Unitad States
Environmental Protection
Agency
Office of Pesticides
and Toxic Substances
Washington, OC 20480
PROPERTY OF THE
OFFICE o
EPA 560/5-83-025
Toxic Substances
Environmental Transport
and Transformation
of Polychlorinated Biphenyis
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Cnjcago, IL 60604-3590
-------
-------
ENVIRONMENTAL TRANSPORT AND TRANSFORMATION
OF POLYCHLORINATED BIPHENYLS
by
Asa Leifer, Robert H. Brink, Gary C. Thorn, and
Kenneth G. Partymiller
U.S. Environmental Protection Agency
Office of Toxic Substances
Washington, DC 20460
-------
DISCLAIMER
This report has been reviewed by the Office of Environmental
Processes and Effects Research, Office of Research and
Development, EPA, and the Office of Pesticide Programs, Office of
Pesticides and Toxic Substances, EPA and has been approved for
publication. Mention of trade names or commercial products does
not constitute endorsement.
-ii-
-------
CONTENTS
Page NO.
ABSTRACT iv
INTRODUCTION 1
Asa Leifer
CHAPTER 1: WATER SOLUBILITY AND OCTANOL/WATER
PARTITION COEFFICIENT OF POLYCHLORINATED
3IPHENYLS 1-1
Gary C. Thorn
CHAPTER 2: VAPOR PRESSURE OF POLYCHLORINATED
BIPHENYLS 2-1
Kenneth G. Partymiller
CHAPTER 3: HENRY'S LAW CONSTANT AND VOLATILITY
FROM WATER OF POLYCHLORINATED BIPHENYLS 3-1
Asa Leifer and Kenneth G. Partymiller
CHAPTER 4: ADSORPTION (SORPTION) OF POLYCHLORINATED
BIPHENYLS TO SOILS AND SEDIMENTS 4-1
Asa Leifer
CHAPTER 5: BIOCONCENTRATION OF POLYCHLORINATED
BIPHENYLS IN FISH 5-1
Asa Leifer
CHAPTER 6: ATMOSPHERIC OXIDATION OF POLYCHLORINATED
BIPHENYLS 6-1
Asa Leifer
CHAPTER 7: HYDROLYSIS AND OXIDATION OF POLYCHLORINATED
BIPHENYLS 7-1
Asa Leifer
CHAPTER 8: PHOTOLYSIS OF POLYCHLORINATED BIPHENYLS 8-1
Asa Leifer
CHAPTER 9: BIODEGRADATION OF CHLORINATED BIPHENYLS 9-1
Robert H. Brink
-111-
-------
ABSTRACT
This report summarizes the environmental transport and
•
transformation of polychlorinated biphenyls and contains nine
separate chapters describing water solubility and octanol/water
partition coefficient/ vapor pressure, Henry's law constant and
volatility from water, adsorption (sorption) to soils and
sediments, bioconcentration in fish, atmospheric oxidation,
hydrolysis and oxidation in water, photolysis, and
biodegradation. In the preparation of each of these chapters,
the emphasis has been on obtaining experimental data on
environmentally relevant rate constants and equilibrium constants
for these processes/properties for individual PCB congeners and
Aroclors. If no experimental data were found, then estimation
techniques were used wherever possible to obtain values for the
rate constants or equilibrium constants for each individual
congener or for groups of congeners (i.e., for monochloro-,
dichloro-, trichloro-, etc., biphenyls). These parameters are in
a form suitable for environmental fate modeling.
Since water solubility (Cs) and octanol/water partition
coefficeint (KQW) are key parameters in chemical fate analyses,
work was sponsored by the Chemical Fate Branch of the EPA's
Office of Toxic Substances to obtain precise values for a number
of individual PCB congeners using the very reliable coupled
generator column-chromatographic method. Based on all the
experimental data, regression equations were developed relating
these two key parameters to the parachor. In addition, a
-IV-
-------
regression equation was developed correlating Cs with KQW so that
if the precise experimental value of one of them is known, then a
precise experimental value for the other one can be obtained for
a specific PCS congener.
Adsorption (sorption) to soils and sediments (Koc) and
bioconcentration in fish (KB) are related to ^ow- Based on the
available experimental data published in the literature, linear
regression equations were developed for a wide range of chemicals
correlating KQC and Kg to KOW- The precise values of KQW were
then used to obtain the best estimate of KQC and KQ for a number
of the individual PCS congeners and all the groups of congeners.
For most of the other transport and transformation
processes, estimation techniques were used to obtain
environmentally relevant rate constants and half-lives. For
example, the structure/reactivity method of Hendry and Kenley was
modified to estimate the second-order rate constant (kOH) for the
reaction of OH radicals with the various PCS congeners in the gas
phase. Using these rate constants, half-lives were estimated for
the various congeners in reasonably polluted air. For aqueous
photolysis in sunlight, the available photolysis data for a few
specific PCB congeners [molar absorptivities and quantum yield]
were used along with solar irradiance data in pure water and at
shallow depths to estimate rate constants and half-lives at 40"
north latitude and for the winter and summer solstices. Based on
an analysis of the available biodegradation literature, some
general conclusions have been deduced about potential
-v-
-------
oiociegradation half-lives for the mono- and dichloro-,
trichloro-, tetrachloro-, penta- and higher chlorinated biphenyls
in various environments [i.e., aerobic fresh and oceanic surface
waters, activated sludge, soil, and anaerobic environment], but
it must be emphasized that these are broad generalizations and
that half-lives in particular environments for specific
chlorinated biphenyls may be much larger due to certain limiting
environmental variables [e.g., low temperatures, low moisture, pH
extremes] or the specific PCB structure.
It must be emphasized that these estimates of rates for
transport and transformation involved simplifying assumptions and
thus these data should not be regarded as precise but rather as a
best estimate based on the available data. Precise
environmentally relevant experimental data are needed to confirm
these rate constants and half-lives.
-vi-
-------
INTRODUCTION
Polychlorinated biphenyls (PCBs) represent a. class of
chlorinated aromatic compounds which have found widespread
industrial application because of their stability and inertness,
excellent dielectric properties, and their excellent solvent
characteristics. There are 209 possible PCB congeners when
biphenyl is chlorinated [Hutzinger et al. (1974)]. Figure A
lists the possible distribution of the chlorines on the two rings
and the number of congeners while Figure B lists a few congeners
and their names. Monsanto, the only U.S. manufacturer, has sold
a number of industrial grade PCBs called Aroclors and the
approximate molecular composition of these products are listed in
Figure C.
The fate of PCBs in the environment is a function of a
number of chemical, physical, and biological processes/
properties. These processes/properties are: water solubility,
octanol/water partition coefficient, vapor pressure, Henry's law
constant, volatility from water, adsorption (sorption) to soils
and sediments, bioconcentration in fish, atmospheric oxidation,
hydrolysis and oxidation in water, photolysis, and
biodegradation. These processes/properties can be divided into
two principal categories: transport and transformation. The
first seven are transport processes/properties while the
remaining ones are transformation processes. The transformations
can be further subdivided into abiotic and biological
processes. Whenever possible, these processes/properties have
been expressed as rate constants or equilibrium constants which
are suitable for environmnental fate modeling.
-1-
-------
;re A: Possible Distribution of Chlorines on the Two
Siphenyl Rings and the Number of Congeners
NUMBER OF
CHLORINE ATOMS
ON RING B (y)
"-ITL
3 &
*<*)-
r 4,
C-lu
»Uo'
-®.'1
4' ^'
NUMBER OF CHLORINE ATOMS ON RING A (x)
0 1
013
1 6
2
3
4
5
2345
6631
18 18 9 3
21 36 18 6
21 18 6
6 3
1
-2-
-------
ure 3: Molecular Structure and Names of a Few Selected
Polychlorinated Bip'henyls
3-chlorobipheny1
Cl
\
2,4'-dichlorobiphenyl
Cl
Ci
•C| 2,4,4',6-tetrachlorobiphenyl
c|
Cl 2,2' ,4,4', 6,6'-hexachlorobiphenyl
Cl
-3-
-------
"iqure C: .Approximate Molecular Composition of Aroclors
(Percent) [Hutzinqer et al. 1974]
Empirical
Formula
Cl2H10
~i_2HaCl
C12H8C12
C12H7C13
C12H6C14
C12H5C15
C12H4C16
C12H3C17
C12H2C13
^12^1^9
Average
:
-------
In describing the environmental fate of PCBs, this report
has been divided into nine separate chapters describing these
physical, chemical, and biological processes/properties in
detail, Figure D. Chapters 1-5 are devoted to a discussion of
the transport processes/properties, Chapters 6-8 are devoted to a
discussion of abiotic transformation processes, while Chapter 9
is devoted to a discussion of biological degradation.
In the preparation of each of the chapters, the emphasis has
been on gathering the available experimental data on these
processes/properties for individual PCS congeners and ^roclors.
If no experimental data were found, then estimation techniques
were used wherever possible to obtain values for rate constants
or equilibrium constants for each individual ?CB congener. In
many cases, these estimation techniques were unable to
distinquish between congeners containing the same number of
chlorines on the two rings. For example, for the PC3s with four
chlorines on the two rings,
ci-
ff
there are 21 possible congeners (Figure \)
-5-
-------
Figure D: A List of the Chapters Describing the Transport and
Transformation of PCBs in the Environment
CHAPTER 1: Water Solubility and Octanol/Water Partition
Coefficient of Polychlorinated Biphenyls
CHAPTER 2: Vapor Pressure of Polychlorinated Biphenyls
CHAPTER 3: Henry's Law Constant and Volatility from Water of
Polychlorinated Biphenyls
CHAPTER 4: Adsorption (Sorption) of Polychlorinated Riphenyls
to Soils and Sediments
CHAPTER 5: Bioconcentration of Polychlorinated Biphenyls in Fish
CHAPTER 6: Atmospheric Oxidation of Polychlorinated Biphenyls
CHAPTER 7: Hydrolysis and Oxidation of Polychlorinated Biphenyls
in Water
CHAPTER 8: Photolysis of Polychlorinated Biphenyls
CHAPTER 9: Biodegradation of Chlorinated Biphenyls
-6-
-------
out only one value of a specific property could be obtained for
ail the possibilities. Thus, all the possible congeners were
grouped together as the tetrachloro PCB congener and a single
value of the specific property was listed.
Since water solubility (Cs) and octanol/water partition
coefficient (Kow) are key parameters, work -was sponsored by the
Chemical Fate Branch of the EPA's Office of Toxic Substances to
obtain precise values for a number of PCB congeners using a
coupled generator column-chromatographic method (Chapter 1).
Based on the experimental data/ regression equations were
developed relating these two key parameters to the parachor.
These equations were then used to estimate and obtain precise
values for other groups of congeners.
Adsorption (sorption) to soils and sediments (Koc) and
bioconcentration in fish (Kg) are related to the octanol/water
partition coefficient (Kow). Based on the available experimental
data published in the literature, linear regression equations
were developed for a wide range of chemicals correlating Koc and
K3 to Kow (Chapters 4 and 5, respectively). The precise values
of Kow developed in Chapter 1 were then used to obtain the best
estimate of KQC and Kg for a number of the individual PCB
congeners and all the groups of congeners (i.e., dichloro,
trichloro, tetrachloro, etc).
-7-
-------
For other transport and transformation processes, estimation
techniques were used to obtain environmentally relevant rate
constants and half-lives. It must be emphasized that these
estimation techniques involved simplifying assumptions and thus
these data should not be regarded as precise but rather as a best
estimate based on the available data.
REFERENCE
Hutzinger 0, Safe S, and Zitlco V. 1974. The Chemistry of
PCBs. CRC Press.
-8-
-------
CHAPTER 1
WATER SOLUBILITY AND OCTANOL/WATER
PARTITION COEFFICIENT OF POLYCHLORINATED BIPHENYLS
by
Gary C. Thorn
Contents
Page No.
I. INTRODUCTION AND SUMMARY 1-1
II. LITERATURE DATA ON THE WATER SOLUBILITY OF PCBs 1-8
III. LITERATURE DATA ON THE OCTANOL/WATER PARTITION
COEFFICIENT OF PCBs 1-10
IV. DETERMINATION OF C AND K FOR SOME
POLYCHLORINATED BIPHENYLS 1-12
V. CONCLUSION 1-22
VI. REFERENCES 1-23
-------
-------
I. INTRODUCTION AND SUMMARY ,
This chapter discusses the water solubility (Cs) and
octanol/water partition coefficient (KQW) of PCBs and gives
experimental and/or estimated values for all PCS congeners.
C_ and K-... are fundamental physicochemical properties that
o W W
significantly affect the transport and transformation of a
chemical in the environment. For hydrophobic chemicals such as
PCBs, the water solubility can affect the rate at which the
chemical is distributed in the environment and can influence the
rate and extent of chemical and biological transformations such
as aqueous photolysis and biodegradation. In addition, the water
solubility can be correlated to the octanol/water partition
coefficient. The partition coefficient is important because it
is useful in predicting biological uptake, lipophilic storage and
soil adsorption (i.e., properties that quantify the distribution
of a chemical between the hydrophobic and hydrophilic
compartments in the environment). With PCBs so widespread in the
environment, their water solubility and octanol/water partition
coefficients have become of critical importance in assessing
their transport and transformation in the environment.
The availability of precise, reliable data on the water
solubility and octanol/water partition coefficient of PCBs is
limited. There are two principal reasons for this lack of
data. First, the highly hydrophobic nature of PCBs makes the
determination of their concentration in water (a measurement that
is required for both Cs and KOW) difficult from an analytical
1-1
-------
standpoint because they are present in such low concentrations.
Second, with the exception of the work of Wasik et al. (1981)
that will be discussed below, all methods for C_ and K use a
o O W *
"shake-flask" procedure that leads to the formation of emulsions
or colloids in the equilibration step. For very hydrophobic
chemicals, such as ?C3s, colloid formation is difficult to
control and leads to erroneously high and/or erratic PCS
concentrations in water.
Wasik et al. (1982a,b) have circumvented the colloid problem
by using a generator column or column leaching method coupled
with a gas chroraatograph/election capture detector. This method
gives reliable, precise values of Cs and KQW for even the most
hydrophobic PCBs. In addition, Wasik has correlated both the C
and KQW of PCBs with the Sugden parachor (Par) for each PCS
congener. Thus, in the absence of experimental values, the Cs
and/or Kow of a PCS can be estimated from its calculated
parachor.
Although there are other methods for estimating these
parameters [e.g., Hansch and Leo's (1979) fragment constant
method and the linear regression correlations summarized by Lyman
et al. (1982)], the Wasik correlations are much more reliable
because they represent the direct correlation of precise Cs or
!<_.. data with the parachor.
O w
Tables 1 and 2 summarize the data on the water solubility
and octanol/water partition coefficient; Cs is given in jug/L (ppb)
«
and moles/liter (M) while Kow is given in log units (log KQW).
1-2
-------
o
in
.u
•5
M
U
3>
V
I
V
c
x*4
U
i
•
9
3
W
u
•«*
3! —
•a —
•5 2
u ~
£
i
u «
U *3
3 £
U
2 -aS
*2 S
u
«
e a —
1 ^J
— "o
fl*
0 ^
u — — .
1) ^ J
£ ^ B^
5 -c
U
3)
3 3
2 I
0
? i i i '
o o a o o
XXX X X
t*l if) n 93 **
(^ - m vO -
CD
O a — irt •
p« O »»-in
>a a^ra m —
-------
-1
.*»
f*- •
"*"
^
3
s
fN
•3
33
y
*•
y
i
j*
.•^
**
^
w
^
••
^
•*
<-»
"t
—
—
'
j*
*•
*"*
i
*
^
3
~
£
w
"
\ ^
n — '
a
H- *S
3 i
I
i ^2
•^ M
\ ~
U
^4
—4 A
-x^ 3"
^^ A
M — •
U «
S »«
b 0
3 S
a
w
a
•*•* -«»
- _: .:
^ a.
51 a.
^ ^^
^
b
V
••4
C S
S at *~
S ^}
3 ^^
Z* 'T?
0
b j;
0 ->
iJ ijj
9 S
5) "^ ^Q
C ? 2,
£ V O-
w ^S-
u
—
™ i
£ ='
0
O — i-
\ \ 1*111
o G o a o a
x x x x x x
,- m <•> tj
(Q C
Z u
— • 2 "5
«- , a S as
v ^ *
j) fM M
4J WJ 2
JJ ^J
£ X 3 53
jj *^f! * ^i
3 *3 > 1
« J3 U ~.
-------
Table 2. Leg Cccanol/watar Partition Coefficients
of PC3 Congeners, at 25 9C
PC3
Congener
Biphenyl
Monoehloro-
2
3
4
Dichloro-
2,2'
2,4'
2,5
2,5
3,4
4,4'
Triehloro-
2, 2', 5
2,4,5
2,4,5
2,4,5
Tetrachloro-
2, 2', 3,3'
2, 2'4', 5
2,3,4,5
2,3,5,6
?«ntachloro-
2,2-4,5,5'
2,3,4,5,5
Hexachloro-
2, 2', 3, 3', 4, 4'
2, 2', 3, 3', 6, 6'
Generator
Column
Method13
3.74
3.39s
4.50
4.38s
4.58s
4.49s
4.90s
5.14s
5.16
4.93
5.29
5.33
5.60s
5.51
5.31s
5.77s
5.47
5.73
5.72
5.92
6.30
6.98
6.65
literature
Values Estimated3
4.04d 4.09
3.75a
4.03f
4.51
4.28
4.94
' 5.189
5.53h
5.37
5.90
4.63?
S.46g
6.24
611
6.67
2,2',4,4'/5,5'
2,2I,4,4',6,6'
7.55a
6.72f
6.34k
1-5
-------
Table 2. Log Octanol/water Partition Coefficients (Cont.d)
?C3
Congener
Heotachloro-
2,2', 3,3', 4,4', 6
Octachloro-
2, 2' , 3, 3',
5, 5', 6, 6'
Nonochloro-
Generator
Column Literature
Methodb values
6.63
7.11
Estimated3
7.10
7.53
7.96
2,2',3,3',4,
5,5',6,6'
Deeachloro-
3.16
3.26
3.39
bWasik, S., et al., 1982.
^oodburn, K.3., 1982.
~3anerjee, S., et al., 1980.
eVeith, G.D., et al., 1979.
*Yalkowsky, S.H. and s.H. Valvani, 1979.
^Sugiura, K. at al., 1973.
hChiou, C.T., et al., 1977.
^Mackay, 0., et al., 1980.
:<:
-------
Both taoles give values as determined by three methods: (1)
estimated fron the parachor; (2) experimentally determined using
Wasi'-c's generator column method, and; (3) experimentally
determined using other methods published in the literature. In
selecting a value from these tables, the value determined by the
generator column method is preferred, if one exists for the
particular congener of interest; if not then the literature value
can be used. If neither of these values are given, the estimated
value should be used. It should be noted, however, that the
estimated value applies to all congeners, i.e., all tetrachloro
congeners have an estimated Cs of 41.8 ppb. This is because the
parachor calculation does not distinguish between different
positions that the chlorine can occupy on the biphenyl ring.
1-7
-------
II. LITERATURE DATA ON THE WATER SOLUBILITY OF FCBs '
A comprehensive summary of the water solubility data on PCBs
has been made by Mackay et al. (1980). It is these data (as
individual values or averages) that are included in Table 1 as
the literature value. These values were measured using various
modifications of the standard shake-flask method along with a
variety of analytical techniques for determining the aqueous
concentration of the PCS.
A comparison of the literature values with those obtained
using the generator column method show considerable
discrepancies. Wide variation also exists between the individual
literature values used in computing the average values given in
Table 1. For example, the average GS value of 1140 ppb for the
2,2' congener represents the average of two significantly
different values - 790 and 1500 ppb. The Mackay summary shows
many other similar discrepancies that lead one to question the
reliability of the procedures used to determine the water
solubility, and of the analytical techniques employed. The major
factor responsible for these discrepancies is the formation and
failure to remove emulsified or colloidal PCB from the aqueous
solution. Another factor is the use of an insufficiently
reliable and sensitive analytical technique, particularly for the
more hydrophobic PCBs, where, for example, values range over a
factor of ten for 2,2',4,4',6,6'-PCB. Thus, it seems clear that
an alternate method is needed for determining PCB water
solubility — a method that avoids the problems of colloid
1-8
-------
formation and removal, and contains a reliable and precise
analytical procedure. To fulfill this need the Chemical Fate
Branch of EPA's Office of Toxic Substances sponsored research
work at the National Bureau of Standards (Wasik et al. 1981) to
develop a coupled generator column/chromatographic method that
could be used to obtain reliable and precise measurements of C_
o
(and KQW) particularly for very hydrophobic chemicals such as
PC3s. The results of this research are discussed in Section IV.
1-9
-------
III. LITERATURE DATA ON THE OCTANOL/WATER PARTITION COEFFICIENT
OF PCBs
As shown in Table 2, few experimental values of the
octanol/water partition coefficient for PCBs have been deter-
mined. Of those values published in the literature, most have
been estimated from retention time on a high pressure liquid
chromatograph (Sugiura, 1978). Because so little data are avail-
able on the partition coefficient of PCBs, most values must be
obtained from the several methods available for estimating log
KQW. The most widely used method for estimating this parameter
is that of Hansch and Leo (1979), who have compiled over 10,000
log KOW values for a wide variety of chemicals. Their method
sums the "fragment constants" for each atom in a molecule to give
the log KQW for the entire, intact molecule. This method has
been computerized by Chou and Jurs (1979).
Another method for estimating log KQW is through correlation
with water solubility. There are a number of equations
correlating log KQW with solubility. These are reviewed and
summarized by Lyman et al. (1982). Each equation has been
obtained for certain classes of compounds thus making the
correlation as good as possible. Two of these equations are:
log KQW = -1.085 log Cg + 4.538 (1)
and
log KQW = -0.73 log Cs + 5.30 (2)
1-10
-------
Equation (I) was developed to correlate the two parameters for a
t
large nunoer of pesticides and PCBs and, therefore, should give
more "accurate" values of log KQW than equation (2) which applies
to chlorinated benzenes.
Although these estimation techniques enable the log KQW of
any PCS congener to be calculated, most values are only accurate
to within a factor of ± 1 log units, due to scatter around the
regression line and to the fact that these equations do not
contain a lattice energy correction term for solids. In
addition, it should be noted that none of the estimation
techniques distinguish between congeners; that is, all dichlor -
PCB's have the same log KQW, regardless of the positions of the
chlorines on the two phenyl rings. This problem has recently
been addressed by Woodburn (1982), who developed an additional
" ir " factor that accounts for the position of the chlorine.
However, his technique does not have the broad applicability or
the advantage of obtaining precise experimental data afforded by
the generator column method discussed in the next section.
1-11
-------
IV. DETERMINATION OF C,. AND KOM FOR SOME POLYCHLORINATED
BIPHENYLS BY THE GENERATOR COLUMN METHOD
Because of the lack of reliable, precise data on the K and
Cs of PCBs/ EPA's Chemical Fate Branch recently sponsored a
laboratory research program at the National Bureau of Standards
to experimentally determine the values of these parameters for a
series of 16 representative PCBs, ranging from the mono- to
decachloro congeners. The measurements were made using the
generator column method developed by Wasik et al . (1981) which is
ideally suited for obtaining reliable, precise data on very
«
hydrophobic chemicals such as PCBs. The principal advantage of
this method is that it avoids the formation of solute emulsions
and colloids that introduce considerable error in both the K_..
ow
and Cs measurements. The method accomplishes this by eliminating
the agitation used to attain solute-solvent equilibrium in the
shake-flask method. Instead, a column is used to generate
equilibrium saturated solute-solvent solutions by the aqueous
leaching of the solute from an appropriate column support; the
concentration of the solute is then determined using a gas
chromatograph equipped with an electron capture detector.
The results of this work have been reported (Wasik et al .
1982) and the values for the Cs and KQW at 25 °C are given in
Table 3. As a part of this study Wasik was able to show that
both parameters are correlated to the solute parachor, Par. The
parachor was first demonstrated to be directly related to K... and
o w
Cs by McGowan (1966) through the equation:
log KQW = k'-m Par
1-12
-------
where k1^ is a constant that is characteristic of the
octanol/water solute system. The parachor is an ideal parameter
-o correlate with K__. and C_ because it can be easily calculated
O " 3
by summing the individual parachor values of the atoms and
structural units that make up the molecule. An example of how
the parachor is calculated for 1-chlorobiphenyl is given below:
Atom or Structure Sugden Parachor
C 4.8
H 17.1
Cl 54.3
double bond 23.2
six-membered ring 6.1
For 1-chlorobiphenyl
12 C 57.6
9 H 153.9
1C 1 54.3
6 double bonds 139.2
2 six-membered rings 12.2
TOTAL 417.2
*
From Dreisbach compilation [1955].
The major disadvantage of the parachor estimation method is
that it does not distinguish between congeners. For example, the
1-13
-------
two monochlorobiphenyl congeners, 1-CBP and 2-CBP have the same
•
parachor (417) and therefore, the same estimated water
soluoility. However, the experimental data (Table 1) show that
these two congeners have very different water solubilities: 5020
and 2400 ppb, respectively. As a consequence, the log Cs - Par
correlation does not give unique values for each congener, but
the same value for both congeners. The same problem applies to
the log KQW - Par correlation.
The parachor for each of the 16 PCB's along with the
experimentally determined values of log KQW and Cs are given in
Table 3. Correlation of K w and Cs with the parachor using
linear regression analysis gave the following results that are
shown graphically in Figures 1 and 2.
For the octanol/water partition coefficient:
log KQW = 0.0116 Par - 0.33
with a coefficient of determination, r = 0.936; for the water
solubility given in moles/liter:
log 1/C_ = 0.0182 Par - 2.78
with a coefficient of determination, r = 0.954,
1-14
-------
Table 3. Aqueous Solubilities and Octanol/Water Partition Coefficients of
Polychlorinated Biphenvls at 25°C as Measured tJsinq the Generator Column
Compound
1 .
2.
3.
4.
Biphenyl
2-Chlororbiphenyl
2,5-Qichlorbi-
phenyl (liquid)
2 , 6-Oichlorobi-
?arachor
330.0
417.2
454.4
454.4
Cs
(4.35 _+_
(2.68 +_
(8.70^
(6.23 +
0.14)
0.03)
0.21)
0.13)
(M)a
X 10~5
x 10~5
x 10~6
x 10"6
*»«.
