were not considered in the analysis.
Few studies have provided information on characteristics of the dry
cleaning industry. As a result, little is known about the location of
facilities, the extent to which dry cleaners are used by the local
population, perc use per facility, the extent and effectiveness of emi --'~r
controls, and the distribution of facilities within urban areas. The1-0^-
a number of assumptions had to be made.
Exposures were projected only for cities with populations of 25,000 or
more. Limiting the analysis in this way seems justifiable because the
population density of smaller cities is usually much lower.
Little information exists that documents differences in the density of
dry cleaners in various areas of the United Staes, although it is
35
-------
Table-15
VARIATION IN PERC EMISSIONS
BASED ON TYPE OF LOAD
FOR DRY CLEANING
Type of Load
Empty
Curtains
Rugs
Max
Min
Mean
Max
Min
Mean
Perc Emissions*
at Outside Air Vent
(ppm)
11
70
33
57
371
5
133
Source: MRI (1976b) as cited in Lapp et al. (1977)
* The carbon absorption beds may be underdesigned for the type of load.
36
-------
Table 16
PERC EXPOSURE DATA FOR EMPLOYEES IN
COMMERCIAL DRY CLEANING PLANTS
Number
Machine operator 5
Presser 2
Counter 5
Miscellaneous 7
Mean
of People Time-Weighted
Sampled Average (ppm)
Mean Breathing
Zone Sample
(ppm)
Mean
Peak
(ppm)
37.2 (+24.96) 20.47 (+18.21) 214.90 (+_179.41)
11.43 (+6.82) 4.48 (+0.39) , 51.85 (+64.56)
0.95 (+0.62) 2.61 0+ 1.87)
2.04 (+2.29) 27.45 (+46.02)
1.32 (+0.97)
3.03 (+2.09)
Source: Tuttle et al., 1977.
37
-------
strongly suspected that such differences do exist. The available
information is summarized in Table 17. Note that the variations between
cities of different sizes in the same state and between cities in
different states are small and that the cities with larger populations
have a higher density of dry cleaners.
Data from the 1974 County Business Patterns (U.S. Bureau of Census,
1976) were used to estimate the number of dry cleaning operations in
each state by type and by size (number of employees). County Business
Patterns uses size categories ranging from 1-4 employees to more than
500 employees (see Appendix A). Because the number of dry cleaners in
the large size categories is small, all these establishments are grouped
into one category representing more than 50 employees in this report.
These data were used as the basis for estimating emissions in each
state. No other data are available that permit differentiation among
size categories at the state level.
The following assumptions were used in this assessment of exposures
to perc from dry cleaning operations:
1. Dry cleaners act as point-source emitters of perc with
standard Gaussian plume dispersion.
2. Dry cleaners are uniformly distributed throughout the
urban area, and the plumes emitted from each operation do
not overlap.
3. The density of dry cleaners is proportional to population
size because decisions to open a business are generally
governed by the number of potential customers in the area.
4. An "urban area" is defined as a city with a population
greater than 25,000.
5. Consumption of perc depends on the number of employees in
a facility. Each employee accounts for the same amount
of consumption.
6. Dry cleaners in cities with less than 25,000 residents
always have less than 20 employees.
.38
-------
Table 17
DENSITY OF DRY CLEANERS
IN SELECTED CITIES
Location
Michigan
Detroit
Grand Rapids
Kalamazoo
Lansing
Colorado
Denver
Number of
Dry Cleaners
265
51
25
32
1970 Density
Population (Population/Dry Cleaner)
1,355,000
197,649
85,661
134,400
5,113
3,875
3,426
4,200
113
514,678
4,555
Source: Michigan Department of Public Health, 1977.
39
-------
7. No one resides within a circle of radius 200 m around
small operations (less than 20 employees) and within a
circle, of radius 300 m around larger operations (more
than 20 employees).
8. Urban density was uniform for all cities in a state.
(Density was calculated by summing the population and the
land area of cities in each state with more than 25,000
residents and dividing the former by the latter.)
The number of dry cleaners using perc was estimated on the basis of the
data in Table 12. For our estimates, 75% of commercial dry cleaners were
assumed to use perc. The data for coin-operated plants included laudromats
without dry cleaning facilities. Our estimates assume that 752 of
coin-operated plants have dry cleaning equipment and that all coin-operated
dry cleaners use perc. Similarly, the data for industrial plants included
plants without dry cleaning facilities. Our estimates assume that 75% of
industrial plants have dry. cleaning equipment and that 50% of the industrial
dry cleaners use perc.
In addition, the number of dry cleaners in urban areas had to be
estimated. Estimates of the number in urban areas were based on the ratio
between the urban and the rural population in each state according to 1970
census data (Bureau of.Census, 1972 County and City Data Book). This ratio
was then applied to the total number of dry cleaners in the smaller size
categories (less than 20 employees). All of the larger dry cleaners were
assumed to be located in urban areas. For example, the State of
Massachusetts has a ratio of 0.51 (i.e., 51% of the State's population
resides in cities with a population greater than 25,000). The 1974 edition
of County Business Patterns reports that Massachusetts has 171 commercial
dry cleaners (SIC 7216) with 1-4 employees. We assumed that 51% of these
would be in urban areas and that 75% of them use perc. Thus, we estimated
that 65 commercial dry cleaners in the 1-4 size category are point sources
of perc in Massachusetts. Six commercial dry cleaners have more than 50
employees. All were assumed to be located in urban areas, and 75% were
assumed to use perc; thus, there were five point sources of perc in that
size category in Massachusetts.
40
-------
The estimated number of dry cleaners of all three types using perc in
urban areas is shown in Table 18. A state-by-state breakdown is given in
Appendix A.
Emission rates for each type of operation were determined on the basis
of annual perc consumption. We assumed that all perc purchased in a given
year was lost to the atmosphere. We then divided total perc consumption by
the total number of employees for each type of dry cleaning operations. An
emission rate per employee was estimated. The emission rates estimated for
each type of dry cleaner are shown in Table 19.
Dispersion in the vicinity of each plant was determined from Figure 2
and Equation (3.3)- A computer program was written to estimate the number
of people exposed to perc in each state based on population and urban
density data from the 1972 County and City Data Book, which is the most
recent available aggregated data for cities larger than 25,000.
C. Exposures
Estimates of the population exposed to perc from dry cleaning
operations are given in Table 20. More than 352 of the U.S. urban
population is exposed to perc from commercial dry cleaners at concentrations
greater than 0.05 ppb. The estimated population exposures are much lower
for coin-operated and industrial dry cleaners. Both commercial and
industrial dry cleaners caused exposures at concentrations greater than 1.0
ppb. A detailed breakdown of these data by state is given in Appendix A.
