EPA/600/A-92/021
Experimental Investigation of PIC Formation in CFC Incineration
Robert E. Hall and Chun Wai Lee
Combustion Research Branch
Air and Energy Engineering Research Laboratory
United States Environmental Protection Agency
Research Triangle Parte, NC 27711
Garth R. Hassel
Energy and Environmental Research Corporation
18 Mason
Irvine, CA 92718
Jeffrey V. Ryan
Environmental Systems Division, Acurex Corporation
4915 Prospectus Drive
Durham, NC 27713
Abstract
Bench-scale tests were performed to characterize the combustion emissions from
chlorofluorocarbon (CFC-11 and -12) incineration. The destruction efficiencies (DEs)
of the CFCs and the major products of incomplete combustion (PICs) from each
CFC were determined. DEs of at least 99.999% can be attained repeatedly for both
CFCs. Major PICs identified were non-halogenated, with toluene and xylene being
found most frequently. PIC concentrations were independent of the CFC
concentration in the fuel. Sampling was performed one time to screen for
polychlorinated dibenzo-p-dioxins and poly.chlorinated dibenzofurans
(PCDDs/PCDFs) and polyaromatic hydrocarbons (PAHs) while incinerating CFC-
12. Even with a DE of greater than 99.999%, high levels of PCDDs/PCDFs were
formed. The extensive PCDD/PCDF formation apparently occurred in the absence of
entrained particulate matter.
Introduction
Chlorofluorocarbons (CFCs) are implicated in the depletion of stratospheric ozone, and are also
contributors to global warming. As a result of the Montreal Protocol that will curtail the production
of the traditional CFCs and halons, it may be necessary to destroy considerable quantities of these
materials in order to reduce the current "bank" of the older (more stratospheric ozone depleting) CFCs
and halons. A United Nations Environment Programme (UNEP) Ad-Hoc Technical Advisory
Committee on CFC Destruction Technologies has recently been formed to evaluate the most
appropriate destruction technologies for ozone depleting substances (ODS), including CFCs and halons.
Incineration is currently the only widely used CFC destruction technology (1). A sound understanding
of the combustion emission characteristics of CFC incineration is needed for CFC destruction
technology evaluation.
CFCs are thermally stable because of strong carbon-halogen bonds, which are resistant to oxidation.
Thus, CFCs are difficult to destroy by incineration. Furthermore, the halogen free radicals generated as
intermediate species during the combustion of CFCs are well known flame retardants. Formation of
products of incomplete combustion (PICs) becomes possible as the flame propagation is retarded. PIC
formation, in general, is a poorly understood phenomenon associated with waste incineration, which is
further complicated for CFC incineration processes since only limited data on PIC formation are
available.
The release of halogenated aromatic compounds, which have been shown to be toxic in general, from
CFC incineration may cause environmental concerns. It has been shown that the generation of
halogenated aromatics from CFC decomposition is thermodynamically favorable and such compounds
have been identified in a bench-scale CFC decomposition study (2). Formation of a wide variety of
PICs from thermal destruction of chlorinated hydrocarbons (CHCs) has been shown in a theoretical
analysis (3), and observed by experimental investigations (4, 5, 6). The extent of PIC formation in
-------
CFC incineration is not known, and was investigated in this study. The combustion emission
characterization data from a bench-scale CFC incineration study are reported in this paper.
Experimental
A turbulent flame reactor (TFR), which has been used in a previous hazardous waste incineration study
(7), was employed to perform the incineration tests. The TFR (Figure 1) consists of a swirling air/gas
injector burner firing into a 30 cm ID by 60 cm long water-cooled stainless steel enclosure with
refractory quarl placed inside. Following the TFR is an afterburner (AB) section consisting of a 30 cm
ID refractory-lined chamber with three fuel/air injectors near its base. The fuel/air injectors are spaced
120° radially apart and Fired toward the axis of the cylindrical chamber. The TFR/AB system is shown
in Figure 2. The flue gas exits through a 9.1 m long stainless steel duct to a venturi scrubber. The
HC1 and HF gases produced were neutralized by NaOH scrubbing liquor. The TFR was fired at 0.007
GJ/lir at a stoichiometric ratio of 1.1 (2% O2) using propane gas fuel and preheated air at 190 °C.
