Un.ted States
Environmental Protection
Agency 1981
OF
BY
for
of
1-10
by
NC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
May, 1981
FINAL REPORT
on
CONTROL OF EMISSIONS FROM
RESIDENTIAL WOOD BURNING
BY COMBUSTION MODIFICATION
by
John M. Allen and W. Marcus Cooke
Battelle Columbus Laboratories
Columbus, Ohio 43201
Contract Number 68-02-2686
Task Directive 114
Project Officer
Robert E. Hall
Combustion Research Branch
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
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.
-------
ABSTRACT
This report describes an exploratory study of the factors
contributing to atmospheric emissions from residential wood-fired com-
bustion equipment. Three commercial appliances were operated with both
normal and modified designs, providing different modes of burning.
Operating conditions included up draft using a grate, up draft using
hearth, cross draft, down draft, and a high turbulence mode utilizing
a forced draft blower. Fuels used were naturally dried commercial oak
cordwood, commercial green pine cordwood, oven-dried fir brands, and
naturally dried oak cut into reproducible triangular shapes. Continuous
measurements of stack gases included 02, C02, CO, NO, S02, and total
hydrocarbons (FID) as an indication of the total organic species in the
stack gases during batch type operation. Several combustion modification
techniques were identified which have an appreciable effect on emission
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
focus of air supply into primary burning area with high turbulence, and
increase the temperatures in secondary burning regions of the appliances.
-------
CONTENTS
Abstract ii
Figures iv
Tables v
Acknowledgments vi
1. Introduction 1
2. Conclusions and Recommendations 2
3. Wood as a Fuel ' 8
Properties of Wood 9
4. Combustion Characteristics of Wood 10
Evaporation 10
Pyrolysis 10
Char Formation and Burning 11 •
Burning Progression 12
Wood Constituents Resulting in Atmospheric Pollutants .... 12
Production of PAH in Wood Combustion 15
5. Wood Stove Design and Performance 18
Types of Radiant Stoves 18
Secondary Combustion 22
Commercial High Turbulence Burner 23
6. Experimental Program 29
Laboratory Facility 29
Operating Procedures 31
Emission Measurements 31
PAH Measurements 33
Data Processing 35
Fuel Wood Used 35
7. Experimental Program Observations 36
Stove Design Factors 39
Operator Factors 52
Fuel Properties 55
Polycyclic Aromatic Hydrocarbon Emissions 61
Commercial High Turbulence Burner 64
References 66
Appendices
A. Emission Measurement Programs 68
B. Examples of Recorded Test Data 76
C. Polycyclic Aromatic Hydrocarbon Measurements in Flue Gas 82
iii
-------
FIGURES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
B-l
B-2
B-3
B-4
B-5
C-l
C-2
Pyro-Combustion to Produce Benzo(a)pyrene
Generic Designs of Wood Stoves Based on Flow Paths
Limits of Ignition at 20°C
Limits of Ignition at 270°C
Limits of Ignition at 250°C
Wood Fired Residential Boiler Utilizing High Turbulence
Burning
Laboratory Installation of Stove with Stack and Vent System. .
Continuous Gas Sampling System
Particulate and PAH Sampling System
Configuration of Stove Converted to Down Draft Mode
Air Preheat System
CO Emissions for One Stove, Three Modes of Burning
THC Emissions for One Stove, Three Modes of Burning
Effect of Wood Size on CO Emission Factors
Composition of Gas From Up Draft Burning with a Grate
Composition of Gas From Side Draft Burning
Composition of Gas From Down Draft Burning
Composition of Gas From High Turbulence Burning
Weight Loss During Side Draft Burning
Gas Chromatogram of PAH Standard Mixture
Gas Chromatogram of Particulate Associated Organic Compounds
From Seasoned Oak Combustion
Page
16
19
24
25
26
28
30
32
34
42
44
56
57
62
77
78
79
80
81
84
85
IV
-------
TABLES
Number Page
1 Typical Progression of Local Phenomena in Wood Log Burning . 13
2 Emissions and Operating Conditions 37
3 Emissions From Different Modes of Burning Fuel: Oak 4x4. 41
4 Effects of Preheated Primary Air on Emissions 45
5 Effects of Secondary Combustion on Emissions 50
6 Effects of Size of Charge of Emissions 53
7 Effects of Wood Moisture Content 59
8 Effects of Wood Size and Shape on Emissions 60
9 Emissions from Burning Green Pine 63
10 Emissions from High Turbulence Burner 65
C-l Polynuclear Aromatic Species Quantified 86
C-2 Total PAH (mg) Emitted in Wood Stove Emissions 87
C-3 Comparison of Stove Operating Parameters and Emission Factors
for PAH Compounds 89
C-4 Comparison of Reported PAH Emissions 91
v
-------
ACKNOWLEDGMENTS
We wish to acknowledge the assistance of Mr. Robert E. Hall of EPA for
his encouragement and guidance as Task Officer of this program, and his
cooperative position concerning related research.
Appreciation is also expressed to others on the Battelle staff who
contributed significantly to this program: Paul Webb who assembled the
test facility and instrumentation, Dale Folsom who conducted many of the
experimental runs, and Abbott Putnam who developed the limits to combustion
in the secondary combustion region. The authors would also like to express
their appreciation to Marcia Nishioka and Fred Moore for performing the PAH
analyses.
vi
-------
SECTION 1
INTRODUCTION
The use of wood for residential heating is certainly not a new
phenomenon. However, this use of wood has increased appreciably in recent
years and the trend is expected to continue. It has recently become in-
creasingly apparent that (1) the emissions from residential wood stoves can
in some instances constitute a serious environmental problem, and (2) that
the technology required to burn wood in small stoves without significant
emissions is not now available. Because the use of this renewable energy
resource is to be encouraged, it is necessary to develop and apply the
technology necessary to make the stoves environmentally acceptable, rather
than restrict their use on environmental grounds.
This study was initiated to explore the combustion modification
techniques that might be applied beneficially to wood stoves, and more
specifically to identify those techniques which show promise of providing
a significant improvement in emissions when adequately developed and applied,
The focus of the effort is on naturally-drafted, hand-fired radiant stoves,
as these constitute the majority of units used extensively for residential
heating.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
As the result of this study, there appear to be several combustion
modification techniques available to reduce the gas emissions from wood
stoves, with varying degrees of probable effectiveness, and different levels
of acceptability to residential stove builders and operators. These techniques
are briefly summarized below, noting whether a modified stove design: is
required, special operator techniques are required, or specific as a guide
to indicate the more promising combustion modifications which should be
developed in more detail to justify extensive reduction to practice by
stove users. Modifications are presented first which prevent gas composition
formation in the magazine (A), and in the primary combustion region (B),
followed by those which cause previously formed gas combustion to be
destroyed within the primary burning zone (c) or in the secondary combustion
zone (D). Finally, a few add-on devices are identified (E) which may or may
not be incorporated into the stove, but can reduce gas composition from the
complete heating system. In each category the modifications are presented
in decreasing order as to their expected likelihood of being effectively
reduced to practice.
To apply most of the modifications noted below, and all those
involving destruction of previously formed pollutants, will require stove
design alterations. These are generally beyond the ability of most
residential stove operators. A few very significant gas composition
reductions by operator technique modifications can be obtained using
existing stoves, if the operators are sufficiently motivated to change
their procedures or characteristics of the fuel they burn.
A. PREVENTION OF GAS COMPOSITION FORMATION
IN FUEL MAGAZINE
Wood is placed within a wood stove in a quantity to support
burning for a prolonged period. Considering that the magazine receiving
this wood is usually an integral part of the combustion chamber, the entire
-------
charge of wood can become heated, resulting in pyrolysis off-gasing. When
these gases leave the magazine and the stove unburned, they contribute
directly to the organic gas composition of the stove. Several techniques to
reduce this source of gas composition are as follows:
(1) All the wood inventory within the stove should be
kept at a low temperature until active local
burning is started. This will preclude extensive
pre-burning pyrolysis. Primarily design modifi-
cations are needed, possibly with an altered
operator technique.
(2) Small charges of wood should be placed in the
stove at one time in order to prevent extensive
pre-burning pyrolysis. This is an operator
characteristic which at first consideration is
undesirable to an operator who wants long burning
periods with no attention. It should be effective
in most existing stoves as a means of reducing
emissions.
(3) After combustion has been established, only large
pieces of wood should be fired, consistent with
desired burning rate maintenance. This is a fuel
characteristic that may require more frequent
attention or demanding operator techniques. By
this means pyrolysis can be kept to a minimum in
existing designs of stove.
(4) Very low moisture fuel, kiln dried, should be dis-
couraged, particularly when placed in large
quantities in a magazine. Pre-burning pyrolysis
is more rapid and leads to high gas composition.
This fuel characteristic is less of a problem with
conventional air dried cordwood, but imposes
special operator demands when very dry wood is to
be burned in a low-gas composition manner.
(5) Devolatilized wood such as charcoal is a highly
recommended fuel, generally producing low gas
composition of CO and organic materials. This
fuel characteristic is limited by fuel costs
to the consumer, but is suitable for use in some
wood stoves.
-------
B. PREVENTION OF EMISSIONS FORMATION
IN COMBUSTION REGION
The wood that is actively burning in a stove is subjected to a
restricted supply of air as a means of control of the burning rate. This
deficiency of air can result in incomplete burning of the fuel constituents,
permitting the emission of reduced and incompletely oxidized materials into
the stack and ultimately into the environment.
(1) High burning rates should be maintained as an
operator technique to reduce gas composition.
This may become objectionable to the operator,
as being incompatible with comfort heating when
considering the comfort heating demands and the
heating capacity of the stove.
(2) Primary air should be directed specifically to
a limited burning region or volume, to preclude
widespread burning with the available supply of
air. This is a design factor that may be difficult
to accomplish with a fixed geometry design suitable
for a wide range of burning rates, but might be
accomplished with simple discrete geometry changes
made by the operator within design limits.
(3) A high turbulence level should be maintained in
the active burning area, by using stove design
techniques. The use of an air blower with
tailored air ducting should achieve this objective.
A "within-firebox" source of air to the blower
might be used to minimize the hazards of blower
failure and to provide modulated operation during
a batch type of burn. The high turbulence level
assures improved mixing and thorough combustion
with less stratification of air and combustible
pyrolysis products.
(4) Excess fuel should be eliminated from the primary
burning area as a means of preventing severely
fuel-rich combustion. This situation could partially
be achieved by limiting a continuous flow of fuel
feed into the area to match the desired burning rate.
This situation is generally not attainable with stick
wood but could be accomplished with a stove designed
specifically to burn a free flowing form of wood,
such as pellets. This drastic design approach could
essentially provide fuel-controlled burning.
-------
(5) High temperatures should be maintained in the active
burning area to promote rapid and complete burning
of fuel. This design characteristic should minimize
incomplete combustion associated with quenching of
partially burned fuel, and can be improved by use of
insulating refractory materials enclosing the
combustion chamber.
(6) An abrupt reduction in the rate of primary air supply
to an operating stove does not simultaneously reduce
the rate of pyrolysis of the wood being burned. The
sustained gas evolution by pyrolysis, coupled with
the reduced supply of air, increases the gas composi-
tion of unburned or partially oxydized fuels. This
phenomenon should be avoided, whether it be by an
abrupt operator technique or by severe cycling of a
thermostatic control device.
C. DESTRUCTION OF EMISSIONS IN
PRIMARY BURNING ZONE
The flames in the area of active wood burning are fed by two
sources, combustible emissions from the adjacent wood surfaces, and any
combustible products carried into this area from elsewhere in the stove.
If this local burning is complete, it will consume all the organic materials
and combustibles which would otherwise be emitted. The following approaches
should improve the completeness of burning in this primary burning zone.
(1) Increased turbulence in the primary burning
zone will promote the complete burning of
pyrolysis products released directly into
the flame region, thus eliminating some or
all of the products of incomplete burning.
This requires a design change similar to
that noted in technique B-3 above.
(2) Increase the temperature in the primary burning
zone as in technique B-5 above.
(3) Increase the residence time at active combustion
conditions while maintaining temperatures and
turbulence levels. This is a design approach
that may also become practical with a blower
system as in technique B-3 above.
(4) Pyrolysis products from the magazine region should
be ducted only into an active burning region
where adequate air supply and temperatures will
promote complete combustion. This is a design
consideration.
-------
(5) Develop a down draft design of combustion
chamber that will assure that all gas flow
through the wood inventory travels in the
direction contrary to the progression of the
burning through the wood. When down flow
of air is used, a reduction in the bed
area should accommodate reduced burning
rates while maintaining the counter flow
phenomenon.
D. DESTRUCTION OF EMISSIONS IN
SECONDARY COMBUSTION ZONE
Most stoves include a region or zone within the stove where
secondary combustion can occur, and provide a supply of secondary air to
support this combustion. When secondary combustion is achieved, a signi-
ficant reduction in emissions can be obtained. The following procedures
include techniques to increase the liklihood of achieving this combustion,
and extend the range of conditions that will support it.
(1) The temperatures in the secondary combustion
chamber should be kept high to broaden the
range of gas mixtures that will permit com-
bustion. This is a design characteristic of
the stove using insulation and proximity of
the secondary air inlet to the inlet of
the primary combustion chamber to improve
chamber heat retention.
(2) The secondary air should be heated to provide
the same effect as noted in D-l above, and avoid
quenching secondary combustion that might other-
wise occur. This design factor can be acc-
omplished by within-the-stove heating of secondary
air prior to mixing with primary combustion products.
(3) The envelope of gas mixtures that will permit
combustion of primary combustion effluents and
secondary air can be widened by increasing the
combustibles content of primary effluents. This
shift of operation towards gasification in the
primary combustion chamber could ultimately result
in clean burning with reduced emissions if the
secondary combustion was maintained under all
operating conditions. This requires a major
design alteration, possible with the operator
exercising effective control over both primary
and secondary air supply.
-------
(4) An auxiliary ignition source could be pro-
vided in the secondary combustion space, possibly
supplied by an auxiliary fuel, an electric
ignition source (glow plug) or a "pilot light"
fed from the primary combustion itself. Such
a device to assure ignition would be primarily
a design oriented technique, which might be
augmented by stove operator initiation,
monitoring or control.
E. ADD ON SYSTEMS FOR
EMISSIONS CONTROL
Modifications exterior to the fuel and firebox design can also be
effective. Two techniques are suggested which should significantly and directly
reduce emissions. A third add-on device is possible that can indirectly reduce
emissions. These techniques can be applied as separate devices between the
stove and stack, or by major design modification included within a stove
envelope.
(1) Catalytic afterburners are being developed
and applied to wood stoves with at least
one new design stove currently being marketed.
By reducing the temperature at which com-
bustion can be initiated, and by conserving
the heat released to maintain the combustion,
the catalytic unit might significantly reduce
emissions under many but not all conditions of
burning. This constitutes a design modi-
fication although visual monitoring by the
operator would be advisable.
(2) A separately fueled afterburner could be
developed, preferably using a gaseous fuel.
A significant increase in equipment cost,
auxiliary fuel use, and heat delivery rate
will probably result from such a design change.
(3) A greatly increased heat storage capacity can
be added to a stove, permiting a higher burning
rate operating for short time periods, yet still
providing a near-steady delivery of heat to the
living space. This modification permits all
burning to be at high rates, with the corres-
pondingly reduced gas composition. Heat could
be stored as sensible heat such as with water
or solids (stone or metal), or as latent heat
such as with Glauber's salts or other phase
change material.
-------
SECTION 3
WOOD AS A FUEL
Currently wood combustion provides about 1.5 Quads*of energy to
this country yearly, with about 0.3 Quads used for residential heating.
This residential use of wood should increase to 1.0 Quad in the next 20
years, and actions are planned in support of this goal. Practically all
of the residential wood utilization is now based on cordwood which is
widely available, with most of the burning in residential stoves and fire-
places. Wood stoves are increasingly being recognized as the more efficient
device, and the use of wood stoves has increased rapidly in recent years as
the prices of other residential fuels, especially oil, have increased at a
very high rate. Central furnaces will probably become more prevalent as
commercial distribution of wood is developed, especially processed wood such
as pellets. Probably most of this residential wood is now burned in stoves
and enclosed fireplaces, instead of open fireplaces due to the increased
efficiency and longer burn times for stoves and forced air units. It has
been estimated (1) that the^residential heating with wood might grow from the
present 4 million homes to as many as 15 million by 1985. Due to the general
recognition that airtight stoves provide higher efficiencies than open fire-
places, the increase will probably be primarily in this type of stove.
Various other forms of wood are available commercially in limited
quantities and in limited areas of the country. These include sawdust as
an industrial waste product, wood chips, and wood processed into pellets or
densified logs. Although growing, the use of these other forms is still con-
sidered to be small compared to conventional cordwood, for hand fired
appliances.
The environmental aspects of increasing residential wood burning
fall into two arenas: harvesting with resource depletion considerations, and
burning with objectionable atmospheric emission considerations. The former
* 1 Quad = 10 Btu
-------
can be controlled by using proper silvaculture or forestry management tech-
niques which have been demonstrated by the forest products industry and
local farm agents for many years. The objectionable emissions, however,
are not widely recognized, and the technology for reducing or eliminating
objectionable emissions is not well developed. Both health and visibility
aspects of the emissions are significant.
PROPERTIES OF WOOD
Wood is available as a fuel in most parts of the country, usually
in the form of split logs and/or sticks, known as cord wood for residential
applications. The principal variables in cordwood are hard-wood species or
soft-wood species, and air dried (cured) or green (freshly cut). Dry hard-
woods have a slightly lower heating value than dry soft-woods which generally
have a higher resin content. More important to heating value is the moisture
content, which varies between about 5 percent for very dry wood to over 50
percent for freshly cut wood, with 20 percent being typical for air dried wood.
The chemical composition of dry wood as measured in percent carbon,
hydrogen and oxygen is very similar for hardwoods and softwoods. The elemental
content is typically about 40 to 52 percent carbon, about 6 percent hydrogen,
and 40 to 44 percent oxygen. The cellulose content does vary as follows:
Hard Woods Soft Woods
Cellulose 43 43
Hemicellulose 35 28
Lignin 22 29
When the woods are pyrolyzed the cellulosic materials break down most easily
into volatiles whereas the lignin materials are more char forming.
Cellulose is a high molecular weight polymeric carbohydrate con-
taining nonaromatic glycoside rings joined by oxygen. Cellulose
products thus do not directly contain aromatic species, but can contain many
simple hydrated species. The formation of long chain or fused ring species
may require more severe conditions than those required to initiate simple
pyrolysis.
* The moisture content of wood is measured and reported on either of two
bases: weight of moisture per weight of dry wood, and weight of moisture
per weight of moist or as-fired wood. The latter is used in this
discussion.
9
-------
SECTION 4
COMBUSTION CHARACTERISTICS OF WOOD
The combustion of wood is generally recognized as involving three
processes or phases: moisture evaporation, pyrolysis with subsequent space
burning, and surface char burning. These processes occur successively on
any local particle of wood, but in normal combustion systems there is an
overlap with all three processes occurring simultaneously within the com-
bustion chamber. This overlap is especially significant when prolonged
burning eliminates water early and the other two continue.
EVAPORATION
The amount of moisture evaporation depends on the moisture content
of the fuel as fired, which can vary from over 50 percent for some green wood
to essentially zero for oven-dried wood. As the wood is heated this moisture
is evaporated, beginning at the outer surface where the heat is applied. As
heating progresses the moisture from the center of the piece diffuses as vapor
through the outer material. In large pieces this outer material may have been
heated appreciably above the steam point, even to temperatures at which py-
rolysis or char burning occurs. This efflux of water vapor to the surface of
•the wood retards heat flow into the wood, both from the steam's sensible heat
gain and by a thickening effect on the gaseous boundry layer through which
the initial convective heating usually occurs. This type of cooling thus
retards the burning of wood, especially when large pieces are used and when
moisture content of the wood is high.
PYROLYSIS
Pyrolysis of the wood starts as the local temperatures rise above
100°C, with low temperature pyrolysis being endothermic. These product
gases contain H20, C02» CO, CI^COOH, HCOOH, and others. Although the off-gas
10
-------
has a positive heating value, the steam and C02 content of this initial dis-
charge reduces the heating effectiveness. As the local temperature exceeds
about 280°C, the pyrolysis becomes exothermic, temperatures increase with-
out local oxygen consumption, and larger quantities of gas are obtained.
This off-gas contains additional heavy organic molecules and entrained
particles (droplets) of wood tars. The rate of pyrolysis, especially at the
lower temperatures, is closely dependent on the rate of heat transfer to the
pyrolyzing wood.
The quantity of pyrolysis off-gas (including entrained organic
liquids) is increased appreciably by the rate of pyrolysis. Very slow
pyrolysis leaves half of the original dry weight of the wood as a char, whereas
very fast pyrolysis reduces the char residue to 13 percent of the original dry
weight (2) . The total heating value of the off-gas (Btu/lb wood) from rapid
pyrolysis is thus much higher than that from slow pyrolysis. The ignition
temperature of this off-gas from wood pyrolyzing is about 600°C after being
mixed with air. Thus, burning requires either (a) an adjacent flame or hot
surface, such as other localized burning, or (b) a surface layer through which
the gas diffuses which is sufficiently hot that the 600°C ignition temperature
is retained after mixing with adequate air. The latter is likely only when
the gas evolution rate from internal evaporation and pyrolysis is relatively
low and/or the radiant heating of the surface is sufficiently high.
This off-gas burns vigorously in a flame when it passes through the
surface of the wood allowing reaction with the oxygen supporting the combustion.
Should there be insufficient oxygen, insufficient turbulence to assure
thorough mixing, or inadequate residence time at flame temperatures, some
fraction of these pyrolysis off-gases can leave the combustion area in-
completely burned. This incompletely burned fuel can contain either the
original off-gas species or original species thermally altered into other species.
CHAR FORMATION AND BURNING
The third phase of wood burning occurs when the local wood has been
thoroughly pyrolyzed, leaving a porous carbonaceous char. This char burns at its
surface when temperatures are adequate and sufficient oxygen diffuses to the
surface. This char burning is essentially flameless except for (a) slight
11
-------
gaseous hydrocarbon evolution from the small amount of hydrogen residual in
the char, (b) the partial oxidation to CO by oxygen at the exposed active
surfaces and subsequent burning of the CO, and (c) the watergas reaction with
steam from internal moisture vaporization. Although the ignition temperature
of pure carbon is high, fresh charcoal can sometimes ignite at much lower
temperatures than the pyrolysis off gases (2).
The char has a lower thermal conductivity than the original wood.
Thus, slow burning of a large piece of wood, resulting in a thick outer char
layer, is made even slower by the reduced heat flux to the interior. The hot
char layer may also change the effluent gas composition, as noted above, by
reacting with the water vapor to form CO and H2 (water gas reaction), and also
by cracking some of the heavier hydrocarbon gases and liquids from the
pyrolysis zone.
BURNING PROGRESSION
Table 1 shows .the progression of the several phenomena that
typically take place during the progressive burning of a piece of wood. The
sequence of events vary in relative timing at different locations in the
wood depending on many factors. The whole process is accelerated by high in-
cident heat flux, small piece size, and low moisture content. Wood size and
moisture content particularly affect interior change rates. The surface
phenomena are predominantly affected by radiant heat flux, oxygen supply con-
ditions, and gas evolution rates (moisture and pyrolysis products).
WOOD CONSTITUENTS RESULTING IN ATMOSPHERIC POLLUTANTS
There are no constituents in wood in large quantities that are
recognized as being inherently objectional sources of pollution. The S and
N contents are generally less than 0.01 and 0.1 percent, respectively. With
complete conversion to S00 and NO, the emissions would be below 0.025 and
g ^
0.27 lb/10 Btu (12 ppm SO and 276 ppm NO), respectively. Toxic metals such
as lead and mercury 'do not occur in large enough quantities to constitute a
hazard. Wood ash is normally considered a beneficial soil additive, but must
be disposed of with care due to its alkalinity.
12
-------
TABLE 1. TYPICAL PROGRESSION OF LOCAL PHENOMENA
IN WOOD LOG BURNING
LOCATION OF PHENOMENON
DISTANCE FROM SURFACE
NEAR TO SURFACE
Log Introduction
Time Progression
Radiation influx
Convection influx
Radiation influx Radiation influx
Radiation influx
Convection influx Convection influx is Convection influx
reduced by steam reduced by efflux of Unattached luminous
efflux pyrolysis products flame
and steam
SURFACE PHENOMENON
SURFACE MATERIAL
Heat influx
Sensible heating
Heat influx
Steam efflux
Moisture
evaporation
Heat influx
Steam and pyrolysis
products efflux
Increased heat influx
Steam, pyrolysis gas,
and liquids efflux
Low temp, pyrolysis Exothermic pyrolysis
SUBSURFACE MATERIAL
INTERIOR MATERIAL
No jchange
No change
Sensible heating Moisture evaporation Low temp, pyrolysis
No change Sensible heating Moisture evaporation
CORE, CENTRAL MATERIAL
No change
No change
No change
Sensible heating
Time Progression
Burnout
Radiation flux near balance High radiation outflow High radiation outflow High radiation outflow Radiation cea
Intense luminous flame
Decreasing luminous
flame
Reduced flames less
luminous
Diffusion control of
surface burning
Radiation much increased
Large gas efflux
Char formation
Exothermic pyrolysis
Low temp, pyrolysis
Evaporation
Highly radiant surface
Gas efflux reduced
Char burning
Attached blue flame
Highly radiant surface
Pyrolysis products crack- Surface lost
ing and steam dissocia-
tion within char layer
Pyrolysis products
Exothermic pyrolysis
Low temp, pyrolysis
Char burning
Pyrolysis products
cracking
Exothermic pyrolysis
Char burning
No flame
Highly radiant surface
Surface lost
Surface lose
Char burning
Char burning
Surface
Surface lost
Surface lost
Char burnout
13
-------
The organic constituents of wood, cellulose and lignin, are
themselves considered non toxic, but their thermal decomposition or pyrolysis
products contain several objectional species including aldehydes, phenols,
and creosols. In some cases polycyclic aromatic hydrocarbons (PAH) can be
formed. All of these are of concern in the atmospheric emissions.
Two mechanisms for formation of objectionable emissions from wood
burning are
• Pyrolysis product release from the wood which never
encounters conditions of temperature and oxygen
concentration between the pyrolysis site and the
stack which satisfy the requirements of complete
combustion.
• Incomplete burning due to excess fuel and insufficient
oxygen in active areas of burning.
It is to be noted that both mechanisms can occur at high temperatures, and
both mechanisms can occur while appreciable oxygen remains in the flue gas.
When the temperatures are high enough to initiate combustion, and sufficient
oxygen is present, the combustion can go to completion, consuming all of the
organics. The contribution of secondary combustion to eliminate these
objectional species is discussed later.
Organic emissions resulting from incomplete burning of the wood and
its pyrolysis products thus consist of a mixture of chemical species. Some
may leave the stove in condensed form, and additional species becomes con-
densed into and onto particulates or onto the walls of the stack as the
gases cool. The particulate mass in the stack at any point of sampling
depends not only on the quantity and distribution of organic species carried
in the flue gas, but also significantly on the stream temperature at the
point of particulate separation from the gas. The logical mechanisms of ash
carry over into the stack, suspension burning of fine solid wood particles
and aerodynamic entrainment from the bed, are both expected to be ineffective.
Thus, the condensed organic species are presumed to constitute the bulk of
the stack particulates.
14
-------
PRODUCTION OF PAH IN WOOD COMBUSTION
Polycyclic aromatic hydrocarbons (PAH) have been identified in
wood combustion products. Mechanisms proposed for PAH production from pyro-
synthesis and combustion include: (1) polymerization of small organic
fragments under reducing conditions: at temperatures above about 400°C
(750°F) organic compounds are "cracked" to small reactive fragments that
condense into aromatic species with a subsequent gain in free energy over
condensation into their paraffinic analogs, (3-5) and also, (2) a pyrocombustion
synthesis of PAH can occur from condensation and rearrangement of poly-
acetylenes (6). Figure 1 shows a primary pyre-combustion mechanism proposed
for PAH production from "cracked" organic molecules.
Several researchers have also investigated the pyro-combustion
reactions of cellulose and lignin. These reactions in tobacco leaf materials
have been extensively studied and give a valuable literature base to show how
a species similar to wood fuel behaves at flame temperatures. In the tobacco
case, large weight losses are observed at temperatures nearing the ignition
temperatures (400-450°C) (750-840°F)(7^8). Tobacco research has shown that
yields of condensible organic material regularly decrease from 400-1000°C
(750-1800°F) in pyrolyzed tobacco while recovered PAH increase as the temper-
ature is increased in this temperature range (9). Pyrolysis of cellulose at
800°C (1620°F) (approximately the maximum cigarette burn temperature) produces
1000 times more benzo(a)pyrene (BaP) than the amount measured in main stream
cigarette smoke (10)- The production of benzo(a)pyrene and other PAH from the
cellulose in tobacco has led to several studies (10,11,12).
On warming wood there is a loss of water and volatile species
at 60-100°C (140-212°F) and loss of water of hydration at 160-170°C
(320-340*F) (13). During the combustion of wood, cell disruption occurs with
volatilization, sublimation, pyrolysis, fragmentation and chemical con-
densation occurring in and around the flame zone. That significant amounts
of neutral polycyclic compounds are measured in wood flame effluents is
indicative of chemical reactions which have generated the PAH, principally
from starting materials of cellulose and lignin which do not originally
contain the PAH species measured. Lignin pyrolyzed at 700°C (1290°F) in
nitrogen, yields higher levels of phenols and cresols than cellulose under
15
-------
FIGURE 1. PYRO-COMBUSTION TO PRODUCE BENZO(A)PYRENE (2)
-------
the same conditions (14) . The maximum phenol formation from lignin is found
at 500-600°C (930-1110°F) (15). Studies on the pyrolysis of cellulose have
shown that temperatures of 475-420°C (890-790°F) yield low-molecular weight
aldehydes, ketones, and aliphatic acids (16,17).
It is important to note that both pyrolysis and combustion of
cellulose and lignin lead to the production of PAH, aldehydes, ketones, and
cresols. These processes occur at temperatures as low as 400°C (750°F) (3).
Since these compounds are not original constituents of wood, their manufacture
during the combustion process must be considered under the conditions of
available oxygen, temperature, and wood types used in home wood combustion.
Further study of this phenomenon might establish conditions of pyrolysis
and/or burning at which the PAH species are reduced or not formed.
17
-------
SECTION 5
WOOD STOVE DESIGN AND PERFORMANCE
There are many designs of wood stoves for residential heating.
The radiant heater designed for use in the living space and hand fired with
cordwood is the most popular and is the focus of this study. One design of
residential wood burning furnace has been developed with improved performance,
especially relative to efficiency and emissions. This high turbulence burner,
described later, has been included in the study because many combustion
modifications have been effectively included in its design.
TYPES OF RADIANT STOVES
There are several options of air flow and fuel flow patterns within
the stove that have been adopted in commercialized designs. Figure 2 shows
the four generic systems in simplified form. These are described below
according to the draft (flow) of primary air. The significance to air
pollution emissions is also noted qualitatively for each generic system.
Up Draft
In this conventional system, the primary air flow passes upward
through the burning wood. In some stoves the wood, sticks or logs, rest on
andirons or on a horizontal air-permeable grate. More often the fuel support
(hearth) is inherently not permeable or else the grate openings are effec-
tively plugged with ash. The primary air flow is then introduced near the
base of the combustion chamber, i.e., at the wood support level. Buoyance
effects resulting from the heating of gases in the burning region induce
the flow into the bottom of the wood charge where the primary combustion
takes place. The heated gases then move upward through the wood charge.
Such stoves can be referred to as true up-draft or diagonal draft, and are
characterized by the combustion products moving upwards through the burning
18
-------
J
S'— *-
•>
p-*^
p-^>
J
p->-
"^
r
SC
I/}
/tf
Y/ .
B B B
1
k
Underfire Air
Or Up Draft
£- ^
/P \
Q I
l\ ^
(JB B B ]
\
)
)
~r
i
!
SC
/
^. P^
f >
P — Primary Air Supply
S — Secondary Air Supply
E - Exhaust to Stack
B — Primary Burning
>C — Secondary Combustioi
I - }
— *-E S— *-
s- ^
<. J
-*— S P-*-
r -v
\
^ /7) 7
0 >
w -i su
flB B B J
1
Down Draft
i
I
SC
B ft • ,
tiJlLL 1
^/f ^
Cross Draft
S-Flow
FIGURE 2. GENERIC DESIGNS OF WOOD STOVES BASED ON FLOW PATHS.
-------
pieces of wood, and exiting from the upper parts of the combustion chamber.
The products of the primary combustion thus pass through or among the re-
maining or more recently fed wood. This results in heating and pyrolysis
of some of the wood in an oxygen deficient environment. The rate of wood
consumption is effectively controlled by the flow rate of this primary air.
A secondary air supply for secondary burning is often provided
with the intention that it burn the pyrolysis products either in the portion
of the combustion chamber above the charge of wood, or in a separate,dis-
tinct secondary combustion chamber. The primary combustion products may be
cooled or quenched by the newly fired wood or by heat transferred to the side
walls. Even with active flames low in the burning wood, combustion may not
be attained in the secondary combustion space.
Simple box stoves are often: of this design, especially those which
have the primary air inlet at one side (front) of the bottom of an open com-
bustion chamber, and the vent to the stack at the top (rear) of the same open
chamber. Although the natural convection within the burning wood charge
induces the primary air flow upward through the wood charge, some air can
usually bypass this route and support secondary combustion. Simple box
stoves and air-tight stoves often have an additional air inlet higher in the
stove wall or door to assure this supply of secondary air. As described
later this supply of secondary air does not assure secondary combustion,
and may actually quench combustion.
Down Draft
This design approach has an historical basis and is used in some
stoves today in modified forms. The wood is supported on a grate, with the
primary air flow downward through the wood and combustion taking place low
in the charge, i.e., at the grate. The principal characteristic and ad-
vantage is that the gas flow through the wood is in the opposite direction
*
to the progression of burning through the charge of wood. Thus, the
pyrolysis off-gassing is into the primary air stream as it approaches the
hotter, active burning region. This design assures that these pyrolysis
products pass through the burning region, and thus promotes burning of the
organic compounds., The admission of secondary air immediately down stream
of the burning wood thus can aid in the complete burning of pyrolysis gases,
* Some stoves not meeting this description are described and/or advertised
as being "downdraft".
20
-------
as high temperatures can be maintained in this secondary combustion area.
The Vermont Downdrafter stove is a variant of this design, in
which the combustion zone is kept small in a fixed location, promoting
high temperature burning of all the pyrolysis products. A true downflow
combustor is the "Tasso", a residential handfired boiler currently made
in Denmark (.18) .
Side Draft
In the side draft, or cross flow design, the primary air is
supplied to one side of the base of the burning wood charge and leaves the
primary burning area also at the base of the wood charge. Thus, the
evaporated moisture and initial pyrolysis products from the wood prior to
its active burning are not released directly into the primary air flow.
They are released into a more stagnant region which also serves as a fuel
magazine. The convective and radiant heating of the freshly fired wood in
this magazine is less than it would be in the main gas flow from the active
burning region. By directing the off-gas products only through a hot active
burning region, they can be burned when adequate air is supplied and thus
are prevented from 'escaping' directly into the stack. This can be either
in the primary combustion region or a hot secondary combustion region
supplied with additional air.
S-Flow
A large number of small stoves utilize an air flow pattern which
involves primary air supply along the surface supporting the wood charge
from one end of the combustion chamber, with combustion products leaving the
top of the primary combustion chamber at the same side. This design permits
slow end-burning of the wood sticks, starting near the air source and pro-
gressing slowly away from the source. This flow pattern within the primary
chamber is often produced by placing a baffle above the wood charge, directing
the products of primary combustion through a restricted area near the air
source before entering the stack. Secondary air can be supplied here to
support secondary combustion. The location, shape, and size of the baffle
21
-------
varies among different designs, providing higher velocities and turbulence,
and a longer gas flow path but not a longer average resistence time within
a given stove envelope.
Many designs of Scandinavian stoves are of this type. These are
often called cigarette burning designs, and are conducive to very slow burning.
SECONDARY COMBUSTION
The wood pyrolysis products are released into the primary air stream
passing through the wood charge. The pyrolysis rate often exceeds the equiv-
alent air supply as determined by the stoichiometric fuel: air ratio. This
results in fuel gases being carried away from the solid fuel in an unburned
or partially burned form containing appreciable CO and many organic gases.
Most stove designs provide for a secondary air supply to be mixed with the
primary stream after the primary stream leaves the wood. Ideally, in this
manner the secondary air stream should support complete combustion of all the
pyrolyzed fuel, without appreciably increasing the rate of wood consumption.
Figure 3 shows the typical locations for secondary air supply, and
the regions where the secondary combustion is intended to occur in each
generic design.
This ideal situation is hindered by several phenomena which result
in unburned organics and CO leaving the stove unburned:
1. The mixing rate and turbulence intensity are limited
in natural draft stoves impeding complete secondary
burning.
2. The temperatures in the secondary combustion area may be
below that necessary for ignition and burning, especially
at low rates of burning and in locations in the stove
where relatively cold walls quench the flames from the wood.
3. The rate of secondary air flow that would be required to
burn the excess fuel gases in the primary flow varies
during a burn. Early in the burn period when large
amounts of pyrolysis gases are being evolved, insufficient
secondary air leaves unburned fuel in the flue gas. Later
when the primary burning is largely char burning, excess
secondary air might quench the flue gas, resulting in temper-
atures below those required for ignition.
22
-------
Considering that essentially all stack particles are organic
species and therefore combustible, the occurrence of afterburning or
secondary combustion must be recognized as a means of consuming these
particles in addition to the gaseous organic species still in the gas phase
and the carbon monoxide. This phenomena, when achieved in the stove, not
only would reduce emissions, but correspondingly increase the amount of heat
released and recoverable from the fuel. Considering these advantages, most
stove designs provide for the admission of air to support secondary combustion,
and provide a volume or chamber within the stove in which this combustion
can occur.
Figures 3, 4, and 5 show the combinations of temperature, CO,
organics (calculated as being propane), air, and inerts that satisfy the re-
quirements for ignition and burning. The curves on the tri-axial coordinates
bound the region where combustion can occur. It is evident that increased
temperatures widen the bounds making combustion more likely, and that increased
inerts narrow the bounds making combustion less likely. These calculations
discount the effects of moisture on the burning and assume that the propane
ignition characteristics are representative of the mixture of organic species
pyrolyzed from the wood and released from incomplete burning of wood.
Neither assumption is entirely valid, but the results are indicative of the
combustion limitations involved.
COMMERCIAL HIGH TURBULENCE BURNER
A residential wood burning appliance has been developed under a DOE
contract by Professor Hill at University of Maine. The burner demonstrates
that increased efficiencies can be obtained with an improved design of the
combustion and heat recovery system. The burner is a residential boiler
(water heater) for a central heating system that incorporates several factors
beneficial in controlling air pollution emissions. The combustion is sup-
ported by both forced draft of combustion air and induced draft of flue
gases, such that a high turbulence level can be maintained in the burning areas,
yet not develop a firebox pressure exceeding ambient. It operates only at a
fixed, relatively high burning rate ("o 150,000 Btu/hr) comparable to one of
the large radiant heaters. Several characteristics of this design favor low
organic and CO emissions:
23
-------
C3H8
Inert Ratio =
Inert Fraction
Active Fraction
N2 in the inert fraction is the nitrogen
in excen of that corresponding to the
oxygen in the air in the active fraction
The air. CO, and C3Hg are plotted as
volumetric percents of their sum in the
total gas stream. This sum ii considered
to be the active fraction. The inert frac-
tion contains CC>2 and the N2 in excess
of that associated with the air in the
active fraction at room temperature.
ro
-P-
10
C3H8
80
90
100%
FIGURE 3. LIMITS OF IGNITION AT 20 C.
-------
Inert Ratio
Inert Fraction
Active Fraction
N2 in the inert fraction is the nitrogen
in excess of that corresponding to the
oxygen in the air in the active fraction
The air, CO, and C3H8 are plotted as
volumetric percents of their sum in the
total gas stream. This sum is considered
to be the active fraction. The inert frac-
tion contains CC>2 and the IM2 in excess
of that associated with the air in the
active fraction at room temperature.
CO
85
80
70
60
50
40
30
20
10
Ul
10
C3HB
80
90
100%
LIMITS OF IGNITION AT 27O C.
-------
C3H8
Inert Ratio =
Inert Fraction
Active Fraction CO + Air + C3H8
N2 in the inert fraction is the nitrogen
in excesi of that corresponding to the
oxygen in the air in the active fraction
The air, CO, and C3Hg are plotted as
volumetric percents of their sum in the
total gas stream. This sum is considered
to be the active fraction. The inert frac-
tion contains CC>2 and the N2 in excess
of that associated with the air in the
active fraction at room temperature.
N>
10
C3H8
10
20
30
40
50
Air
60
70
80
90
100%
FIGURE 5. LIMITS OF IGNITION AT 520*C.
-------
• High turbulence level in primary burning area
• Primary combustion products isolated from inventory
of wood
• Inventory of wood cooled by water-jacketed chamber
• Burning confined to small portion of the wood inventory
• Refractory lined primary combustion area
• Refractory lined secondary combustion area
• Large pieces of wood utilized to reduce excessive
pyrolysis early during burning cycle
• Burning restricted to high rates, with heat storage
provisions to accommodate lower heating requirements.
Figure 6 shows a schematic sketch of the furnace, showing the
principle components. A commercial production version (Dumont) of this wood
burner was obtained for operation and analysis in the laboratory.
27
-------
" •**_**. **
iiL
8
7 / / 1
/ /
/ / /i
1 - forced draft fan
2 - secondary air
3 - refractory base
4 - stacked wood
5 - refractory tunnel
6 - low velocity ash chamber
7 - fire tube heat exchanger
8 - induced draft fan
9 - water jacket for wood charge
FIGURE 6. WOOD FIRED RESIDENTIAL BOILER
UTILIZING HIGH TURBULENCE BURNING.
28
-------
SECTION 6
EXPERIMENTAL PROGRAM
LABORATORY FACILITY
The test facility assembled at Battelle for this program consists
of three test sites, each with an independent, prefabricated, insulated
chimney exhausting within a high bay area in the building. Thus, the
chimney effects on draft are not affected either by outside wind or by
the building's transient pressure fluctuations. Radiant stoves are mounted
on electronic scales, such that total weight losses associated with fuel
burning are continuously monitored. Figure 7 shows the facility arrangement
and dimensions of a one stove and stack system.
The gases being continuously monitored are CO, C02, NO, 02, S02,
and total hydrocarbons (THC). The Q£ analyzer is a Taylor Servomex Type
OA272 paramagnetic analyzer, the THC analyzer is a Beckman Model 402 high-
temperature flame ionization detector analyzer. The C02, CO, and NO analyzers
are Beckman nondispersive infrared analyzers. The S02 is measured by an
electrochemical analyzer. 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. All instrumentation is kept in a positive air
pressure room where the air is supplied through an activated charcoal filter.
The room is kept closed and is cooled by two air conditioning units.
The permanent gas sampling system pulls the flue gases from the
stack at the point where the hot combustion products leaving the stove enter
the stack. There is a 5/8" tube probe that extends the entire diameter of
the stack. This probe is filled with pyrex wool and has two rows of 1/8"
29
-------
8%'
6'
Frictionless
Rollers
f
PAH Sampling
Sampling Line
(Heated Stainless)
To Instrument
Room
r
\
8"ID
STOVE
Electronic Scale
Platform
Fire Control
Water Line
To Exhaust
Blower
Water
Trap
FIGURE 7. LABORATORY INSTALLATION OF STOVE WITH STACK AND VENT SYSTEM.
30
-------
holes the length of the probe. After the gases are filtered in the probe
they are pumped through a short heated flexible teflon line to a longer
heated stainless steel.line to the control room. The gases then pass
through a Teflon bellows pump, after which the gas stream is split. Part
of the sample stream is sent directly to the THC analyzer, the rest is
passed through a cold trap and cold filter to remove excessive water vapor
and heavy organic vapors before entering the other analyzers. Figure 8
shows a schematic diagram of this system.
OPERATING PROCEDURES
In preparation for a run, the gas sampling lines are flushed with
acetone to remove oils and tars from the sample lines. The pump is also
cleaned. The gas sampling probe is removed from the stack and the glass wool
replaced. The cold traps are cleaned and then placed into operation. The
instruments are purged with nitrogen and spanned with appropriate span gases.
Any adjustments to the instruments are done at this time. Once the sampling
system and the instrumentation is working correctly, a charge of kindling is
placed in the stove and ignited. The kindling burn heats the stove and
generates a bed of hot coals. When 80 percent of the scrap wood is consumed
the pretest burn is started by charging the desired quantity of wood. The
pretest burn is used to assure the stove is operating at the desired con-
ditions. As soon as 90 percent of the pretest wood is burned the actual test
is started while not changing any conditions of the stove. In unusual cases not
requiring hot starts the pretest at the identical operating conditions
is omitted. A test run consists of the burning of a given charge of wood
in a batch process immediately following the pretest burn. The test run
lasts until 95 percent of the added wood is lost as measured by the scale.
At the end of the day, the sampling system is turned off and
nitrogen is passed through all the instruments to recheck their zero and
remove combustion gases from the instruments.
EMISSION MEASUREMENTS
During all tests the millivolt signals from the instruments are
recorded on strip chart recorders and once every minute a Doric Digitrend
31
-------
STOVE
Sample Line
(Heated Stainless)
THC
Beckman 402
FID
Heated Stainless
Dry-Ice
Trap
NO
Beckman 315
(NDIR)
02
Taylor Servomex
(Paramagnetic)
Beckman 315 B
(NDIR)
CO2
Beckman 215 B
(NDIR)
S02
Envirometics
(Electrochemical)
FIGURE 8. CONTINUOUS GAS SAMPLING SYSTEM.
32
-------
220 Datalogger scans the data from the gas analyzers and the thermocouples
in the stove. The Datalogger prints this information and also records it
on a cassette tape by a data link with a Tecktran 815 data cassette recorder.
Notes are taken manually during the runs such as the amount of wood, type
of wood, stove operating conditions, the time of start and end of test,
and other observations. Also a log sheet of data is taken at intervals
during operation in case the Datalogger goes down, and it also shows trends
during the test.
PAH MEASUREMENTS
During some test runs, the flue gas is sampled to determine the
concentration of PAH. Figure 9. shows the sampling equipment used: a modified
EPA Method 5 apparatus. This allows the collection of particulate emissions
using filtration, followed By collection of the gas phase material by adsorption
on the XAD-2 resin. The sampling rate is maintained constant at about 20 1/min
(^0.75 cfm) during the entire sampling period with the probe inlet at the stack
center. The sampling probe Inlet velocities have been appreciably higher than
stack gas velocities. However, the effects of non-isokinetic sampling are
considered insignificant because of the small particle sizes. Using an
electric aerosol analyzer, the mass median diameters of suspended particles
were measured as 0.22 and 0.39 at the PAH sampling location during two runs
in this facility. Fuchs (19) has shown that for particles smaller than 5 mm,
excessive sample withdrawal velocities result in collection efficiencies still
above 95 percent Because the smaller particles follow the gas stream lines.
Following sample collection, the organic materials are extracted
from the column and probe system and analyzed for PAH compounds. The probe
wash and extract from the filter are combined into one sample for analysis,
and the column extract constitutes a second sample. Although the sample
collected on the probe walls and the filter can be defined as particulate,
it must be recognized that they consist essentially of condensed organic
species. The distribution between gaseous organics as retained by the XAD-2,
and particulate organics caught on the filter, must be recognized as depending
primarily on the temperature of the stream at the point of collection on the
filter and walls. Appendix B presents details of the PAH analysis procedures.
33
-------
tf
Heated Probe
U)
Thermostatically
Controlled
Oven
Impinger Train
Or Backhalf Check Valve
"If
• iue Odin I
Bypass Valve Vacl|um
Vacuum Line
FIGURE 9, PARTICULATE AND PAH SAMPLING SYSTEM.
-------
DATA PROCESSING
After each day of testing, taped raw data from on-line gas
composition instruments and the scale weight are calculated and ultimately
read out as volume fraction of the flue gas at each minute of burn, and
weight of wood remaining unburned at each minute. These one-minute recordings
are subsequently machine averaged for the entire test run, and this average
reported as the time-averaged gas composition for that run. The emissions
have thus not been normalized to an arbitrary flue gas 0 and moisture
content, such as 3 percent dry. Such a normalization would in all cases
increase the numerical pollution concentration. These time averages are
then manually converted to emission per weight of wood burned using a
calculated weight flow of flue gas for the entire run and the weight of
wood burned as measured by the platform scale. The volumetric composition
of the flue gas is also machined plotted against time, such that trends,
anomalies, and relative values can be easily identified. The calculated
emissions provide the basis for evaluating the effects of imposed variables
on total emissions for a test, and the flue gas composition plots provide a
basis for diagnosis of the phenomena occurring and demonstrate the real-time
effects of changes imposed on a stove during operation.
FUEL WOOD USED
Four types of wood fuel have been used, as follows:
• White oak cordwood, air dried, 20.7 percent moisture.
Normally split and round pieces were fired together.
• Pine cordwood, not air dried (i.e., green),
42 percent moisture. Normally split and round
pieces were fired together.
• Douglas fir brands 3/4 inch on 1 inch centers,
oven dried, null moisture.
• Oak lumber, nominal 4 inch x 4 inch, cut diagonally
into triangular pieces, thoroughly air dried,
12.4 percent moisture. These 4 x 4s were fired only
in this form, with all surfaces sawed.
The moisture contents of the oak and pine were determined by weight loss upon
storage in a 100+°C oven for 48 hours, and is stated on the wet basis.
35
-------
SECTION 7
EXPERIMENTAL PROGRAM OBSERVATIONS
Table 2 lists the averaged flue gas compositions and emissions
obtained during this program. The burning rate, fuel, and burning mode
have been identified for each test run. The gas compositions which are
mechanically recorded each minute are machine plotted, and several repre-
sentative plots are reproduced in Appendix B. Notes have been added to
identify the burning phenomena occurring at each time. All tests of the
radiant stoves were started when a charge of wood was inserted into hot
stove with a bed of burning embers. The high turbulence burner was started
cold in each test. These averaged data are presented only for tests during
*
which no changes were imposed on the stove operation or controls.
The combustion in wood stoves, and the resulting emissions, can be
modified in many ways. In this experimental program several modifications
were made in stove operation in attempts to correlate specific changes in
emission factors with a specific combustion modification. Synergistic effects
of several simultaneous modifications were not specifically sought. However,
each set of tests conducted to demonstrate the effects of a specific modifica-
tion necessarily depends on the stove design and operating parameters in use
at the time. The effects of a specific modification under different design
and operating conditions can only be assumed, and the synergistic effects
determined by inference.
The modifications conducted in this laboratory program were of three
types; stove design, fuel properties, and operator techniques. The uses of
non-wood fuel and combustion additives were not considered. The use of processed
wood, such as pellets, was also not considered, as this fuel can most effectively
be burned in equipment specifically designed for its use.
A group of 32 test runs was also conducted for TVA in a similar manner and the
detailed results of these tests will be separately reported by that agency,
including overall thermal efficiency calculations.
36
-------
TABLE 2. EMISSIONS AND OPERATING CONDITIONS
Test
Run
Number
77
78
91a
92
93
94
95
Wood
Specie
Oak
Oak
Oak
Pine
Pine
Oak
Oak
Piece
Shape
Up
Split
Split
4x4
Split
Split
4x4
4x4
Burning
Rate
Ib/hr
Draft Burning
14.4
31.8
16.5
20.7
21.6
20.9
27.1
Fuel
Time Averaged
Gas Composition, as measured
Burned Percent by Volume
Ib/test CO C02 02
Mode with Grate
25.3
33.9
21.0
18.2
30.3
24.0
25.8
Up Draft Burning Mode
18
19
20
Oak
Oak
Oak
4x4
4x4
4x4
5.8
14.4
19.5
6.3
10.2
14.4
2.5
2.3
3.6
3.9
2.0
2.7
5.4
(Modified
11.0
10.9
13.8
15.3
13.3
10.6
13.0
Parts Per Million
THC
NO
S02
Riteway)
9.
9.
6.
5.
8.
9.
6.
1
3
8
6
2
2
1
6700
6400
7100
9100
7100
5900
9900
110
80
120
180
100
90
130
60
50
50
70
50
40
50
with Hearth (Defiant)
0.1
0.2
0.3
Side Draft Burning Mode
37
36
34
35
38
82
83
84
85
86
87
88
89
90
Oak
Oak
Oak
Oak
Oak
Oak
Oak
Pine
Oak
Oak
Oak
Oak
Oak
Oak
4x4
4x4
4x4
4x4
4x4
Split
4x4
Split
Solit
Split
Split
Split
Split
Split
6.3
11.8
17.4
19.2
12.3
16.7
18.7
18.4
16.5
13.0
16.6
15.6
14.6
15.0
5.9
9.8
12.9
41.9
26.4
24.0
23.4
26.1
15.2
15.0
14.9
14.6
14.6
15.5
High Turbulence
51
52
53
91
42
43
60
61
63
64
80
81
65
70
71
76
46
47
48
49
50
68
69
72
73
66
67
Oak
Oak
Fir
Oak
Oak
Oak
Oak
Oak
Oak
Oak
Oak
Oak
Fir
Fir
Oak
C
Fir
Oak
Oak
Fir
Pine
Oak
Oak
Oak
Fir
Down
Oak
Oak
Split
Split
Split
Split
Split
Split
4x4
Split
Split
Split
4x4
Split
Split
Split
Split
Split
Split
Split
Split
Split
Split
Down
Split
Split
Split
Split
21.9
22.1
20.9
24.1
Down Draft
7.3-
10.3
18.4
19.1
12.8
11.3
13.7
20.4
10.7
18.0
9.8
8.0
14.0
13.2
15.6
14.5
10.1
Draft Burning
15.1
16.0
21.6
24.6
37.3
35.9
21.6
39.3
Burning
13.6
16.2
15.4
16.2
15.8
14.9
16.2
25.5
15.0
16.8
14.7
8.0
14.0
23.8
23.5
12.0
12.4
Mode -
13.4
12.0
15.8
9.0
1.8
0.6
1.2
2.5
1.8
0.5
2.5
0.5
0.5
1.5
1.1
0.9
0.4
0.8
Burner
0.3
0.6
1.6
0.1
1.9
5.9
8.5
19.
15.
12.
6
4
2
1800
1600
1100
40
80
120
20
20
10
(Defiant)
6.2
10.1
11.9
11.7
9.6
10.1
13.0
8.6
13.0
12.5
12.9
13.9
13.8
12.5
(Dumont)
9.6
6.2
8.3
2.1
15.
10.
8.
7.
11.
10.
6.
11.
9.
9.
9.
8.
8.
9.
10.
14.
11.
18.
1
7
1
8
1
6
6
9
4
7
4
5
8
9
9
2
2
9
8300
1900
2800
5300
7600
1800
6200
1800
1200
3400
2400
1100
400
800
300
2000
1900
200
100
50
80
90
120
90
80
70
110
90
80
50
50
60
140
70
50
20
40
20
20
40
60
10
30
10
10
20
20
10
10
10
20
40
10
Mode (Modified Defiant)
0.2
0.2
0.5
1.0
0.1
0.3
0.4
0.1
0.4
0.4
0.1
+
0.3
0.3
0.3
0.4
0.1
3.4
3.4
9.4
8.7
5.4
4.9
7.4
8.5
6.3
6.9
2.5
7.5
5.0
5.9
6.0
8.1
2.3
Heated Air Supply
0.2
0.2
0.1
0.7
8.3
9.9
7.0
12.2
Draft Burning Mode, Reduced Area and Insulated
Split
Split
12.9
13.0
13.8
14.6
0.1
0.0
4.5
5.3
17.
17.
11.
11.
15.
15.
13.
12.
14.
13.
18.
13.
15.
15.
14.
12.
18.
5
6
3
6
5
8
2
3
3
9
7
4
8
1
9
6
9
(Modified
12.
10.
13.
8.
Bed
14.
15.
3
6
9
0
3100
3100
1000
1600
600
900
600
200
900
2000
1500
100
1100
3200
1000
600
1200
70
80
80
210
110
50
40
90
30
40
20
60
20
80
220
130
100
60
60
20
20
10
10
10
<10
10
20
10
<10
<10
10
10
10
20
Defiant)
300
200
1300
1600
(Modified
4
4
900
800
260
220
70
80
Defiant)
70
200
<10
<10
10
20
<10
<10
37
-------
TABLE 2. (Continued)
Test
Run
Number
77
78
91a
92
93
94
95
18
19
20
37
36
34
35
38
82
83
84
85
86
87
88
80
90
51
52
53
91
42
43
60
61
63
64
80
81
65
70
71
76
46
47
48
49
50
68
69
72
73
Excess
Air,
percent
80
80
40
30
70
80
30
2050
290
140
340
110
60
60
120
110
40
130
90
90
80
70
80
100
110
220
120
920
560
560
120
130
290
330
180
140
220
200
880
180
320
280
250
150
1000
140
100
200
61
Stack Emissions
Temp. lb/1000 Ib Wood
°F THC CO NO Comments
Up Draft Burning Mode with Grate
710 34 220 1.1 Small pieces
720 33 200 0.8 Small pieces
950 32 280 1.0 PAH sample
970 44 330 1.6 Oven dried wood
760 25 120 0.7 Green Pine
990 34 270 1.0
960 42 400 1.0
Up Draft Burning Mode with Hearth
560 113 90 4.9 Low rate
760 19 33 2.0 Medium rate
1080 8 37 1.7 High rate
Side Draft Burning Mode
270 112 420 2.6 Low rate
580 12 63 0.6 Medium rate
680 15 100 0.8 High rate
660 26 220 0.8 Large load
450 53 210 1.6 Thermostat control
620 10 50 0.9 All split wood
700 28 190 0.6 PAH sample
590 9 40 0.6 PAH sample, all split
640 6 44 1.1
630 18 140 0.9
620 13 100 0.2
700 5 75 0.1
670 2 34 0.1
620 4 74 0.1
High Turbulence Burner
330 2 30 1.6 Many small pieces
250 18 97 1.2 4 large pieces, 1 unburned
290 16 222 0.8 Brands only
340 5 25 1.2 Large pieces, normal run
Down Draft Burning Mode
340 56 67 2.5 Low burn rate
370 56 64 2.7 Medium burn rate
800 7 61 1.0 Sawed 4x4
770 10 110 2.5 Small split pieces
610 7 19 2.2 4" round wood
570 11 48 1.1 Medium splits
710 5 59 0.6 PAH sample
740 2 6 1.1 Frequent additions
640 9 74 0.6 14% moisture
660 16 49 0.6 28% moisture
410 34 39 0.9 31% moisture
720 1 3 2.0 Charcoal fuel
620 16 67 0.6 Secondary air valred
510 33 54 1.6 Normal bed area
500 10 44 3.8 Reduced bed area
600 5 60 2.3 Reduced bed area
330 26 28 4.1 Reduced bed area
Down Draft Burning Mode - Heated Air
630 2 20 3.3 370 F primary air
680 1 15 2.3 370 F primary air
600 11 19 1.1 440 F secondary air
980 10 71 0.8 440 F secondary air
Down Draft Burning Mode, Reduced Area and Insulated Bed
66
67
380
280
540 11 25 1.6 Bed channeling
600 8 4 3.9 Ho bed channeling
38
-------
The following paragraphs describe various modifications investi-
gated with the observed effects on emissions. For some modifications the
emission factors resulting from complete tests are presented, and in some
modifications, visual and instantaneous results are more appropriate.
This section is a discussion of stove combustion modifications.
In each case the objective or basis of the modification is presented,
followed by experimental technique, observed results, and appraisal of
the results.
The very significant effects of burning rate on flue gas composition
and emissions must be considered in any comparison between stoves or between
the same stove in different tests. These effects will be discussed later in
this section. Most of the data reported here are from tests conducted at
medium and high rates for the size of stoves used. The two basic radiant
stoves modified and used in this study have manufacturers nominal ratings
of approximately 13 pounds wood per hour, assuming 20 percent moisture in
the wood and 60 percent overall thermal efficiency. The test program included
operation at burning rates both greater and smaller than these nominal ratings.
Sustained burning at hold-fire conditions (closed dampers) would naturally be
much below these burning rates.
STOVE DESIGN FACTORS
An operator has little control of the design of the stove except
at the time of stove purchase. At that time the operator decides which
type and design to buy, depending on personal preference and the designs
that are available. Some stoves however, have limited capabilities to
operate in different modes, such as by use of an internal damper.
Air Flow Path in Stove
Three of the four generic designs of air flow path were incor-
porated into one stove (Defiant) for a direct comparison of emissions. The
up-draft mode with the hearth was achieved by opening the damper thus
39
-------
permitting the primary combustion products and pyrolysis products to pass
directly upward and out of the stove. In the side draft mode all primary
combustion products passed under the side arch or baffle, and through the
more extensive internal passages before leaving the stove. In both of these
modes the fire rested on an impervious hearth and primary air was supplied
through tuyeres at the hearth level. The Defiant was also modified to burn
in the down draft mode as shown in Figure 10. In this mode the fire bed was
entirely supported on a grate, with the bed area reduced by about one third
by a retaining firewall composed of bricks. The primary combustion products
passed downward through the grate, under the arch, and then through the
same passages as used in the side flow design. In this mode, the primary
air was supplied through an opening in the end door.
The up-draft mode with the grate was obtained in another stove
(Riteway) modified to provide all air supply upward through the grate and
leaving the top of the combustion chamber.
The results of the tests are shown in Table 3. As can be seen
from the data, there is no large difference between the three modes of
operation of one stove. Both gas composition and emissions are shown for
the three different modes of burning for various burning rates. The down-
draft mode at high burning rates is consistently low in THC emissions.
When the burning rate is low the thermal buoyancy forces result in a local
updraft flow (countercurrent) within parts of the bed, overcoming the
overall average down flow through the bed. This results in channeling and
flow recirculation within the bed, and rapid pyrolysis of wood throughout
the bed. Extensive pyrolysis products can thus be released into the counter
stream and ultimately pass with the channelled down flow out of the bed without
burning. The effects of Bed channeling on the CO and the emissions is also
evident when comparing tests 66 and 67 (>ee Table 2). Using several windows
in the stove permitted observations of gas flow through the bed. In test 66
extensive counterflow and channeling were observed, whereas in test 67
uniform downflow throughout the bed was maintained.
The up draft burning mode using the grate produced higher THC and
CO emissions than the other modes at comparable burning rates, although the
PAH emissions do not support this trend.
40
-------
TABLE 3. EMISSIONS FROM DIFFERENT MODES OF
BURNING FUEL: Oak 4x4
Test
Run
Number
37
36
34
83
84
42
67
80
60
18
19
20
Gas Composition
THC, CO,
ppm percent
Side Draft
8300 1.8
1900 0.6
2800 1.2
6200 2.5
1800 0.5
Down Draft
3100 0.2
800 0.0
600 0.4
1000 0.5
Up Draft Mode
1800 0.1
1600 0.2
1100 0.3
Emissions
lb/1000 Ib wood
PAH THC CO
Mode (Defiant)
ND 112 420
ND 12 63
ND 15 100
0.035 28 190
0.045 9 40
Mode (Defiant)
ND 56 67
ND 84
0.050 5 59
ND 7 61
with Hearth (Defiant)
ND 113 90
ND 19 33
ND 8 37
Excess
Air,
percent
340
110
60
40
130
560
280
180
120
2050
290
140
Burning
Rate,
Ib/hr
6.3
11.8
17.4
18.7
18.4**
7.3*
13.0
13.7
18.4
5.8
14.4
19.5
Up Draft Mode with Grate (Riteway)
77
91a
94
95
78
6700 2.5
7100 3.6
5900 2.7
9900 5.4
6400 2.3
ND 34 220
0.018 32 280
ND 34 170
ND 42 400
ND 33 200
80
40
80
30
80
14.4*
16.5
20.9
27.1
31.8*
High Turbulence (Dumont)
91
51
* = Oak
200 0.1
300 0.3
Logs
0.011 5 25
ND 2 30
920
110
24.1*
21.9*
** = Green Pine
ND = Not
determined
41
-------
Gas Flow Sequence in Down Draft Mode
1.
2.
3.
4.
5.
Primary air enters through
door at left side
Rises to go over top of brick
firewall staying in front of
the closed bypass damper.
Descends on the right of the
firewall, passing first down-
ward through the inventory of
unburned wood.
Continues descent downward
through the burning wood.
Continues downward through
the grate.
6. Passes horizontally under the
baffle into the secondary
combustion chamber.
7. Mixes with secondary air as it
rises through the secondary
combustion chamber.
8. Passes behind the closed bypass
damper to rear center of stove
9. Passes upward through top rear
of stove into fluepipe.
Oval
Flue Pipe
Bypass
Damper
Thermo-
couples
Front
Side
FIGURE 10. CONFIGURATION OF STOVE CONVERTED TO DOWN DRAFT MODE.
42
-------
The Dumont high turbulence burner produced lower emissions than
the up draft or side draft modes when burning oak fuel. The emissions
were comparable to the lowest values obtained with good down draft burning,
as obtained in test 67.
Primary Air Preheat
One purpose of preheating the primary air is to increase the rate
and completeness of the primary combustion of the wood. The combustion
products will hopefully be more completely oxidized because the air entering
the combustion zone is heated, thus raising the temperature of the reaction
and thereby the reaction rate. The combustion products are also less cooled
or quenched by the air supply.
Using a stove operating in the downdraft mode, tests were conducted
in which the incoming primary air was preheated to 188°C (370°F). Air flow
rate into the stove from a pressurized source was controlled so that the
rate would be similar to other tests, i.e., 850 1/min (^30 scfm). The pre-
heating was done by passing the primary air through an electrically heated
pipe and blowing this into the inlet of the stove. This arrangement is
schematically shown in Figure 11. The heated air was the only primary
air source of the stove, and used a large pipe in the stove's door side
to reduce the air velocity and to distribute the air evenly in the stove.
Oak logs, round and split, were used during these tests.
Even with the low emissions inherent with downdraft burning,
reductions in emissions were obtained when the primary air was preheated
with the outside source. The time averaged gas compositions from four entire
tests were converted to emissions and are shown in Table 4. It is not certain
if this effect would be observed to the same degree with other modes of
burning, as the heated incoming primary air could promote early and extensive
pyrolysis of the wood thus adversely affecting emissions. However, in down-
draft burning these pyrolysis products cannot easily bypass the active
burning area at the downstream edge of the bed, especially in the early
stage, when there is a good bed of hot coals that the gases must pass
through.
43
-------
1
Filter
i | i
1
1
1
1
V
••M
T
Rotometer
3/8 Inch Copper Tube
(Insulated)
Electrically Heated Pipe
FIGURE 11. AIR PREHEAT SYSTEM.
-------
TABLE 4. EFFECTS OF PREHEATED
PRIMARY AIR ON EMISSIONS
Down Draft Burning of Oak Logs
Ambient Heated
Primary Air Primary Air
(80°F) (370°F)
Test Run Number
Burning Rate Ib/hr
Gas composition
as measured,
CO, percent
THC, ppm
Emissions,
lb/1000 Ib wood
CO
THC
48
15.6
0.3
1000
44
10
64
11.3
0.3
900
48
11
68
15.1
0.2
300
20
2
69
16.0
0.2
200
15
1
45
-------
The preheating of primary air was thus shown to offer possi-
bilities in reducing emissions in a wood burning stove. A main design
problem to overcome is the location of the heat-transfer surfaces to preheat
the air, while keeping the pressure losses low enough to prevent serious re-
strictions to the flow of air.
Secondary Air Preheat
Preheating of the secondary air is important to prevent cooling
or quenching of the combustion products before leaving the stove so that
secondary combustion can occur. An increase in the secondary air temperature
broadens the flammability limits allowing initiation of secondary combustion.
The heating of the secondary air was accomplished by using the same
preheater used for the primary air. The air supply temperature was 228°C (442°F)
and the flow was controlled to about 2 I/sec (5 scfm). The heated air was
introduced through the stove's own internal secondary air ducting before
entering the secondary combustion chamber. This ducting provided inherent
and additional heating of the secondary flow.
The preheating of the secondary air is assumed to have no effect
on emissions unless secondary combustion has been established or can be
initiated by the preheat. This is evident from a test burning wood which
emitted large quantities of combustibles early in the burning cycle (dry
fir brands). Secondary combustion started only when the heated secondary
air was introduced into the stove. It is uncertain whether the preheat or
just an increase of total air was the cause of secondary combustion. The
secondary combustion was evident from the increased intensity of the flames
in the secondary combustion chamber observed through a pyrex window.
The preheating of secondary air can achieve reductions in emissions
by supporting secondary combustion, however, the achievement of secondary
combustion is so strongly dependent on the gases present that the effect
of preheating by itself is overshadowed.
46
-------
Secondary Combustion
Secondary combustion is the combustion of the combustible materials
not burned in the primary combustion or immediate vicinity of the wood.
These materials can result from quenching of the primary combustion products
or from pyrolysis of wood without burning downstream of the primary com-
bustion region. This secondary combustion is achieved by mixing the gases
from the wood and from the primary combustion with suitable oxygen at a
temperature sufficient to ignite the mixture. In some stoves this is
attempted directly above the burning wood and in other stoves the primary
combustion zone is separated from the secondary combustion zone by a physical
barrier, with the secondary air entering at a different location than the
primary air. For secondary combustion to take place the gas composition must
be within the flammability limits of the gases. Therefore the addition of
secondary air could dilute the combustion gases to a fuel concentration below
the flammability limit, or quench the mixture to below ignition temperatures.
The flammability limits calculated for CO and propane with air
are shown in Figures 3, 4, and 5. Propane was chosen as being representa-
tive of wood pyrolysis gases realizing that the composition of the actual
pyrolysis gases is extremely complex and varies during burning in actual
cases. As can be seen from the triangular graph, the flammability region
is very narrow and would explain the difficulty in obtaining and maintaining
secondary combustion especially at low temperatures. The amount of inerts
from completed combustion also affects the flammability as shown in Figures
3, 4, and 5 by making the range narrower. As the temperature of the combus-
tion products increases, the flammability limits broaden, making secondary
combustion easier to initiate.
The stove used for these observations has a separate secondary
combustion zone. For the observation and determination of secondary com-
bustion a pyrex window and a thermocouple were placed in the secondary
combustion chamber. Different woods were burned so that different burning
rates and emissions could be tested for secondary combustion. The tests
were conducted in the downdraft mode permitting flames to persist the short
distance from primary to secondary zones, but essentially precluding
burning ember carry over. Secondary air was supplied through internal
stove ducting which provided air preheat before mixing with the primary
products.
47
-------
During the test program there were very few occurrences of dis-
tinct secondary combustion when it was not being deliberated promoted. Even
when secondary combustion was attempted, if and when it was obtained, it
occurred only at high burning rates and for only a short period of time (less
than five minutes). During one test the secondary air flow was intermittently
stopped. With the secondary air inlet closed, emissions were high (CO: 2.3
percent, THC: 8200 ppm, and the oxygen: 3.6 percent). The inlet was then
suddenly opened. One-minute data extracts from this test are shown in
Table 5. When the secondary air inlet was opened an audible noise was heard,
and the initiation of secondary combustion was indicated by the thermocouple
located in the secondary combustion chamber. The initial pulse of secondary
air was presumably heated above the equilibrium inlet temperature, as the
supply duct was surrounded by the hot gas. Along with the indicated temperature
increase of about 100°C (170°F), there was a decrease in emission levels.
Later during that test when secondary combustion was still apparent, the
secondary air inlet was suddenly closed. The results were a decrease in
temperature in the secondary combustion zone and a sharp increase in
emissions. Secondary combustion was not reobtained upon reopening of the
air inlet after several .additional minutes.
High temperature of the primary combustion products encourages
secondary combustion by preventing quenching by the cold secondary air
with primary products above 800°C (1500°F), large quantities of unheated
secondary air still support secondary combustion. This was confirmed when
a sight glass (2 1/2 inch, diameter) in the secondary combustion chamber
inlet was removed, permitting a large influx of room air. The associated
initiation and maintenance of secondary combustion was very evident: CO
concentrations dropped reversibly from 4 percent to below 1 percent, and
THC concentrations dropped reversibly from 9000 ppm to about 1000 ppm as
the window was opened and closed. However, these tests were run at very
high burning rates to obtain the temperatures desired; the burning rates
were over 20 pounds per hour. This is appreciably higher than a homeowner
would normally operate a stove.
48
-------
The concept of secondary combustion is very effective for the
reduction of emissions. However, it is very difficult to attain when the
stove is operated at lower burning rates more typical of consumer operation,
where both the temperature and the turbulence level are lower.
Afterburning with Auxiliary Fuel
The purpose of an afterburner is to burn the emissions before they
exit the stove. This is done by introducing fuel into the flue gas already
containing sufficient air to burn the fuel and the combustible emissions.
The fuel source could be external to the stove and could be a gas such as
methane, propane or butane.
To observe the effects of an afterburner, a quarter inch stain-
less steel tube with 1/6 inch holes every inch was placed in the stove
directly below the secondary air distributor tube while the stove was operating
in the downdraft mode. Propane was- used as the auxiliary fuel and was fed
through a rotometer to monitor flow. The flowrate was from 2 to 3 cubic
feet per minute into the secondary combustion chamber.
The introduction of propane directly into the secondary combustion
chamber usually resulted in at least partial burning as indicated by gas
temperature increases. Three example observations are cited in which at
least some burning of propane occurred:
1. When the initial flue gas temperature was low
(450°F, 230°C) and the oxygen content high
(18.9 percent 02), the introduction of propane
increased the CO from 0.2 to 1.2 percent and the
THC from 3100 to 7400 ppm. With this low
initial temperature in the flue, the combustion
of propane was evidently not completed as the
gas left the stove, even though the oxygen content
remained above 10 percent.
2. At a higher temperature in the secondary combustion
chamber (810°F, 430°C), the introduction of a larger
amount of propane was more effective. The CO level
of 0.5 percent did not change appreciably, but the
THC was reduced from 4100 to 1600 ppm, and the
oxygen content was reduced from 16.0 to 6.4 percent.
The propane burning was apparently essentially
completed.
49
-------
TABLE 5. EFFECTS OF SECONDARY COMBUSTION ON EMISSIONS
Burning Fir Brands at Approximately 30 Ib/hr
Prior to
Secondary
Burning
After
Secondary
Burning
Initiation
During
Secondary
Burning
After
Secondary
Air Flow
Stopped
Time, hr. min.
Flue Gas Composition,
as Measured
02, percent
C02» percent
CO, percent
THC, ppm
Temperature in °C
2nd Comb Zone, (°F)
15:10
3.6
15.1
2.3
8200
613
(1135)
15:13
4.6
14.6
0.1
680
712
(1314)
15:59
7.6
13.0
0.1
430
656
(1212)
16:01
5.37
14.3
1.51
4100
617
(1143)
50
-------
3. At very high temperatures in the secondary combustion
chamber (1400°F, 760°C), the introduction of propane
was even more effective. The CO level was reduced
from 1.5 to 0.3 percent and the THC was reduced from
2400 to 400 ppm. The quantity of propane required to
produce these reductions was small, resulting in only
a 90°F (50°C) temperature increase and a slight 02
reduction to 6 percent.
The stove conditions during these observations were not very stable,
with appreciable gas composition and temperature fluctuations. Because of
this, the data noted above are each averages of 5 one-minute readings taken
just before or just following the initiation of propane introduction. Though
crude and preliminary data, these do show that auxiliary fuel can in some
situations reduce the emission of CO and organics.
The use of an afterburner with an external air source would reduce
emissions if the gas and air flow rates were high (compared to the total
wood emissions flow rate). This combustion would be very uneconomical if
it is placed near the exit of the stove. Another option would be to have
a control system that would regulate the amount of fuel and air. This would
be an expensive installation and might have high maintenance costs.
Thermostatic Control
Many stoves on the market today have a thermostatic control. The
thermostat consists of a coil of a strip of bimetal linked with the primary
air damper. This permits the operator to set the stove at desired and re-
producible operating condition for unattended operation. As the stove heats
up the coil will deform, lowering the damper and cutting back on the primary
air, thus reducing the temperature.
A stove with a thermostat was loaded with a large charge of wood
and monitored for two hours. The thermostat was set at the medium setting
and allowed to operate without any disturbance.
During the first hour everything was fairly steady. After which
time the control system started to cycle. The thermostat closed the primary
air damper, lowering the oxygen level and raising CO and THC emissions,
51
-------
and lowering the outlet temperature of the stove. After 10 minutes the damper
opened allowing temperatures to increase and emissions to decrease. The
thermostat again started to close the damper as the wood was depleted. Even
though the emissions and temperature were starting to oscillate the last hour
of the test, the burning rate was steady.
OPERATOR FACTORS
During the operation of a wood burning stove, there are factors
that effect emissions over which the operator has direct control. These
are the amount of wood placed into the stove, the way the wood is placed in
the stove, and the burning rate at which the stove is operated. The effects
of mass of charge and burning rate on emissions have been reported by
Butcher (19).
Amount of Wood Charged
The amount of wood placed into the stove at one loading will have
a large influence on the amount of pyrolysis products leaving the stove.
The emissions go up when more wood is within the stove, even at the same burning
rate, due to the increased quantity of wood exposed to the heat and thereby
subject to preburning pyrolysis.
To demonstrate the effect of increased wood inventory in a stove
two tests were run. In one test, 12.9 pounds were burned and the other
test burned 41.9 pounds of wood. The same triangular wood pieces were used
(oak 4 x 4's) and the stove was operated at the same air setting during the
same day. The tests were run using the sidedraft mode of burning discussed
earlier in the text.
Table 6 shows the comparison of the two tests, demonstrating that
as the wood inventory is increased the emissions increase. The burning rates
for the two tests were essentially the same but the large charge resulted
in approximately doubling the emissions.
52
-------
TABLE 6. EFFECTS OF SIZE OF CHARGE ON EMISSIONS
Side Draft Burning of Oak 4 x 4's
Test Run Number
Mass of Charge, Ib
Burning Rate, Ib/hr
Excess Air, percent
Stack Temperature, °F
(°c)
Gas Composition as Measured,
CO, percent
THC, ppm
Emissions, lb/1000 Ib wood
CO
THC
34
12.9
17.4
62.9
680
(360)
1.2
2800
100
15
35
41.9
19.2
58.1
660
(350)
2.5
5300
220
26
53
-------
Any approach that keeps the wood from pyrolyzing while it resides
in the stove prior to active burning should reduce emissions. The only course
open to the operator to reduce emissions from this mechanism is to add small
quantities of wood more frequently.
Wood Placement in Stove
How densely and where the wood is placed in the stove can affect
emissions and are controlled by the operator at the time of loading. The
wood placement in the stove controls the path of the primary air within
the stove. If the wood is packed densely, the primary air cannot reach the
internal surfaces of the wood in the pile, allowing the pyrolysis of the
wood to take place without sufficient air to burn. The location of the
wood with respect to the entrance of the primary air can allow the air to
bypass the combustion zone and exit the stove resulting in high excess air
and high emissions.
There were no quantitative tests conducted to determine the effect
of location or loading density on emissions. However, it was observed in
tests where the primary air entered through openings in the side of the
stove and the wood was placed over a bed of hot coals in a manner that reduced
the surface area exposed to the incoming air, the wood burning rate was low
and the excess air was high, as indicated by 02 content in the flue. The
wood that was left unburned late in the burning cycle was the wood in the
area where the wood had been the most densely packed. In stoves where the
wood was placed directly on a hearth with no provision for air to get under
the wood, it was found that the wood had to be piled loosely so that air
could get through the wood to allow for complete combustion of the pyrolysis
products and to achieve the desired burning rates.
Burning Rate
The rate at which the wood is consumed in a stove is known as the
burning rate and in this report refers to actual weight of wood as fired, not
the dry weight. Higher stack temperatures usually accompany higher burning
54
-------
rates, and this observable characteristic may be used to correlate burning
rates on different stoves. Most stoves are designed to operate over a
wide range of burning rates. In this program, as in conventional service,
the wood burning rates in the naturally drafted stoves were arbitrarily con-
trolled by restricting the primary air inlet to the stove, i.e., closing the
air inlet damper to reduce the rate of burning. At the lower burning rates
the oxygen concentration in the flue gas in many stoves is higher than at
higher burning rates, indicating a larger amount of excess air for the fuel
that actually burned. This seemingly anomalous trend presumably results
from a larger fraction of the total air bypassing the active burning at the
low burning rates. Thus, at the lower rates the primary burning is even more
severely air limited than the stack gas analy&es would indicate. This results
in the lower temperatures and the increased emissions of combustibles.
It became evident early in the program that the emissions of CO and
organics (as indicated by THC) were very dependent on burning rate. Figures
12 and 13 show this effect for three modes of burning in one stove (Defiant).
The data plotted for the downdraft mode have not been connected by lines to
form a curve, but the similar effect is evident. This effect has been noted
by others (19). Very high emissions associated with low burning rates (hold-
fire conditions) were observed in this program and a related TVA program
when average burning rates were maintained below about 10 pounds per hour.
Presumably, the same effect would be observed in smaller stoves when
operated at correspondingly lower rates of burning.
FUEL PROPERTIES
An operator has some control over the fuel properties of the wood
that is used. The wood can be bought seasoned or the operator can season
it. Where and how the wood is stored is also under the influence of the
operator. The size of the wood can be decreased by splitting or sawing.
The type of wood is controlled by its availability in the area. The effects
of the fuel properties are discussed in the following paragraphs.
55
-------
500
400 .
o
§
,fl
H
O
O
O
01
13
•H
x
o
I
i-i
cd
o
300 .
200 -
100 -
LEGEND
= Up Draft Mode
= Side Draft Mode
= Down Draft Mode
T
10
T
15
20
25
30
35
Burning rate, Ib/hr
FIGURE 12. CO EMISSIONS FOR ONE STOVE,
THREE MODES OF BURNING.
56
-------
'•a
o
o
O
o
o
M
S
O
o
o
!-i
•a
>•,
o
H
100 -
75 -
50 '
25
LEGEND
d = Up Draft Mode
O= Side Draft Mode
A= Down Draft Mode
A
10 15 20
Burning rate, Ib/hr
25
30
35
FIGURE 13. THC EMISSIONS FOR ONE STOVE,
THREE MODES OF BURNING.
57
-------
Moisture Content
The moisture in wood is dependent upon 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. This
reduces the rate of pyrolysis until the moisture is released from the wood.
For investigating the effects of moisture, oven dried Douglas Fir
brands were soaked in water for different periods of time to provide dif-
ferent moisture contents. The three different moisture contents were
0, 14, and 28 percent on a wet basis. All tests were run in the downdraft
mode of burning with a bed of hot fine coals on the grate, two to three
inches thick, at the start of the test. There was also a test run using
oak soaked in water to increase moisture content from 21 to 31 percent,
and split pine logs oven dried to remove all free moisture (see Table 2).
Table 7 shows the effects of the augmented moisture concent on
the emissions, with no distinct trends evident from these limited tests.
Wood Size and Shape
The size of wood has a large effect on the rate of pyrolysis. The
smaller pieces of wood result in: 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. The heat transfer co-
efficient is generally twice as high for with-the-grain transfer as opposed
to across the grain transfer. Therefore split wood will become heated and
pyrolyze quicker than unsplit wood of the same size under the same conditions.
The effects of size and configuration were investigated in tests
burning cured oak in the downdraft mode. The stove was run at the same
conditions for each test, so the burning rates were different, dependent
only on the wood. The smallest wood used was oak kindling which was four
to six inches long and had an equivalent diameter of about one inch. It
was all split, there were no round sticks used. The other three configu-
rations had the same length (11 to 14 inches) with an equivalent diameter
as indicated in Table 8. The differences were; the split wood was
triangular and was split from large round pieces of oak without bark, the
58
-------
TABLE 7. EFFECTS OF WOOD MOISTURE CONTENT
Burning Mode
Fuel Used
Moisture Content, % wet basis
Burning Rate, Ib/hr
Gas Composition as measured
CO, percent
THC, ppm
Down
Oak
21
11.3
0.3
900
Draft
Logs
31
9.8
0.1
1500
Down Draft
Fir Brands
0
14.0
0.3*
1100*
14
10.7
0.4
900
28
18.0
0.4
2000
Up Draft
Pine
0
20.7
3.9
9100
42
21.6
2.0
7100
CO, percent
THC, ppm
Emissions , lb/1000 lb wood
CO
THC
0.3
900
48
11
0.1
1500
39
34
0.3*
1100*
67
16
0.4
900
74
9
0.4
2000
49
16
3.9
9100
330
44
2.0
7100
120
25
* Secondary air rate was changed several times during this test, but no
evidence of secondary burning was observed on the continuous recording
monitors.
59
-------
TABLE 8. EFFECTS OF WOOD SIZE
AND SHAPE ON EMISSIONS
Fuel: Air Dried Oak
Test Run Number 61 64 60 63
Wood Size, vL -vL-3/4 2.8. 4
equivalent (4x4
diameter, inches triangular)
Shape split split sawed rounds
Burning rate, Ib/hr 19.1 11.3 18.4 12.8
Gas Composition
as measured
CO, percent 1.0 0.3 0.5 0.1
THC, ppm 1600 900 1000 600
Emissions
lb/1000 Ib wood
CO 110 48 61 19
THC 10 11 7 7
60
-------
4x4 wood was triangular and was sawed from 4 x 4's cut diagonally, the
rounds were round oak logs, the bark was left on. All tests burned about
the same amount of wood (14 to 16 pounds).
Table 8 shows the emission data. As expected, the split kindling
resulted in the highest emissions, the larger round pieces had lowest
emissions. These data are shown in Figure 14, together with the same
emission data from test 58 burning dry fir brands.
Wood Type
The differences in ultimate analyses of wood on a dry basis
(carbon, hydrogen and oxygen) are within one to two percent for the majority
of all wood species. The more important differences between woods arise
from moisture, density, and heating value. The inherent difference between
soft wood and hard wood is the amount of resins that are contained in soft
woods, which increases their heating value on a weight basis.
In two popular stove configurations, updraft and sidedraft, burning
of green pine resulted in slightly lower emissions than the dried oak 4x4
fuel. The one test conducted with the same pine fuel after being oven dried,
to nil moisture content resulted in emissions only slightly higher than the
oak 4x4 fuel. The relatively lower emissions for the green pine presumably
result from the delayed pyrolysis associated with the moisture content,
whereas the dried pine pyrolyzed at a higher rate resulting in the increased
emissions. Table 9 summarizes these comparisons.
Charcoal was burned in the down draft mode to demonstrate the
effects of a wood based fuel with very low volatile matter (Test Run 76).
As expected, the emissions for both total hydrocarbons and carbon monoxide
were the lowest of all test runs as shown in Table 1.
POLYCYCLIC AROMATIC HYDROCARBON EMISSIONS
The emissions of PAHs from the three burning modes in the naturally
drafted radiant stoves are shown in Table 10. The differences between the
three burning modes while burning oak was not great. The emissions for down-
61
-------
140 .
120 .
•a
g 100 J
O
O
O
W
B
O
•H
w
eo
•H
W
80 .
60
40
20 .
LEGEND
Oak, Round ,
Oak, Split
Oak, 4x4
Oak Kindling
Fir Brands
10 20 30 40 50 60 70 80
Surface Area Wood/Volume Wood, I/ft
90 100
FIGURE 14. EFFECT OF WOOD SIZE ON CO EMISSION FACTORS.
62
-------
TABLE 9. EMISSIONS FROM BURNING
GREEN PINE
Burning mode
Fuel Used
Updraft, with grate
Oak
4x4
Burning rate, Ib/hr 20.9
Gas Composition,
as Measured
CO, percent 2.7
THC, ppm 5900
Emissions,
lb/1000 Ib wood
CO 270
THC 34
Green
Pine
21.6
2.0
7100
120
25
Dried
Pine
20.7
3.9
9100
330
44
Side Draft
Oak
4x4
18.7
2.5
6200
190
28
Green
Pine
18.4
0.5
1800
40
9
63
-------
draft burning (50 lb/10 Ib wood) was about 3 times the emissions of updraft
burning with a grate (18 lb/10 Ib wood). The side draft burning was
intermediate, in PAH emissions, with wet pine producing a slightly higher
factor than dry oak. This difference is reversed for the emissions for
Benzo(a)pyrene (BaP), a very toxic constituent of the total PAH mixture.
The BaP emissions measured for updraft burning (1.4 lb/10 Ib wood) were
higher than that from sidedraft burning, and much higher than from down-
draft burning (0.38 lb/10 Ib wood). The sidedraft mode was intermediate
again, with wet oak and dry pine producing 1.2 and 0.52 lb/10 Ib wood burned,
respectively.
When burning split oak logs in the high turbulence burner, appre-
ciably lower emissions were observed than with either of the more conventional
stove designs. Details of the measurements, techniques, and experimental
observations of PAH emissions are presented in more detail in Appendix C.
COMMERCIAL HIGH TURBULENCE BURNER
The commerical stick burner was made to burn very efficiently by
utilizing a high turbulence level and other techniques as noted earlier (see
page 23). It was operated as prescribed by the manufacturer burning stick
wood. When the burner was fired with 4 very large split logs, one jammed in
the hopper and remained partially unburned. When fired with the recommended
size of split logs, operation was reliable with low emissions. When fired
with smaller sizes of split logs the emissions increased only slightly.
Table 10 shows the measured emissions when the burner was run until the
blower was automatically turned off. When operating properly the emissions
are appreciably less than from the two typical radiant stoves. Even when
burning brands to deliberately maximize early hydrocarbon emissions or
when faulty fuel performance was encountered, the THC emissions were not
as high as from the more conventional radiant stoves. The low CO emissions,
when properly burning the more appropriate split oak logs, is also much
lower than from the radiant heaters.
These data demonstrate how appreciable reductions in emissions can
be obtained with the application of several combustion modification techniques.
Unfortunately, many of these techniques are not applicable to the naturally
drafted radiant stoves, without extensive system redesign.
64
-------
TABLE 10. EMISSIONS FROM HIGH TURBULENCE BURNER
High Turbulence Burner
Fuel
Burned
Burning Rate, Ib/hr
Excess Air, %
Gas Composition
as Measured,
CO, %
THC, ppm
Emissions,
lb/1000 Ib wood,
CO
THC
PAH
Stack Inlet Temperatures,
°F
(°C)
Large
Split
Oak
Logs
22.1*
220
0.6
2000
97
18
ND
247
(119)
Split
Oak
Logs
**
24.1
920
0.1
200
25
5
0.011
343
(173)
Small Split
Oak Logs
( 4" edges)
21.9
110
0.3
300
30
2
ND
331
(166)
Dry
Fir
Brands
20.9
120
1.6
1900
222
16
ND
285
(141)
Comparable
Updraft
Stove With
Grate
Dry Oak
4x4
16.5
40
3.6
7100
280
32
0.018
950
(510)
Comparable
Side Draft
Stove
Dry Oak
4x4
18.7
42
2.5
6200
190
28
0.035
703
(393)
** =
ND
This run terminated with 31% of fuel unburned. One log jammed in the
hopper and did not fall into the burning region.
Sizes of split logs recommended by manufacturer.
Not determined.
-------
REFERENCES
(1) Hartman, D. L. Proceedings of the DOE Residue and Waste Fuels
Utilization Program Contractor Review Meeting. U.S. Depart-
ment of Energy, December 1979.
(2) Shelton, J., and A. B. Shapiro. The Woodburners Encyclopedia.
Vermont Crossroads Press, 1976.
(3) Badger, G. M., J. K. Donnelly, and T. M. Spotswood. Some Reactions
of Branched-Chain Carboxylic Acids. Aust. J. Chem., 19:1023-1033, 1965.
(4) Badger, G. M. Mode of Formation of Carcinogens in Human Environ-
ment. Nat. Cancer Inst. Monogr., 9:1-16, 1962.
(5) Hoffman, D., and E. L. Wynder. Chemical analysis and Carcinogenic
Bioassays of Organic Particulate Pollutants. In: Air Pollution,
Vol. II, 2nd Ed., Chapter 20, A. C. Stern, Ed. Academic Press,
New York, 1968. pp. 187-247.
(6) Bauer, S. H. Comments. In: Tenth Symposium on Combustion.
Combustion Institute, Pittsburgh, Pennsylvania, p. 511.
(7) Burton, H. R., and D. Burdick. Thermal Decomposition of Tobacco.
I. Thermogravimetric Analysis. Tob. Sci., 11:180-185, 1967.
(8) Tibbitts, T. W. Ignition Temperature of Tobacco. Method of Deter-
mination and Relation to Leaf Burn. Tob. Sci., 6:170-173, 1962.
(Pub. in Tobacco 155(19):30-33, 1962).
(9) Grimmer, G., A. Glaser, and G. Wilhelm. The Effect of Temperature
and Flow Rate of Air and Nitrogen Atmospheres on the Formation of
Benzo[a]pyrene and Benzo[e]pyrene from Pyrolysis of Tobacco. Beitr.
Tabakforsch., 3(6):415-421, 1966. (In German).
(10) Robb, E. W., W. R. Johnson, J. J. Westbrook, and R. B. Seligman.
Model Pyrolysis. Study of Cellulose. Beitr. Tabakforsch., 3(9):
597-604, 1966.
(11) Nekomm, S., J. Bonnet. On the Combustion of Organic Material and
Origin-of Carcinogenic Substances in Tobacco Smoke and in Food.
Oncologia, 13:266-278, 1960.
(12) Pyriki, C. Bull. Inf. CORESTA, No. 1:11, 1960.
66
-------
(13) Edmonds, M. D., M. T. Core, A. Bavley, and R. F. Schwenker.
Applications of Differential Thermal Analysis and Thermogravi-
metric Analysis of Tobacco. Tob. Sci., 9:48-53, 1965.
(14) Scholtzhauer, W. S., I. Schmeltz, and L. C. Hickey. Tob. Sci.,
11:31, 1967.
(15) Kato, K., F. Sakai, and T. Nakahata. Thermal Decomposition of
Tobacco Lignin. Nippon Senbai Kosha Chuo Kenkyusho Kenkyu Hokoku,
107:171-175, 1965.
(16) Holmes, F. H., and C. J. G. Shaw. The Pyrolysis of Cellulose and
the Action of Flame-Retardants. I. Significance and Analysis of
the Tar. J. Appl. Chem., 11:210-216, 1961.
(17) Schwenker, R. F., Jr., and L. R. Beck. Study of the Pyrolytic De-
composition of Cellulose by Gas-Chromatography. J. Polymer Sci.,
Part C:Polymer Symposia, No. 2, 331-340, 1963.
(18) Woodstove Directory, Energy Communications Press, Inc., 1980.
(19) Fuchs, N. A. The Mechanics of Aerosols. MacMillan Co., New York,
1964.
(20) Butcher, S. S., and E. M. Sorenson. A study of Wood Stove Particu-
late Emissions. J. Air Pollut. Control Assoc., 29(7):724-728, 1979.
The additional references listed below provide definitive information about
the characterization, measurement, and significance of wood stove emissions.
(1) DeAngelis, D. G., D. S. Ruffin, and R. B. Resnik. Preliminary
Characterization of Emissions from Wood-Fired Residential Combustion
Equipment. EPA-600/7-80-040, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, March 1980.
(2) DeAngelis, D. G., D. S. Ruffin, J. A. Peters, and R. B. Resnik.
Source Assessment: Residential Combustion of Wood. EPA-600/2-80-042b,
U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina, March 1980. 99 pp.
(3) Hall, R. E., and D. G. DeAngelis. EPA's Research Program for Control-
ling Residential Wood Combustion Emissions, J. Air Pollut. Control
Assoc., 30(8):862-867, 1980.
(4) Cooper, J. A. Environmental Impact of Residential Wood Combustion
Emissions and Its Implications. J. Air Pollut. Control Assoc.,
30(8):855-861, 1980.
67
-------
APPENDIX A
EMISSION MEASUREMENT PROGRAMS
In the course of this project a search has been conducted to
identify programs in this country in which actual measurements of stove
emissions have been or are being made, wherein some correlation is possible
with store design and operating conditions. The programs are few, and
there has been no consistency in the experimental techniques used for
• Stove installation
• Stove operation
• Sampling of emissions
• Analysis of collected samples
• Interpretation of the results
The programs have used different fuels burned in different stoves
with different firing techniques, all of which have generally been adequately
described.
Major problems exists 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.
Monsanto Research Corp.
This organization has conducted the most extensive analysis of
emissions from residential wood burning equipment. The 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 gasses (02> C02, SOX, NOX, 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 relatively high burning rates, and
did not demonstrate the effects of burning rate on emissions.
68
-------
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 gases samples combined)
and bioassays of these specimens were found to be both mutagenic and
cytotoxic.
This program was managed by D. G. DeAngelis of Monsanto under
EPA IERL/RTP sponsorship.
California Air Resources Board
This agency has 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.
These tests were conducted by Peter Kosel of the Stationary Source
Control Division of the Air Resources Board.
Bowdoin College, Main
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 quantaties
of fresh air. No gas composition or excess air measurements have been made.
The implications 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.
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.
This program has been led by Professor S. S. Butcher, and the
results published in the APCA journal. Support has been received from the
State of Maine.
69
-------
Canadian Combustian Research Laboratory
This laboratory has conducted emissions analyses and efficiency
tests on wood stoves of different generic designs. Continuous measurements
are made of CO, C02, 02, NOX, and total hydrocarbons. Split cordwood
has been burned at relatively low rates (< 50,000 Btu/hr) while overall
thermal efficiencies have been determined by the stack gas (indirect)
method.
The results to date 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 up
draft stoves were observed.
This study is led by ACS Hayden and R. W. Braaten far the
Department of Energy, Mines and Resources, Ottawa, Canada.
Del Green Associates, Woodburn, OR
This company has just started 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 pollution measurements
in a few residences heated by wood burning within the living space.
This program is being led by Norman Edmisten with cooperation
from OMNI Environmental.
70
-------
Auburn University - Auburn Alabama
The Mechanical Engineering Department of Auburn has conducted
research on residential wood burning equipment for several years. This work
has focussed on the measurement and characterization of stove performance,
and the safety of wood stove operation including creosote problems. They
have developed extensive facilities and capabilities for measuring stove
efficiency and heat output rate. Their emission measurements are normally
limited to 02, C02 and CO as manually determined from grab samples by Orsat.
Auburn's program for improving the efficiency, safety, and utility
of woodburning units has been sponsored by DOE and has been led by Professors
Glennon Maples, David Dyes and Timothy Maxwell. They have also been active
in developing wood stove evaluation techniques in conjunction with the ASHRAE
and the Fireplace Institute, and conduct the Fireplace Institute's ex-
perimental evaluation program.
Argonne National Laboratory, DOE
This laboratory has an ultimate objective to develop emission factors
for residential wood stoves. Their experimental program includes an in-
vestigation of stove design characteristics that affect emissions, and the
development of techniques and instrumentation to measure and characterize
these emissions.
Their experimental program involves three stoves in a laboratory
facility with on-line CO and C02 measurements. Stack samples are obtained
with cold traps, and analysis of POMS is with GC/MS.
Their preliminary measurements have focused on low burning rates
(air inlets 10 to 25% open). With above-bed temperatures below 400°F, no
polycyclic species have been detected. More recently traces of naphthalenes
have been observed.
John Harkness is leading this program at Argonne as an in-house
study for DOE. No publications describing emissions have been issued to
date.
71
-------
Tennessee Valley Authority
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 brands, using the UL batch burning procedure in which
a test consists of burning a fresh charge of brands to 95% 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,
Sattelle 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.
The results of some of these tests have been used in support of this EPA
program at Battelle.
The TVA program has been actively managed for the Solar Applications
Branch of TVA by Dr. Jerry Harper. No publication of these studies has been
made to date.
Virginia Polytechnic Institute
This university has an active program related to control of emissions
from wood stoves, Their three phases include:
(1) Data aquisition facility development
Their facility includes on line continuous measurement
of 02, C02, CO, NOX, and stack gas flow measurement.
They plan to incorporate a hot FID instrument for total hydrocarbons,
(2) Catalyst bed afterburner
This development is focussed on a bed type unit for stack
installation rather than within the stove.
(3) Staged combustion studies
This effort is to determine experimentally what conditions
are necessary and sufficient to assure secondary combustion
without auxiliary fuel. This phase is focusing on gas
composition, temperatures, and mixing rates.
72
-------
This program has not resulted in any published results, although
catalysis tests have been run.
Dr. Dennis Jaasma is conducting this program primarily supported
by University funding although a limited industrial grant has been received.
They are seeking support for continuation of the program.
National Bureau of Standards
The center for Fire Research, NBS, in conducting a study of "Gaseous
and Liquid Combustions Products from Wood Burning Devices". The objective is
to provide data and techniques to support combustion modification efforts.
The emphasis is to be placed on detailed physical and chemical characterizaion
of the emission products as affected by time during a burn cycle, type of wood,
moisture content, and excess air. Both within firebox phenomena and within
chimney effects are of concern.
This study is to be led by Dr. Richard Gann, NBS, on funding
provided by DOE. No published data has been generated by this program.
Thermocore, Inc. Lancaster, Pa.
This company has attempted to improve stove thermal efficiency
(including combustion efficiency) by means of altered air supply, focussing
on preheating of the air, and alternative paths for air entry into the stove.
Their experimental program included extensive temperature measurements with
thermocouple recordings maintained throughout batch type test runs. Their
emission measurements were limited to Orsat analyses made at several times
during the burning of a large charge of wood. No combustibles were measured
by Orsat, and hydrocarbon emissions were estimated by visual plume
observations.
Yale Eastman managed this program under the sponsorship of the
Philadelphia office of the DOE. Their report has been submitted to DOE, but
not released in any form for publication.
73
-------
New York University of Plattsburgh
This program of stove development has focussed on overall
efficiency and particulate emissions as affected by stove design, operating
procedures, and controls development (thermostatic). The emission measure-
ments have been limited to particulates as measured by collection on a
cooled but not temperature-controlled filter from stack gasses 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 one minute
sampling periods at 10 minute intervals through out run periods of several
hours. Stove operation has generally been at very low burning rates, i.e.,
stove exit temperatures 140 to 500°F.
This program has been conducted by Professor Stockton Barnett with
personal and internal funding, and outside support is being sought. No
results have been formally published.
Vermont Environmental Control Agency
This agency has been conducting a field measurement program of
the ambient environment in an area where residential heating is extensively
supplied by wood stoves. The community has no large industrial sources, and
an extensive inventory has been conducted to determine the true extent of
wood burning for residential heating. The principal emphasis is on the
determination of the contribution of wood burning to the total particulate
loading, as determined by analysis of the particulates caught on hi-volume
samplers placed around the community. Microscopic observations of the
particulates are being made to determine their source or origin. Some
emission measurements have been made at residences using a Rader High Volume
Sampler and an EPA Method 5 sampler. The organic materials penetrating the
Method 5 filter have been measured by extraction from the back half impingers,
but these materials have not been analyzed to determine their specie composition.
Cedric Sanborn is conducting this program with state agency support.
74
-------
Institute of Man and Resources.
This organization has conducted a field demonstration program
of ten different residential furnaces. The emission measurements in-
cluded CO, NOX, S02, 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 ex-
tracted with benzene and dried to determine the condensable organic content
which varied from 13 to 51 percent of the total particulate matters
collected.
The emission test program was conducted for the institute by
Atlantic Analytical Services Ltd., Saint John, N.B. and led by Alex Graham.
The overall program received support from both provincial and federal
Canadian funds, and is coordinated by Rob Brandon.
75
-------
APPENDIX B
EXAMPLES OF RECORDED TEST DATA
Attached are several typical plots of gas composition measurements through-
out test runs. The high concentrations of organics, as indicated by the total
hydrocarbon (THC) measurements, are usually accompanied by high CO concentrations
as in the up draft and side draft burning modes (Figures B-l and B-2). These
result primarily from the rapid early pyrolysis of the wood after being placed
in the stove. The high rate of pyrolysis gas generation, exceeding the local
availability of oxygen for complete burning, lasted about one quarter hour in
the side draft mode of burning, and for nearly an hour in the up draft mode with
airflow through the grate. The latter stages of burning in these two tests
consisted primarily of burning the residual char with sufficient local air
supply, thus resulting in lower emissions of organics as indicated by THC
measurements.
These effects were not observed in the down draft mode, Figure B-3, as the
new charge of wood was not as rapidly heated by combustion products in the test
shown here, even though a similar charge of wood was placed in the pre-heated
stove. The high turbulence burner, Figure B-4, also demonstrated this lack of
early pyrolysis, because the wood moved progressively from the cooled upper chamber
down into the active primary burning area as burning progressed. The small peak
in organic emissions at approximately 0.5 hours into the high turbulence mode test
was presumably due to an erratic motion of wood into or within the primary
burning area. In these two types of burning, the composition of the fuel actually
burning did not change appreciably during the test.
Figure B-5 shows the weight loss record during the side draft burning. It is
evident that the weight loss rate was higher during the early stages of burning
and generally reduced throughout the test. The rate of heat released, however,
did not fall off proportionally, as the early weight loss was associated with
relatively low heating value pyrolysis products, whereas the later burning rates
were associated with fuel which was primarily char with a higher heating value.
76
-------
6/10/80 RITEWAY OAK LOGS RUN 77
25.000 r
20.000 -
I
Z
O 15.000 -
Q C02
e eo
Z
O
u
10.000
5.000
o.ooo
25.000 r
-i 20.000
0.000
0.000 .313 .625 .938 1.250 1.563 1.875
TIME (HOURS)
0.000
FIGURE B-l. COMPOSITION OF GAS FROM UP
DRAFT BURNING WITH A GRATE
Burning split oak logs
Burning rate 14.4 Ib/hr (6.5 Kg/hr).
77
-------
6/17/80 DEFIANT (SIDE DRAFT) OAK RUN 87
o
o
25.000 r
20.000 -
15.000 -
10.000 -
5.000
o.ooo
o coa
o cc
25.000 r
-I 10.000
8.000
6.000
X
o
4.000 i
a.
o_
2.000
0.000
0.000 .250 .500 .750 1-000 1.250 1.500
TIME (HOURS)
FIGURE B-2. COMPOSITION OF GAS FROM SIDE
DRAFT BURNING
Burning split oak logs
Burning rate 16.6 Ib/hr (7.5 Kg/hr).
78
-------
5/15/80 DEFIANT (DOWN DRAFT) ROUND OAK LOGS N.D-63
o
o
25.000 r
20.000 -
15.000 -
10.000
5-000
0.000
25.000 r
o.ooo I 1 1 1 1 1 1 i
50.000
40.000
30.000
o
X
20.000
0_
a.
10.000
0.000 .250 .500 .750 1.000 1.250 1.500*
TIME (HOURS)
FIGURE B-3. COMPOSITION OF GAS FROM DOWN
DRAFT BURNING
Burning small split oak cordwood
Burning rate 19.1 Ib/hr (8.7 Kg/hr).
0.000
79
-------
6/18/80 DUMONT OAK LOGS RUN 91
i
z
<
o:
z
o
o
25.000 r
20.000
15.000
10.000
5.000
o.ooo o
o 002
o co
-S B-
LfTj L~ I —
I
z
<
o:
z
o
o
25.000 r
20.000
15.000
10.000
5.000
o.ooo
OXTQEH
TOT.HC
-e -1-
50.000
40.000
30.000
o
I
20.000 *7
z
0_
0.
10.000
0.000 .313 .625 .938 1.250 1.563 1.875
TIME (HOURS)
FIGURE B-4. COMPOSITION OF GAS FROM HIGH
TURBULENCE BURNING
Burning split oak logs
Burning rate 24.1 Ib/hr (11 Kg/hr) .
o.ooo
80
-------
6/17/80 DEFIANT (SIDE DRAFT) OAK RUN 87
25.000 r
20.000
15.000
u_
o
CO 10.000
CD
5.000
0.000
Q
WEIGHT
J I
I I I
J I I I I
0.000 .250
.500 .750 1.000 1.250 1-5
TIME (HOURS)
FIGURE B-5. WEIGHT LOSS DURING SIDE DRAFT BURNING
Burning split oak logs
Burning rate: 16.6 Ib/hr (7.5 Kg/hr).
81
-------
APPENDIX C
POLYCYCLIC AROMATIC HYDROCARBON
MEASUREMENTS IN FLUE GAS
Particulate emissions and polycyclic aromatic hydrocarbons (PAH)
samples were collected using a Modified Method 5 sampling apparatus. Both
the location and configuration of the sampling system were described pre-
viously in the experimental section (Section 6). This apparatus allowed the
collection of particulate emissions using filtration followed by collection
of gas phase PAH by adsorption on XAD-2 resin. Operation of the sorbent
*
sampler and the extraction procedure have been described elsewhere. The
analysis of PAH compounds is complicated by the structural similarity of many
members of this class of compounds. Several compounds exhibit positional
isomers, e.g., benzo(a)pyrene and benzo(e)pyrene, which coelute on most
chromatographic systems. Such isomeric groups often show drastic differences
among their members in physiological behavior. For instance, benzo(a)pyrene
is known to be a potent animal carcinogen while benzo(e)pyrene is only weakly
active. In order to resolve these isomeric groups, high resolution, gas
capillary column chromatography coupled to quadrupole mass spectrometry was
chosen to separate PAH compounds from these complex mixtures. Single ion
monitoring was used to quantify PAHs. Single ion monitoring improved the
sensitivity of mass spectral measurements along with providing electronic
isolation of fragment or parent ions characteristic of the compound of
interest. An additional parameter was used to verify compound identity, the
gas chromato graphic retention time. From the elution of a set of PAH standards
* Neher, M. B., Jones, P. W., and Perry, P. J. Performance Evaluation
of a Solid Sorbent Hydrocarbon Sampler, Research Report EPRI EP-959,
Electric Power Research Institute, January, 1979.
82
-------
of known purity, a retention time window was determined and quantification
was performed only on those compounds eluting at the proper time. Chromato-
graphic and mass spectrometer stability were verified by analysis of the
reference standards at the beginning and end of each run of samples. In
addition, standards were run in triplicate at five concentration levels to
establish the linearity and precision of the analytical curve. The capillary
column used for this analysis exhibited a very high chromatographic efficiency
as shown by the chromatogram for a test mixture of 21 PAH standards in
Figure C-l. The extracts from both filter and resin catches proved to contain
a complex mixture of organic species as shown in the GC/FID trace of the
particulate associated PAH from a radiant wood stove burning seasoned oak
brands and operating in the up draft mode (see Figure C-2).
The specific compounds quantified are listed in Table C-l. It
should be noted that several isomeric groups were measured together, e.g.,
methylanthracenes. Since PAH species often exhibit a simple mass spectral
fragmentation pattern consisting of the parent ion, this single ion can be
searched at retention times corresponding to known members of the compound
class and area responses grouped together for quantification.
The measured levels of individual PAH offer a very interesting
comparison of the effect of burn mode on distribution of PAH formed during
combustion (see Table C-2). PAH production during combustion of wood is an
especially interesting phenomena since neither lignin nor cellulose contain
PAH compounds in the raw fuel. The data of Table C-2 show total PAH in
milligrams emitted during combustion of a weighed charge of wood. A portion
of the flue gas was sampled at the flue gas outlet and the measured PAH
concentration was multiplied by a factor representing the total flue gas
effluent. Thus, Table C-2 gives the total PAH emissions discharged by the
test stoves.
Examining the data in Table C-2, several effects are noticed. The
lower molecular weight PAH are found predominantly in the gas phase while the
larger molecules tend to be associated with particulate matter. During
sampling, the nozzle and filter holder were thermostatically controlled to
remain at or above 250°F (121°C), but the flue gas entering the sampler was
regularly higher than this temperature. Thus the absolute distribution of
PAH between gas and particulate samples is a function of sampling conditions.
83
-------
00
01
o> a
G 0)
QJ 43
rH 4-1
cd 43
43 0.
4-1 Cd
43 C
PL, 0)
n) u
53
43
4J
(X
10
C
01
O
pi
0)
/^ Anthrac
01
a
0)
u
cd
43 01
•I—) C
0) (3 01 01
(3 n) H 01 C
01 ^ >-, C 01 C
43 -, 0
QJ (3
|3 cd
01 M
0 O
cd 3
^1 r-l
43 Fn
4J
C
id
1-1
t^.
43
4->
01
g
1
er>
OJ O C Cd Cd P. r-
ai t3N OJV-'^I^P
C Ol(3 43O43^3^
01 iHOO) 4JN4JCJQ
M !^ idpQ (3 |3 a 1 C
>, 43M cdQJQJ cdfOx-
FL, 4J 43
43 4J
a c
cd cd
13 r-l
•H >-.
Pd C
1 QJ
- 43
r-l f^
" 1
iH CT\
l^
t-l C PQ rH •• T
O QJ r^, CM
3 t-i C - 4:
rH >-, OJ rH
in Cu
01 s-^ s~~.
C r^
0) v-^
03 0
!>> N
U f3
43 QJ
0 pq
,
0)
o
N
c
QJ
pq
(
43 v— • 6
QJ CL, O •*-
13 -H f3 C
01 O OJ N
rH 1 13 C
P>-> O C Q
r-l rH M PC
OJ "
PM cr>
k
QJ
a
o
M
O
FIGURE C-l. GAS CHROMATOGRAM OF PAH STANDARD MIXTURE.
-------
00
FIGURE C-2. GAS CHROMATOGRAM OF PARTICULATE ASSOCIATED ORGANIC COMPOUNDS
FROM SEASONED OAK COMBUSTION.
-------
TABLE C-l. POLYNUCLEAR AROMATIC SPECIES QUANTIFIED
Compounds in Order of Gas Chromatographic
Elution on SE-52
Naphthalene
Acenaphthene
Acenaphthylene
Fluorene
Phenanthrene
Anthracene /• v
Methylanthracenes/Methylphenanthrenes
Fluoranthene
Pyrene
Methylpyrenes/Methylfluoranthenes
Benzo(a)anthracene
Chrysene ,, ,
Methylchrysenes
Dimethylbenzo(a)anthracene
Benzofluoranthenes(b)
Benzo(a)pyrene
Benzo(e)pyrene
Perylene
Indeno(1,2,3-cd)pyrene
Benzo(g,h,i)perylene
Coronene
Dib enzo(a,h)pyrene
(a) Summation of all isomers identified in sample excluding
internal standard: 9-methylanthracene.
(b) Summation of all isomers identified in sample.
86
-------
TABLE C-2. TOTAL PAH (rog) EMITTED IN WOOD STOVE EMISSIONS
Burn Mode
Fuel
Species Measured
Naphthalene
Acenapthene
Acenaphthylene
Fluorene
Phenanthrene
Anthracene
Methylanthracenes/
Me thy Iphenanthrenes
Fluoranthene
Pyrene
Methylpyrenes/Methyl-
fluoranthenes
Benzo (a) anthracene
Chyrsene
Methylchyrsenes
Dimethylbenzo (a) anthracene
Benzof luoranthenes
Benzo(e)pyrene
Benzo(a)pyrene
Perylene
Indeno (1 , 2 , 3-cd) pyrene
Benzo (g,h,i)perylene
Coronene
Dibenzo (a ,h) pyrene
Total
Total Emission
Side Draft
Oak-Seasoned
Particulate Gaseous
1.2
0.12
0.06
0.08
0.24
0.09
0.07
2.6
2.8
0.79
5.9
5.6
0.56
0.29
14
3.5
5.0
0.51
15
22
10
4.2
95
358
45
58
26
19
62
18
3.2
18
8.6
0.53
1.0
1.1
0.14
-
0.96
0.24
0.27
0.10
0.29
0.20
0.11
-
263
Side Draft
Pine-Green
Particulate Gaseous
0.22
0.27
0.06
0.12
2.2
0.35
0.13
2.9
3.1
0.63
3.0
3.2
0.54
0.28
21
7.1
13
1.8
38
30
16
9.3
154
502
80
89
7.3
26
100
19
7.4
11
6.1
0.82
0.22
0.21
0.12
-
0.15
0.08
0.08
0.09
_
-
-
-
348
Up Draft Down Draft
Oak-Seasoned Oak-Seasoned
Particulate Gaseous Particulate Gaseous
0.16
0.10
0,04
0.07
0.35
0.09
0.07
0.30
0.28
0.16
0.39
0.37
0,16
0.09
1.3
0,42
6.3
0.14
1.3
1.2
0.51
0.39
14
If
11
18
6.4
8.2
20
4.6
1.4
9.1
6.8
1.9
13
8.2
1.0
0.25
17
3.8
6.5
0.68
3.9
3.2
0.55
0.23
146
.0
0.07
0.12
0.05
0.09
0.17
0.08
-
0.12
0.07
-
0.16
0.11
_
„
0.11
0.07
0.07
0.09
0.17
0.13
0.17
0.12
2.0
3
52
51
4.8
3.4
96
11
2.0
68
38
2.6
6.1
8.5
0.26
0.13
3.7
0.48
0.20
0.09
0.15
0.09
_
349
51
High Turbulence
Oak-Seasoned
Particulate Gaseous
0.36
0.60
0.27
0.46
0.99
0.43
0.33
0.73
0.46
0.30
0.96
0.73
—
0.93
0.50
0.50
0.50
1.1
0.83
0.86
12
1
62
31
1.9
2.4
40
2,2
0.99
15
9.6
1.5
1.8
1.9
0.60
0.86
0.43
0.40
0.46
0.80
0.50
0.63
175
87
-------
In normal home operations, the cooling of flue gas on leaving the stove
distribution system will cause an association and condensation of PAH with
the discharged particulate matter. Thus stove emissions measured after
discharge to the ambient air would be expected to have increased particulate-
associated PAH concentrations.
As a general trend, observed in many combustion generated PAH
samples, the amounts of high molecular weight PAHs are generally lower than
the amounts of low molecular weight compounds. Phenanthrene was found to be
a predominant species, a situation analogous to high phenanthrene concentra-
tions found in coal combustion products. Benzo(a)pyrene and benzo(a)pyrene
were produced in equivalent amounts except in the case of the up drafted
stove burning seasoned oak, a situation where the benzo(a)pyrene content was
appreciably higher.
The measured gas concentration and emissions of the PAHs are shown
in Table C-3. The operating conditions of the stoves were shown in Table 2
of the report. Comparing stove burning modes, it is observed that PAH
production was similar for different modes, with up draft producing slightly
less total PAH than side or down draft. Green pine generated slightly higher
levels of PAH than seasoned oak when the two fuels were burned in the same
stove. The most significant reduction in total PAH was observed in the high
turbulence furnace which evidently produced a highly efficient burn condition
and greatly reduced PAH emissions.
Benzo(a)pyrene average concentrations and emissions were separately
determined because of the special interest in this specie and its relatively
high level of toxicity. The side draft mode produced slightly lower emissions
of this specie when burning oak than the up draft burning mode. The
down draft and high turbulence burning modes, however, resulted in signifi-
cantly lower average emissions than the more conventional side draft and up
draft modes. These benzo(a)pyrene average emissions were lower by one to two
orders of magnitude.
A recent study of PAH emissions from residential wood combustion
*
was reported by EPA. Data from that program can be directly compared with
* DeAngelis, D.G., Ruffin, D.S., Reznik, R.B., and Milliken, J.O.,
Preliminary Characterization of Emissions from Wood-Fired Residential
Combustion Equipment, EPA-600/7-80-040, March, 1980, pg 60.
88
-------
TABLE C-3. COMPARISON OF STOVE OPERATING PARAMETERS
AND EMISSION FACTORS FOR PAH COMPOUNDS
Burn
Mode
Fuel
Side
Draft
Oak(a)
Side
Draft
Pine(b)
Up
Draft
Oak(a)
Down
Draft
Oak(a)
Turbu-
lence
Oak
-------
data from this study (see Table C-4) since test stoves were operated with
similar fuels and similar operating parameters. In Table C-4, specific
compounds and compound classes which were measured in both studies are
compared as parts-per-million (W/W) of discharge based on the weight of
fuel burned.
The data from these two studies show similar patterns for production
of PAH although the earlier report found generally higher levels of total PAH.
Both studies found that the lower molecular weight PAH were produced in
larger amounts than larger molecules and both studies showed that anthracene
and phenanthrene are a large fraction of the PAH produced. This finding
suggests that benzene may be produced in significant amounts during
residential wood combustion and future studies of wood combustion products
should include this important suspect carcinogen in measurements.
The apparent difference in levels of PAH measured in the two
studies can possibly be explained by differences in stove operating parameters
or possibly by the difference between the packed column gas chromatographic
technique used in the earlier report compared to the glass capillary method
used in this study. The packed column can make integration of peak areas
difficult due to coelution and peak overlap. Since these two studies were
not conducted identically, the observed differences in PAH emissions are well
within the variability of this type of emission measurement, and the vari-
ability in operating parameters for the stoves tested. Operating parameters
are expected to be the major contributing factor to observed differences in
PAH content, but consideration should be given to the lack of multiple tests
required to establish the precision of reported data.
90
-------
TABLE C-4. COMPARISON OF REPORTED PAH EMISSIONS (rag/kg or lb/10 Ib Wood Fuel)
Stove
Fuel
Baffled Non-Baffled*3) Non-Baffled(a) Side Draft(b) Side Draft(b) Up Drafc(b) Down Draft(b> High Turbulence^)
Oak(c) Oak(c) Pine(d) Oak(e) Pine
-------
TECHNICAL REPORT DATA
(Pleat read lauructions on the reverse before completing)
.REPORT NO.
EPA-600/7-81-091
2.
3. RECIPIENT'S ACCESSION NO.
.TITLE AND SUBTITLE
Control of Emissions from Residential Wood Burning
by Combustion Modification
5. REPORT DATE
May 1981
6. PERFORMING ORGANIZATION CODE
.AUTHORIS)
8. PERFORMING ORGANIZATION REPORT NO.
J.M.Allen and W.M.Cooke
. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle, Columbus Laboratories
505 King Avenue
olumbus, Ohio 43201
10. PROGRAM ELEMENT NO.
C9BN1B
11. CONTRACT/GRANT NO.
68-02-2686, Task 114
2. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 6/79-11/80
14. SPONSORING AGENCY CODE
EPA/600/13
s.SUPPLEMENTARY NOTES iERL_RTp project officer is Robert E. Hall, Mail Drop 65, 919/
541-2477.
6. ABSTRACT
The report describes an exploratory study of factors contributing to atmos-
pheric emissions from residential wood-fired combustion equipment. Three commer-
cial appliances were operated with both normal and modified designs, providing dif-
ierent burning modes: updraft with a grate, updraft with a hearth, crossdraft, down-
draft, and a high-turbulence mode utilizing a forced-draft blower. Fuels were nat-
urally dried commercial oak cordwood, commercial green pine cordwood, oven-dried
fir brands,, and naturally dried oak cut into reproducible triangles. Continuous mea-
surements of stack gases included O2, CO2, CO, NO, SO2, and total hydrocarbons
(FID) as an indication of the total organic species in the stack gases during batch
:ype operation. Several combustion modification techniques were identified which
lave an appreciable effect on emission factors and, therefore, can be developed and
applied to reduce emissions in consumer use. The more promising design modifica-
tions 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 tur-
bulence, and increasing the temperatures in the secondary burning regions of the
appliances.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATi Field/Croup
Pollution
Residential
Buildings
Combustion
Stoves
Wood
Flue Gases
Dust
Sampling
Carbon Monoxide
Nitrogen Oxides
Aromatic Polycyclic Hydro-
carbons
Pollution Control
Stationary Sources
Combustion Modification
Wood Stoves
Particulate
13 B
13M
21B
ISA
11L
11G
14B
07B
07C
118. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
101
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 U-73)
92
------- |