Nine states California, Illinois, Maryland, Michigan, New Jersey, New
York, Ohio, Pennsylvania, Texas accounted for more than 70% of the total
exposures from each type of dry cleaning operation. New York had the
largest exposures, acounting for 22% of all exposures from commercial dry
cleaners, 26% of those from coin-operated dry cleaners, and 18% of those
from industrial dry cleaners.
As previously discussed, very few data are available on U.S. dry
cleaning operations. Changes in life styles and textile materials have
41
-------
Table 18
ESTIMATED NUMBER OF URBAN DRY CLEANERS USING
PERC BY SIZE OF OPERATION
Type of
Operation
By
Numbers* of Operations (Number of Employees)
1-4 5-9 10-19 20-49 50 Total*
Commercial
4.200 1,900
850
700
120 7,800
Coin-operated 3,700
Industrial
20
500
10
120
20
60
100
10 4,400
110
260
* Rounded to two significant figures,
Source: SRI estimates
42
-------
Table 19
ESTIMATED EMISSION RATES FOR EACH
TYPE OF DRY CLEANING OPERATION
Emission Rates (g/s)
Type of
Operation
Commercial
Coin-operated
Industrial
Size (.employees J
Average Number
of Employees
1-4
2
0.11
0.072
0.075
5-9
6
0.25
0.17
0.18
10-19
14
0.53
0.36
0.38
20-49
35
1.2
0.84
0.88
50
70
2.5
1.7
1.8
Source: SRI estimates
43
-------
Table 20
POPULATION EXPOSED TO PERC FROM DRY CLEANERS
Operation
Commercial
Coin-operated
Industrial
"
Population Exposed* to Perc
0.06 - 0.10
20,000,000
2,600,000
2,900,000
0.11 - 1.00
11,000,000
790,000
1,800,000
(ppb)**
1.10 :r~4TOO
41,000
4,000
Total*
Exposed
Population
31,000,000
3,400,000
4,700,000
*Rounded to two significant figures; totals do not sum vertically because
of double counting due to overlap of sources.
**
Annual average concentration; to convert to 8-hour worst case, multi-
ply by 25; to convert g/m3, multiply by 6.7.
Source: SRI estimates
44
-------
affected Che industry in recent years. The exposure projections in this
report have therefore been based on many assumptions and can only be
considered approximations. The number of people who may be exposed is
likely to slowly decrease if the number of dry cleaning plants operating in
the United States continues its current decline. The addendum to this
report contains a curuory assessment of the possible number of people
exposed to perc while washing or dry cleaning clothes at coin-operated
facilities.
45
-------
V METAL CLEANING OPERATIONS
A. Sources
1. General
Metal cleaning can be divided into three major categories: cold
cleaning, open-top vapor degreasing, and conveyorized degreasing. A cold
cleaner is a tank of solvent, usually with a cover for nonuse periods.
Inside this tank is a work surface or basket suspended over the solvent. An
open-top vapor degreaser resembles a large cold cleaner. However, the
solvent in it is heated to its boiling point, creating a zone of solvent
vapor contained by a set of cooling coils. Both the cold cleaner and the
opentop vapor degreaser are used to clean individual batches of parts at a
time; thus, they are termed "batch loaded." A conveyorized degreaser is
loaded continuously by means of various types of conveyor systems; it may
operate as either a vapor degreaser or a cold cleaner (Office of Air Quality
Planning and Standards, 1977).
The EPA Office of Air Quality Planning and Standards (OAQPS) (1977)
estimates that 1,220,000 cold cleaning units were operating in the United
States in 1974, about 70% of them devoted to maintenance and servicing
operations and the remainder used in manufacturing operations. In the same
year, an estimated 21,000 open top vapor degreasers and 3,700 vapor
degreasers were operating. Projected growth in the degreasing industry is
shown in Table 21. It is estimated that these degreasing operations used
726,000 metric tons of solvents of all types in 1974. Estimated 1974
consumption by solvent type is shown in Table 22. Estimated 1974 and 1975
consumption of halogenated degreasing solvents is shown in Table 23. The
use of perchloroethylene decreased in cold cleaning operations and increased
in vapor degreasing operations during this period. Perchloroethylene
accounts for about 15Z of the solvents used in vapor degreasing.
Consumption of perchloroethylene in degreasing is estimated at 54,000 metric
46
-------
Table 21
PROJECTED GROWTH IN SOLVENT
METAL CLEANING INDUSTRY, 1974-1985
Year
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
Cold Cleanera
1,220,000
1,250,000
1,280,000
1,310,000
1,340,000
1,380,000
1,410,000
1,450,000
1,490,000
1 , 530 ,000,
1,570,000
1,620,000
Number of Open Top
Vapor Degreaaera
21,000
24,000
27,000
29,000
32,000
35,000
39,000
42,000
46,000
50,000
54,000
58,000
Number of
Conveyorized
Degreaaera
3,700
3,900
4,200
4,400
4,600
4,900
5,100
5,400
5,700
6,000
6,400
6,700
Source: OAQPS (1977)
47
-------
Table 22
NATIONAL DECREASING SOLVENT CONSUMPTION (1974)
Solvent Consumption (10^ metric tons)
Solvent Type Cold Cleaning Vapor degreasingAll degraasing
Halogenated:
Trichloroethylenu 25 128 153
1,1,1 Trichloroethane 82 80 162
Perchloroethylene 13 41 54
Methylene Chloride 23 7 30
Trichlorotrifluoiroethane 10 20 30
153 276 429
Aliphatics , 222 222
Aromatic s:
Benzene 7
Toluene 14
Xylene 12
Cyclohexane 1
Heavy Aromatics 12
46 0 ' 46
Oxygenated:
Ketones:
Acetone 10
Methyl Ethyl Ketone 8
Alcohols:
Butyl 5
Ethers 6
29 0 29
Total Solvents: 450* 276 726
Source: OAQPS (1977)
* Includes 25,000 metric tons from conveyorized cold cleaning degreasers.
tIncludes 75,000 metric tons from conveyorized vapor degreasers.
48
-------
Table 23
U.S. HALOGENATED SOLVENT CONSUMPTION
BY TYPE OF DECREASING OPERATION (1974 and 1975)
Consumption, 10 Metric Tons
10
Cold Cleaning
Solvent
Trichloroethylene
1,1, 1-Trichloroethane
Perchloroethylene
Methylene chloride
Trichlorotri f luoroethane
1974
25
82
13
23
10
1975
5.4
100
4
0.4
NA
Vapor Degreasing
1974
128
80
41
7
20
1975
112
62
45
9
20
NA = Data not available.