Propane and air were replaced by natural gas and compressed oxygen, respectively, in the afterburners in
order to achieve the target AB conditions of 982 °C and 7% excess air. Incineration tests were
performed by metering one of the CFCs (11 or 12) at a time into the primary flame. The CFCs were
certified as 99.7% pure.
A continuous emission monitoring (CF.M) system, which included O2, CO2, CO, NOx as NO, and
total unburned hydrocarbons (THCs), was used to monitor the flue gas downstream of the afterburner.
Combustion gas samples were taken upstream and downstream of the afterburner simultaneously using
EPA Method 18, "Measurement of Gaseous Organic Compound Emissions by Gas Chromatography."
The sampling procedure involves drawing combustion gas into a tedlar bag by evacuating a sealed
chamber containing the bag. Sampling points are indicated in Figure 2. Because of the high HCi and
HF concentrations in the combustion gases, an impinger of 1 N NaOH solution followed by an
impinger of deionized water were placed upstream of the tedlar bag, per Method 18. The tedlar bag gas
samples were analyzed for the CFCs to determine the destruction efficiencies (DEs) of the CFCs. An
attempt was made to quantify many of the major PICs in the samples. A known volume of the tedlar
bag gas sample was transferred to an organic trapping sorbent, desorbed, and cryogenically concentrated
on column for gas chromatographic/mass spectrometer (GC/MS) analysis. The analytical method used
ensured adequate detection levels suitable for determination of CFC DEs.
A semivolatile organics sampling train was run downstream of the afterburner one time at the baseline
condition (8.3 vol.% CFC-12) with the objective of screening for PCDDs/PCDFs and PAHs.
Sampling was performed according to California Air Resources Board (CARB) Method 428,
"Determination of Polychlorinated Dibenzo-p-dioxin (PCDD), Polychlorinated Dibenzofuran (PCDF),
and Polychlorinated Biphenyl Emissions from Stationary Sources." The combined components of the
train were screened for PCDDs and PCDFs using a high resolution gas chromatography/low resolution
mass spectrometry (HRGC/LRMS) method similar to EPA SW-846, "Test Methods for Evaluating
Solid Waste," Method 8280. A sample fraction was also screened for the presence of PAHs. A gas
chromatograph equipped with a flame ionization detector (FID) was configured for this analysis. The
retention times of 16 chromatographably resolved PAHs were compared to the retention times of
analytes present in the sample.
Results and Discussion
The flammability limits of the two CFCs were first determined in order to perform two incineration
tests at concentrations close to the two CFCs' flammability limits. Flammability limits were
determined by slowly increasing the flow of CFC into the steady stream of propane fuel primary flame
until the flame extinguished. The flammability limits of both CFCs (39 vol.% for CFC-11 and 58
vol.% for CFC-12) are significantly higher than the concentrations of CFCs (1 to 5%) normally
injected into a commercial incinerator. Because the flammability limits were higher than anticipated,
such high CFC concentrations for test durations would have the unacceptable risks of high acid gas
emissions and damage to the facility. The CFC-12 incineration tests were decided to be performed at
concentrations of 1.3, 8.3, and 14.0 vol.% in fuel. The CFC-11 tests were performed at concentrations
of 1.3 and 13.0 vol.%.
The test conditions for CFC incineration are summarized in Table 1. It was found that the THC
concentrations were below detection limits under all test conditions. The NO concentrations were
varied and showed no apparent dependence on any of the measured parameters. The concentrations of
2
-------
HC1, CI2, HF, and F2 were not measured in the present study. They were estimated by using the
NASA (CET-85) thermodynamic equilibrium code. Results of the equilibrium analysis under CFC-12
test conditions are shown in Figure 3. The TFR conditions were 787 °C, 2% O2 and the AB
conditions were 982 °C, 7% O2. The results indicate that fluorine favors combination with hydrogen
more so than does chlorine, as the concentrations of species F2 and F were consistently less than 10"11
mole fraction. On the other hand, chlorine tends to form significant concentrations of both
monoatomic and diatomic chlorine. Monoatomic chlorine is thermodynamically favored over diatomic
chlorine at higher temperature.
