Human Exposures to Atmospheric
Arsenic: Final Report

-------
                              Final Report
HUMAN EXPOSURES
TO ATMOSPHERIC ARSENIC
September 1978
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development and
Office of Air Quality Planning and Standards
Washington, D.C. 20460

-------
HUMAN EXPOSURES TO
ATMOSPHERIC ARSENIC
Final Report
'\.
September 1 978
~"
Prepared for:

U.S. Environmental Protection Agency
Office of Research and Development and
Office of Air Quality Planning anq Standards
Project Officers:

Alan P. Carlin
Joseph D. Cirvello
Project Monitors:

J. S. Cooper
Ken Greer
Contracts 68-01-4314 and 68-02-2835
SRI Projects EGU-5794 and CRU-6780
Prepared by: Benjamin E. Suta

Center for Resource and
Envtronmental Systems Studies
CRESS Report No. 50

-------
NOTICE
This is a final.report. It has been released by the U.S. Environ-
mental Protection Agency (EPA for public review and comment and does
not necessarily reflect Agency Policy. This report was provided to EPA
by SRI International, Menlo Park, California, in partial fulfillment
of Contract Nos. 68-01-4314 and 68-02-2835. The contents of this report
are reproduced herein as received by SRI after comments by EPA. The
opinions, findings, and conclusions expressed are those of the authors
and not necessarily those of EPA. Mention of company or product names
is not to be considered as an endorsement by the EPA.
iii

-------
- , . .
CONTENTS
LIST OF ILLUSTRATIONS.
.....
...,.,..
LIST OF TABLES '. .
. '. . .
. . .. .
. . '. .
. . . .
I INTRODUCTION.
I I SUMMARY
.. . . . . . .
. .. ..
. . . .
. . . .
. .' .
. . . .
.....
......
. . . .
. . . . . .
. . . .
. . . .
III SOURCES AND BEHAVIOR OF ARSENIC IN THE ENVIRONMENT., ~
" IV BACKGROUND CONCENTRATIONS OF ARSENIC IN 'THE

ATM.OSPHERE. . . . . . . . . . . . . . . .
. .,.
......
V ARSENIC EXPOSURES FROM COAL-FIRED POWER PLANTS. . .. . ...
 A.
,-' B.
 C.
 D.
 E.
General. . .
......
. . . .
.....
Coal-Fired Power Plants.
. . ... ..
..........
. . . .
Power Plant Atmospheric Emissions.
Population Exposures
. . . . . . .
Summary. .
.....
. . . .
.....
VI ARSENIC EXPOSURES FROM NONFERROUS SMELTERS. . .
A.
General. . .
. . . .
. . . . . . .
. . . p. . . -. .
. . . . .
......
.......
C.
Primary Nonferrous Smelter Locations. . .

Sampling of Atmospheric Arsenic Concentrations Near
Nonferrous Smelters. . . . . . . . .
B.
D.
Atmospheric Emissions.
Human Exposures. . . .
. . . .
.....
E.
.....
F.
Summary. . . . . . .
. . . . .
G.
Secondary Nonferrous Smelters. .
VII ARSENIC EXPOSURES FROM PESTICIDE MANUFACTURERS.
A.
General. . .
. . . .
......
. . . .
B.
C.
Pesticide Manufacturers. . . .
Pesticide Plant Emissions.
Population Exposures.
. . . .
D.
E.
. . . .
. . . . . .
Summary. .
.....
. . . . . . . . . . .
v
.....
. . . .
. . .
.....
. . . .
......
. . . .
. . . . . .
......
vii
ix
1
3
9
15
33
33
33

33

38
39
41

41

41
46

46

57
64

65
69
69

69
69
73

76

-------
VIII ARSENIC EXPOSURES FROM COTTON GINS.
. . .
A.
General. . . .
. . . .
. .
. . .
B.
Cotton Gins. .
. . .
. . . . .
. . . .
C.
Cotton Gin Emissions. .
. . . .
. . . .
D.
Population Exposures. . .

SUlIID18.ry. . . . . . . . . .
. . .
. . . . . . . .
E.
IX ARSENIC EXPOSURES FROM GLASS MANUFACTURERS.
A.
General. . . . . . . . .
. . . .
. . .
. . . .
. . .
. . . .
. . .
B.
Glass Manufacturers. .
. . . . .
. . .
. . . . .
. . .
. . . . .
. . .
C.
Glass Manufacturing Emissions.
Population Exposures. . . . .
. Summary. .
. . . .
. . . . .
. . .
. . '.
D.
E.
. . . . .
. . . .
. . . . .
. . . . .
. . .
. . .
. . . . .
. . .
. . . .
. . . .
. . .. .
. . . .
X SECONDARY HUMAN EXPOSURES RESULTING FROM ATMOSPHERIC
ARSENIC EMISSIONS. .. . . . . . . .. . . . . . . .
A.
General. .. .
. . . .
. . .
. . . .8 .
. . .
. . . .
B.
Exposures.
. . .
. . . . .
. .
.....
. . . .
Appendix--DISPERSION.ESTIHATES OF ATMOSPHERIC ARSENIC
CONCENTRATIONS FOR SELECTED SOURCES. . . . . ... .
BIBLIOGRAPHY.
. ..
. . .
. .
. .. .
vi
. .
. . . . . .
79
79
80
80
84
87
89
89
90
92
9S
96
99
99
99
103
105

-------
III-l
III-2
IV-l
VI-l

VI-2
VI-3
VI-4
IX-l
ILLUSTRATIONS
The Generalized Geochemical Cycle for Arsenic.

The Proposed Biological Cycle for Arsenic. . .

Atmospheric Arsenic Concentrations for Exposed
Urban Populations. . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
Domestic Primary ~onferrous Smelters. . . . .
. . . .
Atmospheric Arsenic Concentrations as a Function of
Distance from the El Asarco Smelter. . . . . . . . . .

Atmospheric Arsenic Concentrations as a Function of
Distance from the Tacoma Smelter. . . . . . . . . . .
Atmospheric Arsenic Concentrations as a Function of
Distance from the East Helena Smelter. . .

Probability Di~tribution of Particle Sizes Present
in Glass Furnace Effluents. . . . . . . . . . .
vii
12
13
31

42
48
50
52
94

-------
I!-l
III-l
III-2
IV-l
IV-2
V-I
V-2
V-3
V-4
V-5
V-6
VI-l
VI-2
VI-3
VI-4
VI-5
VI-6
VI-7
VI-8
TABLES
Summary of H0man Papulation Ex?osures to Arsenic
for Selected Emission Sources. . . . . . . . . .

Arsenic Pollutant Sources.
. .. . .
.....
.....
Environmental Dissipations of Arsenic (1974) .
. . . .
1974 Atmospheric Arsenic Concentrations for U.S.
Loea t ions. . . . . . . . . . . . . . . . . . . .
Locations Having Annual Average Atmos~heric Arsenic
Concentrations in Excess of 0.01 ~g/m . . . . . ~ .
Coal Use by Source (1974).
..........
. . . .
Arsenic Released from Coal Consumption--1974
.....
State Totals of Coal-Fired Power Plants and Fuel
Capacity, Based on EPA Listing. . . . . . . . . . . .

Arsenic Concentration of Airborne Fly Ash Particles
Emitted by Coal-Fired Power .Plants. . . . . . .
Estimated Annual Average Atmospheric Arsenic
Concentrations for Coal-Fired Power Plants Based
on Dispersion Modeling. . . . . . . . . .

Maximum Average Annual Atmospheric Arsenic
Concentrations in the Vicinity of Large
Power Plants. . . . . . . . . . . . . . .
......
Primary U.S. Copper Smelters.
.....
.......
Primary U.S. Lead
Smelters. . . . . . . . .
.....
Primary U.S. Zinc Smelters
..............
EPA Atmospheric Arsenic S~ples Taken Near
Nonferrous Smelters. . . . . . . . . . . . . .
. . . .
Average Atmospheric Arsenic Concentrations for
El PasQ~ Texas. . . . . . . . . . . . . . . .
. . . .
Atmospheric Arsenic Concentrations for Tacoma,
Washington. . . . . . . . . . . . . . . . . . .
Atmospheric Arsenic Concentrations Recorded for
The Helena Valley Environmental Pollution Study.

Atmospheric Dispersion Modeling for Arsenic
Concentrations in the Vicinity of Copper Smelters
due to. Stack Emissions Alone. . . . . . . . . . . . .
ix
5
10

11
16
29
34
34
35
37
38
40

43

44

45
47
47
49
51
53

-------
VI-9
VI-10
VI-ll
VI-12
VI-13
VI-14
VI-15
VI-16
VI-l7
VII-l
VII-2
VII-3

VII-4
VII-5
VII-6
VII-7
VIlI-l
VI II - 2
VIII-3
VIII-4
Atmospheric Dispersion Modeling for Arsenic
Concentrations due Solely to Fugitive Emissions
in the Vicinity of Copper Smelters. . . . . . .

Atmospheric Dispersion Modeling for Arsenic
Concentrations in the Vicinity of a Lead Smelter
Atmospheric Dispersion Modeling for Arsenic
Concentrations in the Vicinity of a Zinc Smelter. . .

Estimated Human Population Exposures to Atmospheric
Arsenic Emitted by Copper Smelters. . . . . . . . . .
Estimated Human Population Exposures to Atmospheric
Arsenic Emitted by Lead Smelters. . . . . . . . . . .

Estimated Human Population Exposures to Atmospheric
Arsenic Emitted by Zinc Smelters. . . . . . . . . . .

Comparison of Measured and Predicted Atmospheric
Arsenic Concentrations Near Copper Smelters. . .
Comparison of Measured and Predicted Atmospheric


Arsenic Concentrations Near Lead and Zinc


Sme 1 t e r s . . . . . . . . . . . . . . . . . . . .
Estimated Arsenic Retained in or Added to Copper,
Lead, and Zinc Metals. . . . . . . . . . . . . .
Arsenical Compounds and Their Use. . . .

Annual Use of Arsenic in Major Arsenical
Pesticides. . . . . . . . . . . . . . .
. . . .
.......
Manufacturers of Arsenical Pesticides. .
. . . .
Annual Atmospheric Arsenic Emissions for Pesticide
Manufacturers,. . . . . . . . . .. . . . . . . . . . .
Estimated Arsenic Use and Atmospheric Emissions for
Major Hypothetical Arsenical Pesticide
Manufacturers. . . . . . . . . . . . . . . . . .
Estimated Arseni~ Concentrations for Various
Distances for a Manufacturer Having Emis~~os of
207 lb/yr Arsenic. . . . . . . . . . . . . . . . . . .
Estimated Population Exposed to Atmospheric Arsenic
Emit ted from Pesticide Manufacturing. . . . . . . . .
Cotton Gins in the United States,
1972 . . . . .
Comparison of Atmospheric Arsenic Concentrations


Near Cotton Gins Based on Modeled and Measured


Me t hod s . . . . . . . . . . . . . . . . . . . . .
. . .
Assumed Distribution of Production Rates for
Cotton Gins. . . . . . . . . . . . . . . . .
.....
Distribution of Emission Controls for Cotton Gins. . .
x
S4
SS
SS
59
61
62
63
64
66
70
71
72
73
74
75
76
81
82
83
83

-------
VIlI-5
VIlI- 6
VIlI-7
VllI-8
IX-I
IX-2
IX-3
IX-4
X-I
Estimated Number of Gins Processing Machine
Stripped Cotton in Te~as and Oklahoma. . . .
. . . . .
Estimated Atmospheric Arsenic Concentrations (~g/m3)
for Various Distances from Gins Processing
Arsenic Desiccated Cotton. . . . . . . . . . . .
Estimated Human Population Exposures to Arsenic
Emitted from Cotton Gins During Ginning Season.

Estimated Annual Average Human Population
Exposures to Arsenic from Cotton Gins. . . . . .
Estimated Production of Pressed and Blown Glass
by Number of Plants. . . . . . . . . . . . . . .
. . .
Stack Emissions for Glass Manufacturing. .
......
Atmospheric Arsenic Concentrations for Glass
Manufacturers Based on Dispersion Modeling.
.....
Estimated Human Population Exposures to
Atmospheric Arsenic Concentrations from Glass
Manufacturing. . . . . . . . . . . . ~ . . . .
. . . .
Arseni~ Concentrations in Drinking Water of
Cities Near Nonferrous Smelters. . . . . . .
.....
xi
83
85
86
87
91
92
95
97
101

-------
I
INTRODUCTION
The primary objective of this study has been to quantify the expo-
sure of the United States population to arsenic in the atmosphere. The
seven primary sources of atmospheri~ arsenic evaluated "are coal-fired
power plants; copper, lead, and zinc smelters; pesticide manufacturers;
cotton gins; and glass manufacturers. In addition, general ambient
background concentrations were analyzed to place the contributions by
the seven sources in proper perspective. To estimate exposures, we have
located and characterized plants for the seven sources, indicated at-
mospheric environmental concentrations of pollutants resulting from plant
activities, and estimated the human population exposed to various levels
of these pollutant concentrations.
Substantial evidence, both direct and indirect, indicates that
arsenic is carcinogenic. Current U.S. Environmental Protection Agency
(EPA) policy states that there is no zero risk level for carcinogens.
To determine what regulatory action should be taken by EPA on atmospheric
emissions of arsenic, three reports have been prepared: (1) a health
effects assessment, (2) a population exposure assessment, and (3) a risk
assessment document based on the data in the first two assessments. This
is the document on population exposure assessment. On their review of
earlier drafts of the three reports, the EPA Science Advisory Board,
recommended that
... data essential for determining the potential hazard of
arsenic emissions from certain point sources, such as glass
manufacturing plants and cotton ginning operations, be collected
and made available for critical analyses. There is also a need
for additional monitoring data from regions with primary smelters.
1

-------
II
SUMMARY
Arsenic occurs in the atmosphere both as a result of man's activities
and from natural sources. Anthropogenic sources can be divided into two
groups: (1) from arsenic occurring naturally in many commercial raw
materials and released when those materials are processed, and (2) from
arsenic used to manufacture commercial products or from arsenic added to
commercial products. Products such as arsenical pesticides may cause
atmospheric contamination as a result of their manufacture and of their
use. Examples of man-made sources include nonferrous smelters, pesti-
cide manufacture and use, the combustion of fossil fuels (primarily
coal), glass manufacturing, cotton ginning, and the lead alloy industry.
Natural sources of atmospheric arsenic include volcanic action, hot
springs, decay of plant matter and other microbial actions, .and the
weathering of minerals within soils.
Arsenic exists in several forms in the atmosphere. Some of these
forms are more toxic than others. For example, trivalent arsenic is the
most toxic form, whereas pentavalent arsenic is only slightly toxic.
Arsine, a poisonous gas primarily associated with occupational hazards,
is extremely toxic, whereas metallic arsenic is nontoxic.
\
Residence times for arsenic in the atmosphere have not (insofar as
we have been able to determine) been reported in the literature. Arsenic
absorbed onto particulate matter returns to earth with the particulates.
It has been hypothesized that when volatile arsenicals are released to
the atmosphere, they are oxidized in the presence. of light and return to
the earth by way of dry deposition or precipitation (Union Carbide, 1976).
Most compounds of arsenic, when heated in the presence of air, are
converted to arsenic trioxide. Because arsenic trioxide sublimes at
193°C, it is easily suspended on small particles in the air (Sullivan,
1969). For this reason, atmospheric emission of arsenic from processes
in which heat is involved are generally in the form of arsenic trioxide.
This would include smelters, glass manufacturers, and coal-fired power
plants. Because arsenic acid is used to desiccate cotton, the arsenic
emissions from cotton gins are presureably the residues of the acid. The
form of arsenic released during pesticide manufacturing depends on the
stage in the manufacturing process at which it is released. Arsenic
trioxide is used as raw material for their manufacture. More complex
organic and inorganic compounds may be emitted at later stages in a
process, depending on the product being manufactured.
This study characterizes human population exposures to seven major
man-made arsenic emission sources. To assess the importance of the
contributions of these human activities, it is necessary to estimate
3

-------
the background concentrations for locations in which these activities
are not present. Atmospheric arsenic concentration data for 1974 for
267 locations representing a resident population of more than 58,000,000
people were evaluated." The annual average concentrations for all sites
ranged from below detection lUnit to 0.083 ~g/m3. The annual average
concentration for all locations was 0.003"~g/m3. The average concentra-
tion for eight locations near nonferrous smelters* was 0.030 ~g/m3, and
the average concentration for eight locations in remote rural areas was
0.0004 ~g/m3 assuming a concentration of zero for sample~ reported as
below detection limit. The lower detection limit for an individual
arsenic sample is 0.001 ~g/m3.
Estimating human population exposures from the seven selected emis-
sion sources has necessitated reliance on very limited data. Because of
the paucity of measured atmospheric arsenic data near emission sources,
it was necessary to approximate concentrations through the use of disper-
sion modeling. The dispersion modeling used was developed by EPA for
representative point source emissions; it is described in Appendix A.
All estimates given in this report are subject to considerable u~cer-
tainty in regard to:
(1)
(2)
( 3)
( 4)
( 5)
( 6)
The quantity of arsenic emissions
Ar~enic consumption levels
Source locations
Control technologies
Deterioration of installed control equipment over time
"The physical characteristics (e.g., stack height) of arsenic
emission sources
( 7)
Inherent inaccuracies in the modeling approach.
As a result, the overall accuracy of the exposure estUnates could not be
directly assessed. However, the dispersion modeling results have been
compared to ambient measured concentrations for nonferrous smelters and
cotton gins. Generally, the modeling results agree with the measured
concentrations within a factor of 3 to 4; however, variations for some
results are much larger.
Several alternative exposure estimates were made for each of the
seven arsenic sources evaluated. These alternatives are based on various
assumptions designed to illustrate the uncertainties involved in such
estUnates. Table 11-1 summarizes the results of this study for one se-
lected set of exposure estimates for each of the sources evaluated. The
exposure concentrations shown in the table are annual averages. Expo-
sures for selected times may be much higher or lower than the annual
averages. Population exposures for concentrations below 0.003 ~g/m3 are
*
Smelter locations were separated because of their obviously high concen-
trations.
4

-------
     Table 11-1  
   SUMMARY OF HUMAN POPULATION EXPOSURES TO ARSENIC 
   FOR SELECTED EMISSION SOURCES 
      Emission Sotirce 
 Average Annual       
 Concentrationa Copper Lead Zinc Cotton Pesticide Glass
 (II g/m3) Smeltersb Smeltersc Smeltersd Ginse Manufacturerf Manufacturerg
 3.0-5.9     5  
 1.0-2.9     100  
 . 0.60-0.99 24,420    200  
 0.30-0.59 19,380    700  
 0.10-0.29 137,450 800 9,000 2,000  1,440
 0.060-0.099 119,860 2,600 18,000 3,000  10,140
 0.030-0.059 286,560 5,100 101,000 5,900  75.180
 0.010-0.029 313,380 38,000 170,000 20,000 60 ~12 ,OliO
VI  
 0.005-0.009 8,810 46,000 110,000 56,000 800 363,870
 0.003-0.004 12,960 67,000 13,500 135,000 11,900 583,360
aAverage omnidirectional concentrations. With the exception of cotton gin exposures, 24-hr worst-case
exposures can be estimated by multiplying the annuai averages by 12.5. The 24-hr worst-case exposures
for cotton gins may be obtained by multiplying the concentrations by 81.

bBased on EPA's estimate of stack emissions. Assumes fugitive emissions are 5% of input for a11 but Tacoma.

cBased on an emission of 0.8 lb of arsenic per each ton of lead produced. Fugitive emissions are esti-
mated to be 101. of stack emissions.
d
Based on an emission of 1.3 lb of arsenic per ton of zinc produced by pyrometallurgical smelters and
no stack em~ssions at electrolytic smelters. Fugitive emissions assumed to be 101. of
stack emissions.
e
Annual average exposure, assuming that ginning exposures occur during 151. of the year and that there
are no exposures during the remainder of the year.

fAssumes that all large plant pesticide emissions are well controlled.

gAssumes that 15% of pressed and blown glass is manufactured with arsenic, that only certain manu-
facturers use arsenic in all of their pressed and blown glass production, and that the size distri-
bution of manufacturers who use arsenic is proportionate to the size output given in Table IX-I.
It is assumed that 90% of the manufacturers are well-controlled and that 10% are poorly controlled.