Source: Mitre Corporation (1978)
-------
cons in 1974 and 49,000 metric tons in 1975. Consumption in 1976 and 1977
is estimated to have remained stable at 50,000 metric tons. Future
consumption of perchloroethylene in degreasing is difficult to estimate
because possible regulations to reduce emissions and limit worker exposure
to its vapors could affect it; however, SRI believes that its use as a
degreaser will remain stable or will increase at no more than 2Z annually.
At this rate of increase, 50,000 - 55,000 metric tons would be used in 1982.
2. Service/Maintenance Industry Degreasing
Solvent degreasing is used in both the service/maintenance and
manufacturing industries. The primary solvent degreasing users in the
service/maintenance industry include service stations, automotive repair
shops, oil well operations, maintenance for railroads, and maintenance for
civilian and military aircrafts.
The automotive repair industry includes service stations, car and
truck dealers, general and specialized auto maintenance shops, and small
engine repair facilities. The solvents used in these shops are almost
exclusively of the Stoddard or mineral spirits type, and they are used at
room or slightly higher temperatures (Leung et al., 1978a).
Oil wells also require periodic maintenance, including organic
solvent degreasing. For obvious reasons, petroleum products are widely used
for degreasing in this industry (Leung et al., 1978a).
The railroad industry follows a schedule of routine maintenance to
assure adequate vehicle performance. Locomotives must be cleaned and
degreased regularly to facilitate inspection and maintenance. According to
the Eureka survey of California railroads (Leung et al., 1978a), the
majority of maintenance cleaning is done with alkaline cleaners. Relatively
small amounts of chlorinated solvents (including perchloroethylene) are used
primarily in cleaning electrical parts such as generators. Heavy
maintenance may involve a more extensive use of chlorinated solvents; for
Southern Pacific's Sacramento heavy maintenance, the solvent was
1,1,1-trichloroethane.
50
-------
Civilian and military aircraft require periodic scheduled
maintenance to assure continuous safe performance. Degreasing is an
important aspect of the maintenance. The Eureka survey (Leung et al.,
1978a) found no perchloroethylene used on military aircraft in California.
The Eureka survey of civilian aircraft maintenance was less comprehensive.
It found that one large commercial airline uses several vapor degreasers,
but the type of solvent was not reported. The survey also found that two
small airports servicing private aircraft use Shell or Chevron solvent. The
formulations were not specified.
3. Manufacturing Industry Degreasing
The metal-working industry is the major user of solvent metal
cleaning. These industry categories are included in eight (2-digit) SIC
codes (25 and 33-39). Examples of industries in these classifications are
automotive, electronics, appliances, furniture, jewelry, plumbing, aircraft,
refrigeration, business machinery, and fasteners. In all of these
industries, organic solvents are frequently used for metal cleaning.
However, solvents are also used for metal cleaning in non-metal working
industries such as printing, chemicals, plastics, rubber, textiles, glass,
paper, and electric power. In these industries, organic solvents are
usually used for maintenance cleaning of electric motors, fork lift trucks,
printing presses, and other kinds of equipment (OAQPS, 1977).
The Eureka survey (Leung et al. 1978a) collected data on the
degreasing solvents used in industries covered by 45 industrial 3-digit SIC
codes in the state of California. These SIC codes are listed in Table 24.
Questionnaires were sent to 1,505 randomly selected manufacturers, or about
10.3Z of the 14,404 appropriate California manufacturers. Responses were
received from about 841 companies for a survey response of 56Z.
Approximately one-third of these manufacturers used an organic solvent for
degreasing.
SRI obtained the work sheets for this survey (Leung et al., 1978b)
and subsequently analyzed them to determine perchloroethylene use in
industrial degreasing. Manufacturers in 17 of the 45 SICs surveyed used
51
-------
Table 24
CATEGORIES OF MANUFACTURERS USING DECREASING SOLVENTS
en
ro
Subcategory of
Manufacturer
25 Furniture and Fixtures
251 Household Furniture
252 Office Furniture
253 Public Building & Related Furniture
254 Partitions and Fixtures
259 Miscellaneous Furniture and Fixtures
33 Primary Metal Industries
331 Blast Furnace and Basic Steel Products
332 Iron and Steel Foundries
333 Primary Nonferrous Metals
334 Secondary Nonferrous Metals
335 Nonferrous Rolling and Drawing
336 Nonferrous Foundries
339 Miscellaneous Primary Metal Products
34 Fabricated Metal Products
341 Metal Cans and Shipping Containers
342 Cutlery, Hand Tools, and Hardware
343 Plumbing and Heating, Except Electric
344 Fabricated Structural Metal Products
345 Screw Machine Products, Bolts, etc.
346 Metal Forgings and Stampings
347 Metal Services
348 Ordnance and Accessories
349 Misc. Fabricated Metal Products
35 Machinery, Except Electrical
Subcategory of
Manufacturer
351 Engines and Turbines
352 Farm and Garden Machinery
353 Construction and Related Machinery
354 Metalworking Machinery
355 Special Industry Machinery
356 General Industrial Machinery
357 Office and Computing Machines
358 Refrigeration and Service Machinery
359 Misc. Machinery, except Electrical
36 Electric and Electronic Equipment
361 Electric Distributing Equipment
362 Electrical Industrial Apparatus
363 Household Appliances
364 Electric Lighting and Wiring Equipment
365 Radio and TV Receiving Equipment
366 Communication Equipment
367 Electronic Components and Accessories
369 Misc. Electrical Equipment & Supplies
37 Transporation Equipment
371 Motor Vehicles and Equipment
372 Aircraft and Parts
373 Ship and Boat Building and Repairing
376 Guided Missiles, Space Vehicles, parts
379 Miscellaneous Transportation Equipment
38 Instruments and Related Products
381 Engineering and Scientific Instruments
382 Measuring and Controlling Devices
Source: Leung et al. (1978a).
-------
perc. The fraction of Che manufaccurers in these SIC codes that use perc
and their estimated annual consumption are given in Table 25.
