United States
Environmental Protection
Agency
Industrial Environmental
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-84-046 May 1984
&ER& Project Summary
Study of the Effectiveness of a
Catalytic Combustion Device on a
Wood Burning Appliance
John M. Allen, William H. Piispanen, and Marcus Cooke
A wood stove, incorporating a
catalytic combustor, was operated
while burning air-dried oak at low bur-
ning rates. Gas composition was mea-
sured continuously both at the entrance
and the catalyst (after the gases had left
the burning wood) and at the exit of the
stove (after the gases had traversed the
catalyst). These continuous monitors
showed a reduction of about 3:1 in
volatile hydrocarbons measured by a
flame ionizatlon detector, and a com-
parable reduction in carbon monoxide.
These monitors showed no significant
corresponding reduction when the ac-
tive catalyst was replaced with an iden-
tical ceramic element, but without the
noble metal catalytic coating.
Grab samples collected from the flue
gas earn/ during the wood burning cycle,
analyzed by MS and GC, confirmed the
lower concentrations of volatile com-
bustible species when the active
catalyst was used.
This Project Summary was developed
by EPA's Industrial Environmental Re-
search Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
The increased use of catalytic combustion
devices on wood-burning appliances has
raised a number of concerns regarding emis-
sions from these sources into residential
areas. The purpose of this study was to
determine the effectiveness of catalytic com-
bustors in reducing emissions of hydrocar-
bons and CO, and also to determine the
species of compounds which are emitted
during the operation of a typical catalyst-
equipped appliance. The stove used in this
program had a honeycomb, ceramic catalytic
device, referred to as a catalyst, internally
within the stove.
While the catalyst is believed to be effec-
tive in promoting combustion, actual opera-
tion had not been thoroughly characterized
with respect to stove operating variables. Of
particular concern to this study were:
• Operating conditions where the catalyst
is most effective in reducing emissions.
• Effectiveness of the catalyst in low burn-
rate operation.
• Combustion products formed in the
catalytic combustion process.
For this program a catalyst-equipped
wood-burning appliance was tested in Bat-
telle's combustion research laboratory. Con-
tinuous samples of the combustion gases,
before and after the catalyst, were analyzed
by on-line gas analyzers for 02, CO2, CO,
THC, NO, and NO,. Temperatures at various
points in the gas stream and also within the
catalyst bed were monitored during the test.
A sampling train (EPA Method 5) with a
packed column of XAD-2 resin was used to
collect paniculate and vapor-phase organics
from the gas stream. A grab sample of com-
bustion gas leaving the stove was taken for
analysis of low molecular-weight species (>
100 amu) by direct mass spectrometry and
selective gas chromatographic analysis.
On-line gas analyses indicate that the
catalytic device is effective in reducing both
total hydrocarbons (THC) and carbon mon-
oxide (CO) at temperatures where the
catalyst is active (above about 425°C
(800°F)). At lower temperatures there is less
effect on THC emissions. When a blank
catalyst (i.e., ceramic substrate without the
-------
noble metal coating) was used in place of the
active catalyst, there was a noticeable
decrease in the catalyst temperature, and the
emission levels increased.
Test Methods
A test program was conducted at Bat-
telle's Columbus combustion test facility to
evaluate a catalytic-combustor-equipped,
wood burning appliance. A Webster Oak
stove was installed in the facility and, with
the catalyst operating, the stove was condi-
tioned for 100 hours by burning split white
oak and pine. During these burns, the con-
trols were adjusted to obtain effective burn-
ing at low wood consumption rates. Figure
1 is a schematic diagram of the stove.
Sampling Methods and
Equipment
Aspirated thermocouples and rake-type
sample probes in the stove sampled combus-
tion gas both entering the catalyst (primary)
and in the flue (secondary). Thermocouples
monitored: (1) secondary air temperature in-
side the air duct within the combustion
chamber, (2) several stove skin tempera-
tures, (3) stack wall temperature, (4) sam-
ple train temperatures, and (5) ambient air
temperature. Also, the viewing window over
the catalyst was replaced with a fitted insert
that permitted vertical traversing of the
catalyst to measure temperatures within the
catalyst bed. During the preliminary burns,
the point of maximum temperature was
determined in the catalyst and the bare ther-
mocouple was locked at that point.
The test facility provided continuous
sampling and monitoring of the stove
parameters shown in Figure 2. The data col-
lection and processing system is limited to
collecting single point data. Due to limita-
tions on data storage, most tests were
designed for 110 scans at 2-minute intervals
(total time, 220 minutes).
This program included six test burns. Dur-
ing five of the six tests, a sample for volatile
gas analysis was taken at a point near the
gas exit to the chimney by drawing a sam-
ple into a preevacuated 3-liter glass bulb.
Also during monitoring, flue gas was sam-
pled continuously from the stack at a point
75 cm above the flue's entrance into the in-
sulated stack. For condensible organics, an
extractive sampler (similar to an EPA Method
5 train) was used with a packed column of
XAD-2 resin following the heated filter. This
train includes a glass-lined sampling probe
(sealed into the flue), a heated filter holder
- 200°C (390°F) - with a fiberglass filter,
an XAD-2 resin adsorption chamber at 20 °C
(68°F), a series of ice-cooled scrubbers, a
drying cartridge filled with silica gel, and gas
flow monitoring and gas moving devices.
Test Procedures
Tests were conducted by starting a fire in
the appliance, using red oak splits. During
this time, the sampling equipment was equil-
ibrated and leak-checked, and monitoring in-
struments were zeroed and spanned. The
data acquisition program was initiated by
recording: instrument and sample codes in
the data system, wood type and moisture,
barometric pressure, scan interval, test dura-
tion, start time, weight and type of wood
charge, and test title. Also, the sample pump
and line heaters were turned on to condition
the system.
After the initial wood charge had burned
to glowing coals, the bed was raked flat and
the electronic scale zeroed. A preweighed
charge of wood was placed in the combus-
tion chamber at exactly 1 minute prior to
data acquisition with the primary setting at
"air full open." Analyzers were then checked
to verify that flow rates and line pressures
were within specified limits.
Upon initiation of data acquisition, the
continuous sampler was started. The stove
operator monitored catalyst temperature
using a digital readout. When the catalyst
temperature reached 425°C (800°F), the
primary air supply was adjusted to a pre-
determined set point. This time delay was
normally less than 5 minutes. At no other
time during the burn was the air readjusted
or fuel added to the stove. Data acquisition
and extractive sampling continued until the
predetermined termination time was
reached.
Post-Test Procedures
After sampling was complete, the extrac-
tive sampling system was removed from the
H
LS
H
A Primary air supply, with controls in door
B Secondary air supply, with no controls
C Secondary air distributor
D Ceramic catalytic element
E Baffle
F Baffle bypass, opens when loading door is open
G Exhaust to chimney
H Ceramic hearth
J Gas sample probe for gas leaving stove
K Gas sample probe for primary combustion products entering catalyst
L Distribution pipe, heating secondary air supply
Figure 1. Schematic diagram of the catalytic stove.
-------
stack and all openings capped with clean
aluminum foil. The system was then trans-
ported to a chemistry laboratory for recovery
according to the wood stove sampling pro-
tocol. Analyzers were all purged with N2 gas
and then post-test zero and span checks
were made to determine instrument drift or
malfunction during testing.
Test Results
The test program included six stove tests.
Of these, five were run with the normal
catalyst in place, and one (Test 6) was run
using a ceramic substrate without the cat-
alytic coating. Tests were designed for
medium to low burn rates and run as close
to reproducible conditions as possible, ex-
cept Test 3, which was conducted with a
lower rate of primary air. All tests used four
pieces of wood, together weighing about 10
Ib (4.54 kg).
Continuous Sampling
Stove operating data for Tests 2 through
6 are included in Table 1. The operational
data for the Modified Method 5 tests are in-
cluded in Table 2. Note that Modified
Method 5 tests and grab samples were taken
Duplicate system for flue gas
Legend: Data Supply to Automatic Logger
2 Weight of stove
3 Primary Oz, %
5 Primary COa %
7 Primary CO. %
13 Primary THC, ppm
24 Air supply, room temp.
25 Secondary air temp.
26 Stove top metal temp.
27 Primary flue gas (entrance to catalyst)
29 Base of chimney, temp.
30 Top of chimney, temp.
32 Exhaust hood, temp.
A Primary gas sample time. temp.
B Primary gas pump. temp.
Figure 2. Schematic diagram of sampling system.
only on Tests 2, 3, 4, and 6. Test 5 was not
sampled for organics. Continuous gas analy-
ses were conducted at locations before and
after the catalyst. The data were recorded
every 2 minutes and then averaged over the
entire test. The results are presented in Table
3. The first test was considered a shakedown
run: reliable data were not obtained.
Emission Factors
The test results were compiled using Bat-
telle's wood stove data reduction program.
These results were then presented graphi-
cally using a Techtronix Plot 50 program.
The emission factors for CO and THC as
grams per kilogram of wood burned are
shown in Figures 3,4, and 5. These presen-
tations show the effect of secondary com-
bustion in the catalyst on CO and THC
emissions.
Grab Sample Results
The composition of volatile gases is sum-
marized in Table 4. Test 6, made with a blank
catalyst, clearly shows the effect of the
catalyst during Tests 1 through 4. Without
the catalyst, higher concentrations of CO,
THC, and H2 were found. Also elevated
benzene levels were found in the blank run.
Benzene and ethanol are not gases at room
temperature and atmospheric pressure;
therefore, these data are approximates —
true concentrations may be much higher. All
results are reported on a dry basis.
Discussion of Test Results
Wood combustion is generally character-
ized by initially fast surface combustion in
which moisture and organic pyrolysis pro-
ducts are evolved, during which an outer
char layer is formed. During the middle por-
tion of typical wood burning, pyrolysis pro-
ducts are evolved. Slow burning takes place
in this portion of the bum, with pyrolysis pro-
ducts and interior moisture passing through
the char layer simultaneously. Late in a
typical burn, residual fuel is essentially all
char; therefore, hydrocarbon emissions are
very low and surface burning takes place.
The organic emissions, which are products
of incomplete combustion, generally are
more concentrated during the early portion
of a test burn.
These effects are observed by examining
the emission factors shown in Figures 3, 4,
and 5. Pyrolysis products, shown by THC
emissions, usually have initially high values
early in the test and lower values as the wood
becomes charred. Irregularities are presum-
ably associated with shifting wood and in-
herent irregularities within the pieces of
wood.
-------
Table 1. Operating Conditions for Stove Tests
Weight of Average (actual) Average Cata- Average
Test Wood Burn Plate lyst Temp. Stack Temp.
No. Ib kg Ib/hr kg/hr °F °C °F °C Condition
2 10.0 4.5 3.5 1.6 941 505 364 184 High air
3 9.4 4.3 2.4 1.1 775 413 343 173 Low air
4 10.4 4.7 2.6 1.2 919 493 376 191 Med. air
5 10.4 4.7 2.3 1.0 898 418 349 176 Med. air
6 10.2 4.6 2.1 1.0 676 358 318 154 Uncoated
catalyst
Table 2. Results of Modified Method 5 Samples
Flue Gas Average Paniculate
7p?f Volume Sampled Sample Time Moisture Filter Temp. Emissions3
No. dscf dscm min. percent H,0 °F °C gr/dscf g/dscm
2 99.85 2.83 135 6.7 407 208 0.0130 0.0298
3 109.17 3.08 191 6.0 405 207 0.0256 0.0586
4 131.15 3.71 310 4.9 410 210 0.0268 0.0613
6 120.53 341 219 3.3 406 208 0.0186 0.0426
aCorrected to 12 percent C02.
Table 3. Average Flue Gas Analysis
j. t Oi, percent CO2r percent CO, percent THC, ppm PAH, mg/m3
No Primary Secondary Primary Secondary Primary Secondary Primary Secondary Secondary
2 11.2 17.4s 9.0 3.0s 0.9 0.1a 4320 1135s 3.4s
3 13.4 - 6.7 - - - 2922 1001 3.4
4 12.7 173 7.6 3.4 0.7 0.1 5089 2184 2.0
5 13.8 15.8 7.2 5.0 0.7 0.2 4029 1279
6 12.1 16,3 8.1 4.2 0.8 0,5 3550 3186 3.2
a Averaged only over last 105 minutes of test run. Early secondary readings not valid during Test 2
Table 4. Volatile Flue Gas Composition
Test No. 12346 Instrumental
Volume Percent (Dry Basis) Reproducibility
Nitrogen 77.3 77.2 77.4 77.5 76.5 ± 0.2
Oxygen 14.1 8.85 13.8 12.0 10.1 ± 0.1
Carbon dioxide 7.11 12.5 7.46 9.17 10.1 ±0.05
Carbon monoxide 0.35 0.38 0.20 0.25 1.30 ±0.03
Argon 0.94 0.94 0.94 0.95 0.94 ± 0.02
Methane 0.13 0.09 0.15 0.07 0.16 ±0.03
Ethanola 0.03 0.01 0.01 0.01 0.03 ± 0.01
Hydrogen 0.05 0.05 0.05 0.05 0.15 ± 0.05
Benzene 0.01 0.01 0.01 0.01 0.02
Parts Per Million By Volume
Ethylene or 200 214 175 149 732 ± 10
acetylene
Ethane 115 50 100 51 117 ± 10
Propylene 20 18 22 14 45 ±5
Propane 17 0.2 12 4 45 ± 1
Unknown No. 1° 2 18 13 1 4
Unknown No. 2 2 1 1 1 2
Unknown No. 3 45 20 25 17 39
l-Butene 11 1 1 1 9 ±2
n-Butane 11111
Hydrogen sulfide 0.5 0.5 0.5 0.5 0.5
Carbonyl sulfide 0.5 0.5 0.5 0.5 0.5
Carbon disulfide 0.5 0.5 0.5 0.5 0.5
Sulfur dioxide 11111
a Primarily ethanol.
^Calculations based on normal sensitivity.
Effect of Stove Operation
on Combustion
A wood heating appliance is generally
operated at a fixed rate of combustion air.
although thermostatically controlled air
dampers are available. Decreasing pyrolysis
and combustion rates during the burn cycle
results in: (1) increasing excess air levels, and
(2) decreasing combustion gas temperatures.
These are compound effects since increased
excess air thermally dilutes the combustion
gas temperature. This effect can be observed
in Figure 5 where excess air and combustion
gas temperatures are plotted relative to time.
Catalyst Performance
With a catalytic converter in the stove, the
catalyst bed heats up very quickly: as it
reaches about 260°C (500°F), the catalytic
action becomes effective and secondary
combustion starts. Even when the gas tem-
perature entering the catalyst goes slightly
below 260°C (500°F), the catalyst is still ef-
fective and maintains ignition with at least
partial burning of the combustible material.
Figure 6 shows a typical test with peak
temperatures for gas leaving the burning
wood at about 315°C (600°F) as they enter
the catalyst. This temperature falls off
throughout the remainder of the burn.
reaching a temperature just above 175°C
(350°F) at the end of the test burn. This,
referred to as the primary gas flow, is subse-
quently mixed with heated secondary air
upon entering the catalyst. Figure 7 shows
the typical catalyst temperature over a burn
cycle. These temperatures rise after heated
air is added to the primary combustion pro-
ducts, and secondary combustion begins.
This temperature goes much higher and
stays high throughout about two thirds of
the test period, indicating active catalytic
combustion.
When the catalyst was replaced by an
identical ceramic substrate that had not been
coated with the catalyst, the early tempera-
tures were slightly lower and the blank cat-
alyst temperature dropped to a lower value
late in the test run. Figure 7 shows the dif-
ferences dramatically: the upper curve is with
the activated catalyst, and the lower curve
is with the blank catalyst.
Effect of Catalytic Action
on THC and CO Emissions
Secondary combustion by catalytic activa-
tion of combustion products can be initiated
at temperatures around 260°C (500°F). The
measured effect on emissions is a reduction
in THC and CO levels measured before and
after the catalyst. Figure 8 shows the effect
of the catalytic action on hydrocarbon emis-
sions. Emissions are reported in pounds of
total hydrocarbons per thousand pounds of
-------
200-
ISO-
• '
Time, Woi/rs
Time. Hours
-1*
O X
80.
160
' I
072
Time, Hours
Entering Catalyst
Figure 3. Emission factors with active catalyst —Test 4.
20Q.
160.
120-
,ES
uj r-
40-
0/
77/776, A/ours
0/2
Time, Hours
3
75-
1
1
2
Time, Hours
Entering Catalyst
Time, Hours
Leaving Stove
wood burned. Therefore, the hydrocarbon
concentrations going into the catalyst
(shown by the upper curve) and the hy-
drocarbon emissions leaving the stove
(shown by the lower curve) are normalized
to the same reporting units. There is a signifi-
cant reduction in hydrocarbon emissions
through two-thirds of the burn period using
an active catalyst. After that period,
presumably the wood has been reduced to
char, and burning after that point has low
hydrocarbon emissions. These data corres-
pond to hydrocarbons with boiling points
less than about 177°C (350°F). Total
hydrocarbon data are measured in the FID
as methane equivalents and are not abso-
lutely quantitative. They do represent valid
comparative values for evaluating catalyst
performance.
CO emissions are also effectively reduced
by catalytically initiated combustion. In most
cases, CO emissions downstream of the
catalyst were 30 to 50 percent of the emis-
sions entering the catalyst (Figures 3 and 4).
Neither CO nor THC emissions were sig-
nificantly reduced in the one test which used
an inactive catalyst (Figure 5).
In examining the effectiveness of the
catalyst in reducing CO and THC, two com-
bustion processes must be considered: (1)
the catalytic surface reactions which are
diffusion-limited (for hydrocarbons the
assumed reaction is directly to CO2 and H20
without formation of CO); and (2) the gas-
phase reactions which are kinetically limited
and are assumed to react by oxidation to CO,
H20, and H2, followed by downstream con-
version to C02 and H20. Therefore, the ex-
istence of CO in the catalyst outlet is
evidence of gas-phase reactions. In a cat-
alyst-equipped appliance, the suspected
reaction is surface catalysis of hydrocarbons
and gas-phase reaction of CO emissions. An
unknown effect in wood combustion is the
action of the catalyst on hydrocarbons with
oxygen-containing substituents; e.g.,
aldehydes, ketones, phenolic groups, and
hydroxy groups. For most oxygenated com-
pounds, bond energies would indicate that
the CO group would be eliminated and then
oxidized to CO2 by gas-phase reactions.
Figure 4. Emission factors with active catalyst —Test 5.
-------
180
Time, Hours
125
Time, Hours
Entering Catalyst
Figure 5. Emission factors with inactive catalyst—Test 6.
Leaving Stove
£ *uu
Q)
.C
-------
(871) 1600
(760) 1400
S_
2- (649) 1200
Q)
a
m
1
| (538) 1000
K
5
jt
1 (427) 800 <
0
(316)600
(304)400
CO
**+ co° 0*° o ° o ° Active Catalyst, Test 5
+** <** *b *
-c^, ^^ o + o Inactive Catalyst, Test 6
** Txd**^> .. o
%
+ CO
\v
•»• o
t_ O
•* o
v*\+* 0.5 1 1.5 2 2.5 3
7»ne,
Figure 7. Temperatures within catalyst.
75 r-
50
I
0.5
1 1.5
Time, Hours
+ Gas into catalyst
O Gas leaving stove
2.5
Figure 8. Hydrocarbon emission rates with active catalyst —Test 5.
-------
J. Allen, W. Piispanen. and M. Cooke are with Batelle-Columbus Laboratories,
Columbus, OH 43201.
Michael C. Osborne is the EPA Project Officer (see below).
The complete report, entitled "Study of the Effectiveness of a Catalytic
Combustion Device on a WoodBurning Appliance," (Order No. PB84-171 545;
Cost: $8.50, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
PS
------- |