STANFORD RESEARCH INSTITUTE
Menlo Park, California-94025 • LLSWU
EPA Report Number
July 1977
ENVIRONMENTAL PATHWAYS OF SELECTED CHEMICALS
IN FRESHWATER SYSTEMS
Part I: Background and Experimental Procedures
by
J. H. Smith, Wv R. Mabey, N. Bohonos,
B. R. Holt, S. S. Lee, T.-W. Chou,
D. C. Bomberger, and T. Mill
SRI International
Menlo Park, California 94025
Contract No. 68-03-2227
Project Officer
George Baughman
Environmental Research Laboratory
Athens, Georgia 30601
V
ENVIRONMENTAL RESEARCH LABORATORY .
OFFICE OF RESEARCH AND DEVELOPMENT
•U.S: ENVIRONMENTAL PROTECTION AGENCY
BORGIA 30601
-------
DISCLAIMER
This report has been reviewed by the Athens Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views
and policies of the U.S. Environmental Protection Agency, nor does mention
of trade names or commercial products constitute endorsement or recommenda-
tion for use.
ii
-------
FOREWORD
Environmental protection efforts are increasingly directed towards
prevention of adverse health and ecological effects associated with specific
compounds of natural or human origin. As part of this laboratory's research
on the occurrence, movement, transformation, impact, and control of environ-
mental contaminants, the Environmental Processes Branch studies the micro-
biological, chemical, and physico-chemical processes that control the trans-
port, transformation, and impact of pollutants in soil and water.
Delineation of the environmental pathways followed by aquatic pollutants
is a key element in assessing the effects of low concentrations of pollutants.
Based on concepts developed over a number of years at this laboratory, the
extramural work reported here provides an approach that can be used to predict
potential exposure of aquatic organisms.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
iii
-------
ABSTRACT
This research program was initiated to develop environmental exposure as-
sessment procedures that can be used to predict the pathways of potentially
harmful chemicals in freshwater environments.
The fundamental premises on which the environmental exposure assessment
approach is based are that (1) the overall rate of disappearance of a chemical
from the aquatic environment is controlled only by the dominant transformation
and transport processes, (2) these processes can be studied independently in
the laboratory, and (3) the laboratory data can be extrapolated to environ-
mental conditions.
Laboratory procedures have been developed for measuring the rates of
volatilization, photolysis, oxidation, hydrolysis, and biotransformations as
well as the sorption partition coefficients on natural sediments and on a mix-
ture of four bacteria. Two models have been used to extrapolate the
laboratory results to the environment. The one-compartment model assumes that
the aquatic system is a single, well-mixed reactor in which chemicals are
transformed, degraded, and/or transported. It can be used to analyze acute
discharges such as spills and to establish priorities for in-depth laboratory
studies. The nine-compartment computer model is used to study the effect of
the transport and transformation processes studied in the laboratory program
on the distribution of a chemical in ponds, streams, and eutrophic and oligo-
trophic lakes.
This report is Part I of a two-part report and describes the environ-
mental exposure assessment models and the laboratory procedures. Part II will
report the results ot' using these procedures to study eleven chemicals:
p-cresol, benz[a]anthracene, benzo[a]pyrene, quinoline, benzolf]quinoline,
9H-carbazole, 7H-dibenzo[c,g]carbazole, benzolb]thiophene, dibenzothiophene,
methyl parathion, and mirex.
This report was submitted in partial fulfillment of Contract No.
6S-03-2227 by SRI International under the sponsorship of the U.S. Environ-
mental Protection Agency. This report covers the period from June 30, 1975,
'.o June 30, 1977.
iv
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tjbles vi
Abbreviations and Symbols viii
Acknowledgments x
1. Introduction 1
2. Conclusions 6
3. Recommendations 7
4. Environmental Assessment 9
4.1 One-Compartment Model 10
4.1.1 Assumptions of the One-Compartment Model ... 10
4.1.2 Mathematical Formulations 11
4.2 Nine-Compartment Computer Model 12
4.2.1 Assumptions 14
4.2.2 Mathematical Formulations 18
4.3 Environmental Parameters 23
5. Physical Properties 24
5.1 Solubility 24
5.2 Absorption Spectra 25
5.3 Volatilization Rate 27
5.3.1 Background 27
5.3.2 Experimental Procedures 29
5.4 Sorption of Organic Substrates 31
5.4.1 Background 31
5.4.2 Sorption on Clays and Sediments 33
5.4.3 Biosorption and Desorption 41
5.4.4 Discussion 43
6. Chemical Transformation 44
6.1 Background . 44
6.2 Photochemistry 45
6.3 Free Radical Oxidation 49
6.4 Hydrolysis 51
7. Biodegradation 56
7.1 Background 56
7.2 Development of Enrichment Cultures 57
7.3 Biodegradation Rates 60
7.4 Isolation and Identification of Major Biodegradation
Metabolites 67
8. References 68
Appendix A. Flow of Water and Sediments between
Compartments in the Computer Model 72
M'pendix 15. Theory of Volatilization of Organic
Substrates from Water 75
-------
FIGURES
Number Page
1.1 Flowchart for Technical Approach 3
4.1 Sorption and Transformation Routes Simulated 13
4.2 Schematic of Assumed Flows Between Compartments in an Aquatic
System 16
4.3 Physical Configurations of the Pond, River, and Lake
Simulations 17
4.4 Flowchart for the Environmental Assessment Model 22
5.1 Substrate Solubility versus Partition Coefficient on Coyote
Creek Sediments (K ) and on a Mixed Population of Bacteria
(Kb) P 35
6.1 pH Dependence of k, for Hydrolysis by Acid, Water, and Base-
Promoted Processes 52
B. ] Schematic of the Two-Film Model of Volatilization from the
Surface of Water Bodies 71
TABLES
>iumher pa
4.1 Ratio of Bacteria to Sediments in Natural Water Bodies .... 21
4.2 Physical Dimensions and Water Quality Characteristics
Assumed in the Environmental Analysis '. 23
5.1 Nominal Wavelengths and Wavelength Intervals for uv and
Visible Absorption Spectra 25
'•I' Oxygen Keaeration Rates in Representative Water Bodies .... 28
^•1 Sources and Characteristics of Ca-Montmorillonite Clay and
Natural Sediments 34
1-4 Recommended Experimental Plan for Isotherm Measurements ... 36
vi
-------
Number
Page
A. L. Flow of Water and Solids Between Compartments in
the Pond Model
A. 2. Flow of Water and Solids Between Compartments in
the River System
A. 3. Flow of Water and Solids Between Compartments in
the Eutrophic and Oligotrophic Lake Systems ......... 74
Vll
-------
ABBREVIATIONS AND SYMBOLS
AA -i ,4 '-Azobis (A-cyanovaleric acid)
H Henry's law constant
! Light flux (photons time"1 liter"1)
v Partition coefficient for sorption on biota
i)
'•; Partition coefficient for sorption on sediments
P
.< Concentration of substrate at which v = % v (mass ml"1)
s m
:-\ Holes liter-1
v Mass of chemical in biota in compartment i
X Total mass of substrate in aqueous phase (£) of compartment i before sorp-
tion
::'. Total mass of substrate in aqueous phase (£) of compartment i after sorp-
Lion
" Mass of suspended sediment
:! Total mass of substrate in the suspended sediment of compartment i before
sorption
.'i^ Total mass of substrate in the suspended sediment of compartment i after
sorption
"i^. Mass of water
Vapor pressure of pure substrate (torr)
:- Gas constant
w Substrate concentration (mass per unit volume)
'M Substrate concentration (moles per liter)
\ Substrate concentration in the aqueous phase of compartment i
r Temperature (°K)
Voluiiie of compartment i
Bacterial mass or cell count (cells ml"1)
Microbial population in compartment i (cells ml"1)
(••-'11 count was used in biokinetic studies and biomass was used in biosorptions.
vili
-------
Y Biomass or cell yield per mass of substrate utilized (cells yg sub-
strate)
Z, Solar radiance intensity (photons cm~3 sec"1 nm"1)
A
e Efficiency of production of R02- from AA
f . Total transformation rate in the aqueous phase in compartment i
k. Rate constant for acid-catalyzed hydrolysis (M~l sec"1)
A
kD Rate constant for base-catalyzed hydrolysis (M"1 sec"1)
D
kN Rate constant for neutral hydrolysis (sec~l)
k First-order rate constant for light absorption by chemical (see"1)
3
k Rate constant for biodegradation (ug cell"1 hr"1)
k£ Pseudo-first-order rate constant for biodegradation (hr"1)
k, 2 Second-order rate constants for biodegradation (ml cell"1 hr"1)
k. Rate constant for hydrolysis (sec"1)
k Rate constant for decomposition of AA
k Rate constant for oxidation (M~l time"1)
k Rate constant for photolysis (time"1)
k" Oxygen reaeration rate (time"1)
kS Rate constant for volatilization (ug ml"1 time"1)
m
, . Mass of biota in compartment i
MW£ Mass of water in compartment i
r Biodegradation rate (ug ml"1 time"1)
r Hydrolysis rate (ug ml"1 time"1)
h
r Oxidation rate (ug ml"1 time"1)
r Photolysis rate (ug ml"1 time"1)
r Volatilization rate (ug ml"1 time"1)
t, Half-life (time)
•5
e Absorption coefficient (M cm"1)
A Wavelength (nm)
X Wavelength of an absorption maximum (nm)
M Specific growth rate (hr"1)
li u = Maximum specific growth rate (hr"1)
m max
-------
ACKNOWLEDGMENTS
The assistance and counsel of the Project Officer for this work, Mr.
George Baughman of the Environmental Research Laboratory, Athens, Georgia, has
been invaluable in this work., as uere the frequent discussion with his staff:
Drs. Richard Zepp, Lee Wolfe, David Brown, Charles Steen, and Ms. Doris
Paris, as well as James Hill and Ray Lassiter, who are also located at the
Athens, Georgia, laboratory.
The contributions of the SRI International staff members who carried out
the laboratory work are gratefully acknowledged. The following persons parti-
cipated in the following tasks:
Environmental Assessment: T. Peyton, B. Suta, E. C. Walters
Physical Transport: D. Haynes, B. Kingsley, D. Stivers, and M. Zinnecker
Chemical Degradation: D. Hendry, A. Baraze, B. Lan, and H. Richardson
Biodegradation: R. Spanggord, E. Shingai, H. G. Shan, S, Sorenson, G.
Shepherd, R. Langley, D. Donaldson, D. Watkins, and
D. Brajkovich
Substantial assistance in report preparation was provided by C. Reeds.
-------
1. INTRODUCTION
This study was designed to develop objective, well-documented procedures
that can be used to predict the pathways of potentially harmful chemicals in
freshwater systems before extensive damage occurs or major investments are
made in production facilities. Although either field or laboratory studies
might provide the data necessary for environmental assessment, field experi-
ments are limited to those chemicals already present in the aquatic environ-
ment and are costly because of the large number of samples that must be col-
lected and analyzed. Laboratory studies, on the other hand, are relatively
inexpensive and more easily controlled, and the results are potentially more
'imenable to generalization to different environmental conditions.
Two important laboratory approaches that are now under development are:
the microcosm or ecosystem study (Isensee et al., 1973; Metcalf et al., 1971;
Taub, 1973) and the integration of independent transformation and transport
processes (Wolfe et al, 1976; Paris et al., 1975; Hill et al. , 1976). The use
of microcosms can provide an overall assessment of complex interactions
in a specific environment, but provides little or no basis for extrapolating
the results to other kinds of environments because the relative rates of many
of the component processes cannot be determined.
We have used the second approach, which we call environmental exposure
analysis. Many of the concepts of this approach were first suggested to us
by the staff of the Athens Environmental Research Laboratory. The approach
uses the results of laboratory measurements of specific physical, chemical,
and biological processes in a computer model that integrates the data with
hydrologic parameters of selected aquatic systems. This approach can provide
information on environmental exposure in many kinds of aquatic environments.
This study was designed to achieve three objectives:
o Develop laboratory procedures for a general environmental exposure
analysis of a chemical, based on measurements of the rates of
physical, chemical, and microbiological transformations believed to be
important for that chemical in natural freshwater ecosystems.
• Develop an integration procedure for extrapolating the laboratory data
to a variety of natural waters.
• Demonstrate the procedures using a series of selected organic
chemica]s.
:'-irt J of this report describes the theory and methods of the laboratory
••-•viburements and computer modeling used for environmental exposure assessment.
-------
Part II describes the results of specific laboratory and modeling studies
with eleven organic chemicals selected for this program.
The scope of the environmental assessment was limited to transport and
transformation processes that might occur under steady-state environmental ex-
posure, such as would result from the continued release from manufacturing
plants, agricultural field runoff, or desorption from contaminated sediments.
Laboratory experiments were performed on homogeneous water solutions of se-
lected chemicals below their solubility limits (usually at less than 1 ug
ml~l) to measure their rates of volatilization, oxidation, hydrolysis, photol-
ysis, and microbiological transformations using adapted mixed cultures, as
well as their partition coefficients for sorption to sediments and biomass,
under conditions representative of, or extrapolatable to, freshwater aquatic
systems. The results of the laboratory studies were integrated with simple
one- and nine-compartment computer models to predict the pathways of the
chemicals in ponds, streams, and lakes (see Part II).
A potential shortcoming of this approach to environmental assessment
is that it may not measure important transformation or transport pro-
cesses that occur in a natural aquatic system. To minini2e this possi-
bility, we have compared idealized laboratory experiments in pure water
with experiments using natural sediments and waters. Nonetheless, the
possibility for more complex interactions in natural systems does exist,
and, if they occur, could lead to incorrect estimates of persistence,
distribution, and pathways.
Three potentially important pathways were deliberately omitted: chemical
and biochemical transformations that might take place in or on sediments, bio-
degradation by bacteria not obtained in mixed culture systems by enrichment
procedures or by microorganisms other than bacteria, and biomagnification.
Tlie uffect of these omissions will be discussed in specific sections of Part
II, but we believe that these omissions will probably not significantly affect
Llie general conclusions.
Figure 1.1 shows the sequence of the various phases of this study, begin-
ning with selection of the chemicals, followed by literature review, laboratory
programs, and environmental assessment. Eleven organic chemicals were
selected for study. Nine of these were aromatic compounds typical of those
likely to be found in effluent streams from fossil fuel processing plants.
These compounds are p-cresol, benz[a]anthracene. benzo[a]pyrene, quinoline,
bcnzoff]quinoline, 9H-carbazole, 7H-dibenzofc,g,Jcarbazole, benzofb]thio-
plicnc, and dibenzothiopliene. The two other compounds, methyl parathion and
nuri'x, arc pesticides that have been used extensively in field applications
':UTC runoff to streams and ponds is likely. In most cases, the literature
r''»«n on these compounds were insufficient to allow us to decide which e.nviron-
H-;iL,il processes might be important for detailed study. Therefore, screening
stt-djL't, were cosiducted to obtain an estimate of the relative importance of
L .uli process. Pathways that appeared to be important were studied in detail
""> "Mam cafe data and to identify products. To maximize the amount of rele-
:i>. data produced and minimize the cost, processes that did not .appear to be
s!3'"licant were not carried past the screening stage.
-------
POLLUTANT
SELECTION
LITERATURE
SURVEY
PREPARE
WORK PLAN
DEVELOP ANALYTICAL
METHODS
BIODEGRADATION
SCREENING
CHEMICAL
SCREENING
PHYSICAL TRANSPORT
SCREENING
PRELIMINARY ENVIRONMENTAL
ASSESSMENT: SELECT AREAS
FOR DETAILED STUDY
BIODEGRADATION
STUDY
CHEMICAL
STUDY
PHYSICAL TRANSPORT
STUDY
ASSESS
ENVIRONMENTAL
FATES
PREPARE
FINAL REPORT
SA-4396-74
FIGURE 1.1 FLOWCHART FOR TECHNICAL APPROACH
3
-------
Brief descriptions of the program elements shown in Figure 1.1 are listed
below. Details are given in the sections noted in parentheses.
Screening studies were designed to measure:
-- Solubility in water at 20 to 25°C (Section 5.1).
— Absorption spectra at wavelengths greater than 290 nm (Section
5.2).
— Volatilization rates under high turbulence conditions (Section
5.3).
— Sorption partition coefficients (Section 5.4) for a Ca-
montmorillonite clay, one high organic content natural sediment,
and a mixture of four species of bacteria.
— Photolysis rates in sunlight and monochromatic light above 300 nm
(Section 6.2).
— Oxidation rates in air-saturated water using a free-radical
initiator at 50°C (Section 6.3).
— Hydrolysis rate at constant pH and temperature (Section 6.4).
—Biodegradation susceptibility, by attempting to develop within
6 weeks enrichment cultures that would degrade substrate when it
was the sole carbon source (Section 7.2).
• Preliminary assessments, using a one-compartment model with rate con-
stants based on screening studies, were used to decide which processes
should be studied in detail (Section 4.1).
• Detailed studies were designed to:
— Measure volatilization rates under several low turbulence condi-
tions (Section 5.3).
— Measure sorption partition coefficients on additional natural
sediments (Section 5.4).
— Measure photolysis rates and quantum yield in pure water and in
natural waters and identify major products (Section 6.2).
— Measure oxidation rates and identify major oxidation products
(Section 6.3).
— Measure hydrolysis rates at several temperatures and over a pH
range 3 to 10 and identify major hydrolysis products (Section 6.4).
— Measure biodegradation rates and identify major metabolites
(Section 7.3).
-------
• Final environmental assessments were made with the one-compartment
model and the nine-compartment computer model.
— The one-compartment model was used to refine the preliminary as-
sessments, using rate constants obtained in detailed studies. This
assessment provided a comparison of the relative transformation
rates of the substrate under different aquatic conditions (Section
4.1).
*
— The nine-compartment computer model was used to predict in detail
the transport, distribution, and steady-state concentrations in
four representative aquatic environments (Section A.2).
The following chapters present and evaluate the laboratory procedures used
in this study and the procedures for integrating laboratory measurements. These
evaluations are important in the application of the methods and conclusions of
the assessment and must be considered in any critical evaluation of the en-
vironmental exposure analysis.
>•- umber of compartments can be increased to 99 if necessary. Individual
"• •'.vrmtnus represent various parts of the water column and sediment layer
•• representative pond, lake, or river.
-------
2. CONCLUSIONS
1. The approach described in this report is a simple, a_ priori method for
evaluating many of the possible environmental pathways of chemical pol-
lutants in natural aquatic environments. This technique can provide use-
ful predictions of the potential environmental exposure of chemicals in
freshwater systems before they are introduced into the environment.