-------
not given because they are assumed to equal the average urban background
concentration. Population exposures are not given for concentrations
below 0,01 ~g/m3 for some copper smelter altern~tive estimates because
to do so would have required extrapolation of modeling results beyond 20
km from the source. At these greater distances, the accuracy of the
modeling results becomes increasingly uncer~ain.
Population exposures for coal-fired power plants are not shown in
Table 11-1 because realistic worst-case annual average arsenic exposures
for these plants have been estimated to be less than 0.003 ~g/m3 for the
largest poorly controlled power plant"s and less than 0.001 ~g/m3 for most
power plants.
In using the EPA estimate of current stack emissions for copper
smelters and assuming that fugitive emissions are 5% of input except for
Tacoma, it is estimated that 923,000 people are exposed to annual average
arsenic concentrations of 0.003 to 1 ~g/m3. Another estimate of copper
smelter exposures assumed an arsenic emission of 4.9 1b/ton of concentrate
with fugitive emissions as 10% of stack emissions. These assumptions
resulted in an estimated 950,000 people exposed to annual average concen-
trations of 0.01 to 1.0 ~g/m3. "
It is estimated that 160,000 people are exposed to annual average
concentrations of 0.003 to 0.3 ~g/m3 due to lead smelter emissions. This
estimate assumes an arsenic emission factor of 0.8 lb/ton of concentrate
and that fugitive emissions are 10% of stack emissions. An alternative
estimate assumed that 25% of the arsenic in the ore concentrate is emitted
to the atmosphere as stack emissions with 10% additional fugitive emis-
sions. This assumption resulted in an estimated 163,000 people exposed
to annual average concentrations of 0.003 to 0.1 ~g/m3.
It is estimated that 442,000 people are exposed to annual average
concentrations of 0.003 to 0.3 ~g/m3 from zinc smelter emissions. This
estimate assumes an arsenic emission factor of 1.3 1b/ton of concentrate
for pyrometallurgical zinc smelter stacks and no emissions from electro-
lytic smelter stacks. Fugitive emissions for all smelters are assumed
to be 10% of stack emissions.
Arsenic exposures from secondary smelters processing copper and zinc
scrap are estimated to be insignificant by comparison with nominal urban
background concentrations. Arsenic is added to lead used in batteries
and battery cables. This lead is extensively recovered in blast furnace
operations that can vaporize and release the arsenic content. The
secondary lead industry is highly fragmented, and production figures for
battery lead recovery by individual smelters are not available; however,
it is estimated that such smelters might produce annual average environ-
mental atmospheric concentrations on the order of 0.003 ~g to 0.020 ~g/m3.
Because most secondary smelters are located in fairly densely populated
areas, an average of several thousand people could be exposed to emissions
from each such smelter.
6

-------
It is estimated that approximately 13,000 people are exposed to
arsenic emissions from pestici'de manufacturing. The annual average ex-
posure concentrations range from 0.003 to 0.03 ~g/m3. These exposures
are judged to be relatively minor because it is assumed that large pesti-
cide manufacturers have already installed control equipment.
It is estimated that approximately 1.4 million people are exposed
to atmospheric arsenic concentrations of 0.003 to 29 ~g/m3 during the
cotton ginning session. Annual average population exposures, however,
are much smaller because the ginning season occurs for only a few weeks
a year. Based on annual average exposures, it is estimated that approxi-
mately 223,000 people are exposed to atmospheric arsenic concentrations
of 0.003 to 6.0 ~g/m3. These exposure estimates assume that all Texas
and Oklahoma machine-stripped cotton is desiccated with arsenic acid.
During some years, however, it is not necessary to desiccate all machine-
stripped cotton. For those years, the estimated number of exposed people
should be proportionately reduced.
Because of uncertainties in determining which glass producers use
arsenic and in estimating the amount of arsenic used, attempts to estimate
the people exposed to arsenic from glass manufacturing are marked by
considerable uncertainty. This uncertainty is illustrated by the several
alternative exposure estimates based on different assumptions. However,

all alternative estimates indicated that 15 thousand people or more
are exposed to annual average arsenic concentrations in excess of
0.003 ~g/m3. The exposure estimate given in Table 11-1 shows approxi-
mately 1.2 million people exposed to arsenic concentrations of 0.003 to
.30 ~g/m3.
As indicated in Table II-I, the exposures estimated in this report
are subject to considerable uncertainty; they thus require further moni-
toring and sampling data and evaluation for a more complete assessment.
Despite the insufficiency of data, however,_the population exposure from
these sources can be substantially greater than nominal background con-
centrations. Potential health effects from the estimated exposures
will be addressed in another report being prepared by the EPA Cancer
Assessment Group.
The primary purpose of this report was to estimate human inhalation
exposures to atmospheric arsenic emissions. However, atmospheric emis-
sions of arsenic are eventually deposited on the earth or in water.
Deposits on the earth may enter waters through runoff. Plants may absorb
this arsenic and they may in turn be consumed by humans or by animals that
are eaten by humans. Humans may drink the arsenic-contaminated waters or
consume aquatic organisms with an increased arsenic content caused by
residing in such waters. Arsenic bioaccumulates in various organisms
found in the environment, but the compound arsenic is "ot bion~gnified in
the food chain. Arsenic is ubiquitous in nature, and small amounts of
it are found in most food and water supplies. Elevated levels of arsenic
that appear to be unrelated to atmospheric emissions ar~ found in seafood
and local drinking waters. The effects of atmospheric fallout of arsenic
7

-------
onto land from the sources evaluated in this report must be considered
minor, in light of the fact that large quantities of arsenic are released
on croplands as pesticides and are aissipated as wastes onto land or into
water. Moreover, contamination of earth and water from the atmospheric
sources evaluated in this report must accumulate over a relatively long
period of time to reach significant levels. The most likely significant
route of human exposure to arsenic from these secondary sources resulting
from atmospheric arsenic emissions, is eating vegetables raised in
arsenic-contaminated soils.
8

-------
III
SOURCES AND BEHAVIOR OF ARSENIC IN THE ENVIRONMENT
Arsenic occurs in the atmosphere both as a result of man's activities
and from natural sources. Man-made sources include nonferrous smelters,
pesticide manufacture and use, the combustion of fossil fuels (prtmari1y
coal), glass manufacturing, cotton ginning, and th~ lead alloy industry.
Tables 111-1 and 111-2 provide rough estimates of all environmental
contribution from these and other sources. The estimates given on these
two tables were made by different authors and differ considerably for
some sources. Natural sources of arsenic include volcanic action, hot
springs, decay of plant matter and other microbial actions,and the
weathering of minerals within the soils.
Arsenic is mobile in the environment. Microbes in the soils, waters,
and sediments methylate and reduce arsenic to arsine, which is volatile
and can enter the atmosphere. The arsines may then be oxidized to less
toxic products, that can return to the earth and'may be recycled. Most
foods and beverages contain small arsenic concentrations. Seafoods,
particularly marine invertebrates, usually contain more arsenic than
other foods. Although arsenic is bioaccumulated by various organisms,
especially in marine environments, there appears to be no significant
biomagnification of this element within food chains (Union Carbide, 1976).
The geochemical cycle' for arsenic is presented in Figure 111-1. As
arsenic cycles through the environment, it is converted ,from one oxidation
state to another. Because of multiple oxidation states and the tendency
to form soluble complexes, arsenic geochemistry is intricate and not well
characterized (Boyle and Jonasson, 1973). ' '
The biological cycle for arsenic (Figure 111-2) is understood more
completely than the geochemical cycle. In sediments, arsenate is micro-
bially reduced to arsenite, which is then methylated to methylarsenic
acid. Methylarsenic acid is reduced and methylated to form dtmethy1arsinic
acid. This 'acid is reduced to dimethylarsine, which is methylated to
yield trimethylarsine. Arsines, which are volatile and very toxic, leave
the sediments to pass through the water. When the arsines enter the at-
mosphere, they are readily oxidized to less toxic products such as
cacodylic acid (CA) (McBride and Wolfe, 1971; Wood, 1974). If the CA
returns -to water (by way of d~ fallout or precipitation), it can be re-
duced by microbes, repeating the latter portion of the cycle, or it can
return to sediment either by precipitation. with or by adsorption on
hydrous oxides (Ferguson and Gavis, 1972). If the CA returns to land,
soil organi~s can alter it to arsine and recycle it to the atmosphere,
or they can convert it to C02 and As043- The As043- is then bound to
hydrous oxides in the soil. Thus, the environmental fate of arsenic
appears to be metabolization to iporganic arsenate which is bound in in-
soluble compounds in the soil (Woolson and Kearney, 1973).
9

-------
Table IlI-l
ARSENIC POLLUTANT SOURCES
Source
Mining

Phosphate rock

Primary copper
Roasting
Reverberatory furnaces
Converters
Material handling

Primary zinc
Roasting

Primary lead
Sintering
Blast furnace
Reverberatory furnace

Gray iron foundry
Cotton ginning and
Nonferrous alloys
Phosphoric acid
Glass manufacture
burning
Wood preservatives
Miscellaneous arsenic chemicals
Arsen~c pesticide production

Pesticide, herbicide, fungicide use

Power plant boilers
Pulverized coal
Stoker coal
Cyclone coal

Industrial boilers
Pulverized coal
Stoker coal
Cyclone coal
All oil

Residential/commercial coal
Incineration
Total
"nnual
Environmental
Release
( ton)

2

Negligible
900
400
1150
250
1390
285
80
11
97
,345

Negligible
Negligible
638
Negligible
3
196

2925
429
49
15
19
67
9
Negligible
6
---1
9268
*
Sum exceeds loot because of rounding error.
Source:
Union Carbide (1977)
10
Percent of
Total
Release
0.03
Negligible
9.71
4.32
12.41
2.70
15.00
3.08
0.87
0.12

1.05

3.73

Negligible
Negligible
6.89
Negligible
0.04
2.12

31.56
4.63
0.53
0.17
0.21
0.73
0.10
Negligible
0.07
0.03
*
100.00

-------
Table III-2
ENVIRONMENTAL DISSIPATIONS OF ARSENIC (1974)
(Metric Tons per Year)
Airborne
Emissions
Primary zinc
Primary lead
190
240

4,800 *
( 200)

50
130
Primary copper
Other nonferrous metals
Lead alloys
Phosphates
Water and wastewater treatment
1
Boric and boric acid
Manganese ore
Iron and steel
10

32

650 *
( 170-340)

108

130

2,300
Coal
Petroleum
Pesticide production
Pesticide use
Feed additive production
Feed additive use
.2
Glass manufacturers
210
Waterborne
.Effluents
0.4
32 *
(1,550)
110
3.9
* .
These estimates are from Holt and Moberly (1976).

t'rhe American .Iron and Steel Institute (1978) has stated
estimate is too high by several orders of magnitude.
Source:
OTS (L976)
11
Dissipated
to Land
120

800

21,800
(7,460) *

50
300
415

67
1,080

37,350t

1,800
(3,000)*
10,490
407
that this

-------
INHALATION OF OUST
r -AND GASEOUS FORMS
I OF ARSENIC

I
I
I
BIOSPHERE
DEGRADATION
PLANTS = ANIMALS
DEGRADATION
AND
SOLU nON
VAPORI Z A TlON
PRECIPITATION
PRECIPITATION
WATER ~ SEDIMENTS
SOLUTION
PEOOSPHERE
SOILS
GLACIAL MATERIALS
ATMOSPHERE
HYDROSPHERE
....
N
I
I
I
L --- ~~ - -
CHEMICAL PRECIPITATION
AND SEDIMENTATION
OF SOLIDS
SOLUTION AND
MECHANICAL
WEATHERING
CHEMICAL
PRECIPITATION
LITHOSPHERE
ROCKS
ARSENIC - BEARING
DEPOSITS
SOLUTION AND
MECHANICAL
WEATHERING
PRECIPITATION AND
CONSOLIDATION OF SOLIDS
Source: Union C.tbitle 119171
FIGURE 111-1. THE GENERALIZED GEOCHEMICAL CYCLE FOR ARSENIC

-------
Air
CHJ
I
HO- A5+ - CH,
II
o
(OJ]
Water
.-
IN
CHI
I
CHI - As'- - CHI +
T,imeth~IAlsin.
CH,
I
H - As'--CH,
o.melhwl~rsin.
OH CHI' CH,
I I ~ I
HO- Ass+ -OH~ A5'+ -OH/HO- As' -OH/HO- AS+-CH.
II ~cteria II BJCteri~ II ~ct.,i. II
o 0 0 0
Arsenate
Arsenite
Meth,larsenic
. Itld
Oimelhyl.rsinic
~cid
Sediment
Source:
Union Ca.lJide 119711
fiGURE 111.2. THE PROPOSED BIOLOGICAL CYCLE FOR ARSENIC
'.

-------
Union Carbide (1976) could not locate residence times for arsenic
in the atmosphere in the literature. It is hypothesized that when vola-
tile arsenicals are released to the atmosphere they are oxidized in the
presence of light and return to earth by way of dryfall or precipitation
(Sullivan, 1969; Woolson and Kearney, 1973).
14

-------
IV
BACKGROUND CONCENTRATIONS OF ARSENIC IN THE ATMOSPHERE
Natural and human activities produce a persistent low-level concen-
tration of arsenic in the atmosphere. However, elevated concentrations
are found near certain human activities. To assess the importance of
the' contributions of some of these activities, it is necessary to esti-
mate the background concentrations for locations in which these activi-
ties are not present.
The National Air Sampling Network (NASN) routinely monitors sus-
pended particulate concentration levels in urban and nonurbanareas,
generally reporting them as quarterly composites for stations in the
network. The composite, which pools all samples collected during the
quarter, assists in generating sufficient material for laboratory
analysis..
The arsenic samples are analyzed by flame less atomic absorption
which has a lower detection limit of 0..001 IJ.g/m3 and a mean precision,
based on duplicates from the same filter, of i37% (Shearer, 1975).
The quarterly and annual average arsenic concentration data for
1974 for individual NASN locations are given in Table IV-I. For some
locations, quarterly.data were available from two sources. In the
generation of Table IV-l these two observations were averaged. When
1974 quarterly data were not available for a location, the most recent
available quarterly data (from 1973; 1972, or 1971) were used.
Eight of the NASN sites are located.in areas that have nonferrous
smelters \.Jithin approximately 50 miles. Another eight of the sites are
located in remote national parks or forests.
Table IV-2 lists the 15 NASN locations having average annual arsenic
concentrations in excess of 0.010 IJ.g/m3. Of these 15 locations, 4 are
situated within approximately 10 miles of nonferrous smelters and an
additional 4 are situated within 50 miles of nonferrous smelters.
Average annual arsenic concentrations for locations ranged from 0
to 0.083 IJ.g/m3. The average annual concentration for all NASN locations
was 0.003 IJ.g/m3. The average annual concentration for the eight smelter
sites was 0.030 IJ.g/m3, and the average annual concentration for the eight
remote areas was 0.0004 IJ.g/m3. Hence, concentrations increase by an
order of magnitude for urban over rural and another order of magnitude
for smelter cities over urban.
The 1970 populations were obtained for all cities and towns listed
in Table IV-I. These populations indicate that the arsenic concentrations
15

-------
 Table IV-1      
1974 ATMOSPHERIC ARSENIC CONCENTRATIONS    
FOR U.S. LOCATIONS (~g/m3)     
 1st  2nd  3rd  4th  Yr1y.
Loca tion Qtr. Qtr. Qtr. Qtr. Av'1..
ALABAMA         
Birmingham 0.022  0.015  0.000  0.000  0.009
Gads en  0.020  0.007  0.000  0.000  0.007
  *      * 
Huntsville 0.000  0.005  0.000  0.000  0.001
  *   0.007   * 
Mobile 0.000  0.017   0.000  0.006
Montgomery 0.011  0.000  0.000  0.000  0.003
ALASKA         
    *    * 
Anchorage 0.000  0.000  0.000  0.000  0.000
  *  *    * 
Fairbanks 0.000  0.000  0.000  0.000  0.000
ARIZONA         
  *      * 
Apache County 0.000      0.000  0.000
Coconino County 0.032t O.ooot O.ooot 0.006t 0.010
        * 
Douglas 0.011  0.014  0.000  0.044  0.017
Grand Canyon Nat. Park 0.000  0.000  0.000  0.000  0.000
      *   
Maricopa County 0.000  0.000  0.000    0.000
Phoenix' 0.013  0.000  0.000  0.009  0.006
 O.ooot O.ooot  ...  ... 
Superior 0.000' 0.000'. 0.000
Tucson 0.000  0.000  0.000  0.000  0.000
Note: Readings below the lower detection limit of 0.001 are reported as 0.000.

*
1973 data.

t 1972 data.
**
1971 data.
- - No data available.
16

-------
  Table IV-I (Cor.tinued)    
  1st 2nd 3rd 4th Yrly.
Location Qtr. Qtr. Qtr. Qtr. Avg.
ARKANSAS        
Little Rock 0.008 0.006 0.007  0.000 0.005
Montgomery County 0.000 - 0.000 0.000  0.000 0.000
    * O.ooot  
Texarkana    0.000 0.000 0.000
West Memphis 0.016  0.012 0.000  0.000 0.007
CALIFORNIA        
Anaheim  0.000 0.000 0.000  0.000 0.000
Berkeley  0.000  0.000 0.000  0.000 0.000
Burbank  0.000  0.000 0.000  0.000 0.000
Fresno  0.000  0.000 0.000  0.000 0.000
      * * 
Glendale  0.000  0.000 0.000  0.000 0.000
Long Beach  0.014  0.000 0.000  0.000 0.004
Los Angeles  0.000  0.000 0.000  0.000 0.000
Oakland  0.000  0.000 0.000  0.000 0.000
Ontario  0.000  0.011  t O.ooot 0.003
  . 0.000 
Pasadena  0.000  0.000 0.000  0.000 0.000
Riverside  O.OOOT O.ooot O.ooot  0.000
Sacramento  0.000  0.000 0.000  0.000 0.000
San Bernardino 0.000  0.000 0.006  0.000 0.002
San Diego  0.000  0.000 0.000  0.000 0.000
San Francisco 0.000  0.000 0.000  0.000 0.000
San Jose  0.000  0.000 0.000  0.000 0.000
Santa Ana  0.000  0.000 0.000  0.000 0.000
Torrance  0.000  0.000 0.000  0.000 0.00J
COLORADO        
   * O.ooot O.ooot  
Denver  0.008   0.003
   * * * * 
l1esa Verde Hat. Park 0.000  0.000 0.000  0.000 0.000
17

-------
 Table IV-1 (Continued)     
 1st  2nd  3rd 4th  Yr1y.
Location Qtr. Qtr. Qtr. Qtr. Avg.
CONNECTICUT         
        * 
Bridgeport 0.000  0.000  0.000 0.000  0.000
Hartford 0.000  0.000  0.000  0.000  0.000
New Haven 0.000  0.000  0.000  0.000  0.000
Waterbury 0.000  0.000  0.000  0.000  0.000
DELAWARE         
 0 .000+ 0.000+  *   
Kent County 0.000    0.000
    *    * 
Newark 0.000  0.000  0.000  0.000  0.000
Wilmington 0.000  0.000+ 0.000+ 0.000  0.000
DISTRICT OF COLUMBIA         
Washington 0.000  0.000+ 0.000+   0.000
FLORIDA         
     0.000+  * 
Hardee County 0.000  0.000  0.000  0.000
  *  *     
Jacksonville 0.000  0.000  0.000  0.000  0.000
Miami 0.000  0.000  0.000  0.000  0.000
  * 0.000+     
St. Petersburg 0.000  0.000  0.000  0.000
  *       
Tampa 0.000  0.000  0.000  0.000  0.000
GEORGIA         
Atlanta 0.007  0.000  0.000  0.000  0.002
  *       
Columbus 0.000  0.000  0.000  0.000  0.000
  *  *  * 0.000+ 
Savannah 0.000  0.000  0.000  0.000
HAWAII         
Ha1awa Heights 0'. 000+ 0.000+ 0.000+ 0.000+ 0.000
      ...   
Hawaii County 0.000  0.000  0.000' 0.000  0.000
Volcanoes Nat. Park   0.000  0.000  0.000+ 0.000
Honolulu 0.000  0.000  0.000  0.000  0.000
18

-------
  Table IV-l (Continued)    
  1st 2nd  3rd 4th Yrly.
Location Qtr. Qtr. Qtr. Qtr. Avg.
IDAHO         
       *  
Boise City 0.005  0.000  0.000  0.000. 0 . 001
       * * 
Butte County 0.006  0.000  0.000  0.000 0.002
ILLINOIS         
     t  *  
Chicago  0.000  .0.000  0.000  0.000 0.000
   *  t   O.ooot 
East St. Louis 0.000  .0.000  0.009  0.002
   *    *  
Effingham  0.000    0.000   0.000
   *  *  * O.ooot 
Joliet  0.000  0.000  0.000  0.000
~oline  0.000  0.000  0.000  0.000 0.000
     * O.ooot  
North Chicago 0.000  0.000  0.000 0.000
..  *  *  * O.ooot ..
Peoria  0.000  0.000  0.000  0.000
  O.ooot  *  *  
Rockford  0.000  0.000  0.000 0.000
Rock Island   0.000  0.000  0.000 0.000
Springfield 0.000  0.000  0.000  0.000 0.000
INDIANA.         
~st Chicago 0.000  0.000  0.000  0.000 0.000
   *      
Evansville 0.000  0.000  0.000  0.000 0.000
   *  * * O.ooot 
Fort Wayne 0.000  0.000  0.000  0.000
Gary  0.000  0.000  0.000  0.000 0.000
Hammond  0.000  0.000  .0.000  0.000 0.000
Indianapolis O.ooot 0.000  0.000  O.ooot 0.000
   *     O.ooot 
Monroe County 0.000  0.000  0.000  0.000
Muncie  0.010       
   *   * O.ooot 
New Albany 0.000  0.000  0.000  0..000
   ...   O.ooot O.ooot 
Parke County 0.000' 0.000  - 0.000
South Bend 0.000  0.000  0.000  0.000 0.000
   *      
Terre Haute 0.000  0.000  0.000  0.000 0.000
19

-------
 Tab le IV-l (Continued)    
 1st  2nd  3rd 4th Yrly.
Location Qtr. Qtr. Qtr. Qtr. AVR.
IOWA        
  *  * *   
Cedar Rapids 0.000  0.000  0.000 0.000 .0.000
     .  
Davenport 0.005  0.005  0.012 0.000 0.006
Des Moines 0.000  0.000  0.000 0.000 0.000
Dubuque 0.000  0.000  0.000 0.000 0.000
Waterloo 0.005  0.000  . 0.000 0.000 0.001
KANSAS        
Kansas City 0.000  0.000  0.000 0.000 0 . 000
Topeka 0.004  0.000  0.000 0.000 0.001
Wichita 0.000  0.000  0.000 0.000 0.000
KENTUCKY        
Ashland 0.014  0.000  0.000 0.000 0.004
       * 
Bowling Green 0.000  0.000  0.000 0.000 0.000
Covington 0.000  0.000  0.000 0.000  0.000
  *  * O.ooot   
Lexington 0.000  0.000  0.000  0.000
Louisville O.ooot 0.000  0.000 0.000  0.000
LOUISIANA        
Ba ton Rouge 0.000  0.000  0.000 0.000  0.000
Iberville Parish 0.000  0.000  0.000 0.000  0.000
     *   
New Orleans 0.000  0.000  0.000 0.000  0.000
Shreveport 0.000  0.000  0.000 0.000  0.000
MAINE        
 0.000     0.000 * 
Acadia Nat. Park  0.000  0.000  0.000
Portland 0.000  0.000  0.000   0.000
MARYLAND        
Baltimore 0.025  0.000  0.005 0.000  0.008
   O.ooot   * 
Calvert County   0.000 0.000  0.000
20

-------
   Table IV-l (Continued)       
    1st  2nd   3rd  4th  Yrly.
 Location  Qtr. Qtr.  Qtr. Qtr. Avg.
MASSACHUSETTS            
    0.000+       *  
Boston   0.000   0.000  0.000   0.000
            *  
Cambridge    0.000   0.000  0.000   0.000
            *  
Fall River    0.000   0.000  0.000   0.000
            *  
New Bedford    0.000   0.000  0.000   0.000
     *      0.000+  
Springfield  0.000  0.000   0.000   ' 0.000
     *  ,   *  *  
Worcester ' 0.000  0.000 ' , 0.000  0.000   0.000
MICHIGAN             
Dearborn  0.000  0.000   0.000  0.000   0.000
Detroit        0.000  0.000  '" O. 000
     *  *   *    
Flint  0.006  0.000   0.000  0.000   0.002
     * O~OOO * '0;000 * 0.000+  
Grand Rapids  0.000     0 :000 '
Lansing   0.000  0.000   0.000  0.000+ " 0.000
            "
Saginaw   0.000  0.000   0.000  0.000   ,9.,000-
     *      0.000+ "
Trenton   0.000  0.000   0.000  " O. 000 '
MINNESOTA            "
Duluth  -0.000  0.000   0.000  0.000   -0.000 .
     * 0.000+    0.000+  
Minneapolis  0.006   0.000   ,,-9.002
Moorhead  0.000  0.000   0.000  0.000   0.000
St. Paul'  0.000  0.000   0.000  0.000   0.000
MISSISSIPPI            
Jackson  .0.000  0.000  ' .0.000  0.000   '0.000-
Jackson County  0.000  0.000   0.000  0.000   0.000
MISSOURI             
   -        *  
Kansas City 0.005  0.000   0.000  0.000   0.001
St. Louis  0.009  0.007   0.000  0.000   0.004
      0.000   *    0.000
Shannon County  0.000    0.000  0.000  
21

