EPA-560/1-77-002
PRE-SCREENING
FOR ENVIRONMENTAL HAZARDS -
A SYSTEM FOR SELECTING
AND PRIORITIZING CHEMICALS
APRIL 1977
PHASE 1 REPORT
ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF TOXIC SUBSTANCES
WASHINGTON, D.C.
-------
PRE-SCREENING FOR ENVIRONMENTAL HAZARDS --
A SYSTEM FOR SELECTING AND PRIORITIZING CHEMICALS
REPORT TO THE
OFFICE OF Toxic SUBSTANCES
U,S, ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D,C, 20460
APRIL 1977
-------
EPA REVIEW NOTICE
This report has been reviewed by the Office of Toxic
Substances, EPA, and approved for publication. Approval
does not; signify that the contents necessarily reflect
the views and policies of the Environmental Protection
Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for
use.
This document is available to the public through the
National Technical Information Service, Springfield,
VA. 22151.
-------
TABLE OF CONTENTS
Page
List of Tables iv
List of Figures v
SUMMARY 1
I. INTRODUCTION 2
A. BACKGROUND 2
B. PURPOSE AND SCOPE 4
C. DEVELOPMENT OPTIONS 6
II. OVERVIEW OF THE PROPOSED SYSTEM 20
A. INTRODUCTION 20
B. EVENTUAL ENVIRONMENTAL LEVELS 20
C. LEVELS OF CONCERN 23
D. RANKING METHODS 25
III. EVENTUAL ENVIRONMENTAL LEVELS 26
A. INTRODUCTION 26
B. ESTIMATION OF EVENTUAL ENVIRONMENTAL LEVELS 26
C. ESTIMATION OF EMISSION RATES 29
D. ESTIMATION OF PHYSICOCHEMICAL PROPERTIES 31
E. TRANSPORT RATES BY DIFFUSION 33
F. COMPARTMENT DATA 35
IV. LEVELS OF CONCERN 46
-------
TABLE OF CONTENTS (Continued)
Page
V. RANKINGS 50
A. SUBSTANCES OF CONCERN 50
B. PRIORITY ORDERING 50
VI. SOME EXAMPLES 51
A. INTRODUCTION 51
B. DATA ON THE COMPOUNDS 51
C. RESULTS 54
D. RELATIVE PRIORITY 60
E. SENSITIVITY ANALYSIS 60
VII. RECOMMENDATIONS 70
APPENDIX I—DISTRIBUTION OF A POLLUTANT IN THE ENVIRONMENT 1-1
APPENDIX II—CHEMICAL CLASSES WITH POTENTIAL FOR CAUSING II-l
BIOLOGICAL DAMAGE
APPENDIX HI—ESTIMATION OF DIFFUSIVITIES III-l
APPENDIX IV--SAMPLE COMPOUND DATA IV-1
-------
LIST OF TABLES
Table
No. Page
I Relation of System Concepts and Criteria 19
II U.S. Compartment Data 36
Ilia Water Flows Between Compartments—Case 1 38
I lib Water Flows Between Compartments—Case 2 39
I lie Water Flows Between Compartments—Case 3 40
11 Id Water Flows Between Compartments—Case 4 41
IVa Data Sheet No. 1 52
IVb Data Sheet No. 2 53
Va Steady-State Concentrations of Benzene 55
Vb Steady-State Concentrations of Bis(2-Chloroisopropyl) 56
Ether
Vc Steady-State Concentrations of Chlorodifluoromethane 57
Vd Steady-State Concentrations of Methyl Chloroform 58
Ve Steady-State Concentrations of Trichlorofluoromethane 59
Vla-b Steady-State Concentrations of Benzene 61 - 62
VIc-h Steady-State Concentrations of Benzene 64 - 69
iv
-------
LIST OF FIGURES
Figure
No. . Page
1 Schematic Diagram of a Hierarchical System 13
2 Pollution Chain Diagram for the Environmental Stressor 14
Lead
3 Structure/Toxicity Correlation Systems 16
4 Schematic Diagram of the Proposed System 21
5 Schematic Diagram of the First Branch of the System 22
6 Schematic Diagram of the Second Branch of the System 24
7 Schematic Diagram of Intercompartment Flows of 28
Emitted Chemical
8a Schematic Diagram of Flows For "Case 1". 42
8b Schematic Diagram of Flows For "Case 2". 43
8c Schematic Diagram of Flows For "Case 3". 44
8d Schematic Diagram of Flows For "Case 4". 45
9 Schematic Representation Of Intersecting Dose-Response
Curves 47
10 Range Of Actual Maximum Allowable Concentration
Related To Calculated Values. 48
-------
SUMMARY
This report presents the results of a study whose objective was the
development of a conceptual scheme for ranking chemicals emitted into
the environment in order of their hazard potential. Although the major
focus is on preliminary screening of chemicals prior to full commercial
production, the scheme we recommend could also be used for evaluating
the potential hazard of chemicals already in production.
We have explored a number of alternatives for ranking chemicals so that
subsequent research efforts may be properly focused. The method we rec-
ommend for subsequent development has the potential of fulfilling the
needs. Basically, the method consists of selecting chemicals for further
attention by comparing the concentration of each chemical that may be
expected in the environment to the concentration levels of that chemical
which are of concern.
The proposed method for estimating the eventual environmental level is
based on a multi-compartment model of the environment. In order to
provide estimates with moderate effort the model is substantially sim-
plified. The emphasis has been on ensuring that the model does not
underestimate the eventual levels, and that overestimation is kept
within reasonable bounds.
The proposed method for estimating levels of concern was also selected
on the basis of simplicity and accuracy. A test using available infor-
mation on tolerable air concentrations indicates that the estimated
levels would be adequate for preliminary screening.
The method also provides the capability of ranking the selected chemicals
into more refined priorities by estimating the time horizon during which
regulatory action would prevent significant deleterious effects.
Before the conceptual scheme can be implemented additional research is
required. The main area of research to be pursued is the development of
improved methods of predicting levels of concern from available informa-
tion on chemical composition. The performance of the system with a
number of test chemicals also needs investigation.
-------
INTRODUCTION
A. BACKGROUND
The hazards posed to the total environment by certain substances is now
widely appreciated. These substances affect the environment, threaten
the integrity of ecological niches, or endanger man by a variety of
modes of action ranging from direct effects, through effects of their
decomposition products, bioaccumulation in prey-predator chains, syner-
gism, and interaction products.
It is generally recognized that the potential for damage could often be
anticipated if there were adequate data on the toxicity of the substances
involved in a variety of relevant species. Yet, the collection of an
adequate data base on chronic toxicity would entail substantial expendi-
tures and extended periods of time.
In an economy that relies heavily on new materials to improve the quality
of life, decrease the cost of goods and thereby improve their distribution
to people of all income levels, the potential delays and costs associated
with thorough testing prior to production is viewed with substantial con-
cern. This concern is quite justified if we realize that the eventual
distribution of substances in the environment is not easy to predict and,
in fact, it may take decades before it can be estimated accurately enough
to relate these levels to the toxicological information. A further cause
for concern is that in some compartments of the environment the concen-
trations may continue to increase well beyond the time at which release
of the substance has been discontinued.
In spite of and because of these difficulties, it is essential that some
methodology be developed that will allow an orderly review of environ-
mental contaminants and will lead to the selection of some subset of
these as being of sufficient concern as to warrant the development of an
adequate toxicological data base or, in extreme cases, a reduction in the
level of emissions into the envimoment.
The recently enacted Toxic Substances Control Act requires the testing of
chemical substances and mixtures which "may present an unreasonable risk
of injury to health or the environment." Test data might be required, for
example, to establish potential risks of acute toxicity, subacute toxicity,
chronic toxicity,, persistence, carcinogenicity, mutagenicity, teratogenicity,
behavioral disorders, etc. The Act recognizes the need for prioritizing
chemicals for testing, and provides for the establishment of a list of
chemicals, not to exceed 50 at any time, for which test data are most ur-
gently required.
-------
Prioritization is of key importance to protecting health and the en-
vironment, without imposing both major economic burden on the chemical
industry and a major administrative review burden on the EPA. The
National Institue of Occupational Safety and Health "Registry of Toxic
Effects of Chemical Substances" (formerly called "The Toxic Substances
List") 1976 Edition, for example, includes nearly 22,000 different
chemicals, and the numbers expected to be included in subsequent editions
is currently estimated at about 100,000 unique toxic substances. De-
velopment of a full battery of health and environmental effects test
data for all of them is clearly impractical within a realistic time
frame. The alternative described in this report is aimed at the de-
velopment of an objective prioritization methodology capable of (1)
classifying chemical substances with respect to the probable risk they
present to human health and/or the environment; and (2) identifying
the kinds of test data that would assest in determining whether or not
the probable risks are "unreasonable."
If such a methodology is to be effective in reducing the amount of
data which must be developed, while at the same time directing data
development efforts to the most crucial problem area, it should have
the following characteristics:
(1) The screen should "pass" a significant fraction of
chemical substances, on the grounds that they have
such a low probability of presenting unreasonable
risks under current and projected conditions of use,
that additional data development does not appear to
be worthwhile. (This assumes of course that a large
number of chemical substances can defensibly be cat-
egorized in this way.)
(2) Ideally the screen should also provide some indication
of the nature of the probable risk for substances that
do not "pass" (i.e., indicate whether the risk is to
air, water, and/or ground pollution, and whether it
is carcinogenicity, teratogenicity, mutagenicity,
chronic toxicity, etc., to humans; phytoxicity; per-
sistance; bioaccumulation; synergism, etc.)
-------
The list of substances which "pass" the screen would include substances
such as those on the FDA's GRAS list and substances which, though poten-
tially damaging at some concentrations, are not likely to reach concen-
trations which pose unreasonable risks to health or environment. They
would be substances, which on the basis of current knowledge and
perceptions, appear to be sufficiently safe to require no further testing
at the moment. As new knowledge develops and perceptions change, the
list would have to be reexamined and reevaluated.
To say that a chemical substance is not hazardous to human health and/or
the environment is to imply that the substance does not induce a whole
variety of potential adverse effects traditionally associated with
chemicals. To say that a chemical substance is hazardous is to imply
that the substance exhibits at least one adverse human health or
environmental effect. If only one such effect is suspected, then that
is the effect for which data development should be prescribed. Even if
there is a high probability that a chemical substance may produce several
adverse effects, it may not be necessary to document all of them. If,
for example, a substance is a suspected human carcinogen and also may
lead to fires in landfills, data development could probably most usefully
be focused on the question of carcinogenicity.
A reasonable and defensible chemical screening system is not a substitute
for experimental data. A screening system is nothing more than a system-
atic mechanism for reviewing available data, and for prioritizing future
data needs, so that resources (which are always limited) may be directed
as early as possible into the most crucial problem areas. This report is
concerned with the development of an objective screen based on the amounts
of chemical substances released into the environment and the kinds of
problems associated with the projected levels of such substances in the
environment. Prioritization of the problems identified with respect to
their need for attention and with respect to the kinds of data that should
be sought is a subjective matter which is well beyond the scope of the
present effort.
B. PURPOSE AND SCOPE
The overall goal of this project is to design, implement and install
within the Office of Toxic Substances an identification system for
selecting and ranking chemicals, chemical classes or use classes by
objectively assessing their environmental hazards. The project has been
organized into two phases - the first focused on systems design and the
second directed towards implementation. This report summarizes the
results of Phase I.
-------
Phase I encompassed the following tasks:
• Task 1 - System Design Criteria - Criteria were developed to
guide the design and development of an environmental hazard
identification system that would meet the needs of the Office
of Toxic Substances.
• Task 2 - Analysis of Information Needs - Minimum input param-
eters were defined that should enable a system to select and
rank chemicals potentially hazardous to the environment.
• Task 3 - Formulation of System Concepts - A number of poten-
tial system concepts were formulated to serve as informal
models against which to evaluate selected systems.
• Task 4 - Evaluation of Existing Identification Systems -
Prior work had shown that none of the many existing systems
were readily adaptable to meeting the specific needs of the
Office of Toxic Substances.1 Several of the more promising
approaches, however, were evaluated against the design
criteria, information requirements, and system concepts
developed in Tasks 1-3 in order to define their shortcomings
more precisely.
• Task 5 - Resolution of Information Gaps - The problem of
data availability, not just for specific chemicals, but in
whole areas of health and environmental concern, was con-
sciously and seriously considered, but not entirely resolved.
• Task 6 - Proposed System - A basic system with a number of
variations of increasing complexity, has been developed for
selecting and prioritizing environmental hazards.
• Task 7 -System Test Methodology - A methodology for testing
the applicability and reliability of the basic system, and
for evaluating the potential benefits of the more complex
variations is presented in the final section of this report
and concludes the Phase I effort.
1 Literature Search and State-of-the-Art Study of Identification Systems
for Selecting Chemicals or Chemical Classes as Candidates for Evalua-
tion (EPA-560/1-74-001), Environmental Protection Agency, Office of
Toxic Substances, Washington, D.C., November 1974.
-------
C. DEVELOPMENT OPTIONS
1. System Design Criteria
Some of the criteria developed to guide the design of a workable environ-
mental hazard identification and prioritization system include:
(1) Objectivity. Policy decisions with respect to potentially toxic
substances in the environment must of necessity be subjective. The sub-
jective decisions, however, are generally required to be reasonable and
defensible in the legal sense. This usually means that they must stem
from an even handed interpretation of objective facts. The desired
identification system must be objective, in terms of input requirements
and procedures or rules to be followed in selecting and ranking chemicals.
Use of the output results in decision making is subjective and need not,
in fact cannot, be addressed by the objective system sought. The neces-
sity of objectivity implies that the system would accept only a modicum
of external judgment or personal interpretation. To the extent that it
may be desirable or necessary to distinguish hazards to target populations,
OTS has established the following order of importance (descending): a) man,
b) economically significant animals and plants, and c) ecologically im-
portant species and then, presumably, the inanimate environment.
(2) Reproduci'bility. The identification system(s) must be capable
of producing identical results (at any given point in time) when operated
by different people. This is not a trivial problem due to the plethora
of information sources and the likelihood that some judgment may be
required even in the most objective system.
(3) Credibility. The system must possess demonstrable credibility
in selecting and prioritizing potential chemical environmental hazards.
The results should be statistically credible, i.e., at most a small per-
centage of the substances ranked as non-hazardous should turn out to give
rise to major health or environmental problems; and at most a small
percentage of the chemicals ranked as highly hazardous should in fact
prove to be benign.
(4) Specificity. The system should classify chemicals, chemical
classes, or use classes into their probable major hazard categories,
e.g., carcinogenicity, mutagenicity, teratogenicity, oral toxicity,
dermal toxicity, inhalation toxicity, aquatic toxicity, bioaccumulation.
It may also be desirable to rank chemicals according to the perceived
risk presented to different target populations, i.e., man/animals,
plants, and the inanimate environment.
-------
(5) Discrimination. The system must possess the ability to scale
chemicals according to their associated risks of environmental hazard.
A process of discrimination is needed to distinguish among different
potential hazard levels and thereby achieve a prioritization. In general,
it may be expected that the simpler systems will yield coarser gradations.
(6) State-of-the-Art Concepts. The system should utilize only proven
state-of-the-art techniques, methodologies, information sources, etc.,
and not attempt to incorporate heretofore untested or incompletely
developed approaches. For example, a new and unknown theory relating
chemical structure to-biological activity should be incorporated into the
system only as a last resort because its merit would not be known before-
hand. The same consideration would hold, for example, in deciding whether
to include a new information center under development and not yet opera-
tional .
(7) Response Time. A realistic time must be established for the
system to return results once it is set in motion.
(8) Personnel Requirements. The system must be designed in confor-
mance with anticipated personnel skill levels and numbers. A system
demanding unavailable technical skills must be avoided.
(9) Expansion. The system should be devised so that it may evolve
without undue hardship as new information sources and techniques become
available in the future.
(10) Built-in Hierarchy. In recognition of the many potential infor-
mation sources and voluminous data (not all of which is necessarily
pertinent), it would be desirable to develop a hierarchical system which
would produce results of increasing specificity and credibility the
further the process was followed. That is, an early indication of
probable chemical toxicity (but one with limited credibility) might be
achieved by following the recommended process to a predetermined point.
Succeeding stages of evaluation requiring more information and analysis
would lead to improved predictions of hazard classes and levels.
(11) Gaps in Knowledge. The system should identify the existence of
information gaps which, if filled, would permit improved predictions.
-------
(12) Statistical Confidence. If possible, hazard predictions should
be accompanied by statements of statistical confidence, however approxi-
mate these might be. The measure of statistical confidence can be viewed
as an indicator of the need for additional information. A hazard evalu-
ation accompanied by a low confidence level indicates that more data may
be required to yield a stronger statement.
(13) Input Data Requirements. The system should be capable of
selecting and prioritizing potential chemical environmental hazards on
the basis of data'normally provided by the manufacturer.
(14) Ability to Deal with Degradation Products. Chemically induced
health and environmental effects may be due not only to manufactured
chemical substances, but also, and sometimes entirely, to degradation
products. Where such products and their properties are known, the system
should be capable of handling them in a normal way. When the routes of
degradation of a chemical substance are unknown or very complicated, it
is unlikely that any simple identification and priorization system will
be able to flag the potential hazards accurately.
(15) Ability to Deal with Synergism. The goal of the project
(cf. section I.B.) is to design an objective system for selecting and
ranking chemicals, chemical classes or use classes, based on their
environmental hazards. It is implied that any selection or ranking
algorithms that may be developed will be applied to individual chemical
substances or chemically related groups of substances. Environmental
hazards, however, may result from or be amplified by synergistic inter-
actions between or among unrelated chemical substances. There is very
little data on the importance of synergism in the environment, and even
if there were more, it would not be easy to incorporate synergistic ef-
fects into an objective system design. For the subjective regulatory
viewpoint, the problem of synergism would be even more difficult to deal
with.
2. Analysis of Information Needs
The primary rationale for an early warning or pre-screening system for
environmental hazard identification and prioritization stems from the
desirability of reducing the requirements for extensive experimental
data development. A complete experimental evaluation of the potential
environmental impact of a chemical substance would involve toxicological,
pharmacological, and metabolic studies in a number of species (e.g., mice,
rats, dogs, rabbits, domestic animals, fish, wildlife, lower aquatic
organisms, plants); transport mechanism and persistence studies in air,
water, and various soil types; potentiation studies; bioaccumulation
studies; degradation studies and evaluation of the hazardous effects of
degradation products. Not only would such an experimental program be
time consuming and expensive, but so much data would be developed that
it would be difficult to sort out what the real problems are. Technical
resources would be more effectively utilized in developing data in
8
-------
particular areas for specific chemical substances where there is good
reason to believe that serious health or environmental effects may be
found. Choosing productive areas to work on is not an easy task for
anyone. Nonetheless, it is not possible to do everything, and the
choice is made, with greater or lesser degrees of success, by individ-
uals, corporations, and agencies. An early warning system should be
an effective tool for helping to guide data development efforts towards
major problems.
The more data that must be developed experimentally as input to an
early warning system, the less useful it can be as a planning tool for
focusing future technical effort, i.e., the more effort required to
develop routine input data, the less effort available to investigate
specifically identified potential problem areas.
