FATE OF PCB'S
IN ROLLINS ENVIRONMENTAL
SERVICES PLANT FIRE
MAY 1978
CONTRACT NO. 68-02-2165
TRW.
OBFBHSB AND SPACE SYSTEMS GROUP

-------
FATE OF PCB'S
IN ROLLINS ENVIRONMENTAL
SERVICES PLANT FIRE
MAY 1978
CONTRACT NO. 68-02-2165
TRW.
DtmatAMOtMcrtYtmumouf

-------
28055-6018-RU-00
FATE OF PCB'S IN ROLLINS ENVIRONMENTAL SERVICES PLANT FIRE
May 1978
Contract No. 68-02-2165
Task 54
EPA Project Officer: F. E. Briden
Prepared for
Industrial and Environmental Research Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Prepared by
M. K. O'Rell
Program Manager

-------
TABLE OF CONTENTS
Page
1.	Executive Summary	1
2.	Introduction	2
3.	Background	5
3.1	Description of the Accident	5
3.2	Structure of PCB's	8
3.3	Stereochemistry of Chlorinated Biphenyls	11
3.4	Photochemistry of PCB's	12
4.	Chemical Fate of PCB's	14
4.1	Formation of Dibenzofurans	15
4.1.1	Dibenzofuran Via Hydroperoxide Intermediate	15
4.1.2	Formation of Dibenzofuran Via OH* Attack	17
4.1.3	Formation of Dibenzofuran Via Oxygen
Radical Attack	17
4.2	Dibenzo-p-dt'oxin Formation from PCB's	19
4.3	Formation of Fluorene	23
4.4	Formation of Polycyclic Aromatic Hydrocarbons	23
5.	Dispersion of PCB's and Decomposition Products	28
6.	Commercial Scale Incineration of PCB's	30
7.	References	31
ii

-------
1. Executive Summary
This study indicates that under certain conditions that may have
existed at the Rollins Environmental Services fire polychlorinated •
biphenyl compounds could conceivably serve as precursors for the formation
of other environmentally undesirable compounds. The most probable class
of compounds formed are dibenzofurans. Favorable reaction pathways to
these compounds to these compounds exist under oxygen deficient combustion
conditions and low temperature. Conversely, it seems improbable that di-
benzo-p-dioxins are formed under similar conditions. If dibenzo-p-
dioxins were to be formed, they would most probably come from the di-
benzofurans that are initially generated from the PCB's. Formation of poly-
cyclic aromatic hydrocarbons (PAH) is possible via pyrolysis of PCB's
in oxygen deficient conditions that could have existed in the Rollins
fire.
1

-------
2. Introduction
This study was conducted for the Industrial and Environmental Research
Laboratory of the Environmental Protection Agency, Research Triangle Park,
North Carolina, as Task 54 on Contract No. 68-02-2165. Dr. Ronald Venezia
was the EPA Task Officer. The study was performed in the Applied Chemistry
Department of the Chemistry and Chemical Engineering Laboratory, Applied
Technology Division, TRW Defense and Space Systems Group, Redondo Beach,
California. Dr. Ray Maddalone, Head, Environmental and Process Chemistry
Section, was Program Manager and Mr. Michael O'Rell was the Task Manager.
The objective of this study was to investigate the fate of polyclilorinated
biphenyl compounds (PCB's) that were stored in a tank farm at a chemical
disposal site when a catastrophic fire struck the facility in December 1977.
The PCB's were selected for study because they are not only toxic and environ-
mentally long lived, but tliey may also serve as precursors for the formation
of new compounds wliich are even more toxic. This study attempts to define
what new products may have been formed under the conditions that existed at
the time of the accident.
Two compounds of particular concern that may have been formed are the
polychlorinated dibenzofurans and the polychlorinated dibenzo-p-dioxins. Both
of these compounds are highly toxic when partially chlorinated. Dioxins have
been shown to have pronounced fetotoxic and some teratogenic effects even in
9
1
10
9
DIBENZOFURAN
OIBENZO-P-DIOXIN
2

