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)

-------