EPA-600/2-77-228
December 1977
LABORATORY EVALUATION OF HIGH-TEMPERATURE
DESTRUCTION OF POLYCHLORINATED BIPHENYLS
AND RELATED COMPOUNDS
by
D.S. Duvall
W.A. Rubey
University of Dayton Research Institute
Dayton, Ohio 45469
Grant No. R803540-01-0
Project Officer
Richard A. Carnes
Solid and Hazardous Waste Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
.A-iror.•!^tital Protection Agency
;,:brar? (5PL-16)
. oorn Street, Room 1670
., .:,.., 60604
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
11
-------
FOREWORD
The Environmental Protection Agency was created because of
increasing public and government concern about the dangers of
pollution to the health and welfare of the American people.
Noxious air, foul water, and spoiled land are tragic testimony
to the deterioration of our natural environment. The complexity
of that environment and the interplay between its components re-
quire a concentrated and integrated attack on the problem.
Research and development is that necessary first step in
problem solution and it involves defining the problem, measuring
its impact, and searching for solutions. The Municipal Environ-
mental Research Laboratory develops new and improved technology
and systems for the prevention, treatment, and management of
wastewater and solid and hazardous waste pollutant discharges
from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize the
adverse economic, social, health, and aesthetic effects of pollu-
tion. This publication is one of the products of that research;
a most vital communications link between the researcher and the
user community.
Surplus, unwanted, and highly hazardous industrial chemicals
pose safety hazards to the public and are potential sources of
environmental contamination. The study reported herein presents
thermal decomposition data for several organic chemicals based on
a specially designed laboratory technique.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
111
-------
ABSTRACT
Polychlorinated biphenyls (PCB's) are extremely stable
synthesized compounds which have reached global distribution in
the environment. These compounds have been found to be partic-
ularly dangerous to certain species and ecosystems. Accordingly,
further production of PCB's in the United States will cease by
October 31, 1977, and existing stocks of these hazardous materi-
als will eventually be destroyed. However, as there is consider-
able disparity as to what constitutes acceptable thermal disposal
criteria for PCB's, this study was undertaken to provide basic
data on the thermal destruction of these compounds.
A specialized laboratory technique incorporating a two-stage
quartz system was utilized for determining the thermal destruc-
tion properties of PCB's and related compounds. With this system,
a small sample was first converted to the gas phase, then exposed
to high-temperature destruction conditions. Critical parameters
of temperature and residence time were accurately measured. When
PCB's were exposed for one second to a series of high-temperature
air environments, it was found that initial decomposition
occurred at approximately 640°C; greater than 95% molecular de-
struction was obtained at 740°C; and 99.995% molecular destruc-
tion was found at 1000°C. Also, it was determined that PCB's
(and certain related compounds) thermally decompose to low
molecular weight products, as yet unidentified.
This report was submitted in fulfillment of Grant No.
R803540-01-0 by the University of Dayton Research Institute under
the sponsorship of the U.S. Environmental Protection Agency. This
work was done during the period August 15, 1976 through
January 15, 1977.
IV
-------
CONTENTS
FOREWORD iii
ABSTRACT iv
FIGURES vi
TABLES vii
ACKNOWLEDGEMENTS viii
I . BACKGROUND 1
II . INTRODUCTION 4
III. CONCLUSIONS 6
IV. RECOMMENDATIONS 7
V. STABILITY OF PCB' s 8
VI. RATIONALE OF LABORATORY APPROACH 10
VII. EXPERIMENTAL PROCEDURES 13
VIII. RESULTS AND DISCUSSIONS 18
REFERENCES 42
APPENDIX
I. DESCRIPTION AND OPERATION OF QUARTZ
TUBE APPARATUS 45
II. ERROR ANALYSIS OF MEAN RESIDENCE TIME
DETERMINATION 55
III. RESIDENCE TIME VARIATION 58
REFERENCES FOR APPENDICES 62
V
-------
FIGURES
Number Page
1 Residence time of gas phase molecules 12
2 Description of chemical compounds 14
3 Thermogravimetric analysis of Aroclor #1262 . . 16
4 Programmed temperature GC analysis
conditions 19
5 Scatter of thermal destruction data 21
6 Thermal destruction profiles for individual
compounds in the combined samples 22
7 Thermal destruction profiles 23
8 Chromatograms produced by Aroclor
#1221 samples 25
9 Chromatograms produced by Aroclor
#1232 samples 26
10 Chromatograms produced by Aroclor
#1242 samples 27
11 Chromatograms produced by Aroclor
#1248 samples 28
12 Chromatograms produced by Aroclor
#1254 samples 29
13 Chromatograms produced by Aroclor
#1260 samples 30
14 Chromatograms produced by Aroclor
#1262 samples 31
15 Comparison of thermal destruction of
Mirex and PCB's 32
16 Open tubular column gas chromatographic
conditions 34
17 Chromatograms of Aroclor #1242 samples .... 35
18 Chromatograms of 2,5,2',4', 5'-
Pentachlorobiphenyl samples 37
19 Chromatograms of composited Aroclor mixtures
plus 2,5,2',4',5'-Pentachlorobiphenyl 38
20 Effect of residence time at 704°C 41
VI
-------
TABLES
Number Page
I Weight Percent Remaining After One Second
Exposure to Respective Temperature 20
II High-Temperature Exposure Destruction Percentages . . 39
III Weight Percent Remaining After 704°C Exposure
for Respective Residence Times 39
VII
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance of R.A.
Carnes, the EPA project officer. Also, they wish to express
their appreciation to R.A. Grant for his help in the glassware
design and fabrication, to R.G. Keil, and B.L. Fox for their
critical reviews, and to L.G. Wallick for her technical editing
of the report.
vnx
-------
SECTION I
BACKGROUND
The original objective of this research project was to
determine, for a number of selected pesticides, the temperature
at which molecular destruction was effectively completed at a
residence time of <2 seconds. In order to achieve this objec-
tive, it was first necessary to develop the laboratory apparatus
and techniques capable of providing such data.
PYROLYSIS GAS CHROMATOGRAPHIC APPROACH
The laboratory facility for acquiring thermal destruction
data centered about a pyrolysis gas chromatography setup utiliz-
ing a Chemical Data Systems Pyroprobe 100 and a Varian 2400 gas
chromatograph. The basic approach was to deposit a microgram
size sample of a pesticide upon a uniform, platinum ribbon. The
ribbon was then placed in a chamber immediately upstream of a gas
chromatographic column assembly. After admitting air into the
chamber, the platinum ribbon and pesticide sample were subjected
to a defined thermal pulse. The thermal decomposition products
were then swept with an inert carrier gas (helium) into the
chromatographic instrument and analyzed.
A considerable amount of time and effort was spent in
developing the pyrolysis gas chromatographic apparatus. A unique
feature of this technique was that both the temperature of the
platinum ribbon and the time duration at that temperature could
be very accurately measured. In order to obtain these data, a
special electronic circuit was assembled. This circuit simulta-
neously sensed the voltage (E) across the ribbon, along with the
current (I) through it, and thereby generated an output pro-
portional to E/I. This output represented the average resistance
of the platinum ribbon during a timed interval, thereby providing
a direct measurement of the ribbon temperature.
The approach of using a pyrolysis gas chromatographic
technique, however, was not successful. While preliminary work
had indicated that the technique was promising for samples of low
volatility, more detailed work showed otherwise. In short, it
was found that the thermal degradation of a pesticide at elevated
temperatures on the platinum ribbon produced variable and incon-
sistent results. The basic problem was that the pesticide
-------
vaporized (at least partially) before it reached degradation
temperatures -- even at the extremely rapid temperature rise of
50°C per millisecond.
TWO-STAGE QUARTZ TUBE SYSTEM
At this point in the research project, the pyrolysis gas
chromatography approach was suspended, and research efforts
turned to a two-stage quartz tube system. In this system, the
sample was vaporized in a low-temperature chamber, and then
carried in a stream of flowing air into a high-temperature de-
struction zone. This system was successfully developed, and is
described in detail later in this report.
Coincidental with our developing a workable system to eval-
uate the thermal destruction properties of organic hazardous
wastes, the Kepone problem emerged. The serious environmental
problems concerning Kepone manufacturing operations in Hopewell,
Virginia, have been widely publicized. Large-scale disposal and
environmental cleanup problems were associated with Kepone found
in soil, water, sewage sludge, etc. Shortly after the severity
of this contamination problem became apparent, a Kepone Task
Force was formed within the Environmental Protection Agency (EPA).
As task force discussions progressed, it was concluded that high-
temperature controlled incineration offered the best potential
for successful disposal of this hazardous pesticide. However,
thermal degradation data for Kepone were not available, and would
have to be generated.
For this reason, the project officer requested that the
University of Dayton alter its schedule of work under Grant No.
