United States
Environmental Protection
Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-87/104 Jan. 1988
&EPA Project Summary
Development and
Assessment of Methods for
Estimating Protective Clothing
Performance
Rosemary Goydan, Arthur D. Schwope, Todd R. Carroll, Hsiao-Show Tseng,
andR. C. Reid
Methods were developed for estima-
ting the ability of protective clothing
polymers to act as a barrier to chemical
transport. The methods were devel-
oped for use by the U.S. Environmental
Protection Agency (EPA), Office of
Toxic Substances (OTS) to assess the
ability of protective clothing to reduce
chemical exposure risks as required in
Premanufacture Notification (PMN)
reviews. Two methods were focused
on:
1. Development of a mathematical
model based on Fickian diffusion
theory. The model emphasizes
prediction of cumulative permea-
tion and is based on refinements
of existing theoretical approaches
for estimating two critical
parameters; the solubility, S, and
the diffusion coefficient, D. Per-
meation model predictions were
compared to well-documented
permeation data from the litera-
ture. The model's range of applic-
ability and its limitations are
described.
2. Critical review of the applicability
and reliability of test methods for
measuring permeation resist-
ance, degradation resistance,
liquid immersion weight change,
and chemical sorption/desorp-
tion of protective clothing
materials.
Efforts were initiated to develop an
integrated system that will enable
accurate and efficient evaluation of
protective clothing requirements and
recommendations. Specific accomp-
lishments include: computerization of
the permeation estimation model,
development of a hierarchy that ranks
chemical resistance tests according to
their ability to generate data needed to
assess PMN clothing requirements,
and recommendations for improving
existing test methods.
This Project Summary was devel-
oped by EPA's Water Engineering
Research Laboratory, Cincinnati, OH.
to announce key findings of the
research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering
information at back).
Introduction
Section 5 of the Toxic Substances
Control Act (Public Law 94-469) requires
prospective manufacturers to submit
PMN's, which are reviewed by OTS prior
to manufacture or import of new chem-
icals. OTS is permitted only 90 days to
review each of the approximately 1,800
PMN's submitted annually. While many
substances are not subjected to all
aspects of the review process, those that
are judged potentially toxic require
detailed assessments of the potential for
their environmental release and human
exposure during manufacture, process-
ing, and end use. If concerns are raised
to warrant regulation; engineering con-
trols, work practice restrictions, or
protective clothing and equipment are
investigated as a means to reduce
exposure risks.
The submitter of the PMN often recom-
mends protective clothing as the means
to limit dermal exposures. Occasionally,
the type of clothing is specified; more
often it is not. In either case, OTS must
-------
have means of assessing the exposure
reduction provided by protective clothing.
Two options are available to OTS for
estimating the adequacy of protective
clothing: (1) OTS can request that PMN
submitters test the clothing materials
and submit the resultant data, or (2) OTS
can estimate clothing performance using
available information and predictive
models. Because of the volume of PMN's
and the limited time permitted for each
review, the development of a reliable
model is desirable. If a model is to be
used, it must estimate exposure protec-
tion using only the limited chemical
property data available for PMN substan-
ces. If testing is to be specified, OTS must
have an awareness of relevant test
methods and the limitations of the data
that are produced.
The overall assessment of protective
clothing requirements must consider the
potential health effects of the PMN
substance, the probable exposure con-
ditions, and the effectiveness of clothing
in limiting exposures. This study focused
on predicting the barrier effectiveness of
clothing materials. Figure 1 provides an
overview of the recommended system for
judging clothing effectiveness. Clothing
performance depends directly on the
exposure scenario, and the major factors
that define the scenario are highlighted.
Three general methods for assessing
clothing performance were pursued—
one to develop Pick's law predictive
models, a second based on the analysis
of existing performance test data, and a
third involving the identification of
appropriate chemical resistance test
methods.