For chemicals that are either produced commercially or under considera-
tion for commercial production, the manufacturer can usually supply a
data sheet which includes:
e Common and/or trade name;
• Chemical class and/or structural formula;
e Physical properties (e.g., melting point, boiling point,
vapor pressure, solubility, etc.);
t Chemical properties (e.g., reactions with air and mois-
ture, if any, and other relevant reactions); and
e Suggested applications.
A large chemical company has reported that they would normally make some
additional measurements during the course of development of a new
product specifically to provide some preliminary indications of potential
environmental impacts.2 One parameter that might be experimentally
determined is the octanol/water partition coefficient, which appears to
be correlated with bioconcentration in the environment. Another is
five-day biological oxygen demand (BODs), which provides some indication
of the possibility of microbial decay in aquatic or soil environments.
In addition, some initial toxicological tests would be carried out.
These might include, for example, a determination of acute oral, inhala-
tion, dermal, and/or ocular toxicity to rats or mice.
2 Papers of a Seminar on Early Warning Systems for Toxic Substances
(EPA-560/1-75-003), Environmental Protection Agency, Office of Toxic
Substances, Washington, D.C., July 1975, p. 167.
-------
Most chemically induced health and environmental effects (with the pos-
sible exception of cancer) are concentration dependent. Even a rough
assessment of possible environmental hazard, therefore, requires some
knowledge of the concentration levels to which potentially affected
populations might be exposed. The primary data likely to be available
that might be related to environmental concentrations are planned pro-
duction and major uses. Manufacturers usually have such data, but would
generally be reluctant to release it unless required by law to do so.
On the basis of a structural formula, a few easily obtainable physical
and chemical properties, a five-day BOD, an indication of acute toxicity,
and some estimate of production and use, it is not possible to predict
with certainty the human or environmental hazards of a chemical or
chemical class.. If only the minimal data base is available, then, the
real question is whether a system can be developed which will identify
potentially hazardous chemicals, chemical classes or use classes with
sufficient accuracy to justify its implementation.
3. Formulation of System Concepts
The one guiding principle used in formulating system concepts was
simplicity. There are several reasons for assigning paramount importance
to simplicity. One is that the current state of knowledge about factors
that produce environmental hazards is so primitive as to preclude any
but the simplest of identification systems. Another is that to a simple
or even simplistic system, refinements and embellishments may be added
as needed. Every additional refinement however will usually entail
greater efforts at data collection, information processing, and finally
interpretation of results. It is clear that any contemplated system
must not entail greater effort at selection and ranking than would be
involved in the experimental pre-screening tests themselves. For example,
it might be of interest to develop for each selected chemical to be
prioritized a comprehensive statement of the conditions of exposure per-
taining to the principal plant, animal, and inanimate populations at
risk. The enormity of this undertaking alone would seem to overwhelm the
basic objective of developing a tool for selecting pre-screening candi-
dates.
A prior report, Literature Search and State-of-the-Art Study of Identifi-
cation Systems for Selecting Chemicals or Chemical Classes as Candidates
for Evaluation/* reached the conclusion that none of the numerous exist-
ing systems reviewed are readily adaptable to the problem at hand. We
concur in that conclusion, and only discuss here a number of systems
concepts that address the early warning problem directly.
3 Op. cit.
10
-------
(1) Systems Based on Clinical and Epidemiological Studies. One
approach to early warning is to monitor the concentrations of toxic sub-
stances in the environment, track various health indices (urine analysis,
blood analysis, pulmonary function, etc.) in the exposed populations, and
look for correlations.1* Apart from obvious practical difficulties
(e.g., deciding which of thousands of substances to monitor and selecting
relevant health indices), the results are not likely to be obtained early
enough to avert serious problems.
(2) Laboratory Model Ecosystems. Model terrestrial-aquatic systems
have been used to evaluate bioconcentration, bio-and chemical degradation,
and toxicity of chemical substances to fish, snail, mosquito, daphnia,
and algae. The experiments are quite time consuming, and the toxicity
results are not readily extrapolated to man.
(3) Cost-Risk-Benefit Analysis. Most of the identification system
concepts for selecting and ranking chemicals focus on the risk of hazard-
ous health and environmental effects. The higher the risk is, the higher
the priority that is assigned to the chemical for further attention.
Another possible prioritization concept is based on an evaluation of the
benefits that accrue from accepting a risk, and the costs of reducing or
eliminating the risk.5 This concept is fundamentally sound, but large
scale implementation is almost totally impractical. Definition and
quantification of environmental risks requires that a great deal more be
known about the fate of a chemical in the environment than could be
estimated from a few simple tests. If non-economic costs and benefits
are included in the analysis (and they must be to assure credibility),
their identification and quantification is also extremely difficult.
(4) Delphi. Industry, government, university, and public interest
leaders or experts in the fate of toxic substances in the environment
frequently must (and do) make decisions on priority problems without the
benefit of a great deal of objective data. They rely very heavily on
their own experience in interpreting and weighting the data that are
available, and generally have some rationale for reaching an essentially
subjective conclusion. Prioritization systems based in the Delphi tech-
nique rely on the development of consensus among experts acting independ-
ently as part of a panel. The National Science Foundation recently as-
sembled a Delphi panel to select and prioritize organic compounds hazardous
to the environment.6
4 Op. cit., pp. 5, 82, and 154.
5 Ibid.s p.93.
6 Final Report of NSF Workshop Panel to Select Organic Compounds Hazard-
ous to the Environment (AEN 74-14553A02), Institute of Environmental
Medicine, New York University Medical Center, October 1975.
11
-------
The panel members identified chemicals of concern solely on the basis of
their own experience, and refined their judgments based on available data
in production, import, uses, disposal, and toxicity. Through successive
examinations by the panel, individually and together, in progressively
greater detail, a prioritized list was developed. The system is certain-
ly not objective. It is probably not reproducible (a different panel
might reach a different consensus). It does have a modicum of credibility,
at least among those who value and respect expert opinion. While experts
are sometimes wrong, they generally establish their reputations as
"experts" because they are more often right.
(5) Hierarchical Testing and Evaluation. A hierarchical system for
assessing the probable impacts of chemical substances on the environment
has been proposed by Robert J. Moolenaar of the Dow Chemical Company.7
From some simple measurements on the basic physical, chemical and bio-
logical properties of a chemical, problem areas (if any) for more detailed
experimental evaluation are identified. Strategies are devised for avert-
ing potential environmental problems confirmed by the detailed evaluation.
Actual environmental problems that may not have been anticipated in
earlier stages of testing are assessed by field observations and monitor-
ing of the effects of handling, use and disposal.
The basic system is shown schematically in Figure 1. Interpretation of
the basic property measurements to determine the additional data needed
for environmental impact assessment can be done fairly objectively. In
the illustration shown in Figure 2, the substance has a high vapor
pressure, a low solubility in water, does not degrade readily, does bio-
concentrate, and is acutely toxic to mammals via the inhalation route.
Such a substance could be an air pollution problem, but is unlikely to
be a water pollution problem. Further work might involve subacute and
chronic inhalation toxicity studies in rats, mice and/or dogs (including
pathological examination for cancer), and bioconcentration studies in a
model ecosystem,. The fundamental screening parameter is taken to be
environmental degradation; i.e., if a substance degrades rapidly to non-
toxic products in air and water, it is not likely to present an unreason-
able risk, even if its octanol/water partition coefficient is high.
The initial (Level I) property measurement can be done relatively rapidly
and relatively inexpensively. Certain .types of problems will be missed.
For example, the need for detailed toxicology will only be triggered for
substances which are acutely toxic and/or which bioconcentrate. Chronical-
ly toxic substances which do not bioconcentrate would not be flagged.
(6) Stressor Matrix System. David L. Morrison of Battelle has
developed a systematic approach for organizing and evaluating the multi-
tude of data which relates to environmental hazard identification.8 From
7 Op. oit.f p. 167.
B.Ibid., p. 175.
12
-------
FIGURE 1
SCHEMATIC DIAGRAM OF A HIERARCHICAL SYSTEM
MEASUREMENT OF BASIC PROPERTIES
Reactions with water
Reactions with air
Biological oxygen demand
Is degradation to
non-toxic substances rapid?
ye
\
No further work
Solubility in
water
Will the
stance move
with the water?
Vapor
Pressure
Will the sub-
stance move
with the air?
Octanol-Water
Partition
Coefficient
T
Is the substance
likely to
bioconcentrate?
Acute
Toxicity
Level I
Measurements
vesJToxic to plants? tac.^
Toxic to fish?
yes
Production and use data
^4 _^ ^j ^
' • Subacute and chronic toxicity via inhalation
I • Bioconcentration
^es_
-. ves
-------
FIGURE 2
POLLUTION CHAIN DIAGRAM FOR THE
ENVIRONMENTAL STRESSOR LEAD
Stressor
(effect)
i
Dose
i
lotion
1
Ingestbn
Cont
Neighborhood environment
Leod mining,
smelting, refining
14
-------
information on sources of a chemical substance; distribution of emissions
in the atmosphere, hydrosphere, lithosphere, and biosphere; transport
paths; and effects or consequences of exposure, a pollution chain diagram
may be constructed. An example of such a diagram for lead is shown in
Figure 2. The various pathways shown can be analyzed mathematically, if
sufficient data are available, to yield an integrated exposure to humans.
The system does not address the question of what exposure levels repre-
sent "unreasonable risk." However, it is not entirely clear that such a
judgment can be made objectively. It is interesting to note that the
diagram neglects at least three paths:
• lead shot uptake by bottom-feeding birds, which
leads to lead poisoning;
• lead vapor in shooting galleries which leads to
high occupational exposures; and
• lead pipes, glasses.
(7) Structure/Toxicity Correlation. One approach considered in this
program was based on the hypothesis that there is a correlation between
chemical structure or substructure and toxicological effects. The screen-
ing and ranking procedure, as shown schematically in Figure 3, would
begin by identifying the principal substructures, functional groups, or
elements contained in the chemicals of interest that are frequently
associated with health or environmental hazards. A search would then be
made via CHEMLINE for chemical analogs containing the substructure of
interest. Sources of toxicity and related data, such as MEDLINE and
Chemical Abstracts, would next be consulted to identify the hazardous
characteristics associated with the structural analogs. Finally correla-
tion equations or algorithms would be developed relating substructures
in analogous compounds to various classes of toxicological hazard, and
to degrees of hazard within each class. The substructure correlations
would be used to classify and rank the chemicals of interest.
Reasoning by analogy is far from perfect, but certain physical and
chemical properties can be fairly well predicted from a knowledge of
structure, and in specific instances some toxic effects have been shown
to be related to structure. Conceptually, the structure/toxicity ap-
proach is appealing in that, if it could be implemented, it would provide
guidance on the types of toxicity to anticipate and hence on productive
areas for subsequent testing and research. Implementation, however,
depends upon the existence of toxicological data related to carcinogenic-
ity, mutagenicity, teratogenicity, central nervous system action, etc.,
for a wide range of chemicals of different structural characteristics.
If the data exist, they have certainly not been compiled in a readily
retrievable form. Services such as MEDLINE, TOXLINE, CAS, provide liter-
ature references, not data. The data are yet to be retrieved from the
literature, compiled, tabulated and correlated. Furthermore, a brief
review of the literature suggests that the range of data required to
develop meaningful structural correlations for even major toxicological
hazards is in fact not available.
15
-------
Commercialize
Chemicals
Surveillance
and Monitoring
Biologically
Active
Substructures
Chemicals
of Interes
Pre-
Commercialized
Chemicals
Production and Use
Statistics Filter
Substructure
Decomposition (CAS)
Chemical
Data
Base
(CAS)
Substructure
(Analog)
Search (CAS)
Ranking
Procedure
FIGURE 3
STRUCTURE/TOXICITY CORRELATION SYSTEMS
16
-------
Apart from the practical difficulties of establishing structure/toxicity
correlations for predictive purposes, there are also important conceptual
problems. These center around the normally observed concentration depen-
dence of toxic effects. For example, certain heavy metals, chlorinated
hydrocarbons, cyanides, sulfides, etc., have been associated with health
and environmental hazards of various kinds. Assessment of whether a
chemical with given structural characteristics is likely to give rise to
major environmental problems however, depends critically on how much of
the chemical is released and how it is distributed in the environment
following release. The structure/toxicity approach is deficient in not
taking into account production, use and emissions.
(8) Structure/Toxicity/Release. It would be possible to modify the
structure/toxicity correlations to incorporate release rates or some
surrogate thereof. However, the resultant rankings might be misleading
for substances of different persistences.
(9) Quantity/Toxicity. The relative hazard of a chemical with an
environmental regime is sometimes evaluated as some function of the product
of a quantity (or concentration) index and a toxicity index.9 The assump-
tion is that materials of moderate toxicity present in the environment
in large quantities are comparable in hazard potential with highly toxic
materials present in small quantities. This basic premise may not be
entirely credible. The quantity and toxicity indices may be defined very
crudely, with a severe loss in discrimination,10 or quite precisely, with
an attendant increase in data requirements.
(10) Environmental Levels/Levels of Concern. After assessing the
above concepts with respect to the desired system design criteria, we
developed the selection and ranking systems described in detail in the
remainder of this report. From input data likely to be readily available
or attainable, the system evaluates steady state concentrations of toxic
substances in the atmosphere, hydrosphere, and lithosphere. Ranking is
based on the ratio of eventual concentration levels in each environmental
compartment to the level of concern in that compartment.
9 Control of Hazardous Material Spills (R802610), Proceedings of the
1974 National Conference on Control of Hazardous Material Spills,
American Institute of Chemical Engineers, New York, August 1974,
p. 25. See also "Facility Test Plans," Vol. II, Destructing Chemical
Wastes in Commercial Scale Incinerators3 Environmental Protection
Agency, Office of Solid Waste Management Programs, Washington, D.C.,
March 1975.
10 Ibid.
17-
-------
In Table I, ten system concepts summarized above are rated with respect
to the systems evaluation criteria cited in section C.I. No system meets
all the criteria. Except for credibility, the criteria most difficult to
satisfy are short response time, minimal personnel requirements, minimal
input data requirements, and adequate handling of degradation products
and synergistic effects.
Credibility and reproducibility are most difficult to achieve in the
area of levels of concern. For any system that is aimed at preliminary
screening, especially, the "levels of concern" must be set in ways that
leave much to be desired from a scientific point of view. A scientific
definition such as,
the level of concern in any compartment is the
concentration in that compartment which would,
under chronic exposure cause damage y, per year
could be adopted, but serious problems of measurement would exist if y
is small. For example, if y is 1000 excess deaths per year (a level
which we would consider very high for purposes of pre-screening), the
concentration which causes an increase in mortality of five per million
per year would have to be established. This would require large-scale
experiments of long duration. It cannot be expected that adequate data
would be available for estimating the level of concern so defined for
many chemicals, much less for chemicals not yet in widespread use. These
problems exist whenever standards are set.
18
-------
TABLE I
RELATION OF SYSTEM CONCEPTS AND CRITERIA
CRITERIA
1. Objectivity
o
3
T3
O
0.
(/>
1 OL c
-------
II. OVERVIEW OF THE PROPOSED SYSTEM
A. INTRODUCTION
The system consists of two parallel branches, the results of which are
eventually merged and used for ranking, as shown in Figure 4. The gen-
eral concept is to estimate, on the one hand, the levels of pollutant
that will be encountered in the environment, and on the other, the levels
which can be tolerated in the environment. A comparison of these two
sets of numbers then leads to a preliminary ranking of the potential pol-
lutants in priority order. More refined rankings are possible and will
be discussed in Chapter V.
This simplified description immediately raises important questions, even
on the premise that environmental levels and tolerable levels can be
estimated. The primary questions are:
« How will the system predict the concentration of degra-
dation products?
• How will the system predict the toxicity (or tolerable
levels) of degradation products?
• How will the system allow for interactions and syner-
gistic effects?
As far as interactions and synergistic effects are concerned, we have no
way of handling them. Effects such as those of phosphates in water or
freons in the upper atmosphere, for example, are notoriously difficult
to predict, and it is unlikely that simple models could encompass enough
of these situations to be worthwhile. We, therefore, do not propose a
system which will attack this aspect of the problem. On the other hand,-
the system has the capability of handling degradation products, at least
in an approximate way by estimating the environmental levels on the
basis of the longest-lived degradation product and the tolerable levels
on the basis of the degradation product of highest putative toxicity.
B. EVENTUAL ENVIRONMENTAL LEVELS
The first branch of the system is shown in Figure 5. It has as its pur-
pose the computation of the levels of pollutant that will be attained in
the environment. We do not expect that the computed levels will corre-,
spond closely to what would be found in the environment after decades or
centuries, but we feel that estimation to within one or two orders of
magnitude will go a long way toward meeting the objectives. Closer
estimation would require detailed information on the modes of emission
of each product and on the distribution (geographical and temporal) of
20
-------
FIGURE 4
SCHEMATIC DIAGRAM OF THE PROPOSED SYSTEM
ESTIMATION OF
EVENTUAL
ENVIRONMENTAL
LEVELS BY
COMPARTMENT
BASED ON DATA ON:
RELEASE
PERSISTENCE
PHYSICAL PROPERTIES
AND COMPARTMENT DATA
ESTIMATION OF
LEVELS OF
CONCERN BY
COMPARTMENT
BASED ON DATA ON:
CHEMICAL COMPOSITION
STRUCTURE
PHYSICAL PROPERTIES
RANKINGS
21
-------
FIGURE 5
SCHEMATIC DIAGRAM OF THE FIRST BRANCH OF THE SYSTEM
PRODUCTION AND
USE DATA
SECONDARY
EMISSION
DATA
PHYSICO-
CHEMICAL
PROPERTIES
COMPOSITION
AND
STRUCTURE
DATA
ESTIMATED
RELEASES
PERSISTENCE
DATA
COMPARTMENT
DATA
EVENTUAL
ENVIRONMENTAL
LEVELS BY
COMPARTMENT
22
-------
these emissions, and it is unlikely that these would be known within
orders of magnitude when a product is in the early stages of commercial-
ization.
This branch of the system will include several types of data inputs:
9 quantification of the total eventual industrial rate of
production and allocation of this production to modes of
use with three different ranges of emission;
e quantification of the emission of the product from non-
industrial sources (e.g., natural production by plants,
production as a result of chemical reactions of other
materials in the environment, unintended or by-product
emissions);
• quantification of the gross geographic distribution of
the emissions;
a estimation of the half-life for degradation of the
product into non-toxic final products, at least in four
different ranges, in water and air; and
e quantification of basic physico-chemical constants,
such as the solubility in water, the partition coeffi-
cient between fat and water, the vapor pressure, etc.
Based on these data and fixed data on regional water and air flows, the
system would compute the eventual levels in various compartments (air,
surface water, plants, animals).
C. LEVELS OF CONCERN
The second branch of the system is shown in Figure 6. It has the aim of
estimating the levels of the pollutant which are of concern. Again, we
do not expect that the computed levels will correspond closely with
toxicity data on any specific compound. We feel that, given the current
state of the art, estimation of levels of concern which are within two
or three orders of magnitude of those dictated by toxicological data
would be adequate. Moreover, as shown in Chapter IV, we believe that
this kind of accuracy can be achieved with relatively simple methods.
