United States Office of Research and EPA-600/R-02-093
Environmental Protection Development M^/omKo
Agency Washington, DC 20460 NOVemoe
<>EPA Primary Particles
Generated by the
Combustion of Heavy
Fuel Oil and Coal
Review of Research Results
from EPA's National Risk
Management Research
Laboratory
Prepared for
Office of Research and
Development
Prepared by
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions leading to
a compatible balance between human activities and the ability of natural systems to
support and nurture life. To meet this mandate, EPA's research program is providing
data and technical support for solving environmental problems today and building a
science knowledge base necessary to manage our ecological resources wisely,
understand how pollutants affect our health, and prevent or reduce environmental risks
in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's
center for investigation of technological and management approaches for preventing
and reducing risks from pollution that threaten human health and the environment. The
focus of the Laboratory's research program is on methods and their cost-effectiveness
for prevention and control of pollution to air, land, water, and subsurface resources,
protection of water quality in public water systems; remediation of contaminated sites,
sediments and ground water; prevention and control of indoor air pollution; and
restoration of ecosystems. NRMRL collaborates with both public and private sector
partners to foster technologies that reduce the cost of compliance and to anticipate
emerging problems. NRMRL's research provides solutions to environmental problems
by: developing and promoting technologies that protect and improve the environment;
advancing scientific and engineering information to support regulatory and policy
decisions; and providing the technical support and information transfer to ensure
implementation of environmental regulations and strategies at the national, state, and
community levels.
This publication has been produced as part of the Laboratory's strategic
long-term research plan. It is published and made available by EPA's Office of
Research and Development to assist the user community and to link researchers with
their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/R-02-093
November 2002
Primary Particles Generated by the
Combustion of Heavy Fuel Oil and Coal
Review of Research Results from EPA's National
Risk Management Research Laboratory
Prepared by:
C. Andrew Miller
William P. Linak
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
Abstract
Researchers at the U.S. Environmental Protection Agency's (EPA's) Office of Research
and Development (ORD) have conducted a series of tests to characterize the size and
composition of primary particulate matter (PM) generated from the combustion of heavy
fuel oil and pulverized coal. These tests, conducted at ORD's National Risk
Management Research Laboratory (NRMRL) in Research Triangle Park, NC, burned
four heavy fuel oils and seven coals in three small combustion systems and measured size
distributions and composition of the particles formed in the combustion process. The
research found that, for heavy fuel oils, particle composition and size are dependent upon
the combustion environment in the combustor. In the coolest test unit, unburned carbon
in the fly ash was very high (70-90% by mass), and PM mass emissions were also high,
with the majority of the mass being in the coarse (> 2.5 jim in aerodynamic diameter)
size range. In the hottest test unit, the PM mass emissions were approximately 50% of
those from the coolest unit, unburned carbon levels were approximately zero, and the
entire PM mass was in particles smaller than 1 jim in aerodynamic diameter. For coal,
the NRMRL research identified a previously unreported trimodal particle size
distribution, with an ultrafine (~ 0.1 jim in aerodynamic diameter) mode, a coarse (> 5
|im in aerodynamic diameter) mode, and a central mode of particles between 2 and 4 jim
in aerodynamic diameter. The central mode has a composition similar to the coarse
mode, but significantly different from the ultrafine mode. It is hypothesized that the
composition of the ultrafine coal particle fraction is similar to the fine particle fraction in
heavy fuel oil, which has been shown to generate increased toxicological responses
following pulmonary exposure in laboratory animals. These findings may have
implications for future PM control strategies, given the influence of operating conditions
and the water-soluble composition of the smallest particle fractions of PM from metal-
bearing fossil fuels
-------
Table of Contents
Page
Abstract ii
List of Figures iv
List of Tables iv
Executive Summary 1
Findings 1
Implications 2
Future Directions 3
Section 2, Introduction 5
Section 3, Background 7
Toxicological Studies 7
Ambient Particles 8
Sources of PM 10
Previous Studies Characterizing Fossil Fuel Particles 12
Particle Formation in Coal 12
Particle Formation in Fuel Oil 13
Section 4, Particle Characterization Research atNRMRL 15
Experimental Equipment 15
Fuels 18
Instrumentation and Analytical Methods 20
Test Conditions 23
Computer Models 24
Section 5, NRMRL Research Results 25
Primary PM from Combustion of Heavy Fuel Oils 25
Phase I Tests 25
Phase II Tests 26
Data from Other Studies 30
Further Evolution of Research Directions 30
Primary PM From Combustion of Coals 31
Model Predictions 35
Chemical Equilibrium Modeling 35
Aerosol Nucleation and Coagulation Modeling 35
Section 6, Summary and Implications 37
Summary 37
Implications 37
Section 7, Future Directions 39
Section 8, Publications from NRMRL's Combustion-Generated PM Research Program 41
Section 9, References 43
in
-------
List of Figures
Figure Page
1. Composition of ambient PM25 in the Eastern and Western U.S 9
2. Major sources of directly emitted PM2 5 in the U.S. in 1998 10
3. 1999 consumption of coal and residual oil in the U.S. by sector 11
4. Schematic of NRMRL's firetube package boiler 16
5. Schematic of NRMRL's Rainbow horizontal refractory lined furnace 17
6. Schematic of NRMRL's package boiler simulator (PBS) 18
7. Schematic of NRMRL's down-fired innovative furnace reactor (IFR) 19
8. Schematic of NRMRL dilution sampling system 23
9. Particle size distributions for No. 6 fuel oil in the North American Package
Boiler and Rainbow furnace 27
10. Stack concentrations of PM as a function of fuel ash content for the four fuel oils burned in
the firetube boiler and fly ash loss on ignition for the four fuel oils burned in the firetube
boiler and the Rainbow refractory lined furnace 28
11. Particle size distributions for three coals burned in NRMRL tests 32
12. Stack concentrations of PM as a function of fuel ash content and fly ash
loss on ignition for the seven coals burned in the innovative furnace reactor 33
List of Tables
Table Page
1. Composition of fuels burned during NRMRL's PM research studies 21
2. Nominal test conditions for NRMRL's combustion tests 23
3. PM emission rates and emission factors and measures of unburned
carbon for oils burned in NRMRL tests 25
4. PM emission rates and emission factors and measures of unburned carbon
for coals burned in NRMRL tests 31
IV
-------
Section 1
Executive Summary
In 1997, the U.S. Environmental Protection Agency (EPA) promulgated revisions to the
National Ambient Air Quality Standards (NAAQS) for particulate matter (PM) to add
primary standards for ambient concentrations of PM smaller than 2.5 jim in aerodynamic
diameter (PM^.s). The motivation for this revision was the body of epidemiological
studies that associated elevated concentrations of ambient PM2.5 with increases in human
mortality and morbidity. To support the implementation of the revised standard, the
National Risk Management Research Laboratory (NRMRL) began a series of tests to
characterize the particles directly emitted from the combustion of heavy fuel oils and
pulverized coals. This research was initially intended to provide data for improved
emission inventories. During the course of the program, the characterization data for
inventory improvement has been generated, and a portion of the research has also
evolved to include a series of toxicological studies in collaboration with EPA's National
Health and Environmental Effects Research Laboratory (NHEERL).
The main focus of this report is the physico-chemical characterization of primary PM
from fossil fuel combustion. This work was conducted at NRMRL's combustion
research facilities in Research Triangle Park, NC, using several of the Laboratory's
research combustors. Using state of the art measurement systems, NRMRL researchers
have identified several characteristics of combustion-generated primary particles and
have associated these characteristics with fundamental particle formation mechanisms to
allow these results to be extended to other high temperature systems. The key findings of
this research are provided below.
The collaborative NRMRL-NHEERL research is addressed briefly here, but is currently
on-going and outside the scope of this report.
Findings
The key findings from NRMRL's research to characterize primary PM generated from
the combustion of heavy fuel oil and pulverized coal are:
• Modest changes in combustion conditions, such as slight reductions in excess air and
reducing oil feed temperature, do not significantly change the mass emissions of PM
from heavy fuel oil combustion.
• Significant changes in combustion conditions due to different combustor designs result
in substantial changes in PM characteristics and mass emissions when burning heavy
fuel oil.
• Incomplete combustion of the carbon in the heavy fuel oil results in the formation of
large (>20 jim diameter) particles that contribute substantially to the total PM mass
emission rates for certain boiler designs.
-------
• The submicron (<1 jim in aerodynamic diameter) particles contain metals and sulfur
that are largely in water soluble form, in contrast to the supermicron (>1 jim in
aerodynamic diameter) fraction of particles that contains less soluble thiophenic sulfur
and less soluble metal species.
• A previously unreported particle mode between 0.8 and 2.0 jim contributes
considerably to the mass emissions of PM2.s from the combustion of pulverized coal.
• The composition of particles in the "fine fragmentation" mode between 0.8 and 2.0
|im is similar to that of coarse mode particles from pulverized coal combustion. This
suggests that these 0.8-2 jim particles are formed by coarse ash particles that fragment
during the combustion process. The composition also suggests that these particles will
have metals with relatively low water solubility and possibly relatively low
bioavailability.
• The ultrafine (<0.1 jim in aerodynamic diameter) fraction of coal-generated particles
has composition similar to that of the ultrafine mode of particles from heavy fuel oil
combustion, suggesting that the pulmonary toxicity of ultrafine particles from coal
combustion may be similar to the heavy fuel oil ultrafine particles.
Implications
Particle composition varies considerably with size, with potentially significant
implications for particle toxicity. NRMRL's research, consistent with previous work,
has identified significant variation in particle composition as a function of particle size.
The submicron (especially the ultrafine) particle fractions were consistently composed of
metal and other sulfates that are highly water soluble, meaning that they are more likely
to interact with biological tissues when inhaled. The formation mechanisms that generate
these particles govern both size and composition. Therefore, the question of whether size
or composition plays a dominant role in causing adverse health effects may be
inappropriate, since the two (size and composition) cannot be separated in real-world
systems. It may be more relevant to focus on particles with similar formation processes
rather than on particles with specific characteristics that cannot be naturally separated. In
short, multiple mechanisms are involved in the formation of PM2.5 from combustion
sources, and these mechanisms will directly impact the role of particles in causing
adverse health effects.
The design and operation of a source can result in significant changes in the
physico-chemical and toxicological characteristics of the particles generated. This
finding has scientific and, potentially, regulatory implications. In the context of
toxicological research, more emphasis needs to be placed on how particle samples are
generated and collected. PM toxicological studies often do not adequately describe the
source characteristics (including system design), the conditions at which the system was
running, or the methods used to extract particle samples. PM from heavy fuel oil, as
demonstrated here, is not a homogeneous material with consistent characteristics. In
order to develop an adequate understanding of the health effects associated with PM and
-------
the causal mechanisms leading to those effects, studies must provide enough information
to allow the reader to understand the representativeness of the particles used in the tests.
From a regulatory perspective, these results may eventually be used to identify source
types that have the potential for emitting particles having characteristics closely
associated with health risk. As additional understanding is gained regarding links
between health effects and particle characteristics, it may someday be possible to develop
regulatory approaches to reduce PM emissions from a more limited range of source types,
or to minimize emissions of particles with certain characteristics. If enough information
on adverse health effects due to water-soluble ultrafme particles is developed, for
instance, limits on emissions of these particles may replace or supplement limits on total
mass emissions.