Predictive Models
Permeation theory was used to esti-
mate chemical permeation through
clothing materials under continuous
exposure to pure substances. The model
development effort focused on predicting
the cumulative mass-permeated as a
function of time and not solely a break-
through time. OTS needs more than a
single breakthrough time to judge
whether clothing performance will be
acceptable. OTS requires estimates of
the time at which the cumulative amount
permeated reaches a limit of unaccep-
table human risk. The time when this
toxicity limit is reached may be well
Exposure Scenario
• Physical state of chemical
—Liquid, solid, gas
—Neat
—Mixture
• Temperature
• Duration
• Type of work
• Type of clothing
(D
Predictive Models
• Permeation mechanisms
—Pick's law
—Solubility (S)
—Equation of state
—Group contributions
—Statistical correlations
—Diffusion Coefficient (D)
—Empirical correlations
—Concentration
dependence
• Paniculate penetration
Protective
Clothing
Performance
Chemical Resistance
—Permeation
—Degradation
—Penetration
Physical Properties
—Abrasion
—Cut
—Torsional modulus
—Tear
—Tensile
— Thickness
—Puncture
Paniculate penetration
(3)
Data Bases
• Permeation
—EPA "Guidelines"
—NIOSH
—CPCbase
—GlovES
• S Values
• D Values
• Physical Properties
—Must be developed
beyond the time when breakthrough i
detected.
Fickian diffusion behavior wa
assumed and the classic mathematics
relationships were used for the estima
tion of the rate, J, and cumulativ
amount, Q, of chemical that permeate
a polymer film at any time following th
initiation of the exposure. These relation
ships require two key parameters: thi
diffusion coefficient, D, and the solubility
S.
Five generic polymers were empha
sized: butyl rubber, natural rubber
Neoprene,* nitrile rubber, and low
density polyethylene (LDPE). This repor
summarizes the progress to date; refine
ments to the permeation estimatioi
model are continuing.
Solubility Estimation
There are no well tested, pureh
theoretical approaches that providi
accurate predictions of S for the system;
of interest. A theoretical equation of statt
(EOS) approach shows promise althougl
the technique requires further refine
ment. Three group contributior
approaches were investigated and th<
Oishi and Prausnitz approach was deter
mined the most accurate and broadh
applicable. Empirical correlations with
the functional groups of the solutf
molecule were also explored.
Theoretical Approach—Kumar
Equation of State
A statistical mechanics-based lattice
model EOS was recently used with
success by Kumar et al. to model the
phase behavior of supercritical fluid;
containing polymer molecules. A simpli-
fied form of the Kumar EOS was usec
here to estimate chemical/polymei
solubilities at room temperature anc
pressure. S was predicted using thi;
approach for natural rubber, butyl rubber
and Neoprene. In general, the predictions
are accurate to an order of magnitude
but tend to underestimate values
reported in the literature. An importam
advantage is that this approach requires
only three pure component properties 01
the solute to estimate S: the moleculai
weight, vapor pressure, and liquid den-
sity at room temperature.
Figure 1. Schematic representation of an integrated system for judging the suitability of
protective clothing recommendations.
"Mention of trade names or commercial product
does not constitute endorsement or recommenda
tion for use.
-------
jroup Contribution Approaches
Group contribution approaches require
only that the generic structure of the
solute and polymer be known. Most
methods proposed in the literature stem
from an activity coefficient correlation
method called UN1QUAC (universal
quasi-chemical activity coefficients). In
this approach, x. the activity coefficient,
is expressed as a sum of y, and includes
terms for combinatorial, residual, and
free volume contributions.
Solubilities were predicted using three
group contribution approaches and
compared with measured values for butyl
rubber, LDPE, natural butyl rubber,
Neoprene, and nitrile rubber, when
applicable. Table 1 presents results for
natural rubber in which predictions by
the Oishi and Prausnitz, the Holten-
Andersen, and the l!yas and Doherty
techniques are compared with solubility
data. The UNIFAP (UNIQUAC functional
group activity coefficients to polymer
systems) approach by Oishi and Praus-
nitz and the Holten-Andersen approach
are both very accurate for a wide range
of chemicals; the predictions are accu-
rate within an order of magnitude and
many are within a factor of two (i.e., two
times higher or half the measured value).
Predictions are less accurate using the
llyas and Doherty method and can only
be made for a more limited set of
chemicals. Results for butyl rubber,
LDPE, Neoprene, and nitrile rubber are
provided in the report.