B. Methodology
1. Emissions-Cold Cleaners
Emissions from cold cleaning occur during: (1) bath evaporation,
(2) solvent carry-out; (3) agitation; (4) waste solvent evaporation; and (5)
spray evaporation. The emission rates vary widely by operation. On the
basis of national consumption data, OAQPS (1977) calculated an average
national emission rate for all degreasing solvents of 0.3 metric ton per
unit per year. OAQPS (1977) estimated that cold cleaners used for
maintenance and manufacturing emit approximately 0.25 and 0.5 metric tons
per year, respectively. Data from Safety Kleen Corporation show that only
0.17 metric tons are emitted for their cold cleaners each year. However,
emissions from Safety Kleen's operations are expected to be lower than those
of others because most of Safety Kleen's waste solvent is distilled and
recycled by the company.
2. Emissions-Open-Top Vapor Degreasers
Unlike cold cleaners, open-top vapor degreasers lose a relatively
small proportion of their solvent in the waste material and as liquid
carry-out. Rather, most of their emissions are the vapors that diffuse out
of the degreaser. Open-top vapor degreasing emissions, like those from cold
cleaners, vary widely, depending on the operation. OAQPS (1977) estimates
that an average open-top vapor degreaser emits about 2.5 kg per hour per
2
m of surface area at the opening. These estimates are derived from
national consumption data on vapor degreasing solvents and from seven EPA
tests. Assuming that an average open-top vapor degreaser would have an
open-top area of about 1.67 m , a typical emission rate would be 1.2 g/s.
3. Emissions-Conveyprized Degreasing
Conveyorized degreasers may be of several types and may operate
with either cold or vaporized solvents. About 85Z of the conveyorized
degreasers are vapor types, and 15Z are nonboiling degreasers (OAQPS,
1977). Most of the nonboiling conveyorized degreasers are board cleaners.
53
-------
Table 25
ESTIMATED PERCHLOROETHYLENE USED
FOR DECREASING IN THE MANUFACTURING INDUSTRY
Amount of Perc Used
Fraction Using Perc per Plant (gal/yr)
SIC Number of Employees Number of Employees
Code* 0-20 >20 0-20 20-100 >100
331 0.00 0.11 l,800f 10,000
336 0.00 0.06 45 f 250
339 O.Op 0.14 12,000 67,200f
342 0.00 0.09 2,655f 14,752
343 0.00 0.14 2,750 15,400f
344 0.00 0.08 350 l,960f
345 0.00 0.18 863 f 4,795
347 0.08 0.11 330 5,262 17,044f
349 0.00 0.14 1,310 7,336 f
352 0.00 0.11 90 f 500
361 0.00 0.07 149 f 825
362 0.17 0.00 219
364 0.00 0.07 73f 403
366 0.00 0.11 176f 980
367 0.06 0.00 700
371 0.00 0.05 110 616f
372 0.10 0.22 10 490 25,703
industries covered by SIC codes not listed are estimated not to use
perchloroethylene in degreasing.
t_
These amounts have been extrapolated on the basis of the number of
employees in the plant.
Source: Based on data from Leung et al. (1978b)
54
-------
The average emission rate from a. conveyorized vapor degreaser is about 25
metric tons per year (0.8 g/s), whereas the average for nonboiling
conveyorized degreasers is almost 50 metric tons per year (1.6 g/s).
However, more recent designs for nonboiling conveyorized degreasers are far
more efficient than o.lder designs. It is estimated that the vapor degreaser
currently contribute about 75% of the conveyorized degreaser emissions in
the United States. Nonboiling types contribute the remaining 252 (OAQPS,
1977). It is also estimated that 75,000 metric tons were emitted from
conveyorized nonboiling degreasers (OAQPS, 1977). Evaporation, carry-out
emissions, and exhaust emissions are the primary sources.
4. Perchloroethylene emissions
The Mitre Corporation (Mitre, 1978) estimates that more than 98% of
the perc purchased annually for degreasing is emitted to the environment.
During 1975, more than 90% of the perc was used in vapor degreasing (Table
23). OAQPS (1977) estimates that only 10-20% of the virgin solvent from
conveyorized vapor degreasers, and 20-25% of the virgin solvent from
open-top vapor degreasers, is disposed of as waste solvent. Most
conveyorized vapor degreasers distill and recycle their own solvent. Used
solvents from-open-top degreasers are usually transferred to another system
or company for distillation. The preferable methods for minimizing waste
solvent evaporation into the atmosphere are distillation plants and special
incineration plants. Disposal in landfills after evaporation is also used,
but this is a less desirable method. Current waste disposal practices are
such that most of the waste can evaporate into the atmosphere. A large
fraction of perc waste is indiscriminately dumped into drains or onto the
grounds surrounding the facilities that used it. Some waste solvent is
stored in open containers; it evaporates. A small amount of waste solvent
finds its way to municpal or chemical landfill where little attempt is made
to encapsulate it (OAQPS, 1977).
5. Exposure Estimates
The results of the Eureka survey show that almost all the perc used
for metal cleaning is consumed by manufacturing industries. Consequently,
the population exposure estimates given in this report cover only emissions
55
-------
from manufacturing industries. An analysis of the Eureka survey data has
provided estimates of the fraction of manufacturers of a particular type
(SIC) that use perc for degreasing and has also provided estimates of the
amount of perc purchased by these users. These data, categorized by
manufacturer size (<20, 20-100, and >100 employees) for the industries
that use perc were sunmarized in Table 25.
Data on the number of manufacturers in each state categorized by
SIC code and by plant size, have been obtained from the 1976 Bureau of
Census report on 1974 County Business Patterns. These data are too lengthy
to be included in this report; however, the national totals are summarized
in Table 26. The number of manufacturers that use perc for metal cleaning
was estimated by multiplying the number of such manufacturers by the
fraction that used perc in the Eureka survey (Table 25). This procedure
resulted in an estimate of the number of plants in each state (by plant size
and SIC) that use perc. Table 25 also provides an estimate of the amount of
perc purchased annually for each of these plants. All of these data were
used to project the national consumption for metal cleaning in
manufacturing. This approach yielded an estimate of annual perc consumption
of 43,000 metric tons. Independent estimates indicate that 49,000 metric
tons of perc is purchased annually for all metal cleaning (Table 23), so the
national industrial manufacturing projections based on the Eureka,
California survey are remarkably accurate. Some perc is undoubtedly used in
nonmanufacturing industries, which would account for some of the
difference. In estimating atmospheric exposures, it must be remembered that
approximately 10-20% of the perc is eventually disposed of as waste and that
less than 100% of the perc in the waste enters the atmosphere causing human
exposures. Therefore, although the projections of perc use based on the
Eureka survey slightly underestimate the perc used for all degreasing, they
are assumed to be a close approximation of the amount of perc being emitted
to the atmosphere as a result of metal cleaning in industrial manufacturing.