2. Calculations based on first-order kinetics and a homogeneous water body
are useful for rapid assessment oflaboratory screening studies and can
be used to estimate both the volatilization and transformation rates of
the substrate in solution and the importance of sorption by sediments fol-
lowing spills or long-term exposure.
3. The nine-compartment computer model developed during this study for use in
extrapolating laboratory data to typical water bodies can predict the
persistence, distribution, and pathways of chemicals in ponds, lakes, and
streams. This model sacrifices the simplicity of the single compartment
model for realism, but it is still much simpler than many of the computer
models now in use and it allows the users to adjust parameters selectively
to conform to their best judgment or knowledge.
'». Laboratory procedures have been developed to measure the rate constants
for volatilization, photolysis, oxidation, hydrolysis, and biodegradation
and the sorption partition coefficients on natural sediments and bacteria
at substrate concentrations from 0.1 to 1000 ng ml"1. The procedures have
been designed so that the rate constants can be extrapolated to the en-
vironmental conditions simulated by the one-compartment model and nine-
compartment computer models.
-------
3. RECOMMENDATIONS
The work done under this program has laid the basis for further work in
this field. We have developed the following recommendations as a result of
our experience in developing an environmental exposure assessment model.
1. Verify this environmental exposure assessment model by comparing the
predicted pollutant concentrations with those measured in an ecosystem
and, if possible, in the field.
2. Develop procedures for measuring and expressing the rates of biodegradation
and chemical transformations of substrates sorbed on sediments. These pro-
cesses should be included in the environmental exposure assessment model.
3. Investigate further the observation that natural waters and humic acid had
varying effects on the photolysis rates of different substrates. These
investigations could focus on the role of natural substances as photosensi-
Lizers, quenchers, and free radical photoinitiators as well as on how other
natural waters affect the reaction rates and products of specific sub-
strates.
'i. Intensify efforts to identify the major products of chemical transformation
and the metabolites from the biodegradations. The latter may require
changes in fermentation schedules, use of cell-free enzyme systems, or use
of mutated organisms. The toxic properties of these products and meta-
bolites should be measured, perhaps by the B. Ames mutagenic assay, by
unscheduled DNA synthesis, or by animal culture systems.
5. Continue work on refinement of the nine-compartment model and the labora-
tory procedures by incorporating the following recommendations:
a. Expand the usefulness of the nine-compartment model by making the
following modifications:
• Develop sets of input data that define compartment size, flow rates,
microbial populations, and the like for a larger variety of water
bodies to make it easier for nonspecialists to use the model.
Introduce equations describing interactions among phenomena such as
pH, temperature, light intensity, and turbidity to allow more real-
istic and sophisticated simulations.
•'• ••• refinements have been listed in roughly the order of discussion In the
1 • M- i.
-------
b. Obtain additional data concerning the relationship of substrate solu-
bility and jn-octanol partition coefficient and sediment organic con-
tent to the sorption partition coefficients on natural sediments.
These correlations may provide simpler techniques for estimating
partition coefficients for some types of compounds under different
environmental conditions.
c. Develop improved experimental procedures for measuring volatilization
rates of very volatile materials under conditions of low turbulence.
d. Use radiolabelled substrates for measurement of sorption partition
coefficients, biodegradation rates, and metabolites to improve the
precision of the analytical techniques at very low, environmentally
realistic concentrations.
e. Increase the variety of organisms in conducting biosorptions, inclu-
ding phytoplankton and protozoa. Selected protozoa could be added
after a predetermined sorption period, and the viability of the pro-
tozoa could serve as an indicator of potential biomagnification
dangers of a pollutant and its metabolites.
f. Develop refinements to the pvocedures described in this report to
increase the likelihood of obtaining cultures capable of degrading
recalcitrant substrates:
Increase the number of locations for sampling and take samples
during different seasons.
Maintain incubation temperatures at the temperature of the sample
at the time of sampling.
With recalcitrant compounds, conduct some screening with pure
cultures or with young enrichment cultures isolated from aquatic
sources. Analog enrichment procedures could be used in the iso-
lation of these cultures, and fermentation conditions could be
somewhat different from those existing in the environment.
Replace the buffering salts used in the enrichment culture pro-
cedure by automatic addition of alkali or C0a to more realisti-
cally simulate environmental conditions.
-------
4. ENVIRONMENTAL ASSESSMENT
The objectives of the environmental assessment were to:
• Estimate the probable concentrations and distributions of
selected chemicals in aquatic systems resulting from continu-
ous discharge of low concentrations of these chemicals in indus-
trail waste and surface runoff.
• Assess the relative importance of photolysis, hydrolysis,
oxidation, volatilization, sorption, and biotransformation
in the removal of selected chemicals from solution in natural
freshwater systems, exclusive of transfer through food chains
or transformation on sediments.
To accomplish these objectives, it was necessary to make two fundamental
assumptions concerning the various possible environmental transport and
transformation processes in aquatic systems. These are:
• The overall rate of disappearance of a pollutant from
solution is controlled only by the transformation and
transport processes that were studied separately in the
laboratory and by hydraulic and hydrological processes of the
aquatic systems.
Each transformation and transport process can be studied
independently by laboratory experiments, and the results
of these experiments can be extrapolated to natural waters.
On the basis of these assumptions, laboratory procedures were used
co acquire data for, discrete physical, chemical, and biological processes
Hint, arc believed to be important in aquatic systems by using solutions
of the selected chemicals in pure water below their solubility limits.
The data from laboratory studies were integrated by a computer model that
can simulate streams, ponds, and stratified lakes by suitable combinations
o: compartments and hydrologic parameters.
Tn its simplest application, the model is fixed as one compartment,
•>:ui all data on transport and transformation processes are put in the form
"i .simple first-order relations in which only the concentration of the
>'> nical is a variable. Rate constants and reactive environmental inter-
"••duucs are lumped together as constants typical of a specific water body.
•'•!•• simple and preliminary assessment provides a good method of evaluat-
'•"'•'. l he relative importance of different transformation and transport pro-
.mcl thereby eliminating additional laboratory studies on subordinate
-------
More elaborate analyses, based on our multicompartment computer model,
•..•:v:i allows for the heterogeneity of actual water bodies, give somewhat
OiMtor accuracy and greatly facilitate computation of the pollutant concen-
::.i:ions in different parts of the simulated water bodies.
The following subsections present the assumptions that are specific
to c:ic mathematical formulations of the models used and discuss the compu-
:.itional approaches in detail. Subsection 4.1 discusses the computation
.1: overall transformation rates, and subsection 4.2 discusses the multi-
wv.-;Mrtment model developed for this study. The assumptions that are in-
Jcacndent of the mathematical formulations are presented in subsection. 4.3.
.. 1 ONE-COMPARTMENT MODEL
Analysis of. the data under the assumption of first-order kinetics, re-
forrod to hereafter as the one-compartment model, assumes that the system
m A single, completely mixed reactor from which the chemicals disappear
:-irjut;li transformation and transport. This model allows analysis of acute
lilijhnrges such as spills or deliberate use of pesticides and was used to
fsc.ibl Lsh priorities for detailed laboratory studies.
...1.1 Assumptions of the One-Compartment Model
The equations of the one-compartment model make the following
i-»Mi~i|>cions:
The water body is homogeneous with respect to all physical,
chemical, and biological properties.
Chemical, physical, and biological properties (other than
changes in the concentrations of the pollutant and solid
masses within the compartments) remain constant.
• The effects of physical, chemical, and biological variables
such as temperature, pH, and species composition are included
implicitly in the rate factors used in the simulations, but
are otherwise excluded.
Exogenous environmental parameters such as sunlight intensity
are constant for a given water body.
• The pollutant is introduced as a pulse at time zero.
• Sorption occurs only between solids and solution and between
solution and biota; no sorption occurs directly between
the biota and solids.
• The sorption equilibrium is rapidly established compared with
all other transformation and transport processes. The relative
proportions of sorbed and dissolved chemical are those calcu-
lated from the equilibrium constant or partition (sorption
partition coefficient) for the chemical between water and a
natural sediment.
10
-------
• A portion of the microbial population is acclimated to the spe-
cific substrate at all times. The rate of microbial transfor-
mation is a function of the number of acclimated microbes and
is first order with respect to substrate concentration.
• The microbial yield factor is constant.
The assumption that an acclimated microbial population is present
implies that the pollution is chronic or that it consists of repeated, dis-
crete releases of the chemical. Since several hours or days may be required
for acclimation, the model tends to overestimate the loss rate of chemicals
from solution in cases of acute pollution caused by spills. Also, the assump-
tion that the water body is homogeneous precludes appraisal of the distribution
of the chemical within large, incompletely mixed water bodies, such as large
stratified lakes. However, this model is a useful approximation of transport
and transformation in small water bodies such as ponds and has been especially
useful as a preliminary assessment tool to establish priorities for the de-
tailed laboratory studies.
A. 1.2 Mathematical Formulations
Given the assumptions listed above, each process may be described by
a first-order or pseudo-first-order rate law:
R.J = kjtS] (4.1)
where RJ is the transformation or transport rate for process j , ki is the
first-order or pseudo-first-order rate constant for process j, and [S] is
the concentration of substrate. If k. is a pseudo-first-order rate constant,
then J
k. = k? [E] (4.2)
J ZJ
where ^i is the second-order rate constant for process j , and [E] is the
concentration of the environmental component. We have assumed that the net
rate of loss of substrate from the water body is
R = JR.- S k. [S] (4.3)
j J j 3
The half-life of a chemical in any first-order process (t ,_) is
U1/2> . - IjU (4.4)
Also, since all the transformation and transport processes have been expressed
as first- or pseudo-first-order rate expressions, it is possible to calculate
an overall or net half-life for the pollutant, since
11
-------
In 2
Z k.
j J
To be realistic, allowance should be made for flow of a pollutant out of the
system by adding a rate constant for the movement of water through the system.
Thus, equation (4.5) becomes
t = ln2
1/2 k, + E k. (4.6)
d J J
where kj is the dilution rate constant for the system as a whole, kj is de-
fined as the mass of chemical per unit time in the outflow divided by the
total mass of chemical in the system. Equation (4.6) shows that the effec-
tive half-life is determined by two terms: the dilution rate constant and
the sum of transformation rate constants. If the system is assumed to be
completely mixed, the system dilution rate than becomes the outflow rate
divided by the total volume. Two different cases of water movement are dis-
cussed: In the case of zero or low dilution rate, equation (4.6) becomes
- ln 2
tl/2 ~ Z k. (4'7)
In the case of rapid dilution rate, equation (4.6) becomes
'1/2 ' V C4'8)
We have further assumed that the value for exogenous parameters [E]
for any particular process will differ in different water bodies and will
affect the values of kj. The details of the differences we have assumed are
given in subsection 4.3.
4.2 NINE-COMPARTMENT MODEL
*
The nine-compartment computer model was designed to explore the impact
of water body heterogeneity on the transformation and transport mechanisms
covered in the laboratory phases of this project (Figure 4.1). Transforma-
tion of sorbed chemicals and accumulation of chemicals within food chains
were not considered, but the model does permit assessment of the variations
in concentrations to which specific segments of food chains will be exposed.
*
The computer program for this model has been modified to accommodate up to
99 compartments, to allow for three-dimensional simulations, and to allow
continuous flows between compartments. However, for the sake of consistency
of methodology, the nine-compartment, two-dimensional, batch-flow version
described here has been used throughout this study. The assumptions under-
lying the two versions of the model are otherwise unchanged.
12
-------
VOLATILIZATION
HYDROLYSIS
PHOTOLYSIS
OXIDATION
P' • nvmATtnw
1
1
)i • PHfYmi YQI C
fH- BIODEGRADATION
Note : Items to the right of the dotted line are not used in current versions of the model
FIGURE 4.1 TRANSPORT AND TRANSFORMATION ROUTES SIMULATED
-------
4.2.1 Assumptions
The model contains nine compartments of arbitrary size and composition,
which are used to represent segments of the water column or sediments. By
selectively removing compartments and appropriately adjusting the constants
for water quality and water movement parameters, we can use the model to sim-
ulate streams, ponds, or stratified lakes.
Inputs can be made to any compartment, allowing simulation of atmospheric
inputs, offshore outfalls, or accidental spills. For this study, we have
assumed that inputs are restricted to a single surface compartment.
Transfers between compartments are assumed to be dominated by rates of
water or solid particle movement and hence are assumed to be specific to the
ecosystem rather than to the chemical. Solid and solution flows are allowed
between compartments, and they can flow at different rates, but the biota re-
main in place. Inflows of new solution and solids are allowed as are equiva-
lent outflows of solution and solids. Estimates of these rates of transfers
between compartments are based on published data.
Sorption and desorption are allowed within compartments in the model
and can occur between solution and solid particles and between solution and
biota. Although exchange of sorbed substrate between solid particles and
biota can be important, it has been excluded because it is not being measured
in our current project. No distinction is made between organic and inorganic
particles.
Calculations are made on a time-sequenced basis. We assume that flows
and mixing occur at the end of each time interval and that volatilization,
sorption, and transformations occur within each time interval. A short time
interval, generally less than 0.2 hour, is used in the simulations.
The model assumes that a number of base conditions are constant within
each compartment but may vary between the nine compartments. These conditions
include temperature, pH, light, mass and species of biota, and so on, the
effects of which are included implicitly in the input, transformation, and
soiption rate factors. Changing base conditions, such as day to night, can
be approximated by several sequential computer runs in which the output of
one run regarding substrate concentration in the daytime is input to the next
run where rate factors are modified to represent the nighttime conditions.
An initial concentration of the substrate can be arbitrarily assumed in
cither the solutions, solids, or biota in any of the compartments. Loading
raLes of the chemicals and suspended solids can be input as the specifications
of the system of study.
Rate constants for chemical and biological transformation and volaril-
izat-ion measured in the laboratory are used directly when feasible or, when
necessary, they are used to estimate the rate constants for combinations of
temperature, light, and acidity that were not appraised in the laboratory.
-------
Laboratory data for hydrolysis, oxidation, and volatilization are used
directly or adjusted by empirical equations or coefficients. Data for sorp-
tion are used directly even though there is some uncertainty in any extrapo-
lation of the complex and poorly understood phenomena collectively known as
sorption. Environmental parameters related to sorption, volatilization,
photolysis, and biodegradation are adjusted subjectively to approximate the
net effect of the numerous differences between laboratory and field environ-
ments. A simple coefficient is used to adjust the photolysis rate constant.
Biotransformation rates are adjusted by varying the number of bacteria as-
sumed to be acclimated to the chemical of interest.
The complete structure of the model is presented schematically in Fig-
ure 4.2. The arrows indicate direction of flows and the rectangles represent
the three-dimensional compartments. The model can direct flows between any
pair of compartments (for example, from 4 to 8); however, those illustrated
are the ones typically used. Compartments 1, 2, and 3 are generally used to
represent surface waters; 4, 5, and 6 deep waters; and 7, 8, and 9 sediments.
However, these compartments can be used in other ways if appropriate. The
configurations of compartments actually used in this study and the dimensions
assumed are shown in Figure 4.3.
Each compartment of the computer model can be considered as a completely
mixed batch reactor. Transformation of a pollutant follows its transforma-
tion kinetics during the simulation time interval, and masses of the pollutant
interchange among the compartments and between the aqueous and solid phases
within compartments between each simulation time step.
The nine-compartment model lacks the assumptions of overall homogeneity,
irreversible sorption, episodic discharge of pollutants, and first-order ki-
netics for biodegradation used in the one-compartment model, but shares the
remaining assumptions listed in subsection 4.1.1. In addition, the nine-
compartment model assumes that:
Inputs, outputs, and transfers between compartments are limited
to nonliving solids and solutions.
• Movements between compartments occur in discrete time steps.
• The contents of each compartment are thoroughly mixed after
each intercompartment transfer.
Volatilization, sorption, desorption, and transformation occur
simultaneously within each compartment within each time interval.
Chemical reactions are assumed to be pseudo-first order in pol-
lutant concentration; biodegradation is assumed to follow Monod
kinetics.
• Degradation is assumed to occur only in the liquid phase or in
(or on) microorganisms.
15
-------
VOLATILIZATION LOSSES
INFLOWS
OUTFLOWS
FIGURE 4.2 SCHEMATIC OF ASSUMED FLOWS BETWEEN
COMPARTMENTS IN AN AQUATIC SYSTEM
16
-------
POND
0-05" [SMIMlMlSi
oH
1000 m —:—-I I- [000 m
-lOOOra-
^:£ 3
="£ RIVER
100 m
|- 500m-
|- 100 m »]
500 atj
{=»
5 m
i
J_
<^:=r
:/---: ^>
T
/ (71 _
^$sg9g$$jg
2 IOm
/
i=r
t *
IOm
3 5m
r::> LAKE
I 1 WATER COMPARIMEKT
lx\'-SS^l SEDIMENT COMPARTMENT
SA-4396-13
FIGURE 4.3 PHYSICAL CONFIGURATIONS OF THE POND, RIVER, AND
LAKE SIMULATIOMS
17
-------
In the current version of the computer program, bottom sediments
are half water and half solids by volume.
• Outflows of solids are based on the ratio of solution in the
compartment from which the outflow occurs.
Inflows of solids are equal to outflows plus losses to the
sediments.
The model also requires explicit assumptions for the environmental components,
including physical dimensions and water quality, which are presented in
Section 4.3.
A.2.2 Mathematical Formulations
The mass of pollutants in the aqueous phase and suspended solid phase
in each compartment during the simulation time interval (At) is determined
by the concentration of the external inflow and outflow, and interflows among
the compartments. The mass of pollutants in the aqueous phase of compartment
i after a time step can be written as follows:
MU = I(At)i + Z Mfcii + MU ~ E MMi (4'9)
in J out J
where I(At) is the external input of pollutant to compartment ± during the
simulation time interval, M is the original mass in the compartment, £ Mn..
X.1 in Aj i
is the mass of pollutant in the aqueous phase added from the jth adjacent com-
partments, and I M . is the mass of pollutant in the aqueous phase that
out ^
flowed out of compartment i to compartment j.
For the pollutant adsorbed on the suspended solids, a mass balance
similar to equation (4.9) is written as follows:
Ms. = IS(At). + E M + M - I M (4.10)
in J out
vhere IS(At) is the external input of solids to compartment i during the sim-
ulation time interval, E M .. is the mass of pollutant in the suspended solids
in ® J
•iddcd from the jth adjacent compartments, and Z M .. is the mass of pollutant
in the suspended solids that flowed from compartment i to compartment j.