Future Directions
NRMRL's research to characterize primary particles from the combustion of pulverized
coal and heavy fuel oils has been completed. There remain numerous research
opportunities in the broader area of combustion-generated particles, which will be
followed as appropriate. Immediate efforts will focus on characterizing emissions from
small off-road diesel engines, which will involve more organic analysis as compared to
the significant levels of inorganic analysis conducted for the previous studies. Several
efforts associated with the collaborative studies linking the physico-chemical properties
of combustion-generated particles to measures of toxicity during direct inhalation
exposures will continue. This area of work shows considerable promise for generating
data that can be used to better understand the links between sources and adverse health
effects.
The capabilities developed during the work described in this report will continue to
generate state of the art research. NRMRL's flexible and unique facilities, in
combination with ORD's on-site expertise, provide unique opportunities for leading edge
multidisciplinary research in PM, air toxics, and other air programs.
-------
(Blank Page)
-------
Section 2
Introduction
In 1997, the U.S. Environmental Protection Agency (EPA) promulgated revisions to the
National Ambient Air Quality Standards (NAAQS) for particulate matter (PM) to add
primary standards for ambient concentrations of PM smaller than 2.5 jim in aerodynamic
diameter (PM^.s). The motivation for this revision was the body of epidemiological
studies that associated elevated concentrations of ambient PM2.5 with increases in human
mortality and morbidity. Although the specific mechanisms of health damage that
resulted in these associations had not been demonstrated, numerous toxicological studies
provided considerable evidence for several plausible mechanisms. Several of the particle
characteristics hypothesized to cause at least part of these adverse health effects are
typical of particles formed during the combustion of metal-containing fossil fuels such as
coal and heavy fuel oil.
As part of the broader effort by EPA's Office of Research and Development (ORD) to
provide scientific tools and information to support implementation of the new standard,
the National Risk Management Research Laboratory (NRMRL) initiated a study to more
fully characterize the primary particles generated from fossil fuel combustion. NRMRL's
facilities and expertise developed for studies of waste incineration and hazardous air
pollutant characterization provided an opportunity for ORD to quickly generate data on
emissions rates, particle size distributions (PSDs), and composition as well as
information on particle formation mechanisms in fossil fueled combustion systems.
Based on available data, a relatively modest program was anticipated to develop data on
PM from fossil fuels, beginning with heavy fuel oils and later moving to coals.
As the program developed, opportunities for collaboration with ORD's National Health
and Environmental Effects Laboratory (NHEERL) and unexpected results combined to
stretch the program beyond the originally anticipated scope. The research has met its
original goals, and the evolution of the program has resulted in clearer insights into
particle formation processes while expanding ORD's capabilities to address complex,
multidisciplinary environmental problems.
This document will provide a summary of the scientific motivations for this research,
briefly describe how the work was conducted and the quantitative results were obtained,
discuss the broader significance of these results, and identify some potential future work
that may be derived from these efforts.
-------
(Blank Page)
-------
Section 3
Background
The fundamental motivation for this work is the epidemiological evidence that associates
adverse human health effects with increased levels of ambient PM concentrations, and
EPA's promulgation of a NAAQS for PM2.5. A number of studies have linked increases
in ambient concentrations of PM to both short-term (acute) and long term (chronic)
adverse human health effects, including premature death. These epidemiological studies
examined statistics on mortality and morbidity as measured by data such as hospital
admissions and sought to identify possible correlations with ambient air pollution data.
In general, the epidemiological studies concluded that there is, in fact, a direct correlation
between ambient fine PM concentrations and increases in human mortality and morbidity
(Dockery et al. 1993, Burnett et al. 1995, Schwartz and Morris 1995, Lippmann et al.
2000, Samet et al. 2000). These studies were summarized by EPA (Bachmann et al.
1996, U.S. EPA 1996) and reviewed by EPA's Clean Air Science Advisory Committee
(CAS AC), both of which determined there was sufficient evidence linking ambient fine
PM concentrations and acute adverse cardiopulmonary health effects to warrant a
revision of the previously existing standard (Wolff 1996). In July 1997, EPA revised the
PM NAAQS to add to the existing standards for PM smaller than 10 |im in aerodynamic
diameter (PMio) two primary standards for PM2.5, a 65 |ig/m3 24-hour average and a 15
|ig/m3 annual mean, while retaining the previous 24-hour and annual PMio standards of
150 and 50 |ig/m3, respectively (Federal Register 1997). Although the PMio standards
were challenged in court and remanded to EPA for modification (U.S. Supreme Court
2001), the fundamental purpose of the standards—to reduce ambient concentrations of
specific size fractions of PM for the purpose of protecting human health and the
environment—has not changed.
lexicological Studies
Numerous theories have been proposed to explain the mechanisms by which exposure to
ambient particles causes adverse health effects. Among the particle characteristics that
have been suggested as playing significant roles in the causal mechanisms are total
particle mass; particle acidity, morphology, number, and/or size; presence of biogenic
materials, organic compounds, oxidants, or transition metals; or interaction with other
pollutants in the atmosphere. Studies of the effects of these different particle character-
istics have been, and are currently being, conducted in clinical human exposure studies,
studies of laboratory animals exposed to particles with different characteristics (in vivo
studies), and biochemical studies of the response of cells to different compounds found in
PM2 5 (in vitro studies). Of particular interest here are a series of studies that examined
the effects of PM generated by the combustion of fossil fuels in stationary sources.
Research conducted at NHEERL showed that laboratory animals exposed to fine PM
generated from the combustion of fossil fuels demonstrated significant adverse health
-------
impacts. This research concluded that a major factor leading to these adverse health
impacts was the amount of water-soluble transition metals such as copper (Cu), iron (Fe),
nickel (Ni), vanadium (V), and zinc (Zn) present in the particles (Dreher et al. 1996a,
1996b, 1997; Hatch et al. 1982, 1985). Other, more recent, work has also indicated that
pulmonary exposure to transition metals leads to adverse health effects through a number
of different mechanisms including changes in cardiac function (Lighty et al. 2000,
Fernandez et al. 2000, Godleski et al. 2000).
Particle size has also been proposed as playing a significant role in causing adverse health
effects. Particles smaller than 0.1 jam in aerodynamic diameter (ultrafme particles) can
have a greater biological effect than an equal mass of larger particles of the same
substance. Ultrafme particles can be deposited deep in the lungs where they may be
retained. Furthermore, ultrafme particles can pass through cells lining the lung, and their
higher surface area (compared to an equal mass of larger particles) can result in a faster
release of toxic compounds that may be associated with the particles (Oberdorster et al.
1995, Kuschner et al. 1997, Baggs et al. 1997).
Health damage has also been found to occur at significantly higher rates after exposure to
particles on which acid or organic compounds have condensed compared to damage
caused by exposure to any of the components by themselves. Particles on which sulfuric
acid had condensed caused the same effect on lung capacity as a ten-fold higher dose of
sulfuric acid alone in the form of a respirable mist. The level of this effect also depended
upon the type of particle, since the acid could react with particle constituents to mitigate
the effects (Amdur et al. 1986).
In summary, substantial evidence indicates that the particles containing transition metals,
ultrafme particles, or acid aerosols are of importance to the issue of PM-related health
effects. Each of these characteristics is associated with at least a fraction of the particles
formed by the combustion of heavy fuel oils and coals.
Ambient Particles
Ambient particles are a mix of primary and secondary particles. Particles formed in
combustion processes are considered to be primary because they are emitted directly
from the exhaust in particulate form. Primary particles also include those particles
formed from the physical nucleation or condensation of compounds that are in the vapor
phase in the exhaust, but quickly nucleate or condense at ambient temperatures. In
contrast, secondary particles are formed when vapor phase compounds chemically react
in the ambient atmosphere to form particles. These particle-forming reactions often
occur many miles downwind of the source of the precursor gases.
A majority of the fine particle mass in the ambient atmosphere is in the form of
secondary nitrates and sulfates formed from precursor gases, sulfur dioxide (862) and
nitrogen oxides (NOx). The percentage of 862 and NOx that form solid secondary
-------
particles varies by location and is influenced by factors such as local meteorology and
composition of the air mass (including presence of water as either humidity or droplets,
oxidants such as ozone, ammonium ions, and other pollutants or biogenic compounds). A
considerable fraction (25-50%) of the total ambient PM mass across the United States is
composed of these compounds (U.S. EPA 1996). Other constituents include organic
carbon (condensed organic compounds and secondarily formed organic aerosols),
elemental carbon (soot or other unburned carbon from combustion sources), particles of
geological origin such as wind-blown dust from agricultural or construction activities,
and inorganic compounds such as trace metals from the combustion of metal-containing
fuels. Figure 1 illustrates representative compositions of ambient PM2.5 in the Eastern
and Western United States (U.S. EPA 1996).
rates
Elemental Carbon
Crustal
Organic Carbon
Sulfates
Eastern U.S.
Nitrates
Sulfates
Elemental Carbon
Organic Carbon
Crustal
Western U.S.
Figure 1. Composition of ambient PM25 in the Eastern (top) and Western (bottom) United States (U.S. EPA
1996).
-------
Generally, ambient concentrations of primary particles from fossil fuel combustion are
very low. Typical fractions of transition metals in ambient particle samples are 5% or
less of the total sample mass. Nevertheless, these particles remain of interest for at least
two reasons. First, it is unclear what the effects of long-term exposures to ambient
concentrations of these compounds are, given the evidence that they are bioactive (based
on testing at much higher concentrations than in the ambient environment). Second, it is
not clear what the effects of these compounds may be in combination with sulfates,
nitrates, and other ambient PM constituents.
Sources of PM
Much of the current information on sources of primary PM2.5 is derived from existing
emission inventories for PMio combined with data on size distributions to estimate PM2.5.
Direct measurement of PM2.5 emissions is, therefore, important to improving the
accuracy of inventory estimates.
Figure 2 shows the relative amounts of directly emitted PM25 in the U.S. in 1998. Wood
>. 400-
OT
3 350
300-
250-
to
200-
V)
150-
CD
CL
00
100-
50
INI
IIII IIII
ii ii ii ii ii ii innnnn
O
*•*=
to
E
I
8
T3
O
O
e:
CD
0)
CC
(D
tn
CD
.1
D)
T3
CO
O
tr
05
CD
CD
O
g
Q_
"co
O)
c
k_
3
CO
CD
Q_
O
CD
"w
or
O
JC
LU
i
c
o
to
3
"E
o
O
M
3
c
.a
E
o
O
® O
O ^£
•« £
ra
o
a:
o
c
o
E
o
O
"ai
"o
M
to
to
O
ce
O
(O
O
a,
CO
nj
o
fg
CD
«
-------
combustion from the residential sector is the largest source of primary PM2.5 (U.S. EPA
2000). Point source emissions of primary PM2 5 from fossil fuel combustion remain of
considerable interest even though they do not emit the greatest amount of PM2.5 mass. Of
most interest within this broad category are sources that use coal or heavy fuel oils. This
is because of the presence of metals in the emissions and the relatively large size of the
sources. Control of these large stationary sources is often more cost effective than
control of smaller, more dispersed sources.
Coal and heavy fuel oils are used primarily in industrial and utility boilers to generate
steam for process heat and electricity generation (see Figure 3). Coal use is heaviest in
Texas and the Ohio River Valley, but is much more evenly distributed across the country
than is heavy fuel oil use. Over 90% of coal is consumed by utilities in the generation of
electricity, with a much smaller amount of electricity being generated from the
combustion of heavy fuel oil. Heavy fuel oil is also used in utility and industrial plants
and to a smaller extent in the transportation sector, almost entirely in ocean-going ships
(U.S. Department of Energy 2000a, b, c). Heavy oil use largely occurs in states where
the fuel can be transported by ship or barge or where there are local oil resources that can
be used near the production sites (U.S. Department of Energy 2000b).
Residential/Commercial
Industrial
Utility
Vessels
Figure 3. 1999 consumption of coal and residual oil in the U.S. by sector (U.S. Department of
Energy, 2000a, b).
11
-------
Previous Studies Characterizing Fossil Fuel Particles
There has been considerable work to characterize PM generated by the combustion of
fossil fuels, particularly pulverized coal (Linak and Wendt, 1994). The focus of much of
this previous work has been the prediction of ash composition for control of corrosion,
fouling, slagging, and other adverse effects on combustor materials and heat transfer
surfaces (Raask 1985). The fundamental particle formation and evolution mechanisms
described in these studies are directly applicable to environmental investigations.
Particle Formation in Coal
Particles are generated during the combustion of pulverized coal through several different
mechanisms that determine not only particle size but also the composition of the
particles. In coal, 90-98% of the particle mass exiting the furnace is inorganic material
that was present in the original fuel. The inorganic material may exist in the coal in
several ways, ranging from tightly bound within the combustible organic material to
distinct particles of inorganic matter (Gluskoter 1978). In the high temperatures of the
combustion zone, the inorganic material may remain relatively intact, melt and form
relatively large (1-30 jim) liquid droplets, or vaporize. The path depends upon the
specific compound, how it is present in the fuel, the temperature, and the concentrations
of other compounds in the combustion zone. In general, approximately 95% of the
inorganic material present in coal does not vaporize.
Work by Gallagher et al. (1990), Flagan and Taylor (1981), Markowski et al. (1980),
Neville et al. (1982, 1983), Okazaki et al. (1983), and Quann et al. (1982) typify the
investigations that identified the governing mechanisms of particle formation in coal
combustion systems and their effect on particle size and composition, particularly trace
metal composition. These mechanisms are discussed briefly below.
Inorganic species that vaporize may undergo complex chemical reactions depending on
the temperature and presence of other species. As a result of chemical reaction or falling
temperatures, partial pressures approach and exceed critical supersaturation pressures,
causing the inorganic species to either homogeneously nucleate and grow to form a
submicron aerosol in the range of 0.01 to 0.05 |im in diameter or to heterogeneously
condense on existing particle surfaces. Because submicron particles typically offer the
major fraction of the available surface area, condensation tends to result in an enrichment
of the inorganic vapor phase elements in the submicron particle fraction. Nucleation and
coagulation also enrich the submicron particle fraction in vaporized elements, since these
processes form particles that are smaller than 1 jim in diameter. Thus, the submicron and
ultrafme particle size fractions can contain significantly higher percentages of those
elements that pass through a vapor phase, compared to the larger particles that are not
formed by nucleation or condensation processes.
Because growth by collision and coagulation with other particles is directly dependent on
the number concentrations of the available particles, nuclei tend to grow very quickly for
12
-------
a short time (1 to 5 s) and then, as number concentrations fall, particle growth slows
considerably, causing the aerosol to accumulate into a mode between 0.05 to 0.5 jim
diameter. Additional coagulation between nuclei and the larger accumulation mode
particles does not noticeably affect the particle size distribution (PSD) because of the
relatively small incremental mass added during each coagulation event.
In general, then, the results of previous investigations demonstrated that inorganic
material, especially metals, can partition to different particle sizes due to the mechanisms
governing particle formation in high temperature reactive processes such as coal
combustion.
Particle Formation in Fuel Oil
Although the fundamental particle formation processes during the combustion of heavy
fuel oils are the same as those for pulverized coal, there are distinct differences between
the two fuels. In contrast to coals, oils do not typically contain significant extraneous or
included mineral matter. The metals in heavy fuel oils are generally inherently bound
within the organic molecule, which may be the case for only a small portion of the metals
in higher rank coals. Unlike coal, interactions between volatile metal species and non-
volatile minerals within the heavy fuel oil droplet are much less likely in heavy fuel oils.
Typical fuel oils contain Fe, Ni, V, and Zn, in addition to aluminum (Al), calcium (Ca),
magnesium (Mg), silicon (Si), and sodium (Na). Transition metals [Fe, manganese (Mn),
and cobalt (Co)] and alkaline-earth metals [barium (Ba), Ca, and Mg] may also be added
for the suppression of soot (Bulewicz et al. 1974) or for corrosion control (Feldman
1982).
Hersh et al. (1979), Piper and Nazimowitz (1985), and Walsh et al. (1991) showed that,
in contrast to pulverized coal, the majority of the sampled fly ash mass from residual fuel
oil combustion in power plants is likely to lie below 1 |im in diameter, although larger
particles can form with poor carbon burnout. Furthermore, Walsh et al. (1991) have
shown that Fe, Mg, and Ni are concentrated at the center of the submicron particles,
while Na and V are associated with a "halo of sulfate residue." Bacci et al. (1983) found
substantial enrichment of both Ni and V in the submicron particle size fraction of samples
collected at a large oil-fired power plant. Walsh et al. (1994) indicated similar behavior
for Ni and V. Again, these results demonstrate the compositional differences between
coarse and fine PM fractions due to the mechanisms governing particle formation in
fossil fuel combustion systems. However, none of these data showed detailed PSDs in
the submicron range or provided in-situ measurements of the coarse (supermicron) mode.
13
-------
(Blank Page)
14
-------
Section 4
Particle Characterization Research at NRMRL
With these results representing the state of the science at the inception of the NRMRL
PM program, NRMRL researchers began a project to characterize PM generated by the
combustion of heavy fuel oil. Heavy fuel oil was chosen because less research had been
done on this fuel than on coal and because of its high transition metal content. In
addition, heavy fuel oil is also often easier to burn than pulverized coal, and the
fundamental mechanisms governing particle formation could be more easily studied than
if coal were used. Even though heavy fuel oil use is much lower than coal use, the
broader applicability of findings associated with particle formation in systems burning
fuels containing transition metals (including both heavy fuel oil and coal) made heavy
fuel oil an attractive first step in the research program.
Thus, initial work at NRMRL was proposed to provide accurate information on the mass
emissions rate and size-specific composition of PM generated during the combustion of
fossil fuels, with the intent of testing heavy fuel oil followed by pulverized coal. As the
project progressed, the research expanded to incorporate a collaborative effort between
NRMRL and NHEERL and incorporated data from other test programs, as appropriate, to
allow the broadest possible evaluation of PM generated from fossil fuel combustion.
This section will discuss the equipment and instrumentation used in the research program
and the evolution and results of the characterization studies.
Experimental Equipment
NRMRL's experimental work was conducted at EPA facilities in Research Triangle Park,
NC, using several different combustion systems and a range of instruments and methods
for measuring particle size distributions and composition. NRMRL's variety of
combustion equipment allowed the research to generate data simulating the behavior of a
range of commercial and industrial systems using several different liquid and solid fuels.
Four combustors were used in NRMRL's PM characterization experiments: a firetube
package boiler, a refractory-lined horizontal tunnel furnace, a waterwall package boiler
simulator, and refractory-lined vertical (down-fired) tunnel furnace. Each unit has
characteristics that simulate different types of industrial combustion systems and that
result in somewhat different particle characteristics.
Initial experiments were conducted using NRMRL's North American firetube package
boiler (NAPB), which is a practical heavy fuel oil combustion unit of a design used in
commercial, institutional, and small industrial applications. A schematic of the NAPB is
shown in Figure 4. In a firetube boiler, the hot combustion gases generated by the flame
pass through tubes surrounded by water, which then generates low pressure steam for use
in space and water heating and light process needs. The NAPB is rated at 2.5 x 106
Btu/hr (732 kW), and is capable of firing natural gas and No. 2 through No. 6 fuel oils.
15
-------
Sampling location lor
Methods 5 and 60,
SMPS, APS,
inertlal impaclors
To air pollution
control system
Sampling location tor
dilution sampler, in-
situ light scattering
-C
Sampling location
for gas samples
o
o
Stack
Third pass:
20 tubes, 6.4
cm (2.5 in) I.D
Boiler end
view
Natural gas
First pass:
1 tube, 46 cm
(18 in) I.D.
Second pass:
24 tubes, 6.4
cm (2.5 in) I.D.
APS - Aerodynamic particle sizer
P - Pressure
SMPS - Scanning mobility particle sizer
T-Temperature
Figure 4. Schematic of NRMRL's firetube package boiler
The NAPB is equipped with several sampling ports located at the exit of the boiler.
Systems of this design are characterized by a high rate of heat transfer from the
combustion gases to the water, resulting in significant amounts of unburned carbon in the
boiler exhaust as temperatures rapidly drop and quench the combustion reactions.
At the opposite end of the spectrum in terms of combustion gas temperatures is
NRMRL's "Rainbow" horizontal tunnel furnace, shown schematically in Figure 5. This
unit is less than one-tenth the size of the NAPB as measured by its fuel input rate of
200,000 Btu/hr (59 kW). The Rainbow furnace is a refractory-lined research combustor
designed to simulate the time/temperature and mixing characteristics of large gas- and
oil-fired combustion systems. Fuel oil and combustion air are introduced into the burner
section through a variable-air swirl burner. Gas and aerosol samples are taken from the
stack, where the temperature is approximately 745 °F (670K). This system generates
combustion gas temperatures that are more representative of large industrial and utility
16
-------
To air pollution control system
Stack
— - — Gas sampling port
— In-situ particle sizer sampling port
-Aerosol sample port
Moveable-block
burner
Natural
gas
Refractory sections
Atomizing air
Transport air '—Combustion air
Figure 5. Schematic of NRMRL's Rainbow horizontal refractory lined furnace.
boilers compared to the temperatures in the NAPB. The Rainbow furnace is also capable
of burning natural gas and a range of fuel oils.
In between the NAPB and the Rainbow in terms of combustion gas temperatures is
NRMRL's Package Boiler Simulator (PBS), illustrated in Figure 6. The PBS simulates
the performance of a small watertube boiler, in which the combustion gases pass through
a large volume surrounded by tubes containing the fluid to be heated (usually water).
The PBS, like the NAPB and the Rainbow, can burn natural gas and several types of fuel
oils, including oil-water emulsified fuels. The PBS has a maximum fuel input rate of
3xl06 Btu/hr (880 kW) and generates combustion gases that are hotter than those in the
NAPB, but not as hot as those in the Rainbow.
Coal combustion experiments were conducted using a down-fired, refractory-lined
furnace referred to as the Innovative Furnace Reactor (IFR) (see Figure 7) operated at
170,000 Btu/hr (50 kW). The IFR, although small, is designed to simulate the time-
temperature profiles characteristic of large coal-fired boilers. The IFR is capable of
burning natural gas and coal that has been pulverized to the same size range as used in
17
-------
Fuel Pump
To Flue Gas Cleaning System
Extractive
Sampling Ports
/\ Combustion
) Air Blower
Continuous
Emission
Monitors
Burner Transition
Section Section
Data
Acquisition
System
Figure 6. Schematic of NRMRL's package boiler simulator (PBS).
utility boilers. The IFR is refractory lined and generates combustion gases that are
slightly cooler than large utility boilers, although levels of unburned carbon are similar to
those found in full-scale utility units (Linak et al. 2000a).
Fuels
Several different fuels were used in the course of NRMRL's PM characterization tests.
In addition, PM size distributions and/or compositions were measured during other
NRMRL research programs characterizing emissions from different fuels, and results
from those programs provide additional insight into the mechanisms governing formation
of particles during fossil fuel combustion. The use of the different fuels provided
NRMRL researchers the opportunity to identify both the differences and similarities
among particles from different fuels.
Fuel oils are typically easier to burn than coals, due mostly to their higher volatility, their
lack of significant inorganic content, and their ability to readily form small fuel particles.
As fuel oils become heavier, they become more viscous, less volatile, and more difficult
to effectively atomize. To efficiently burn heavy fuel oils, they must be heated, and
higher pressures are required, often along with compressed air or steam, to form droplets
small enough to burn to completion.
18
-------
Variable swirl ^
coal burner-^_ o .^
- Coal/transport air
Axial air
Tangential air
Refractory sections
20 cm (8 in)
Cooling coil
PM sampling locations
o o
o
o
To baghouse & scrubber
Ash pot
Figure 7. Schematic of NRMRL's down-fired innovative furnace reactor (IFR).
One No. 5 fuel oil and three No. 6 fuel oils were tested as part of NRMRL's initial PM
characterization tests. The No. 6 fuel oils were chosen to represent a range of sulfur and
metal contents, and the No. 5 fuel oil was chosen to represent a relatively high sulfur fuel
oil that could be used to test the effects of changes in the combustion conditions typically
experienced on a single system. (Miller et al. 1998).
In addition to the fuel oils, several U.S. coals were also burned during NRMRL's PM
characterization tests. Coals were chosen to represent a range of different regions of the
U.S. as well as different sulfur and metal contents. The experiments burned West
Kentucky, Pittsburgh, West Virginia (Pocahontas), Ohio, and two Utah bituminous coals;
Montana and Power River Basin (Wyodak) subituminous coals; and a North Dakota
19
-------
lignite. Compositions, including trace metal content, of all fuels used in the NRMRL test
program are shown in Table 1.
Two other test programs conducted by NRMRL also measured PM characteristics during
testing to evaluate the performance of liquid fossil fuels with different properties from
conventional fuel oils. Several emulsified fuel oils and two samples of Orimulsionl were
burned, and PM composition and size distributions were measured. Emulsified fuels
contain at least two distinct materials that do not physically mix and are blended to form
an emulsion. The most common emulsified fuels are water-in-oil emulsions in which
small discrete droplets of water are suspended in a continuous phase of fuel oil. A more
recent emulsified fuel that is being used internationally is Orimulsion, which is a
bitumen-in-water emulsion in which the water is the continuous phase that contains
discrete droplets of bitumen, a heavy naturally occurring petroleum fraction (Miller and
Srivastava 2000, Miller et al. 2001). The presence of water in emulsified fuels has been
shown to reduce emissions of total PM and NOx under proper operating conditions, but
can result in the formation of smaller particles (Miller 1996, 1998).
Each of these fuels was burned under closely controlled conditions that simulated actual
industrial or utility plant operation. PM and other pollutants were measured, with the
emphasis on PM, using one or more of the measurement methods noted below. In most
cases, the tests were conducted several times at the same condition to evaluate variability
in the measurements. Data were collected and analyzed with the goal of determining
differences in PM characteristics among the different fuels and why such differences
occurred.
Instrumentation and Analytical Methods
NRMRL's PM characterization studies used a variety of instruments and methods to
measure combustion conditions and PM characteristics. Continuous emission monitors
(CEMs) were used to measure concentrations of combustion gases, including carbon
dioxide (CO2), carbon monoxide (CO), NOX, oxygen (O2), SO2, and total hydrocarbons
(THCs). Signals from the CEMs were sent to a computerized data acquisition system
(DAS) for continuous recording. In some cases, other process parameters, such as flue
gas temperature, were also recorded using the DAS. All flue gases were treated
downstream of the measurement points using air pollution control equipment for
reduction of pollutant emissions.
Orimulsion is an emulsified fuel produced from Venezuelan bitumen, water, and a trace
amount of additives.
20
-------
Table 1. Composition of fuels burned during NRMRL's PM research studies.
Content
C, %
H, %
N, %
S, %
0°, %
Mois-
ture, %
Ash, %
Fuel Oils
No. 5
86.36
10.82
0.33
1.73
0.34
0.35
0.07
LowS
No. 6
85.99
11.29
0.43
0.53
1.24
0.50
0.02
Medium
SNo. 6
86.46
10.98
0.43
0.93
0.67
0.50
0.03
HighS
No. 6
85.45
10.35
0.35
2.33
0.92
0.50
0.10
Emulsified Fuels
No. 6
77.83
10.16
0.24
1.70
10.07
9.00
0.096
No. 2
57.40
8.77
0.48
0.009
2.42
30.93
0.003
Orimul-
sion 400
58.12
7.14
0.17
2.23
3.35
28.92
0.07
Bituminous Coals
Ohio 5,6,7
Blend
71.00
4.81
1.37
2.62
8.09
2.33
9.69
Pitts-
burgh 8
72.57
4.68
1.36
2.78
7.35
2.60
8.66
Utah
65.10
4.58
1.36
0.90
12.40
5.97
9.69
West Ken-
tucky
65.28
4.25
1.39
2.89
11.71
6.97
7.51
Subbituminous
Coals
Mon-
tana
57.50
3.52
0.91
0.74
15.56
11.36
10.41
Wyodak
(PRBa)
51.19
3.64
0.72
0.32
12.29
25.81
6.03
Lignite
North
Dakota
38.57
2.60
0.42
0.63
12.52
35.88
9.21
Higher Heating Value, Btu/lb
10,230
10,460
10,390
10,150
16,604
12,786
12,596
12,685
Transition Metal Content, (ig/c
Cu
Fe
Ni
V
Zn
4
50
34
180
39
0.56
23
17
35
4.11
0.78
19
22
70
3.7
3.5
21
30
220
74
NAC
NA
59
440
NA
0.19
56
0.19
2.77
1.0
<0.005
22
59
262
0.37
NA
NA
NA
NA
NA
12,978
11,289
11,291
9526
8869
6392
2.64
8060
2.19
ND
4.32
3.37
2000
ND°
4.25
ND
3
9210
6.35
13
30.8
2.89
2560
2.39
4.55
ND
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Powder River Basin
' Calculated by difference
: Not available
' Not detected
-------
PM emissions were sampled using a variety of techniques. Standard EPA Method 5 and
Method 60 sampling and analytical procedures were used to determine total PM and
metal concentrations. NRMRL also uses a number of techniques for measuring particle
size distributions, including a scanning mobility particle sizer (SPMS), a time-of-flight
aerodynamic particle sizer (APS), a light-scattering method, and three different inertial
impaction systems. These systems rely on different techniques for measuring particle
size distributions, measure different size ranges, and provide data over different sampling
time spans. This variety of approaches gives NRMRL researchers valuable internal
verification of results and data over a broader range than would be possible with only a
single instrument. The different instruments also allow NRMRL researchers to collect
size-segregated samples for later analysis, providing further detailed information on the
characteristics of particles that can shed light on particle formation mechanisms (Linak et
al. 2000a).
Using the SMPS, APS, and light-scattering instruments, it is possible to obtain near-real
time particle measurements over a range of 0.01 to 100 jim diameter. Because most
current interest is in the smaller particles, only the SMPS and APS are typically used,
providing particle size data over a 0.01 to 20 jim diameter range (0.01 to 1.0 jim diameter
for the SMPS and 0.5 to 20 |im diameter for the APS) (Linak et al. 2000a).
Extracted samples were also collected using three different designs of cascade impactors,
including an eight-stage atmospheric pressure impactor, a ten-stage Micro-Orifice
Uniform Deposit Impactor (MOUDI), and a custom-made eleven-stage Berner-type low
pressure impactor. The use of different impactor designs provided a wider range of data
than would otherwise have been available. The atmospheric impactor collected adequate,
size-segregated mass to allow elemental analysis to be conducted on the different size
fractions, while the other two systems expanded the lower range of particle sizes that
could be collected.
Following the initiation of collaborative studies with NHEERL, samples were also
collected using a large, custom-made dilution sampler (see Figure 8). Samples were
extracted from the exhaust of the combustor being used and were then passed through a
cyclone designed to remove coarse (>2.5 jim) particles. To simulate the effect of a plume
mixing in the ambient atmosphere, this system then diluted the particle-laden flue gases
with filtered ambient air to uniformly cool the sample to approximately ambient
temperatures within about 3 s. Because of the relatively low surface area associated with
the coarse particle fraction, little change in particle composition or size distribution
occurred due to the removal of the coarse particles. The diluted and cooled particles were
then captured on a large (25.5 in diameter) Teflon coated glass fiber filter, which
provided adequate collection area for the relatively large samples needed for NHEERL
toxicity testing. As the program progressed, this sampler was also used to dilute flue gas
samples used in direct inhalation tests by removing the filter and directing the diluted and
cooled gas stream to inhalation exposure chambers.
22
-------
Non-
isokinetic
Sampling
'Probe
Filter
Flow
Sample Control
Blower
Dilution
Air
Intake
90 cfm
@ Pressure Gauge
(r) Temperature Sensor
Figure 8. Schematic of NRMRL dilution sampling system.
Test Conditions
For the majority of tests, the combustion conditions were designed to simulate operating
conditions typical of full-scale utility and industrial boilers. Most tests were conducted at
excess air levels that were as low as possible, consistent with acceptable CO and
unburned carbon in the ash. Several tests were conducted to evaluate the impacts of
reducing excess air levels or to modify the fuel oil spray parameters. These tests were
designed to determine whether slightly misoperated systems would significantly change
the characteristics of PM compared to systems that were operated closer to optimum
conditions. The greatest variation in combustion conditions occurred when the tests were
conducted on combustors of different design (see descriptions above). No tests were
conducted to simulate lowNOx combustion conditions such as overfire air injection,
flame staging, low NOx burners, or reburning. The nominal test conditions are presented
in Table 2.
Table 2. Nominal test conditions for NRMRL's combustion tests.
Test
Target Stack O2, % dry
Target SR
Oil Temperature, °C
Fuel Oils
No. 5 and No.
6, Baseline
3.5
1.2
120
No. 5, Low Oil
Temperature
3.5
1.2
77
No. 5, Low SRa and
Low Oil Temperature
2.5
1.1
77
Emulsified Fuels
No. 2
2.5
1.1
Ambient
Orimulsion
3.5
1.2
40
Pulver-
ized
Coals
3.5
1.2
NA
1 Stoichiometric ratio
23
-------
Computer Models
In addition to the experiments, two computer models were used to assist in understanding
the behavior of the fuels and their constituents. One model was designed to simulate the
formation mechanisms of particles formed from vaporized materials. This model was
used to predict the growth rate and size distribution of the particles, and these predictions
were then compared to experimental results. The other model was used to determine the
equilibrium concentrations of different chemical species formed by the fuels' trace metals
in the high temperature combustion zone. As with the previous model, these predictions
were compared to measurements, where possible, and were used to provide insight into
what species were likely to be present in the emitted PM. The chemical species present
in the PM are thought to be of importance with respect to the potential toxicity of the
particles, as noted above.
24
-------
Section 5
NRMRL Research Results
Primary PM From Combustion of Heavy Fuel Oils
Phase I Tests
Initial experiments were conducted with the goal of understanding how PM emissions
and size-specific composition changed with changes in fuel composition and operating
conditions typical of small boilers commonly used in commercial, institutional, and light
industrial applications. The test results indicated that modest changes in combustion
conditions did not result in significant changes in PM stack concentrations or
characteristics. Reducing the fuel oil feed temperature from 120 °C to 77 °C and
reducing excess air to achieve 2.5% 02 in the stack (compared to 3.5%) changed the total
PM emission rate (lb/106 Btu or mg/kJ) by less than 5%. Similarly, changes in the loss
on ignition (LOT) for particles larger than 2.5 jim in aerodynamic diameter were seen
during the same tests. Increases of up to 18% in PM concentration (mg/m3) were noted
during the same tests, and the LOT values for particles smaller than 2.5 urn in
aerodynamic diameter also increased by roughly 20% (Miller et al. 1998).
The mass emission rates were different, but not significantly, for the four fuel oils burned
(a No. 5 fuel oil and low, medium, and high sulfur No. 6 fuel oils). The differences in
total mass emission rates correlated positively with unburned carbon levels (see Table 3).
These results are consistent with particle composition, which showed high levels of
carbon in the coarse particle fraction due to incomplete combustion in the relatively cold
combustion environment of this firetube boiler design. The high carbon levels added
significant mass to the total PM concentrations, resulting in the correlation between LOT
and PM mass concentration.
Table 3. PM emission rates and emission factors and measures of unburned carbon for oils burned in
NRMRL tests
Fuel type
Combustor
Fuel
PM, mg/m6
PM, lb/10bBtu
PM, lb/1000gal
Total LOI,%
Filter LOI,%
Cyclone LOI,%
Unburned
carbon in fly
ash, %
Fuel Oils
Firetube Boiler
No. 5
197.7
0.143
21.0
NAa
75.3
95.5
NA
LowS
No. 6
219.5
0.161
24.2
NA
86.6
96.9
NA
Medium S
No. 6
243.5
0.142
21.3
NA
79.0
93.6
NA
HighS
No. 6
183.6
0.123
18.6
NA
65.8
90.3
NA
Refractory
Lined Furnace
High S No. 6
93
0.052
7.17
19.6
NA
NA
~0
Emulsified Fuels
Firetube Boiler
No. 6
160
0.11
15.4
NA
NA
NA
NA
No. 2
3.4
0.0027
0.26
NA
NA
NA
NA
Watertube Boiler
Simulator
Orimulsion 400
160
0.12
12.5
NA
NA
NA
NA
1 Not Available
25
-------
The PSDs were consistently bimodal, with a fine mode slightly smaller than 0.1 jim and a
coarse mode near 40 jim. This indicates two particle formation processes— vaporization
followed by nucleation and condensation forming the fine particle mode and mechanical
processes forming the coarse mode. This is consistent with the previously published
results noted earlier. These findings are also consistent with particle images viewed
under a scanning electron microscope (SEM). Many of the particles from the package
boiler appeared as large (>20 jim in diameter), porous cenospheres. These cenospheres
are formed when the lighter fractions of the oil burn off, leaving the less volatile and
more slowly burning char. When conditions are not hot enough or when there is not
enough oxygen present, these remaining spheres composed of the heavier fractions form
a char, which passes unburned out of the boiler in the form of PM. These unburned
carbon cenospheres significantly increased the PM mass above what would be expected
based on the fuel ash content alone (Miller et al. 1998).
The key findings from the initial phase of research are:
• Modest changes in combustion conditions, such as slightly reducing excess air and
reducing the oil feed temperature, did not significantly change PM mass emissions;
• PM mass emissions from a firetube boiler do not positively correlate with oil ash
content, but more closely correlate with unburned carbon levels in the ash; and
• Incomplete combustion of the carbon in the fuel results in the formation of large (>
20 |im diameter) particles that contribute substantially to the total PM mass emission
rates for these boiler designs.
As this initial study was being developed, discussions with NHEERL identified an
opportunity for collaboration through providing PM samples for toxicity testing. These
samples were collected from the large dilution sampler shown in Figure 8, and the
particles were physically removed from the filter for later toxicological tests. Although
the physico-chemical properties of the samples varied only slightly, particle samples from
these initial tests showed significant differences in biological response as measured by
indications of pulmonary damage following intratracheal instillation of the particles into
laboratory animals. These indications were most pronounced in the fine (<2.5 jim)
particle fraction and in the oils with the highest sulfur and transition metal contents.
Phase II Tests
As a consequence of these initial studies and based on understanding of particle
formation mechanisms, the research turned toward examining the effects of large changes
in combustion condition by comparing particles formed in two combustion systems of
significantly different design. By using a hotter furnace, more complete carbon burnout
would be expected to form particles that were almost entirely in the ultrafine (<0.1 jim)
fraction, since the unburned carbon that formed the larger supermicron particles would be
completely combusted in the hotter unit. The smaller particles were also expected to be
at least as water soluble as the fine fraction generated in the initial study and would
26
-------
provide NHEERL with the ability to examine the toxicological mechanisms in more
detail.
Therefore, the second set of heavy fuel oil experiments was conducted on the refractory-
lined Rainbow tunnel furnace. This unit generated particles that had very little unburned
carbon, in contrast to the substantial amount of carbonaceous cenospheres in the firetube
boiler. The dominant particle formation route in the Rainbow furnace was through
vaporization and subsequent nucleation and condensation of inorganic material. This is
illustrated by the single particle size mode near 0.1 jim (see Figure 9). The smaller
particles generated in the Rainbow were also found to be highly water soluble, similar to
1000
1E+5
9E+4-
00
100
Diameter, Dp (|am)
Figure 9. Particle size distributions for No. 6 fuel oil in the North American Package Boiler (open circles)
and Rainbow furnace (closed circles). The inset shows details of the particle size distributions
between 0.01 and 1 urn in diameter.
27
-------
the fine fraction of particles from the firetube boiler. Because of the improved carbon
burnout, the total PM emission rate from the Rainbow was approximately half that from
the firetube boiler, even when using the same fuel at nominally similar combustion
conditions (as determined by excess 02) (Linak et al. 2000b).
The emission rates from the two combustion systems were similar to values reported
from commercially operating equipment, with the firetube boiler representing small
commercial and industrial boilers and the Rainbow simulating larger industrial and utility
boilers. Comparing the results from these first two series of tests, it is interesting to note
that more complete carbon burnout resulted in significantly less total PM mass emissions,
but a higher mass of the fine particles that appear to be more bioactive.
The relationship between carbon burnout (as measured by LOT) and PM mass emissions
is clearly seen in Figure 10, which also shows the relationship between fuel oil ash
content and PM mass emissions for the fuels and combustors used in NRMRL's
experiments. The relationship between ash content and PM emissions indicates a slight
inverse relationship; as ash levels in the fuel increase, PM mass emissions decrease. It is
D Baseline, No. 5 Oil A MS No. 6 Oil
O Low Oil Temp, No. 5 Oil o HS No. 6 Oil
Low Stoichiometric Ratio, No. 5 Oil • HS No. 6 Oil, Refractory Lined Furnace
+ LS No. 6 Oil
250
CO
O)
E
200-
~ 150-1
CO
o 100
o
O
W 50-
250-
200-
150-
100-
50-
ppppppppp
ooooooooo
Fuel ash content, %
10 20 30 40 50 60 70 80 90
Fly ash LOI, %
Figure 10. Stack concentrations of PM as a function of fuel ash content for the four fuel oils burned in the
firetube boiler (left) and fly ash loss on ignition (right) for the four fuel oils burned in the firetube
boiler and the Rainbow refractory lined furnace
28
-------
unclear whether there is an underlying mechanism resulting in this relationship, but it
seems unlikely that this will hold in general. For the NAPB boiler data only (i.e.,
ignoring the single point from the Rainbow furnace), the relationship between ash content
and PM concentration shows a slightly negative correlation.
In contrast, the relationship between LOT and PM concentration is positive. Physically
this makes sense because particles with high LOT contain mass that would have been
converted to CC>2 under optimal conditions rather than remaining with the particles and
contributing to the emitted mass. Although this positive correlation between LOT and PM
is also evident from the samples taken from the firetube boiler, the most pronounced
difference is between the Rainbow furnace and the data from the firetube boiler clustered
in the upper right of Figure 10. This difference in PM mass emissions between the two
systems clearly illustrates the impact of system design on emissions, with the hotter
Rainbow generating approximately half the mass emissions compared to those from the
firetube boiler.
The key findings from the second phase of the research are:
• Significant changes in combustion temperature and residence time due to different
combustor designs result in substantial changes in PM characteristics and mass
emissions;
• Combustion of heavy fuel oil under conditions simulating large industrial and utility
boilers generates a nearly unimodal submicron PSD; and
• The submicron particles contain metals and sulfur that are largely in water soluble
form, in contrast to the supermicron fraction of particles that contains less soluble
thiophenic sulfur and metal species.
The level of unburned carbon also appears to influence the chemical species that are
formed. Detailed examination of the particles from the NAPB indicate that
approximately 50% of the sulfur in the particles is not oxidized, although essentially all
of the sulfur in the PM collected from the Rainbow furnace is in the form of sulfates
(SO4) (Huffman et al. 2000, Pattanaik et al. 2000). The form of sulfur is significant,
since the sulfate form is much more soluble in water than the unoxidized organic
(thiophenic) form of sulfur. Much of the sulfate in these submicron particles is combined
with metals, resulting in metals with relatively high solubility in water. As noted earlier,
one of the characteristics of particles from heavy fuel oil that is suspected of leading to
adverse health effects is the presence of water-soluble transition metals, and metal
sulfates are much more water-soluble than most other metal species that are formed in
combustion systems. The combination of transition metals and high sulfate levels
indicates that the Rainbow PM is likely to contain much higher levels of water-soluble
transition metals than the PM from the NAPB, even though both are generated by
combustion of the same fuel.
29
-------
Data from Other Studies
In addition to the conventional heavy fuel oils, two water-in-oil emulsified fuels and
Orimulsion, a bitumen-in-water fuel, were also tested under other programs. PM size
distributions and mass emissions were tested during these programs and provide
additional information of relevance here. For the emulsified fuels tested, the presence of
water in the fuel changed the way the fuels burned and resulted in noticeable changes in
PM emissions and size distributions. The water in the emulsified fuel oils created
"microexplosions" as the fuel is rapidly heated, thereby forming much smaller fuel
droplets than could be formed by the mechanical processes of the nozzle. These smaller
droplets burn more quickly, thereby reducing the amount of unburned carbon and
reducing both the mass of PM emissions and the size of the emitted particles (Miller et al.
2001, Miller 1996). Even though the physical arrangement of Orimulsion is different
from the water-in-oil emulsified fuel oils, the results were very similar when compared to
a non-emulsified fuel oil. This is most likely due to the small size (8-20 jim diameter) of
the bitumen droplets in the Orimulsion, which is much smaller than the fuel oil droplets
generated by the oil nozzle (Miller and Hall 2000). The net effect of the microexplosions
and subsequent smaller fuel droplets is to reduce the time needed to volatilize and
combust the fuel. This reduces the impact of rapid temperature drop in systems such as
the firetube boiler and results in particles with lower unburned carbon and higher
fractions of submicron particles compared to non-emulsified fuels burned under similar
conditions.
Further Evolution of Research Directions
Given the successful generation, sampling, and instillation of particles in these first two
series of tests, the research then moved forward along two different, but complementary,
routes. The physico-chemical characterization studies of heavy fuel oil were essentially
complete, but there was substantial interest in expanding the collaboration on determining
toxicity by providing a source of particles for joint NMRML-NHEERL studies of direct
inhalation of the particles by animals. This work has begun with initial tests using heavy
fuel oil. This route of research is anticipated to continue for the foreseeable future, with
the emphasis on toxicological rather than physicochemical characterization. Results from
these collaborative studies are still being analyzed, but the capability to generate well
characterized, real-world particles for real-time inhalation exposure studies has been
demonstrated and can be applied to other source types as research needs dictate.
The second route of research stemming from the original heavy fuel oil studies was
testing of pulverized coals. Although there has been much more research conducted to
characterize PM generated by the combustion of pulverized coal, no physico-chemical
characterization studies were known to have been conducted with the goals of
quantifying the presence of water soluble transition metals and generating well-
characterized primary PM samples for toxicity studies aimed at better understanding the
causal mechanism of health damage associated with exposure to ambient PM2.s.
30
-------
Primary PM From Combustion of Coals
Nine different U.S. coals were burned during NRMRL's PM characterization tests.
These coals were chosen to represent different regions of the U.S. and have different
sulfur and trace element contents (see Table 1) as well as different ash chemistry. The
initial tests focused on particle size, mass, and composition measurements for the coals
burned in the down-fired IFR.
PM concentrations generated by the combustion of coal were much higher than for the
fuel oils, as shown in Table 4. This is expected because of the significantly higher
amount of inorganic ash present in the coal (8-20%) compared to the fuel oils (usually
much less than 1%). Uncontrolled mass concentrations for the coals tested were over 20
times higher than those for the heavy fuel oil, which is similar to the ratio of inorganic
ash contents of the coals versus the fuel oils. However, most of the coal PM mass is in
the larger (>5 jim diameter) size fractions, with only about 4 to 7% of the total PM mass
associated with PM2.5. The unburned carbon values measured in EPA's experiments
were slightly higher than, but not atypical of, many utility-scale boilers and considerably
less than the LOT values for the PM from heavy fuel oil burned in the NAPB (Linak et al.
2000a). However, because of the much higher levels of inorganic material in coal
compared to that in heavy fuel oil, it can be misleading to directly compare fractions of
unburned carbon or LOT for the two fuel types. For instance, 5% unburned carbon for a
coal-generated PM may translate into more total mass of unburned carbon than 80%
unburned carbon for PM from the combustion of heavy fuel oil. The much higher
inorganic content in coal also results in unburned carbon having a much lower effect on
PM mass emissions for coal than for oil, since coal PM mass emission rates are
dominated by the inorganic material present in the fuel.
Table 4. PM emission rates and emission factors and measures of unburned carbon for coals burned in
NRMRL tests.
Fuel Type
Fuel
PM, mg/nnj
PM, lb/10bBtu
PM, Ib/ton
Unburned carbon
in fly ash, %
Bituminous Coals
Ohio
5,6,7
Blend
4477
3.70
95.2
NAa
Pitts-
burgh 8
3565
2.71
70.3
NA
Utah
4323
3.32
74.9
10.9
West
Ken-
tucky
3807
3.00
67.9
10.2
Subituminous Coal
Montana
4374
3.30
62.8
0.5
Wyodak
(PRB(a))
3441
2.63
46.6
NA
Lignite
North
Dakota
5582
6.02
77.0
NA
a Not available
NRMRL's experimental results demonstrate some of the different mechanisms involved
in particle formation during coal combustion. Figure 11 presents typical particle size
distributions for several of the coals tested. The three distinct modes indicate different
formation mechanisms. The smallest mode occurs near 0.1 jim in diameter and is due to
the vaporization and subsequent nucleation and accumulation similar to that in the
31
-------
ultrafme particles from heavy fuel oil combustion. These particles tend to be enriched in
metals that have volatilized and nucleated and to have high levels of sulfates. The second
mode occurs between 0.8 and 2.0 jim in diameter and has not been commonly identified
in previous studies. This mode is believed to be composed of fragments of ash particles
originally contained within the pulverized coal particles. The third mode begins near 10
|im diameter and is formed largely by coalescing inorganic material that is produced as
the carbonaceous coal char burns away. The two largest modes have similar composition,
indicating similar mechanical mechanisms of formation (Linak et al. 2001, Shoji et al.
2000). The similarity in composition supports the hypothesis that the 0.8-2 jim particles
were originally part of the larger (>10 jim) particles, which fragmented to form the
1E+4
1E+6
D Western Kentucky
O Montana
A Utah
100
Diameter, Dp (urn)
Figure 11. Particle size distributions for three coals burned in NRMRL tests.
32
-------
0.8-2 jim particles. Both these larger modes significantly differ in composition from the
submicron mode.
As is the case for PM from heavy fuel oils, the particles larger than 1 jim are formed by
primarily mechanical means and will contain metals in forms that are tightly bound
physically and chemically. This will result in water solubility that is significantly lower
than that seen for PM from heavy fuel oils. The submicron particles, formed via
vaporization, nucleation, and condensation, will be composed in part of metal and other
sulfates that have a much higher solubility in water, and are therefore expected to have a
higher level of bioavailability.
Figure 12 illustrates the relationships between coal ash content and PM mass
concentrations, and LOT and PM mass concentrations. Note the contrast between these
relationships and those for the heavy fuel oils (see Figure 10). For coal, the fuel ash
content has a significant effect on PM concentrations, while LOT has minimal, if any,
effect. As noted above, this is due to the much higher ash content in coals that
overwhelms any significant effect from LOT.
6000-
CO
5000-
- 4000-
o
£ 3000-
0>
o
g 2000-
o
^ 1000-1
0-
• Pittsburgh #8 T Utah Bituminous H North Dakota Lignite
+ W. Kentucky Bituminous E Wyodak (PRB) Subbituminous * Ohio Bituminous (Blend)
A Montana Subbituminous
1 i '
4
' i '
6
' i '
8
' i ' ' ' i ' ' ' i '
10 12 14 16
' i ' '' i ' '' i ''
10 12 14 16
Fuel Ash Content, %
LOI, %
Figure 12. Stack concentrations of PM as a function of fuel ash content (left) and fly ash loss on ignition
(right) for the seven coal burned in the innovative furnace reactor.
The presence of the "fine fragmentation" mode between 0.8 and 2.0 jim means that the
PM2.5 from coal combustion will be composed of particles formed by different
mechanisms, and with different chemical compositions. The fine fragmentation mode
33
-------
makes up the majority of coal-generated PM2.5 mass and has characteristics similar to the
coarse (>2.5 jim) particle fraction. These particles will contain significant amounts of
silica and alumina compounds that tie up many of the other trace elements, including the
transition metals. On a per-unit-mass basis, therefore, PM2 5 from coal combustion is
likely to contain less water-soluble transition metals compared to PM2.5 from heavy fuel
oil. Even so, there will remain in coal-generated PM an ultrafine fraction that is expected
to be more similar in composition and physical characteristics, and possibly in
toxicological characteristics, to the ultrafine fraction in PM2 5 from heavy fuel oil
combustion.
As was the case for PM from heavy fuel oil, these results emphasize that particle size and
composition are a consequence of the formation processes. The larger particles (>0.8
jim) are formed by processes that do not involve vaporization and are therefore less likely
to undergo significant chemical change. On the other hand, the ultrafine particles that do
vaporize are more likely to undergo the same types of chemical reactions that favor the
formation of water soluble metal sulfates.
The hypothesis that the submicron fraction of the coal generated PM may be more
bioactive than the larger particle fractions has led NRMRL researchers to begin efforts to
collect larger quantities of these particles. Because they make up such a low percentage
of the total particle mass, however, such collection requires different particle generation
and collection techniques. One approach is to begin with micronized coal, which is
ground to a much finer size (~ 5 jim) than standard pulverized coal (~ 70 jim). This
approach is expected to generate a substantially higher fraction of the ultrafine particles
of interest, making collection of significant quantities of these particles for NHEERL
toxicity studies much more efficient. Initial efforts in this area are promising and are
ongoing at the writing of this report.
The key findings from this segment of the NRMRL research program are:
• A previously unreported particle mode between 0.8 and 2.0 jim contributes to the mass
of PM25 from the combustion of pulverized coal;
• The composition of particles in this "fine fragmentation" mode is similar to that of the
coarse mode particles, suggesting that the mode is formed by coarse ash particles that
fragment during the combustion process. The composition also suggests that these
particles will have metals with relatively low water solubility and therefore relatively
low bioavailability; and
• The ultrafine fraction of coal-generated particles has composition similar to that of the
ultrafine mode of particles from heavy fuel oil combustion, suggesting that the
pulmonary toxicity of ultrafine particles from coal combustion may be similar to the
heavy fuel oil ultrafine particles.
34
-------
Model Predictions
In addition to the experimental program, NRMRL researchers also used computer models
to help explain the behavior and speciation of particles. These models were used to
simulate the major chemical and physical processes governing the formation and
speciation of particles and to provide data to which the experimental results could be
compared. This approach provided guidance in interpreting the experimental results and
was especially useful in gaining a better understanding of the metal species likely to be
present in the ultrafme PM.
Chemical Equilibrium Modeling
Chemical equilibrium calculations allow one to identify chemical species that will form
with adequate time at a given temperature. While they do not account for the effects of
incomplete mixing or chemical reactions that do not have adequate time to go to
completion (kinetics), equilibrium results can be used to determine what chemical species
are likely to be present in a given environment and how changes in temperature or
presence of other compounds can influence the speciation in the flue gases.
Thermochemical predictions for a simulated heavy fuel oil were determined using the
Chemical Equilibrium Analysis (CEA) computer code for calculating complex chemical
equilibrium compositions (McBride et al. 1993) with thermochemical data taken from the
literature (see Linak et al. 1999, 2000b). Of interest is the predicted partitioning between
vapor and condensed phases as well as elemental speciation. Using experimental data
from NRMRL tests (Miller et al. 1998), CEA predicted that seven inorganic elements of
potential interest to toxicologists were likely to form vapor phase species in the high
temperature regions of combustion systems. All seven of these elements were also
predicted to form sulfates at temperatures typical of exhaust stacks (Linak et al. 1999,
2000b). These results were consistent with the study of Huffman et al. (2000) that
analyzed for several species in the fine and coarse fractions of oil-generated PM samples
collected during the NRMRL experiments. Taken together, the computational and
analytical results demonstrate that the formation of water soluble sulfate species is
favored by the conditions present in large industrial and utility boilers.
Aerosol Nucleation and Coagulation Modeling
In addition to modeling the chemical processes, NRMRL researchers also used a multi-
component aerosol simulation (MAEROS) code (Gelbard and Seinfeld 1980) to model
the physical evolution of particle sizes. The purpose of these calculations was to
determine whether the experimental data were consistent with the calculated predictions
to support the interpretation of the experimental results. A secondary objective was to
examine the effects of particle coagulation on particle size distributions, specifically to
evaluate how changes in the vaporized inorganic material would impact PSDs.
Beginning with initial conditions based on NRMRL experimental results, MAEROS
predicted PSDs similar to those measured during the heavy fuel oil tests. The
35
-------
consistency of the model and experimental results indicated that NRMRL's interpretation
of the particle formation mechanisms based on the experimental data is reasonable (Linak
etal. 1999, 2000b).
36
-------
Section 6
Summary and Implications
Summary
In summary, NRMRL's combustion PM research has found the following:
• Increased unburned carbon results in increased PM mass emissions and increased
particle size when burning heavy fuel oil.
• Use of emulsified fuel oils reduces both the PM mass and the particle size compared
to a non-emulsified fuel oil.
• Unburned carbon does not affect PM mass emissions or particle size distributions
nearly as strongly for coal as for heavy fuel oil.
• For all fossil fuels tested, the submicron particles tend to have higher concentrations
of metals and sulfates than particles larger than 1 jim in diameter.
• The water solubility of metals is significantly higher for the total PM from heavy fuel
oils than for the total PM from pulverized coal, although the similar compositions of
the ultrafine fractions of both fuels suggests the difference for this fraction is much
smaller.
• A fine fragmentation mode between 0.8 and 2.0 jim exists in PM generated from the
combustion of pulverized coal. The particles in this mode are of similar composition
to the particles in the coarse (>2.5 jim) mode.
Implications
Particle composition varies considerably with size, with potentially significant
implications for particle toxicity. NRMRL's research, consistent with previous work,
has identified significant variation in particle composition as a function of particle size.
The submicron particle fractions were consistently composed of metal and other sulfates
that are highly water soluble, meaning that they are more likely to interact with biological
tissues when inhaled. The formation mechanisms that generate these particles govern
both size and composition. Therefore, the question of whether size or composition plays
a dominant role in causing adverse health effects may be inappropriate, since the two
(size and composition) cannot be separated in real-world systems. It may be more
relevant to focus on particles with similar formation processes rather than on particles
with specific characteristics that cannot be naturally separated. In short, multiple
mechanisms are involved in the formation of PM2 5 from combustion sources, and these
mechanisms will directly impact the role of particles in causing adverse health effects.
The design and operation of a source can result in significant changes in the
physico-chemical and toxicological characteristics of the particles generated. This
finding has scientific and potentially regulatory implications. In the context of
toxicological research, more emphasis needs to be placed on how particle samples are
generated and collected. PM toxicological studies often do not adequately describe the
source characteristics, including system design, the conditions at which the system is
running, or the methods used to extract particle samples. PM from heavy fuel oil, as
37
-------
demonstrated here, is not a homogeneous material with consistent characteristics. In
order to develop an adequate understanding of the health effects associated with PM and
the causal mechanisms leading to those effects, studies must provide enough information
to allow the reader to understand the representativeness of the particles used in the tests.
From a regulatory perspective, these results may eventually be used to identify source
types that have the potential for emitting particles having characteristics closely
associated with health risk. As additional understanding is gained regarding links
between health effects and particle characteristics, it may someday be possible to develop
regulatory approaches to reduce PM emissions from a more limited range of source types
or to minimize emissions of particles with certain characteristics. If enough information
on adverse health effects due to water-soluble ultrafme particles is developed, for
instance, limits on emissions of these particles may replace or supplement limits on total
mass emissions.
38
-------
Section 7
Future Directions
NRMRL's research to characterize primary particles from the combustion of pulverized
coal and heavy fuel oils has been completed. There remain numerous research
opportunities in the broader area of combustion-generated particles, which will be
followed as appropriate. Immediate efforts will focus on characterizing emissions from
small off-road diesel engines, which will involve more organic analysis, as compared to
the significant levels of inorganic analysis conducted for the previous studies. Several
efforts associated with the collaborative studies linking the physicochemical properties of
combustion-generated particles to measures of toxicity during direct inhalation exposures
will continue. This area of work shows considerable promise for generating data that can
be used to better understand the links between sources and adverse health effects.
The capabilities developed during the work described in this report will continue to
generate state of the art research. NRMRL's flexible and unique facilities, in
combination with ORD's on-site expertise, provide unique opportunities for leading edge
multidisciplinary research in PM, air toxics, and other air programs.
39
-------
(Blank Page)
40
-------
Section 8
Publications from NRMRL's
Combustion-Generated PM Research Program
The 10 articles listed below represent key publications generated from the work described
above (NRMRL authors noted in boldface). In addition to these publications, an
additional 12 papers were prepared and presented at technical conferences between 1997
and 2002.
Huffman, G.P., Huggins, F.E., Shah, N., Huggins, R., Linak, W.P., Miller, C.A.,
Pugmire, R.J., Meuzelaar, H.L.C., Seehra, M.S., and Manivannan, A. (2000).
"Characterization of fine particulate matter produced by combustion of residual fuel oil,"
Journal of the Air & Waste Management Association, Vol. 50, 1106-1114.
Linak, W.P., Miller, C.A., Wendt, J.O.L., and Dreher, K. (1998). "Fine particulate from
residual fuel oil combustion: Physical, chemical, and health effect characteristics," 17th
Annual Conference of the American Association for Aerosol Research, June 22-26, 1998,
Cincinnati, OH.
Linak, W.P., Miller, C.A., and Wendt, J.O.L. (1999). "Fine particle emissions from
residual fuel oil combustion: characterization and mechanisms of formation," 5th
International Conference on Technologies and Combustion for a Clean Environment,
July 12-15, 1999, Lisbon, Portugal.
Linak, W.P., Miller, C.A., and Wendt, J.O.L. (2000). "Comparison of particle size
distributions and elemental partitioning from the combustion of pulverized coal and
residual fuel oil" Journal of the Air & Waste Management Association, Vol. 50, 1532-
1544.
Linak, W.P., Miller, C.A., and Wendt, J.O.L. (2000). "Fine particle emissions from
residual fuel oil combustion: characterization and mechanisms of formation,"
Proceedings of the Combustion Institute, Vol. 28, 2651-2658.
Linak, W.P., Miller, C.A., Seames, W.S., Wendt, J.O.L., Ishinomori, T., Endo, Y., and
Miyamae, S. (2002). "On trimodal particle size distributions in fly ash from pulverized
coal combustion," Proceedings of the Combustion Institute, Vol. 29, in press.
Miller, C.A., Linak, W.P., King, C., and Wendt, J.O.L. (1998). "Fine particle emissions
from heavy fuel oil combustion in a firetube package boiler," Combustion Science and
Technology, Vol. 134, 477-502.
41
-------
Pattanaik, S., Huggins, F.E., Huffman, G.P., Linak, W.P., and Miller, C.A. (2000).
"XAFS spectroscopy analysis of the molecular structure of metals and sulfur in fine
particulate matter (PM) derived from the combustion of residual oil," ACS Fuel
Chemistry Division., Vol. 46.
Shoji, T., Huggins, F.E., Huffman, G.P., Linak, W.P., and Miller, C.A. (2000). "XAFS
spectroscopy analysis of the molecular structure of metals in fine parti culate matter (PM)
derived from the combustion of coal," ACS Fuel Chemistry Division, Vol. 46.
Wendt, J.O.L., Fernandez, A., Witten, M.L., Wang, S., Riley, M.R., Okeson, C.D.,
Linak, W.P., and Miller, C.A. (2001). "On the generation and subsequent health effects
of fuel combustion generated particulates: the roles of zinc and vanadium," 13th
International International Flame Research Foundation Members' Conference, May 15-
18, 2001, Noordwijkerhout, The Netherlands.
42
-------
Section 9
References
Amdur, M.O., Sarofim, A.F., Neville, M., Quann, R.J., McCarthy, J.F., Elliot, J.F., Lam,
H.F., Rogers, A.E., and Conner, M.W. (1986). "Coal combustion aerosols and SC^: an
interdisciplinary approach," Environmental Science and Technology, Vol. 20, 138-145.
Bacci, P., Del Monte, M., Longhetto, A., Piano, A., Prodi, F., Redaelli, P., Sabbioni, C.,
and Ventura, A. (1983). "Characterization of the particulate emission by a large oil fuel
fired power plant," Journal of Aerosol Science., Vol. 14, 557-572.
Bachmann, J.D., Damberg, R.J., Caldwell, J.C., Edwards, C., and Koman, P.O. (1996).
"Review of the National Ambient Air Quality Standards for particulate matter: policy
assessment of scientific and technical information," EPA-452/R-96-013 (NTIS PB97-
115406), Research Triangle Park, NC, July 1996.
Baggs, R.B., Ferin, J., and Oberdorster, G. (1997). "Regression of pulmonary lesions
produced by inhaled titanium dioxide in rats," Veterinary Patholology, Vol. 34, 592-597.
Bulewicz, E.M., Evans, D.G., and Padley, P.J. (1974). "Effect of metallic additives on
soot formation processes in flames," 15th Symposium (International) on Combustion,
1461-1470, The Combustion Institute, Pittsburgh.
Burnett, R.T., Dales, R., Krewski, D., Vincent, R., Dann, T., and Brook, J.R. (1995).
"Associations between ambient particulate sulfate and admissions to Ontario hospitals for
cardiac and respiratory disease," American Journal of Epidemiology, Vol. 142, 15-22.
Dockery, D.W., Pope, C.A., III, Xu, X., Spengler, J.D., Ware, J.H., Fay, M.E., Ferris,
E.G., Jr., and Speizer, F.E. (1993). "An association between air pollution and mortality in
six U.S. cities," New England Journal of Medicine, Vol. 329, 1753-1759.
Dreher, K., Costa, D., Hoffman, A., Bonner, J., and Osornio-Vargas, A. (1996a).
"Pulmonary toxicity of urban air particulate matter (PM)," Air & Waste Management
Association Meeting on Measurement of Toxic and Related Air Pollutants, Research
Triangle Park, NC, May 1996.
Dreher, K., Jaskot, R., Richards, J.H., Lehmann, J.R., Winsett, D., Hoffman, A., and
Costa, D. (1996b). "Acute pulmonary toxicity of size-fractionated ambient air particulate
matter," American Journal Respiratory and Critical Care Medicine, Vol. 153, 4-A15.
43
-------
Dreher, K., Jaskot, R., Lehmann, J.R., Richards, J.H., McGee, J.K., Ohio, A.J., and
Costa, D.L. (1997). "Soluble transition metals mediate residual oil fly ash induced lung
injury," Journal of Toxicology and Environmental Health, Vol. 50, 285-305.
Federal Register (1997). 62 FR 38652, July 18, 1997.
Feldman, N. (1982). "Control of residual fuel oil particulate emissions by additives,"
19th Symposium (International) on Combustion, 1387-1393, The Combustion Institute,
Pittsburgh.
Fernandez, A., Davis, S.B., Wendt, J.O.L., Young, S., and Witten, M.L. (2000).
"Systematic generation and resuspension of combustion generated fine particles for
inhalation health effects studies," presented atNARSTO 2000, Tropospheric Aerosols:
Science and Decisions in an International Community, Queretaro, Mexico, October 23-
26, 2000.
Flagan, R.C. and Taylor, D.D. (1981). "Laboratory studies of submicron particles from
coal combustion," 18th Symposium (International) on Combustion., 1227-1237, The
Combustion Institute, Pittsburgh.
Gallagher, N.B., Bool, L.E., Wendt, J.O.L., and Peterson, T.W. (1990). "Alkali metal
partitioning in ash from pulverized coal combustion," Combustion Science and
Technology, Vol. 74, 211-221.
Gelbard, F. and Seinfeld, J.H. (1980). "Simulation of multicomponent aerosol
dynamics," Journal of Colloid Interface Science, Vol. 78, 485-501.
Gluskoter, HJ. (1978). An introduction to the occurrence of mineral matter in coal, ash
deposits and corrosion due to impurities in combustion gases, R.W. Bryers, ed.,
Hemisphere Publishing, Washington, DC.
Godleski, J., Verrier, R.L., Koutrakis, P., and Catalano, P. (2000). "Mechanisms of
morbidity and mortality from exposure to ambient air particles," Health Effects Institute,
Research Report 91, Health Effects Institute, Cambridge, MA.
Hatch, G.E., Boykin, E., Miller, F.J., and Graham, J.A. (1982). "Effects of fly ash and its
constituents on sensory irritation in mice," Fundamental and Applied Toxicology, Vol. 2,
77-81.
44
-------
Hatch, G.E., Boykin, E., Graham, J.A., Lewtas, J., Pott, F., Loud, K., and Mumford, J.L.
(1985). "Inhalable particles and pulmonary host defense: In vivo and in vitro effects of
ambient air and combustion particles," Environmental Research, Vol. 36, 67-80.
Hersh, S., Piper, B.F., Mormile, D.J., Stegman, G., Alfonsin, E.G., and Rovesti, W.C.
(1979). "Combustion demonstration of SCR II fuel oil in a utility boiler," ASME Winter
Annual Meeting, 79-WA/Fu-7, December 2-7, 1979, New York, NY.
Huffman, G.P., Huggins, F.E., Shah, N., Huggins, R., Linak, W.P., Miller, C.A.,
Pugmire, R.J., Meuzelaar, H.L.C., Seehra, M.S., and Manivannan, A. (2000).
"Characterization of fine particulate matter produced by combustion of residual fuel oil,"
Journal of the Air & Waste Management Association, Vol. 50, 1106-1114.
Kuschner, W.G., Wong, H., D'Alessardro, A., Quinlan, P., and Blanc, P.O. (1997).
"Human pulmonary responses to experimental inhalation of high concentration fine and
ultrafine magnesium oxide particles," Environmental Health Perspectives, Vol. 105,
1234-1237.
Lighty, J.S., Veranth, J.M., and Sarofim, A.F. (2000). "Combustion aerosols: Factors
governing their size and composition and implications to human health," Journal of the
Air & Waste Management Association, Vol. 50, 174-227.
Linak, W.P., Miller, C.A., Wendt, J.O.L., and Dreher, K. (1998). "Fine particulate from
residual fuel oil combustion: physical, chemical, and health effect characteristics," 17th
Annual Conference of the American Association for Aerosol Research, June 22-26, 1998,
Cincinnati, OH.
Linak, W.P., Miller, C.A., and Wendt, J.O.L. (1999). "Fine particle emissions from
residual fuel oil combustion: characterization and mechanisms of formation," 5th
International Conference on Technologies and Combustion for a Clean Environment,
July 12-15, 1999, Lisbon, Portugal.
Linak, W.P., Miller, C.A., and Wendt, J.O.L. (2000a). "Comparison of particle size
distributions and elemental partitioning from the combustion of pulverized coal and
residual fuel oil" Journal of the Air & Waste Management Association, Vol. 50, 1532-
1544.
Linak, W.P., Miller, C.A., and Wendt, J.O.L. (2000b). "Fine particle emissions from
residual fuel oil combustion: characterization and mechanisms of formation,"
Proceedings of the Combustion Institute, Vol. 28, 2651-2658.
45
-------
Linak, W.P., Miller, C.A., Seames, W.S., Wendt, J.O.L., Ishinomori, T., Endo, Y., and
Miyamae, S. (2001). "On trimodal particle size distributions in fly ash from pulverized
coal combustion," Proceedings of the Combustion Institute, Vol. 29, in press.
Linak W.P. and Wendt, J.O.L. (1994). "Trace metal transformation mechanisms during
coal combustion," Fuel Processing Technology, Vol. 39, 173-198.
Lippmann, M., Ito, K., Nadas, A., and Burnett, R.T. (2000). "Association of particulate
matter components with daily mortality and morbidity in urban populations," Health
Effects Institute Report No. 95, Health Effects Institute, Cambridge, MA.
Markowski, G.R., Ensor, D.S., Hooper, R.G., and Carr, R.C. (1980). "A submicron
aerosol mode in flue gas from a pulverized coal utility boiler," Environmental Science
and Technology, Vol. 14, 1400-1402.
McBride, B.J., Gordon, S., and Reno, M.A. (1993). "Coefficients for Calculating
Thermodynamic and Transport Properties of Individual Species," NASA Technical
Memorandum 4513.
Miller, C.A. (1996). "Hazardous air pollutants from the combustion of an emulsified
heavy fuel oil in a firetube boiler," EPA-600/R-96-019 (NTIS PB96-168281), Research
Triangle Park, NC, February 1996.
Miller, C.A. (1998). "Environmental Technology Verification: Verification testing of
emissions from the combustion of A-55® Clean Fuels in a firetube boiler," EPA-600/R-
98-035 (NTIS PB98-142169), Research Triangle Park, NC, April 1998.
Miller, C.A. and Hall, R.E. (2000). "Generation and control of air pollutants from
Orimulsion® combustion," Fall 2000 Meeting of the American Flame Research
Committee, Irvine, CA, September 18-21, 2000.
Miller, C.A. and Srivastava, R.K. (2000). "The combustion of Orimulsion and its
generation of air pollutants," Progress in Energy and Combustion Science, Vol. 26, 131-
160.
Miller, C.A., Dreher, K., Wentsel, R., andNadeau, RJ. (2001). "Environmental impacts
of the use of Orimulsion®: Report to Congress on phase 1 of the Orimulsion® technology
assessment program," EPA-600/R-01-056a (Vol. 1) and -056b (Vol. 2), Washington, DC,
July 2001.
46
-------
Miller, C.A., Linak, W.P., King, C., and Wendt, J.O.L. (1998). "Fine particle emissions
from heavy fuel oil combustion in a firetube package boiler," Combustion Science and
Technology, Vol. 134, 477-502.
Neville, M., and Sarofim, A.F. (1982). "The stratified composition of inorganic
submicron particles produced during coal combustion," 19th Symposium (International)
on Combustion, 1441-1449, The Combustion Institute, Pittsburgh.
Neville, M., McCarthy, J.F., and Sarofim, A.F. (1983). "Size fractionation of submicron
coal combustion aerosol for chemical analysis" Atmospheric Environment, Vol. 17,
2599-2604.
Oberdorster, G., Gelein, R.M., Ferin, J., and Weiss, B. (1995). "Association of particulate
air pollution and acute mortality: involvement of ultrafme particles?" Inhalation
Toxicology, Vol. 7, 111-124.
Okazaki, K., Ohtake, K., and Nishikawa, T. (1983). "Flame structure and ash particle
characteristics of one-dimensional pulverized coal combustion," 1983 ASME-JSME
Thermal Engineering Conference, Hawaii, March 1983.
Pattanaik, S., Huggins, F.E., Huffman, G.P., Linak, W.P., and Miller, C.A. (2000).
"XAFS spectroscopy analysis of the molecular structure of metals and sulfur in fine
particulate matter (PM) derived from the combustion of residual oil," ACS Fuel
Chemistry Division, Vol. 46.
Piper, B. and Nazimowitz, W. (1985). "High viscosity oil evaluation, 59th street station
unit 110, vol. 1," KVB report to Consolidated Edison Co., 21640-1, March 1985.
Quann, R.J., Neville, M., Janghorbani, M., Minis, C.A., and Sarofim, A.F. (1982).
"Mineral matter and trace-element vaporization in a laboratory pulverized coal
combustion system, Environmental Science and Technology, Vol. 16, 776-781.
Raask, E. (1985). Mineral impurities in coal combustion: behavior, problems and
remedial measure?,, Hemisphere Publishing, New York.
Samet, J.M., Zeger, S.L., Dominici, F., Curriero, F., Coursac, I, Dockery, D.W.,
Schwartz, J., and Zanobetti, A. (2000). "The national morbidity, mortality, and air
pollution study, part II: morbidity, mortality, and air pollution in the United States,"
Health Effects Institute Report No. 94, Part II, Health Effects Institute, Cambridge, MA.
47
-------
Schwartz, J., and Morris, R. (1995). "Air pollution and hospital admissions for
cardiovascular disease in Detroit, Michigan," American Journal of Epidemiology, Vol.
142, 23-35.
Shoji, T., Huggins, F.E., Huffman, G.P., Linak, W.P., and Miller, C.A. (2000). "XAFS
spectroscopy analysis of the molecular structure of metals in fine particulate matter (PM)
derived from the combustion of coal," ACS Fuel Chemistry Division, Vol. 46.
U.S. Department of Energy (2000a). "Electric power annual 1999: volume 1," DOE/EIA-
0348(99)71, Washington, D.C., July 2000.
U.S. Department of Energy (2000b). "Fuel oil and kerosene sales 1999," DOE/EIA-
0535(99), Washington, D.C., September 2000.
U.S. Department of Energy (2000c). "Coal industry annual 1997," DOE/EIA-05 84(99),
Washington, D.C.
U.S. EPA (1996). "Air quality criteria for particulate matter," EPA-600/P-95/001cF
(NTIS PB96-168224), Research Triangle Park, NC.
U.S. EPA (2000). "National air pollutant emissions trends: 1900-1998," EPA-454/R-00-
02 (NTIS PB2000-108054), Research Triangle Park, NC, March 2000.
U.S. Supreme Court (2001). "Whitman v. American Trucking Associations," Cases 99-
1257 and 99-1426, February 27, 2001.
Walsh, P.M., Wei, G., and Xie, J. (1991). "Metal oxide and coke particulates formed
during combustion of residual fuel oil," Proceedings of the 10th Annual Meeting
American Association for Aerosol Research, Paper 7P.36, Traverse City, MI, October
1991.
Walsh, P.M., Rovesti, W.C., Fangmeier, B.A., Markoja, R.L., Brown, J.P., Hopkins,
K.C., Lange, H.B., Freeman, R.F., Olen, K.R., Washington, K.T., Patrick, S.T.,
Campbell, G.L., Harper, D.S., Teetz, R.D., Bennett, T.E., and Moore, S.P. (1994). "Size
distribution of metals in particulate matter formed during combustion of residual fuel
oil," Electric Power Research Institute Proceedings of the 1993 Fuel Oil Utilization
Workshop, EPRI TR-103990, 3-132-3-149.
48
-------
Wolff, G.T. (1996). Closure by the Clean Air Scientific Advisory Committee (CASAC)
on the staff paper for paniculate matter, EPA-SAB-CASAC-LTR-96-008, Washington,
DC, June 13, 1996.
49
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. RE PORT NO.
EPA-600/R-02-093
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Primary Particles Generated by the Combustion of Heavy Fuel Oil
and Coal: Review of Research Results from EPA's National Risk
Management Research Laboratory
5. REPORT DATE
November, 2002
6. PERFORMING ORGANIZATION CODE
7. AUTHORS
C. Andrew Miller and William P. Linak
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
None, In-house
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES
Project Officers are C. Andrew Miller, MD E-305-01, phone (919) 541-2920, and William P. Linak, MD
E-305-01, phone (919) 541-5792.
16. ABSTRACT
Researchers at the U.S. Environmental Protection Agency's Office of Research and Development (ORD) have
conducted a series of tests to characterize the size and composition of primary particulate matter (PM) generated from
the combustion of heavy fuel oil and pulverized coal. These tests, conducted at ORD's National Risk Management
Research Laboratory (NRMRL) in Research Triangle Park, NC, burned four heavy fuel oils and seven coals in three
small combustion systems and measured size distributions and composition of the particles formed in the combustion
process. The research found that, for heavy fuel oils, particle composition and size are dependent upon the combustion
environment in the combustor. In the coolest test unit, unburned carbon in the fly ash was very high (70-90% by mass),
and PM mass emissions were also high, with the majority of the mass being in the coarse (>2.5 |im in aerodynamic
diameter) size range. In the hottest test unit, the PM mass emissions were approximately 50% of those from the coolest
unit, unburned carbon levels were approximately zero, and the entire PM mass was in particles smaller than 1 |jm in
aerodynamic diameter. For coal, the NRMRL research identified a previously unreported trimodal particle size
distribution, with an ultrafine (~ 0.1 |jm in aerodynamic diameter) mode, a coarse (>5 |j,m in aerodynamic diameter)
mode, and a central mode of particles between 2 and 4 |im in aerodynamic diameter. The central mode has a
composition similar to the coarse mode, but significantly different from the ultrafine mode. It is hypothesized that the
composition of the ultrafine coal particle fraction is similar to the fine particle fraction in heavy fuel oil, which has been
shown to generate increased toxicological responses following pulmonary exposure in laboratory animals. These
findings may have implications for future PM control strategies, given the influence of operating conditions and the
water-soluble composition of the smallest particle fractions of PM from metal-bearing fossil fuels.
17.
KEYWORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Combustion
Mortality
Morbidity
Fuel Oil
Coal
Particle Size Distribution
Pollution Control
Stationary Sources
13B
12B
05K
06E
11H.21D
08G
14G
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
54
Release to Public
20. SECURITY CLASS (This Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77 ) PREVIOUS EDITION IS OBSOLETE
50
forms/admin/techrpt.frm 7/8/99 pad
------- |