At the simple level which we propose, the input into this part of the
system would consist of information on the presence or absence of a num-
ber of functional groups in the compound in question.
23
-------
FIGURE 6
SCHEMATIC DIAGRAM OF THE SECOND BRANCH OF THE SYSTEM
DATA ON
COMPOSITION
OF
CHEMICAL
PHYSICOCHEMICAL
PROPERTIES
OF
CHEMICAL
COMPOSITION-TOXICITY
RELATIONSHIPS
LEVELS OF CONCERN
BY COMPARTMENT
24
-------
D. RANKING METHODS
The preliminary ranking would be carried out by comparing the levels of
concern with the eventual environmental levels. We currently envision
estimates of environmental levels on five primary compartments: air,
atmospheric moisture, surface water, ground moisture, and groundwater;
and two secondary compartments: vegetable matter and flesh. The levels
of concern could be developed on the basis of these compartments or in
terms of combinations of these (for example, human intake could be speci-
fied in terms of air breathed, water drunk, and flesh and vegetable
matter consumed).
For each compartment, the ranking system would calculate the ratio
of the eventual environmental level to the level of concern. The com-
pound would then be ranked on the basis of the highest of these ratios.
In view of the requirement that the resulting ranks be conservative, we
would envision any compound for which the largest ratio is of the order
of 10~3 to be classified as definitely of concern, assuming that the
predictors of levels of concern are set so as to be unbiased.
The compounds classified as being of concern would then be ranked by con-
sidering the actual ratios of eventual levels to levels of concern as
well as the time horizon available for action. This time horizon is
defined as the last time at which emissions could be discontinued without
exceeding a threshold level (for example, one one-thousandth of the level
of concern) in any compartment. The second ranking variable is important
because it sets a definite limit on how long we can afford to wait for
results from the research effort undertaken as a result of the ranking.
25
-------
III. EVENTUAL ENVIRONMENTAL LEVELS
A. INTRODUCTION
The estimation of the levels of a pollutant that will eventually be en-
countered in the environment can be conducted in a number of ways.
Ideally, the whole chain from release of the substance through dissolu-
tion, evaporation, sorption, transport, and degradation would be con-
sidered in detail; this would require voluminous data on physicochemical
characteristics of each compound to be considered and extensive computa-
tions. At the other extreme, simple projections or guesses could be
provided, but these would fail to meet a basic requirement of objectivity.
B. ESTIMATION OF EVENTUAL ENVIRONMENTAL LEVELS
We propose that a simple multiple compartment model be used. Schemati-
cally, this model is shown in Figure 7. For present purposes we do not
provide a compartment for the upper atmosphere; this is neither by
oversight nor because of the difficulty in dealing with this compartment:
it reflects the recognition that most of the damage in this compartment
is due to compounds of low molecular weight and reasonable chains of
degradation products would have to be predicted and their interactions
with the higher atmosphere would have to be estimated before a reasonable
assessment of potential damage could be made.
Compartments for animals and vegetable matter are not shown in Figure 7.
We propose to compute the eventual concentration in these by applying
the octanol/water partition coefficient11 to the average fat content as
a reasonable approximation.
The quantitative model assumes that each of the compartments behaves as a
completely mixed, flow reactor, with flows between compartments. In most
cases, the flows are bulk flows determined by water flows and concentra-
tion in the compartment in which the flow originates. In the case of
flow between ground or water and air, diffusional and convective pro-
cesses are invloved as well as bulk flows (through rainfall). On this
basis, as discussed more fully in Appendix I, the conditions in compart-
ment "X" at a time "t + At" can be related to conditions in the system
at time "t" by the equation.
11 See Neely, W.B., D.R. Branson and G.E. Blau, "Partition Coefficient to
Measure Bioconcentration Potential of Organic Chemicals in Fish,"
Environmental Science and Technology 8, 1974, pp. 1113-1115.
26
-------
(t +At) = WxCx(t) + At[Px(t)
(F + r ) C (t) - K W C (t)
y y+x y+x' y x x xv '
- C (t) (F + r X1 ti\
xv ' y x x->y x+y)J (1)
where W is the weight of compartment x, in kg,
/\
C is the concentration of pollutant in compartment
x x, in kg/kg,
•
F ^ is the bulk flow of pollutant solution from com-
x~*y partment x to compartment y in kg/yr,
r is the diffusional flow of pollutant from compart-
x^y ment x to compartment y in kg/yr/unit concentration,
•
P is the pollutant emission rate into compartment x,
x in kg/yr,
K is the first order reaction constant in compartment
x x in yr-1
The eventual environmental levels, C (06),in the various compartments
can then be expressed as a rector by the equation
so that the eventual environmental levels can be computed from the
emission rates by inverting the matrix M whose elements are:
mxx ' -kxWx '
mxx
(2)
27
-------
FIGURE 7
SCHEMATIC DIAGRAM OF INTERCOMPARTMENT FLOWS
OF EMITTED CHEMICAL
(man and biosphere omitted)
EMITTED CHEMICAL
BULK FLOW
DIFFUSIONAL OR CONVECTIVE FLOW
28
-------
The data required for the computation are:
(1) emission rates
• total emissions
t allocation to compartments
(2) compartment data
9 size
e f1ows
(3) physicochemical data
• half life in water and air
• air/water partition coefficients
• octanol/water partition coefficients
(4) transport rates for water/air and ground/air
C. ESTIMATION OF EMISSION RATES
1. Total Emissions
At a naive level, the rate at which a substance is produced could be
used as an estimator of the rate of emission of that substance into the
environment, either directly or by applying a proportionality constant
to it. Such a simple procedure would fail to recognize the fact that
some chemicals are produced primarily for conversion to other chemicals,
in which case only a small fraction will be emitted in the original form;
whereas others are used in ways that are, directly or indirectly, dis-
persive. For example, phosgene is produced primarily as an intermediate
in chemical synthesis, whereas freon is (or was) produced mainly for use
in aerosols (which are directly dispersive) and for refrigeration equip-
ment (from which it is dispersed by leaks or eventual destruction of the
equipment).
A variety of ways for estimating emission rates were considered. We
believe that for present purposes it is sufficient to allocate total
production into three ranges of usage:
(a) Low emission uses, comprising use as chemical inter-
mediates in th,e same or proximal plants. We estimate
that in this type of use emissions would not exceed
5% of production. The use of a factor of 3% for
purposes of estimation would lead, at most, to a 50%
underestimate of emissions in extreme cases and would
be highly conservative for most chemicals in this use
class.
29
-------
(b) Intermediate emission uses, comprising uses involving
substantial handling and transportation prior to trans-
formation of the compound. We estimate that in this
type of use emissions would range from 5% to at most
50% of production. For estimation, a factor of 30%
would be used; it would be about as conservative as
that for low emission uses.
(c) High emission uses, comprising uses in which the com-
pound is not modified chemically. For these, long-term
production rate will make up for losses in use, so the
uses may be considered to be completely dispersive.
A factor of 100% will be applied to production to es-
timate emissions.
In the absence of specific allocation of the product to these three emis-
sion classes, the system should allocate all production to the high
emission class.
In addition to these three categories of use, total "emissions" from
sources other than intended production will have to be estimated. This
category includes "emissions" from unintended production, such as pro-
duction as a by-product, and from production occurring in the environ-
ment through natural processes or as the result of reactions between
other emitted chemicals. These have to be estimated and provided to the
system.
2. Allocation of Emissions
The emissions in each category, computed as outlined above, must next be
allocated to the various compartments. We currently foresee only three
primary compartments into which emissions would be considered:
• air
• ground
• surface water.
The input would, therefore, have to specify the fraction of compound
from each class of emission that goes into each of these three compart-
ments.
Multi-regional compartment models could be considered. At present, the
development of compartment data on a regional basis is a definite bottle-
neck to such a refinement because data on inter-regional flows are meager.
The allocation of emissions to regions could be handled by requiring
30
-------
input of estimated distribution or, in the absence of specific input, by
allocating the emissions to regions on the following basis:
(a) For emissions from the low emission rate class, the
allocation to regions should be on the basis of con-
centration of industrial production.
(b) For emissions from the high emission rate class, the
allocation should be on the basis of population.
(c) For emissions from the intermediate emission rate
class, the allocation should be on the average of
the preceding two.
D. ESTIMATION OF PHYSICOCHEMICAL PROPERTIES
The methodology for computing eventual environmental levels involves the
use of a number of physicochemical properties. Principally, we are con-
cerned with
(1) Effective first-order reaction constants in the
various compartments;
(2) Air/water partition coefficients
(3) Octancl/water partition coefficients.
These properties of the compounds will have to be entered into the sys-
tem or computed.
1. Effective First-Order Reaction Constants
The main purpose of the system is to rank chemicals in terms of the
damage that their release may cause to the environment. Hence, a com-
pound that disappears rapidly from the environment is still of concern
if its degradation products persist and lead to damaging environmental
levels. For this reason, the system requires information on the
effective rather than actual reaction constants. Probably the easiest
framework in which to think of these parameters is to discuss them in
terms of the time at which one-half of the compound has degraded to
basically nontoxic materials or materials that are found in nature in
substantial concentrations (H20, C02, NaCl). The half-life can then be
converted to an effective reaction constant by assuming first-order
chemical kinetics (that is, the reaction constant would be 0.693 divided
by the half life.)
31
-------
We believe that a crude classification of the half-lives will be ade-
quate for our purposes. We suggest the following classification to
be provided as input:
(a) extremely reactive-- compounds with half-lives of up
to one week;
(b) highly reactive—compounds with half-lives between
one week and one month;
(c) moderately reactive—compounds with half-lives
between one month and one year;
(d) persistent--compounds with haIf-lives between one
year and ten years;
(e) inert—compounds with half-lives in excess of ten
years;
In the absence of better data or estimates, the longest half-life of the
range would be used to determine the effective reaction constant to be
used,and for the last group, an effective half-life of 100 years would
be assumed.
The system should, of course, be capable of taking more accurate data
if this is available for any or all of the compartments.
As a step for future improvement, we recommend the development of methods
for predicting effective reaction constants from the kinds of data that
are generally available at the time that OTS wishes to evaluate the
potential future damage. We believe that methods could be found that
lead to predictions of half-lives within an order of magnitude, and these
could be used to increase the objectivity of the process and reduce the
burden of input preparation.
2. Air/Water Partition Coefficients
Air/water partition coefficients are needed in order to compute trans-
port rates and equilibria at air/water interfaces. Unless better in-
formation is available on a substance, these partition coefficients will
be estimated by the ratio of the vapor pressure of the substance to the
water solubility; a temperature of about 15°C should be used for
evaluating these properties.
3. Octanol/water Partition Coefficients
The eventual concentration in flesh is an important parameter in itself
and is needed to estimate human consumption. We propose to estimate this
eventual concentration by considering flesh as a mixture consisting of
50% water at the concentration in streams and 50% protein and fat at a con-
centration equal to the stream water concentration times the octanol/water
partitioning coefficient. For vegetable matter, the same procedure would
be used, but the effective fat would be assumed to be 10% of the weight.
32
-------
In the absence of specific data on partitioning coefficients, the methods
advocated by Leo and co-workers should be used.12 Primarily, the method
would involve the computation of the logarithm of the partitioning
coefficient from additive contributions of individual substructures of
the molecule of interest. This computational method would lend itself
to automation, but manual operation appears preferable at least initially
in view of the possible departures from additivity discussed by Leo and
co-workers. Automation could be pursued at a subsequent stage when the
usefulness of the system has been established.
E. TRANSPORT RATES BY DIFFUSION
1. Between Liquid and Vapor Phases
We have chosen not to rely too heavily on transport equations derived
from aerodynamics, because they may be unreliable in predicting the
transport from water surfaces or ground moisture to air and vice versa.
As an alternative, we have opted to depend, on estimates based on the
rate of evaporation from water bodies and ground moisture, using the
functional form of the aerodynamic formulations for extrapolation.
On this basis, the transport between surface water (sw) and the atmos-
phere (a) can be expressed as
T = r C
sw+a sw-*a sw (1)
T = r C
a+sw a-*sw a (2)
with
rsw+a = 1000 • A • M- 1-7^1 .. v /. ^
Vsw = 100° 'A 'M j^Ws '7 (4)
°w
12 Leo, A., Hansch, and D. Elkins, "Partition Coefficients and Their
Uses," Chemical Reviews, 1971, Vol. 71, No. 6, pp. 525-616
33
-------
where
T ^ is the rate of transport from x to y in kg/yr
2
A is the total surface area in m
M is the molecular weight of the contaminant
D is the diffusivity of the contaminant in air in m2/sec
ca
D is the diffusivity of water vapor in air in m2/sec
Wo
V is the vapor pressure of the contaminant in atmospheres
C^ is the concentration of the contaminant in surface water
SW (kg/kg)
C is the concentration of the contaminant in air (kg/kg)
a
Based on an effective surface area for lakes (1) and streams (s) of
1.4 x 1011 m2 and 0.25 x 10um2, respectively, we have
• 7.5 x 1015 . M . ca /3 V / (5)
r a = 0.28 x 1015 M • / Dca V/3 V (6)
\ "n— i P
* wa /
r..n = 0.98 x 10's • M YDcaY/3 (7)
• M ./5c.V
\D I
> wa'
r_c = 0.18 x 1015
a"*~s
wa
For the transport from ground moisture (gm) to air we have
,^,m
a-*gm
^ wa
= 2.7 x 1016 M/Dca
The method which we recommend for estimating diffusivities is discussed
in Appendix III.
34
-------
Because of the very large surface area available, the transport between
air (a) and atmospheric (am) moisture is rapid and the diffusional pro-
cesses may be ignored. For our purpose we will use
(12)
ra-»-am ~ x
2. Between Ground Moisture and Soil
We have not found any reliable correlation which permits the prediction
of soil adsorption from readily available data. The total surface area
of active ingredients is equally difficult to estimate. For these
reasons we recommend that the transport from ground moisture to soil be
estimated as proportional to the diffusivity of the contaminant in water,
D , the fraction of organic material in the soil, f , and the con-
centration in ground moisture, C :
•
T = 1 1 x lO1^ D f C
gm+soil ' cw o gm (13)
while the transport from soil to ground moisture can be estimated as di-
rectly proportional to the diffusivity of the contaminant in water, in-
versely proportional to the octanol/water partitioning coefficient,
P . and the concentration is soil C :
*
Cp (14)
s
The method for estimating D is discussed in Appendix III.
CW
These equations imply that at equilibrium the ratio of concentration in
the organic fraction of soil to concentration in ground moisture is
equal to the octanol/water partition coefficient
F. COMPARTMENT DATA
Estimates of the various compartments, in terms of size and flows, are
shown in Table II. For the most part, these are based on data from the
U.S. Geological Survey. In the case, of air, we assumed a column one
mile high and an average flow of two miles per hour through a 1000 mile
front.
35
-------
OJ
ov
COMPARTMENT
Air (1 mile, 2 mph)
Surface Water
Ground water
Ground Moisture
Atmospheric Moisture
Soil
TABLE II
U.S. COMPARTMENT DATA
SUB-COMPARTMENT
Effective
area (m2)
-
Lakes 1.4 x 1011
Streams 0.25 x 10n
Shallow
Deep
0-1 m 8 x 1011
1-5 m
5-10 m
10-15 m
15-30 m
30-50 m
00
0-1 m
1-5 m
5-10 m
10-15 m
15-30 m
30-50 m
INVENTORY
1015kg
16.2
18.8
0.05
63.7
63.7
0.6
0.4
0.2
0.2
0.2
0.2
0.18
15.2
60.9
76.1
76.1
228.5
304.4
ANNUAL
FLOW
1015kg
710
0.19
1.86
0.31
0.006
3.1
4.8*
TIME OF
RESIDENCE
yr
0.002
100
0.3
200
0.04
Net of short-term reevaporation of approximately 1.2 x 1015kg per year.
-------
We have not been able to locate adequate data for individual basins
or regions except for selected flows (such as water run-off). For the
contiguous States, estimates of the cross-flows of water between com-
parments, are given in Tables III a, b, c, d. The first two tables
differ in the short-circuiting from ground moisture to ground water,
the last two differ from the first two by allowing back flow from
deeper to shallower soil layers. These differences are illustrated in
Figures 8a, b, c, d which show schematically the flows in the four
cases.
The flows from air particulates are not specified in the tables. These
flows will depend largely on the size of particulates. We suggest that
flows from air particulates to ground, lakes, and streams be specified
as part of the input data.
37
-------
00
FLOW
FROM:
l.AIR MOISTURE
2.LAKES
3.STREAMS
H.SOIL MOISTURE Q-1M
5.SOIL MOISTURE 1-5/f
6.SOIL MOISTURE 5-10M
7. SOIL MOISTURE 10-15'W
fi.SOIL MOISTURE 15-30^
9.SOIL MOISTURE 30-SOW
IQ.GROUtW VATER(SHALLOW)
11.GROUND VATER(DEEP)
12.OCEAN
TABLE Ilia
WATER FLOWS BETWEEN COMPARTMENTS
(in 1015kg/yr)
1
.000
.170
.030
2. 800
.000
.000
.000
.000
.000
.000
.000
1.800
2
1.800
.000
.160
.100
.050
. 020
.010
.005
.005
,000
.000
.000
3
.100
1.690
. 000
.100
.050
. 020
.010
.005
.005
.000
. 000
.000
CASE ]
4
2.900
.080
.120
.000
.000
.000
.000
.000
.000
.000
.000
.000
5
.000
.050
.050
.100
.000
.000
.000
.000
.000
.000
.000
.000
6
.000
.025
.015
.000
.100
.000
.000
.000
.000
.000
.000
.000
7
.000
.015
.005
.000
.000
.100
. 000
.000
.000
.000
.000
.000
8
.000
.010
. 000
.000
.000
.000
.100
.000
.000
.000
.000
.000
9
.000
.010
.000
.000
.000
.000
.000
.100
.000
.000
.000
. 000
10
.000
.100
.100
.000
.000
.000
.000
.000
.100
.000
.000
.000
11
.000
.000
.000
.000
.000
.000
.000
.000
.000
.100
.000
.000
12
.000
.000
1.500
.000
.000
.000
.000
.000
.000
.200
.100
.000
-------
TABLE Ilib
WATER FLOWS BETWEEN COMPARTMENTS
(in 1015kg/yr)
GO
IO
CASE 2
FLOW
FROM:
l.AIR MOISTURE
2.LAKES
3.STREAMS
H.SOIL MOISTURE 0-1W
5. SOIL MOISTURE 1-5/f
ft. SOIL MOISTURE 5-lOAf
7 .SOIL MOISTURE 10-157*
B.SOIL MOISTURE 15-30M
9.SOIL MOISTURE 30-507-f
in.GROUND WATER{SHALLOV)
11.GROUND VATER(DEEP)
12.OCEAN
10
11
12
.000
.170
.030
2. 800
.000
.000
.000
.000
.000
.000
. 000
1.800
1.800
.000
.160
.100
.050
.020
.010
.005
.005
. 000
.000
.000
.100
1.690
.000
.100
.050
.020
.010
.005
.005
.000
. 000
. 000
2.900
. 080
.120
.000
.000
.000
.000
.000
.000
.000
.000
.000
. 000
.050
. 050
.100
. 000
. 000
.000
. 000
.000
. 000
. 000
.000
.000
. 025
.015
. 000
. 080
. 000
. 000
. 000
. 000
. 000
.000
. 000
.000
.015
.005
. 000
.000
. 050
. 000
. 000
. 000
.000
. 000
.000
.000
.010
.000
.000
. 000
. 000
. 020
.000
. 000
.000
. 000
.000
. 000
.010
. 000
.000
. 000
. 000
.000
. 010
.000
. 000
. 000
. 000
. 000
.100
.100
. 000
.020
.020
.010
.000
.000
.000
. 000
.000
. 000
.000
.000
.000
. 000
.010
. 020
. 010
.010
. 250
.000
. 000
.000
.000
1.500
.000
. 000
. 000
.000
.000
. 000
.000
. 300
.000
-------
TABLE lilc
WATER FLOWS BETWEEN COMPARTMENTS
(in 1015kg/yr)
CASE 3
FLOW \
FROM: \TO:
l.AIR MOISTURE
2. LAKES
3. STREAMS
H.SOIL MOISTURE 0-lW
5. SOIL MOISTURE 1-5/4
&.SOIL MOISTURE 5-10W
7 .SOIL MOISTURE 10-15W
B.SOIL MOISTURE 15-30W
9. SOIL MOISTURE 30 -SOW
1 0 . GROUND VA TER ( SHA L LOU )
1 1 . GROUND VA TER ( DEEP )
12. OCEAV
1
.000
.170
. 030
2.800
.000
.000
. 000
. 000
.000
.000
.000
1.800
2
l.ROO
.000
. 160
.100
.050
.020
.010
.005
.005
.000
.000
.000
3
.100
1.P90
.000
.100
.050
.020
.010
.005
. 005
. 000
.000
.000
4
2.900
. 080
.120
. 000
.010
. 000
.000
.000
.000
.000
.000
.000
5
.000
.050
.050
.110
.000
.010
.000
.000
.000
.000
.000
.000
R
. 000
.025
.015
.000
.110
.000
.010
.000
.000
. 000
.000
.000
7
.000
. 015
.005
.000
.000
.110
.000
.010
.000
.000
.000
.000
8
. 000
.010
.000
.000
.000
. 000
.110
. 000
.010
.000
. 000
.000
9
. 000
.010
. 000
. 000
.000
. 000
. 000
.110
.000
. 000
. 000
.000
10
.000
.100
.100
. 000
.000
. 000
. 000
. 000
.100
.000
. 000
.000
11
.000
. 000
. 000
. 000
.000
.000
.000
. 000
. 000
. 100
. 000
. 000
12
.000
;ooo
1.500
. 000
.000
.000
.000
.000
.000
.200
.100
.000
-------
1.