Because many transformation processes may occur simultaneously in the
•••:--MLic system, a function f^ is defined as the total transformation rate of
i.'i1 pollutant, on a mass basis, in the aqueous phase in compartment i.
dMi±
£ii ' -dT ' (rP + rh + ro + rv + rb}ivi (
18
-------
where
M' . is total mass of the pollutant in
aqueous phase (£,) of compartment i
r is photolysis rate
r, is hydrolysis rate
r is oxidation rate
o
r is volatilization rate
v
r, is biodegradation rate
V. is the volume of compartment i.
The photolysis rate is expressed as follows:
. (4.12)
where k is the photolysis rate constant (Section 6.2} and [S]^ is the chem-
ical concentration in the aqueous phase in compartment i. Note that kp is a
function of quantum yield, absorption spectrum, and solar irradiance; *p will
vary with time of day, season, and location.
The hydrolysis rate is written as follows:
. (4.13)
-hero k^ is the hydrolysis rate constant, which depends on the temperature,
lii1"], and {OH~] in the aquatic system.
The oxidation rate is expressed by the equation:
ro = Will (4'14>
-:ic>rc kOx is a pseudo-first-order rate constant equal to k( [RO^] ^n which
ciit: concentration of R02' is fixed at 10~9 M (Section 6). Oxidation by other
o\ldants such as HO- or 63 would follow similar kinetic relationships but were
not measured here.
The volatilization rate is proportional to the difference between the
''<.TMc.il. concentrations in the aqueous and the air phases:
r = k ([S)n. - [S] ) «
V V &X £
the mass transfer rate constant and [S]g is the chemical concen-
'•: 10.i in the air phase. In a normal atmospheric environment [S] is usually
'•""••ill Lliat It can be assumed equal to zero. Therefore, equation (4.15)
•' • • lo
(4.16)
19
-------
Clearly, k§ should depend on the surface area, wind speed, air and water
temperature, and so on. Laboratory measurements of k| for chemicals have
been described elsewhere (Hill et al.» 1976). The ratio of the gas transfer
constant of chemical to that for oxygen is constant for a wide range of tur-
bulence conditions. The gas transfer constant of chemicals in various water
bodies can be estimated if the oxygen reaeration constants in the correspond-
ing water bodies are also known (Section 5.3).
The biodegradation rate is described by the following equation:
b Y (Kg + [S]£i) Ai ^
where umax, Ks, and Y are the kinetic constants of the Monod expression (Stumra-
Zollinger and Harris, 1971; Monod, 1949). Mmax is defined as the maximum growth
rate, K is the half-saturation rate, which is defined as the pollutant
concentration at one-half of the maximum growth rate, and Y is the yield
factor, which describes the efficiency of converting chemical mass into
microbial mass. In the model, [X-jJ, the microbial mass or concentration,
is considered to be an environmental parameter and is assumed to be constant.
Sorption of the chemical on the suspended solid particles and the sus-
pended biota (Section 5.4) is assumed to be an equilibrium process.
The partition coefficient for distribution between the suspended sedi-
ment phase and the aqueous phase is defined by:
= mwi Msi
*" msi MU
where m^ is the mass of water and ms^ is the mass of suspended sediment in
compartment i. Msi and Mj^ are the masses of chemical in suspended sediment
phase and the aqueous phase, respectively.
Similarly, the partition coefficient for distribution between biota
•n-nd aqueous phase is:
m . MR.
mBi ii
v.-liore mB^ and Mg^ are the masses of biota and chemical in biota, respectively.
"•nice it is assumed that no biodegradation or chemical transformation processes.
t-.ikc place on the surface or inside the sorbent phase, the total mass of
Pj'Uiitant in a compartment before and after sorption will be the same.
'•'^roforc, the relationship of the total mass of chemical in the aqueous
:>:MSC (M-ii), the suspended sediment (Msi), and biota (Mg^) before (primed)
i ic! .iFter (not primed) sorption can be written as:
M' + M' . + M; . = M.. + M . + M_ . (t
Hi si TJi £,1 si Bi v
20
-------
The mass distribution after sorption, therefore, can be calculated by solving
equations (4.18), (4.19), and (4.20).
The relative amounts of bacteria and sediments found in the water
column and sediment layer of eutrophic and oligotrophic water bodies are
summarized in Table 4.1.
TABLE 4.1. RATIO OF BACTERIA TO SEDIMENTS
IN NATURAL WATER BODIES
Weight bacteria/weight sediment3
Water column Sediments
Entrophic 1 x 10"s 1.6 x 10~2
Oligotrophic 1 x 10~7 to 1 x 10~6 1.6 x 10"* to 1.6 x 10~3
•a
Dry weights.
Sources: Tables 28 and 31 of Kuznetsov (1970), pp.
592-595 of Wetzel (1975), and an empirical
bacteria density to dry weight conversion fac-
tor of 4 g (dry weight) per 1013 cells.
The estimates show that even if the biosorption is ten times as much
as the sorption on sediments, the bacteria contribute little to the total
amount of substrate sorbed. The model developed in the study, therefore, ex-
cludes the biosorption from the calculations. The mass distribution after
sorption can be directly calculated from equations (4.18) and (4.20) with
omission of the Ml. and Mg. terms in the equations.
To integrate equation (4.11), the predictor-corrBctor method (ES0DEQ)
(Rollins, 1968) is used. ES0DEQ uses a four-point Adams-Bashforth-Moulton
predictor-corrector method to carry out its integration. The predictor-
corrector can be executed only if four points are available. To estimate
the first four points (including the initial point), the fourth-order Runge-
Kutta method is chosen because it is quite accurate over small intervals.
ES0DEQ also provides a feature whereby the integration can be carried out
only with the Runge-Kutta method by selecting an appropriate control index.
The functional relationships of the computer program subroutines are
JJlustrated in Figure 4.4. The sequence for the time interval At is:
(1) Calculate the mass of chemical transformed.
(2) Calculate the distribution of chemical between the sediment
and aqueous phases, according to the equilibrium partition
coefficient.
21
-------
D.
IU
fc
IU
Z
X
UJ
o
o
INPUT DATA
1. SYSTEM PARAMETERS
2. TRANSFORMATION RATES
• CHEMICAL
• PHYSICAL
• BIOLOGICAL
DETERMINE THE MASS
CHANGES IN EACH
COMPARTMENT DUE
TO TRANSFORMATION
PROCESSES
USE
NUMERICAL
INTEGRATION
SUBROUTINE
DETERMINE
RATE OF
CHANGES
CALCULATE MASS
BALANCE ACCORDING
TO CHANGES OF FLOW
CALCULATE MASS CHANGES IN
AQUEOUS AND SOLID PHASES
DUE TO SORPTION
PLOT RESULTS
( STOP J
FIGURE 4-4. FLOWCHART FOR THE ENVIRONMENTAL ASSESSMENT MODEL
22
-------
(3) Calculate the new masses in each compartment based on the
assumed inflow-outflow conditions.
4.3 ENVIRONMENTAL PARAMETERS
The constants used to define the volumes of the compartments, the fluxes
of materials, and the inflow of pollutant at a concentration of 1 ug ml"1 are
given in Table 4.2 (input pollutant concentrations are set at 1 ug ml'1 or
below the solubility, whichever is lower). The values used to define the water
quality characteristics that directly affect the transformation and transport
processes studied in the laboratory are also given in Table 4.2. The values
of the constants presented in this table are based on values given in Wetzel
(1975) and Leopold et al. (1964), except for the reaeration rates, which are
discussed in Section 5.3.1.
TABLE 4.2. PHYSICAL DIMENSIONS AND WATER QUALITY CHARACTERISTICS
ASSUMED IN THE ENVIRONMENTAL ANALYSIS
_ Eutrophic Oligotrophic
Parameter River Pond lake lake
Poysteal dimension
Total water volume (m3) 9 x 10s 2 x 101* 5.5 x 106 5.5 x 106
Inflow (m3 hr"1) l.OxlO6 20 9.7 x 102 9.7 x 102
Moan residence time (hr) 8.3x1CT1 1.0 x 103 5.7 x 103 5.7 x 103
PoUutant inflow (kg hr"1)3 l.OxlO3 0.02 0.97 0.97
aLor quality
Total bacteria (cells ml"1)
Active bacteria (cells ml"1)
pH
Sc-diTient loading (i^g ml"1)
5'hotolysis activity index'3
Oxvgen reaeration rate (hr""1)
(Rl)..-] (M)
106
105
7
100
0.5
0.04
10"«
106
10s
8
300
0.2
0.008
10-9
106
105
8
50
0.2
0.01
10- 9
102
10
6
50
1.0
0.01
10"9
:he flow rates between compartments are given in Appendix A.
to account for differences in light transmission through different
of water. Distilled water has an index value of 1.
23
-------
5. PHYSICAL PROPERTIES
Four physical properties of each substrate were measured as part of this
study: solubility in water, ultraviolet (uv) and visible absorption spectra,
volatilization rate constant, and sorption partition coefficients. The solu-
bility in pure water must be known because the substrate must be in solution
for meaningful rate constants and sorption partition coefficients to be
measured. The solubility is also used to estimate volatilization rates. The
uv/visible absorption spectrum is used to estimate the photolysis rate. Vola-
tilization, photolysis, and sorption are possible important environmental
pathways for the substrate.
5.1 SOLUBILITY
The general procedure described by Campbell (1930) is simple to use to
-icasure the solubilities of solids in the ug ml"1 range. A small amount of
die solid substrate is placed in an all-glass apparatus, which is immersed in
a water bath and shaken gently. This apparatus has two compartments separated
in a glass frit. When the flask is inverted, the aqueous solution is filtered
through the frit to remove solid substrate. The filtration step can be
tarried out without removing the apparatus from the water bath. After equili-
bration in the water for several days, the sample is filtered in the apparatus
and the filtrate is analyzed for the substrate.
Generally, at least three measurements are made. Also, measurements are
-ado on samples that have been heated to 35° to 40°C, allowed to equilibrate,
and then cooled in the water bath. Since the concentrations are fairly high,
potential problems that are encountered with low-solubility materials, such as
adsorption onto the frit during filtration and the possibility that finely di-
vided particulate substrate is not removed during filtration, are not -likely
LO be significant.
Solutions of compounds having a solubility in the ng ml-1 range were pre-
pared by the procedure described by Haque and Schmedding (1975). The sub-
strate is dissolved in an organic solvent and put on the walls of a 5-gallon
fL8.8-liter) carboy. The carboy is rotated slowly on its side while the solvent
cv.iporat.es, so that a thin film of substrate coats the wall of the carboy. A
l.-irj'o Teflon-coated magnetic stirring bar is added, and the carboy is filled
-nil the purest water available. Care must be taken to prevent the substrate
'run codling the bottom of the carboy so that it is not dislodged by the stir-
: !ng bar. The solution is allowed to stir gently for at least a week to assure
i'i.'-t equilibration has taken place. Samples of water are withdrawn with a
•l-i^s siphon and analyzed for the substrate.
24
-------
Tn several cases, we found that even with these precautions particulate
-utter, presumably substrate, could be observed in the carboy, and the sub-
-.Lratc concentration was reduced by centrifugation at 10,000 rpm. The solu-
bility measurements of low-solubility (less than 0.1 |jg ml"1) compounds should
nc made on centrifuged samples.
5.2 ABSORPTION SPECTRA
The absorption spectrum of the substrate is measured to determine if
photochemical transformation or degradation is possible. If the substrate
LIOCS not absorb light in some region of the solar spectrum, then direct photo-
chemical transformation in the environment is not possible. Sensitized
photolyses may also be possible, but their importance is quantitatively as-
-,cssed by experiments with humic acid (Section 6.2).
Zepp and Cline (1977) and Wolfe et al. (1976) have developed a computer
program that will calculate the direct photochemical transformation rate of a
.substrate, provided the substrate absorption spectrum and transformation
(luantum yield and the solar spectrum are known (see Section 6.2 for details).
This program requires the average molar_ extinction coefficient and the solar ir-
radiance for specific wavelength intervals, which are listed in Table 5.1.
TABLE 5.1. NOMINAL WAVELENGTHS AND WAVELENGTH INTERVALS
FOR UV AND VISIBLE ABSORPTION SPECTRA
Nominal
wavelength
(nm)
Wavelength
interval
(nm)
297.5 ± 1.25
300.0 to 320.0 ± 1.25
323.1 + 1.9, - 1.85
330.0 ± 5.0
340.0 and higher ±5.0
These numbers represent the wavelength intervals used by the computer
(Wolfe et al., 1976). The actual precision of a measured wavelength
is about ± 0.5 nm.
25
-------
Most measurements of absorption spectra reported in the literature have
focused on the location and intensity of the absorption maxima. Since the
solar irradiance rises rapidly between 295 and 350 nm, the low absorption
"tail" that is often present in molecules at the longer wavelengths can have a
significant contribution to the photolysis rate. Therefore, it is necessary
to measure the magnitude of the absorption tail as well as the more incense
portions of the absorption spectrum.
It is well known that the absorption spectra of many compounds are
slightly different in polar and nonpolar^solvents. Since the substrate will
be in water in the aquatic environments, we have used pure water as the sol-
vent, for measurement of the absorption spectra whenever possible. It is im-
portant to ensure that the absorption tail of low-solubility substrates can be
-icnsured accurately; the concentration of the substrate should be in the 10~z
10 10"6 M range. Therefore, it is often necessary to increase the solubility
;>y adding a water-soluble cosolvent, such as acetonitrile. Since acetonitrile
i.s also very polar, at concentrations less than about 20% by volume, acetoni-
trile does not significantly affect the absorption spectrum. Also, the ab-
sorption spectium should be measured in 1-cm and 10-cm cells; the 10-cm cell
LS necessary to maximize the precision of the measurement of the absorption
uill at low substrate concentrations. The general procedure has been to pre-
pare a substrate solution of a minimum of about 10~5 M in water, using
acetonitrile to dissolve any solid, undissolved substrate. Obviously, some
trial and error is required to minimize the amount of acetonitrile. To obtain
satisfactory spectra with 10-cm cells, we have always run solvent versus sol-
vent to obtain a baseline. "*" Then, the cell containing solvent in the sample
beam is refilled with the substrate solution, and the absorption spectrum is
.'K-asured. The molar extinction coefficients e. are obtained from Beer's law.
A
Absorbance = - log — = e IS (5.1)
J. A
O
Uicre IQ is the incident light flux and I is the transmitted light flux in the
spoctrophotometer. If 1, the cell path length, is in centimeters and if S,
the substrate concentrations, is in M, then the molar extinction coefficient E
i^ in units of cm"1 M"1. The average molar extinction coefficient for each
nominal wavelength is calculated from the average of the molar extinction co-
ofEJcients at the lower and upper limits of the wavelength interval (Table 5.1)
r
This may not be true for substrates that are sorbed onto sediments or biota,
especially since absorption spectra are known to shift in some cases when
nolccules are adsorbed.
'A1 h.ive used Gary model 14 and 15 spectrophotometers in these studies. How-
fVL-r, any high quality, uv/visible spectrophotometer that will accept 10-cm
colls can be used. SuitabJe standards should be prepared to assure that the
'"imc absorbancc is obtained in the 10-cm cell with a solution that is onc-
l entli the concentration used in the 1-cm cell.
26
-------
,.j VOLATILIZATION RATE
.3. L Background
To assess the importance of volatilization as a pathway for pollutant
>ubstrates in natural water bodies, it is desirable to have an expression of
the form
X (5.2)
s
v-iere kv is the volatilization rate of the substrate S. Compounds of low
-inlecular weight and high vapor pressure, such as vinyl chloride (Hill et al.,
1976), have been shown to volatilize rapidly as one might expect. However,
•jomc high molecular weight, low solubility compounds, such as DDT (Acree et
.1] . , 1963), also volatilize at an appreciable rate, since the Henry's Law con-
-.trint for these compounds is very high because the activity coefficients are
,ilso very high. The details are discussed in Appendix B.
Several authors have suggested ways to estimate volatilization rates of
compounds from water, using theoretical considerations and laboratory measure-
meats. Mackay and Wolkoff (1973) have suggested simple equations that can be
used to estimate the volatilization rate of an organic solute from a water
body under certain conditions. When their assumptions for the evaporation
rate of water, etc., are used and a 1-m depth for homogenous mixing is used,
Lticir equation (10) for the volatilization half-life of substrate in a repre-
sentative lake reduces to
0.108 S ,.
Vdaye) -- ^-jj-5^ (5-3)
s s
where Pg is the vapor pressure (torr) , Mg is the molecular weight of the sub-
strate, and Ssat is the solubility of the substrate (yg ml 1) . While esti-
mates using this equation are simple to make, the assumption of a 1-m depth
for homogeneous mixing is a serious deficiency. In most water bodies, mass
transfer across the boundary layers, which is not accounted for in equation
(5.3), is the rate-determining step for volatilization.
Mackay and Leinonen (1975) recognized the problem with assuming homo-
geneous mixing and developed equations that included the mass transfer across
the boundary layers (equations B.3, B.4, and B.5 in Appendix B of this
report) . To use this method it is necessary to measure the mass transfer
coefficients and the Henry's Law constants. Mackay and Cohen (1976) have
described several methods for measuring these values in the laboratory.
However, these maasurements require a special apparatus and some experi-
mental care. The principal difficulty with this approach is that there is
no way to relate the mass transfer coefficients determined in the laboratory
to those in the real body of water with varying wind and water flow condi-
tions. There is no convenient way to measure the mass transfer coefficients
directly in the real water bodies.
27
-------
Tsivoglou has made an important observation. He showed that the ratio
,.- [.lie volatilization rates of several low-molecular weight gases from water
constant over a wide range of turbulence conditions (Tsivoglou et al.,
,i03; Tsivoglou, 1967). Thus, for compounds A and B, the ratio
= constant (5.4)
kv
It is convenient to choose the substrate for A and oxygen for B. The law of
r.icroscopic reversibility requires that the rate of volatilization equal the
rate of dissolution into the liquid for identical conditions. The term,
oxygen reaeration rate, is commonly used to express the rate at which oxygen
from the atmosphere dissolves in oxygen-deficient water. The oxygen reaera-
cion rate is defined by
(I0»]8ftt- 10,1) (5.5)
where [02] is the oxygen concentration, [02]sat is the oxygen concentration
when the water is saturated, and kQ is the oxygen reaeration rate constant.
The oxygen reaeration rate has the additional advantage that it has been
measured for many different water bodies. Values for representative water
bodies are given in Table 5.2.