R803540-01-0 in order to conduct a study of the thermal destruc-
tion properties of Kepone. The motivation for this work was a
request from EPA ORD KEPONE TASK FORCE pointing out the need for
thermal degradation data to assist cleanup operations of the
hazardous environmental situation caused by Kepone pollution in
Hopewell, Virginia. Accordingly, the study was designed to pro-
vide data from which to establish requirements for the thermal
disposal of Kepone. The specific objective of the laboratory
effort was to identify destruction temperature characteristics
of the vaporized pesticides at preselected residence times. A
major part of the program involved examination of the gas phase
destruction behavior of the selected pesticides for completeness
of destruction and identification of decomposition products while
in a controlled-flow, high-temperature air environment. (Note
that this system differs from a pyrolysis, or oxygen deficient
technique.)
The findings of this laboratory study clearly pointed out
that the complicated chemical nature of pesticide thermal de-
struction products is strongly dependent on the temperature and
residence time that the pesticide encounters. Therefore, it is
-------
most important to acquire precise analytical data which identify
and quantify decomposition products relative to exposure temper-
ature and residence time. The laboratory approach is ideally
suited for obtaining these basic data, which are needed to insure
logical decision making in planning the thermal disposal of
pesticides.
The Kepone work was handled as a high-intensity, short-turn-
around project. For comparison purposes, the chlorinated
pesticides Mirex and DDT were also studied at the same time. By
Spring of 1976, substantial data had been developed and a special
technical report covering the work on all three pesticides had
been submitted to the EPA Project officer for distribution to in-
volved parties. This report, EPA Report 600/2-76-299 "Laboratory
Evaluation of High Temperature Destruction of Kepone and Related
Pesticides," has formed the data base in support of pilot plant
incineration studies of Kepone and Kepone contaminated sludge.
POLYCHLORINATED BIPHENYLS (PCB's)
Following the completion of the Kepone work, a program was
initiated, at the request of the EPA Project officer, to simi-
larly study the thermal decomposition properties of polychlori-
nated biphenyls. The need for this work is well justified by
the serious disposal problems related to the widespread environ-
mental contamination of PCB's. At the present time, some 750
million pounds of PCB's are in use within the United States, with
another 750 million pounds already having entered the environ-
ment.
Over the past several years, increased attention has been
focused on PCB's as it has been established that they persist and
accumulate in the environment. PCB's, even at low part-per-
million levels, cause adverse effects in fish and other aquatic
life. Also, laboratory tests have shown PCB's to cause gastric
disorders, skin lesions, reproductive failures, and tumors in
mammals. For these reasons, the EPA now considers PCB's to be a
significant hazard to human health as well as to the environment.
The remainder of this report is concerned with the PCB
thermal destruction studies carried out at the University of
Dayton Research Institute.
-------
SECTION II
INTRODUCTION
Polychlorinated biphenyls (PCB's) were introduced
commercially in 1929. They were initially utilized as dielec-
tric fluids in transformers and capacitors; however, since that
time PCB's have found a multitude of applications. Further, it
was not until 1966 that PCB's were identified as widely distri-
buted environmental contaminants [1].
PCB mixtures have unusual and almost unique properties.
They are clear liquids possessing both thermal stability and
resistance to biological degradation. Also, as a class of
chemical compounds, PCB's are non-flammable, and have low vapor
pressure, low water solubility, high solubility in most organic
compounds, and high dielectric constants. As a result of their
unusual material properties, PCB's have been used over the past
40 years for many applications other than as dielectric fluids.
PCB's have been formulated into varnishes, waxes, synthetic
resins, epoxy and marine paints, sealants, printing inks,
textile coatings, glues, cutting oils, carbonless reproducing
paper, and protective coatings for wood, metal, and concretes.
PCB's have also been used as plasticizers, solvents for adhesives,
hydraulic fluids, flame retardants, heat exchanger fluids,
petroleum additives, and pesticide extenders.
Initially, PCB's were not considered to be very toxic. In-
deed, they demonstrated a rather low acute toxicity in animal
studies [2], However, they have since been shown to be potent
chronic toxicants [3]. PCB's readily undergo biomagnification
by successive members in a food chain, e.g., plankton to fry to
fish to birds, etc. Also, because of their resistance to bio-
degradation, PCB's persist for long periods of time in water and
biota. Related research has shown that water containing even
very low levels of PCB's (<10 ppb) inhibits the growth of shrimp,
oysters, and plankton [4,5]. Also, when PCB's are ingested in
animals and humans, they are preferentially stored in the body's
fatty tissues. This is a serious matter as recent studies [6-8]
have linked PCB exposure to liver and pancreatic cancer, along
with melanoma skin cancer. For a thorough understanding of the
wide ranging problems associated with the global distribution
of PCB's in the environment, the reader is referred to the re-
search works of Risebrough et al. [9], Gustafson [10], Jensen
-------
[11], Fishbein [12], Selikoff [13], Carnes et al. [14], Ruopp et
al. [15], and others [16-19].
Production of PCB's in the United States will cease by
October 31, 1977 [20]. Also, measures are underway to restrict
the importing of PCB's and products containing PCB's. Meanwhile,
the research continues for acceptable substitutes for PCB's
within the electrical industries [21] . In view of the per-
sistency, chronic toxicity, and structural stability of various
PCB molecules, adequate disposal of these unwanted materials
represents a challenging task. As the most likely method of
permanent disposal is through the use of specialized high-
temperature incineration, the University of Dayton Research
Institute (UDRI) was requested by the EPA Project Officer to
provide some basic thermal destruction data on PCB's. According-
ly, a laboratory investigation was undertaken to determine the
high-temperature destruction properties of PCB's.
The specific objectives of this investigation were:
(1) to determine the degree of molecular thermal destruc-
tion as a function of exposure temperature;
(2) to determine the effects of residence time at high-
temperature; and
(3) to determine whether other toxic decomposition products
might be produced as a result of the high-temperature exposure.
-------
SECTION III
CONCLUSIONS
1. PCB's are more thermally stable than Mirex — a very
thermally stable pesticide.
2. Commercial PCB's undergo decomposition in air between
640 and 740°C at a residence time of ^1 second. PCB's subjec-
ted to a very high temperature exposure (1000°C for 1 second in
air) yield a destruction percentage of 99.995.
3. Residence time is a strong factor in the high-tempera-
ture destruction of PCB's.
4. Upon thermal stressing in air, PCB's decompose to low-
molecular weight products. (These products were not identified
in this study.)
5. For a given thermal exposure, the lower molecular
weight PCB's are less thermally stable than the higher molecular
weight PCB ' s.
6. Compounds related to PCB's--biphenyl, dibenzofuran,
dibenzo-p-dioxin, and hexachlorobenzene--demonstrate thermal
destruction properties comparable to PCB mixtures.
-------
SECTION IV
RECOMMENDATIONS
This study has experimentally measured the destruction
temperature - residence time relationships for various PCB com-
pounds and mixtures of PCB's. To further clarify the thermal de-
struction properties of these particular compounds, the follow-
ing actions are recommended:
1. Determine the thermal destruction and effects of
residence time at temperatures greater than 1000°C.
2. Make complete analyses of degradation products and
effluents. (Further research is required to design and develop
a laboratory system capable of providing this information.)
3. Investigate the thermal destruction behavior in moist
air.
4. Study the flame mode of decomposition and compare re-
sults to the non-flame mode of destruction.
5. Eventually test the scale-up validity of PCB thermal
decomposition.
-------
SECTION V
STABILITY OF PCB's
The stability of various PCB compounds in the environment
has been the subject of numerous investigations. At the present
time, it is generally agreed that PCB compounds will persist in
the ecosystem for a long period of time. However, certain
mechanisms of degradation have been identified [22-25]. Various
PCB compounds can be dechlorinated by ultraviolet light, while
others are vulnerable to microbial attack. It has generally been
observed that the lower molecular weight PCB's are more readily
biodegraded, while the heavier compounds exhibited greater over-
all stability.
PCB's can withstand long-term heating (200-300°C) in air
without decomposing. This high thermal stability has led to wide
industrial applications; however, this same property also helps
account for widespread occurrence of PCB's in the ecosystem. For
example, it is well known that many PCB's have entered the
environment through inadequate incineration and indiscriminate
burning of wastes. At the present time, most of the conventional
incineration processes do not produce sufficient exposure
temperature or residence time to break down these stable com-
pounds. Consequently, the PCB's have been merely vaporized, and
thus further dispersed.
In the available literature pertaining to the thermal dis-
posal of PCB's, considerable disparity is found in the prescribed
conditions for thermal disposal. In addition, since this thermal
disposal information was gathered using a variety of incineration
equipment and methods, there are many variables which complicate
interpretation. To illustrate the broad range of specified
incineration conditions, two extreme examples are presented.
PCB's contained in wet sludges were reported to have been
destroyed in an after-burner when exposed to 593°C for a period
of 0.1 second [26]. At the other extreme, it is contended [27]
that PCB's should be exposed to 1316°C for a time duration of
2.5 seconds. There are many other specified conditions in the
literature [28-31] which fall between these extremes.
In view of the widely differing accounts of PCB thermal
stability, a thorough laboratory investigation was undertaken.
