Physical and Chemical Characterization of Fine Particulate from the Combustion of Residual Fuel Oil

PHYSICAL AND CHEMICAL CHARACTERIZATION OF FINE
PARTICULATE FROM THE COMBUSTION OF RESIDUAL FUEL OIL

C. Andrew Miller*, William P. Linak
Air Pollution Prevention and Control Division, MD-65
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711 US A

Jost O.L, Wendt
Department of Chemical and Environmental Engineering
University of Arizona
Tucson, AZ 85721 USA

Kevin Dreher

Experimental Toxicology Division, MD-82; National Health and Environmental Effects Research Laboratory;
U.S. Environmental Protection Agency; Research Triangle Park, NC 27711

Abstract

Combustion is a significant source of fine particulate matter (PM) emissions, and heavy oil combustion is
suspected of producing PM emissions with potentially significant toxicity. Ongoing collaborative research
between EPA's National Risk Management Research Laboratory and National Health and Environmental Effects
Research Laboratory is concerned with the characterization of PM emissions from a practical boiler burning
heavy fuel oils. The purpose is to identify hypothetical mechanisms that might relate both the combustion
process and the fuel burned to the size segregated characterization of the fine PM formed, and ultimately to its
propensity to cause pulmonary injury. The data presented, therefore, help suggest specific fundamental issues
which define directions for future research in this area.

In this initial study, samples of PM were taken from the stack of a commercial 732 kW (2.5xl06 Btu/hr) rated
firetube boiler burning four different heavy fuel oils, including two grades and three sulfur contents. Submicron
and supermicron particle size distributions (PSDs) were measured using an in-stack cascade impactor, a scanning
mobility particle sizer (SMPS), and an in-situ light scattering system. Size-classified bulk samples were also
collected using a high volume dilution sampler. In addition to chemical and physical characterization, these
samples were examined for pulmonary toxicity. Finally, EPA Method 5 (total PM) and Method 60 (metal
analyses) samples were extracted and analyzed. Measured PSDs show evidence of a submicron accumulation
mode between 0.07 and 0.08 |im in diameter. PM less than 2.5 (im in diameter (PM2.5) comprised between 30
and 50% of the total PM emissions and contained all of the accumulation mode plus a significant portion of a
broad coarse mode. Small particles less than 2.5 jam in diameter contained significant quantities of metals and
sulfates, while the coarse mode was composed primarily (70-95%) of cenospheric carbon. These and related data
are interpreted in the light of possible new mechanisms governing the partitioning of toxic metals from heavy
oil combustion.

Introduction

Research studies correlating short term exposure of ambient levels of fine particulate matter (PM) with adverse
health effects and the recent revision of the National Ambient Air Quality Standards (NAAQS) for PM by the
U.S. Environmental Protection Agency (EPA) to reduce concentrations of fine PM in the environment have
emphasized the importance of issues associated with fine PM.1 A number of epidemiological studies have been
conducted to identify correlations (if any) between ambient PM concentrations and adverse human health effects.
Many of these studies have concluded that there is a direct correlation between ambient fine PM concentrations
and increases in human mortality and morbidity 2-4 These studies have been summarized by EPA5-6 and
reviewed by EPA's Clean Air Science Advisory Committee (CASAC), 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.7 In July 1997, EPA revised the NAAQS by
amending the existing standards, based on PM less than 10 Jim in diameter (PMjq), by adding standards based on
PM less than 2.5 p.m in diameter (PM2.5)1 The revised PM NAAQS adds two new primary PMj 5 standards: a
65 fig/m3 24-hour average, and a 15 Hg/m3 annual mean, while retaining the previous 24-hour and annual PM10
standards of 150 and 50 Hg/m3, respectively.

Although the evidence of adverse effects at fine PM concentrations at or below the current PM NAAQS has been
determined to be adequate to warrant a revision of the PM NAAQS, it has also been determined that additional
research is necessary to better quantify these effects and the biological mechanisms through which they occur.8
The EPA summary concluded that the increased adverse health effects noted were most closely associated with
the fine PM "generated largely by combustion processes," and highlighted aerosol acidity, ultrafine particles, and
transition metal ions as potential factors influencing mechanisms for these health impacts. One source of PM
that is characterized by this combination of factors is the combustion of heavy fuel oils (e.g., No. 5 and 6 fuel
oils).

Research conducted at EPA's National Health and Environmental Effects Research Laboratory (NHEERL) has
shown that laboratory animals exposed to fine PM generated from the combustion of heavy fuel oil demonstrated
significant adverse health impacts. This research concluded that a major factor leading to these adverse health
impacts was the amounts of water-soluble transition metals such as copper (Cu), iron (Fe), nickel (Ni),
vanadium (V), and zinc (Zn) present in the particles.9'11 Physical characterization and chemical analyses of PM
emissions, and particularly the behavior of trace metals, from pulverized coal combustion have been the subject
of numerous studies;12 however, only limited information regarding the physical and chemical characterization of
PM emissions from heavy fuel oil combustion is available. Therefore, a primary objective of this research was
to characterize PM emissions from heavy oil, and to determine ash particle size and metal partitioning
mechanisms. To coincide with the new PM NAAQS, the focus of this paper is on the collection and
characterization of PM greater than and less than 2.5 ftm in diameter. However, this classification may not
correspond to any natural partitioning caused by mechanisms which describe aerosol particle size distributions
(PSDs) in combustion environments. A further objective of this work was to correlate the physicochemical
characteristics of heavy oil PM with the toxicological properties of the same PM. The detailed health effects
results are now being evaluted, and will be reported at a later date.

Mechanisms of ash transformation, elemental partitioning, and particle formation and evolution during coal
combustion have been attributed to multiple mechanisms in which volatile inorganic species vaporize from
within a burning coal particle to either react with other non-volatile species present in the coal particle or be
released to the vapor environment.12 Depending upon the temperature, pressure, and presence of other reactive
species, these vapor-phase species may then undergo complex chemical reactions. Partial pressures may
approach and exceed critical supersaturation pressures leading to the homogeneous nucleation of inorganic
species and the formation of an aerosol (fume) in the range of 0.05 to 0.5 |lm in diameter, or to heterogeneous
condensation of the vapor onto available surfaces. The magnitude of the supersaturation and presence of existing
surfaces determines whether nucleation or condensation dominates. Regardless of which mechanism dominates,
condensation tends to cause submicron enrichment of vapor-phase inorganic species because submicron particles
typically offer the major fraction of the available surface area. Collision and coagulation of nuclei result in
particle growth, and this process is directly dependent on the number concentrations of the available particles.

These nuclei-nuclei interactions dominate due to the very large nuclei number concentrations available, even
though a relatively small mass of vaporized material may be involved (typically less than 5% of the coal ash).
As a result, nuclei tend to grow very quickly for a short time and then, as number concentrations fall, particle
growth slows considerably causing the aerosol to accumulate into a mode between approximately 0,05 and 0.5
fiiri in diameter. This material and characteristic behavior of the PSD has been termed the accumulation mode.