Results of DEs are shown in Table 2. When no CFC was detected, the detection limit was used to
calculate the DE and a ">" sign was placed in front of the DE. Note also that only 99.99% DE could
be determined for the low CFC concentration tests. This is a mathematical, not physical, limitation.
The generally high DEs in a primary flame observed in the present study were also reported for the
thermally stable organics using the same TFR (7).
Table 3 summarizes the PIC identification results. PIC concentrations were generally low. Some
PICs most likely reflect interaction of the CFC with the propane fuel, such as hexafluoropropene,
trichlorofluoroethane, and chloroform. Transformation of CFCs to CHCs as PICs was evidenced.
Toluene and xylene were the two most frequently occurring PICs. The non-halogenated aromatic
structures of these PICs may be indicative of their formation from the pyrolysis of the propane fuel
alone. Benzene and its derivatives, chlorinated species in particular, have been found to be the most
dominant PICs produced from thermal destruction of chlorinated organics (4, 5). The high halogen
input, chlorine in particular, with CFC incineration may promote aromatics formation. It is well
known that chlorine addition in hydrocarbon flames enhances aromatics formation leading to sooting
(8). Reaction pathways for the formation of aromatics from chlorinated alkanes have been proposed
(9). The majority of the PICs identified in the present investigation were non-halogenated. The most
interesting exception was the presence of CFC-11 as a PIC of CFC-12. The transformation of CFC
and the formation of aromatics in CFC incineration observed in the present study are evidences that
complex reaction pathways play an important role in the thermal incineration of CFCs.
Overall, PICs formed at least as frequently in the secondary flame as in the primary flame. For CFC-
12 the greater number of PICs appeared when it was at its highest concentration in the fuel. However,
CFC-11 tests yielded the seemingly odd result of many PICs at low CFC concentration and no detected
PICs at high CFC concentration. The non-halogenated PICs may have formed as a result of flame
anomalies unrelated to the presence of the CFCs. Comparison of PIC results to the data in Table 1
shows that such anomalies were not signaled by temperature or CEM readings.
Table 4 list the PCDD/PCDF emissions. Figure 4 presents the distribution of isomer groups. The
PCDD/PCDF levels were surprisingly high. It has been suggested that PCDD/PCDF generation tends
to be promoted by large (> 10 |im) particulate matter (10). Entrained particulate matter for the tests in
this study using gaseous fuels should have been negligible. The filter on the sampling train used to
collect the dioxin sample was observed immediately after the test to have turned a light beige color,
with no large particulate discemable. A more thorough examination of the filter was not possible
because, when the filter housing cooled down during the few minutes between testing and sample
recovery, the condensate from the residual flue gas in the housing apparently contained enough HF to
dissolve the glass filter. Only remnants of the partially dissolved filter remained at the time of sample
recovery. The contents of the filter housing were combined with the rinse of the sampling glassware.
The presence of PCDDs/PCDFs from the field blank analysis was not detected, indicating no
contamination from components of the sampling apparatus, recovery agents, or the ambient air. One
could speculate that the HF in the flue gas could have etched the sampling glassware, particularly the
last piece before the XAD module because condensation can occur there, releasing dioxin buildup from
previous sampling. This argument assumes that the standard glassware preparation and sample
recovery methodologies are inadequate. The field blank was taken before the actual sample was taken,
and the same set of glassware was used for both the field blank and the sample. It is extremely
unlikely that the microgram levels of PCDDs/PCDFs detected from the sampling run were caused by
residues on glassware that had previously survived treatment using the same recovery method. To
dispel the possibility of an acid etching artifact during sample analysis, a test was performed where HF
was added to a method blank and analyzed. HF was added to a commonly used sample extraction
apparatus and refiuxed as an ordinary sample. Acid etching of the extraction glassware was similar to
the etching observed during extraction of the original sample. However, no PCDD/PCDF material was
present in the final analyzed extract.