Table 1. Solubility Predictions for Natural Rubber/Solute Systems Using Group Contribution
Approaches
Chemical Name
Acetic Acid
Acetic Anhydride
Acetone
Benzene
Benzyl Alcohol
Butylamine
Carbon Tetrachloride
Cyclohexane
Cyclohexanone
Dimethylaminopropylamine
Dimethylethanolamine
Dimethylformamide
Ethanol
Ethyl Acetate
2-Ethyl-1 -Butanol
Ethylene Dichloride
Eth yle nediamine
n-Heptane
n-Hexane
Isopropyl Alcohol
Methanol
Methyl Aery/ate
Methyl Chloroform
Methyl Ethyl Ketone
Methyl Isobutyl Ketone
Methyl Methacrylate
n-Pentanol
t-Pentanol
n-Propanol
n-Propyl Acetate
Tetrachloroethylene
Tetralin
Toluene
Trichloroethylene
o-Xylene
Average*
Experimental
S (g/cm)
0.078
0.039
0.095
3.2
0.15
1.4
8.3
2.8
2.4
1.1
0.17
0.039
0.007
0.43
0.42
2.1
0.087
1.6
1.3
0.040
O.OO2
0.52
4.5
0.46
1.2
1.1
0.12
0.39
0.088
1.3
7.0
4.4
3.5
7.5
3.8
Estimated S (g/cm)*'
UNIFAP
0.062
0.012
0.10
1.6
0.028
0.76
7.3
4.1
2.4
0.74
0.020
0.052
0.019
0.34
0.11
0.21
0.038
3.5
3.4
0.048
O.OO8
0.15
4.4
0.18
1.1
0.28
0.088
0.082
0.047
0.40
4.2
4.6
3.7
4.8
5.0
Holt-And%
na
0.009
0.13
1.7
0.019
na
2.4
0.93
0.62
na
na
na
0.019
0.75
0.068
2.9
na
2.0
1.8
0.035
OO08
0.14
5.0
0.24
0.55
0.37
0.085
0.085
0.036
0.22
na
1.6
1.2
na
0.79
Ily-Doh*
0.002
0.009
0.090
6.5
na
0.33
na
5.8
9.8
na
0010
na
6.2
0.52
0.007
na
6.5
4.5
45
0014
57
na
na
0.14
0.26
0.27
0.17
0028
0.015
1.4
na
38
1.4
na
1.5
"Experimental solubility data are average values for that chemical in natural rubber in the
temperature range from 20° to 30°C.
**(na) indicates that the required group interaction parameters are not available for this solute/
polymer system.
%Holten-Anderson approach.
V/yas and Doherty approach.
The Oishi and Prausnitz method is the
recommended approach. It can be used
to estimate S for the five polymers
emphasized here. In some cases, the
Holten-Andersen approach is more
accurate but its applicability is more
limited. At present, the Holten-Andersen
approach can treat natural rubber, butyl
rubber, LDPE, and solutes from chemical
classes including ketones, alcohols,
esters, ethers, simple chlorinated com-
pounds, and most hydrocarbons. The
llyas and Doherty approach is more
limited and demonstrates only marginal
accuracy. For cases in which the UNIFAP
technique cannot be applied, the Kumar
EOS approach is recommended although
the approach is generally less accurate.
Chemical functional groups that cannot
be treated by the UNIFAP method include
nitriles, tertiary amines, phosphorous
containing compounds, and compounds
containing fluorine.
Diffusion Coefficient Estimation
Models to estimate D for solute/
polymer matrix systems are much less
advanced than solubility estimation
procedures. There are no broadly appli-
cable theoretical models to estimate D
in concentrated polymer solutions. At
present, a useful approach is an empir-
ical correlation of experimental diffusion
coefficients with physical properties of
the solute molecule.
Theoretical Approach—Theoretical
approaches to diffusion coefficient pre-
diction generally involved application of
free volume theory. Although these
models provide a good qualitative repre-
sentation of variations in D with temper-
ature and solute concentration, they are
difficult to apply and require physical
property data that are not generally
available. Consequently, these ap-
proaches are not suitable for PMN review
evaluations where simplicity and broad
applicability are essential requirements.
Empirical Approach—Because of the
limitations of theoretical approaches, the
project focused on correlating diffusion
coefficient data with properties of the
solute. The goal was a simple, broadly
applicable technique that requires only
the physical property data typically
available in a PMN. Published diffusion
coefficients were obtained by literature
review. Additional values were calcu-
lated from permeation versus time data
reported in the literature.