-------
the absence of any noticeable acute toxicity (Reference 1). The toxicity
of 2,3,7,8-tetrachlorodibenzofuran is similar to that reported for 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) (Reference 2) The toxicities of these
two compounds are given in Table I.
The formation of polycyclic aromatic hydrocarbons (PAH's) from PCB's
also is a concern because they are known carcinogens (see Table I). PAH
formation is favorable under pyrolysis (no oxidizing species present) and
oxygon deficient conditions which were expected to have existed as a result
of the fire.
The study was divided into three major elements of activity. The first
activity was dedicated to the definition of the probable conditions to which
the PCB's were exposed. The second activity addressed the question of what
new products could be formed given the conditions that were defined in the
first activity. The approach employed to define the new products was to
allow the PCB's to react with the simple radical species likely to exist in
combustion processes. Particular emphasis was placed on the formation of
dibenzofurans and dibenzo-p-dioxins. Part of this activity was also directed
to a definition of how the PCB's and reaction products may have been dispersed
in the environment during the accident. The last activity summarized results
of commercial scale incineration of PCB materials for comparison to the results
of this study.
Using this methodology it was concluded that formation of chlorinated
dibenzofurans from PCB's was highly probable during this accident. Conversely,
it was very unlikely that dibenzo-p-dioxins were formed. Conditions were
favorable for the formation of fluorenes and PAH's.
3

-------
Table I. Toxicities of Selected Compounds
Compound	Structure	Toxicity
2,3,7,8-Tetrachloro
Dibenzofuran
5yg/kg/day
chicks died within
8-15 days
(Reference 2)
2,3,7,8-Tetrachloro
dibenzo-p-dioxin
Benzo(a)Pyrene
0.6pg/kg dose
killed 50% of
guinea pigs
(Reference 1)
2yg/kg induces a
carcinogenic response
LDrn(subcutaneous,rat)
50 mq/kg
(Reference 3)
FT uorene
No evidence that fluorene
alone is carcinogenic
4

-------
3. Background
This section presents the details of the accident that occurred at
the Rollins disposal site and some aspects of the chemistry of polychlorinated
biphenyls that will provide a basis for the discussions in Section 3.
3.1 Description of the Accident
In December 1977, a series of explosions and fire occurred at a chemical
waste disposal site at Bridgeport, New Jersey, operated by Rollins Environmental
Services. The fire was in the tank farm which contained a variety of organic
compounds listed in Table II. Figures 1 and 2 present two pictures of the
disposal site after the fire which were furnished by M. Polito, Environmental
Protection Agency, Region II, Edison New Jersey. Of particular interest in
this study are the two tanks containing PCB's. One tank contained 4300 gallons
of a mixture of PCB's and aromatics (composition of aromatics and ratio of
PCB's to aromatics unknown) and the second tank contained 11,000 gallons of
a mixture of water and PCB's.
No detailed report of the accident exists at this time and it is impossible
to model with any degree of certainty the conditions to which PCB's were
exposed as a result of this accident. However, it is safe to assume that all
possible conditions conceivable during an uncontrolled fire existed during
this accident. For these conditions, temperatures may have ranged from those
associated with a "cool flame" (lower temperature limit of 275-300°C) up to
those expected in more idealized combustion processes (>1500°C). It is also
expected that conditions existed in which combustion occurred in "oxygen rich"
and in "oxygen deficient" environments. The oxygen deficient conditions would
5

-------
Tank
1
2
3
4
5
6
7
12
13
20
21
Table II. Contents of Tank Farm at Rollins Site+
Contents	Quantity,
Gallons
Waste Solvent, benzene sulfonic acid,	16,800
cresylic acid, toluene, waste nonene
Trichloroethane, methylene chloride,	15,400
nitrobenzene, phenol, methyl carbitol,
methyl cellosolve, isopropyl alcohol,
methyl alcohol, chlorobenzene, nonenes,
water
Sludge residue	NA
PCB, Aromatics	4,300
Acetone, toluene, alcohol solvent with	900
fatty acids
Sludge (Constituents unknown)	5,800
Nitrobenzene, phenol, methyl carbitol,	5,000
methyl cellosolve, isopropyl alcohol,
methanol, chlorobenzene, oils, kerosene,
ethyl acetate, ethyl alcohol waste
solvent
Otto fuel*, allyl chloride, nitrochloro-	15,000
benzene solvents
Hexane, paint solvents, 5% ethylene/	4,200
propylene polymer, Otto fuel, aniline,
ethanol
Methanol, water and BEB (composition	15,000
unknown)
DMT (dimethyl terephthalate)	13,000
6