2.
3.
5.
6.
7.
8.
9.
10.
11.
12.
FLOW
FROM:
AIR MOISTURE
LAKES
STREAMS
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
MOISTURE
MOISTURE
SOIL
SOIL
GROUND
GROIIHD
OCEAH
0-lAf
1-5M
5-10W
10-15/f
15-30W
30-SOW
WATER(SHALLOW)
VATER(VEEP)
TABLE 11 Id
WATER FLOWS BETWEEN COMPARTMENTS
(in 1015kg/yr)
CASE 4
2 3 1 5 R
10
11
12
. 000
.170
.030
2.800
.000
.000
.000
.000
.000
.000
.000
1.800
1.800
.000
.160
.100
.050
.020
.010
.005
.005
.000
.000
.000
.100
1.690
. 000
.100
.050
.020
.010
.005
.005
.000
.ono
. 000
2.900
. 080
.120
.000
.010
.000
.000
.000
.000
.000
.000
. 000
. 000
.050
.050
.110
. 000
.010
. 000
.000
.000
.000
.000
.000
. 000
.025
.015
. 000
.090
.000
.005
.000
.000
.000
.000
.000
.000
.015
.005
. 000
. 000
.045
. 000
.002
.000
. 000
.000
. 000
.000
.010
.000
. 000
.000
. 000
.012
.000
.000
.000
.000
. 000
. 000
.010
.000
.000
.000
.000
. 000
.005
.000
.000
.000
.000
. 000
.100
.100
.000
.020
.020
. 010
. 000
.000
. 000
.000
.000
. 000
. 000
.000
. 000
.000
.020
. 020
.005
.005
.250
.000
.000
. 000
. 000
1.500
.000
.coo
.000
.000
.000
.000
.000
.300
.000
-------
FIGURE 8a - SCHEMATIC DIAGRAM OF FLOWS.FOR "CASE 1"
ATMOSPHERE
AND
HYDROSPHERE
SOIL
MOISTURE
(1ST LAYER)
SOIL
(1ST LAYER)
SOIL
MOISTURE
(2ND LAYER)
SOIL
MOISTURE
(LAST LAYER)
SOIL .
(2ND LAYER)
SOIL
(LAST LAYER)
GROUND
WATER
42
-------
FIGURE 8b- SCHEMATIC DIAGRAM OF FLOWS FOR"CASE 2"
ATMOSPHERE
AND
HYDROSPHERE
SOIL
MOISTURE
(1ST LAYER)
SOIL
(1ST LAYER)
SOIL
MOISTURE
(2ND LAYER)
SOIL
MOISTURE
(LAST LAYER)
SOIL
(2ND LAYER)
SOIL
(LAST LAYER)
GROUND
WATER
43
-------
FIGURE 8c - SCHEMATIC DIAGRAM OF FLOWS FOR "CASE 3"
ATMOSPHERE
AND
HYDROSPHERE
SOIL
MOISTURE
(1ST LAYER)
SOIL
(1ST LAYER)
SOIL
MOISTURE
(2ND LAYER)
SOIL
MOISTURE
(LAST LAYER)
SOIL
(2ND LAYER)
SOIL
(LAST LAYER)
GROUND
WATER
44
-------
FIGURE 8d- SCHEMATIC DIAGRAM OF FLOWS FOR "CASE 4"
ATMOSPHERE
AND
HYDROSPHERE
SOIL
MOISTURE
(1ST LAYER)
SOIL
MOISTURE-
(2ND LAYER)
SOIL
MOISTURE
(LAST LAYER)
SOIL
(1ST LAYER)
SOIL
(2ND LAYER)
SOIL
(LAST LAYER)
GROUND
WATER
45
-------
IV. LEVELS OF CONCERN
The second "branch" of the proposed system must calculate the levels of
a contaminant that are of concern. In principle, the system can cope
with a variety of levels of concern. For example, levels of concern
for surface water, air,and flesh might be specified as concentrations,
and a level of concern for total human intake (specified in terms of
human intake of water, air, and flesh) might be specified in terms of
the amount taken in. The various levels of concern need not be con-
sistent across compartments,13 as long as the disparities are known
and can be adjusted for in the ranking process.
Ideally, the levels of concern should oe set on the basis of potential
damage caused by chronic exposure. For example, the level of concern
could be defined as the concentration or intake of the compound which
causes an excess mortality of 10~9 per person year (one excess death
every five years at current population levels) due to chronic exposure.
Unfortunately such data are unavailable (and unobtainable).
Alternatives which readily present themselves, such as use of acute LD50
or chronic levels which produce substantial excess in mortality, are far
from the ideal because dose-response curves of different compounds may
intersect, as shown in Figure 9. The main interest is in preventing
even low levels of excess mortality and the ranking should, as far as
possible, reflect this interest; if the intersection represented in
Figure 8 is at 50% mortality, the compounds should be assigned the same
level of concern.
On a practical level it was necessary to establish whether any kind of
toxicity levels could be predicted to within two or three orders of
magnitude on the basis of simple information such as chemical formula.
Our main emphasis has been in resolving this issue as a prelude for more
refined methods based on more definitive data. Appendix II discusses
the methods by which we attempted to develop crude predictors of toxicity.
Because of the tentative nature of the work, we did not look for the
most comprehensive sets of data on the most relevant levels of concern.
Most of our work was done using the "allowable levels" of air concentra-
tion for compounds on the "toxic substances list." In summary, we found
that crude methods exist which could "narrow" the spread of "allowable
levels" for a given calculated level to a ratio (of highest to lowest
"allowable level") of 101* (Figure TO). The geometric mean between these
13Though obviously they should be consistent across compounds being
considered.
46
-------
FIGURE 9
SCHEMATIC REPRESENTATION OF INTERSECTING
DOSE-RESPONSE CURVES
COMPOUND
A
LU
V)
•z.
o
Q_
W
LU
COMPOUND
B
DOSE
47
-------
10s
10"
gg 101
_J _J
<
< <
31
O
I
10 -
FIGURE 10
RANGE OF ACTUAL MAXIMUM ALLOWABLE CONCENTRATION
RELATED TO CALCULATED VALUES
8 10 12 14 16 18 20 22 24 26 28 30 32 34 36
CALCULATED MAC
48
-------
two limits will generally be within a factor of 100 of the "allowable
level," and in the majority of cases it should do considerably better.
The use of formal statistical methods would reduce the dispersion between
empirical "allowable levels" at a given calculated level, but it was not
considered effective to develop better estimators until the conceptual
design of the system had been reviewed.
Based on these findings, it appears feasible to develop relatively simple
methods of predicting levels of concern from data which are generally
available when a product goes into commercial production. In refining
the crude methods developed in this phase, several alternatives are
possible.
One alternative is to rely on the same form of relationship that we used
in our crude work, namely
Level of Concern = f{^-} (15)
where the L is the sum of contributions, £-j, of the different functional
groups of the compound. The refinement would arise from improved values
of the functional group contributions which would give better predictions
than the crude values we assigned. This methodology could be used for
levels of concern in various compartments if empirical data are available.
Another alternative is to use Equation (15) to predict the level of con-
cern for the basic members of each homologous series and use Hansch
approaches, which rely on octanol/water partition coefficients to extrap-
olate to other members of the series. Since the octanol/water partition
coefficient is used in the estimation of eventual environmental level,
no additional data requirements are needed in the operation of the system.
Other approaches, based on extrapolating from the substance of known
level of concern whose chemical structure most resembles that of the
compound in question could also be developed.
The ultimate choice between approaches should be based on three consider-
ations:
(1) The cost of developing and operating the system;
(2) The dispersion of empirical values about the
predicted values; and
(3) The relationship between the predicted levels
of concern for a compound and the sum of the
predicted levels for degradation products.
On this basis we believe that the development of improved values for
functional group contributions (shown in Equation 15) is the best choice.
-------
V. RANKINGS
A. SUBSTANCES OF CONCERN
The levels of concern and the eventual environmental levels will be com-
pared to classify substances as "substances of concern" and others.
The classification will be in terms of the ratio of the eventual environ-
mental level (or function of eventual environmental levels) to the
corresponding level of concern. These ratios will be computed for each
level of concern (e.g., for each compartment, or for human consumption,
etc.) which has been specified or estimated. A number of classification
methods could be considered, but we feel that for this initial classifi-
cation it is sufficient to use the highest of the ratios. For example,
if a substance has ratios of 10 5 for water, 10 3 for air, and 10 ** for
human consumption, the classification would be based on the ratio for
air (10~3) which is the highest.
The threshold which would divide the "substances of concern" from others
can be set at a very low level to ensure, as far as possible, that no
potentially harmful compounds are missed. Based on the uncertainties
inherent in the calculations, we believe that the threshold ratio should
be no higher than 10 2. Lower ratios could be used and would be justi-
fiable if a ratio of the order of 10 2 to 10~3 results in a substantial
fraction of compounds being ranked as not of concern.
B. PRIORITY ORDERING
Compounds which are classified as substances of concern can be ordered
into priority ranks in a variety of ways which require little or no
additional effort other than computation. From the point of view that
the output of this system is intended to guide research activities, it
would appear fruitful to use the framework developed for estimating the
eventual environmental levels to estimate the time scale available for
such research activities.
To tie in closely with the potential for meaningful action, the system
could compute the "action date," the latest date at which the emission
of the compound could stop without driving the eventual environmental
levels above a threshold related to the levels of concern. The action
date could then be used alone, or in conjunction with the ratio of
eventual level to level of concern, to create a priority list.
50
-------
VI. SOME EXAMPLES
A. INTRODUCTION
In order to examine the behavior of the proposed system we are present-
ing some examples. The substances we have chosen for illustrative
purposes are: benzene, bis (2-chloroisopropyl) ether, chlorodifluoro-
methane, methyl chloroform, and trichloromethane. The level of concern
desired as discussed in Chapter IV and Appendix II will be used for all
compartments.
For all these cases we have assumed that particulates have a half life
of about eight months in the atmosphere. Also, we have assumed that the
six soil layers have organic fractions of 10%, 5%, 3%, 3%, 3%, and 3%,
respectively.
Sensitivity analyses are presented for the case of benzene.
B. DATA ON THE COMPOUNDS
The necessary data was collected using the forms shown in Tables IVa
and b. The information sheets for the five compounds are reproduced in
Appendix IV. Most of the data were obtained from reports published by
the Office of Toxic Substances. In general, production data may be
obtained from the U.S. Tariff Commission reports or from the SRI direc-
tory of manufactured chemicals. The user and percentage of distribution
to each use would ordinarily have to be provided by the manufacturer.
Some judgment is required with respect to whether a use should be classi-
fied as low, intermediate or high emission. Emissions from sources
other than manufacturer and use of the manufactured chemical are diffi-
cult to identify. In the case of benzene, for example, the major
emission source is motor vehicle exhaust. Unless this were known from
other studies, as it was for this example, it might have been overlooked.
The probable distribution of emissions to air, ground and surface Water
was estimated on the basis of the uses and volatility of the compounds.
This distribution will ordinarily involve some judgment, and a range of
possible distributions should be used to check the sensitivity of the
results to these estimates.
Vapor pressure at room temperature, solubility in water, and the octanol/
water partition coefficient were obtained from a variety of published and
unpublished sources. First-order reaction constants were estimated if
enough information was available; otherwise judgment was used to classify
the compounds by reactivity and the method outlined in Chapter III was
used.
51
-------
TABLE IVa. DATA SHEET NO. 1
Chemical:
Emission Data:
Production Rate kg/yr.
Natural Production kg/yr.
Uses % of Prod. % Emission Emission Rate kg/yr.
I. Low Emission
• Production 100% 3%
3%
3%
3%
II. Intermediate Emission
30%
30%
30%
30%
III. High Emission
100%
100%
100%
100%
IV. Natural Sources 100%
Total
Distribution of Emissions
To Air kg/yr
To Air Particulates kg/yr
To Lakes kg/yr
To Streams kg/yr
To Ground kg/yr
52
-------
TABLE IVb. DATA SHEET NO. 2
en
CO
Chemical:
Basic Data:
Molecular Weight
Molar Refractivity
Vapor Pressure*
Water Solubility*
Octanol/Water Partition Coeff.*
* at 20°C
atm
gm/gm
Reaction Constants
Compartment Type
Air
Particulate
Air Moisture
Water
Adsorbed to Soil
Reactivity
Extreme High Moderate Persistent Inert Half Life Reaction Rate
-------
C. RESULTS
The eventual environmental levels are shown in Tables V a, b, c, d, e.
These tables give, for each compound, the computed eventual environmental
levels in various compartments. For air and particulates, the concentra-
tion is in kilograms of compound per billion kilograms of dry air. For
air moisture it is in kilograms of compound per billion kilograms of
moisture. For lakes, streams, ground water, and soil moisture the con-
centration is in kilograms of compound per billion kilograms of water in
the corresponding compartment. For the soil compartments the concentra-
tion is per billion kilograms of soil in the compartment.
Each table shows four "cases," corresponding to the four matrices of
bulk flows presented in Tables Ilia, b, c, d. It is clear that the
differences in bulk flow affect the eventual levels only to a minor
degree in most compartments; the concentrations in deep ground water are
a factor of two higher when short-circuiting into deep ground water is
allowed (cases 2 and 4).
1. Benzene
The eventual environmental levels are about 2400 parts per billion in the
air (including vapor and parti oil ate) and about 6 ppb in air moisture
and surface waters. Using the method outlined in Chapter IV and
Appendix II, the MAC for benzene is 8 mg/m3 or 7000 ppb. Using a safety
factor of 100 to arrive at a level of concern we obtain 70 ppb. This
level is exceeded for the respirable air. The levels in meat (about 400
parts per billion) and vegetable matter (about 90 parts per billion) are
substantial.
2. Bis (2-chloroisopropyl) ether
Because of the low emission rate, bis (2-chloroisopropyl) ether is not
expected to reach one part per billion in any compartment. The MAC
obtained by the method of Chapter IV is about 500 parts per billion; if
we use a safety factor of 100, the level of concern would be 5 ppb, much
higher than the estimated environmental levels even if the octanol /water
partition coefficient is applied.
-3. Chlorodifluioromethane
For this compound the MAC is about 3000 ppb, so the level of concern is
30 ppb. The eventual concentration in air is 24 ppb, quite close to the
level of concern, hence, the geographic distribution of emission would
be important.
54
-------
TABLE Va: STEADY-STATE CONCENTRATIONS OF BENZENE
MOL. J*T.:78
ML. VOL.:26.2
VAP. PRESS.:Q.I
HATER SOL. :0.0018
OCT/WAT PART.-A35
BULK RATE FLOU FROM PART. TO LAKES :0.01
STREAMS:Q.O\
SOIL :0.98
en
en
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS,
SOIL MOISTURE 0-1.M
SOIL MOISTURE l-5tf
SOIL MOISTURE 5-10W
SOIL MOISTURE 10-15W
SOIL MOISTURE IS~>3QM
SOIL MOISTURE 30-50M -
GROUND »ATER(SHALLOd)
GROUND UATER(DEEP)
OCEAN
SOIL 0-lAf
SOIL 1-5.V
SOIL 5-10W
SOIL 10-15M
S0/Z/ 15-30AJ
SOIL 30"SOW
ASSUMED VALUES
REACTION
CONSTANT
0.01380
0.01380
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06931
.0.03466
0.02079
0.02079
0.02079
0.02079
EMISSION
RATE(E6)
520.00000
0.00000
48.80000
8.00000
2.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
F-ORG
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.1.0000
0.05000
0.03000
0.03000
0.03000
0.03000
STEADY -STATE CONCENTRATIONS (PPB)
CASK 1
2395. 672
6.341
22.990
6.510
6.406
6.257
5.JI?
4.548
3.912
2.6^5
1.787
0.312
0.007
1.837
0.062
0.050
O.OU5
0.038
, 0.026
0.018
CASE 2
2395.463
6.341
22.990
6.510
6.406
6.257
5.117
4.472
3.'553
1.504
0.834
0.326
0.047
1.828
0.062
0.050
0.044
0.035
0.015
0.008
CASE 3
2395.634
6.341
22.990
6.510
6.406
6.257
5.139
4.555
3.881
2.678
1.834
0.313
O.OO7
1.837
0.062
0.051
0.045
0.038
0.026
0.018
CASE 4
2395.411
6.341
22.990
6.510
6.406
6.256
5.136
4.491
3.467
1.293
0.785
0.326
0.053
1.828
0.062
0.051
0.044
0.034
0.013
0.008
-------
TABLE Vb: STEADY-STATE CONCENTRATIONS OF BIS-(2-CHLOROISOPROPfL) ETHER
MOL. ivT.:!71
ML. I/OL.:41.3
VAP. PRESS.:0.001
MTER SOL. iQ.0017
OCT/tfAT P/J/tT.:5.899999976
BULK RATE FLOU FROM PART. TO LAKES
:0.01
STREAMS :0'.0*.