TABLE 5.2. OXYGEN REAERATION RATES IN REPRESENTATIVE WATER BODIES
Values used in
Literature values this study
(day"1) (day"1) (hr"1)
Pond
River
Lake
0.11
0.2b
0.10
- 0.23a
, 0.1 - 9.3C
- 0.30a
0.19
0.96
0.24
0.008
'0.04
0.01
^letcalf and Eddy (1972).
Grenney et al. (1976).
r*
Langbein and Durum (1967); taken from Table 2 for rivers such as the
Allegheny, Kansas, Rio Grande, Tennessee, and Wabash.
Therefore, if the ratio of the substrate volatilization rate to the
oxygen reaeration rate constant can be measured in the laboratory, the
28
-------
rate of the substrate in a real water body for which the
reaeration rate is known can be estimated:
(kS) L „ = (k°) . J (kS/k°)n , (5.6)
v'water body v'water body v' v' laboratory v '
.:L11 et al. (1976) used this procedure successfully to estimate the volatili-
sation rate constants for vinyl chloride.
The purpose of Appendix B is to tie all the theoretical work together
.inci co show that the simple relationship
k o
•
0 S
.here d and d are the molecular diameters of 02 and S (assuming that they
.-ire spherical), can be used to estimate values of k/k at the low turbulence
values likely to be found in natural water bodies. In Part II of this report,
this theory will be compared with the laboratory measurements.
5.3.2 Experimental Procedures
Volatilization rates were measured using the method described by Hill
et al. (1976). A solution of the substrate in pure water is prepared at a
concentration that is below its saturation value. About 1 liter of this solu-
tion is placed in a 2-liter beaker equipped with a stirring bar. The solution
is purged with nitrogen to remove most of the dissolved 02. At the start of
the experiment (t = 0), the concentration of substrate is measured"' and the 02
concentration is measured with an 02-analyzer. Successive substrate and 02
measurements are made at regular time intervals.
The substrate concentration versus time data are fit to an exponential
decay curve of the form
[St] = [Sje (5.8)
which is the integrated form of the first-order rate expression
Usually, an aliquot was removed and saved for subsequent extraction or direct.
ana]ysis.
29
-------
The oxygen concentration data are fit to the integrated form of equation
(5.10)
^-fcv^aat- l°2]0 C5.10)
which is
t0*}t= I0*lsat-(t02]sat- r°*0e~kvt (5'11}
•*here [02]sat is the saturation concentration of 02 in water at the tempera-
ture of the measurement and is a constant because the concentration of 02 in
:he air is constant.
0 S
To calculate values of kv and kv, we used the linear least squares
routine supplied with the Hewlett-Packard Model 65 calculator. This program
i;ives a linear least squares fit to InS versus t, plus the variance of
the parameter estimates. The value of k^/k^ was calculated from these values.
Other curve fitting procedures could also be used.
Experimental problems arise if ky is either very low or very high. If
kv is low, evaporation of water becomes significant. When that happens, the
Fixing rate in the beaker changes because the stirring bar, which rotates at a
nearly constant rate, imparts more turbulence to the solution as the amount of
vratcr decreases. Therefore, the value of k^ must be measured several times
and the variance in k~ is significantly larger.
If kv is very high, as was the case for benzo[b]thiophene , the sub-
strate volatilization rate can be comparable to the 02 reaeration rate. In
that case, the stirring rate must be reduced to bring k§ within a range that
can be measured. If the stirring is too low, the solution becomes inhomogene-
ous and the data do not fit the theoretical scheme above. These problems were
overcome by using a reasonable stirring rate but reducing the liquid surface
area exposed to the atmosphere. A 1-liter Erlenmeyer flask, filled nearly to
the rim, was used in the volatilization experiments with benzo[b]thiophene.
'•otc that the natural logarithms of [S] or ([02] safc - [02]t) must be entered,
1bviously, the exponential curve-fitting routine could also bo used, but it
'!r»L>s not provide the variance of the parameter estimates.
30
-------
SORPT10N OF ORGANIC SUBSTRATES
, :. . L Background
Sorption of organic substrates onto sediments and biota can be a very
.urtant phenomenon in the aquatic environment. The sediments can act as
'^a for sorbed materials, removing them from the water column. However, the
Derate can also be released (desorbed) from the sediments at a later time.
•^ this way sorbed material can also be a source of pollution. Sorption of
: lutants by sediments often results in high concentrations of low-solubilitj
•vLLutants in a part of the water column where uptake by biomagnif ication may
significant.
Three types of sorbents have been evaluated in this study: a montmo-
::ll(>nite clay, several natural sediments, and a mixture of bacteria. The
l.iv was used as a reference sorbent because it could be readily obtained and
it-pared by other workers. The sediments were collected from a variety of
, >-irtes chosen to represent different types of freshwater bodies in different
IJ-L.S oC the United States. The bacteria cultures were obtained from the
'..-'•-rican Type Culture Collection (ATCC) . They were chosen because they are
n-.nvsentative of the types of microorganisms found in freshwater bodies and
•..id not been exposed to the types of substrate being studied.
The low proportion of bacteria to other materials (such as clays, de-
•nius, humic substances) in both suspended and bottom sediments (see Table
...1) suggests that the bacterial population does not contribute significantly
:.) the total amount of sorption. The effects of bacteria in sediments on the
. >iption partition coefficients are already included because the procedures we
•vp used to collect and preserve the sediments are such that any bacteria
••riginally present will be in the sediment samples used to measure the sorp-
:i.m properties.
The separate biosorption measurements are useful, even if bacterial ad-
- >rption is not an important fate, because biosorption is often the first step
of biomagnif ication and is therefore important environmentally. Biosorption
studies indicate whether biomagnification could be an important pathway for
a particular compound. Sorption takes place, in general, when solutions
containing a dissolved substrate contact a solid phase surface. If the
Lotal amount of substrate is increased, the amount of substrate that is
sorbed is also increased.
Hie term "sorption" includes any type of process whereby the substrate is
I'livsjcally or chemically bound to a solid surface. The term "adsorption"
implies to us that the process that holds the substrate on the sediment is
•'trictly physical, such as the Van der Waals type of attraction. We have
ust'd Iho term "sorption" (or "biosorption" when the solid sorbent is a micro-
Tp.anism) throughout this report to avoid questions about the details of the
•iirption mechanism.
31
-------
Experimental data for sorption have generally been found to fit one of
-V.J niathematical forms: the Langmuir and Freundlich isotherms. First, Sw
,,j s are defined as:
weight substrate in solution ,.
w ml solution
weight substrate sorbed ,,. ._.
s g sorbent
3t equilibrium. The substrate weights must be in the same units (e.g., ng,
.^). For a dilute aqueous solution, 1 ml of solution equals 1 g of solution,
weight substrate in solution ,,.
w g solution
Langmuir isotherm equation is defined as
abS
s 1 + bSw
»iiere a and b are constants and
(5.15)
a = X (5.16)
c
K
b = / (5.17)
c
where Xc is the sorption capacity of the sorbent and K is a partition coeffi-
cient. Data for gas-solid sorption generally and data for organic substrates
sorbed on clay minerals usually fit the Langmuir isotherm. However, natural
sediments are not homogeneous—sorbed complex organic material such as humic
substances is already present on the clay particles—and sorption by natural
sediments usually fails to fit the Langmuir isotherm.
32
-------
Data for sorption of multiple substrates from solution on nonuniform
,uriaces generally fit the Freundlich isotherm, which is an empirical equa-
KS 1/n (5-18)
w
At low substrate concentrations, n is often very nearly equal to 1. If n = 1
and S and Sw are in the same units, the units of each side of equation (5.18)
cancel and K becomes a partition coefficient, as defined by
Ss = Vv (5.19)
equation has been used to describe the laboratory sorption data obtained
in this study and is equivalent to equation (4.18), which is used in the en-
vjronmental assessment models.
5 . 4 . 2 Sorption on Clays and Sediments
5.4.2.1 Clay and Natural Sediments Selected for This Study — The mont-
;norillonite clay used in these studies was a Wyoming montmorillonite obtained
from Dr. William John, Department of Geology, University of Missouri, Columbia,
Missouri. Clay suspensions (about 1% by weight) were prepared by soaking a
measured weight of clay in distilled-deionized water for at least one week.
The clay suspension was then passed through an ion exchange column that had
been presaturated with calcium ions to convert the clay from the sodium form
to the calcium form. This was necessary because the Na-montmorillonite sus-
pension could not be centrifuged but the Ca-montmorillonite could. The parti-
cle size was less than 1 ym.
The procedures that were used to prepare and store the natural sedi-
ments were designed to preserve the sediments i-i their natural state as well
as possible. Natural sediments were screened to remove large rocks, twigs,
and other debris. The mesh sizes of the screens used were 4, 16, and 28 per
2.54 cm. Following the screening, the sediment was mixed, using a Humbolt
splitter to be sure the sediment was uniform. A small volume of each sediment
was mixed with two volumes of 0.1 M calcium chloride and the pH was recorded.
The remainder of the screened and split sediment was stored in 100-ml Nalgene
bottles at 4°C until use. The sediments were never allowed to dry out, be-
cause the drying would change the characteristics of the clay-organic complexes
in the sediments. Similarly, no attempts were made to kill or remove the bac-
teria in the sediments. Therefore, the contribution of the sediment bacteria
*Hamaker and Thompson (1972) point out that the exponent in equation (5.18),
1/n, is ". . .an archaic remnant of an attempt to give the Freundlich
equation physical meaning and is retained only because its use is embed-
ded in the literature." We have continued this practice.
33
-------
co the total sorption properties of the sediment has been included in our data.
It is very difficult Co make reproducible transfers of whole sediments
L: the sediments are in suspension. The sand fractions settle rapidly, and
they constitute a significant portion of the total mass of many sediments.*
Therefore, we used only the sediment particles smaller than 100 urn. Our ex-
perience indicates that less than 100-um sediment can be reproducibly trans-
ferred as a slurry. The less than 100-um sediment was prepared from as-
ruceived sediment by screening and a final settling of 30 seconds to remove
coarse sand.
The characteristics of the Ca-montmorillonite clay and natural sediments
used in these studies are given in Table 5.3. The organic carbon (OC) values,
^pressed as percent carbon by weight, were determined using the Walkley and
Black procedure, which involves oxidation of the organic material by chromate
followed by back-titration with ferrous ammonium sulfate (Hesse, 1971). Other
nethods of determining OC values, such as combustion and determination of
evolved C02, could have been used and would most likely give different OC
values. However, the trend in the organic carbon levels in the sediments
s.nould not change.
5.4.2.2 General Procedures—Screening studies for sorption were made to
estimate the magnitude of the partition coefficient Kp. This value was used
in the one-compartment model to provide an estimate or the importance of sorp-
non as an environmental pathway. Biosorption studies were not carried out on
-.ubstrates that were not strongly sorbed on the natural sediments and/or were
rapidly biodegraded (see Section 7). The screening isotherm measurement was
vi-ually made on the Coyote Creek sediment, which was arbitrarily selected be-
i.iuse we had collected a large sample and because it has an intermediate or-
.•.inic content.
To set up these screening isotherms, it is necessary to make an
intimate of the partition coefficient, based on the solubility of the substrate
•n water. In general, as the substrate solubility decreases, the value of the
partition coefficient increases. For instance, Bailey et al. (19681 studied
'.lie sorption of several series of organic herbicides including amines and
i-ids on montmorillonite clays. They found that the log Kp was related to
^olubility in water "within an analog series basic in chemical character." In
'-iieir case, cation exchange and surface acidity of the clay and the pKa of the
'•orbicide determined the sorption within a chemical family. The chemicals
^tidied in this program do not fit within the families studied by Bailey et al.
they do not have acid or basic character (except £-cresol) . The sorp-
of the compounds studied here is probably due entirely to Van der Waals
of sorption.
••'i'niduc Lbici transfers of dried, whole sediments could be made, but the nature
': the sediment would probably be significantly altered by the drying process.
-------
10'
10E
10s
LU
u
iu
8
cc
<
Q.
10
10"
E I | I Illlll—| I I |||!!|—f I III III]—I I Mlllll—! l!ll(l!| I I illlll| I I Illllll I I
O MEASURED Kp
A MEASURED Kb
I I Mlllll I I I Illill I I Illllll I I I II
mi
10-10
10'9
10-8
10'7 10"6 10'5
SOLUBILITY (g m|-1 )
10"
10-3
FIGURE 5-1. SOLUBILITY VERSUS PARTITION COEFFICIENT ON COYOTE CREEK SEDIMENTS (Kp)
AND ON A MIXED POPULATION OF BACTERIA (Kb)
-------
TABLE 5.3. SOURCES AND CHARACTERISTICS OF
Ca-MONTMORILLONITE CLAY AND NATURAL SEDIMENTS
Source
Location and description
Sediment OCC
pH (%)
Cation
exchange
capacity
(meq/100 g)
Navarro River
DCS Moines
River
Oconee River
Covote Creek
Ca-montmorillonite clay
Mendocino County, California
An unpolluted river that
drains redwood forests,
orchards, and pasture
Iowa
Georgia
San Jose, California
A eutrophic, polluted
stream
Searsville Pond Woodside, California
A small eutrophic but
unpolluted pond
6.7
7.1
6.2
6.5
6.7
0.05
0.5
0.8
0.8
1.9
5.0
69.0
4.5
10.5
8.5
13.5
34.5
'Organic carbon, Walkley and Black value, corrected for recovery by multiplying
experimental value by 1.33.
Figure 5.1 is a plot of the logarithm of the partition coefficient
data obtained on this project versus the logarithm of the substrate solu-
bility in water at about 20°C. The values of Kg were obtained from the
Coyote Creek sediment (Table 5.3); the values of Kb were obtained from the
mixed population of bacteria described in Section 5.4.3. While there is
some scatter of the data, the correlation is surprisingly good. This
correlation can be used to estimate the order of magnitude of the sorption
partition coefficient for other compounds. Compounds that interact with
sediments via an ion exchange mechanism probably would not fit this plot.
Many experimental designs for sorption studies are possible. In most
cases, the clay and sediment isotherm measurements v?ere made at two sediment
loadings and two substrate concentrations. Biosorption studies were made at
one level of the mixed bacteria culture and two levels of substrate. Replicate
flasks were used at each level, and at least three analyses of each flask
were made. Suitable blanks of both sorbent and substrate were carried
through the experimental steps and analyzed. Contact times of 1 to 16
36
-------
hours were used. The equilibrium partitioning is probably reached in
about 1 to 2 hours. There were no experimental problems in the sediment
sorption studies with the 16-hour time, except for £-cresol, which biode-
graded rapidly during the experiment. In that case, a 1-hour exposure was
used. For the biosorption studies, the partitioning time was about 1 hour.
At longer times, sorption by the glassware was a problem with low-solubility
substrates.
As the experiments progressed, the experimental plan for the isotherm
measurements evolved into the experimental plan described in Table 5.4. We
consider this to be the minimum number of data points that will permit a sound
statistical analysis of the data.
TABLE 5.4. RECOMMENDED EXPERIMENTAL PLAN FOR ISOTHERM MEASUREMENTS
Number of flasks3
Substrate No Low High
concentration sediment sediment sediment
None
Low
High
1
2
2
2
2
2
2
2
2
3Four replicate measurements of the substrate concentration at equili-
brium in each flask should be made.
The mechanics of the isotherm measurements were generally the same
whether the sorbent was Ca-montmorillonite clay, a natural sediment, or our
bacterial mixture. A solution* of the substrate in water was prepared. An
aliquot of the substrate solution and an aliquot of a suspension of the
sorbent were mixed and allowed to shake for a specified period of time. The
mwss of dry sorbent used was determined, normally by a gravimetric procedure,
in a separate experiment. A portion of the mixture was centrifuged to
separate the substrate remaining in solution and the sorbent. The supernatant
and usually the sorbent were analyzed separately to measure the substrate
concentration.
5.A.2.3 Statistical Analysis of Isotherm Data—The statistical
analysis of the isotherm data on clays and sediments was considered carefully.
The importance of using true solutions below the substrate solubility limit
cannot be overemphasized, especially with low-solubility substrates.
37
-------
The simplest, and often adequate, procedure is to fit the data to the
Freundlich isotherm equation with n = 1,
S = K S
s p w
(5.19)
The value of Sw was always measured. If Ss was not measured independently,
then it was calculated from
S = (S - S )V /m
s o w' w s
(5.20)
where So is the initial concentration of substrate used (determined from the
concentration in flasks without sorbent), Vw is the volume of solution (in
ml), and ms is the mass of sediment (in grams) added to the flask. Pre-
liminary estimates of Kp were obtained by two methods: when only Ss was
measured, every concentration measurement was used, and when both Ss and Sw
were measured, average Ss and Sw were calculated for each flask. In both
cases the data were then fit to an equation of the form
y = bx
(5.21)
(notice the similarity to equation 5.19) using a linear least squares regres-
sion method. The regression equations are:
b =
yx
Zx.y
i-'i
Ex.
(n - I)'
-1
(5.22)
(5.23)
(5.24)
95% confidence interval = ftn_1)0 | syx(Ixi2)T (5.25)
where b = Kp is the slope, x^ and y^ are the individual or average measure-
ments from each flask of Sw and Ss, respectively, Syx is the standard error,
r' is the correlation coefficient, and tn_ljQ is the t-value from Student's
38
-------
t-test for n measurements at a - 0.05 confidence. These expressions are in the
form suitable for use in hand calculators, such as the HP-65.
It is also possible to test that the linear Freundlich isotherm (n = 1,
equation 5.19) does pass through the origin. To do this, the data are fit to
linear equation of the form
S = K S + a (5.26)
s p w o
where a_ is the intercept. The HP-65 Stat-Pac routine is used to estimate the
95% confidence intervals about a0. If these confidence intervals include the
origin, then equation (5.19) can be used. In the data we have tested so far,
the linear Freundlich equation, passing through the origin gives the best fit
of the data, based on the values of the correlation coefficient.
The isotherm data have also been fit to other equations, using linear
regressions available in the HP-65 Stat-Pac routines. The Freundlich isotherm
Ss = KSs (5.18)
becomes
y = axb (5.27)
The Langmuir equation can be written as
W
and then fit to the form y = ax + b. However, this is not a good statistical
analysis because the variables 1/SW and 1/SS are not normally distributed even
if S and S are.
s w
Strictly speaking, these linear least squares methods are not
statistically correct. When only Sw is measured, Ss is calculated using equa-
tion (5.20) and a linear least squares method then has the response variable
Sw on both sides of the regression equation. This makes the confidence limits.
for the parameter values invalid. The linear least squares procedure is also
inefficient because it does not use all the data to estimate S .
39
-------
The limitations of the linear method can be removed by stating the
problem as :
When sorbent is present
A «s
s p w
• When sorbent is not present
S = S. ; i = h or Z, (5.30)
w i
The hat on a^variable indicates that its value is estimated by the regression
procedure. S^ is equivalent to SQ and represents the original amount of sub-
strate present in each flask,^§^ being the concentration in flasks with the
high amount of substrate and §£ being the concentration in flasks with the
low amount of substrate. Hs represents the amount (grams) of sorbent present.
In this procedure, §h or Sj, is estimated using all the flasks as in the
linear method.
Since each flask is handled separately and at a different time, the
above procedure was modified to include "flask effects" as an estimated
parameter. Flask effects include such things as biases due to instrument
drift and systematic errors on the part of the analyst. The resulting problem
formulation is suitable for input to a nonlinear regression program that esti-
mates values for Sh, S^, a0, Kp, and the flask effects. The actual results of
the nonlinear regressions are comparable to the estimates obtained from the
linear least squares regressions, except that the nonlinear approach gives
smaller confidence limits for the parameter estimates.
When both Ss and Sw are measured, the linear least squares procedures
using average values for Ss and Sw are not statistically correct because again
response variables appear on both sides of the regression equation. In addi-
tion, the averaging procedure throws away valuable information about experi-
mental variance. To deal more correctly with this situation, we use eight
simultaneous nonlinear regression equations. Four regressions use only the
substrate concentrations measured on the sediment from the various flasks.
S = K S. i = 1, 2, 3, 4 (5.31)
s pi
and four regressions use only the concentrations measured in the supernatant,
Sw>
S = S. i = 1, 2, 3, 4 (5.32)
w i
-------
where S^ is the estimated value of substrate concentration in a particular
flask. With this formulation, the response variables appear only on the left-
hand side of the regression equation. The subscript "i" on the right-hand
side is the independent variable, since it indicates the conditions used to
set up the flask. For example, i = 1 indicates the two duplicate flasks that
contained high sediment and high substrate concentrations. The common para-
meters Kp and Si tie the simultaneous regressions together and assure that
the resulting estimate for Kp is conditioned on both the sediment and super-
natant concentrations measured.
The method is superficially similar to the simple linear least squares
procedure used to estimate K_. The regression equations (5.32) use the super
natant concentrations to estimate a single concentration that best represents
the concentration in the flasks for each different substrate and sediment
level. This representative concentration is then used with the concentration
of supernatant on the sediment in equations (5.31) to estimate Kp. An impor-
tant difference between the two approaches, however, is that, with the non-
linear approach, the supernatant concentration, S^, that represents a parti-
cular substrate and sediment level is not necessarily the average supernatant
concentration. The values of S^ determined by the method are almost always
very close to the average except when concentration measurements are highly
scattered.
An important feature of nonlinear approach is that it does not require
that the same number of observations of both Ss and Sw be made in each flask.
The formulation also does not require that individual measured values of Sg
and Sw from a particular flask be paired (see equation 5.19). This is an
important feature because any pairing of data points is artificial and could
bias the estimate of K.
5. A. 3 Biosorption and Desorption
Biosorption and desorption on biomass are important because they may
affect biomagnification up the food chain and the "available" concentration
of a substrate for biodegradation. They may also affect the viability or
growth of micro- and macroorganisms that may participate in biodegradative
reactions. All these factors have been demonstrated in various studies with
polyaromatic hydrocarbons, which are of much concern as potential carcinogens.
Our biosorption and desorption studies were conducted with mixtures of
four species of gram-positive and gram-negative aquatic-origin bacteria that
had frequently been used in various microbiological assays and had no record
of functioning as degraders of the types of compounds studied. The mixtures
contained equal optical densities of Azotobacter beijerinckii ATCC 19366,
Bacillus cereus ATC 11778, Escherichia coli ATCC 9637, and Serratia marcescens
ATCC 13880. In the early stages of this program, Flavobacterium capsulatum
ATCC 14666 was used, but because this organism was difficult to centrifuge to
a compact pellet and clear supernatant, it was replaced with the above indi-
cated Serratia marcescens.
41
-------
The test organisms were transferred several times in Trypticase-Soy
broth at 25°C before they were used for sorption studies. Sixteen-hour cul-
tures were either in the late logarithmic or early stationary growth phases.
.\c this stage, each culture was harvested by centrifuging, washed with 0.05%
potassium phosphate buffer (pH 7.0), resuspended and diluted with this buffer
until the suspension had an optical density of 2 to 4.
Appropriate aliquots of suspensions of each of the four organisms were
combined and diluted with buffer to form a mixture containing equal optical
densities of each organism and a mixture that, when mixed with the solution of
the substrate, resulted in the desired concentration and organisms. With
substrates having a low solubility, the density of the bacterial mixture
was lower than with more soluble substrates.
In some instances, biosorption studies were also conducted with heat-
killed cells. Consequently an aliquot of the above mixture of organisms was
heated at 100°C for 15 minutes, cooled, and centrifuged. The resulting pellet
was resuspended in fresh buffer to the original volume, and an appropriate
volume of this suspension was diluted as above with a solution of the sub-
strate under study to yield the corresponding cell densities and substrate
concentrations.
Biosorption studies were conducted by incubating the viable and heat-
killed cell mixtures in Corex centrifuge tubes or bottles for 1 hour at 25°C.
Cells were maintained in suspension by placing the containers in roller drums
or on a rotary shaker. The tubes or bottles were centrifuged for 10 minutes
at 12,000 or 16,000 G, respectively, and the supernatants were carefully de-
canted. The supernatants were extracted with an organic solvent (usually
ethyl acetate). The solvent extract was dried and then assayed directly or
concentrated before assay. To assay sorbed substrate in the pellets, water
was added and this suspension was solvent extracted as above. With some sub-
strates that were tenaciously retained by the cell pellets, the water-
suspended cells in the presence of some solvent were slowly frozen and thawed
three times before extraction.
Desorptions were conducted only if ths sorption partition coefficients
were 10,000 or more. The cell pellets from replicate sorption studies were
suspended in volumes of buffer or buffer and solvent equivalent to those used
for sorptions, incubated with shaking at 25°C for 3 hours, and centrifuged.
Both supernatants and pellets were analyzed by the procedures used for sorp-
tion determinations.
In some cases, corrections were made for adsorption on glassware of sub-
strate and cultures containing sorbed substrate. Separate controls consisted
of extraction of tubes from which the incubated suspensions in test substrate-
solutions were decanted in lieu of separation of cells by centrifuging.
Dry weights of cells used in sorption studies were determined by
weighing the pellet obtained after cells from aliquots of mixed viable
or heat-killed bacterial suspensions were centrifuged, washed with dis-
tilled water, and dried for 16 hours at 90-95°C.
42
-------
The biosorption partition coefficients of chemicals between bacteria
and buffer were determined as
yg substrate per g dry wt of cells
p ~ ^g substrate per ml in supernatant ^~
The results obtained in these tests had good consistencies and may be
regarded as indicators of sorption of the compounds on the bacteria. It
would be interesting to compare these results with data that would be obtained
(,'ith algae and protozoa or with bioaass from natural reservoirs. The latter
would present problems in separation of biomass from inorganic or humic mate-
rials that would also be centrifuged.
j.4.4 Discussion
The major problem in extrapolating the sorption partition coefficients
obtained by these laboratory procedures to environmental conditions is that
the composition of sediments and bacterial mixtures that would exist in the
natural system change dramatically with the time of the year. Therefore, the
composition of sediment and bacterial samples collected at one location are
likely to be different even if they ate collected only several days apart.
In an attempt to overcome this problem, we have measured the sorption parti-
tion coefficient on several sediments and on a mixture of bacteria. On the
basis of these and other studies, VJE estimate that the -value of K- for a
specific sediment or group of organisms should not vary by moce than a factor
of 3 in the environment. Also, there is considerable evidence, both from our
studies and from studies reported in the literature, that there is often good
correlation between the magnitude of Kp and the total organic content of the
sorbent.
-------
6. CHEMICAL TRANSFORMATION
6.1 BACKGROUND
The chemical processes of photolysis, free radical oxidation, and hydroly-
sis described in this section can be important transformation processes for
some chemicals in aquatic environments. Emphasis in this work has been on ob-
taining reliable kinetic data for use in estimating how rapidly these processes
will occur in freshwater systems. Laboratory studies were conducted in pure
water and in natural waters from three natural sources in California: Lake
Tahoe (oligotrophic), Coyote Creek (eutrophic), and a pond near Searsville
Lake (eutrophic). Chemical properties of these waters are given in Appendix
A of Part II. All natural water samples were filtered through a 0.45-um fil-
ter to minimize sorption problems and to ensure homogeneous solutions for
kinetic studies. Kinetic data for reactions in these natural waters were com-
pared with data for reactions in pure water to determine what effect, if any,
the dissolved natural substances have on the rates of specific chemical pro-
cesses.
When comparison with other processes under study indicated that a
chemical transformation was an important pathway, the primary reaction
products were identified or characterized by several procedures, including
isolation by chromatography followed by spectrometric analyses. Knowledge
of the reaction products is essential to understanding the chemical pro-
cess and to any subsequent hazard evaluation that might be based on these
data, whenever possible, quantitative material balances were obtained to
ensure that other unsuspected physical, chemical, or biological processes
were not occurring simultaneously.
As a further check on the validity of the kinetic measurements, control
or blank experiments were carried out along with the kinetic experiments.
Where preliminary studies indicated the importance of biological transforma-
tions, sterile conditions were maintained during chemical experiments. Pre-
cautions were taken to exclude losses through volatilization, and glassware
was continually checked to identify any losses through adsorption. In a very
few cases, competing processes could not be excluded entirely, and corrections
for these processes wore made in the kinetic data obtained. These problems
emphasize the need for good material balances and control experiments in order
to obtain reliable chemical kinetic data and relationships for environmental •
processes.
It is important to recognize that the procedures described below are in-
tended to describe experimental laboratory procedures for reliably evaluating
chemical transformation processes in the solution phase of aquatic systems.
-------
Although the immediate application of this methodology is for environmental
exposure assessment, the data and procedures also serve as basis for investi-
gations of the more complex features of aquatic environments, including the
effects of suspended solids and sediments and dissolved natural organics in
some waters on these environmental transformation processes.
6.2 PHOTOCHEMISTRY
The cutoff for the solar spectrum by the upper atmosphere is at about
290 nm, and in aquatic systems only absorption of photons of this or longer
wavelengths can result in photochemical transformations. These transforma-
tions may occur through direct photolysis of compounds that absorb light
above 290 nm or through photosensitized reactions involving other light
absorbing organic substances found in natural waters. Although the kinetics
and mechanism of direct photolysis of. compounds can usually be evaluated
using present theory and experience, the details of sensitized photolyses in
which organics in natural waters act as sensitizers are largely undefined.
An excellent discussion of environmental photochemistry has recently been
published (Wolfe el al., 1976).
The rate of absorption of light, IA (rate constant ka), by a chemical
at one wavelength is determined by: the molar absorbtivity e (also called
the molar extinction coefficient), a term 1^ proportional to the intensity
of the incident light, and die concentration of substrate [S] at concentra-
tions of S where only a small percentage of the light is absorbed (Zepp and
Cline, 1977)
IA = ElA[S] - ka[S] (6.1)
where k = el,. The rate of direct photolysis of a chemical (rate constant
kp) is then obtained by multiplying T by the quantum yield , which ia the
efficiency for converting the adsorbed light into chemical reaction, measured
as the ratio of moles of substrate transformed to einsteins of. photons absorbed.
- dT= V[sl = Vs1 (6-2)
and
k = k <)> (6.3)
The simplest: and most direct method of using laboratory experiments (as
contrasted to field studies) to estimate environmental photolysis rates is to
expose an aqueous solution of a chemical to outdoor sunlight and monitor its
rate of disappearance. However, the data obtained are of limited use because
45
-------
-untight intensity varies with Che time of day, season, latitude, weather con-
ditions, and light scattering. Thus, any outdoor experiment has questionable
in application to other conditions of sunlight irradiation.
Another method for estimating environmental photolysis rates has been de-
.t-ribed by Wolfe and coworkers (1976) and by Zepp and Cline (1977). This
1>t-ot-edure Calculates the rate constant kp from values of e and <£ measured in
mboratory experiments; the sunlight intensity (lx) data as a function of time
jf day, season, and latitude are available in the literature. Thus photolysis
rates can be estimated for different environmental conditions.
Both the molar extinction coefficient (e) of a chemical and the sunlight
intensity (I) vary as functions of wavelength. The average value of the
absorption coefficient, e*, for a specified wavelength interval centered at
wavelength X (see Table 5.1 for wavelengths X and the wavelength intervals)
is determined from the absorption spectrum of the compound (Section 5.2).
The absorption rate constant, ka, for a compound absorbing in the solar spec-
trum is then obtained by summing the product of EX and 1^ over all wavelengths,
where E > 1 M~x cm"1 and I\ is the solar intensity over the wavelength inter-
val centered at X for a selected latitude and season or time of day.
k <= Il.e, (6-A)
a A A
When e^ is expressed as the molar absorption coefficient (M~l cm~l).
, 2.303 VT ,, ,s
ka=~T-ZVx (6>5)
where J = 6.02 x 1020 is a conversion constant that makes the units of I and t
compatible.
The rate constant for photolysis (kp) is equal to the product of ka and
the quantum yield 4>; generally does not vary significantly with wavelength.
(6.6)
Assuming the reaction is first order in the chemical, the half-life for pho-
tolysis is given by
*& tf, ?^
r— (.0.1)
P
The computer program used to make these calculations gives a plot of the
half-life of the chemical toward photolysis as a function of the month of the
year. These half-lives are based on the average values of kp for a full day's
photolysis. For half-lives of less than several days, variations in sunlight
intensity during the day must also be considered. For these shorter half-
lives, the computer program provides data for half-lives as a function of the
time of day.
Laboratory measurements of $ were carried out using a merry-go-round ap-
paratus (Moses et ai. , J969) to achieve even irradiation of all reaction tubes.
-------
We used a 450-watt medium pressure Hg lamp because it provides intense lines
at 313 and 366 nm that are easily isolated by filters (Calvert and Pitts, 1966).
Samples of chemicals dissolved in water were placed in borosillcate tubes
in the merry-go-round and irradiated for periods from several hours to days at
each wavelength. Tubes were withdrawn at various intervals and analyzed for
starting chemical. Light intensity at each wavelength was measured periodi-
cally using an o-nitrobenzaldehyde actinometer (Pitts et al., 1964) in tubes
similar to those used for reaction mixtures. In most cases these photolyzed
solutions of chemicals were also used for product analyses (Section 6.1).
In cases where the laboratory photolyses proceeded slowly under monochro-
matic light, photolyses were also carried out using only the borosilicate
glassware as a filter. Although borosilicate cuts off wavelengths below 280
nm, it does not isolate a single wavelength and quantum yield measurements
cannot be made. However, the greater light intensity transmitted by the boro-
silicate filter allowed the photolyses to be carried out rapidly to beyond two
half-lives to establish that the reaction was first order in substrate. This
information was needed for extrapolation of the rate data in the environmental
assessment.
Outdoor photolyses using sunlight were also carried out with each
chemical to validate the computer calculation of half-life in sunlight based on
measured values of ex and . Sunlight photolyses require some attention to
placement of apparatus. Ideally, the photolysed solutions should be in a lo-
cation free of excessive reflections from walls and windows and without morn-
ing and afternoon shadows. Although large diameter dishes with flat trans-
parent tops (petri dishes, for example) are preferred, we used 11-mm-O.D.
borosilicate tubes held in a rack at a 60° angle to the horizon. Test tubes
are much more readily sealed for long exposure and, judging from the good
agreement between computed and measured values for tij in sunlight for most
chemicals, are quite satisfactory for this purpose.
When possible, photolyses were carried out in pure water or in filtered
natural waters with no cosolvent. In many cases, however, it was necessary to
include up to 1% acetonitrile by volume as cosoivent because the solubility of
the chemical in water alone was very low. Acetonitrile was chosen because it
would not act as sensitizer at wavelengths in the solar spectrum or,take part
in any free radical reactions that might occur in photolyses. Initial con-
centrations of chemicals in the photolyzed solutions were usually 1 yg ml"1
or less, and for each chemical some photolyses were carried out to at least
two half-lives to verify that the reactions were first order in substrate
as predicted by equation (6.1).
For systems that absorb less than 5% of the incident light (< 0.02 ab-
sorbance) at wavelength A, may be calculated from the photolysis rate con-
stant k obtained from the slope of the first-order plot of the photolyses data
using the integrated form of equation (6.1).
In (S /S) = k t (6.8)
o P
((.. = k /2.3E>I^ (6.9)
A p A "
47
-------
where £ is Che pathlength, S and S0 are substrate concentrations at times t
and t0, and I\ and t\ are as defined above. The value of 1^ may change from
time to time and requires periodic calibration using the act: in a metier to
correct for such changes.
When comparisons were made between measured and calculated half-lives for
chemicals in sunlight at 4Q°N. latitude (about that of Menlo Park, California)
using procedures described above, we found excellent agreement, usually within
a factor of two. These results support our assertion that laboratory measure-
ments of this kind can provide reliable estimates of half-lives toward photol-
ysis in sunlight.
In some aquatic environments, however, rates of photolysis may differ
significantly from those measured in pure water owing to the presence of
naturally occurring light absorbers, quenchers, or sensitizers. In water,
naturally occurring materials such as humic or fulvic acids, which have high
optical densities,, may absorb sunlight and effectively screen the chemical
from being photolyzed. The presence of particulate materials in water may al-
so result in light scattering. In both cases the photolysis rate of the
chemical would be slower than in pure water because the physical processes
reduce the light available for reaction.
Materials present in natural waters may also either accelerate or retard
photolyses of substrates through chemical processes. Acceleration of photoly-
sis rates for some pesticides in natural Maters has been demonstrated (Wolfe
et al., 1976). In most experiments it was not determined whether the rate
acceleration w^s due to a photosensitized reaction or to a photoinitiated free
radical process.
Both processes occur through absorption of light by the natural sub-
stance, which then interacts with the chemical. In the photosensitized reac-
tion, the excited-state energy from the sensitizer is transferred to the
chemical, which then undergoes reaction; the identity of the sensitizer is
maintained. In the photoinitiated reaction, the natural substance that ab-
sorbed the light reacts with the chemical and both materials are transformed.
If either process is more rapid than direct photolysis of the chemical, the
rate of photolysis will be accelerated. However, in natural waters with
significant optical densities, an acceleration of photolysis due to either
mechanism may be somewhat offset by the screening capability of the water.
Results obtained in this project also indicate chat in some cases the
presence of natural substances in water may make the photolysis rate slower
Chan it would be in pure water. The reason for this effect is not known, but
it is not due to a screening effect, since the natural waters in which the
observations were made had absorbances of < 0.02 at 366 nm where the photoly-
'•>cs were carried out.
Since the presence of natural water can either accelerate or retard the
aiiotolysis of a chemical, half-lives based on pure water photolyses must be
interpreted with some caution. If experiments in natural waters give faster
Photolysis rates than in pure water, the photolysis rate in pure water is use-
iul as a conservative value (i.e., maximum half-life). When the photolyses in
-------
natural waters are slower than in pure water, the half-life estimate obtained
for the pure water photolysis should be used with an appropriate qualification
chat in some cases longer half-lives may occur and that more experiments may
be needed to determine how much slower the photolysis is likely to be.
6.3 FREE RADICAL OXIDATION
Oxidation of organic compounds by free radical processes may be important
under some environmental conditions. The most general reaction scheme for
radical oxidation with an azo initiator is
kd
RaN2 _£»2R-+N2 initiation (6.10)
(6.11)
+ X- oxidation (6.12)
2R02' "(2 RO) + 02 (6.13)
2RO' termination
products
k.
RO- + XH • ROH + X- oxidation (6.14)
kd
RO' •- cleavage products (6.15)
The rate of oxidation of compound XH is then
r = -d[XH]/dt = k [R02][XH] + k._[RO'][XH] (6.16)
OX OX rvu
To evaluate the potential importance of oxidation under environmental
conditions, we need to be able to evaluate equation (6.16) for specific com-
pounds in specific environments. Values for rate constants k and k are
known reliably for many organic compounds in organic solvents (Hendry et al.,
1974) but have rarely been measured in water or for most of the organic com-
pounds studied here. For these reasons we developed a simple screening ex-
periment that provides a reasonably reliable method for evaluating kox in
water for compounds of interest and a reliable method for evaluating relative
reactivities toward R02'for a series of compounds.
As a source of R02'we have chosen a commercially available azo initiator,
4,4-azobis(4-cyanovaleric acid) (AA)
[H02CCH2CH2C(CN)(Me)]aN2
49
-------
AA is slightly soluble in water and decomposes at 50°C with a rate constant
•>f 1.9 x 10~6 sec~l (t!j = 100 hours) to give two carbon radicals, a fraction
>i which is rapidly converted to peroxy radicals in the presence of oxygen.
(6.17)
(6.18)
The rate of production of R02 is d [R02-]/dt = 2ekd[R2N2] where e is the
fraction of radicals R- that are available for oxidation. In a separate study
(Mill et al., 1977) e has been evaluated as 0.6, which is very similar to
values of e found for other azo compounds in organic solvents (Denisov, 1974).
ider conditions where only a small concentration of [XH] is oxidized compared
with the total concentration of R02* generated, the instantaneous concentration
of R0a- (steady-state) depends only on the rates of initiation and termination.
Moreover, under these conditions the only fate of RO- is to cleave (reaction
6.15); therefore, equation (6.16) simplifies to:
rox = ko
Experiments were carried out by heating an air-saturated aqueous solu-
tion containing about 10"* M AA and the chemical below its solubility limit
for up to 100 hours at 50°C, a half-life for AA. In some cases analyses for
the chemicals were made at periodic intervals, and in other cases replicate
analyses were made at 100 hours. A first-order plot of the data (log concen
tration versus time) at two or more times gave a straight line corresponding
the relation:
ln([XH]./[XH] ) = -k [R0a-]t (6.20)
o
with a slope equal to -k [R0a' ] •
With these concentrations of AA and XH used in most experiments, the
value of [ROa'l may be calculated from the steady-state assumption that the
fates of initiation and termination are equal:
[R0a-] = (2ek,[AA] /2kv)T (6.21)
-Jiere k = 1.9 x 10~6 sec"1, e = 0.6, and 2kx = 2 x 107 M~l sec"1.
•• A. Howard, National Research Council of Canada, private communication,
1(J77.
50
-------
With [AA] = 7.5 x 10~s M, [ROa'J is then 2.9 x 10~9 M. From the slope of the
line described by equation (6.20):
k 50° = slope/(2.9 x 10~9) (6.22)
25° 50°
The value of ko may be calculated from the value of kox by assuming
that the activation energy for the reaction of R0a" with XH has an average
value of 10 kcal mole"1(41.8 kJ mole)"1, which corresponds to a factor of nine
in rate. Thus:
7S° 50°
kox =0'1Lkox <6'23>
The rate of oxidation and half-life at 25°C may then be calculated from this
value of kQX and an estimate of the concentration of R02' present in aquatic
environments.
For purposes of this study, we have assumed that [ROa'l in aquatic en-
vironments is 10" 9 M. This assumption is untested, but when combined with an
experimental value of kOX}it places a probable upper limit on the rate of oxi-
dation and thus on the importance of oxidation under environmental conditions
when compared with competing physical, chemical, and biological transformations.
6.4 HYDROLYSIS
Hydrolysis of organic compounds usually results in introduction of a hy-
droxyl function (-OH) into a chemical, most commonly with the loss of a
leaving group (-X). These reactions
RX + H20 -ROH + HX (6.24)
n n
R-C-X + H20 -R-fl-OH + HX (6.25)
may be catalyzed by acids or bases (rate constants k^ or kfi, respectively) or
both. The kinetics of hydrolysis can. be expressed as
^ = kh[S] = kB[OH-][S] + kA[H+][S] + kN'[H20][S] <6-26>
where k^ is the measured first-order rate constant at a given pH. The Last
term is the neutral reaction with water (second-order rate constant kjj1), and
51
-------
in water it can be expressed as a pseudo-first-order rate constant k[j. Since
with few exceptions, hydrolysis reactions are first order in chemical, the
half-life of a chemical toward hydrolysis may be expressed as
t^ = In 2/(kB[QH-] + kA[H-] + kj,) = In 2/kh (6.27)
From equations (6.26) and (6.27), it is clear that when kg and/or kA $ 0,
k, will depend on pH. From, the autoproto lysis water equilibrium,
1 = Kw = 10 (6.28)
equation (6.26) may be rewritten
k
BKw
The contribution of each term to kj, will depend on the acidity (or pH) of the
solution.. Three regions may be defined:
Acid kA[H+] > kN + -j|^ , log 1^ = log kA + log[H+] (6.30)
= log k. - pH
A
k K
Base
] , log k = log k - log [H] (6i31)
Neutral \ > k(H } + -- , log k = log k (6.32)
These expressions assume that the catalyzed processes are first order in [H ]
or [OH~], Such behavior is almost always the case in the range of pH 2 to 12
and frequently extends to greater extremes.
The dependence of k^ on the pH of the solution is conveniently shown by a
l>lot of log kh as a function of pH (Figure 6.1). From expressions (6.30),
(6.31), and (6.32)j it is seen that in the pH range where the base-catalyzed
process is dominant, a slope of +1 is found; a slope of -1 is found in the
acid-catalyzed region. The neutral hydrolysis is pH independent and shows a
of zero.
The present knowledge of the thoretical and experimental aspects of hy-
drolysis reactions makes laboratory studies of hydrolysis rates useful for en-
52
-------
(Jl
TA-327522-29R1
FIGURE 6.1. pH DEPENDENCE OF kh FOR HYDROLYSIS BY ACID, WATER, AND BASE-PROMOTED PROCESSES
-------
vironmental assessments. Precautions must be taken, however, to ensure that
experimental artifacts are not introduced into the kinetic data. For example,
the use of buffer salts to maintain constant pH is a necessary and acceptable
practice, but some precautions should be taken to preclude effects of buffer
catalysis and/or ionic strength.
Another problem may arise in the use of organic cosolvents in water to
increase solubility. Large amounts of organic cosolvent will usually make the
hydrolysis rate slower than it would be in pure water.
For chemicals that are moderately stable toward hydrolysis, it will ex-
pedite laboratory procedures to work at higher reaction temperatures (30° to
100°C). From data obtained at several temperatures, the rate constant at any
environmentally relevant temperature can be calculated, using temperature re-
lationships such as the Arrhenius equation
k = A exp (-E/RT) (6.33)
where E is the activation energy and A is the preexponential factor. However,
such kinetic relationships hold only for rate constants for individual pro-
cesses (k^ or kg or kN) _ If the temperature dependence of an observed rate
constant kn is measured at a pH where two processes contribute to the overall
rate (e.g., k., and kg), k, is then the sum of the rate expressions:
kh = 1^ + kB[OH~] (6.34)
= ^ exp(-EN/RT) + Ag exp(-EB/RT)
Thus the measured temperature dependence of k^ is a complex function of T and
useful only at the one pH. The pH profile indicates where each rate process
is dominant and thus where temperature dependence measurements are most use-
fully made. Arrhenius parameters obtained in the pH regions of the pH profile
where only one process is important (slope of -1, 0, or +1 for acid, neutral,
and base processes, respectively, in Figure 6.1) then allow for calculation of
rate constants for individual processes any any temperature, and the constants
may then be combined in equation (6.29) to estimate the rate constant kh that
will be relevant to any particular environmental pH and temperature.
For compounds that hydrolyze extremely slowly around 25CC and thus have
very long half-lives, a semiquantitative estimate of k^ or fc^. may suffice for
assessment purposes if other environmental processes are relatively fast. In
those cases, the rate constant for hydrolysis at some high temperature where
the rate is reasonably fast may be measured and then extrapolated to 20°-25°C
using the rule of thumb that the value of kh will change by a factor of two
for each 10° change in temperature. This rule of thumb assumes that the
energy of activation for hydrolysis is about 18 kcal mo]e~ (75.3 kJ mole"1).
5/t
-------
Since the actual values of E^ range from 15 to 28 kcal mole"1 (factors of 1.8
to 3), the extrapolated rate constants are only semiquantitatively correct.
Hydrolysis data can be used in environmental assessments at selected pHs
and temperatures with considerable confidence, provided the chemical is dis-
solved in the water rather than suspended or emulsified. Although catalyses
of hydrolysis by metal ions and nucleophiles are known, the concentration of
such catalytic substances in the water column are so low that the rates of
these catalyzed hydrolyses are insignificant compared with rates of the neu-
tral and H* and OH~ catalyzed processes. Moreover, the concentrations of the
active metal ions or nucleophiles available for reaction may be lower yet, due
to complexation and association with natural substances present in natural
waters. It is the availability of the catalytically active form and not the
mere presence of the species that would result in any contribution to the hy-
drolysis rate of a substrate.
55
-------
7. BIODEGRADATION
7.1 BACKGROUND
The techniques used in evaluating the biodegradability of organic sub-
strates have varied extensively, and it is doubtful that any one procedure can
be used to indicate susceptibilities to biodegradation in aquatic or soil en-
vironments. The phenomena are too complex and varied with some of the sub-
strates that are difficult to degrade. Alexander (1965) introduced the term
"recalcitrance" to define the characteristic of a compound that resists micro-
bial biodegradation and presented some explanations of this microbial falli-
bility. There has been much research, elucidation of metabolic pathways, and
theoriEing on characteristics that are involved in biodegradation or recal-
citrance of organic products. An additional complexity is that some readily
biodegradable substrates can resist biodegradation when small quantities are
strongly sorbed on soil or clay particles, particularly if these substrates
are deposited or sorbed in locales or microenvironraents into which micro-
organisms cannot penetrate.
Microorganisms are highly susceptible to frequent enzymatic reorientation
in response to environmental change or alteration in substrate availability.
Since the discovery of plasmids, many microbial degradations have been attri-
buted to enzyme systems synthesized by these DNA particles. The phenomena of
repression, derepression, induction or enrichment by analogs, and availability
or lack of availability of other substrates and nutrients play important and
differing roles with various culture-substrate combinations.
The phenomenon of cooxidation or cometabolism can also be very important
in biodegradations. This involves the metabolism of a nongrowth-promoting
substrate only when it is present with a growth-promoting substrate. Interpre-
tations of these terms have been broadened to include growth-promoting sub-
strates that do not necessarily have chemical structures very close to the
substrate under study. In nature, organisms are exposed to a large variety
of chemicals, and cometabolism can be very important.
Techniques frequently used in biodegradation studies involve pure cultures
obtained from random isolations, culture collections, or enrichment cultures.
Enrichment cultures are frequently mixtures of organisms that are developed
by adding and incubating a water, soil, compost, or other natural substance
In a medium initially or finally containing the substrate under study as the
sole carbon source. In nature, of course, constantly changing mixed culture
systems are invariably involved.
56
-------
Analog enrichment or induction refinements involve the addition of more
readily metabolizable compounds, chemically related to the enrichment substrates.
If metabolism of the substrate depends on the presence of the analog, a coraeta-
bolic process is generally involved. However, in some cases the added chemical
functions as a hydrogen or oxygen donor, or as an organic carbon substrate.
Examples of the former are some dehalogenations or reductions of nitro groups.
Under anaerobic conditions, it is not unusual to isolate systems that can
reduce nitro groups or halogenate organic compounds. Another type of complex
biodegradation is the metabolism of hydrocarbons by sulfate-reducing organisms.
These conditions are not analog enrichment processes, but depend on the presence
of a biologically reducible inorganic substrate and hydrogen donor organic com-
pounds. These organic compounds are eventually converted to products that may
be assimilated for microbial growth.
Not to be overlooked are the phytoplankton and protozoa that may be involved
in environmental metabolism, but have received less research attention because
of their complexities.
Most of our current knowledge of metabolic mechanisms has been derived with
pure cultures and their mutants under growth conditions, as resting cells, or
with their enzyme preparations. Mixed culture systems present complications in
maintenance of their component character because of varying growth rates and a
host of antagonistic and synergistic relationships. Included in these complex
phenomena are high biosorptive characteristics for some substrates. Under these
conditions, the "available" concentrations of substrates may be reduced to such
a degree that organisms capable of metabolizing a substrate do not have suffi-
cient organic carbon available for growth before they die.
In natural waters, there are normally many types of microorganisms and these
may vary with the water body, season, and organic substrates being introduced.
The studies described in this report were designed to reflect many of these
factors. In the screening studies, we attempted to obtain biodegrading systems
by enrichment procedures. If we obtained one or more biodegrading systems,
detailed studies were conducted to determine the biodegradation rate character-
istics of one of these systems. The specific procedures used are described
in Sections 7.2 and 7.3. Isolation and identification of major biodegradation
metabolites are described in Section 7.4.
7.2 DEVELOPMENT OF ENRICHMENT CULTURES
The objective of the biodegradation studies was to develop a rapid inex-
pensive experimental approach that would approximate natural conditions if
selected compounds were introduced into freshwater environments. Under the
provisions of the contract, the enrichment studies included:
Enrichment techniques under aerobic conditions.
• Enrichment studies to be completed within six weeks. Within this
period, subtransfers were made to develop enrichment culture systems
that could utilize selected substrates as the sole carbon sources.
57
-------
The use of a biodegrading enrichment system in kinetic and metabolite
studies without isolation of a pure culture that utilizes the substrate.
Identification of major metabolites.
If an enrichment culture that could degrade the substrate as a sole carbon
source was not developed within six weeks, no metabolite or kinetic studies
were carried out.
It is understandable that any isolation procedure, whether it is an enrich-
ment procedure or the isolation of single colonies from natural habitats, favors
isolation of certain types of organisms and does not express the total micro-
bial potential or populations in the natural habitat. Enrichment procedures
such as those that were used in this work favor specific types of microbial
populations, and many types of organisms that are present in the environmental
sample cannot survive the competitive aspects of the process.
The principal natural aquatic reservoirs used as representative sources
for cultures were:
• A eutrophic pond near Searsville Lake in Woodside. California.
Coyote Creek, a eutrophic stream in San Jose, California.
• Aeration effluent from the Palo Alto, California, sewage treatment
facility. The organic matter treated is primarily of domestic origin.
Aeration effluent from the South San Francisco treatment plant. Approx-
imately 45% of the biological loading in this facility is of miscellan-
eous industrial origin.
• Aeration effluent from the treatment plant in the Shell Oil Refinery,
Martinez, California. This plant treats wastes that could have a great
similarity to those that might be expected from a coal liquefaction
plant.
• Aeration effluent from the sewage plant of the Monsanto Chemical Company
installation in Anniston, Alabama, where parathion and methyl parathion
are produced.
• Lake Tahoe, California, a large, deep, cold oligotrophic lake between
California and Nevada.
Water samples were settled for approximately one hour and the super-
natants were screened through fine-mesh polyester cloth. Four volumes of
water sample were added to one volume of sterile 0.05% NlUNQs or (NH^aSCU
and 1.0% KHaPCWKzHPOi solution. This salt solution was added to provide
adequate nitrogen and to buffer the fermentation at pH 7.0.
In our enrichment procedures we used 4-liter water samples in a final
5-liter volume in the 9-liter bottle fermentors because we felt that large
samplings could facilitate the development of biodegrading systems. The ad-
vantages of large volume samplings over traditional small samples became
apparent in four experiments when 50-ml aliquots were also incubated in 250-ml
shaker flasks in rotary shakers. This concept was supported by one experiment
with p-cresol and three experiments with methyl parathion as substrates. In
58
-------
these experiments, biodegrading systems were more rapidly developed in the 9-
liter bottle fermentors. ''Dagley (1976) expressed his view that, in enrichment
studies, samples may be generally too small.
The 9-liter bottles with the 1-liter of additive solutions were sterilized
by autoclaving, 4-liter water samples were added, and then sterilized fittings
were introduced into the 9-liter fermentors. These fittings included facilities
for the introduction of sterile air through ceramic diffusers at the bottoms
of the bottles, sampling ports, addition and pressure relief ports, and air
exhausts through sterilizing filters. Inlet air was first humidified and par-
tially sterilized by bubbling through 1% H3P04 and then sterilized by passage
through pyrex glass wool packed filters. Incubation was at 25°C. Lake Tahoe
water samples were transported in ice-water baths and incubated at 15 °C. Fermen-
tors were shaken several times daily because the equivalent of 0.1 volume of air
per minute was not adequate to maintain total suspension of some samples.
Occasionally, when it was apparent from previous studies with other water
samples that enrichment cultures could be readily obtained, the first step of
the enrichment process was conducted in cotton-plugged 2.6-liter Fernbach flasks
containing 1.2 liters of 4:5 diluted water sample with proportional amounts of
buffer and NH^+ salts. Adding the substrate as a powder might have introduced
microorganisms and presented problems in obtaining fine suspensions. If the
substrate had been added in a solution of a metabolizable solvent, additional
carbon source would have been added. Preliminary studies indicated that
dimethyl sulfoxide (DMSO) was not digested or inhibitory under aerobic conditions
and that the concentrated DMSO solutions were self-sterilizing. Consequently,
it was frequently convenient to add the substrate in a DMSO solution to obtain
either a fine suspension or a solution of substrate in water.
In some of our enrichments, compounds with structures similar to the test
chemicals were used in anticipation that they may be inducers of desired
enzymes.
When possible, a rapid uv absorption assay, verified by gc or hplc, was
used to monitor the breakdown of the test compound. In other cases, gc or hplc
was used alone. When degradation was apparent, 2.5-ml aliquots from the 9-li^er
bottle fermentors or the Fernbach flasks were transferred to 250-ml Erlenmeyer
shaker flasks containing 50 ml of basal salts medium at the original level, a
twofold or threefold increased level of compound, with or without other nutrients
including 50 ug ml"1 glucose with 10 yg ml"1 Difco Bacto yeast extract or with
peptone. When degradation of a test substrate was nearly complete, successive
transfers (1% to 2% by vol) were made to basal salts media with the same and
higher (i.e. , two levels) concentrations of substrate and lower amounts of other
carbon nutrients. Eventually, no other added carbon source was used.
Each liter of basal salts medium contained: 1.4 g K2HP04, 0.6 g
0.5 g (NH4)2S04, 0.3 g NaCl, 0.1 g MgS04'7H20, 0.02 g CaCl2-2H20, 0.005 g
FoS04 -71120, and 1 ml of trace elements solution. The trace elements solution
contained 0.1 g H3B03, 0.05 g each of CuSO<.-5H20, MnSO<,'H20,
, and CoC^IO liter"1.
59
-------
Even after 10 to 15 serial transfers, the enrichment culture systems were
usually mixtures of organisms. They were centrifuged and then suspended in
sterile 57, DMSO-H20 for 30 minutes; aliquots were preserved by freezing and
storing them in the vapor phase of a liquid nitrogen storage tank.
7.3 BIODEGRADATION RATES
Once a biodegrading system was developed for a specific substrate, kinetic
rate constants were determined so that a quantitative comparison might be made
between the different pathways governing loss of a specific pollutant in aquatic
environments. Four procedures for measuring kinetics were investigated and
used with these mixed culture systems and with the substrate serving as the
sole carbon source:
Batch fermentations with low-level inocula of washed biodegrading cells
Continous chemostat fermentations
Cascade batch fermentation
Batch fermentations with large iricrobJal populations and low substrate
levels.
Classical kinetic expressions were applied to the laboratory data to
describe the rate of growth of an organism and utilization of a substrate
when it was the growth-rate-limiting carbon source. These procedures were
used successfully to obtain biodegradation rate constants, which were used
in our environmental assessment models to compare the importance of bio-
degradation with the other transport and transformation pathways.
The Monod kinetic equations (Monod, 1949; Stumm-Zollinger and Harris, 1971),
can be expressed as
U S
V = ,„ m. ^ (7.1)
^Jx ~ &J
dS _ iv - ^ SX = , _ SX _
dt " YX " Y (K + S) *b (K + S)
s
*
(7.3)
where S is the concentration of substrate, v is the specific growth rate, (Jm
is the maximum growth rate, Y is the cell yield, X is the biomass per unit
60
-------
volume, and Ks is the concentration of substrate supporting a half-maximum
growth rate (0.5ym). The utilization rate constant, kb, is conventionally
defined as
T
It is implicit in these kinetic analyses that um, KS, and Y are constants.
The similarity of equation (7.2) to the Michaelis-Menton (1913) equation
, .
(7>5)
for the enzymatic decomposition of a substrate is apparent. In this equation,
E0 is the maximum concentration of available enzyme, and Km is the substrate
concentration that produces half the maximum enzymatic velocity. In cellular
metabolism, a much more complex situation exists.
Although Monod kinetics are based on the use of a pure culture and the rate
of disappearance of a growth-rate limiting single substrate, these kinetic
expressions can be used to obtain useful rate constants with mixed culture
systems. These rate constants can be derived by various procedures in which
there are specific limitations regarding relative values of X, Y, S, and Ks.
In the following paragraphs, the various limits of the Monod expressions that
must be built into the experimental plan will be examined in order to obtain
simple relationships between the experimental variables X, S, and t.
For many of the more common substrates, Ks is on the order of 10"* yg ml"1
(Pirt, 1975), and this, in general, is considerably higher than the concentration
that would be expected for chemicals in natural waters. If
So ** Ks (7<6)
equation (7.2) reduces to
dj> _ m v_ _ . yc /7 7\
~ Ji. ~ v V - ~ KO \' ' ' )
where k, ? is a second-order rate constant equal to
v k.
k = —— = — (7.81
b2 Y K K k '
s s
the disappearance in substrate is first order in both X and S.
61
-------
In batch fermentations with low-levels of inocula. the initial conditions
were chosen so that
Xo «YS (7.9)
With a small inoculum, there is generally a lag phase of growth to a biomass
concentration that may be designated as Xa, and then a logarithmic or expo-
nential phase of growth develops. During this rapid growth phase, So does not
change significantly and the biomass concentration X at time t in this phase
may be expressed as
In X = pt + In X (7.10)
o
If In X data obtained during this period are plotted against t, then u is the
slope of the line. A small inoculum facilitates a longer exponential phase
of growth and facilitates a more accurate calculation of p. This is particu-
larly the situation during the early exponential phase when S has not changed
significantly and there is less complication due to metabolites. These batch
fermentations were conducted with different S values, and \i was determined
for each value of S .
These v values and the corresponding Rvalues were used to calculate Ks
and iim. Inverting equation (7.1) and multiplying by SQresults in the following
equation:
S K s
•S--& + -S- (7.11)
V V V
m m
It becomes apparent that when SQ/U is plotted versus So>the slope of the line is
l/um and the intercept on the S axis is -Ks- This procedure was used by
Lineweaver and Burk (1934) to determine K^j and E0 in the Michaelis-Menton
equation (7.5). In most instances, by using the S and p data from batch fermen-
tations with low-level inocula, it was possible to obtain Ks and pm values, and
then to calculate kj, and kjj2 using equations (7.4) and (7.8).
In some cases, there weie significant increases in biomass concentrations
(X) before utilization of substrate was initiated and other kinetic
analyses were used. Equation (7.2) was integrated by Stratton and McCarty
(1967) and it can be written as
K i ,
S ' £ In (X + YfiS) + $ In X
X + YS Y ^" v o ' Y ~ o
L o o
K SX \\
F-TTs- ^^--V- -^ °'12)
o o o
62
-------
If the initial experimental conditions were
X « YS (7.9)
o o
then
X « YiS (7.13)
o
and equation (7.12) reduced to
or
- In YAS + In X = - - t (7.15)
Y Y o i
(7.16)
A plot of InAS versus t should be a straight line with a slope um and an intercept
(at t = 0) of ln(Xo/Y). This behavior was observed in the batch fermentations
with low-level inocula.
The value of Y can be calculated since
In AS = M t + ln o
However, this procedure did not provide a value for Ks to calculate k|j2 by
equation (7.8). Stratton and McCarty (1969) developed a graphical procedure
that can be used to determine Ks in batch degradations. This procedure depends
on determining the time periods when there are equal utilizations or degradations
of substrate (AS), with different So levels. Their equation (6) is
dS /dt - dS /dt
I(dsm/dt)/s]n] - t(dsn/dt)/sn]
where Sm and Sn are the concentrations of substrate present at times when AS
values were equal for two fermeiuations with different levels of substrate. The
slopes of the S versus t curves at these times correspond to the dSn/dt and
dSm/dt values. This procedure was used to determine KS in some batch degradations
63
-------
with low-level inocula and when it was more convenient to determine the time
periods necessary for equal utilizations of substrates. By these methods, values
for ym, Y, and Ks were calculated from batch fermentations using low-level
inocula.
In continuous f er mentations conducted in chemostats, when equilibrium was
established, the dilution rate or (residence time)"1 was equivalent to u at
the concentrations of substrates present in the chemostats or in the overflow
from the chemostats. If the dilution rates were changed and equilibria were
established, new u and S values were obtained. The values of Ks and um veTs
calculated by the Lineweaver-Burk plot procedure using equation (7.11). Then
kjj and kjj2 were calculated using equations (7.4) and (7.8).
In the cascade fermentations, low levels of inocula and relatively high S0
levels were used. Equations (7.16) and (7.17) were used to calculate k{j values,
and Ks values from other procedures to convert kfo to k{,2- 1° this procedure,
there was no culture selection as occurs in sequential transfers of enrichment
systems on substrate/basal salts media.
Batch fermentations with large microbial populations and low substrate
levels were observed to be pseudo-first-order reactions with respect to S (plots
of Ins versus t were linear for each case tested). In these fermentations, the
microbial populations would not change significantly if all the substrate were
utilized for growth purposes. This is a particularly useful procedure with sub-
strates that have low solubilities and/or critical limitations in analyses.
The experimental data obtained under these conditions can be described
by the equation (7.19)
where k^ is a pseudo-first-order rate constant. A choice of different XQ
would give a different value of k^. Based on equation (7.7), which is
(7.7)
k^2 can be calculated from kj, and Xo by equation (7.20) (assuming Xo is a con
stant) '
k'X (7'20)
Note that the values of ym, Y, and Ks cannot be determined by this procedure,
but in fact they are not required since equation (7.7) is a satisfactory rate
expressions for biodegradation in natural waters.
-------
The procedures used in the laboratory techniques were carried out in the
following ways.
Batch fermentations with low-level inocula were conducted in shaker flasks
incubated at 25°C in rotary shakers. The enrichment culture systems were grown
on substrate/basal salts media or substrate in 0.1 strength nutrient broth (Difco).
The cells were removed by centrifugation, washed three times with 0.05% potassium
phosphate buffer at pH 7, and rested in buffer for 2 to 18 hours at room temper-
ature. These cells were then added at appropriate levels to sterile substrate/
basal salts medium. In some experiments, several lots of inoculated media were
prepared with different concentrations of substrate; 800-ml volumes of inoculat-
ed media were incubated in 2-liter Erlenmeyer flasks. During the fermentation,
duplicate samples were removed from these shaker-incubated flasks for analyses.
Continuous chemostat fermentations were conducted in 350-ml working volume
New Brunswick chemostats. Inocula in the exponential phase, grown in shaker
flasks containing substrate/basal salts media, were transferred to sterile
chemostats containing substrate/basal salts media to bring the liquid volumes
to capacity. Fermentations in the chemostats were initiated with aeration (350
ml air min"1), stirring, temperature control, and continuous feed of substrate/
basal salts medium. Initially, feed rates were very slow to prevent washout of
cells. The feed media were at higher concentrations of substrates in basal
salts medium than the concentrations anticipated in the chemostat. These feed
media were introduced into the chemostats at different rates until equilibria
were established. At equilibrium, the samples from the overflow had reached
a steady state with respect to substrate and biomass concentrations. Because
there were, at times, attachments of microbial cells to the wall and other parts
of the chemostats (aerator, sampling tube, temperature control units, and
stirrers), chemostats were thoroughly shaken and contents were transferred once
or twice daily to other similar chemostats.
Cascade batch fermentations were initiated with freshly developed degrading
systems from eutrophic waters. When the substrates were almost totally degraded
in the original 9-liter bottle fermentors, small aliquots from these fermentors
were transferred to 250- or 500-ml Erlenmeyer flasks containing fresh water
samples (from the same sources), NH^"*" salt, buffer, and substrates. -These
flasks were incubated at Z5°G in shakers. Cell counts and substrate levels were
monitored. Sequential transfers were made daily to new flasks containing fresh
water samples, salts, and substrate. In this procedure it was difficult to
determine the volume of inoculum needed for the sequential transfers to develop
essentially total decomposition of substrate in 24 hours, and it v?as difficult
to follow cell counts at critical times in the fermentations.
Batch fermentations with large microbial populations and low substrate levels
were conducted with cells grown on substrate/basal salts media. The fermentations
producing the inocula were monitored for substrates to be certain that the
substrates were consumed. In each case, several shaker flasks had to be used
co produce sufficient cells, and these organisms were undoubtedly in the late
exponential or early stationary phase. Cells were separated by centrifuging at
room temperature and high speeds, resuspended in basal salts medium, and centri-
fuged. They were resuspended in basal salts medium and incubated at 25°C in a
shaker flask for 4 hours, centrifuged, and agdin resuspended in basal salts
65
-------
medium. The optical densities of these cell suspensions were used as guides for
the dilutions to be used in the kinetic studies. Appropriate aliquots were
added to substrate/basal salts media. The relative quantities of cells and
substrate were such that if all the substrate was utilized, there would be
insignificant increases in cell mass. The inoculated media were vigorously
stirred at 25°C in siliconized tissue culture spinner flasks. Cell counts and
substrate levels were determined at short time intervals.
Each of the kinetic evaluations by the above four procedures has its par-
ticular advantages and shortcomings when mixed culture systems are used to
develop rate constants. The batch fermentations with large microbial popula-
tions require the least time and for this reason avoid the relative changes of
individual microbial components from the initiation of the experiment. However,
they cannot reflect on the character of the culture mixture several transfers
prior to the kinetic study. This is also the procedure most adaptable to
fermentations with very low substrate levels in which changes in biomass con-
centrations may be difficult to follow by other methods.
The cascade batch fermentation represents a procedure most similar to that
existing in nature. In this procedure, the biodegrading cultures were returned
to an environment that contained all types of predator-prey relationships orig-
inally present in the water source (assuming that daily water samplings were
similar). There is a problem in maintaining a supply of unchanged inflow water
sample and selecting inoculum levels. If realistic low levels of substrate
are used, it would be virtually impossible to determine the biomass attributed
to the metabolism of the compound.
Even if the enrichment mixed culture system consisced of only primary
utilizers of the pollutant being evaluated, the batch fermentations with low-
level inoculations or continuous fermentations in chemostats could suffer from
changes in relative components in the culture mixture during the course of the
experiments. Also, in both these procedures, with realistic low levels of sub-
strate, the biomass determinations as discussed by Pirt (1975) would be subject
to considerable error. Cell counts appeared to be the best alternative. In
the continuous culture procedure, the solubility of the compound in the feed
stock must be sufficient for the dilution it is subject to on addition to the
fermentation vessel.
In continuous fermentations at high feed-rates, some metabolites that are
utilized at slower feed-rates may be lost from the fermentors and thus not
contribute to cell growth, but the analyses for substrate consumed would not
indicate this. If substrate levels were high enough for accurate carbon anal-
yses or if uniformly labelled lt"C- substrates were used, corrections could be
made or at least a better understanding would be obtained from carbon balance
studies.
The batch fermentations with low-level inocula and the continous fermenta-
tions in chemostars arc excellent and indeed are the procedures of choice if
a pure culture is used with a substrate that provides normal cell growth and
when the level of substrate is adequate to produce sufficient biomass to be
measured with precision. With the compounds tested in these studies, we invar-
iably developed organisms that ware much smaller than those that would develop
in good microbiological media or as they are normally found in eutrophic
66
-------
containing a mixture of digestible carbon substrates, and this was another reason
for using cell counts as indices of bioraass.
7.4 ISOLATION AND IDENTIFICATION OF MAJOR BIODEGRADATION METABOLITES
During the course of uv, gc, or hplc analyses of extracts from biodegra-
dation studies, there was constant surveillance for evidence of metabolites.
Most extractions were conducted under acidic or neutral conditions, and the
types of major metabolites expected would have been acidic or neutral and
extracted with our solvents. When nitrogen heterocyclic substrates were used,
some extractions were made under slightly alkaline and neutral conditions.
If there was evidence for metabolites, mass spectometric analyses were
applied to the gc or hplc fractions containing the products. When possible,
these spectra were compared with those of authentic reference samples of the
anticipated metabolites to positively establish their structures.
67
-------
8. REFERENCES
Acree, F., M. Beroza, and M. C. Bowman. 1963. Codistillation of DDT with
Water. J. Agr. Food Chem. 11:278-280.
Alexander, M. 1965. Blodegradation: Problems of Molecular Recalcitrance and
Microbial Fallibility. Adv. Appl. Microbiol. 7:35-80.
Bailey, G. W., J. L. White, and T. Rothberg. 1968. Adsorption of Organic
Herbicides by Montmorillonites: Role of pH and Chemical Character of
Absorbate. Proc. Soil Sci. Soc. Amer. 32:222-234.
Calverc, J. G., and J. N. Pitts, Jr. 1966. Photochemistry, John Wiley &
Sons, Inc., New York, pp. 780-782, 795-798.
Campbell, A. N. 1930. An Apparatus for the Determination of Solubility. J.
Chem. Soc. London, Part 1:179-180.
Dagley, S. 1976. The Contribution of Microbial Metabolism to the Carbon
Cycle. In a seminar on "Degradation of Hazardous Compounds: An Inter-
disciplinary Overview. Amer. Soc. Microbiol. 76th Meeting, Atlantic
City, New Jersey, May 2-7.
Dal Nogare, S., and R. S. Juvet, Jr. 1962. Gas-Liquid Chromatography.
Interscience, New York.
Denisov, E. T. 1974. Liquid Phase Reaction Rate Constants. English
translation. IPI/Plenun Data Company, New York, pp. 109ff.
Grenney, W. J., D. B. Porcella, and M. L. Cleave. 1976. Water Quality Rela-
tionships to Flow-Streams and Estuaries. In: Methodologies for the De-
termination of Stream Resource Flow Requirements: An Assessment, C. B.
Stalnaker and J. L. Arnette, eds. Utah State University, Logan, Utah.
Hadden, N., F. Baumann, F. MacDonald, M. Munk, R. Stevenson, D. Gere, F.
Zamaroni, and R. Majors. 1971. Basic Liquid Chromatography. Varian
Aerograph, U.S.A.
Hamaker, J. W., and J. M. Thompson. 1972. Adsorption. Chapter 2 In: Organ-
.ic Chemicals in the Soil Environment. Vol. 1. C.A.I. Goring and J. W.
Hamaker, eds. Marcel Dekker, Inc., New York. pp. 49-143.
Haquc, R., and D. Schroedding. 1975. A Method of Measuring the Water Solubil-
ity of Hydrophobia Chemicals: Solubility of Five Polychlorinated Bi-
phenyls. BuJl. Environ. Contam. Toxicol. 14:13-18.
68
-------
Hendry, n. (;. . 'I. JliLl, L. Piszkiewic/!, J. A. Howard, and H. K. Eigenmann.
197A. A Ci.-iLic.-il Review of H-Atom Transfer in the Liquid Phase:
Clilcrlne Mom, AlkyJ , Trichloromethyl, Alkoxy, and Alkylperoxy] Radicals.
J. Pliys. Chora. Ref. Data 3:937-978.
Hesse, P. R. 1971. .\ Textbook of Soil CheraicaJ Analysis. Chemical Publish-
ing Company, Inc., New York. Chapter 11.
Hill IV, J., eL nl.. 1976. Dynamic Behavior of Vinyl ChJoridc in Aquatic Eco-
systems. U.S. linvironmentul Protection Agency, EPA-600-13-76001.
Isensce, A. R. , P. C. Kearney, E. A. Woolsen, G. E. Jones, and U. P. Williams.
J973. Distribution of Alkyl Arsenicals in Model Ecosystems. Environ.
Sci. Technol. 7:841-845.
Kuznetscv, S. 1. 1970. The Microflora of Lakes and Its Geochemical Activity.
University of Texas, Austin, Texas.
Langbein, W. B-, and W. H. Durum. 1967. The Aeration Capacity of Streams.
Geological Jurvey Circular 542.
Leopold, L. B., M. G. Wolman, and J. P. Miller. 1964. Fluvial Processes in
Geomorphology. W. A. Freeman and Company, San Francisco, California.
522 pp.
Lineweaver, H. and D. Burk. 1934. The Determination of Enzyme Dissociation
Constants, J. Amer. Chem. Soc. 56:658-666.
Mackay, D., and Y. Cohen. 1976. Prediction of Volatilization Rate of Pollu-
tants in Aqueous Systems. Symposium on Nonbiological Transport and
Transformation of Pollutants on Land and Water, Hay 11-13. National
Bureau of Standards, Gaithersburg, Maryland.
Mackay, D. , and P. J. Leinonen. 1975. Rate of Evaporation of Low Solubility
Contaminants from Water Bodies to Atmosphere. Environ. Sci. Tech. 9:
1178-1180.
Mackay, D., and A. W. Wolkoff. 1973. Rate of Evaporation of Low Solubility
Contaminants from Water Bodies to the Atmosphere. Environ. Sci. Tech.
7:611-614.
Metcalf and Eddy, Inc. 1972. Wastewater Engineering: Collection, Treatment,
Disposal. McGraw-Hill, New York, New York; p. 671.
Metcalf, R. L., G. K. Sangha, and I. P. Kapoor. 1971. Model Ecosystem for
the Evaluation of Pesticide Biodegradability and Ecological Magnifica-
tion. Environ. Sci. Technol. 5:709-793.
Michaclis, L., and M. L. Menton. 1913. Kinetics of Invertase Action. Bio-
chcm. Z. 49:333-369.
69
-------
Mill, T., D. C. Hcndry, W. R. Mabey, H. Richardson, B. Y. Lan, and A. Barazc.
1977. Oxi.dar.jon oC Organic Compounds in Dilute Aqueous Solution.
Abstract. I7jrd Meeting of the American Chemical Society, New Orleans,
March 20-25. PIIYS. 222.
Moelwyn-Huglies, E. A. 1971. The Chemical Statics and Kinetics of Solutions.
Academic Press, New York.
Honod, J. L949. The Growth of Bacterial Cultures. Am. Rev. Microbiol. 3:
371-394.
Moses, F. G., R.S.H. Liu, and B. M. Monroe. 1969. The "Merry-Go-Round"
Quantum Yield Apparatus. Hoi. Photochem. 1:245-249.
Paris, D. F,, D. L. Lewis, J. T. Barnett, and G. C. Baughman. 1975. Microbi-
al Degradation and Accumulation of Pesticides in Aquatic Systems. Envi-
mental Protection Agency, EPA-660/3-75-007.
Pirt, S. J. 1975. in Principles of Microbe and Cell Cultivation. John Wiley
& Sons, Inc., New York.
Pitts, J. N., Jr., J.K.S. Wan, and E. A. Schuck. 1964. Photochemical Studies
in an Alkali Halide Matrix. I. An £-Nitrobenzaldehyde Actinometer and
Its Application to a Kinetic Study of the Photoreduction of Benzophenone
by Benzhydrol in a Pressed Potassium Bromide Disk. J. Amer. Chem. Soc.
86:3606-3610.
Rollins, A. L. 1968. Predictor-Corrector Method for Evaluating Systems of
First Order Ordinary Differential Equations. Computer Center, University
of California, Davis, California.
Stratton, F. P., and P. L. McCarty. 1967. Prediction of Nitrification Effects
on the Dissolved Oxygen Balance Df Streams. Environ. Sci. Tech., 1:405-
410.
Stratton, F. E., and P. L. McCarty. 1969. Graphical Evaluation of the Kinet-
ic Parameters for Bacterial Growth. Can. J. Microbiol. 15:1201.
Stumm-Zollinger, E., and R. H. Harris. 1971. Kinetics of Biologically
Mediated Aerobic Oxidation of Organic Compounds in Aquatic Environments,
Faust, S. L., and J. V. Hunter (eds.). Marcel Dekker, Inc., New York,
pp. 555-598.
Taub, F. B. 1973. Biological Models of Freshwater Communities. Ecol. Res.
Series, EPA-660/3-73-008.
Tsivoglou, E. C., et al. 1965. Tracer Measurements of Atmospheric
Reaeration-I. Laboratory Studies. J. Water Pollution Control Federation
37:1343-1362.
70
-------
Tsivoglou, E. C. 1967. Measurement of Stream Reaeration. U.S. Department of
the Interior, Washington, D.C. June 1967.
Wetzcl, R. G. 1975. Limnology. W. B. Saunders Company, Philadelphia, Penn-
sylvania, 743 pp.
Wolfe, N. L., R. G. Zepp, G. L. Baughman, R. C. Fincher, and J. A. Gordon.
1976. Chemical and Photochemical Transformation of Selected Pesticides
in Aquatic Systems. U.S. Environmental Protection Agency,
EPA-600/3-76-067.
Zepp, R. G., and D. M. Cline. 1977. Rates of Direct Photolysis in Aquatic
Environments. Environ. Sci. Technol. 11:359-366.
71
-------
Appendix A
FLOW OF WATER AND SEDIMENTS BETWEEN
COMPARTMENTS IN THE COMPUTER MODEL
72
-------
TABLE A.I. FLOW OF WATER AND SOLIDS BETWEEN
COMPARTMENTS IN THE POND MODEL
Water compartment
1
1
7
Water
(m3hr *)
0.1
Solids
(kg hr '
3.75
Solids compartment
7
Water
(m3hr *)
0.1
Solids
(kg hr
3.75
73
-------
TABLE A. 2. KLOU OF WATLR AND SOLIDS B£TWr.i:N
COMPARTMENTS IN THE RIVER SYSTEM
\ To
From >v
compartments
1
2
3
7
8
9
Water
(m'hr"1)
_
0
0
4.0
0
0
1
Solids
(kg hr"1)
_
0
0
1.08 x 10*
0
0
W.itcr co
2
Water
(ra'hr"1)
1.01 x 106
-
0
0
4.0
0
apjrtmcnLb Solids comp.irLim.-nL
Solids
(kg hr"')
1.01 x 10s
-
0
0
1.08 x 10*
0
3
Hater
(m'hr"1)
0
1.01 x 10*
-
0
0
4.0
Solids WaLcr
(KE hr l) (re'hr l)
0 4.0
1.01 x 10s 0
0
0
0 0
1.08 x 10* 0
7
Solids
(ks hr"1)
1.08 x 10*
0
0
-
0
0
Water
(m'hr"1)
0
4.0
0
0
-
0
8
So I id i
(kf> hr ')
0
l.OS x 10*
0
0
-
0
9
Water
(T'hr"1)
0
0
4.0
0
0
—
Solics
(i.(> IT'')
0
0
1.08 < JO*
0
0
~
TABLL A. 3.
FLOW OF WATI R AND SOLfUS BLTWhEN COMPARTMCTTS IN
Till. CUTKOrillC AND Ol.lCOTROrillC LAkK SYSltHS
\To
1
From \
compart&icncv.
1
2
3
5
7
8
9
Water
(m'hr"1)
_
l.l x 10*
0
0
0.46
0
0
Solids
(kR hr'1)
498
0
0
375
0
0
Mater
2
Water
(m'hr"1)
J.I x 10*
-
1.1 x 10*
800
0
0 .
0
Solids
(kE hi'1)
5.46 x 103
-
4.98 x 103
40
0
0
0
compartments
3
Water
(m'hr"1)
0
1.1 x 10*
-
0
0
0
0.46
Solids
kkfi hr"1)
0
546
-
0
0
0
375
Water
(m3hr">)
0
800
0
-
0
4.63
0
5
Solids
(kg hr"1)
0
40
0
-
0
3750
0
Mater
(m'hr"1)
0.46
0
0
0
-
0
0
Solids
7
Solids
(kR hr"1)
375
0
0
0
-
0
0
compartment
S
Water
(ra'hr"1)
0
0
0
4.63
0
-
0
Solids
(ks hr"1)
0
0
0
3750
0
-
0
h'aicr
(•n'ri"1;
0
0
0.46
0
0
0
-
9
Solids
1 Cke hr"1)
n
0
J75
0
0
0
-
-------
Appendix B
THEORY OF VOLATILIZATION OF ORGANIC SUBSTRATES FROM WATER
The theory of volatilization of slightly soluble organic substances from
aqueous solutions and oxygen reaeration in water has been developed by several
authors.l! They assumed a two-film model in which the rates of diffusion in
air and in water control the rate of transfer of both oxygen and the substrate
across the interface between air and water. (Oxygen and the substrate are
represented by the superscripts 0 and S, respectively, in the equations in
this appendix.)
Figure B.I illustrates the major features of the two-film model of mass
transfer. The water phase is assumed to be well-mixed so that any volatile
component is at a uniform concentration Cg, except in the vicinity of the in-
terface. A stagnant liquid film or concentration boundary layer of thickness
6L separates the bulk of the water phase from the actual interface. Since
turbulence levels in this film are low, any movement of a volatile component
through this film is due to diffusion alone. The concentration of a volatil-
izing component decreases across this film from the bulk concentration Cg to
the interface concentration Cg^. This concentration decrease is the driving
force for mass transport.
On the air side of the interface is a stagnant gas film or concentration
boundary layer of thickness of 6Q, where diffusion is again the only mass
transport mechanism. The partial pressure PSI on the air side of the inter-
face is related to the molar concentration [S^] on the water side of the
interface by Henry's law:
PSI = HC[S.] =HXS. (B.i)
where H and H are the Henry's law constant expressed in concentration (M) or
mole fraction units, respectively, and
Hc: H i55o = 1'8 x 10~3 H (B'2)
P. S. T.iss and P. G. Slater. "Flux of Gases Across the Air-Sea Interface,
Nature 247:181-184 (1974); D. Mackay and Y. Cohen, "Prediction of Volatiliza-
tion Rate of Pollutants in Aqueous Systems," Symposium on Nonbiological
Transport and Transformation of Pollutants on Land and Water, May 11-13,
1976, National Bureau of Standards, Gaithersburg, Maryland.
75
-------
AIR/WATER
INTERFACE
UJ
cc
CO
CO
UJ
DC
Q-
_l
<
f-
oc
a.
cc
o
O
1-
cc
H-
2
UJ
O
2
O
O
WATER
DIFFUSION
TRANSPORT v
CONVECTIVE TRANSPORT
CS
STAGNANT LIQUID FILM
(CONCENTRATION ^
BOUNDARY LAYER)
i
^\
^V
cs.
_,_
}
/
'
PS.
V
\
\
Nj
^
AIR
DIFFUSION
"TRANSPORT
CONVECTIVE TRANSPORT
PS
STAGNANT GAS FILM
^(CONCENTRATION BOUNDARY
LAYER)
DISTANCE
SA-4396-G5
FIGURE B.I. SCHEMATIC OF THE TWO-FILM MODEL OF VOLATILIZATION FROM THE SURFACE
OF WATER BODIES
-------
If we denote the rate at which substrate is being transported across these
films by N , Ln moles dm"2 hr~l, then
b
Ng = K^([S] - [SJ) liquid film (B.3)
KS
Ns=^|(PSi- Ps) 8^ mm (B.4)
Combining equations (B.I), (B.3), and (B.4),we obtain:
(B.5)
where:
kv Overall mass transfer coefficient (hr~l)
A Interfacial area (dm2)
V Liquid volume (dm3)
D Molecular diffusion coefficient (dm2 hr"1)
Hc Henry's law constant (torr M"1)
KL Liquid film mass transfer coefficient (dm hr"1)
K_ Gas film mass transfer coefficient (dm hr~L)
R Gas constant
X Temperature (°K)
A similar equation can be written for oxygen transport.
0 A " — *~1
In a liquid with dilute concentration of oxygen or substrate and when the
amount of material being transferred across the interface into air is small,
the two- film model assumes that
where D is the diffusion coefficient of oxygen or substrate in water and 5L is
the thickness of the mass transfer film or boundary layer on the liquid side
77
-------
of the interface. This relation develops as a simplification of Pick's Jaw of
diffusion. Tn a similar manner, it can be shown that
K =•£- (B.8)
G 6G
where SQ is the thickness of a mass transfer film on the gas side of the gas-
liquid interface and D is the diffusion coefficient of o-:ygcn or substrate in
air. High turbulence in the liquid causes 5L to be thin, and similarly, high
turbulence in the gas phase causes
-------
radon/oxygen, krypton/oxygen, C02/oxygen, and N2/oxygen. For example,
Tsivoglou showed that
-^- = 1.22 + 0.06 experimentally (B.12)
k
v
=1.25 theoretically
over a range of k from 0 to 0.6 hr"
At high levels of liquid turbulence, S^ becomes very small and as a
consequence K^ becomes very large and the gas phase resistance becomes the
rate-controlling step. Based on the findings of Tsivoglou, this occurs when
k§ » 0.6 hr"1, although the actual point where the transition occurs is un-
known. When gas phase resistance is rate controlling, equation (B.9) becomes
AHcKG
and
c c c e c c
k; H;K* HbKb Hbo
-£--^-£--^4 (B.1A)
0 00 00 0
v c G H KG H D0
where here D refers to the diffusion coefficient in air. If it is assumed
that the diffusion coefficients are still inversely proportional to the mole-
cular diameters
HSd°
there is also a transition region where both liquid phase resistance and gas
phase resistance control the transport rate. In this transition region, the
ratio kS/kO must be expressed as the ratio of equations (B.5) and (B.6), and
does not reduce to a simple form.
If data on the diffusion coefficients or molecular diameter for the sub-
strate are not available, molecular diameters can be estimated from the
79
-------
critical voJumu (Vf,), wlurh is a commonly t.ibuKit.'d ,)h-.sn.il con&L.int.'' If
the critical volume cannot be Found, J volume for .1 du.si.l> r.jJjted compound
can be used. The crJtic.iJ vulu.nc is t\.o or three tHin.-s the molecular volume.
From the IPO 1 ecu Kir volume, the molecular diami-ti-r can be calculated by assum-
ing that the molecule as spherical:
, 3 V V
MU c c fr
— = K °r 3^ (I
0 °
where N is Avogndro's number. A widely accepted value for d is 2.98 A.
Values for the Henry's law constant can be estimated from solubility and
vapor pressure following the procedure of Mackay and Wolkoff.* Based on
thermodynamic principles, they determined that
wo
where P is the substrate vapor pressure in pure form and Swo is its
solubility Ln water. If data for the substrate are not available, data for a
related compound can be used.
One source is the American Chemical Society Advances in Chemistry Series,
Vol. 15 (1955).
' R. D. Present, Kinetic Theory of Gases (McGraw-Hill, New York, 1958).
D. Mackay and A. W. VIolkoff, "Rate of Evaporation of Low Solubility Con-
taminates from Water Bodies to the Atmosphere," Environ. Sci. Tech. 7:611-
614 (1973).
80
-------
'
t read
TECHNICAL REPORT DATA
I fa^-bcnons on the reverse before completing)
RfcPOflT MO
3. RECIPIENT'S ACCESSIOWNO.
TITLE AND SUBTITLE
ENVIRONMENTAL PATHWAYS OF SELECTED POLLUTANTS IN
FRESHWATER SYSTEMS
Part I: Background and Experimental Procedures
5. REPORT DATE
August, 1977
6. PERFORMING ORGANIZATION CODE
AUTHQR(S)
J. H. Smith, W. R. Mabey, N. Bohonos, B. R. Holt, S,
Lee. T-W. Chou. D. C. Bomberger. and T. Mill
3. PERFORMING ORGANIZATION REPORT ',;
S.
PERFORMING ORGANIZATION NAME AND AOORCSS
Stanford Research Institute
333 Ravenswood Avenue
Menlo Park, California 94025
10. PROGRAM ELEMENT NO.
INE 625
11. CONTRACT/GRANT NO.
68-03-2227
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Environmental Research Laboratory
College Station Road
Athens. Georgia 30601
13. TYPE OF REPORT AND PERIOD COVEREC
Final. 6/30/75 to 4/30/77
14. SPONSORING AGENCY CODE
S. SUPPLEMENTARY NOTES
16. ABSTRACT
This research was initiated to develop environmental exposure assessment proce-
dures that can be used to predict the pathways of potentially harmful chemicals in
freshwater environments. The approach is based on three premises: (1) the overall
rate of disappearance of a chemical from the aquatic environment is controlled only by
the dominant transformation and transport processes, (2) these processes can be studied
independently in the laboratory, and (3) the laboratory data can be extrapolated to
environmental conditions.
Laboratory procedures have been developed for measuring the rates of volatiliza-
tion, photolysis, oxidation, hydrolysis, and biotransformation as well as the sorption
partition coefficients on natural sediments and on a mixture of four bacteria. Two
aodels have been used to extrapolate the laboratory results to the environment. The
one-compartment model assumes that the aquatic system is a single well-mixed reactor
from which chemicals are transformed, degraded, and/or transported. It can be used to
analyze acute discharges such as spills and to establish priorities for in-depth labo-
ratory studies. The nine-compartment computer model is used to study the effect of
transport and transformation processes studied in the laboratory on the distribution of
a chemical in ponds, streams, and eutrophic and oligotrophic lakes. Part II of this
report describes the application of these procedures to environmental assessment of the
flj.fit:r1b"flon 3nd facg of elgven organic compounds. —
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Croup
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (Thu Rtport)
unclassified
21 NO. OF PAGES
90
20 SECURITY CLASS (This page)
unclassified
CPA Form 2220-1 (3-73)
81
------- |