3.74 +_
4.50 +_
5.16 +_
4.93 +
r
0.01
0.01
0.01
0.02
biohenyl
5. 2,4,5-Trichloro-
biphenyl 491.6
6. 2,4,6-Trichloro-
biphenyl 491.6
7. 2,3,4,5-Tetrachloro-
biphenyl 528.8
9. 2,2',4',5-Tetra-
chlorobiphenyl 528.8
9. 2,3,4,5,6-Penta-
chlorobiphenyl 566.0
10. 2,2',4,5,5'-Penta-
chlorobiphenyl 566.0
11. 2,2',3,3',6,6'-
Hexachlorobiphenyl 603.2
12. 2,2',3,3'/4,4'-Hexa
chlorobiphenyl 603.2
13. 2,2'/4,4l,6,6t- 603.2
Hexachlorobiphenyl
14. 2,2',3,3',4,4'6- 640.4
Heptachlorobiphenyl
15. 2,2',3,3',5,5',6,6'- 677.6
Octachlorobiphenyl
16. 2,2',3,3',4,5,5',6,6' 714.8
-Nonchlorobiphenyl
17. 2,2',3,3',4,4',5,5',6,6'
-Decachlorobiphenyl 752.0
(6.32 + 0.31) x 10'
,-7
(8.76 + 0.47) x 10
,-7
(7.17 + 0.34) x 10
~8
(7.84 +_ 0.36) x 10~10
(1 .13 + 0.05) x 10~9
(9.15 •)• 0.50) x 10
-10
5.51 + 0.1 1
5.47 0.03
5,72 + 0.07
(5.63 + 0.33) x 10"8 5.73 + 0.09
(1.63 + 0.08) x 10~8 6.30 +• 0.05
(5.92 •*• 0.27) x 10~8 5.92 +• 0.01
(1.67 -t- 0.08) x 10~8 6.65 + 0.06
6.98 +_ 0.04
7.55 + 0.21
(5.49 + 0.24) x 10~9 6.68 + 0.20
7.1 1 + 0.30
(3.88 + 0.17) x 10"11 8.16 + 0.22
(1 .49 -i- 0.19) x 10"'
-11
8.26 -)• 0.10
iThe uncertainty is the standard deviation for three replicate measurements.
-------
-------
0)
«3
*J
s
'"
C3
0.
i.
a
T3
.-3
S_
a
• ' v
C1
c
CJ
3
-------
Figure 3 shows the correlation between the number of
chlorines on the biphenyl and log KQW. The resulting regression
equation:
log KQW = 0.43 ncl + 4.08
has a coefficient of determination, r2 = 0.936. This figure also
shows the error that would result if Hansch and Leo's fragment
constant method were used to calculate log KOW. The slope, or u
value, of their curve differs significantly. Recently however,
Woodburn (1982) has shown that a series of ir factors could be
developed for PCBs based on chlorine position. Use of these new ir
factors is limited to estimating log KQW values of less than 6.
Thus, it appears that the parachor parameter is, at the
present time, the most reliable method for estimating the
octanol/water partition coefficient and water solubility of those
PCBs for which no experimental data are available. As an equally
good alternative one can estimate either Cs or log KQW if the
other is known, from the regression equation derived from
correlating these two parameters. Using the data in Table 3, the
form of the regression equations are:
log KQW = 0.594 log 1/CS + 1.80
that can be rearranged to give:
log 1/CS = 1.68 log KQW - 3.03
1-18
-------
SOT
-------
For both equations, Cs is expressed in moles/liter and the
coefficient of determination is 0.980. An advantage of using
these regression equations is that, unlike the parachor-Cs-Kow
correlations, unique values are obtained for each PC3 congener.
Finally, if the melting point of the PCS congener is known,
the water solubility can be estimated from the following
equation:
log 1/CS » 2.82 log K^ + 0.0147 t^ - 1.91
where Cs is given in moles per liter and t is the melting point
in *C of the PCS; the coefficient of determination for this
equation is, r2 = 0.97.2. Inclusion of the melting point term
corrects for heat of fusion or crystal lattice energy that
increases as the solubility of a substance decreases. As
indicated by the value of r2, this equation does not necessarily
give a better estimation of the water solubility than the other
equations because only one of the PCBs used in developing the
equation was not a solid (i.e., was a liquid). Thus, the melting
point factor has already been "included" in the other two
regression equations.
Table 4 summarizes the various regression equations that
have been described in this paper for estimating the water
solubility and octanol/water partition coefficients of PCBs.
1-20
-------
Table 4. Summary of Regression Equations for Estimating the
water Solubility and OctanoI/Water Partition
Coefficients of Polychlorinated Biphenyls.
Water Solubility (given in moles/liter)
log l/Cg = 0.0182 Par - 2.78 r2 - 0.954
2
log 1/CS = 1.68 log KQW - 3.03 r2 - 0.980
log 1/C, = 2.32 log Knw + 0.0147 tm("C) - 1.91 r2 * 0.972
Octanol/Water Partition Coefficient
1
-------
V. CONCLUSION
A summary of the available and most recent data on the
water solUbility (Cs) and octanol/water partition coefficients
(KOW) of polychlorinated biphenyls (PCBs) has been made. Those
?C3s for which no experimental data -exist can have their C£ and
KOW estimated using regression equations recently developed at
the National Bureau of Standards (NBS). In addition, the MBS has
used a new more reliable and precise method to experimentally
determine the C_ and K__. for 16 PCBs. The data and information
5 \j W
contained in this report can be used to obtain the Cs and KQW for
ail PCB congeners.
1-22
-------
VI. REFERENCES
Banerjee S, Yalkowsky S, and Valvani S. 1980. Water
solubility and octanol/water partition coefficients of
organics. Limitations of the solubility-partition
coefficient correlation. Environ Sci and Tech 14: 1227-1229.
Chiou C, Freed V, Schmedding D, and Kohnert R. 1977.
Partition coefficient and bioaccumulation of selected organic
chemicals. Environ Sci and Tech 11:475-478.
Chou JT and Jurs PC. 1979. Computer assisted computation of
partition coefficients from molecular structures using
fragment constants. J Chem Inf Comput Sci 19:172-175.
Dexter RN and Parlou SP. 1978. Mass solubility and aqueous
activity coefficients of stable organic chemicals in the
marine environment: Polychlorinated biphenyls. Marine Chem
6:41-53.
Dreisbach RR. 1955. Physical Properties of Chemical
Compounds. American Chemical Society/ Washington, DC.
Hansch C and Leo A. 1979. Substituent Constants for
Correlation Analysis in Chemistry and Biology. John Wiley
and Sons. New York.
Karickhoff SW, Brown DS and Scott TA. 1979. Sorption of
hydrophobia pollutants on natural sediments. Water Res
13:241-248.
Lyman WJ, Reehl WF and Rosenblatt DH. 1982. Handbook of
Chemical Prooerty Estimation Methods. McGraw-Hill Book Co.
New York, NY*.
Mackay Df Mascarenhas R and Shiu WY. Aqueous solubility of
polychlorinated biphenyls. Chemosphere 9:257-264.
McGowan JC, Atkinson PN and Ruddle LH. 1966. The Physical
Toxicity of Chemicals. V. Interaction Terms for Solubilities and
Partition Coefficients. J Appl Chem 16:99-102.
Sugiura K, Ito N, Matsmoto N, Mihara Y, Murata K, Tsukakoshi
Y and Goto M. 1978. Accumulation of polychlorinated
biphenyls and polybrominated biphenyls in fish: Limitation
of "Correlation between partition coefficients and
accumulation factors." Chemosphere 9: 731-736.
Veith GD, Austin MM and Morris RT. 1979. A rapid method
for estimating log P for organic chemicals. Water Res
13:43-47.
1-23
-------
Wasik SP, Tewari YB, Miller MM and Martire DE. 1981.
Octanol/Water partition coefficients and aqueous
solubilities of organic compounds. National Bureau of
Standards, MBSIR 81-2406. Washington, DC.
Uasik SP, Tewari YB, Miller MM and Purnall JH. 1982a.
Measurements of the octanol/water partition coefficient by
chromatographic methods. NBS J Res 87:311-315.
Wasik SP, Tewari YB, Miller MM, Martire DE and G'hodbane S.
1982b. Water solubility and octanol/water partition
coefficient of polychlorinated biphenyls and other selected
substances - Task 1C" Final Report to Office of Toxic
Substances, Chemical Fate Branch, USEPA, Washington, DC.
Woodburn KB. 1982. Measurement and application of the
octanol/water partition coefficient for selected
polychlorinated biphenyls. Master's Thesis. University of
Wisconsin - Madison.
Yalkowsky SH and Valvani SC. 1979. Solubilities and
partitioning. 2. Relationships between aqueous
solubilities, partition coefficients, and molecular surface
areas of rigid aromatic hydrocarbons. J Chem Eng Data
24:127-129.
1-24
-------
CHAPTER 2
VAPOR PRESSURE OF POLYCHLORINATSD BIPHENYLS
by
Kenneth G. Partymiller
Contents
Page No,
I. INTRODUCTION AND SUMMARY 2-1
II. MEASURED VALUES OF PCS VAPOR PRESSURE 2-6
III. ESTIMATED VALUES OF PCS VAPOR PRESSURE 2-7
IV. REFERENCES 2-10
2-i
-------
-------
T. INTRODUCTION AMD SUMMARY
A knowledge of the vapor pressure of various PCB congeners
is important in predicting the behavior and fate of these
compounds in the environment. A knowledge of the vapor pressure
of the various PCBs will allow the estimation of rates of
evaporation of PCBs from spills, the rate of volatilization of
PC3s from water, and possible maximum levels of PCRs in the air.
Very little data exists in the literature on the vapor
pressure of Aroclors (mixtures of PCBs). Even less data are
available on the vapor pressure of individual PCBs. Vapor
pressure data on Aroclor mixtures are presented in Table 1. This
table also lists the main PCBs which are present in each Aroclor
and their percent in the mixture. Table 2 lists all measured
vapor pressure data for individual PCS congeners which could be
found in the open literature.
Due to the apparent paucity of vapor pressure data on PCBs,
it was necessary to consider other methods for obtaining the
vapor pressure of the various congeners. Lyman et al. (1982)
describe a number of estimation techniques in a handbook which
are useful for estimating physicochemical properties of organic
chemicals. Several methods for estimating vapor pressure are
included in Chapter 14 [Lyman et al. (1982)]. These methods
require a boiling temperature as an input. However, this
property also had to be estimated since none of these data are
available in the published literature. Therefore, the boiling
temperature was estimated by the method recommended by Lyman et
al. (1982). Both of the above mentioned estimation techniques
2-1
-------
Table 1. Vapor Pressures of Aroclors at 258C
Vapor Pressure
Aroclor (Torr)
1016 4 x 10~4 (estimate)3
1221 6.7 x 10~3 (estimate)3
1231 4.06 x 10~3 (estimate)3
1242 4.06 x 10~4 (measured)5
1243 4.94 x 10"4 (measured)5
1254 7.71 x 10"5 (measured)5
1260 4.05 x 10"5 (measured)5
Primary PCBs
and Percent
57%
51%
31%
24%
28%
49%
40%
36%
43%
38%
41%
Present
Composition0
C12H7C13
C12H9C1
C12H9C1
C12H8C12
C12H7C13
C12H7C13
C12H6C14
C12H5C15
C12H5C15
C12H4C16
C12H3C17
1USEPA (1979) .
^Monsanto (1974) Extrapolated to 258C from higher temperatures.
cTable C, of the Introduction, lists the complex composition of
the Aroclors.
2-2
-------
Table 2. Measured Vaoor Pressures of PC3s
PC3 Tempera far e( °C)
2' ,3,4
2,2' ,5,5'
2 , 2 ' , 4 , 5 , S '
3,3'
-2,2' ,4,4' ,6,6'
25
25
30
25
25
30
25
25
30
25
25
0
1
1
5
1
3
9
7
1
2
1
.3
.0
.3
.5
.9
.6
.2
.2
.3
.0
.2
X
X
X
X
X
X
X
X
X
X
X
Vapor Pressure (Torr)
10
10
10
10
10
10
10
10
10
10
-4
«B <1 .
-4
-5'
-Si
-Si
-61
-5<
-4(
10-5<
(a)
(b)
(b)
(a)
(b)
(b)
!a)
;b)
r.b)
!a)
!a)
(extrapolated from 30"C)
(extrapolated from 30°C)
(extrapolated from 30°C)
a..
Westcott and aidleman (1981).
Westcott et al. (1981).
2-3
-------
are insensitive to location of chlorine atoms on the biphenyl
rings so that, for example, all pentachloro biphenyls will be
predicted to have identical vapor pressures and boiling points.
Table 3 lists estimated vapor pressures of all possible PCS
congeners (with the above mentioned caveat) as estimated, using
estimated boiling temperatures, by the method preferred for PCBs.
In summary, it is apparent that there is only limited
experimental data on vapor pressures of a few PCBs. Furthermore,
the estimation techniques used to estimate the vapor pressures of
the various PCS congeners have two limitations:
1) these estimation techniques cannot distinguish between
congeners of PCBs containing the same number of
chlorines; and
2) the estimation techniques for the vapor pressure of the
PCB congeners containing 7 to 10 chlorine atoms are not
reliable since these PCBs have vapor pressures below the
useful limit of the estimation technique.
It is thus obvious that carefully conducted experiments must
be performed on a series of PCB congeners to obtain reliable
vapor pressures. With these data, one should be able to develop
empirical correlations for determining the vapor pressure of all
PCB congeners.
2-4
-------
Table 3. Estimated Vapor Pressures of PCS Congeners
4 of C1 ' s
0
1
2
3
4
5
6
7
3
9
1 0
Es
Vapo
(To
7
2
1
*
8
3
7
1
(3
(7
( 1
(4
tima ted
r Pressure
rr 3 25»C)
.3
.3
.1
.5
.7
.6
.7
.6
.6
.9
.0
x 1-2
x 10
x 1 O""2
x 1 0
x 1 0
x 1 0~7
x 10~7
x 10)
x 1 O"9 )
. x 1 o" )
x 10-10)b
Physical State
Liquid
Liquid
Liquid
Solid
Solid
Solid
Solid
Solid
Solid
Solid
Solid
a?redicted by vapor pressure estimation method 2 as described in
Lyraan at al. (1982).
Values in parentheses are outside of the useful ranae of the
estimation technique and must be used with caution.
-------
II. MEASURED VALUES OF PCS VAPOR PRESSURE
The vapor pressure data on Aroclors (PCS mixtures) is given
in Table 1 [USEPA (1979), Monsanto (1974)]. These data were
either estimated or measured. It should be noted that Monsanto
was the only US manufacturer and supplier of PCBs for a number of
years.
The PCS vapor pressure data listed in Table 2 has "been
determined by two techniques. Westcott et al. (1981) measured
the vapor pressure of a few PCBs by the gas saturation method at
temperatures of 30, 35, and 40°C and the data were then
extrapolated to 25°C using the Clapeyron or Antoine equation.
The gas saturation technique which was used is very similar to
the one published by EPA as an OTS Test Guideline [USEPA
(1982)]. In the method used by Westcott et al. (1978), air was
used as the carrier gas rather than the nitrogen recommended in
the EPA procedure. The second technique used by Westcott
[Westcott and Bidleman (1981)] utilized a capillary gas
chomatographic technique which determines vapor pressure by
measuring the PCB retention times relative to a reference
compound of known pressure.
2-6
-------
III. ESTIMATED VALUES OF PCS VAPOR PRESSURS
Several of the Aroclor vapor pressure values listed in
Table 1 were estimated [USEPA (1979)]. The exact source and/or
method of estimation of these values is unknown. These values
appear to be in agreement with expected values.
Vapor pressures of the pure PCS congeners were estimated by
the methods recommended by Lyman et al. (1982). These methods
require a boiling temperature as an input. Boiling temperatures
were not found in the literature and thus they were estimated by
the method recommended by Lyman et al. (1982). The recommended
methods [Lyman et al. (1982)] for the estimation of both vapor
pressure and boiling temperature are not sensitive to position of
chlorine atoms on the biphenyl rings so that all congeners will
have identical estimated vapor pressure and boiling
temperature. Table 4 lists the calculated boiling points of the
PCS congeners. Lyman et al. (1982) recommended two methods for
estimating vapor pressures and these are designated as Method 1
and Method 2. Method 1 is applicable only to liquids or super-
cooled liquids. Method 2 is applicable to both solids and
liquids. Method 1 has a lower vapor pressure limit of 10"3 Torr
while Method 2 is useful down to 10 Torr. Values less than
10~7 Torr are out of the useful range of the estimation
techniques but have been included in both Tables 3 and 4 as they
are only the available numbers. These values must be used with
caution. Method 2 makes corrections for variations in AH /AZ
with temperature (AHv is the heat of vaporization while AZ is the
2-7
-------
compressability factor) while Method 1 assumes that this ratio is
invariant. For the above three reasons, Method 2 values are
preferred. Values estimated by both techniques are listed in
Table 4.
2-8
-------
Table 4. Estimated Boiling Tenoeratures and Vapor Pressures of ?CBs
* of C^'s
0
1
2
3
4
5
&
7
3
9
10
aZstimated
k~ -i »^
(1982).
Boiling Tenn . '
( »C at 1 atai)
247
263
238
308
326
344
361
378
39 S
410
426
by method of Meissner
7aoor Pressure
Method 1(b)
5.2 x 10~2
1.7 x 10~2
5.S x 10'3
(1 .8 X 10'3)8
(5.9 x 1Q-4)8
(1 .9 x 10~4)8
(-6.4 x 10'5)a
(2.1 x 10~5)a
(6.3 x 10~6)8
(2.1 x 10"6)a
(6.6 x 10"7)a
as described by
by vapor pressure estimation method 1
Method 2(c)
7.3 x 10~2
2.8 x 10"2
1 .1 x 10~2
8.5 x 10"5
3.7 x 10"S
7.6 x 10"7
1.7 x 10~7
(3.6 x 108-)8
(7.6 x 10'9)8
(1.9 x 10~9)8
(4.0 x 10"10)
Lyman et al.
as described
Physical
State (d)
Liquid
Liquid
Liquid
Solid
Solid
Solid
Solid
Solid
Solid
Solid
8 Solid
(1982).
by Lyman et al.
\_ __ * _ ._ _ *
(1982).
3ased upon estimated melting teaperatures.
out of range of usefulness of the estimation technique.
2-9
-------
V. REFERENCES
Lyman WJ, Reehl WF, and Rosenblatt DH. 1982. Handbook of
Chemical Property Estimation Methods, McGraw-Hill. New
York.
Monsanto. 1974. Aroclor Plasticizers. Technical Bulletin,
0/PL-306A.
USEPA. 1979. Water-related Environmental Fate of 120
Priority Pollutants. Washington, D.C. EPA-440/4-79-029a.
USEPA. 1982. Chemical Fate Test Guidelines. EPA 560/6-82-
003. NTIS publication PB 82-233008. Test Guideline CG-1600,
Westcott JW and Bidleman TF. 1981. J Chromatogr 210:331-6.
Westcott W, Simon CG, and Bidleman TF. 1981. Environ Sci
Tech. 15:1375-1378.-
2-10
-------
CHAPTER 3
HENRY'S LAW CONSTANT AND VOLATILITY FROM WATER
OF POLYCHLORINATED BIPHENYLS
by
Asa Leifer and Kenneth G. Partymiller
Contents
Page No
I. INTRODUCTION AND SUMMARY 3-1
II. MEASURED VALUES OF PCB VAPOR PRESSURE 3-7
A. Background 3-7
B. Henry's Law Constant 3-9
C. Rates and Half-Lives of Volatilization from
Water . „ 3-12
III. REFERENCES 3-17
3-i
-------
I. INTRODUCTION AND SUMMARY
Volatilization from water bodies to the atmosphere can be a
significant environmental pathway for hydrophobic chemicals such
as polychlorinated biphenyls. Volatilization rate data is needed
to estimate the amount of chemical that enters the atmosphere and
the change in concentration of pollutants in the water bodies.
The mass transfer of chemical from water to the atmosphere is
dependent upon chemical and physical properties (e.g., water
solubility, vapor pressure, and thus Henry's law constant), the
presence of other pollutants in the water body, and environmental
properties (e.g., water body depth, flow rate, and turbulence?
and the wind speed above the water).
Henry's law constant (H) is an important physical property
of the chemical which is used in the calculation of rates of
volatilization. Under equilibrium conditions, the value of H
gives the direction of transfer. Chemicals having H in the range
3 x 10 to 1 x 10"^ atm m mole" are considered to be
moderately volatile. Chemicals with higher or lower values of H
are considered to be very volatile or nonvolatile, respectively.
Mackay and Leinonen (1975) estimated H from the vapor
pressure and water solubility of Aroclors 1242, 1248, 1254, and
1260 and the results are summarized in Table 3, Section II.B.
Doskey and Andren (1981), based on some experimental data,
reported H for the Aroclor 1016 and this result is summarized in
Table 3. The Aroclors are complex mixtures of PCBs and the
composition of each Aroclor is defined in the Introduction,
Figure C.
3-1
-------
Mackay and Leinonen (1975) developed a model and equations
to predict the liquid-phase mass transfer coefficient (KL) and
che half-lives (t l/ ). These equations were used to estimate K^
and t I/ at a depth of 100 cm for the Aroclors 1242, 1248, 1254,
and 1260, Table 3, Section II.B. The half-lives ranged from 10-
12 hours and thus they are quite volatile from pure water.
Doskey and Andren (1981), based on experimental work published in
the literature, estimated KL for the Aroclors 1016, 1221, 1242,
and 1254 and these results are summarized in Table 4, Section
III.3.
No literature data has been found on the determination of
rates and half-lives of volatilization for any of the PCS
congeners. Hence, estimation techniques were employed to obtain
these data. One of the best methods of estimating volatilization
half-lives from water bodies is described by Lyman et al. (1982).
The method is based on the two-film concept for estimating the
flux of volatile chemicals across an air-water interface and the
work of Mackay and Leinonen (1975). Detailed procedures are
described by Lyman et al. (1982) to estimate the overall liquid-
phase mass transfer coefficient (KL) and half-lives (t y ) under
environmental conditions.
Sates of volatilization from pure water and half-lives were
estimated for all the PC3s listed in Tables 1 and 2. The
environmental conditions were: depth in water body = 100 cm;
river flow rate = 100 cm sec ~^; wind speed = 300 cm sec"^-. The
volatilization half-lives for the PC3s listed in Table 1, in
3-2
-------
1
1
a
3
> —
-. a
j a
1 b
u 3
•* a
« a
• 41
c £
0 *
-4 b
*•> 3
S|
- >
p»
-4 -^
^ ^"
3 *
o a
> 0
•^
~ ^i
0 -i
•u .4
fl **4 •
J -3
*> z
5 S
^ w
2 a
a jj
4J 3
e
« -0
• I 21
a 2
« W
c s
0 a
CJ -9
3 i
ifl
"•!
H
M Xt
*£
"*" ^^
-
9
>*
«
S
—
CN a
^x b
9 •- i
4J •«»
^,
1
b
-, • --
* t
5
*^
_^
^
i
4)
>4
0
a="a
S
-8
£ 5
> -
**
(W
'a
u
« a •*
° 1
b
4>
a
jyi
I
|
•
(N 33 C
03 P> ^"
*~
•J1 ffi O
• • *
CO CS iT>
T T T
o o o
^ ^ ^
XXX
o tfl a>
« • •
•9 m -
r> 03 03
i i i
0 2 2
XXX
(N W —
. . •
«.9
X
S
i
W3
US
«
V
*
?J
M
•
e
e
•a
3
b
0
JJ
Q
b
S
4)
^
b
4)
JJ
Oi
a
U
§
b
u
a
4)
3
Measured val
*
O
4
in
CN
U
<8
>Q
U
•«rf
**
Vl
3
M«
O
a
b
0
ij
>a
<8
•
M
b
2
*
6
9
O
b
«d
a
aaured value
£
»
* i
v^
a
in
M
a
•s
0
b
3
a
a
0
b
&
b
0
1
3.
•
O
5,
fM
U
«
JJ
C
«
^J
a
£
0
u
2
fl
^-4
a
V
b
C
2
5
*j
18
«-«
^j
3!
a
u
O
a
m
p»
4J
a
a
c
0
MR4
U
—4
*4
a
•»• -^
^
i >.
tw — >
^ ~4
fl 0
- 0
^4
^3 43
2 >
a >O
S. C
•^ —4
« 3
a ,
4) 4}
JZ
41 JJ
= T3
*
^M^0
« 1
33 U
01 4)
— a
"" e
• o
18 O
* 2
o
«
« V
n>
e»
s e
recommended
' river flowl
s iT
— 41
44 -J
>. S
2 U
4» 0
-------
a
3 — •
> a
-«4 JJ
_: w
1 3
tu a
•- a
S u
o w
- 2
a Sf
M >
— 4 «
JJ 4.!
0 —
> jj
1 3
JJ "5
a s
_s -a
"3 a
3.2
a
"3 -.
C —
a -*
u "^
e —
2 S
9]
s u
2 i
3
a u
*>• a
|i
U 03
« '—
2
^i a
^» u
i
w
** ^ *" *
e
5
~
i
®
0
= =
1
**
^..^
j a
a- a
> "-•
inut m ^*tfi
1 (I II 11 I t I
o oo oo oo o oo
X XX XX XX X XX
P* P. CO *JD ^C f*4 *~
o
ii i i i T
o o o o o o
.- • — — .- f- —
X X X • X X X
f^ «T " 9\ O tV
• • • • « •
1*1 — T — CM
fN ?nf^ ^^ in in p*»in
1 II II in 1 1 II
o -oo oo 10 o oo
*•
X XX XX XX X XX
a « « a * r.* -' i-' -
i
i -
- ^
0 0 » ^O
0 u 1 u 1 0 •* -
w . 0 0 m 0 vfl fc • -
— 8! 3 ^. in 2 m* — «.