-------
Table II. Contents of Tank Farm at Rollins Site (Cont'd)
Tank No.
Contents
Quantity, (total)
Gal Ions
22
23
24
25
PCB and water	11 ,000
Unknown	1,100
Nitrobenzene, phenol, methyl carbitol,	Unknown
methyl cellosolve, isopropyl alcohol,
methanol, chlorobenzene, waste nonene,
bromoethylbenzene
Waste fluorocarbon, fluorinated solvents, 11,500
trichloroethylene, methylene chloride
* Otto fuel - propylene glycol dinitrate 76%, Di-n-butyTsebatate 22%,
nitrodiphenyl amine 2%.
+ Information supplied by M. Polito, EPA Region II, Edison, N. J.
Figure 1. Rollins Environmental Services Plant
7

-------
Figure 2. Rollins Environmental Services and Surroundings
lead to partially combusted and pyrolysis products rather than complete
combustion products (i.e., CO^, ^0, HC1).
3.2 Structure of PCB's
PCB's are prepared by the chlorination of biphenyl. The chlorination
process produces a mixture of biphenyl compounds containing a varying number
of chlorines and the degree of chlorination (chlorine content) is dictated by
the end use of the product. It is also apparent upon examination of the
8

-------
biphenyl ring structure (see below) that a number of different isomers is
possible for each degree of chlorination. For example, monochlorobiphenyl has
three different isomers. The chlorine may be located at the 2,3 or 4 position
(positions 5 and 6 are identical to positions 2 and 3 for monosubstituted
bi phenyl).
BI PHENYL
The number and position of chlorines on the rings is important because
it determines the structure of the products that are formed from PCB's. An
extensive analytical study of PCB's was conducted by Sissons and Welti
(Reference 4) in which they separated and identified the major constituents
of various commercial PCB's marketed under the trade name Aroclor. Many
different PCB mixtures have been sold in the United States and one repre-
sentative of this family is Aroclor 1254. Aroclor 1254 contains 54% by
weight chlorine and is therefore representative of a biphenyl compound con-
taining five chlorines. Other significant chlorinated species also present
in 1254 are tetra and hexachlorobiphenyls.
In a study conducted by TRW on destroying PCB's by incineration
(Reference 5) the composition of a PCB containing an average of 3 chlorines
(42% by weight chlorine) was determined. The major constituents as shown in
Table III are dichlorobiphenyl, trichlorobiphenyl, and tetrachlorobiphenyl.
Lesser amounts of monochlorobiphenyl and pentachlorobiphenyl also were
detected. These results confirm those of Sissons and Wilte. They observed
9

-------
that the major constituents in a particular PCB mixture are the compound
representative of the degree of chlorination (e.g., pentachloro for 54% w/w
chlorine) and compounds containing one more and one less chlorine atoms
(i.e., tetrachloro and hexachloro).
Table III. Composition of PCB Containing 42% (w/w) Chlorine4
Compound
Estimated Concentration,
(percent w/w)
Bi phenyl
0.2
Monochlorobi phenyl
2
Di chlorobiphenyl
18
Trichlorobiphenyl
41
Tetrachlorobiphenyl
37
Pentachlorobi phenyl
2
+ Data from Reference 5
The major constituents of Aroclor 1254 as determined by Sissons and
Welti are shown in Table IV. Based on this work it was concluded that the
most abundant chlorine substitution occurs at 2,5; 3,4; 2,3,4; 2,3,6; 2,4,5
and 2,3,4,5. Rarely does substitution occur at 2,4,6, and infrequently at
3; 3,5 and 2,3. For hexasubstituted biphenyls, the distribution is similar
to that observed for Aroclor 1254, although 2,3,5, 2,3,4,6 and 2,3,5,6 are in
greater evidence.
10