SOIL :0.98
01
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE 0-1A/
SOIL MOISTURE 1-5M
SOIL MOISTURE 5-10A/
SOIL MOISTURE 10-15/4
SOIL MOISTURE 15-3QM'
SOIL MOISTURE 30-50M
GflOlWZ? WATER(SHALLOW)
GROUND rfATER(DEEP)
OCEAN
SOIL 0-ltf
SOIL 1-5M
SOIL 5-10W
SOIL 10-15M
SOIL 15-30tf
SOIL 30-SOW
ASSUMED VALUES
REACTION
CONSTANT
0.00690
0.00690
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06930
0.06931
0.03466
0.02079
0,. 02079
0.02079
0.02079
EMISSION
RATE(Eb)
0.01000
0.00000
0.00000
0.15000
0.14000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
F-ORG
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
b.ooooo
0.00000
0.10000
0.05000
0.03000
0.03000
0.03000
0.03000
STEADY-STATE CONCENTRATIONS (PPS)
CASE 1
0.706
0.081 •
0.000
0.083
0,082
0.079
0.06&
0.060
0.052
0.038
0.027
0.004
0.000
0.023
0.001
0.001
0.000
0.000
0.000
0.000
CASE 2
0.705
0.080
0.000
0.083
0.082
0.0^9
0.066
0.059
0.048
0.022
0.013
0.004
0.001
0.023
0.001
0.001
0.000
0.000
0.000
0.000
CASE 3
0.706
0.081
0.000
0.083
0.082
o.o-'g
0.066
0.060
0.052
0.038
0.0217
0.004
0.000
0.023
0.001
0.001
. 0.000
0.000
0.000
0.000
CASE 4
0.7Q5
0.080
0.000
0.003
0.081
o.o^g
0.066
0.059
0.04"
0.019
0.012
0.004
0.001
0.023
0.-001
0.001
0.000
0.000
0.000
0.000
-------
TABLE Vc: STEADY-STATE CONCENTRATIONS OF CHLORODIFLUORO'-IETHANE
on
MOL. MT. :86.5
ML, VQL.:12.5
VAP. PRESS.:?.5
HATER SOL. :0.003
OCT/VAT PART.:12
6£/LK RATE t'LOU FROM PART. TO
COMPARTMENT
AIR--
AIR MOISTURE
PARTICULATE
LAK£S
STREAMS
SOIL MOISTURE 0-ltf
SOIL MOISTURE 1-5M
.SOIL MOISTURE S-lOAf
SOIL MOISTURE 10-15W
S0/L MOISTURE 15-SOW
SOIL MOISTURE 30-SOW
G/KWWD UATER(SHALWV)
GROUND VATER(DEEP)
OCEAN
SOIL 0-lAf
SOIL 1-5M
SOIL 5-10Af
SOIL IQtlSAf
50/L 15-3Q/V
SOIL 30-5QM
T. TO 6MSS :0.01
STREAMS :0i 01
SOIL :0.98
ASSUMED VALUES
REACTION
CONSTANT
0.03470
0.03U70
0.03470
0.00500
0.00500
0.00500
0.00500
0.00500
0.00500
0.00500
0.00500
0.00500
0.00500
O.'OOSOO
0.00070
0.00035
0.00021
0.00021
0.00021
0.0002V
EMISSION
RATE(ES)
13.20000
0.00000
0.00000
0.00000
0..00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
O.-OOOOO
0.00000
. 0.00000
0.00000
0.00000
F-ORG
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.10000
0.05000
0.03000
0.03000
0.03000
0.03000
STEADY-STATE CONCENTRATIONS (PPB)
CASE 1
23.481
0.001
0.000
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.000
0.000
. 0.001
0.00.1
0.000
0.000
0.000
0.000
0.000
CASE 2
23.481
0.001
0.000
0.001
0.001-
0.001
0 . 001
0.001
0.001
0.001
0.000
0.000
0.000
0.001
0.001
0.000
0.000
0.000
0.000
0.000
CASE 3
23.481
0.001
0.000
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.000
0.000
0.001
0.001
0.000
0.000
0.000
0.000
0.000
CASE 4
23.481
0.001
0.000
. 0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.000
0.000
0.000
0.001
0.001
0.000
0.000
0.000
0.000
0.000
-------
TABLE Vd: STEADY-STATE CONCENTRATIONS OF METHYL CHLOROFORM
MOL. vr.
ML. VOL.-.26.3
VAP. PRESS. :0.1.3
HATER SOL. :0.0044
OCT/XAT PART.:310
BULK RATE PLW FROM PART. TO LAKES :0.01
STREAMS:0.01
SOIL :0.98
en
oo
COMPARTMENT
AIR
AIR MOISTURE .
PANICULATE
LAKE'S
STREAMS
SOIL. MOISTURE Q-1M
SOIL MOISTURE 1-5/V
SOIL MOISTURE 5-10tf
SOIL MOISTURE 10-15M
SOIL MOISTURE 15-30Af
S0JZ, MOISTURE 30-5,OM
GffOW/0 UATER(SHALLOW)
GROUND VATER(DEEP)
OCEAN
SOIL 0-ltf .
SOIL 1-5M
£07£ 5-10W
SOIL 10-15M
SOIL 15-30W
SOIL 30-50M
ASSUMED VALUES
REACTION
CONSTANT
0.06930
0.06930
0.06930
0.00690
6.00690
0.00690
0.00690
0.00690
0.00690
0.00690
0.00690
0.00690
0.00690
0.00690
0.00069
0.00035
0.00021
0.00021
0.00021
0.00021
EMISSION
RATE(E6)
185.00000
0.00000
1.00000
1.. 00000
0.50000
2.00000
0.00000
0.00000
0. '00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
F-ORG
0.00000
0.00000
0.00000
0.00000
0.00000
o.ooooo
0.00000
0.00000
0.00000
Q. 00000
0.00000
0.00000
0.00000
0.00000
0.10000
0.05000
0.03000
0.03000
0.03000
0.03000
STEADY -STATE CONCENTRATIONS (PPB)
CASE 1
168.027
0.492
0.171
0.505
O.U97
0.485
0.4'*'»
O.U1U
0.380
0.280
0.197
0.162
0.030.
0.36U
0.465
O.H12
0.370
0.340
0.250
0.176
CASE 2
168.015
0.492
0.471
0.505
0.497
0.4R5
0.443
0.410
0.358
0.166
0.086
0.175
0.078
0.358
0.465
0.412
0.3T.6
0.320
0.148
0.077
CASE 3
168.025
0.492
0.471
0.505
0.497
0.485
0.444
0.414
0.376
0.279
0.201
0.163
0.030
0.364
0.465
0.413
0.3^0
0.336
0.249
0.180
CASE 4
lod.015
0.492
0.471
0.505
0.497
O.'iflS
0.444
0.410
0.349
0.141
0.0~9
0.175
0.081
0.359
0.465
0.412
0.3&7
0.312
0.126
0.071
-------
TABLE Ve: STEADY-STATE CONCENTRATIONS OF TRICHLOROFLUOROMETHANE
MOL. UT.-.1.37- •
ML. VOL.:21.9
'VAP. PRESS.:1
JATER SOL. :0..0011.
OCT/UAT PART. :340
BULK RATE FLOrf FROM PART.
TO LAKKS :0.01
STREAMS :0..0-\
SOIL :0.98
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE 0-1M
SOIL MOISTURE l-5tf
SOIL MOISTURE 5-1OM
SOIL MOISTURE 10-15M
50IL MOISTURE 15-3 a?/
SOIL MOISTURE 30-5CW
GflOWVZ? UATSR(SKALLOU)
GROUND WATER(DEEP)
OCEAN
SOIL 0-1,V.
SOIL 1-5M
SOIL 5-10W
SOIL 10-ISA/
SOIL 15-30M
SOIL 30-SOW
ASSUMED VALUES
REACTION
CONSTANT'
0.03470
0.03470
0.03U70
0.00510
O.OOS10
O.OOS10
0.00510
0.00510
O.OOS10
0.00510
0.00510
0.00510
0.00510
0.00510
0.00069
0.00035
0.00021
' 0.00021
0.00021
0.00021
EMISSION
RATE(E
-------
4. Methyl Chloroform
For methyl chloroform the level of concern is 70 ppb. The air concen-
tration is nearly 170 ppb, so the substance is of concern.
5. Trichlorofluoromethane
The level of concern of 30 ppb is exceeded in the air compartment by a
factor of 7, thus it is a substance of substantial concern.
D. RELATIVE PRIORITY
Based on the computation benzene and trichlorofluoromethane would be ranked
as of most concern since eventual levels exceed the levels of concern by
substantial margins. Methyl chloroform would also be of direct concern.
Chlorodifluoromethane would be of concern if its major source of emission
where concentrated. Bis (2-chloroisopropyl) ether would be considered of
little concern.
E. SENSITIVITY ANALYSIS
The sensitivity of the eventual environmental levels to the elements of
the matrix of bulk flows was assessed by using four matrices. As men-
tioned earlier, only the eventual levels in deep ground water are
affected to a major extent. It follows that the flows need not be spec-
ified with great accuracy except when there is concern about deep ground
water.
The sensitivity of the eventual environmental levels to other parameters
was explored for the case of benzene.
• Decreasing the emission into air from 520 to 320 million kilo-
grams per year (Table Via) caused a nearly proportional change
in eventual levels in air and water compartments. Particulate
levels were not affected.
e Increasing the reaction rate in all water compartments tenfold
(Table VIb) led to an 13% reduction in eventual levels in the
water compartments and in the vapor phase. There was no effect
on particulates.
60
-------
Table Via.
STEADY-STATE CONCENTRATIONS OF BENZENE
MOL. wT.:7&
MOL. VOL.-.26.2
VAP. PRESS.:Q.I
tiATER SOL. :0.0018
OCT/WAT PART.-.13S
BULK RATE FLO'S FROM PART. TO LAKES :0.01
STREAMS-.O.Ql
SOIL :0.98
CTt
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE 0-ltf
SOIL MOISTURE 1-5/4
SOIL MOISTURE 5-10AJ
SOIL MOISTURE 10-15/4
SOIL MOISTURE 15-30M
SOIL MOISTURE 30-SOW
GROUND UATER(SaALLOM)
GROUND UATER(DEEP)
OCEAN
SOIL 0-1A7
SOIL IrSM
SOIL 5-10#
SOIL 10-15/f
SOIL 15T30AJ
SOIL 30*50«
ASSUMED VALUES
REACTION
CONSTANT
.01380
.06900
.01380
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.03U50
.02070
.02070
.02070
.02070
EMISSION
RATE(E&)
320.00000
.00000
48.80000
8.00000
2.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
. 00000
.00000
.00000
.00000
.00000
.00000
.00000
F-ORG
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
:oodoo
.00000
.00000
.00000
.10000
.05000
.03000
.03000
.03000
.03000
STEADY-STATE CONCENTRATIONS' (PPB)
CASE 1
1602. U7U
U.2U2
39.88U
U.355
4.285
u.185
3.U25
3.0UU
2.619
1.792
1.197
.210
.005
1.232
.OUl
.034
.030
.026
.018
.012
CASE 2
1602.334
4.241
39.88U
u.355
4.285
u.185
3.424
2.993
2.379
1.007
.558
.219
.031
1.226
.041
.034
.030
.02U
.010
.006
CASE 3
1602. uug
U.2U2
39.88U
u.355
4.285
u.185
3.U39
3.0U9
2.599
1.794
1.228
.210
.005
1.232
.GUI
.034
.030
.026
.018
.012
CASE u
1602.299
u.241
39.88U
U.3F.U
u.235
u.185
3.U37
3.006
2.322
.866
.526
.219
.036
-1.226
.OUl
.03u
.030
.023
.009
.005
-------
Table VIb:
STEADl^STATE CONCENTRATIONS OF BENZENE
MOL. W.:78
MOL. POL.: 26. 2
VAP. PRESS. :0.1
tfATER SOL. : 0.001 8
OCT/MT PART.: 135
BULK RATE b'LOU FROM PART.
TO LAKES :0.01
STREAMS :0. -01
SOIL :0.98
en
ro
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATB
LAKES
STREAMS
SOIL MOISTURE Q*iM
SOIL MOISTURE l^SAf
SOIL MOISTURE 5-1QM
SOIL MOISTURE 10tl5Af
SOIL MOISTURE 15-T.30W
SOIL MOISTURE 30«BOW
OtfOOT/Z? UATER(SHALLOU)
GROUND XATER(DEEP)
OCSAfi
SOIL O^IM
SOIL 1-5/4
SOIL SnlQM
SOIL lOwlS/tf
SOIL 15i30W
SOIL 30-5Wf
ASSUMED VALUES
REACTION EMISSION
CONSTANT RATE(ES) F-
-------
• Increasing the reaction rate for air moisture by a factor of
10 (Table Vic) led to very small reductions of the concentra-
tion in most compartments.
t Reducing the octanol/water partition coefficient to practical-
ly zero (and thereby eliminating adsorption into soil) had
practically no impact on eventual levels in water and air. (Table VId)
• Reducing the rate of precipitation of particulates by 50%
(Table Vie) nearly doubled the concentration of particulates
in air. Doubling the rate (Table Vlf) almost halved the
concentration. Only minor effects were seen in other
compartments.
• Increasing the vapor pressure by a factor of 5 lead to a 3%
increase in the concentration of vapor in air and an 30%
reduction in the water compartments (Table VIg). Decreasing
the vapor pressure by 50% reduced the concentration of vapor
in air and nearly doubled the concentration in water compart-
ments by 3% (Table Vlh).
In view of these results it appears that the eventual levels are not
sensitive to soil adsorption characteristics and relatively insensitive
to reaction rates. This is fortunate because these parameters are
relatively difficult to obtain with any accuracy. Adsorption parameters
need not be refined further, reaction rate constants should be within
one order of magnitude in order to be useful.
63
-------
Table Vic:
STEADY-STATE CONCENTRATIONS OP BENZENE
MOL. UT. :78
MOL. VOL.-.26.2
VAP. PRESS.:0.1
JATER SOL. :0.0018
OCT/HAT FART.:135
BULK RATE FLW FROM PART.
TO LAKES :0.01
STREAMS:0.01
SOIL :0.98
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE 0-1M
SOIL MOISTURE l-5Af
SOIL MOISTURE 5-1QM
SOIL MOISTURE 10-ISM
SOIL MOISTURE 15-30M
SOIL MOISTURE 30-SOW
GROUND WATBR(SHALLOV)
GROUND HATER(DEEP)
OCEAN
SOIL 0-ltf
SOIL l-5Af
SOIL 5-10A/
SOIL 10-ISA/
SOIL 15-30M
SOIL SO-SO/-/
ASSUMED VALUES
REACTION
CONSTANT
.01380
.69000
.01380
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.03450
.02070
.02070
.02070
.02070
EMISSION
RATE(E6)
520.00000
.00000
48.80000
8.00000
2.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
F-ORG
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.10000
.05000
.03000
.03000
.03000
.03000
STEADY-STATE CONCENTRATIONS (PPB)
CASE 1
2465.858
6.527
39.884
6.701
6.594
6.440
5.270
4.685
4.030
2.757
1.842
.322
.007
1.896
.064
.052
.046
.040
.027
.018
CASE 2
2465.642
6.527
39.884
6.701
6.593
6.440
5.269
4.606
3.661
1.550
.859
.337
.048
1.887
.064
.052
.046
.036
.015
.008
CASE 3
2465.819
6.527
39.884
6.701
6.594
6.440
5.292
4.692
3.999
2.760
1.890
.323
.007
1.896
.064
.052
.046
.040
.027
.019
CASE 4
2465.588
6.526
39.884
6.700
6.593
6.440
5.289
4.625
3.573
1.332
.809
.337
.055
1.887
.064
.052
.046
.035
.013
.008
-------
Table VId:
STEAD*-STATS CONCENTRATIONS OF BENZENE
MOL, tvT. :73
MOL. VOL.:26.2
VAP. PRESS.:Q.I
HATER SOL. :0.0018
OCT/'JAT rAfff.:lE~6__
BULK KATE FLOU FROM PART. TO LAKES :0.01
STREAMS:0.01
SOIL :0.98
tn
COMPARIMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE 0-1M
SOIL MOISTURE 1-5A?
SOIL MOISTURE' 5-10Af
SOIL MOISTURE 10-15M
SOIL MOISTURE 15-30W
SOIL MOISTURE 30-SOW
GtfOMWP XATER(SHALLOW)
GROUND H'ATE'R(DEEP}
OCEAN
SOIL 0-lfS
SOIL l-5Af
SOIL 5-10«
SOIL 10-15/4
50JL 15-30M
30-SOW
ASSUMED VALUES
REACTION
CONSTANT
.01380
.06900
.01380
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.03450
.02070
.02070
.02070
.02070
EMISSION
RATE(E&)
520.00000
.00000
48.80000
8.00000
2.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
F-ORG
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.10000
.05000
.03000
.03000
.03000
.03000
STEADY -STATE CONCENTRATIONS (PPB)
CASE 1
2169.882
6.538
39.884
6.712
6.605
6.451
5.760
5.480
5.095
4.658
4.305
.375
.008
1.902
.000
.000
.000
.000
.000
.000
CASE 2
2469.677
6.537
39.884
6.712
6.604
6.450
5.759
5.438
4.840
3.742
3.093
.345
.065
1.891
.000
.000
.000
.000
.000
.000
CASE 3
2469.863
6.538
39.884
6.712
6.605
6.451
5.777
5.485
5.096
4.666
4.338
.376
.008
1.902
.000
.000
.000
.000
.000
.000
CASE 4
2469.633
6.537
39.884
6.711
6.604
. 6.450
5.775
5.446
4.774
3.475
2.934
.345
.069
1.891
.000
.000
.000
.000
.000
.000
-------
Table Vie:
STEADY-STATE CONCENTRATIONS OF BENZENE
MOL. UT.-.7Q
MOL. VOL.:26.2
VAP. PRESS.:O.I
tfATER SOL. :0.0018
OCT/HAT PART.:135
BULK. RATE PLOW FROM PART.