-------
   Table IV-I (Continued)     
   1st  2nd  3rd  4th  Yr1y.
Location Qtr. Qtr. Qtr. Qtr.  Avg.
MONTANA           
    *  *  *  * 
Glacier Nat. Park 0.000  0.000  0.000  0.000  0.000
   O.ooot 0.014t  **  ** 
Helena   0.103  0.091 0.052
NEBRASKA           
Lincoln  0.000  0.000  0.000  0.000  0.000
0mah8   0.000  0.000  0.008  0.000  0.002
        *  * 
Thomas County 0.000  0.000  0.000  0.000  0.000
NEVADA           
Las Vegas  O.ooot 0.000    0.000  0.000
Reno   0.000  0.000  0.000  0.000  0.000
White Pine County 0.000  0.000  0.000  0.000  0.000
NEW HAMPSHIRE         
    *  *    * 
Belknap County 0.000  0.000    0.000  0.000
Boscawen      O.ooot   
          * 
Concord   0.000  0.000  0.000  0.000  0.000
Coos County 0.000  0.000  0.000  O.ooot 0.000
NEW JERSEY          
Bayonne     0.010  0.000  0.000  0.003
        *  * 
Camden   0.000  0.005  0.000  0.013  0.005
Elizabeth' 0.016  0.013  0.000  0.000  0.007
G1assboro  0.000  0.000  0.000  0.000  0.000
   o .000 t 0.008   *   
Jersey City  0.000  0.000  0.002
Newark   0 .010  0.000  0.000  0.1>00  0.003
     0.007   *   
Paterson  0.000   0.000  0.000  0.002
Perth Amboy 0.019t 0.010  0.007  0.022t 0.015
Trenton  0.007  0.000  0.000  0.000  0.002
22

-------
Table IV-I (Continued)    
   1st  2nd 3rd  4th Yrly.
Location Qtr.  Qtr. Qtr.  Qtr. Avg.
NEW MEXICO         
Albuquerque  0.000 0.000 0.000 0.000 0.000
Arriba County  O.ooot O.ooot O.ooot O.ooot 0.000
NEW YORK         
Albany   o.ooa  0.000 0.000 0.000 0.002
Buffalo   0.007  0.000 0.000  0.000 0.002
Jefferson County 0.000  0.000 0.000  0.000
New York City  0.000  0.008 0.000  0.000 0.002
Niagara Falls  0.009  0.000 0.000  0.000 0..002
Rochester   0.015  0.000 0.000  0.000 0.004
Syracuse   0.000  0.000 0.000  0.000 0.000
Troy   O.ooot O.ooot O.ooot O.ooot O~OOO
Utica     . 0.000   0.000 0.000
Yonkers   O.ooot O.ooot 0.000  0.000 0.000
NORTH CAROLINA        
Cape Hatteras Nat. Park O.ooot O.ooot    0.000 .
Charlotte   0.000  0.000 0.000  0.000 0.000
Durham   0.000  0.005. ' 0.000  0.000 ~). 001
Greensboro   0.000  . 0.000 0.000  0.000 0.000
Winston-Salem  0.007  0.007' . 0.000  .'0. OlJ .0.007
OHIO         
    *     
Akron   0.019  0.007 0.008   0.011
    *     
Canton   0.008  o.ooa 0.006  0.006 0.007
Cincinnati   O.ooot 0.006 O.ooot 0.000 0.002
Cleveland   0.005t 0.006t   o.ooat 0.006
   O.ooot O.ooot  * O.ooot 
Columbus   0.000  0.000
     *  *  
Dayton   0.011' 0.000 0.013  0.006 0.008
   *   * O.ooot 
Ironton   0.013  0.000 0.000  0.003
   *  * O.ooot 
Portsmouth   0.000  0.000 0.000  0.000
23

-------
 Table IV-I (Continued)     
 1st  2nd  3rd  4th  Yrly.
Location Qtr.  Qtr.  Qtr.  Qtr.  Avg.
OHIO (Continued)         
  *  *   O.ooot 
Steubenville 0.021  0.000  0.020  0.010
Toledo 0.014+ 0.000  0.000  0.000+ 0.004
      *   
Youngstown 0.015  0.009  0.006  0.019  0.012
OKLAHOMA         
       0'.000 * 
Cherokee County 0.000  0.000  0.000   0.000
Oklahoma City 0.012  0.000  0.006  0.000  0.005
    *     
Tulsa 0.000  0.000  0.000  0.000  0.000
OREGON         
    *  * 0.000+ 
Curry County 0.000  0.000  0.000  0.000
Portland      * 0.000+ 
0.000  0.000  0.000  0.000
PENNSYLVANIA         
Allentown 0.009  0.000  0.000  0.000  0.002
  *       
Altoona 0.012  0.000  0.015  0.023  0.013
        * 
Bethlehem 0.005  0.000  0.006  0.000  0.003
Cambria. County 0.005*       
Clarion County 0.000  0.000  . 0.000  0.000+ 0.000
Clear field County O.ooot 0.000+ 0.000+   0.000
  *       
Erie 0.000  0.007  0.006  0.009  0.006
  *   0.000+  * 
Harrisburg 0.000  0.000  0.000  0.000
Haz1eton 0.015  0.000  0.000  0.000  0.004
Indiana County 0.000  O.ooot O.ooot 0.000  0.000
Lancaster City 0.000  0.000  0.000  O.ooot 0.000
Philadelphia 0.017  0.042  0.000  0.031  0.023
Pittsburgh 0.017  0.009  0.000  0.000  0.007
  *  *  * 0.000+ 
Reading 0.000  0.000  0.000  0.000
24

-------
  Table IV-l (Continued)     
  1st  2nd  3rd  4th  Yr1y.
Location  Qtr.  Qtr.  Qtr.  Qtr.  Avg.
PENNSYLVANIA (Continued)         
Scranton  0.016  0.000  0.000  0.015  0.008
Warminster    0.000  0.000    0.000
     *  *  * 
West Chester    0.000  0.000  0.000  0.000
Wilkes-Barre  0.014  0.000  0.000  0.000  0.004
   *  ..  *   
York  0.000  0.000  0.000  0.000  0.000
PUERTO RICO          
         * 0.003
Bayamon  0.000  0.010  0.000  0.000 
  O.ooot      * 
Catano  0.000  0.014  0.000  0.004
       *  * 
Guayanil1a  0.000  0.000  0.000  0.000  0.000
Guayani11a County    0.000      
       *   
Ponce  0.000  0.000  O.qoo  0.000  0.000
       *   
San Juan  0.010  0.000  0.000  0.000  0.003
       *   
San Juan County      0.000    
RHODE ISLAND          
   *  *    * 
East Providence  0.000  0.000  0.000  0.000  0.000
     *     
,Providence  0.000  0.000  0.000  0.000  0.000
  O.ooot O.ooot  *  * 
Washington County  0.000  0.000  0.000
SOUTH CAROLINA          
   *       
Columbia  0.000  0.000  0.000  0.000  0.000
Greenvi11e  0.000  0.000  0.000  0.000  0.000
        * 
Richland County  0.000  0.000  0.000  0.000  0.000
SOUTH DAKOTA          
Black Hills Nat. Forest 0.000  0.000  0.000  0.000  . 0.000
.Sioux Falls  0.000  0.000  0.000    0.000
25

-------
 Table IV-l  (Continued)     
 1st  2nd  3rd  4th  Yrly.
Location Qtr.  Qtr.  Qtr.  Qtr.  Av'1..
TENNESSEE         
  *  *   O.ooot 
Chattanooga 0.000  0.000  0.000  0.000
Knoxville O.ooot 0.000  0.012  0.000  0.003
Memphis 0.000  0.000  0.000  0.000  0.000
        * 
Nashville 0.014  0.000  0.000  0.000  0.004
TEXAS         
Amarillo   0.000      
Austin 0.000  0.000      0.000
Beaumont   0.000      
Corpus Christi   0.000  0.000    0.000
      *   
Dallas 0.010  0.006  0.000  0.000  0.004
El Paso 0.067  0.070  0.005  0.132  0.069
Fort Worth 0.000  0.000  0.000  0.000  0.000
Houston 0.000  0.000  0.000  0.000  0.000
Lubbock   0.000      
Matagorda County 0.000  0.000  0.000  0.000  0.000
    *  *   
Pasadena 0.000  0.000  0.000  0.000  0.000
San Antonio 0.000  0.000  0.000  0.000  0.000
  *  *  *   
Tom Green County 0.000  0.000  0.000  0.000  0.000
Wichita Falls 0.000    0.000    0.000
UTAH         
  *  *  *  * 
Ogden 0.000  0.015  0.021  0.014  0.013
Salt Lake City 0.031 t O.OlOt 0.034t 0.035t 0.028
VERMONT         
Burlington 0.000  0.000  0.000  0.000  0.000
      *   
Orange County 0.000  0.000  0.000  0.000  0.000
26

-------
  Table IV-l (Continued)     
  1st  2nd  3rd  4th  Yrly.
Location Qtr.  Qtr.  Qtr.  Qtr.  Avg.
VIRGINIA          
      0.000+ -... 
Danvil1e  0.018  .0. 000  0.000' 0.005
   *  *  *  * 
Fairfax County 0.000  0.000  0.000  0..000  0.000
Hampton  0.000  0.000  0.000  0.000  0.000
   *  *     
Lynchburg 0.012  0.000  0.000  0.000  0.003
New Kent County       0.000+ 
   *    *  * 
Newport News 0.000  0.000  0.000  0.000  0.000
Norfolk  0.000  0.000  0.000  0.000  0.000
   *    *   
Portsmouth 0.000  0.000  0.000  0.000  0.000
Richmond  0.016  0.011  0.000  0.000  0.007
         * 
Roanoke  0.000  0.000  0.000  0.000  O!OOO
Shenandoah Nat. Park 0.000+ 0.000  0.000  0.000+ 0.000
     *  *   
Wythe County 0.000  0.000  0.000  0.000  0.000
WASHINGTON          
     *  *  * 
King County 0.026  0.024  0.042  0.000  0.031
       *  * 
Seattle  0.029  0.040  0.060  0.025  0.039
       *  * 
Spokane  0.000  0.000  0.000  0.000  0.000
     *  *  * 
Tacoma  0.042  0.099  0.190  0.000  0.083
WEST VIRGINIA         
Charles tOrt  0.058  0.054  0.075  0.037  0.056
       *  * 
Huntington 0.000    0.006  0.Q09  0.005
       *   
South Charleston 0.044  0.020  0.021  0.009  0.024
WISCONSIN          
   *  ...  * 0.000+ 
Door County 0.000  0.000' 0.000  0.000
     *  * 0.000+ 
Eau Claire 0.000  0.000  0.000  0.000
   *  *  * 0.000+ 
Kenosha  0.000  0.000  0.000  0.000
     *  *   
Langlade County   0.000  0.000    0.000
Madison  0.000  0.000  0.000  0.000  0.000
27

-------
 Table IV-l (Concluded)     
 1st  2nd  3rd  4th  Yrly.
Location Qtr.  Qtr.  Qtr.  Qtr.  Av~.
WISCONSIN (Continued)         
  * O.ooot     
Milwaukee 0.000  0.000  0.000  0.000
Racine 0.011  0.000  0.000  0.000  0.003
 O.ooot   O.ooot  * 
Superior 0.000  0.000  0.000
WYOMING         
  *  *  *  * 
Casper 0.000  0.000  0.000  0.000  0.000
  * O.ooot O.ooot O.ooot 
Cheyenne 0.000  0.000
  *  *  *  * 
Grand Teton Nat. Park 0.000  0.000  0.000  0.000  0.000
  *  *  *  * 
Yellowstone Nat. Park 0.000  0.011  0.000  0.000  0.003
28

-------
Table IV-2
LOCATIONS HAVING ANNUAL AVERAGE ATMOSPHERIC
ARSENIC CONCENTRATIONS IN EXCESS OF 0.01 ~g/m3
Location
Average Annual
Concent~ation (~~/m3)
Akron, OH
0.011
Youngston, OH
0.012
Ogden, UT
*
0.013
A1toona, PA
0.013
Perth Amboy, NJ
0.015
Douglas, AZ
4-
0.017'
Philadelphia, PA
0.023
South Charleston, WV
0.024
Salt Lake City, UT
*
0.028
King County, WA
*
0.031
Seattle, WA
*
0.039
Helena, :-rr
4-
0.052'
Charleston, \N
. 0.056
El Paso, TX
4-
0.069'
Tacoma, WA
4-
0.083'
'*
Nonferrous smelter within 50 miles.
-'-
~onferrous smelter within 10 miles.
29

-------
listed iri Table IV-I represent an exposed population of more than
58,000,000. The statistical distribution of people to exposures i~
given in Figure IV-I. Also shown in Figure IV-I is the statistical dis-
tribu~ion of people to exposures excluding.the eight nonferrous smelter
locations.
Exposures weighted by city population give an average population
exposure of 0.004 ~g/m3. The population weighted average exposure for
the eight nonferrous smelter locations is 0.030 ~g/m3 (the same as the
unweighted average).
30

-------
.10
.09

.08

.07
.06
.05
.04
M 
E 
- 
en 
:t 
I 
Z 
0 
i= 
c( 
a: 
I- 
Z 
w 
(J 
Z 
0 
(J 
~ 
z 
w .01
II)
a:: .009
c(
(J .008
a:: 
w .007
:I:
~ 
II) 
0 .006
~
I- 
c( .005
 .004
 .003
.03
.02
.002
;001
40
BASED ON NASN DATA
REPRESENTING A POPULATION
OF 58 MilliON PEOPLE
INCLUDING 8
NONFERROUS
SMELTER
LOCA TIONS
""
'EXCLUDING 8
NONFERROUS
SMELTER
LOCATIONS
50 60 70 80 90 95 98 99.8 99.9 99.99
PERCENT OF POPULATION EXPOSED TO INDICATED CONCENTRATION OR LESS
FIGURE IV-'. ATMOSPHERIC ARSENIC CONCENTRATIONS FOR
EXPOSED URBAN POPULATIONS
31

-------
v
ARSENIC EXPOSURES FROM COAL-FIRED POWER PLANTS
A.
General
Arsenic is one of' a number of volatile trace elements contained in
coal. ,The average'~rsenic-content of U.S. coals h~s been reported to be
5.44 ppm, with average estimates of 10 ppm for eastern, coal, 5 ppm for
coal from interior states, and 1 ppm for coal from western states (Davis,
1971). Ruch et al. (1974) found an average arsenic concentration of'14
ppm for 101 coal samples taken mostly from Illinois. Magee et a1. (1973)
reported arsenic concentrations of 3-59 ppm for Appalachian coal, 9-45
ppm for interior .coal, and 73 ppm for dne sample taken from the Four
Corners area of Utah, Arizona, New Mexico, and Colorado. Mage~ et ale
did not report'average conc~ntrations. The variation in concentrations
as reported by different authors is probably indicative of the variation
among coal samples and test procedures.
The somewhat higher arsenic content of ~astern coals (a~ reported
by Davis, 1971) is partially offset by their higher heating content for
the generation of an equivalent unit of power. Appalachian coals give
about 12,500 Btu/lb, interior coals about 11',000 Btu/lb, and western
coals about 9,500 Btu/lb (Lee et al.~ 1977).
T
-------
Table V-l
COAL USE BY SOURCE (1974)
Source
Electric power generation
Coke plants
Other manufacturing and mining
Retail ~ealers
% Use
71

17
11
1 '
Millions of Metric
Tons Der Yea r

347.9

83.3

53.9

4.9
.Total
100
490.0
Source:
Bureau of Hines (1975).
Table V-2
ARSENIC RELEASED FROM COAL CONSUHPTION--1974
(metric tons)
   Airborne Land
Reference  Total Emi ss ions Disposal
Holt and Moberly (1976) 3350 170-340 3000
OTS (1976)  2450 . 650 1800
34

-------
   Table V-J  
 STATE TOTALS OF COAL-FIRED POWER PLANTS AND 
 . FUEL CAPACITY, BASED ON EPA LISTING 
   Total Fuel Capacity Fuel
 State Plants (1000 ton/vr) ~
Alabama  10 17,849.5 B
Alaska  2 164.1 B
Arizona  2 6,286.5 B
Colorado  9 7,231.2 B
Delaware  1 782.7 B
. District of Co It:mb i a 1 0.02 B
Florida  6 6,067.9 - B
Georgia -  7 8,775.1 B
Illinois  25 31,736.0 'B
Indiana  25 28,207.2 B
Iowa  22 6,459.8 B
',Kansas  7 3,506.8 B
Kentucky  16 22,955.8 B
~1ary land  6 5,083.1 B
Nichigan  27 20,250.1 B
Minnesota \  18 8,981.6 B
 2 671. 5 L
Hississippi  2 1,687.0 B
Nissouri  17 18,678.6 B
Nontana  2 2,060.6 B
Nebraska  4 1,996.4 B
Nevada  2 4,849.9 B
New Hampshire 1 750.9 B
New Jersey  3 2,604.2 B
New Nexico  2 9,667.6 B
35

-------
   Table V-3 (c~nc1uded) 
    Total Fuel Capacity Fuel
 State  Plants (100 ton/vr) ~
New York  10 6,017.4 B
North Carolina 13 20,961.8 B
North Dakota  5 6,320.4 L
Ohio   34 49,265.4 B
  I 26 36,088.9 B
Pennsylvania  
  J 3 1,349.7 A
South Carolina 9 5,550.7 B
 Dakota I  2 233.5 B
South    
  1 2,258.8 L
Tennessee  8 18,558.0 B
Texas I  1 332.7 B
 2 12,018.4 L
Utah   4 1,263.8 B
Vermont  1 8.7 B
Virginia  6 5,666.6 B
\~ashington  1 4,106.9 B
West Virginia  12 27,433.2 B
\-lisconsin  19 10,842.9 B
\~yoming  5 8,982.5 B
Total  381 434,564.4 
A - Anthracite coal
B - Bituminous coal
L - Lignite coal
36

-------
conventional control equipment. NaLusch et ~1. (1974) found the arsenic
content of particles emitted by coal-fired L,ower plants to increase as
particle size decreases (Table V-4). Klein et ale (1975) showed that
the arsenic concentrations in fly ash from coal-fired power plant out-
lets are 100 times the concentration in slag.' .
Table V-4
ARSENIC CONCENTRATION OF AIRBORNE
FLY ASH PARTICLES EMITTED BY
. COAL-FIRED"POWER PLANTS
Particle 
Diameter Arsenic
 (~m) (ppm)
> 11. 3 ' 680
7.3-11.3 800
4.7- 7.3 1,000
3 . 3- 4. 7 900
2.1- 3.3 1,200
1.1- 2.1 1,700
Source:
Natusch et ale (1974~.
It has been estimated that 73% of the arsenic is captured either in
bottom fly ash or collected fly ash and that 27% is .re1eased in uncon-
trolled stack emissions (Davis, 1971). Ba.sed on a mass balance of trace
impurities, Bolton et ale (1975) concluded that electrostatic precipita-
tors were 95% to 98% efficient in the recovery of arsenic. A we11-
controlled power plant was found to discharge 0.2 g/min of arsenic to
the atmosphere (Klein et a1., 1975); The power plant consumed 110 ton/hr
of coal at peak; because the arsenic flow into the plant with the coal
was estimated as 6 g/min, 3.3% of the arsenic input was released as stack
emissions. Ferguson and Gravis (1972) have calculated that 2.5 g of
arsenic are released into the atmosphere for every ton of .coal consumed.
Youngblood (1978) used dispersion modeling to estimate the atmo-
spheric arsenic concentrations in the vicinity of p~er plants. Three
- -,....",
37

-------
cases were modeled corresponding to the stack characteristics of 25-,
250-, and 1000-MW power plants. All three plants were assumed to have
abnormally high arsenic emissions' of 100 g/s to enable the model's
detection of significant atmospheric arsenic concentrations. The re-
sults of the modeling of the three cases are given in Table V-5.
Table V-5
ESTIMATED ANNUAL AVERAGE
ATMOSPHERIC ARSENIC CONCENTRATIONS FOR
COAL-FIRED POWER PLANTS BASED ON
DISPERSION MODELING
Distance From
Plant (km)

0.3
0.8
1.3
2.0
3.0
5.0
8.0
12.0.
16.0
20.0
Atmospheric Arsenic
. Concen tra tion
(~g/m3)
Plant BX
Plant AX
0.096
2.784
2.928
2.296
1. 680
1. 240
0.920
0.656
0.496
0.392
**
0.480
0.526
0.438
0.395
0.326
0.222
0.156
0.135
0.122
Plant CX
**
0.016
0.196 .
0.260
o. 198
0.140
0.108
0.080
0.066
0.054
* .
All three plant types were assumed to have an
arsenic emission rate of 100 g/s. Plant type A
corresponds physically to a 25-MW power plant;
plant type B to a 250-MW power plant; and plant
type C to a' 1000-MW power plant.
**
Negligible.
Source:
Modified from Youngblood (1978)
D.
Population Exposures
In estimating population exposures, it is assumed that the average
arsenic content of coal is 5.44 ppm, with a worst case of 14 ppm. Be-
cause a listing of the type of emission controls at each power plant is
unavailable, two cases are evaluated: Case 1 assumes that all power
plants are well-controlled, with 3.3% of the coal's arsenic emitted
38

-------
(arsenic emissions of 0.16 g/ton for S.44 ~pm coal); Case 2 assumes
that all power plants are poorly controlled, with 27% of the coal's
arsenic emitted (arsenic emissions o~ 1.3 g/ton for S.44 ppm coal).
The power plant coal capacity from the EPA plant listing is used to
estimate arsenic ~missions. This assumption should lead to an over-
estimate of exposures because power plants infrequently operate at
capacity, and some plants are operated intermittently for standby re-
serve or for peak demand.
The dispersion ~odeling results for the three cases given in Table
.V-S are applied by first estimating the atmospheric arsenic emissions
for each power plant in terms of grams per second. . The ratios of these
estimated emissions to 100 g/s (the emission on which the modeling re-
sults are based) are calculated and used to proportionately scale the
atmospheric arsenic concentrations from modeling given in Table V-So
The 2S-MW power plant characteristic dispersion curve was used to scale
concentrations for all power plants smaller than 100 MW; the 2S0-MW dis-
persion curve was used to scale concentrations for all plants of 100 to
SOO MW; and the 1000-MW dispersion curve was used to scale concentrations
for all power pldnts larger than SOO MW.
When this procedure was applied, it became apparent that no environ-
mental arsenic exposure for S.44 ppm coal would exceed 0.001 ~g/m3. Con-
sequentl", a number of worst-case exposures were estimated, leading to
the conclusion that the very worst-case annual average exposure does not
exceed 0.003 ~g/m3 (the national average urban exposure). These worst-
case analyses are given in Table V-6. The largest coal consuming power
plants were selected from the EPA listing for each of the three plant
types defined in Table V-5. These largest plants consumed 380,000,
3,200,000, and 7,500,000 tons of coal per year, respectively, for the
25-, 250-, and 1000-MW dispersion curve classifications. Well controlled
and poorly controlled cases were evaluated by assuming average arsenic
in coal (5.44 ppm) and high arsenic in coal (14 ppm). The results given
in Table V-6 indicate that most maximum annual exposures are less than
0.001 ~g/m3 and the very worst exposure is 0.0021 ,~g/m3. Additional
calculations assuming that all the arsenic in the coal is released to
the atmosphere lead to estimated very worst case maximum annual-average
environmental arsenic concentrations of 0.003 ~g/m3 for 5.44 ppm coal and
0.008 ~g/m3 for 14 ppm coal.
E.
Summary
Because the realistic worst-case annual average environmental arsenic
exposures for coal-fired power plants are less than 0.003 ~g/m3 for all
power plants and less than 0.001 ~g/m3 for most power plants, it must be
concluded that power plant emissions do not add appreciably to nominal
urban background concentrations.
39

-------
Table V-6
MAXIMUM AVERAGE ANNUAL
ATMOSPHERIC ARSENIC CONCENTRATIONS
IN THE VICINITY OF LARGE POWER PLANTS
Conditiona
Average coal arsenic-poorly controlled
b
Plant Type and
Consumption
( ton/vr)
A - 380,000
B - 3,200,000
C - 7,500,000
A - 380,000
B - 3,200,000
C - 7,500,000
A - 380,000
B - 3,200,000
Average coal arsenic-well controlled
High coal arsenic-poorly controlled
C - 7,500,000
High coal arsenic-well controlled
380,000
A-
B - 3,200,000
C - 7,500,000
Arsenic
Concentrationc
(ug/m3)
0.000 5
0.000 7
0.000 8
0.000 0
.0.000 1
0.000 1
0.001 2
0.001 8
0.002 1
0.000 1
0.000 2
0.000 3
a .
Average coal arsenic is 5.44 ppm, high coal arsenic is 14 ppm, poorly
controlled allows 27% arsenic emissions, and well .controlled allows
3.3% arsenic emissions.
b
Plant types A, B, and C correspond to the dispersion curves given in
Table V-5.

cFrom distance of highest concentration given in Table V-5.
40

-------
A.
VI
ARSENIC EXPOSURES FROM NONFERROUS SMELTERS
General
Ores of two types have elevated arsenic concentration: sulfide.
deposits, associated with copper, lead, zinc, and other ores; and sedi-
mentary deposits such as iron ore, phosphate rock, borax ore, ma~ganese
ore, and fossil fuels. The very high temperatures required to smelt
metallic ores generally result in the release of a large portion of the
naturally occurring arsenic to the atmosphere. Both ele~ental arsenic
and its common oxide, ASZ03' are extremely volatile at typical smelting
~emperatures (OTS, 1976). .
Three factors, aside from the inherent volatility of ASZ03' con-
tribute to the generally high losses of this material to the atmosphere
during smelting:
B.
(1)
A.~203 is slow
cool. Hence,
precipitators
peratures are
(OTS, 1976).
to condense as higher temperature flue gases
it may pass through baghouses or electrostatic
as a supersaturated vapor even if their tem-
below the equilibrium sublimation temperatures
( 2)
Because dust collection devices such as electrostatic pre-
cipitators and baghouses are routinely operated at elevated
temperatures, they remain well above the dew point of the flue
gas (OTS, 1976).