-------
Table IV. Major PCB Constituents in Aroclor® 1254+
Number	,
of Chlorines	NMR Determined Structure'
4	2,5-2',5'
4	2,3-2',5'
4	2,5-31,41
5	2,5-21 ,31 ,61
5	2,3-2',3',6'
5	2,5-2',4',5'
5	2,4-2',4',5'
5	2,3-2',4',5'
5	2,5-2',3',4'
5	3,4-2',3',6'
5	3,4-2',4',5'
6	2,3,6-2',4',5'
6	2,4,5-2',4',5'
6	2,3,4-21,4',5'
+Data from Reference 4
^Nuclear magnetic resonance spectroscopy
3.3 Stereochemistry of Chlorinated Biphenyls
Another major consideration in determining the potential products that
could be formed from PCB's is the stereochemistry of the PCB's. The presence
of the chlorine atoms in the 2 and 2' positions have a strong influence on the
spatial relationship of the two phenyl rings to each other. Biphenyl prefers
to be a planar molecule to permit maximum orbital overlap and consequently,
maximized stabilization energy through electron delocalization. (The stabili-
zation energy of biphenyl is 83 kcal/mole). However, addition of substituents
11

-------
in the ortho positions (2,2',6,6' in numbering of biphenyl ) causes the two
phenyl rings to rotate in relationship to each other resulting in an inter-
planar angle between the rings that varies from 60° to 90°, depending on the
substituent. In particular, in 2,2'-dichlorobiphenyl, the angle is estimated
to be between 62° and 74° and based on dipole moment studies, the rings are so
inclined in their equilibrium positions as to place the chlorine atoms on the
same side of the rings (Reference 6). This configuration has been attributed
to attractive London forces between the chlorine atoms. As will be discussed
in Section 3, this preferred configuration makes it possible to remove two
chlorines from the PCB during the formation of a new product. If the chlorines
were located on opposite sides of the rings, then only one chlorine atom might
be eliminated during product formation.
3.4 Photochemistry of PCB's
The results of studies conducted on the photochemistry of PCB's give
additional insight into the certain preferred reactions that PCB's undergo.
It has been shown that all PCB compounds containing ortho (2-position) chlorines
yield products arising from the loss of these positional isomers in preference
to other isomer positions (i.e., 3 and 4 positions). The reason given for
preferential reaction of the ortho substituents is steric hindrance. Crowded
conditions created by ortho substituents results in a greater interplanar
angle (twisting) (see Section 3.3) with a subsequent decrease in its double
bond character. Products obtained in greatest yields arise from the loss
of ortho (2 position) chlorines thereby relieving bond strain. For example,
in solvents containing reactive hydrocarbons the radical formed by cleavage
of the carbon-chlorines bond at the 2 position abstracts hydrogen from the
solvent and results in dechlorination (Reference 7). Chlorines in the meta
12

-------
positions (3 and 5 positions) are cleaved in the absence of ortho chlorines.
Chlorines in the para position (4 position) were not cleaved after 20 hours
of irradiation.
13

-------
4. Chemical Fate of PCB's
Presented in this section are discussions concerning the potential
formation of new chemical compounds from PCB's as a result of the conditions
that were likely to exist during the fire and explosions. Particular emphasis
was placed on the formation of dibenzofurans and dibenzo-p-dioxins, two
extremely toxic classes of compounds. Other compounds considered in this bLudy
included fluorene and polycyclic aromatic hydrocarbons (PAH's).
To maintain the scope of this study within the allocated technical
effort, formation of new products from PCB's was limited to the following
reaction pathways:
1)	Reaction of PCB's with the species that are expected to exist
during combustion (e.g., OH', 0*, H*, CH^*) and
2)	Pyrolysis of PCB's .
No considerations were given to potential reactions which may have occurred
between the PCB's and other compounds that were present in the tank farm,
because of the limited scope of this study.
The discussions in Section 3 cited evidence that the normal isomer dis-
tribution of polychlorinated biphenyl compounds is such that there is a high
concentration of the diortho chloro isomer (2,2'-isomer). For convenience,
a general structure for PCB's will be used as shown below in which the 2,2'-isomer
is shown. It will also be recalled that Aroclor 1254 is representative of PCB s
GENERALIZED PCB STRUCTURE
14