TO LAKES ;0.005
STREAMS:0.005
SOIL
cr>
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE 0-1/f
SOIL MOISTURE 1-5/4
SOIL MOISTURE 5-l(W
SOIL MOISTURE . 10 -15Af
SOIL MOISTURE 15-30M
SOIL MOISTURE 30-50/V
SfflM/tffl HATER(SHALLOU')
GROUND *ATSR(DEEP)
OCEAN
SOIL 0-1M
SOIL 1-5W
SOIL 5-10/4
SOIL 10-ISM
SOIL 15-30M
SOIL 30-SOW
U
ASSUMED VALUES
REACTION
CONSTANT
.01380
.06900
.01380
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.03450
.02070
.02070
.02070
.02070
EMISSION
RATE(E&)
520.00000
.00000
48.80000
8.00000
2.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
F-ORG
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.00000
.10000
.05000
.03000
.03000
.03000
.03000
STEADY-STATS CONCENTRATIONS (PPB)
CASE 1
2442.257
6.465
67.444
6.637
6.531
6.379
5.219
4.640
3.991
2.730
1.824
.319
.007
1.878
.063
.052
.046
.039
.027
.018
CASE 2
2442.043
6.464
67.444
6.637
6.530
6.378
5.219
4.562
3.626
1.535
.851
.333
.048
1.869
.063
.052
.045
.036
.015
.008
CASE 3
2442.218
6.465
67.444
5.637
6.531
6.379
5.241
4.647:
3.960
2.734
1.872
.320
.007
1.878
.063
.052
.046
.039
.027
.018
CASE 4
2441.990
6.464
67.444
6.536
6.530
6.378
5.238
4.581
3.539
1.319
.801
.333
.055
1.869
.063
.052
.045
.035
.013
.008
-------
Table Vlf:
STEADY-STATE CONCENTRATIONS OF BENZENE
CT>
MOL. '*/T. :78
MOL. VOL.-.2&.2
VAP. PRESS.:0.1
WATER SOL. : 0.0018
OCT/'JAT PART. :13S
BULf. RATE FLO* FROM PART
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE Q-IM
SOIL MOISTURE 1-5W
SOIL MOISTURE 5-10Af
SOIL MOISTURE 10-lSAf
SOIL MOISTURE 15-30W
SOIL MOISTURE 30-SOW
GROUUD dATER(SRALLOU)
Gl
OCEAN
SOIL 0-1W
SOIL l-5Af
SOIL 5-10H
SOIL 1 OrISM
SOIL 15-30M
SOIL 30-SOW
GROUND dATER(DEEP)
'T. TO LAKES :0.02
ST/?ff/lW5:0.02
SOIL ;1.96
ASSUMED VALUES
REACTION
CONSTANT
.01380
.06900
.01380
. 06900
.06900
.06900
.06900
.06900
.06900
. 06900
.06900
.06900
.06900
.06900
.06900
.03450
.02070
.02070
.02070
.02070
EMISSION
RATE(E6)
520.00000
. 00000
48.80000
8. 00000
2.00000
.00000
. 00000
.00000
. 00000
.00000
.00000
.00000
. 00000
. 00000
. 00000
. 00000
.00000
. 00000
.00000
. 00000
F-ORG
. 00000
.00000
.00000
.00000
. 00000
.00000
. 00000
. 00000
.00000
. 00000
.00000
. 00000
. 00000
.00000
. 10000
.05000
. 03000
.03000
. 03000
.03000
STEADY-STATE COtJCEXTF.ATIONS (PFB)
CASE 1
2486. 323
6. 581
21. 947
6.757
6.649
6.494
5.313
4.723
4.063
2.780
1.857
.325
.007
1.912
.064
.053
.047
.040
.027
.018
CASE 2
2486. 106
6. 581
21. 947
6.756
6.648
6.493
5.313
4.644
3.591
1. 562
. 866
. 339
. 049
1.902
. 064
.053
.046
.036
.015
.009
CASE 3
2486. 284
6.581
21. 947
6.757
6.649
6.494
5.335
4.731
4.032
2.783
1.906
. 325
. 007
1.912
.064
.053
.047
.040
.027
.019
CASE 4
2486.051
6. 531
21.947
6.756
6.648
6.493
5.332
4.664
3.602
1. 343
. 815
.339
.056
1.903
.064
.053
.046
.036
.013
.008
-------
TABLE VIg:
STEADY-STATS CONCENTRATIONS OF BENZENE
MOL. VT.:78
MOL. VOL.:26.2
VAF. PRESS.:0.5
4ATER SOL. : 0.0018
OCT/dAT PART .:135
Bi/Z/A" fl/lffi1 FLOW
PART. TO LAKES :0.01
SOIL :0. 98
oo
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
GROUND y/ATER(
GROUND dATER(
OCEAN
SOIL 0-ltf
SOIL 1-5A/
SOIL 5- 10/4
SOIL 10-15/4
SOIL 15-30/0
SOIL 30-50.V
0-l.V
1-5/V
5-1 0/4
10-15/4
15-30M
30-5 0/4
SHALLOW)
DEEP)
REACTION
CONSTANT
. 01380
. 06900
.01380
. 06900
. 06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.06900
.03450
.02070
.02070
.02070
.02070
ASSUMED VALUES
EMISSION
RATE(SS')
520. 00000
. 00000
48. 80000
8. 00000
2. 00000
.00000
. 00000
. 00000
. 00000
. 00000
.00000
.00000
. 00000
. 00000
. 00000
.00000
.00000
. 00000
.00000
.00000
F-ORG
. 00000
. 00000
.00000
. 00000
. 00000
.00000
.00000
. 00000
.00000
.00000
. 00000
.00000
. 00000
.00000
.10000
. 05000
.03000
.03000
.03000
.03000
STEADY-STATE COilCStlTRATIOHS (PPS)
CASE 1
2532. 815
1. 341
39. 884
1.377
1. 355
1.323
1. 083
. 962
.828
.566
. 378
. 066
. 001
. 390
.013
.011
.010
. 008
.006
.004
CASE 2
2532.770
1. 341
39. 884
1. 377
1.355
1.323
1. 083
. 946
.752
. 318
.175
. 069
. 010
. 388
.013
.011
.009
.007
.003
.002
CASE 3
2532. 807
1. 341
39. 884
1. 377
1. 355
1. 323
1. 087
. 964
. 821
. 567
. 388
. 066
. 001
. 390
.013
. Oil
.010
. 008
.006
.004
CASE 4
2532.759
1.341
39.884
1.377
1.355
1.323
1.087
.950
.734
.274
. 166
.069
.011
.388
.013
.011
.009
.007
.003
.002
-------
TABLE Vlh:
STbADX-STATS CONCENTRATIONS OF
'AOL. vlT . :78
VOL. VOL.:26.2
VAP. PRESS.;O.OS
HATER SOL. :0.0018
OCT/'JAT PART. : 135
BULK RATE FLOS FROM
COMPARTMENT
AIR
AIR MOISTURE
PARTICULATE
LAKES
STREAMS
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
SOIL MOISTURE
GROUND
GROUND dATER(DESP)
OCEAN
SOIL 0-1/4
SOIL
SOIL
SOIL
5 - 1
10-
15-
30-
15M
30/4
5 OA<
SOIL
SOIL
1-5A7
5-10/4
10-15/4
15-30/4
30-5 0/4
'?. .T-3 L/5XSS :0.01
SfffS/JMSiO. 01
SOIL :0.98
ASSUMED VALUES
REACTION
CONSTANT
. 01380
. 06900
. 01380
. 06900
. 06900
.06900
. 06900
.06900
.06900
.06900
.06900
.06900
.06900
. 06900
.06900
.03450
.02070
.02070
. 02070
.02070
SVISSION
RATE(E6)
520.00000
. 00000
48. 80000
8.00000
2.00000
. 00000
.00000
. 00000
. 00000
.00000
. poooo
.00000
. 00000
. 00000
. 00000
.00000
. 00000
. 00000
. 00000
. 00000
F-ORG
. 00000
.00000
.00000
.00000
.00000
.00000
. 00000
. 00000
.00000
. .00000
. 00000
.00000
. 00000
. 00000
. 10000
.05000
..03000
. 03000
. 03000
. 03000
STEADY-STATE CONCENTRATIONS (PP3)
CASE 1
2393. 525
12.671
39. 884
13. 009
12.801
12. 503
10. 230
9. 094
7. 824
5.352
3. 576
.626
. 014
3.682
. 124
. 101
.090
.077
.053
.035
CASE 2
2393. 119
12.669
39. 884
13. 007
12.799
12. 501
10. 223
3.941
7. 106
3.008
1.667
. 553
. 094
3.662
. 124
. 101
.088
.070
. 030
. 016
CASE 3
2393.451
12.671
39. 884
13. 009
12. 801
12.503
10. 272
9. 108
7.762
5.358
3. 670
.628
. 014
3.582
. 124
. 102
.090
. -.077
.053
. 036
CASE 4
2393.017
12.66S
39. 884
13. 006
12.798
12.500
10. 265
8.973
6. 935
2.585
1.570
.653
. 107
3.663
. 124
.101
.083
. 069
.026
. 015
-------
VII. RECOMMENDATIONS
The subsystem used to compute eventual levels is ready for implementa-
tion. The data requirements for this subsystem are quite manageable, and
the results are relatively insensitive to data which may not be available.
The area which may require additional effort is the development of a more
definitive matrix of bulk flows and of similar matrices that can be used
for a "worst case" analysis or for regional analyses! Since these
matrices constitute fixed data for the system, they may be modified at
any time that better information becomes available.
The refinement of soil adsorption relationships is of low priority
because of the relative insensitivity of the results to this relation.
The subsystem used to compute levels of concern requires refinement and
expansion to compartments other than air. One possible approach is to
base the levels of concern on the amount taken in by humans. This
amount can be approximated from the eventual levels by a linear trans-
formation:
Daily Human Uptake = B(Ca1r + Cpart + C^^^ ^
+ DCwater+ M('5 Cwater + "5 Po/wCwater)
. ^ soil moisture+ o/w soil moisture)
where B is the kg/day of air breathed
D is the kg/day of water drunk
M is the kg/day of meat consumed
V is the kg/day of vegetable matter consumed
and C . is a suitable average of the concentration in lakes and streams,
This would provide a useful summary index.
It is essential, however, that regardless of the indices chosen the level
of concern be based on the concentrations that are likely to impose undue
risk to health and environment under chronic exposure conditions rather
than on concentrations that cause easily measurable damage: if the
damage is easily measurable the risks are intolerable.
Inasmuch as the system must be able to cope with new chemicals, the most
attractive alternative would be based on correlating structure with
damage measured by epidemiologic methods under actual conditions. While
• this would provide a valid and defensible beginning, we believe that the
70
-------
available information would not be sufficient to establish a correlation
between structure and effect. Delphi methods and experimentation could
be used separately or jointly to provide a reasonable relation between
structure and effect; if the substances to be used in the assessments
are selected on the basis of their contribution to the analysis, such
an effort could yield a useful tool for screening at a reasonable cost.
Such an approach should be considered. Other approaches, which the Office
of Toxic Substances is already pursuing are providing useful results for
measurable damage. Methods for extrapolating from these approaches would
be useful.
71
-------
APPENDIX I
DISTRIBUTION OF A POLLUTANT IN THE ENVIRONMENT
1-1
-------
APPENDIX I
DISTRIBUTION OF A POLLUTANT IN THE ENVIRONMENT
The distribution of a pollutant among the compartments of the environ-
ment is determined by intercompartment flows and compartment concentra-
tions. For purposes of the present analysis, it appears adequate to
divide the environment into five primary compartments. The ocean is
considered a residual compartment in which the pollutant is absorbed or
decomposed so that flows back from the ocean can be neglected. A
schematic diagram of the flows is given in Figure 1-1, which is identi-
cal to Figure 7 in the text of the report. The total amount of material
in a compartment at time t + At must be equal to the amount in that
compartment at time t plus the amount created in or flowing into the
compartment in time At minus the amount degraded in or flowing out of
the compartment in time At. From these considerations it is possible to
write equations for each compartment. For example, for streams we
have
= WsCs(t)
where W is the weight of water in compartment x, in kg,
A
C is the concentration of pollutant in compartment x,
in kg/kg,
is the bulk flow of pollutant solution from compart-
ment x to compartment y in kg/yr,
r is the diffusional flow of pollutant from compart-
^ ment x to compartment y in kg/yr/unit concentration,
P is the pollutant emission rate into compartment x,
x
in kg/yr,
k is the first order reaction constant in water,
in kg/kg/yr,
1-2
-------
FIGURE 1-1
SCHEMATIC DIAGRAM OF INTERCOMPARTMENT FLOWS
OF EMITTED CHEMICAL
(man and biosphere omitted)
EMITTED CHEMICAL
BULK FLOW
DIFFUSIONAL OR CONVECTIVE FLOW
1-3 .
-------
and the subscripts are
1 for lakes
s for streams
gm for ground moisture
gw for ground water
a for air
am for atmospheric moisture
o for ocean
Considering all compartments, the equations corresponding to Equation 1-1
can be summarized by
WxCx(t + At) = WxCx(t) +-At[Px(t)
(1-2)
where in many cases F and r are zero or negligible. Defining the
variable
Zx - MxCx{t + 4t)-HxCx{t)-Px{t)it
the above equation can be expressed as
Vx)Cy(t)-kxWxCx(t}
or, in matrix form:
Z = (At) • M • C (1-4)
where M is a matrix of coefficients.
1-4
-------
At steady state C(t + At) = C(t), so that
X A
Zx = -px(a>)At f1"5)
where Px(°°) is the eventual (steady state) rate of emissions into com-
partment x. Under these conditions Equation 1-4 becomes
-»• -»-
P(oo) = -M • C(») (1-6)
and the vector of steady state concentrations becomes
C(oo) = -M"1 ?<») (1-7)
so that the eventual environmental levels can be computed from Equation
1-7 by inverting the matrix M.
The elements of the matrix M can be computed as follows:
mxx = -kxwx - £ *W- £ Vy
• _ •
mxy = W + Vx
where k is the first order effective reaction
constant for compartment x
F is the bulk flow rate of solution from
y compartment x to compartment y
r is the diffusional flow rate for the
y pollutant from compartment x to com-
partment y per unit concentration in
compartment x
•
The values of Fx+y and rx^y for compartment pairs for which they are
not negligible are given in the text of the report.
1-5
-------
APPENDIX II
CHEMICAL CLASSES WITH POTENTIAL FOR
CAUSING BIOLOGICAL DAMAGE
n-i
-------
APPENDIX II
CHEMICAL CLASSES WITH POTENTIAL FOR CAUSING BIOLOGICAL DAMAGE
A. INTRODUCTION
Anyone with a background in chemistry recognizes that some chemical
classes are more hazardous than others. The student in freshman chem-
istry laboratory, for example, will handle strong mineral acids with
far more circumspection than metal oxides or organic solvents. Still,
systematic correlation of biotic hazard with chemical type has defied
many excellent scientists, whose attempts have proved expensive and
unsatisfying.
If we are to protect the biota of the environment including man from
chemical risk, we must either (1) exhaustively test all chemicals for
potential environmental damage--a prohibitively expensive alternative,
or (2) devise some sort of screening methodology which will enable us
to restrict our precautionary investigations only to suspect chemicals.
The necessary condition for the development of any sort of screening
methodology is that there be a workable and defensible correlation
between chemical class and environmental damage. One such correlation
is presented below.
B. DOSAGE
All chemical substances are toxic. All chemical substances can damage
the environment. With respect to chemical substances found in the
natural environment, damage to the biosphere can result from either low
concentration levels (boron deficiencies in many U.S. soils) or exces-
sive levels (mercury pollution of Minamata Bay). That is to say, the
dose-response curve is not a monotonically increasing function passing
through the origin: it will, for many naturally occurring substances,
have at least one point at which the first derivative is zero. The com-
plicated relation of dose level to effect is frequently cited as a
reason for the hopelessness of trying to relate environmental harm to
chemical class. While it is a real difficulty, it is not insurmountable,
since all organisms tend to have the same basic biochemistry. Thus,
the dosage problem can be avoided in part'by restricting comparisons to
the same species or by normalizing for body mass. Persistence, biodeg-
radation, and biomagnification are further complicating factors. These
difficulties do not preclude the possibility of workable screening, but
they should serve as a warning that no screening system can be perfectly
reliable and that any screening system should always be applied with
its limitations strongly in mind.
II-2
-------
C. SELECTION RULES
One possible screening methodology is to promulgate "selection rules;"
i.e., a list of chemical substituents which, when present, make a
chemical suspect. The following group of selection rules has been de-
veloped and evaluated:
A chemical compound may damage the biosphere if it contains or
reacts with water to form compounds containing:
(1) Be, B, P, and/or any element of atomic number
greater than 22 (except Br~ and I salts and
"noble" gases);
(2) A covalently bound halogen;
(3) A proton with large K :
a
(4) A base with large K. ;
(5) An -0-0- linkage;
\
(6) ^t)=0, >C-OH, -C-O-C- groups, but not -C=0
(7) R'-S-R and R'-S-S-R groups where R and R1 can
be the same or different and where R and/or
R1 is or contains any element of atomic number
less than 22;
(8) N bonded to N (except for N ), 0, S, and/or C;
(9) R,X=C^ and/or R-CsC- groups where R is not -C-
(note that this rule includes aromatics); andx
- (10) A heterocyclic ring (saturated and/or unsatu-
rated).
Application of these selection rules screens out all the hazardous chem-
icals on the OTS's Appendix I list except for Al, silicones, CaCl?., and
ethyl acetate, and the inclusion of these on the list, it should be noted,
is questionable. Spot application testing indicates that these rules are
capable of catching 80%-90% of the chemicals in the much longer CHRIS
list; again, it is questionable whether those chemicals this screening
misses (acetic acid, ammonium nitrate and sulfate, amyl acetate, butane,
etc.) are really of major environmental concern. These tests indicate
that the selection rules proposed are effective at screening out hazard-
ous substances. For the screen to be useful, however, it must pass a
high percentage "innocent" chemicals.
II-3
-------
To test the "fineness" of the screen, the selection rules were applied
to a "neutral" chemical list; i.e., a general list of chemicals not
specifically hazardous. When applied to 47 inorganic chemicals (every
20th compound on pp. 528-620, 44th ed. of the Chemical Rubber Handbook),
the selection rules passed only 7, yet it seemed to be working well.
When every 100th organic chemical on pp. 770-862 was checked, only 1 out
of 17 passed. Again, the screen seemed to be working well. It is clear
there are few "innocent" chemicals. Applied to the Chemical and Engineer-
ing News (June 2, 1975, p. 32) list of 50 biggest volume chemicals
(Table II-l) 18 out of 50 passed and the selection rules seemed to be work-
,ing; i.e., those that passed would probably be judged "innocent" by most
experts. It should be noticed in this instance, where the list is not
one specifically of hazardous substance, that a substantial fraction
(about 36%) of the chemicals listed passed the screen. Of those that did
not some are already under investigation (e.g., vinyl chloride, styrene,
dimethyl, terephthalate), and others could very well prove to be problems
in the future (e.g., cumene, vinyl acetate).
A workable screening should, of course, err on the side of being too con-
servative. The inconvenience of having to examine further "innocent"
compounds caught by the screen is far less worrisome than major leakage
of hazardous substances through the screen. However, if a screen were
too indiscriminate, it would obviously be worthless. Lists of "innocent"
chemicals are relatively hard to come by. One such listing is the GRAS
("generally recognized as safe") food additive list. Examination and
application of the above selection rules to this list is rather unsettl-
ing, since many of these substances do not pass the screening. Many
flavoring substances, for example, contain^C=0, H^C=0, or -C-OH groups
turpenes, heterocyclic rings, and even phenolic groups. Some of these
may, in fact, merit reinvestigation. On the other hand, sugars fail
rules 6 and 10, and proteins fail 6, 7, 8, and 9. In fact, there is a
very large class of substances of biological origin which would fail the
screening. There are two possible approaches to this problem:
(1) Some kind of specification of dosage level, and
(2) Restricting the applicability of selection rules
6, 7, 8, and 9 (and 10 where the heterocyclic
element is not oxygen) to compounds of molecular
weight under 150-200 (exclusive of any elements
of At. No. greater than 22).