The nonferrous metals industry generally recycles ~ollected
flue dusts until concentrations of valuable metals build up
sufficiently for economical processing. At £ach stage of re-
cycle, the volatile AsZ03 has another opportunity to escape
collection (OTS, 1976).
( 3)
Primary Nonferrous Smelter Locations
Currently 16 copper, 6 lead, and 7
are in operation in the United States.
is shown on Figure VI-l..
zinc primary nonferrous smelters
The location of these smelters
1.
Copper Smelters
The 16 primary copper smelters and their annual production figures
are listed in Table VI-l. Seven of the smelters are located in Arizona,
two in New Mexico, with one each in Washington, Utah, Texas, Tennessee,
Nevada, Montana, and Michigan. Secondary copper smelters account for
41

-------
h r-,-
-£j --
~' ""j-.- /"-,
, -- . .
. , ' ----- ( \
i. 't' -----,.-..-- j\ , ~
; I .--.--..,--- \. , .
',- , ", I, _..,- ~
'----. \ ' I, .....-.. y
~'-, -J. , , ., ,
I (.. 0 I I ~ --1- )\
" I I. ,0 ,.. , ! \
/ --' ....---- , . \.
" ',. r .--...--.-.--- -I'" \. i
i ,-;,r--.-.---., ( U '), i .
'- I , -'~'~'-'-'1 " '.... ~.'
-.-- ; i It'
~ ~'- -.' , t I. ""\, Ir-'--:-
, -._,~ , _.~ .
! '- ~.~ -, I ,-.--.--.. .......\ --.-.-. .' i

;' :~'~-'-'-j '. t--.-.-.-.-.-.-......-,) '",.-.--.- \ -.-. r~.
, ," ) -~._.....- ,
I 0; 0 l._.- - I \ I
, . --.-.. I J. r \, -'f
, , . ~ -.- - -_._._.~ .. , ., "-'-.-'-.- ""'\-.....
"''-' I , j I J ,. . \-._,-' - -.. .I~ ' \
\ . '. . ,--.---....". i \ ) .,. r.ro. ~.' ,
" , t.. , J " ,,\-
, ! ! ~_.--_._._._._...._._....- \ I l .: / It.~-..
, , I' ., C I ... " _I'
\' ." , -,..", ,,~ .
, i I ..... . J ,- . . "' . V'
..; , I I., ,.J \) -..-;.
. "~.-._._.- i' . ...... ~4''''''' ,.....-- -.- ;?-.
l.. '\' -.- -'""1-.- - -.-' I \ ,oJ ,.1 -.-.-.-.--.-. ~~"\.
"' N , -'-.-.-.-.' I -- --- -.-" -,
, ' , -''''''-'-'--'--'-'-'-'.'-'-'-'-'1- ~ ---.--.-.-.-./ 7i%)
, I I . -----. --'-'--'_'_'JL._..r .... Y.
"'L ' r.. . .-,'
. '. i ,. I \ t 0' - -'.'-.
" I .. I f_.-."",
. .' / I: " --'-'i,'-'-'-,\-'- f,,\. ".
I , I ~ . "
_: 0 , . -- I ,...
.... 0 000; : '...,-..-~,4. I \ ....
", 0;0 i t 1 I \ ' .
............. . i , .--'-'-- . I \
...: 0 I -.--. ) I
'-.-- r'-' 0 -'-'-.-.J I \
-. \. , I
',. .. , '.-.- r
" . \ \....._., i \"-.- ..--.---"

"'J~ ~J.-.-.~'t"'~
\ ..-?~
'\ {
'--~
~
I'.)
.
LEAD SMELTER
COPPER SMELTER
ZINC SMELTER
o
.
l
o
SCALE IN MILES
. .
1:>0 200 300 400
...,
Source:
NIOSH. 1976,
FIGURE VI-1.
DOMESTIC PRIMARY NONFERROUS SMELTERS

-------
Table VI-1
PRIMARY U.S. COPPER SMELTERS
Company/Location
Annual Production
(metric tons per year)
Phelps-Dodge
Morenci, AZ

Kennecott
Hayden, AZ
160,000
73,000
Cities Services
Copperhill, TN

Anaconda
Anaconda, MT
17,000
170,000
Kennecott
Hurley, NM

Kennecott
McGill, NV
77 ,000
61,000
Kennecott
Garfield, UT
250,000
Nagma
San Manuel, AZ

\.Jhite Pine
White Pine, MI
100,000
73,000
Inspiration
Miami, AZ
. 99,000
Phelps-Dodge
Ajo, AZ

ASARCO
E1 Paso, TX
65,000
66,000
ASARCO
Hayden, AZ

ASARCO
Tacoma, WA
120,000
99,000
Phelps-Dodge
Douglas, AZ

. Phe 1 ps- Dodge
Hidalgo, NM
120,000
91,000
Source:
OAQPS (1974)
43

-------
about 30% of the U.S. copper demand (Environmental Sciences and
Engineering, 1976) and are situated close to a source of scrap or near
inexpensive transportation. They are believed to produce much less
arsenic pollutant than the primary smelters.
The primary copper industry employs a pyrometa11urgical process.
Primary copper smelters conventionally produce blister copper after
roasting, smelting, and converting. In some cases, the blister copper
is purified by fire refining. If further purification is desired, an
electrolytic process is used to produce cathod~ copper. Copper ptoduc-
tion in 1974 is estimated as 1,470,000 metric tons with 1,440,000 metric
tons from domestic sources and 30,000 metric tons from foreign sources
(OTS, 1976).
2.
Lead Smelters
The domestic lead industry is comprised of six smelters. Their
locations are described in Table VI-2. T~ree plants are located in Missouri,
with one each in Montana, Idaho, and Texas. The smelter in E1 Paso, Texas
produces both copper and lead, and the smelter in Idaho produces both lead
and zinc. A lead refinery owned by ASARCO is located in Omaha, Nebraska.
Table VI-2
PRIMARY U.S. LEAD SMELTERS
Company/Location
Annual Production
(metric tons per year)
AS ARC 0
Glover, MO
74,000
AS ARC 0
East Helena, MT

ASARCO
E1 Paso, TX
35,000
50,000*
Bunker Hill
Ke 11ogg, ID

St. Joe Minerals
Hcrcu1aneum, MO
120,000
180,000
Missouri Lead Operating
Boss, MO
130,000
*
The Bureau of Mines (1978) lists the ASARCO-E1 Paso
capacity as 80,000 metric tons per year.
Source:
OAQPS (1974)
44

-------
The six lead smelters employ pyrometall~~gical smelting and use
domestic or foreign sulfide ores. Lead prodJction in 1974 is estimated
to be 608,000 metric tons, of which 526,000 metric tons was from domestic
concentrates and 82,000 metric tons was from imported concentrates (OTS,
1976) .
3,
Zinc Smelters
The domestic primary zinc industry is comprised of seven smelters that
produce zinc. Their locations and" production capacities are described in
Table VI-3. Two plants are located in Texas, two in Pennsylvania, and one
each in Idaho, Illinois, and Oklahoma. An eighth smelter owned by ASARCO
and located in Columbus, Ohio roasts concentrates to make calcine.
Table VI'-3
PRIMARY U.S. ZINC SMELTERS
Company/Location
Annual Production
(metric tons per year)
ASARCO
Amarillo, TX

ASARCO
Corpus Christi, TX
. *
45 , 000
99,000
AMAX
Sauget, IL

Bunker Hill
Kellogg, ID
64 ,000 t
99,000
Ne~l Jersey Zinc
Palmerton, PA

St, Joe Minerals
Monaca, PA
100,000
210,000
National Zinc
Barte1svi11e, OK.
46,000
J..
"Production for the old Amarillo smelter .
~ .
Estimate from Environmental Sciences and
Engineering (1976)",
Source:
OAQPS (1974)
45

-------
Plans to expand production include one new smelter in Tennessee
scheduled to come on line in 1979. In add~tion, the old horizontal
retort smelter of National Zinc at Bartlesville, Oklahoma was replaced
with a new electrolytic plant. ASARCO may expand its Corpus Christi
plant (Environmental Sciences and Engineering, 1976).
Primary zinc production in 1977 was estimated at 411,114 metric
tons, of which 299,825 metric tons were from domestic concentrates and
11,289 metric tons were from imported concentrates. It is further esti-
mated that 194,500 metric tons were produced pyrometallurgically and
216,614 metric tons were produced electrolytically (Bureau of Mines,
1978).
c.
Samplin~ of Atmospheric Arsenic Concen~rations Near Nonferrous
Smelters
From 1973 to 1975, EPA analyzed the atmospheric arsenic concentra-
tions near a number of the nonferrous smelters. Most locations had one
monitoring site; however, there were 10 monitoring sites in El Paso,
Texa~. Concentration data were analyzed by neutron activation analysis
and atomic absorption. The neutron activation analysis has a lower de-
tection limit of 0.003 ~g/m3 and the atomic absorption has a lower de-
tection limit of 0.001 ~g/m3 (Shearer, 1975). Data for the ~ndividual
locations are summarized in Table VI-4. E1 Paso concentration data are
summarized in Table VI-5. In addition, the concentrations as a function
of distance from the El Paso smelter are given in Figure VI-2.
Roberts et ale (1976). collected atmospheric arsenic concentrations
for four monitoring points in Tacoma during November and December 1975
(Table VI-6). These data are a1so plotted on Figure VI-3 as a function
of distance from the ~elter.
As part of the Helena Valley, Montana, air pollution study (U.S.
EPA, 1972), atmospheric arsenic concentrations were recorded for five
monitoring sites near the East Helena smelter. Data from this program
are summarized in Table VI-7 and plotted as a function of distance from
the smelter on Figure VI-4. The plot on Figure VI-4 ~1so shows the one
NASN site from Table VI-4. .
D.
Atmospheric Emissions
Although some atmospheric monitoring data for the vicinity of non-
ferrous smelters are available, these data are insufficient to charac-
terize population exposures for all nonferrous smelters. It is there-
fore necessary to use atmospheric dispersion modeling to estimate
exposures in the vicinity of those smelters for which few or no moni-
toring data exist. Youngblood (1978) performed dispersion modeling to
46

-------
Table VI-4
*
EPA ATMOSPHERIC ARSENIC SAMPLES TAKEN NEAl' NOt-FERROUS SMELTERS
  Sampling Site    3 
  Location in Test  uf1./m As
  Relation to Stack Methodt ~ Averaf1.e Ranl!;e
Hurley, NM 2.0 lan ESE NAA  NA  
   AA  NA  
Anaconda, MT 5. 9 lan NE NAA 28 0.269 0.006-0.854
   AA 7 0.269 . 0.073-0.719
East Helena, MT 2.6 lan S NAA 52 0.043 0.000-0.185
   AA 11 0.047 0.002-0.159
Carfie1d, UT 6.4 lan ESE NAA 42 0.347 0.002-1. 355-
   AA 10. 0.380 0.093-0.883
Ajo, AZ  0.2 lan W NAA 32 0.009 0.000-0.053
   AA 11 0.014 0.001-0.044
Douglas, AZ 0.5 kI>: NE NAA 44 0.022 0.000-0.102
   NA 11 0.030 0.011-0.075
Hayden, AZ 0.9 lan SSE ~AA  NA  
   AA  NA  
Miami, AZ 3 . 8 lan NNE NAA 39 0.050 0.003-0.192
   AA 8 0.044 0.000-0.114
Morenci, AZ 0.2lanNW NAA  NA  
   .AA  NA  
San Manuel, AZ 1.5 lan N NAA 39 0.020 0.000-0.073
   AA 9 0.035 0.004-0.075
McGill, NV 3.3 lan SSW NAA 48 0.090 0.004-0.315
   AA 11 0.065 0.004-0.176
Kellogg, ID 0.8 lan W NAA 50 0.299 0.024-0.B48
   AA 11 0.220 0.024-0.546
*
24-hr samplea recorded during 1973 and 1974.

tNAA - Neutron activation analysis, AA - atomic absorption.

Source: . Based on data given by Shearer (1975)
Table VI-5
AVERAGE ATMOSPHERIC ARSENIC CONCENTRATIONS
FOR EL PASO, TEXAS
SampL.r:g Site   * (u.g/m~
 Annual Average
Lucation in    Most Recent
Relation to Stack 1973 1974 ..illl. 12 Months
0.27 lan W 0.666 0.293  0.300
0.46 lan NW 0.902 0.568 0.556 0.582
1. 75 lan ESE 0.098 0.074.  0.062
1. 84 lan NE 0.037 0.037  0.037
2.30 lan SE 0.103 0.072 0.055 0.056
3.14 lan N 0.098 0.096 -- 0.095
3.35 lan SSE 0.150 0.115  0.112
5. 34 lan ESE 0.023 0.032 0.029 0.026
6.70 lan SSE 0.073 0.089 0.059 0.060
6.81 lan NNE 0.076 0.081  0.076
*
There were generally four to eight 24-hr samples
available per month; however, 12 months of data were
not available for every year.
Source:
Cooper (l978A)
47

-------
 .8   
 .6 .  
 .4   
CO)    
!    
'"    
=l    
I    
Z .2   
Q   
I-    
c(    
II:    
I-    
Z    
IU   . 
(.)   
Z .1   
8  . 
~ .08   .
z    
IU    
CI) .06 .  .
a: 
c(  . 
(.)    
a:    
IU .04 .  
:%:  
A.    
CI)    
0    
~   .
I-  
c(    
 .02   
.01
.1
.2
.4
.6
.8
2
4
6
8 10
20
DISTANCE FROM THE SMELTER-km
FIGURE VI.2. ATMOSPHERIC ARSENIC CONCENTRATIONS AS A FUNCTION
OF DISTANCE FROM THE EL ASARCO SMELTER
48

-------
Table VI-o
ATMOSPHERIC ARSENIC CONCENTRATIONS
FOR TACOMA. WASHINGTON
~
\D
Sample Site Location Number of  \J g/m3
Relative to Smelter Samp1es* Average Range
0.28 km W 8 3.004 0.318-9.353
0.40 km SSW 9 1. 426 0.025-2.843
0.49 km SW 8 0.515 0.188-0.985
0.94 km S 7 0.314. 0.190-0.379
*
24-hr samples collected during November and December 1975.
Source:
Roberts et al. (1976)

-------
 10
 8
 6
 4
cot 
E 
CI 
:to 
I 
Z 
0 
~ 2
< 
a: 
~ 
z 
Y.I 
U 
Z 
0 
U 
~ 
z .s
Y.I
U) 
a: 
< .6
~ .
a:
Y.I 
% .4
~
U)
o 
:E 
~ 
< 
 .2
.1 .
.1
4
DISTANCE FROM THE SMELTER-km
FIGURE VI-3. ATMOSPHERIC ARSENIC CONCENTRATIONS AS A FUNCTION
OF DISTANCE FROM THE TACOMA SMELTER
50

-------
Table VI-7
ATMOSPHERIC ARSENIC CONCENTRATIONS RECORDED FOR
TilE HELENA VALLEY ENVIRONMENTAL POLLUTION STUDY
VI
......
   Sample   3
Sampling Site Location Test Size As Concentration (p glm )
Relative to Smelter Method* (days)  Average Range
1. 29 km, 340 GF 76  0.006 0.000-0.070
  MF 28  0.007 0.000-0.020
4.02 km, 1050 GF 87  0.011 0.000-0.090
  MF 25  u.006 0.000-0.020
0.64 km, 1120 GF 85  0.060 0.000-0.260
  MF 23 .. 0.082. 0.000-0.400
7.24 km, 2740 GF 82  0.005 0.000-0.070
  MF 8  0.009 0.000-0.010
 o GF 34  0.084 0.000-0.260
0.80 km, 2 
  MF     
*
GF - glass fiber, MF - membrane filter.
Source:
U.S. EPA (1972)

-------
.08
'"
~
= .D6
:L
I
Z
52 .eM
~
<
II::
~
Z
w
C,,)
Z
o
. C,,) .02
C,,)
z
w
en
II::
<
S:! .01
II::
w
~ .008
en
o
~ .006
<
.D04
.002
.1
.2
. NASN SITE
. HELENA VALLEY
STUDY SITES .
.1
.
.
.
.
.2
.6 .8. 1.0 2
DISTANCE FROM THE SMELTER-km
8 10
.4
6
4
FIGURE VI-4. ATMOSPHERIC ARSENIC CONCENTRATIONS AS A FUNCTION
OF DISTANCE FROM THE EAST HELENA SMELTER
52
20

-------
assist in estimating atmospheric arsenic con~entrations near nonferrous
smelters. Both stack and low-level fugitivt... emissions must be considered.
Table VI-8 shows results of the dispersion modeling for hypothetical
conditions for four copper smelter stacks. Table VI-9 shows estimated
atmospheric concentrations for three levels ~f fugitive emissions at
copper smelters. Tables VI-lO and VI-II shew estimated atmospheric con-
centrations for stack and fugitive emission~ for hypothetical lead and
zinc smelters.
Table VI-8
ATMOSPHERIC DISPERSION MODELING FOR ARSENIC CONCENTRATIONS
IN THE VICINITY OF COPPER SMELTERS DUE TO STACK EMISSIONS ALONEa
Distance from
Smelter (km)
Ajo Main
(E=102. 9) c
Tacoma Main
(E=222.9)C
Arsenic Concentration (~g/m3)b

Magma Ma iri Magma Ac id
(E=385.7)c (E=1.63)c
Annual Average Atmospheric
.0.3
0.8
1.3
2.0
3.0
5.0
8.0
12.0
16.0
20.0
0.210
0.205
0.200
0.130
0.086
0.073
0.058
0.047
0.039
- Negligible -
0.190 0.092
0.180 0.220
0.160 0.150
0.130 0.120
0.100 0.110
0.071 0.096
0.053 0.071
0.043 0.055
0.040 0.045
0.008
0.007
0.005
0.004
0.003
0.002
0.001
0.001
0.001
aStack characteristics correspond to the stacks at Ajo, Tacoma; and
Nagma; however, arsenic emissions were arbitrarily selected for dis-
persion modeling.

bOmnidirectional average obtained by dividing 24-hr downwind concentra-
tion dispersion modeling results by 12.5.

cArsenic emission rat~ in lb/hr.
Source:
Based on Youngblood (1978)
53

-------
Table VI-9
ATMOSPHERIC DISPERSION MODELING FOR ARSENIC CONCENTRATIONS DUE SOLELY
TO FUGITIVE EMISSIONS IN THE VICINITY OF COPPER SMELTERS
  Annual Average Atmospheric Arsenic a)
  Concentration6 (~g/m )
 Distance from . Case 1 Case 2 Case 3
 Smelter (km) (E=20.6)b (E=89. 2) b (E=115. 7) b
 0.3 5.200 22.500 29.200
 0.8 2 . 100' 8.800 12.000
 1.3 1.040 4.500 6.000
 2.0 0.570 2.400 3.200
VI 3.0 0.320 1.400 1.800
~
 5.0 0.160 0.680 0.880
 8.0 0.080 0.340 0.430
 12.0 0.044 0.190 0.240
 16.0 0.030 O. 120 0.160
 20.0 0.021 0.088 0.110
aOmnidirectional average obtained by dividing
dispersion modeling results by 12.5.
b
Arsenic emission rate in lb/hr.
24-hr downwind concentration
Source:
Based on Youngblood (1978)

-------
Table VI-10
ATMOSPHERIC DISPERSION MODELING FOk ARSENIC
CONCENTRATIONS IN THE VICINITY OF A LEAD SMELTERa
Distance from Smelter (km)

0.3
0.8
1.3
2.0
3.0
5.0
8.0
12.0
16.0
20.0
Atmospheric Arsenic Concentration
(uS!:/m3)
Stack
(E=18. 3) b
Annual AveraS!:ec

Negligible
0.014
0.014
0.015
0.014
0.007
0.006
0.005
0.004
Q.004
Fugitive
(E=5.5) b
Annual Avera!l;ec
1.380
0.540
0.260
0.150
0.083
0.040
0.020
0.011
0.008
0.005
as tack characteristics correspond to those of St. Joe in
Herculaneum.

bAssumed arsenic emission rate in

cOmnidirpctional average obtained
concentration by 12.5.
lb/hr.
by dividing the 24-hr downwind
Source:
Based on Youngblood (1978)
Table VI-11
ATMOSPHERIC DISPERSION MODELING FOR ARSENIC
CONCENTRATIONS IN THE VICINITY OF A ZINC SMELTERa
Distance from Smelter (km)