-------
having 5 chlorine groups. Therefore, in the general structure given above
X + Y is equal to 3. The location of the three chlorines is not given in
the general structure but the preferred isomer distribution for a penta-
chlorobiphenyl was given in Section 3-
Presented below are the results of an evaluation of the products that
may have formed as a result of the accident.
4.1 Formation of Dibenzofurans
A number of similar pathways exist for the formation of dibenzofuran
from PCB's. The difference between the routes is dependent on the selection
of the attacking group.
4.1.1 Dibenzofuran via Hydroperoxide Intermediate.
This route to bibenzofurans requires fairly mild conditions which may have
existed in cool flames. However, more vigorous combustion processes are thought
to have been prevalent in the fire. In the absence of highly reactive radical
species (very unlikely in combustion processes) it would be expected that the
first step is cleavage of a carbon chlorine bond because it is the weakest
bond present. The C-Cl bond energy is 81 kcal/mole, considerably lower than
the bond energies measured for the C-H and the C-C (between the phenyl rings)
bonds which are 103 kcal/mole and 100 kcal/mole, respectively. The radical
formed has a number of reaction pathways available to it. It can recombine
with the chlorine radical to reform the original compound or it may abstract
a hydrogen radical which results in dechlorination. Alternatively, the radical
may react with a second identical radical to form a dimer. However, the
expected reaction is that of combination of the radical with oxygen to give a
radical with the general structure R00* as shown in Figure 3. This free radical
may be fairly stable or it may be merely a collision complex which breaks down
15

-------
PCB WHERE
X 4 Y 3
X	Y
CHLORINATED DIBENZOFURAN
Figure 3. Hydroperoxide Route to Chlorinated Dibenzofuran
16

-------
within the period of a single molecular vibration (picoseconds). As reported
by Fish (Reference 8), it is believed that at temperatures of about 300 to 4Q0°C
RO2* radicals are quite stable but at temperatures in excess of 450°C the
complex breaks up more rapidly. The next step, assuming the radical
remains intact, is abstraction of a hydrogen from the surrounding environ-
ment to produce a hydroperoxide. Cleavage of the 0-0 bond then occurs, a
common decomposition process of hydroperoxides. Recalling that there is
another chlorine in the same vicinity on the second ring, and if rotation
of the phenyl rings is slow relative to the 0* insertion reaction, it
is expected that the oxygen adds to the second ring forming the basic
dibenzofuran ring structure. Subsequent loss of chlorine gives the chlori-
nated dibenzofuran as shown in Figure 3.
4.1.2	Formation of Dibenzofuran via 0H° Attack
There are more likely routes to the chlorinated dibenzofurans which are
expected to predominate in combustion processes. It was mentioned at the
beginning of this section that some of the reactive species present in com-
bustion environments are 0* and OH*. Consider the formation of dibenzo-
furan from PCB and OH*. The OH* species adds to an electropositive carbon on
the biphenyl ring (carbons in the 2 and 2' positions are electropositive because
of the chlorine substitution). The resultant structure is shown in Figure 4.
The presence of the hydrogen provides an easy route for elimination of HC1 and
results in the same intermediate species proposed in the hydroperoxide route
to dibenzofurans. The remainder of the reaction pathway is the same as for
the hydroperoxide route.
4.1.3	Formation of Dibenzofuran via Oxygen Radical Attack
Similar arguments can be made for the formation of the dibenzofuran ring
system from PCB's and 0#. The oxygen radical is also a common species during
17

-------
Figure 4. Route to Chlorinated Dibenzofurans Via OH* Attack
18

-------
combustion processes. In this case the oxygen radical adds to the electro-
positive carbon in the same fashion as does *0H. Elimination of CI* gives
the same intermediate that has been proposed for the two previous mechanisms.
The reaction pathway then proceeds as has been outlined in the last two cases
to give a dibenzofuran ring system. The reaction pathway is shown in Figure 5.
It would appear based on these discussions that dibenzofuran compounds
are formed under certain combustion conditions. However, it must be assumed
that residence times are extremely short and only permit the oxygen insertion
reaction sequence to occur. Longer residence times permit additional reactions
to occur and result in destruction of the dibenzofuran system, or at a minimum,
removal of a number of the remaining chlorine atoms. (See Section 6 on Commercial
Scale Incineration of PCB's).
4.2 Dibenzo-p-dioxin Formation from PCB s
A second group of compounds that is of interest in this study is	composed
of the dioxins. A direct route for dioxin formation from PCB's is not	evident
because the sequence requires the formation of two carbon-oxygen bonds	and
cleavage of the carbon-carbon bond between the rings as shown below.
PCB
DIBENZO-P-DIOXIN
One potential route to dioxin from PCB's starts with the cleavage of
the C-C bond joining the phenyl rings. This would be a very energetic reaction
because the bond energy of the C-C bond is 100 kcal/mole. The favored
19

