United States
Environmental Protection
Agency
Research and Development
Air and Energy Engineering
Research Laboratory
Research Triangle Park, NC 27711
EPA/600/SR-94/070 July 1994
EPA Project Summary
Pilot-Scale Evaluation of the
Potential for Emissions of
Hazardous Air Pollutants from
Combustion of Tire-Derived Fuel
Paul M. Lemieux
Experiments were conducted in a 73-
kW (250,000-Btu/hr) rotary kiln incin-
erator simulator to examine and char-
acterize emissions from incineration of
scrap tire material. The purposes of
the project are to (1) generate a profile
of target analytes for full-scale stack
sampling, not statistically defensible
emissions factors for the controlled
combustion of scrap tire material and
(2) where possible, give insight into
the technical issues and fundamental
phenomena related to controlled com-
bustion of scrap tires. Wire-free crumb
rubber, sized to < 0.64 cm (< 1/4 in.),
was combusted at two feed rates, two
temperatures, and three kiln oxygen
concentrations. Along with continuous
emissions monitoring for oxygen (O,),
carbon dioxide, carbon monoxide (CO),
nitric oxide, sulfur dioxide, and total
hydrocarbons (THCs), samples were
taken to examine volatile and semi-vola-
tile organics, polychlorinated p-
dibenzodioxins and dibenzofurans
(PCDD/PCDF), and metal aerosols. In
addition, a continuous polycyclic aro-
matic hydrocarbon (PAH) analyzer was
used in all the tests. Samples were
analyzed with emphasis on the 189 haz-
ardous air pollutants listed in the 1990
Clean Air Act Amendments, but other
compounds were also identified where
possible.
Results indicate that, if burned in a
steady-state mode, tire-derived fuel
(TDF) combustion will result in very
low emissions of CO, THCs, volatile
and semi-volatile organics, and PCDD/
PCDF. Metal emissions were also very
low, except for arsenic (As), lead (Pb),
and zinc (Zn). Uncontrolled stack con-
centrations of As and Pb were 37.16
and 65.96 jjg/Nm3, respectively. Uncon-
trolled Zn emissions were considerably
higher, 35,465 ug/Nm3. Results also in-
dicate that organic emissions can in-
crease significantly when TDF is fired
in a non-steady mode. The continuous
PAH analyzer appeared to track tran-
sient operation well and gave concen-
tration results in the same range as
those using EPA standard semi-vola-
tile organic sampling methodologies.
This Project Summary was developed
by EPA's Air and Energy Engineering
Research Laboratory, Research Tri-
angle Park, NC, to announce key find-
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
Approximately 240 million vehicle tires
are discarded annually in the U.S. Viable
methods for reclamation exist. One attrac-
tive option for use of scrap tires is burn-
ing, either alone or with another fuel (such /
as coal) in a variety of energy-intensive
processes, such as cement kilns and util-
ity boilers. Another potentially attractive
option is the use of ground tire material as
a supplement to asphalt paving materials.
Congress has passed a law, the Intermodal
Surface Transportation Efficiency Act of
1991, which mandates that up to 20% of
all federally funded roads in the U.S. in-
clude as much as 9 kg (20 Ib) of rubber
derived from scrap tires per 907 kg (1 ton)
of asphalt by 1997. In spite of these ef-
forts, less than 25% of the total amount of
Printed on Recycled Paper
-------
discarded tires are reused or reprocessed,
and the remaining 175 million scrap tires
are discarded in landfills, above-ground
stockpiles, or illegal dumps. Disposal in
above-ground dumps increases the po-
tential for uncontrolled tire fires, which pro-
duce large quantities of air pollutants, many
of which are known or suspected carcino-
gens. In addition, these reclamation ef-
forts do little to affect the estimated 2
billion tires already stockpiled.
Tires can be burned whole or can be
shredded or chipped before burning. Tires
that have been processed into smaller
pieces are called tire-derived fuel (TDF).
Three main industries use either whole
tires or TDF either as a sole fuel or a fuel
supplement. These industries are
• Electric utilities that use TDF and
whole tires as supplemental fuel in
power generation. One company is
using whole tires as its sole source of
fuel in power generation.
• Cement manufacturing companies
using tires and TDF to supplement
their primary fuel (usually coal) for
firing cement kilns. Some of the com-
panies are using tires or TDF directly
in the kiln, and some are using tires
or TDF in the precalciner (prior to the
kiln).
• Pulp and paper companies using tires
or TDF as supplemental fuel in their
waste-wood products boilers.
TDF can be processed to remove steel
belts and the metal bead that surrounds
the wheel rim. TDF with the metals re-
moved is termed wire-free, and TDF with
the wire remaining is termed wire-in. TDF
can be purchased in a variety of sizes,
some as small as < 0.7 cm (< 0.25 in.).
TDF that is very small is termed crumb
rubber. TDF has a higher heating value
than coal and contains about as much
sulfur as medium sulfur coal.
The main environmental concern of us-
ing whole tires or TDF as supplementary
fuel is the potential for increased air emis-
sions. Pollutants of concern include crite-
ria pollutants—carbon monoxide (CO),
sulfur dioxide (SO2), nitrogen oxides (NOX),
and particulates—metals, and unburned
organics. Title III of the 1990 Clean Air
Act Amendments (CAAA) lists 189 haz-
ardous air pollutants (HAPs) of concern.
These include volatile organic species such
as benzene, polycyclic aromatic hydrocar-
bons (PAHs) such as benzo(a)pyrene,
metal species such as lead, and several
individual compounds such as polychlori-
nated p-dibenzodioxins and polychlorinated
dibenzofurans (PCDD/PCDF).
Past field data have shown that, for the
most part, emissions of most criteria pol-
lutants from coal-fired boilers and cement
kilns are reduced:when a fraction of the
fuel input is replaced with tires or TDF.
This includes SOxi (which drops if the pri-
mary fuel is a high sulfur eastern coal),
and NO (since tires have very little fuel
nitrogen). Particulates have generally in-
creased slightly. Emissions data for other
pollutants, however, are either very lim-
ited or nonexistent
A significant data gap exists in the da-
tabase of HAPs that can be formed from
combustion of tires or TDF. This makes it
difficult for state and local air pollution
agencies to granf air quality permits al-
lowing a facility to Supplement,its fuel with
tires or TDF, since stack sampling is quite
expensive, especially when a list of target
analytes does not exist. It was for this
reason that EPA's Control Technology
Center (CTC) funded this project. The pur-
poses of this project are to (1) generate a
profile of target analytes for full-scale stack
sampling, not statistically defensible emis-
sions factors for the controlled combus-
tion of scrap tire;material and (2) where
possible, give insight into the technical
issues and fundamental phenomena re-
lated to controlled combustion of scrap
tires.
Several issues: are of concern with the
use of TDF in combustion devices:
• The effect ori products of incomplete
combustion (PICs) of the mode of tire
feeding (e.g.:, whole tires vs. shred-
ded tires), i
• The potential for the formation of
classes of air toxics not normally found
in the stacks of combustion devices
burning conventional fuels.
• The impact of TDF-generated particu-
late on operation of existing particu-
late control devices.
• Potential operational problems due to
differences in feed characteristics.
• Potential operational problems due to
differences in the residues that are
generated.
This project addresses the first two is-
sues listed abovjB. As much as possible,
the last three issues are eliminated from
the scope of the project by 1) utilizing a
very uniform feed so as to enable as close
to steady-state operation as possible and
2) using a grade of TDF that will not
generate significant residue.
Experimental Equipment
A single laboratory-scale combustor was
used to perform all the tests, and the tests
were performed, in as wide a range of
operating conditions as possible, to simu-
late the process'conditions in a variety of
combustion units. In addition, it was de-
cided that the scrap tire material be co-
fired with natural gas (rather than coal or
wood waste) as the primary fuel. By using
natural gas as the primary fuel, it was
hoped that the effect of the TDF could be
isolated, rather than adding the additional
experimental complications inherent with
burning an additional heterogeneous fuel
like coal or wood.
The tests were performed in EPA's ro-
tary kiln incinerator simulator (RKIS) at
Research Triangle Park, NC. Figure 1
shows the simulator schematically. A small
afterburner (43.8 kW, 150,000 Btu/hr)
mounted at the base of the secondary
combustion chamber established near-iso-
thermal operating conditions throughout
the unit. Sample ports are at various loca-
tions.
Measurements made on the RKIS are
not intended to be extrapolated directly to
full-scale units. It is, for example, very
difficult to scale up some of the important
gas-phase mixing phenomena from the
simulator, where, for instance, stratifica-
tion is not significant, to a full-scale unit,
where stratification is significant. In addi-
tion, there are significant differences be-
tween kilns and other combustion devices,
and this study does not address those
issues, although some of the information
from this study can be applied to other
types of combustors. The purpose of the
simulator is to individually examine the
fundamental phenomena that occur in full-
scale units and to gain an understanding
of the qualitative trends that would be
found in a full-scale rotary kiln. In no way
should it be inferred that the concentra-
tions of pollutants from this apparatus
would be the same as those from full-
scale units.
TDF, consisting of wire-free crumb rub-
ber, sized < 0.64 cm (< 0.25 in.), was
introduced into the kiln via a vibrating
feeder. The feeder enabled TDF to be fed
into the RKIS continuously at rates from 0
to 2 kg/hr.
Continuous emission monitors (CEMs)
measured oxygen (O), carbon dioxide
(CO2), CO, nitric oxide (NO), and total
hydrocarbons (THC) both before and after
the secondary combustion chamber (SCC),
as well as SO2 at the SCC exit. In addi-
tion, a continuous photoelectric PAH ana-
lyzer sampled the gases at the stack exit.
VOCs were collected by a Volatile Or-
ganic Sampling Train (VOST) system.
Semi-volatile organics and bulk particu-
late were collected by isokinetic sampling
protocols with a Modified Method 5 (MM5)
train. Metal aerosols were collected by a
Multiple Metals Train (MMT), except that
the potassium permanganate (KMnO4) so-
-------
To Flue Gas Cleaning System
Secondary Combustion Chamber
Kiln Section Transition Section
Figure 1. Rotary kiln incinerator simulator.
lution was omitted. The TDF samples were
analyzed for metals before the experi-
ments, and the absence of mercury (Hg)
in the analysis made it unnecessary to
configure the sampling train for Hg sam-
pling. PCDD and PCDF were sampled
using the MM5 train with the protocols
laid out in EPA Method 23 and analyzed
by high resolution gas chromatography/
low resolution mass spectrometry (HRGC/
LRMS), using a Hewlett-Packard 5890/
5970 Gas Chromatography/Mass Selec-
tive Detector (GC/MSD) system and meth-
ods adapted from EPA Method 23 and
RCRA Method 8280.
Results
Continuous Emission Monitor
Samples
All CEM data were averaged over the
course of the run to yield a single number.
Tables 1 and 2 list average values from
the CEMs: Table 1 taken at the kiln exit
sample port, and Table 2 taken at the exit
of the secondary combustion chamber.
Note that the SO2 analyzers exhibited un-
reliable and erratic operation. The SO2
data did not meet quality control specifi-
cations, although they are reported in
Table 2, except where data were not avail-
able.
Volatile Organic Samples
In general, the volatile organic com-
pounds (VOCs) detected were fairly close
to practical quantitation levels. A number
of compounds identified in combustion
samples were not present in the field (com-
bustion) blanks. However, several of the
compounds found in combustion samples
.were also present in the field blanks at
similar levels; primarily chloromethane,
acetone, methylene chloride, and benzene.
Benzene is a common PIC from combus-
tion in general and could be attributed to
PICs from the natural gas flame found in
the field blanks. However, benzene is also
a breakdown product of Tenax, and ac-
etone and methylene chloride are ubiqui-
tous in laboratory environments. In addi-
tion, some samples contained .trichloro-
fluoromethane, which is a chlorofluorocar-
bon commonly used in air conditioners.
Table 3 summarizes the results from the
volatile organic samples, by averaging the
emission values from all compounds that
were present in concentrations greater
than the quantitation level on at least one
VOST tube. On compounds where one or
more VOST tubes had concentrations be-
low the quantitation level, the quantitation
level was used for averaging. The results
from the trichlorofluoromethane, acetone,
and methylene chloride are considered
suspect and are not reported in Table 3.
Semi-Volatile Organic Samples
The results from the semi-volatile or-
ganic compound (SVOC) analyses do not
seem to indicate the presence of SVOCs
in detectable concentrations. The PAH
analyzer gave readings on all steady-state
tests ranging up to 3289 ng/Nm3, which
converts into 3.3 ng/Nm3. This concentra-
tion is below the method detection level
for individual PAH compounds in the
SVOC analysis. Considering that other
-------
past experiences with the PAH analyzer
gave good agreement with conventional
SVOC analyses, it can be surmised that
the results reported from the PAH ana-
lyzer compare favorably with the PAH con-
centrations that were actually present in
the stack. Overall, it appears that, when
combusted in an efficiently operated facil-
ity, emissions of SVOCs from TDF com-
bustion are not significantly different than
from natural gas.
The results of the PCDD/PCDF analy-
ses indicate that insignificant amounts of
PCDDs and PCDFs were emitted during
these tests. The results from the TB6 -
17% TDF test reveal that hexachloro-
dibenzofuran was present at a concentra-
tion essentially equal to the method de-
tection limit. Similarly, the results from the
combustion background test CTB7 - 0 %
TDF) revealed that tetrachlorodiben-
zodioxin was present at a concentration
also essentially equal to the method de-
tection limit. The method blank did not
detect either of these target analytes.
Table 2. OEM Data Taken at SCC Exit
Table 1. CBM Data Taken at Kiln Exit
Run
No. (TB)
r
2'
3'
4*
5'
6
7
8*
9>
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
(<&)
8.28
7.17
7.35
8.51
9.23
7.64
9.82
6.45
8.38
7.66
3.68
5.71
7.62
7.85
3.10
5.07
7.53
5.40
3.55
8.32
4.24
6.06
5.33
4.91
7.59
3.70
3.18
6.39
8.23
5.49
CO.
(%}
7.33
7.45
7.70
7.14
6.31
7.77
5.85
8.38
7.32
7.09
9.34
8.22
7.15
7.25
9.97
8.86
7.43
8.92
10.00
7.35
9.59
8.71
8.08
9.03
7.55
9.71
10.24
8.54
7.39
8.94
CO
(ppm)
33
36
16
31
44
20
17
30
700
30
38
36
35
35
42
40
35
44
48
41
47
43
28
50
45
51
59
54
54
53
NO
(ppm)
52
73
58
54
39
55
32
60
53
39
53
47
40
46
62
57
50
62
64
54
64
62
56
66
64
65
68
66
58
68
THC
(ppm)
3
3
0
1
0
1
2
1
43
-1
0
1
0
1
2
1
1
1
1
1
1
1
0
1
1
2
1
2
1
1
Run
No. (TB)
1*
2*
3*
4"
5"
6
7
&
9*
10
11
12
13
14
15
16
17
18
19
20
21
'22
23
24
25
26
27
28
29
30
(°fa
7.59
5.43
5.62
6.78
7.50
6.06
8.10
5.20
8.85
6.23
2.82
4.49
6.17
6.60
2.48
' 4.07
6.28
4.30
2.61
7.12
3.20
5.01
4.11
4.40
6.85
3.68
3.47
5.97
7.36
5.39
,C02
(%)
8.62
9.44
9.62
9.07
8.44
9.45
7.80
10.07
8.11
8.86
10.77
9.94
9.01
8.99
11.18
10.35
9.15
10.46
11.33
9.00
11.01
10,18
9.94
10.30
9.02
10.60
10.83
9.68
8.88
9.93
CO
(ppm)
16
14
12
20
16
15
9
18
70
13
17
16
16
16
20
19
17
19
20
18
20
20
13
19
18
20
20
20
20
19
NO
(ppm)
49
64
48
62
52
60
42
58
46
" 28
36
35
30
35
46
44
41
50
50
45
50
50
32
36
37
36
37
39
36
37
THC
(ppm)
-2
-1
-2
2
1
1
2
1
6
0
0
-1
0
0
0
0
0
0
0
0 .
0
0
-1
-1
-1
0
0
0
0
0
so,
(ppm)
12
81
59
67
14
51
35
76
26
t
t
t
t
42
•t
103
106
t
96
82
74
78
7
26
22
42
108
106
100
84
PAH
(ng/nf)
1437
2278
2284
3289
1941
982
429
1957
214000
410
630
404
426
767
1553
1313
702
1939
2364
1474
1937
1734
925
1581
618
1550
1787
1632
1130
1551
Electrical noteo problems on the OEMs.
Non-continuous food tests.
* Electrical noise problems on the CEMs.
+ Non-continuous feedtests.
'Analyzer not working.;
Metals Samples
Table 4 lists the concentrations of met-
als for the two tests where sampling oc-
curred. Conversion of metal species con-
centrations to emission factor units and
comparing the resulting estimated emis-
sions to emission factor data from com-
bustion of coal and oil in utility boilers
suggest that, except for zinc, uncontrolled
metal emissions from TDF combustion are
similar in magnitude to those for coal and
oil. j.
Conclusions
Experiments ;were performed on a
bench-scale RKIS to examine HAPs from
combustion of TDF. Both steady-state and
transient testing!were performed so that
continuous vs. incremental TDF feeding
could be evaluated. Samples were ana-
lyzed continuously by GEM for O, CO,
CO2, NO, THCs, SO , and PAHs. VOST,
MM5, Method 23, and MMT samples were
collected to analyze for VOCs, SVOCs,
PCDD/PCDF, and metal aerosols, respec-
tively. X-ray diffraction and X-ray fluores-
cence techniques were used to identify
species in the fly ash. A regression analy-
sis of the GEM data helped determine
pollutant emission trends.
Several VOCs were identified, particu-
larly chloromethane, benzene, and sty-
rene. The concentrations of those VOCs
were affected by the amount and mode of
TDF feeding. Emissions of benzene, in
particular, are particularly sensitive to tran-
sient upsets of the combustion process.
Comparison of calculated emission fac-
tors to those found in the literature for
conventional fossil fuel combustion indi-
cates that VOC emissions from TDF com-
bustion are comparable to those from coal
and oil combustion.
No significant amounts of SVOCs were
identified. The PAH analyzer indicated
PAH concentrations on the same order as
the detection level of the SVOC analytical
methods, except for the test where TDF
was batch fed to the RKIS. The PAH
analyzer indicated considerably higher con-
centrations of PAHs during batch feeding;
however, these elevated PAH levels were
not detected with the MM5 samples. It is
possible that the short duration of the tran-
sients, coupled with the mandatory
isokinetic sampling protocols, prevented
-------
Table 3. Summary of VOC Concentrations (ng/L)
Compound
1, 1, 1 .trichloroethane
2-methyl propane
2-methyl-2-propanol
Benzene
Bromomethane
Carbon disulfide
Chlorobenzene
Chloromethane
Ethyl benzene
Heptane
Hexane
lodomethane
m,p-xylene
Nonane
o-xylene
Styrene
Toluene
TBS
0%TDF
(blank)
0.55
2.36
0.52
1.65
0.49
0.52
0.52
0.59
0.52
0.52
0.49
0.52
1.52
0.68
0.45
0.65
0.97
TB1
7% TDF
(steady-state)
0.88
5.40
0.51
2.93
0.51
0.81
0.51
1.68
0.51
0.67
0.58
0.51
0.98
1.72
0.51
1.85
1.18
TBS
17% TDF
(steady-state)
1.00
4.38
4.10
2.83
0.58
0.52
0.52
8.81
0.61
0.56
0.55
0.52
2.40
0.96
0.72
1.62
1.05
TB8
19% TDF
(ramp)
0.52
1.70
0.52
17.00
2.82
0.52
0.52
55.03
0.52
0.52
0.52
0.54
0.61
0.52
0.52
1.62
0.80
TB9
15% TDF
(batch)
0.47
0.50
0.50
47.31
0.83
2.04
0.48
11.17
1.07
0.50
0.51
0.50
3.85
0.59
1.13
1.69
2.78
sufficient amounts of pollutants from be-
ing sampled.