0.3
0.8
1.3
2.0
3.0
5.0
8.0
12.0
16.0
20.0
Atmospheric Arsenic Concentration
(u,S!:/m3)
Stack Fugitive
(E=34.7)b (E=6.9)b
Annual AverageC Annual AveraRec
Negligible
0.070
0.048
0.041
0.035
0.022
0.018
0.014
0.011
0.009
1.700
0.680
0.350
0.180
0.100
0.052
0.026
0.014
0.010-
0.007
aStack characteristics correspofid-to those of New Jersey Zinc.
bAssumed arsenic emission rate in lb/hr.
cOmnidirectional average obtained by dividing the 24-hr downwind
concentration by 12.5.
Source:
Based on Youngblood (1978)
55

-------
The difficulty in applying the atmospheric modeling results to
specific smelters is in estimating their stack and fugitive emissions.
Ore arsenic and nonferrous metal contents, smelting processes, and emis-
sion controls vary considerably for each smelter.
The arsenic content of copper concentrates varies widely for copper
ores, depending on the area in which they are mined. For this study it
is assumed that material processed by smelters in Arizona, New MexicQ;
Tennessee, and Nevada contains 0.015% arsenic, material smelted in
Montana contains 0.96% arsenic, whereas material processed in Texas,
Michigan, and Washington contains 0.80%, 0.4%, and 5.2% arsenic, respec-
tively (Holt and Moberly, 1976 and OTS, 1976).
Generally, it is expected that smelters that process ore concen-
trates with high arsenic content will have high arsenic emission. How-
ever, because of emission controls and smelting processes this ~s not
always the case. These differences are evaluated in more detail later
in this chapter.
Missouri lead concentrates are assumed
and lead concentrates from other states are
arsenic. All zinc concentrates are assumed
(Holt and Moberly, 1976; and OTS, 1976).
to contain 0.05% arsenic,
assumed to contain 0.1%
to contain 0.05% arsenic
Davis and Associates (1971) list average atmospheric arsenic emis-
sion factors to be 4.9 Ib/ton of copper, 0.8 Ib/ton of lead, and 1.3
Ib/ton of zinc. NAS (1977) states that in 1974 the average emission
factor for copper smelters was 2.1 Ib/ton. Using the Davis emission
factors, OTS (1976) estimates that 15% of the arsenic in copper concen-
trates, 25% in lead concentrates, and 35% in zinc concentrates is em~tted
to the atmosphere during smelting.
OAQPS (1974) lists stack particulate emissions for all primary non-
ferrous smelters. They do not, however, estimate the arsenic content on
the particulates. Davis (1971) presents data from a copper smelter in
Northern Chile showing the arsenic concentration on dust from the stack
to be 10 times higher than in the ore concentrate; Davis (1971) also
reports on another South American copper smelter in which the arsenic,
concentration on the flue dust was more than 9 times that of the ore
concentrate. Weisenberg and Serne (1976A) present data showing the
arsenic concentration of the particulates at Tacoma to be 10 times
higher th~n that of the ore concentrate. For four other U.S. copper
smelters, Weisenberg and Serne (1976B, C, D, E) show arsenic concentra-
tion on the particulates to be 5 to 25 times that 'of the ore concentrate.
Recently, the EPA made rough estimates, of c,urrent and "best possible"
arsenic emission~ for copper 'smel ter . stacks' -(EPA, "f978). However, no
estimates w~'re made of f~gitive' emissions; These' emission estimates were
updated in September 1978 (Verveart, 1978). Preliminary sampling results
by EPA of the El Paso smelter indicates that fugitive emissions may be
on the order of 5% of arsenic input. This may, however, be an over-
estimate for the Tacoma smelter which processes arsenic trioxide. ,EPA
56

-------
currently plans to test the Tacoma smelter t~r fugitive arsenic emissions
(Verveart, 1978).
The copper smelter industry has recently estimated the arsenic
emissions for the ASARCO-Tacoma, ASARCO-Hayoen, and Phelps Dodge-Ajo
copper smelters as 21, 6, and 24 lb/hr, respectively (Pendleton, 1978).
These emissions estimates are about five times smaller than the EPA
estimates used in this report.
In estimating atmospheric arsenic concentrations near nonferrous
smelters it is difficult to determine the relationship between fugitive
and stack emissions. Weisenberg and Serne (1976A) report 80 ton/yr
stack and 90 ton/yr low level arsenic trioxide emissions for the Tacoma
smelter in 1975. The Tacoma smelter is probably an unusual case because
of the high arsenic content of the ore and because of the additional
arsenic trioxide processing facilities. Cooper (1978B) suggests using
a range of 10%-20% of stack emissions to represent fugitive emissions.
E.
Human Exposures
BecauSE of the many unresolved uncertainties in arsenic emissions
for nonferrous smelters, population exposures are estimated by using
several assumptions. These assumptions have been selected to be broad
enough to attempt to span the actual exposures. In addition, the re-
corded ambient concentration data are used to'help select the "best
estimate" of exposures.
Populations residing near nonferrous smelters and subject to arsenic
exposures are estimated by two methods, depending on population distribu-
tion near the smelter. For rural locations, the populations residing
at various distances from the smelter were estimated by consulting maps
and measuring distances between the smelter and towns. For urban loca-
tions, the populations residing at various distances from the smelter
were estimated by using maps in which the population has been allocated
to l-km and S-km grids. Resident populations were assumed to be uni-
formly distributed within each grid. The 1970 census population was
used for urban and rural locations. Because of the uncertainties in the
dispersion modeling, exposures were estimated only for populations re-
siding within 20 km of each smelter.
1.
Copper Smelters
Population exposures to arsenic emissions from copper smelters were
made by. .using. six cases:
Case A
An arsenic emission of 4.9 Ib/ton of copper, as reported
by Davis (1971), for all smelters. Fugitive emissions
were estimated to be 10% of stack emissions.
57

-------
Case B
Case C
Case D
Case E
Case F
Case G
15% of the arsenic in the copper concentrate is emitted
to the atmosphere (OTS, 1976). The assumed arsenic con-
centrations of the ores, as previously reported for each
smelter, were used. The ore concentrate feed rates re-
ported by OAQPS (1974) were used. Fugitive emissions
'were estimated as 10% of stack emissions. The Tacoma
smelter was treated differently by assuming 80 ton/yr
stack emissions and 90 ton/yr low-level arsenic trioxide
emissions.
An.arsenic concentration of the particulates 10 times the
concentration of the ore concentrate. The assumed arsenic
concentrations of the ores, as previously reported for
each smelter, were used. .Particulate emission rates as
reported by OAQPS (1974) were used. Fugitive emissions
were estimated as 10% of stack emissions. The Tacoma
smelter was treated in the same manner as in Case B.

Uses the recent EPA estimate of current stack emissions
and assumes that fugitive emissions are 5% of arsenic
input.
Uses the recent EPA estimate of current stack emissions
and assumes that fugitive emissions are 5% of arsenic
input for. all smelters but Tacoma. The fugitive arsenic
emissions for the Tacoma smelter are taken as 90 T/yr.

Uses the recent EPA estimates of current stack emissions
and assumes that fugitive emissions are 10% of stack
emissions.
Uses industry estimated emissions for the ASARCO-Tacoma,
ASARCO-Hayden, and Phelps Dodgp.-Ajo smelters. EPA
estimates of current stack emissions were used for all
other smelters. Fugitive emissions are assumed to be 5%
of arsenic input.
Each set of assumptions leads to the generation of a set of emission
factors for the copper smelters. These emission factors are used to
scale the dispersion estimates to make population ~~T 'sure estimates.
The estimated population exposures for ~opper smelters are given in
Table VI-12.
2.
Lead Smelters
Three sets of assumptions were used in estimating population" expo-
sures to arsenic emissions from lead smelters:
. .Case A
The arsenic emission factor of 0.$ l~ton as reported by
Davis (1971) for all smelters. Fugutive emissions were
estimated to be 10% of stack emissions.
58

-------
Table VI-12
ESTIMATED HUNAN POPULATiUN EXPOSURES TO ATMOSPHERIC ARSENIC EMITTED BY COPPER SMELTERS
Annual Av"rage
Atmuspherlc Arsenic
CuncentraLlona
(\
-------
Case B
Case C
25% of the arsenic in the lead concentrate emitted to
the atmosphere (OTS, 1976). Lead concentrates from
Missouri were assumed to contain 0.05% arsenic, and con-
centrates from other states were assumed to contain 0.1%
arsenic. Fugitive emissions were estimated to be 10% of
stack emissions.

An arsenic concentration of the particulates 10 times the
arsenic concentration of t~e ore concentrate. The arsenic
content of the ores are assumed to be the same as in Case
B. The particulate emission rates reported by OAQPS (1974)
were used. Fugitive emissions were estimated as 10% of
stack emissions.
The exposure estimates are given in Table VI-13.
3.
Zinc Smelters
Five sets of assumptions were used in estimating population exposures
to arsenic for zinc smelters:
Case A
Case B
Case C
Case D
C~se E
An arsenic emission of 1.3 Ib/ton of zinc for each smelter
(Davis, 1971). Fugitive emissions are estimated to be
10% of stack emissions.

Same as Case A, except only the pyrometallurgical smelters
have arsenic in their stack emissions (OTS, 1976). Fugi-
tive emissions are assumed to be 10% of the Davis emission
factor (1.3 lb/ton) for stack emissions.
36% of the arsenic in the zinc concentrates is emitted to
the atmosphere (OTS, 1976). All zinc concentrates are
assumed to contain 0.05% arsenic. Fugitive emissions are
estimated to be 10% of stack emissions.

Same as Case C, except only the pyrometallurgical smelters
have arsenic in their stack emissions (OTS, 1976). Fugi-
tive emissions are assumed to be 10% of the. stack emissions.
.Arsenic concentrations of the particulates 10 times the
arsenic concentration of the ore concentrate. Fugitive
-emissions are estimated to be 10% of stack emissions.
The exposure estimates are given in Table VI-14.
4.
Comp~rison of Dispersion Modeling to Ambient Data
Some inferences can be drawn about which of the dispersion modeling
cases to use by comparing the estimated concentrations for these cases
with the ambient.data recorded in Tables VI-4, -5, and -6. Table VI-15
compares measured and predicted atmospheric arsenic concentrations near
copper smelters.
60

-------
Table VI-13
ESTIMATED HUMAN POPULATION EXPOSURES TO ATMOSPHERIC ARSENIC
EMITTED BY LEAD SMELTERS
Annual Average
Atmospheric Arsenic
Concentrationa
- (~~/m3)

0.10-0.29
0.060-0.099"
0.030-0.059
0.010-0.029
0.005-0.009
0.003-0.004
(J\
to-'
Assumed Emissions
Case A
0.8 lb/ton
10% Fu~itiveb

800
2.600
5.100
38.000
46.000
61.000
Case B
25% Arsenic Emitted.
1070 Fu~itivec
Case C
Particulate
Concentration lOx.
10% Fu~itived
2.300
1 . 200
26.000
51.000
10.000
550
1.100
800
aAnnual "omnidirectional average.
b
Assumes 0.8 lb of arsenic emitted to the atmosphere for each ton of lead pro-
duced. Fugitive emissions were assumed to be 10% of stack emissions.

cAssumes 25% of arsenic in concentrates emitted in the atmosphere. Fugitive
emissions were assumed to be 10% of stack emissions.
d "
Assumes arsenic concentration of particulate emissions is 10 times the concen-
tration of the learl ore concentrates. Fugitive emissions were assumed to be
10% of stack emissions.

-------
Table VI-l4
ESTIMATED HUMAN POPULATION EXPOSURES TO ATMOSPHERIC ARSENIC'
EMITTED BY ZINC SMELTERS
    Assumed Emissions 
 Annual Average    Case D 
 Atmospheric  Case B Case C 36% Arsenic Case E
 Arsenic Case A  1. 3 Ib/ton at 36% Arsenic Emitted by Particulate
 Concentrationa 1. 3 -1 b/ ton Pyrometallurgical, . Emit ted Pyrometa11urgical, Concentrate lOx,
 (~g/m3) 1010 Fugitiveb 10% Fugitivec 10% Fugitived 10% Fugitivee 1070 Fugitivef
 0.10-0.29 52,000 9,000 3,000 3,000 
 0.060-0.099 22,000 18,000 6,000 6,000 
 0.030-0.059 155,000 101,000 47,000  
0\ 0.010-0.029 1,182,000 170,000 189,000 119 ,000 2,700
N 0.005-0.009 1, 811, 000 . 110,000 344,000 144,000 33,000
 0.003-0.004 800,000 13 , 500 1,176,000 13 7 , 000 10,000
a
Annual omnidirectional average.
b
Assumes 1.3 Ib of arsenic emitted to the atmosphere for each ton of zinc produced. Fugitive emissions
were assumed to be 10% of stack emissions.
c .
Assumes 1.3 Ib of arsenic e~itted to the atmosphere for each ton of zinc produced by pyrometallurgical
smelters and no stack emisslons at electrolytic smelters. Fugitive emissions were assumed to be 10%
of stack emissions.
d
Assumes 36% of arsenic in concentrates emitted to the atmosphere.
be 10% of stack emissions.

eAssumes 36% of arsenic in concentrates emitted to the atmosphere by pyrometallurgical smelters and no
stack emissions at electrolytic smelters. Fugitive emissions were assumed to be 10% of stack emissions.
f
Assumes arsenic concentration of particulate emissions is 10 times the concentration of the zinc ore
concentrates. Fugitive emissions assumed to be 10% of stack emissions.
Fugitive emissions were assumed to

-------
'Table VI-IS
COMPARISON OF MEASURED AND PR~DICTED ATMOSPHERIC ARSENIC CONCENTRATIONS
NEAR COPPER SMEL TERS(J (~g/m3)
Case A
:. .' b '
Predicted Arsenic,Concentration
Case B Case C Case D ' Case E
------
, , Place
Measured Arsenic
Concentration
Anaconda, MT
Garfield, UT
0.27
0.35

0.02

0.05
0.25
0.14

.10.80

, '0.12
0.24

0.06
0.09
0.02
1.85
0.24
, 4.57
0.02
Douglas, AZ
'Miami, AZ
San Manuel, AZ
, McGill, NV
0.02
0.09
0-
W,
Tacoma, WA
(1 km)
. c
El Paso, TX
(1 km)
c
E1 Paso, TX
. (7 km)
  .  
0.24  0.49  46.37
0.15 , 0.32  4.05
    ".
0.04 .. 0.08 ... 1.01
 : ;  
0.24 '
0.00
0.97
0.00
0.00
0.00
0.79
, ,
0.22
0.05 :
0.68 0.,68 
0.00 0.00 
0.11 0.11 
0.01' 0.01 :
0.04 0.04 "
0.00 0.00 
17.45 1.91 
1.,05 "
1.05
Case F.
0.02

0.00

0.09

Q.OO..
0.01,

"0. OO~
'"," '..I)
. (',
...:, ',.
'0. '87. "
i ..} $-
, .
\ ,',
, - ~ "
; 9. :in :'
" .
"
J
. '. :.
O'~ 06 " 0.01'
, ~... I, , ~
'~:I .
a ' .;. . " "" i:' .
Annual average' concentrations. Both measured and predicted concentration'~ :ar'e ~estC. .'
mated at the same distance from: the sme1te~'. ' . :' . ,. ,,:. ~r :':
b . ,; .
See T.a'.b1e VI-12 for a description :of the various cases~

C' \'" , , ' "
In addition, arsenic ,from ~ead smelting is a1so emitted to the environmentiri: E1Paso.
0.06
i' "1'.
. . ,t~ '
. ',' :., ,p). t:'
.. .. ':' ," f. ,

-------
These comparisons are not expected to be precise because the ambient
data are usually recorded for one monitoring point for a relatively few
days and because the predicted dispersion values are only rough approxi-
mations. Based on.these comparisons it. appears that Cases A, C, and E
give predictions nearer the actual concentrations than do any of the
other cases. Generally, a comparison of the measured and predicted values
by both Cases A and C agree within a factor of 4. Case A used the Davis
emission factor of 4.9 lb/ton, and for Case C the particulate arsenic
concentrations are 10 times that of the ore concentrate.
Few ambient monitoring data are available for comparison with dis-
persion modeling estimates for lead and zinc smelters. . table VI-16 shows
the comparisons for locations in East Helena and Kellogg. Based on these
comparisons, it is difficult to distinguish between the case that uses
the Davis emission factor and the case that uses a percent of the arsenic
in the ore concentrates. It is fairly clear that the estimates are too
low for the case that uses an arsenic concentration on the particulates
of 10 times that of the ore concentrate. "
table VI-16
COMPARISON OF MEASURED AND PREDICtED ATMOSPHERIC ARSENIC
CONCENtRAtIONS NEAR LEAD AND ZINC SMELtERS (~g/m3)
        *
 Measured Arsenic Predicted Arsenic Concentration
Place Concentration Case A Case B Case C Case D Case E
East Helena, M1'        
(lead) 0.03 0.12 0.01  0.01   
Kellogg, ID        
(lead) 0.30 0.33 0.09  0.00   
Kellogg, ID        
(zinc) 0.30 0.09 0 . {)O  0.18  0.00 0.00
*
See Tables VI-13 and VI-14 for a description of the cases.
F.
Summary
When compared, concentrations predicted by dispersion modeling (based
on selected assumptions) and of concentrations recorded by monitoring are
shown to generally agree within a factor of approximately 4. When all
the factors involved are considered, this is considered fairly close
agreement.
Population exposures were not estimated for conc~ntrations less than
0.10 ~g/m3 for "copper smelter emissions for some of the cases beca~se
64

-------
this would have entailed extrapolat~ng the .;ispersion modeling results
beyond 20 km)from the smelters. At. distances beyond 20 km from the
source, the dispersion estimates are considered to be unreliable.
. .
With Case A assumpti9ns ',(Davis emission factors), it is estimated
that 950,000 people are exposed to emissions from copper smelters, re-
sulting in annual average concentrations of 0.01 to 1.0 ~g/m3. The EPA
estimate of current stack emissions resulted in an estimated 923,000 ex-
posed to annual average concentrations .of 0.003 to 1. ° ~g/m3.
It is estimated that 160,000. people ~re exp9sed to annual average
concentrations of 0.003 to 0. 3 ~g/m3 due to .le'ad' smeiter emissions. This
estimate assumes that the' arsenit emission factor is 0.8 Ib/ton of con-
centrate and that fugitive emissions are 10% .of stack emissions. An
alternative estimate assumed that 25% of the arsenic in the ore concentrate
is emitted to the atmosphere as stack emissions" with 10% additional
fugitive emissions. This assumption resul~ed in an estimated 163,000
people exposed to annual average concentrations ofO.003'to 0.1 ~g/m3.
. ....
"
It is estimated that 442,000 people are exposed to annual average
concentrat ions of 0.003 to '0. 3 ~gjm3 of emissions trom Zinc smelters.
This estima';e assumes an arsenic emission factor of 1.3, l'b./ ton of con-
centrate fo~ pyrometallurgical zinc smelter stacks and no emissions from
electrolytic smelter stacks. . .. .
G.
Secondary Nonferrous Smelters
1.
General
Small quantities of arsenic. ar.e retained in .copper, lead,.. and
zinc after the refining process. ' In addition, arsenic -is .added. to. these
metals in the manufacture of .some products. When copper, lead,:.and zinc
scrap metals are salvaged and reprocessed, arsenic may be released during
the secondary smelting.
. ".... .". u
Table VI-17 contains estimates of the amount of arsenic retained
in or added to .copper, lead, and zinc. .Arsenic .is.added to copper for
the manufacture of admiralty brass and auto radiators, and it is added to
lead for the .manufacture o'f lead shot, bearing metals, and batteries and
battery cables. A total of 755 kkg of arsenic is estimated to be retained
in or added to .these metals each year.
2.
Secondary ~melter Locations
Total annual production figures for the secondary copper, lead,
and zinc industry are given as 513,308, 698,698, and 182,665 tons per
year (1974) by the American Metal Market (1977). Environmental Sciences
and Engineering (1976) list locations of secondary smelters. However, no
65

-------
Table VI-17
ESTIMATED ARSENIC RETAINED IN OR ADDED
TO COPPER, LEAD, AND ZINC METALS
Source
Retained in primary copper
Added to copper for admiralty brass
Added to copper for auto radiators
Retained in primary lead
Added to lead for lead shot
Added to lead for bearing metals
Added to lead for batteries, cables
Retained in primary zinc
Total
Source:
OTS (1976)
Arsenic
( kkSt/ vear)
43
7
100
20
60
25
495
-2
755
figures on individual plant capacity or production are available. The
Environmental Sciences and Engineering lists include 50 secondary copper
smelters, 40 secondary lead smelters, and 6 secondary zinc smelters.
3.
Secondary Smelter Emissions
Because arsenic is added during the manufacture of certain
products from nonferrous metals, arsenic emissions during secondary
smelting will depend, in part, on the material being refined. The
arsenic content of copper, lead, and zinc products in which arsenic has
not been added is estimated as 9, 33, and 10 ppm, respectively. These
concentrations are on the order of 30 to 1,000 times lower than the
arsenic concentrations in the original ore concentrates. Assuming that
arsenic emission factors for secondary smelting 'are 100 times less* than
those for primary smelting, secondary smelting emissions would result in
insignificant environmental atmospheric arsenic concentrations (average
annual exposures of less than 0.003 jJ.g/m3). . .
*
This assumption probably leads to an overestimate

the temperatures used for refining processes will

arsenic.
66
of emissions because
rarely release the

-------
Bec~use of the value of copper scrar, almost all auto radiators
are recycled. Auto radiator copper contains about 0.3% arsenic and is
normally fire refined. Fire refining removes li~tle.if any arsenic from
elemental copper because arsenic in elemental copper is very difficult
either to oxidize or to volatize. Some of this secondary refined product
is later electrolytically refined (to remove arsenic and other impuri-
ties), and some is used directly in products that require arsenic (OTS,
1975). Hence, atmospheric emissions of arsenic from secondary copper
refining of auto radiators are estimated to be near zero.
Lead from bearing materials and batteries is extensively re-
cycled. Lead tin bearings contain about 0.6% arsenic. The arsenic con-
tent of the battery alloys ranges from 0.15% for arsenical lead to 0.5%
for antimonial lead. These concentrations are approximately the same as
the arsenic content of the original lead concentrates. Bearing material
is kettle refined and should produce insignificant atmospheric emissions
because the melting points of lead (327°C) and babbitt materials (260-
270°C) are low compared to the vaporization temperature of arsenic (6l3°C)
(OTS, 1976). Antimonial lead is recovered in lead blast furnaces; thus,
the atmospheric arsenic emission factors of secondary smelting of this
type of lead may be similar to those for primary lead smelting. The
American Bureau of Metal Statistics (1978) estimates that, in 1977,
324,000 metric tQns of lead were recovered from old antimonial lead
scrap and old cable covering scrap. About 3,000 metric tons were re-
covered -t primary smelters and the remainder was recovered at secondary
smelters.
4.
Potential Human Exposures from Secondary Smelters
Humans are expected to be exposed to atmospheric arsenic from
secondary smelting of lead from battery materials. Arsenic exposures
from other nonferrous secondary metal smelting are estimated to be in-
significant by comparison with normal urban background concentrations.
Production data for secondary smelters of battery lead are not
available in the literature. It is therefore not possible to estimate
accurately the number of people exposed to arsenic from this source. It
is estimated that a smelter producing 6.5 kkg/year of lead from batteries
could produce annual average atmospheric concentrations of 0.003 ~g/m3
at distances of 3.5 km from the smelter and concentrations of 0.020 ~g/m3
near the smelter boundry. Most secondary lead smelters are located in
fairly densely populated areas (average population density of 1,100
people/km2). It is therefore likely that an average of several thousand
people could be exposed to arsenic emissions from each such secondary
smelter. These estimates are spectulative and require further data for
verification.
67