Because of the relatively small incremental mass added during each coagulation event, additional coagulation
between nuclei and the larger accumulation mode particles does not noticeably affect the PSD. The majority of
the inorganic material present in coal (typically greater than 95%, and composed of extraneous material, micron-

2	PROTECTED UNDER INTERNATIONAL COPYRIGHT

ALL RIGHTS RESERVED.

NATIONAL TECHNICAL INFORMATION SERVICE

U.S. DEPARTMENT OF COMMERCE

-------
sized mineral inclusions, and scavenged volatile species) does not vaporize, but coalesces as the carbonaceous
material recedes and fragments to form particles in the size range of approximately 1 to 30 [im in diameter. This
material and characteristic behavior of the PSD has been termed the coarse mode.

Heavy fuel oils, in contrast to coals, do not typically contain substantial extraneous or included mineral matter,
although the refining process may introduce extraneous materials. The metals in oils are generally inherently
bound within the organic molecule, whereas only a small portion of the metals may be inherent in higher rank
coals. Although inherently bound metals are not necessarily volatile under combustion conditions, non-volatile
elements in heavy oils can contribute to a residual supermicron fraction.13 However, size dependent chemical
analysis indicates that non-volatile metals in heavy oils do contribute substantially to the submicron particle
mass. Additionally, the interactions between volatile metal species and non-volatile minerals prior to the
metals' escaping to the gas phase seen in coals may not be as prevalent for oils, unless there is a mechanism by
which these materials are drawn into proximity with one another during vaporization. Typical fuel oils contain
Ni and V, in addition to aluminum (Al), calcium (Ca), Fe, magnesium (Mg), silicon (Si), and sodium (Na). In
addition, transition metals [Fe, manganese (Mn), and cobalt (Co)] and alkaline-earth metals [barium (Ba), Ca,
and Mg] may be added for soot suppression.14 Metals may also be added to help catalyze the combustion of
residual coke cenospheres or for corrosion control 13

In contrast to the coarse fraction common to PM emissions from the combustion of pulverized coal, the
majority of the sampled flyash mass from residual fuel oil combustion in power plants is likely to lie below 1
|im in diameter (although substantial fractions of larger particles can occur with poor carbon burnout).15 17
Furthermore, studies 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 derived from the acid," and V is known to be a
catalyst for sulfur dioxide (SCh) oxidation to sulfur trioxide (SO3).17 Arsenic (As), Cu, and sometimes Mn have
also been found in the fine PM fractions from heavy fuel and residential heating oil combustion.18 A study of
PM emissions from a large oil-fired power plant found substantial enrichment of both Ni and V in the
submicron PM size fraction.'9 Another study found similar behavior, and suggested that V partitioning depends
on the presence of MgO.20 None of these studies reported detailed PSDs.

Experimental Facilities

North American Package Boiler

Experiments were performed at EPA's Environmental Research Center in Research Triangle Park, NC, using a
North American three-pass firetube package boiler (NAPB), which is a practical, commercially available heavy
fuel oil combustion unit. A schematic of the NAPB is shown in Figure 1. The burner used in the tests (North
American model 6121-2.5H6-A65) has both a natural gas ring and an air-atomizing center nozzle capable of
firing No, 2 through No. 6 fuel oils, and is rated at 732 kW (2,5 x 10® Btu/hr), The main firetube has an inside
diameter of 46 cm (18 in), and each of the 24 second pass tubes and 20 third pass tubes has an inside diameter of
6,4 cm (2.5 in). The boiler can generate up to 1090 kg/hr (2400 Ib/hr) of saturated steam at gauge pressures up
to 103.4 kPa (15 psig). and has 27.9 m2 (300 ft2) of heat transfer surface. An industrial cooling water system
provides a simulated boiler load by extracting heat from the steam through a heat exchanger. Consistent fuel
viscosity is maintained by controlling oil temperature using an in-line electric heater. Oil and atomizing air
pressures are independently controlled to ensure proper oil atomization. The NAPB is instrumented with
continuous emission monitors (CEMs) for carbon dioxide (CO2), carbon monoxide (CO), nitrogen oxides
(NOx), oxygen (O2), sulfur dioxide (SO?), and total hydrocarbons (THCs), which provide signals to a
computerized data acquisition system for continuous recording of CEM measurements and steam and flue gas
temperatures. The flue gases from the unit pass through a manifold to a facility air pollution control system
(APCS). Although this type of boiler normally operates under forced draft, the imposition of an induced draft
(necessary for the APCS operation) did not introduce any significant effects on boiler emissions or operating
conditions.

Several sampling ports are located at the exit of the boiler (see Figure 1) for extractive sampling. Temperatures

3

-------
To air pollution control system -

Sampling location tor
Method 5, Method 60, electrical mobility,
inertia! impactors

*fJh

Sampling location for
dilution sampler, in-situ light scattering

1 m

Sampling location for
gas samples,

Third pass: 20 tubes 6.4 cm I.D.

Fuel oil

Airbloweri

First pass: 1 tube 46cm I.D.

View port

Figure t. Schematic of North American Package Boiier.

at these sampling ports ranged from 180 to 280 °C. The exhaust duct [20.3 cm (8 in) diameter steel pipe] is
sufficient in length and free of flow disturbances so that PM can be sampled at locations that meet Method 1A
PM sampling requirements.21

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.22.23 Metal analyses were
performed using acid digestion and inductively coupled plasma (ICP) mass spectrometry. PSDs were measured
by electrical mobility and inertial impaction for sampled aerosols and light scattering for in-situ analyses. A
large dilution sampler was also used to collect relatively large amounts of size classified PM for use in
NHEERL's animal exposure studies. Finally, PM samples were collected on silver filters for scanning electron
microscope (SEM) analysis.

Total PM and Metal Measurements

Total PM emission levels were determined using an EPA Method 5 train, with three measurements taken per
test condition. EPA Method 60 was also used to measure As, antimony (Sb), beryllium (Be), cadmium (Cd),
chromium (Cr), Cu, Fe, Mn, Mg, mercury (Hg), Ni, V, and Zn concentrations in the flue gas. Only one
Method 60 procedure was conducted per test condition.

Aerosol Particle Size Distribution: Sampling and Analysis

Extracted samples were taken for electrical mobility analyses using an isokinetic aerosol sampling system based
on the modified designs of a research sampler.24.25 The sampling system dilutes and cools the aerosol sample

-------
using filtered nitrogen immediately after sampling to minimize in-probe gas and aerosol kinetics. Dilution
ratios of approximately 5:1 are directly measured for each experiment and can be verified independently by the
measurement of a nitric oxide tracer gas. Diluted samples were directed to a Thermo-Systems-Inc. scanning
mobility particle sizer (SMPS), which classifies and counts particles using principles of charged particle
mobility through an electric field. The SMPS yields 54 evenly spaced (logarithmically) channels over a 0.01 to
1.0 |xm diameter range.

Extracted samples were also taken with an Andersen, Inc. eight-stage, 25 L/min {0.9 fWmin), atmospheric
pressure in-stack cascade impactor. The impactor can collect nine (including the after filter) size-specific samples
less than approximately 10 p.m in diameter for subsequent gravimetric and chemical analysis. A modified
California Air Resources Board (GARB) Method 501 procedure was followed for the impactor sampling.26 The
standard CARB method calls for averaging data from seven sample sets to characterize the PSDs. Due to the
research focus of these experiments (as opposed to the regulatory focus of the CARB method), only three data
sets were taken for each experimental condition. The CARB method also calls for in-situ placement of the
impactor within the boiler exhaust stack and a straight sampling nozzle. In these experiments, a Method 5
goose neck (or button hook) nozzle was used, and the impactor was located external to the boiler stack due to the
small size of the boiler's existing exhaust stack. In addition, the entire impactor and sampling nozzle were heated
to maintain gas temperatures.

In-situ light scattering PSDs were also taken using an Insitec, Inc. laser doppler velocimeter, to allow an optical
size distribution to be taken for comparison with the SMPS measurements by electrical mobility and the
impactor measurements by inertial impaction. The Insitec determines particle size by measuring light scattering
intensity of particles which pass through a sampling volume established within the combustor stack by a laser
focused through a set of quartz optical access ports. The working range of this device was approximately 0.3 to
30 urn in diameter, providing a slight overlap and extension of the PSD data collected by the SMPS.

Samples were also collected for analysis using a SEM equipped with an energy dispersive x-ray (EDX)
spectrometer for morphological information of individual particles. PM was extracted from the stock location
using the same sampling system and dilution as used by the SMPS described above, but was directed through a
stainless steel filter holder containing a 47 mm silver membrane filter. Sampling times of approximately 30-60
s provided a sufficient quantity of PM for analysis.

Dilution Sampler

The large dilution sampling system is capable of isokinctically sampling 0.28 m3/min (10 ft3/min) of flue gas
by using a modified Source Assessment Sampling System (SASS) cyclone and then diluting the sample with
clean ambient air [2.8 m3/min (100 ft3/min)] by means of a perforated cone assembly. The SASS cyclone
produces 50 and 95% PM collection efficiencies at approximately 1.8 and 2.5 }lm diameter, respectively. The
sampled gases are cooled to approximately ambient temperature by rapid uniform dilution within a 3 s residence
time. The resulting PM is collected on large [64.8 cm (25.5 in)] Teflon coated glass fiber filters, transferred to
sampling jars, and made available for subsequent chemical, physical, or biological analysis. This sampling
system was originally designed and constructed to sample PM and condensable emissions from municipal and
hazardous waste incineration facilities.27

Experimental Approach

The fuels and test conditions were chosen such that a range of metal and sulfur contents could be examined, and
the effects of changes in oil viscosity (which changes the resulting spray droplet size distribution) and excess air
on PM emissions and characteristics could be evaluated. A single No. 5 fuel oil was used for three operating
conditions, and three No, 6 fuel oils were examined under a single operating condition to evaluate the role of oil
composition. The three combustion conditions examined for the No. 5 fuel oil were: (1) baseline [120 °C oil
temperature and target stoichiometric ratio (SR) = 1.2 (nominal stack O2 concentration = 3.5% dry )]; (2) low oil
temperature (77 °C) and (baseline) SR = 1.2: and (3) low oil temperature and low SR = 1.1 (nominal stack O2

-------
concentration = 2,5% dry). The three No, 6 fuel oils had sulfur contents of 0.53,0.93, and 2.33% corresponding
to the low, medium, and high sulfur oils, respectively, and were burned at constant combustion conditions (120'
°C oil temperature, nominal SR = 1.2 corresponding to stack O2 concentration = 3.5% dry). A target boiler load
of 586 kW (2,0x106 Blu/hr) was attempted for all six experiments. Actual experimental conditions are
summarized in Table 1. In addition to sulfur content, concentrations of the transition metal contents of each of
the fuel oils were also determined. Table 2 presents ultimate and elemental analyses and physical properties of
the four fuel oils. The V contents of the four fuel oils follow the same sequence (from low to high) as the
sulfur contents of the fuel oils, and vary from 35 to 220 |J,g/g. Fe levels in the three No. 6 fuel oils are similar,
at about 20 p.g/g. Zn concentrations are approximately 4 [lg/g for the low and medium sulfur No. 6 fuel oils,
39 pg/g for the No. 5 fuel oil, and 74 (J.g/g for the high sulfur No. 6 fuel oil. The Cu content varied between
0.56 ng/g for the low sulfur No. 6 fuel oil and 4 jxg/g for the No. 5 fuel oil, and was again closely correlated
with the sulfur content of the four fuel oils. Note that the No. 5 fuel oil shows higher concentrations of ash and
several elements (As, Cu, Fe, Ni) than all three of the No, 6 fuel oils, and higher S, Pb, V, and Zn
concentrations than two of the three No. 6 fuel oils, even though it is often considered a higher quality fuel
because of its viscosity and pumpability characteristics.

The effect of spray droplet size was determined for the No. 5 fuel oil by changing oil temperature and, therefore,
oil viscosity. At 120 °€ (248 °F), the modal diameter of the No, 5 fuel oil spray was approximately 50 pm,
while at 65 °€ (149 °F) it was greater than 100 Jim, as measured by optical light scattering techniques. Fuel oil
and air atomizing pressures used for the Delavan Airo Combustion air atomizing oil nozzle (model 30615-84)
were maintained at approximately equal pressure between 200 and 240 kPa (29 and 35 psig) during boiler
operation. Other variables (except excess air) were also kept constant, by design or by necessity. Table 2 also

Table 1, Combustion conditions and PM emissions for the six experimental conditions.

No. 5 Oil

Low Sulfur

Medium

High Sulfur

Baseline

Low Temp.

Low SRa

No, 6 Oil

Sulfur

No. 6 Oi!

Low Temp.

No. 6 Oil

Actual boiler loadb, kW

590

598

596

606

622

619

Actual stack 02c, % dry

3.40

3,56

2.46

3.60

3.45

Actual SR

1.18

1.19

1.12

1.20

1.19

Oil feed temperature, °C

120

PM emisslonsd

mg/m3

197.7

221.0

221,0

219.5

243.5

183.6

(std. dev.)

(4.9)

(6.2)

(22.6)

(51.1)

(59.9)

(6.2)

PM emission rate

lb/108 Btu

0.145

0.151

0.143

0.142

0.161

0.123

kg/MJ

6.23x10-5

6.49x10-5

6,15x10-5

6.11x10-5

6.92x10-5

5.29x10 5

LOIe, wt.%

Filter

64.1

78.5

75.3

65.8

79.0

86.6

Cyclone

88.3

87.7

93.5

90.3

95.8

96.9

^Stoichiometric ratio

^Target boiler loads were 586 kW (2.0x108 Btu/hr) for all experiments

cTarget stack 02 was 3.5% for all experiments, except for 2.5% for the low SR experiments

dpM emissions are based on the average of three replicate measurements

°Loss on ignition

-------
presents kinematic viscosities for the four fuel oils at 38 and 99 °C (100 and 210 °F). As expected, these data
show the No. 5 fuel oil to have significantly lower viscosities compared to the No. 6 fuel oils at the same
temperatures. It is interesting to note that the high sulfur No. 6 fuel oil has a notably lower viscosity compared
to the other two No. 6 fuel oils.

Table 2. Analyses of the four fuel oils examined.



No. 5 Oil

Low Sulfur
No. 6 Oil

Medium Sulfur
No. 6 Oil

High Sulfur
No. 6 Oil



Ultimate Analysis, wt.%





Carbon

86.36

86,00

86.48

85.49

Hydrogen

10.82

11.29

10.98

10.36

Nitrogen

0.33

0.43

0.43

0.35

Sulfur

1.73

0.53

0.93

2.33

Oxygen8

0.34

1.24

0.67

0.92

Moisture

0.35

0.50

0.50

0.50

Ash

0.07

0.02

0.03

0.10



Elemental Analysis, ng/g





Arsenic

2

0.2

0.2

0:1

Beryllium

<1

<0.3

<0.3

<0.3

Cadmium

0.1

0.50

0.60

0.60

Chromium

0.5

1.08

0.96

1.05

Copper

4

0.56

0.78

3.5

Iron

50

23

19

21

Lead

3

0.80

0.58

4.5

Mercury

Jb

0.06

0.12

0.10

Nickel

34

17

22

30

Selenium

<2

<0.1

<0.1

<0.1

Vanadium

180

35

70

220

Zinc

39

4.11

3.70

74

Specific grav.

0.9541

0.9554

0.9652

0.9916

Kinematic viscosity, cSt
@38 °C
@99 "C

81.5
8.4

865
41

843
37

560
25

HHVC, kcal/kg

10,230

10,460

10,390

10,150

BDetermined by difference
bNot available
cHigher heating value

7

-------
Results and Discussion

Total PM Emissions

Table 1 presents the total PM emissions (triplicate averages and standard deviations) from the six experimental
conditions. Emission rates for the No, 5 fuel oil were similar for all three experimental conditions, at approximately
6.0x10-5 kg/MJ (0.14 lb/106 Btu), indicating that the effects of oil spray droplet diameter and in SR examined were
not sufficiently large to significantly affect the total PM emission rates. There was somewhat more variation in the
PM emissions for the three No. 6 fuel oils. Emission rates for the three No. 6 oils ranged from 5.3x10"5 kg/MJ
(0.12 lb/106 Btu) for the high sulfur No. 6 fuel oil to 6.9x10 5 kg/MJ (0.16 lb/106 Btu) for the medium sulfur No. 6
fuel oil. The high sulfur No. 6 fuel oil had the lowest measured PM emissions but the highest ash content (0.1
wt.%), which is consistent with findings that V catalyzes carbon oxidation.28 The PM emission factors obtained
here were approximately twice those of oil fired utility boilers,29 with the difference likely because of poor
cenospherc burnout in this small firetube boiler. Given the differences in scale and design between a utility boiler
and the firetube boiler used here, such an increase is not unexpected.

As measured by loss on ignition (LOI), the predominant element in the total PM mass emission for all the fuel oils
was carbon, with carbon contents of the PM ranging between 64 and 97% by weight. Table 1 presents the percent
LOI for each of the experimental conditions. Although the LOI values are quite high, the calculated combustion
efficiencies in each case (total carbon consumed divided by total carbon input) were at least 99.7%. Samples were
taken from the cyclone and filter catches of the large dilution sampler. The LOI for the cyclone samples was
consistently higher than for the filter samples, indicating that a larger fraction of carbon was associated with PM
larger than approximately 2 urn in diameter.

Size-Fractionated PM Emissions

PSDs of the PM measured in the boiler stack are shown in Figures 2, 3, 4, and 5 for the six experimental
conditions. Figures 2 and 3 present typical volume PSDs taken by electrical mobility (SMPS) and in situ light
scattering methods for the No. 5 and 6 fuel oil experiments, respectively. Figures 4 and 5 present mass fraction
PSDs determined from cascade impactor samples (gravimetric) for the No. 5 and 6 fuel oil experiments, respectively.
Figures 2 and 3 illustrate that there is a distinct submicron mode at approximately 0.07-0.08 Jim in diameter for all
experimental conditions, as well as broad supermicron modes extending past 50 (im in diameter. Although these
types of measurements do not easily allow one to accurately determine absolute number and volume distributions
from these types of measurements, valuable information indicative of changes between the experimental results are
often indicated by the relative differences between PSDs. The submicron PSDs for the three No. 5 fuel oil
experimental conditions (baseline, low oil temperature, and low SR and low oil temperature) are nearly identical, as
seen in Figure 2, and the supermicron PSDs are similar for the baseline and low oil temperature conditions. Large
repeatable increases in the larger particle volume distribution are seen for the low SR and low oil temperature
conditions, although similar large increases were not seen for the corresponding total PM emissions (Table 1). Note
that the in-situ particle sizer measures light scattering, which depends on particle cross section and refractive index
(presumably related to particle volume and morphology), rather than particle mass. The SF.M analysis of the PM
structure shows many large porous spherical structures and fragments which, along with the measured LOIs, support
the hypothesis that these low density structures are likely residual fuel char formed from individual fuel oil droplets.
This may explain the apparent inconsistencies between the total PM mass and light scattering measurements;
although these particles exhibit large effective cross sections for optical light scattering, they are extremely porous
and contribute relatively little to the PM mass. Note that the large supermicron volume increase seen in Figure 2 is
a result of only a very small increase (<1 /cm3) in the corresponding supermicron particle number concentration.

The cascade impactor data presented in Figure 4 and the SMPS data in Figure 2 are consistent, although the impactor
data are much less resolved. Hie PSDs for all three conditions burning the No. 5 fuel oil are bimodal, with a
submicron mode somewhere less than 0.3 p.m in diameter (after filter) and a supermicron mode increasing from less
than 1 pm in diameter. While the light scattering techniques determine particle size based on effective optical cross
section, inertia! impaction devices determine aerodynamic particle size usually calibrated for spherical unit density
particles. Particles such as low density hollow cenospheres which vary significantly from this ideal condition may

-------
Dp
-------
%
CO

Q.
Q
o>
o

0.01 0,1 1 10 100

Dp frtm)

Figure 3. Particle volume distributions measured by electrical mobility arid light scattering for No. 6 fuel oil
combustion: low sulfur; medium sulfur; high sulfur,

four submicron distributions consistently correlate inversely with sulfur content and may indicate an effect of sulfur
on the submicron PSD.

The bimodal PSDs presented here are consistent with a mechanism of metal vaporization/nucleation/coagulation/
condensation and formation of residual carbon cenospheres. The combustion process involves the atomization of fuel
oil, producing a broad initial droplet size distribution ranging from <10 to >100 um in diameter. As the fuel
droplets are heated, the lighter organic components vaporize, diffuse away from the fuel droplet, and combust in a
diffusion flame surrounding the droplet. The heavier organic fractions remain to form a char that continues to react
as oxygen diffuses into the solid structure. These spherical particles (cenospheres), composed primarily of carbon and
ash, may be comparable to the original fuel droplet in size, and are often hollow and, therefore, low in density.

These cenospheres may fragment or remain intact to form the coarse mode (supermicron) PSD. The metal
components may vaporize and nucleate or condense (depending on temperature, speciation, volatility, and physical
structure within the organic matrix) as noted earlier. Condensation on submicron particles is preferred because these
particles typically dominate the available surface area. The distinctive submicron peaks seen in Figures 2 and 3
indicate particles formed by nucleation, coagulation, and condensation of materials that have vaporized in the high
temperature region of the combustor. These metal transformation and particle formation mechanisms in combustion
systems have been reviewed by Linak and Wendt.,230

Trace Metals

Table 3 presents trace metal emissions for the six experimental conditions examined. Even though the data presented
are based on analyses of single samples, data reproducibility is indicated by comparing the metal analyses for the
three No. 5 fuel oil experiments, since the same fuel was used for all three conditions. Although these three
experiments examined different combustion conditions, the changes in combustion condition are not likely to have

-------
0.3.

0.2.

(a) No. Soil - baseline

0.1.

0.0.
® 0.3.

CJ>
to

(b) No. S oil - low oil temp.

I 0.2.

.§

c
o
c

S3
Q

52
cd

2 0.0M

0.1.

0.3.

02m

0.1."

0.0.

0.1 1 10

Aerodynamic particle diameter (urn)

100

Figure 4. Particle mass fraction distributions measured by gravimetric inertial impaction for No. 5
fuel oil combustion: (a) baseline; (b) low oil temperature; and (c) low SR and low oil
temperature.

a significant effect on the total trace metal emissions. Except for the baseline Cd value, the No. 5 fuel
oil data indicate remarkably good metal data reproducibility. Major metal species (V, Ni, Fe, Mg, and
Zn) correlate approximately with the metal concentrations determined from the four fuel oils (Table 2).
However, even the highest metal constituent, V, is at most 5% of the total PM emissions.

Comparing V and Ni emissions with V and Ni concentrations in the fuel oils indicates a mass balance
closure of approximately 50% for those species, with the closure for some other metal species (e.g.,
As) being significantly lower. Ash deposition on the boiler tubes is believed to account for much of
the missing ash. The samples were also analyzed for Hg but, except for small amounts detected in the

-------
0.3.

0.2.

B
w

a
,|

c
o

0.1

0.0
0.3

0.2

(a) LowS No. 6 oil

§ 0.1

•s
5

0.0
0.3

(b) Medium SNo. Boil

c) High S No. 6 oil

1 10

Aerodynamic particle diameter (nm)

100

Figure 5. Particle mass fraction distributions measured by gravimetric inertia! impaction for No.

6 fuel oil combustion: (a) low sulfur; (b) medium sulfur; and (c) high sulfur.

front half of the Method 60 train for the low sulfur No. 6 fuel oil sample, Hg was not detected in any flue gas
sample. Table 3 presents the upper limit of measured mercury emissions based on method detection levels.

Tables 4 and 5 present trace metal and carbon concentrations (as determined by LOI) for the filter and cyclone samples
of the dilution sampler for No. 5 and 6 fuel oils, respectively. Four columns represent data for each element for each
experiment. The first column reports metal concentrations in micrograms of metal per gram (jxg/g) of filter sample,
containing particles which have passed through the dilution sampler cyclone with a 50% particle cutoff efficiency
(Dpso) of 1.8 |im diameter (which corresponds to a Dpgs of 2.5 |im diameter). The second column provides the
composition of the particles retained by the cyclone (i.e., greater than 1.8 fim in diameter). The third column
presents the ratio of concentrations in the filter sample to those in the cyclone sample, with values of this ratio
greater than 1.0 indicating element enrichment in small particles and values less than 1.0 indicating element

-------
depletion. The fourth column presents values of the filter to carbon (F/C) ratio that have been modified to account
for the level of carbon in each sample.

Carbon was in all cases the most abundant element in both filter and cyclone samples, which can greatly affect the
elemental enrichment in small particles. Because the metal contents are presented in |lg/g, changing only the LOI
level (assumed to represent the carbon content) can have a substantial impact on the reported metal concentrations,
with higher levels of carbon resulting in a lower relative amount of trace element. The LOIs of the filter samples
(although greater than 60%) were always less than those of the cyclone samples, and this fact alone led to apparent
enrichment of other elements in the smaller particles (see Miller et al.3* for further discussion of this issue). The
modified F/C ratios are near 1.0 for most cases, indicating little enrichment or depletion of elements by means other
than through the dilution by carbon. Analysis of the impaetor after filters (Dp < 0.3 jim in diameter) showed that
this material, which better isolates the mass associated with the accumulation mode, consisted of between 50 and
90% metals and sulfates. Average analyses indicate approximately 21% sulfates, 18% Mg, 14% V, 10% Fe, 2% Ni,
1.5% Zn, and smaller percentages of other metals collected on the after filter.

With the possible exception of Zn for the No. 5 fuel oil, Tables 4 and 5 show that in general the partitioning of the
trace metals appears to be primarily influenced by the carbon content. However, Figures 2 and 3 (and Figures 4 and
5) indicate distinct PSD accumulation modes, which are normally attributed to partitioning processes associated with
vaporization followed by nucleation, coagulation, and condensation. The modified F/C values observed are most
likely due to the fact that the tail of the coarse PM contributes substantially to the mass of the filter fraction. Very
low levels of carbon were noted on the after filters (Dp < 0.3 |im diameter), which showed much higher
concentrations of the metals, indicating that vaporization/condensation accumulation is the dominant process in

Table 3. Trace metal emissions for the six experimental conditions (jig/m3).

No. 5 Oil
Baseline

No. 5 Oil
Low Temp.

No. 5 Oil
Low SRa
Low Temp.

Low
, Sulfur
No. 6 Oil

Medium
Sulfur

No. 6 Oil

High Sulfur
No. 6 Oil

Antimony

5.7

5.3

6.7

2.6

2.5

• 7,7

Arsenic

3.4

4.3

3.6

7.4

9.3

6.3

Beryllium

0.10

0.11

0.12

0.05

0.06

0.09

Cadmium

3.7

0.40

0.36

0.19

0.21

3.5

Chromium

7.2

8.2

6.5

5.5

6.6

Copper

170

Iron

530

500

430

700

710

740

Lead

Magnesium

250

240

200

220

210

1200

Manganese

Mercury

<1.8

<2.0

<1.9

<1.8

<2.1

<2.2

Nickel

1400

1300

1100

, 470

710

1200

Vanadium

8600

8300

6900

1600

3700

9800

Zinc

210

180

3300

aStoichiomctric ratio
t>Not available

-------
Table 4. Trace element concentrations and enrichment ratios of filter and cyclone samples from dilution sampler for No. 5 fuel oil experiments (|J.g/g).a

No. 5 Oil
Baseline

No. 5 Oil
Low Oil Temperature

No. 5 Oil
Low SRb, Low Oil Temperature

Filter

Cyclone

F/C
Ratio0

Mod. F/C
Ratiod

Filter

Cyclone

F/C Ratio

Mod. F/C
Ratio

Filter

Cyclone

F/C Ratio

Mod. F/C
Ratio

Antimony

34.5

4.86

7.1

2.3

18.1

3.81

4.8

2.7

30.2

3.9

7.7

2.0

Arsenic

18.7

1.70

3.6

8.1

1.39

5.8

3.3

16.3

1.46

2.9

Beryllium

0.44

0.20

2.2

0.7

0.22

0.20

1.1

0.6

0.34

0.20

1.7

0.4

Cadmium

2.75

0.69

4.0

1.3

1.48

0.49

3.0

1.7

2.51

0.45

5.6

1.5

Chromium

60.5

33.3

1.8

0.6

47.0

29.1

1.6

0.9

41.4

24.5

1.7

0.4

Copper

233

58.1

4.0

1.3

131

43.1

3.0

1.7

174

34.5

5.0

1.3

Iron

4220

1110

3.8

1.2

2270

820

2.8

1.6

2650

697

3.8

1.0

Lead

. <

Magnesium

1770

101

5.7

211

183

1.2

0.7

485

166

2.9

0.8

Manganese

89.9

23.6

3.8

1.2

47.8

20.1

2.4

1.4

60.0

19.3

3.1

0.8

Nickel

10600

2200

4.8

1.6

5710

1660

3.4

2.0

7330

1350

5.4

1.4

Vanadium

58600

13000

4.5

1.5

33700

9910

3.4

2.0

42000

8320

5.0

1.3

Zinc

2750

6530

0.4

0.1

1430

5130

0.3

0.2

2520

4390

0.6

0.2

LOIf, mg/g

641

883

0.7

N/A9

785

877

0.9

N/A

753

935

0.8

N/A

aCyclone preseparator produces 50 and 95% particle collection efficiencies at approx. 1.8 and 2.5 jam diameter, respectively
bStoichiometric ratio
cFilter/cyclone mass ratio

dModified filter/cyclone mass ratio -- corrected for the loss of mass on ignition which is presumed to represent the sample carbon content
eNot available
fLoss on ignition
9N/A - not applicable

-------
Table 5. Trace element concentrations and enrichment ratios of filter and cyclone samples from dilution sampler for No. 6 fuel oil experiments (jj.g/g).a

Low Sulfur
No. 6 Oil

Medium Sulfur
No. 6 Oil

High Sulfur
No. 6 Oil

Filter

Cyclone

F/C
Ratiob

Mod. F/C
Ratio0

Filter

Cyclone

F/C Ratio

Mod. F/C
Ratio

Filter

Cyclone

F/C Ratio

Mod. F/C
Ratio

Antimony

23.4

4.90

4.8

1.4

24.2

2.9

8.3

0.9

48.6

8.20

5.9

1.4

Arsenic

49.9

11.0

4.5

1.3

49.8

4.9

10.

1.1

35.9

8.60

4.2

1.0

Beryllium

0.40

0.10

4.0

1.1

0.47

0.11

4.3

0.4

0.46

0.15

3.1

0.7

Cadmium

0.50

0.21

2.4

0.7

1.26

0.46

2.7

0.3

19.3

1.84

2.4

Chromium

32.6

27.5

1.2

0.3

44.7

46.9

1.0

0.1

60.2

41.3

1.5

0.3

Copper

123

33.8

3.6

1.0

159

36.8

4.3

0.5

1050

222

4.7

1.1

Iron

5100

1410

3.6

1.0

4460 ,

1510

3.0

0.3

3850

2300

1.7

0.4

Lead

114

21.5

5.3

1.5

164

22.4

7.3

0.8

990

94.2

2.4

Magnesium

1450

428

3.4

1.0

1450

436

3.3

0.3

6190

2220

2.9

0.6

Manganese

93.3

34.1

2.7

0.8

84.5

37.1

2.3

0.2

73.2

42.8

1.7

0.4

Nickel

4840

863

5.6

1.6

7470

1230

6.1

0.6

8020

2270

3.5

0.8

Vanadium

14700

4510

3.2

0.9

35300

7560

4.7

0.5

58900

19900

3.0

0.7

Zinc

1600

328

4.9

1.4

1840

422

4.4

0.5

21000

2740

7.7

1.8

LOId, mg/g

658

903

0.7

N/A®

790

978

0.8

N/A

866

969

0.9

N/A

aCyclone preseparator produces 50 and 95% particle collection efficiencies at approx. 1.8 and 2.5 nm diameter, respectively
bFilter/cyclone mass ratio

cModified filter/cyclone mass ratio ~ corrected for the loss of mass on ignition which is presumed to represent the sample carbon content
dLoss on ignition
eN/A - not applicable

-------
formation of these small particles. These results are consistent with full scale power plant data of Bacci et al.,19 who
showed significant enrichment of Ni and ¥ on small particles, since large power plants are expected to emit
significantly less carbon than was the case for this small firetube boiler.

PM Metal Solubility

As noted earlier, health effects research has shown that the presence of water-soluble transition metals (such as Cu,
Fe, Ni, V, and Zn) are strongly suspected of leading to adverse health effects when they are inhaled.9-11 The results
presented here have shown that the fine fraction (< 1.8 Jim in diameter), and particularly the after filters (< 0,3 fim in
diameter), have relatively high levels of V, Fe, Ni, and Zn. Tests by NHEERL have shown these metals to be
highly water-soluble, as shown in Figure 6. These data show the transition metals {V, Fe, Ni, and Zn) that are
soluble in distilled deionized water, as a percentage of the total mass of those metals in the fine particle fraction for
the six test conditions. Also shown on Figure 6 are the transition metals in the coarse fraction for the high sulfur
No. 6 oil and the baseline No. 5 oil test condition, again as a percentage of the total mass of the transition metals in
the particles.

The high level of water-soluble transition metals in these samples suggests that, based on the hypothesis that such
metals play a key role in causing acute adverse health effects, the PM from these (and similar) oils may be important
from a health perspective. Note that, although the solubility is higher for the coarse fractions, the total amount of
water-soluble metals is substantially lower since the total metals in the coarse PM fraction is also substantially
lower. For the No. 5 oil, the total amounts of V, Fe, Ni, and Cu in the coarse fraction are roughly a quarter of the
total amounts of those metals in the fine fraction (Table 4). Therefore, even the higher percentage of water-soluble
metals for the coarse fraction would result in lower total amounts of water-soluble metals than for the low SR
condition, which has the lowest solubility of the fine fraction samples. In general, one would suspect that the fine
fraction of the PM from these oils under the combustion conditions used in these tests is likely to exhibit health
effects behavior similar to the exposure tests conducted previously.®-1'

Conclusions

Approximately 30 to 50% of the PM mass emissions from heavy fuel oil combustion reside in particles smaller than
1.8 jam in diameter. Of all PM, 80 to 90% consisted of carbon, even though combustion efficiencies were greater
than 99.7%. Emission rates for this small firetube boiler were approximately twice those from large utility scale
oil-fired boilers, mostly due to the high carbon content of the PM.

PSDs measured by three different experimental techniques show definite evidence of a bimodal distribution,
consisting of an accumulation mode at 0,07-0.08 (Am in diameter and a broad coarse mode extending beyond 100 p.m.
The coarse mode consists primarily of low density, porous, carbonaceous cenospheres, PM2.5 samples contain both
the accumulation and a significant fraction of the coarse mode, and their composition is determined by a hybrid of
two very different processes. Therefore, the enrichment caused by metal vaporization/condensation processes is
diluted by variations in the major constituent, carbon. This enrichment is dependent upon carbon oxidation processes
rather than on variations in metal volatility. For fuel oils, the metals are for the most part inherently bound within
the organic structure and are not associated with large inorganic ash inclusions as in the case of coal. Thus, when
carbon burnout is poor, substantial portions of the trace metals will remain in the unbumed cenospheres rather than
undergoing the vaporization/nucleation/condensation processes, and the total fraction of metals will appear to be less
on a percentage basis than when carbon burnout is more complete. Thus, the carbon oxidation may play a larger role
than do metal volatility variations that are a primary factor in determining metal partitioning in coal combustion.

Processes governing fine PM emissions from residual oil combustion are different from those from pulverized coal in
a number of respects. Most important is the fact that coal contains significantly more ash, resulting in substantially
lower carbon fractions for similarly efficient combustion processes. This suggests significant differences in the
pathways followed by the trace metals contained in the two fuel types and can confound mechanistic interpretation of
size speciated data.

-------
120

100-

80-

cr-

_3
O

ffi
©

60-

40-

20-

No. 5 Oil

No, 6 Oil

111

.£
_

5
o

E ©

OJ cd
A 0Q

(f)

g>
I

13
©

3
a=
3
C/D

3. 3

m co
c\i r

a CD

Figure 6. Percent solubility of metals from the fine fraction (< 2.5 urn) in distilled water for each of the six
test conditions, and from the coarse fraction from the baseline No. 5 oil condition and the high
sulfur No. 6 oil.

The transition metals were highly water-soluble for all conditions, with the highest solubility exhibited by the coarse
particle fractions tested; For the No. 6 oils, the water solubility of the transition metals was greater than 80% for all
the PM samples tested, and for the No. 5 oil, the water solubility was greater than 40% for all samples tested, and
averaged over 60% for the four samples. The high level of transition metal solubility in water suggests that the PM
from these oils, under these conditions, are likely to exhibit similar toxicity behavior as samples of residual oil fly
ash from full scale utility boilers.

Future research should focus on isolation and chemical analyses of the accumulation mode without any confounding
effects of the coarse mode. Furthermore, the toxicity attributable to both modes should be investigated together and
separately. The relationship between toxicology and PM speciation is critical and under investigation. Models that
allow prediction of metal partitioning during oil combustion will require additional information on processes that
govern how organically bound metals evolve.

Acknowledgments/Disclaimer

Portions of this work were conducted under EPA Purchase Order 7D0222NATX with J.O.L. Wendt and EPA

-------
Contract 68-D4-00G5 with Acurcx Environmental Corp. (now ARCADIS Geraghty & Miller). The authors
gratefully acknowledge the contributions of P. Groff of EPA's Air Pollution Prevention and Control Division,
and C. King, C. Elmore, and D. Janek of ARCADIS Geraghty & Miller to the experimental efforts. The
research described in this article has been reviewed by the Air Pollution Prevention and Control Division, U.S.
Environmental Protection Agency, and approved for publication. The contents of this article should not be
construed to represent Agency policy nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.

References

1. Federal Register, 62 PR 38652, My 18, 1997.

2. Dockery, D.W., Pope, C.A., III, Xu, X., Spengfer, J.D., Ware, J.H.. Fay, M.E., Ferris, B.G., Jr., and
Speizer, F.E., "An association between air pollution and mortality in six U.S. cities," New England J.
Med., 329, 1753-1759 (1993).

3. Burnett, R.T., Dales, R., Krewski, D., Vincent, R., Dann, T., and Brook, J.R., "Associations between
ambient particulate sulfate and admissions to Ontario hospitals for cardiac and respiratory disease," Am. J,
Epidemiology, 142, 15-22 (1995).

4. Schwartz, J., and Morris, R., "Air pollution and hospital admissions for cardiovascular disease in Detroit,
Michigan," Am. J. Epidemiology, 142, 23-35 (1995),

5. Bachmann, J.D., Damberg, R.J., Caldwell, J.C., Edwards, C,, and Koman, P.D., "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 (1996).

6. U.S. Environmental Protection Agency, "Air Quality Criteria for Particulate Matter," EPA-600/P-95/00lcF
(NTIS PB96-168224), U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Research Triangle Park, NC (1996).

7. Wolff, G.T., Closure by the Clean Air Scientific Advisory Committee (CASAC) on the staff paper for
particulate matter, EPA-SAB-CASAC-LTR-96-008, U.S. Environmental Protection Agency, Washington,
DC, June 13, 1996.

S. National Research Council, "Review of EPA's Particulate Matter Research Program," National Academy
Press, Washington, DC, 1998.

9. Dreher, K., Costa, D., Hoffman, A., Bonner, J., and Osornio-Vargas, A., "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 1996a.

10. Dreher, K„ Jaskot, R„ Richards, J.H., Lehmann, J.R., Winsett, D„ Hoffman, A„ and Costa, D„ "Acute
pulmonary toxicity of size-fractionated ambient air particulate matter," Amer. Jour, Respir. Grit. Care Med.,
153(4:A15) (1996b).

11. Dreher, K., Jaskot, R., Lehmann, J.R., Richards, J.H., McGee, J.K., Ghio, A.J., and Costa, D.L., "Soluble
transition metals mediate residual oil fly ash induced lung injur)'," J. Toxicol. Environ. Health, 50, 285-305
(1997).

12. Linak, W.P., and Wendt, J.O.L., "Trace metal transformation mechanisms during coal combustion," Fuel
Process. Techno!.. 39, 173-198 (1994).

13. Feldman, N„ "Control of residual fuel oil particulate emissions by additives," 19th Symp, (Int.) on Comb.,
1387-1393, Combustion Institute, Pittsburgh, PA (1982).

14. Bulewicz, E.M., Evans, D.G., and Padley, P.J., "Effect of metallic additives on soot formation processes in
flames," 15th Symp. (Int.) on Comb., 1461-1470, Combustion Institute, Pittsburgh, PA (1974).

15. Hersh, S,, Piper, B.F., Mormile, D.J., Stegman, G„ Alfonsin, E.G., and Rovesti, W.C., "Combustion
demonstration of SCR II fuel oil in a utility boiler," ASME Winter Annual Meeting, 79 WA/Fu-7, New
York, NY, December 2-7,1979.

16. Piper, B„ and Nazimowitz, W,, "High viscosity oil evaluation, 59th street station - unit 110," Vol. 1, KVB
report to Consolidated Edison Co., 21640-1 (March 19S5).

17. Walsh, P.M., Wei, G., and Xie, J., "Metal oxide and coke particulates formed during combustion of residual
fuel oil," 10th Annual Meeting American Assoc. for Aerosol Research, 7P.36, Traverse City, MI (October

-------
1991).

IS, Buerki, P.R., Gaelli, B.C., and Nyffeler, U.P., "Size-resolved trace metal characterization of aerosols emitted
by four important source types in Switzerland," Atmos. Environ., 23(8), 1659-1668 (1989).

19. Bacci, P., Del Monte, ML, Longhetto, A., Piano, A., Prodi, F., Redaelli, P., Sabbioni, C., and Ventura, A.,
"Characterization of the particulate emission by a large oil fuel fired power plant," J. Aerosol Sci., 14(4),
557-572 (1983).

20. 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., "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, EPRITR-103990, 3-132-3-149 (1994).

21. EPA Test Method 1A - Sample and Velocity Traverses for Stationary Sources with Small Stacks or Ducts,
in 40 CFR Part 60 Appendix A, Government Institutes Inc., Rockville, MD, July 1994.

22. Garg, S., EPA Method 0060 - Methodology for the Determination of Metals Emissions in Exhaust Gases
from Hazardous Waste Incineration and Similar Combustion Processes, in Methods Manual for Compliance
with the BIF Regulations: Burning Hazardous Waste in Boilers and Industrial Furnaces, EPA/530-SW-91-
010 (NTIS PB91-120006), pp. 3-1 through 3-48, Washington, DC, December 1990.

23. EPA Test Method 5 - Determination of Particulate Emissions from Stationary Sources, in 40 CFR Part 60
Appendix A, Government Institutes Inc., Rockville, MD, July 1994.

24. Scotto, M.A., Peterson, T.W., and Wendt, J.O.L., Hazardous waste incineration: the in-situ capture of lead
by sorbents in a laboratory down flow combustor, 24th Comb. (Int.) Symp., 1109-1118, Comb. Inst,
Pittsburgh, PA (1992).

25. Linak, W.P., Srivastava, R.K., and Wendt, J.O.L., Metal aerosol formation in a laboratory swirl flame
incinerator," Combust. Sci. Technol, 101(1-6), 7-27 (1994).

26. CARB Method 501 - Determination of Size Distribution of Particulate Matter Emissions from Stationary
Sources. State of California Air Resources Board Stationary Source Test Methods: Volume 1 - Methods for
Determining Compliance with District Nonvesicular (Stationary Source) Emission Standards. Adopted
March 23,1988; amended September 12,.1990.

27. Steele, W.J., Williamson, A.D., and McCain, J.D., "Construction and operation of a 10 cfm sampling
system with a 10:1 dilution ratio for measuring condensable emissions," EPA/600-8-88 069 (NTIS PB88-
198551), Research Triangle Park, NC, April 1988.

28. Bachman, K.C., and Siegmund, C.W., "The effect of fuel oil composition on particulate emissions," Joint
Meeting of the Pulverized Fuel and Oil Panels of the International Flame Research Foundation, Amsterdam,
September 13-14, 1979.

29. U.S. Environmental Protection Agency, Compilation of Air Pollutant Emission Factors Volume 1:
Stationary Point and Area Sources, AP-42, 5th Ed. (NTIS PB95-196028), U.S. EPA, Office of Air Quality
Planning and Standards, Durham, NC, November 1996.

30. Linak, W.P., and Wendt, J.O.L., "Toxic metal emissions from incineration: mechanisms and control," Prog.
Energy Combust. Sci., Vol. 19, 145-185 (1993).

31. Miller, C.A., Linak, W.P., King, C., and Wendt, J.O.L., "Fine particle emissions from heavy fuel oil
combustion in a firetube package boiler," Comb. Sci. and Tech., Vol. 134, 477 (1998).

-------
t«td i/rij t dtti -d o °n TECHNICAL REPORT DATA

INxt IVJJ \,.L—rt 1 1 i o Ou (Please read Instructions on the reverse before completing)

,REPORTNf00/A-98/135

3. RECIPIENT'S ACCESSION NO.

4. TITLE AND SUBTITLE

Physical and chemical characterization
of Fine Particulate from the Combustion of
Residual Fuel Oil

S. REPORT DATE

6. PERFORMING ORGANIZATION COOE

7. author^ . Miller and W. Linak (EPA, APPCDJ, J.
Wendt (Univ. of AZ), andK.Dreher (EPA.ETD)

I llll III Willi III"™°'

PB99-121477

9. PERFORMING ORGANIZATION NAME AND ADDRESS

University of Arizona, Tucson, AZ 85721.
EPA,Experimental Toxicology Div, .National Health
and Environmental Effects Research Laboratory,
Research Triangle Park, NC 27711.

10. PROGRAM ELEMENT NO.

,fefrrdGW2l2NATX

(Univ. of AZh EPA 68"D4~
0005 (Acurex)

12. SPONSORING AGENCY NAME AND ADDRESS

EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 277E

13. TYPE OF REPORT AND PERIOD COVERED

Published paper; 9/96-6/98

14. SPONSORING AGENCY CODE

EPA/600/13

15.supplementary notes EpA project officer is C. Andrew Miller, Mail Drop 65, 919/541-
2920. For presentation at the Conference on Air Quality: Mercury, Trace Elements,
and Particulate Matter, McLean. VA, 12/1-4/98.

i6. abstract paper gives results of research concerned with the characterization of
particulate matter (PM) emissions from a practical boiler burning heavy fuel oils.
The purpose is to identify hypothetical mechanisms that might relate both the com-
bustion process and the fuel burned to the size-segrated characterization of the fine
PM formed, and ultimately to its propensity to cause pulmonary injury. The data
presented, therefore, help suggest specific fundamental issues which define direc-
tions for future research in this area. In this initial study, PM samples were taken
from the stack of a commercial 732 kW (2. 5 MBtu/hr) rated firetube boiler burning
four different heavy fuel oils, including two grades and three sulfur contents. Sub-
micron and supermicron particle size distribution® (PSDs) were measured using an
in-stack cascade imp actor, a scanning mobility particle sizer, and an in-situ light
scattering system. Size-classified bulk samples were also collected using a high
volume dilution sampler. In addition to chemical and physical characterization,
these samples were examined for pulmonary toxicity. Finally, EPA Method 5 (total
PM) and Method 60 (metal analyses) samples were extracted and analyzed. Meas-
ured PSDs show evidence of a submicron accumulation mode between 0.07 and 0.08
micrometer in diameter.

17. KEY WORDS ANO DOCUMENT ANALYSIS

a. DESCRIPTORS

b.IDENTIFIERS/OPEN ended terms

c. cosati Field/Group

Pollution
Particles
Residual Oils
Combustion
Boilers

Respiratory Diseases

Pollution Control
Stationary Sources
Particulate

13 B
14G

11H, 2 ID
2 IB
13A
06E

18. distribution statement

Release to Public

19. SECURITY CLASS {This Report)

Unclassified

21. NO. OF PAGES
19

20. SECURITY CLASS (This page)

Unclassified

22. PRICE

EFA Form 2220-1 (9-73)

-------