Correlations of measured D values
with properties representative of solute
size and shape were investigated. These
-------
properties include molecular weight,
molecular connectivity, surface/volume
ratios, and the acentric factor. The best
correlations were with molecular weight,
although, for the polymers with less
extensive data sets, the correlations are
not well defined.
In Figure 2, D's are plotted for organic
liquids and gases in natural rubber
versus the molecular weight of the
solute. In general, the data show a
consistent trend with molecular weight.
For the straight-chain hydrocarbons and
other approximately linear molecules,
the D values decrease approximately
linearly by two orders of magnitude as
the molecular weight increases from 10
to 1,000 on the log-log scale. However,
the values for branched or cyclic mole-
cules lie above the general trend.
The development of correlations
requires a sufficiently large set of D
values. Few diffusion coefficient values
were found for butyl rubber, Neoprene,
and nitrile rubber, (less than 10 in each
case), and the degree of confidence in
the correlation equations is less for these
polymers. In general, the approach is not
accurate for cyclic molecules. Analysis
of additional permeation data sets con-
tinues and should enable significant
improvements to the preliminary corre-
lations presented here.
Prediction of Permeation
Behavior
A permeation estimation model proto-
type was developed based on the above
estimation techniques. The model was
coded in FORTRAN and designed to run
on personal computers at OTS. The
permeation model uses Pick's law with
the assumption of a constant diffusion
coefficient. The overall accuracy of the
model is determined by the accuracy to
which D and S can be predicted and the
applicability of simple Fick's law equa-
tions to describe the permeation
behavior.
Permeation data from the literature
were used to judge the accuracy of the
model in predicting the full permeation
curve. Four sources of good quality, well
documented data were used. In total, th
data set for model validation include
data for approximately 2 to 7 solutes fc
each of the five polymers of interesi
Predictions were generated using th
computer model; two approaches wer
used when applicable—one using th<
molecular weight correlation for D am
UNIFAP for S and a second using thi
molecular weight correlation for D am
Kumar EOS forS.
The permeation data and the mode
predictions are summarized in terms o
breakthrough time and steady-stati
permeation rate estimates; graphs com
paring the model predictions with th<
data over the entire permeation curve an
provided in the report for each of the 2{
polymer/solute combinations evaluated
Values of the breakthrough time wen
calculated at a given minimum detectei
flux or minimum detected amount per
mealed. The results for natural rubbe
are summarized in Table 2. Figure ;
compares the model predictions
10-'
I 1 1 1 1 1 1 1 I 1 1 1 ! 1 1 1 1
-
_
A
2
9
\ 9
T
\
\
%n
v
• te
— A
: o
-
_
i l i i i i ill 1 1 — 1 1 U 11
1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
( D ) Acetone
( A ) Benzene
( O ) Butane
( rj ) Cyclohexane
( A ) n-Decane
(A) n-Dotriacontane
\ ^v /
(3 ) Ethane
( ® ) n-Hexadecane
( B ) Hexane
( A ) Isobutane
( O) Isopentane
( • ) Methane
A ( • ) Pentane
( C ) Polybutadiene Oligomer
( + ) Propane
( v ) b-Propiolactone
( y ) Tolulene
\
x§ ^
O
c
1 i 1 1 1 1 1 1 1 1 1 1 1 1 U.J 1 J U-i_L
1 L
-
_
_
—
-
I
-
-
_
-
~-
-
J_1_L
^ A
70"
u
Q
10'
10'
10
W2 103
Molecular Weight
Figure 2. Correlation of diffusion coefficients with solute molecular weight for solute/natural rubber systems.
4
-------
Table 2.
Comparison of Permeation Data with Permeation Model Predictions for Natural
Rubber
Breakthrough Time
Steady State Permeation
Rate (ug/cm2-min)
Thickness Measured Basis* Predicted"*
(cm) (min) (ug/cm*-min) (min) Measured Predicted
Natural Rubber
Acetone
Cvclohexane
Isopropanol
b-Propiolactone
Toluene
0.06
0.06
0.06
0.03
0.06
17
19
5340
25
10
0.76
1.8
1.0
19.7%
109.7
8
20
8
13
11
115
7
21
77
34
525
1.0
3.6
782
34
2
959
59
16
1
22
792
126
*These values are the permeation rate detected at breakthrough and were used as the basis
for predicting breakthrough times.