-------
VII
&~SENIC EXPOSURES FROM PESTICIDE MANUFACTURERS
A.
General
Arsenical pesticides are used mainly as insecticides and herbicides.
Other uses include fungicides, rodenticides"acaricides, and nematocides.
Table VII-l ~ists 25 of the main arsenical pesticides, and~heir uses~
, '
Bornstein (197-5) estimates that 7070 of the domestic .arsenic .con-
sumption (25.5 million .lb) is used an~ually for, pesticidemanu~acturing.
It is estimated that ;well over ,95% of this arsenic is used in 'the manu-
facture of six pesticides: lead arsenate, arsenic acid, cOpp'er arsenate,
MSMA, DSMA, and cacodylic acid. The remaining 1.5 million lb of arsenic
is used by the industry for the manufacture of other pesticides.' Table
VII-2gi~es an estimate-of arsenic used by major ,arsenical" pesticides.
(Because of s~e differences in production years, the total arsenic given
in Table VII-2. slightly exceeds the total estimated by Borns~ein.)
B.
Pesticide Manufacturers
Producers capabl~ of manufacturing of arsenical pesticides include
20 companies at 27 plants. These plants are listed in Table VII-3,
along with the ~rseriical pesticides that each produces~'This-list of
producers differs slightly from a similar list pr~pared by..'the EPA Office
of Toxic Substanc'es that includes 32 manufacturing plants. . However,
most of the production is attributable to only a few firms it:1~luded in
both lists. Data on the production of these pesticides by ~anufacturer
are difficult to obtain, because such information is proprietary. Most
of the arsenic acid is manufactured by the Pennwalt Corporation; which
was expect~d to have a production capacity of 12.5 million lb for 1977.
MSMA is produced primarily by two firms: The Ansul Corporation, with an
annual capacity of 25 million lb; and Diamond Shamrock, wit~ an annual
capacity of 9 million lb. Two basic manufacturers produce cacodylic
acid: The Ansul Corporation and Vineland Chemical (OPP, 1975A). In
1973, three companies produced DSMA: Ansul, W. A. Cleary, and Vineland
(OPP, 1975B). City Chemical is the major producer of copper arsenate.
L. A. Chemical and Woolfolk are the major producers of calcium arsenate,
and lead arsenate is primarily produced by Woolfolk, L. A. Chemical, and
Dimension Pigments.
C.
Pesticide Plant Emissions
Bornstein (1975) estimated atmospheric emissions for arsenical pesti-
cide manufacturing. These emissions are given in Table VII-4. They
indicated that the total arsenic emissions from these activities range
69

-------
Tab~e VlI-l
ARSENICAL COMPOUNDS AND THEIR USE
C om1)o\,U1d
o - Arsanilic acid (sodium arsanilate)
I - Arsenic acid (arsenic pent oxide; arsenic oxide)
I - Arsenic disulfide (tri and penta)
I - Arsenic iodide (arsenous iodide)
I - Arsenic penta fluoride
I - Arsenic thioarsenate
I - Arsenic tribromide (arsenous bromide)
I - Arsenic trichloride (arsenous chloride, arsenic
butter)
I -Arsenic trifluoride (arsenous fluoride)
I - Arsenic trioxide (arsenous oxide, arsenous .acid)
I - Arsine (arsenous hydride)
o - Cacodylic acid (dimethylarsinic acid)
I - Calcium arsenate
I - Calcium arse~ite
I - Copper arsenate
I - Copper arsenite

o - Fluor chrome arsenate ~henol

I - Lead arseaate

o - Methanearsonic acid
(MSMA and DSMA)
(mono- and disodium salts)
o - Heth~nearsonic acid (calcium salt)
o - Hethanearsonic acid (ammonium salts)
o - Par-s green (copper acetoarsenite)
I - Sodium arsenate
I - Sodium arsenite (arsenous acid, sodium salt)
I - Zinc arsenate
Notes:
o - Organic
I - Inorganic
Source:
Bornstein (1975)
70
Use
Feed additive
Cotton defoliant
Textile printing, tanners
paint pigment, medicinals

Antiseptic

Laboratory research
Scavenger
Medicinals
Herbicide
Laboratory research

Precursor in production of
other arsenical compounds
Semiconductor ~ndustry
Herbicide
Insecticide
Insecticide
Insecticide - wood preservative
Insecticide
Wood preservative
Insecticide
Herbicide
Herbicide
Herbicide
Insecticide
Wood preservative
Herbicide
Tree debaT. ....ng
Herbicide

Insecticide

-------
Table VII-2
ANNUAL USE OF ARSENIC IN MAJOR ARSENICAL PESTICIDES
Product
Arsenic trifluoride
1974 Production
(lb active ingredient)

<100
Arsenic in Product
(lb)
Arsenic acid
<10,000

1,500,000a.

<1 ,000

11, 700, OOOa

3,00Q,000

36,000,000

9,000,000
c
6,000,000
. <9, 500

800,000a

<400
'. . a
2,900,000

500 , 000

16,700,000

3,600,000
. . c
3,200,000
Arsine
Cacodylic add

Calcium arsenate
b
Copper arsenate
Lead arsenate
MSMA
DSMA
Arsenic penta fluoride
Total
10
27,700,000
a1973 production.
b
Chromated copper

c197l production.
larger.
arsenate.
Current arsenic acid production is probably much
Source:
Modified from Bornstein (1975), OPP (1975A), and OPP (1975B)
between 4 and 151 ton/yr, depending on assumptions about the degree of coa-
troIs used by the industry. Many of the large producers of pesticides
have alre.:1dy installed control equipment so that emissions are in fact
much lower th~n the 151 ton/yr. These data indicate emission factors of
0.63 lb/ton for controlled manufacturers and 24 1b/ton for uncontrolled
manufacturers. W. E. Davis and Associates (1971) estimated atmospheric
arsenic emissions at 20 lb/ton of arsenic processed for pesticide manu-
facturing.
The most common means of controlling emissions from the pesticide
manufacturing industry are baghouses employing cotton sateen bags. In
some applications, water scrubbers are used to control dust emissions.
Inertial separators such as cyclones and mechanical centrifugal separators
are not recommended because collection efficiencies are too low for
smaller particles. In the manufacture of liquid pesticides, air pollu-
tion control problems usually entail collection of dust in a wet air-
stream. For these cases, wet scrubbers are employed (Bornstein, 1975).
71

-------
Manufacturer
Abbott Labs
North Chicago, IL

Ansul Chemical
Marinett, WI

Arico Inc.
Sant"a Clara, CA

Blue Spruce Co.
Bound Brook, NJ
City Chemical
Jersey City, NJ

Diamond Shamrock
Green Bayou, TX

D~ensional Pigments
Bayonne,.NJ

Fleming Labs Inc.
Charlotte, NC
G. D. Searle Co.
Cucamonga, CA
East Rutherford, NJ
Gloucester, .MA
Joliet, IL
La Porte~ TX
Morrow, GA
Newark, CA

Los Angeles Chemical
South Gate; CA
Osmoro Wood Preserving
Memphis, TN

Pennwalt
Bryan, TX
Tulsa, OK
Tacoma, WA
Rohm and Haas
Myertown, PA

Vine land Chemical
Vineland, NJ

W. A. Cleary
Somerset, NJ
Woolfolk
Fort Valley, GA
Source:
SRI estimate
Table VII-3
MANUFACTURERS OF ARSENICAL PESTICIDES
Compound
Arsanilic acid
Carodylic acid, MSMA, and ~SMA
Arsine
Sodium arsenite

Arsenic iodide, copper arsenate, copper arsenite,
zinc arsenate
MSMA and DSMA
Lead Arsenate
Arsanilic acid, arsine
Arsine
Arsine
A rs ine
Arsine
Arsine
Arsine
Arsine

Arsenic acid, cacodylic acid, calcium arsenate, calcium
arsenite, lead arsenate, paris green, sodium arsenite
Arsenic acid
Arsenic acid
Arsenic pentafluoride, arsenic trifluoride
Sodium arsenite
Arsanilic acid
. Cacodylic acid, MSMA and DSMA,MSMA calcium
MSMA and DSMA .
Arsenic acid, calcium arsenate, lead arsenate,
sodium arsenite
72

-------
Table VII-4
ANNUAL ATMOSPHERIC ARSENIC EMISSIONS
FOR PESTICIDE MANUFACTURERS
Final product packaging equipment
Total
1
<4
Uncontrolled
Emissions
(ton/vr)

7

1

136

7

151
Operation
Raw material handling and equipment
Controlled
Emissions
(to~/yr)

01
Reactor
Product purification equipment
<1
1
Source:
Estimates based on Bornstein (1975)
A rou~h estimate was made of the amount of arsenic used in the manu-
facture of arsenical pesticides for the assumed nine largest,producers
(Table VII-5). Because no published data are available onO'the amount manu-
factured by producers, these estimates were based on information relating
to which company produces the material, production capabilities, and other
qualitative information relating to producer size. Although an effort was
made to estimate capacities of actual manufacturers, these estimates must
be regarded as hypothetical until better data are available. Atmospheric
arsenic emissions were estimated by using controlled and uncontrolled
emission factors of 0.63 and 24 Ib/ton.
The total annual arsenical pesticide production assigned to the nine
plants listed in Table VII-5 accounts for most of total annual production
for all plants in the United States. As the analysis later indicates,
environmental emissions for producers smaller than these listed in Table
VII-5, are estimated not to add significantly to the ambient background
arsenic concentrations. For these reasons the information listed in
Table VII-5 should be sufficient for parametrically assessing whether
emissions from arsenical pesticide producers constitute a health hazard.
Evaluation of exposures for specific producers will require more detailed
information.
D.
Population Exposures
Because atmospheric arsenic concentration data were unavailable for
the vicinity of the pesticide manufacturing plants, it was necessary to
estimate atmospheric concentrations by using atmospheric dispersion
modeling. Youngblood (1978) used dispersion modeling to estimate the
atmospheric arsenic concentrations for two manufacturers with emissions
of 2 X 10-5 lb/yr and 207 lb/yr of arsenic. The lower emission rate
73

-------
Table VII-5
. ESTIMATED ARSENIC USE AND ATMOSPHERIC EMISSIONS FOR MAJOR
HYPOTHETICAL ARSENICAL PESTICIDE MANUFACTURERS*
    Arsenic Emissions
   Arsenic Processed  (lb/yr)
Manufacturer  (lb/yr) Controlled Uncontrolled
Plant A   3,200,000 1,010 38,400
Byr~n, TX    
Plant B   14,200,000 4,470 170,000
Marinette, WI   
Plant C   4,200,000 1,320 50,000
Green Bayou, TX   
Plant D   1.400.000 440 16,800
Vineland, NJ    
Plant E   1,200,000 370 14,400
Somerset, NJ    
Plant F   2,900,000 910 34.800
Jersey City, NJ   
Plant G   167,000 50 2,000
Fort Valley, GA   
PlantH   167,000 50 2,000
South Gate, CA   
Plan t I   167,000 50 2,000
Bayonne, NJ    
*
These are crude estimates, subject to refinement as better data become
avail~ble .
resulted in negligible atmospheric arsenic concentrations for all dis-
tances from the plant. The results for the plant with emissions of 207
Ib/yr of ~rsenic are given in Table VII-6. The concentrations given in
Table VII-6 were scaled proportionately to estimate concentrations for
manufacturers having the emission rates listed in Table VII-5. This
procedure indicates that pesticide manufacturers whose emissions are
well controlled and whose production rates are less than those for the
manufacturers listed in Table VII-5 would produce maximum annual average
74

-------
Table VII-6
ESTIMATED ARSENIC CONCENTRATIONS FOR VARIOUS DISTANCES
FOR A MANUFACTURER HAVING EMISSIONS OF 2(7 LB/YR ARSENIC
Distance from 24-hr Maximum Downwind Annual Omidirectional
Plan t (km) Concentration (~g/m3) Average (ng/m3)
0.30 0.0110 0.880
0.45 0.0130 1.040
0.60 0.0120 0.96.0
0.75 0.0100 0.800
1.00 0.0076 0.608
1. 25 0.0058 0.464
1. 60 0.0042 0..336
2.50 0.0023 0.184
4.00 0.0012 0.096
6.00 0.0007 0.056
Source:
Modified from Youngblood (1978)
environmental con~entrations of less than 0.001 ~g/m3. Because these
concentrations arE b~low average urban background concentrations, th~
exclusion from this analysis of the smaller pesticide manufacturers is
justified.
A series of 14 concentric geographic rings was drawn around each of
the 9 major manufacturing plants listed in Table VII-5. The radii of
the rings ranged from 0.4 to 20 km. It was assumed that no residential
population .occurred within 0.4 km of each plant. Residential population
was estimated for each of the geographic rings by using the 1970 county
population density. The average annual atmospheric arsenic concentra-
tion was estimated for the midpoint of each geographic ring through use
of the dispersion modeling technique previously described. Two bounding
cases were evaluated: (1) a1l plants with well controlled emissions and
(2) all plants with uncontrolled emissions. Exposures were not estimated
for concentrations less than 0.003 ~g/m3. The atmospheric concentrations
75
~

-------
and estimated exposed populations are given i~ Table VII-7 for the con-
trolled and uncontrolled cases.
Table VII-7
ESTIMATED POPULATION EXPOSED TO ATMOSPHERIC ARSENIC
EMITTED FROM PESTICIDE MANUFACTURING
Number of People Exposed
Average Arsenic
Concentration*
(~g/m3)
Assuming Emissions
Well Controlled
Assuming Emissions
Uncontrolled
0.600-0.999
0.300-0.599
0.100-0.299
0.060-0.099
0.030-0.059

0.010-0.029
0.005-0.009

0.003-0.004
60
800
11,900
20
35
20,800
13 ,800
119,000
296,000
413,000
74,000
*
Annual omnidirectional average.
Assuming that the emissions from all major arsenical pesticide
manufacturing plants are well controlled results in an estimated 12,760
people exposed to average arsenic concentrations that range from 0.003
to 0.020 ~g/m3. Assuming that the emissions from all major arsenical
pesticide manufacturing plants are uncontrolled results in an estimated
936,000 people exposed to average arsenic concentrations from 0.003 to
0.80 ~g/m3. Actual exposures should fall somewhere between the uncon-
trolled and well controlled bounds. Bornstein (1975) has indicated that
many of the large pesticide manufacturers have already installed control
equipment. The estimated population exposures given here are for the
larger manufacturers; hence, the actual exposures are probably close to
the estimates for well controlled plants.
E.
Summary
It is estimated that approximately 13,000 people are exposed to
arsenic concentrations that range from 0.003 to 0.020 ~g/m3. Because
76
I

-------
the general urban population exposu~es aver1~e about 0.003 ~g/m3, it
appears that emissions from arsenica: pest~ .ide manufacturers do not
present a significant increased environmental health risk. A possible
exception would occur if one of the larger manufacturers (processing
more than 200,000 lb/yr arsenic) has poor emission controls.
77

-------
VIII
ARSENIC EXPOSURES FROM COTTON GINS'
, '
A.
General
Arsenic acid is used as a desiccant to aid in cotton harvesting.
During ginning, arsenic from the cotton is released to the atmosphere
and thus constitutes potential occupational and environmental exposures.
Cotton production in Texas and Oklahoma is of two basic types,
distinguished by the cotton planted and the manner of harvesting. In
the irrigated regions of Texas and Oklahoma, farmers successfully grow
high quality cotton that can be ha~vested with a spindle picker. Chemi-
cal defoliants, such as chlorate and phosphorus materials, are frequently
used to remove the to~ leaves in preparation for spindle harvesting
(USEPA, 1976). '
In the drylan~ areas of these two states, the High Plains, Rolling
Plains, and Blae: Prairie areas, the spindle-picked cotton varieties
cannot be grown because of lack of moisture. Production in the dryland
areas de rends on harvesting only once a year with ~ stripper harvesting
machin~. Stripper harvesting consists of taking up the cotton boll and
burr and frequently the leaves and stems, in a rake-like action. This
harvesting requires that the moisture of the cotton be very low, lest
the cotton be stained thereby bringing a lower price. To ensure low
'moisture content, the cotton plants are dried out before stripping
(USEPA, 1976). '
The cotton matures late in the season in much of the dryland area
of Oklahoma a~d the High Plains and Rolling Plains of Texas. In these
areas the growers can frequently wait for the first frosts to dry the
cotton plants bef~~e harvest. Because cotton degrades in quality if
harvest is delayea after it reaches maturity, late frosts necessitate
the use of ~ chemical desiccant. Furthermore, in areas south of Dallas
where the cotton matures in late August or September, a chemical desiccant
must be used practically every year. In regard to both cost and effec-
tiveness, arsenic acid is the preferred chemical desiccant (USEPA, 1976).
During 1973, approximately 85% of the Texas cotton was machine stripped
~lnd 15% was spindle picked (USEPA, 1976).
Part of the desiccant adheres to the parts of the cotton plant that
are t.:1ken.'during harvesting. Because this' desiccant can then be released
to the atmosphere during ginning, it presents a potential health hazard
to cotton gin employees and to the surrounding residential population.
79

-------
B.
Cotton Gins
Table VIII-l estimates the cotton gins active in the United States,
during 1972 and their employees; 3517 gins were active, with an esti-
mated 35,000 workers. Of " these gins, 1040 are in Texas and 118 in
Oklahoma.
Cotton ginning is not a year-around operation. Bureau of the Census
Bulletin 202 (1965) states that ginning of each crop begins in mid-July,
continues through the autumn and winter, and is substantially completed
by the following February. In most locations the ginning season lasts
2-4 months. Many of the gins are operated 24 hours a day during ginning
season.
C .
Cotton Gin Emissions
Durrenberger (1974 and 1975) monitored particulates around 16 Texas
cotton gins that processed machine-stripped cotton. The purposes of his
study were twofold: (1) to determine emission levels of suspended par-
ticulates, arsenic, and cellulose and (2) to develop a method of pre-
dicting cotton gin emissions as a function of control equipment. Three
classes of control equipment were defined: uncontrolled, controlled,
and well controlled.
Atmospheric particulates were monitored around each gin.' Durrenberger
found that, if particulate concentrations were expressed as the ratio of
atmosph~ric concentration to the amount of trash processed per hour, gins
processing similar cotton and with similar controls could be grouped to
give average atmospheric particulate concentration as a function of dis-
tance from the gin. These functions,were derived for distances as far
as 300 m from the gin. To obtain atmospheric arsenic concentrations it
is necessary to multiply the particulate concentrations by their arsenic
content. Durrenberger (1974) found that the arsenic content of trash
from arsenic desiccated cotton was 0.2%. This is in agreement with the
concentrations found by Aboul-Ela and Miller (1965). Hence, the atmo-
spheric arsenic concentrations can be estimated by multiplying the
particulate concentrations by 0.002.
Becau~e the validity ,of Durrenberger-'s methods does not extend
beyond 300 m from the gins, a method to estimate concentrations for
greater distances is required. Youngblood (1978) mathematically modeled
the expected concentrations for distances of 300 m to 20 km from two
types of gins: well controlled and poorly controlled. The poorly con-
trolled gins were assumed to have particulate emissions of 490 lb/hr for
a 10 bale/hr gin; the well-controlled gins were assumed to have particu-
late emissions of 26 lb/hr for a 10 bale/hr gin as noted by Herring
(1973). These modeling data were for 24-hr downwind concentrations,
which were converted to annual downwind concentrations by dividing by
5.0; the annual downwind were converted to average omnidirectional con-
centrations by further dividing by 2.5. Youngblood's estimates were
80

-------
Table VIII-l
COTTON GINS IN THE tmITED STATES, 1972
  Number of   
Location Cotton Gins Estimated Workers *
Alabama  226 1,582 
Arizona  ll5 1,725 
Arkansas  397 4,367 
California 234 3,510 
Florida  4  28 
Georgia  153 1,071 
Illinois     
Kentucky  1  II 
Louisiana. 148 1,628 
r-'ississippi 458 4,122 
Missouri  ll4 1,254 
t-4evada  1  15 
New Nexico 53  742 
North Carolina 91  637 
Oklahoma  ll8  826 
South Carolina 165 1,155 
Tennessee  196 1,372 
Texa s  1040 II ,440 
virginia  --1  21 
United States 3517 35,480 
*
Estimates based on number of workers by state
given in Wellford (1963).
Source:
Active gins from Bureau of the Census (1972)
further modified by assuming that the arsenic concentration of the pa~-
ticulates is 2000 'ppm, rather than the 300 ppm he used. The ,results of
the m~dified modeling procedure .and the results of Durrenberger's method,
which is based on ambient data, are compared on Table VIII-2. The re-
sults for the two methods are almost identical for uncontrolled gins
and within a factor of 3 for controlled gins.
81

-------
Table VIII-2
COMPARISON OF ATMOSPHERIC ARSENIC CONCENTRATIONS
NEAR COTTON GINS BASED ON MODELED AND
MEASURED METHODS* (~g/m3)
Environmental
Controls

Uncontrolled
Concentration
Estimation
Method
Measured
Cotton Production
5 bale/day 20 bale/day

0.250 1.000

0.256 1.024
Modeled
Well-controlled
Measured
0.037
0.014
0.148
0.054
Modeled
*
Modeled results are based on Youngblood (1978), and
measured results are based on Durrenberger (1974); both
are for concentrations 300 m from the gins.
Production data for cotton gins were analyzed to determine the fre-
quency distribution of gins as a function of bales of cotton produced
per hour. This analysis is summarized in Table VIII-3 and indicates that
four representative production rates should be used: 5, 7, 14. and 20
bale/hr. In addition, the frequency distribution of types of emission
controls ~t gins was evaluated (Table VIII-4). Based on this evaluation,
it was assumed that 50% of the gins are well controlled, 20% are uncon-
trolled, and that 30% have some type of controls. For these estimates
we have used the definition of "controlled" and "well-controlled" given
by Durrenberger (1974): Controlled means controlled with cyclones and
bur hoppers. Well controlled means having controls in addition to cyclones
and bur hoppers.
To determine the number of gins processing desiccated cotton at the
various production rates and with different controls. it was assumed:
.
That there are 1040 gins in Texas and 118 gins in Oklahoma.