^3 31 1-4 1 1 0 ^ » U ••— >rn
O 1 ^* ««u «%3 »«^ «<2] ^«
Z ?N £| CNfN r* ?M(N &* *Nr*J Zt
-------
a
S3
>
••4
T
«
e
0
^j
«
M
.«4
•»4
JJ
Id
O
"3
jj
(0
- *s
« 1
2u C
«*•
1 1
*•*
jj ^^
c
fl
^
c
0
o
3
id
a
£
I
"
^1
D
^•*
.2
«
5«
(N 31
•v u
^ „
|
U
s
o
^^
1
1
U (*9
•a fi
s
Id
I
> -
^^
f*1
's
<•">
1 -t 1)
1^ >•
Id w ^
~ O -4
3 5 -^
dj ^-* 3
id > 9>
jj e 3
• ca -^ i
— . o 3 "
c «
S 2 2 S
3 ^ *•• G*
8W
.1 1
^ ^*» (Q fl
a (N^ >
13 flS 1 C
S 3 a °
c a *o
5) « 4)
^^ id C3 id 43
J S
_ . . jj C 7
* » «- i« e
8*4 m s a j;
jj a M > jj
6 id JJ -C 3 33
•C 4) C 0 <3 *JJ
4J *C >w 4) ^4 y 33
W 3) 3J S ^u dj
3 JJ C S i u
a « • a 0 14 jj £.
id SJJ i- oo •—
a-H ffjj a> 3w
S JJ
u Q g jj a 3
« • o a a
in w a >* T3 a ^
O4 a 3 a .£ j2 JJ
in .c au "O'j^
(B AI *U ^^ O
jj a i -* o 3
jj ^. 2 T* •- "?i
>4 a w 3 44 "3s
•4 W C B* -4 0 -o
-4 3 1J — fl -U au
3 a >n — — *
^4 a^"c<^i» ""^s
o w a a a o >d ^
aa.jjjjjj'- > «
'•& a uua S3
id jj u ^ a *4
-------
which H was estimated from the measured water solubility, and
•
vapor pressure, ranged from 5.7 hours to 14 hours. Thus,, these
?C3s are very volatile from pure water under these specific
environmental conditions.
The volatilization half-lives for selected PCBs for each
class of congeners (i.e., monochloro, dichloro, , decachloro)
under the specific environmental conditions are listed in
Table 2. It must be emphasized that H, used in these
calculations, was obtained from measured water solubility data
but from estimated values of the vapor pressure for these
congeners. For the mono to the hexachloro congers (exclusive of
the 2,2',3,3',6 ,6' congener) t 1^ ranged from 5-32 hours and thus
they are very volatile to moderately volatile. The half-lives
for the hepta to the decachloro congeners ranged from 31-210
hours and these are moderately volatile. However, the latter
results must be used with caution since the vapor pressures of
these congeners fell outside the useful range of the vapor
pressure estimation techniques (Chapter 2) and thus H and t ]/ are
uncertain. In general, the data indicate that all PCBs are very
•volatile to moderately volatile in pure water. However, reliable
volatilization data are needed on selected PCS congeners to
confirm these results.
3-6
-------
II. DISCUSSION OF RESULTS
A. Background
Volatilization from water bodies to the atmosphere is
recognized as a significant environmental pathway for solutes
such as gases and some hydrophobic organic pollutants such as
hydrocarbons and chlorinated hydrocarbons. Volatilization rate
data are necessary to estimate the amount of chemical that enters
the atmosphere and the change in concentration of pollutants in
the water bodies. The mass transfer of a chemical from water to
the atmosphere is dependent upon (1) the chemical and physical
properties of the chemical; (2) the presence of other pollutants
present in the water body; and (3) environmental properties such
as the water body flow velocity, depth/ and turbulence and the
atmospheric conditions above the water body (e.g./ wind speed).
Henry's law constant H is an important physical property of
a chemical used in calculating rates of volatilization. Under
equilibrium conditions, the value of H immediately gives, the
direction of transfer. In most cases, the distribution of
resistance to mass transfer between the atmosphere and a water
body, and hence the overall rate of volatilization, depends on H.
Henry's law constant is defined as the ratio of the vapor
pressure of a solute above an aqueous solution to the
concentration of the solute in the water body at a fixed
temperature at equilibrium. Mathematically, it is defined by the
equation
H = VP/CS , (1)
3-7
-------
where V? is the vapor pressure of the solute above the water
phase and Cs is the water solubility. Mackay et al. (1979)
describes one of the best methods for measuring H directly.
Henry's law constant can also be estimated from the measured
vapor pressure and the water solubility of a solute at a given
temperature. H is often expressed in the units atm m3 mole "^ so
that in equation 1, the vapor pressure (VP) is expressed in
atmospheres and the water solubility (Cs) is expressed in moles
m~3. In addition, it is often convenient to express H as a
dimensionless parameter H' defined by the equation
H1 = H/RT , (2)
where H is Henry's law constant in atm m^ mole""^, R is the gas
constant and is equal to 8.21xlO~5 atm m^ mole "^ °K~1 and T is
the absolute temperature in °K.
A class of chemicals having a value of the H less than
3x10'^ atm m^ mole~^ is considered to be nonvolatile from
water. Examples of chemicals that fall in this class are 3-
bromo-1-propenol, Dieldrin, and ionic organic compounds. Another
class of chemicals having H in the range 3x10 lxlO~3 atm m^ mole'l are considered to be
highly volatile from water. Examples of chemicals that fall in
this class are biphenyl, methylene chloride, o-xylene, and
ethylene.
3-8
-------
One of the best methods of estimating half-lives of
volatilization from water bodies is described in detail by Lyman
at al. (1982) . The method is based on the two-film concept for
estimating the flux of volatile chemicals across an air-water
interface and the work of Mackay and Leinonen (1975) . Detailed
procedures are described by Lyman et al. (1982) to estimate half-
lives for volatilization of chemicals in water bodies under
environmental conditions . The detailed procedures require the
following minimum data: (1) physical properties - vapor
pressure, aqueous solubility, and molar mass; (2) environmental
properties — wind speed, current speed, and depth in a water
body. The half-life for volatilization under environmental
conditions is given by
tu
(3)
where Z is the depth in a water body in cm and K- is the overall
liquid-phase mass transfer coefficient in cm hr~ . The method of
estimating KL under environmental conditions is described, and
examples are given to illustrate how to use the method.
B. 'Henry's Law Constant
Mackay and Leinonen (1975) estimated Henry's law constant
from the vapor pressure and water solubility of Aroclors 1242,
1243, 1254, and 1260, Table 3. Doskey and Andren (1981), based
on experimental work of Paris et al. (1978), reported H for the
Aroclors 1016 and 1242, Table 3. Aroclors are complex mixtures
of PC3s, and the composition of each grade is defined in Figure C
of the Introduction.
3-9
-------
No literature data have been found on the direct
determination of Henry's law constant for any of the PC3
congeners. However, H can be estimated from the experimental
values of the water solubility and vapor pressure in equation 1.
H was estimated from the experimental data for a few PCBs
[(21,3,4,-trichloro), (2,2',5,5'-tetrachloro), 2,2',4,5,5'-
pentachloro), (2,2',4,4',6,6'-hexachloro)] from Chapters 1 and 2 and
the results are summarized in Table 1. H for the tri-, tetra-, and
pentachlorobiphenyls are in the range 1.9-5.5 x 10~4 atm m^ mole"*,
and thus these PCBs are moderately volatile. H for the
hexachlorobiphenyl is 1.4 x 10~2 atm m3 mole~^, and thus this PCB is
highly volatile.
For a number of PCB congeners, the water solubility was
determined experimentally (Chapter 1). However, experimental
values of the vapor pressure for these congeners were not
available; hence they were estimated by the procedure described
in Chapter 2. Table 2 lists the values of H for selected PCBs
for each class of congeners (i.e., monochloro, dichloro,
trichloro, , decachloro) using the measured water
solubilities and estimated vapor'pressures. H for all the
congeners are in the range 10~3 to 10~5 atm m^ mole""1- and thus
they are moderately volatile. However, the values of H for the
heptachloro to the decachloro congeners must be used with caution
since the estimated vapor pressure of these congeners fell
outside the useful range of the vapor pressure estimation
technique (Chapter 2).
3-10
-------
Cable 3. Estimated Henry's Law Constant, Rates, and
Half-Lives for Several Aroclors
3H 1 "l-l 2
Aroclor (ata m mole ) (ca hr ) hrs
1016 (1.4 ± 0.7) x 10"2
1242 5.73 x 10~4 5.7 12.1
1248 3.51 x 10~3 7.2 9.5
1254 2.76 x 10~3 6.7 10.3
1260 7.13 x 10~3 6.7 10.2
3-11
-------
C. Rates and Half-Lives of Volatilization £rom Water
Mackay and Leinonen (1975) developed a model and equation
to predict the liquid-phase mass transfer coefficient KL and the
half-life, and this model formed the basis of the method of Lyman
et al. (1982) as described in Section II.A. These equations were
used by Mackay and Leinonen to estimate K^ and t Lc at a depth of
1 meter (100 cm) for the Aroclors 1242, 1248, 1254, and 1260,
Table 3. The half-lives ranged from 10-12 hrs, and thus these
Aroclors are quite volatile.
Doskey and Andren (1981) , based on experimental data
published in the literature, estimated KL for the Aroclors 1016,
1221, 1242, and 1254, and these results are summarized in
Table 4.
Paris et al. (1978) studied the volatilization of Aroclors
1016 and 1242 in the presence of sediments in three pond waters
by the reaeration rate method developed by Smith et al. (1979,
1980). Thus, volatilization and adsorption were taking place
simultaneously. These researchers determined the ratio of the
rate coefficients k/k2» where k is the rate coefficient for PCB
loss and k2 i-s tne reaeration rate constant, and compared these
results with the theoretically estimated ratio.
3-12
-------
91
U
0
O
0
u
<<
4)
0
C/l
u
0
'4-1
0!
U
4)
^
U
i
•g
" ~
i-^
^
10
xj
c
1)
g
-*4
w
8,
>s
<"?
1
O
X
CN —
1 •
O
0
X
" C>
in i •
• o m
(N i-
44
^1
-H X
r^* ^^
-l
u
0
w
<
!u
T i,
« -2 * *
« JJ C C
•i c « a)
•2 u V4 u
S U 11
en < <
— 14 <4
o o«
u 2 >. >.
10
-------
The results are given in Table 5/ and the agreement is quite
good. These researchers concluded that for water containing no
sediments, volatilization would be an important pathway. For
water containing sediments, in which Aroclors are highly
adsorbed, these Aroclors would be transported within the aquatic
environment in association with the sediments.
No literature data has been found on the determination of
t
rates and half-lives of volatilization for any of the PCS
congeners. Rates of volatilization (KL) and half-lives (t I/, }
were estimated for all the PCB congeners listed in Tables 1 and 2
as described by the method outlined in Section II.A. These
values are summarized in Tables 1 and 2. The environmental
conditions were: depth of water body (2) = 100 cm; river flow
rate = 100 cm sec"1; wind speed = 300 cm sec"1. The
volatilization half-lives for the PCBs listed in Table 1, in
which H was estimated from the measured water solubility and
vapor pressure, ranged from 5.7 hours for the congener
2 ,2', 4 ,4',6 ,6'-hexachlorobiphenyl to 14 hours for the congener
2,2',4,5,5'-pentachlorobiphenyl. Thus, these PCBs are very
volatile under these specific environmental conditions.
The volatilization half-lives for selected PCBs for each
class of congeners (i.e., monochloro, dichloro, trichloro, ,
decachloro) under the specified environmental conditions are
listed in Table 2. It should be noted that H, used to estimate
t I/ , was obtained from measured water solubility data and
estimated values of the vapor pressure for these congeners.
3-14
-------
u
1
V
ij ^ .
a
3~
IU«-
0 1 ^
*» ffi
— S
3 -S
-H
1 §
u
^, 93
a w
a
4J
M*
U3
^a in ^
o o* o
i i i
•» «*
• • •
t* & V
a & *
•^
a
a.
a
1 3
a. = -a
e
e > o
3 I* &
a. o
x *
a u o
S-t «]
= =
•
91
0
^4
JJ
4
••4
i.
a
u
4J
•a
i
u
w
O
a
jj
P^
•s
to
a
as
«
-------
For most of the congeners, t 1«. ranged from 5-32 hours for the
monochloro to the hexachloro cpngeners (excluding the congener
2 , 2' , 3 , 3'3 , 6 , 6'-hexachlorobiphenyl) . Thus, these PC3s range from
very volatile to moderately volatile chemicals. The other PCS
congeners from heptachloro to decachloro had half-lives ranging
from 31-210 hours and are thus moderately volatile chemicals.
However, these results must be used with caution since the vapor
pressure of these congeners fell outside the useful range of the
vapor pressure estimation techniques (Chapter 2) and thus H and
t I/ are uncertain. In general, the data indicate that all PC3s
are very volatile to moderately volatile in pure water. However,
reliable volatilization rate data are needed for selected PCS
congeners to confirm these results.
3-16
-------
III. REFERENCES
Doskey PV and Andren AW. 1981. Modeling the flux of
atmospheric polychlorinated biphenyls across the air/water
interface. Environ Sci Tech 15:705.
Lyman WJ, Ruhl WF, and Rosenblatt DH. 1982. Handbook of
chemical property estimation methods. McGraw Hill Book Co.
N. Y.
Mackay D and Leinonen PJ. 1975. Rates of evaporation of
low-solubility contaminants from water bodies to
atmosphere. Environ Sci Tech 9:1178.
Mackay D and Wolkoff AW. 1973. Rate of evaporation of
low-solubility contaminants from water bodies to
atmosphere. Environ Sci Tech 7:611.
Paris DF, Steen WG, and Baughman GL. 1975. Role of
physico-chemical properties of Aroclors 1016 and 1242 in
determining their fate and transport in aquatic
environments. Chemosphere 4:319.
Smith JH and Bomberger DC. 1979. Prediction of
volatilization rates of chemicals in water. Water:
1978. AICHE Symposium Series, 190, 75:375.
Smith JH, Bomberger DC, Haynes DL. 1980. Prediction of
the volatilization rates of high volatility chemicals from
natural water bodies. Environ Sci Tech 14:1332.
•3 _i
-------
-------
CHAPTER 4
ADSORPTION (SORPTION) OF POLYCHLORINATED BIPHENYLS
TO SOILS AND SEDIMENTS
by
Asa Leifer
Contents
Page No
I. INTRODUCTION AND SUMMARY 4-1
II. DISCUSSION OF RESULTS 4-7
A. Background 4-7
B. Experimental Data on Soil/pediment
Adsorption (Sorption) and Mobility of
Polychlorinated Biphenyls 4-9
C. Estimation of Adsorption (Sorption) to
Soils and Sediments from the Octanol/
Water Partition Coefficient and Water
Solubility 4-15
III. REFERENCES 4-21
4-i
-------
-------
I. INTRODUCTION AND SUMMARY
The extent to which an organic chemical partitions itself
between water and soil or sediment is determined by several
physical and chemical properties of both the chemical and the
soil or sediment. In most cases, it is "possible to express the
tendency of a chemical to be adsorbed (or sorbed) in terms of the
equilibrium constant K^ derived from the Preundlich isotherm
equation. It is generally accepted that for a given soil or
sediment, the sorption of neutral organic molecules can be well
correlated with the organic matter content in the soil or sedi-
ment [Karickhoff (1981), Karickhoff et al. (1979), Lyman et al.
(1982)]. Consequently, the adsorption constant K^ can be con-
verted to the more generally useful constant "
-------
deficient in that several of the compounds were crystalline
solids at 25°C and a crystal energy correction term must be added
to the correlation [Karickhoff (1981)].
Karickhoff carried out two adsorption (sorption) studies
and correlated log KQC wi.th log KQW [Karickhoff et al. (1979),
Karickhoff (1981)]. In the research work published in 1981,
Karickhoff chose five reference compounds (benzene, naphthalene,
phenanthrene, anthracene, and pyrene) and the KQW and KQC data
were fitted to give the equation
log KQC = log KQW - 0.386 , (1)
with a correlation coefficient (r ) of 0.994. Karickhoff also
expressed the correlation in the more traditional form and
obtained the equation
log KQC = 0.989 log KQW - 0.346 , (2)
with a correlation coefficient (r ) of 0.997. It should be noted
that for these five reference compounds, kQW and KQC varied by
only three orders of magnitude.
Karickhoff (1981) listed the published experimental values
of KQW and KOC for 17 hydrocarbons and chlorinated hydrocarbons
(including two hexachloro PC3s), six chloro-s-triazines, three
carbamates, four organophosphates, six phenyl ureas and six
miscellaneous compounds. The predicted Values of KQC were within
4-2
-------
0.43 log units or within a factor of three except for certain
classes of compounds. That is, compounds for which solute
speciation could be expected were excluded. The phenyl urea
class was also excluded for either speciation or for poor
experimental data.
In developing a regression equation which would have more
widespread applicability to more different hydrocarbons and
chlorinated hydrocarbons, it was decided to use all ?2 hydro-
carbons including Karickoff's five reference compounds and the
four high molar mass chlorinated compounds (i.e., Di")T, methoxy-
chlor, and the two hexachloro-biphenyls). Linear correlation
analysis on all 22 compounds gave the relationship
log Koc = 0.942 log KQW - 0.144 , (3)
with a correlation coefficient (r2) of 0.985. This equation is
more applicable to a larger set of hydrocarbons and chlorinated
hydrocarbons (N=22) and both KQW and KQC vary by 4.5 and 4.7
orders of magnitude, respectively, in comnarison to equations 1
and 2 which only apply to a data set of 5 compounds and both KQW
and *
-------
dicnlorobiphenyls, trichlorobip'henyls, etc.] using the method
outlined in Chapter 1. These values were used to estimate log
KQ from equation 3. All these data are summarized in Table 1.
Inspection of all the data indicates that PCBs are strongly
adsorbed (sorbed) to soils and sediments high in organic content
and are immobile.
4-4
-------
Table 1. Log KQC for the PCS Congeners from Estimated and
Experimental Values of Log KQW Using Ecuation 3
PCS Congener
Monochloro-
2-
3-
4-
Dichloro-
2,2'-
2,4'-
2,5-
2,6-
3,4-
4,4'-
Trichloro-
2,2',S-
2,4,5-
2,4',5-
2,4,6-
Tetrachloro-
2,2',3,3'-
2,2',4',5-
2,3,4,5-
2,3,5,6-
Pentachloro-
2, 2', 4,5, 5'-
2,3,4,5,6-
Hexachloro-
2, 2'3, 3', 4, 4'-
2, 2', 3,3', 6, 6'-
2, 2', 4, 4', 6,6'-
log KQW
^st. Exp.
4.51
4.44a
4.58
4.49
4.94
4.90
5«14
5.16
4.93
5.29
5.33
5.37
5.60
5.66a
5.79
5.47
5.80
4.63b
5.73
5.72
5.46b
6.24
5.29
6.30
6.67
6.98
6.65
7.55
I*OQ K-. -,
oc
4.10
4.04
4.17
4.09
4.51
4.47
4.70
4.72
4. SO
4.34
4.88
4.92
5.13
5.19 .
5.31
5.01
5.32
4.22
5.25
5.24
5.00
5-. 7 3
5.43
5.79
6.14
6.43
6.12
6.97
6.34C 5.83
4-5
-------
Table 1. Log K__ for the PCS Congeners (Continued)
QG
log KOW
PC3 Congener Sst. Sxp. Log K
'oc
Heptachloro- 7. 10
2,2' ,2,3', 4,4' ,6- 6.68
Octachloro- 7.53
2,2', 3,3', 5,5', 6,6'- 7.11
Nonachloro- 7.96
2, 2', 3, 3', 4,5,5', 6, 6'- 8.16
Decachloro-
2,2', 3,3', 4,4', 5,5', 6,6'- 3.26
6.54
6.15
6.95
6.55
7.35
7.54
7.64
Almost all experimental data were obtained from the coupled column
generator—chromatographic method, Chapter 1.
a. These values represent the average of the experimental
results of NBS and Woodburn.
b. These values were measured by the reverse-phase liquid
chromatographic method.
c. This value was measured by the conventional shake-flask method.
method discussed in Chapter 1.
All estimated by KQW values for the PCS Congeners were obtained by
the method discussed in Chapter 1.
4-6
-------
A. Backq round
The extent to which an organic chemical partitions
itself between water and soil or sediment is determined by
several physical and chemical properties of both the chemical and
the soil or sediment. In most cases, it is possible to express
the tendency of a chemical to be adsorbed (or sorbed) in terms of
the equilibrium constant, Kd, in the Preundlich isotherm equation
x/m = KdC1/n , (4)
where x/m is the a gm of chemical adsorbed/gm of soil or
sediment, C is the concentration of chemical in the acqueous
solution, and n is a constant. In general, 1/n is close to unity
[Karickhoff (1981) , Karickhoff et al (1971) , Lyman et al. (1982) 1
and equation becomes
x/m - KdC . (5)
It is generally accepted that Cor a given soil or
sediment type, the sorption of neutral organic molecules can be
well correlated with the organic matter content in the soil or
sediment [Karickhoff (1981) , Karickhoff et al. (19?*) , Lyman et
al. (1982)1 and the adsorption constant K^ can be converted to
the more general and useful constant KQC defined by the
relationship
4-7
-------
Kd/oc
where oc is the fractional mass of organic carbon in the soil or
sediment.
The conventional method for measuring adsorption
coefficients is to determine an adsorption isotherm with one or
more soils or sediments with well defined characteristics.
Specific soil/solution ratios of soil or sediment are prepared in
water using at least six different initial concentrations of the
chemical being studied. The mixture is generally shaken For at
least 48 hours to achieve equilibrium and the concentrations in
the water and on the soil or sediment are then measured. The
amount adsorbed (x/m) and the equilibrium solution concentration
are fitted to equation 4 to determine Kd and n. Using the
fractional mass of organic carbon in the soil or sediment, KQC is
obtained from equation 6.
Helling and Turner (1958) introduced an alternative
method for measuring the mobility of chemicals on soils which is
related to adsorption. This method is called soil thin-layer
chromatography (soil TLC) and is analogous to conventional TLC
with the use of soils instead of silica gels, oxides, etc. as the
adsorbent phase. The term designating movement in the soil TLC
method is R^, defined as the furthest distance traveled by a
chemical on a soil TLC plate divided by the distance traveled by
the solvent front (arbitrarily set at 10.0 cm in soil TLC
studies). 3ased on considerable research work, a relationshio
-------
has been established between Kd, Rf, and mobility on soils, and
the results are summarized in Table 2. The details of this work
and soil TLC are discussed in depth in an EPA test guideline and
support document [US EPA (1981)].
3. Experimental Data on Soil/Sediment Adsorption
(Sorption) and Mobility of Polychlorinated Biphenyls
Griffin et a'l. (1978) measured Rf values for the
Aroclors 1242 and 1254 on four soils and the results are given in
Table 3. The principal PCB components in these two commercial
products are (49 percent trichloro, 25 percent tetrachloro, 16
percent dichloro) and (48 percent pentachloro, 23 percent
hexachloro, 21 percent tetrachloro), respectively. Figure C of
the Introduction gives the complete composition of these two
Aroclors. The Rg values with water as the solvent is in the
range 0.02-0.03. For landfill leachates as the solvent, the
range was 0.02 to 0.04. Thus, these two Aroclors are highly
adsorbed and are immobile, Table 2. However, it should be noted
that this method does not give precise measures of adsorption for
Rf values less than 0.1.
Haque et al. (1974) reported the adsorption
coefficients for Aroclor 1254 on Woodburn soil. The values for
the constants n and K^ from equation 4 are given in Table 4. The
composition of Aroclor 1254 has been described previously. Haque
and Schmedding (1976) continued these studies with the PCB
congeners (2,4'-dichloro), 2,2',5,5'-tetrachloro), and
2,2' ,4,4' ,5,5'-hexachloro) on Woodburn soil and humic acid at 24
± 2°C and the values of n and K
-------
e
o
3 JJ
-4 _
0 *"4
CO •— 4
=1
0
en ->
| S
OT
g C
4) «
| .
2 *"
e. TJ
??
a jj
c e
JJ "y
«S ,-
4} U
fl£ 4)
-8
(Q
U -
« o
A) "2
C -
^ fl
F fi»
(N
4)
2
fl
s-
"3
fl
«2
5,
2
^
(8
o
5
Q
•o
a
s-
4)
0
Hj
^
a
4)
o
«
91
Q
a
to
(Q
0
—4
^4
Q
2
s
M
^
s
U
o
(N
S1
JJ
CP
s
U
£^
^
09
O
1-4
,
U
I
in
^
o
^J
•
o
J
o
a
3
a
a
s
—4
^
^4
8-
V
2
^
U
>
o
o
"
19
4>
d
AJ
J
^4
^4
0
a
0
4J
C
••4
^
1
«
4)
«
U
§**
JZ
U
'o
£
Q
I
OJ
•w
^
2
ui
^i
o
0
•
t*
"o
o
s
4J
c
^
c
0
-4
TI
w
JJ
c
4)
O
e
§
•o
a,
^
e
(Q
e
^
jC
o
«-4
O
^^
^
•^
1
5
S
0
o
^
o
^4
3
9
S
U
ul
8-
U
c
^4
i-4
1
0
£
s
0
JJ
|Q
S
4)
U
C
0
o
4)
.s
0
E
S
H*
O
0
*
o
m
-------
Table 3. R« Values for Two Aroclors on ?our Different Soils
with the Solvents Water and Landfill Leachate
Rf
(Landfill
Aroclor Soil Type (Water) Leachate)
1242
Ottawa Sand
Catlin loam C
Ava silty clay loam B-
Catlin silt loam A,,
0.03
0.02
0.02
0.02
0.03
0.03
0.02
0.04
1254 Ottawa Sand 0.03 0.03
Catlin loam C 0.02 0.03
Ava silty clay loam B. 0.02 0.02
Catlin silt loam A_ 0.02 0.04
4-11
-------
-
jjjj
a
0
3
e
c
VI
u
a
s
0
u
23
O
V
u
_
?
^J
S
"V
^
in
0
r-4
0
J*
0
3)
jj
C
^
a
CJ
B?
rtj U
.£ **•
• j
0 «
M 2
** —
5 "o
£ ^
Ii
* w
*
™
M
^5
si*
-
«
*F
c
•e
^
c
J8
u
0
a
•e
I
C
«
2
ai
en ^ T
O C C
X XX
t O 0* 00 (**•
fM r» r^ T *•
*• CD ^3 ^? ^^
00 — 03 C OC
O ^™ o *™ en
<*•< r?