This study was designed to determine the high temperature de-
composition properties of the various PCB's and related compounds.
-------
The information obtained from these experiments will be of vital
assistance in developing future pilot scale thermal decomposition
tests.
-------
SECTION VI
RATIONALE OF LABORATORY APPROACH
In order to determine the high-temperature destruction
characteristics of an organic molecule, it is necessary to obtain
precise information on certain crucial parameters. It is impor-
tant that the system temperature be accurately determined, along
with the residence time, i.e., the time interval during which the
molecule experiences the destructive temperature. Another
critical parameter that must be fulfilled in such an evaluation
is that there be an adequate supply of air so that the destruc-
tion takes place in an oxidative environment, as opposed to an
atmosphere containing insufficient oxygen. Lastly, an important
factor is the composition of the high-temperature environment,
e.g., water content, presence of other gases, and interfering
solid materials.
The laboratory approach to establishing a material's high
temperature, non-flame destruction characteristics has certain
distinct advantages. First, one can examine the undiluted
sample; therefore, no interferences are encountered from other
materials. Second, the composition of the high-temperature en-
vironment can be precisely established by using compressed air of
known quality and employing in-line filters to remove water, oil,
and other foreign materials. Next, by using a technique whereby
a small sample is gradually vaporized and then passed through a
high-temperature zone, an excess of oxygen is assured, thus
avoiding the possibility of a pyrolytic reaction occurring.
Further, it is possible to evaluate the behavior of the pure
sample on the molecular level. By vaporizing the sample prior to
its exposure to the high-temperature environment, one can be
assured, based on the kinetic theory of gases [32], that the
molecules do indeed experience the actual average temperature.
Finally, the laboratory evaluation of a sample's destruction
characteristics can be accomplished quickly and economically with
minimum environmental risk.
It is important to note that during incineration of waste
materials, a certain amount of energy must be applied just to
change the organic compounds from their usual state (whether it
be solid or liquid) to the gas phase. These phase transitions,
apart from requiring additional energy, also require an undefined
amount of exposure time to the high-temperature source. There-
10
-------
fore, it is almost certain that in the case of waste materials
containing PCB's, some molecules do not encounter the prescribed
incineration temperature. In short, even though the temperature
within a chamber is known, it does not necessarily mean that all
substances within that chamber have indeed attained thermal
equilibrium. However, if all of the substances in the chamber
are in the gas phase, one can safely assume that the individual
molecules have (with sufficient time exposure) achieved the
average temperature of that environment.
For similar reasons, the laboratory evaluation of a sample's
residence time is considerably simplified over that of measuring
residence time in a large scale unit. When a gas is passed
through a long narrow-bore flow path (in the laminar flow region),
radial dispersion can be neglected and the main factor affecting
the residence time distribution is longitudinal diffusion. How-
ever, for flow paths that are of large diameter, mixing chamber,
or multichamber configuration, radial dispersion is the major
factor affecting the variation in residence time of the trans-
ported molecule [33]. Figure 1 depicts the residence time
relationships for gas phase compounds passing through the two
different flow paths. The exact contours of these generalized
profiles are dependent upon the nature of the flow, i.e., laminar
or turbulent, and other factors such as wall composition [34] and
surface finish. However, in general, residence time can be
accurately measured with a narrow-bore flow path, whereas for the
large diameter flow path, molecular residence time can, at best,
only be approximated. This situation can be illustrated by
observing in Figure la that some molecules can pass through the
large diameter flow path in a very short time; while some
molecules, as evidenced by the asymptotic behavior of the right-
hand side of the same curve, can remain in the high-temperature
region for an extended period of time.
The approach for obtaining high-temperature destruction data
on the PCB's utilized a discontinuous system where the thermal
stressing of the sample and product analysis are performed sepa-
rately. Using this approach, various PCB samples and related
compounds were evaluated in a series of tests where the sample
was first vaporized and then transported through a narrow-bore,
high-temperature zone by a controlled flow of air. In this way,
average temperature and residence time were firmly established.
These tests were conducted at a series of temperatures ranging
from ^300°C to 1000°C. Also, the effluent from each high-
temperature test was passed through a trapping medium and the
collected fraction was subsequently analyzed by gas chromato-
graphy.
11
-------
(a)
CONCENTRATION
RESIDENCE TIME
(b)
CONCENTRATION
RESIDENCE TIME
Figure 1. Residence time of gas phase molecules, (a) Large
diameter, mixing chamber, or multichamber flow
path. (b) Narrow-bore flow path.
12
-------
SECTION VII
EXPERIMENTAL PROCEDURES
SELECTION OF SAMPLES
The primary objective of this study was to obtain basic
experimental data pertaining to the thermal destruction prop-
erties of PCB's. To accomplish this objective, we examined
commercial PCB mixtures and pure samples of those PCB isomers
that have received previous study [35] and are widely dispersed
in the environment [36]. Accordingly, the isomers selected for
study were 2,5, 2',5,' -tetrachlorobiphenyl and 2,5,2',4",5'-
pentachlorobiphenyl. Also, to better understand the PCB thermal
destruction process, other compounds were included in this study.
Biphenyl was included as it is the skeleton member of the PCB
family. Although decachlorobiphenyl is not present in signifi-
cant quantities in the various commercial PCB mixtures, the
thermal destruction behavior of this compound was considered of
vital importance. It was anticipated that this compound would
represent the "worst case" with respect to thermal exposure
needed for destruction. Another chlorocarbon, hexachlorobenzene,
was included in this study for comparison purposes. It was
known from previous work that hexachlorobenzene possesses very
high thermal stability.
In the past, commercial mixtures of PCB's that were manu-
factured in Europe were found to contain small concentrations of
highly toxic chlorinated dibenofuran compounds [37]. Also, low
levels of chlorinated dibenzofurans were found to occur when
water solutions of certain PCB's were subjected to photolysis
[38]. Therefore a sample of the relatively low-toxicity
dibenzofuran was included for examination of thermal destruction.
This non-chlorinated compound could be easily examined with the
present laboratory system. In addition, it was hypothesized that
the thermal destruction properties of chlorinated and non-
chlorinated dibenzofuran would be closely related. (This
hypothesis will be subjected to thorough testing in the near
future.) For similar reasons, and in view of the recent interest
in the chlorinated dibenzo-p-dioxins [39-41], it was also decided
to obtain thermal destruction data on dibenzo-p-dioxin, a non-
chlorinated compound.
The specific compounds examined in this study are listed
and described in Figure 2. The biphenyl, hexachlorobenzene, di-
13
-------
COMPOUND
STRUCTURE EMPIRICAL FORMULA
Biphenyl
C12H10
2,5,2',5',-
Tetrachlorobiphenyl
C/
C/
C12H6C14
Z,5,Z',4',5',-
Pentachlorobiphenyl
Decachlorobiphenyl
C/ C/
c/ c/ c/
c/ c/ c/
C12H5C15
C12C110
Poly chlorinated
Biphenyls
Hexachlorobenzene
C/n
C/ C/
\ _ /
C/-\f )V-C/
) (
C/ C^
tA
6 6
Dibenzofuran
Dibenzo-p-dioxin
C12H8°2
Figure 2. Description of chemical compounds.
14
-------
benzofuran, dibenzo-g-dioxin, and pure PCB isomers were obtained
from Analabs, Inc., North Haven, Connecticut. The seven PCB
mixtures used in this study were Aroclor numbers 1221, 1232, 1242,
1248, 1254, 1260, and 1262, respectively. This series of Aroclor
samples was obtained from the National Environmental Research
Center, Research Triangle Park, North Carolina.
THERMAL DESTRUCTION TESTING
The temperature necessary to vaporize a PCB mixture was
determined by conducting a thermogravimetric analysis with the
heaviest of the selected PCB mixtures, specifically Aroclor #1262.
This analysis was conducted using a Fisher Series 100 thermo-
gravimetric instrument with a flow of 60 ml/min of dried air
passing through the sample region. The temperature was then
programmed from ambient to 915°C at the rate of 10°C per minute.
A thermogravimetric tracing for Aroclor #1262 is shown in
Figure 3.
Following this thermogravimetric work, quantities of each of
the selected pure compounds and commercial PCB mixtures were
weighed and dissolved in a suitable volatile solvent (90% n-
hexane + 10% acetone). Each sample was then subjected to a
series of preliminary thermal destruction tests while in flowing
dry air. These preliminary tests used the same instrumentation
and procedure as a previous study involving the high-temperature
destruction of the pesticides Kepone, Mirex, and DDT [42]. The
description of the instrumentation and associated procedures are
presented in Appendix I.
The results of the preliminary testing clearly revealed that
the entire series of samples (those described in Figure 2) pos-
sessed considerable thermal stability. Furthermore, these pre-
liminary tests showed only small differences between the thermal
destruction behavior of the least stable and the most stable of
these compounds.
In view of the similar thermal destruction behavior of these
compounds, we combined the pure compounds and examined them in
the high-temperature apparatus as a highly diluted mixture in
flowing air. By employing this combined sample technique, slight
differences in thermal destruction behavior were readily measured.