**AII predictions used the molecular weight correlation for D. Predictions using UNIFAP for S
are reported on the first line followed by predictions using the Kumar EOS S on the second
line. I—) indicates that predictions were not possible.
§This value is the cumulative mass permeated, in units of fig/cm2, detected at breakthrough.
experimental data for natural rubber/
acetone. In most cases, the permeation
predictions using a constant D are
accurate within factors of 2 to 5 and are
often well within the range of experimen-
tal values reported in the literature.
Test Methods and Test Data
Interpretation
For instances when the predictive
models do not apply or clothing perfor-
mance data for the chemical of interest
do not exist, OTS may need to request
protective clothing performance data to
facilitate a thorough review. This is often
the case for mixtures. Thus, it is very
important to have well defined test
methods and data reporting require-
ments. Chemical resistance and physical
property test methods were reviewed
relative to OTS needs, with the emphasis
on chemical resistance testing.
Chemical Resistance
Chemical resistance test methods
were compared on the basis of the types
of results obtained, their approximate
cost, the relative skill level required to
perform them, and their inherent limita-
tions. Based on this comparison, a testing
hierarchy was developed, as discussed
below.
Permeation Tests—In a permeation
test, the chemical of interest is placed
on one side of the clothing material and
the other side is monitored for the
appearance of the chemical. From the
results, the total amount (mass) of
chemical permeating a known surface
area of the clothing at any given time
can be calculated. The cumulative per-
meation (mass/area) or the correspond-
ing permeation rate (mass/area/time)
can be used along with the estimated
frequency, duration, and exposed body
surface area to estimate dermal expo-
sures for specific workplace activities.
Although the test is straightforward
and a standard method (ASTM F739)
exists for its performance, variable
results can be obtained under different
testing conditions for the same chemi-
cal/clothing material. Consequently, in
describing or interpreting the results of
a permeation test, there is a certain,
minimum amount of information
required. This information includes the
breakthrough time, the steady state
permeation rate, the clothing material
thickness and surface area, the analytical
sensitivity, the collection medium flow
rate (open-loop systems) or volume
(closed-loop systems), and temperature.
Immersion Tests—In an immersion
test, the clothing material is exposed on
one or both sides of the chemical of
interest for some period of time. The
change in weight or in other physical
characteristics is measured. In tests in
which the weight is accurately monitored
as a function of time, D and S of the
chemical in the material can be calcu-
lated. Such tests are referred to as
sorption/desorption tests and may be
performed using either liquid or vapor
exposures.
In addition to weight, other physical
characteristics of the clothing material
can be used to assess the overall
resistance to a chemical or solution of
chemicals. These characteristics include
dimensions, puncture resistance, tear
resistance, elongation resistance,
strength, and so forth. Such tests are
referred to as degradation tests since
significant detrimental changes in the
clothing due to the chemical are of
interest.
Testing Hierarchy
Both the chemical resistance and the
physical properties of the clothing mate-
rial must be considered in judging its
acceptability. While the physical property
requirements are specific to the partic-
ular application, chemical resistance is
a more general concept. Therefore, a test
hierarchy was proposed for assessing
chemical resistance only (Table 3). The
testing hierarchy ranks chemical resist-
ance tests according to their ability to
generate data that can be used directly
to estimate the effectiveness of the
clothing material in reducing exposure
risks.
Physical Characteristics
In combination with chemical resist-
ance, protective clothing must possess
certain physical characteristics. These
characteristics can be classified under
three headings: physical properties of the
base materials and components of
construction, type and quality of con-
struction, and human factors. Physical
property test procedures to measure
these characteristics were reviewed;
however, the applicability and limitations
of these procedures have not been
established relative to clothing require-
ments. The present state of the art is such
that the ideal set of physical character-
istics for any particular application may
be difficult to define and quantify. Efforts
in this regard should be pursued.
Integrated System
Development
The overall objective of this project is
to improve OTS's capability to rapidly and
effectively assess the adequacy of the
protective clothing recommendations
-------
Replicates of
Experimental data,
Bhown etal.. 1985
O (UNIFAP S/CORR D)
20
740
Figure 3. Comparison of model predictions for permeation flux as a function of time with experimental data for acetone/natural rubber.