0 0
^ ^*
X X
P> O CC C^ ^
• O Cf. • m
vO • • 00 •
f-( ^ pH
•H ^ -^
O 0 0
W M T W ^
•H *H
C CO C U
w w < w <
£ £ U 5 O
0 0 "g 0 'g
1 Si i i
o o o o o
(N {N •
in
•
^
•>
•*
tN
«T
4J
C
^
1
"S
S
c
c
7]
S
*
a
Q.
I
ep
C
.#4
e:
5
39
JjJ
*
-------
In general, the dichloro congener showed some adsorption to soil
and huraic acid and. some mobility on these soil profiles may be
expected. The tetrachloro congener exhibits less mobility while
the hexachloro congener should be immobile.
Karickhoff et al. (1979) determined Koc for
2,2'4,4'6,6'-hexachlorobiphenyl from adsorption isotherm
experiments using Doe Run Pond and Hickory Run Pond sediments at
25°C. The average KQC value for both of these sediments is
listed in Table 5.
Karickhoff (1981) , using the experimental data of Haque
and Schmedding (1976), calculated KQC for the congener
2,2',4,4',5,5'-hexachlorobiphenyl and found that log KQC was
equal to 5.62.
Paris et al. (1978) carried out a simple adsorption
experiment for Aroclor 1016 and 1242 on three different pond
sediments at 25°C and characterized adsorption by the equation
K » CA/CW , (7)
where CA is the concentration of the component on the sediment
and Cw is the concentration of the component in water at
equilibrium. These results are summarized in Table 5.
4-13
-------
'J,
JJ
5
e
.j
T*
A*
at
w
^
0
2
n
a
o
c.
4)
C
«
vu
0
Constants
e
0
•- <
jj
u
0
m
<**
<
in
4)
13
e*
m
l
O
^«
X
0
^
(*1
1
o
X
vv
X
£
Sediment T
ffi
U
0.
c
a
«N
*~
ffi O P» c» in t-
n d ci o r* -i e >H c H
0 -«^ 0 •*•* 0 "*
a = -c a. = T: cu z:
C>0 C>.0 C>-
sue. 2 w a, s^<
a: 0 os o tf o
j2 •< j< <: ^
uuc
-------
C. Estimation o£ Adsorption (Sorption) to Soils and
Sediments from the Octanol/Water Partition Coefficient
and Water Solubility'
Since there are very little experimental data on the
adsorption to soils and sediments, estimation techniques were
used to evaluate adsorption (sorption) to these substrates.
Using thermodynamic concepts, Karickhof (1931) showed
that for liquids
log Koc = -a log Xsol + 8 , (8)
where Xsol is the mole fraction water solubility and a and 0 are
constants. However, for crystalline solids at room temperature,
a crystal energy must be added. That is
log Koc * -alog Xsol + "crystal energy term" + g. (9)
Chiou et al. (1979) reported the correlation of KQC
with the water solubility for 15 chlorinated hydrocarbons
including the PCS congeners (2,4'-dichloro),
(2,2' ,5,5'-tetrachloro) and (2 ,2' ,4 ,4' ,5 ,5'-hexachloro) by the
regression equation
log Koc = 4.040 - 0.557 log S , (10)
where S is the solubility in u moles/L and with a correlation
coefficient (r2) of 0.988. Though the correlation is good,
several of the compounds are crystalline at 25°C and a crystal
4-15
-------
energy correction term must be added to equation 10 as described
above.
Karickhoff (1981) clearly showed the improvement in the
correlation for five liquid and solid hydrocarbons [benzene,
naphthalene,.phenanthrene, anthracene, and pyrene] when using
equation 9 in comparison to equation 8. Linear regression
analysis of the KQW and KQC data in equation 8 gave
log Koc = -0.594 log Xsol - 0.917 , (11)
with a correlation coefficient (r^) of 0.945. Linear regression
analysis of the same data in equation 9 containing the crystal
energy term gave
log KQC = -0.921 log Xsol - 0.00953 (mp-25)-1.405, (12)
where mp is the melting point of the compound (in °C) and the
correlation coefficient (r^) was 0.995. Thus, the crystal energy
term significantly improves the correlation.
Octanol/water partitioning provides a much better
estimator for soil or sediment-water partitioning than does water
solubility. This is because in octanol/water partitioning, the
partitioning already involves the distribution of monomers
between the two phases. However, in water solubility, the
formation of a saturated solution involves the equilibration
between monomers in solution with the crystalline solid. As a
4-16
-------
result, crystal energy contributions enter into the formation of
saturated solutions in water but do not affect the distribution
of the species between octanol and water or soil/sediment and
water since the molecules are already present in the monomeric
state.
Karickhoff et al. (1979) carried out adsorption
isotherm experiments at 25aC with 15 aromatic compounds using Doe
Run and Hickory Pond sediments and determined KOC. These
researchers correlated KQC with Kow using linear regression
analysis to give the equation
log Koc = 1.00 log KQW - 0.21 , (13)
with a correlation coefficient (r^) of 1.00.
Using thermodynamic concepts, Karickhoff (1981) showed
that
log KQC * log KQW + A , (14)
where A is a constant. Five compounds [benezene, naphthalene,
phenanthrene, anthracene, and pyrene] were chosen as reference
compounds based on the following criteria:
1. they form a hydrocarbon framework for a large
percentage of aromatic organic compounds and thus
constitute a good reference set from which to
extrapolate to other compounds;
4-17
-------
2. they hydrophobically sorb with little potential for
sorption involving solute dipoles or hydrogen
bonds;
3. kQW and KQC span three orders of magnitude; and
4. there is good agreement for the experimental values
of KQW and KOC published by other research
scientists.
•
Fitting the experimental data for these five reference compounds
to equation 14 gave
log KQC = log KQW - 0.386 , (15)
with a correlation coefficient (r2) of 0.994.
A more general equation correlating KQC with KQW
traditionally has taken the form
log KQC = C log KQW + D , ^ (16)
where C and D are constants. Applying linear correlation
analysis to the experimental data for the five reference
hydrocarbons gave
log KQC = 0.989 log KQW - 0.346 , (17)
with a correlation coefficient (r2) of 0.997.
4-18
-------
Karickhoff listed the published experimental values of
Kow and Koc for a number of pesticides and other organic
compounds including: (a) 17 hydrocarbons and chlorinated hydro-
carbons, including two hexachloro PC3s [the two hexachloro-
biphenyl congeners are (2 ,2',4 ,4',6 ,6'-) and (2 ,2',4 ,4',5,5'-)].
The log KQW values for these two congeners are 6.34 and 6.72,
respectively (Table 2, Chapter 1); the log KQC values for these
two congeners are 6.08 and 5.62, respectively (Section II.3 of
this Chapter)]; (b) six chloro-s-triazines; (c) three
carbamates; (d) four organophoshates; (e) six phenyl ureas; (f)
six miscellaneous compounds including heterocyclics, a ketone,
and a brominated quinone. The predicted values of Koc from
equation 15 were then compared to the experimental values.
Compounds for which solute speciation could be expected were
excluded (e.g., organic bases with pKa greater than 3). The
phenyl ureas were excluded as a class because the predicted KQC
values were an order of magnitude less than the measured
values. This could be due to speciation but Karickhoff
attributed the deviation to poor experimental data. Overall, the
agreement between estimated and measured KQC for the remaining
compounds agreed to within 0.48 log units or within a factor of 3.
In developing a regression equation which has more
widespread applicability to more different hydrocarbons and
chlorinated hydrocarbons, it was decided to use the Kow and KQC
data for all twenty-two hydrocarbons including the five reference
compounds and the four high molar mass chlorinated compounds
(DDT, methoxychlor, and the two hexachlorobiphenyls). Linear
4-19
-------
correlation analysis or> all the data for the se.t N=22 using
equation 15 gave
log Koc = 0.942 log KQW - 0.144 , (18)
with a correlation coefficient (r^) of 0.986. This equation is
now applicable to be a larger set of hydrocarbons and chlorinated
hydrocarbons (22) and KQW and KQC vary by 4.5 and 4.7 orders of
magnitude, respectively, in comparison to equation 17 which
applies to a data set of 5 compounds and both KQW and KQC only
vary by 3 orders of magnitude.
Precise values of Kow and PCBs have been obtained by
the coupled column generator-chromatographic method, Chapter 1.
These values (log KQW) are summarized in Table 1, Section I, for
the various PC3 congeners measured and those estimated for the
various congeners of a given type [e.g., dichloro-, trichloro-,
etc.]. These values were substituted into equation 18 to
estimate log KQC and the results are summarized in Table 1. All
PCS congeners are immobile.
4-20
-------
III. REFERENCES
Chiou CT, Peters LJ, and Freed VH. 1979. A physical
concept of soil-water equilibria for nonionic organic
compounds. Science 206:831.
Griffin R, Clark R, Lee M, and Chian Ee 1978. Disposal and
removal of polychlorinated biphenyls in soil. In: Land
and disposal of hazardous wastes proceedings. 4th annual
research symposium. EPA-600/9-78-016:282.
Haque R, Schmedding D, and Freed VH. 1974. Aqueous
solubility, absorption, and vapor behavior of
polychlorinated biphenyl Aroclor 1254. Environ Sci Tech
8:139.
tiaque R. and Schmedding D. 1976. Studies on the
adsorption of selected polychlorinated biphenyl isomers on
several surfaces. J Environ Sci Health 811:129.
Karickhoff S, Brown DS, and Scott TA. 1979. Sorption of
hydrophobia pollutants on natural sediments. Water Res
1-3:241.
Karickhoff SW. 1981. Semi-empirical estimation of
hydrophobic pollutants on natural sediments and soils.
Chemosphere 10:833.
Lyman WJ, Reehl WF, Rosenblatt DH. 1982. Handbook of
chemical property estimation methods. McGraw tiill Book
Co., N.Y.
Paris, DF, Steen WC, and Baughman GL. 1978. Role of
physico-chemical properties of Aroclors 1016 and 1242 in
determining their fate and transport in aquatic
environments. Chemosphere 4:319.
U.S. Environmental Protection Agency. 1982. Chemical fate
test guidelines. EPA 560/6-82-003. NTIS publication
PB 82-233008. Test Guideline CG-1700.
4-21
-------
-------
CHAPTER 5
3IOCONCSNTRATION OF POLYCHLORINATSD BIPHENYLS IN FISH
by
Asa Leifer
Contents
Page No.
I. INTRODUCTION AND SUMMARY 5-1
II. DISCUSSION OF RESULTS 5-5
A. Background 5-5
B. Experimental Data on Bioconcentration
and Ecological Magnification of
Polychlorinated Biphenyls 5-8
C. Estimation of Bioconcentration from the
Octanol/Water Partition Coefficient 5-13
III. REFERENCES 5-19
5-i
-------
I. INTRODUCTION AND SUMMARY
One very important aspect of the environmental fate of
PC3s is the prediction of the extent to which these chemicals
will achieve concentrations in fish which may be several orders
of magnitude greater than the concentration of PCBs in the water
phase. A convenient equilibrium constant which describes this
process is the bioconcentration constant, Kg, defined as the
ratio of the concentration of PCS in fish to the concentration in
the water.
There is only a small amount of published data on the
experimental determintion of log Kg of PCBs in fish. What
experimental data are available is discussed in Section II.B. and
all these results are summarized in Table 2 for the Aroclors and
Table 3 for individual PCS congeners.
Since there is only a small amount of data available on
the bioconcentration of most of the PCB congeners in fish,
estimation techniques were used to estimate log Kg. Veith et al.
(1981) developed a regression equation correlating log KQ with
log KQW for 122 different chemicals and 13 different species of
fresh and marine water fish. This equation is
log KB = 0.79 log KQW - 0.40 , (1)
and had a correlation coefficient (r2) of 0.86. In a recent
publication by Mackay (1982), the data of Veith et al. (1979)
were carefully analyzed, suspect data were eliminated, and the
regression equation
-------
log KB = log Kow - 1.32 (2)
was developed with a correlation coefficient of 0.95 for
approximately fifty chemicals. One of the major differences
between equations 1 and 2 is that the slope of equation one is
0.79 while the slope in equation two is 1.0. Furthermore,
equation two is only valid for log Kow less than 6.0. Clearly,
there is a need for precise and reliable Kow and Kg data in the
region log KQW > 6.0 to test equation two.
At present, the best correlation is given by equation one,
since it is based on the most extensive data set (122 chemicals)
for 13 different fresh or marine water fish. Thus, this equation
was chosen to estimate log Kg for the various PC3 congeners and
the results are listed in Table 4, Section II.C. Whenever
precise experimental log KQW data were available for specific PCS
congeners (see Chapter 1), these values were used to estimate log
Kg. In addition, log KQW was estimated for all other congeners
of a given type [i.e., monochlorobiphenyls, dichlorobiphenyls,
trichlorobiphenyls, etc.] using the method outlined in
Chapter 1. These values were then used to estimate log Kg. All
these data are given in Table 4 and are summarized in Table 1.
For comparison purposes, the average experimental values of log
Kg for each group of congeners, for all aquatic species (obtained
from the data listed in Table 3), are listed in Table 1.
Comparison of both sets of data for each group of congeners
indicates that the agreement, in general, is very good and
5-2
-------
further substantiates equation one and the data,for the PCBs
presented in Table 4. It should be noted that there is an
experimental spread in the data (Table 3} and thus the data are
probably good to ± 0.5 log units or within a factor of 3.
All the data listed in Tables 1 to 4 indicate that PCBs
have the potential to bioconcentrate to a large extent in fish.
However, one must consider transformation of PCBs before making a
final assessment of bioconcentration potential. For the higher
chlorinated species, where transformation is most likely slow,
these estimated values of log KB should be reasonably reliable.
5-3
-------
Taole 1. Comparison of Estimated Log KB Values of Groups of PCS Congeners
(from Estimated Values of Log KOW and Equation 1} with the
Average Experimental Values.
PC3 Congeners
Monochloro-
Dichloro-
Trichloro-
Tetrachloro-
Pentachloro-
Hexachloro-
Heptachloro
Octachloro-
Nonochloro-
Decachloro-
Estimated
Log KOW
4.51
4.94
5.37
5.80
6.24
6.67
7.10
7.53
7.96
8.26
Log
Estimated
3.16
3.50
3.84
4.18
4.53
4.87
5.21
5.55
5.89
6.13
KB
Experimental
3.26
3.61
3.56
3.95
4.14
5.27
a
a
a
5.10
a. No experimental data was found in the literature.
5-4
-------
II. DISCUSSION OF RESULTS
A. 3ackaround
One very important aspect of the environmental fate of
a chemical is the ability of a chemical to partition from the
water phase into a hydrophobic phase. For example, consider the
environmental scenario in which a fish is in a water body
containing a dissolved organic chemical. If the chemical is
hydrophobic, it will have a tendency to partition out of the watei
phase and into the fatty tissue of the fish (i.e., the
hydrophobic phase). Thus, over a period of time, the hydrophobic
chemical will bioconcentrate in the fish. A convenient
equilibrium constant which describes this environmental process
is the bioconcentration constant, Kg, defined by the relationship
KB = CB/CM , (4)
where C3 is the concentration of the chemical in the fish and CM
is the concentration of the chemical in water, at equilibrium.
Since Cg and C^ are usually expressed in the same concentration
units, KQ is a dimensionless constant. Hydrophobic chemicals
have high values of Kg and will thus bioconcentrate in fish.
Since Kg can span several orders of magnitude, a convenient
expression for bioconcentration is log Kg.
The bioconcentration process is often viewed as a
balance between two kinetic processes, uptake and depuration
[Veith et al. (1979), Mackay (1982)]. If one defines the uptake
5-5
-------
• and depuration as first-order processes with rate*contants K^ and
K9/ respectively, then the increase in concentration of a
chemical in the fish (C3) with time (t) is given by the
differential equation
(dCB/dt) = K1C_M - K2CB , '(5)
where C^ is the concentration of the chemical in water.
Integrating equation 5 for the boundary conditons t=0, CB=0 and
t, CB, yields
CB = (K1CM/K2) [1 - exp(-k2t)] . (6)
If KB is defined by the ratio of the uptake and depuration rate
constants
KB = K1/K2 , (7)
then equation 6 becomes
CB = KBCM (1 - exp (-K2t)] . (8)
After a long time (i.e., as t •»•« ), which corresponds to.
equilibrium conditions, equation 8 yields
C= IT c* i Q ^
B ^S^M ' ^y'
which is the expression for bioconcentration defined in equation
5-6
-------
4. Values of Kg can be obtained from long term exposure
measurements to give C« or by kinetic measurements. It must be
emphasized that this approach is only valid for chemicals which
are not degraded to any appreciable extent in the aquatic
environment by chemical or biological processes. This is true
for the higher chlorinated biphenyls.
Mackay (1982) proposed an alternative approach to
bioconcentration by viewing the fish as an inanimate volume of
material that approaches thermodynamic equilibriun with the water
as defined by the chemical potential or fugacity of the biocon-
centrating-persistent solute. In this case, the differential
equation (eq. 5) can be reformulated to give a first-order
process in which Ca approaches KgCM after long exposure (*«•)/
but has the attractive feature that KQ can be related to
physical-chemical properties. Support for this approach has been
demonstrated by the fact that K« is well correlated with the
octanol/water partition coefficient (KQW). Traditionally,
correlations of KB and KQW have taken the form
log KB =« n log KQW + b , (10)
where n and b are constants evaluated from a large body of
experimental data on Kg and K^ for a wide variety of
chemicals. [Veith et al. (1979, 1981) and Mackay (1982)].
5-7
-------
B. Experimental Data on Bioconcentration and Ecological
Magnification of Polychlorinated Biphenyls
Hansen (1975) reported the uptake of two Aroclors, 1016
and 1254, in oysters, shrimp, and fish in a laboratory estuarine
environment. The principal PCS components in these two
commercial products were: (57 percent trichloro, 20 percent
dichloro, and 21 percent tetrachloro) and (48 percent
pentachloro, 21 percent tetrachloro, and 23 percent hexachloro),
respectively. Figure C of the Introduction gives the complete
composition of these Aroclors. Considerable bioconcentration was
observed for these two Aroclors and the log Kg is given in
Table 2. Bioconcentration factors were estimated from Escambria-
Bay data for Aroclor 1254 for oysters, shrimp, and fish,
Table 2. Again, the data indicated considerable bioconcentration
in these estuarine species. A comparison of the field data with
the laboratory data indicated that the laboratory experiments
underestimated bioconcentration. More experimental data is
needed to verify these results.
Veith et al. (1979) reported laboratory biocon-
centration data in fathead minnows in fresh water for several
Aroclors [1016, 1254, 1248 and 1260], Table 2. The principal PCS
components of the first two Aroclors have been given above and
the principal PCS components of the last two Aroclors are (40
percent tetrachloro, 18 percent trichloro, and 36 percent
pentachloro) and (41 percent heptachloro, 38 percent hexachloro,
and 12 percent pentachloro), respectively. Figure C of the
Introduction gives the complete composition of these two
5-8
-------
Table 2. Experimental Bioconcentration Factors for Several
Aroclors in Aquatic Species
ArocLor Lab. Field Aquatic Species Reference
1254 5.00 - 5.02 >5a Oyster Hansen (1975)
1254 4.42 5.36a Shrimp Hansen (1975)
1254 4.51 5.82a Fish Hansen (1975)
1016 4.53 - Fish Hansen (1975)
1016 4.63 Fathead Minnow Veith et al. (1979)
1048 4.85 Fathead Minnow Veith et al. (1979)
1254 5.00 Fathead Minnow Veith et al. (1979)
1260 5.28 Fathead Minnow Veith et al. (1979)
1254 4.60 Brook Trout Veith et al. (1981)
1254 4.43 Spot Veith et al. (1981)
aCalculated from Escambria Bay data.
5-9
-------
Aroclors. Since the log kQW values reported in Table 2 are
greater than 4.4, bioconcentration is considerable in these fresh
water species.
Veith et al. (1981) reported laboratory biocon-
centration data for Aroclor 1254 in the fresh water species Brook
Trout and Spot, Table 2. Since the log kQw values reported in
Table 2 are greater than 4.4, bioconcentration was substantial.
Sugiura et al. (1978) measured the uptake of several
PCBs [(4-chloro), (2,2'-dichloro), (4,4'-dichloro), (2,2',3,3'-
tetrachloro), (2,3,5,6,-tetrachloro), (3,3',4,4'-tetrachloro) and
(3,3'/S/S'-tetrachloro)] in killifish (pryzias latipes) and
calculated K« from the data. However, these researchers used
Tween 20* at 1 or 100 ppm to prepare solutions of PCBs. They
claimed that Tween 20R had no effect on the uptake of these PCBs
which is quite surprising since the Tween concentrations we're
considerably greater than the concentration of the PCBs.
Sanborn et al. (1975) measured the uptake of -three
chlorinated biphenyls [(2,2',5-trichloro-), (2,2',5,5'-
tetrachloro-, and (2,2'/4,5,5'-pentachloro-)] in green sunfish
[Leponis cyanellus Raf] in tap water in the laboratory and found
substantial bioconcentration for all three congeners. Table 3
lists the log KB for these three PCBs.
An extensive literature review has been recently made
by Nabholz (1983) on the determination of bioconcentration or
ecological magnification (EM) in the laboratory for various PC3
congeners and the results are summarized in Table 3 together with
the literature references. It should be noted that EM is
C .1
-------
Table 3. Experimental Bioconcentration Factors for Several
PCB Congeners in Aquatic Species
PCS Congener
log
Species
Reference
Monochloro-
2-
3-
4-
Dichloro-
2,3-
2,5-
Trichloro-
2,2', 5-
2,4', 5-
2',3,4-
Tetrachloro-
2,2', 3,5'-
2, 2', 4,4'-
2,2', 5,5'-
2,3,4,5
3.30
3.02
3.72
2.95
3.60
3.26
2.98
3.08
4.14
1.73
3.76*
*
• 2.91
4.30
3.81*
4.63
3.79
4.04
3.98^
4.60*
4.02*
4.69^
4.07*
3.87
2.66
4.29
3.57
3.94
3.90
unknown fish sp.
unknown fish sp.
Golden Orfe
Carp.
Brown Trout
Guppy
unknown species
Oyster
Goldfish
Green Sunfish
Snail
Mosquito Larvae
Goldfish
Gambusia (fish)
Goldfish
Oyster
Oyster
Rainbow Trout
Snail
Mosquito Larvae
Goldfish
Gambusia ( fish)
Oyster
Green Sunfish
Golden Orfe
Carp
Brown Trout
Guppy
Moolenar (1982)
Moolenar (1982)
Sugiura et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Moolenar (1982)
Vreeland (1974)
Bruggeman et al. (1981)
Sanborn et al. (1975)
Metcalf et al. (1975)
Metcalf et al. (1975)
Bruggeman et al. (1981)
Metcalf et al. (1975)
Bruggeman et al. (1975)
Vreeland (1974)
Vreeland (1974)
Branson et al. (1975)
Metcalf et al. (1975)
Metcalf et al. (1975)
Bruggeman et al. (1981)
Metcalf et al. (1975)
Vreeland (1974)
Sanhorn et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
5-11
-------
Table 3. Experimental Bioconcentration Factors for Several
PCS Congeners in Aquatic Species (Continued)
?C3 Congener
log Kg
Species
Reference
2,3',4',5-
3, 3', 4,4'-
4.62
3.85
3.90
3.24
3.63
4.15
Goldfish
Rainbow Trout
Golden Orfe
Carp
Brown Trout
Guppy
Bruggeman et al. (1979)
Stalling et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Sugiura et al. (1979)
Pentachloro-
2, 2', 3,4,5'-
2,2', 4,5,5'-
Hexachloro-
2, 2', 4,4', 5,5'-
Decachloro
2,2(,3,3'f4,4'
5, 5', 6', 6'
4.43
4.78*
4.24*
•*
4.09
3.18
5.00
5.02
4.62
4.68
5.67
6.03
5.20
5.93
5.97
4.34
4.99
Oyster
Snail
Mosquito Larvae
Gambusia ( fish)
Green Sunfish
Snail
Mosquito Larvae
Gambusia (fish)
Oyster
Benthic Amphipod
Benthic Amphipod
Benthic Amphipod
Benthic Amphipod
Snail
Mosquito Larvae
Gambusia ( fish)
vreeland (1974)
Metcalf et al. (1975)
Metcalf et al. (1975)
Metcalf et al. (1975)
Sawfaorn et al. (1975)
NRC (1979)
NRG (1979)
NRC (1979)
Vreeland (1974)
Lynch and Johnson (1982)
Lynch. and Johnson (1892)
Lynch and Johnson (1982)
Lynch and Johnson (1982)
NRC (1979)
NRC (1979)
NRC (1979)
*log (Ecological Magnification) or log (EM).
5-12
-------
determined from laboratory model ecosystems and is defined as the
increase in concentration of the PCB over the concentration in
the water and the uptake is bioconcentration and biomagni-
fication. These EM data are starred in Table 3. The average
values of the log bioconcentration for the different species
within a class of congeners is summarized in Table 1. Inspection
of the data in Table 3 indicates that there is a spread in the
results and thus these data are probably good to _+0.5 log units
or within a factor of 3. The spread in' these data can be
attributed to: (1) both EM and bioconcentration are reported and
EM data involves both bioconcentration and biomagnification and
(2) the inherent difficulties in measuring precise values of
bioconcentration or ecological magnification.- Inspection of all
the data indicates that bioconcentration in all cases is
substantial.
C. Estimation of Bioconcentration from the Octanol/Water
Partition Coefficient
Since there is only a small amount of experimental data
on the bioconcentration of individual PCB congeners in fish,
estimation techniques were used to estimate the bioconcentration
factor Kg from the octanol/water partition coefficient. In
Section IIA., it has been shown that KB is related to KQW
(equation 10). Veith et al. (1979) determined log Kg for 30
different chemicals in fathead minnows (Pimephales promelas) and
found a relationship between log Kg and log KQW. This
relationship is given by the equation
5-13
-------
log KB = 0.85 log KQW - 0.70 ,(11)
where the constants in equation 10 were found to be n=0.85 and b=
-0.70. The correlation coefficient (r^) for the linear
regression analysis was 0.90.
Veith et al. (1981) expanded the work to include 122
different chemicals with 13 different species of fresh and marine
water fish. Linear correlation analysis of these data gave the
equation
log KB = 0.79 log KQW - 0.40 , (12)
with a correlation coefficient (r^) of 0.86. An important
conclusion of the evaluation of the relationship of Kg to Kow is
that the prediction limits of the regression did not change
^
significantly (i.e., r only changed from 0.90 to 0.86) by
quadrupling the number of test chemicals and increasing the
number of species from one to thirteen.
In a recent publication by Mackay (1982), the physical-
chemical factors influencing bioconcentration were studied in
detail. Using purely thermodynamic concepts, Mackay showed that
log KB = log Kow + log A , (13)
5-14
-------
where A is a constant. In this case the slope n is equal to
one. Using equation 13 and the data by Vieth et al. (1979),
Mackay eliminated suspect chemicals and found that
log KB = log Kow - 1.32 , (14)
with a correlation coefficient (r2) of 0.95. Mackay concluded
that this correlation will satisfactorily predict KB using one
constant and that indeed n=l.
Since the reliability of KB and Kow measurements is
often suspect, it is apparent that it is difficult to determine n
precisely and reliably. Mackay indicated log Kow values in
excess of 6.0 may be suspect since many were measured using HPLC
technique and extrapolated outs-ide the range of values measured
under equilibrium conditions. However, the HPLC technique is one
of the best and most precise methods for estimating high log KQW
values Cor hydrophobia compounds. This method measures the
retention time of a chemical on a hydrophobic column to estimate
the value of log Kow. Given the nature of the experimental
procedure for measuring KB, it is apparent that the KB data is
far less precise. Clearly, there is a need for precise and
reliable KQW and KB data in this region to test the linearity
hypothesis proposed by Mackay.
At present, the most extensive set of data and the most
reliable correlation equation is the one proposed by Veith et al.
(1981), given by equation'12, and this one was chosen to estimate
log K3 from KQW data.
5-15
-------
Very precise values of KQW for PCBs have been obtained
*
from the coupled column generator-chromatographic method,
Chapter 1. These values are summarized in Table 4 for the
various PCS congeners measured and for those estimated for all
the other congeners of a given type [i.e., monochloro, dichloro,
trichloro, etc.]. These values were•substituted in equation 12
to estimate log KB and the results are summarized in Table 4.
The log KB values for all the congeners of a given type
(i.e., monochloro, dichloro, trichloro, etc.) are given in Table
4 and have been summarized in Table 1, Section I, for convenience
of presentation of the data. In addition, the average
experimental values of log KB for each group of congeners, for
all aquatic species (obtained from Table 3) , are also listed in
Table 1. Comparison of both sets of data for each group of
congeners indicates that the agreement, in general, is very good
and further substantiates equation 12 and the data for the PCBs
presented in Table 4. It should be noted that there is a spread
in the experimental data (Table 3) and thus the data are probably
good to ± 0.5 log units or within a factor of 3.
All the data indicates that PCBs have the potential to
bioconcentrate to a large extent in fish. However, one must
consider the transformation of -PCBs before making a final
assessment of bioconcentration potential. For the higher
chlorinated species, where transformation is most likely very
slow, these estimated values should be reasonably reliable.
5-16
-------
Table 4. Log K3 for the PCB Congeners from Estimated and
Experimental Values of Log K Using Equation 12.
PCS Congener
Monochloro-
2-
3-
4-
Dichloro-
2,2'-
2,4'-
2,5-
2,6-
3,4-
4,4'-
Trichloro-
2, 2', 5-
2, 4, 5-
2,4',5-
2,4,6-
Tetrachloro-
2,2'3,3'-
2,2', 4', 5-
2,3,4,5-
2,3,5,6-
Pentachloro-
2, 2', 4,5, 5'-
2,3,4,5,6-
Hexachloro-
2,2', 3,3', 4,4'-
2,2', 3,3', 6,6'-
2,2',4,4',6,6'-
log KOW
Est. Exp.
4.51
4.44a
4.53
4.49
4.94
4.90
5.14
5.16
4.93
5.29
5.33
5.37
5.60
5.66a
5.79
5.47
5.30
4.63b
5.73
5.72
5.46b
6.24
5.92
6.30
6.67
6.98
6.55
7.55
6.34C
Log KB
3.16
3.11
3.22
3.15
3.50
3.47
3.66
3.68
3.50
3.78
3.81
3.34
4.02
4,07
4.17
3.92
4.18
3.26
4.13
4.12
3.91
4.53
4.28
4.58
4.87
5.11
4.85
5.57
4.61
5-17
-------
Table 4. Log Kg for the PC3 Congeners from Estimated and
Experimental Values of Log KQW Using Equation 12 (Continued)
log KOW
?C3 Congener Sst. Exp.
Heotachloro- 7.10
2,2', 3, 3', 4,4', 6- 6.68
Octachloro- 7.53
2,2I3,3I,5,5',6,6I- 7.11
Nonachloro- 7. 96
2, 2', 3, 3', 4, 5, 5', 6,6', 8.16
Decachloro-
2,2',3>3',4,4'/5,5<,6,6'- 8.26
Log KB
5.21
4. 88
5.55
5.22
5.89
6.05
6.13
Almost all experimental data were obtained from the coupled column
generator—chmomatographic method, Chapter 1.
a. These values represent the average of the experimental
results from NBS and Woodburn.
b. These values were measured by the reverse-phase liquid
chromatographic method.
c. This value was measured by the conventional shake-flask
method.
All estimated by log K values for the PCS congeners were obtained by the
method discussed in Chapter 1.
5-18
-------
III. REFERENCES
Branson DR, Blau GE, Alexander HC, Neeley WB. 1975.
Bioconcentration of 2,2',4,4'-tetrachlorobiphenyl in rainbow
trout as measured by an accelerated test. Amer Fish Soc
104:785-792.
Bruggeman WA, Martron LBJM, Kooiman D, Hutzinger 0. 1981.
Accumulation and elimination kinetics of di-, tri-, and
tetrachlorobiphenyls by goldfish after dietary and aqueous
exposure. Chemosphere 10:811-832.
Hansen D. 1975. PCBs: Effects on accumulation by
estuarine organisms. National Conference on Polychlorinated
Biphenyls. Chicago, Illinois. pp. 282-283. EPA-560/6-75-
004.
Mackay D. 1982. Correlation of bioconcentration factors.
Env Sci Technol 16:274.
Metcalf RL, Sanborn JR, Lu P-Y, Nye D. 1975. Laboratory
model ecosystem studies of the degradation and fate of
radiolabled tri-, tetra-, and pentachlorobiphenyl compared
with DDE. Arch Environ Contam Toxic 3:151-165.
Moolenar RJ. 1982. Environmental behavior of MCBs.
Unpublished Manuscript. The Dow Chemical Co., Midland, MI.
Nabholz JV. 1983. Bioconcentration factors for selected
polychlorinated biphenyl isomers in aquatic organisms.
Memorandum to Irwin Baumel, Director, Health and
Environmental Review Division/Office of Toxic
Substances/0.S. Environmental Protection Agency, Washington,
D.C.
National Research Council. 1979. Polychlorinated
biphenyls. Washington, D.C. National Academy of Sciences.
Sanborn JR, Childers WF, and Metcalf RL. 1975. Uptake of
three polychlorinated biphenyls, DDT, and DDE by the green
sunfish, Leponis cyanellus RAF. National Conference on
Polychlorinated Biphenyls.cFIcago, Illinois. PP. 236-
242. EPA-560/6-75-004.
Sugiura K, Ito N, Matsumoto N, Mihara Y, and Goto M.
1978. Chemosphere 7:731.
Sugiura K, Washino T, Haltori M, Sato E, Goto M. 1979.
Accumulation of organochlorine compounds in fishes.
Difference of accumulation factors by fishes. Chemosphere
6:359-364.
5-19
-------
Veith GD, DeFoe, DL, and Bergsdedt BV. 1979. Measuring and
estimating bioconcentration factor of chemicals in fish. J
Fish Res Board Can 36:1040.
Vei'th GD. 1981. State-of-the-art report on structure-
activity methods development (II). Draft EPA report.
Vreeland V. 1974. Uptake of chlorobiphenyls by oysters.
Environ Poll 6:135-140.
5-20
-------
CHAPTER 6
ATMOSPHERIC OXIDATION OF POLYCHLORINATED BIPHENYLS
by
Asa Leifer
Contents
Page No.
I. INTRODUCTION AMD SUMMARY 6-1
II. OXIDATIOM OF POLYCHLORINATED BIPHENYLS
3Y HYDROXYL RADICALS 6-8
III. STRUCTURE-REACTIVITY RELATIONSHIPS FOR
HYDROXYL RADICAL REACTION WITH
POLYCHLORINATED BIPHENYLS 6-9
IV. CALCULATION OF THE SECOND-ORDER RATE
CONSTANTS (fcQH) FOR POLYCHLORINATED BIPHENYLS
USING STRUCTURE-REACTIVITY RELATIONSHIPS 6-13
A. 2-Chlorobiphenyl 6-14
1. OH Addition to the Rings..... 6-14
a. Ring (A).... .. . 6-14
b. Ring (B) 6-14
c. Total kadd< 6-14
a com.
2. H Abstraction on the Rings 6-15
a. Ring (A) 6-1*
b. Ring (B) 6-16
c. Total kabs. • 6-16
3. Total kOH 6-16
B. 2,2-Dichlorohiphenyl 6-17
1. OH Addition to the Rings 6-17
a. Rings (A) and (B) 6-17
b. Total kj«;> 6-17
2. H Abstraction on the Rings 6-13
a. Rings (A) and (3) 6-13
b. Total ' 6-13
3. Total kOH 6-18
-------
-> . 4
1 .
9 .
3.
c . Total kadd '
' ioc
-------
I. INTRODUCTION AND SUMMARY
If poly chlorinated biphenyls (PCBs) are transported into the
atmosphere (i.e., the troposphere), then there are two principal
"lodes of oxidation that can occur. ?C3s may be transformed by
reaction with the oxidants hydroxyl radicals or ozone. However,
because of the chemical structure of PCBs, the dominant oxidative
reaction occurs with hydroxyl radicals (OH). Consequently, this
chapter is devoted to the determination of the rate of OH
reaction with PCBs.
There are no experimental data published in the literature
on the rates of transformation of PCBs in the atmosphere with OH
radicals. Cupitt (1980), using structure-reactivity
relationships developed by Hendry and Kenley (1979), estimated
the secoad order rate constant ('
-------
reactivity framework involves the calculation of the two major
pathways for the reaction of OH with ?CBs: H abstraction (
and addition to the double bonds in the aromatic rings (ka ' )
3 arom.
The second-order rate constant k is then the sum of
and '
-------
Table 1. Second-Order Rate Constants (!
-------
Table 1. Continued
Structure
1012kQH(cm3 molec'1 sec'1 )
todays)
O
2.1
3. 8
7.
7.0
11
Cl.
0.41
20
9.
2.1
3.8
Cl,
10,
0.65
12
Cl,
11,
0.26
31
6-4
-------
Table 1. Continued
btructure
12.
1012kou(cm3 raolec'1 sec'1)
0.64
todays
13
13.
0.22
36
14.
0.13
62
15.
0.21
38
Cl-
1*. /7'\v//i\N
0.085
94
17.
0.069
120
6-5
-------
Table 1. Continued
Structure
13.
"c"1 )
0.039
210
19.
0.024
330
20.
0.0049
1700
a. The half-life was calculated based on an average value of
[OH] = 1 x 10° molecules era"3 for the global concentration of
OH radicals produced by sunlight in reasonably polluted air.
6-6
-------
(4) For unsymmetrical chlorinated biphenyls with all the
chlorines on one ring, kOH is dominated by the rate of
addition of OH to the double bonds of the unsubstituted
ring. Therefore, kOH and tl/ are approximately the same
regardless of the number of chlorines on the one ring.
(5) For a given group of chlorinated congeners (e.g., the
tetrachlorobiphenyls) the totally unsymmetrical
chlorinated congener has the largest RQ^J and the
smallest CL/. As the number of chlorines becomes more
symmetrically distributed between the two rings, kQH
decreases and OA increases. For the congeners with the
most symmetrical distribution of chlorines on the
rings, KQ^ has the lowest value and tl/ has the largest
value (and therefore these congeners are the most
persistent).
(5) For symmetrically distributed chlorinated congeners
with both rings containing high numbers of chlorines
(e.g., Clx, Cly, with x + y = 7, 8, 9, 10), as the
total number of chlorines increase, kOH decreases
rapidly and tl/ increases rapidly. For example, the
nonachlorobiphenyl has a half-life of 330 days while
the decachlorobiphenyl has a half-life of 1700 days.
Hence, these compounds are reasonably persistent.
For a large number of the PCB congeners especially those
containing a small number of chlorines and those with all or most
of the chlorines on one ring, atmospheric transformation occurs
6-7
-------
at a reasonable rate by reaction with OH radicals. These results
are based on the structure-reactivity relationships developed by
Hendry and Kenley (1979) and Mill et al. (1982). It should be
emphasized that these results are tentative and must be verified
experimentally. Experiments should be carried out to measure kOH
by flash photolysis and flow through techniques on selected PCBs
at room temperature and elevated temperatures. The OH radicals
should be produced by the flash photolysis of HONO or ^0 in the
presence of H20 or H2. The OH decay should be monitored by
fluorescence spectroscopy or resonance absorption.
II. OXIDATION OF PQLYCHLORINATED BIPHENYLS
BY HYDROXYL RADICALS
The reaction of FCBs with OH can be treated mathematically
in the following manner:
PCB + OH + Products (1)
= kQH[OH] [PCB] , (2)
where KQJJ is the second-order rate constant in cm molecule"-'-
sec.""-'- and [OH] and [PCB] are the concentrations of hydroxyl
radicals and polychlorinated biphenyls, respectively. Since very
low concentrations of PCBs exist at any given time in the
atmosphere and a steady-state concentration of OH radicals is
6-3
-------
produced by sunlightj in polluted air, the hydroxyl radical
concentration can be treated as a' constant and equation 2 becomes
a pseudo first-order rate equation.
- d[p,°.B] = k[PCB] , (3)
dt
where k = knu[OH] . (4)
Uti
The pseudo first-order rate constant k is in the units of
reciprocal time (usually in seconds). Since equation 3 is a
pseudo first-order rate equation, the half-life (i.e., the time
to reduce the initial concentration of PCBs by one half) is
0.693 0.693 (5)
V2 R kQH[OK] *
III. STRUCTURE-REACTIVITY RELATIONSHIPS FOR
HYDROXYL RADICAL REACTION WITH
POLYCHLORINATED BIPHENYLS
Hendry and Kenley (1979) developed a structure-reactivity
method for estimating values of the second-order rate constant,
kQ9/ for the reaction of OH radicals with an organic molecule.
There are three major reaction pathways in the gas phase: (1) H
atom abstraction; (2) addition to olefinic bonds; and (3) addition
to aromatic rings. Since the structure of PCBs is
6-9
-------
(6)
where x and/or y is equal to 0 to 5, only H abstraction and
addition to aromatic rings need to be considered. Each of these
reaction pathways has an intrinsic reactivity constant for each
reaction center, k_h_ and k*~. ' . These reactivity constants
owo • 3 JTOITl•
are modified by substituents at the reaction center (a position)
and adjacent to the reaction center (3 position) and these
substituent constants are denoted by a and 8. Thus, the general
expresssion for the second-order rate constant, kQH' ^-n terms of
the two reactivity constants is
abs.
arom.
(7)
The rate of hydrogen atom abstraction is affected by
substitution on the sane and adjacent functional groups. The
total reactivity rate constant for hydrogen abstraction, ^abs.'
may be expressed as the summation of the rate constants for each
reactive hydrogen according to equation
'abs.
(8)
6-10
-------
where ' Hendry and Kenley (1979) developed
cl D C
the values of kH for various hydrogens on different functional
groups and the values of a and 3 for various substituents. The
term n^_ represents the number of times the same type of hydrogen
with the same a and 3 substituents appear in the molecule. The
development of the values of these constants was based on a
detailed study of an extensive list of published rate constants
6-11
-------
for each kind of reaction or composite reactions which were
dissected into the contributory constants for each pathway
[Table 4, Hendry and Kenley (1979)].
The rate of addition of OH to aromatic rings is given by the
equation
fl •
arom. ~ ^—aAl Al '
where kAl is the reactivity of the ltn aromatic ring towards OH
and depends on the degree of substitution on the ring (e.g.,
alkyl, methoxy, or aldehyde groups). The term o, is a factor
which takes into account the effect of halogen atoms substituted
on .the ring and a., represents the product of a, for each ltn
halogen atom on the ring. Hendry and Kenley (1979) developed
values for kA and a, based on a detailed analysis of published
" A
rate constants for aromatic compounds and these results are
summarized in Table 6 of the Hendry and Kenley report.
Further work was carried out by SRI to further validate the
method of Hendry and Kenley [Mill et al. (1982)]. Detailed
kinetic studies were carried out for the model compounds
2-chlorobutane, 2,3-dichlorobutane, 2-chloropropene,
3-chloropropene, chlorobenzene, and p-dichlorobenzene and the
precisely measured second-order rate constants C<
-------
con-pounds had to be adjusted to obtain a better fit between the
experimental and the estimated values from the structure-
reactivity method. SRI found that k.^ = 2.0 x 10~12 cm3 molecule'1
sec."1 for an aromatic ring (Hendry and Kenley listed a value of
1.4 x 10~12 cm3 molecule'1 sec.'1) and acl = 0.30 (Hendry and
Kenley listed a value for acl < 1.0). All the updated values for
tne reactivity constants for hydrogen abstraction, OH addition to
aromatic rings, and OH addition to olefinic double bonds are
summarized in the report by Mill et al. (1982), Tables 9-12, of
the Section on Oxidation in Air. Detailed calculations are given
for each of the model compounds to illustrate the application of
the Hendry and Kenley method of estimating !
-------
A. 2-Chlorobipheny1
The structure of 2-chlorobipheny1 is:
C/-O
A a
where A and B designate each rinq.
1. OH Addition to the Rings
a. Ri ng A
Since ring A contains one chlorine, a-, = 0.30;
k^ = 2.0 x 10~12 crn-3 raolec.~^ sec" . Using these results in
equation 9 yields
= 0-30(2.0 x 10~12) = 0.60 x 10~12
a
b. Ri ng B
Since ring B does not contain a chlorine
2.0 x 10~12
add. m kadd. .. kad|,
arora. arSm. arom.
6-14
-------
kadd< = 0.60 x 10"12 * 2.0 x 10~12
arora.
;
-------
X, = 0.034 x 10~12 cm3 molec."1 secT1 (11)
aos.
b. Ri ng B
There are five hydrogens on the ring with no a substituents,
no halogen substitution, and only g hydrogens. Therefore, n = 5;
3„ = 1; kH = 0.01 x 10~12 cm3 molec."1 sec."1 .
k' - 5(1)2(0.01 x 10"12) '
aos •
* = 0.05 x 10"12 cm3 molec.'1 sec."1. (12)
aos.
c. Total kabs<
k . = k'' + k", = 0.034 x 10~12 + 0.050 x 10"12
abs. abs. abs.
. = 0.084-x 10~12 cm3 molec."1 secT1 (13)
aos.
3. Total
v „ = k K * kadd>
OH abs. arom,
Using equations 10 and 13 yields
kQH = 0.084 x 10~12 + 2.60 x 10~12
k_u = 2.68 x 10"12 cm3 molec."1 sec.'1 . (14)
Un
6-16
-------
X
3. 2,2-Dichlorobiphenyl
:he structure of 2,2-dichlorobiphenyl is
Cl
where A and B "designate each ring,
1. OH Addition to the Rinqs
a. Rings A and B
Since rings A and B are identical and the same as the ring
in Section IV.A.I.a, the same
-ads"
!< Til is obtained. Therefore,
arom.
0.60 x 10~12
b. Total *aronu
karom. = 1*20 * 10~12 cm3 molec-"1 sec."1 (15)
6-17.
-------
2. H Abstraction on the Rings
a. Rings A and B
Since rings A and B are identical and the same as the ring
in Section IV.A.2.a., the same k,w_ is obtained (equation 11).
Therefore
k' = 0.034 x 10~12
abs.
b. Total kabs<
•
-------
C. 2, 4-Dichlorobipheny1
The structure of 2,4-dichlorobipheny1 is
C\
1. OH Addition to the Rings
a . Ri ng A
Since ring A contains two chlorines, a-, = 0.30;
}CA = 2.0 x 10~12 cm3 molec.'1 sec.'1. Using these results in
equation 9 yields
(0.30)2(2.0 x 10~12:
=
a
- 0.18 x 10~12 cm3 molec.'1 sec.'1
b. Ri ng B
Since ring B does not contain a chlorine
ifoii. = 2-
° x
cm3 molec."1 sec.'1
6-19
-------
c. Total kadd<
arom.
-arom. -arom. "— °'18 * 10~"12 + 2'° x 10"12
kadd. = 2<18 x 10-12 Cm3 molec.-l sec.'1 (18)
arom.
2. H Abstraction on the Rings
a. Ri ng A
Ring A has the structure
(i) Ha; n = 1; there are two 3 chlorines and no
a substitution; 3C1 = 0.4; kH = 0.01 x 10"12,
Using these results in equation 8 yields
= 0.0016 x 10~12 cm3 raolec."1 sec.'1 »
(ii) HV.J; n = 1; there is no a substitution, only
one 3 hydrogen, and only one 3-chlorine;
3a = 0.4; 3H = 1; kH= 0.01 x 10"12
6-20
-------
/ -12
!< 12) = 1(1) (0.4) (0.01 x 10 )
aos.
k/1(2) = 0.004 x 10~12
aos.
(iii) HC; n = 1; there is no a substitution and
only a 3 hydrogen; SH = 1; kH = 0.01 x 10~12
'n) = K1M0.01 x 10"12)
aos.
/
:at
k' 13) » 0.01 x 10
ibs.
kX = O.OOT6 x 10~12 -i- 0.004 x 10~12 +• 0.01 x 10"12
abs«
x, = 0.0156 x 10"12 cm3 molec.'1 sec.'1 • (19)
abs.
(b) Ri ng B
This ring is the same as in Section IV.A.2.b. Therefore,
from equation 12
k'X. = 0.05 x 10"12 cm3 molec."1 sec."1 . (12)
abs.
(c) Total kabs>
Te a V ->r ]f '
abs. abs. abs,
6-21
-------
Using equations 12 and 19 yields
k . = 0.015 x 10"12 + 0.050 x 10~12
abs.
k , = 0.066 x 10~12 cm3 raolec."1 sec."1 . (20)
abs.
3. Total k,-
k - k + kadd'
OH ~- abs. arom.
Using equations 18 and 20 yields
0.07 x 10~12 + 2.18 x 10"12
k.u » 2.25 x 10"12 cm3 molec."1 sec."1 . (21)
OH
The rate constants for these PCB congeners are summarized in
Table 1.
V. CALCULATION OF THE HALF-LIFE OF POLfCHLORINATHD 8IP4EMYLS
Hydroxyl radicals are formed as a result of a complex set of
chemical reactions in the atmosphere in the presence of sunliaht. The
concentration is a function of the solar light intensity (which
is a function of time of day, or zenith angle, and season of the
year)/ latitude, pollutant concentrations, temoerature, and altitude.
3ased on the research work of several scientists, a global
6-22
-------
average for the OH concentration for reasonably polluted air in
trie troposphere is 1 x 10^ molecules cn~^ [Hendry and Kenley
(L979), Sprung (1977), Crutzen and Fishman (1977)]. Using these
results in equation 5 yields a half-life of
0.693
0.693
k_aU x 106molecules cm"3] [8. 64 x 104sec. day"1]
On
. ,, , 8.02 x 10 12 .....
tlx(day) = - * - -y - -j- . (22)
2 k_,,(cni molec. sec. )
"
Consider the calculation of the half-life of 2-chloro-
biphenyl. From Section VI. A. 3, the second-order rate constant,
'
-------
VI. OISCUSSION OF RESULTS
Section III discusses in detail the general framework for
calculating the rate constants for the two major pathways for the
reaction of hydroxyl radicals with chlorinated biphenyls:
H abstraction (kabs ) and addition to double bonds in the
aromatic rings C^-Q^ )• Tne second-order rate constant C
-------
That is, changing the positions of the chlorines on the ring did
not change the value of kOH. Hence, all these congeners can be
grouped together as the tetrachloro PCB congener using the
general expression:
/ ' " // V> kOH = 2.1 x 10~12 cm3 molec."1 sec.'1.
(2) The dominant reaction' pathway is addition of OH
radicals to the double bonds of the aromatic ring
^add. ^ This leads to cleavage of the ring or the
arom.
formation of hydroxy PCBs.
(3) As the number of chlorines on the rings increase,
ka " decreases and consequently knH decreases. Since
arom. u"
knH decreases, q/, increases.
(4) For unsynmetrical chlorinated biphenyls
where x = 1, 2, 3, 4, 5,
it is evident that kOH is dominated by the rate of
attack on the unsubstituted ring and this process
controls the value of kQH. For example, for the
congener with x = 1, knH = 2.7 x 10~12 CTTI^
6-25
-------
."^- and q./ = 3.0 days while for the congeners with x
= 3 or 5, kOH = 2.1 x IQ""1-2 cm3 molec."-1- sec."1 and tU,
= 3.8 days. In other words, the presence of chlorines
on one ring deactivates this ring relative to the
unsubstituted ring and the overall kOH is dominated by
the rate of reaction on the unsubstituted ring.
(5) Consider the pentachlorobiphenlys:
Cl
X
5
4
3
1012k_,,(cm3 molec.~l sec."1.)
Un
0 2.1
1 0.65
2 0.26
todays)
3.8
12
31
(a) The unsymmetrically chlorinated congener with x =
5 and y = 0 has the largest KQ^ and the smallest
tLu. This occurs because the rate of addition is
dominated by the extremely large reactivity of the
unsubstituted ring (rule 4).
(b) As the chlorines become more symmetrically
distributed between the two rings, kQH decreases
and tl/ increases.
(c) For the congener with the most symmetrical
distribution of chlorines between the rings [i.e.,
6-26
-------
three chlorines on one ring and two chlorines on
the other ring], KQH has the lowest value and tl^
is a maximum. This general pattern is true for
all groups of chlorinated biphenyls: e.g.,
monochloro-, dichloro-, trichloro- etc. biphenyl
classes of congeners.
(6) The data, for the symmetrically distributed chlorinated
congeners with high numbers of chlorines can be
summarized as follows:
1012kOH(cm3 molec.'1 sec.'1) todays)
4
4
5
5-
3
4
4
5
0.088
0.039
0,424
0.0049
94
210
330
1700
As the total number of chlorines increases, kQ^
decreases rapidly and tL/ increases rapidly. The half-
life for the last two congeners is fairly large and
thus these two congeners are reasonably persistent.
VII. REFERENCES
Crutzen, PJ and Fishman J. 1977. Average concentration of OH in
the troposphere and budgets of CH4, CO, H2, and CH3 CC13.
Geophys Res Letters 4:321-324.
Cupitt LT. 1980. Fate of toxic and hazardous materials in the
air environment. EPA-600/3-80-084.
Hendry DG and Kenley RA. 1979. Atmosphere reaction products of
organic compounds. EPA-560/12-79-001.
6-27
-------
••[ill T, -tfinterle JS, Davenport JE, Lee GC, Mabey WR, Barich VP,
Harris W, and Bawol R. 19R2. Validation of estimation
techniques for predicting transformation of chemicals in the
environment. Unpublished [Draft Final Report for an EPA contract
witn SRI].
6-28
-------
CHAPTER 7
HYDROLYSIS AND OXIDATION OF
POLYCHLORINATED BIPHENYLS IN WATER
by
Asa Leifer
Contents
Page No.
I. INTRODUCTION AND SUMMARY 7-1
II. REFERENCES 7-2
7-i
-------
I. INTRODUCTION AND SUMMARY
There are no experimental data published in the literature
on the hydrolysis of polychlorinated biphenyls (PCBs) under
environmental conditions. However, all the PCS congeners contain
chlorines which are attached directly to the aromatic ring and as
a result they should not hydrolyze under environmental conditions
[Mabey and Mill (1978)]. Furthermore, PCBs are so hydrolytically
stable that even under severe acidic and basic conditions,
hydrolysis does not occur [Gustafson (1970), Hutzinger et al.
(1974)]. Hydrolysis, therefore, is not an important
environmental transformation process.
PCBs are extremely resistant to oxidation [Hutzinger et al.
(1974)]. Gustafson references a Monsanto, technical bulletin on
PCBs which states "they can be heated to 140°C under 260 psi of
oxygen pressure without showing any evidence of oxidation as
judged by the development of acidity or formation of sludge."
Oxidation, therefore, is not an important environmental
transformation process.
7-1
-------
II. REFERENCES
Gustafson CG. 1970. PCBs - prevalent and persistent. Env
Sci and Tech 4:814.
Hutzinger 0, Safe S, and Zitko V. 1974. The chemistry of
PCBs. CRC Press, Inc.
Mabey WR and Mill T. 1978. Critical review of hydrolysis
of organic compounds in water under environmental
conditions. J Phys Chem Ref Data 7:383.
7-2
-------
CHAPTER 8
PHOTOLYSIS OF POLYCHLORINATED BIPHENYLS
by
Asa Leifer
Contents
Page No,
I. INTRODUCTION AND SUMMARY 8-1
II. DISCUSSION OF RESULTS " 8-6
A. Ultraviolet Absorption Spectra... 8-6
B. Photolysis Data 8-10
III. REFERENCES 8-24
IV. APPENDIX: DETAILED REVIEW OF THE AVAILABLE
PHOTOLYSIS LITERATURE 3-26
8-i
-------
-------
I. INTRODUCTION AND SUMMARY
Two important factors must be considered when studying the
photolysis of polychlorinated biphenyls (PCBs) in solution and
estimating rates of photolysis in the environment. These factors
are: (1) the absorption of ultraviolet light by the PCBs in the
solar region (X, greater than 290 ran) and (2) the quantum yield
( 290 nm), highly chlorinated
8-1
-------
biphenyls absorb most strongly, PCBs lacking ortho substitution
are intermediate, and PCBs having one or two ortho chlorines are
the least absorbing. Nevertheless, all the PCBs are weak
absorbers at X > 290 nm.
A number of papers have been published on the photolysis of
PCBs in solution. Unfortunately, most of these experiments were
carried out in nonaqueous media. However, an understanding of
these data can be very useful and with caution, one can use these
results to obtain some insight into the photolysis of PCBs in the
environment; that is, photolysis in aqueous media in sunlight.
Bunce et al . (1978) published a paper on the photolysis of some
PCBs in the solvent system water-acetonitrile (1:4) containing
oxygen. Quantum yields were reported for several PCBs along with
molar absorptivities (S3Q0)« Using the method of Zepp and Cline
(1977) and Mill et al. (1982), direct photolysis rate constants
(kg) and half-lives (t/^ were calculated for several PCBs and
two Aroclors at 40° north latitude on the summer and winter
solstices at shallow depths (less than 0.5 meters) and under
clear sky conditions. All the results are summarized in
Table 1. Inspection of the data indicates that in general, as
the chlorine content increases, the photolysis rate constant
increases and the half-life decreases. Decachlorobiphenyl
photolyzes rapidly on the summer and winter solstices. A few of
these PCBs decompose at a moderate rate on the summer solstice.
However, it must be emphasized that these results must be used
with caution since the solvent is predominantly acetonitrij.e (75
percent) and therefore does not correspond to environmental
conditions.
8-2
-------
•**
-U
p*
*r>
W
Z
c
••4
3
«n
C
fl
U
^
2
c
cH
V
JJ
u
0
IU
31
a
>
•w
j
i
IU
*-«
2
V
<3
31
u
fl
JJ
a
0
u
4)
JJ
^
a.
91
•U
01
>.
*•*
0
JJ
0
£
JJ
£
c
•*J
1-4
jj
5
•
4)
f-t
&
«J
« 2
•^ 53
jj R
« "V
^-1 *->
0 JJ
CO
^j
QJ ^^
Wi»
C t
•** en
3 >
T3
t3
^^
a
0*
.X
01
>•
IQ
« 2
^J S
•U 'V
BS t-
0 *
en
«. ^^
=7
I'o,
5 S1
wi R}
*Q
^^
W
Oa
^
JC
Q
^
•
•
OOOS9° *
o o in ^ o vo • .
S «- m »- — « O
r- <~ V
^™
,n
1 1 1 1 1
0 ,C 0 0 0
x x x x x m ^
M •» O * <*> "* *
• • • c * ^ ^
!fl Vfl CN Vf) VO O A
vo
o
S - s a ? 5 s
^ CN CM CN (*)
o o o o o
^ *™ ^
x x x x x
m o o . o <* o «-
* • • • • * ^*
(»> ^» ^- rn co u^ ^
.0
S
p>4
1 U
0 <0
to- ta
Irt (rt 1 1 ff
o ^^. ** **
0 i-4 ^ 4) 33
S € a Y » o
1 i— 4 .^ *• * *• .W
1 P"» *^ '"^ r "
o — i- m . o
U 0 4J VJ U fH
o -H i .2 ^ ,2 •£
•S 7 * o ." o S
0 i « 0
C (0 1-4
0 JJ D JJ
•H C J= C
JJ 4] JJ (8
•H a . c
09 I* —4
0 JJ « S
a e *j o
E 4i « a
-C A) /*]
* 2 5-°
• jj
• ~. • 01
"0 O W 3
c >j 4) e
3 0 -« S C3
>u ^ 3
JJ IQ -
0 X W
C 4> 0 M
^ -* 0
01 O r*
IQ JJ 0 U
3 C u 0
4) <; u
CO U <
>P iw 4)
« « J3 .
— a JJ M
u en u £
0 m 4> e
T1 •£
o ~ a>
O 0 -H • •
U Vl £ 01 01
< O b b
^ 11 1) 01
Q f) U
0 4) o b us
a u o -u
a jj f>4 a
f. — C £ 4)
F ^r M o £
•
J3
i
CO
-------
Based on all the available PCB photolysis data in the
literature, dechlorination at 300 ± 10 nm is the predominant
reaction in nonaqueous solvents including methanol containing
oxygen, a solvent which is somewhat similar to water. Thus,
PC3s, especially the more highly chlorinated congeners and those
that contain ortho chlorines, photodechlorinate. All PCBs
containing ortho chlorines yield products arising from the loss
of these ortho chlorines. In their absence, meta chlorines are
cleaved. Para chlorines do not cleave to any significant
extent. The differences in photolability are due to the fact
that almost all ortho chlorinated biphenyls have high quantum
yields ($~ 0.1 or greater). PCBs lacking ortho chlorines
9 ^
generally have low quantum yields ($~ 10 ^ to 10 J) [Bunce
(1982)]. These results are significant, for if aqueous
photolysis is an important transformation process, then
photodechlorination results in the formation of lower chlorinated
congeners and congeners with less chlorine content are more
readily biodegradable (Chapter 9). Furthermore, PCBs with ortho
chlorines are the most readily removed by photolysis and thus the
resulting dechlorinated PCBs are more readily biodegradable.
Hence, over a period of time, a combination of photolysis and
biodegradation could remove PCBs from the environment.
More reliable photolysis data is needed on selected PCBs to
estimate rates of photolysis in aqueous media in sunlight and to
predict the transformation products and the mechanism of this
transformation process. Photolysis experiments should be carried
out in water - acetonitrile (99:1) to determine the uv absorption
8-4
-------
spectra, the molar absorptivities (e^ ), and the quantum yield
(?) .
Hutzinger et al. (1974) discusses the photolysis of PCBs in
the gas phase but there are no relevant data published to predict
rates of photolysis in the atmosphere or the nature of the
decomposition products.
8-5
-------
II. DISCUSSION OF RESULTS
•
A. Ultraviolet Absorption Spectra
As a prelude to the discussion of the photolysis of
PC3s in the environment and specifically to environmentally
relevant photolysis rates, it is necessary to consider the
absorption of light by polychlorinated biphenyls (PCBs) in
aqueous media. Information on the absorption of light by PCBs
can be obtained from the ultraviolet (uv) absorption spectrum.
Unfortunately, very little data is available on the uv spectra of
PCBs in aqueous media and one has to resort to data from the uv
spectra in nonaqueous solvents [Hutzinger et al. (1974)]. An
analysis of these data indicates that there are several very
useful correlations on uv absorption and PC3 structure. With
caution, one can use these results to obtain insight into the
behavior of the absorption of uv light by PCBs in aqueous media.
The uv spectrum of biphenyl contains two important
absorption maxima: one band is at 202 nm (e = 44000) and is
designated as the main band; the other absorption maxima is at
242 nm (s = 17000) and is called the < band. The K band is
attributed to the conjugated biphenyl system with contributions
from both rings. The effects of chlorine substitution on the
rings on \ are given in Tables 2 and 3 [Hutzinger et al.
TOcL X
(1974)]. Table 2 lists X „ for the PCBs with one or no
ma X
chlorine in the ortho position while Table 3 lists Xmax for the
PCB congeners with two or more ortho chlorines. \n analysis of
8-6
-------
Table 2. UV spectra of Chlorcbiphenyls (None or One Orthe Chlorine)
Chlorinated biphenyl
4
3
2
4,
3,
2,
2,
2,
2,
3,
2,
2,
2,
2,
2,
3,
Congener
4'
3'
4
4,4'
3', 4
3' ,4' ,5
3' ,4,4'
3' ,4,4'
4, 4', 5
3,4,4'
3, 4', 4, 5
3, 3', 4, 4'
3', 4,4', 5, 5'
0V Maxima and Extinction
Coefficients (x 10~3)
"Main band" < band
(nm)
199
205
204
200
204
205
210
214
222
(43.
(42.
(39.
(41.
(42.
(42.
(44.
(42.
(51.
3)
3)
2)
9)
2)
5)
9)
0)
7)
(no)
253
248
240
258
248
255
250
246
248
260
253
257
250
253
253
265
(20.5)
(16
.0)
(10.2)
(22
(23
(12
(14
(12
(11
(22
(15
(15
(12
(2.
(2.
(27
.9)
.4)
.3)
.3)
.0)
.3)
.9)
.9)
.1)
.6)
S)
5)
.7)
8-7
-------
Table 3. Ultraviolet Spectra of Chlorobiphenyls (Two or More Ortho Chlorines)
Chlorinated biphenyl
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
Congener
21
2', 5
2',4,4'
2', 5, 5'
2', 6, 6'
2', 4, 4', 5, 5'
2', 4, 4', 6, 6'
2', 3, 4, 5, 5', 6
2', 3, 3', 4, 4' ,5,5'
2', 3, 3', 5, 5', 6, 6'
2', 3, 3', 4, 4', 5, 5', 6, 6'
uv Maxima and ac-tinetion
Coefficients (x 10~3)
"Main band" « band "£ bands"
(nm)
208
197
207
204
197
211
202
214
210
210
216
(36
(62
(51
(43
(88
(45
(93
.0)
.5)
.2)
.3)
.9)
.5)
.1)
(100)
(57
(91
.5)
.6)
(108)
(nm) (nm)
230 (6.6) 273
266.
267
275
283
220 (29.4) 273
282
214 (34.9) 276
284
272
280
282
290
267
275
288
268
277
286
297
285
294
285
295
291.
301.
,5
(1
(1
(0
(1
(0
(1
(1
(0
(0
(1
(1
(0
(0
(0
(1
(1
(1
(0
(0
(0
(2
(2
5
5
54)
(.74)
.10)
.17)
.32)
.49)
.83)
.32)
.25)
.78)
.65)
.60)
.12)
.50)
.59)
.46)
.37)
.76)
.82)
.63)
.69)
.59)
.04)
.31)
(1 .10)
(1.22)
Shoulder
8-8
-------
these data indicate that these absorption maxima are affected by
the location of chlorines on the rings, the number of chlorines
on the rings and, in particular, by the degree of substitution at
the positions ortho to the Ph-Ph bond (i.e., at the positions
2,2', 6,6').
Consider the spectra of PCBs with less than two
chlorines ortho to the Ph-Ph bond, Table 2. For the monochloro-
biphenyls, the main absorption band only changed slightly in
comparison to the same band in biphenyl. However, the 3- and 4-
chloro groups induced a bathochrochromic or red shift (i.e., a
shift to higher wavelengths) of the K band with the 4-chloro
groups showing the larger shift. The < band for the 2-chloro
congener is shifted to a slightly lower wavelength (i.e., a blue
shift) with a lowering of e. This effect has been attributed to
some steric inhibition of resonance between the two phenyl rings.
Similarly, for the 4,4'- and the 3,3'-congeners, the
magnitude of the red shift of the < band is greater for the para
disubstituted derivative.
For the more highly substituted PCBs, both the main
absorption and < bands shift to the red and exhibit appreciably
more absorption tailing in the solar region beyond 290 nm.
Upon the introduction of two or more chlorines in the
ortho positions to the Ph-Ph bond, major changes in the uv
spectrum occur, Table 3. The < band shifts to lower wavelengths
and the molar absorptivity e is lowered markedly; on the other
8-9
-------
hand, the e value for the main band increases significantly. In
addition, these highly ortho substituted congeners exhibit a
series of weak absorption maxima, called 3 bands, between 268 and
302 nm with tailing absorptions in the region beyond 290 ran. In
the same manner as encountered in 2-chlorobiphenyl, it is
generally accepted that these highly hindered PCBs are hindered
to free rotation about the Ph-Ph bond and this results in the
loss of coplanarity between the two rings. Thus, these 3 bands
have been attributed to the contributions from the individual
phenyl rings.
The significance of these changes in the absorption
spectrum with respect to photolysis in the environment is that in
the solar region of the spectrum (X > 290 nm), highly chlorinated
biphenyls absorb most strongly, PCBs lacking ortho substitution
are intermediate and PCBs having one or two other chlorines are
the least absorbing. Nevertheless, all the PCBs are weak
absorbers at X > 290 nm.
B. Photolysis Data
A number of papers have been reviewed which pertain to
the photolysis of PCBs in solution [Safe and Hutzinger (1971),
Hustert and Korte (1972), Hutzinger et al. (1972), Ruzo et al.
(1972), Ruzo et al. (1974), Nordblom and Miller (1974), Hutzinger
et al. (1974), Ruzo et al. (1975), Safe et al. (1976), Wagner and
Schere (1977), Bunce and Kumar (1978), Bunce (1982)].
Unfortunately, almost all of these photolysis experiments were
3-10
-------
carried out in nonaqueous media. However, a study of these data
can provide some potential insight into the photolysis of PC3s in
the environment in sunlight. The following paragraphs highlight
these results and Appendix IV discusses these papers in greater
detail.
All the experiments on the photolysis of PC3s were
carried out in the laboratory by irradiation in the spectral
region 290-310 run; and these results are environmentally relevant
with respect to solar radiation since the wavelengths of
photolysis occur beyond the solar cutoff (X > 290 nm). As
discussed in Section II.A., many of the PCB congeners, especially
those which contain several chlorines on the rings, have weak
absorption tails beyond 290 nm. Furthermore, the PCB congeners
with more than one chlorine in an ortho position exhibit weak
absorption maxima in the region 270-300 nm (B absorptions) and
the absorptions tail beyond 290 nm. As a result, these PCBs all
have the potential to undergo photolysis in the solar region at
wavelengths greater than 290 nm.
Based on a review of all the available literature, PCBs
undergo photodechlorination when exposed to radiation in the
spectral region 290-310 nm. For example, Ruzo et al. (1974)
carried out a series of photolysis experiments with the tetra-
chloro PCB congeners [(3,3 ' ,4,4'), (2,2',4,4'), (3,3',5,5'),
(2,2'3,3'), (2,2'5,5') and (2,2',6,6')] in degassed cyclo-
hexane. The major reaction undergone by these congeners was the
stepwise dechlorination to yield tri- and dichlorobiphenyls as
8-11
-------
the major products. All PCBs containing ortho chlorines yielded
products arising from the loss of the ortho chlorines. In their
absence, meta chlorines were cleaved. Chlorine in the para
position did not cleave to any significant extent. Similar
experiments were carried out by photolyzing these compounds in
undegassed and degassed methanol at 290 - 310 run. Again, the
major reaction undergone by these compounds was stepwise
dechlorination to yield tri- and dichlorobiphenyls as the major
products. The order of removal of the chlorines was ortho
chlorines > meta chlorines » para chlorines. The differences in
photolability are due to the fact that almost all ortho
chlorinated biphenyls have high quantum yields ($ ~ 0.1 or
greater). PCBs lacking ortho chlorines generally have low
quantum yields ($ ~ 10"2 to 10"3} [Bunce (1982)1. Minor amounts
of the methoxylated products were also observed «3 percent of
reacted PCBs in all cases). It should be noted that methanol is
a hydroxylated solvent, similar in some respects to water. Thus,
one might expect a similar pattern of photodechlorination in
water, especially for the PCB congeners with chlorines in the
ortho positions. Furthermore, these results were applicable to
methanol containing oxygen. Water in the environment usually
contains oxygen.
In order to elucidate the mechanism of photo-
dechlorination described above, Ruzo et al. (1974) carried out
more detailed photolysis experiments: the quantum yield ($ ) was
determined for several congeners in cyclohexane, Section IV,
Table 9; the quantum yield of intersystem crossing (. ),
1. SC
8-12
-------
corresponding to the conversion of the excited singlet state to
the excited triplet state, was measured for the congeners
(3,3',4,4'), (2,2'4,4')/ (3,3',5,5') and was found to be 1 ±0.05;
and the triplet lifetimes (t) were measured in cyclohexane and
methanol for all the congeners [Section IV, Table 9]. The
lifetimes of all the ortho congeners were approximately three
times smaller than the congeners containing no ortho chlorines.
Because of the presence of the ortho chlorines, the ground state
is non-planer while the excited triplet state is planar [Wagner
(1967), Wagner and Scheve (1977)]. Triplet lifetime shows a
definite variation between the ortho congeners and the others.
This is undoubtedly due to the greater steric hindrance to the
preferred excited state geometry. Crowded conditions created by
the ortho chlorines result in a greater twisting of the Ph-Ph
bond with the subsequent decrease in its double band character.
The products obtained in the greatest yield arise from the loss
of ortho chlorines; thus, the strain on that bond is,relieved.
In some recent work, Bruce (1982) reported the results
of photolysis experiments using a series of compounds of the type
CJ
Compounds with these structures were more photolabile than the
analogous compounds lacking the ortho methyl substituents, but
were not as photolabile as the ortho chlorine compounds. Bunce
8-13
-------
concluded that the relief of strain when an ortho chlorine
departs must be important. However, the extraphotolability
arises because the ortho chlorine substituent raises the energy
of the excited state and hence provides the extra energy for
dissociation.
Based on all the data, Ruzo et al. (1974) postulated
the mechanism of photodechlorination of PCBs. As an example, the
mechanism is illustrated in Figure 1 for the congener 2,2'4,4'-
tetrachlorobiphenyl (congener II, Table 7, Section IV). The
dechlorination products obtained from the uv irradiation result
from the cleavage of the C-C1 bond in the ortho position to form
a biphenyl free radical. This free radical then abstracts a
hydrogen from the solvent to form the dechlorinated product and
HC1. Photolysis of congener II in methanol yielded a small
amount of the methoxylated products (Table 8, Section IV). These
methoxylated products would be formed by nucleophilic
displacement of chlorine by methanol. In the latest publication
by Bunce (1982), this researcher summarized the status of the
photolysis of PCBs and supported the mechanism postulated by Ruzo
et al. (1974).
Hutzinger et al. (1972) attempted to carry out photo-
lysis experiments on selected PCB congeners under environmental
conditions (i.e., photolysis in sunlight). Unfortunately, these
experiments were poorly designed and no useful data were
obtained.
3-14
-------
Figure 1. Mechanism of ?hotodeconposition of 2, 2',4, 4'
Tetrachlorobyphenyl
C!
ci—^/>—f "7^~cl
'Cl
I)
„
„
M«OH
-HC1
RH
Cl
OM«
HCJ
8-15
-------
Bunce et al. (1978) carried out photolysis experiments
on selected PC3s to assess the impact of solar degradation of
PC3s in the aquatic environment. These researchers obtained
extinction coefficients at 300 run and quantum yields at 254 nm
(at less than 10 percent conversion) for a series of PCB
congeners and two Aroclor samples in the oxygenated solvent
system water-acetonitrile (1:4). All their results are
summarized in Table 4. In general, the quantum yield for complex
molecules in solution is wavelength independent so that the
quantum yield at 254 nm is the same in the spectral region
greater than 290 nm [Zepp and Cline (1977)]. These data can be
used to estimate rates of photolysis by the method of Zepp and
Cline (1977) and Mill et al. (1982) to see if photolysis would be
important in an aquatic environment in sunlight.
Zepp and Cline (1977) published a paper on the rates of
direct photolysis in aquatic environments. The rates of direct
photochemical processes in a water body are affected by solar
spectral irradiance at the water surface, radiative transfer from
air to water, and the transmission of sunlight in the water
body. It has been shown that in dilute solution (i.e., the
absorbance of a chemical is less than 0.02 in the reaction cell
at all wavelengths greater than 290 nm) at shallow depths
(>0.5m), the kinetic expression for direct photolysis of a
chemical at a molar concentration C is
- dC/dt m 9EkaC » *PEC , (1)
8-18
-------
3S
...
S3
U
0.
4)
S
0
e
0
15
u
ia
c
"
0
—
5
^
u
0
JJ
0
^
a.
t
^
4)
,-4
i
rtj
•e-
_
r>
i
S
u
,-
1
z
~-
0
0
u
to
u
0.
r-
n p. vo
o ,
C
«
^
a
1-4 -*4
>. J3
= 0
-4 «) W
>. f 0
C fl4 1-1 ft
j) -«t j; >,
-4 JS J3 O C
>. a. o IB o
c -H u u —
4) ^Q 0 4J ^
.£ 0 "^ 4) »* f «0
CU W «C 4J «2 ^ vC
^100 1 0
0 — i m £ o 0
>-4 ^J V0 . O ~4 i-t
£ 1 » - fl O 0
O rr
^4
4J
a
^4
0
0)
JD
fl
W
-------
with
pE
where <(>„ is the reaction quantum yield of the chemical in dilute
tL
solution and is independent of the wavelength, and ka = £ k ^,
the sum of k . values of all wavelengths of sunlight that are
0 A
absorbed by the chemical. The term kg represents the photolysis
rate constant for water bodies in sunlight in the units of
reciprocal time. Integrating equation 1 yields
ln(CQ/C)= kp£t , (3)
where C is the molar concentration of chemical at time t during
photolysis and CQ is the initial molar concentration. By
measuring the concentration of chemical as a function of the time
t during photolysis in sunlight, k_g can be determined using
equation 3. In addition, equation 3 can be solved for the
condition Ct = CQ/2 and the half-life of the chemical is given by
°-693/kpE . (4)
Furthermore, under the same conditions cited above
[i.e., for homogenous chemical solution with absorbance less than
0.02 in a reaction cell at all wavelengths greater than 290 nm
At an absorbance of 0.05, equation 5 is in error by only 11%.
8-18
-------
and at shallow depths (less than 0.5 m)], the first-order direct
photolysis rate constant, ^n£' ^s
where 4>e is the quantum yield which is independent of the
wavelength, e is the molar absorptivity in the units molar
A
cm , and L, is the solar irradiance in water in the units 2.303
A
x 10"3 einsteins cm"2 day ~l [Mill et al. (1981, 1982a,
1982b)]. L. is the solar irradiance at shallow depths for a
water body under clear sky conditions and is a function of
latitude and season of the year.
Calculations were carried out for all the PCBs listed
in Table 4 to see if sunlight photolysis would be important. The
following assumptions were made in the calculations.
1. The molar absorptivity reported by Bunce et al.
(1978) in the solvent water-acetonitrile (1:4) was
used. It was assumed that the molar absorptivity
decreased uniformly as the wavelength increased:
S30S = ) £300? 6310 s £305'* £315
£310
2. A latitude of 40 °N was chosen since this latitude
is approximately in the center of the Uni'ted
States.
3. Calculations were made on the summer and winter
solstices.
8-19
-------
4. The solar irradiance values (L^ ) were obtained
from Mill et al. (1982) and interpolations were
made to estimate L^ at 300, 305, 310, 315, ..... nm
on the summer and winter solstices.
5. The quantum yield reported by Bunce et al. (1973)
in the solvent water-acetonitrile (1:4) was used.
6. The calculations correspond to water bodies under
clear sky conditions and at shallow depths.
The following calculation for 2,4-dichlorobiphenyl
illustrates the method of determining the rate constant (kpE)
half-life (t]/2E^ for the summer and winter solstices. Table 5
summarizes the data for 1. *v^x f°r fcne summer and winter
solstices. Substituting the values of J! c^ L^ from Table 5 in
equation 5 yields
Summer solstice: k_E = 0.040 days'*
Winter solstice: k_E = 0.0064 days""*
Substituting these results in equation 4 yields
Summer solstice: tl/2£
Winter solstice: tl/2E
The results for all the PCBs are summarized in Table 1 of
Section I. Inspection of the data indicates that in general, as
the chlorine content increases, the photolysis rate constant
increases and the half-life decreases. Decachlorobiphenyl
8-20
-------
Table 5. Summary of Photolysis Data for 2,4-Oishlorobiphenyl at 40° Morth
Latitude
Simmer Solstice
X (ran) e, (M"1cm'1} L*, s, L-, (day
X
300 2S.O 0.66 X 10~3
305 12.3 3.4 x 10~3
310 6.2 0.99 X 10~2
315 3.1 2.0 x 10~2
320 0.0 —
winter Solstice
X (an) sx (M"1ca"1) I,* ^ s^ L^ (day
300 25.0 0.35 x 10~4
305 12.3 0.33 X 10~3
310 6.2 1.6 x 10'3
315 3.1 4.5 x 10~3
320 0.0
£ e\L\ * 0.029
*The units of L^ are in 2.303 x 10~3 einsteins cm"2 day~1.
8-21
-------
photolyses rapidly in the summer and winter solstices. A few of
these PC3s decompose at a moderate rate on the summer solstice.
However, it must be emphasized that these results must be used
with caution since the solvent is predominantly acetonitrile (75
percent) and therefore does not correspond to environmental
conditions. It should be noted that E and the molar absorptiv-
ities were obtained in oxygenated solvent.
Based on all the photolysis data available,
dechlorination is the predominant reaction. Thus PC3s,
especially the more highly chlorinated congeners and those that
contain two or more chlorines in the ortho positions, photo-
dechlorinate. This is a significant result, for if aqueous
photolysis is an important transformation process in the environ-
ment, then photolysis results in the formation of lower
chlorinated congeners and the lower chlorinated congeners
biodegrade more readily (Chapter 9). Furthermore, PCBs with
ortho chlorines biodegrade very slowly (Chapter 9). However,
ortho chlorines are the most readily removed by photolysis and
thus the resulting dechlorinated PCBs are more readily
biodegradeable. Hence, over a period of time, a combination of
photolysis and biodegradation could remove PCBs from the
environment. Since this could represent a viable mechanism for
the removal of PCBs from the environment and a number of the PC3
congeners have the potential to undergo photolysis" at a moderate
rate in the summer, it is important that laboratory studies be
carried out in the solvent water-acetonitrile (99:1). Reliable
molar absorptivities and quantum yields are needed in this
8-22
-------
solvent saturated with oxygen to verify if PCps can transform
photolytically in the environment.
8-23
-------
III. REFERENCES
Bunce NJ. 1982. Photodechlorination of PCBs: current
status. Chemosphere 11:701.
Bunce NJ, Kumar Y, and Brownlee BG. 1978. An assessment
of the impact of solar degradation of polychlorinated
biphenyls in the aquatic environment. Chemosphere No.
2:155.
Hustert K and Korte F. 1972. Beitrage zur okologischen
chemie XXXVIII. Synthese polychlorierter biphenyle und
ihre reaktionen bei uv - bestrahlung. Chemosphere No.
1:17.
Hutzinger 0, Safe Sf and Zitko V. 1972. Photochemical
degradation of chlorobiphenyls (PCBs). Env Health Persp
1:15.
Hutzinger 0, Safe S, and Zitko V. 1974. The chemistry of
PCBs. Chapter 6. Photodegradation of chlorobiphenyls.
Chapter 10. Ultraviolet spectroscopy of chlorobiphenyls.
CRC Press,
Mill T, Mabey WR, Bomberger DC, Chou T-W, Hendry DG, and
Smith JH. 1982. Laboratory protocols for evaluating the
fate of organic chemicals in air and water. Chapter 3.
EPA 600/3-82-022.
Nordblom GD and Miller LL. 1974. Photoreduction of 4,4'-
dichlorobiphenyl. J Agri Food Chem 22:57.
Ruzo LO, Zabik, MJ, and Scheutz RD. 1972. Polychlorinated
biphenyls: Photolysis of 3,4,3',4'-tetrachlorobiphenyl and
4,4'-dichlorobiphenyl in solution. Env Cont and Tox 8:217.
RUZO LO, Zabik MJ, and Scheutz RD. 1974. Photochemistry
of bioactive compounds. Photochemical processes of
polychlorinated biphenyls. J Am Chem Soc 96:3809.
Ruzo LO, Safe S, and Zabik MJ. 1975. Photodecomposition
of unsymmetrical polychlorobiphenyls. J Agri Food Chem
23:595.
Safe S and Hutzinger 0. 1971. Polychlorinated
biphenyls: photolysis of 2,4,6,2',4',6'-
hexachlorobiphenyl. Nature 232:641.
Safe S, Buncs NJ, Chittim B, Hutzinger 0, and Ruzo LO.
1976. Chapter 3. Photodecomposition of halogenated
aromatic compounds. In: Identification and analysis of
pollutants in water. L.H. Keith, Editor. Ann Arbor Press.
8-24
-------
Wagner PJ. 1967. Conformational changes involved in the
singlet-triplet transitions of biphenyl. J 'Am Chem Soc
39:2820.
Wagner PJ and Scheve BJ. 1977. Steric effects in the
singlet-triplet transitions of methyl- and
chlorobiphenyls. J Am Chem Soc 99:2888.
Zepp RG and Cline DM. 1977. Rates of direct photolysis in
aquatic environment. Environ Sci and Technol 11:359.
8-25
-------
IV. APPENDIX; DETAILED REVIEW OF THE AVAILABLE PHOTOLYSIS
LITERATURE
Safe and Hutzinger (1971) carried out some of the earliest
experiments on the photolysis of PCBs. Specifically/ the PCS
congener 2,2',4,4',6,6'-hexachlorobiphenyl was photolyzed at 310
nm in the solvents hexane and methanol. Although the structures
of the products were not determined, the mass spectra of the
products indicated that dechlorination took place.
Laboratory experiments were carried out by Ruzo et al.
(1972) on the photolysis of 4,4'-dichlorobiphenyl (DCS) and
3,3',4,4'-tetrachlorobiphenyl (TCB) in hexane at wavelengths
greater than 286 nm and A at 310 nm. DCB decomposed to a
max
small extent to 4-chlorobiphenyl and no biphenyl was detected in
the reaction products. The absence of biphenyl is not surprising
since 4-chlorobiphenyl shows no tailing beyond 290 nm while DCB
exhibits marginal tailing absorption at X > 290 nm which results
in a low yield of 4-chloribiphenyl. The photolysis of TCB
yielded stepwise dechlorination: TCB decomposed to 3,4,4'-
trichlorobiphenyl which photolyzed to 4,4'-dichlorobiphenyl.
Thus, the meta chlorines were the most labile and were removed in
a stepwise sequence to form DCB. A similar pattern was observed
for the photolysis of several hexa- and tetrachlorobiphenyls
[Hustert and Korte (1972)1.
Several unsymmetrical PCBs were photolyzed in degassed
cyclohexane in the wavelength region 300 ± 10 nm [Ruzo et al.
(197S)]. GC/MS analysis of the products indicated the loss of
8-26
-------
one or two chlorines followed by hydrogen abstraction from the
solvent. The products and yields are listed in Table 6. The
quantum yield was determined for each PCS at less than 10 percent
conversion to avoid sensitization or quenching of the reaction by
the photoproducts and these results are summarized in Table 6.
The reactivity of the PCBs depends upon the position of the
chlorines on the rings: ortho chlorines cleaved first and at a
considerably faster rate than meta chlorines; para chlorines did
not cleave.
In another set of photolysis experiments, a series of
symmetrical tetrachlorobiphenyls were photolyzed in degassed
hexane and methanol and in methanol containing oxygen at 300 ± 10
nm [Ruzo et al. (1974)]. The compounds studied are listed in
Table 7 and are designated as compounds I-VI, The photodegra-
dation products are listed in Table 8. The major reactions of
compounds I-VI in cyclohexane were stepwise dechlorinations to
yield tri- and dichlorobiphenyls as the major products. Very
little monochlorobiphenyl was detected after 20 hours of
radiation (less than 1 percent of reacted PCS). All PCBs
containing ortho chlorines yielded products with the loss of
ortho chlorines. In the absence of ortho chlorines, meta
chlorines were cleaved. Para chlorines were not cleaved after 20
hours of photolysis.
In methanol, dechlorination was also found to be the major
reaction, Table 8. However, minor amounts of methoxylated
8-27
-------
Table 5. Photoprcducts and Quantum Yields of Unsymmetrical PCB's in
Cyclohexane
PCS
2,4, 6-Trichloro-
biphenyl (I)
2,4, 5-Trichloro-
biphenyl (II)
2,3,4, 5-Tetrachloro-
biphenyl; (III)
2,3,5, 6-Tetrachloro-
biphenyl (IV)
2' , 3 , 4-Trichloro-
biphenyl (V)
Product
0.02 2,4-Dichloro
4-Chloro
0.05 3,4-Dichloro
4-Chloro
0.04 3,4,5-Trichloro
3,4-Oichloro
<0.01 2,3,5-Trichloro
3,5-Dichloro
0.02 3,4-Dichloro
Tr
Min.
1 .25
0.85
1 .80
0.85
3.60
1 .80
2.20
1.50
1.80
%
Yield*
35
15
98
2
95
5
50
50
100
Based on total product formation.
3-2S
-------
Table 7. Tatrachlorobiphenyls
PC3 Designation
3,3',4,4'-Tatrachlorobiphenyl I
2, 2',4,4'-Tetrachlorobiphenyl II
3,3',5,5'-Tetrachlorobiphenyl III
2,2' ,3,3'-Tetrachlorobiphenyl IV
2,2',5,5'-Tetrachlorobiphenyl V
2,2',6,6'-Tetrachlorobiphenyl VI
8-29
-------
Table 3, ?hotcproducts in Hexane and in Methanola
?C3
Dechlorinated oroducts
Methoxylated products
II
III
IV
VI
3,4,4'-Trichlorobiphenyl
4,4'-Qichlorobiphenyl
2,4,4'-Trichlorobiphenyl
4,4'-Dichlorobiphenyl
4-Chlorobip henylc
3,3',S-Trichlorobiphenyl0
2,3,3'-Trichlorobiphenyl
2,2',3-Trichlorobiphenylc
3,3'-Oichlorobiphenyl
2,3,5'-Trichlorobiphenyl
3,3'-Oichlorobiphenyl
3-Chlorobip henylc
2,2'6-Trichlorobiphenyl
2,2'-Dichlorobiphenyl2
Tr i chlorome thoxyb ip henyl
Tr i chlorome thoxyb ip henyl
Dichlorodiaiethoxybip henyl°
Tr ichloromethoxyb ip henyl
Di chlorodime thoxyb ip henylG
Trichlorome thoxyb ip henyl
Dichlorodiae thoxyb ip henylc
Tr ichlorome thoxyb ip henyl°
^ondegassed solutions: Twenty hours irradiation.
Only major products are shown.
cCoroound represented less than 1 percent of reacted starting material.
8-30
-------
products were observed (less than 3 percent of reacted PCS in all
cases). These results are important because methanol is a
hydroxy solvent, similar in some respects to water, and thus one
might expect that dechlorination would be the major photolytic
process in aqueous media. In addition, the photolysis
experiments were carried out in methanol containing oxygen. In
general, water bodies in the environment contain oxygen.
Quantum yields for reaction ( $ ) were determined in
degassed cyclohexane for compounds I-VI at 300 ran, Table 9.
Maximum conversion (of PCS reacted) was kept below 10 percent to
avoid sensitization or quenching of the reaction by the
photoproducts.
Quantum yields of intersystem crossing ($•--)»
corresponding to the conversion of the singlet excited state to
the triplet excited state, were determined by measuring the
phosphorescence emission intensity of biacetyl at 516 nm
sensitized by either benzophenonone or PCS. Compounds I, II, and
III were tested and $,_„ was found to be 1 ± 0.05 compared to
benzophenone (4»;sc = 1 ) • Thus the conversion to the triplet
state is extremely efficient.
Quenching studies were performed on the photolysis of
compounds I-VI in degassed methanol and cyclohexane. Degassed
solutions of PCS containing various concentrations of 1,3-
cyclohexadiene (Et < 55 kal/mol) were irradiated to conversion
< 20 percent. Based on an analysis of the data, the triplet
lifetimes (T) were calculated and are listed in Table 9. The
8-31
-------
Table 9. Summary of Triplet State Reactivities of Polychlorobiphenyls in
Cyclohexane
?ca x
I 0.005
II 0.100
III 0.002
IV 0.007
7 0.010
VI 0.006
T, 10" 8 1/T, 107 Ky, IO7 fy 107
-1 * -1a -1b
sec sec sec sec
2.20 4.54 0.023 4.52
0.78 12.82 1.282 11.54
1.91 5.23 0.010 5.22
0.77 ' 12.99 0.091 12.90
0.67 14.92 0.149 14.77
0.70 14.28 0.086 14.20
8-32
-------
values of t were essentially the same within experimental error
in both solvents.
The properties of biphenyl and its derivatives in the
ultraviolet indicate that the ground state is nonplanar. The
excited triplet state of biphenyl is planar [Wagner (1967),
Wagner and Scheve (1977)]. Based on an analysis of the data, it
was postulated that the triplet excited states of the PCBs
studied are planar [Ruzo et al. (1974), Bunce (1982)]. Triplet
lifetimes show a definite variation between ortho substituted
PCBs and the others, Table 9. This is undoubtedly a result of
the greater steric hindrance to the preferred excited state
geometry. Crowded conditions created by the ortho substituents
result in a greater twisting of the ?h-Ph bond with a decrease in
its double bond character. The products obtained in the greatest
yields arise from the loss of ortho chlorine; thus, the strain in
the Ph-?h bond is relieved.
Ruzo et al. (1974) postulated the following mechanism based
on all results
hu 1 * i sr 7 f P—PI 1
i a-1• i i •*• r p_r11 LSC j i f—v.11 , , .
Lf-v_lJ —-= Lr-v.IJ L _,— , (n)
* VH *
—^~ [p-cn , a)
—^— (products) , (8)
8-33
-------
where °[?-Cl), 1(P-C1]*, and 3[P-C1]* represent the PCB in its
ground state, excited singlet state, and excited triplet state,
respectively. Ia is the quanta of light absorbed by the ground
state .PCB and ^. is the quantum yield for conversion of the
excited singlet state to the excited triplet state. Kinetic
analysis of reactions 6-8 yielded
(9)
F.rom the experimental data of r~ , 4> , and $|sc/ k^ and ^
have been calculated and these results are summarized in
Table 9. The value of kr is much greater for II, IV, V, and VI
than for I and III. Thus, the presence of ortho chlorines
decreased the lifetime and increased the reactivity of the
excited triplet state. As a result, the ortho substituted PCBs-
react much more rapidly. This has been ascribed directly to the
destabilizing effect of ortho substitution. The large
differences in kr between compound II and the others may be
attributed to the increased double bond character of the Ph-Ph
bond as a result of the para chlorine. Since it has been
postulated that a para substituent may increase conjugation
between the phenyl rings by election donation, the excited state
of PCB II may be represented by structures 1 and 2.
8-34
-------
a q
Vci
^ •
'a 'a
This effect would bring about greater conjugation with increasing
driving force to planarity resulting in a faster chlorine
cleavage.
Ruzo et al. (1974) postulated the mechanism of the
photodechlorination of 2,2'4,4'-tetrachlorobiphenyl (II) and this
mechanism is depicted in Figure 1, Section II.B. The
dechlorination products obtained from the uv irradiation result
from the cleavage of the C-C1 bond in the ortho position to form
a biphenyl free radical species. This free radical species then
abstracts a hydrogen atom from the solvent. HC1 was detected as
a reaction product, in support of this mechanism. Photolysis of
compounds I-VI in methanol yieided a small amount of the
methoxylated products, Table 8. These methoxylated products
would be formed by nucleophilic displacement of chlorine by
methanol.
The photolysis of 4,4'-dichlorobiphenyl was performed in
degassed methanol and isopropyl alcohol at 310 nm [Nordblom and
Miller (1974)]. The reaction yielded exclusively 4-
chlorobiphenyl which was stable. Photolysis in CH^OD did not
lead to the incorporation of deuterium indicating that the
hydrogen atom is donated from the methyl group and is typically a
free radical reaction where the weaker C-H bond is broken in
preference to the 0-H bond.
3-35
-------
All the data on the photolysis of PC3s in solution can be
summarized as follows. PCBs with chlorines in the ortho position
decompose more readily than congeners without ortho
substitution. This is a direct result of the fact that, in
general, ortho substituted biphenyls have higher quantum yields
than PCBs without ortho chlorines. The ortho substituted PCBs
have a nonplanar configuration due to steric hindrance of the
»
ortho chlorines. The mechanism of photodecomposition involves
the triplet state which is planar. Crowded conditions created by
the ortho substituents result in greater twisting of the Ph-Ph
bond with a decrease in its double bond character. The products
obtained in the greatest yields arise from the loss of ortho
chlorines, thereby relieving the strain in the Ph-Ph bond. Thus,
the ortho substituted PCBs decompose photolytically in a stepwise
fashion by removal of the ortho chlorines.
8-36
-------
CHAPTER 9
BIOPEGRAPATION OF CHLORINATED BIPHENYLS
by
Robert H. Brink
Contents
Page No.
I. INTRODUCTION AND SUMMARY 9-1
II. MONITORING EVIDENCE 9-3
III. BIODEGRADATION RATES 9-4
A. General .. 9-4
B. Anaerobic . 9-5
C. Aerobic Aquatic 9-5
D. Biological Waste Treatment 9-8
S. Soil 9-11
IV. ANAEROBIC BIODEGRADATION 9-12
V. PURE CULTURE STUDIES.... 9-13
VI. SORPTION ' 9-15
VII. VOLATILIZATION 9-16
VIII. OTHER FACTORS 9-17
IX. REFERENCES 9-19
9 -i-
-------
-------
I. INTRODUCTION AND SUMMARY
A review of the available literature on the degradation of
chlorinated biphenyls by microorganisms discloses many
conflicting findings and conclusions. There are, however, some
discernable patterns. It is quite clear that there are numerous
aerobic "nicroorqanisms in the environment that are capable of
degrading most of the chlorinated biphenyls present in commercial
?C3 products and that such organisms are widely distributed in
the environment. It is also evident that rates of biodegradation
are related to both the degree of chlorination of the. biphenyl
structure and the positioning of those chlorines on the biphenyl
rings.
There is no evidence for biodegradation of PCBs under
anaerobic conditions and this might be quite important given the
high degree of sorption to solids and the likelihood that many of
those solids will reside in the environment under anaerobic
conditions.
"lith respect to the degree of chlorination, as a broad
generalization biodegradation proceeds more slowly as more
chlorines are added to the biphenyl. Given the information in
section III on biodegradation rates, it is possible to arrive at
some general conclusions about potential biodegradation rates in
various environments, but it must be emphasized that these are
9-1
-------
broad generalizations and that half lives in particular
environments for specific chlorinated biphenyls may be much
larger due to certain limiting environmental variables (e.g. low
temperatures, low moisture, pH extremes) or the specific PC3
structure.
Half Lives Resulting from Biodegradation
Aerobic
Surface Waters
Fresh
Oceanic
Activated Sludge
Soil
Anaerobic
Mono- & dichloro
Trichloro Tetrachloro
Pentachloro
and hicher
2-4 days 5-40 days 1 wk-2+mos. >1 year
several months — ——— >]_ year
1-2 days 2-3 days 3-5 days
6-10 days 12-30 days ' >1 year
12-30 days
oo
*• It is not clear how long the highly chlorinated PCBs would last
under activated sludge treatment but there appears to be no
significant biodegradation during typical residence times.
These half-life approximations also must be tempered by the
knowledge that positioning of the chlorine atoms on the biphenyl
rings can be important. It has been shown, for example, that
(1) PCBs containing all of the chlorines on one ring are degraded
faster than PCBs containing all of the chlorines distributed over
both rings, (2) PCBs containing chlorine on 2 or more ortho
9-2
-------
positions degrade very slowly, (3) the resistance of tetrachloro
?C3s is jraater,wnen there are 2 chlorines on each ring and, (4)
the initial bicdegradation step occurs on the biphenyl ring with
the fewest chlorines.
Biodegradation possibilities are also complicated by the
face that much of the PCBs released to the environment is likely
to become tightly bound to sewage solids and sediments that are
under anaerobic conditions where biodegradation will not be
significant and by the possibility that a substantial portion of
the more highly-chlorinated, less water-soluble PCBs will
evaporate into the atmosphere.
II. MONITORING EVIDENCE
Among the most convincing evidence for the persistence of
the more highly chlorinated biphenyls in the environment is
actual monitoring data on various samples. There is a large
numoer of reports, mostly on samples of biota, that might be
cited and those presented here are not intended as a
comprehensive listing.
Tucker et al. (1975) noted that the PCBs generally found in
the environment are those containing 5 or more chlorine atoms; per
molecule. This is strong evidence that the less highly
chlorinated biphenyls degrade more rapidly because the less
highly chlorinated isomers constituted about 65% of all of the
?C3s manufactured.
9-3
-------
er et al. (1978) identifiea more than 3U PCBs in
-arine fish and found the ratios of ten major PCB components
(pentacnloro and higner) in the fish were the same as the ratios
of those congeners found in Aroclor 1254 and Aroclor 1260. They
speculated that this is true because the differences in the
degradation rates of these highly chlorinated PCBs are too small
to produce any changes in their relative occurrence during the
time that they have been in the environment (up to 40 years).
tfszolek et al. (1979) analyzed lake trout in 1970 and again
in 1978, from the same lake. They found PCBs similar to Aroclor
1254, but with a higher proportion of more chlorinated congeners
at about 13 ppm at both sampling times. Moein et al. (1976)
reported no reduction in Aroclor 1254 concentration over a 2-year
period in a soil contaminated by a spill of transformer fluid.
III. 8IOOEGRADATION RATES
A. General
Biodegradation studies that report rates of degradation come
in many sizes and shapes. Some used commercial mixtures, some
pure congeners and some used both. PCS concentrations employed
range from 5 ug/1 up to 500 mg/1. Many analyzed only for the
disappearance of parent compound(s), some for potential
intermediates, and a. few for mineralization to CL>2 and water.
Studies have been conducted using lake water, sea water, soils
and sewage, as well as various synthetic media, with a variety of
9-4
-------
microbes and culture conditions. This hodge-podge of approaches
-nakes it difficult to compare results and leads to skepticism
ibout so Tie of the procedures and conclusions. Nevertheless, it
is possible to arrive at some conclusions with respect to
biodegradation rates. Those conclusions, presented below, are
tiosti/ based on studies that used mixed microbial populations
obtained from the environment and not those that employed pure
cultures. The pure culture work is discussed in section V.
8. Anaerobic
Biodegradation of the PC3s under anaerobic conditions is
probably very slow or nonexistent. This is discussed in more
detail in section IV.
C. Aerobic Aquatic
Biodeyradation of mono-, di- and trichlorinated biphenyls is
probably moderately fast in most surface waters. Clark et al.
(1979), using bacteria isolated from soils in shake flask
cultures, found 10U% primary biodegradation (loss of parent
compound) in less than 5 days for monochlorooiphenyl and 9U to
9y% degradation of dichlorobiphenyl, 42 to 87% degradation of
trichlorobiphenyls and 6 to 61% degradation of tetrachlorobiphenyls
after 5 days incubation. They also examined the biodegradation
of Aroclor 1242 and presented data on the percent biodeyradation
of 33 congeners (identified by gc peaks) showing very substantial
loss of most of the mono-, di- and trichlorinated biphenyls
.9-5
-------
vitnin a -id-hour incubation tine. Some of tne trichloro isotners
did not aegrade rapidly and these may be isomers with chlorines
in trie ortho positions. Baxter et al. (1975), in shake flask
studies, found that most dichlorinated biphenyls had half lives
of less than 10 days and the trichlorobiphenyls were half gone in
2'J to 40 days. Wong and Kaiser (1975), using lake water in
stoppered shake flasks found that Aroclor 1221 was degraded
completely, in about one month, to lower molecular weight
metabolites. They also demonstrated that bipnenyl degraded
faster than 2-chlorobiphenyl which, in turn, degraded faster than
4-chlorobiphenyl. Shiaris and Sayler (1982), however, have shown
that the biodegradation of the lesser chlorinated biphenyls in
natural waters may lead to the accumulation of chlorobenzoic acid
transformation products.
While the evidence is that the less chlorinated biphenyls
degrade readily in aerobic freshwater, the same may not be true
for seawater. Carey and Harvey (1978) found very low rates of
biodegradation of 2,5,2'-trichlorobiphenyl with only 1 to 4% loss
after 25 days in stoppered shake flasks. And Reichardt et al.
(1981), using closed bottles of seawater at 10°, estimated the
half-lives of biphenyl, 2-chlorobiphenyl, 3-chlorobiphenyl and
4-chlorobiphenyl. The biphenyl half-life, in their seawater, was
about 3 months, and that for the monochlorobiphenyls was about 8
months. Observations of considerably slower biodegradations in
the oceans are not confined to biphenyls and may be related to
the low concentrations in seawater of certain essential elements,
»
especially nitrogen.