II-4
-------
TABLE II-l
SELECTION RULES APPLIED TO 50 TOP CHEMICALS
Rank
1974 1973
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
1
2
3
4
5
6
7
9
8
10
11
12
13
14
15
18
16
17
20
19
11
2!1
23
24
25
26
27
30
28
29
31
32
33
34
37
40
38
35
36
41
39
44
42
43
45
46
47
50
47
49
Sulfuric acid
Lime, except refractory dolomite
Oxygen, high and low purity
Ammonia, synthetic anhydrous
Ethyl ene
Sodium hydroxide, 100% liquid
Chlorine, gas
Nitrogen, high and low purity
Nitric acid
Sodium carbonate, synthetic and natural
Ammonium nitrate, original solution
Phosphoric acid
Benzene
P ropy 1 ene
Ethyl ene di chloride
Toluene
Urea, primary solution
Methanol , synthetic
Styrene
Formaldehyde, 37% by weight
Xylenes
Ethyl benzene
Vinyl chloride
Hydrochloric acid
Ethyl ene oxide
Ammonium sulfate
Butadiene (1,3-), rubber grade
Terephthalic acid, crude
Carbon black
Ethyl ene glycol
Carbon dioxide, all forms
Cumene
Sodium sulfate, high and low purity
Dimethyl terephthalate
p-Xylene
Cyclohexane
Phenol
Aluminum sulfate, commercial
Acetic acid
Acetone
Calcium chloride, solid and liquid
Isopropanol
Ethanol , synthetic
Sodium tripolyphosphate
Propylene oxide
Acetic anhydride
Titanium dioxide
Sodium silicate
Adipic acid
Vinyl acetate
PRODUCTION
(Bi.1JJ.0-n-s_..°.f.J..b) SELECTION
1974 1973 RULE FAILED
64.71
40.75
32.12
31.40
23.52
21.73
21.24
17.17
16.37
15.10
15.09
14.26
11.07
9.82
7.70
7.49
7.40
6.86
5.94
5.85
5.79
5.70
5.60
4.81
3.89
3.88
3.66
3.43
3.35
3.11
2.91
2.87
2.75
2.74
2.68
2.34
2.32
2.32
2.26
2.06
2.00
1.91
1.90
1.87
1.74
1.71
1.58
1.54
1.51
1.40
63.45
39.76
32.45
30.19
22.33
21.44
20.80
16.52
26.88
15.04
14.31
13.70
10.65
9.88
9.29
6.94
7.27
7.06
5.98
6.42
5.66
5.69
5.35
5.03
4.17
3.97
3.64
3.20
3.49
3.28
3.14
2.67
2.61
2.56
2.33
2.12
2.28
2.50
2.43
1.99
2.16
1.84
1.96
1.92
1.75
1.67
.1.57
1.45
1.57
1.50
3
4
4
2
3
3
9
2
9
6,8
6
9
6
9
9
2
3
6
9
9
6
9
9
9
3,9
6
6
6
1
6
1
9
II-5
-------
D. TLV'S, LD50'S AND CHEMICAL TYPE
In order to quantize possible correlations between toxicity and chemical
type, TLV's in ppm were compared to molecular weight within types of
compounds such as aliphatic alcohols, aldehydes, and monochlorides. Two
difficulties were immediately evident: (1) there are not enough TLV's
quoted for members of a homologous series so that any trends can be
traced further than the first two or three members of the series, and
(2) TLV's are not empirical numbers but rather represent a rather arbi-
trary and unhappy attempt to quantize judgmental assessments of toxic
levels.
LD50 values were also examined for a possible quantitative correlation.
Because of the importance of body mass (really a dilution factor) com-
parisons were made only for the same organism. Oral values were used
since that seemed the most probable route of exposure in real environ-
mental situations. However, since oral values encompass transport
through biomembranes as well as specifically toxic biochemical effects,
one could argue that injection values might be preferable. The results
(Figure II-l is one example) are not very compelling, but do illustrate
that with many important exceptions, toxicity tends to decrease with
increasing chain length in a homologous series of aldehydes. Another
disquieting observation is that the LD50 values for organic compounds
all tend to be similar, suggesting that what one is looking at is not
real biochemical toxicity, but simply the limits of the liber to accom-
modate these substances.
E. THE 1974 TOXIC SUBSTANCES LIST AND CHEMICAL TYPE
In order to correlate toxicity with chemical class an extensive list of
substances is needed with some quantitative measure of toxicity. Such
a list is Table G-l, Appendix I, in the 1974 "The Toxic Substances List"
(Federal Register 37, No. 202, October 18, 1972) which tabulates about
400 chemicals. This list, for all its difficulties, at least seems to
boast some sort of official imprimatur. Chief among the difficulties is
the fact that it is an air pollution list, giving maximum respiration
exposure levels in mg/m3. As a consequence the hazard of heavy metals,
particularly if they are in the form of fine powder or dusts, and of
substances which have been notorious causes of respiratory attack in
past industrial settings, are given a very magnified (and from the stand-
point of water pollution, distorted) prominence. We find that silver
metal, copper, and even carbon black, calcium oxide, and mineral oil
appear as much more hazardous than cyanide, benzoyl peroxide, nitric and
hydrochloric acids, sulfur dioxide, Malathion, pyridine, and phenol.
Data in the EPA's Water Quality Criteria Book should be more relevant
but it is too sketchy for the present purposes. Data are not given for
many different chemical classes, and those data that are given are for
different species. The "toxicity factors" and "harmful quantities"
listed in the Batelle Memorial Institute study on penalties for hazardous
II-6
-------
LD5Q FOR RATS (ORAL)
HCHO
CH3CHO
C2H5CHO
C3H7CHO--
C H CHO
4 9
C5HUCHO-
o
X
I—I
o
t—I
—I
<
CO
o
O
73
3=
t—*
O
-<
O
-------
spills, might be used but the orientation of the study is largely eco-
nomic. To demonstrate a methodology for estimating environmental levels
of concern, we have based our initial work on the air standards listing.
A matrix was formed with the approximately 400 toxic substances listed
in the left-hand column in order of increasing allowable level (ranging
from 0.001 mg/m3 for soluble rhodium salts to 9,000 mg/m3 for carbon
dioxide. Across the top were ranged chemical types) based on the selec-
tion rules. (The key to Figure II-2.)
One or more entries were checked for the chemical types exemplified by
each substance. The resulting completed matrix was examined to see if
chemical types tended to cluster in certain MAC ranges. Sophisticated,
rigorous techniques are available for quantitatively establishing non-
randomness but their application here did not appear to be warranted.
Figure II-2 is a highly abbreviated representation of the results. The
matrix clearly evinced the tendency of some classes of chemicals to
cluster in a characteristic acceptable level range and it provided tan-
gible, semi-quantitative evidence to support the existence of certain
toxicity trends.
1. Toxic Heavy Metals
Heavy metals and their compounds, both soluble and insoluble, especially
if they are in finely divided form such as dusts or mists, are very in-
jurious when inhaled, with the majority falling in the range of 0.001 mg/m3
(for soluble Rh salts) to 1 mg/m3 (for Yttrium). Metals with less strin-
gent allowable levels include tin (2 mg/m3); tantalum, manganese,
soluble compounds of molybdenum, ZnO fume, and zirconium compounds (all
with 5 mg/m3); iron oxide fume (10 mg/m3); and Ti02» MgO fume, and
molybdenum (all with 15 mg/m3).
2. Strong Acids and Bases
Sulfuric, oxalic, and phosphoric acids all have an allowable level of
1 mg/m3 but otherwise strong acids and bases show little tendency to
cluster.
3. Isocyanates
Isocyanates appear to be highly toxic substances but not many examples
are included in the list. They range from 0.05 mg/m3 for methyl isocy-
anate down to 0.2 mg/m3 for methylene biphenyl isocyanate.
II-8
-------
FIGURE 11-2
CHEMICAL TYPE KEY
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
Heavy metals
Stong acids and
-N=C=0
Organic phosphorus
Phenols
P-S-0-N-halides
Halogenated hydrocarbons
-N , -N=N-, -(f-N<
R-N02
0-R
DN
-CN
COO©
R-O-RT R-O-O-R^
R-S-S-, RSH, R-S-K
^C=0 (but not -COO)
Cl-O-COOR(H), =C-COOR(H)
ROH
.0
R-C -OR
II-9
-------
ACCEPTABLE
LIMIT RANGE
fl.l
>0.5
FIGURE 11-2
TOXICITY-CHEMICAL TYPE MATRIX
SOMF. EXAMPLE
SUBSTANCES
CIII'MICAI. I VIM,
A K i; i) i i r, M i ,) K i M N o r
Sol. Rh salts (0.001)
Nickel carbonyl (0.007)
Methyl isocyanato (0.05)
Tetraethyl lead (0.075)
Endrin (0.1)
Org. Sn compds. (0.1)
Strychnine(O.lS)
Arsine (0.2)
I
Hafnium (0.5)
Chlorodane(O.S)
Sb and compds. (0.5)
Bromine (0.8)
>5
Sulfuric acid (1)
DDT (1)
Hydrzine (1.3)
Allyl chloride (3)
Benzoyl peroxide (5)
Nitric acid (9)
p-Nitrogniline (6)
Nitrogen dioxide (9)
>50
Methyl amine
Sulfur dioxide
Phenol (19)
Ammonia (39)
(12)
(13)
I
>100
Naphthalene (50)
Carbon monoxide (55)
Acetonitrile (70)
Ethylene oxide (90)
>500
Ethyl acrylate (100)
Ethyl butyl ketone(230)
Methanol (260)
Dioxane (360)
-•1000
Propyl alcohol (500)
Turpentine(560)
1,2-Dichlorethane (790)
t-Butyl acetate (950)
>5000
Ethyl ether (1,200)
Ethanol (1,900)
n-Heptane (2,000)
11-10
-------
4. Organic Phosphorus Compounds
Phosphings and organic phosphorus compounds also appear to be highly
toxic with the substances listed tending to "cluster" a bit diffusely
between 0.1 mg/m3 for triorthocresyl phosphate and 0.4 mg/m3 for phosph-
ings. Of course, it should be noted that organic substances can
exemplify more than one suspect chemical class.
5. Phenols
Phenolic substances range over a broad spectrum of maximum allowable
limits with phenol itself being at 19 mg/m3.
6. P-S-0-N-halide
Compounds with covalent bonds among P,S,0,N, and halides can be very
dangerous. They tend to concentrate between 0.2 mg/m3 (for ozone) and
3 mg/m3 (for chlorine gas). A second clustering appears to occur between
6 mg/m3 (for sulfur monochloride) and 30 mg/m3 (for nitric oxide). Par-
ticularly toxic is oxygen difluoride (0.1 mg/m3).
7. Halogenated Hydrocarbons
Halogenated hydrocarbons, which includes many pesticides, rather to our
surprise do not seem to tend to cluster but are widely scattered through-
out the entire allowable level range from 0.1 mg/m3 for Endrin to
7,600 mg/m3 for 1,1,2-trichloro, 1,2,2-trifluoroethane. Some interesting
trends are evident, however. Bromination is less troublesome than
chlorination and fluorination markedly reduces toxi'city even when heavier
halogens remain in the compound. While not noted in the matrix, we might
observe in passing that aliphatic hydrocarbons are not very toxic and do
not start to make their appearance until quite high allowable levels.
8. Amine, Azo, and Amide Compounds
Some of these N-compounds are quite dangerous; p-phenylene diamine, for
example, appears at 0.1 mg/m3. Most of the compounds do not start to
appear in what might be called a "diffuse cluster," until about 1 mg/m3
(for ethylene imine). They then continue to occur with some frequency
until triethyl amine at 100 mg/m3.
9. Nitro-Compounds
Nitro-compounds appear to tend to cluster between 1.5 mg/m3 (for trini-
trotoluene) to 6 mg/m3 (for p-nitroaniline) but examples can be found
both above or below this range: picric acide at 0.1 mg/m3 and nitro-
toluene at 30 mg/m3.
11-11
-------
10. Phenyl-benzyl Compounds
Unless halogenated or containing some other active groups, phenyl-,
benzyl-, toluene, and related compounds are not among the most dangerous
substances. They begin to appear with allowable levels at about 1 mg/m3
and continue to reappear down to about 500 mg/m3 in a diffuse "cluster."
11. Heterocyclic Nitrogen
These substances appear to be thinly scattered throughout a wide range
from strychnine (0.15 mg/m3) to N-ethylmorpholine (94 mg/m3) with
pyridine at 15 mg/m3. It should be noted that this listing does not
include many exceedingly deadly compounds of biological origin probably
because of their non-volatility and relative rarity.
12. Cyanide-Mitriles
Cyanides, rather surprisingly, do not fall high on the list. They form
a somewhat thinly populated cluster ranging from 3 to 70 mg/m3.
13. Fused Rings
Fused ring compounds range thinly over the list and cluster about 50 mg/m3,
Those few high on the list are there because of some active group such as
a halogen. Naphthalene is at the bottom of this range at 50 mg/m3.
14. Ethers, Oxides, and Peroxides
Carbon chains and rings interrupted by one or two oxygens range widely
from about 0.1 to 740 mg/m3 with a noticeable increase in frequency at
around 100 mg/m3. Again, those high on the list are there by virtue of
other active substituents such as chlorine. There also does appear to
be some discernable tendency for peroxy-compounds to fall higher than
oxy-compounds and for oxygen containing heterocyclic rings to fall higher
than simple ethers.
15. Organic Sulfur Compounds
Of the five organic S-compounds examined on the list, four fell between
12 and 35 mg/m3. The only substance above this range, perchloromethyl
mercaptan (0.8 mg/m3) is chlorinated.
11-12
-------
16. Carbony'l Compounds (ketones, aldehydes, etc.)
With several important exceptions, carbonyl compounds do not appear very
dangerous and for the most part are concentrated in the 100-800 mg/m3
range.
17. Carboxylic Groups with Substituents on the Alpha Carbon
No significant tendencies are evident from the matrix for these substances.
18. Alcohols
With few exceptions, alcohols begin to appear on the matrix list at
50 mg/m3 and they recur with considerable frequency down to 750 mg/m3.
Ethanol is at 1,900 mg/m3.
19. Esters
Simple aliphatic esters begin to appear as a cluster at 250 mg/m3 with
a particularly dense concentration in the matrix at 525-900 mg/m3.
Aliphatic hydrocarbons were not included in the matrix, but, as noted
earlier, it is apparent that they are relatively innocent, not beginning
to appear until about 1000 mg/m3.
In conclusion, the foregoing matrix analysis does show semi-quantitatively
that there are important correlations between chemical types and struc-
ture and that certain chemical types tend to cluster in certain allowable
level ranges.
F. OTHER SCREENS
The methodologies discussed above by no means exhaust the methodologies
for pre-screening potentially environmentally damaging chemical sub-
stances.
Inasmuch as the biosphere has responded to subtances present in the
natural environment, a screen could be formulated on the basis of natural-
ly occurring substances. That is, any chemical in ested into the natural
environment in excess of levels found in the natural unpolluted environ-
ment could be considered suspect. Such a screen would immediately
preclude man-made chemicals not found in the environment. This type of
screen has the advantage of a built-in level specification and it would
probably be well accepted by the environmentalists. It does have one
particularly serious difficulty, however. This is that some chemical
substances occur in the unpolluted natural environment at very injurious
levels. These include mercury, fluoride, turpenes, and a host of toxins
of biological origin.
11-13
-------
A second screen might be based on whether a candidate compound is
structurally analogous to a compound known to be biologically active.
Any effective 'drug or medication would be suspect, for example. (In
this connection it is most surprising that aspirin, so effective, appears
to be so relatively harmless.) Substances which resemble a biosubstance
are particularly troublesome, since the biochemistry of the organism may
fail to distinguish them from their analogues. Still, they cannot per-
form the analogous biological function and, thus, can be most disruptive.
A screen of this type would be most useful for preliminary assessment of
the relative safety of complex organic compounds which are possible can-
didates as food additives, drugs, etc.
G. ASSESSMENT OF SOVIET WORK AND A NEW METHODOLOGY
For some years now Soviet scientists have apparently invested consider-
able effort in the problem of relating biological activity, specifically
toxicity, to chemical structure. A recent summary of this work by
Ljublina and Filov1 has attracted attention in the United States. For
calculating indices of toxicity these authors present three methodologies:
(1) Zaeva (1964)
(2) Zahradnik (1962)
(3) Ljublina (1965-1967)
In the Zaeva equation maximum allowable concentration (MAC) in mg/m3 for
organic substances are calculated by the expression:
(1) MAC = 1000M/Z£i
where M is the molecular weight and zs,i is the sum of "biological activi-
ty values" which have been assigned to various chemical bonds. The
Zahradnik equation is useful only for estimating the relative toxicity
of members of a homologous series and is not applicable for comparisons
among widely different chemical types of organic compounds--our present
concern.
Ljublina and co-workers at the Leningrad Institute of Industrial Hygiene
and Occupational Diseases developed several empirical relationships
correlating physiochemical properties, such as molecular weight, density,
refractive index, and melting point with toxicity indices. The approach
is a statistical one and it is unclear, even a posteri, why these
1 Ljublina, E.I. and V.A. Filov, Methods Used in the USSR for- Establish-
ing Biological Safe Levels of Toxic Substances, World Health Organiza-
tion, Geneva, 1975.
11-14
-------
correlations might be obtained. Furthermore, a large correction factor
must be applied to the computation; this correction factor depends on
chemical type, and thus the whole approach tends to bring us back there-
by to the concepts underlying the Zaeva equation.
We have applied the Zaeva equation to the calculation of the MAC's for
the air inhalation toxic substance list. While there may be some slight
correlation, the results are not very satisfying (Figure II-3). There
are several difficulties in applying the Zaeva equation. For example,
in what is an apparent misprint in the text they list two different
values by £i for the -C-C- bond: 15.4 and 173.7. Which does one use?
The circle-points in Figure II-3 are based on one value, the square
products on the other. We found the correlation is improved if one
deliberately misuses the Zaeva equation, counting a bond which appears
n-times in a compound only once rather than n-times (Figure II-4). We
did find it sometimes possible to obtain an impressive correlation within
a homologous series. Another problem is that there is not necessarily a
relationship between Soviet MAC's and U.S. MAC's, both being judgmentally
based.