The data scatter associated with the subtle variations in thermal
exposure, as observed in the previous individual tests, was
eliminated. With this combined sample technique, each species
experienced an identical thermal exposure.
This combined sample technique required a slightly different
sample vaporization procedure. A process of step-wise pro-
grammed heating over a 10 minute interval was used to slowly
vaporize the combined sample in order of volatility. By using
this procedure, a highly diluted, multicomponent sample was
15
-------
SAMPLE 25mg of AROCLOR " 1262
i nn
1 \J W
.0 80
0^
O
Z
z 60
<
S
UJ
o:
K ^O
x
o
UJ
5
20
0
Tc= 915 °C
F/ -
/
/
/A /
75 ^C \ /
\ /X
\ /
-
—
/
/ ~~
/
/
/
/ ~
/
/
/
/
/
1 / Wtlon I
, 7
/ TEMPERATURE — i
/7
/ ,185 °C —
1 /x/
— /
/
T0 = 23 °C / x
"D^START OF —
*— TEMPERATURE PROGRAM
1 1 1 1 1
IUUU
900
800
700
600^
UJ
Ql
500^
h-
<
or
UJ
400Q-
2
UJ
t-
300
200
100
0
0 30 60 90 100 150
TIME, MIN.
Figure 3. Thermogravimetric analysis of Aroclor #1262
16
-------
passed through the high-temperature destruction region of the
quartz tube. This gradual vaporization procedure produces a
high dilution of organic compounds in the flowing air, thereby
minimizing the high-temperature interaction of sample and de-
composition products. The entire group of samples was examined
for high-temperature destruction behavior using this procedure.
RESIDENCE TIME CONSIDERATIONS
The average high-temperature exposure of a sample molecule
can be measured using the quartz-tube apparatus described in
Appendix I. Also, accurate measurements of mean residence time
can be obtained (see Appendix II) if accurate data are available
on the outlet gas volume flow, the average exposure temperature,
and the differential pressure for the quartz tube system.
Even though the mean residence time can be accurately deter-
mined, additional information is needed concerning the distri-
bution of the residence times of different molecules. The
residence time distribution of molecules can be studied in the
same manner as if they entered the quartz tube as a concentrated
narrow pulse (Appendix III). Also, the behavior of different
molecular weight substances at different temperatures has been
investigated. It was found that different molecular weights pro-
duced only a very small effect on the respective residence time
distribution; however, the residence time distribution was
strongly dependent upon temperature, as described in Appendix III.
Even so, a fairly small tolerance can be applied to the mean
residence time and still envelop the rapidly migrating molecules.
17
-------
SECTION VIII
RESULTS AND DISCUSSIONS
From the results of the preliminary high-temperature de-
struction tests, it was determined that almost no thermal break-
down occurred at temperatures below 550°C. Therefore, the second
series of tests using the slow evaporation of combined sample
compounds was started at rv550°C. It was also learned from the
earlier tests that intense thermal decomposition occurred in this
region.
DETERMINATION OF THERMAL DESTRUCTION PROFILES
A series of tests was designed to cover the approximate
range of 550°C to 800°C. These tests were conducted using a
total of 40 micrograms of combined sample and a sample vaporiza-
tion of 10 minutes for each test. By controlling the air flow
through the folded quartz tube, mean residence time in the high-
temperature zone was maintained at 1.0 + 0.1 seconds. The trap-
ped effluent products were then subjected to the programmed
temperature gas chromatographic analysis as described by the
instrumental conditions listed in Figure 4.
The results of these analyses are presented in Table I,
where it is seen that the actual temperature range covered by
this series of experiments starts at 562°C and extends to 801°C.
The data obtained from this series of tests on combined samples
were plotted on a single graph as shown in Figure 5 (note that
weight percent remaining is displayed logarithmically). The same
data were also plotted on two other graphs (Figures 6 and 7) hav-
ing an expanded temperature axis; however,the thermal destruction
profile for each compound was plotted using its respective data
points.
From Figure 6, it is interesting to observe that decachloro-
biphenyl and hexachlorobenzene have approximately the same
thermal stability. Also, from the same figure, it is seen that
biphenyl, dibenzofuran, and dibenzo-p-dioxin have somewhat
similar thermal stabilities. Figure"? shows the two extremes of
the PCB family, namely the non-chlorinated biphenyl and the
chlorine saturated decachlorobiphenyl. It is probable that the
two compounds represent the thermal stability boundaries of the
PCB family. Indeed, Figure 6 indicates that the tetrachloro-
18
-------
GAS CHROMATOGRAPHIC CONDITIONS
Instrument Tracer 550
Analyst W. Rubey
Date October 15, 1976
Column:
tubing material Pyrex
tubing length 1.8
tubing bore 3.5
meters
mm
Stationary Phase OV-1 Silicone
weight 3%
Support
mesh
Chromosorb-W, HP Grade
" 100/120
Carrier Gas
inlet pres
linear velocity
outlet flow
Chart Advance
Read Out
Sample
size
Helium
1.8
10
16
0.5
abs atmo
cm/sec
ml/min
cm/min
1.0mv full scale
PCB Compounds
40 yg total
Detector
range
HFID
10-yAFS
none
attenuation see chromatograms
Gas Flows, ml/min
hydrogen 50
air 350
split ratio
Temperatures, °C
detector 320
injector RT
column:
initial 40
final 270
program rate
Sample Solvent
concentration
14
'C/min
n-Hexane
5 UQ/ul
Figure 4. Programmed temperature GC analysis conditions
19
-------
TABLE I. WEIGHT PERCENT REMAINING AFTER ONE
SECOND EXPOSURE* TO RESPECTIVE TEMPERATURES
Compound
Biphenyl
Dibenzofuran
Dibenzo-p-dioxin
Hexachlorobenzene
Tetrachlorobiphenyl
(2, 5, 2', 5')
Pentachlorobiphenyl
(2, 5, 2', 4', 5')
Decachlorobiphenyl
562
100
100
100
100
100
100
100
590
100
100
98. 6
100
97. 2
96. 9
96. 2
T2'
629
78. 1
86.4
80.8
94. 4
83. 3
87. 5
80.7
Exposure
679
28. 8
48. 2
31.5
77. 8
55.6
57. 8
65. 4
Temperature (°C)
694
5. 0
14. 5
6. 8
55.6
33. 3
39. 1
53. 8
708 722 742 801
0.36 0.08
1. 16 0. 23
0.34
33.3 13.9 4. 2 < 0. 1
9. 0 0. 69
13.7 1.12
35.6 15.4 5. 8 < 0. I
*t = 1. 0 ± 0. 1 sec
20
-------
640°C 740°C
100
UJ
ct
UJ
o
cc
LiJ
CL
I
O
UJ
10
1.0
O.I
O BIPHENYL
O DIBENZOFURAN
O DIBENZO-p- DIOXIN
A HEXACHLOROBENZENE
t, TETRACHLOROBIPHENYL
O PENTACHLOROBIPHENYL
O DECACHLOROBIPHENYL
tr - 1.0 ±0.1 sec
0.01
I
I
I
I
0 100 200 300 400 500 600 700
EXPOSURE TEMPERATURE,°C
800
900
1000
Figure 5. Scatter of thermal destruction data,
21
-------
100
< 10
5
UJ
o:
u
ir
UJ
x
o
UJ
L°
0.1
0.01
50
-o—o-
BIPHENYL
DIBENZOFURAN
DIBENZO-p-DIOXIN
tr = 1.0 ±0.1 sec
-DECACHLOROBIPHENYL
HEXACHLOROBENZENE
500 550 600 650 700 750 8C
EXPOSURE TEMPERATURE °C
850 900 950 1000
Figure 6. Thermal destruction profiles for individual
compounds in the combined samples.
22
-------
100
10
LU
cc
o
ct
I
o
O.I
0.01
BIPHENYL
Tr = 1.0 ± O.I SBC
DECACHLOROBIPHENYL
2,5, 2',4', 5'-
PENTACHLOROBIPHENYL
2,5, 2',5'-
TETRACHLOROBIPHENYL
50
500 550 600 650 700 750
EXPOSURE TEMPERATURE , °C
800 8 50
900
950
1000
Figure 7. Thermal destruction profiles
23
-------
biphenyl has a somewhat lower thermal stability than the penta-
chlorobiphenyl, which in turn has a significantly lower thermal
stability than the decachlorobiphenyl.
EXAMINATION OF COMMERCIAL PCB MIXTURES
In Figure 5, it was observed from the distribution of the
thermal decomposition data points that a temperature zone cover-
ing these data points is surprisingly narrow. Therefore, it was
decided to assign a low-temperature threshold to signify the
initiation of thermal breakdown, and an upper temperature value
which would represent practical completion of the thermal de-
struction of PCB's (<5% remaining). The values selected for
these two temperatures were 640°C and 740°C.
Each of the commercial PCB mixtures was then examined at
these two temperatures. First, 40 microgram samples of each of
the Aroclor mixtures were analyzed by injecting onto the Tenax-
GC trapping media and conducting a programmed temperature GC
analysis. Then, samples of each of the seven mixtures were
subjected to a thermal exposure of 640°C for a one-second resi-
dence time in flowing dry air. Finally, each mixture was sub-
jected to one-second exposures of 740°C. The trapped effluents
from these thermal exposures were subsequently subjected to GC
analysis as described in Figure 4, and the chromatographic traces
resulting from these examinations are shown in Figures 8 through
14. (Note that in this series of chromatograms, signal
amplification was adjusted according to hydrocarbon content of
the PCB's).
In almost all cases there is very little difference between
the respective chromatograms obtained from the unheated sample
and the same Aroclor mixture which had been exposed to 640°C.
However, the chromatograms obtained from samples exposed to 740°C,
in all cases, show either an absence of chromatographic peaks, or
an occurrence of peaks at locations other than what would be
representative of that particular PCB mixture. (The newly-formed
substances were not identified during this program). From this
information it is highly probable that the thermal stability of
all of the PCB compounds would fall between the extremes as
depicted in Figure 5; that is bounded by the thermal destruction
behaviors of biphenyl and decachlorobiphenyl.
THERMAL STABILITY COMPARISON OF PCB's AND MIREX
Although the various PCB compounds thermally decompose with-
in a fairly narrow region, as a group they possess considerable
high-temperature stability. For comparison, the thermal destruc-
tion profile for Mirex, obtained in a previous study [42], was
redrawn and directly compared with a thermal destruction plot of
a representative PCB, 2,5,2',4',5'-pentachlorobiphenyl. The
profiles of Mirex and this PCB are shown in Figure 15. It is
24
-------
CD
o
-------
u_
<£
O)
I
o
X
10
740°C
640°C
UNHEATED
0
10
15 20 25
TIME , MIN
30
35
Figure 9. Chromatograms produced by Aroclor #1232 samples
26
-------
C/)
u>
O
X
CO
740°C
640°C
UNHEATED
0
10 15 20
TIME , MIN
25
30
Figure 10. Chromatograms produced by Aroclor #1242 samples
27
-------
O
X
10
740°C
640°C
UNHEATED
_L
0
10
15 20
TIME, MIN
25
30
Figure 11. Chromatograms produced by Aroclor #1248 Samples
28
-------
en
01
'o
X
to
740°C
640°C
UNHEATED
10 15 20
TIME, MIN
25
30
35
Figure 12. Chromatograms produced by Aroclor #1254 samples,
29
-------
(f)
'2
X
00
740°C
640°C
UNHEATED
_L
0
10 15 20
TIME, MIN
25
30
35
Figure 13. Chromatograms produced by Aroclor #1260 samples.
30
-------
CO
'o
X
CO
740°C
640°C
UNHEATED
_L
_L
0
10
15 20
TIME, MIN
25
30
35
Figure 14. Chromatograms produced by Aroclor #1262 samples
31
-------
100
? 80
UJ
en
UJ
O
cr
UJ
CL
O
UJ
60
40
20
tr = 1.0 SEC
200 400 600 800
TEMPERATURE, °C
1000
Figure 15. Comparison of thermal destruction of Mirex and
PCB's.
32
-------
readily apparent from this figure that PCB's are indeed con-
siderably more stable than Mirex at high temperatures.
In addition, Mirex has also been examined by other investi-
gators using pilot-scale incineration equipment [43]. With this
high-temperature disposal equipment it was difficult to satis-
factorily incinerate Mirex at temperatures of 900-1000°C. Based
upon our experimental information, the thermal disposal of PCB' s
should require more severe thermal exposures than Mirex.
THERMAL FRAGMENTATION OF PCB's
In an attempt to obtain a more thorough understanding of the
actual thermal decomposition of PCB' s, a test was conducted where
a sample of Aroclor #1242 was subjected to a 690°C thermal ex-
posure for one second. This test produced approximately 50%
destruction of the Aroclor sample. A second test was conducted
using the same Aroclor #1242 sample and thermal exposure
conditions; however, this time a charcoal trap was substituted
for the Tenax-GC trap. The contents of the charcoal trap were
then chemically desorbed using carbon disulfide. This desorbed
sample was first subjected to a gas chromatographic analysis
which centered attention on the molecular weight range of 78
(benzene) to 189 (monochlorinated biphenyl). It was determined
from this analysis that the desorbed sample contained only very
low levels of organic compounds falling within the above mole-
cular weight range. Next, the same sample was subjected to a
controlled evaporation of the solvent, after which the residue
was dissolved in n-hexane and subjected to a gas chromatographic
analysis using a high-resolution glass open tubular column [44].
An unheated sample of Aroclor #1242 in n-hexane was also
subjected to the same GC analysis. The gas chromatographic
conditions for these two analyses are presented in Figure 16,
while Figure 17 shows the chromatograms obtained. The chromato-
gram shown in Figure 17a represents the unheated Aroclor #1242;
Figure 17b represents the collected effluent from an equal
quantity of Aroclor #1242 that was subjected to a 690°C, one-
second thermal exposure in air. Essentially, both samples con-
tain the same compounds. However, through comparison of the
normalized integrated responses, it was again determined that
approximately 50% of the heated Aroclor #1242 was thermally de-
composed. Upon numerical comparison of the individual peaks in
these two chromatograms, it was generally found that proportion-
ally more of the lower molecular weight PCB compounds decomposed
than the higher molecular weight compounds.
The main conclusion derived from this thermal testing of
Aroclor #1242 in air is that PCB compounds fragment to low mole-
cular weight products, generally lighter than benzene, upon
reaching a destructive temperature. This same type of fragmenta-
tion was obtained through another series of hiqh-temperature
destruction tests, the results of which are illustrated in
33
-------
GAS CHROMATOGRAPHIC CONDITIONS
Instrument
Analyst
Date Aug. 12, 1977
Varian 1860-1
W. Rubey
Column:
tubing material
tubing length
tubing bore
soft glass
15
0.25
meters
mm
Stationary Phase SE-30 Silicone
weight
Support
mesh
Carrier Gas
inlet pres
Helium
1.9
1.0
linear velocity
outlet flow ^0.6
Chart Advance
Read Out
Sample Aroclor 1242
size
20
abs atmo
cm/sec
ml/min
cm/min
1.0
mv full scale
5.0 yl
Detector
range
HF1D
10-11
attenuation 8_
Gas Flows, ml/min
hydrogen 30
air
300
split ratio
Temperatures,
detector
injector
column:
initial
final
45 to 1
°C
290
260
100
230
program rate
Sample Solvent
concentration
4.0
'C/min
n-hexane
0.8 yg/yl
Figure 16. Open tubular column gas chromatographic conditions
34
-------
c/>
u.
'g
X
00
(a) -
CO
I
LU
cc
I UNHEATED 'AROCLC^R *I242
SOLVENT
X(n-HEXANE)
QUANTITATIVE
STANDARD
(lOng.n-TETRADECANE)
LJ
_L
0 5 10 15 20 25 30 35
TIME , MIN
o
(b) m
LU'
CO
o
a.
co
LU
CE
SOLVENT
(n-HEXANE)
QUANTITATIVE
STANDARD
(IOng,n-TETRADECANE
\
AROCLOR *I242 AFTER
690°C EXPOSURE FOR 1.0 SEC
IN FLOWING DRY AIR
10 15 20 25
TIME, MIN
30 35
Figure 17. Chromatograms of Aroclor #1242 samples,
35
-------
Figure 18. The bottom GC trace represents the chromatographic
analysis of an unheated 2,5,2',4',5'-pentachlorobiphenyl sample,
which contains an impurity of approximately 0.25%. When 40
micrograms of this substance are subjected to thermal exposure at
increased temperatures, e.g., those in Figure 18, it is observed
that only small levels of trapped decomposition products (<1%)
are formed. (The PCB isomer peak was allowed to go off-scale
during these tests, so that attention could be centered upon the
decomposition products formed during the thermal exposures.) It
is apparent from this series of chromatograms that decomposition
products that can be trapped by Tenax-GC are of very low
concentration. Despite their low level, the same pattern of
decomposition products is observed at exposure temperatures of
651° through 727°C. When this same PCB isomer is exposed to an
even higher temperature, 754°C, the PCB isomer is no longer
present in the chromatogram and only extremely low levels of
decomposition products are observed. Again, the information
obtained from these tests agrees with the earlier results, i.e.,
once the PCB molecule encounters a thermally destructive environ-
ment, it essentially fragments into small decomposition products.
THERMAL DESTRUCTION AT THE HIGHER TEMPERATURES
The extent of thermal destruction at temperatures above
800°C was next investigated in this laboratory program. Subse-
quently, a sample was prepared which contained equal parts by
weight of each of the seven Aroclor mixtures and the 2,5,2',4',
5'-pentachlorobiphenyl. (A chromatogram of this composited
sample, dissolved in hexane, is shown in Figure 19.) High-
temperature exposure tests were then conducted at 900°C and
1000°C with a mean residence time of 1.0 seconds. An additional
test was performed at 1000°C using a 2.0 second mean residence
time. The respective trapped effluents were then analyzed. By
using the peak height of the pentachlorobiphenyl isomer as a
quantitative measure, destruction percentages for these high-
temperature exposures were calculated and tabulated in Table II.
It is evident that increased residence time has a pronounced
effect upon the destruction percentage at 1000°C. Another
observation was made while analyzing the chromatograms produced
by this composited Aroclor sample. It was noted that the lower
molecular weight PCB compounds, those emerging to the left of
the pentachlorobiphenyl in Figure 19, were almost totally absent
in the collected effluents. However, traces of the heavier
PCB's, those with retention times to the right of the penta-
chlorobiphenyl, were still in evidence after the 1000°C exposure.
Thus, this would again substantiate the finding that the higher
the molecular weight of a PCB, the greater its thermal stability.
RESIDENCE TIME CONSIDERATIONS.
To further examine the effect of residence time, tests were
conducted at a fixed temperature which was less than required
36
-------
o
X
to
z
o
a
CO
UJ
ce
J
754°C
I 0 SEC RESIDENCE TIME IN
FLOWING DRY AIR
727°C
692°C , i_ ^
65 1°C
PCB ISOMER — -
UNHFATFO
1.
I
A
A
/ IMPURITY, ~0 25 °/
10
15
~20
25
TIME, MIN
—r—
30
35
—i—
40
45
Figure 18. Chromatograms of 2,5,2',4',5', -
pentachlorobiphenyl samples.
37
-------
CO
o
X
LJ
CO
o
0.
CO
u
cr
COMPOSITE PCB SAMPLE
equal parts by weight of Aroclor
1221, 1232, 1242, 1248, 1254,
1260, 1262, and 2, 5, 2',41, 5', -
Pentachlorobiphenyl
2,5,21,4',51-
PENTACHLOROBIPHENYL
0
10
15 20
TIME , WIN
25
30
35
Figure 19. Chromatogram of composited Aroclor mixtures
plus 2,5,2',4',5', - pentachlorobiphenyl.
38
-------
TABLE II. HIGH-TEMPERATURE EXPOSURE DESTRUCTION PERCENTAGES
Exposure Temperature, T
CO
900
1000
1000
Residence Time, t
(sec)
1. 0
1.0
2. 0
Destruction Percentage*
99.988
99.995
99.9995
*Destruction Percentage = 100
1 -
wt trapped sample
wt input sample
TABLE III. WEIGHT PERCENT REMAINING AFTER 704°C EXPOSURE
FOR RESPECTIVE RESIDENCE TIMES
Compound
Biphenyl
Dibenzofuran
Dibenzo-p-dioxin
Hexachlorobenzene
Tetrachlorobiphenyl
(2,5, 2', 5')
Pentachlorobiphenyl
U,5,2',4',5')
Decachlorobiphenyl
Residence
0.27
8. 1
35. 1
45. 3
88.4
78. 5
81. 1
84.7
0.
0.
3.
1.
37.
14.
18.
37.
Time,
95
70
8
2
2
0
5
3
Sec
3.
0.
0.
0.
11.
2.
3.
16.
84
07
93
07
3
6
4
1
39
-------
for complete thermal decomposition. Specifically, a combined
sample containing pure compounds was examined using a 704°C
exposure, but with different air flow rates passing through the
quartz-tube apparatus, thereby obtaining different residence
times at the fixed temperature. The effluent samples obtained at
these different residence times were then analyzed. The results
are presented in tabular form in Table III. These same data are
presented in log-log form in Figure 20. From these data, it is
readily apparent that residence time is a very strong factor
affecting the high-temperature destruction of PCB's and related
compounds.
40
-------
lOO
LU
o:
o
a:
LJ
Q_
10
1.0
0.01
O.I
T2 = 704°C
DECACHLOROBIPHENYL
HEXACHLOROBENZENE
PENTACHLJOROBIPHENYL
(2,5,2',4, 5,)
TETRACHLOROBIPHENYL
(2,5, 2', 5\)
DIBENZOFURAN
DIBENZO-p- DIOXIN
BIPHENYL
0.3 0.5 1.0 3.0
(tr )
RESIDENCE TIME , SEC
5.0
10
20
Figure 20. Effect of residence time at 704°C.
41
-------
REFERENCES
1. Jensen, S., New Scientist, 32, 612 (1966).
2. Miller, J.W., U.S. Pub. Health Rep., 59, 1085 (1944).
3. Umeda, G., Ambio, 1, (4) 132 (1972).
4. Duke, T.W., Lowe, J.I., Wilson, A.J., Bull. Environ. Contam.
Toxicol. , 5_, 171 (1970).
5. Hansen, D.J., Parrisa, P.R., Lowe, J.I., Wilson, A.J., and
Wilson, P.O., Bull. Environ. Contam. Toxicol., 6_, 113 (1971).
6. Kimbrough, R.D., and Linder, R.E., J. Natl.. Cancer Inst.,
3_3, 547 (1974) .
7. Bahn, A.K., Rosenwaike, I., Herrmann, N., Grover, P.,
Stellman, J., O'Leary, K., N. Engl. J. Med., 295, 450 (1976).
8. Chemical Week, p. 29, Sept. 22, 1976.
9. Risebrough, R., Reiche, P., Peakall, D.B., Herman, S.G., and
Kirven, M.N., Nature, 220, 1098 (1968).
10. Gustafson, C.G., Environ. Sci. Tech., 10, 814 (1970).
11. Jensen, S., Ambio, 1, (4) 123 (1972).
12. Fishbein, L., J. Chromatog., 68, 345 (1972).
13. Selikoff, I.J., Ed., Environ. Res. , 5_, 3 (1972).
14. Carnes, R.A., Doerger, J.U., and Sparks, H.L., Arch. Environ.
Contam. and Toxicol., 31, (1) 27 (1973).
15. Ruopp, D.J., and DeCarlo, V.J., International Conference on
Environmental Sensing and Assessment, Vol. 1, Paper 2-3.
16. Veith, G.D., and Lee, G.F., Water Research, £, 265 (1970).
17. Hutzinger/ O., Safe, S., and Zitko, V., Analabs Res. Notes
12, (2) 1972.
42
-------
18. PCB's and the Environment, HEW Report, COM-72-10419, March
1972.
19. Environmental Quality - 1976, The Sixth Annual Report of the
Council on Environmental Quality (Washington, D.C.: U.S.
Government Printing Office, 1976), pp. 44-46.
20. Chemical and Engineering News, p. 8, Oct. 11, 1976.
21. Chemical Week, p. 34, Feb. 25, 1976.
22. Safe, S., and Hutzinger, O., Nature, 232, 642 (1971).
23. Ruzo, L.O., Zabik, M.J., and Schuetz, R.D., Bull. Environ.
Contam. Toxicol. , 8_, 217 (1972).
24. Herring, J.L., Hannan, E.J., and Bills, D.D., Bull. Environ.
Contam. Toxicol., 8_, (3) 153 (1972).
25. Ahmed, M., and Focht, D.D., Bull. Environ. Contam. Toxicol.,
10., (2) 70 (1973) .
26. Sebastian, F.P., Kroneberger, G.F., Lombana, L.A., and
Napolean, J.M., Latest Developments on PCB Decomposition
in BSP Multiple Hearth Furnaces, paper presented at the
Annual AICHE continuing education workshop, Nov. 1, 1972,
at Midland, Michigan.
27. Sewage Sludge Incineration, EPA Task Force Report PB-211-323,
March, 1972, pp. 36-37.
28. Whitmore, F.C., Destruction of Polychlorinated Biphenyls in
Sewage Sludge During Incineration report for EPA under
Contract 68-01-1587, 1975.
29. World Environment Report, pp. 3-4, Aug. 2, 1976.
30. Karlsson, L., and Rosen, E., Chemica Scripta, _!, 61 (1971).
31. Solid Waste Report, p. 61, April 12, 1976.
32. Kennard, E.H., Kinetic Theory of Gases, McGraw-Hill, New
York, 1938, Chap. I.
33. Sternberg, J.C., Advances in Chromatography, Vol. 2,
(Giddings, J.C., and Keller, R.A., eds.) Marcel Dekker, New
York, 1966, Chap. 2.
34. Sutton, R., and Harris, W.E., Can. J. Chem., 45, 2913
(1967).
43
-------
35. Chen, J.T., JAOAC, 5_9, 993 (1976).
36. Sissons, D., and Welti, D., J. Chromatog., 60, 15 (1971)
37. Zitko, V., and Choi, P.M.K., Bull. Environ. Contain. Toxicol.,
10., (2) 120 (1973) .
38. Crosby, D.G., and Moilanen, K.W., Bull. Environ. Contam.
Toxicol. , 10., (6) 372 (1973).
39. Chemical and Engineering News, p. 27, Aug. 23, 1976.
40. Gribble, G.W., Chemistry, 47, (2) 15 (1974).
41. Buser, H.R., Anal. Chem. , 4_8, 1553 (1976).
42. Duvall, D.S., and Rubey, W.A., Laboratory Evaluation of High
Temperature Destruction of Kepone and Related Pesticides,
EPA Report, EPA-600/2-76-299, Dec., 1976.
43. Ferguson, T.L., Bergman, F.J., Cooper, G.R., Li, R.T., and
Honea, F.I., Final Report EPA Contract No. 68-03-0286,
July, 1975.
44. Kaupcik, J., Leclercq, P.A., Simova, A., Suchanek, P.,
Collak, M., and Hrivnak, J., J. Chromatog., 119, 271
(1976).
44
-------
APPENDIX I
DESCRIPTION AND OPERATION OF QUARTZ TUBE APPARATUS
A schematic of the high-temperature quartz tube apparatus
is shown in Figure Al, and a photograph in Figure A2. An
important component in this apparatus is the folded quartz tube
which is contained in a Lindberg furnace, type 55035A. Two
separate quartz tubes of different bore size are available with
this apparatus. Both tubes have an effective length, when
positioned in the furnace, of 84 cm. The average inside
diameters of the finished, folded tubes are 0.80 mm and 2.14 mm,
respectively. The study reported on here made principle use of
the 0.80 mm inside diameter folded tube.
Upstream of the installed, folded quartz tube is a special
sample holder which is shown in Figure A3a. Using a 10 micro-
liter syringe and following suggested procedures [45], a known
quantity of diluted sample can be readily deposited on the sand
blasted region of the sample holder. The solvent quickly
evaporates, leaving the sample on the rough Pyrex surface. This
device is then inserted into a chamber upstream of the high-
temperature quartz tube. Next, this chamber is gradually heated
to the extent necessary to vaporize the sample, thereby sweeping
the gas phase molecules through the high-temperature tube.
An effluent trap is located at the outlet of the high-
temperature tube. A sketch of the basic trap is shown in Figure
A3b. This long Pyrex tube (22.5 cm) is connected to the quartz
tube outlet with a special Swagelok reducer union containing a
Teflon sleeve. Vespel ferrules were used to produce a gas-tight
seal. The outlet of the trapping medium tube approaches ambient
temperature.
A variety of different trapping media can be used for
collecting organic products. Of special interest, however, is
Tenax-GC, which has been found to be an excellent trapping
medium for various organic compounds and their respective
decomposition products (M.W. of 150 to 550). Tenax-GC, a porous
polymer, is commercially available in 35/60 and 60/80 mesh sizes.
Also, this material has a desirable feature in that it can be
readily, thermally desorbed [46-53]. The research reported on
here utilized traps prepared with the 60/80 mesh Tenax-GC
45
-------
HIGH TEMPERATURE REGION
VENT
A. COMPRESSED AIR, BREATHING QUALITY GRADE
Q TWO STAGE PRESSURE REGULATOR
C • "HYDROPURGE" FILTER
D FLOW CONTROL VALVE
E PRESSURE TRANSDUCER
F SAMPLE HOLDER, PYREX
G HEATED INLET CHAMBER
H QUARTZ TUBE
1 HEATED OUTLET CHAMBER
J EFFLUENT TRAP, TENAX-GC OR CHARCOAL
K FLOW METER
Figure Al. Schematic of quartz tube apparatus
46
-------
TRANSDUCER
CONTROL UNIT
DIGITAL READ-OUT
OF TEMPERATURE
AND PRESSURE
LINDBERG
FURNACE
FLOW METER
TRANSDUCER
POWER SUPPLY
PRESSURE
REGULATOR
FLOW CONTROL
VALVE
PRESSURE
TRANSDUCER
SATURATOR
HYDRO-PURGE
FILTER
COMPRESSED
AIR
HEATED
OUTLET
Figure A2. Photograph of quartz tube apparatus
47
-------
GAS-TIGHT SEAL
(3-SILICONE 0-RINGS)
ADJUSTABLE
SET-SCREW
RETAINER
SAMPLE DEPOSIT REGION
2.0cm BY 0.4cm DIAMETER
( SAND BLASTED)
PYREX ROD
25cm BY 0.6cm
(a) sample holder
-PYREX TUBE
GAS FLOW
DIRECTION
SILANE
TREATED
GLASS
TRAPPING MEDIUM
/ SILANE
TREATED
^GLASS
/ WOOL
3 Omm-
I 8mm-
—4D8ci
I 5cm
- 20cm HOScml-
225cm
(b) effluent trap
Figure A3. Sample holder and effluent trap,
48
-------
particles. A 2.0 cm length of this Tenax-GC packing was suffi-
cient to obtain quantitative recoveries of pesticides and other
related organic compounds. However, acetone and benzene are not
quantitatively retained at the typical trap inlet temperature of
approximately 300°C. Charcoal [54,55] has also been used as a
trapping medium, however, it must be chemically desorbed as
opposed to the thermal desorption of Tenax-GC.
Other important components in this high-temperature destruc-
tion apparatus are a Porter flow control valve (model number
VCD-1000), a differential pressure transducer, two Blue M model
APH-500 temperature indicators, which continually display the
temperatures of the inlet and outlet chambers, and a specially
fabricated soap-bubble flowmeter for measuring the system's
outlet flow rate. The gas used with this apparatus was breathing
quality compressed air passed through an Applied Science
Laboratories Hydro-Purge filter. The average temperature in the
high-temperature furnace is continually monitored by the digital
readout from a compensated chromel-alumel thermocouple.
Prom the sketch shown in Figure A4, it is seen that the
pressures, temperatures, and volumes in this system can be
related according to the ideal gas law as follows:
Vopo (A.I.I)
where V is volume, p is pressure, T is temperature, and T is
temperature averaged over the length of the tube. The subscript
2_ denotes the quartz tube interior, while the subscript o
represents the ambient outlet conditions.
Due to the extremely low pressure drop posed by the open
quartz tube, and the comparatively high pressure drop presented
by the packed small particle effluent trap, the differential
pressure p^ can be expressed as
Pd= P2 ~ PO = PI-PO ' (A. 1.2)
Now, volume flow F can be written as
F = ~ , (A.I.3)
where t is time, equation (A.I.I) can be rewritten as
f 2 = ——— ' (A.1.4)
49
-------
GAS FLOW
=3>
INLET
EFFLUENT
TRAP
V2 «P2
OUTLET
Figure A4.
Sketch of pressure, volume, and
temperature relationship in test
system.
50
-------
or
^r Lp2 = Fotpo , (A.I.5)
T2 To
where r and L are the tube radius and length, respectively.
Therefore, it is seen that time, or, in this case, residence time
can be written simply as
- = /CrfL^ /V\ fl + :o_l f (A.I>6)
r
where tr represents the calculated mean residence time of a
substance in the quartz tube interior.
Special equipment is required to measure the outlet flow,
the temperature of the quartz tube, and the upstream differential
pressure. Specifically, the outlet flow is measured with a small
bore soap-bubble flowmeter having an in-line watar saturator
[56]. This device was fabricated in the University Glass Shop
and incorporated a calibrated volumetric pipette in the design.
The average temperature of the quartz tube was determined by
measuring the axial temperature profile in the Lindberg furnace
and then determining the location of a point within the furnace
which would represent the average temperature. This point was
found to be at a location 9 cm from the axial midpoint of the
furnace. The temperature profile for the contained, folded
quartz tube is shown in Figure A5. This thermal information was
obtained with a chromel-alumel thermocouple, which was used to
measure the temperature at 1 cm increments along the tube furnace
axis. The average temperature that a compound would encounter
during its gas phase traverse through the quartz tube is
represented by ^2 as shown in Figure A5.
The pressure within the quartz tube is difficult to measure.
A low pressure gage, such as a Bourdon gage, customarily exhibits
high degrees of inaccuracy, along with a large mechanically
induced hysteresis. Consequently, a very precise pressure
transducer is used to measure the differential pressure at the
inlet to the quartz tube. This pressure transducer (see Figure
A6) is extremely accurate, linear, and exhibits no hysteresis.
The output from the pressure transducer is displayed on a
digital millivoltmeter. The system differential pressure varies
with the particle size of the trapping medium and the linear gas
velocity.
51
-------
1200
1000
800-
z>
I—
<
600-
400
200
-TEMPERATURE OF OUTER SURFACE OF FOLDED QUARTZ TUBE-
— 9cm-
-MIDPOINT OF FURNACE
T2, AVERAGE TEMPERATURE OF
FLOWING GAS IN QUARTZ
TUBE
- LOCATION OF T THERMOCOUPLE
I i I
__ TUBE FURNACE
TEMPERATURE PROFILE
0 10 20 30
40 50
DISTANCE, cm
60 70 80 90
Figure A5. Temperature profile of quartz tube furnace
STATHAM STRAIN GAGE
DIFFERENTIAL PRESSURE
TRANSDUCER
Figure A6. Electrical schematic of pressure transducer,
52
-------
Once the sample molecules or fragments and products thereof
have passed through the high-temperature quartz tube, they are
trapped in or on the sorbinq medium of the effluent trap. This
trap is then removed from the high-temperature apparatus and a
quantitative standard is deposited in the trap. The contents are
then subjected to a programmed-temperature gas chromatographic
analysis using a modified Tracor 550 instrument.
