PB88-19S2SO
Formation of PICs (Products of
Incoaplete Combustion) from the
Combustion of Chlorinated Materials
(U.S.) Environmental Protection Agency
Cincinnati, OH
Apr 88
-------
EPA/600/0-88/062
April 1988
THE FORMATION OF PICs PROM THE COMBUSTION CP CHLORINATED MATERIALS
Laurel J. Staley
Marta K. Richard*
George L. Huff SB n
Thermal Processes Research Staff
Thermal Destruction Branch
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protect Ion'Agency
Cincinnati, Ohio 45268
and
Barry Dellinger
University of Dayton Research Institute
Dayton. Ohio 45469
HAZARDOUS HASTE ENGINEERING RESEARCH LABORATORT
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCT
CINCINNATI, OHIO 45268
-------
1. RE*OPT WO. 2.
EPA/600/D-88/062
4- Tltiif AWO SUBTITLE
THE FORMATION OP PICs FROM THE COHBVSTIOIf
OF CHLORINATED MATERIALS
* -^CKSrel1 J. Staley, Marts K. Richards, George L. Huffman
and Barry Del linger
• . PERFORMING OROAMiyATiriM •••*** AMn ADDRESS
Sam* as (12)
12. SPONSORING AGENCY NAME AND AOORESS
HAZARDOUS HASTE ENGINEERING RES. LAB.
OFFICE OF RESEARCH AMD DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION ACEHCT
CINCINNATI, OH 45268
a. a££pteMrs ACCESSION NO.
liiff" 1 95230JIS
April 1988
I »«R»ONMINO O*OAMI2*TtOM COO*
S. PER^ORKIMO OMOAMIZATION. RE*O*T NO.
*
1O. rROQRAM ELEMENT MO.
11. COW^HACT/CRANT NO.
U. TYPt Of REPORT AMO fCRIOD COVERED
1*. S»'ONSORINC ACENCY COOt
EPA/600/12
18 SUf^LCMENTAMV NOT!»
»6. ABSTRACT
The formation of unwanted Product* of Incomplete Combuatloo (PICs) fro*
Che Incineration of hazardous VMte* under sab~optimal conditions Mist be
avoided if incineration it to be effectively need to treat hazardous «aate*.
Studies of the thermal behavior of chlorinated C|S (methanes) and €2* (eth-
anes and ethenes) in non-flame, flat-flame, and shock tube devices have
yielded results which offer a plausible explanation for the formation of
certain PICs. To determine iritether the same results are obtained in a larger,
more complicated device, two mixtures of chlorinated methane* and two mix-
tures of chlorinated ethanes and ethenes were burned under "high* and "low"
oxygen conditions in a Turbulent Flame Reactor (TFR). Results of these
tests indicate that the earlier test results predict some of the emissions
observed from the TFR. Soot formation affects the emissions of volatile
organic chemicals in ways not predicted by the earlier studies however*
it.
KtV WQftOS AMD OOCUMCMT ANALYSIS
1. DESCRirrORS
IS. DISTRIBUTION STATEMENT
Release to Public
fclOewTIFIiHSWtN SNOCD TCRWS
IS. SECURITY CLASS fl** tttf^rtf
On classified
10. SECURITY CLASS fTft* p«jr|
Unclassified
c. COSATI ricHVGnmy
21. MO. O* PAGES
flu»2 iS'QS^
«m HIS.1 <•**.
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
ii
-------
THE FORMATION OF PICS FROM THE COMBUSTION OF CHLORINATED MATERIALS
Laurel J. Stale?*, Harta K. Richards. George L. RuffBan, U.S. EPA and
Barry Dellinger, University of Dayton
ABSTRACT
The formation of unwanted Products of Incomplete Combustion (PICs) from
the incineration of hazardou* wastes under sub-optimal conditions must be
avoided if Incineration is to be effectively used to treat hazardous wastes.
Studies ot the thermal behavior of chlorinated Cjs (methanes) and C2« (eth-
anes and ethenes) in non-flame, flat-flame, and shock tube devices have
yielded results which offer a plausible explanation for the formation of
certain PICs. To determine whether the same results are obtained in a larger,
more complicated device, two mixtures of chlorinated methanes and two mix-
tures of chlorinated ethanes and ethenes were burned under "high" aad "low"
oxygen conditions in a Turbulent Flame Reactor (TFR). Results of the*-?
tests indicate that the earlier test results predict some of the emissions
observed from the TFR. Soot formation affects the,emissions of volatile
organic chemicals in ways not predicted by the earlier studies however.