To facilitate exposure calculations, each manufacturing plant thought
to use perc for degreasing was assigned to 1 of 13 use groupings based on
the amount of perc purchased annually. These 13 use groupings are given in
56
-------
Table 26
FACILITIES OF SELECTED SIC CODES* AND SIZES
IN THE UNITED STATES
SIC
Code
331
336
339
342
343
344
345
347
349
352
361
362
364
366
367
371
372
< 20
2,795
923
590
956
363
6,026
1,388
3,245
2,652
906
275
674
877
369
1,397
1,776
467
Number of Employees
20-100
2,322
614
389
506
183
3,221
853
1,203
1,619
511
151
405
594
582
903
883
311
>100
1,868
230
52
360
166
938
240
117
641
281
138
398
399
500
693
752
270
* These are the industries that were found by the Eureka survey to use perc.
Source: 1974 County Business Patterns
-------
given in Table 27, which also lists estimates of perc emissions (g/s)
and of the number of U.S. plants in each use grouping. The number of
plants in each use grouping was actually calculated on a state-by-state
basis for estimating population exposures.
Dispersion modeling was used to estimate annual average
atmospheric perc concentrations as a function of distance from
manufacturing plants for each of the 13 perc use groupings. The
dispersion model has previously been described (refer to Chapter III;
Figure 2), and involves,the use of the Equation (3.3). Because it was
assumed that no one resides within 0.5 km of any of the manufacturing
facilities, exposure concentrations at distance of less than 0.5 km from
a plant were not estimated. With this exclusion, it was estimated that
no population was exposed at concentrations greater than 4 ppb. The
radii to the 5 selected concentration levels for each of the 13 size
perc use groupings is given in Table 28.
In estimating exposure populations, it was assumed that all of
the manufacturers are located in cities with 25,000 or more residents.
The validity of this assumption is somewhat reinforced by the fact that
no population exposures greater than 0.01 ppb are estimated for the four
smallest plant size groups (see Table 28). The smaller manufacturing
plants are more likely to be located in the smaller cities.
Census data were obtained on the 1970 population density of
all cities with more than 25,000 residents. A uniform urban population
density within each state was assumed.
C. Exposures
The number of people exposed to perc at annual average
concentrations of 0.06-0.10, 0.11-1.00, and 1.01-4.00 ppb was estimated
for each state on the bases of concentration radii given in Table 28,
the urban population densities in the state, and the estimated number of
plants in each of the 13 perc use groupings. The resulting estimates of
population exposures to perc emissions from metal cleaning are given in
58
-------
Table 27
ESTIMATED PERC EMISSIONS FROM METAL CLEANING
IN MANUFACTURING PLANTS OF VARIOUS SIZES
Estimated
Perc Purchased* Assumed Perc Number of
(gal/yr) Emissions (g/s) Plantst
10 - 29 0.003 47
30 - 59 0.009 37
60 - 99 0.015 98
100 - 299 0.029 254
300 - 599 0.087 645
600 - 799 0.136 122
800 - 999 0.174 225
1,000 - 2,999 0.291 409
3,000 - 5,999 0.872 175
6,000 - 7,999 1.356 90
8,000-9,999 1.744 0
10,000 - 29,999 3.875 231
30,000 - 59,000 8.718 0
60,000 - 79,999 ' 13.562 7
*Purchased perc includes virgin and recycled perc as projected by the
Eureka survey results.
Estimated number of U.S. plants that use perc in degreasing.
Source: SRI estimates.
59
-------
Perc Emission
Table 28
2STIMATED ANNUAL AVERAGE ATMOSPHERIC
CONCENTRATIONS OF PERC AS A FUNCTION OF
DISTANCE FROM PLANTS USING PERC AS A DEGREASER
Distance (km) from Plant to Indicated Concentration
(g/s)
0.003
0.009
0.015
0.029
0.087
0.136
0.174
0.291
0.872
1.356
1.744
3.875
8.718
13.562
0.01 ppb
*
*
*
*
0.90
1.22
1.44
2.04
4.27
5.76
6.82
11.70
20.24
27.28
0.05 ppb
*
*
*
*
*
*
*
0.69
1.44
1.94
2.30
3.94
6.82
9.20
-------
Table 29. It is estimated that 32 million people are annually exposed as a
result of metal cleaning operations to atmospheric perc at concentrations of
0.05 to 4.0 ppb. The states in which more than 1 million people are exposed
at these concentrations are California, Illinois, Michigan, New Jersey, New
York, Ohio, and Pennsylvania.
61
-------
Tab la 29
ESTIMATED POPULATION EXPOSURES TO
ATMOSPHERIC PERC EMISSIONS FROM INDUSTRIAL DECREASING
Population* Exposed to Perc (ppb)
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Hawaii
.Idaho
Illinois
Indiana.
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
0.06-0.10
150,000
0
4,600
32,000
3,500,000
71,000
700,000
0
0
130,000
120,000
0
0
2,700,000
570,000
31,000
130,000
170,000
6,600
420
180,000
370,000
1,600,000
, 180,000
5,400
120,000
0
12,000
0
230
1,500,000
0
2,400,000
77,000
0
1,700,000
25,000
105,000
3 , 200 , 000
97,000
32,000
0
66,000
0.11-1.00
88,000
0
1,500
18,000
2,100,000
40 , 000
410,000
0
0
76,000
69,000
0
0
1,600,000
340,000
17,000
82,000
93,000
2,100
0
94,000
200,000
96 , 000
100,000
1,800
71,000
0
5,700
0
0
380,000
0
1,400,000
43,000
0
960,000
14,000
60,000
1,900,000
51,000
18,000
0
38,000
1.01-4.00
270
0
0
50
42,000
100
5,100
0
0
230
220
0
0
15,000
4,800
50
260
230
0
0
200
500
9,800
240
0
190
0
0
0
0
13,000
0
3,700
120
0
2,700
40
170
18,000
100
40
0
110
62
-------
Table 29 Continued
Population Exposed* to Perc (ppb)
State
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
U.S. Total
0
20
.06-0.10
290,000
0
0
8,600
130,00
0
55,000
0
,000,000
0.11-1.00
160,000
0
0
3,600
80,000
0
32,000
0
11,000,000
1.01-4.00
430
0
0
0
250
0
90
0
120,000
, *Rounded to two significant figures.
-4-
'Annual average concentrations; to convert to 8-hour worst case,
multiply by 25; to convert to Mg/nH, multiply by 6.7.
Source: SRI estimates.