3
-------
The lack of contamination and particulate in the incineration system implies that a significant
PCDD/PCDF formation mechanism exists that is not catalyzed by entrained particulate matter. The
PCDDs/PCDFs probably formed either in the primary flame and survived the afterburner via cold
pathways along the walls and near the burner jets or in the long section of the 10 cm stainless steel
duct, which was about 6 m in length from the afterburner exit to the point of sampling. Flame-mode
PCDD/PCDF formation may have been enhanced by the relatively low TFR flame temperature. Note
from Figure 2 that the TFR wall at the location of the pyrometer port is no longer refractory, but is
water-cooled stainless steel. Although the low flame temperature did not prevent 99.999% DE of the
CFC, it may have had an influence on PCDD/PCDF formation.
Formation of PCDDs/PCDFs through gas-phase homogeneous reactions remains controversial. Based
on the results of numerical calculations on the kinetics of a model PCDD formation mechanism, it has
been suggested that the probability of gas-phase formation of PCDDs is likely to be very low (11).
However, a recent study of dioxin formation reaction mechanism based on fundamental thermochemica!
principles demonstrated that the gas-phase formation is possible (12). Elementary reaction pathways
leading to the formation of PCDDs at both high and low temperature regimes in incineration are
identified by the study. It has been predicted that if PCDDs are already present in the gas phase, the
formation of PCDFs from PCDDs in the gas phase is more likely (11). The much higher levels of
PCDFs compared to those of PCDDs measured in this study are apparently consistent with the gas-
phase formation mechanisms.
Considering the possibility of heterogeneous formation mechanisms, the temperature of the flue duct
ranged from about 620 °C at its entrance to 110 °C at the sampling point PCDD/PCDF formation
has been shown to be enhanced in the temperature range of about 200-370 °C and in the presence of
some metals, especially copper. The flue duct in the experimental system had previously been exposed
to parts per million levels of copper, which may have condensed inside the duct. The long, narrow,
metal duct might provide the conditions, including sufficient surface area without entrained particulate,
to catalyze PCDD/PCDF formation. No PAH was detected in the PCDD/PCDF sample.
Conclusions
Based on bench-scale test results, destruction efficiencies of at least 99.999% are attainable for both
CFC-11 and CFC-12 even from a relatively low temperature flame. The use of an afterburner was not
necessary to attain high DE of CFC-11 and CFC-12. The results were repeatable. Most PICs
identified were non-halogenated species. PIC concentrations were independent of the concentration of
CFCs in the fuel. The wide variety of PICs, ranging from aliphatic to aromatic species, observed in
the present study is a strong indication that CFC destruction during the thermal incineration occurs
through complex reaction pathways. DEs of greater than 99.999% still allow for high generation of
PCDDs/PCDFs. Extensive PCDD/PCDF formation when incinerating CFC-12 was apparently
independent of entrained particulate matter. Further studies are needed to investigate the levels of
PCDDs/PCDFs generated in pilot and commercial scale CFC incinerators.
References
(1) Dickerman, J. C., Emmel, T. E., Harris, G.E. and Hummel, K. E., "Technologies for CFC/Halon
Destruction," EPA-600/7-89-011 (NITS PB90-116955), EPA, Air and Energy Engineering Research
Laboratory, Research Triangle Park, NC, October 1989.
(2) Wakabayashi, et al, "Decomposition of Halogenated Organic Compounds by r.f. Plasmas at
Atmospheric Pressure," paper presented at the 9th International Symposium on Plasmas Chemistry,
September 4-8, 1989, Pugnochiuso, Italy.
(3) Chang, D. P. Y„ Mourningham, R. E. and Huffman, G. L„ "An Equilibrium Analysis of Some
Chlorinated Hydrocarbons in Stoichiometric to Fuel-rich Post-flame Combustion Environments," J.
Air Waste Manage. Assoc., 1991,41: 947-944.
(4) Graham, J. L., Hall, D. L. and Dellinger, B., "Laboratory Investigation of Thermal Degradation of
a Mixture of Hazardous Organic Compounds," Environ. Sci. Technol., 1986,20:701-710
(5) Tayor, P. H. and Dellinger, B., "Thermal Degradation Characteristics of Chloromcthane Mixtures,"
Environ. Sci. Technol., 1988, 22:438-447.