That 85% of the gins in Texas and Oklahoma process desiccated
cotton.
.
.
The type of control at a gin is independent of the production
rate.
The estimated number of gins in each production rate-control class is
shown in Table VIII-5.
82

-------
Table VIII-3
ASSUMED DISTRIBUTION OF PRODUCTION
RATES FOR COTTON GINS*
Production Rate (bales/hr) Percent of
Avera e Ran e Gins
5 4-7 50
7 8-11 30
14 lZ-15 10
20 >15 10
* .
Based on an analysis of data for 17 cotton gins.
Table VIII-4
DISTRIBUTION OF EMISSION CONTROLS
FOR COTTON GINS*
Control
Percent of
Gins
High-efficiency cyclone
Inline filter
50
4

28
Other controls
No controls
18
*
Based In a 1972 EPA survey.
Table VIII-5
ESTIMATED NUMBER OF GINS PROCESSING MACHINE STRIPPED
COTTON IN TEXAS AND OKLAHOMA
Production
(ba Ie s Ihr)

5
Uncontrolled
Emission Controls
Controlled Well Controlled
7

14
98 148 246
55 89 148
20 30 50
20 30 50
83 
20

-------
Atmospheric arsenic concentrations as a function of distance from
the gin were. calculated for well controlled, controlled, and uncontrolled
gins and for production Tates of 5, 7, 14, and 20 bales of cotton per
day for each class of control. These are given in Table VIII-6. The
methods given by Durrenberger, were used to estimate the concentrations
for distances as far as 300 m and Youngblood's methods modified as pre-
viously described, were used for distances greater than ~OO m. These
concentrations were nut extended to distances beyond which the concentra-
tion would be less than 0.003 ~g/m3.
Data on particle sizes of particulate emissions--composed of plant
foliage, lint, motes, and dirt--from cotton gins are not well documented.
Particulate size analysis of gin trash data extracted from the USDA by
Herring (1973) indicate that 99% of the particulates was 25 ~ or larger;
and EPA tests indicate that particulates in 'controlled emissions are
preponderantly composed of particle sizes greater than 3~. It is ex-
pected that the larger particles settle out near the cotton gin. A 1965
study by the Texas State Department of Health shows that at a distance
of 200 ft downwind from the gin, 50% of the particles are larger. than
5 ~m; whereas, at 1000 ft, this fraction drops to 20%.
D.
Population Exposures
This study did not attempt to locate and count the resident popula-
tion within the vicinity of all gins processing arsenic desiccated cotton
as these data are not readily-available. It is known that most of the
gins are located in rural areas. Maps given by Durrenberger (1974),
the Texas Department of Health (1965), and others indicate that residen-
tial housing exi,sts within 300 m of some of the gins. In this exposure
study, it is assumed that residential populations are uniformly distributed
around the gins with a density of 40 people per square mile. This density
is consistent with the average statewide population densiti~s of Texas
and Oklahoma. It is further assumed that there are no residences within
50 m of any of the gins.
The resident population was estimated for 19 concentric rings about
each gin, for radii up to 20 km. These populations were multiplied by
the number of gins in each of the nine production rate-emission control
groups given in Table VIII-5. It is assumed that none of the gins are
colocated. Arsenic concentrations from Table VIII-6 were assigned to
the estimated exposed populations. These were then used to derive the'
exposure estimates shown in Tables VIII-7 and VIII-8. Table VIII-7 gives
average population exposures during the ginning season. Table VIII-8
gives annual average population estimates, based on the assumption that
the gin~ing season lasts approximately 8 weeks per year.
84

-------
       Table VIII-6        
    ESTIMATED ATMOSPHERIC ARSENIC CONCENTRATIONS 11 3   
     (IJ g/m )   
      FOR VARIOUS DISTANCES FROM GINS       
     PRqCESSING ARSENIC DESICCATED COTTON     
 Distance from Uncontrolled   Controlled     Well Controlled 
 Gin (meters)  Emisstons    Emissions     Emissions 
   st 7t 14t 20t st 7t 14t   20t st 7t 14t 20t
  SO 7.249 10.148 20.297 28.995 4.789 6.704 13.409 19.155 1.804 2.526 5.051 7.216
  75 5.498 7.697 15.395 21. 992 3.600 5.040 10.079 14.399 1. 312 1.837 3.674 5.248
  100 4.211 5.895 11. 790 16.843 2.550 3.570 7.141 10.201 0.861 1. 205 2.411 3.444
  150 1. 898 2.658 5.315 7.593 1. 349 1. 888 3.777 5.396 0.312 0.436 0.872 1.246
  200 0.951 1.332 2.663 3.805 0.763 1.068 2.135 3.050 0.152 0.212 0.425 0.607
  250 0.500 0.700. 1. 401 2.001 0.349 0.488 0.976 1. 394 0.062 0.086 0.172 0.246
0)  300 0.250 0.350 0.700 1.000 0.201 0.281 0.563 0.804 0.037 0.052 0.103 0.148
UI 
  450 0.147 0.205. 0.411 0.587 0.076 0.107 0.213 0.304 0.008 0.011 0.021 0.031
  600 0.096 0.134 0.269 0.384 0.050 0.069 0.139 0.198 0.005 0.007 0.014 0.020
  750 0.067 0.093 0.187 0.267 0.033 0.046 0.093 0.132 0.003 0.005 0.009 0.013
 1000 0.043 0.060 0.119 0.171 0.023 0.032 0.065 0.093  0.003 0.007 0.009
 1250 0.032 0.045 0.090 0.128 0.017 0.023 0.046 0.066   0.005 0.007
 1600 0.021 0.029 0.058 0.083 0.010 0.014 0.028 0.040   0.003 0.004
 2500 0.010 0.015' 0.029 0.042 0.007 0.009 0.019 0.026    0.003
 4000 0.005 0.007 0.014 0.020 0.003 0.004 0.007 0.011    0.001
 6000 0.003 0.004 0.008 0.012   0.005 0.007    
 9000   0.004 0.006    0.003    
 14000    0.003          
 20000              
 11 Bast'!d on annual average metero1ogical conditions.         
 t               
  Bales per hour production.            
 - Indicates a concentration less  . 3'         
  than 0.003 \Jg/m .         

-------
Table VIII-7
ESTIMATED HUMAN POPULATION EXPOSURES TO ARSENIC
EMITTED FROM COTTON GINS DURING GINNING SEASON
Assumed  50% well controlled,  
Concentration * 30% controlled, 100% well 100%
(IJ g/m3)  20% uncontrolled Controlled Uncontrolled
20-24  5  10
15-19  15  20
10-14  40  30
6-9  60 10 70
3-5  310 80 540
1-2  1,050 430 2,160
0.6-0.9  620 300 860
0.3-0.6  1,700 670 3,800
0.1-0.2  8,800 2,040 25,000
0.06-0.09  11.500 990 32,000
0.03-0.05  30,000 4,500 82,000
0.01-0.02  208,000 5.100 645,000
0.005-0.009  410,000 27,000 1,057.000
0.003-0.004  747,000 66,500 2.290.000
*
These are average exposures during the ginning season. Annual averages
would be approximately 10% to 25% of these values and are estimated in
Table VIII-8.
86

-------
Table VIII-8
*
ESTIMATE~ ANNUAL AVERAGE HUMAN POPULATION
EXPOSURES TO ARSENIC FROM COTTON GINS.
Annual Average
Concentration*
(~g/m3) .
No. of
People
+
Exposed'
3.0-5.9
1.0-2.9
0.60-0.99
0.30-0.69
0.10-0.29
0.06-0.99
0.030-0.059
0.010-0.029
0.005-0.009
0.003-0.004
5

100

200

700

2,000

3,000
5,900
20,000

56,000
135,000
*
These annual average concentrations are based
on the assumption that the ginning season
occurs during 15% of the year.
+ '
Exposures are based on the assumption that 50%
of the gins are well controlled, 30% are moder-
ately controlled, and 20% are uncontrolled.
E.
Summary
It is estimated that approximate'1y 1.4 million people are exposed to
atmospheric arsenic concentrations of 0.003 to 29 ~g/m3 during the ginning
season. Annual average population exposur~s are ~~ch smaller because the
ginning season occurs for a few weeks out of a year. Based on annual
average exposures, it is estimated that approximately 223,000 people are
exposed to atmospheric arsenic concentrations of 0.003 to 6 ~/m3. These
exposure estimates ,aSS\Ul1e that aU. Texas ,and o.klahoma tI,lachine-stripped
cotton is desiccated with arsenic acid. During some years; however, it
is not necessary to desiccate all machine-stripped cotton. For those
years, the number of exposed people on Tables VIII-7 and -8 should be
reduced proportionately.
87

-------
IX
ARSENIC EXPOSURES FROM GLASS Mk~UFACTURERS
A.
General
Arsenic is added to the glass batch during manufacturing for three
purposes: (1) to assist in freeing the glass from small bubbles or
"seeds," (2) to ~iminish the perceived color of the glass by oxidizing
Fe (II) (green) to Fe (III). (light orange brown), and (3) to stabilize
the selenium added to the glass batch for color balancing (Maasland,
1975) .
The use of arsenic in the manufacture of glass has been decreasing
in recent years, particularly because of the increased use of cerium
oxide as a decolorizing agent. Cerium oxide cannot be used if arsenic
trioxide is present in quantities greater than 8 oz/ton of sand because
solarization will very rapidly occur (Shult et al., 1970).
The ingredients used in glass manufacturing are batch weighed and
mixed before they are charged into the furnace. In the furnace, the
mixture of materials is held in a molten state at about 2800°F until it
acquires the homogeneous character of glass. It is then gradually cooled
in other sections of the furnace to about 2200°F to make it viscous
enough to form. In a matter of seconds, while at a yellow hot tempera-
ture, the glass is drawn -from the furnace and worked on forming machines
by a variety of methods, including pressing, blowing in molds, drawing,
rolling, and casting (Research Triangle Institute, 1972).
Most U.S. furnaces use the regenerative system for heat recovery.
Regener~tive firing systems consist of dual chambers filled with brick
checkerwork. While the products of combustion from the melter pass
through and heat one chamber, combustion air is preheated in the opposite
chamber. The functions of each chamber are interchanged by reversing
the flow of air and combustion products. Reversals occur every 15 to 20
minutes as required for maximum conservation of heat (Research Triangle
Institute, 1972). The use of regenerative furnaces for the production
of glass is slowly declining because of the increased use of recuperative
and electri~ furnaces.
Apparently all of the arsenic mixed in the batch does not remain in
the finished glass. Ideally, the arsenic dissolves in the melt and is
the last volatil~ compound to distill out. The arsenic ~apor condenses
to a submicron fume as it cools. Some vapor may condense"'on the brick
checke~ork in the chamber, and some fumes may be caught by tmpingement.
If a baghouse is in use, still more may be caught (Maasland, 1975). Owens
Corning data suggests that 70% to 99% of the .arsenic remains in the glass
(Mosely, 1978).
89

-------
B.
Glass Manufacturers
The glass industry is divided into three four-digit SIC categories:
SIC 3211
SIC 3221
SIC 3229
Flat Glass
Glass Containers
Pressed and Blown Glass.
Flat glass manufacturers produce ~heet (window) glass, plate glass,
laminated glass, and other flat glass. Flai glass production is approxi-
mately 3.2 million ton/yr of finished products. Total production is
approximately 4.4 million tons, the difference being that off-quality
glass is recycled. During 1975, arsenic was estimated to be used in
less than 10% of the flat glass made (Reznik, 1975). The use of arsenic
in the mdnufacture of flat glass has continued to decline. As a conse-
quence, arsenic exposures from flat glass manufacturing have been ex-
cluded from this study.
Glass container manufacturers produce glass containers for commer-
cial packing and jottling, and for home canning. Glass container pro-
duction is approximately 13 million tons annually, with an additional
10% to 20% recycled. Reznick (1975) estimated that in 1975 less than
15% of this industry used arsenic. As with flat glass manufacturing,
the use of arsenic for container manufacturing has continued to decline.
As a consequence, arsenic exposures from container glass manufacturing
have been excluded from this study.
Pr~ssed and hlown glass manufacturers produce glasses and glassware
by pr~ssing and blowing or shaping glass produced in the same establish-
ment. Textile glass manufacturing is included in this industry, but
glass wool insulation is not. Production of pressed vehicular lighting,
beaco~s, and lanterns is also included. Other manufactured items include
table, kitchen, art, and novelty glassware; tumblers; stemware; lighting
~nd electronic envelopes and blankets; and electric light bulb blanks.
Reznik (1975) estimated production at 4.5 million tons annually. Ex-
cluding textile fibers, Schorr et ale (1977) estimated finished products
at -1.6 million tons for 1974. Total production is -"proximately 3.3
~illion tons, with about 52% of production ~sed ~~ recycled cullet or
ending ilS w
-------
Research Triangle Institute (1972) listed glass manufacturers by
SIC and state. They attempted to include complete mail~~g addresses,
sales, and a percentage share of the industrj~ ~ut .not all of the infor-
mation was available for every plant. They li~t 219 plants for SIC 3219.
Schorr et al. (1977) provide a ~ore recent listin6 of manufacturing plant
addresses for SIC 3219 excluding the glass textile manufacturers. Their
list of 176 plants includes names and locations; however, no information
is given about manufacturing volume. . Annual glass productio~ rates for
the 176 plants listed by Schorr et al. (1977) were estimated by using a
co~bination. of (1) the information presented by Research Triangle
Institute (1972), (2) production sales, employment,. and-location data
given by Standard and Poor's Register (1977) and Dun &-Bradstreet (1977),
and (3) the statistical distribution of number of firms vs. value of
shipments as reported by the Bureau of the Census (1975). . The estimated
distribution of plants by production volume is given in Table IX-l.
Table IX-l.
ESTIMATED PRODUCTION OF PRESSED AND
BLOWN GLASS BY NUMBER OF PLANTS*
Numb~r of  4-
Plant Production (Annual) I
Plants -103" kg Tons
71 145 160
17 500 550
11 890 990
9 2,080 2,290
11 4,950 5,450
22 11,280 12,430
12 37,760 41,620
16 76,230 84,030
7 139,000 153,000
*
Excludes textile fiber manufacturers.
-+-
This is the total annual p~oduction, of
~lich approximately 48% is converted to
finished products.
No data were located to indicate which of the plants still use
3rse~ic in their manufacturing process because such data could only be
obtai~ed through a survey of the manufacturing plants, an activity be-
yond the scope of this study. Potential population exposures to arseni=
from 31ass manufacturing will be explored through a series of paracetri=
analyses in which assumptions concerning arsenic use are varied.
91

-------
C.
Glass Manufacturin~ Emissions
Potential significant sources of atmospheric emissions are the raw
materials handling operations, the glass furnace, and the forming opera-
tions. Of these, the glass furnace is usua~ly the major source (RTI,
1972). The rate of emissions from the glass melting furnace usually
varies considerably, depending on the composition of the glass produced,
the design and operating characteristics of the furnace,. and the emission
controls installed. The arsenic volatizes during glass melting and is
thought to condense on the particulates and thus be released to the at-
mosphere on the particulates. As with coal emissions, the arsenic con-
centration on smaller particulates probably exceeds that on larger par-
ticulates. RTI (1912) states that the particulate emission rates for
container glass furnaces is usually between 0.8 and 3.0 g/kg (1.5 and
6.0 lb/ton). Table IX-2 shows particulate and arsenic emission rates
for the manufacture of lead and opal glass reported by Reznik (1975) .
These particulate emission rates varied between 2.1 and 4.3 g/kg (4.1
and 8.6 lb/ton). The arsenic concentration of the particulates ranged
from 6'7.. to 9% and averaged 7%. Particulate emissions for press~d and
blown glass manufacturing have been analyzed by Schorr et a1. (1977).
They found particulate emission factors for 19 soda/lime glass manu-
facturers to vary from 0.49 to 12.57 g/kg (0.97 to 25.14 lb/ton). The
average emission factor was 5.22 g/kg (10.44 lb/ton). The particulate
emission factor for a lead glass manufacturer was 4.52 g/kg (9.04 lb/ton).
Another manufacturer reported uncontrolled emissions for lead glasses to
be about 15 g/kg (30 lb/ton) and about 25 g/kg (50 Ib/ton) for borosili-
cate glasses (Schorr et al., 1977).
Table IX-2
STACK EMISSIONS FOR GLASS MANUFACTURING
  Feed Rate Particulates Arsenic % As in
Product Case (lb /hr) (lb/ton) (lb/ton) Particulate
2/:' lead glass 175 4.1 0.27 6.6 
Opal glass 85 6.8 0.40 5.9 
Opal glass 100 8.6 0.74 8.6 
*
Emissions are based on feed rates rather than glass manufactured.

There is ~ 10%-15% loss of feed due to volatilization. .
Source:
Reznik (1975).
92

-------
On the basis of a worst case en6ineering analysis, Schorr et al.
(1977) estimated the highest particul~te emi~sion factor for borosilicate
glass to be 25 g/kg (50 lb/ton), for o?al glass to be 5 g/kg (10 lb/ton),
and for lead glass to be 15 g/kg (30 lb/ton). The accuracy of these
estimated emissions was reported as :tlOO%. Schorr et a1. (1977) then
used a weighted average particulate emission factor of 8.7 g/kg (17.4
lb/ton) for all pressed and blow glass. This results in a worse case
arsenic emission factor of 0.6 g/kg (1.2 lb/ton) assuming that the
particulates contain 7% arsenic.
Cuffe (1978) analyzed recent particulate emission data on controlled
glass furnaces. Particulate emissions for 10 furnaces equipped with elec-
trostatic precipitators (ESPs) ranged from 0.03 to 4.79 g/kg (0.06-9.58
lb/ton) and averaged 0.16 g/kg (0.31 lb/ton). No data were available for
particulate emis~ions for all-electric melting furnaces or for furnaces
equipped with baghouse controls. Preliminary test results, however,
sh:Jwed that arsenic is removed by ESPs and baghouses in the same propor-
tion ~s total particulates are removed. Particulat€ emissions from one
furnace using electric boosting were 1 g/kg (2 lb!ton), but the amount
of arsenic used in this process is very small. .
The arse~ic emission factor for well controlled pressed and blown
glass manufac:urers who use arsenic in their process is estimated at
0.015 g/kg (0.03 lb/ton) (Cuffe, 1978). This estimate assumes that all
furnaces' are equipped with emission controls equivalent to ESPs and that
the particulates contain 7% arsenic. It also includes an allowance for
possible emissions of arsenic as vapor. This estimated emission factor
is significantly lower than the one presented in the Davis report (Davis,
1971), which gives a particulate emission factor of 1.0 g/kg (2.0 lb/ton)
of glass and estimates the arsenic trioxide concentration of the particu-
lates to be 7.7io (6.7% arsenic). Maasland (1975) implies that the Davis
estimate is unreliable because in its development, arsenic was reported
in the particulate catch of only five melters, one of which was melting
amber glass at an unspecified rate.
Particle size distributions for particulates emitted from a flint
and from an amber glass furnace are shown on Figure IX-I. The geometric
median particle diameter of the flint glass particles was 0.13 ~, and
the geometric standard deviation was 1.5 ~m. Corresponding values for
the amber glass furnace effluent particles were 0.11 and 1.7 ~m.
Youngblood (1978) used atmospheric dispersion modeling to estimate
the 24-hr downwind arsenic concentrations for two sizes of glass manu-
facturers that had ars~nic emissions of 0.016 and 0.047 gis, respectively.
These 24-hr concentrations were converted to annual worst case by dividing
by 5. They were then converted to omnidirectional by dividing the result
by an additional 2.5. These estimated 24-hr downwind and annual concen-
trations are given in Table IX-3: Concentrations for manufacturers of
other sizes were estimated by proportionately scaling the values given
in Table IX-3 on the basis of the ratio of arsenic emissions.
93

-------
 I 0.2
 .iI 
 e 
 I 
 a: 
 III 
 I- 
 III 
 ~ 
 4( 0.1
 a
\0 III 
.p. ... 0;08
U
 t= 
 a: 
 4( 0.06
 CL.
0.6
0.5
0.4
0.3
0.04
0.03
0.0~05 0.1
2
10
PERCENT BY NUMBER EQUAL to OR LESS THAN INDICATEO SitE
FIGURE IX-1. PROBABILITY DISTRIBUTION OF PARTICLE SIZES PRESENT IN GlASS FURNACE EFFLUENTS

-------
Table IX-3
ATMOSPHERIC ARSENIC CONCENTRATIONS FOR GLASS MANUFACTURERS
BASED ON DISPERSION MODELING
D is tance    
from 0.016 g/s Emission 0.047 g/s Emission
Plant 24-hr Downwind Annual Average 24-hr Downwind Annual Average
(km) . (u.g/m) (u.g/m) (u.g/m) (u.g/m)
. 0.3 0.069 0.0056 0.120 0.0096
0.8 0.036 0.0029 0.046 0.0036
1.3 0.026 0.0020 0.038 0.0031
2.0 0.019 0.0016 0.029 0.0024
3.0 0.015 0.0011 0.024 0.0020
5.0 0.011 0.0009 0.020 0.0015
8.0 0.009 0.0007 0.015 0.0013
12.0 O. (',..:: 0.0006 0.013 0.0011
16.0 0.005 0.0004 0.012 0.0010
20.0 0.004 0.0003 0.009 0.0007
Source:
Modified from Youngblood (1978)
D.
Population Exposures
Population exposures were estimated for atmospheric arsenic emissions
from glass manufa~turing. These exposure estimates were made by using
the plant sizes su~arized in Table IX-l and by assuming an arsenic emis-
sion factor of 0.015 g/kg of glass produced for well controlled manu-
facturers and 0.6 g/kg for poorly controlled manufacturers. The esti-
mated arsenic emissions for each plant were used to proportionately scale
the concentrations for the dispersions modeling results given in Table
IX-3. Because it is not known which plants use arsenic or which plants
are controlled, several alternative situations are evaluated.
At-risk populations were estimated for a number of concentric geo-
graphic rings about each manufacturing plant by using the 1970 county
population density for each plant. It was assumed that no population
r~sides within O.~ km of a plant.
95

-------
The esti~ated population exposures are listed in Table IX-4. Be-
cause ~f the uncertainties concerning the use of arsenic, four cases are
c\aluated using different assumptions. These cases are:
{3)
tl)
A random use estimate. This case assumes that 15% of the
pressed and blown glass is manufactured with arsenic and -that
all plants are well controlled. It is further assumed that
all glass produced at certain plants is made with arsenic.
Fifteen percent of the plants were randomly selected from each
of the Categories in Table IX-l. For this case it is estimated
that approximately 19,000 people are exposed to concentrations
ranging from 0.003 to 0.009 ~g/m3.

A largest plant estimate. This case assumes that 15% of the
pressed and blown glass is manufactured with arsenic and that
all plants are well .controlled. It is further assumed that
this production is made by the largest manufacturing plants.
For this case, it is estimated that approximately 65,000
people are exposed to concentrations ranging from 0.003 to
0.009 ,...g/m3.
( 2)
( 4)
A uniform use estimate. This case assumes that 15% of the
pressed and blown glass is manufactured with arsenic. It is
assumed that all manufacturing plants are well controlled and
use arsenic for 15% of their production. For this case, it
is estimated that no exposures exceed 0.003 ~g/m3.