Emission levels of PCDD and PCDF
were found to be similar in magnitude to
the combustion blank that consisted of a
natural gas flame. Those congeners of
PCDD and PCDF identified in all samples
were on the same order of magnitude as
the method detection level.
Elevated levels of arsenic, lead, and
zinc were found in the stack gas. Zinc
was present in significant concentrations.
Analysis of the fly ash residue indicates
that most of the particulate matter was
SiO2, Al Si2O13, and Zn2SiO4. Comparison
of calculated emission factors from TDF
combustion to those found in the literature
for conventional fossil fuel combustion sug-
gests that, except for zinc, the magni-
tudes of metal emissions are similar to
those from coal and oil, although the dis-
tributions are significantly different, espe-
cially with respect to emissions of mer-
cury and selenium, which are significantly
higher from coal combustion.
The PAH analyzer tracked transient kiln
operation during periods of good combus-
tion more effectively than the CO ana-
lyzer, and with faster response. It may be
significantly more effective for process con-
trol applications due to its sensitivity. Re-
gression analysis of PAH analyzer mea-
surements indicated that an approximately
fivefold increase (over natural gas emis-
sions) in PAH emissions occurs while in-
creasing the TDF fuel input fraction from
0 to 20%.
Regression analysis of CO emissions
from the steady-state tests did not signifi-
cantly correlate with kiln operating condi-
tions. A slight increase in CO emissions
with increasing TDF feed rate was found.
The results suggest that burning TDF in
batches, such as during the feeding of
whole tires, has the potential to form sig-
nificant transient emissions. This phenom-
enon could be exacerbated in a system
that exhibits significant vertical gas-phase
stratification or operates at low excess air
levels, such as cement kilns. The size of
the facility, however, will certainly impact
the intensity of transient emissions result-
ing from batch charging of tires or TDF
since, for an extremely large facility, a
steady stream of whole tires may roughly
approximate steady-state operation. Even
so, the potential for generation of large
transients should hot be ignored, espe-
cially in smaller facilities.
Data gaps still exist, since this limited
study was performed on a small combus-
Table 4. Stack Concentration (fig/m3) of Metals
from TDF Combustion
Metal
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Nickel
Selenium
Zinc
TB70% TDF TB6 1 7% TDF
(blank) (steady-state)
0.18
1.12
nd*
0.41
0.65
8.05
2.82
0.71
0.83
286.94
2.11
37.16
0.05
1.06
3.88
65.96
5.79
3.51
4.50
35465
tor, under controlled conditions. The fol-
lowing issues might be addressed in fu-
ture research:
• The effect of TDF particle size and
feeding mode on HAP emissions. This
Study was done using a single TDF
particle size and included only limited
testing on different feeding modes.
• Emissions of HAPs from combustion
of wire-in TDF. It would be logical to,
assume that emissions of metals from
combustion of wire-in TDF may be
significantly different than from TDF
.that has had the wire removed. Com-
bustion temperature would likely af-
fect metals emissions significantly,
since the partitioning of metals be-
tween the bottom ash residue and
the fly ash would change.
• Characteristics of other TDF-gener-
ated residues, such as bottom ash,
especially in regard to teachability of
metals, and slag composition and
quality.
• Emissions of HAPs from co-firing of
TDF with other solid fuels, such as
coal, biomass-derived fuels, munici-
pal solid waste, or refuse-derived fuel.
• The characteristics of TDF-generated
fly ash, including the particle size dis-
tributions and speciation of the met-
als, especially as a function of halo-
gens or sulfur that might be present
due to co-firing of other fuels.
• Some basic research, on a very small
scale, to examine the chemistry of
TDF pyrolysis and combustion.
• Tests on other types of facilities (such
as a vertically fired unit), including
studies examining TDF combustion
in suspension vs. bed-burning phases.
Overall, it appears that, except for zinc,
potential emissions from TDF combustion
are not significantly different from emis-
sions from combustion of conventional fos-
sil fuels, when burned in a well-designed
and well-operated combustion device. If
particulate loading is unacceptable be-
cause of zinc emissions, the emissions
would have to be controlled by an appro-
priate particulate control device.
"None detected.
tVu.S. GOVERNMENT PRINTING OFFICE: 1994 - 550-067/80278
-------
------- |