Introduction
Public acceptance of hazardous waste incineration is complicated by the
fact that toxic byproducts (also known as Products of Incomplete Combustion
or PICs) can eometimes form from materials fed to incinerators. Although
they are formed in small quantities, these materials can sometimes (though
not always) be more toxic than the original chemicals fed to the incinerator
for disposal. This is a cause for concern. To address this concern studies
have been undertaken to determine how best to minimize the formation of
these compounds.
The Problem
Studying the formation of PICs in hazardous waste incinerators is com-
plicated by three factors. First, the waste stream fed to most incinerators
is a multi-component, and sometimes even a multi-phase, mixture of organic
and inorganic materials. Second, the combustloo chemistry of each component
of the feed stream consists of multi-step free radical chain reaction mechan-
isms which can result In the formation of a variety of reaction products.
The relative amounts of reaction products which ultimately form depends on
the conditions existing within the incinerator at the time the waste is
incinerated. Third, the conditions existing within the Incinerator are
neither precisely known nor controlled. That is, while incinerator operating
temperature and oxygen level may be known well enough for process control
purposes, localized fluctuations in these and other parameters which may
result in PIC formation may not be discemable using conventional measurement
techniques. Sorting through these complications first requires an under-
standing of the combustion chemistry of the most likely components of the
hazardous waste feed streams fed to incinerators.
Chlorinated organic compounds are most often found in hazardous waate
streams fed to incinerators. Short chain chlorinated hydrocarbons, specifi-
cally chlorinated methanes and ethenes, have been studied most extensively in
non-flame thermal destruction units, flat-flame burners and shock tubes
(1,2,3,4,5,6,7,8). Each of these three experimental devices has rather
uniform and predictable temperature and oxygen concentration profiles. Their
-------
predictability and controllability makes these devices significar.tly differ-
ent from field-scale combustion devices for reasons mentioned above. Never-
theless, results from these studies provide a logical starting point for this
effort.
The Approach
To see how well the results of studies of chlorinated compounds using
these small-scale devices compared to a more realistic combustion environ-
ment, four mixtures of chlorinated methanes, ethanes and ethenes were burned
in a Turbulent Flame Reactor (TFR) under "high" and "low" oxygen conditions.
The gaseous emissions were collected and characterised. These data were then
compared with the results of the earlier studies to determine if temperature
and concentration gradients and also localised fluctuations affected the type
and quantity of PIC* formed.
Tables 1 and 2 list the experimental mixtures used and also the test
conditions evaluated. The large amount of heptane added to each mixture was
required to sustain combustion in the TFR. Although the presence of this
material undoubtedly affected the combustion of the, test chemicals fed, it
does represent a somewhat more realistic situation than in the smaller de-
vices. The "high" and "low" oxyge>: levels were chosen because they repre-
sented the limits of flams stability for the device. Even though the "low*
oxygen condition was still on the average an oxidatlve condition (1151 theo-
retical air), fuel/air mixing was apparently inefficient enough to result in
excessive scot formation.
Table 1. Experimental Test Mixtures and Composition
Composition
Holes
Properties
MollCT/H wtl Cl Moll Cl
(excluding
heptane)
Mixture lAt CH2C12
CHC13
CC14
C7H16
Mixture IB:
78.47
113.642
2500
.66
.74
25.00
.25
2.49
2.80
94.46
6.4
6.65
.87
GHC13
CC14
C7H16
55.05
118.0
11.8
2500
.6476
.9874
.07662
25.00
2.42
3.70
.3065
93.57
2.0
6.03
.78
Mixture 2A:
C2H2Cl4
C2HC13
C2C14
C7Hift
8.311
83.5
107.7
2500
.04947
.635
.648
25.00
.18
2.41
2.46
94.95
6.4
6.18
.84
Mixture 2B: C2H2Cl4
C2HClj
C2C14
7.8816 .0469 .70
39.01 .2966 1.13
157.63 .9196 3.61
2500 25.00 95.08
12.48
6.40
.84
2.
-------
Table 2. Experimental Test Conditions
Temperature Average CO
Mixtures 0? Level Z (Exhaust) *P Level (ppm)
1A 4 IB 6.0-6.3 930 5SO
2 1A 4 IB 4.0 888 50
3 2A 4 2B 6.6-6.8 940 450
4 2A & 2B 3.5 910 200
Figure 1 shovs the TPR and Figures 2 and 3 show the sampling trains
employed. Limitations In both the TFR and In the sampling trains affected
the type and amount of Information obtainable. The TFR, which is described
in detail elsewhere, was designed to enhance the formation and emission of
combustion byproducts (9). Turbulent diffusion flames with large fuel,
oxygen, and temperature gradients enhance the formation of soot and other
combustion byproducts. The TFR's water-cooled walls cooled the post-
flame combustion gases Inhibiting further breakdown of these byproducts.
While low temperature and low oxygen concentrations represent conditions that
are likely to enhance PIC formation, the TFR exaggerates these conditions
resulting in higher levels of these emissions than might ordinarily occur.
The TFR's exhaust gases were sampled for CO, C02 and 02 using the
continuous monitors described in Figure 2. Since the sampling was done only
at the exhaust duct, only concentrations at this point were determined. The
non-uniformities in combustion gas composition within the TFR, which may be
responsible for the formation and emission of organic materials, could not be
measured using these devices.
The Tenax sampling system described In Figure 3 could only reliably
sample materials whose boiling points ranged from 40*C to 160*C. Other
materials Including many polycyclic aromatic hydrocarbons (PAHs) and semi-
volatile organic compounds could not be sampled using this method. Soot was
not sampled.
Results Observed
Several qualitative, yet key, results were obtained from this study.
These are as follows:
1. The observed emissions could be explained, in part* by combining the
results of the studies mentioned above.
2. The formation of soot significantly affected the formation of vola-
tile emissions.
3. PIC formation, and indeed, organic emissions in general, were not
directly correlated with exhaust temperature or carbon monoxide emissions.
Characterization of Emissions
Figures 4 and 5 show the mass emissions for the chlorinated methanes
under "high" and "low" oxygen conditions. Figures 6 and 7 show the emissions
for the chlorinated ethanes and ethenes under similar conditions. All four
sets of results show some similarities. First, the predominant volatile
emla^lon was the fully chlorinated derivative of the test chemicals fed. In
the case of the chloromethanes, carbon tetrachloride was the main emission.
3.
-------
ACCESS PORTS
COOUNO WATER INLET I
AIR INLET
SAMPLING PORTS
COOUNG WATER OUTLET
VIEWING PORTS
REFRACTORY QUARL
VARIABLE VANE WINDBOX
WASTE/FUEL INLET
FIGURE 1 Turtutent Flame Reactor
-------
TFR WAU
PROBE
GAS DRIER
(CoS04)
IMPINCER
(.1 NORMAL PHOSPHORIC ACID)
CO MONITOR
(NDIR)
C02 MONITO
(NDIR)
02 MONITOR
(PARAftlABNCnC)
FIGURE 2 Contlnuoua Monitors
-------
I
TFRWALL
SAMPLE OAS
(CoS04)
TENAX SAMPLE TRAP
(COOLED)
OR/CAS METER
FIGURES TenaxSamplngTrain
-------
1A1
CCL4
IBS
•AUPl' NAMEB (1A1-MK 1A SAMPU 1)
S3 CtHt \ZA CTNIf KS C6HSO GS1 C2HCU C3 C2CI4
FIGURE 4 Mass Emteskxis, Chtoromethanes
Ugh Oxygen Testa
-------
1A1
1A2
1A3
SAMPLE NAMES (1A1-MIX1A SAMPLE 1)
oz GZ9 C7 tza ca4 CSS TCC
1B3
FIGURE 6 Mass Emissions, Chtoromothanes
Low Oxygen Tests
-------
Ml
CCL4
283
SAMPLE NAMES (2A1-MTX2A SAMPLE 1)
CZ3 ctHi S3 ciHsa an CIMCO
caai
FIGURE 6 Mass Emissions, Chlorinated C2s
High Oxygen Tests
-------
IB -
H-
11-
1«-
11 -
10-
§ -
7-
• -
• •
5.
3-
f.
1 -
.8
1A)
. tt
"i M
iH
TCL
M
. i
PI .^ _ ,. ,.
1A4 MA !*• 3A7 2t1 213 113
SAMPLE MAUCS (•.§. JA3-Utt 2A SAMPLE 2)
•I oa* QQI R Q 67 KS on * EZ1 ice en PP>
RQURE7 Mass Errisslons, Chlorinated C2s
Low Oxygen Tests
-------
In Che case of Che chlorinated €28 It was tetrachloroethylene. Perehlorln-
a ted compounds are ofCen observed as stable intermediates in the combustion
of chlorinated hydrocarbons as Indicated by the results of some of the
studies mentioned above (6,7,8,10).
The prevalence of perhalogenated organic compounds in Che exhaust re-
sults. In part, from how chlorine atoms affecC Che combosCion of hydrocar-
bons. Chlorine abstraction, followed by molecular fragmentation, is an
iaportant pathway for the combustion of chlorinated hydrocarbons under oxida-
tive conditions (11)* Even when present in relatively small amounts, chlor-
ine, in the form of Cl- and CIO- radicals, can become the most abundant reac-
tive species present in the combustion environment. This is true even under
oxidative conditions where OH would be expected to predominate (3,12). The
presence of chlorine radicals further alter* the combustion of chlorinated
hydrocarbons by preferentially abstracting hydrogen from chlorinated hydro-
carbons. In the case of the chlorinated Cj and €2 compounds studied here,
these reactions result in the formation of fully chlorinated intermediates.
If these chlorinated intermediates fall to react further it is possible
that they will react with the abundant Cl-and CIO-radicals Co ultimately form
a fully chlorinated and quite stable intermediate. Non-flame studies of
chloromethanes support this theory (3). In these studies, carbon tecrachlor-
ide appeared Co be more thermally stable when treated in a mixture Chan when
treated separately, a result that was attributed to carbon Cecrachloride
reforming from the other mixture components. A similar set of chemical
reactions seems to also be occurring when chlorinated f^s are burned under
oxidative conditions.
Under oxygen deficient conditions Che observed emissions were quite
different and appeared to be greatly affected by the formation of soot. A*
mentioned earlier, soot was the major emission under these conditions.
Although we did not quantify it, Che sooC formed in such large amounts Chat
it rapidly coated all of Che viewports. Unexpectedly, Che amounts of vola-
tile emissions produced under these conditions were much less than under
oxidative conditions. Benzene, a side product of soot formation, was among
the most abundant volatile emissions observed. Chlorinated hydrocarbons tend
Co form sooC more readily Chan other organic chemicals (13). This is be-
cause, under oxygen deficient conditions, chlorinated hydrocarbons Cend Co
decompose via Che elimination of HC1 from the parent molecule. Repeated HC1
elimination results in Che formation of unsatorated radicals which can con-
dense into large polycycllc aromatic hydrocarbons and, ultimately, soot (13).
Chlorocarbons with H:C1 ratios of 1 form soot Che easlesc (ibid).
The sooting tendency of these chlorocarbons mas observed during Che
tests on mixture IB which contained a relatively high level of methylene
chloride (Ct^C^) for which H:C1-1. Even under oxidative conditions large
quantities of soot were formed similar to Che oxygen deflcienC condlclons.
Results of the volatile organic sampling showed Chat benrene was also pro-
duced in very large amounCs.
When chlorinated C2& were tesCed In Che TFR under "low" oxygen con-
ditions, trichloroethylene was observed in significant amounts. This was
likely caused by Che elimination of HC1 from the 1,1,2,2-tetrachloroethane
Chat was present in test mixtures 2A and 2B.
Under conditions leading to the formation of large amounts of soot,
volatile emissions decreased. This was unexpected. Short chain pyrolysls
products were expected to be observed. It is unclear at this point whether
the feed materials formed soot instead of short chain pyrolysls products or
11.
-------
whether the short chain products that were formed adsorbed onto the soot that
was formed. There is evidence from other researchers that suggests that
soot does adsorb volatile compounds present in the combustion environment
(14,15).
Finally, Figures 8 and 9 show the volatile emissions observed plotted
against the carbon monoxide level and exhaust temperature respectively. No
correlations are visually apparent. There are two explanations that account
for this* First, it is likely that a number of factors not necessarily
related to either exhaust temperature or carbon monoxide level account for
the eaisslons observed. Secondly, fluctuations in temperature or in carbon
monoxide level which would bear on the emission of specific volatile organic
chemicals, are perhaps too transient and localized to be detected by the
measurement techniques employed in this study.
Conclusions
In summary, this study supported several conclusions. First, the re-
sults of studies of the combustion of chlorinated Cjs and C2S in non-flame,
flat flame, and shock tube devices were consistent with the emissions from the
TFR. Second, the formation of large amounts of soot affected the emissions of
volatile organic chemicals in unexpected ways that are not folly understood.
Third, the emissions observed did not seem to correlate with temperature or
carbon monoxide level in the exhaust*
More research is needed in this area* Farther studies of the combustion
chemistry of individual compounds and simple mixtures would aid in the
understanding of PIC formation. In addition, studies of the effect of soot
formation on the emission of volatile organic chemicals is needed in order to
determine the fate and transport of these materials. Finally, additional
evaluations of temperature, carbon monoxide level, and other potential com-
bustion indicators is needed to determine if there is a practical way to
continuously monitor the combustion process. Taken together, these studies
will help to evaluate the extent to which incineration can be applied as a
disposal technique for hazardous waste.
References
1. Dellinger, B., J.L. Torres, V. Rubey, D. Rail and J-L. Graham. Deter-
mination of the ThermalDecomposition Properties of 20Selected Organic
Compounds. EM 600/52-84-138, October, 1984.
2. Miller, D., V.A. Cundy and R.A. Matula. "Incinerabllity Characteristics
of Selected Chlorinated Hydrocarbons'*. Proceedings of the Ninth Annual
Research Symposium; Incineration and Treatment of Hazardous Hastes. EPA
600-9-84-015, July, 1984.
3. Taylor, P. and 8. Bellinger. "A Mechanistic Study of the Thermal Degra-
dation of A Hazardous Waste Mixture. I, POHC Degradation". Submitted to
Environmental Science and Technology, October, 1986.
4. Frenklach, M., J.P. Hsu, D.L. Miller and R.A. Matula. "Slock Tube
Pyrolysis of Chlorinated Hydrocarbons: Formation of Soot". Combustion
and Flame, Vol. 64, 1986.
5. Malanchuk, M. "Report on Non-Flame Hazardous Waste Thermal Destruction".
Intetoal Report, D.S. Environmental Protection Agency, Cincinnati, Ohio,
December, 1986.
6. Chang, W.D., S.B. Karra and S.N. Senkan. "Molecular Beam Mass Spectro-
meter Study of Trichloroethylene Flames". Environmental Science i Tech-
nology Vol. 20, No. 12, 1986.
12.
-------
0-
12-
» n "
2 10<
7-
9-
I •
2 *•
2-
1 •
0-
<
oa4
•
•
9
• •
9 « x « •
» » » •
t. ;« '"% ^._ . /; ;j ;
^J ». ^ ^. »^.x » 1^ » ^A i • A^ •
~^_y^yw^_ ^^^^»^^^^^^^^^^^^^^_^^^^^^^^^^^^^^^^^^^^^^^^^l^^^^^p^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1 200 400 000 M
CO (PPM •0*02)
4 C«H« • 07H10 4 C«19a X C2HCO 9 COO
FIGURE 8 CO Level va. Fraction Remalnin{
(Fraction RemoWng-Exhaust Rate/Feed Rate)
-------
^»
I
14
1*
S 10
ii
X •
* J
1
1
0
• oou
r
MM
-..*•• .
MM
« CMCO V 0104
RQURE 0 Exhaust Temperature vs. Fraction Remaining
(Fraction RemaWng-Exhaust Rate/Feed Rate)
-------
7. Sewer, D.W., J« Morse and V, Cundy. "PICs - A Consequence of Stable
Intermediate Formation During Hazardous Waste Incineration (Dichloro-
«w thane)". Hazardous Haste and Hataidous Materials Vol. 4, pp. 473-486,
1985.
8. Rose, D. aid S.M. Senkan. "On the Combustion of Chlorinated Hydrocar-
bons, I. TCE". Cofflbuitton Science 6 Technology Vol. 35, pp. 187-202,
Cordon and Breach Cclence Publishers Inc., Great Britain, 1983.
9. Overmoe, B., S. Chen and H.R. Seeker. Development of Laboratory-Scale
Reactors for Hazardous WasteIncineration. Final Report to EPA, Con-
tract 68-02-3633, October, 1983.
10. Hall, D,L., J.L. Graham, N. Stamatlades and B. Dellinger. 'The Thermal
Decomposition Characteristics of Chloroform", Fourth International Sym-
posium on Environmental Pollution, HI ami Beach FL, October 3-4, 1983.
11. Westbrook, C.K. "Inhibition of Hydrocarbon Oxidation in Laminar Flamea
and Detonations by Halogenated Compounds'. Proceedings of the nine-
teenthSymposium (International) on Combustion. The Combustion Insti-
tute, pp. 127-141, August, 1982.
12. We8tbrook, C.K. "Chemical Kinetics of Hydrocarbon Oxidation in Caseous
Detonations". Combustion and Flame Vol. 46 No. 2, May, 1982.
13. Welssman, M* and S. Benaen. "Pyrolysia of Methyl Chloride, A Pathway
in the Chlorine-Catalyzed Polymerization of 014"* International Journal
of Chemical Kinetics Vol. 16, pp. 307-333. John Wiley and Sons, 1984.
14. Van Dell, R. and G.V. Boggs. "Particulate Associated Partial Combustion
Products of Dlchlorobenrene". Unpublished.
IS. Newman, J.S. and J. Seleclak. "Characterization of Partlculates from
Diffusion rlamet". Combustion andFlame, Vol. 67, pp. 55-65, January,
1987.
15.
------- |