4
-------
(6) Fisher, E, M„ Kashland, C. P., Hall, M. J., Sawyer, R. R. and Lucas, D., "Experimental and
Numerical Study of the Thermal Destruction of C2H5CI," Twenty-Third Symp. (Intl.) on Comb./The
Combustion Institute, Pittsburgh, 1990, pp. 895-901.
(7) Kramlich, J. C., Heap, M. P., Seeker, W. R. and Samuelsen, G. S., "Flame-mode Destruction of
Hazardous Waste Compounds," Twentieth Symp. (Intl.) on Comb./The Combustion Institute,
Pittsburgh, 1984, pp. 1991-1999.
(8) McKinnon, I. T. and Howard, J. B., "Application of Soot Formation Model: Effects of Chlorine,"
Combust Sci. and Tech., 1990, 74:175-198.
(9) Tsang, W., "Mechanisms for the Formation and Destruction of Chlorinated Organic Products of
Incomplete Combustion," Combust. Sci. and Tech., 1990,74:99-116.
(10) Lanier, W. S., von Alten, T. R. and Kilgroe, J. D., "Assessment of Trace Organic Emissions
Test Results from the Montgomery County South MWC in Dayton, Ohio," presented at the American
Flame Research Committee 1990 Fall International Symposium on NOx Control, Waste Incineration
and Oxygen Enriched Combustion, San Francisco, October 8-10,1990.
(11) Shaub, W. M. and Tsang, W„ "Dioxin Formation in Incinerator," Environ. Sci. TechnoL, 1983,
17:721-730.
(12) Ritter, E. D. and Bozelli, J. W., "Reactions of Chlorinted Benzenes in H2 and H2/Q2 Mixtures:
Thermodynamic Implications on Pathways to Dioxin," Combust. Sci. and Tech., 1990,74:117-135.
5
-------
Table 1. Test Conditions
Vol.%
Dry Flue Gas Composition
Primary Flame
Secondary Flame
in
02
CO2
CO
NO
THC*
Pyrometer
Pyrometer
TC-1
TC-2
TC-3
CFC
Fuel
%
%
ppm
ppm
ppm
°C
99.9996
AB exit
76
383.3
33826
<63
<0.0021
>99.9994
12
8.3
TFR exit
76
383.3
24569
<63
<0,0015
>99.9996
AB exit
76
383.3
33917
<63
<0.0021
>99.9994
12
8.3
TFR exit
76
383.3
24569
<63
<0.0015
>99.9996
AB exit
76
383.3
34099
<63
<0.0021
>99.9994
12
14
TFR exit
133
667.2
24682
93
0.0023
99.9997
AB exit
133
667.2
33948
<63
<0.0021
>99.9997
12
1.3
TFR exit
11
56.79
24439
<63
<0.0015
>99.9997
AB exit
11
56.79
33686
<63
<0.0021
>99.9996
11
13
TFR exit
127
725.8
24671
<63
<0.0016
>99.9998
AB exit
127
725.8
33990
<63
<0.0021
>99.9997
11
1.3
TFR exit
11
64.52
24439
181
0.0044
99.993
AB exit
11
64.52
33524
<63
<0.0021
>99.997
6
-------
Table 3. Summary of Products of Incomplete Combustion
CFC-12
Vol.% CFC in
8.3
8.3
8.3
14
1.3
Fuel •
Primary Flame
CFC, ng/L
<63
<63
<63
93
<63
DE, %
>99.9996
>99.9996
>99.9996
99.9997
>99.997
PICs, ng/L
ND
CFC-11, 81
ND
CFC-11, 81
chloroform.