-------
Figure 5. Route to Chlorinated Dibenzofurans Via 0* Attack
20

-------
reaction site is at the carbon-chlorine bond as was discussed in Section
A.1.1 because the bond energy is 81 kcal/mole. Should cleavage of the C-C
bond occur in any case, the two phenyl radicals would quickly separate in
relationship to each other and recombination after reaction with oxygen
(or an oxygen containing radical) is highly unlikely. This sequence is shown
below. One potential product from the cleavage of the C-C bond is a chlorinated
phenol which could subsequently react with a second chlorinated phenol to form
a dioxin via the normal condensation route. However, it is also very
unlikely that this sequence of reactions takes place in a combustion
environment. Abstraction of chlorine remains the thermodynamically
favored reaction.
CI
21

-------
A more probable route to dioxiri is the transformation of the dibenzofuran
into dioxin via an oxygen insertion reaction. It was shown in the previous
sections that the formation of dibenzofuran can occur in a number of ways and
represents an intermediate for dioxin formation. As shown below, cleavage of
the C-C bond with oxygen radical ultimately results in a dibenzo-p-dioxin.
tor this illustration, a hydroxy! radical has been used as an example.
This route requires that dibenzofurans first be formed from PCB's in order to
obtain dioxins. The relative concentrations of dibenzo-p-dioxin to dibenzo-
furan is dependent upon the rate of the second oxygen insertion. A slow rate
for the second oxygen insertion would result in a low concentration of dioxins
relative to dibenzofuran. Conversely, a fast second reaction results in a
higher concentration of dioxin relative to dibenzofuran. However, since the
second reaction is more energetic than the first, the concentration of dioxin
from this reaction sequence is expected to be very low or nonexistant.
Based on these two proposed pathways it appears unlikely that significant
amounts of dioxins are formed.
22

-------
4.3 Formation of Fluorene
Other radical species present in a combustion environment are the hydro-
carbon fragments generated by decomposition of the fuel molecules. These
species include 'CHg, and the like. It is also possible that these
species add to the PCB's and result in another new compound. In this case,
a fluorene ring structure (see below) would result. A possible route to the
rluorenes is presented in Figure 6.
FLUORENE
In this pathway the carbon radical again adds to an electropositive
carbon atom. The radical formed then splits out a chlorine radical and in
the process it abstracts a hydrogen radical from the methyl group resulting
in HC1 and a hydrocarbon radical. The C^* radical then adds to the second
ring forming the fluorene ring system in an analogous fashion to the dibenzo-
furan formation. Elimination of a chlorine radical results in the final
chlorinated fluorene compound. The other hydrocarbon radicals are also expected
to result in similar compounds.
4.4 Formation of Polycyclic Aromatic Hydrocarbons.
The formation of polycyclic aromatic hydrocarbons is also a concern
because they are known carcinogens. These compounds are expected to be formed
by pyrolysis of PCB's. (Pyrolysis refers to thermal degradation in the absence
23

-------
FLUORENE
Figure 6. Route to Fluorene Via 'CH^ Attack
24

-------
or near absence of oxidizing species). Presented in this study are the results
obtained on a previous TRW study (Reference 9) on the thermal degradation of
chlorinated hydrocarbon species in which the TRW Chemical Analysis Program was
applied to determine not only the equilibrium product distribution from pyrolysis,
but also the secondary thermodynamically feasible reaction products. The
pyrolysis cases examined assumed the complete absence of oxidizing species and
covered the temperature range 399° to 1093°C for three cases:
Case 1: No restriction placed on and solid carbon is allowed to be formed.
Case 2: Solid carbon is not allowed to be formed.
Case 3: Both solid carbon and polynuclear aromatic hydrocarbons are not
allowed to be formed.
The reason for examining all three cases was to identify pyrolysis products
that were less thermodynamically favored but more likely to be formed from the
kinetic standpoint.
The PCB's are somewhat different in structure from the chlorinated
hydrocarbon wastes that were used for these calculations. However, the
results are informative and show support for PAH formation. The results for
Case 1, Case 2 and Case 3 are shown in Tables V, VI and VII, respectively.
Of particular interest to this study is Case 2 wherein PAH's are formed.
The major equilibrium products are CH^,	^2^4' ^6* Pyrene» benzo(a)
pyrene, benzo(e)pyrene, H£ and HC1. These PAH's are reasonably stable at
high temperatures and may survive the fire and escape into the environment.
It was also intended to use the TRW Chemical Analysis Program to
predict formation of dibenzofuran and dibenzo-p-dioxin from PCB's. This was
not feasible because the model requires accurate data describing the thermo-
chemical behavior of potential products and these were not available for
these two compounds. The data required are the enthalpy and entropy as a
function of temperature.	«