Table 3. Test Data Priority for Estimating the Chemical Resistance of Protective Clothing
Polymers
Chemicals Test Methods
Single- Component
Liquid/Gases
Mult/component
Solutions
1. ASTM MethodF739 Permeation Test
2, Permeation Cup (if chemical has sufficient vapor pressure)
3. Weight change from liquid immersion.
4. D and S from vapor or liquid sorption/desorption test.
1. ASTM MethodF739 Permeation Test
provided in PMN submittals. This
involves assessing the feasibility of
developing a model, test methods, and
test data interpretation guidelines to
meet OTS needs for accuracy and effi-
ciency. This assessment will serve as the
basis for developing an integrated system
for reviewing PMN protective clothing
requirements based on both predictive
model estimates and clothing perform-
ance data.
The current PMN review process was
analyzed in detail with OTS. Based on
this review, a permeation estimation
computer model was developed and
evaluated by OTS for incorporation into
the standard review process. The com-
puter model was designed for use on IBM
personal computers. The model requires
input of basic physical properties (i.e.,
molecular weight, liquid density, vapor
pressure and/or the functional groups)
of the PMN substance. Permeation
predictions are output in either tabular
or graphic form. This computer model
prototype represents one component of
the overall system recommended for use
in reviewing clothing recommendations.
Existing databases of chemical resist-
ance information were also reviewed.
These were reviewed to assess the
feasibility of using data for existing
chemicals with structures and properties
similar to those of a PMN substance to
estimate clothing performance for a PMN
substance.
Conclusions and
Recommendations
A system for estimating clothing
performance should consider the cumu-
lative amount of chemicals that cross the
protective clothing boundary. A theoret-
ical model developed in this project was
successful in estimating the cumulative
permeation to within a factor of five. If
-------
'greater accuracy is required in PMN
review, additional refinement of the
models is necessary. Alternatively, OTS
can estimate protective clothing perform-
ance on the basis of data for existing
chemical analogues. Where there is no
such chemical, the PMN chemical/
clothing combination must be tested.
The recommended approaches to
estimating permeation behavior as a
function of time are those based on Pick's
law. These require the estimation of the
diffusion coefficient and the solubility of
the chemical in the polymer. Techniques
for estimating D and S were developed
and used in a permeation model proto-
type based on Pick's law. In most cases,
the permeation predictions were well
within the range of experimental values
reported for a specific permeant/polymer
system. In other cases, the predictions
were less consistent with the experimen-
tal data and this result may be because
certain assumptions of the methodology
were not valid. For example, the assump-
tion of a constant diffusion coefficient is
likely to be invalid for polymers that are
swollen to a high degree by the
permeant.
For PMN substances that are mixtures
and those with no chemical analogues
for which there are permeation data,
testing is required. Permeation testing is
preferred and should be performed
according to ASTM Method F739 or a
similar procedure under well controlled
and documented experimental condi-
tions. Degradation and sorption/desorp-
tion test methods provide either funda-
mental parameters for use with
predictive models or general indications
of clothing performance. However, the
results of these tests require interpre-
tation by persons with some training in
polymer permeation science.
To achieve the goal of developing an
accurate and efficient method for judging
the suitability of PMN protective clothing
recommendations, OTS should continue
its efforts to develop a three-part,
integrated system. The system should
incorporate (1) predictive models for
estimating chemical resistance, (2)
existing chemical resistance and phys-
ical property databases, and (3) the
identification and specification of
appropriate test methods.
The full report was submitted in
fulfillment of Contract No. 68-03-3293
by Arthur D. Little, Inc., under the
sponsorship of the U.S. Environmental
Protection Agency.
Rosemary Goydan. Arthur D. Schwope, Todd R. Carroll, and Hsiao-Show Tseng
are with Arthur D. Little, Inc., Acorn Park, Cambridge, MA 02140; and Robert
C. Reid is with Massachusetts Institute of Technology, Cambridge, MA 02139.
Michael D. Royer is the EPA Project Officer (see below).
The complete report, entitled "Development and Assessment of Methods for
Estimating Protective Clothing Performance." (Order No. PB 88-133 657/
AS; Cost: $25.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Water Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAIC
EPA
PERMIT No G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-87/104
0000329 PS
TKET
------- |