A mixed controls estimate. This case assumes that 15% of the
pressed and blown glass is manufactured with arsenic and that
YO% of the plants are well controlled and that 10% are poorly
controlled. It is further assumed that all glass produced at
certain plants is made with arsenic. For this case it is
estimated that approximately 1.2 million are exposed to con-
centrations ranging from 0.003 to 0.29 ~g/m3.
~o attempt was made in these exposure estimates to assess the effect
of several plants located in close proximity. In effect, it was assumed
that all plants are not collocated. This assumption has .little effect on
the estimated exposures because significant exposures were not found
more than a fcw kilometers away from even the largest plants for the
well controlled cases.
E.
Sununa ry
Population exposures to arsenic from glass manufacturing are esti-
mated to involve a few thousand people and fairly low conccntratio!1s when
it is assumed that all plants are well controlled. The mixed controls
case shows these estimates are very much dependent on the emission
factors that arc assumed.
96

-------
Table IX-4
ESTIMATED HUMAN POPULATION EXPOSURES TO ATMOSPHERIC
ARSENIC CONCENTRATIONS FROM GLASS MANUFACTURING
Arsenic
Concentrationa
(~g/m3)
15% of
Plants Use
Arsenicb
Largest
Plants Use
Arsenicc
People Exposed, Assuming
All
Plants Use
Arsenic 15%
of Timed
Mixed Controls
15% of
Plants Use
Arsenice
0.100-0.299
0.060-0.099
1,400
10 , 140
0.030-0.059
0.010-0.029
0.005-0.009
0.003-0.00.,
370
15,510
1, 770
63,060
75,180
212,040

363,870

583 , 360
a
Average an~ual concentration.

bAssumes that 15% of pressed and blown glass is manufactured with
arsenic, that only certain manufacturers use arsenic, and that the
size distribution of manufacturers who use arsenic is proportionate to
the size output given in Table IX-I. All pl~nts are assumed to be well
controlled.

cAssumes that 15% of pressed and blown glass is manufactured with
arsenic and that only the largest plants use arsenic. All plants are
assumed to be well controlled.
d
Assumes that 15% of pressed and brown glass is manufactured with
arsenic and that all manufacturers use arsenic 15% of the time. All
plants are assumed to be well controlled.

eAssumes that 15% of pressed and blown glass is manufactured with
arsenic, that only certain manufacturers use arsenic, and that the size
distribution of manufacturers who use arsenic is proportionate to the
size output. given in Table IX-I. It is assumed that 90% of the manu-
facturers are well =ontrolled and that 10% are poorly controlled.
97

-------
x
SECONDARY HUMAN EXPOSURES RESULTING FRO~1
ATMOSPHERIC ARSENIC EMISSIONS
A.
General
Atmosphe~ic arsenic emission's eventually are deposited on the earth
or on bodies of water. Plants may absorb arsenic, and they may in turn
be consumed by humans or by animals that are eaten by humans. Humans
may drink the contaminated waters or consume aquatic' organisms with in-
creased arsenic concentrations caused by residing in contaminated waters.
Humans are not only exposed by inhalation to atmospheric arsenic emis-
sions put may also be subjected to second~ry exposures by ingestion of
these emissions.
The primary purpose of this report was to evaluate human
exposures to selected atmospheric emissions. T~is discussion
exposure routes is intended to suggest the rela'tive magnitude
from various sour~es.
inhalation
of secondary
of exposures
B.
Exposures
To view these secondary exposures in their proper perspective, it
is necessary to point out that arsenic is ubiquitous in the environment.
Arsenic ranks twentieth among the elements in abundance in the earth's
crust, and arsenic at low concentrations is distributed. throughout the
natural world. The earth's crust contains arsenic at about 5 ppm, with
a range of 0.1 to 50 ppm. Untreated soils usually have arsenic levels
of 5 to 10 ppm, whereas soils to which arsenical pesticides have been
applied contain as much as 165 ppm arsenic. At that concentration,
arsenic may reduc~ growth rates and inhibit generation of many plant
species if it is p~esent in available form. The arsenic content of un-
contaminated water is low, averaging from 0.002 to 0.003 ppm in seawater
and about 0.0004 ppm in rivers (Union Carbide, 1976).
Arsenic concentrations in plants vary from less than 0.01 to about
5 ppm (drY-weight). Differences in arsenic content reflect differences
in plants and in environmental and edaphic factors in particular geo-
graphic regions. Plants growing in arsenic-contaminated soils generally
have higher residues than plants grown in normal soils. However, concen-
trations in some nontreated plants are as high or higher as those found
in plants treated with arsenic or grown in arsenic-contaminated soils.
Natural variations in plants, plant species, available soil arsenic, and
growing conditions are all partly responsible for these discrepancies.
There appears to be little chance that animals would be poisoned by con-
suming plants that contain arsenic residues from contaminated soils, be-
cause less than toxic concentrations cause plant injury (NAS, 1977).
99

-------
Arsenic is present in all living organisms. Marine fish may con-
tain up to 10 ppm; coelenterates, some mollusks, and crustaceans may
contain even higher arsenic "concentrations. Freshwater fish may contain"
up to 3 ppm, although most values are less than 1 ppm. Domestic animals
a?d man generally contain less than 0.3 ppm on a wet-weight basis (NAS,
1977).
The contribution
sources considered in
arsenic released into
(Table 111-1). These
chance to affect food
plants, and smelters.
of atmospheric fallout to soils of arsenic from the
this report seems irisignificant because 32% of the
the environment comes from pesticides and herbicides
products are applied to croplands and have more"
crops. than do emissions from power plants, glass
Ecological damage directly associated with the arsenic emitted from
nonferrous smelters other than the Tacoma copper smelter is poorly docu-
mented. Both acceptable and unacceptable levels of arsenic (based on
Food and Drug Administrations standards) have been found in foods and
forage grown near nonferrous smelters (Environmental Science and Engi-
neering, 1976). The Tacoma smelter has undoubtedly produced high past
atmospheric arsenic emissions. Soil arsenic concentrations of more than
300 ppm have been found near the smelter, with concentrations generally
decreasing to 10-30 ppm 5 miles away. This is primarily an urban area;
however, vegetation from the area near the smelter (within a l-mile
radius) has had arsenic ~oncentrations of up to 30 ppm, and vegetation
from the remainder of the city has had concentrations of 1-9 ppm (Ratsch,
1974).
The H~lena Valley Study (EPA, 1972) points out that th~ average
arsenic content of soil is normally about 5 ppm, and the upper 4-inch
layer of soil outside the Helena Valley has a geometric mean arsenic
content of 6 ppm. Yet the concentration in the upper 4-inch layer within
a mile of the Helena Valley smelter complex averages 50 ppm and is some-
times as high as 150 ppm.

The annual arsenic deposition within 56 km2 (34.80 mi2) of a coal-
fired power plant (amount of coal processed annually is not given) has
been estimated at 2.8 mg/mi2, assuming 99% efficiency in electrostatic
precipitators (Rancitelli et al., 1974). This would result in an annual
increase in arsenic concentration in the soil of 0.2 ppm, assuming that
the arsenic has no mobility and that the top 3 inches of the soil are
uniformly contaminated.
What people eat greatly affects the amount of arsenic they will
consume per day. For most people, ingestion of arsenic in food varies
from 10 to 100 ~g per day, with an estimated average of about 60 ~g per
day (Suta, 1977). The commodity of greatest concern is fish; purchased
and nonpurchased fish account for about 50% of the average daily consump-
tion of arsenic in food. An estimated average of another 7 ~g per day
is consumed in drinking water (Suta, 1977).
100

-------
The pocential for arsenic toxicity in drinking wacer is limited be-
cause arsenic forms insoluble sediment complexes, becomes less toxic when
it is oxidized, and is removed by municipal water treatment (Environ-
mental Sciences and Engineering, 1976). Furthermore, many communities
get their drinking water from sources that are unaffected by major air-
borne arsenic emissions. Arsenic concentrations in drinking waters of
six cities near nonferrous smelters are given in Table X-I. Most of
these arsenic concentrations are below the test detection level and all
are near the U.S. median arsenic concentrations in drinking water of
< 0.005 ppm (Suta, 1977)..
Table X-I
ARSENIC CONCENTRATIONS IN DRINKING WATER
OF CITIES NEAR NONFERROUS SMELTERS
(ppm)
  City   Sample Size Ran ge
S31t Lake City, Utah 4  <0.005-<0.010
n Paso, Texas  4  <0.005-0.012
Anaconda, Montana 1  <0.005
East Helena, Montana 14  0.000-<0.030
But t e, Montana  2  <0.005-<0.010
White Pine, Michigan 1  <0.005
Source:
Kent (1976)
Arsenic bioaccumulates in various organisms in the environment, but
it is not biomagnified. Macroinvertebrates and fishes high in the food
chain contain low concentrations of arsenic even when they feed on lower
level organisms with high concentrations (Union Carbide, 1976).
The most likely route of secondary human exposure to atmospheric
emissions of arsenic is eating vegetables grown in contaminated soils.
An average person in the United States consumes about 200 grams of
vegetables per day, although the amount varies with dietary preferences.
If these vegetables were contaminated with arsenic at average U.S. con-
centrations, the average person would consume about 1.5 ~g of arsenic
per day from vegetables. The arsenic concentrations in vegetables grown
in the Tacoma area could lead to daily consumption of 200 to 2,000 ~g
of arsenic by persons who eat only locally grown Froduce.
101

-------
Appendix
DISPERSION ESTIMATES OF ATMOSPHERIC ARSENIC
CONCENTRATIONS FOR SELECTED SOURCES
The arsenic exposure concentrations given in this report are based
on rough dispersion modeling estimates supplied by EPA (Youngblood,
1978). These dispersion calculations assumed source emission rates that
are rough estimates of actual emissions. The sources of arsenic that
were considered are power plants, pesticide plants, cotton gins, glass
plants, and nonferrous smelters (copper, lead, and zinc). Although the
calculations were made under well-defined assumptions about source charac-
teristics and meteorological data, thp.y are intended to be applicable to
a wide range of sources and geographical/climatological situations. Thus,
they are to be used as rough approximations. The effects of complex
terrain were not considered in the calculations. In specific cases, the
source characteristics, local meteorological conditions, and the presence
of complex terrain may make the estimates inapplicable. This is particu-
larly true with regard to smelters in complex terrain, for which esti-
mated concentrations will be very much higher in situations of plume
impaction.
The dispersion modeling provided estimates of maximum 24-hour
average concentrations for various distances out to 20 km from each
source. These 24-hour maximum estimates were converted to maximum
annual values by dividing by 5. These values were then further divided
by 2.5 to smooth the maximum annual values with respect to direction.
This procedure leads to upper-limit estimates for elevated point sources
such as smelter stacks.
The 24-hour maximum concentration estimates were derived by two
techniques. For the elevated point sources (power plants, smelter stacks,
and glass plants), the Single Source (CRSTER) Model was used. In the
case of the glass plants, a version of CRSTER was used that allows the
consideration of building wake effects. For the low-level sources
(pesticide plant, cotton gin, ~nd the fugitive emissions from the smelter),
the PAL dispersion model was used.
For the CRSTER executions, meteorological data for a I-year period
from Phoenix, Arizona (surface), and Tuscon, Arizona (upper air), were
used becau~e Arizona~meteorological conditions roughly typify those at
many smelter sites. However, the choice was also a matter of convenience
because the intended application of the results does not call for site-
specific considerations. Within the range of uncertainty inherent in
the exposure study, the 24-hour maximum concentration estimates from
CRSTER would not differ significantly over a variety of choices for the
103

-------
meteorological data base. The CRSTER output was screened to eliminate
spurious concentration estimates stemming from anomalous meteorological
events.
In regard to the PAL executions, I-hour estimates were made for
various distances along the downwind plume centerline under the assumed
meteorological conditions of neutral stability (Pasquill-Gifford "D")
and a wind speed of 4 m/s (9 mph). These I-hour estimates were multi-
plied by 0.25 to convert them to maximum 24-hour estimates.
The emission rates used in the dispersion modeling were largely
contrived. This, however, does not affect the utility of the results
because, for a given source, the results are directly proportioned to
the emission rate. T~us, the estimates can be scaled up or down according
to the actual emission rate in any given situation.
104

-------
BIBLI(X;RAPHY
Aboul-Ela, M. M., and C. S. Miller, St~di~s of Arsenic Acid and Residues
in Cotton, MP 771, Texas A&M University (College Station, Texas,
1965) .
American Bureau of Metal Statistics, Non-Ferrous Metal Statistics, 1977
(New York, 1978).
American Iron and Steel Institute, letter from E. F. Young, Jr. to A. Saz
of EPA concerning atmospheric arsenic exposures (June 22, 1978).
American Metal Market, Metal Statistics (Fairchild Publishing, New York,
1977).
Bolton, N. E., J. A. Carter,
Hulett, and W. S. Lyon,
Fired Stelm Plant," Oak
Tennessee (1973).
J. F. Emery, C. Feldman, W. Fulkerson, L. D.
"Trace Element Mass Balance Around a .coal-
Ridge National Laboratory, Oak Ridge,
Bornstein, M. 1., "Assessment of Arsenical Pesticide Plants to Determine
Sources, Levers, and Control Technology for Potential Arsenic
Emissions," GCA Corporation, Bedford, Massachusetts, GCA-TR-75-20-G
(July 1975).
Boyle, R. W., and 1. R. Jonasson, "The Geochemistry of Arsenic and Its
Use as an Indicator Element in Geochemical Prospecting," J. Geochem.
Explor., Amsterdam, 1:251-296 (1973).
Bureau of the Census, "Cotton Production and Distribution," Bulletin 202
(1965).
Bureau of the Census, "Cotton Ginning in the United States:
1972" (1973).
Crop of
Bureau of the Census, "1972 Census of Manufacturers--Glass Products,"
U.S. Department of Commerce, MC72(2)-32A (January 1975).
Bureau of Mines, "Commodity DatA Summaries 1975" (1975).
Bureau of Mines,. letter from L. E. Meirotto to D. S. Barth (U.S. Environ-
mental Protection Agency) concerning a~senic exposures, U.S. Depart-
ment of the Interior (June 14, 1978).
Cooper, J., U.S. Environmental Protection Agency, Private correspondence
( 19 7 8A ) .
105

-------
Cooper, J., U.S.. Environmental Protec~10n Agency, Private correspondence
(February 7 1978B). .
Cuffe, S. T., "Emissions of Arsenic from Glass Melting Furnaces," memo-
randum to J. R. O'Connor, U.S. Environmental Protection Agency
(July 17, 1978).
Davis, W. E., and Associates, "National Inventory of Sources and Emis-
sions: Arsenic--1968," APTD-1507 (May 1971).
Dun & Bradstreet, Reference Book of Manufacturers (New York, Spring 1977).
Duncan, L. J., E. L. Keitz, and E. P. Krajeski, "Selected Characteristics
of Hazardous Pollutant Emissions," Volume II, MITRE Corporation,
MTR-640l (1973).
Durrenberger, C. J.,"Cotton Gin Report," Texas Air Control Board (May
1974).
Durrenberger, C. J., "Particulate and Arsenic Emissions of Texas Cotton
Gins Processing Machine Stripped Cotton," Texas Air Control Board
(December 1975).
Environmental Science and Engineering, Inc., "The Ecological Effects of
Arsenic Emitted from Nonferrous Smelters" (February 1976).
Ferguson, J. F., and J. Gavis, "A Review of the Arsenic Cycle in Natural
Waters," Water Res., ~:1259-1274 (1972).
Herring, W. 0., "Status of Gin Investigation," U. S. Environmental
Protection Agency (October 16, 1973).
Holt, B. R., and J. W. Moberly, "Environmental Mass Balance of Arsenic,"
Stanford Research Institute (1976).
Kent, G., "Data on Arsenic in Drinking Water Supplies," unpublished data,
U.S. Environmental Protection Agency (September 1976).
Klein, D. H., A. W. Anders, J. A. Carter, J. F. Emery, C. Feldman, W.
Fulker30n, W. S. Lyon, J. C. Ogle, U. Talmi, R. I. Van Hook, and
N. Bolton, "Pathways of Thirty-Seven Trace Elements Through Coal-
Fired Power Plant," Envir. ScL Tech., .2,(10) :973-979 (1975).
Lee, H.,. T. O. Peyton, R. V. Steel, and R. K. White, "Potential Radio-
active Pollutants Resulting from Expanded Energy Programs," Stanford
Research Institute (April 1977).
Haasland, D. B., "Arsenic Emissions from Glass Production," Draft
Memorandum, U.S. Environmental Protection Agency (March 5, 1975).
106

-------
Magee, E. M., H. J. Hall, and G. M. Varga, Jr., "Potential Pollutants
in Fossil Fuels," Esso Research and Engineering Co., GRU. 2DJ. 73
. (June 1973).
McBride, B. C., and R. S. Wolfe, "Biosynthesis of Dimethylars ine by
Methanobacterium," Biochemistry, lQ.:4312-.4.317 (1971).
Mosely, G. H., Corning Glass Works, Correspondence with John R. O'Connor,
U.S. Environmental Protection Agency, concerning arsenic emissions
during glass manufacturing (August 28, 1978).
National Academy of Sciences, "Arsenic," Washington, D.C. (1977).
National Institute for Occupational Safety and Health, "Occupational
Exposure to Inorganic Arsenic--New Criteria--1975" (1975).
Natusch, D.F.S., J. R. Wallace, and C. A. Evans, "Toxic Trace Elements:
Preferential Concentration in Respirable Particles," Science,
183(4121):202-204 (January 1974).
Office of Air Quality Planning and Standards, "Background Information
for New Source Performance Standards: Primary Copper, Lead and
Zinc Smelters," U.S. Environmental Protection Agency, EPA-450/2-74-
0028 (October 1974).
Office of Pesticide Programs, "Initial Scientific Review of Cacodylic
Acid," U.S. Environmental Protection Agency, EPA-540/l-75-02l
(December 1975A).
Office of Pesticide Programs, "Initial Scientific Review of MSMA/DSMA,"
U.S. Environmental Protection Agency, EPA-540/l-75-020 (December
1975B).
Office of Toxic Substances, "Technical and Microeconomic Analysis of
Arsenic and Its Compounds," prepared by Vesar, Inc., EPA 560/6-
76-016 (April 1976).
Pendleton, R. W., Jr., Phelps Dodge Corporation, correspondence with
Douglas Costle, U.S. Environmental Protection Agency, concerning
copper smelter emissions (August 3, 1978).
Rancitelli, L. A., K. H. Abel, and W. C. Weimer, "Trace Pollutant Emis-
sions in Fossil Fuel Consumption," in Battelle Pacific Northwest
Laboratory Annual Report for 1973, Part 3, Atmospheric Sciences
(1974).
Research Triangle Institute, "A Screening Study to Develop Background
Information to Determine the Significance of Glass Manufacturing,"
(Research Triangle Park, North Carolina, December 1972).
Reznik, D., Monsanto Research Corporation, letter to Jo Cooper, U.S.
Environmental Protection Agency (October 7, 1975).
107

-------
Roberts, J. W., R. D. Pollack, ~1. J. Svoboda, and H. A. Watters, "Ambient
Air Sampling for Total Arsenic Near the Tacoma Smel ter," Puget Sound
Air Pollution Control Agency (March 1976).
- Ruch, R. R., H. J. Gluskoter, and N. F. Shimp, "Distribution of Trace
Elements in Coal," in Environmental Aspects of Fuel Conversion
Technology, Proceedings ofa symposium, St. Louis, Missouri, May
1974 (October 1974).
Schorr, J. R., D. T. Hooie, M. C. Brockway, P. R. Sticksel, and D. E.
Niesz, "Source Assessment: Pressed and Blown Glass Manufacturing
Plants," Battelle-Columbus Laboratories, EPA-600/2-77-005 (January
1977).
Shearer, S. P., "Arsenic Data," memorandum to R. Neligan, U.S. Environ-
mental Protection Agency (July 11, 1975).
Shutt, T. C., A. P. Herring, and J. L. Drobnick, "Technology and Economics
of Decolorizing Systems," paper presented at 31st Annual Conference
on Glass Problems, Columbus, Ohio (November 20, 1970).
Standard and Poor, Register 6f Corporations (New York, 1977).
Stockham, J. D., "The Composition of Glass Furnace Emissions," J. Air
Poll. Cont. Assoc., 11(11):713-715 (1971).
Sullivan, R. J., "Preliminary Air Pollution Survey of An:enic and Its
Compounds," U.S. Department of Health, Education and Welfare,
Raleigh, North Carolina, APTD-69-26 (1969).
Suta, B. E., "Population Exposures to Arsenic," Stanford Research Institute
(1977).
Texas State Department of Health, "Air Pollution S.tudy of Cotton Gins
in Texas" (Austin, Texas, April 1965).
Union Carbide Corporation, "Review of the Environmental Effects of
Arsenic," Oak Ridge National Laboratory, Oak Ridge, Tennessee,
ORNL/EIS-79 (September 1976).
U.S. Environmental Protection Agency, '''Helena Valley, Montana Area
Environmental Pollution Study," No. AP-9l (1972).
U.S. Environmental Protection Agency, "Draft of Standards Support and
Environmental Impact Statement~ Volume I: Proposed National Emis-
sion Standards for Arsenic Emissions from Primary Copper Smelters,"
Office of Air Quality Planning and Standards, Research Triangle
Park, North Carolina (June 1978).
U.S. Environmental Protection Agency, "Arsenical Pesticides--Internal
Review," Criteria and Evaluation Division, Office of Pesticide
Programs (1976).
108

-------
U.S. Public Health Service, R3dio1ogical Health Handbook, U.S. Department
of Health, Education and Welfare (1960).
Vervaert, A., Private communications, U.S. Environmental Protection
Agency (March 1978).
Weisenberg, 1. J., and J. C. Serne, "Design and Operating Parameters fur
Emission Control Studies: ASARCO, Tacome, Copper Smelter," Pacific
Environmental Services, EPA-600/2-76-036k (February 1976A).
Weisenberg, 1. J., and J. C. Serne, "Design and Operating Parameters for
Emission Control Studies: Phelps Dodge, Ajo, Copper Smelter,"
Pacific Environmental Services, EPA-600/2-76-036f (February 1976B).
Weisenberg, 1. J., and J. C. Serne, "Design and Operating Parameters for
Emission Control Studies: Kennecott, McGill, Copper Smelter,"
Pacific Environmental Services, EPA-600/2-76-036c (February 1976C).
Weisenberg, 1. J., and J. C. Serne, "Design and Operating Parameters for
Emission Control Studies: Magma, San Manuel, Copper Smelter,"
Pacific Environmental Services, EPA-600/2-76-036e (February 1976D).
Weisenberg, I. J., and J. C. Serne, "Design and Operating Parameters for
Emission Control Studies: Kennecott, Hurley, Copper Smelter,"
Pacific Environmental Services, EPA-600/2-76-036d (February 1976E).
Welford, D. S., "Measurements of the U.S. Cotton Industry," National
Cotton Council of America (1963).
Wood, J. M., "Biological Cycles for Toxic Elements in the Environment,"
Science, 183:1049-1052 (1974).
Woolson, E. A., and P. C. Kearney, "Persistence and Reactions of C-14-
Cacodylic Acid in Soils," Envir. Sci. Tech., 2.:47-50 (1973).
Youngblood, P. L., "Rough Dispersion Estimates for Arsenic from Various
Sources," U.S. EPA Draft Memorandum (March 1978).
~[j'@[9)~uiry F!)1
~~,%. il"D~"
W1r~ fM~ L.
Lu
109

-------