63
-------
BIBLIOGRAPHY
1. Appleby, Alan, "Atmospheric Freons and Halogenated Compounds," Rutgers
University, report to the Environmental Sciences Research Laboratory,
U.S.. Environmental Protection Agency, NTIS PB-262 432, pp. 10-12,
107-127, 264-271 (November 1976).
2. Battelle-Columbus Laboratories, "Determination and Evaluation of
Environmental Levels of Methyl Chloroform and Trichloroethylene,"
report to the U.S. Environmental Protection Agency (March 1977).
3. Dilling, W. L., N. B. Tefertiller, and G. J. Kallos, "Evaporation
Rates and Reactivities of Methylene Chloride, Chloroform,
1,1,1-Trichloroethane, Trichloroethylene, Tetrachloroethylene, and
Other Chlorinated Compounds in Dilute Aqueous Solutions,"
Environmental Science and Technology, Vol. 9, No. 9, pp. 833-838
(September 1975).
4. Dusoleil, S., P. Goldfinger, A. M. Malieu-Van der Auwera, G. Martens,
and D. Van Der Auwera, Trans Faraday Society, Vol. 57, p. 2197 (1961)
5. Evatovich, Mark, San Diego Air-Pollution Control District, personal
communication, San Diego, California. (October 1978).
6. Fisher, William E., International Fabricare Institute, Research
Division, Silver Spring, Maryland, paper presented at the EPA
Hydrocarbon Control Workshop, Chicago, Illinois (20 July 1977).
7. Franklin Institute Research Laboratories, "Preliminary Study of
Selected Potential Environmental Contaminants," report to the U.S.
Environmental Protection Agency, NTIS No. PB-243 910 (July 1975).
%
8. Fuller, B. B., "Air pollution Assessment of Tetrachloroethylene,"
Mitre Corporation, report to the U.S. Environmental Protection
Agency, Contract No. 68-02-1495 (February 1976).
9. Goldfinger, P., G. Huybrechts, and G. Martens, Trans Faraday Society,
Vol. 57, p. 2210 (1961).
10. Greer, Jack K., Jr., Pollutant Strategies Division, Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency,
personal communication (December 1978).
11. Hawley, G. E. (editor), The Condensed Chemical Dictionary, Van
Nostrand Reinhold Company, Ninth Edition (1977).
12. Horowitz, A., and L. A. Rajbenback, Journal of the American Chemical
Society, Vol. 9, p. 4105 (1968).
64
-------
13. Huybrechts, J., T. Olbrechts, and K. Thomas, Trans Faraday Society,
Vol. 63, p. 1647 (1967).
14. International Fabricare Institute, "Dry Cleaning Solvent Emissions in
the State of California 1974-1975," Joliet, Illinois (June 1975).
15. Kleeberg, Charles, Emission Standards and Engineering Division, Office
of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, personal communication (October 1978).
16. Kuriyang, T., Chemical Abstracts, Vol. 67 (1967).
17. Lapp, Thomas W. , Betty L. Herndon, Charles E. Mumma, and Arthur D.
Tippit, "An Assessment of the Need for Limitations on
Trichloroethylene, Methyl Chloroform, and Perchloroethylene," draft
final report to the Office of Toxic Substances, U. S. Environmental
Protection Agency, Contract 68-01-4121, pp. 47-51, 133-135
(September 1977).
18. Leung, S., R. Johnson, C. S. Liu, G. Palo, R. Peter, and T. Tanton,
"Alternatives to Organic Solvent Degreasing", Eureka Laboratories,
Sacramento, CA, ARB A6-206-30 (May 1978a).
19. Leung, S., R. Johnson, C. S. Liu, G. Palo, R. Peter, and T. Tanton,
"Alternatives to Organic Solvent Degreasing Supplements," Eureka
Laboratories, Sacramento, CA, ARB A6-206-30 (May 1978b)
20. Lillian, D., and H. B. Singh, "Absolute Determination of Atmospheric
Halocarbons by Gas' Phase Covlimetry," Analytical Chemistry, Vol. 46,
No. 8, pp. 1060-1063 (1974).
21. Mara, Susan J., and Shonn S. Lee, "Assessment of Human Exposures to
Atmospheric Benzene," SRI International, Menlo Park, California,
report to the U.S. Environmental Protection Agency, NTIS No. PB 284
203/AS (May 1978).
22. Michigan Department of Public Health, Bureau of Environmental and
Occupational Health, "Computer Listing of Dry Cleaners Licensed in
the State of Michigan," Lansing, Michigan (November 1977).
23. Midwest Research Institute, "Test of Industrial Dry Cleaning
Operations," draft final report, U.S. Environmental Protection
Agency, Contract No. 68-02-1403 - Task 21 (April 1976a).
24. , "Source Test of a Dry Cleaner," draft final report, U.S.
Environmental Protection Agency, Contract No. 68-02-1403 Task 23
(May 1976b).
25. Mitre Corporation, "Development of Standards of Performance for
Solvent Metal Cleaning (Degreasing)," draft document prepared for
the U.S. Environmental Protection Agency (July 1978).
65
-------
26. Reid, F. H., and W. R. Halpin, Amer. Ind. Hygiene Assoc., Vol. 29, p.
390 (1968).
27. Scott Environmental Technology, Inc., "A Survey of Perchloroethylene
Emissions from a Dry Gleaning Plant," report for the U.S.
Environmental Protection Agency, No. 76-DRY-l, Research Triangle
Park, North Ciurolina (March 1976).
28. Simmonds, P- G., S. L. Kerrin, J. E. Lovelock, and R. H. Shair,
"Distribution of Atmospheric Halocarbons in the Air Over the Los
Angeles Basin," Atmospheric Environment, Vol. 8, No. 3, pp. 209-216
(1974).
29. Suta, B. E., "BESTPOP: A Fine-Grained Computer System for the
Assessment of Residential Population," SRI International, Menlo
Park, California (1978).
30. Tuttle, T. C., G. D. Wood, C. B. Grether, and B. L. Johnson, "A
Behavioral and Neurological Evaluation of Dry Cleaners Exposed to
Perchloroethylene," report to the National Institute of Occupational
Safety and Health, Center for Disease Control, Cincinnati, Ohio
(June 1977).
31. U.S. Department of Commerce, Bureau of the Census, "County Business
Patterns - 1970," CBP-70 Series, Government Printing Office,
Washington, D.C. (December 1972).
32. ___, "1972 County and City Data Book," Government Printing
Office, Washington, D.C. (1973).
33. , "County Business Patterns - 1974," CBP-74 Series,
Government Printing Office, Washington, D.C. (December 1976).