Next we used the Zaeva equation, but substituted for Zaeva's "biological
activity values," which we had found ambiguous in application, £-j's based
up the matrix we generated earlier (Figure II-2). Since in Figure II-2
the MAC level at which clustering occurs can easily be arranged according
to types of substituents and/or bonds, so as to regulatory decrease, fc-j's
can be assigned. Although the decreasing clusters of MAC's in the matrix
(Figure II-2) make it clear that correlation is possible, the correlation
obtained was only slightly improved over the original Zaeva equation
(Figure II-5A). After examining the results, however, it became apparent
that the presence of the molecular weight in the numerator of expression
(1) in effect nullifies the additive toxicities in the denominator. Ac-
cordingly, we formulated a new expression:
(2) MAC = 100/E£i
where, based on the matrix (Figure II-2), the following x,i values are
used:
£i = 10 -N=C=0
9 +-C1, -6=0
8 <)>-OH, 4>-CH3, <)>-, R-C1, fused rings (sat. or unsat.),
P-S, -N=N=
"~^s»
7 *-N02, R-NOit C-SH, C=0, -N, -0-0-, =(rC=0
6 -CsN, N-0, heterocyclic N- and 0- rings
5 R-OH, R-Br, C-O-C, C^C
11-15
-------
FIGURE II-3
THE ZAEVA FORMULA
X
Q
| ! J
1—
1
•^
400
350
300
250
200
150
100
50
n-j
o
o
o
o
o
o
i
o
0 ° 0 o
o
o
KO o
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
C_3
150-
100-
50-
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
TOXIC SUBSTANCES LIST MAC (mg/m3)
11-16
-------
FIGURE II-4
MODIFIED ZAEVA FORMULA CORRELATION
2400
2000-
1600 -
1200
«=c
o
800-
400
400 800 1200
TOXIC SUBSTANCES LIST MAC (mg/m3)
1600
2000
240Q
11-17
-------
FIGURE II-5A
FIRST CORRELATION
t— «
1
00
1
1
1 — 1
X
o
H-
O
LiJ
H-
_l
ID
O
_J
CJ
20
18
16
14
12
10
8
6
t
4.
2
o
0 0
o
o °
o
o
o
0 °
0 0 o
00
0 0
00 .0 . 0
0
0
0 0
o
o
o
0.1 0.5 1.0
MAC (in mg/m3)FROM THE TOXIC SUBSTANCES LIST
5.0
10
-------
4 R=0, C=C (non-aromatic
R 0-H
3 R-C=0, C=0
P-R
2 C=0 , R-F
Where: = aromatic (benzene) ring
R = aliphatic HC backbone
This approach immediately gave much more promising results (Figure II-5B)
and so was applied to 100 substances from the toxic substances list
(Figure II-6). A number of features of this approach should be noted:
1. Since we were simply exploring, trying to find a
basis for correlation, the analysis was done in
an "eye balling" fashion. The whole procedure,
however, is readily amenable to rigorous mathe-
matical analysis.
2. The list used is a peculiar one, highly distorted
by virtue of being an air pollution list. The
scatter can be due to the list as well as the
calculation.
3. The scatter envelope is narrowest in the most
critical region—i.e., for the more highly
toxic substances.
4. There are many ways open for refining and sharp-
ening up equation (2) and its application.
Finally, it should be noted in Figure II-6 that we were only looking for
a correlation; thus, the value of MAC is not equal to the value of MAC
from the toxic substances list. This can be readily accomplished by
introducing a logarithmic term and a normalization constant.
11-19
-------
FIGURE II-5B
INITIAL SECOND CORRELATION
0.1 0.5 1.0
MAC (in mg/m3) FROM THE TOXIC SUBSTANCES LIST
5.0
11-20
-------
ro
5
o
o
o
FIGURE II-6
SECOND CORRELATION
0.05 0.1 0.5 1.0 5
MAC (in mg/m3) FROM THE TOXIC SUBSTANCES LIST
10
50
100
1000 10,000
-------
APPENDIX III
ESTIMATION OF DIFFUSIVITIES
in-i
-------
APPENDIX III
ESTIMATION OF DIFFUSIVITIES
A. INTRODUCTION
A variety of methods have been proposed for the estimation of diffusiv-
ities. These have been reviewed by Reid and Sherwood.1 The better
methods depend: on the estimation of molecular volumes or of Lennard-Jones
parameters. Boiling points and specific volumes at the normal boiling
point can be used in these estimations. Since boiling points may not be
available, especially for organic compounds of higher molecular weights,
alternatives appeared necessary. We have found that methods based on the
molar refractivity are adequate for most practical purposes.
The molar refractivity depends on the refractive index, molecular weight,
and density of the substance, properties which are relatively easy to
measure. When it cannot or has not been determined experimentally, it
can also be derived from the chemical formula by using tables of additive
values for various atoms and bonds. Finally, the molar refractivity
enters into the estimation as a cube root, and for compounds of high re-
fractive index this is not a strong function of the index; hence, for
compounds of very high molecular weight, only small errors are introduced
by assuming a refractive index of 1.7.
B. DIFFUSIVITY IN WATER
The diffusivity in water solution may be estimated from the following
empirical relationship:
fi v in-10 T
cw .
cw ysoln (R - 0.855)
where T is the temperature in °K
Msoln is the viscosity of the solution, in poises
R is the molar refractivity of the solute to the V3 power
M\V3
n is the refractive index
M is the molecule weight
p is the density
1 Reid, R.C. and T.K. Sherwood, The Properties of Gases and Liquids,
McGraw-Hill, New York, 1958.
III-2
-------
The data on which the relationship is based is given in Table III-l. For
very dilute solutions soln may be replaced by the viscosity of water.
A sample of substances not included in the development of the empirical
equation is given in Table II1-2. For these substances the refractivity
was computed as follows: for low molecular weight R was computed by
adding the element refractivities; for high molecular weights it was com-
puted by assuming an index of refraction of 1.7. The correspondence is
generally good,, considering the range of extrapolation involved.
C. DIFFUSIVITY IN AIR
The best method of estimating the diffusivity of gases and vapors in air
is based on the relation presented by Hirschfelder:2
Dca = 0.0002628
where T is the absolute temperature, °K
M! is the molecular weight of the gas or vapor
M2 is the molecular weight of air
P is the pressure in atmospheres
o 12 1>s tne arithmetic average of the Lennard-Jones
distance parameter for air and gas or vapor
n "-^is a function of T* = . kT
E! is the Lennard-Jones energy parameter for the
gas or vapor
e2 is the Lennard-Jones energy parameter for air.
Although this expression is quite accurate, Lennard-Jones parameters are
not readily available; therefore, the equation cannot be used directly.
The expression is, nonetheless, very useful since c^ and EJ can be
estimated by a number of methods. Our approach has been to rely on the
observations that:
• n(ltl)can be approximated by a reasonable function
2 Hirschfelder, J.O., C.F. Curtis and R.B. Bird, Molecular Theory of
Gases and Liquids, John Wiley and Sons, New York, 1954.
III-3
-------
TABLE III-l
DIFFUSION PARAMETER AND REFRACTIVE RADIUS
Bromine
Ethanol
Oxygen
Urea
OF SELECTED SUBSTANCES
T Y -in?
n
lide 278
: Acid 257
Alcohol 281
le 256
'1 Alcohol 328
i Dioxide 161
ne 165
•ium Oxide 107
»1 253
-ol 344
ien 99
10! 197
ien 152
is Oxide 169
i 128
10! 290
"ic Acid 414
233
R
2.36
2.35
2.57
2.63
2.81
1.89
2.26
1.55
2.35
2.74
1.25
2.02
1.64
1.98
1.59
2.60
2.98
2.36
III-4
-------
TABLE 111-2
COMPARISON OF CALCULATED AND EXPERIMENTAL
DIFFUSIVITIES IN WATER
Substance
Ammonia
1,2-Propanediol
Resorcinol
Phenol
Pyrogallol
Hydroquinone
Pentaerythritol
Nicotene
Lactalbumen
Lactoglobulin
Serum Albumen
Serum Globulin
Urease
R
1.79
-------
• n(1>1) is a slowly varying function of T* so even a
crude approximation to ej will give useful
values3
• both ai and E± are correlated with the molar refrac-
tivity radius R, as shown in Figures Ill-la
and Ill-lb. Note that substances with devia-
tions on the high side of Figure Ill-la have
deviations on the low side in Figure Ill-lb
(and vice versa), so the errors are partly
compensating.
After appropriate substitutions we find
1 , 1
T1-5 Mi29 (III-3)
Dca = 0.00289 — (R+2.75)2.0
with 0 = 0.81 - 0.066 ln(y) + 0.62(y)-°-88
(III-4)
The diffusivities for a number of gases and vapors in air are given in
Table III-3.
Since the relations between R and the Lennard-Jones parameters were not
derived from the diffusivities, the errors shown in Table III-3 are
representative of estimation errors.
3 A 100% error in zl will affect a*1*1* and hence D12 by about 20%.
III-6
-------
a:
CO UJ
~
33 00
CD LU
I-H Z
u- o
O
o;
ffij-
ITT' -
T
_. ..i
-t
hH4±
-f-
.5
•--i i
j.J i 4-
1J -
I I
e-
PH
bi
!M
Wl
a :
CH3C1 : i i ;•
.. '.
)|He|p,t
cTNfne
1.5 2.0 2.5 3.
R(cc/mo1e)V3
3.5
III-7
-------
a:
o
•-3
o
-------
TABLE II1-3
CALCULATED AND OBSERVED DIFFUSIVITIES IN AIR
Suhr.liiiicu
Ace Lie Acid
Ainiixm hi
Annmmirt
An! HIM-
HlMI/IJIII!
llrumlnr
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Dioxide
Carbon Disulphide
Chlorine
D1 ethyl ami ne
Diphenyl
Ethanol
Ethyl Acetate
Ethyl Butyrate
Ethyl Ether
Helium
Hydrogen
Hydrogen
Methanol
Naphthalene
Nitrobenzen
n-Octane
Oxygen
Oxygen
Pentane
Propanol
Propyl Acetate
Sulphur Dioxide
Toluene
Valeric Acid
Mater
Water
Mater
Water
Mater
Water
Water
Water
Water
Water
a' in nitrogen
' in argon
M
fid
VI
17
•1:1
/i>
IM)
44
44
44
44
44
44
44
44
44
76
76
73
154
46
88
116
74
4
2
2
32
128
123
170
32
32
88
60
102
64
92
102
18
18
18
18
18
18
18
18
18
18
I)
7.3S
1 . 7')
l./'l
:i.i3
7. '17
i'.57
1.H9
1.89
1.89
1.89
1.89
1.89
1.89
1.89
1.89
2.81
2.26
2.89
3.74
2.35
2.80
3.16
2.82
0.81
1.25
1.25
2.02
3.34
3.20
3.40
1.59
1.59
2.99
2.60
3.00
2.17
3.15
2.99
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1
/<)H
m
7'lll
:».m
?
O.M;'
0.152
0.166
0.265
0.538
0.882
1.289
2.007
2.552
0.082
0.106
0.093
0.152
0.130
0.081
0.067
0.094
0.757
0.655
0.741
0.166
0.073
0.077
0.070
0.182
0.213
0.087
0.109
0.095
0.114
0.082
0.086
0.253
0.359
0.742
1.260
1.701
2.192
2.920
3.314
3.684
3.811
'''"'!)!,-..
0.1:1:1
II . 1 '.III
().?:«.
O.ll/.l
0 . O'Hi
ll.O'H
(1. 1.16
0.151
0.164
0.273
0.555
0.915
1.32
1.97
2.45
0.088
0.124
0.105
0.160
0.132
0.071
0.057
0.090
0.641b
0.674a
0.760a
0.162
0.061
0.689
0.060
0.175
0.206
0.070
0.100
0.092
0.122
0.084
0.070
0.260
0.353
0.849
1.556
2.190
2.626
3.250
3.940
4.480
4.490
III-9
-------
APPENDIX IV
SAMPLE COMPOUND DATA
IV-l
-------
Chemical: Benzene (C6H6)
Emission Data:
Production Rate 3810 x 106 kg/yr.
Natural Production kg/yr.
Uses % of Prod. % Emission Emission Rate kg/yr.
I. Low Emission
• Production & Intermediate 100% 3% 114.3 x 106
• • . 3%
3%
3%
II. Intermediate Emission
30%
30%
30%
30%
III. High Emission
• Spills 100% 10 x 106
. Vehicle Exhaust 100% 454.5 x 10G
100%
100%
IV. Natural Sources 100%
Total 578.8 x 106
Distribution of Emissions
To Air 520 x 106 kg/yr
To Air Particulates 48.8 x 106 kg/yr
To Lakes 8 x 106 kg/yr
To Streams 2 x 106 kg/yr
To Ground kg/yr
IV-2
-------
Chemical: Bis (2-chloroisopropyl) ether (C1-CH2-CH)?0
CH3
Emission Data:
Production Rate as byproduct 11 x 106 kg/yr.
Natural Production kg/yr.
Uses % of Prod. % Emission Emission Rate kg/yr.
I. Low Emission . ' .
• Production 100% 3% ' 3 x 105
3%
3%
3%
II. Intermediate Emission
30%
30%
30%
30%
III. High Emission
100%
100%
100%
100%
IV. Natural Sources 100%
Total 3 x 105
Distribution of Emissions
To Air 0.1 x 10s kg/yr
To Air Particulates kg/yr
To Lakes 1.5 x 105 kg/yr
To Streams 1.4 x 105 kg/yr
To Ground kg/yr
IV-3
-------
Chemical: Chlorodifluoromethane (CHC1 F2)
Emission Data:
Production Rate 40 x 106
Natural Production
kg/yr.
kg/yr.
Uses
I. Low Emission
• Production
II. Intermediate Emission
. Refrigerant.
III. High Emission
IV. Natural Sources
of Prod. % Emission Emission Rate kg/yr.
100% 3% 1-2 x 106
3%
3%
3%
100%
30%
30%
30%
30%
100%
100%
100%
100%
100%
12.0 x 106
Total
13.2 x 106
Distribution of Emissions
To Air
To Air Particulates
To Lakes
To Streams
To Ground
13.2 x 106
kg/yr
kg/yr
kg/yr
kg/yr
kg/yr
IV-4
-------
Chemical: Methyl Chloroform (CH3CC13)
Emission Data:
Production Rate 245 x 10G
Natural Production
kg/yr.
kg/yr.
Uses
I. Low Emission
• Production
• Vinylidene Chloride
•
*
II. Intermediate Emission
• Exports
• Miscellaneous
9
»
III. High Emission
. Solvent
IV. Natural Sources
of Prod.
100%
9%
13%
11%
67%
% Emission
3%
3%
3%
3%
30%
30%
30%
30%
100%
100%
100%
100%
100%
Emission Rate kg/yr.
7.3 x 106
0.7 x 106
9.5 x 106
8.0 x 106
164 x 106
Total
189.5 x 106
Distribution of Emissions
To Air
To Air Particulates
To Lakes
To Streams
To Ground
185 x 106
1 x 106
1 x 106
0.5 x 106
2 x 106
kg/yr
kg/yr
kg/yr
kg/yr
kg/yr
IV-5
-------
Chemical: Trichlorofluoromethane (CC13F)
Emission Data:
Production Rate 140 x 10G kg/yr.
Natural Production kg/yr.
Uses % of Prod. % Emission Emission Rate kg/yr.
I. Low Emission
• Production 100% 3% 4.1 x 10G
3%
3%
3%
II. Intermediate Emission
• Refrigerant 3% 30% 1.3 x 106
• Foaming Agent 15% 30% 6.3 x 106
30%
30%
III. High Emission
. Aerosol Propellent 82% 100% 114.8 x 106
100%
100%
100%
IV. Natural Sources 100%
Total 126-5 x 10*
Distribution of Emissions
To Air 99.8 x 106 kg/yr
To Air Particulates 11.1 x 106 kg/yr
To Lakes 5 x 106 kg/yr
To Streams 7.6 x 106 kg/yr
To Ground 3 x 106 kg/yr
IV-6
-------
TABLE IVb. DATA SHEET NO. 2
Chemical: Benzene
Basic Data:
Molecular Weight-
Molar Refractivity
Vapor Pressure*
Water Solubility*
Octanol/Water Partition Coeff.*
* at 20°C
f Reaction Constants
--J
Compartment Type
Air
Particulate
Air Moisture
Water
Adsorbed to Soil
78
26.2
0.1
0.0018
135
atm
gm/gm
Reactivity
Extreme High Moderate Persistent Inert Half Life Reaction Rate
50
50
10
10
1
0.0138
0.0138
0.069
0.069
0.69
-------
TABLE IVb. DATA SHEET NO. 2
00
Chemical: Bis (2-Chloroisopropyl ) ether (Cl
Basic Data:
Molecular Weight-
Molar Refractivity
Vapor Pressure*
Water Solubility*
Octanol/Water Partition Coeff.*
* at 20°C
0
Reaction Constants
Compartment Type
Air
Particulate
Air Moisture
Water
Adsorbed to Soil
171
41.3
0.0010
0.0017
5.9
atm
gm/gm
Reactivity
Extreme High Moderate Persistent Inert Half Life Reaction Rate
100
100
10
10
T
0.0069
0.0069
0.069
0.069
0.69
-------
TABLE IVb. DATA SHEET NO. 2
Chemical: Chlorodifluoromethane (CHC1 F2)
Basic Data:
Molecular Weight- 86.5
Molar Refractivity 12.5
9 5
Vapor Pressure* ' atm
Water Solubility* 0.003 gm/gm
Octanol/Water Partition Coeff.* 12
* at 20°C
Reaction Constants
Reactivity
Compartment Type Extreme High Moderate Persistent Inert Half Life Reaction Rate
Alr 20 0.03
Particulate 20 0.03
Air Moisture 20 0.03
Water 137 0.005
Adsorbed to Soil / 1000 0.001
-------
TABLE IVb. DATA SHEET NO. 2
-------
TABLE IVb. DATA SHEET NO. 2
Chemical: Trichlorofluoromethane (CC13F)
Basic Data:
Molecular Weight, 137
Molar Refractivity 21.9
Vapor Pressure* ] atm
Water Solubility* 0.0011 gm/gm
Octanol/Water Partition Coeff.* 340
* at 20°C
Reaction Constants
Reactivity
Compartment Type Extreme High Moderate Persistent Inert Half Life Reaction Rate
Air 20 0.03
Particulate 20 0.03
Air Moisture 20 0.03
Water 137 0.005
Adsorbed to Soil / 1000 0.001
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-560/1-77-002
2.
3. RECIPIENT'S ACCESSIOI*NO.
4. TITLE AND SUBTITLE
PRE-SCREENING FOR ENVIRONMENTAL HAZARDS- A System
For Selection and Priortizing Chemicals
5. REPORT DATE
APRTI 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Emilio C. Venezian
8. PERFORMING ORGANIZATION REPORT NO.
78486
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
10. PROGRAM ELEMENT NO.
2 LA 328
11. CONTRACT/GRANT NO.
68-01-3208
12. SPONSORING AGENCY NAME AND ADDRESS
Office Of Toxic Substances
U.S. Environmental Protection Agency
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
Phase I Report
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A number of alternatives for pre-screening chemicals for their potential to
inflict environmental hazards were considered. A system design concept which
takes into account both the toxicity of the chemical and the eventual levels
which it can be expected to reach in the environment was selected for further
analysis. Although neither toxicity nor eventual levels can be predicted
with great accuracy, the accuracy attainable by simple methods appeared adequate
for selecting and prioritizing chemicals for additional investigation. A
specific design which relies on data which is usually available was developed
to the point of testing the feasibility of collecting the necessary data and
performing the required computations on five chemicals.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
18. DISTRIBUTION STATEMENT
Document is available to the public
through the National Technical Informa-
tion Service. Springfield. Va. 22151
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
124
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |