The impact of turbulent mixing on the oxidation of a chlorinated hydrocarbon

Article abstract of DOI:10.1021/es9702935

Combustion of chlorinated wastes can lead to the formation of hazardous byproducts. Rates of mixing of fuel and air in combustion systems can have an impact on the composition of the byproducts. Methyl chloride and methane were burned in a turbulent diffusion flame in a combustion wind tunnel with a coflow of air. Reynolds numbers were varied from 3500 to 7200. A water- cooled sampling probe was used to obtain gas samples from within the flame at a number of locations and at various Reynolds numbers. The postflame gases and particulate matter were trapped above the flames with sorbent tubes and filters. The samples were desorbed and analyzed for aromatic species and other trace products of incomplete combustion. Destruction of the methyl chloride was essentially complete for all the Reynolds numbers that were studied. Small amounts of low molecular weight chlorinated compounds were fOUnd within the flame and in the postflame gases. The major chlorinated species in the postflame gases was chloronaphthalene. Low Reynolds number flames were found to Yield larger amounts of aromatic and chlorinated aromatic species than the high Reynolds number flames. Fluoranthene was present in greater amounts on the soot particles at lower Reynolds numbers, suggesting that the rate of mixing of reactants could have an impact on the toxicity of the combustion byproducts. Combustion of chlorinated wastes can lead to the formation of hazardous byproducts. Rates of mixing of fuel and air in combustion systems can have an impact on the composition of the byproducts. Methyl chloride and methane were burned in a turbulent diffusion flame in a combustion wind tunnel with a coflow of air. Reynolds numbers were varied from 3500 to 7200. A water-cooled sampling probe was used to obtain gas samples from within the flame at a number of locations and at various Reynolds numbers. The postflame gases and particulate matter were trapped above the flames with sorbent tubes and filters. The samples were desorbed and analyzed for aromatic species and other trace products of incomplete combustion. Destruction of the methyl chloride was essentially complete for all the Reynolds numbers that were studied. Small amounts of low molecular weight chlorinated compounds were found within the flame and in the postflame gases. The major chlorinated species in the postflame gases was chloronaphthalene. Low Reynolds number flames were found to yield larger amounts of aromatic and chlorinated aromatic species than the high Reynolds number flames. Fluoranthene was present in greater amounts on the soot particles at lower Reynolds numbers, suggesting that the rate of mixing of reactants could have an impact on the toxicity of the combustion byproducts.

Full text of DOI:10.1021/es9702935

Environ. Sci. Technol. 1998, 32, 1265-1268
of chlorine in the waste stream poses some significant
The Impact of Turbulent Mixing on
the Oxidation of a Chlorinated
Hydrocarbon
problems for the operation of incinerators and other thermal
treatment systems. Consequently, a considerable amount
of attention has been given to issues related to the combustion
of chlorinated hydrocarbons. Senkan and co-workers (7, 8)
have studied chlorinated premixed flames for some years.
They used experimental results in laminar premixed flames
to verify numerical modeling studies of complex chemistry.
Senser et al. (9) investigated dichloromethane-methane-air
premixed flames stabilized on a flat flame burner. They found
that chloroform was an important intermediate species.
However, their analysis did not extend to PAH species.
Detailed models of the oxidation of chlorinated hydrocarbons
are now available. However, in general, these models do
not include pathways to larger hydrocarbons and PAH. They
have not been applied to calculations of the nonpremixed
combustion of chlorinated hydrocarbons.
G O S U YA N G , A . D A N I E L J O N E S , A N D
I A N M . K E N N E D Y*
Department of Mechanical and Aeronautical Engineering and
Facility for Advanced Instrumentation University of
California, Davis, California 95616
Combustion of chlorinated wastes can lead to the formation
of hazardous byproducts. Rates of mixing of fuel and air
in combustion systems can have an impact on the
composition of the byproducts. Methyl chloride and methane
were burned in a turbulent diffusion flame in a combustion
wind tunnel with a coflow of air. Reynolds numbers were
varied from 3500 to 7200. A water-cooled sampling probe
was used to obtain gas samples from within the flame at a
number of locations and at various Reynolds numbers.
The postflame gases and particulate matter were trapped
above the flames with sorbent tubes and filters. The
samples were desorbed and analyzed for aromatic species
and other trace products of incomplete combustion.
Destruction of the methyl chloride was essentially complete
for all the Reynolds numbers that were studied. Small
amounts of low molecular weight chlorinated compounds
were found within the flame and in the postflame gases.
The major chlorinated species in the postflame gases was
chloronaphthalene. Low Reynolds number flames were
found to yield larger amounts of aromatic and chlorinated
aromatic species than the high Reynolds number flames.
Fluoranthene was present in greater amounts on the soot
particles at lower Reynolds numbers, suggesting that the
rate of mixing of reactants could have an impact on the
toxicity of the combustion byproducts.
Halogens are well-known to inhibit combustion reactions.
Hence, flame stability may become a problem under condi-
tions of high rates of fuel/ air mixing. On the other hand,
rapid mixing of reactants is desirable to lessen the amounts
of soot and associated polycyclic aromatic hydrocarbons that
are formed. Although the role of mixing in the release of
hazardous combustion byproducts has been acknowledged
(10), little attention has been directed toward nonpremixed
or diffusion flames. Some early studies (11, 12) examined
the impact of a halogen on a diffusion flame in a global sense;
they reported critical fuel flow rates at extinction in a
counterflow diffusion flame. More recently, Yang and
Kennedy (13) performed experiments in counterflow diffu-
sion flames of methyl chloride and trichloroethylene in
methane. They measured critical strain rates in these flames
at extinction. They reported a relatively simple dependence
of critical strain rate on chlorine loading. Yang and Kennedy
(
14) interpreted measurements in a laminar trichloroethylene
flame in terms of a diffusion flamelet model. Their results
are potentially useful in applications to modeling combustion
in incinerators.
These studies were undertaken in laminar flames. There
is a paucity of information on the formation and emission
of PAH and chlorinated PAH in turbulent flames. A notable
exception is the work of Brouwer et al. (15) who investigated
the impact of turbulent mixing on the formation of PICs by
injecting methyl chloride into the flow in a plug flow reactor.
The results showed that the rate of turbulent mixing and the
rate of chemical reactions were both important in the
formation of PICs in this flow.
Introduction
The combustion of hydrocarbon fuels has been recognized
for some time as a source of material that may represent a
hazard to human health. Combustion under fuel rich
conditions leads to the formation of polyaromatic hydro-
carbons (PAH) and other products of incomplete combustion
One well-known impact of the chlorine is to enhance the
formation of soot and PAHs (16-18). This process is affected
by the mixing rate between the fuel stream and the air stream
and by the rate of the chemical kinetics (15, 19). High
Reynolds numbers improve the mixing rate of the reactants
and hence minimize the time available for the production
of particulates of soot and PAHs. However, high mixing rates
may also lead to flame instability and, consequently, may
result in flame lift-off or blow-off with a reduction in
destruction efficiency. An ideal window of mixing rates may
exist for which high destruction efficiencies are maintained
with minimal emissions of PICs. The present investigation
was undertaken with a view to providing basic information
on the formation of undesirable combustion byproducts in
a well-defined chlorinated, turbulent diffusion flame. Similar
information is not available in the literature, particularly from
an experiment that is sufficiently well-defined to be of use
to turbulence modelers.
(
PICs). These compounds may be emitted from combustion
sources either in the gas phase or as a condensed or absorbed
phase on soot particles. For example, Kozinski (1) recently
reported a large number of PAHs that were emitted along
with soot from the combustion of a heavy fuel oil. Some of
the PAHs that are formed during hydrocarbon combustion
are known to exhibit mutagenic activity (2-5). In particular,
Howard et al. (2) noted the significant mutagenic activity of
fluoranthene, a common PAH formed during hydrocarbon
combustion. When chlorine is added to a hydrocarbon flame,
the potential exists for the creation and emission of chlo-
rinated PAHs; chlorinated dioxins and furans are particularly
toxic (6).
Many hazardous waste sites contain large amounts of
chlorinated solvents such as trichloroethylene. The presence
*
Corresponding author.
S0013-936X(97)00293-9 CCC: $15.00
Published on Web 03/21/1998

1998 Am erican Chem ical Society
VOL. 32, NO. 9, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1 2 6 5

FIGURE 2. Centerline profiles of mean mole fractions of major
species, Re ) 4500.
cluding HCl. Filters were weighed to determine the mass of
collected particulates and transferred to a separate extraction
thimble containing Na
Soxhlet extracted for 16 h with 250 mL of dichloromethane
CH Cl ); an internal standard solution containing terphenyl-
14 was added. The volume of each extract was reduced to
2 3
CO . Sorbents and filters were each
(
2
2
d
a volume of 10 mL using a rotary evaporator, and the extract
was divided into two equal aliquots, one of which was
archived at -20 °C. One aliquot was reduced to a volume
of 200 µL under a stream of nitrogen and analyzed by GC/
MS using a VG Trio-2 mass spectrometer operated using 70
eV electron ionization. A 15 m DB-5 column (0.25 mm i.d.,
FIGURE 1. Wind tunnel.
Experiments
3
The fuel was a mixture of 50% methyl chloride (CH Cl by
0
2
.25 µm film) was programmed from 40 °C (2 min hold) to
85 °C at 8 °/ min. Preliminary experiments identified
volume) and methane. It was supplied to a 5 mm diameter
nozzle that was located on the centerline of a vertical
combustion wind tunnel (Figure 1). A 1 mm wide annulus
supplied H to stabilize the hydrocarbon flame. The total
2
hydrocarbon fuel flow rates varied from 2.9 L/ min to 6 L/ min
to achieve Reynolds numbers between 3500 and 7200. The
numerous combustion byproducts so that sensitive detection
of target analytes could be achieved using GC/ MS and
selected ion monitoring. Selected ion monitoring was
performed for molecular ions of phenylacetylene (m/ z 102),
styrene (m/ z 104), chlorophenylacetylene (m/ z 136), chlo-
rostyrene (m/ z 138), naphthalene (m/ z 128), acenaphthylene
H
2
flow rate was held constant at 0.4 L/ min in all cases. This
flow rate of H was found to be just sufficient to hold the
flames attached to the burner for all the Reynolds number
conditions. Tests were performed in which the flow of H
2
(
m/ z 152), chloronaphthalenes (m/ z 162), chloroacenaph-
thylene (m/ z 186), phenanthrene (m/ z 178), pyrene and
fluoranthene (m/ z 202), benz[a]anthracene and chrysene
2
was increased by small amounts; the flame and the con-
centrations of species within the flame were not affected
measurably. The air speed through the tunnel was 5 m s
for all the measurements. The axial pressure gradient along
the tunnel was approximately zero.
Gas samples from within the flame were obtained with
a water-cooled, stainless steel probe that was operated in an
approximately isokinetic manner. The probe opening was
(
m/ z 228), and the benzofluoranthenes, benzopyrenes, and
perylene (m/ z 252) in addition to the internal standard (m/ z
44). Quantitative analyses are based upon relative response
-
1
2
factors determined experimentally for the various analytes.
Ions were also monitored at m/ z 236 (chloropyrenes and
chlorofluoranthenes) and m/ z 286 (chlorinated benzopy-
renes, benzofluoranthenes, and perylene), but no monochlo-
rinated analogues of these PAHs were observed. Limits of
detection were approximately 0.1 ng/ sample.
2
.5 mm. Samples were stored in a glass sample bottle and
back filled with He so that condensation of water vapor did
not occur. The samples were analyzed later on a gas
chromatograph for the major stable species in the flame such
Results and Discussion
The methyl chloride/ methane flames were not fully turbulent
at a Reynolds number of 3500. The emissions from this flame
were used as an indication of the possible results of low
mixing rates in a transitional flow. A flame with a Reynolds
number of 4500 was turbulent in the sense that the flame
length did not shorten significantly as the Reynolds was
increased. At a Reynolds number of 7500, the flame was
only barely attached to the burner and the flame noise was
significant. The flame was sensitive to any perturbations
under these conditions. As a result, in-flame sampling was
not attempted for the highest Reynolds number conditions.
2 2 2
as N , O , CO , CO, etc. Measurements of HCl were not
attempted in view of the considerable uncertainty that
surrounds sampling of this species with a metal probe in
high-temperature gases. Furthermore, the results are re-
ported on a dry basis.
Samples of the postflame gases and particulates were
obtained by locating the sampling probe 500 mm above the
visible flame tip. The samples were drawn through a sorbent
tube that was prepared by packing 100 mm lengths of Pyrex
glass tubing (12 mm o.d.) with 3.5 g of Carbotrap C. Glass
wool plugs were inserted into both ends. Particles were
collected on a Teflon filter.
The water-cooled sampling probe was used to extract
gases from within the flames at different axial and radial
locations. The mole fractions of stable species along the
centerline of a flame at a Reynolds number of 4500 are shown
After sample collection, all sorbent and packing materials
were removed from the glass tubes and added to a cellulose
extraction thimble containing 10 g of anhydrous sodium
in Figure 2. It should be noted that H
2
O and HCl were not
2 3
carbonate (Na CO ) which neutralized trapped acids, in-
measured; the mole fractions of the reported species sum to
1
2 6 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 9, 1998

TABLE 1. Material Extracted from Soot Particulate in
picograms per liter of Sampled Gas
Reynolds number,
mass of particulate (mg/L)
7200, 6500, 5600, 5000, 4500, 3500,
compound
0.1
0.0
2.4
8.8
15.9 45.3
phenylacetylene
styrene
31
72
17
0
9
3
115 1287 671
184 1267 101
chlorophenylacetylene
chlorostyrene
naphthalene
0
0
18
72
8
0
30
0
40
0
0
0
0
57
3
0
0
11
50
0
0
605 161
87 116
0
0
0
266 3900 126
1
-chloronaphthalene
753 2030 2180 1600
33
7
acenaphthylene
chloroacenaphthylene
fluoranthene
pyrene
benz[a]anthracene
chrysene
benzo[b]fluoranthene
benzo[k]fluoranthene
benzo[a]pyrene
benzo[e]pyrene
perylene
61
22
38
20
68
33
FIGURE 3. Centerline profiles of mean mole fractions of minor
species, Re ) 4500.
43 425
788 554 1740
247 174 1210
0
103 120 100
135 144 369
0
128 132 286
20
6
33
20
18
55
33
27
31
14
55
14
45
81
14
53
27
7
34
29
1
. Consequently, a chlorine balance cannot be closed in this
system and the presence of HCl as a sink of much of the fuel
chlorine can only be surmised from the experience of other
investigators. For example, Senser et al. (9) determined HCl
0
0
concentrations from a Cl balance in a premixed CH
4
/ CH
2
-
0
0
20
8
0
5
Cl / air flame. After the fuel had been oxidized, the con-
2
centrations of HCl were far greater than any of the other
chlorinated species that they detected in flames with Cl/ H
ratios of 0.06 and 0.33, values that bracket the Cl/ H ratio in
the present experiment. Therefore, it is reasonable to
suppose that HCl is a major sink of Cl in the present
experiment. In fact, mass spectrometer scans of the flame
samples did not reveal any abundant chlorinated species in
addition to those that were analyzed quantitatively and are
reported herein.
TABLE 2. Material Collected from Gas Phase on Sorbent
Tubes in picograms per liter of Sampled Gas
Reynolds number
compound
7200 6500 5600 5000 4500 3500
phenylacetylene
styrene
61
15
0
0
34
0
188
0
1472 342
197 693 609
Measurements of methyl chloride provide an indication
of the destruction efficiency of this species under different
chlorophenylacetylene
chlorostyrene
0
0
0
31
0
0
0
0
0
0
0
naphthalene
97 330 1502 261 840 3650
484 489 559 434 405 1874
conditions of mixing. The mole fraction of CH Cl at X/ D )
3
-
5
1-chloronaphthalene
acenaphthylene
chloroacenaphthylene
fluoranthene
pyrene
benz[a]anthracene
chrysene
benzo[b]fluoranthene
benzo[k]fluoranthene
benzo[a]pyrene
benzo[e]pyrene
perylene
1
50 was approximately 2 × 10 . The visible flame tip was
26
0
9
0
19
0
13
0
18
0
8
39
95
89
13
60
11
11
12
11
10
at around 200 nozzle diameters. Thus, it appears that the
original methyl chloride was reduced by the necessary 99.99%
at a Reynolds number of 4500. However, Figure 3 indicates
221
94
0
62
37
23
25
0
0
0
0
0
96 100
65
46
13
16
20
10
12
0
87
20
70
0
81
19
88
53
0
-
5
that benzene (8 × 10 mole fraction) and chlorobenzene (6
-5
×
10 mole fraction), as well as some aliphatic hydrocarbons,
0
were present toward the end of the flame in concentrations
greater than the methyl chloride which decomposed quite
rapidly. A similar efficiency in the destruction of the methyl
chloride was apparent in the other flames that were studied.
The results that were obtained from collecting particulates
145
109
92
0
0
0
0
0
20
11
11
0
0
(
soot) and semivolatile organics above the flames are
presented in Tables 1 and 2 for the complete range of
Reynolds numbers. It should be noted that the partitioning
between particulate and gas phases was not absolutely certain
because the postflame gases passed through the filter on
which the soot particles were trapped. It was possible that
some of the semivolatile organics could have been absorbed
onto the trapped soot. The results are reported in terms of
picograms of compounds per liter of sampled gas. Unfor-
tunately, it was not possible to report the results in terms of
picograms per gram of fuel burned; to do so would have
required the postflame gases and soot to be uniformly
distributed across the wind tunnel. This was not the case
at 500 mm above the flame. Alternatively, some form of
sampling rake could have been employed to yield a repre-
sentative sample, but this was difficult to implement with
quasi-isokinetic sampling in a sooty gas stream for a
reasonable size vacuum pump. The results of postflame
sampling therefore represent typical mass concentrations
on the axis of the flow above the flame.
et al. (21) found that naphthalene was the most abundant
PAH that was formed in a jet-stirred reactor supplied with
CH Cl with pyrene and cyclopentapyrene present in small
3
amounts. In general, greater amounts of PAH were found
to be associated with the soot samples as the Reynolds
numbers were decreased. The results reflected, in large part,
an increase in the amount of soot that was emitted from the
flames as the rate of mixing was decreased (see the particulate
masses reported in Table 1). It should be noted that some
amount of soot was collected at all Reynolds numbers. The
reported mass of soot at Re )6500 does not imply the absence
of soot altogether. It reflects the inability to measure a very
small amount of soot that was below the measurable range
of mass. Some of the PAH, in particular the benzofluoran-
thenes and the benzopyrenes, were only found in significant
amounts on the soot samples; very little or nondetectable
amounts were extracted from the sorbent tubes. In view of
the reported mutagenicity of fluoranthene by Howard et al.
(2), the levels of this compound that were detected at low
Reynolds, particularly on the soot samples, suggested that
the mixing conditions may well have an impact on the toxicity
Naphthalene, styrene, fluoranthene, and its isomer,
pyrene, were the predominant PAHs found on the soot and
in the gas phase. Ciajolo et al. (20) found a similar distribution
of PAHs in a slightly sooting premixed methane flame. Marr
VOL. 32, NO. 9, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1 2 6 7

1
of the total flame emissions, i.e., low rates of mixing favor
Literature Cited
greater biological risk. Only two chlorinated PAHs were
detected in significant amounts in the soot extracts. They
were chlorostyrene and chloronaphthalene. At the lower
Reynolds numbers, 1-chloronaphthalene was the main
aromatic compound found in the soot extracts.
In the samples of the postflame gases, 1-chloronaphtha-
lene was the only chlorinated aromatic of significance. The
amount of this compound in the gas phase appeared to be
almost independent of Reynolds number until the flame
approached transitional behavior at a Reynolds number of
(1) Kozinski, J. Combust. Flame 1994, 96, 249.
(2) Howard, J. B.; Longwell, J. P.; Marr, J. A.; Pope, C. J.; Busby, W.
F.; Lafleur, A. L.; Taghizadeh, K. Combust. Flame 1995, 101, 262.
(
3) Lafleur, A. L.; Longwell, J. P.; Shirname-More, L.; Monchamp,
P. A.; Peres, W. A.; Plummer, E. F. Energy Fuels 1990, 4, 307.
4) Lafleur, A. L.; Longwell, J. P.; Marr, J. A.; Monchamp, P. A.;
Plummer, E. F.; Thilly, W. G.; Mulder, P. P. Y.; Boere, B. B.;
Cornelisse, J.; Lugtenburg, J. Environ. Health Persp. 1993, 101,
(
1
46.
(
(
(
5) Lam, F. W.; Longwell, J. P.; Howard, J. B. Twenty-Third
Symposium (International) on Combustion, 1477, The Combus-
tion Institute, Pittsburgh, PA, 1990.
6) Blankenship, A.; Chang, D. P. Y.; Jones, A. D.; Kelly, P. B.;
Kennedy, I. M.; Matsumura, F.; Pasek, R.; Yang, G. Chemosphere
1994, 28, 183.
3
500. Styrene and naphthalene were the other dominant
PAH species. Styrene appeared to decrease with Reynolds
number while naphthalene appeared to increase with
Reynolds number. Although there is considerable variability
in the samples, a general trend to a smaller amount of gas
phase byproducts is apparent with increasing Reynolds
numbers. However, the concentrations of some biologically
significant PAHs increased sharply at the highest Reynolds
number (7200), at which condition the flame was marginally
stable and barely attached to the nozzle. Fluoranthene and
the benzofluoranthenes and benzopyrenes exhibited sig-
nificantly elevated concentrations at the highest Reynolds
number. The results suggest that excessive rates of mixing
may lead to incomplete oxidation of some the intermediate
species that are formed on the fuel side of the flame. Local
extinction of the flame front due to high strain rates, and
concomitantly high scalar dissipation rates (13), may con-
tibute to this behavior.
7) Chang, W. D.; Senkan, S. M. Environ. Sci. Technol. 1989, 23,
4
42.
(
(
8) Senkan, S. M. Combust. Sci. Technol. 1984, 38, 197.
9) Senser, D. W.; Cundy, V. A.; Morse, J. S. Combust. Sci. Technol.
1
987, 51, 209.
(10) Kilgroe, J. D.; Nelson, L. P.; Schindler, P. J.; Lanier, W. S. Combust.
Sci. Technol. 1990, 74, 223.
(
(
(
(
11) Ibiricu, M. M.; Gaydon, A. G. Combust. Flame 1964, 8, 51.
12) Friedman, R.; Levy, J. B. Combust. Flame 1963, 7, 195.
13) Yang, G.; Kennedy, I. M. Combust. Flame 1993, 92, 187.
14) Yang, G.; Kennedy, I. M. Combust. Sci. Technol. 1994, 101, 347.
(15) Brouwer, J.; Sacchi, G.; Longwell, J. P.; Sarofim, A. F. Combust.
Flame 1994, 99, 231.
(
16) Senkan, S. M.; Robinson, J. M.; Gupta, A. K. Combust. Flame
983, 49, 305.
1
(
(
17) Morse, J. S.; Cundy, V. A. Combust. Sci. Technol. 1994, 95, 333.
18) Frenklach, M.; Hsu, J. P.; Miller, D. L.; Matula, R. A. Combust.
Flame 1986, 64, 141.
Acknowledgments
(
(
19) Brouwer, J.; Sacchi, G.; Longwell, J. P.; Sarofim, A. F. Combust.
Sci. Technol. 1994, 101, 361.
20) Ciajolo, A.; Danna, A.; Barbella, R. Combust. Sci. Technol. 1994,
This research has been supported by the NIEHS Superfund
Basic Research Program with Grant 5P42ES04699 from the
National Institute of Environmental Health Sciences, NIH
with funding provided by EPA. Its contents are solely the
responsibility of the authors and do not necessarily represent
the official views of the NIEHS, NIH, or EPA.
1
00, 271.
(21) Marr, J. A.; Allison, D. M.; Giovane, L. M.; Yerkey, L. A.;
Monchamp, P.; Longwell, J. P.; Howard, J. B. Combust. Sci.
Technol. 1992, 85, 65.
(1) Although the issue of toxicity has not been considered
directly in the present study by the use of bioassays, an earlier
study by this group has shown that soot formed from a methane/
trichloroethylene flame elicited a strong response in a bio-assay
sensitive to trichloro-dibenzo-dioxin (TCDD) (6).
Received for review March 31, 1997. Revised manuscript
received January 22, 1998. Accepted January 29, 1998.
ES9702935
1
2 6 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 9, 1998