34. U.S. Environmental Protection Agency, Environmental Monitoring and
Support Laboratory, "Monitoring for Perchloroethylene in Selected
Cities," draft report (December 1978).
35. , Office of Air Quality Planning and Standards, "Control of
Volatile Organic Emissions from Solvent Metal Cleaning,"
EPA-450/2-77-022 (November 1977).
36. Williams, I. H., Analytical Chemistry, Vol. 37, p. 1723 (1965).
66
-------
Appendix A
NUMBER OF DRY CLEANERS IN URBAN AREAS AND
EXPOSED POPULATION FROM DRY CLEANERS, BY STATE
A-l
-------
Table A-l
ESTIMATED URBAN COMMERCIAL DRY CLEANERS
THAT USE PERC, BY STATE
Number of Plants
by Size of Operation (Employees)
Total
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of
Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Mis souri
Montana
Nebraska
1-4
67
1
38.
24
566
62
65
4
38
106
65
8
3
217
82
43
50
32
75
3
35
128
169
51
29
96
8
21
5-9
28
1
17
10
247
35
29
2
26
71
44
2
3
126
41
15
17
18
29
1
27
64
92
25
14
35
3
12
10-19
14
12
5
101
13
14
2
15
27
20
4
2
59
18
5
4
5
14
1
18
34
51
9
4
18
1
5
20-49
9
1
8
2
44
10
17
4
5
27
25
2
1
36
14
5
4
9
6
5
21
21
30
11
7
9
2
4
50+
2
2
2
2
8
1
1
1
5
2
2
8
5
1
4
2
2
5
4
3
2
2
Number of Plants
120
5
77
43
966
121
126
12
85
236
156
18
14
446
160
69
75
68
126
10
103
252
346
99
54
160
14
44
A-2
-------
Table A-l (Concluded)
Number of Plants
by Size of Operation (Employees)
Total
State
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total U.S. Plants
1-4
9
7
163
19
683
63
6
198
59
31
126
20
18
5
74
381
13
2
77
65
11
69
4
4,194
5-9
4
3
58
7
228
39
2
107
19
11
68
10
11
2
38
144
6
47
25
7
34
2
1,906
10-19
5
1
21
5
74
19
2
47
10
2
36
4
7
1
18
68
2
1
26
7
3
11
1
846
20-49
14
2
23
2
31
23
2
41
6
5
45
3
14
1
19
65
8
28
5
5
10
~M^MB
691
50+
--
2
10
3
1
7
3
7
2
2
1
10
1
2
2
..MMH
119
Number of Plants
32
13
267
33
1,026
147
13
400
97
49
282
39
52
9
150
668
30
3
180
102
26
126
7
7,756
Source: SRI estimates.
A-3
-------
Table A-2
ESTIMATED URBAN COIN-OPERATED DRY CLEANERS
THAT USE PERC, BY STATE
Number of Plants
by Size of Operation (Employees)
Total
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of
Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Mis souri
Montana
Nebraska
1-4
61
4
54
27
269
59
33
2
25
195
45
14
9.
308
138
47
50
46
37
6
45
83
175
43
22
90
7
18
5-9
4
1
3
2
47
8
6
1
5
20
5
.2
1
' 59
32
4
6
6
2
2
8
10
53
8
14
2
2
10-19
2
2
12
2
2
1
'3
1
1
14
5
1
2
1
>- __ .
2
5
10
2
4
~
1-
20-49 50+
1
1
1
4 2
1
__
2 1
3
4 2
Y
1
1
2
4
3
4
2 TIT
I
1
Number of Plants
63
5
65
30
334
70
41
3
31
221
54
17
10
387
176
53
58
53
42
8
59
101
242
55
23
109
9
21
A-4
-------
Table A-2 (Concluded)
Number of Plants
by Size of Operation (Employees)
Total
State
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total U.S. Plants
1-4
13
11
83
29
398
59
4
209
71
22
109
23
20
7
105
410 ,
13
2
68
44
14
64
5
3,695
5-9
1
1
8
3
32
5
1
42
6
4
13
4
1
5
25
2
3
7
2
11
499
10-19
2
8
1
9
1
2
5
2
5
1
2
2
1
5
119
20-49 50+
1
2 1
6 2
i
3
2 1
2
1
2 1
1
1 "
^M^B ^^^»
57 12
Number of Plants
15
12
96
32
446
66
5
263
78
28
130
27
23
8
112
443
16
2
78
54
17
81
5
4,382
Source: SRI estimates,
A-5
-------
Table A-3
ESTE1ATED .URBAN INDUSTRIAL DRY CLEANERS
THAT USE PERC, BY STATE
Number of Plants
by Size of Operation (Employees)
Total
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Florida
Georgia
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
~J> 5-9 10-19 20-49
'- -- - 2
1. 1
2
3 2 3 12
11 1
1
11 4
5
1 5
1 2
2
. i __ 3
, 3
^Ka _ ^ ^
11 2
12 6
11 1
. »<» 1
^^ <«^» O
^w »« ^^ *7
]_
1 2
232 7
50+
2
2
1
13
1
1
6
3
7
4
2
2
3
3
4
5
1
1
3
1
4
1
6
Number of Plants
4
4
3
33
4
2
12
8
13
7
4
4
5
6
3
8
14
4
2
5
2
1
1
7
1
20
A-6
-------
Table A-3 (Concluded)
Number of Plants
by Size of Operation (Employees)
Total
State
North Carolina
Ohio
Oklahoma
Oregon
Pennsylvania
South Carolina
Tennessee
Texas
Virginia
Washington
West Virginia
Wisconsin
Total U.S. Plants
1-4 5-9
_
2 1
1
__
1 2
16 11
10-19
2
1
1
1
2
18
20-49
6
3
2
1
3
2
6
1
3
_1
98
50+
4
6
1
7
2
11
3
1
1
112
Number of Plants
10
14
3
2
12
4
1
22
4
3
1
2
255
Source: SRI estimates.
-------
Table A-4
ESTIMATED POPULATION EXPOSED TO PERC FROM
COMMERCIAL DRY CLEANERS IN URBAN AREAS
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of
Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Population Exposed*to Perc (ppb)**
0.06-0.10 0.11-1.00 1.01-4.00
Total
Exposed
Population*
110,000
- 45,000
110,000
55,000
1,500,000
240,000
210,000
74,000
500,000
290,000
330,000
59,000
16,000
1,800,000
270,000
60,000
61,000
290,000
170,000
20,000
750,000
540,000
920,000
220,000
68,000
180,000
29,000
61,000
27 ,000
60,000
30,000
770,000
120,000
110,000
44,000
260,000
160,000
180,000
34,000
8,000
970,000
150,000
31,000
29,000
160,000
84,000
12,000
430,000
290,000
490,000
120,000
36,000
94,000
, 15,000
300
400
300
300
2,400
300
200
900
700
400
400
4,300
1,000
100
1,400
400
1,200
1,400
1,500
700
400
170,000
72,000
170,000
85,000
2,300,000
360,000
320,000
120,000
760,000
450,000
510,000
93,000
24,000
2,800,000
420,000
91,000
90,000
450,000
250,000
32,000
1,200,000
830,000
1,400,000
340,000
100,000
270,000
44,000
A-8
-------
Table A-4 (Concluded)
State
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota ,
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total Exposed*
Population Exposed*to
0.06-0.10 0.11-1.00
120,000
78,000
9,800
890,000
280,000
4,700,000
280,000
55,000
1,100,000
44,000
70,000
1,900,000
130,000
110,000
13,000
130,000
800,000
84,000
3,800
200,000
110,000
55,000
220,000
7,200
67,000
47,000
5,400
460,000
140,000
2,200,000
160,000
32,000
600,000
24,000
36,000
1,100,000
69,000
62,000
6,700
69,000
430,000
49,000
1,800
110,000
51,000
31,000
120,000
3,100
Perc (ppb)**
1.01-4.00
500
1,100
10,000
600
300
2,400
200
4,400
600
300
100
1,600
200
200
600
^^»
Total
Exposed
Population*
190,000
130,000
15,000
1,400,000
420,000
6,900,000
440,000
87,000
1,700,000
68,000
110,000
3,000,000
200,000
170,000
20,000
200,000
1,200,000
130,000
5,600
310,000
160,000
86,000
430,000
10,000
20,000,000 11,000,000 41,000
31,000,000
Source: SRI estimates.
*Rounded to two significant figures.
**Annual average concentrations; to convert to M8/m > multiply by 6.7; to
convert to 8-hour worst case, multiply by 25.
A-9
-------
Table A-5
ESTIMATED POPULATION EXPOSED TO PERC FROM
COIN-OPERATED DRY CLEANERS IN URBAN AREAS
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of
Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Population Exposed*to Perc (ppb)**
0.06-0.10 0.11-1.00 1.01-4.00
13,000
1.200
15,000
7,000
180,000
40,000
11,000
1,800
33,000
44,000
26,000
4,300
1 1,900
360,000
47,000
11,000
13,000
37,000
18,000
1,000
93,000
52,000
150,000
27,000
7,000
30,000
3,000
3,200
200
4,200
2,100
61,000
14,000
2,600
400
7,000
12,000
9,800
1,100
200
120,000
12,000
2,800
2,500
13,000
6,700
200
39,000
18,000
49,000
9,700
2,200
7,700
500
«MI»
--
Total
Exposed
Population*
16,000
1,400
19,000
9,100
240,000
54,000
14,000
2,200
40,000
56,000
36,000
5,400
2,100
480,000
69,000
14,000
16,000
50,000
25,000
1,200
132,000
70,000
200,000
37,000
9,200
38,000
3,500
A-10
-------
Table A-5 (Concluded)
Population Exposed*to Pare (ppb)**
0.06-0.10 0.11-1.00 1.01-4.00
7,100
4,800
1,100
110,000
4,700
700,000
16,000
1,600
140,000
4,700
10,000
160,000
8,900
8,100
5,700
12,000
84,000
4,400
400
10,000
22,000
5,200
35,000
1,100
1,500
1 , 900
90
39,000
400
200,000
3,800
200
38,000
500
2,900
55,000
1,100
3,400
2,600
1,300
16,000
1,100
1 , 700
6,600
1,300
11,000
^^ ~
Total
Exposed
Population*
3,500
6,700
1,200
150,000
5,100
900,000
20 , 000
2,000
180,000
5,200
13,000
220,000
10,000
12,000
8,300
13,000
100,000
5,500
400
12,000
29,000
6,500
46,000
1,100
Total Exposed*
2,600,000
790,000
3,400,000
Source: SRI estimates.
*Rounded off to two significant figures.
**Annual average concentrations; to convert to
convert to 8-hour worst case, multiply by 25.
m^ , multiply by 6.7; to
A-ll
-------
Table A-6
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Florida
Georgia
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
Ohio
ESTIMATED POPULATION EXPOSED TO PERC FROM
INDUSTRIAL DRY CLEANERS IN URBAN AREAS
Population Exposed*to Perc (ppb)**
0.06-0.10 0.11-1.00 1.01-4.00
Total
Exposed
Population*
22,000
21,000
15,000
320,000
27,000
16,000
62,000
61,000
300,000
53,000
22,000
14,000
63,000
'52,000
100,000
79,000
160,000
21,000
14,000
41,000
12,000
3,100
5,600
150,000
8,900
540,000
67,000
150,000
13,000
13,000
9,100
200,000
15,000
9,900
37 , 000
38,000
170,000
33,000
14,000
7,500
39,ooa
32,000
66,000
47,000
94,000
12,000
9,000
26,000
7,000
1,800
3,600
93,000
5,700
310,000
41,000
88,000
30
30
20
400
30
20
90
70
400
90
30
80
70
200
100
200
30
20
60
--
10
200
20
700
80
200
35,000
34,000
24,000
520,000
72,000
26,000
99,000
99,000
470,000
86,000
36,000
22,000
100,000
84,000
170,000
130,000
250,000
33,000
23,000
67,000
19,000
4,900
9,200
240,000
15,000
850,000
110,000
240,000
A-12
-------
Table A-6 (Concluded)
State
Oklahoma
Oregon
Pennsylvania
South Carolina
Tennes see
Texas
Virginia
Washington
West Virginia
Wisconsin
Population Exposed*to Perc (ppb)**
0.06-0.10 0.11-1.00 1.01-4.00
6,100
8,600
300,000
21,000
1,200
130,000
23,000
17,000
13,000
20,000
3,700
3,300
180,000
13,000
78,000
14,000
9,800
8,100
13,000
10
500
30
200
40
30
30
Total
Exposed
Population*
9,800
12,000
480,000
34,000
1,200
210,000
37,000
27,000
21,000
33,000
Total Exposed*
2,900,000 1,800,000
4,000
4,700,000
Source: SRI estimates.1
*Rounded off to two significant figures.
**Annual average concentrations; to convert to Mg/» , multiply by 6.7; to
convert to 8-hour worst case, multiply by 25.
A-13
------- |