-------
Table V. Equilibrium Product Distribution from the Pyrolysis of Military
Standard Pesticide Formulations - Case 1: Solid Carbon
Allowed to be Formed
Major Equilibrium Products:
CH^, , HC1, and graphite; and CO, CO^, and
H^O for formulations containing oxygen.
Minor Equilibrium Products:
C2H4 (< 2 ppm)
(<2 ppm)
(<9 ppm)
H (<4 ppm)
Table VI. Equilibrium Product Distribution from the Pyrolysis of Military
Standard Pesticide Formulations - Case 2: Solid Carbon Not
Allowed to be Formed
Major Equilibrium Products:
CH4, C2H2, C2H4, CgH0, pyrene, benzo(a)pyrene
benzo(e)pyrene, H2, and HC1; and
CO, CO2, and H2O for formulations containing oxygen
Minor Equilibrium Products (< 500 ppm):
CH3 (< 12 ppm)
c2m6 (< ^36 PPm)
Propyne (< 28 ppm)
Allene (< 10 ppm)
Propylene (<16 ppm)
1,3- butadiene (< 2 ppm)
Toluene (< 11 ppm)
Styrene (< 2 ppm)
Naphthalene (< 189 ppm)
Fluoranthene (< 393 ppm)
Acenaphthalene (< 14 ppm)
Anthracene (< 4 ppm)
Phenanthrene (< 45 ppm)
Perylene (< 426 ppm)
Triphenylene (< 11 ppm)
Chrysene (< 20 ppm)
Coronene (<13 ppm)
H (< 3 ppm)
Methyl chloride (< 28 ppm)
Monochlorobenzene l< 2 ppm)
26

-------
Table VII. Equilibrium Product Distribution from the Pyrolysis of
Military Standard Pesticide Formulations - Case 3: Solid
Carbon and PAHs Not Allowed to be Formed
Major Equilibrium Products:
CH^, CjjH,), C2H4»	toluene, sty rone, H^, MCI, and
monochlorobenzene; arid
CO and CC^ for formulations containing oxyger
Minor Equilibrium Products (<100 ppm):
CH3 (<21 ppm)
C2H5 <<184 ppm)
Allene (<126 ppm)
Propyne (<369 ppm)
Propylene (<136 ppm)
1.2	- butadiene (<2 ppm)
1.3	- butadiene (<33 ppm)
Ethyl benzene (<142 ppm)
0 - xylene (<424 ppmt
m - xylene (< 918 ppm)
p - xylene l<413 ppm)
Propenylbenzene-cis (<12 ppm)
Propenylbenzene-trans (<14 ppm)
0 - methyl styrene <*15 ppm)
m - methyl styrene <<39 ppm)
p - methylstyrene '<21 ppm)
Cyclopentadiene (<22 ppm)
H (< 3 ppm)
Methyl chloride (<39 ppm)
H2O for formulations containing oxygen (<48 ppml
27

-------
5. Dispersion of PCB's and Decomposition Products
An additional consideration in this study is the ultimate fate of the
PCB's (unreacted) and of the chemical compounds formed from the PCB's as a
result of the accident. Basically, it is expected that the high degree of
chlorination of the PCB's and chlorinated decomposition products would cause
them to volatilize and move with the smoke rather than burn. As a result, some
of these materials may have been widely dispersed over a large surrounding area
in the smoke plume. The PAH's are relatively stable and are also expected to be
dispersed via the same pathway.
Very little information exists on the physicochemical properties of the
chlorinated products that may have been formed. All compounds are listed as
insoluble in water although concentrations in some cases may be in the ppm range.
A brief summary of other physical properties are given below.
Aroclor 1254 has a boiling point (bp) range of 365-390°C and a specific
gravity of Vl.5. Chrysene, a 4-ring PAH, possesses a melting point (mp) of
256°C and a bp of 448°C. Benzopyrene melts at 177°C and boils at 500°C and
fluorene boils at 295°C. 3-Chlorodibenzofuran has a mp of 101°C. Dioxins melt
at 300-350°C and are reported to be stable up to 800°C.
Based on these limited data, it is expected that the chlorinated sub-
stances (high density) would settle to the bottom of bodies of water and enter
the soil. Of course a finite concentration (although small) of these materials
would exist in water. The depth of penetration of the substances into the soil
is dependent upon soil compositions.
28