9-6
-------
•v'ith respect to those ?C3s with 5 or more chlorines, it
appears that biodegradation is very slow in all environments
including surface waters. Oloffs et al. (1972) incubated Aroclor
1260 in river water for up to 12 weeks and found no evidence of
biodegradation. They did, however, observe significant losses by
evaporation. Shiaris et al. (1980) used reservoir water and
found no apparent biodegradation of Aroclor 1254 during 2 months
incubation. They used sealed vessels and did observe that
significant amounts of the Aroclor sorbed tightly to the glass
vessel walls and to the suspended solids, 'only about 20% of the
?C3s, initially added to a concentration of 0.1 mg/1, remain in
the aqueous phase. Wong and Kaiser (1975) compared Aroclors
1221, 1242 and 1254 and found that the microorganisms in lake
water samples could use 1221 and 1242 for growth but were not
able to utilize 1254. In contrast to most reports, Sayler et al.
(1977), using a pure culture of Pseudomonas sp., reported 70 to
35% oiodegradation of 2,4,5,2,'4,',5'-hexachlorobipnenyl in 10 to
15 days. This finding does not seem to be consistent with the
evidence from other studies.
There is little information on the biodeyradation of
tetrachlorobiphenyls in surface waters. In other media the
t^trachloro congeners on the average, have biodegradation rates
that are intermediate to those with fewer chlorines, most of
which degrade quite readily, and those with 5 or more chlorines,
which are quite persistent. The data presented by Clark et al.
(1979) show that most of the tetracnloro congeners did not
degrade significantly in 43 hours in shake flasksi The work by
9-7
-------
Carey and Harvey (1978) included 2,5,2',5'-tetracnlorobiphenyl
and they found little, if any, biodegradacion after 25 days in
seawacer. Given the results of studies using other media, it is
prooably safe to assume that the biodegradation rates of the
tetrachloro congeners are highly dependent upon the positioning
of the chlorines on the biphenyl rings.
D. Biological ^aste Treatment
Studies using activated sludge or sewage organisms and
simulating biological waste treatment processes have shown that
the biodegradation rates of PCBs are dependent upon both the
degree of chlorination and the location of the chlorines. As
might be expected, however, the rates of biodegradation, for the
readily biodegradable PCBs, are higher under waste treatment
conditions than in surface waters or soil.
Tucker et al. (1975) studied primary biodegradation by
activated sludge microorganisms using Aroclors 1U16, 1221, 1242
and 1254 plus a non-coramercial mixture, MCS-1043, containing
about 30% cnlorine. After several weeks of acclimation, the PCBs
were tested in semi-continuous activated sludge units operated on
two 43-hour and one 72-hour cycles per week. They observad 1UO%
biodegradation of biphenyl, 80% for 1221, 55% for MCS-1043, 35%
for 1016, 25% for 1242 and 15% for 1254 in 48 hours. They also
claim no significant losses of 1221, 1043 or 1016 through
9-8
-------
volatilization. Zitco (1979) noted that the data of Tucker
et al. (1975) show a decreasing rate of biodegradation with
increasing chlorine content that has the following relationship-
R = 106 (± 6) - 1.7 (± 0.2)0
wnere R = % degraded in 48 hours and D = % chlorine
The evidence, however, indicates that those PCBs with 5 or
more chlorines degrade too slowly to allow any practical
application of that relationship to them. Also, it must be noted
that individual tri- and tetrachloro congeners degrade at rates
that are highly dependent upon the location of the chlorines on
the biphenyl rings.
In contrast to the work of Tucker et al. (1975), Kaneko
et al. (1976), using semi-continuous activated sludge units,
following 3 months of acclimation, found no biodegradation of
;
-------
sludge solids. However, they used relatively short test periods
DC 6 hours and there* is no discussion of prior acclimation, which
-".ay be important.
Tulp et al. (1978) described the use of activated sludge
inocula in shake flask and PCBs at 5U rag/1 (well above the water
solubility). Some of the flasks were supplemented with 500 mg/1
additions of other carbon sources such as glucose, peptone or
hu-nic acid. They reported that the microbial populations did not
degrade any of the PCBs during 14 days of incubation when there
were three or more chlorines on the biphenyl rings. They also
noted that the additions of other carbon sources dramatically
reduced the biodegradation of 4,4'-dichlorobiphenyl. There are
too few details on their experimental work to permit a good
evaluation of their findings, but it is interesting to note that
their test PCBs with three or more chlorines were the 2,4'5-
trichloro-, 2,2',5,5'-tetrachloro-, 2,2',3,4,5,5'-hexachloro- and
decachlorobiphenyls. Other evidence (Furukawa et al. 1978b)
shows that those congeners with chlorines in any two ortho
positions are degraded poorly.
Liu (1981), using a bench-scale ferraentor and sewage
inoculum, found that the half-life of Aroclor 1221 was highly
dependent on the rate of mixing in the fermentor. His data show
that the half-life was a logarithmic function of impellor speed
between 0 and 800 rpra. At the top speeas, tne half-life of 1221
was aoouc 2 days. Liu also claimed no more tnan 10% loss of
Aroclor 1221 tnrough volatilization, over a 10-day test period.
9-10
-------
£. soil
As in water and sewage sludge, the biodegradation'of PCBs in
most soils appears to be rapid for the less chlorinated ones and
increasingly difficult with increasing chlorination. In their
review on the fate of PCBs, Pal et al. (1980) categorized
decomposition rates in soils in three groups. Group 1 is for
chlorinated biphenyls with 2 or fewer chlorines per molecule and
Baxter et al. (1975) have shown that these degrade rapidly with
half-lives of about 8 days. The second group contains the tri-
and tetrachloro PCBs which have half-lives of 12 to 30 days. The
third group, those with 5 or raore chlorines, have half-lives in
excess of one year. \s with the biodegradation of any chemical
in soils, biodegradation rates will vary greatly and depend upon
the nature and viability of the raicrobial populations, the
presence of other degradable organic matter, the moisture and
oxygen content of the soils, pti, temperature and other
environmental variables.
Griffin et al. (1978) also described the fate of PCBs in
soils but the section on biodegradation is not easy to follow and
presents data indicating a high percentage of biodegradation of
tetrachloro PCS congeners (up to 99%) in only 20 hours. This
seems unlikely. Fries and Marrow (1982), on the other hand,
reported that only about 20% of labelled biphenyl and
monochlorobiphenyls could be accounted for as 14CO2 after 98 days
in silt loam soil.
9-11
-------
IV. ANAEROBIC BIODEGRADATION
There is no evidence in the Literature that the PCBs are
degraded by microorganisms under anaerobic conditions. Carey and
Harvey (1978), Kaneko et al. (1976) and Pal et al. (1980) discuss
anaerobic studies with PCBs, and there is no indication of
anaerobic biodegradation. This seems somewhat surprising since
dehalogenation is a commonly observed reaction for other organics
under anaerobic conditions, for example with DDT and heptachlor.
On the other hand/ when DDT is transformed anaerobically to DDE
or ODD, the dehalogenation removes only one chlorine from the
ethane group and the products are more stable than the original
DDT. It may be that, even if there is some reductive
denalogenation with PCBs, the transformation product would be
very stable and that the investigations conducted to date have
not looked for those kinds of transformations. At any rate, the
•
resistance to biodegradation under anaerobic conditions is
probably quite significant. It is likely that much of the PCBs
released to the environment are rapidly bound to particulate
matter and stored under anaerobic conditions in sewage sludges
and sediments. Mclntyre et al. (19815) found that about 33% of
Aroclor 1260 found in digested sludge was retained on that sludge
after chemical conditioning and dewatering steps and, as
discussed in section VI, there is considerable evidence that PCBs
can sorb rapidly to the sludge solids in sewage and waste
treatment plants. Overall, it appears from the evidence that an
important fraction of the PC3s released to the environment will
9-12
-------
3econe tigntly bound to particulate matter in sewage sludges, in
sediments and in flooded soils, where anaerooic conditions will
.prevent: or greatly slow degradation by microorganisms.
V. PURE CULTURE STUDIES
Much of the literature on the biodegradation of PCBs
describes studies made using pure cultures of microorganisms.
Those studies are of considerable value in elucidating the
potential pathways of biodegradation and the relative rates'of
Die-degradation of various isomers. They do not provide much of
value in assessing the environmental biodegradation rates of
specific congeners.
Lunt and Evans (1970), using pure cultures of gram-negative
bacteria/ described the transforttiaton of biphenyl into
2,3-dihydroxybiphenyl. Ahmed and Focht (1973) subsequently
demonstrated the biodegradation of 3-chloro-, 4-chloro-/
2,2' dichloro- and 4,4'-dichlorobiphenyl by Achromobacter sp. ana
proposed a hypothetical pathway going from the PCS to a
dihydroxychlorobiphenyl followed by ring opening and degradation
to chlorinated benzoic acids. Other pure culture studies
("urukawa and Matsumura 1976, Furukawa et al. 1978a, Yagi and
Sudo 1980, Ballschmiter 1977, Ohmori et al. 1973, and Wallnofer
et al. 1973) tend to confirm this general pathway and have
supplied additional details. The potential pathways of aerobic
biodegradation are not very relevant to this review and will not
oe described in a«y detail. It does appear, however, that the
9-13
-------
jihydrox/lation requires oxygen and occurs on ti\e less
chlorinated ring when there is uneven distribution of the
cniorines. It -nay be that the appearance of chlorines in two or
•nore ortho positions sterically hinders the dihydroxylation
step. It also seems tnat the dihydroxylation may be accomplished
by an electrophilic forra of oxygen and that the electron-
withdrawing nature of the chlorines supresses that initial
biodegradation step.
Pure culture studies have also helped in demonstrating not
only that increasing levels of chlorine lead, in general, to
decreasing rates of biodegradation, but also that the
biodegradability is influenced by the positioning of the
chlorines on the biphenyl rings. Purukawa et al. (1978b) studied
31 PC3 congeners and demonstrated that (1) the resistance of
tetrachloro PCBs is greater when there are cwo chlorines on. each
ring, (2) PCBs containing chlorine on 2 or more ortho positions
(of either ring or both) are very resistant, (3) PCBs containing
all of the chlorines on one ring are degraded faster than those
with the same number of chlorines distributed over both rings,
and (4) hydroxylation and ring fission occur preferentially on
the biphenyl ring with the fewest chlorines. Furukawa et al.
(1979), using Alcaligenes and Acinetobacter sp., have also shown
chat the positioning of chlorines has an effect on the-metabolic
pathways and kinds of degradation products formed. Liu (1982),
using a Pseudomonas species in a closed fermentor, also obsorvaa
9-14
-------
tnat the number of chlorines and the position of the chlorines on
trie rinys are important factors in the relative biodegradation
races of ?Cas.
VI. SORPTION
The preceding sections on biodegradation in various
environments are complicated by the fact that a large proportion
of the PCBs released to the environment will sorb tightly to the
surfaces of sewage solids, suspended matter in surface waters and
various sediments and may not be available for degradation by
microorganisms in sewage treatment plants or in natural waters.
While the subject of adsorption of PCHs is covered in detail
elsewhere in this review, it is important to keep this phenomenon
in mind when considering biodegradation potential and to consider
some of the findings of those who were primarily investigating
biodegradation.
Furukawa et al. (1978a), Bourquin and Cassidy (1973),
Gresshoff et al. (1977), Kaneko et al. (1976) and Mclntyre et al.
(1981b) all noted, in connection with microbial studies, the
highly sorptive nature of the PCBs. Gresshoff et al. (1977)
speculated that much of the PCBs in the environment would tend to
adsorb to rocks or sand or soil surfaces and to resistant
organisms. They go on to suggest that those sorbed PCBs might be
remobilized by other organic pollutants, such as oil spills.
Colwell and Sayler (1977) noted that PCBs in the environment will
be partitioned into suspended sediments, oils and surface
9-15
-------
films. Studies at sewage treatment plants have shewn that PCBs
are principally removed form wastewater during sludge settling
steps. Mclntyre et ai. (1981b) demonstrated that PCBs will be
closely associated with the settled solids in sewage treatment
and that those PCBs will be retained on the solids during
chemical conditioning and dewatering steps. Fifty percent or
more of the PCBs in raw sewage may be associated with the solids
removed in primary sedimentation. (Mclntyre et al. 1981a,
Garcia-Gutierrez et al. 1982). Shiaris et al. (1980), in a study
on extraction techniques for residual PCBs, found that when 0.1
mg/1 preparations of Aroclor 1254 were incubated for 4 to 8
weeks, most of the PCB became tightly bound to vessel walls and
particulates in the water. Marinucci and Bartha, in a study on
the accumulation of Aroclor 1242 in percolators containing a
shredded marshgrass (Spartina sp.) demonstrated that PCB
accumulation in the litter was significantly enhanced by the
presence of litter - decaying microbes and concluded that a
significant fraction of the PCB in the litter was contained in
the microbiota.
VII. VOLATILIZATION LOSSES
There are many questions that come to mind when reviewing
che literature on PCB biodegradation. An important one concerns
the possibility that losses due to volatilization may have been
reported as losses due to biodegradation. Baxter et al. (1975)
and Tucker et al. (1975) assert that their studies contained
checks on volatilization losses and that there were no
9-15
-------
signicicanc Losses co the air. Liu (1981) demonstrated no more
than 10% loss of Aroclor 1221 due to volatilization during 10
.-lays of stirring in a bench-top ferraentor. It should be noted,
however, that their claims for no significant volatilization
losses are limited to the less-chlorinated, more water-soluble
PC3s. On the other hand, Kaneko et al. (1976) and Oloffs (1972)
reported very high levels of evaporative losses of Kanechlor-500
and Aroclor 1260 in their studies. Many of the biodegradation
studies in the literature (e.g. Furukawa et al. 1978b, Sayler
et al. 1977, Tulp et al. 1978) are not described in sufficient
detail to allow the reader to determine whether or not the
investigators accounted for potential evaporative losses.
VIII. OTHER FACTORS
There are several other factors in the literature on PCB
biodegradation that raise questions about the validity of certain
studies or present the reviewer with conflicting conclusions.
Most of them do not alter significantly the general conclusions
presented above, but they should be kept in mind by those who
might attempt to obtain better evaluations of biodegradation
possibilities or more reliable rate predictions.
Among the more interesting factors is the indication that
PCSs can affect the metabolic processes of microorganisms.
Kaneko et al. (1976) and Wong and Kaiser (1975) reported the
stimulation of microbial respiration by PCBs at concentrations as
low as 1 ug/1. Kaneko et al. (1976) suggested that the PCBs
9-17
-------
mi-^nt act as uncouolers o£ oxidative phosphorylation. These
findinys cast some doubt on the work of others who used
respirometrie techniques to study ?C3 biodegradation (e.g. Ahmed
and Focht 1973 and Sayler et al. 1977).
Other unresolved factors include the influence of other
degradable organic matter. Clark et al. (1979) and Yagi and Sudo
(1980) found that PC3s degraded better in the presence of other
substrates (acetate, meat extract or peptone). Tulp et al.
(1978), on the other hand, found that the presence of other
carbon sources (glucose, peptone, glycerol, yeast extract or
huntic acid) led to dramatically reduced biodegradation. Some
researchers have observed faster biodegradation by. pure cultures
than oy mixed cultures (Tulp et al. 1978 and Sayler et al. 1977)
while others have noted the opposite (Clark et al. 1979). Liu
(1981) reported the isolation of a Pseudomonas sp. that degraded
Aroclor 1221 ten times faster than sewage organisms, and he
proposed the use of that strain to seed biological treatment
plants.
One area that needs to be investigated more fully is the
role that acclimation may play in enhancing the rate and extent
of biodegradation of those PCBs that are relatively
biodegradable. It would also be very interesting to investigate
anaerobic processes more fully to find out if reductive
cecniorinations do occur and what would happen to PCBs in
environments exposed to alternating aerobic and anaerobic
conditions.
9-13
-------
[X. REFERENCES
Ahmed M, Focht DD. 1973. Degradation of polychlorinated
ciphenyls by two species of Achromobacter. Can J Microbiol
L9~:47-52.
Ballschmiter K, Unglert -Ch, Neu H J. 1977 Abbau von chlorierten
aromaten: mikrobiologischer abbau der polychlorierten biphenyle
(?C3). III. Chlorierte Benzoesauren als metabolite der PCB.
Chemosphere 1:51-56.
Ballschmiter K, Zell M, Meu HJ. 1978. Persistence of PCB's in
the ecosphere: will some PCB-components "never" degrade?
Chemosphere 2:173-176.
Baxter RA, Gilbert PS, Lidgett RA et al. 1975. The degradation
of polychlorinated biphenyls by micro-organisms. Sci of Total
Environ 4:53-61.
Bouquin AW, Cassidy S. 1975. Effect of polychlorinated biphenyl
formulations on the growth of estuarine bacteria. Appl Microtaiol
29:125-127.
Carey AE, Harvey GR. 1978. Metabolism of polychlorinated
biphenyls by marine bacteria. Bull Environ Contam Toxicol
20": 527-534.
Clark RR, Chian ESK, Griffin RA. 1979. Degradation of
Polychlorinated Biphenyls by mixed microbial cultures. Appl
Environ Microbiol 37:680-635.
Colwell R, Sayler G. 1977. Effects and interactions of
polychlorinated biphenyl (PCB) with estuarine microorganisms and
shellfish. SPA-600/3-77-070.
Fries GF, Marrow GS. 1982. Metabolism of chlorinated biphenyls
in soil. Paper presented at third annual meeting of the Society
for Environmental Toxicology and Chemistry, Arlington VA,
Mov. 14-17, 1982.
curukawa X, Matsumura F. 1976. Microbial metabolism of
polychlorinated biphenyls. Studies on the relative degradability
of polychlorinated biphenyl components by Alkaligenes sp. J Agric
Food Chem 24:251-256.
Furukawa K, Matsumura F, Tonomura K. 1978a. Alcaligenes and
Acinetobacter strains capable of degrading polychlorinated
bipnenyls. Agric Biol Chem 42:543-548.
Furukawa '<, Tonomura K, Kamibayashi A. 1978b. Effect of
chlorine substitution on the biodegradability of polychlorinated
biphenyls. Appl Environ Microbiol 35:223-227.
9-19
-------
Furu.owa K, Tomizura M, Kamibayashi A. 1979. Effect of chlorine
substitution on bacterial metabolism of various polychlorinated
oipnenyls. Appi Environ Microbiol 38:301-310.
Garcia-Gutierrez A, Mclntyre AE, ?erry R ec al. 1982. The
benavior of polychlorinated biphenyls in the primary
sedimentation process of sewage treatment: a pilot plant
study. Sci of Total Environ 22:243-252.
Grasshoff PM, Mahanty HK, Gartner E. 1977. "ate of
polychlorinated biphenyl (Aroclor 1242) in an experimental study
and its significance to the natural environment. Bull Environ
•Contam Toxicol 17:686-691.
Griffin R, Clark R, Lee M et al. 1973. Disposal and removal of
polychlorinated biphenyls in soil. In: Land Disposal of
Hazardous Wastes. EPA-600/9-78-016, pp. 169-181.
rierost E, Scheunert I, Klein W et al. 1977. Fate of PC3s-^4C in
sewage treatment - Laboratory experiments with activated
sludge. Chemosphere 11:725-730.
Kaneko M, Morimoto K, Nambu S. 1976. The response of activated
sludge to a oolychlorinated biphenyl (KC-500). Water Res
10:157-163.
Liu D. 1981. Biodegradation of Aroclor 1221 type PC3s in Sewage
Wastewater. Bull Environ Contain Toxicol 27:695-702.
Liu D. 1982. Assessment of continuous biodegradation of
commercial PCS formulations. Bull Environ'Contam Toxicol
29:200-207.
Lunt D, Evans WC. 1970. The microbial metabolism of biphenyl.
aiochera J 118:54-55.
Marinucci AC/ BArtha R. 1982. Biomagnification of Aroclor 1242
in decoraoosing Spartina litter. Appl Environ Microbiol
44:669-677.
Mclntyre AE, Perry R, Lester JN. 1981a. The behaviour of
polychlorinated biphenyls and organochlorine insecticides in
primary mechanical wastewater treatment. Environ Pollution
2:223-233.
Mclntyre AS, Lester JN, Perry R. 1981b. The influence of
chemical conditioning and dewatering on the distribution of
polychlorinated bipenyls and organochlorine insecticides in
sewage sludges. Environ Pollution 2:309-320.
9-20
-------
Moein GJ, Smith AJ Jr, Stewart ML. 1976. Follow-up study of the
distribution and fate of PC3s and benzenes in soil and
groundwater samples after an accidental spill of transformer
fluid. In: Prcc of 1976 Mat'l Conf on Control of Haz Mat'l
Spills. In f orma t ion ^Transfer Inc. Rockville, MD. pp. 363-372.
Ohraori T, Ikai T, Minoda Y et al. 1973. Utilization of
oolyphenyl and polyphenyl-relatsd compounds by microorganisms.
Agr Biol Chem 37:1599-1605.
Oloffs PC, Albright LJ, Szeto SY. 1972. Fate and behavior of
five chlorinated hydrocarbons in three natural waters. Can J
Microbiol 18:1373-1398.
Pal D, Weben JB, Overcash MR. 1980. Fate of Polychlorinated
oiphenyls (?C3s) in soil-plant systems. In: Residue Reviews,
vol. 74, Springer-Verlag New York Inc. pp. 52-69.
Reichardt PB, Chadwick 8L, Cole MA et al. 1981. Kinetic study
of the biodegradation of biphenyl and its monochlorinated
analogues bv a mixed marine microbial community. Environ Sci
Technol 15:75-79.
Sayler GS, Shon M, Colwell RR. 1977. Growth of an estuarine
Pseudomonas sp. in polychlorinated biphenyl. Microbiol Ecology
3:241-255.
Shiaris MP, Sherril TW, Sayler GS. 1980. Tenax-GC extraction
technique for residual polychloriaated biphenyl and polyarotnatic
'hydrocarbon analysis in biodegradation assays. Appl Environ
Microbiol 39:165-171.
Shiaris MP, Sayler GS. 1982. Biotransfomation of PCB by
natural assemblages of freshwater microorganisms. Environ Sci
Tecnnol 16:367-369.
Tucker ES, Saeger VW, Hicks 0. 1975. Activated sludge primary
biodegradation of polychlorinated biphenyls. Bull Environ Contam
Toxicol 14:705-713.
Tulp MTh;M, Schmitz R, Hutzinger 0. 1978. The bacterial
metabolism of 4,4'-dichlorobiphenyl and its suppression by
alternative carbon sources. Chemosphere 1:103-108.
Wallnofer PR, Engelhardt G, Safe S et al. 1973. Microbial
hydroxylation of 4-chlorobiphenyl and 4,4'-dichlorobiphenyl.
Chemospnere 2:69-72.
Wong PTS, Kaiser KLE. 1975-. Bacterial degradation of
polychlorinated biphenyls. II. Rate studies. Bull Environ
Contain Toxicol 12:249-256.
9-21
-------
wszolek PC, Lisk DT, Wachs T et al. 1979. Persistence of
co lychlorinated biphenyls and 1,1-bis (p-chlorophenyl) etaylene
(p,p'-DDE) with age in lake trout after 3 years. Environ Sci
Technol 13:1269-1271.
0, Sudo R. 1980. Degradation of polychlorinated biphenyls
by microorganisms. J Water Poll Cont Fed 52:1035-1043.
Z i tko V. 1979. Role of biodegradabili ty in the environmental
evaluation of polychlorinated biphenyls and chemicals in
general. In: Biotransformat ions and fate of chemicals in the
Environment. Maki AW, Dickson KL, Cairns, JC Jr, eds.
Washington DC: American Society for Microbiology, pp 120-125.
9-22
-------
-101
REPORT DOCUMENTATION
PAGE
l._ REPORT NO.
0/5-33-025
3. Recipient's Accession No.
I «. T.tte and Subtitle
Environmental Transport and Transformation of Polychlorinated
Bichenvls
i. Report Date
December 1983
ifer, Pcbert H. Brink, Gary C. Than, and
T11 PT-
4. Performing Organization Reot. No.
9. Performing Organization Nam* ana Address
U.S. Environmental Protection Agency
Office of Pesticides and Toxic Substances
401 M Street, S.W. .
Washington, D.C. 20460
10. Proiect/Task/Work Unit No.
II. ContraeUQ or Gr*nt(G> No.
(0
(G)
12. Soonwring Organization Nam* and Address
U.S. Environmental Protection Agency
Office of Pesticides and Toxic Substances
401 M Street, S.W.
Washington. D.C. 20460
13. Type at Report & Period Covered
14.
IS. Supplementary Note*
It. Aestrsct (Limit: 200 words)
1
This report summarizes the environmental transport and transfonration of poly-
chlorinated biphenyls and contains nine separate chapters describing water solubility
and cctanol/water partition coefficient, vapor pressure, Henry's law constant and
volatility from water, adsorption (sorption) to soils and sediments, bioconcentration
in fish, atmospheric oxidation, hydrolysis and oxidation in water, photolysis, and
bicdegradaticnc In the preparation of each of these chapters, the emphasis has been
on obtaining experimental dkte on environmentally relevant rate constants and
equilibrium constants for these processes/properties for individual PC3 congeners
and Arochlors. If no experinental data were found, then estimation techniques
were used wherever possible to obtain values for the rate constants or equilibrium
constants for each individual congener or for groups of congeners (i.e., for mono-
chloro-, dichloro-, trichloro-, etc., biphenyls). It must be emphasized that these
estimates of rates for transport and transformation involved simplifying assumptions
and thus these data should not be regarded as precise but rather as a best estimate
based on the available data.
17. Document Analysts a. Oe«eriptan
». tder.tir.en/ooen-Eiwed Terms Toxic Substances, water solubility and octanol/water partition
coefficient of PC3s, vapor pressure of PCBs, volatilization of PCBs from water,
soil and sediment sorption of PCBs, fish bioconcentration of PCBs, photolysis of
PCBs, atmospheric oxidation of PCBs by OH radicals, hydrolysis and oxidation of
PCBs in aqueous media, bicdegradation of PCBs.
e. COSAT1 Field/Grouo
Aoeilaoillty Statement
Peleased Unlimited
19. Security Cla»i (This fteoort)
Unclassified
20. Security Class (This Pi«el
21. No. of Pices
204
22. Price
See Instructions «n Reverse
OPTIONAL FORM 272 (4-77)
(Formerly NTIS-35)
Department o4 Commerce
-------
U S Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
-------
|