The injector of this gas chromatograph has been modified so
that it will accept the Pyrex tube containing the trapping medium
and its sorbed sample. Modifications to the chromatograph are
such that thermal desorption of the trapping medium can be
conducted while it is positioned just upstream of the column
packing (see Figure A7). This is accomplished by means of a
cartridge heater embedded in an aluminum block which surrounds
the inlet region of the glass column. By heating this inlet
region to 270°C for 10 minutes, the trapped effluent is
completely desorbed. Since this desorption is made with the
chromatographic oven at room temperature, the products are swept
from the trapping medium and deposited on the inlet portion of
the packed column.
After a suitable thermal desorption time (10 minutes), the
chromatographic oven is closed and the column temperature is
programmed, thereby separating and subsequently detecting the
various organic products. With this trapping and thermal
desorbing technique, essentially quantitative collection and
transfer has been achieved for organic compounds in the molecular
weight range of 150 to 550.
With this apparatus, the effect of residence time on high-
temperature degradation can be readily determined. Samples can
be subjected to successive tests at a constant ^2 but with
different volume flow rates, that is, different residence times.
The trapped effluents from these tests can_then be analyzed, thus
permitting an evaluation of the effect of t on thermal destruc-
tion at a given temperature.
53
-------
EXTERNAL POWER
SUPPLY AND
TEMPERATURE
CONTROL
INJECTOR'
CAP
INSULATION
/
ALUMINUM BLOCK
WITH EMBEDDED
CARTRIDGE HEATER
HELIUM
CARRIER
-INSULATION
VENT
GLC COLUMN
X(PYREX)
I.8M, OV-I
Figure A7. Interior of modified Tracer 550
54
-------
APPENDIX II
ERROR ANALYSIS OF MEAN RESIDENCE TIME DETERMINATION
Earlier, equation (A.I.6) described the mean residence time
of a substance as it passed through a high-temperature quartz
tube. This residence time equation,
(A.II.l)
contains many variables, some of which have a pronounced effect
on tr and are difficult to measure.
Upon investigation, it is found that fused quartz has a very
small coefficient of expansion (approximately 0.59 x 10 ).
Therefore, for a given tube, the radius r and the length L can be
assumed temperature invariant. Thus, the effective volume of a
specific quartz tube, i.e.,
Vt = trr L
remains essentially constant.
(A.II.2)
Further, the laboratory temperature TQ and pressure p can
However, the outlet ambient
be easily and accurately measured.
o,
differential pressure p^ require special measurement equipment.
flow rate F, the average oven temperature T2, and the system
The maximum change in tr, which can be attributed to
addition of errors in FO, T^, and p^, can be expressed as
/ \ f ^ / \
Atr =
J8tr J AFQ
+
/9tr\
V*2/
AT2
+
1 rl
\9Pd/
Apd
•
(A.II.3)
From equations (A.II.l) and (A.II.2), we find
3F
1 +
(A.II.4)
55
-------
3t,
3T-
1 +
(A.II.5)
VA
(A.II.6)
3P,
F0T2P0
Now the maximum errors typically associated with the selected
measurement methods are
AFQ = FQ x 10~2cm3sec~1
AT2 = 5°K
= 5 x 10~3atm
Next, an example case is presented where
£„ = 1.0 sec
o
o
2
o
1.235 cm sec
296°K
873°K
= 0.974 atm
= 0.2 atm
= 3.021 cm3
By substituting into equation (A.II.3) the data for this
case and the typical measurement errors, the worst condition is
found to be
Thus,
At,
At,
tr
Atr
(flow)
0.00999 +
0.020
At,
(temperature)
0.00575
or, the maximum relative error in tr is 2.0%.
At
(pressure)
0.00426
56
-------
With this instrumentation, the_flow measurement possesses
about twice the effective error in tr as the temperature or
pressure measurement. Also, upon further examination, it is
observed that Atr/tr is essentially unaffected by the broad
temperature range and the various residence times.
Consequently, this error analysis shows that the individual
measurement errors encountered in determining tr are small and in
the same domain.
57
-------
APPENDIX III
RESIDENCE TIME VARIATION
The longitudinal dispersion of a gas as it passes through an
open cylindrical tube has been thoroughly studied [57,58].
Consequently, in the absence of a retentive or interactive wall
surface, the variation in transport of like molecules can be
expressed by
H =
dx
(A.III.l)
where H is the height equivalent to a theoretical plate (as
commonly utilized in chromatography), ox is the distribution
standard deviation in distance units, x is distance, Dg is the
intermolecular diffusion coefficient, v is the average linear
velocity of the gas, and r is the radius of the open cylindrical
tube.
From equation (A.III.l) we can write
•x=L / 2D \ /-x=L
^^••i
v
f
dx
r
dx
(A.III.2)
and, at distance x=L
, x 2
2D L
g
v
vr2L
24D
(A.III.3)
Now, for like molecules that are randomly entering and
passing through a narrow bore tube, the variation in residence
time can be represented by a Gaussian distribution with a time-
based independent variable. Thus, it is necessary to convert
a to a time-based sigma value. This is readily accomplished as
(CV
= (at}
9
(A.III.4)
58
-------
Equation (A.III.3) can now be rewritten as
2D L
(atr =
A
24vD
(A.III.5)
or
f
2D L
g
(v)3
, A
24vD
(A.III.6)
The intermolecular diffusion coefficient has been equated
by Reid and Sherwood [59] as
D = D
g 1,2
Pd'
(A.III.7)
where P is pressure, T is temperature, d is distance between the
molecular centroids upon impact, M, and M_ are the respective
molecular weights, and b is a constant.
Due to the absence of high- temperature diffusion data for
MOO MW compounds in air, it was necessary to select a model
compound for which Dg data did exist. Calculated D data could
then be obtained with the use of equation (A. Ill . 7) "for a broad
range of compounds.
The selected model compound was naphthalene. This substance
is low enough in molecular weight that D~ data are available,
e.g., Dg for naphthalene in air = 0.0611 cm2 sec"1 at 298°K and
1.0 atm [59], and it is a common household pesticide. Also,
naphthalene is structurally similar to other fused-ring
compounds, e.g., aldrin, carbaryl, dichlone, dieldrin, endrin,
and the halowaxes.
Next, from the diffusion expression,
(A. III. 7), it is observed that
.e.
equation
3b
T(M,
M1M2
(A.III.8)
Therefore, even with fixed values of P, r, L, and residence
time, a will continue to vary with temperature and M- (the
molecular weight of the larger component).
59
-------
To clarify this situation, residence time distribution data,
as represented by cr^, have been calculated and plotted for the
specific case where molecules experience a 1.0 sec residence time
while passing through a tube 84 cm in length by 0.8 mm inside
diameter with an internal pressure of 1.2 atmospheres. These
calculated at data which are presented in Figure A8a encompass
the temperature range of 298°K to 900°K, and the molecular weight
range of 128 (naphthalene) to 546 (Mirex).
Another graph was prepared, again from calculated data,
where a^_ was plotted versus residence time for naphthalene at
700°K and 1.2 atmospheres. This information is presented in
Figure A8b.
From information contained in Figures A8a and A8b, a number
of conclusions can be drawn. First, from Figure A8a, it is
observed that molecular scatter, or variation in residence time,
is small. As an example, it is seen that at 700°K the residence
time for 95% of the naphthalene molecules would fall within
1.0 ± 2a.(. seconds, that is, between 0.96 and 1.04 seconds.
Secondly, it is observed from this same figure that the molecular
weight of the vaporized compound has only a small effect on at-
This is evidenced by the narrowness of the molecular weight band.
Thirdly, it is observed that molecular scatter is strongly
dependent upon temperature. In fact, although the intermolecular
diffusion coefficient increases with temperature, the residence
time distribution, as represented by at, decreases with temper-
ature .
Figure A8b shows the variation of at with tr for
naphthalene. Upon further investigation it is determined that
the term 4a^./t varies only by a factor of two over the residence
time range of 0.3 to 2.0 seconds. Thus, the relative change in
at with Er is tolerable.
Lastly, it is important to note that at is strongly
dependent upon the tube radius r. It is seen from equation
(A.III.6) that as r increases, the residence time distribution
broadens. Thus, for a small variation in molecular transport
time, a quartz tube with a narrow bore is required.
60
-------
(a)
O.O4
0.03
-------
REFERENCES FOR APPENDICES
45. Fairbairn, J.W., and Ralph, S.J., J. Chromatog., 33, 494
(1968) .
46. Novotny, M. , Lee, M.L., and Bartle, K.D., Chromatographia, 1_,
333 (1974) .
47. Zlatkis, A., Lichtenstein, H.A., and Tishbee, A., Chromato-
graphia,
------- |