-------
An additional consideration in the fate of these materials is their stability
(or lack thereof) to photochemical decomposition. The PAH's are unstable under
solar radiation. For example, benzo(a)pyrene has a chemical half-life of less
than 1 day with solar radiation and several days without solar radiation
(Reference 10). The dioxins and dibenzofurans also are photochemically unstable
with half-lives of days to weeks (References 11 and 12). More recent evidence
(Reference 7) shows that even PCB's undergo some dechlorination in photochemical
experiments but the half-lives have not been measured. These data suggest that
the products formed will not persist in the environment for long periods of time
(except the PCB's).
29

-------
6. Commercial Scale Incineration of PCB's
In previous work (Reference 5) incineration tests with PCB's were conducted
at a disposal facility operated by Rollins Environmental Services, Inc., in Deer
Park, Texas. The incineration system consisted of a rotary kiln and a liquid
inaction afterburner. Number 2 oil was utilized as auxiliary fuel in the
destruction of the waste. The kiln flame temperature was 1250°C and
afterburner temperature was 1330°C. The calculated residence time was
3.2 seconds.
Samples of the combustion products were taken from the hot duct leading
from the afterburner. No PCB's were found and destruction efficiency for total
organic was calculated to 99.98%. (Destruction efficiency for total organics
compares the input rate of combined waste and aujxiliary fuel to the emitted
rate of all organic material found in the combustion samples). Thus, long
residence times (3 seconds) at temperatures in excess of 1250°C result in
complete destruction of PCB's. Conversely, the conditions required for forma-
tion of dibenzofurans and fluorenes are very short residence times and lower
temperatures (the lower minimum temperature is expected to be 350-400°C in
order to initiate the radical reactions).
30

-------
7. References
1.	Report on 2,4,5-T, a report of the Panel on Herbicides of the President's
Science Advisory Committee, USGPO, Washington, D.C., 1971.
2.	Goldstein, J.A., McKinney, J.D., Lucier, G.W., Moore, J.A., Hickman, P.,
and Bergman, H. Pharmacologist Ij5, 239 (Abstract No. 278) (1974).
3.	Cleland, J.G., and Kingsbury G.L. Multimedia Environmental Goals for
Environmental Assessment Volume II. MEG Charts and Background Information.
EPA-600/7-77-1366 U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina, 1 977.
4.	Sissons, D. and Welti, 3. Structural Identification of Polychlorinated
Biphenyls. J. Chromat., 60: 15-32, 1971.
5.	Ackerman, D.G., Clausen, J.F., Johnson, R.J., Tobias, R.F., and Zee, C.A.
Destroying Chemical Wastes in Commercial Scale Incinerators. Published
under EPA Contract No. 68-01-2966, June 1977.
6.	Weissberger, A., Sangewald, R., and Hampson, G.C. Trans. Faraday Soc., 30,
1 (1934); Hampson, G.C., and Weissberger, A. J. Am. Chem. Soc., 58, 2111
(1 936).
7.	Ruzo, L.O., Zbik, M.J., and Schuetz, R.D. J. Am. Chem. Soc., 9£ (12) 3809
(1 974).
8.	Fish, A. Angew. Chem. Internat. Edit., 7(1) 45 (1968).
9.	Shih, C.C., Tobias, R.F., Clausen, J.F., and Johnson, R.J. Thermal
Degradation of Military Standard Pesticide Formulations. Published under
Army contract DADA 17-73-C-3132, December 1974.
10.	Chemical Industry Institute of Toxicology. First Priority Chemicals.
Chemical Industry Institute of Toxicology; Annual Report. Research Triangle
Park, NC (1976).
11.	Crosby, D.G. and Wong, A.S. Science, 195, 1337 (1977).
12.	Safe, S. and Hatzinger, 0. Nature (London) 232, 641 (1971).
31

-------