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

-------