United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-82-003
May 1983
Air
Emission Factor
Documentation
ForAP-42:
Section 1.10,
Residential
Wood Stoves
-------
EPA-450/4-82-003
Emission Factor Documentation For AP-42
Section 1.10, Residential Wood Stoves
by
Pacific Environmental Services, Inc.
Contract No. 68-02-3511
EPA Project Officer: William H. Lamason, II
Prepared for
U.S. Environmental Protection Agency
Office of Air, Noise and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
May 1983
-------
Disclaimer
This report has been reviewed by the Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency, and
approved for publication as received from Pacific Environmental
Services, Inc., Durham, NC. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for
use. Copies of this report are available from the Air Management
Technology Branch, Monitoring and Data Analysis Division, Office of
Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711.
ii
-------
Table of Contents
Section Page
1.0 Introduction 1
1.1 Identification of Pollutants 2
1.2 Variables Affecting Emissions
Characterizations 4
2.0 Procedure for Revision of Emission Factors 7
2.1 Nitrogen Oxide Emissions 7
2.2 Sulfur Oxide Emissions 9
2.3 Carbon Monoxide Emissions 9
2.4 Particulate Emissions 12
2.5 Volatile Organic Compound (VOC) Emissions .... 14
2.5.1 Polycyclic Organic Materials (POMs). ... 16
3.0 Emission Factor Ratings 19
4.0 References 21
APPENDIX A - Emission Measurement Programs 24
APPENDIX B - Test Results Extracted From References
and Used in Emission Factor
Determinations 37
ill
-------
Number
List of Tables and Figures
Page
Figure 1
Histogram 2-1
Histogram 2-2
Figure 2
Histogram 2-3
Histogram 2-4
Table 1
Table 2
Example Generic Designs of Wood
Stoves Based on Flow Paths . . ,
Nitrogen Oxide Emissions ...
Carbon Monoxide Emissions. . .
Carbon Monoxide Concentration in
Flue Gas as a Function of Time .
Particulate Emissions,
Volatile Organic Compound (VOC)
Emissions
Low Molecular Weight Hydrocarbon
Emissions
Polycyclic Organic Material (POM)
Emissions
3
8
10
11
13
15
17
17
iv
-------
1.0 INTRODUCTION
Emissions from residential wood combustion and associated air
quality impacts have only recently been the subject of intensive
investigation. A wide range of emission rates has been reported;
many results are not comparable because operating conditions were
not equivalent. The purpose of this effort is to assemble and
organize emissions data from past and ongoing research, screen this
data for quality and recentness, and incorporate it into the AP-A2
emission factor file. Improved emission factors will provide a
more realistic assessment of the impact of residential wood combustion
on local air quality.
Coal and waste fuels were not included in computing emission
factors because of the relative scarcity of test data available.
More research is being done in this area and possibly the next
effort at revision will be supported by a data base enabling the
generation of representative emission factors for coal and waste
fuel burning stoves. Coal and wood waste logs burn at significantly
higher temperatures than cordwood. The performance of various
heaters within a given type will vary, depending on how a particular
design uses its potential performance advantages. Much of the
available emissions data came from studies conducted on stoves
designed for woodburning.
Wood stoves are used primarily as domestic space heaters to
supplement conventional heating systems, particularly in the
Northeastern and Northwestern United States. The two basic designs
for wood stoves are radiating and circulating. 1 Common construction
materials include cast iron, heavy-gauge sheet metal and stainless
steel. Radiating type stoves transfer heat to the room by radiation
from the hot stove walls. Circulating type stoves have double wall
construction with louvers on the exterior wall to permit the conversion
of radiant energy to warm convection air. Properly designed, these
stoves range in heating efficiency from 50 to 70 percent. 1 Radiant
stoves have proven to be somewhat more efficient than the circulating
type.
The thoroughness of combustion and the amount of heat
transferred from a stove, regardless of whether it is a radiating
or circulatory model, depend heavily on firebox temperature, residence
time and turbulence (mixing). The "three Ts" (temperature, time
and turbulence) are affected by air flow patterns through the stove
and by the mode of stove operation. Many stove designs have internal
baffles that increase the residence time of flue gases, thus promoting
heat transfer. The use of baffles and secondary combustion air may
also help to reduce emissions by promoting mixing and more thorough
combustion. Unless the secondary air is adequately preheated, it
may serve to quench the flue gas, thus retarding, rather than
-------
enhancing, secondary combustion. Secondary combustion air systems
should be designed to deliver the proper amount of secondary air at
the right location with adequate turbulence and sufficient temperature
to promote true secondary combustion.
Stoves are further categorized by the air flow pattern through
the burning wood within the stoves. Example generic designs -
updraft, downdraft, crossdraft and "S-flow" - are shown schematically
in Figure 1.2
In the updraft air flow type of stove, air enters at the base
of the stove and passes through the wood to the stove pipe at the
top. Secondary air enters above the wood to assist in igniting
unbumed volatiles in the combustion gases. Updraft stoves provide
very little gas-phase residence time, which is needed for efficient
transfer of heat from the gases to the walls of the stove and/or
stovepipe.
The downdraft air flow type of stove initially behaves like an
updraft. A vertical damper is opened at the top rear to promote
rapid combustion. When a hot bed of coals is developed, the damper
is closed, and the flue gases are then forced back down through the
bed of coals before going out the flue exit.
The side or cross draft is equipped with a vertical baffle
(open at the bottom) and an adjustable damper at the top, similar
to the downdraft. The damper is open when combustion is initiated,
to generate hot coals and adequate draft. The damper is then
closed. The gases must then move down under the vertical baffle,
the flame is developed horizontally to the fuel bed, and the gases
and flame ideally come in contact at the baffle point before passing
out the flue exit.
The S-flow, or horizontal baffle, stove is equipped with both
a primary and a secondary air inlet, like the updraft stove.
Retention time within the stove is a function of both the rate of
burn and the length of the smoke path. To lengthen the retention
time, gases are kept from exiting directly up the flue by a metal
baffle plate located several inches above the burning wood. The
baffle plate absorbs a considerable amount of heat and reflects and
radiates much of it back to the firebox. The longer gas phase
residence time results in improved combustion when the proper
amounts of air are provided, and it enhances heat transfer from the
gas phase. <-
1.1 Identification of Pollutants
Residential combustion of wood produces atmospheric emissions
of particulates, sulfur oxides, nitrogen oxides, carbon monoxide,
organic materials including polycyclic organic matter (POM), and
mineral constituents. Organic species, carbon monoxide and, to a
large extent, the particulate matter emissions result from incomplete
-------
J
ZUjLHl
sc
s*
Air
Or Up Drill
-l
B B B
P - Primary Air Supply
S - Secondary Air Supply
E - Exhaust to Slack
B — Primary Burning
SC — Secondary Combustion
\
J
sc
f
Side or Cross Draft
Down Draft
SC
B
B
B
SFlow
Figure 1 Generic designs of wood stoves based on flow paths.
-------
combustion of the fuel. Efficient combustion tends to limit
emissions of carbon monoxide and volatile organic compounds by
oxidizing these compounds to carbon dioxide and water. Sulfur
oxides arise from oxidation of fuel sulfur, while nitrogen oxides
are formed both from fuel nitrogen and by the combination of atmospheric
nitrogen with oxygen in the combustion zone. Mineral constituents
in the particulate emissions result from minerals released from the
wood matrix during combustion and entrained in the combustion
gases.
Wood smoke is composed of unburned fuel - combustible gases,
droplets and solid particulates. Part of the organic compounds in
smoke often condenses in the chimney or flue pipe. This tar-like
substance is called creosote. If the combustion zone temperature
is sufficiently high, creosote burns with the other organic compounds
in the wood. However, creosote burns at a higher temperature than
other chemicals in the wood, so there are times when it is not
burned with the other products. Creosote deposits are a fire
hazard, but they can be reduced if the exhaust ductwork is insulated
to prevent creosote condensation, or the exhaust system is cleaned
regularly to remove any buildup.
Polycyclic organic material (POM) , a. minor but potentially
important component of wood smoke, is a group of organic compounds
which includes potential carcinogens such as benzo(a)pyrene (BaP).
POM results from the combination of free radical species formed in
the flame zone, primarily as a consequence of incomplete combustion.
Under reducing conditions, radical chain propagation is enhanced,
allowing the buildup of complex organic material such as POM. POM
is generally found in or on smoke particles, although some sublimation
into the vapor phase is probable.
1.2 Variables Affecting Emission Characteristics
Before discussing the effects of combustion conditions on wood
stove emissions, several comments should be made. First, residential
combustion equipment is difficult to fine tune and the user is
generally not a trained operator. Also, because of the small size
of the combustors, the low flue gas flow rates, and the many types,
sizes, and methods of firing the wood, exact quantitative measurements
are hard to make. Because of the difficult nature of residential
wood heating emissions characterization, the results measured in
different efforts are often contradictory.4
Emissions from any one stove are highly variable and correspond
directly to different stages in the burning cycle. A new charge of
wood produces a quick drop in firebox temperature and a dramatic
increase in emissions, primarily organic matter. When all of the
volatiles have been driven off, the charcoal stage of the burn is
characterized by relatively clean emissions.
-------
Emissions of particulate, carbon monoxide and volatile organic
compounds were found to depend on burn rate. Emissions increase as
burn rates decrease, for the great majority of the closed combustion
devices currently on the market. A burn rate of approximately
three kilograms per hour has been determined representative of
actual woodstove operation. 18,21
Wood is a complex fuel, and the combined processes of combustion
and pyrolysis which occur in a wood heater are affected by changes
in the composition of the fuel, moisture content and the effective
burning surface area. 2,4 The moisture content of wood depends on
the type of wood and the amount of time it has been dried (seasoned).
The water in the wood increases the amount of heat required to
raise the wood to its combustion point, thus reducing the rate of
pyrolysis until moisture is released. Wood moisture has been found
to have little affect on emissions. Dry wood (less than 15 percent
moisture content) may produce slightly higher emissions than the
commonly occuring 30 to 40 percent moisture wood. However, firing
very wet wood may produce higher emissions due to smoldering and
reduced burn rate. The size of the wood also has a large effect on
the rate of pyrolysis. For smaller pieces of wood, there is a
shorter distance for the pyrolysis products to diffuse, a larger
surface area-to-mass ratio, and a reduction in the time required to
heat the entire piece of wood. One effect of log size is to change
the distribution of organics among the different effluents (creosote,
particulate matter and condensible organics) for a given burn rate. 14
These results also indicate that the distribution of the total
organic effluent among creosote, particulate matter and condensibles
is a function of firebox and sample probe temperatures.
Results of ultimate analysis (for carbon, hydrogen and oxygen)
of dry wood types are within one to two percent for the majority of
all species. The inherent difference between softwood and hardwood
is the greater amount of resins in softwoods, which increases their
heating value by weight.
Several combustion modification techniques are available to
reduce emissions from wood stoves, with varying degrees of effectiveness,
Some techniques relate to modified stove design and others to
operator practices. Proper modifications of stove design (1) will
reduce pollutant formation in the fuel magazine or in the primary
combustion zone or (2) will cause previously formed emissions to be
destroyed in the primary or secondary combustion zones.15
A recent wood stove emission control development is the catalytic
converter, a transfer technology from the automobile. The catalytic
converter is a noble metal catalyst, such as palladium, coated on
ceramic honeycomb substrates and placed directly in the exhaust gas
flow, where it reduces the ignition temperature (flash point) of
the unburned hydrocarbons and carbon monoxide. Retrofit cayalysts
tend to be installed in the flue pipe farther downstream of the
woodstove firebox than built-in catalysts. Thus, adequate catalyst
-------
operating temperatures may not be achieved with the add on type,
resulting in potential flue gas blockage and fire hazards. Limited
testing of built-in designs indicates that carbon monoxide and
total hydrocarbon emissions are reduced considerably, and efficiency
is improved, by the catalyst effect. Some initial findings also
indicate that emissions of nitrogen oxides may be increased by as
much as a factor of three. Additionally, there is concern that
combustion temperatures achieved in stoves operating at representative
burn rates (approximately 3 kilograms per hour or less) are not
adequate to "light off" the catalyst. Thus, the catalytic unit
might reduce emissions but not under all burning conditions. 22
In light of the relationships between wood stove emissions and
variables affecting these emission characteristics both quanti-
tatively and qualitatively, the variables employed in various
studies' should be noted (see Appendix A) . The importance of these
relationships is that they suggest a need to carry out environmental
testing over appropriate ranges that can be related to actual stove
usages.
-------
2.0 PROCEDURE FOR REVISION OF EMISSION FACTORS
In the course of this project a search was conducted to identify
programs in which actual measurements of woodstove emissions have
been or are being made, wherein some correlation is possible with
woodstove design and operating conditions. The programs are few,
and there has been no consistency in emission testing protocol for
woodstove configuration or operation, sampling of emissions, analysis
of collected samples, and interpretation of results.
Major problems exist in the characterization of the chimney
and draft systems used which have significant effects on both stove
operation and ultimate emissions, and emission sampling systems
used which arbitrarily separate gaseous from particulate emissions
depending on the temperatures of the point of particulate capture
and retention. Although the measurements and observations of
emissions have been related to stove parameters, these other factors
can materially affect the emissions as measured.
As will be described, there is a wide range in some of the
values listed. Also, some of the numbers listed are only based on
a limited number of measurements and the operating conditions under
which these measurements were made were not always what might be
considered typical.
2.1 Nitrogen Oxide Emissions
Nitrogen oxide formation depends primarily on fuel nitrogen
content, amount of excess air used, combustion temperature, and
design of combustion equipment. NOX emissions from wood stove
combustion were found to range from 0.1 g/kg to 7.1 g/kg for wood
stoves (references 3, 4, 15, 17, 22, and 23). An average NOX
emission factor of 1.4 g/kg (2.8 Ib/ton) was determined from test
data plotted on Histogram 2-1.
A statistical analysis was performed by DeAngelis, et al., to
determine the effect of combustion equipment and wood type on
nitrogen oxide emission rates. The results support the observed
difference in nitrogen oxide emissions between fireplaces and
stoves; however, no significant difference was found to exist
between emissions from the wood types tested. The data reported by
Allen and Cooke15 indicates a trend for NOx emissions to increase
as excess Oo increased for stoves tested.
Recent studies have shown emissions of nitrogen oxides to be
up to three times greater from a catalytic stove in comparison to a
non-catalytic unit.22
-------
4
A
oo
V
* 2?
C-
l-o
C
C-
z-a
M.
g.
-------
2.2 Sulfur Oxide Emissions
Sulfur oxides are formed during combustion by the oxidation of
sulfur in the fuel. The sulfur content of wood is low, typically
ranging from 0.02 to 0.1 weight percent in branches and wood components.
DeAngelis has examined emissions of sulfur oxides during
residential wood combustion, and that study was limited to two
tests using EPA Method 6. Results were close to the detection
limit for the method, and gave emission factors of 0.16 g/kg and
0.24 g/kg. These values are nearly equal to those expected based
on material balance considerations, assuming 100 percent conversion
of fuel sulfur to S02- In both cases, if all the sulfur were
converted to S02, the emissions would be 0.2 g/kg, which is what
was found.
2.3 Carbon Monoxide Emissions
Histogram 2-2 lists carbon monoxide emission data from wood
fired stoves. Reported emissions ranged from 10 g/kg to 420 g/kg
and were found to be dependent on burn rate. For the purpose of
establishing an emission factor, a burn rate of 3 kg/hr was determined
to be representative of actual woodstove operation. 18,21 Data
taken at this burn rate or below were arithmatically averaged to
obtain an emission factor of 130 g/kg (260 Ib/ton). Low and high
burn rate data are plotted separately on Historgram 2-2 in order
that a trend can more easily be seen. Averaging of only high burn
rate data yielded an emission factor of 106 g/kg; a 25 percent
decrease relative to the low burn rate factor.
Findings by Hayden and Braaten ' ' indicate a tendency for lower
carbon monoxide emissions as the firing rate was increased for the
same stove. The excess air level decreased at the same time. They
also found that the side draft type of stove gave the least CO
emissions of the stoves they tested. The data collected by the
California Air Resources Board 17 show less emissions for a Fisher
stove over a Franklin, but these two stoves on the average showed
less carbon monoxide emissions than the stoves tested by DeAngelis,
etal.4 .The stoves tested by DeAngelis were operated at a much
larger heat input rate and, in fact, at a rate which is probably
higher than they would normally use. For these tests on different
stoves, higher loads were giving higher carbon monoxide emission
rates. The study by DeAngelis, et al. ,4 indicates that carbon
monoxide formation is very sensitive to changing fuel bed conditions
which may account for the variability between replicate test results.
The actual carbon monoxide emissions will vary with time as the
wood is burned up and more is added. Figure 2 presents carbon
monoxide concentrations in the flue gas from a wood-burning stove
versus time over one combustion cycle (i.e., one fuel loading
charge). This graph shows greater than an order of magnitude
change in carbon monoxide concentration in 20 minutes.
22
-------
Z.-Z-
A.
c.
•p-
f.
i
O
n
ZZ
-------
70 J3
TIME I WTO IURMNB CTOfc «in
Figure 2 Carbon monoxide concentration in the flue gas from a wood-
burning stove as a function of time.
n
-------
Data collected by Allen and Cooke on a Defiant stove operated
in the side draft mode did not show that this type of operation
emitted less carbon monoxide as found by Hayden and Braaten. For
some stoves, Allen and Cooke did find less carbon monoxide emitted
on a g/kg basis as the burning rate increased. They also found
more carbon monoxide emissions from smaller pieces of wood (air
dried oak). They did not find any trend in carbon monoxide emissions
due to moisture content of the wood. For the airtight appliance,
it seems that CO emissions for cordwood fuel decrease as burn rate
increases while emissions for the oak brands increase with increasing
burn rates.
2.4 Particulate Emissions
Though the data base is still limited, particulate matter has
probably been the most studied emission from stoves. Interpretation
of particulate emissions data is difficult because of the large
amount of condensible organic material present in the exhaust gas.
Test results are strongly dependent on the methods used to collect
samples, and different studies have employed different methods
(i.e., high-volume sampling and EPA Method 5)- The standard method
recognized by EPA for collection of particulate samples is EPA
Method 5. Normally everything collected on or before the filter in
the front half of the sampling train is considered the particulate
catch. The front-half catch is made up both of organic and inorganic
materials. A major frjictipn of the gases passing through the front
filter is cjondensed in the back-half of the Method 5 sampling
train. These condensed organics are made up of hydrocarbons and
oxygenated (C-0) species. As is commonly done, the condensed
or_ganics_ were added to the front filter catch to determine the
emission factor for total particulates.
Q
The Oregon Department of Environmental Quality found approximately
60 percent of the mass of particulate emissions to be condensible
organic materials, which are gaseous in the stack but condense td
form liquid droplets in the atmosphere. In addition, a consistent
emissions pattern was observed in which the greatest mass of particulates
were emitted during the early stages of the burn (i.e., over 50 percent
of the emissions captured by filter occurred during the first
17 percent of the burn period). This is consistent with data
reported by Butcher and Sorenson. °
Thirteen studies have characterized total particulate emissions
from residential wood stoves (Reference nos. 3,4,5,6,8,9,10,13,14,
17,22,24, and 25). As shown in Histogram 2-3, total particulate
emissions (i.e., front and back-half catch combined) ranged from
0.5 g/kg to 97.3 g/kg and were found to be dependent on burn rate.
For the purpose of establishing an emission factor, a burn rate of
3 kg/hr was determined to be representative of actual woodstove
operation. 18,21 Data taken at this burn rate or below were arithmetically
averaged to obtain an emission factor of 21 g/kg (42 Ib/ton).
Low and high burn rate data are plotted separately on Histogram 2-3
12
-------
2-3
I
/a - -
A
z.
M.y
>.4
tef-
».&
1
H.
. /3
\. /7
». zz
,. 24
>. 25"
O n
D T
At
_
O
-------
in order that a trend can more easily be seen. Averaging of only
high burn rate data yielded an emission factor of 9.2 g/kg; a
decrease of approximately 57 percent relative to the low burn rate
factor. The low burn rate factor has been derived and footnoted
to the effect that burn rates greater than 3 kg/hr may result
in particulate emission reductions of 55 to 60 percent.
18
Barnett and Shea measured less particulate matter emissions
from wet wood (30 to 40 percent moisture) than from drier wood
(20 percent moisture). This is based on one test and data from the
other authors do not necessarily support this.
Barnett and Shea were able to modify stove design to lower
particulate emissions. By modifying the draft opening locations
they were able to reduce particulate matter emissions by up to
50 percent. The stove modified had no internal baffles and was
tested over the range of 0.01 to 0.03 kg/min of wood burned. They
also tested other stove designs and found that the low crossdraft
type of internal baffle gave the least emissions. It should be
noted that at high heat input rates, DeAngelis, et al.4, found no
difference in particulate matter emission rates between the two
stoves (baffled and unbaffled) that they tested.
14
A study by Hubble, et al. suggests that one effect of log
size, as well as firebox temperature, is to change the distribution
of organics between the different effluents (i.e., creosote,
particulate matter, and condensible organics).
Although the methods of collection and quantification are not
consistent, each study indicates a degree of variability in emission
factors which can be attributed to the variable nature of the
combustion process. • Factors such as the addition of fresh wood
charges, fuel bed configuration, size of fuel charge, etc., all
have some effect on emissions generated from wood stoves.
2.5 Volatile Organic Compound Emissions
The organic emission data have been reported in several different
forms. It has been reported as total hydrocarbon emissions, condensible
organics, volatile hydrocarbons, and by classes such as POMs.
Hydrocarbon emissions from wood-burning stoves have been measured
in seven studies (Reference nos. 3,4,11,15,17,22 and 23). Measurement
of volatile hydrocarbons was accomplished in three studies 3,4,17 by
collection of a gas sample in an inert gas sampling bag and subsequent
injection into a gas chromatograph with a flame ionization detector
(GC/FID). Continuous flue gas monitors were utilized by the TVA studies 22,23
The studies by Hayden and Braaten TI, Allen and Cooke 15, and TVA,
refer to emissions as "total hydrocarbons" and label the data as a
semi-quantitative measurement of all organics.
The study by Kosel '' indicates that comparison of emission
factors for volatile organic compounds may be difficult because of
the divergent test methodologies involved. Some studies may group
14
-------
A-
8.
c.
. 3
- 4
- n
/)*•
6
f-
13
A e
t> o t>
6
e£
t.
SL
£.£
f
c.
L t>
@L
O
D
- /n
-------
volatile organics with particulate matter, since all test programs
found particulate matter to consist largely of condensed organic
compounds (back-half catch).
The data show a tendency for higher burning rates to produce
lower total hydrocarbon emission rates; however, this is only a
trend and the data do not always show the emissions rate decreasing
as the heat input rate increases. Allen and Cooke 15 found that
smaller pieces of wood produced less total hydrocarbon emissions
but got mixed results for the effects of dry wood.
Histogram 2-4 lists volatile organic compound emissions from
wood stoves. Volatile hydrocarbons represent those hydrocarbons
that will remain as gases when dispersed within the atmosphere.
Those hydrocarbon having high molecular weights will condense when
diluted with ambient air. The measured values ranged from 0.3 g/kg
to 113 g/kg. It became evident early in the review process that
emissions of VOC were very dependent on burning rate. High emissions
associated with low burning rates (hold-fire conditions) were
observed.
The VOC emission factor is based on low burn rate data only
(less than 3 kg/hr). In all, 25 data points were obtained at burn
rates of 3 kg/hr or less. Based on 25 data points, an average VOC
emission factor of 51 g/kg (100 Ib/ton) was determined. Using high
burn rate data (117 data points) an average emission factor of
approximately 26 g/kg was obtained. Therefore, a footnote was
added to the emission factor table stating that if burn rates of
greater than 3 kg/hr were employed, VOC emissions may decrease by
as much as 55 to 60 percent. VOC speciation performed by DeAngelis,
et al.4 found approximately 0.5 g/kg of the low-molecular-weight
hydrocarbon emissions to consist of methane (see Table 1).
2.5.1 Polycyclic Organic Materials (POM's)
Polycyclic organic matter (POM) can be formed in any combustion
process involving compounds of carbon and hydrogen. It is uncertain
whether POM condenses out as discrete particles after cooling or
condenses on surfaces of existing particles after formation during
combustion. POM molecules have a high free energy of sublimation
and several recent publications dealt with the gas-phase solid-phase
equilibrium in source and ambient samples. The implication of this
work is that gas-phase POM would have greater access to the deep
lung than particulate associated POM. The vapor-solid equilibrium
that exists in ambient air between discrete polycyclic organic
compounds and ambient particulate matter is the subject of a great
deal of current research. Most of the mass of POM is associated
with particles of aerodynamic diameters in the respirable size
range, 15,19
16
-------
TABLE 1 LOW-MOLECULAR-WEIGET HYDROCARBON EMISSIONS'
(g/Xg)
Emission scecies
Methane, CJU
Ci to Ca hydrocarbons
Ethylene, CaH»
Ethane, CaH«
Ca to C3 hydrocarbons
Propylene, C3H«
Propane, C3H«
C3 to C« hydrocarbons
Butyl ene, G»H«
Butane, G»Hi0
C« to Cs hydrocarbons
Pentene, CsHio
Pentane, CsHia
Cs to C« hydrocarbons
Hexene, C«Hi3
Hexane, C«Hi»
>C« hydrocarbons
Total
Note: Blanks indicate
a
Fireolace Baffled stove
seasoned oaX seasoned Dine
0.5
0.2
0.5 0.3
0.1
0.08
0.08
0.5
15 0.5
<0.08
2.6
0.5 0.5
19 2.3
emissions not detected.
Nonbaffled stove
Green oaX Green pine
0.02
2.4
0.04
0.6
0.3
0.3 3.0
Determined by GC/FID at test site.
TABLE 2 POM EMISSIONS
(g/kg)
***» «•*-
AathracMM/ptonantAra
MtAyl-anc
Cyalapw>ca-«ichncm««/-ph«aantlira
FluoronclMM
Mthyl-fluoran
ta*> ««"'nttir«c«n««
-b««tr«ainn
•KAyl coolant
ladaao < 1.1. J^d) pyrxx
O.XM3
-------
Emissions of organic material greater than Cg were collected
for analysis by the POM train (a Method 5 procedure modified by the
addition of an XAD-2 resin trap between the filter and impinger
system) and the source analysis sampling system (SASS). 4,14,15
Most organic compounds with boiling points above 120°C, and several
that have boiling points between 100 to 120°C, are collected by the
XAD-2 resin used in the POM train. The SASS train system employs a
set of three cyclones and a filter for particle size fractionation
in addition to a solid sorbent (XAD-2) trap for organic species.
The condensed organic material was extracted from the XAD-2 resin
and from the remainder of the sampling train and submitted for
analysis by GC/MS.
The average factors for wood stove emissions were found by
DeAngelis, et al.4 to range from 0.19 g/kg to 0.37 g/kg. An attempt
was made to identify the organic compounds present in the flue gas
(see Table 2). Allen and Cooke 15 also measured POM emissions from
wood stoves. Their values averaged about 0.03 g/kg (0.01 to 0.05 g/kg
range) which is much less than the values measured by DeAngelis.
Both sets of tests were conducted at fuel feed rates of about 0.12
to 0.18 kg/min. There were no apparent reasons for the differences
in the data.
?2 23
POM emissions data obtained by TVA t}, are in general agreement
with the Monsanto^ results for high gurn rates, but low burn rate
POM emission factors were not obtained in the Monsanto study.
TVA's results also compare closely with a similar POM emission
factor reported in a Battelle study! Swhere again POM emission
factors at low burn rates were not reported. The Argonne Laboratory
study'^reported two POM emission factors, one at a low burn rate
and another at a medium burn rate. Both are much lower than any
POM results obtained by either TVA, Monsanto, or Battelle. The
small recovery of POM material associated with the Argonne test
effort could possibly be associated with the sample recovery system
used by Argonne Laboratory. The TVA, Monsanto, and Battelle studies
added an XAD-2 resin trap to the EPA Method 5 Sampling System
specifically to recover organics from the flue gases. The Argonne
study did not use an XAD-2 trap or any specific mechanism designed
to trap organics. The Argonne study reported very low temperatures
at its impingers, but the impingers alone are not considered adequate
for entrapment of the very diverse range of organics associated
with wood combustion flue gases. Differences in the handling
procedure and type of gas chromatography employed in quantification
of the POMs may have also contributed to the smaller POM recovery
in the Argonne study. TVA observed that there is a general tendency
for POM emission factors to increase with decreasing burn rate.
This does not support the commonly espoused principle held by some
that relates increased POM emission factors with increased thermal
activity, burn rate, and stove exit temperature.
18
-------
3.0 EMISSION FACTOR RATINGS
The data base on emissions from residential wood combustion
equipment is both limited and diverse. The apparent differences in
experimental results are probably due to legitimate but not well
defined differences, including: (1) difficulty in quantifying and
controlling combustion conditions, (2) variety in sizes and designs
of equipment, (3) nonuniformity in properties and condition of
wood, and (4) different firing methods and firing conditions.
Although the important parameters affecting emissions are not well
characterized, the major ones appear to be combustion design,
firing rate, and excess air level.
Findings are indicative rather than conclusive due to
difficulties in comparing tests of diverse methodologies. For an
explanation of emission factor ratings as discussed in this section,
refer to the document "Technical Procedures for Developing AP-42
Emission Factors and Preparing AP-42 Sections", Monitoring and Data
Analysis Division, Research Triangle Park, North Carolina, April 1980.
Most of the tests performed were based on a sound methodology
and reported in enough detail for adequate validation; however, in
the case of particulate emission measurements, several studies were
based on methodologies (i.e., high-volume sampling) that are not
necessarily EPA reference method tests. Some data were excluded
from consideration because of test series representing incompatible
test methods (i.e., comparison of EPA Method 5 front-half with EPA
Method 5 front- and back-half). The EPA Method 5 procedure for
particulates proved to be difficult to adapt to woodstove sampling;
extremely low velocities and sample rates, stack diameters are
generally smaller than the minimum (12 inches) for which the EPA
method is specified, and the labor intensity involved. However,
the manner in which the sources were operated is well documented in
the reports, and most sources were operating within typical parameters
during the test. For these reasons a C rating is assigned to the
particulate emission factors.
Many variations can occur during testing which cannot be
explained. Such variations can induce wide deviations in sampling
results. In the case of VOCs, a large variance between test
results was not explained by information contained in the test
reports and some of the data are suspect. An emission factor
rating of D is therefore assigned to VOCs.
Nitrogen oxide and carbon monoxide emissions sampling were
performed according to EPA Methods 7 and 3, respectively, and
results are fairly consistent among the studies conducted. However,
due to the limited number of data points available, these pollutants
will each be assigned a C rating.
19
-------
Sulfur oxide emission factors were determined from material
balance based on the sulfur content of wood fuel. This value was
further confirmed by testing and is assigned an A rating.
20
-------
4.0 REFERENCES FOR SECTION 1.10
1. H. I. Lips and K. J. Lim, Assessment of Emissions from
Residential and Industrial Wood Combustion, EPA Contract
No. 68-02-3188, Acurex Corporation, Mountain View, CA,
April 1981.
2. D. G. DeAngelis, et al., Source Assessment: Residential
Combustion of Wood. EPA-600/2-80-042b, U.S. Environmental
Protection Agency, Washington, D.C., March 1980.
3. J. A. Cooper, "Environmental Impact of Residential Wood
Combustion Emissions and Its Implications", Journal of the
Air Pollution Control Association, _30.(8): 855-861, August 1980.
4. D. G. DeAngelis, et al., Preliminary Characterization of
Emissions from Wood-fired Residential Combustion Equipment,
EPA-600/7-80-040, U.S. Environmental Protection Agency,
Washington, B.C., March 1980.
5. S. S. Butcher and D. I. Buckley, "A Preliminary Study of
Particulate Emissions from Small Wood Stoves", Journal of the
Air Pollution Control Association, _27_(4): 346-348, April 1977.
6. S. S. Butcher and E. M. Sorenson, "A Study of Wood Stove
Pariculate Emissions", Journal of the Air Pollution Control
Association, _29_(9): 724-728, July 1979.
7. J. W. Shelton, et al. , "Wood Stove Testing Methods and Some
Preliminary Experimental Results", Presented at the American
Society of Heating, Refrigeration and Air Conditioning Engineers
(ASHRAE) Symposium, Atlanta, GA, January 1978.
8. D. Rossman, et al., "Evaluation of Wood Stove Emissions",
Oregon Department of Environmental Quality, Portland, OR,
December 1980.
9. P. Tiegs, et al. , "Emission Test Report on Four Selected
Wood Burning Home Heating Devices", Oregon Department of Energy,
Portland, OR, January 1981.
10. J. A. Peters and D. G. DeAngelis, High Altitude Testing of
Residential Wood-fired Combustion Equipment, EPA-600/S2-31-127,
U.S. Environmental Protection Agency, Washington, D.C.,
September 1981.
11. A. C. S. Hayden and R. W. Braaten, "Performance of Domestic
Wood-fired Appliances", Presented at 73rd Annual Meeting of
the Air Pollution Control Association, Montreal, Canada,
June 1980.
21
-------
12. R. J. Brandon, "An Assessment of the Efficiency and Emissions
of Ten Wood-fired Furnaces", Presented at the Conference on
Wood Combustion Environmental Assessment, New Orleans, LA,
February 1981.
13. B. R. Hubble and J. B. L. Harkness, "Results of Laboratory
Tests on Wood-stove Emissions and Efficiencies", Presented at
the Conference on Wood Combustion Environmental Assessment,
New Orleans, LA, February 1981.
14. B. R. Hubble, et al., "Experimental Measurements of Emissions
from Residential Wood-burning Stoves", Presented at the
International Conference on Residential Solid Fuels, Portland,
OR, June 1981.
15. J. M. Allen and W. M. Cooke, "Control of Emissions from Residential
Wood Burning by Combustion Modification," EPA Contract No.
68-02-2686, Battelle Laboratories, Columbus, OH, November 1980.
16. J. R. Duncan, et al., "Air Quality Impact Potential from
Residential Wood-burning Stoves", TVA Report 80-7.2, Tennessee
Valley Authority, Muscle Shoals, AL, March 1980.
17. P. Kosel, et al., "Emissions from Residential Fireplaces",
GARB Report C-80-027, California Air Resources Board, Sacramento,
CA, April 1980.
18. S. G. Barnett and D. Shea, "Effects of Wood Burning Stove
Design on Particulate Pollution", Oregon Department of
Environmental Quality, Portland, OR, July 1980.
19. J.A. Peters, POM Emissions from Residential Woodburning:
An Environmental Assessment, Monsanto Research Corporation,
Dayton, OH, May 1981.
20. Source Testing for Fireplaces, Stoves, and Restaurant Grills
in Vail, Colorado (Draft), EPA Contract No. 68-01-1999, Pedco-
Environmental, Inc., December 1977.
21. A.C.S. Hayden and R.W. Braaten, "Effects of Firing Rate and
Design on Domestic Wood Stove Performance", Canadian Combustion
Research Laboratory, Ottawa, Canada. Presented at the Residential
Wood and Coal Combustion Specialty Conference, Louisville, KY,
March 1982,
22. C.V. Knight and M.S. Graham, "Emissions and Thermal Performance
Mapping for an Unbaffled, Airtight Wood Appliance and a Box
Type Catalytic Appliance", Proceedings of 1981 International
Conference on Residential Solid Fuels, Oregon Graduate Center,
Portland, Oregon, June 1981.
22
-------
23. C.V. Knight et. al., "Tennessee Valley Authority Residential
Wood Heater Test Report: Phase I Testing", Tennessee Valley
Authority, Chattanooga, Tennessee, November 1982.
24. Richard L. Poirot and Cedric R. Sanborn, "Improved Combustion
Efficiency of Residential Wood Stoves", prepared for U.S.
Department of Energy, Grant no. DE-FG41-80R110340, September 1981,
25. Cedric R. Sanborn, et. al., "Waterbury, Vermont: A Case Study
of Residential Woodburning", Vermont Agency of Environmental
Conservation, August 1981.
23
-------
Appendix A
Emission Measurement Programs
24
-------
APPENDIX A
Emission Measurement Programs
A-l. Acurex Corporation (Reference No. 1)
Lips and Lim, 1981
An assessment is made of emissions and their control from
wood-fired boilers, home stoves, and fireplaces. Although the data
base is limited and diverse, the important parameters affecting
emissions appear to be combustor design, excess air level, and
burning rate. For boilers, optimal emissions control and efficiency
are achieved with 50 to 200 percent excess air, with overfire air
maximized for stokers. Data on other combustion modification
techniques and operational impacts are minimal. On a normalized
basis, residential wood-burning units tend to produce higher emissions
than do boilers. Data on wood-fired stoves indicate that certain
designs, e.g., crossdraft, emit less particulate matter. The
limited data suggest that the low burning rates, typical of home
stoves, are conducive to POM formation, a pollutant of major concern.
As the combustion conditions for fireplaces are even more difficult
to quantify and control, emission trends for those devices are less
well established. However, POM emissions appear to be lower from
fireplaces than from stoves.
A-2. Monsanto Research Corporation (Reference No. 2)
DeAngelis, et al., 1980
This report presents a review of characterization data for
emissions from residential wood combustion and an evaluation of
their potential environmental effects. It describes several types
of residential wood combustion equipment, the 1976 geographic
distribution of wood-fired equipment, fuel characteristics, and
combustion chemistry. Primary and secondary wood heating, as well
as wood burning for aesthetic purposes, are all covered in the
report. Emission control technology and possible future trends of
the source are also discussed.
A-3. Oregon Graduate Center (Reference No. 3)
Cooper, 1980
The primary objectives of this study were to review the available
information on potential environmental impacts of residential wood
combustion, note areas of conflicting data, and suggest areas of
future research.
The study concludes that emissions from residential wood
combustion appliances are a major source of winter air pollution in
a large part of the country. These emissions, if allowed to increase,
will likely have a significant impact on public health, current and
future EPA standards, industrial growth, and EPA regulatory policy.
25
-------
A-4. Monsanto Research Corporation (Reference No. 4)
DeAngelis. 1979
Tests were conducted early in 1979 at Auburn University. The
University staff operated the stoves and measured the efficiencies,
and Monsanto staff conducted the measurement and analysis program.
Two similar stoves and a fireplace were operated burning both pine
and oak cardwood, using both seasoned and green wood. In addition
to the normal combustion gases (0 , CO , SO , NO , and CO), the
emission measurements included particulate matter, condensable
organics, volatile organics, aldehydes, major organic species, and
polycyclic organic species. The tests were all operated at rela-
tively high burning rates, and did not demonstrate the effects of
burning rate on emissions. No significant differences were found
in criteria pollutant emissions or efficiencies between the baffled
and nonbaffled stoves.
The effects of wood type and moisture content were shown to be
generally low, with green pine producing particulate and organic
emissions slightly higher than the other three woods burned. Many
POM species were identified in the stack emissions (particulates
and gaseous samples combined) and bioassays of these specimens were
found to be both mutagenic and cytotoxic.
A-5. Bowdoin College, Maine (Reference No. 5)
Butcher. 1977
This was a preliminary study of particulate emissions from
small wood stoves. The results of this test program were used to
develop the particulate emission factor range originally published
in the AP-42.
Two stoves (baffled and nonbaffled) were tested during this
1977 study. The sampling method used consisted of collecting all
stack emissions for selected time intervals with a high-volume
sampler. For this purpose a 4 foot diameter sheet metal cone,
connected to the high-volume sampler, was positioned over the flue.
The dependence of the emission factor on draft setting indicated an
increase in the emission factor as the available air was reduced.
A-6. Bowdoin College, Maine (Reference No. 6)
Butcher, 1979
This experimental program has focused on particulate emissions
at low burning rates. All emitted particulate matter was collected
on a filter after the flue gas has been cooled and diluted with
large quantities of fresh air. No gas composition or excess air
measurements have been made. The implication of the particulate
collection system is that all organic emissions condensible at
ambient conditions will be collected as particles. Approximately
half of the particulate matter collected has been benzene extractable.
Nearly half of this extractable material is neutral with regards to
acid-base extraction procedures, and this fraction would presumably
contain the polycyclic emissions.
26
-------
This study has shown that the particulate emission factor
varies proportionally with the weight of the initial fuel charge
and with the reciprocal of actual burning rate.
A-7. Williams College (Reference No. 7)
Shelton, et al., 1978
This study summarizes wood stove testing methodologies currently
in practice. Indications are that consideration of the assumptions
behind use of the stack loss method, as usually practiced for
determining the energy efficiency and heat output rate of common
heating appliances, illustrates this method's inadequacy when
applied to small wood (and coal) fueled appliances.
In testing and rating wood stoves, careful attention must be
given not only to installation details, but also to fuel and operation
details. Using a calorimeter room and direct measurement of flue
gas flow, it was found that dry wood does not necessarily burn more
completely than moist wood.
A-8. Oregon Department of Environmental Quality (Reference No. 8)
Rossman, 1980
Source tests for particulate emissions were performed on a
typical welded steel, airtight, thermostatic controlled, wood
burning stove installed at the manufacturer's facility. Field work
was done on September 29 and 30 and October 22, 23, and 24, 1980.
The stove was tested in three phases; at low and high settings
as it is normally sold; at high fire with an add-on stack device
called a "Smoke Consumer"; and at low and high fire conditions with
an air control thermostat developed by Stockton Barnett of State
University of New York. A single sample run was made for each of
the test conditions.
This series of testing was conducted using an EPA Method-5
type sampling train with a final filter instead of a High Volume
sampler. This causes a significant amount of "condensible" material
that passes through the first filter in the sampler to be captured
and weighed.
Several conclusions can be drawn from this study:
— About 40 percent of total measured particulate was consistently
found to be composed of "condensible" matter. The condensible
fraction is defined here as that material collected
following a high efficiency particulate filter. This
material is estimated to be at least 60 percent to 85 percent
organic in nature.
27
-------
— The majority of particulate emissions - greater than
90 percent - were emitted during the first half of each
burn (i.e., start-up emissions).
— The EPA Method 5 sampling procedure adapted for this test
proved to be difficult for the following reasons:
—extremely low velocities and sample rates
—stack diameters are smaller than the minimum (12 inches)
for which the EPA method is specified
—labor intensity and cost
The relatively high emissions from this stove (in comparison
with other data reported in the literature) are consistent
and predictable, given that it is a simple box design and
was tested with a maximum charge of wood and relatively
low heat release rate - all of which are factors known to
increase emissions.
A-9. Oregon Department of Energy (Reference No. 9)
Omni Environmental Services, 1981
This report presents the results of an emission test program
for four wood burning home heating devices conducted in January
1981. The objectives of the study program were to better define
the character and amount of emissions produced by burning wood for
home heating purposes.
The Oregon Method 7 (EPA Method 5 with a back-up filter between
the third and fourth impingers) was used for testing of particulates.
Its advantage is that the particulate catch includes both the
high-temperature solids and condensible materials that are emitted
through the combustion of wood. Other methods reported in the
literature, including high-volume samplers and Method 5 trains
without the back half or condensible catch, do not provide informa-
tion on those materials which form particles at lower temperatures.
These low-temperature condensates are potentially as important to
ambient particulate levels as the high temperature materials. For
example, the average back half catch for the tests conducted here
was 280 percent of the front half catch. In addition, the average
back filter was 136 percent of the front filter catch.
Although the woodstoves were tested with drier fuel and under
more standardized test conditions, it is expected that their opera-
tion at a low to medium burning rate influenced their emission
characteristics towards the high side.
A-10. Monsanto Research Corporation (Reference No. 10)
Peters and DeAngelis, 1981
The objective of this emission testing program was to determine
whether emissions from operating a wood stove at high altitude
28
-------
( 7,000 feet) differ from those at low altitude ( 1,000 feet),
through comparison with the MRC-low altitude data (Reference No. 4).
Accordingly, test conditions duplicated one of the test conditions
from the experimental matrix of the previous study, with altitude
as the only variable.
The wood stove was tested during the week of November 18-21,
1980. The sample collection methods employed were EPA Methods 1
through 5 and POM sampling by a modified Method 5 train with an
XAD-2 resin absorbent trap. Condensible organic emissions were
nearly identical between tests and were very close to the value
obtained under low altitude testing conditions. Total particulate
emissions, carbon monoxide, and polycyclic organic matter were
analyzed, and no statistically significant difference in emissions
was found. Although unconfirmed, it is felt that the lower combustion
zone temperatures of the high altitude test, as indicated by stack
gas temperature, suppressed the POM chemical formation mechanisms.
An even cooler fire should, at some point, result in zero POM
emissions since very high temperatures (at least 400-500°C) and a
chemically reducing environment are necessary for POM formation
from the long-chain molecules present in wood cellulose.
A-ll. Canadian Combustion Research Laboratory (Reference No. 11)
Hayden and Braaten, 1980
This laboratory has conducted emissions analyses and efficiency
tests on wood stoves of different generic designs. Continuous
measurements are made of CO, CO , 0 , NO , and total hydrocarbons.
Split cordwood has been burned at relatively low rates while overall
thermal efficiencies have been determined by the stack gas (indirect)
method.
The results indicate that overall efficiencies fall off at the
higher burning rates, even though the emissions of CO and THC per
pound of wood burned decreases and the combustion efficiency increases.
Their comparison of generically different stoves shows a general
increase in emissions of CO and THC (Ib/lb fuel burned) as the
burning rate is reduced. Large variations in emissions between
different designs of updraft stoves were observed.
A-12. Institute of Man and Resources (Reference No. 12)
Brandon, 1979
This organization has conducted a field demonstration program
on ten different residential furnaces. The emission measurements
included CO, NO , SO , and particulates, with short time emissions
measured under all operating modes of each unit. The particulate
measuring system consisted of diluting and filtering the entire
flue gas stream in the manner used by Butcher. The particulate
samplers operated at temperatures below 100°C, and sampled for
periods up to 10 minutes. The filter catch was extracted with
benzene and dried to determine the condensable organic content
which varied from 13 to 51 percent of the total particulate matters
collected.
29
-------
A-13. Argonne National Laboratory (Reference No, 13)
Hubble and Harkness, 1981
Results from the testing of one generic type of air-tight
stove (horizontally baffled, updraft) are presented as functions of
burn-rates and stove efficiencies.
From the viewpoint of environmental assessment, the results of
this study point out the importance of testing over an appropriate
burn-rate range, one that is related to actual stove usage. Heat-output
requirements for dwellings dictate usage on the lower end of the
burn-rate range associated with commercial stove designs. The
three parameters measured in this study (carbon monoxide emissions,
particulate emissions, and creosote build-up) all increase as the
burn rate is decreased, with the latter two parameters increasing
markedly at the lower end of the burn-rate range tested. The
results show that the worst possible effects associated with these
parameters do actually occur under the conditions at which such
stoves are operated a significant fraction of time. In addition,
the log size used in the stove was shown to have an effect on
emissions, as well as on combustion parameters. Therefore, for
realistic environmental assessments, testing should be done using
log-size ranges actually employed by residential stove users.
A-14. Argonne National Laboratory (Reference No. 14)
Hubble et al., 1981
Organic emissions were measured from an air-tight wood burning
stove operated in a manner consistent with typical residential
heating requirements. Test data are reported for condensible
organic and total organic emissions as a function of burn rate, and
estimates of emission factors of individual organic compounds and
compound classes are given. Test results indicate the following:
o The total organic emissions (particulate matter, creosotes,
and condensible organics) decrease as the burn rate increases.
o The distribution of the organic compounds in the total
emissions (a function of burn rate and of firebox tempera-
tures) shifts from particulate matter and creosote to
condensibles as the burn rate and temperatures increase.
o The total organics do not exhibit a log-size effect, but
the forms of the organics (particulate matter, creosotes,
and condensibles) do exhibit such an effect.
o A comparison of high-burn-rate and low-burn-rate chemical
characterization results indicates that the high burn rate
produces:
30
-------
(a) a. simpler mixture of compounds;
(b) shifts to multi-ring aromatics, and thus higher POM
emission rates;
(c) greater concentrations in the condensible organic
constituent, with corresponding decreases in particu-
late matter and creosote.
o Estimated emission factors for POMs were significantly
lower than other published results.
o The results of this study and previously published studies
do not provide the required information for quantitative
assessment of the impacts of wood stove emissions on public
health and the environment.
A-15. Battelle Columbus Laboratories (Reference No. 15)
Allen and Cooke, 1981
The report describes an exploratory study of factors contributing
to atmospheric emissions from residential wood-fired combustion
equipment. Three commercial appliances were operated with both
normal and modified designs, providing different burning modes:
updraft with a grate, updraft with a hearth, crossdraft, downdraft,
and a high-turbulence mode utilizing a forced-draft blower. Fuels
were naturally dried commercial oak cordwood, commercial green pine
cordwood, oven-dried fir brands, and naturally dried oak cut into
reproducible triangles. Continuous measurements of stack gases
included 0 , CO , CO, NO , SO , and total hydrocarbons (FID). The
THC instrument provides a continuous indication of organic content
in the flue gas. Because the specific composition of these organic
species is constantly changing, a definitive calibration of the
instrument for actual flue gas is not possible; although calibrated
using methane, the instrument reading is interpreted as a semi-quantitative
measurement of all organics.
Several combustion modification techniques were identified
which have an appreciable effect on emission factors and, therefore,
can be developed and applied to reduce emissions in consumer use.
The more promising design modifications include: prevention of
heating the inventory of wood within the stove but not yet actively
burning, focusing the air supply into the primary burning area with
high turbulence, and increasing the temperatures in the secondary
burning regions of the appliances.
A-16. Tennessee Valley Authority (Reference No. 16)
Duncan, et al., 1979
This agency initiated an experimental evaluation of several
wood stoves proposed for customer use in regions of their utility
district. The program was initiated with the measurement of efficiency
by the indirect method, and emissions using on-line instrumentation.
The stoves were fired with kiln-dried fir, using the UL batch
31
-------
burning procedure in which a test consists of burning a fresh
charge of wood to 95 percent completion on a weight basis. Three
rates of burning were maintained for each stove in separate runs.
After a destructive fire occurred at the TVA facility in
Chattanooga, Battelle conducted additional tests for TVA at Battelle's
Columbus Laboratories. These continued tests were modified to
include the burning of air-dried oak in larger triangular pieces,
simulating typical residential split cordwood. A conversation with
Karen Knight of TVA indicates that test results are preliminary and
no published material is yet available.
A-17. California Air Resources Board (Reference No. 17)
Kosel, 1977
This agency conducted tests in 1977 on two free-standing
stoves in both residential and laboratory test installations.
Fuels used were oak, pine, and coal, and relatively high burning
rates were maintained with frequent additions of wood. Particulate
emissions were determined using EPA Method 5. Bag samples were
used for total hydrocarbon and other gaseous determinations at
another location using gas chromotography.
A-18. New York University of Plattsburgh (Reference No. 18)
Barnett and Shea, 1981
This program of stove development has focused on overall
efficiency and particulate emissions as affected by stove design,
operating procedures, and controls development (thermostatic). The
emission measurements have been limited to particulates as measured
by collection on a cooled but not temperature-controlled filter
from stack gases withdrawn through a probe. The velocities of
specimen withdrawal and stack flow are measured by hot wire anemometers,
with sampling velocities greatly exceeding stack velocities.
Particulate collections have been made using 1 minute sampling
periods at 10 minute intervals throughout run periods of several
hours. Stove operation has generally been at very low burning
rates, i.e., stove exit temperatures 140 to 500°F.
A-19. Monsanto Research Corporation (Reference No. 19)
Peters, 1981
This paper presents some of the results of a test program
conducted by the EPA, Monsanto Research Corporation, and Auburn
University to characterize emissions from wood-burning equipment,
specifically, two air-tight stoves and a heat circulating fireplace
while burning four varieties of wood. POM's were collected using a
modified EPA Method 5 train and a SASS train.
32
-------
Over 75 organic compounds were identified and quantified upon
characterization of the organic material present in the flue gas;
22 of these were POM's. The POM's accounted for up to about 35 percent
of the mass of organics identified by GC/MS. The total POM emission
factor was found to be an order of magnitude lower for the fireplace
than for either wood-burning stove; this is consistent with the
carbon monoxide and nitrogen oxide results which indicate more
efficient combustion and/or higher combustion temperature in the
fireplace.
A-20. Pedco Environmental, Inc.
1977
Emission tests were conducted from September 21-30, 1977, on
two fireplaces, a stove, and a restaurant grill exhaust in Vail,
Colorado. Tests were conducted under various operating conditions
to establish the probable range of emission rates. Measurements
were made of total and condensible particulate, carbon dioxide,
carbon monoxide, and gaseous hydrocarbons as well as the velocity
and temperature of the flue gas.
A total of fifteen tests were conducted on the fireplaces. No
significant differences in emission rates of any of the pollutants
were found between the two fireplaces, between dry and green wood,
or between pine and aspen. The condensible portion or "back-half"
catch averaged 75 percent of the total particulate loadings. No
indication was given as to the sampling techniques used to quantify
"hydrocarbons", and the hydrocarbon test results were highly erratic.
A-21. Canadian Combustion Research Laboratory
Hayden and Braaten, 1982
Controlled combustion wood stoves have been shown to have a
significant degree of incompleteness of combustion and corresponding
high emissions. This paper shows that emission levels of carbon
monoxide, unburnt hydrocarbons as measured by a flame ionization
detector and polycyclic organic matter (POM's) are closely related,
and are very sensitive to firing rate below a "critical rate",
which is stove dependent. Above this, the firing rate has much
less effect on emission levels. Two technical strategies to reduce
emission levels by improving combustion are primarily effective in
shifting the critical burn rate to lower levels, with little effect
on emissions at higher burn rates. This makes the determination of
the critical rate one of the more important factors in the evaluation
of wood-fired appliances. Typical home heat demands are shown to
be below the critical burn rate of a typical controlled-combustion
stove for much of the heating season, resulting in field emissions
which may be much higher than those indicated by laboratory tests
above this rate. One way of reducing emission levels is to improve
combustion design. This is also likely to reduce creosote formation,
thus increasing safety; increased efficiency is another likely
benefit.
33
-------
The experimental program described in this paper was designed
to determine the effect of firing rate on emissions, to define a
more reliable technique for measuring unburned hydrocarbons continuously
and to determine if appliances with more sophisticated combustion
designs offered performance with reduced emissions, relative to
conventional wood-fired appliances.
A-22. Tennessee Valley Authority
Knight and Graham, 1981
This paper presents draft results of TVA's current efforts
toward* development of performance maps for several units tested,
i.e., a nonbaffled, airtight wood appliance very common to the TVA
region, and a new box catalytic model that can potentially offer
major improvements in thermal efficiency and reductions in overall
emissions. The two heating appliances were operated over the full
range of supply air control and wood fuel loading with at least
twenty tests being completed for each type of fuel. Continuous
sampling of the flue gases was used to indirectly determine the
thermal efficiency while a Modified EPA Method 5 system was used
for the recovery of an integrated flue gas sample. Carbon monoxide
(CO) nitrous oxides (NO ), sulfur dioxide (SO,,), and total hydrocarbons
(THC) emission factors were developed using a continuous analyzer
system. The remaining emission factors for filterable particulates,
condensible organics, volatile hydrocarbons, and polycyclic organic
materials were developed using a Modified Method 5 System.
A-23. Tennessee Valley Authority
Knight et. al, 1982
This study involved the evaluation of efficiencies, emissions,
and other environmental effects for eight airtight, residential
wood heaters. These models include a fireplace insert, free standing
radiant models and circulators (both free and forced circulation).
Data was gathered at Battelle Laboratories under contract to TVA
and the results were analyzed and models were developed by TVA.
Emissions monitoring equipment consisted of a system of continuous
analyzers, thermocouples, and a sampling train using a Modified EPA
Method 5 system. Continuous monitors for six gas species (C02, CO,
0-, SO-, NO , and THC) were obtained and utilized in arriving at
thermal efficiencies and emission factors for each test. This test
program did not address back half filter catch with regard to
particulate emissions. Fir and oak fuels were tested over a burn
rate range of 4 to 31 pounds per hour (1.8 to 14.1 kilograms per
hour).
34
-------
A-24. Vermont Agency of Environmental Conservation
Poirot and Sanborn, 1981
This document attempts to duplicate actual operating
conditions as closely as possible and also incorporates a complete
burn cycle for each test. Burn rates ranged from 2.2 to 4.8 kg/hr
(4.8 to 10.6 Ib/hr). By incorporating a complete burn cycle within
each test,the effects of inter-cycle variability were kept to a
minimum. Only baseline test results were included in the AP-42
data base, as baseline conditions represent stove configurations as
the manufacturer intended prior to modification of existing secondary
air systems by personnel conducting the emissions testing.
The premise of the study is that the majority of secondary air
systems presently available introduce air at far too low a temperature
to promote true secondary combustion. The intention was to develop
new secondary air systems for two existing wood stove designs which
would deliver the proper amount of secondary air at the right
location with adequate turbulence and sufficient temperature to
promote true secondary combustion.
A-25. Vermont Agency of Environmental Conservation
Sanborn, et. al, 1981
To address the question of how much particulate and condensible
organics an average woodburning appliance emits, testing of a full
range of woodburning appliances took place between December 1979
and March 1981. Emission tests were conducted on fourteen units
which included airtight stoves, parlor stoves, box stoves and wood
furnaces. Tests were made using either EPA Method 5 or a Hi-Volume
sampling method.
It was determined that the average total particulate emission
rate (both front half particle and back half organics) was 19.1 g/kg
with a range of 6.6 g/kg to 42.2 g/kg. The condensible organic
portion (back half recovery) averaged 37 percent of this weight
with a range of 15 to 64 percent. It was also found that a wood
furnace has a lower emission rate than the wood stoves do. The
burning rate for the stoves varied from 1.8 to 4.7 kg/hr. In all
cases, the fuel used was seasoned hardwood.
A-26. Del Green Associates, Woodburn, Oregon (Supplemental Reference)
Neulicht, 1981
This company is conducting a program for Region X, U.S. EPA,
concerned with wood stove utilization in their area. The multi-task
program includes a source sampling task to (1) determine effects of
wood moisture content on emissions, (2) develop simplified testing
procedure, and (3) develop reasonable standards for stove emissions
using 6 stoves in their program. Another task includes indoor air
35
-------
pollution measurements in a few residences heated by wood burning
within the living space.
This document, when finalized, will be obtained and placed in
the background file under "New Additional References" to be included
in any future revision effort.
36
-------
Appendix B
Test Results Extracted From References and Used
in Emission Factor Determinations
37
-------
Co
A774
I i-0
x/7
3
-------
ft
a
z
H8
f&
V/8
4J5
3.£
2--S
3-7-
7.0
z.
-------
REFERENCE NO. 4 (CONT'O)
LOW-MOLECUIAR-WEIGHT HYDROCARBON EMISSIONS'
(g/kg)
Emission species
Methane, CHi*
Ct to Ca hydrocarbons
Ethylene, CjH*
Ethane , CjH«
Cj to Ca hydrocarbons
Propylene, CaH«
Propane, CaHe
Ca to Ci. hydrocarbons
Sutylene, C*Hg
Butane , C*H 1 o
Ca to Cs hydrocarbons
Pentene, CjHio
Fireolace Baffled stove
seasoned oak seasoned pine
0.5
0.2
0.5 0.3
0.1
0.08
0.08
0.5
15 0.5
<0.08
Nonbaffled stove
Green oak Green pine
0.02
2.4
0.04
0.6
Pentane,
Cs to C« hydrocarbons
Hexene, C«His
Hexane, C«H1tt
>C« hydrocarbons
Total
2.6
0 . 5
19
0.5
2.3
0.3
0.3
3.0
Note: Blanks indicate emissions not detected.
a
Determined by GC/FID at test site.
-------
* % C^4rr~&* f^ &&, fr^-Atf 2*6'/I «f
ft^ter^
* 3.f
* 3.4
/*a
/o.a
3/./
2o-o
Z/.4
2-7-3
c/
^A
^3
47
to
42.
35-
zs-
27.
22.
e
S3
42. ^ /3.$ /o./
W
0.4
-------
*
f.fi
/./sr
7.7,9
AT
/o6
//£
CD
Z13
Z1&
&tf
31
Zo
IS"
27
X
4>
-------
o.S
22.-Z.
2-3
S38
/jo
0-iT.~Z.S3
3.o
1.1
27
<* S7*&* **J
/STZ,
2-7
Z.f
7?
5". ^
^. -#Li ^^
.004
. oz>£
SUjOfUcj j \]*Gi&!&. >T^*»TAC
-------
( /faro /* tavr*
&A M* -7#e?
77 22^ /./
3«> A£
236 /.*
/•<
* 7
IS a./
3± *I
*H 0-1
6* 2-1
61 /.&
/» 7° ^ //?
/f J5 25 /f
2* 37 /7 ^
^ ^* ^
37 £3 <3^
Sf t>£ £
£ II t
74 ^ f
^ ^ /^
37 df J4
3 2.0 /
^7 A^ A
5-f /^ 33
-------
/T
&
4-1-3
7V
zz.8
22>
f 3
77
xz-3
*/
-------
. 22.
SB
/<«_
I
2-
/r
X
a
«•
c
7
5
/3
X
27
2f
37
57
^2.
53
<22
/77
^
^f
^
757
£B
4*
^
1
*+
27^
73
/»7
^
^
77
/w
/2T
^
/*«
14
1
/?
/•z.
/a
2»
<7
^7
Jo
n
57
53
A/2.
-M;
A 38
A^"
*2?
^
/?
O^SI
^-77
/?
3f
z£ 0.37
/r -
0*11,
-------
. 23
/sa
/ar
4V
S3.1
77-3
34.0
4;1
23.o
'•7
*/
Z-l
• i)
•af
5*
/»
7
sy
•fe
75-
/ar
/^/
^
/'T?
^7
/r?
275"
77/0
#.*
3*-°
214
ns
^. 3
^/3
^L
27-5-
/•$
z-7
r-T.
^
/^
'•1
2. -2.
•25
-------
28.2.
2&.o
25.?
/S.I.
25"
-7-a
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO.
EPA-450/4-82-003
3. RECIPIENT'S ACCESSION>NO.
FLE AND SUBTITLE
Emission Factor Documentation For AP-42:
Section 1.10, Residential Wood Stoves
5. REPORT DATE
, Mav 1983
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Pacific Environmental Services Inc
1905 Chapel Hill St.
Durham, NC 27707
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3511
12. SPONSORING AGENCY NAME AND ADDRESS
US Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer: William H. Lamason, II
16. ABSTRACT
Emissions from wood combustion in residential stoves and associated effects
have only recently been the subject of intensive investigation. A wide range of
emission rates has been reported, and many results are not comparable because
operating conditions under test are not equivalent. The purpose of this effort
is to assemble and organize emissions data from past and continuing research, to
screen these data for quality and freshness, and to incorporate them into the AP-42
emission factor file. This report discusses variables affecting emissions,
documents how AP-42 emission factors were recalculated, and rates the quality of
the factors. Emission measurement programs reviewed and actual test data used in
the revision are summarized in appendices.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
13. DISTRIBUTION STATEMEN1
19. SECURITY CLASS (This Reporl)
21. NO. OF PAGES
20. SECURITY CLASS (This page I
22. PRICE
EPA Form 2220-1 (9-73)
------- |