unknown
213
hydrocarbons, 44,
pentane, 225
256, 25, 225
toluene, 50
toluene, 488
xylene, 50
bicyclo[4,2,0]-
octa-l,2,5-triene,
Secondary
644
Flame
CFC, ng/L
<63
<63
<63
<63
<63
Overall DE, %
>99.9994
>99.9994
>99.9994
' >99.9997
>99.996
PICs, ng/L
unknown
hexafluoro-
hexafluoropropene
CFC-11, 250
chloroform,
hydrocarbon,
propene, 781
, 2400
213
69
toluene, 31
unknown
unknown
methyl propene, 550
xylene, 19
hydrocarbon,
hydrocarbon, 494
pentane, 581
1000
carbon disulfide/
trichlorotrifluoro -
trichloro-
ethane, 325
trifluoroethane,
toluene, 219
438
ethyl benzene, 113
xylene, 369
ND = not detected (continued)
Table 3. Summary of Products of Incomplete Combustion (continued)
System Blank
CFC-11
Vol. % CFC in Fuel
0
1.3
13
Primary Flame
CFC, ng/L
<63
181
<63
DE, %
NA
999.993
>99.9998
PICs, ng/L
ND
unknown hydrocarbons,
ND
550,806,5600
methyl propene, 713
trichlorotrifluoroethane,
413
cyclohexane, 556
ethylbenzene, 144
xylene, 450
Secondary Flame
CFC, ng/L
<63
<63
<63
Overall DE, %
NA
>99.997
>99.9997
PICs, ng/L
ND
unknown hydrocarbons,
ND
44, 81
toluene, 19
xylene, 31
NA = not applicable
ND = not detected
Table 4. PCDD/PCDF Results for 8.3 Vol.% CFC-12 in Propane
Flue Gas
Generation
Concentration
Rate
(ig/dscm
|ig/g of CFC
Total PCDD
3.13
0.28
Total PCDF
20.70
1.82
PCDD/PCDF
23.80
2.09
7
-------
Cooing Water
Sight Glass
Swf vanes
Eumer
Pftff>£f¥ EPA MetfxxJ
isSarrpfePort
Secenda/y EPA M«tftod 18 ani
Continuous ErrissScn Wcnttcrtng
Sampie Ports
AB Pyrometer Pert
Mlzzcu
Reuacio
-------
TECHNICAL REPORT DATA
i\Hi Jirvi-j ir o Is (Please read Instructions on the reverse before complet*
1. REPORT NO.
EPA/600/A-92/021
2.
3.
4. TITLE ANO SUBTITLE
Experimental Investigation of PIC Formation in CFC
5. REPORT DATE
Incineration
6. PERFORMING ORGANIZATION CODE
7-AUTHOR{Si R.E.Hall and C.W. Lee (EPA/AEERL),
G. R. Hassel (EERC), and J. Ryan (A cur ex)
8. PERFORMING ORGANIZATION REPORT NO.
S, PERFORMING ORGANIZATION NAME ANO ADDRESS
Energy and Environmental Research Corporation
10. PROGRAM ELEMENT NO.
18 Mason
Irvine, California 92718
-
11. CONTRACT/GRANT NO.
68-C0-0094
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 7-9/91
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notes project officer is Chun Wai Lee, Mail Drop 65, 919/541-
7663. Presented at ENVIRONMZXASIA/WATZRMZXASIA92 Conference, Singapore,
2/18-21/92.
i6. abstract paper g£ves results of bench-scale tests to characterize combustion
emissions from chlorofluorocarbon (CFC-11 and -12) incineration. The destruction
efficiencies (DEs) of the CFCs and the major products of incomplete combustion
(PICs) from each CFC were determined. DEs of at least 99.999% can be attained re-
peatedly for both CFCs. Major PICs identified were non-halogenated: toluene and
xylene were found most frequently. PIC concentrations were independent of the CFC
concentration in the fuel. Sampling was performed one time to screen for polychlor-
inated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/PCDF) and poly-
aromatic hydrocarbons (PAHs) while incinerating CFC-12. Even with a DE of grea-
ter than 99. 999%, high levels of PCDD/PCDF were formed. The extensive PCDD/
PCDF formation apparently occurred in the absence of entrained particulate matter.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Incinerators
Halohydrocarbons
Combustion
Emission
Sampling
Fiirarics
Pollution Control
Stationary Sources
Chlorofluorocarbons
Destruction Efficiency
Products of Incomplete
Combustion
Dioxins
13B
07C
21B
14G
14B
18. DISTRIBUTION STATEMENT
Release to Public
IS. SECURITY CLASS (This Report}
Unclassified
21. NO. OF PAGES
ta
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |