&EPA
United States
Environmental Protection
Agency
Environmental Sciences
Research Laboratory
Research Triangle Park
NC 27711
EPA-600/9-78-020b
August 1978
Research and Development
Workshop Proceedings on
Primary Sulfate Emissions
from Combustion Sources
Volume 2
Characterization
>
•:-.«
'.
"i:
I
-
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research arid Development
8. "Special" Reports
9. Miscellaneous Reports
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/9-78-020b
August 1978
Workshop Proceedings on
Primary Sulfate Emissions
from Combustion Sources
Volume 2
Characterization
Sponsorship by
U.S. Environmental Protection Agency
April 24-26, 1978
Southern Pines, North Carolina 28387
Coordination and Editing by
Kappa Systems, Inc.
Arlington, Virginia 22209
Workshop Chairman
John S. Nader
Emission Measurements and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessari-
ly reflect the views and policies of the U.S. Environmental Protec-
tion Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
In general, the technical content of the papers included in this
report have been reproduced in the form submitted by the authors.
Any papers included in the Program and not included herein were not
submitted for publication.
-------
Preface
This volume contains the technical papers presented at the
Workshop on Measurement Technology and Characterization of Primary
Sulfur Oxides Emission from Combustion Sources held in Southern
Pines, North Carolina, April 24-26, 1978. In addition, reports on
deliberations and recommendations of four Working Groups (corres-
ponding to the four sessions of technical presentations) are in-
cluded.
A Working Group was formed for each of the four sessions of
technical presentations and consisted of all the speakers of that
session. Each Working Group met in a Working Group Session, re-
viewed and critiqued the session presentations, and made its
initial report to all attendees in a summary session. The report
summarized what is known and deemed acceptable and what further
research activity needs to be pursued to provide desirable data and
information. At the summary session, the initial report of each
Working Group was discussed and further modified to reflect the
comments and interaction between the Working Groups. The report of
each Working Group resulting from this summary session is presented
in this volume and follows the set of papers presented at the ses-
sion it addresses.
The focus of sulfur pollutants impacting on ambient air quality
has been the criteria pollutant, sulfur dioxide and its oxidation
products, sulfuric acid and sulfate salts. Considerable attention
has been directed to the sulfuric acid and sulfate salts resulting
from the chemical transformation of S02 both temporally and spatial-
ly in the atmosphere. These are referred to as secondary sulfates.
Sulfuric acid and sulfate salts emitted directly as emissions from
combustion sources also impact on the ambient levels of sulfate.
HI
-------
These direct emissions of sulfuric acid and sulfate salts are
referred to as primary sulfates.
Since sulfates at this time are not criteria pollutants and
emission standards are not prescribed, no reference method is
established for their measurement in combustion source emissions.
With the current and ongoing concern about sulfur in fuels, there
is increasing effort in measuring and characterizing sulfur-contain-
ing emissions from combustion sources. There is a need to identify
valid measurement techniques for primary sulfates, specifically
sulfuric acid, and to provide an accurate and consistent base of
characterization data on primary sulfate emissions from the
various combustion processes. There is also the need to determine
what emission data on primary sulfates are available, their ac-
ceptability as valid measurements, and what further research effort
needs to be conducted to provide a good data base for a good under-
standing of the contribution of primary sulfate emissions to ambient
sulfate levels.
The purpose of this Workshop was to help meet these needs.
I am grateful for the active participation of the Workshop
attendees who were invited to present and discuss their activities
and studies in the area of primary sulfate emissions and for their
contributions which made the Workshop an interesting and signifi-
cant accomplishment. In particular, I want to thank the Session
Chairmen (James Dorsey, Kenneth Knapp, James Homolya, and John
Bachmann) and the Working Group Chairmen (Paul Krone, Dale Lundgren,
James Howes, and David Natusch) for their assistance in implement-
ing the Workshop agenda so effectively. I also want to include
my appreciation for the efforts and cooperation of Ann Mitchell
and Wendy Martin of Kappa Systems in coordinating the Workshop and
in editing the Proceedings.
John S. Nader
Workshop Chairman
IV
-------
Contents
VOLUME 1
SECTION 1 — Gas Sampling and Analysis
An Evaluation of a Modified Method 6 Flue Gas Sampling Procedure 3
Russell N. Dietz
Robert F. Wieser
Leonard Newman
Measurements of Sulfur Trioxide at Tennessee Valley Authority 27
Coal-Fired Power Plants Using the Condenser Method
Elizabeth M. Bailey
H. A. Ruddock
Measurement of SO3/H2SO4 Concentration in Kraft Recovery Furnace 41
Stack Gas Using Controlled Condensation
Ashok K. Jain
R, O. Blosser
Howard S. Oglesby
Characterization of Combustion Source Sulfate Emissions with a 53
Selective Condensation Sampling System
James L. Cheney
James B. Homolya
A Specific Method for the Determination of Sulfuric Acid Emissions 63
from Combustion Sources
Paul Urone
Robert A. Lucas
Measurements of Sulfuric Acid Vapor by Infrared Spectroscopy 79
Roosevelt Rollins
Chemical Speciation and Concentration Monitoring of Sulfur Oxides 97
by Laser-Raman Scattering
Richard K. Chang
Robert E. Benner
-------
Report of the Working Croup on Measurement of Gaseous Sulfur 137
Oxides Emissions
Russell N. Dietz, Reporter
SECTION 2 — Particulate Sampling and Analysis
Collection Methods for the Determination of Stationary Source 145
Particulate Sulfur and Other Elements
Kenneth T. Knapp
Roy L. Bennett
Robert J. Griffin
Raymond C, Steward
A Stack Gas Sulfate Aerosol Measurement Problem 161
Dale A. Lundgren
Paul Urone
Thomas Gunderson
Sulfur Oxide Interaction with Filters Used for Method 5 Stack Sampling 179
Edward T. Peters
Jeffrey W. Adams
Particulate Sampling in Process Streams in the Presence of Sulfur 203
Oxides
Kenneth M. Gushing
Primary Aerosol Sulfur Size Distribution Measurements Using a Low 227
Pressure Impactor
Richard C. Flagan
Use of a High-Flow Stack Sampler for Determination of Particulate 241
Sulfate Emissions
A. Jack O'Neal, Jr.
Harold Cowherd
Inorganic Compound Identification by Fourier Transform Infrared 253
Spectroscopy
Robert J. Jakobsen
R. M. Gendreau
William M, Henry
Kenneth T. Knapp
Report of the Working Group on Measurement of Particulate Sulfur 275
Oxides Emissions
Richard C. Flagan, Reporter
Appendix - Participants and Observers 277
vi
-------
VOLUME 2
SECTION 1 — Gas Emissions
An Assessment of Sulfuric Acid and Sulfate Emissions from the 3
Combustion of Fossil Fuels
James B. Homolya
James L. Cheney
Sulfur Oxides Emissions from Boilers, Turbines, and Industrial 13
Combustion Equipment
SkiUman C. Hunter
Paul K. Engel
Some Recent Data on SO3 and SO4 Levels in Utility Boilers 53
Brian W. Doyle
Richard C. Booth
Measurement of Sulfur Oxides from Coal-Fired Utility and 67
Industrial Boilers
William R. McCurley
Daryl G. DeAngelis
Sulfur Oxide Measurements of Utility Power Plant Emissions 87
James E. Howes, Jr.
Effects of Combustion Modification on SO3 Formation in Combustion 99
Arthur Levy
John F. Kircher
Earl L. Merryman
Impact of Sulfuric Acid Emissions on Plume Opacity 121
John S. Nader
William D. Conner
Query: Is There a Connection between the Expansion of Areas of 137
Acid Rain and a Shift from Coal to Oil for Small-Scale Heat Needs?
ArthurM. Squires
Report of the Working Group on Characterization of Gaseous 143
Sulfur Oxides Emissions
ArthurM. Squires, Reporter
VII
-------
SECTION 2 — Particulate Emissions
Characterization of Fly Ash from Coal Combustion
David F. S. Natusch
Sulfur and Trace Metal Particulate Emissions from Combustion 165
Sources
Roy L. Bennett
Kenneth T. Knapp
Inorganic Compounds Present in Fossil Fuel Fly Ash Emissions 185
William M. Henry
Ralph I. Mitchell
Kenneth T. Knapp
Investigation of Particulate Sulfur by ESCA 209
Arthurs. Werner
Sulfur Emissions Sampling and Analysis 219
Ray F. MaddaZone
Operating Parameters Affecting Sulfate Emissions from an Oil-Fired 239
Power Unit
Russell N. Dietz
Robert F. Wieser
Leonard Newman
Report of the Working Group on Characterization of Particulate Sulfur 271
Oxides Emissions
Ray F. MaddaZone, Reporter
Appendix - Participants and Observers 275
VIII
-------
Section 1
Gas Emissions
-------
An Assessment of Sulfuric Acid and Sulfate
Emissions from the Combustion of Fossil Fuels
James B. Homolya
James L. Cheney
U.S. Environmental Protection Agency
ABSTRACT
A series of studies were carried out in which stack gas
emissions from both coal-fired and oil-fired sources
were analyzed for sulfuric acid and total water-soluble
sulfate. The sampling methods included: (1) a modified
EPA Method 6 procedure for S02 and total water-soluble
sulfate; (2) controlled-condensation procedure for sul-
furic acid, sulfate, and S02; and (3) the determination
of the sulfuric acid dewpoint temperature.
These methods were applied to the combustion emissions
from industrial and utility-sized boilers. Our studies
showed that for a given fuel sulfur content, the total
sulfate emissions from oil-fired sources are from three
to ten times greater than from sources burning coal.
It is believed that the higher flame temperatures, the
vanadium and .nickel content, and the lack of particulate
control devices for oil-firing contribute to the observed
increases in emissions. In addition, a number of
studies demonstrated that the available boiler oxygen
in excess of stoichiometric will enhance sulfate forma-
tion .
Based on the work completed thus far, it appears that
the free sulfuric acid content of the oil-fired source
emissions is approximately 60% of the total sulfate
level. The use of fireside fuel additives may alter the
ratio of acid to sulfate. Through the use of dispersion
models incorporating sulfate emission factors based upon
-------
our characterization studies, we found that the emission
of primary sulfate species can have a marked impact on
ambient sulfate concentrations downwind of combustion
sources.
INTRODUCTION
A series of field sampling studies have been carried out to
assess the atmospheric emissions of sulfates from coal-fired and
oil-fired boilers. Emissions sources were selected which would be
representative of a cross-section of boiler designs with respect
to size, operating characteristics, and fuel usage. Flue gas
samples were collected and analyzed for sulfur compounds by: (1) a
modified EPA Method 6 procedure for S02 and total water-soluble
sulfate; (2) a controlled-condensation procedure for sulfuric acid,
sulfate, and S02; and (3) the determination of the sulfuric acid
dewpoint temperature. Results obtained from studies over the past
three years indicate a significantly higher level of sulfates in
the flue gases of oil-fired boilers as compared to coal-fired
sources. In general, the fraction of sulfur oxides emitted as
sulfuric acid and/or sulfate has been found to be related to the
metals content of the fuel as well as the amount of excess air
used for combustion.
BACKGROUND
When characterizing the sulfur oxides emissions from com-
bustion sources, one must be aware that several potential sulfate
species can co-exist in the flue gases. Sulfuric acid can be
found as: (1) a gas phase component at stack temperatures and
water vapor concentrations; (2) a condensed liquid aerosol at
temperatures below the acid dewpoint; and (3) adsorbed on carbon-
aceous particulate matter at stack gas temperatures. The latter
component has been found to exist at temperatures in excess of
275°C (1). In addition, the free acid may react with metal oxides
formed in the combustion flame to yield sulfates such as Na2(S04),
MgSO4, VOS04, and Fe2(S04)3 (2). For discussion we can identify
sulfuric acid and its reaction products as being primary sulfate
emissions and contrasted with sulfate (secondary sulfate) derived
from the transformation of S02 in the atmosphere.
There are several factors which influence the nature and
extent of primary sulfate emissions. These include: (1) fuel
characteristics; (2) boiler design and operation; and (3) emissions
controls. Coal-firing can be characterized as using a fuel which
-------
has a high ash content and is slow-burning. The principal metals
found in coal are iron, silicon, and aluminum. Relatively low
flame temperatures lessen the formation of H2S04 by the combina-
tion of S02 with atomic oxygen in the flame (3), and the high ash
content entrained in the flue gases tends to neutralize acid that
is formed (4)(5). In contrast, oil is a fast-burning fuel with
a low ash content. Principal metals found in fuel oils include
vanadium, nickel, and sodium. An SO2- atomic oxygen reaction may
be enhanced by the higher oil-flame temperatures, and the effects
of vanadium oxides on sulfate formation have been studied in
residual oil-fired boilers (6). Therefore, for a given fuel
sulfur content, one might expect that flue gases leaving an oil-
fired unit may contain elevated levels of primary sulfate with
respect to a coal-fired boiler of comparable size. Also the
differences in fuel ash content (15% for coal versus 0.1% for
oil) may reflect a large fraction of the primary sulfate emitted
as free H2S04 from oil-firing.
The extent of sulfate emissions can be affected by several
boiler design parameters, including the number and type of
burners, residence time and temperature distribution, and the
amount of internal surface area. For a given boiler design, the
boiler oxygen level in excess of stolchiometric has been found
to be a significant factor in the formation of primary sulfates
(7). Excess oxygen appears to enhance the catalytic action of
deposits on metal surfaces. Therefore, the frequency and dura-
tion of sootblowing, as well as operating conditions which alter
the residence time and temperature distribution in the unit, can
influence the sulfate content of the flue gases.
With few exceptions, oil-fired sources are not equipped
with any emissions controls. In the late 1960's many existing
sources switched from coal to oil as a means of compliance with
emissions regulations on sulfur dioxide and particulate matter.
At the time most sources were using only mechanical particulate
collectors which are ineffective for removing any significant
quantity of ash from oil-firing. Coal-fired units are now
equipped with electrostatic precipitators, and limited experi-
ments have shown that their use can reduce sulfate emissions in
excess of 50% (7). Within the last ten years most of the
larger oil-fired boilers are burning a fuel containing corrosion
inhibitors. These additions, usually containing MgO, are
thought to either scavenge the vanadium oxides to form non-
catalytic species or react with S03 to form a low melting ash
which would not deposit in the high temperature sections of the
boiler. The ash would then be retained in the flue gases to
either deposit in an electrostatic precipitator or to be emitted
to the atmosphere.
-------
Primary Sulfate Emissions Characterization Studies
Since many independent factors were thought to influence the
formation of primary sulfate, emissions characterization studies
have been carried out on a large number of various sources en-
compassing a cross-section of boiler sizes, designs, and emissions
controls. The S02 and primary sulfate levels were characterized
for all sources using a modified EPA Method 6 procedure (8). In
addition, as our measurement capability expanded, the components
of the primary sulfate emissions were identified using a sulfuric
acid dewpoint measurement (9) as well as a controlled-condensation
method (10).
Summaries of emissions characterization studies for oil-fired
and coal-fired units appear in Tables 1 and 2, respectively.
Sources are identified with respect to their fuel sulfur and
vanadium contents. The boiler excess oxygen is given and is
identified as the flue gas oxygen level at the air heater inlet
for large boilers and the stack gas oxygen content for packaged
boilers not equipped with air heaters. The far right column ex-
presses the total primary sulfate emissions as a weight percentage
of the total sulfur oxides. In Table 1, Sources 2, 4, 6, 7, and
10 are industrial-sized units with the remainder being utility
boilers. In general, industrial-sized boilers operate with higher
excess oxygen levels and appear to emit a larger proportion of
primary sulfate as compared to utility boilers burning oil of a
similar sulfur content. Source 11 was the only oil-fired unit
studied which was equipped with an electrostatic precipitator.
The precipitator appears to reduce the sulfate emission by a
factor of two. Source 12 is identical with Source 11 with the
exception of not having a precipitator. The combined impact of
low excess oxygen and fuel vanadium content was studied in detail
at Source 15. For a given fuel vanadium content, increasing
oxygen resulted in a measured increase in sulfate emissions. In
addition, the highest primary sulfate emissions occurred with fuel
containing the highest concentration of vanadium.
In Table 2, Source 1 is an industrial-sized unit. Coal is
found to contain appreciably less vanadium than residual oil,
and its combustion characteristics require higher excess air
levels with respect to oil. However, the extent of primary sul-
fate emissions was found to be less than that from the combustion
of oil of a similar sulfur content. In comparing the average of
all the characterization measurements, the oil-fired sources
emitted about 6.5 wt. % of the sulfur oxides as primary sulfate
as compared to 2.1 wt. % of the sulfur oxides for coal.
-------
Table 1. Summary of S02 and SO4 Emission
Measurements from Combustion Sources
Source
A. Oil
1. LILCO Barrett-#20
2. EPA/Beaunit
3. Ponce-South Coast #6
4. IBM-RTP
5. MICHOUD
6. Burlington Industries-
Durham
7. NCSU-Trane (HC1-HC4)
8. Albany-Unit #1
9. Albany-Unit #2
10. NCSU-Riley
11. LILCO-Northport #3
12. LILCO-Northport #2
13. Arthur M. Williams
14. San Juan-Palo Seco #1
15. Anclote I
Anclote II
Anclote III
Sulfur
wt. %
0.3%
0.2
1.0
1.0
1.2
1.2
1.5
1.8
1.8
2.0
2.2
2.2
2.2
2.5
2.5
2.4
2.6
Vanadium,
ppm by wt.
50 ppm
<1
80
70
16
190
200
135
135
375
500
500
447
300
140
593
292
Boiler
Excess 02 , %
1.8%
0
3
6
3* -6
3
5
2.5
2.5
5
1.8
1.9
2
1.2
0.3* -1.0
0.2* -0.6
0.1* -0.5
4 — v inn u/t ?
S02+ S04 x 1UU' wt' *
7%
11
12
9
3-5
5
7
4
5
8 at air heater inlet
10 at air heater outlet
4 at air heater inlet
2 at precip. outlet
5
7
4
4-9
6-12
2-7 at air heater inlet
2-6 at air heater outlet
-------
Table 2. Summary of S02 and S04 Emission
Measurements from Combustion Sources
Source
B. Coal
1. UNC
2. Wilmington
3. KCP&L Hawthorne
4. CPL - River Bend
5. CPL - Cape Fear
6. LG&E Mill Creek
7. CSO - Picway
Sulfur
wt. %
1.7
1.7
1.7
1.9
2.0
3.6
3.3
Vanadium,
ppm by wt .
<15
<15
<24
39
28
99
35
Boiler
Excess 02 , %
7
4
6
5
4
4
5
SCU - x i nn u-t %
S02+ S04 x iUO' wt> /l
2.6
1.4
2.6
0.9
1.3
3.5
2.8
00
-------
We have analyzed the combustion particulate from coal-
firing and oil-firing and find that oil ash characteristically
contains about 30 wt. % sulfate and retains condensed sulfuric
acid (11). Table 3 is a comparison of sulfate emissions data for
two oil-fired sources with differing fuel sulfur and flue gas
oxygen levels. Total sulfate was measured by the modified
Method 6 procedure, and an acid dewpoint probe was used in com-
bination with moisture determinations to provide a calculated
free H2SO4 level. The source with the lowest flue gas oxygen
level yielded the lowest primary sulfate concentration of which
28.1% consisted of H2S04. In contrast, the source with 3% oxygen
in the flue gas produced a higher sulfate level containing 54.8%
H2S04 . By subtracting out the acid component of the total sul-
fate, it appears that the particulate sulfate content of both
sources is nearly identical.
Primary Sulfate Emissions Impact on Ambient Air Environment
Based upon the results of characterization studies over the
past few years, we have concluded that there is a significant
difference in the emissions of primary sulfate between coal-fired
and oil-fired combustion sources. Measurements have indicated
that for a given fuel sulfur content, the total sulfate emissions
from oil-fired sources are from three to ten times greater than
from sources burning coal and contain a large fraction of free
H2SO4. Residual oils are used in large quantities for both
utility and industrial purposes in the Northeast sector of the U.fci,
where elevated ambient sulfate levels have been measured. The
use of oil for electric generation has doubled over the last ten
years. In 1967, oil accounted for about 9% of the energy gen-
erated as compared with 19% for 1977 (12). Much of this pro-
duction occurs in urban centers where oil-firing is used to
maintain ambient air total suspended particulate levels and SO2
standards. Therefore, primary sulfates may prove to have a
significant impact on ambient levels in regions of high emissions
densities and where oil is the principal fuel.
-------
Table 3. Comparison of Sulfate Emissions Data
1.
2.
3.
4.
5.
6.
7.
8.
Fuel Sulfur
Flue Gas 02 , average, dry
Total S04~2, average at Measured
Flue Gas 02 , dry
Free H2 S04 , average at Measured
Flue Gas 02, dry
South Coast total S04~2, average
at 1.28% 02, dry
South Coast Free H SO , average
at 1.28% 02 , dry
H S04
x i on 1
S°4
(Total S04~2 - Free H2S04),
at 1.28% 02, dry
Palo Seco - Unit 1
2.37%
1.28%
5 3
1.28 x 10 /"g/tn
0.36 x 10 5 ,ug/m3
NA
NA
28.1%
0.92 x 105 ("g/m3
South Coast - Unit 6
0.99%
3.13%
1.80 x 105 A'g/m3
0.99 x 105 Mg/m3
1.97 x 10 ,"g/m3
1.08 x 105 ,«g/m3
54.8%
0.89 x 10 j"g/m
-------
REFERENCES
1. Cheney, J. L., and J. B. Homolya. Characterization of
Combustion Source Sulfate Emissions Using a Selective
Condensation Sampling System. In: Proceedings of the
Workshop on Measurement Technology and Characterization of
Primary Sulfur Oxides Emission from Combustion Sources,
Southern Pines, North Carolina, April 1978.
2. Pratt, W. G. Steam, Its Generation and Use. Geo. McKibbin
and Son, New York, New York, 1955. 400 pp.
3. Young, W. E., and A. E. Hershey. Corrosion, 13:725t-732t,
1957.
4. Wickert, K. Brennstoff-Warmes-Kraft, 9:104-118, 1957.
5. Tolley, G. J. Soc. Chem. Ind., 57:369-373 and 401-404, 1943.
6. Anderson, D. R., and T. P. Manlick. Trans. ASME, 80:1231-1237,
1958.
7. Homolya, J. G., H. M. Barnes, and C. R. Fortune. A Charac-
terization of the Gaseous Sulfur Emissions from Coal and Oil-
Fired Boilers. In: Proceedings of the Conference on Energy
and the Environment, Cincinnati, Ohio, 1976.
8. Cheney, J. L., W. F. Winberry, and J. B. Homolya. J. of
Environ. Science and Health, 12(10):549-559, 1977.
9. Cheney, J. L., C. R. Fortune, J. B. Homolya, and H. M. Barnes.
The Application of an Acid Dewpoint Meter for the Measurement
of Sulfuric Acid Emissions. In: Proceedings of the Confer-
ence on Energy and the Environment, Cincinnati, Ohio, 1976.
10. Cheney, J. L., J. B. Homolya, and H. M. Barnes. Measurement
and Identification of Primary Sulfates Emitted from Station-
ary Sources. In: Proceedings of the Annual Meeting of
AICHE, New York, New York, 1977.
11. Homolya, J. B., and C. R. Fortune. The Measurement of the
Sulfate and Sulfuric Acid Content of Particulate Matter from
the Combustion of Coal and Oil. Atmospheric Environment
(in press).
12. Electrical World. 1978 Statistical Report from Energy Infor-
mation Administration/DOE, March 15, 1978. 90 pp.
11
-------
Sulfur Oxides Emissions from Boilers, Turbines,
and Industrial Combustion Equipment
Skillman C. Hunter
Paul K. Engel
KVB, Inc.
ABSTRACT
Measurements of sulfur oxides emissions were made on a
wide variety of combustion devices, including boilers,
gas turbines, and industrial process combustion equip-
ment burning a variety of fuels.
Methods of measurement included continuous monitoring
of sulfur dioxide and wet chemistry measurement of sul-
fur dioxide, sulfur trioxide, and sulfates using the
Southern California air quality management district
method, the Shell Emeryville method, and the Goksoyr-
Ross method. Results from these measurements will be
compared.
Emissions of sulfur trioxide from utility and industrial
boilers are typically observed to vary from 1% to 6% of
total sulfur oxides. Uncertainties in measurement
methods producing high apparent levels of sulfur trioxide
at low total sulfur oxide levels will be discussed.
Sulfur trioxide formation by means of gas phase kinetics,
exclusive of surface catalytic action, was analzyed and
levels typically observed were attributed to combustion
of carbon monoxide in post-flame zones.
Sulfur retention in the ash from western coals will be
compared with eastern coals fired in industrial boilers.
A limited amount of data will be presented on the effect
of combustion modifications performed on an industrial
boiler, and on the emissions of sulfur trioxide and
sulfates.
13
-------
INTRODUCTION
This paper presents an evaluation of the formation of sulfur
oxides in combustion equipment. Sulfur oxides emission is
currently under regulation by air quality control authorities.
Utility plant cold-end corrosion and emission of acid stack plumes
have been identified as sulfur-related problems. Limited avail-
ability of low sulfur fuels demands that full attention be given
to all possible means for controlling these emissions through
plant operating procedures. The purpose of this paper is to
examine the processes involved in sulfur oxide formation, review
experience with test methods, and present test results for sulfur
oxides emissions.
Thermodynamic equilibrium computations show that for practical
fuel/air mixture ratios, the predominant sulfur species present
in combustion gases are sulfur dioxide (S02) and sulfur trioxide
(SO3). At temperatures greater than 2500°F (1640 K), more than
99% of the sulfur is present as S02. At lower temperatures
equilibrium shifts to increasing amounts of S03, so that below
700°F (644 K), more than 99% .of the sulfur can be present in the
form of S03. Limitations in the chemical kinetic rates, however,
are such that only from 1% to 5% of the sulfur in stack gases is
observed in practice to be present as SO3 with the balance as SO2.
Formation of S02 occurs early in the primary flame at rates
comparable to the other combustion reactions. Formation will
occur even in fairly fuel-rich flames so that no practical com-
bustion control techniques have been identified. The only means
for limitation of S02 emission is through control of the fuel
sulfur content or by stack gas treatment.
Formation of S03 is found to occur only in air-rich mixtures
and to be governed by kinetic processes more amenable to combus-
tion control. Kinetic processes within the primary flame were
found to be sufficiently fast so that S03 concentrations quickly
approach equilibrium levels of less than 0.1% of the total sulfur.
However, when combustion gases are cooled, a critical temperature
region between 1500° and 3000°F (1090 and 1920 K) was found within
which S03 is formed by gas phase kinetics, exclusive of any
catalytic activity. Within this temperature range S02 can react
with O atoms to form S03. At the low end of this temperature range
formation ceases, and kinetic computations yield SO3 concentra-
tions between l%-5% of total sulfur as observed in practice for
excess oxygen levels greater than 1%.
14
-------
Kinetic computations confirm that S03 can be reduced by
lowering excess oxygen and suggest that the use of fuel-rich
combustion, together with properly staged excess air addition,
might be an effective method for 863 control. These same tech-
niques have proved effective in control of oxides of nitrogen.
GAS PHASE KINETICS OF SULFUR OXIDES FORMATION
Thermodynamic Equilibrium Considerations
The thermodynamic equilibrium of burned mixtures containing
sulfur indicates that the predominant sulfur compound is sulfur
dioxide (S02). In air-rich mixtures a portion of the sulfur can
be present as sulfur trioxide (SO3). The amount of SO3 expressed
as a percent by volume of the total sulfur is a function of the
mixture temperature and oxygen content. Figure 1 presents this
relationship. At high temperatures, as in a flame, the amount of
SO3 is less than 0.1%. As temperature decreases, the relative
amount of S03 increases to 100% at about 750°F (672 K). A critical
temperature zone exists between 1000° and 2000°F (810 and 1370 K)
where percent SO^ changes rapidly. The fact that complete
conversion of SO3 does not occur in combustion units can only
be explained by the rates at which the pertinent chemical re-
actions proceed. Consideration of the kinetic processes is
therefore important in defining means for limiting S03 formation.
The figure also illustrates that the theoretical percent
80s is a function of the square root of the oxygen content.
The very high level of theoretical percent SO3 suggests that any
form of catalytic activity in the cold regions of a power plant
could result in very substantial levels of SO3.
Gas Phase Kinetic Considerations
A review of measurements in flames, together with computations
obtained with the use of computer programs developed by KVB, has
resulted in an improved understanding of sulfur oxides formation
processes.
Flame profile studies show that SO2 formation occurs within
the primary flame zone at rates comparable to other combustion
reactions. Formation will proceed in both air-rich and fuel-rich
flames at fuel-air ratios up to about 1.4 times stoichiometric.
Formation of S02 is, therefore, unavoidable, and most of the fuel
sulfur will be exhausted in this form. The only practical means
for limiting SO2 stack emission is through control of fuel sulfur
15
-------
100
UJ
(_>
o:
UJ
Q_
CD
X
o
CO
C/5
ce
80
r
PERCENT 0'
BY VOLUME
WET BASIS
60
20
1000 1500
GAS TEMPERATURE, °F
2000
2500
Figure 1. Thermodynamic equilibrium SO, conversion.
16
-------
content or stack gas treatment. There are exceptions to complete
sulfur emissions. These include ash retention of sulfur for
certain coals and process materials absorption, as has been ob-
served in cement kilns.
Several possible reactions forming S03 were defined; reaction
of SOa with 0 atoms and a third body was identified as the most
probable mechanism. Kinetic rate constants were obtained from
the literature, and S02-S03 reactions, together with other re-
actions important in combustion, were evaluated with the KVB one-
dimensional chemical kinetic computer program. Calculations
were performed to examine the primary flame zone and the effects
of post flame cooling and mixing of excess air.
Adiabatic flame zone computations indicated that peaks of
SO3 can occur that coincide with peaks of 0 atom concentration
resulting from CO combustion. However, as the combustion reactions
reach equilibrium, S03 concentrations drop to less than 1 ppm,
so that the adiabatic flame zone does not appear to be a signi-
ficant source of S03.
As post flame gases are cooled, 0 atom concentrations exceed
equilibrium, and SO3 begins to increase at about 3000°F (1922 K).
As temperature drops to the range of 1500° to 2000°F (1090 to
1400 K), O atom production ceases. SO3 concentration dictated
by equilibrium now exceeds actual concentrations, but 0 atoms
are rapidly depleted preventing further S03 increase. SOa forma-
tion ceases and concentration remains constant with further
cooling. The final level of S03 is found to be dependent on 02
content, rate of cooling, and rate and location of excess air
mixing. The effects are critical primarily in a temperature
range of 1500° to 3000°F (1090 to 1920 K). Fuel-rich combustion
does not result in any SO3 formation, and the use of fuel-rich
combustion, together with staged air addition and reduced excess
oxygen, appears to be a possible method of S03 control. Control
of heat transfer in various sections of the plant, although
difficult to implement, is also suggested as a possible reduction
technique.
The results of kinetic analysis described below in general
indicate that gas phase reactions can explain the observed levels
of SO3 independent of any catalytic surface activity. Undoubtedly
both mechanisms contribute to S03 formation.
17
-------
so2 + o + M ^r so3 +
S03 + 0 :£ S02 + 0
S02 + 0 j£- S03
S03 + H 47 S02
S02 + 1 02 +
so + 02 + M 5;
S03 + H2 ^ S02
S02 + N02-iS03
+ OH
S03
S03 •
+ H2(
-t- NO
Formation of S03 in the Adiabatic Flame Zone—Sulfur trioxide
(S03) can be formed by several reactions. Cullis (1) and
Merryman (2) present discussions of the possible and most probable
formation processes. The various reactions considered include:
[ 1 ]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
S02 + H02:£S03 + OH [9]
Reactions involving direct; 02 attack on S02 have been shown
to be very slow (1)(3). Significant rates of flame-produced S03
are only observed in the presence of 0 atoms and can be attributed
to reactions [1] and [3]. Merryman (2) showed that S03 formation
is strongly pressure dependent and concluded that the third body
reaction [1] is the predominant reaction. This conclusion is
supported by many others as discussed by Cullis (1). Reaction
[1] is opposed by reaction [2] which acts to reconvert S03 back
to S02. Sawyer (4) suggests that reaction [9] may be important
in low temperature regions.
Thermodynamic equilibrium considerations indicate that at
flame temperatures the percent S02 converted to S03 is less than
0.1%. However, this also implies that all other species are in
equilibrium, particularly the 0 atoms. This is clearly not the
case in the flame zone. W.hen one species, in particular CO, has
a slower rate of reaction, the remaining much faster reactions
assume a different equilibrium. In particular, 0 atoms increase
by factors of 10 to 1000 times %the value established by 0-02
equilibrium. The driving force for reaction [1] then establishes
a different single-reaction equilibrium for the ratio of S03 to
SO2 concentration, so that the mole fraction of S03 can be in-
creased over the complete equilibrium levels in proportion to the
18
-------
increase in 0 atom concentration above complete equilibrium levels.
Inclusion of reaction [2] in this consideration indicates that
the S03-S02 ratio would tend to stabilize at a value intermediate
between the equilibrium value and the upper limit determined by
excess 0 atom concentration. As the main combustion reactions
reach completion, the 0 atom concentration re-turns to equilibrium
with 02, and flame-produced S03 can be reconverted to SO2 by the
reverse of reaction [1] and by reaction [2], The combined effects
of reaction [1] and [2] are such that a peak in SO3 is observed in
flame profiles as reported by Levy (2) (3) (5) (6) (7) and Medley
(8). The relative rates of these reactions must be considered
to determine whether there is a possibility that the high flame
SO3 levels could persist beyond the flame by quenching of the
reaction rates through cooling or mixing.
Computations have been performed to investigate the S03 for-
mation process in the flame region employing a one-dimensional
gas phase kinetics program developed by KVB. A fuel oil with
basic composition of CnH2n plus 1% sulfur was used. The flame
profile measurements previously discussed have shown that SO3
formation occurs beyond the region of H2 combustion and within the
region of CO combustion. Accordingly, kinetic computations were
initiated as a mixture of CO and H2O and radical concentrations
determined from equilibrium computations with formation of CO2 and
SO3 suppressed. The sulfur was introduced entirely as SO2.
The kinetic mechanism employed, presented in Table 1, is
taken from that of Breen, Bell, and Bayard de Volo (9) with the
addition of reactions [1] and [2] and updated rate constants for
more recent values.
Kinetic calculations were performed at a constant pressure of
one atmosphere for adiabatic combustion with air preheated to
450°F (506 K) to determine S03 concentrations as a function of
time. The fuel to air ratio was varied from 0.87 to 1.3 times the
stoichiometric fuel-air ratio to simulate the range of burner
operation from lean at 3% 02 to fuel-rich. Figure 2 presents the
gas temperature and S03 concentration profiles as a function of
time. The temperature increases monotonically from that for
partial combustion only to CO and H20 up to the adiabatic flame
temperature. SOa is observed to rise to a peak and fall rapidly
back to near equilibrium values within less than 0.001 second.
The peak SO3 level varies from 12 ppm at 3% 02 to 2 ppm for fuel-
rich combustion and is obviously a strong function of 02 content.
A level of 1% sulfur in the fuel produces SO3 concentrations of from
480 ppm to 680 ppm in the flame region. Five ppm of SO3, then,
corresponds to approximately 1% conversion of sulfur to SO3, so
that observed peak SO3 levels vary from 0.3% in fuel-rich flames
19
-------
Table 1. Reaction Set for Combustion with Sulfur
REACTIONS
Third Body Reactions
SO 3 = S02 + 0
N2 = N + N
02 = 0 + 0
H2 = H + H
OH = 0 + H
H20 = OH + H
CO2 = CO + 0
N02= N + 0
NO2 = O + NO
N2O = O + N2
CO/CO2 Reactions
C02 + H = CO + OH
CO + 02 = C02 + 0
H2/O2 Reactions
OH + H = H2 + O
H2O + H = OH + H2
OH + 0 = H + 02
H20 + 0 = OH + OH
Nitrogen Reactions
NO + 0 = 02 + N
N2 + 0 = NO +• N
NO + H = Oil + N
NO + 02 = NO +0
N2 + 0 = N20 + 0
Sulfur Reactions
BACKWARD RATE
A =
A =
A =
A =
A =
A =
A =
A =
A =
A =
A =
A —
A =
A =
A =
A =
A =
A =
A =
A =
A =
1.00E15,
1.E18,
1.9318,
7.5E18,
3.6E18,
1.17E17,
5.1E15,
I.OE20,
I .051-: 15,
i .ooi>;i8 ,
5.6E11,
1.9E13,
1.74E13,
2.19E13,
2.24E14,
5.75E12,
6.43K9 ,
3.10UL3,
4.20131.3,
1 .OOK13,
3.00lil3,
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
N =
CONSTANTS
0.0, B = 0.0
1.0, B = 0.0
0.5, B = 0.0
1.0, B = 0.0
1.0, B = 0.0
0.0, B = 0.0
0.0, B = 3.58
1.5, B = 0.0
0.0, B = -1.87
1.0, B = 0.0
0.0, B = 1.08
0.0, B = 54.15
0.0, B = 9.45
0.0, B = 5.15
0.0, B = 16.8
0.0, B = 0.78
-1.0, B = 6.250
0.0, B = 0.334
0.0, B = 0.9
0.0, B = 0.600
0.0, B = 26.8
REFERENCE
[1]
[9]
19]
[9]
[9]
[10]
[9]
[H]
110]
19]
[10]
[9]
[10]
[10]
[10]
[10]
110]
110]
[12]
[10]
|9J
SO, + 0, = SO, + O A =
k = A T-Ne(B/RT)
R = 0.001987 kcal/mole-0K
B in kcal/mole
1.20E12, N = 0.0, B = 9.5
T = temperature, °K
UJ
k = rate constant, (cm3/mole)nsec
n = 1 for biomolecular reactions
= 2 for third body reactions
20
-------
DC
O
n_
5
GO
4000
3000
FUEL-RICH
0,0% 02
1.0% 02
AIR TEMP.
= 450 °F
(506 K)
1% S FUEL
OIL
GO
GO
-------
to 2.5% conversion at 3% Oz• These peaks are consistent
with observed conversion levels in stack gases. However, SO3
levels decay rapidly back to less than 1 ppm. The time at .which
S03 peaks corresponds to the point at which O atom concentration
becomes sharply reduced. This point is reached when CO approaches
equilibrium. It is possible that turbulent mixing could result
in alteration of the decay process or that distributed mixing in
diffusion flames could sustain high 0 atom concentrations for
longer periods of time. Investigation of these possibilities
represents an area for further analysis. However, based on the
current results, it appears unlikely that flame produced SO3 is
the major source of stack gas SO3. The computed S03 profiles
are similar to measured profiles presented by Levy (3)(5)(7),
Merryman (6), and Hedley (8). However, the computed profiles
decay much more rapidly than the profiles in the referenced works
which extend out to as long as 0.2 seconds before returning to
equilibrium. This difference is attributed to the significantly
lower temperatures in H2S and COS flames used in the referenced
work, but a more detailed examination is warranted.
Conclusions from these calculations are that reduction of
02 level and the use of fuel-rich combustion can be expected to
reduce SO3 levels in the flame. Use of additives to reduce O
atom concentration would also be effective. It appears unlikely
that S03 produced in the high temperature adiabatic part of flames
can escape to the stack gases at concentrations in excess of
flame equilibrium levels, but further investigation with more de-
tailed combustion models and/or test programs is warranted.
Formation of S03 in Cooled Flames—Thermodynamic equilibrium
considerations show that, as gases containing sulfur as S02 and
SO3 are cooled, the percent of sulfur present as S03 can increase
up to 100% at temperatures corresponding to power plant air heater
and stack temperatures. Factors that are expected to influence
the rate of S03 increase include the gas cooling rate and mixing
of excess air. Computations were performed with the KVB one-
dimensional kinetic program to evaluate these two factors.
There is a two-fold effect in gas cooling as heat is trans-
ferred from the flame by radiation and convection. First, at
flame temperatures there is an appreciable amount of dissociation
so that all CO and H2 are not all converted to C02 and H20. As
the gases are cooled, recombination occurs as necessary to
maintain equilibrium at the local gas temperature. If cooling
occurs at a rapid rate, 0 atom concentrations, although decreasing,
22
-------
will exceed equilibrium, and S03 will be produced by
S02 + 0 + M *: S03 + M
As in the adiabatic flame, S03 concentrations in excess of that
for O-02 equilibrium can be produced. The second effect in-
volved is the temperature dependency of the 0 production re-
action rates. The three-body SO2-S03 reaction is not temperature
dependent but the 0 atom reactions are. At low temperatures the
0 atom formation will be slow so that 0 -*~O2 and S03 formation
will deplete O atoms, and SO3 concentrations will stabilize at a
constant level.
Air mixing into hot gases can be expected to produce similar
results, first causing completion of combustion, then cooling with
recombination reactions, and finally quenching the reactions.
Computations with flame cooling have been performed for a
range of fuel-air ratios and cooling rates. Cooling was initiated
at 0.001 second, the point at which the combustion reactions had
essentially reached equilibrium previously as shown for the
adiabatic flame. Based on data of James (13) boiler gases cool
approximately 1000°F (555 K) in the furnace before entering the
superheater. Furnace cooling rates are of the order of 10°F
per foot (18 K/m). In the superheater and air heaters higher
rates of the order of 100°F per foot (180 K/m) are estimated.
Computation was performed with an assumed gas velocity of 100
feet per second (30.5 m/s). Figure 3 presents the profiles of
temperature and SO3 formation as a function of time at 3% 02.
Gas temperature drops 1000°F (555 K) at a rate of 10°F per foot
(18 K/m) simulating furnace conditions, then at a rate of 100°F
per foot (180 K/m) simulating superheater and heat exchanger con-
ditions. S03 level remains below 1 ppm throughout the furnace
and rises sharply to 8 ppm (1-6% conversion) at the point of
increased cooling rates. This level of conversion is consistent
with levels observed in operating plants. For calculations with
faster cooling rates SO3 was higher, indicating a greater O atom
imbalance as expected. Formation of SO3 is quenched when tem-
peratures drop below 1540°F (1111 K) at least for this condition
of about 3% excess oxygen. At this point 0 atoms are depleted.
The influence of oxygen content was also investigated.
Figure 4 shows the computed profiles of temperature and SOs versus
time for four levels of excess O2 and a cooling rate of 1000°F per
foot (1823 K/m). S03 rises most rapidly when gas temperature
drops from 3000°F to 2000°F (1920 K to 1370 K). The effect of
excess oxygen is most pronounced below 1% O2 with only a slight
increase in SO3 at 3% O2.
23
-------
4000
3000
2000
CD
1000
10
10 °F/FT (- 18 K/M)
CD
GO
0
EXCESS 02 = 2,9 %
AIR TEMP, = 450 °F (506 K)
II S FUEL OIL
0
- 100 °F/FT
(- 180 K/M)
=fc
0,2 0,4 0,6 0,8
TIME, SECONDS
1,0
1,2
Figure 3. S02 formation in a cooled gas,
24
-------
c* 4000
o
3000 -
a 2000
Q-
1000
AIR TEMP. = 450
(506
- 1% S FUEL OIL
D_
D_
O
GO
0,01 0,02 0,03
TIME, SECONDS
0 04
Figure 4. Effect of 02 on S03 formation.
25
-------
The computed effect of excess O2 on percent conversion of 862
to S03 was compared with measured data from various boilers (14-20),
shown in Figure 5. The computed conversion is much higher than
test data at zero excess 02. However, the SO3 increase, as 02
is raised, is consistent with the data. A tendency for S03 con-
version to level off above 3% 02 is apparent in both computed
and measured data.
Formation of S03 in Cooled Flames With Staged Air—Fuel-rich
combustion accomplished with a portion of the furnace burners
out of service or air bypassed above the burners are combustion
control techniques currently employed for NOX reduction. The
effect of air addition to a fuel-rich mixture has been computed
by repeating the computed fuel-rich solution of Figure 4, but
mixing additional air sufficient to increase the O2 level to
3%. The point of air mixing initiation and the rate of injection
were varied. Figure 6 presents the temperature and SO3 concentration
profiles versus time. Curves A are for the completely fuel-rich
mixture with no mixing. Curves B are for air injection in high
temperature gases prior to cooling below 3000°F (1920 K). SO3
formation begins near the end of air injection and continues
until the temperature drops below 2000°R (1540°F, 1111 K). Twenty
ppm of S03 is formed, compared with 17 ppm for the case where
the entire flame is at 3% 02 (Figure 4).
Curve C shows the effect of decreasing the rate of air mixing
by adding the same quantity of air over a longer time. The initial
S03 formation rate is slower, but the final level is higher com-
pared to Curve B. Curve D shows the effect of delaying the in-
jection until the temperature has dropped to 2000°F (1111 K). CO
combustion is slower at this condition but does react and causes
a much higher level of S03. With air mixing initiated at 1400°F
(778 K), Curves E, the CO is unable to react and no SO3 is formed.
Nitric oxide levels computed in the foregoing cases showed a
66% NO reduction for Curve B compared with that for air-rich
combustion at 3% 02 (Figure 4) and negligible NO formed for
Curves A, C, D, and E. This is consistent with the effects found
in power plants when fuel-rich combustion is employed so that
heat is lost from the flame prior to and during air addition.
It is interesting to note that formation of NO, as for SO3,
occurs as a result of 0 atom reaction (9). However, the equili-
brium levels of NO at flame temperatures are very high and de-
crease with temperature in contrast to SO3. The NO formation
26
-------
UJ
Q.
o
CO
X
CD
OO
CO
or
UJ
FRIEDRICH (LIGNITE, 0,6% S)
REESE (185 MW, OIL 21 S)
REESE (110 MW, OIL, 2% S)
- KVB ANALYSIS (OIL, II S) -
GILLS
(INDUSTRIAL BOILERS,
OIL, 0,8-3,5% S)
2 4 6
EXCESS OXYGEN, PERCENT
8
10
Figure 5. Effect of 02 on S03 emissions.
27
-------
o:
0
4000
3000
2000
1000
A,E
o
c/o
60
40
20
AIR ADDED OVER TIME SHOWN
0,01 0,02 0,03
TIME, SECONDS
0,04
Figure 6. Air addition in a fuel-rich flame,
28
-------
rates are much slower than for SO3 , and NO levels increase contin-
uously, requiring times of the order of 0.05 seconds to reach levels
observed in power plants as discussed by Breen (9). As temperatures
drop by cooling and bulk gas mixing, the NO is frozen and remains
present at low temperatures. These basic differences between the
equilibrium and kinetic characteristics of NO and S03 indicate that
differing requirements may arise in attempting simultaneous control
of these two emittants. Fortunately the basic technique of reduc-
ing excess air is beneficial in reduction of both.
From the results of air addition computations and comparison
with the work of Barrett (21) and also Hedley (8), it is apparent
that, to minimize S03 , air mixing to high levels of excess oxygen
should be avoided in the temperature range of 2000° to 3000°F
(1367 to 1920 K) . The use of fuel-rich combustion, together with
careful staging of air addition, appears to be a possible technique
for minimizing both NO and SOa formed by gas phase reactions. Fur-
ther work is necessary to define the mixing criteria necessary to
achieve this result. Experimental investigation of the effect of
staged combustion on SO3 has not yet been conducted.
Test measurements of SOa and SOa frequently indicate that the
percent conversion of sulfur to S03 increases as fuel sulfur content
decreases. However, the S03 kinetic reactions are such that S03
formation rates are directly proportional to SOa concentration.
This suggests that for a given flame condition the percent conver-
sion to S03 would be constant and independent of fuel sulfur con-
tent. The kinetic computations described above were re-performed
with fuel oil sulfur contents ranging from 0.5% to 2.5% sulfur.
The computed S03 levels changed in direct proportion to sulfur
content, i.e., percent conversion to SO3 remained constant. The
increase in conversion observed in test data is believed, there-
fore, to be the result either of test method inaccuracy or possible
catalytic surface effects.
TEST METHODS
Methods Evaluated
Over a period of several years, KVB has used three wet chemis-
try methods (other than EPA Methods 6 and 8) for SOa and SO3 mea-
surement: the Shell-Emeryville method (22), the South Coast Air
Quality Management District (SCAQMD) method (23), and the Goksoyr/
Ross controlled condensation method (24). SOa has also been
measured with continuously-recording electronic monitors. This
section compares results obtained with these methods.
29
-------
Shell-Emeryville Method—The Shell-Emeryville method uses
the same chemicals as EPA Method 8. Ninety percent isopropanol
solutions are used to selectively absorb SO3 and a H2O2 solution
follows for the SO2 absorption. The SO2-S03 fractions are titrated
with lead perchlorate to a sulfonzo III end-point. The SO3 results
obtained tend to be somewhat high because of problems associated
with the use of the IPA solution—specifically the nitrogen purge
to remove interfering SO2.
SCAQMD Method—In this method, H2SO4 mist particles are
collected on a Whatman thimble paper filter prior to absorbing
the SO2 in a solution of 5% NaOH. The thimble is maintained
at 165°-200°F (350-370 K) to prevent the condensation of water
vapor but permit the collection of H2S04 mist particles. There
are questions about the efficiency of the paper filter as a means
of collecting SO3. Because there may be incomplete aggregation
of H2SO4 mist particles, the SCAQMD method may produce SO3 measure-
ments on the low side.
Goksoyr/Ross Controlled Condensate Method—This method
has desirable features of separating the SO 2, H2S04(S03), and
particulate matter from flue gases in a clean manner. The
H2S04(S03) is separated from the rest of flue gases, including
SO2, by cooling the gas stream below the dewpoint of H2SO4 but
maintaining the temperature of the gases above the water dew-
point .
Particulate matter is removed by means of a heated quartz
glass filter in a filter holder kept above 500°F (533 K). A
condensation coil where the H2SO4 is collected is maintained
at 140°F (333 K) by a water circulation bath. Sufficient residence
time is allowed to condense all the acid present. The S02 is
then removed in impingers filled with H2O2.
Modified Controlled Condensate Method—This procedure
is similar to Goksoyr/Ross coil but uses an air-cooled coil to
collect the acid. The sampling rate is higher and a larger sample
is taken. One drawback in this procedure is that the control
of the coil temperature is not as effective.
KVB has accumulated considerable experience and data using
these methods in various field tests of combustion and process de-
vices. The results of conversion to S03 have been consistent at
S02 levels above 100 ppm with a lower shelf of measurements at
3 ppm to 4 ppm S03 at lower S02 levels. In some cases the percent
conversion to S03 apparently increases at reduced sulfur dioxide
levels. However, this may be a result of inherent lower limits in
30
-------
the accuracy of measurement methods rather than what is actually
happening as a chemical process.
SO2, SO3, AND SULFATE EMISSIONS DATA
Sources Evaluated
The data shown in this section were excerpted from KVB studies
of:
1. Industrial process exhausts using the Shell-Emeryville
and SCAQMD chemical methods and an SO2 analyzer (25).
2. Industrial boiler exhaust using the Shell-Emeryville
methods (26) and the SASS train (27).
3. Utility boiler exhaust using a modified controlled
condensate method (28).
4. Diesel engine exhaust using the Goksoyr/Ross and Shell-
Emeryville methods (29).
5. Mid-sized coal-fired utility boiler exhaust using the
Shell-Emeryville method (30).
6. Gas turbine exhaust using the Shell-Emeryville method
(28).
Industrial Process Sources
In a survey for the California Air Resources Board, 38 indus-
trial process sources were tested. These were all non-combustion
sources of SOX including glass furnaces, sulfuric acid plants, sul-
fur recovery plants, coke kilns, cement kilns, fluid catalytic
crackers, gypsum kettles, incinerators, lead furnaces, iron cupola,
iron ore sintering machines, blast furnaces, and steel open hearths.
This program included a comparative evaluation of the Shell and
SCAQMD methods for SOX.
Figures 7 and 8 show a comparison of instrumental SO2 measure-
ments versus wet chemical SO2 results for two methods. In Figure 7,
a consistent trend is apparent which shows that the results for the
Shell-Emeryville wet chemical method are approximately 9% lower
than the instrumental reading for SOa. Almost all the points are
above a 45° diagonal which denotes the line of coincident agreement.
31
-------
500
400 ~
Q_
Q_
300 -
200 -
CO
100 -
200 300 400
S02, SHELL METHOD, PPM (WET)
500 600
Figure 7. Comparison of S02 instrumental and SO- Shell
method results.
32
-------
0
200 300 400
S02, SCAQMD, PPM (WET)
500
600
Figure 8. Comparison of S02 instrumental and S02 SCAQMD
method results.
33
-------
A statistical analysis of the Shell-Emeryville data, excluding
the three points at 190 ppm, showed the mean instrumental/wet
measurement ratio to be 1-088. There was a 90% probability that
ratio would lie between 0.97 and 1.21 for any particular experi-
mental measurement. It has been reported that S02 values between
5% and 10% below true values were found using a method with H202
solution in small impingers. In a similar plot for the SCAQMD
method (Figure 8), the agreement between the instrumental and
wet chemical method appears more satisfactory.
During these tests, the wet chemical methods were also checked
by absorbing measured quantities of certified S02 calibration gases
used for calibrating the UV analyzer. Results for S02 using the
Shell-Emeryville were 5% to 6% low in those tests. The field test
results for S02 as measured by the Shell-Emeryville method were
corrected by the ratio of 1.088 to obtain closer agreement with
other methods of measurement. When the SCAQMD method was used, the
S02 results were reported directly. The results reported for the
wet chemical analyses were obtained by average of the replicate
samples (normally three).
The SO3 experimental results for the Shell and SCAQMD methods
that were obtained on common tests are shown on Figure 9. Points
for total sulfate which were obtained with the SCAQMD method are
also shown. S03 results measured by the SCAQMD method were as high
as those measured with the Shell method in only one case. In all
other cases they were lower. These results are consistent with
previous discussion where it was indicated that S03 results obtained
with the Shell method might be high, especially at low total SOX,
and that the S03 results obtained with the SCAQMD method would be
low, if anything. The amount of data obtained is insufficient to
determine which method, in fact, records more accurate values of
SO3. Both methods suffer from a lack of adequate calibration
procedures.
Experimental results for the S03 versus SO2 employing the
Shell method are summarized in Figure 10. For tests where the total
SOX concentration was greater than 50 ppm, there was only one case
where the percentage S03/SOX ratio was as much as 10%. For that
particular device (a coke calcining kiln), a value of 10% is not
unreasonable because of high stack temperature (1800°F, 1260 K).
In three other cases the percentage SO3 was greater than 4%. It was
4% or less for all other cases where the total SOX was greater than
50 ppm. At total SOX values less than 50 ppm, however, the percen-
tage SO3 was greater than 10% in most cases and in some cases greater
than 20%. These excessively high values are believed to be related
to the limited amount of S02 that is dissolved in the isopropyl
34
-------
7^
sox
SOj
sox
A ^L
SHELL
SCAQMD
SOX SCAQMD
0
200
SOX (S02
Figure 9. Comparison of Shell and SCAQMD SOX methods,
35
-------
O5
32
28
Q_
O
CO
oo
GO
cc
20
16
12
4
0
Q 1 1 1 1 1 1 1 1
O SHELL METHOD
O
p
°0
o
: o o °
o o o
OQ)^ -.^^ O O
i i o Q poo, , | ,
1
—
—
-
-
•"
•""
—
p ~
0 100 200
300 400 500 600
SOX, PPM
700 800 900 1000
Figure 10. SCL emissions from industrial processes.
-------
alcohol solutions and which is not removed by purging. This resid-
ual S02 remains to be measured as S03. Of course, the smaller the
total ppm of S02 in the sample gas, the more important this residual
S02 could become.
Although it is not clear at what point the residual S02 may
become important in determining the reported 80s concentrations, it
seems that below total SOX concentrations of 50 ppm-100 ppm the
reported S03 concentrations should be viewed with some skepticism
when obtained by wet chemistry separation.
Industrial Boilers SQ2/S03 Data
SO2/SC>3 data were reported for various industrial boilers under
EPA contract 68-02-1074 (26). The method of analysis was the Shell-
Emeryville method. Figure 11 shows the percent conversion of total
sulfur oxides, SOX, to SOa was typically 1% to 2% except when sul-
fur oxide concentrations dropped below 500 ppm. The steep rise
below 500 ppm was attributed to the measurement method itself.
There appeared to be no strong effect of fuel type other than its
sulfur content. For example, No. 6 oil data decreased with total
sulfur oxides just as with other fuels. For coal, the type of coal
feed did not have a significant effect on the SC>3/SOX, conversion
in the exhaust gases. The coal feed types included spreader stokers,
pulverizers, underfeed stokers, and cyclones.
The S02 emissions for these industrial boilers were related
to fuel sulfur content as shown on Figure 12. The S02 emissions
with fuel oil were generally with •* 100 ppm of that based on 100%
conversion of S to SO2 . However, deviations of up to 300 ppm were
noted. With coal there is much more scatter. For fuel sulfur over
3%, SO2 emissions were only 50% to 75% of 100% conversion. This
is attributed to retention of the sulfur in the ash.
Retention of S in the ash of coal-fired industrial boilers was
investigated to compare the SO2 emissions with western and eastern
coals (30). The results indicated that about 90% of the S was
emitted as SO2 when firing eastern coals, but only about 80% was
emitted with western coals. Figure 13 shows the amount of S retained
in the ash as a function of total fuel sulfur content. Although
there is scatter in the data, there is an indication that the amount
of sulfur that is retained tends to be independent of sulfur content.
Industrial Boiler Sulfate Emissions With The SASS Train
Trace species and organics measurements were made on an indus-
37
-------
1 —
z
UJ
UJ
O_
^
CD
CO
CD
h—
X
CD
GO
Li-
CD
2E
CD
UJ
CD
L.\J
18
16
14
12
10.
i 1 1 I
D NATURAL GAS
O OIL
A COAL
— —
O
_ _
0
l=irN
8
6
4
2
0
-Tjt —
w
-O O
A
oRs A
-$) §P A ^ rO AA "
cr t o ^^ ^ni ^
0 400 800 1200 1600 20
TOTAL SULFUR OXIDES CONCENTRATION,
PPM, DRY AT 3 PERCENT 02
Figure 11. S03 emissions from industrial boilers,
38
-------
.
ca
Q_
Q_
O
OO
2400
2000
1600
1200
800
400
COAL, 100% EMISSION
A / ^
AA/ /°IU
^Y ^ / 100% EMISSION
O
FUEL TYPE
A COAL
O OIL
_L
0,0 1,0 2,0 3,0 4,0 5,0
FUEL SULFUR CONTENT, DRY, PERCENT
Figure 12. SO emissions from industrial boilers.
-------
0.5
CD U,4
UJ
I—
UJ
§ 0,3
<=»
UJ
-------
trial boiler (27). The EPA Source Assessment Sampling System (SASS)
was used and analytical results included sulfates. The data, Table
2, indicate the results for three tests: a baseline test and two
tests with combustion modifications. The modifications were flue
gas recirculation (34%-35%) and reduced excess air. The fuel was
No. 6 oil with sulfur content of 1.0% to 1.2% by weight. The data
indicate that most of the sulfur was recovered in the condensate
and impingers, mostly SO2 that passed through the filter. SO2
emissions were about 2200 mg as SO2/m3, which would be 4400 mg/m3
expressed as SO4 . Total sulfur collected by the SASS was about
one-fourth to one-half of the SO2 emission rate. S0$ emissions were
not measured during these tests, but other tests under similar opera-
ting conditions indicated an SOX to SO3 conversion of about 1.2% or
about 53 mg as SO4/m?. Sulfates collected in the cyclones, filter,
and XAD-2 organic module were from 18 to 20 mg as SO4/m3. About half
of the sulfate was collected on the SASS filter, but a large portion
was also collected in the XAD-2 module. Based on this very limited
test, there was no indication that the use of combustion modifica-
tions caused any significant change in total sulfate emissions or
in the distribution of sulfates by particle size.
SQ2/SO3 Determinations Using a Modified Controlled Condensated
Method in Utility Boilers
The variation of percentage conversion of total sulfur oxides to
SOs with total sulfur oxides is shown in Figure 14. The fuels used
varied in fuel sulfur content from 0.19% to 0.45%. The results did
show a variation with particular units. The percentage conversion
to SO3 averaged 3.4% but was spread over a range of 1% to 6.5%.
Diesel Exhaust S02/SO3 Data
Measurements of S02/S03 were made by Goksoyr/Ross and Shell-
Emeryville wet chemical methods as well as by Dupont UV photometric
instrumentation for SO2 readings.
The results of the Goksoyr/Ross and instrumental S02 measure-
ments were found to correlate well for fuel with medium sulfur
levels. S02 measurements made using the Shell-Emeryville method
were lower than by either of the other methods employed.
Figure 15 shows the variation of SOa level with sulfur dioxide
levels. For nine test conditions, an increase of SO3 level was
observed with increased S02 level. The percentage conversion to
SO3, however, varied inversely with sulfur dioxide level.
41
-------
Table 2. Sulfate Emissions from an Industrial Boiler
with Combustion Modifications
Test No.
Condition
Fuel Sulfur Content,
Excess 02
SOo Emissions, mg SC-2
19-2
Baseline
% 1.17
3.00
/m3 2210
19-3
Low NOx*
1.18
1.8
2310
19-4
Low NOX*
1.02
1.5
2150
Est. S03 Emissions
mg as S04/m3
SASS Train Sulfates,
mg S04/m3
53
55
*34%-35% flue gas recirculation and lowered excess O2
+NES = Not enough sample for analysis
52
10 Mm Cyc
3 urn Cyc
1 Mm Cyc
Filter
Wash
XAD-2 Resin
XAD-2 Rinse
Total, Front Half
Condensate, Impingers
1.2
NES+
NES
9.5
0.6
0.5
6
17.8
2600
1.1
0.74
NES
8.4
1.1
2.1
4.5
17.9
1200
0.6
0.63
NES
9.5
0.53
3.4
5.6
20.3
2000
42
-------
10
LU
CJ
Q±
CO
X
o
GO
GO
OC
LU
1 1 1 1 1 1
O UNIT 1
D UNIT 2
A UNIT 3
V Q V O UNIT 4
ftrO ri ^ UNIT 5
Q V ^ V V UNIT 6
- O ®O
% A 0
1 1 1 1 1 1
%
IN NO
0.21
0,19
0,19
0,19
0,21
0,21
1
S
, 6 OIL
- 0,40
- 0,45
- 0,34
- 0,34 -
- 0,27
- 0,34
—
i
50
100 150
TOTAL SOX,
200
250
300
350
PPM, DRY AT 3 PERCENT 0-
400
Figure 14. Conversion of SOX to S03 vs. total S0x in utility boilers.
-------
Q_
Q_
(=)
X
o
eel
CO
1 1—
O SHELL-EMERYVILLE
A GOKSOYR/ROSS
o o
A
A
I
I
100 200
SULFUR DIOXIDE, PPM
300
Figure 15. Sulfur trloxide vs. sulfur dioxide in diesel
exhausts.
44
-------
With the Goksoyr/Ross method, and fuel with lower sulfur con-
tent, there was an appreciable decrease in S03 level. Although insuf-
ficient data have been taken to make a definitive conclusion, there
does not seem to be a "shelf" value where the method determines the
lower level of sensitivity.
In the diesel studies, acid dewpoint temperature studies were
also conducted. For a fuel with medium sulfur content, the SO3
measurements taken employing the Goksoyr/Ross method were used in
conjunction with a measured moisture to calculate dewpoint tempera-
ture. These calculated dewpoint temperatures correlated closely
with dewpoint temperatures measured with a Land dewpoint meter.
These results increased confidence in the Goksoyr/Ross method.
Gas Turbine SO2/803 Emissions
The exhaust gas temperature from a gas turbine is in the range
of 1000°F (810 K), and there are no low temperature heat exchangers
in the exhaust to provide catalytic surfaces for S03 formation. If
S03 were formed primarily by catalytic action, one would expect to
see much lower SO3 levels from gas turbines as compared with boilers.
This is not the case, however; gas turbines produce 863/SOX ratios
that are quite similar to those from boilers and other combustion
equipment. KVB has measured SC>2/SO3 emissions from a large number of
gas turbines. The majority of that data was obtained for commercial
clients and was not, therefore, available for inclusion in this
publication. However, the data generally fall in the range of 2%
to 6% S03 with scatter much the same as shown in previous figures
for other devices.
An analytical study of the chemical kinetics of SO3 formation
in gas turbines indicated that S02 is converted to SO3 within the
gas turbine combustion chamber in the mid-region of the chamber where
hot combustion gases start to be cooled by the injection of excess
air. The analysis indicates that S03 emissions tend to increase
with combustor size and test results tend to confirm that. The
analysis also indicated that S03 might be reduced by modification
of the rate of injection and mixing of excess air. Gas turbines
require very close control of excess air mixing to provide a speci-
fic shape of temperature profile in the gases entering the turbine
wheel. This restraint severely limits the amount of adjustment
that can be made in the mixing rates and may preclude any changes
to control S03. No experimental work in that direction has been
conducted.
45
-------
CONCLUSIONS
Figure 16 shows a plot of S03, directly in ppm versus SO2, from
the various measurement methods and sources. There is an apparent
lower limit of about 3 ppm to 4 ppm SO3 that is independent of S02.
This limit is not readily apparent when percent SO3/SO2 is plotted.
This value can be attributed to the lower limit sensitivity of the
methods, rather than to a real tendency for increased conversion at
low S02 levels. There was, however, a large degree of scatter that
introduced uncertainty.
The Shell-Emeryville method was seen in some determinations
to give high S03 levels, especially at low fuel sulfur levels.
This can be attributed to the method where both S02 and SO3 can be
absorbed in the IPA solution. If the inert gas purge is not totally
effective, the remaining S02 will be trapped and determined as part
of the S03 fraction.
In the SCAQMD, there may be incomplete aggregation of H 2S04
mist particles. This would lead to low SO3 measurements. The
laboratory analytical procedures require significantly greater
effort than the simple titration of the Shell or controlled con-
densate methods.
Using the Goksoyr/Ross and other controlled condensate methods,
there are advantages of conditioning the flue gases to separate
particulate matter, the S03 fraction, and the S02 fraction. S02/
S03 separation is accomplished by a physical process. The tempera-
ture of the flue gases is reduced below the dewpoint temperature
of H2SO4 causing a condensation. These results indicate that the
controlled condensate methods should produce more reliable and
reproducible data less subject to operator influence experienced
with the wet chemical separation methods.
The percentage of SOX emitted as S03 appears to be largely
independent of the nature of the emission source, i.e., all de-
vices produce about the same level of conversion and show the same
degree of scatter in the data. This scatter makes it very diffi-
cult to assess any effects on SO3 emissions that may be caused by
operational modifications or other changes. However, the controlled
condensate methods appear to offer improved accuracy that may allow
such assessments to be made.
46
-------
25
£ 20
Q_
*»
LU
«=»
>< it:
o J.J
a;
i —
a:
=>
LJ— i n
i J.U
co
5
0
I 1 I i 1 I
O SHELL-EMERYVILLE, INDUSTRIAL PROCESSES
- •SHELL-EMERYVILLE, INDUSTRIAL BOILERS
QSHELL-EMERYVILLE, DIESEL ENGINES
AGOKSOYR/ROSS, DIESEL ENGINES
- DSCAQMD METHOD, INDUSTRIAL PROCESSES
a
-
Sb % eo" 0aD^°o o
1 * ^^ ^HV ^1^ ^^
pO^ i • i i i i n •
) 100 200 "" 300
I 1
o
"
—
-
o
9>c
1 1
aoo
SULFUR DIOXIDE, PPM
Figure 16. SO, vs. SO. for three measurement methods.
J 2
47
-------
Analysis of the gas phase kinetics of SOs formation indicates
that SO3 can be formed, at amounts typically observed, in the
absence of any surface catalysis. The critical temperature range
occurs as gases are cooled from 2540° to 1540°F (1670 to 1110 K).
Control of the rate of cooling and excess air mixing appears to
offer some degree of control over SOs formation that warrants fur-
ther study and as an alternative to the use of additives.
48
-------
REFERENCES
1. Cullis, C. F., and M. F. R. Mulcahy. The Kinetics of Gaseous
Sulfur Compounds. Combustion and Flame, 18:225, 1972.
2. Merryman, E. L., and A. Levy. Sulfur Trioxide Flame Chemis-
try - H2S and COS Flames. Thirteenth (International) Combus-
tion Symposium, The Combustion Institute, 1971. p. 427.
3. Levy, A., and E. L. Merryman. The Microstructure of Hydrogen
Sulfide Flames. Combustion and Flame, 9:229, 1965.
4. Sawyer, R. F. University of California at Berkeley, Con-
sultant to KVB, Personal Communication.
5. Levy, A., and E. L. Merryman. SO^ Formation in H2S Flames.
Trans. ASME J. Engrg Power, 87:374, 1965.
6. Merryman, E. L., and A. Levy. Kinetics of Sulfur-Oxide For-
mation in Flames. J. Air Pollution Control Assoc., 17:800,
1967.
7. Levy, A., and E. L. Merryman. Sulfur-Oxide Formation in
Carbonyl Sulfide Flames. Environmental Sci. and Technol.,
3:63, 1969.
8. Hedley, A. B. Factors Affecting the Formation of Sulfur Tri-
oxide in Flame Gases. J. Inst. Fuel, 40:142, 1967.
9. Breen, B. P., A. W. Bell, and N. Bayard de Volo. Combustion
Control for Elimination of Nitric Oxide Emissions from Fossil-
Fuel Power Plants. Thirteenth (International) Combustion Sym-
posium, The Combustion Institute, 1971. p. 391.
10. Balch, D. L., et al. High Temperature Reaction Rate Data,
V. 1-5, Dept. of Phys. Chem., The University, Leeds, England,
1970.
11. Wray, K. L., et al. Eighth (International) Combustion Sym-
posium, The Combustion Institute, 1960. p. 328.
12. Campbell, I. M., and D. A. Thrush. Trans. Faraday Soc.,
64:1275, 1968.
49
-------
13. James, D. E. A Boiler Manufacturer's View on Nitric Oxide
Formation. Presented to: The Fifth Tech. Meeting, West
Coast Section, The Air Pollution Control Assoc., San Francisco,
CA, October 1970.
14. Reese, J. T., et al. Prevention of Residual Oil Combustion
Problems by Use of Low Excess Air and Magnesium Additive.
Trans. ASME, J. Engrg. Power, 87A:229, 1965.
15. Gills, B. G. Paper 9. Production and Emission of Solids,
SOx, and NOx, from Liquid Fuel Flames. J. Inst. Fuel^ 46:N383,
February 1973.
16. Chaikivsky, M., and C. W. Siegmund. Low Excess-Air Combustion
of Heavy Fuel - High Temperature Deposits and Corrosion. Trans.
ASME J. Engrg. Power, 87A:379, 1965.
17. Lee, G. K., et al. Effect of Fuel Characteristics and Excess
Combustion Air on Sulphuric Acid Formation in a Pulverized -
Coal-Fired Boiler. J. Inst. Fuel, 40:397, September 1967.
18. Lee, G. K., et al. Control of SO in Low-Pressure Heating
Boilers by an Additive. J. Inst. Fuel, 42:67, February 1969.
19. Lee, G. K., et al. Fireside Corrosion and Pollutant Emission
from Crude Oil Combustion. Trans. ASME J. Etigrg. Power, p.
154, 1972.
20. Friedrich, F. D., et al. Combustion and Fouling Characteris-
tics of Two Canadian Lignites. Trans. ASME J. Engrg. Power,
p. 127, April 1972.
21. Barrett, R. E., et al. Formation of 80s in a Non-catalytic
Combustor. Trans ASME J. Engrg. Power, 88A:165, 1966.
22. Anon. Determination of Sulfur Trioxide and Sulfur Dioxide in
Stack Gases, Absorption-Titration Method. Shell Method Series
62/69, Shell Standardization Committee (North America), 1969.
23. Devorkin, H., et al. Air Pollution Source Testing Manual.
Air Pollution Control District, Los Angeles County, CA (Re-
organized as South Coast Air Quality Management District, El
Monte, CA), December 1972.
24. Lisle, L. S., and J. D. Sensenbaugh. The Determination of
Sulfur Trioxide and Acid Dew Point in Flue Gas. Combustion
36:12, 1965.
50
-------
25. Hunter, S. C., and N. L. Helgeson. Control of Oxides of Sulfur
from Stationary Sources in the South Coast Air Basin of
California. NTIS No. PB 261 754, June 1975.
26. Cato, G. A., et al. Field Testing: Application of Combus-
tion Modifications to Control Pollutant Emissions from Indus-
trial Boilers—Phase II. EPA 600/2-76-086a, NTIS No. PB 253
500, April 1976.
27. Cato, G. A. Field Testing: Trace Element and Organic Emis-
sions from Industrial Boilers. EPA 600/2-76-086b, October
1976.
28. KVB, Inc., data obtained from Industrial and Utility Client
programs, 1970-1977.
29. Engel, P. K. Diesel Exhaust Fouling and Corrosion Evaluation
Program, Task II Diesel Exhaust Analysis, Final Report. KVB
21803-775, April 1978.
30. Maloney, K. L. Systems Evaluation of the Use of Low-Sulfur
Western Coal in Existing Small and Intermediate-Sized Boilers.
EPA Contract 68-02-1863, draft report submitted to EPA.
51
-------
Some Recent Data on SO3 and SO4 Levels in
Utility Boilers
Brian W. Doyle
Richard C. Booth
KVB, Inc.
ABSTRACT
A series of recent measurements shed new light on the
typical formation processes of S03 and SO4 in utility
furnaces fired on high-sulfur oil. Measurements of SO3
were made using the controlled condensation (Goksoyr-Ross)
method, and the filter on the sampling end of the probe
was analyzed quantitatively for S04 and other constituents,
Measurements were made in the superheater at several down-
stream locations in the convective section and at the air
heater outlet of two boilers.
The data suggest that most or all of the SO3 is formed in
the hot section of the furnace and that some of it is con-
verted to 864 as the flue gases move through the convec-
tive sections of the furnace. Limited data will also be
presented on the influence of furnace excess air levels
and on the influence of magnesium oxide additives in the
fuel.
INTRODUCTION
Furnaces fired on residual oil emit nearly all ihe sulfur in
the fuel in the form of three oxides, SO2, S03, and SO4. The amounts
of the last two are small and variable compared to SO2, and, to a
large degree, a quantitative understanding of the formation mechanism
of SO3 and SO4 is missing. This must be attributed in part to the
difficulty of rapidly and reliably measuring the amounts of either
of these substances in the flue gas. The work described here sheds
53
-------
some light and raises questions on both the dominant formation
mechanisms and the measurement techniques for S03 and SO4.
The work described here is a portion of a program conducted by
the Niagara Mohawk Power Corporation at one of their power plants.*
The objective of this phase of the work is to minimize emissions of
S03 gas or condensed acid to help control emissions resulting from
corrosion. S03 formation can be formed in the furnace area of the
boiler by gas phase reactions, or it may be formed in cooler areas
of the boiler by the plentiful catalytic oxidation of SO2. The rela-
tive contribution of these potential sources influences the methods
chosen to try to reduce stack SO3 levels as well as the choice of
how to assess the effectiveness of an SO3 reduction program.
Most utility boilers firing high-sulfur oil use additives of
some sort to control fouling and corrosion of the superheat areas
or to reduce corrosion of the air heaters. The first problem re-
lates to fuel ash constituents such as vanadium and the second to
sulfuric acid derived from flue gas S03. Catalytic SO3 formation
mechanisms are sensitive to the surface composition of the tubes
in the convective passes, which tends to be dominated by the thin
layer of accumulated ash and corrosion products. Thus, any change
in an additive which feeds this layer may not be reflected in down-
stream S03 concentration for several weeks. However, any effect
which a fuel additive has on gas phase formed S03 should be apparent
as soon as the additive is started or stopped. The motivation for
the measurements reported here was first to delineate the relative
contribution of gas phase and catalytic SO3 in the subject boilers,
and second to determine the effect of fuel additives on gas phase
formed S03. Both questions have been partially answered, and some
aspects of this work are continuing at the present time.
EQUIPMENT
SO3/SO4 Sampling
Gaseous S03 was measured using the controlled condensation
method (Goksoyr-Ross coil). Flue gas was drawn from the furnace
utilizing an air-cooled Pyrex probe. A quartz wool filter at the
probe inlet was saved after each test and quantitatively analyzed
by wet chemistry for sulfates and magnesium. Pumping and control
of the sampling system utilized components from a Method 5 (EPA)
sampling train.
this
*
Niagara Mohawk's sponsoring role and permission to present
work are gratefully acknowledged.
54
-------
Flue gas extraction utilized a probe, shown schematically in
Figure 1, which was intended to hold the sampling tube at relatively
constant temperature of 300°F to 450°F while sampling flue gases at
temperatures from 250°F to 1400°F. This probe used a Pyrex glass
sampling tube inside a double steel jacket. Cooling air flowed
toward the probe inlet in the outer annulus and back along the inner
annulus, and the air flow rate was manually adjusted to achieve
temperatures of about 500°F or below when sampling high temperature
regions of the boiler. When sampling colder regions, below about
350°F, the flow was set at a low level and the electric resistance
heater was used. Both this heater and the one around the Goksoyr-
Ross coil were thermostatically controlled from thermocouple sensors.
A plug of quartz wool was packed into a bowl at the end of the
pyrex probe to serve as a filter. This filter operated at tempera-
tures near the flue gas temperature. Each filter was saved and
subsequently analyzed microchemically for sulfates and magnesium.
The sulfate analysis utilized a water wash of the filter for soluble
sulfates, followed by alkaline reaction with peroxide to oxidize
any SOa. Sulfates were then precipitated using BaCl and weighed.
The method missed insoluble sulfates and probably reported absorbed
S02 as SO4. Magnesium was determined using atomic absorption. 02
concentrations at the exit of the dry gas meter were measured with
a teledyne portable analyzer.
The condensation coil was a blown glass assembly approximately
12 inches long with both the glass coil and a fritted disk surrounded
by a water jacket. The glass assembly was wrapped with a small
electric heating blanket and fitted into a short length of 3 inch
PVC pipe to provide field durability. Operation between 150°F and
200°F was considered acceptable. Subsequent to each test, the
condensing surfaces, including the fritted disk, were rinsed repeat-
edly with 5% isopropyl alcohol. The Pyrex probe was also rinsed,
and the total rinse was titrated with NaOH. Early experience showed
that roughly one third x»f the total acid catch could accumulate in
the probe.
Sampling utilized the system shown schematically in Figure 2.
The probe and coil were followed by one or two impingers to dry the
sample, then a small pump and a dry gas meter. Most samples were
1.5 ft3 collected over about 20 minutes. Frequently titrations were
done immediately. Utilizing a crew of two people, it was possible
to complete seven or eight tests a day, including the moving of
equipment from one part of the boiler to another.
55
-------
Furnace Wall
en
O)
Electric Heater
Fitting to Condensation Coil
1 1/2" ID Steel Tube
Quartz Wool
© Thermocouples
Air Supply
-Temperature Controller
110 Volt
Figure 1. Probe for constant temperature sampling.
-------
Impingers
Dry Gas Meter
en
-j
Condensation
Coil
Ait Cooled Probe
Meter
Temperature Control
110 V-
Air
Supply
•40"
Figure 2. Sampling configuration.
-------
Boilers
Five boilers are in operation at the site, and data on three
of them are presented here. Units 2 and 3 are fuel converted (coal
to oil) units with nominal 80 megawatt ratings at present. The
design of these units promotes substantial stratification in flue
gas 02 concentrations. This probably influences the production of
S03 and SO4 and is known to make repeatable measurements of furnace
excess air levels either difficult or meaningless. However, delib-
erate changes in furnace air flows could be meaningfully measured
and the effects assessed as long as boiler operation was steady.
Both units are equipped with a substantial number of viewing sample
ports so that access to the flue gases was not difficult at various
points between the superheater and the air heater exit. Figure 3
shows the cross section and various access port locations of Boiler 2.
Unit 5 is a new boiler rated at 850 megawatts. Access to the
flue gas on this boiler was available between the economizer and
the air heater, between the air heater and the electrostatic pre-
cipitator, and downstream of the precipitator.
During this testing all boilers were fired continuously with
residual fuel of 2.3% sulfur content. All the boilers use a mag-
nesium oxide based fuel additive to help control vanadium related
deposits in the superheat sections. The volumeteric pumping rate
for this additive varies from 1/2000 to 1/3000 depending on the
boiler. The accuracy or consistency of MgO feed to the boilers is
not known. Fuel analysis from a single date on two boilers shows
260 ppm vanadium, 60 ppm magnesium, and 30 ppm sodium.
During all testing, the test boiler was maintained at steady
load. Operating parameters relevant to flue gas flow, temperature,
and composition were monitored by the test engineer.
RESULTS
Data were gathered over approximately a three-month period on
the three boilers. Table 1 is a partial tabulation of data for
Boilers 2 and 3. This is a sufficient sampling to indicate the
ranges of the various parameters. Present in much of the data is
a significant scatter that can be attributed to day-to-day varia-
tions in furnace operation (actual excess air levels, choice of
burners, etc.) which in many respects cannot be monitored or con-
trolled. The reported 02 levels are those of the sampled gases and
are not representative of furnace excess air levels. All concen-
trations have been corrected to a standard air dilution of 3% O2.
The sampling and laboratory procedures were of high quality, and
instances of obvious error have been deleted.
58
-------
Figure 3. Cross section of furnace with sampling locations
59
-------
Table 1. Partial Listing of S03/SO4 Data
Test
1A
1C
IE
1G
2A
2B
2E
3A
3C
3F
4B
4C
5A
5C
6A
6B
6F
7A
7B
7E
Location
2-1
2-2
2-3
2-4
2-6
2-6
2-6
2-5
2-1
2-1
2-1
2-1
2-5
2-5
3-2
3-2
3-2
3-4
3-4
3-4
(%?b
7
3
3.9
2.5
9.4
7.5
8.9
8.6
7.5
7.0
4.6
2.0
2.8
1.5
(5.5)
(5.5)
(5.5)
4.8
4.8
5.0
Temp.
1170
1380
1065
865
315
440
380
645
1190
1190
1250
1180
1600
1600
1600
910
910
910
SO
16
19
13
16
15
27
21
11
25
12
25
9
12
8
27
29
42
25
21
26
ppm by
3 S04
13
8
14
10
21
6
16
24
10
15
9
11
.6 13
9
197
218
262
14
10
10
Vol @ 3%
Total
29
27
27
26
26
33
37
35
35
27
34
20
26
17
224
247
304
39
31
36
02
Mg
16
11
10
4
3
1
3
13
4
11
17
12
7
3
262
262
255
10
5
4
Operation
Normal
ii
"
ii
H
"
"
Normal
No Additive
Normal
Normal
Low Air
Normal
Low Air
Location refers to boiler number and locations marked on
Figure 3.
Concentration of the sampled gas.
60
-------
An indication of the manner in which S03 and total sulfate
(S03 + S04) concentrations change as the flue gas moves through the
furnace can be obtained by plotting concentration vs. flue gas tem-
perature, as is done in Figure 4. This clearly shows that the highest
observed SO3 concentrations occur closest to the main furnace com-
bustion zone. Instances of silica catalyzing the oxidation of SO2
at temperatures above 1100°F have been observed elsewhere. While
this may have contributed to measured S03 in the superheat, a test
without the quartz wool plug showed the same SO3 as the previous
test. S03 concentrations tend to decrease or remain constant down-
stream of the furnace section of the boiler.
An apparent increase in total sulfates between the economizer
and the air heater exit may be real or a result of insufficient
sampling in the economizer area. Much of the data scatter can be
attributed to variations in furnace excess air levels which, as
noted earlier, are not easy to delineate on a day-to-day basis.
Tests 4 and 5 of Table 1 were run during a single shift and show
the potential variation in S03 due to changes in excess air levels.
Within the operating range of these units, a variation in SO3
emissions of roughly 3/1 can be achieved with changes in excess air
level. This suggests that in using techniques such as low excess
air to control SO3, it is important to control air to the combus-
tion zone, and that control of air leakage in the convective passes
is less important.
It is reasonable to expect that some sulfur oxides accumulate
on boiler tube surfaces in the process of scale formation. These
deposits are removed periodically by sootblowers. No sootblowing
operations were conducted during a test, so this accumulative
material would not have added to the filtered substances. Thus, a
decrease in SO4, and perhaps SO3, can be attributed to accumulation,
while an increase reflects additional formation.
Magnesium concentrations tend generally to decrease as the gases
move through the convective section of the boiler. This may be
attributed to accumulation on the tubes which is the purpose of the
additive. An anomaly in the data occurred during Test No. 6. It
appears likely that the additive pump was improperly set during
this period of time. This could explain the magnesium but not the
simultaneous large sulfate measurements. Reviewing all the data
and remembering that the precision of measurement for Mg is roughly
5 ppm, one notes that sulfate levels are seldom significantly below
magnesium levels. Figure 5 shows an apparent correlation between
Mg and S04. All the S04 could be magnesium in the form of sulfate,
but it cannot be determined here whether MgS04 formation occurred
61
-------
O)
KJ
JN 60
c
o
•H
4-i
nj
fi
01
o
c
o
CJ
E
(X
40
20
.
f
o
o
0
Furnace
0
0
fc
0
• •
9
^
°0
Superheat
•
so3 + so4
O SO« •
• ,
jl •
•
o
oP o 0
o
J
•to
*
o
o
o
Economizer
*
4
(
0 <
<
i
_
•
^ O
j
)
)
Air Heater
V
•
0
o
•
o
o
o
1400
1200
1000 800 600
Flue Gas Temperature (°F)
400
Figure 4. S03 and S04 data for boiler no. 2.
-------
Data Predomtnantly from
Air Heater Region
IL
I I I I I I
8 10 15 20 30 40 60 80 100
Magnesiumi ppm in Flue gas at 3% 02
150 200
300 400
Figure 5. Correlation of magnesium and sulfates on quartz filter.
-------
in the flue gas or as those gases were drawn through a filter coated
(after a short period of testing) with MgO. In either case, one is
led to the conclusion that the presence of magnesium oxides leads
to the conversion of SO2 to S04 in amounts equivalent on a molar
basis to the amount of magnesium. While these data may condemn the
test procedure, they also call into question how effectively Mg or
MgO reacts with gaseous SO3.
Test 3 exemplifies data which suggest that the MgO additive
may have been responsible for a small suppression in the gas phase
formation of S03. If in fact this is occurring, this and other
data indicate that the effect is no larger than 30% of the zero
additive S03 level. This makes no statement about the effect of
fuel additive on catalytic processes in the convective section.
Since the tubes of these boilers are permanently influenced by the
additive, the S03/S04 levels with no additive cannot be determined.
Boilers 2 and 3 are equipped with tubular air heaters, which
are known to experience difficulties with corrosion and plugging.
However, the typical exit air temperatures are high (350°F), and
the data here give no indication of diminished S03 levels at the
air heater exit. Either the amount of condensation is low or the
condensation occurs during operating conditions different from
those during the testing (i.e., very low load).
Boiler 5 has a rotary air heater with temperatures much lower
than the smaller boilers. Table 2 lists the measurements made on
both sides of the air heater and at the precipitator exit. These
data are substantially more repeatable than the data on the smaller
boilers due probably to more predictable boiler operation. A
coherent pattern is clearly established.
Table 2. S03/SO4 Data on Unit 5
Test
Location
02
Temp.
ppm by Vol @ 3%
S0
SO
Total
Mg
1A
.ID
2A
2E
3A
4A
4D
Precip. Outl.
n it
AH Out
AH In
AH In
AH In
AH In
4.6
5.3
6.2
5.7
4.9
3.4
5.0
320
285
310
695
725
700
700
2.0
6.6
4.1
20
23
18
23
10
11
29
12
10
16
11
12
17
33
32
33
34
34
3
4
15
15
5
5
5
64
-------
As the flue gases pass through the air heater, the S03 level
is diminished by about 15 ppm. Although the data are limited, this
corresponds to a 15 ppm increase in S04 and suggests that the air
heater is condensing the gaseous S03 to form acid. This is not in-
conceivable, given the relatively low exit gas temperatures and
known corrosion problems on this unit. When the gases leave the
precipitator, the SO4 and total sulfates are reduced. A precipitator
is capable of collecting an acid mist, and it appears that much of
the mist created by the air heater is being collected by the precip-
itator. A final piece of data is that the ash from the precipitator
is almost 50% sulfate by chemical analysis. Given the typical fuel
ash levels of 0.1% and 15 ppm of collected sulfate in the precipi-
tator, a mass balance suggests that an ash with 50% sulfate should
be collected in the precipitator.
The cohesiveness of the data in this case suggests that the
test method has considerable value and that much of the scatter
observed in other situations is a real variation in the emissions.
CONCLUSIONS
The controlled condensation technique for measuring SO3 con-
centration appears to yield rapid and consistent results during
field use. Measurement of S04 levels by quartz wool in-stack fil-
tration may suffer from interference by magnesium oxide.
Measurements in the higher temperature regions of the furnace
are practical and useful for delineating formation regions and
mechanisms.
SO3 formation in the furnace section of the boiler is substan-
tial and much greater than the formation by catalytic reactions in
cooler sections. However, this conclusion is drawn from a furnace
with low superheat temperatures which continuously uses significant
quantities of MgO based fuel additive.
The use of low excess air to minimize SO3 formation is effec-
tive on boilers with casing leaks and substantial amounts of O2
present in the convective passes and the air heater.
MgO based fuel additives have relatively little effect on
flame zone SO3 formation.
65
-------
Measurement of Sulfur Oxides from Coal-Fired
Utility and Industrial Boilers
William R. McCurley
Daryl G. DeAngelis
Monsanto Research Corporation
ABSTRACT
The Source Assessment Program sponsored by IERL/EPA
under contract 68-02-1874 involves the characterization
of a wide range of pollutant emissions for selected
industries in the United States. As part of this
program, emissions, including those of sulfur oxides
and particulate sulfates, were measured for industrial
and utility dry bottom boilers firing pulverized bi-
tuminous coals. One typical source in each of the two
categories was sampled before and after the electrostatic
precipitator.
Sulfur emissions from these sources were sampled using
a modified EPA Method 8 train in which a filter was
inserted between the probe and the first impinger to
collect particulate sulfates. Samples for sulfur
dioxide, sulfur trioxide, and particulate sulfate
analyses were collected simultaneously by maintaining
an isokinetic sample flow.
The study results will be compared with the sulfur
balance for each system and with the published emission
factors. In the case of the utility boiler, the results
will be compared with data obtained by a continuous in-
stack monitor. Analytical results obtained by on-site
titrations will be compared with values measured in the
laboratory. Differences in the concentration of sulfur
oxides observed before and after the electrostatic
precipitator at the industrial site will also be discussed
67
-------
Monsanto Research Corporation, under EPA contract No. 68-02-
1874 (Source Assessment), has the responsibility of characterizing
the air emissions, wastewater effluents, and solid wastes discharged
from selected sources and assessing their environmental impact.
This paper describes the measurement of sulfur oxide (SOX) emissions
at two sites typical of two of the source types under study: indus-
trial and utility dry bottom boilers firing pulverized bituminous
coal. The results presented here represent only a small part of
the extensive sampling effort performed at both sites. The primary
objective of the Source Assessment program is to provide the EPA
with sufficient information to decide whether emissions reduction
is necessary. Meeting this objective involves the characterization
of a wide range of pollutants from a large number of stationary
sources in four categories: organic sources, inorganic sources,
open sources, and combustion sources. In the combustion area alone,
56 source types have been identified for assessment (1) by MRC and
TRW. These are listed in Table 1. The relationship of the source
types discussed in this paper with those listed in Table 1 is illus-
trated in Figures 1 and 2 according to fuel use and SOX emissions.
These two source types can be briefly defined as all boilers
(steam generators) which meet each of the following criteria:
1. The primary fuel is pulverized bituminous coal.
2. The operating temperature of the boiler furnace is
kept below the ash fusion temperature so that ash
remaining in the furnace can be removed as a dry
powder (definition of dry bottom).
3. The product of these boilers (steam) is used for
electricity generation for public sale for units in
the utility category or is used for process heating,
space heating, electricity generation for on-site
use, or other miscellaneous uses after being gene-
rated at an industrial site in the case of industrial
boilers.
While coal used by these source types accounts for only 39%
of the fuel used by utilities and 6% of the fuel used in industrial
boilers, these sources as defined are responsible for 63% and 34%
of the SOX emissions from all utility and industrial boilers, re-
spectively (1). In addition, the population of these units, par-
ticularly in the utility area, is expected to increase signifi-
cantly over the next decade in response to governmental policies
directed toward energy self-sufficiency (2).
68
-------
Table 1. Stationary Combustion Sources Identified for Assessment (1)
Combustion System
System
No.
Combustion System
System
No.
Electric Generation
External Combustion
Coal
Bituminous
Pulverized Dry 1
Pulverized Wet 2
Cyclone 3
All Stokers 4
Anthracite
Pulverized Dry 5
All Stokers 6
Lignite
Pulverized Dry 7
Pulverized Wet 8
Cyclone 9
All Stokers 10
Petroleum
Residual Oil
Tangential Firing 11
All Other 12
Distillate Oil
Tangential Firing 13
All Other 14
Gas
Tangential Firing 15
All Other 16
Refuse 17
Internal Combustion
Petroleum
Gas
Internal Combustion/
Gas Turbine
Petroleum
Distillate Oil 18
Gas 19
Internal Combustion/
Reciprocating Engine
Petroleum
Distillate Oil 20
Gas 21
Industrial
External Combustion
Coal
Bituminous
Pulverized Dry 22
Pulverized Wet 23
Cyclone 24
All Stokers 25
Anthracite
All Stokers 26
Lignite
Spreader Stokers 27
Petroleum
Residual Oil
Tangential Firing 28
All Other 29
Distillate Oil
Tangential Firing 30
All Other 31
Gas
Tangential Firing 32
All Other 33
Waste 34
Internal Combustion
Petroleum
Gas
Internal Combustion/
Gas Turbine
Petroleum
Distillate Oil 35
Gas
Internal Combustion/
Reciprocating Engine
Petroleum
Distillate Oil 37
Gas 38
Commercial/Institutional
External Combustion
Coal
Bituminous
Pulverized Dry 39
Pulverized Wet 40
All Stokers 41
Anthracite
All Stokers 42
Lignite
All Stokers
Petroleum
Residual Oil
Tangential Firing 43
All Other 44
Distillate Oil
Tangential Firing 45
All Other 46
Gas
Tangential Firing 47
All Other 48
Refuse
Internal Combustion
Petroleum 49
Gas 50
Residential
External Combustion
Coal.
Bituminous 51
Anthracite 52
Lignite 53
Petroleum
Distillate Oil 54
Gas 55
Refuse
Wood -r>6
69
-------
20 r
UTILITY
INDUSTRIAL
COMMERCIAL/
INSTITUTIONAL
RESIDENTIAL
16.2
15
i_
03
00
o
f— (
z~
0
£ 10
I
C/1
z
o
0
o
1 ' *
z
LU
5
-
"
_
.
-
.
-
-
-
-
11.9
9.0
6.3
^7
w
w
W/v
vs/fr
w^,
8.5
1.4
0.7
KXXXZfl
4.7
0.16
0.2
i 1
Total Ail All Utility Pulv. Total All All Pul». All AIIConm.1 All ResMenlia All Residential
Utility Coal Bituminous Industrial Industrial Bituminous CommJInst In si Coil Combustion Coal Combustion
Combustion Combustion Dry Bottom Combustion Coal Dry Bottom Combustion Combustion
Utility Combustion Industrial
Boilers Boilsrs
Figure 1. Distribution of energy consumption by source type (1),
-------
15
14.5
CO
-2 10
o
'ZZ
•«•-»
o>
en
o
00
O
to
UTILITY
12.7
INDUSTRIAL
COMMERCIAL/
INSTITUTIONAL
RESIDENTIAL
9.1
2.9
2.0
1.0
1.4
1.3
Total All All Utility Put*.
Utility Coil Bituminous
Combustion Combustion Dry Bottom
Industrial
Boilers
Total All All Pulv.
Industrial Industrial Bituminous
Combustion Coal Dry Bottom
Combustion utility
Boilers
AH AMComm.;
Comm.Mnst Insl Coal
Combustion Combustion
AH Residential All Residential
Combustion Coal Combustion
Figure 2.
Distribution of SO emissions by source type (1).
X
-------
SITE DESCRIPTION
The industrial boiler selected for testing was a horizontally-
fired dry bottom unit with a rated firing capacity of 130 GJ/hr and
an output capacity of 45,000 kg of steam/hr. The boiler is fired
with a low sulfur (<1.0%) Appalachian bituminous coal and produces
steam for process and space heating at an industrial site. Partic-
ulate emissions are controlled by a high efficiency (99.0%) electro-
static precipitator (ESP). A schematic of the boiler system showing
the path of the flue gas is presented in Figure 3. Air emissions
were sampled at the inlet and the outlet of the precipitator.
Emission testing for the utility boiler assessment was con-
ducted on a tangentially-fired dry bottom boiler. The boiler has a
design firing capacity of 970 GJ/hr and an output capacity of 590,-
000 kg of steam/hr at 12.4 MPa and 540°C. Steam produced by this
boiler and a similar unit is used to drive a 180 MW electricity
generating turbine. Emissions from this boiler are controlled by a
mechanical collector (MC) and two ESPs in series. The particulate
control system had an overall collection efficiency of 99.97%.
Figure 4 illustrates the boiler system showing the path of the flue
gases and the sampling points. Sulfur content of the coal burned
during sampling ranged from 1.7% to 2.5%.
Results of the analyses of the coals used to fire the indus-
trial and utility boilers tested are listed in Table 2.
SAMPLING AND ANALYTICAL PROCEDURES
Sampling and analytical procedures used for the determination
of sulfur dioxide, sulfur trioxide, and particulate sulfates fol-
lowed the method outlined in the Federal Register for EPA Method 8
(3) with the exception of several modifications. In order to col-
lect a sample for particulate sulfate analysis, an additional glass
fiber filter was inserted between the probe and the first impinger.
This filter was enclosed in a heated box, and the temperature was
maintained at 150°C or above to prevent the collection of any sul-
fur ic acid mist. Leak checks were performed by plugging the inlet
to the particulate sulfate filter. Operation of the sampling train
was as specified in the Method 8 procedure. This included travers-
ing the flue duct and maintaining isokinetic sampling conditions.
At the conclusion of each run, the particulate sulfate filter
was placed in a petri dish and the dish was sealed. The probe and
front half of the filter holder were washed with distilled water
which was bottled and returned to the lab for sulfate analysis.
Remaining portions of the sampling train were washed and bottled
following the Method 8 procedures.
3
72
-------
Sampling
Point
Boiler
Schematic of the industrial boiler system.
Coal Feed
Preheated Air
Sampling
Point
\7\7 \7\7
85 % Efficiency 99.5 % Efficiency
65% Efficiency
Figure 4. Schematic of the utility boiler system.
-------
Table 2. Analysis of Coal Burned at Test Sites
Test run no.
Industrial site
All runs
Utility sitea
MC outlet
1
2
ESP outlet
1
2
3
Heating
value Moisture Ash
MJ/kg content content
(Btu/lb) % %
a.
31.38 8.41 8.23
(13,411)
28.14 1.62 14.59
(12,113)
26.57 1.18 23.11
(11,439)
26.02 1.17 19.48
(11,202)
27.61 1.06 15.89
(11,886)
26.58 2.25 17.40
(11,441)
Volatile Fixed
Sulfur matter carbon Sulfate
content content content content
% or or or
/o fa fa
0.91 b 71.59C 0.09
2.23 35.39 48.40 b
2.45 33.40 42.31 b
1.81 34.35 45.00 b
1.99 35.93 47.12 b
1.72 33.99 46.36 b
^sampled before pulverizer
no data available
ctotal carbon
-------
Particulate sulfate samples from the industrial site were
analyzed for water soluble sulfates. Solid samples collected on
the particulate filter were extracted with hot water. The proce-
dure involved: (1) refluxing the filter and particulate matter in
30 ml of water for a minimum of 30 minutes, (2) filtering the hot
water through Whatman 41 filter paper, (3) washing the filter twice
with 16 ml portions of water, (4) diluting the filtrate and filter
washes to 250 nil with isopropanol, and (5) titrating aliquots of
the 250 ml sample with barium perchlorate (0.01N Ba(C104)2) using
thorin indicator. The combined probe and filter holder wash samples
were also diluted with isopropanol to 500 ml or 1000 ml depending
on the sample size and titrated with barium perchlorate as described
above.
A spectrophotometric procedure was used in the analysis of the
utility particulate sulfate samples. The method, developed by
Bertolacini and Barney (4) and later refined by Schafer (5), is
based on the reaction of sulfate with the barium salt of chlora-
nilic acid (2,5-dichloro- 3,6,-dihydroxy-p-benzoquinone) in
isopropanol to produce barium sulfate and chloranilate ion. The
absorbance of the chloranilate ion is then measured at 530 m/j. and
compared to a standard curve.
Sulfur dioxide and sulfur trioxide were determined in both
the industrial and utility samples by titration of the impinger
solutions with barium perchlorate as specified in Method 8.
RESULTS AND DISCUSSIONS
Results of the emission testing for SOa, SO3, and particulate
sulfates are presented as emission factors in Table 3. Those from
the utility boiler include values from analyses done in the field
and in the laboratory. Emission factors calculated with data ob-
tained from an in-stack continuous SOX monitor at the utility site
are shown for comparison. Also in Table 3 the measured emission
factors are compared with emission factors calculated using the
formula in AP-42 (6). In most cases measured and calculated values
show good agreement.
Table 4 compares the mass emission rates of sulfur measured
at the sites sampled with the feed rate of sulfur to the boiler,
thus providing a partial sulfur mass balance. Emissions of sulfur
species measured before the ESPs, with one exception, account for
from 92% to 112% of the sulfur entering the boiler. Emissions
after the ESP units appear to be somewhat lower, accounting for
from 34% to 91% of the sulfur entering the furnace of the industrial
boiler and from 88% to 97% for the utility boiler.
75
-------
Emissions of S02, SO3, and Partieulate Sulfate from Dry Bottom
Industrial and Utility Boilers Burning Pulveriy,ed IHtuminous Coal
Sampling location
ami test, number
I ndus t r i al site
Uncon tro! led
Run No. 1
Run No. 2
Run No. 3
Average
ESP Outlet
Run No. 1
Run No. 2
Run No. 3
Run No. 4
Average
Utili ty si to
MC Outlnt
Run No_._ 1
Ki f ! d ana 1 ys is
Lab anal ys i s
Continuous ana!yx.or
Kun_No._2
I'icld iinn lysh:
l.ab .-Hiaiysis
KHP Out. 1<-I
Hun No. 1
l.ab analysis
Continuous analy/or
. Run No. 2
Lab analysis
Continuous analy/.cr
Run No. 3
Lab analysis
Continuous analyser
Coal
sul fur
con ton 1.
0.91
0.91
0.91
0.91
0.91
0.91
0.91
0.91
0.91
•iia:!
'tv,
I .81
I .81
1 .99
1.99
1.72
1.72
Hmission
Par tirul ate
sul fate
as S04
0.019a
0.021*
0.027
0.023a
0.024a
0.008
0.025
0.003
0.015a
b
<0 . Ofi?
h
h
<0. <>!>?>
0.0006
i)
<0.0002
b
0.001
b
factors K/kg of coal
iSul fur t r iox ido
and sul fur i c
acid mist.
as S03
0.019
0,017
0.01S
0.018
0.023
0.079
0.12
0.031
0.063
0. 30
0,48
b
0.17
0 .3 1
0.2(i
b
0.26
b
0.22
b
food
Sul fur
dioxide
as SO2
16.8
17.4
16.9
17.0
14.1
6.1
9.9
16.5
1 1.7
49.8
23 .0
44.5
:>:: .4
31.8
39 .2
34.7
40.2
33.1
43.1
AP-42
SO, emission
fan tor
K/.KE
17.3
17.3
17.3
17.3
17.3
17.3
17.3
17.3
17.3
42.4
42.4
42.4
4"::;
34.4
34.4
37.8
37.8.
32.7
32.7
% Agreement Fraction
of measured of SO,
values with as SO-,
AP-42 (1J
97 0.11
101 0.10
98 0.11
98 0.11
82 0.16
36 1.3
58 1.2
96 0.19
68 0.54
119 O.lj
55 2 . 0
ii:; o.:i:i.
10-1 tl.(i-l
94 0.81
113 b
93 0.74
106 b
102 0.66
131 b
water soluble sulfates only
no data available
76
-------
Table 4. Comparison of SO2, S03, and Particulate Sulfate
Emission Rates to the Coal Sulfur Feed Rate
Sampling location
and test number
Industrial site
Uncontrolled
Hun No. 1
Hun No. 2
Run No. 3
Average
ESP Outlet
Run No. 1
Run No. 2
Run No. 3
Run No. 4
Average
Utility site
MC Outlet
Run No. 1
Field analysis
Lab analysis
Continuous analyzer
Run No , 2
Field analysis
Lab analysis
ESP Outlet
Run No. 1
Lab anal ysi :;
Con ti minus unalywr
Hun No. 2
Lah anal yw i w
Continuous final yy&r
Run No. 3
Lab analysi s
Continuous analyzer
Coal
feed rate
metric
tong/hr
3.42
3.42
3.34
3.39
3.34
3.42
3.42
2.86
3.26
30.5
30.5
30.5
34.6
34.6
32.7
32.7
31.7
31.7
33 . 5
33.5
Coal
sulfur
feed rate
kg/hr
31.1
31.1
30.4
30.9
30.4
31.1
31.1
26.0
29.7
681
681
681
847
847
592
S!)2
631
631
576
576
Sulfur emission rates,
kg/hr as sulfur
SO,
0.022
0.024
0.030
0.026
0.026
0.008
0.028
0.003
0.016
a
<0.60
a
a
0.68
0.007
a
<0.003
a
<0.014
a
SO3
0.026
0.024
0.024
0.024
0.031
0.11
0.16
0.036
0.082
3.6
5.9
a
2.3
4.3
3.4
a
a
2.9
a
SO2
28.7
29.7
28.2
28.8
23.6
10.5
17.0
23.7
19.0
760
351
678
906
829
520
641
550
638
554
721
Total
measured
sulfur emission
rate , kg/hr
28.7
29.8
28.3
28.3
23.7
10.6
17.2
23.7
19.1
764
358
678
909
833
523
(Ml
553
638
S57
721
% of
sulfur
accounted
for
92
93
94
78
34
55
91
64
112
53
100
107
98
88
108
88
101
97
125
no data collected
77
-------
Concentration of sulfates on particulate emissions are pre-
sented in Table 5 to compare the results of each test program on
a more uniform basis. Varying particulate collection efficiencies
obscure the comparison of particulate sulfate emission factors.
Results presented in this table show sulfate concentrations ranging
from 0.2 g/kg for uncontrolled emissions to 8.3 g/kg at the ESP
outlet, indicating sulfate enrichment on fine particulates.
Table 5. Particulate Sulfate Concentrations
Sulfate concentration as g
Run No. of sulfate/kg of particulate
Industrial site
Uncontrolled
Average 0.19
ESP Outlet
Average 8.3
Utility site
MC Outlet
1 <1.5
2 <1.5
ESP Outlet
1 4.1
2 <1.5
3 8.1
Discussion of Industrial Boiler Results
The percent distribution of sulfur among the emission species
measured before and after the ESP is shown in Table 6. Measurements
at both locations show >99% of the sulfur in the form of S02. This
agrees with values predicted for combustion temperatures through
78
-------
Table 6. Measured Distribution of Sulfur
Between SO2, S03, and S04
Sulfur Measured distribution of sulfur, %
Specie Before the ESP After the ESP
S02 99.8 99.5
S03 0-083 0.45
S04 0.090 0.084
equilibrium considerations (7), verifying that concentrations char-
acteristic of high temperatures persist even though S0$ is the
favored species at the stack temperature sampled. This departure
from equilibrium is due to a rapid temperature drop at the furnace
outlet which quenches the equilibrium reactions. However, these
reactions do continue at a greatly reduced rate through gas phase
reactions and by catalytic oxidation near metal oxide surfaces.
This is shown by the >500% increase in the percentage of sulfur
recovered as S03 after the ESP. However, it should be noted that
reductions in SO3 emissions were observed after passage through a
precipitator in other tests (8).
Particulate sulfates, which have been linked with the concen-
trations of volatile sulfate forming species (Ca, Mg, Zn, etc.) in
the coal, experienced a 30% reduction after the ESP compared to a
measured particulate collection efficiency of >98%. This indicates
that sulfates were selectively enriched on the smallest particles
and may imply that they were forming and condensing in the flue gas
stream.
Two aspects of the SOx emission results were unusual. First,
the emission factors determined before the ESP show good precision,
but the values measured after the precipitator vary considerably
for all three sulfur species. When averaged, the post-ESP measure-
ments are about 30% lower than predicted by AP-42. This is due to
an average 30% reduction in the S02 emission factor which, according
to reports in the literature, should remain relatively constant from
the boiler furnace to the stack outlet (7). If the S02 measurements
are valid the differences observed before and after the ESP could
be due to either an undetected variance in the coal sulfur content or
to an interaction of the precipitator with the sulfur species in the
flue gas.
79
-------
Since the SOX emission testing at the inlet and outlet to the
ESP were not done simultaneously, it is difficult to draw conclu-
sions as to the effect of the precipitator. However, this possi-
bility should not be discounted.
A review of the literature on SOX emissions showed variations
in SOX values measured before and after precipitators but did not
reveal any well defined trends. However, a consideration of ESP
operating characteristics provides a possible reason for the conver-
sion of S02 to 80s or S04, because the boiler was equipped with a
"hot side" ESP. That is, combustion gases flow directly from the
furnace to the precipitator and then to heat recovery equipment.
Precipitators are used in this configuration for boilers firing low
sulfur coal, as is the case for many western coal-fired units. Two
potential conversion mechanisms based on the input of energy from
the ESP to the combustion gases via the corona discharge (electrical
arcing across the electrodes) are postulated. First, arcing in a
precipitator may cause localized "hot spots" in which the conversion
of SO2 to S03, and/or S04 would occur quite rapidly, as the temper-
ature is a dominant rate controlling factor. Since the gases are
already hot in comparison to those encountered in an ESP in a con-
ventional configuration, it is plausible that this additional heat
input could cause the observed results. The corona discharges also
have been shown to produce ozone (03) which could readily react with
SC>2 to yield SO3 and 02 • This second mechanism has been presented
previously to explain the apparent conversion of N2 to NO in an ESP
(8). The variability of S02 emissions observed after the ESP could
conceivably be explained by both of these mechanisms, as the degree
of arcing is a function of the ash buildup on the electrodes.
Discussion of Utility Boiler Results
Data on SOX emissions collected at the utility site fall into
three categories. Impinger solutions from the two tests after the
mechanical collector were titrated in the field and again when re-
turned to the laboratory. Time constraints prevented field titra-
tion of the samples collected after the ESP. A continuous gas
analyzer was available on site and provided SOX concentrations at
the boiler outlet for all but one run. Test results in Tables 3,
4, and 5 bring out the following points regarding the use of the
modified Method 8 train for combustion source sampling:
1. Particulate sulfate emission factors are in
relatively close agreement to a previous study
conducted on the same source type (9).
80
-------
2. Sulfates emitted in the flue gas are enriched
on the fine particles.
3. Gaseous sulfur emission measurements were in
most cases within 10% of the emission factor
as calculated from AP-42.
4. The ratio of SO3 to total gaseous sulfur
emissions agrees with values presented in
the literature.
These points are discussed below in more detail.
Particulate sulfates collected were, in most cases, below the
detection limit of the analytical method, and the detection limit
will be used for discussion purposes. The emission factor after
the mechanical collector is <0.06 g/kg as compared to a previous
study where the sulfate emission factor was determined to be about
0.2 g/kg for a coal-fired unit in the same source category (9).
This unit burned a 3.5% sulfur coal and had a mechanical collector
of about 50% efficiency as compared to the 2.4% sulfur coal and 65%
efficiency of the unit in this study. Correcting for these factors
brings the two studies in fairly close agreement. The concentra-
tions of sulfate in particulate were <1.5g/kg in this study versus
5.4 g/kg in the comparative study. When corrected for coal sulfur
content, these differ by a factor of 2.
At the mechanical collector outlet, the sulfate species
accounts for about 0.2% of the total sulfur emitted, while at the
outlet of the ESP it accounts for 0.002%. This corresponds to a
removal of about 99%, or about the same as the expected particulate
removal efficiency of 99.9%. For the industrial boiler it was shown
that sulfate removal in the ESP was 30% versus a particulate removal
of 98%, indicating enrichment of sulfates in finer particles. Be-
cause the utility boiler ESP train was preceded by a mechanical col-
lector, most of the larger particles had been removed before the ESP;
therefore, large differences between sulfate and particulate removal
would not be expected.
In Table 5 the concentration of sulfates with respect to par-
ticulates is presented for each test run as gram of sulfate per kilo-
gram of particulate. It appears that there is a 3-fold enrichment
of sulfates on the very fine particles passing through the ESP train.
However, a more sensitive analytical method should be employed to
accurately quantify sulfate emissions.
Emissions of S03 and sulfuric acid mist were about 0.3 g/kg
of coal feed. This amounted to about 0.75% of the gaseous sulfur
81
-------
emissions which is in agreement with previously reported figures
of 1.0%-2.0% (10)(11). Laboratory titration of the S03 impinger
solutions shows about a 70% increase in 80s concentration over field
titration, indicating that possibly some residual SC>2 remained in
the first impinger and was converted to S03 in the time between
field sampling and laboratory analysis.
Emissions of SOa ranged from 30 to 50 grams per kilogram of
coal or, in terms of coal sulfur content, about 20S g/kg. As seen
in Table 3, field titration of SO2 impinger solutions resulted in
slightly higher SC>2 emission factors than laboratory titration of
the same solutions. Laboratory titration of the 862 impinger solu-
tion of Run No. 1 appears to be in error and cannot be explained at
this time. Field titration of Run No. 2 resulted in about a 10%
higher emission factor than lab titration. This is in close agree-
ment with an additional run which was not reported because the coal
feed rate and coal sulfur content were not obtained by the plant
site. Comments by field and laboratory personnel indicate that the
titration end point for this method is difficult to detect and may
account for some difference in the results. It is the authors'
opinion that, whenever possible, field titrations should be performed
to obtain more accurate results.
Continuous monitoring of SOX produced emission factors approx-
imately 20% higher than the modified Method 8 measurement. Both
methods were in close agreement with the value predicted by the
equation in AP-42 .
Table 4 compares the SOX emission rates, as kg/hr of sulfur,
to the sulfur contained in the coal feed also in kg/hr. After re-
duction of particulate emissions in the MC, the sulfate emissions
accounted for almost 0.1% of the coal sulfur. At the ESP outlet,
sulfates were reduced to account for about 0.001% of the coal sulfur.
About 0.5% of the coal sulfur was emitted as S03 . Total sulfur
emitted, as determined by the modified Method 8 train, ranged from 88!
to 112% of the coal sulfur while gaseous sulfur emissions, as deter-
mined by the continuous analyzer, accounted for 100% to 125% of the
coal sulfur. Other sources have indicated that 90% to 100% of the
coal sulfur is emitted in the stack gases (8).
i
OVERALL CONCLUSIONS AND DISCREPANCIES
The EPA Method 8 procedure for determining SO2 and SOa emissions
is not usually employed for characterization of gaseous emissions
from combustion sources due to the high particulate loading. Modi-
fying the Method 8 procedure by inserting a filter between the probe
82
-------
and first impinger prevented particulate sulfates from entering
the S03 impinger and made them available for separate sulfate
analysis.
Sampling at both the industrial and utility sites with this
method produced gaseous sulfur species emission factors usually
within 10% of those calculated by using AP-42. Total sulfur species
emissions determined by this method in most cases accounted for
88% to 112% of the sulfur in the feed coal. The degree of varia-
bility seen in most of the sulfur species emission factors was
minimal for this program in characterizing controlled and uncon-
trolled emissions.
Emissions of SO3 from the utility site before the ESP were
about an order of magnitude higher than from the industrial site
before the ESP. Besides coal sulfur content, differences in dis-
tance from the furnace play an important role in SO3 concentra-
tions. The industrial site sampling was much closer to the boiler
outlet, and, therefore, it would be expected to have lower SO3 con-
centrations. Emission factors for S03 at the ESP outlets were in
close agreement. These values may be lower than that expected at
the stack outlet. The ratio of SO3 to the total gaseous sulfur
species agreed closely with published values and was about seven
times higher for the utility boiler.
Particulate sulfate emission factors are difficult to compare
because of the differences in particulate collection efficiencies
and the lack of information regarding enrichment of SO* on small
particles. Comparison of S04 concentration on particulate was in
fairly close agreement, but it should be recalled that the sulfates
measured at the industrial site are water soluble while the utility
samples were analyzed for total particulate sulfate. Concentration
of sulfates on particulates increased by a factor of 40 after the
ESP at the industrial site and by about a factor of 3 at the utility
site, showing that enrichment of sulfates on fine particles takes
place. Again, because of differences in particulate collection
efficiencies and the presence of the mechanical collector at the
utility site, these values cannot be compared to each other.
83
-------
REFERENCES
1. Surprenant, N., R. Hall, S. Slater, T. Susa, M. Sussman,
and C. Young. Preliminary Emissions Assessment of Conven-
tional Stationary Combustion Systems, Volume II - Final Re-
Port. EPA-600/2-76-046b, (PB 252 175), U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina,
March 1976. 557 pp.
2. National Energy Outlook 1976. FEA/N-75/713, (PB 252 224),
Federal Energy Administration, Washington, D.C., February
1976. 591 pp.
3. Method 8 - Determination of Sulfuric Acid Mist and Sulfur
Dioxide Emissions from Stationary Sources. Federal Register
41(111) :23087-23090, 1976.
4. Bertolacini, R. J., and J. E. Barney II. Ultraviolet Spectro-
photometric Determination of Sulfate, Chloride, and Fluoride
with Chloranilic Acid. Analytical Chemistry, 30(2):202-205,
1958.
5. Schafer, H. N. S. An Improved Spectrophotometric Method for
the Determination of Sulfate with Barium Chloranilate as Ap-
plied to Coal Ash and Related Materials. Analytical Chemistry,
39(14):1719-1726, 1967.
6. Compilation of Air Pollutant Emission Factors. Publication
AP-42-A, U. S. Environmental Protection Agency, Research Tri-
angle Park, North Carolina, February 1976. 216 pp.
7. Wilson, J. S., and M. W. Redifer. Equilibrium Composition of
Simulated Coal Combustion Products: Relationship of Fireside
Corrosion and Ash Fouling. Journal of Engineering for Power,
Transactions of the American Society of Mechanical Engineers,
96(A-2):145-152, 1974.
8. Cuffe, S. T., R. W. Gerstle, A. A. Orning, and C. H. Schwartz.
Air Pollutant Emissions from Coal-Fired Power Plants; Report
No. 1. Journal of the Air Pollution Control Association, 14
(9):353-362, 1964.
9. Cowherd, C., Jr., M. Marcus, C. M. Guenther, and J. L.
Spigarelli. Hazardous Emission Characterization of Utility
Boilers. EPA-650/2-75-066, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, July 1975. 185 pp.
84
-------
10. Gerstle, R. W., S. T. Cuffe, A. A. Orning, and C. H. Schwartz,
Air Pollutant Emissions from Coal-Fired Power Plants; Report
No. 2. Journal of the Air Pollution Control Association,
15(2):59-64, 1965.
11. Cato, G. A., H. J. Buening, C. C. DeVivo, B. G. Morton, and
J. M. Robinson. Field Testing: Application of Combustion
Modifications to Control Pollutant Emissions from Industrial
Boilers—Phase I. EPA-650/2-74-078a, (PB 238 920), U.S. En-
vironmental Protection Agency, Research Triangle Park, North
Carolina, October 1974. 213 pp.
85
-------
Sulfur Oxide Measurements of Utility Power
Plant Emissions
James E. Howes, Jr.
Battelle-Columbus Laboratories
ABSTRACT
Sulfur oxide measurements were performed at one oil-
fired and two coal-fired power plants as part of the
Electric Power Research Institute (EPRI) SURE plume con-
version study being conducted by Battelle's Northwest
and Columbus Laboratories. The measurements included
determination of SO2, H2S04, and particulate sulfate in
the power plant emission streams.
Sulfuric acid sampling was performed primarily with the
controlled condensation method. Some measurements were
also made with a modified version of EPA Method 6. 80%
was sampled with impinger trains containing hydrogen
peroxide. Sulfate in the collected sulfuric acid and S02
samples was determined by barium perchlorate/thorin
titration. Particulate samples were collected by in-stack
filtration to estimate the sulfate content of the parti-
culate emissions.
The methods that were used for the sulfur oxide measure-
ments will be described, with emphasis on the controlled
condensation method for sulfuric acid. Data will be
presented on the relative distribution of sulfur species
in the power plant emissions, with a comparison of the
sulfuric acid measurements by the condensation and modi-
fied Method 6 techniques.
87
-------
INTRODUCTION
The Electric Power Research Institute (EPRI) has embarked on
a major research program known as the Sulfate Regional Experiment
(SURE). The objectives of the SURE program are: (1) to charac-
terize the present air quality on a regional basis using the north-
eastern sector of the United States as the study area, (2) to
determine the relationships between gaseous and particulate emis-
sions from fossil-fired utility power plants and ambient concen-
trations of air pollutants in a regional context, and (3) to develop
the capability for predicting and confirming the effect on regional
air quality of various electricity production, fuel use, and
environmental control scenarios.
As a component of the SURE program, Battelle's Northwest and
Columbus Laboratories have been conducting for the past year a
study of transformation rates and mechanisms in power plant plumes
within the SURE study region. Along with a comprehensive analysis
of the power plant plumes by aircraft, concurrent measurements of
the source emission characteristics have been conducted. This paper
describes the sulfur oxide measurements which have been performed
and presents data on the SO2, H2SO4, and particulate sulfate emis-
sions from the three plants which have been studied.
EXPERIMENTAL PROCEDURES
Power Plant Sites
Studies have been performed at three power plants during the
first phase of the EPRI program. The plants are Gerald Andrus,
Greenville, Mississippi, an oil-fired unit, Breed Power Plant near
Sullivan, Indiana, and B. C. Cobb, Muskegon, Michigan. Both the
latter are coal-fired facilities. Some of the design and operating
characteristics of the three plants are given in Table 1. During
the measurements described in this paper, the plants operated under
stable full-load conditions, normal operating conditions prevailed,
and the units were fired with regular fuel supplies.
Equipment and Procedures
Sulfuric acid measurements were made by the controlled conden-
sation method, and S02 was determined by collection in impingers
containing hydrogen peroxide. A schematic drawing of the sampling
train is shown in Figure 1. A heated Vycor probe equipped with a
quartz wool plug at the inlet end was used to extract the flue gas
88
-------
Table 1. Plant Design and Operating Characteristics
Andrus
Breed
Cobb
oo
CD
Number of units
Generating capacity
Boiler type
Fuel
Fuel source
Nominal fuel S, %
Flue gas data
C02 , %
02> %
H20, %
Temperature, F
Emission control
Sampling Location
1
770
1
420
Cyclone-fired
Fuel oil #6 (w/additive) Coal
Domestic, Gulf Coast
2.5
11
5.5
10
330
Gas recirculation
Stack
Local, strip-
mined
3.8
10
9.4
9
320
Cyclones
Breeching just
prior to stack
Units 1, 2, & 3-60 each
Units 4 & 5-155 each
Tangentially-fired
Coal
Kentucky, deep-mined
3.5
8-11
7-10
7-9
285-310
Electrostatic preclpitator
ESP outlet
-------
Quartz wool
plug
CD
O
Heated Vycor
Probe 1/2" ID x 6.8' long
"Condensation
coil with
control!ed-temp-
erature water i
supply to
jacket
j
Silica
Gel
Midget
Impinger
Train in
Ice Bath
Pump
Flow
Meter
Dry Gas
Meter
Figure 1. S02/H2S04 sampling system.
-------
sample. The probe was preheated at 250°C (482°F) and maintained
at this temperature during the sampling period.
Sulfuric acid was collected in the condensation coil immediately
following the probe. A detailed drawing of the coil is presented in
Figure 2. Water from a constant temperature bath was circulated
through the jacket surrounding the coil to maintain the outlet gas
temperature at 60°C (140°F) during sampling.
SO2 in the sample stream was collected in two serially-con-
nected midget impingers, each containing 15 ml of 3% hydrogen per-
oxide. The remainder of the train consisted of an empty impinger,
a moisture trap (silica gel), a pump, and a dry test meter to mea-
sure the sample volume.
A sampling rate of 1.5 liters/minute (~0.035 cfm) was used
for collection of SC>2 and H2SO4. The sampling period for each test
was 60 minutes. Following sampling, the sulfuric acid retained by
the filter plug and the probe walls was collected by extracting the
plug and rinsing the probe with isopropyl alcohol (IPA). The sul-
furic acid in the condensation coil was recovered by rinsing with
distilled water.
Sulfate analysis of the filter extract/probe washes, coil
rinses, and impinger solutions was performed by barium perchlorate/
thorin titration using the EPA Method 6 procedure.
The flyash samples for particulate sulfate analysis were col-
lected with an in-stack filter assembly. Pallflex quartz paper
which had been acid-washed with 0.1N HC1 was used as the filter
medium. After collection, the filters were stored under a nitro-
gen atmosphere until analysis. The determination of sulfate in
water and 0.1N HC1 extracts of the filter samples was performed by
BCL using ion chromatography and by Brigham Young University using
calorimetry and ion chromatography.
RESULTS
The results of the H2S04 and SO2 measurements at the three
power plants are presented in Table 2. The sulfuric acid data
are based on the sum of the sulfate in the IPA filter extract/
probe wash and the coil rinse. A significant quantity of sulfuric
acid, typically 20%-30% of the total, was found in the IPA filter/
probe washes.
91
-------
<0
-30 mm diameter
coarse glass frit
Water outlet from
constant temperature
water supply
-Water inlet to constant
temperature water supply
10"
Figure 2. Condensation coil used for sulfuric acid collec-
tion.
-------
Table 2. Summary of SO2 and
Measurements
Concentration in Emissions,
ppmv(mg/m3)
Site
Andrus
Breed
Cobb
Unit 1
Unit 2
Unit 3
Unit 5
Date
10/21/77
10/22/77
11/4/77
11/6/77
11/17/77
11/18/77
11/18/77
11/18/77
H2SO4
37.2(149)
45.9(183)
28.9(115)
25.2(101)
6.1(24)
4.7(19)
9.5(38)
15.3(61)
SO2
1287(3430)
1289(3435)
3324(8858)
3218(8576)
2377(6335)
2279(6074)
2285(6090)
2380(6343)
Ratio,
ppm H2S04
ppm SO2
0.029
0.036
0.008
0.008
0.003
0.002
0.004
0.006
93
-------
Measurement of H2SO4 emissions at the Andrus and Breed plants
was also performed using a modified version of EPA Method 6. In
these experiments, considerable quantities of the total sulfate
(40%-60%) were found in the water extract of the filter plug and the
probe rinse. (Probe temperature during the Method 6 sampling was
maintained at 204°C.) Based on total sulfate collected (filter,
probe, and IPA impinger), the H2S04 concentrations in the Andrus
and Breed emissions were estimated at 53 ppm and 33 ppm, respec-
tively. These values are slightly higher than those obtained by
the condensation method; however, this might be due to sulfates
which were water-leached from particulates retained by the filter.
The results of the analysis of the flyash leach solutions for
sulfate are given in Table 3. The range of values shown for the
weight percent of soluble sulfate in the flyash was obtained from
analysis of two Andrus samples and four each from the Breed and
Cobb plants. A slightly higher concentration of sulfate was
observed in the oil flyash acid extract than in the water extract;
however, this may result from run-to-run variation. About the
same sulfate concentrations were found in the water and acid
extracts of the coal flyash samples. Estimates of the particulate
suljate concentrations in the emission are expressed as mg soluble
S04~/m3.
Table 3. Particulate Sulfate Measurements
Site Soluble. Particulate Estimated SO4 Emissions
S04 , weight % mg/m3 (avg)
Andrus 33 (water) 18
46 (HC1)
Breed 0.5-1.9 54
Cobb 6-8 7
94
-------
Table 4 presents a summary of the sulfur emission measurements
from the three power plants. Estimates of the sulfur distribution
in the emissions are given for each plant based on the measurements
of the three emission components.
DISCUSSION
Based on the data which have been presented, the following
observations can be made relative to sulfur emissions from the oil-
and coal-fired power plants.
• The proportion of sulfuric acid in the oil-fired plant
emissions was higher than in the coal-fired plant emissions.
Sulfuric acid comprised about 3% of the total sulfur
emissions from the oil-fired unit versus less than 1%
of the emissions from the coal-fired units.
• Sulfuric acid levels in coal-fired emissions varied
significantly. About 27 ppm was found in the Breed
Plant emissions, while relatively low concentrations
of between 5 ppm and 15 ppm were measured in the emis-
sions from the Cobb Power Plant. Both units use coal
with about the same sulfur content.
• Sulfuric acid was the major component of primary
sulfate emissions (H2SO4 + particulate sulfate) from
both the oil- and coal-fired power plants. In both
types of plant emissions, sulfuric acid accounted for
about 90% of the primary sulfate emissions.
• Flyash from the oil-fired plant contained a higher
proportion of sulfate than the coal-fired flyash.
However, particulate sulfate emission rates from
oil-fired plants may not be significantly higher
than from coal-fired units.
Finally, the preceding observations should be brought into
perspective by pointing out that, by far, S02 accounts for the major
fraction of the sulfur oxide emissions from power plants. There-
fore, transformation of SO2 in the power plant plume and/or the
ambient atmosphere represents a much greater potential for intro-
duction of sulfates into the environment.
95
-------
Table 4. Summary of Sulfur Emission Measurements
Si'-e
-------
ACKNOWLEDGMENTS
The work presented in this paper was funded by EPRI under
Project No. RP 860-1. The EPRI Project Manager responsible for the
overall direction of the program is Mr. Charles Hakkarinen.
The Consumer's Power Company, Indiana and Michigan Electric
Company, and Mississippi Power and Light Company are gratefully
acknowledged for the cooperation of their personnel during the
planning and conduct of the EPRI study.
Dr. Delbert Eatough, Brigham Young University, is acknowledged
for chemical characterization of the flyash Samples.
97
-------
Effects of Combustion Modification on SO3
Formation in Combustion
Arthur Levy
John F. Kircher
Earl L Merryman
Battelle-Columbus Laboratories
ABSTRACT
Primary acid aerosol emissions from the combustion of
coal and oil will be reviewed relative to thermodynamic
and kinetic considerations. In the case of the former,
species distribution will be examined as it relates to
typical ash and fuel compositions. Kinetic aspects
will be examined relative to catalytic and homogeneous
reactions of fly ash constituents under varying fuel/
air ratios.
Results of an experimental study on the effect of
staged combustion on SO2 oxidation will be presented.
Study results will also corroborate a previous obser-
vation regarding enhanced SO3 formation, for example,
that under the same overall fuel/air condition, more
SOs is produced in staged combustion than in a single-
step system. The experiment results suggest that the
enhanced SO3 production may be of more concern as a
corrosion-deposit promoter than as a pollutant. The
oxidation reactions are dependent on post-flame
temperatures, on the extent to which the secondary
air mixes with combustion products, and on the oxida-
tion kinetics of CO in the second stage combustion.
99
-------
INTRODUCTION
In recent years there has been increasing evidence that sul-
fates in the atmosphere may be of more concern as a health and
environmental hazard than sulfur dioxide. Part of this concern is
reflected in the fact that S02 levels in the atmosphere have been
on the decline, while sulfate levels remain unchanged (1)(2).
Historically, it has generally been stated that only about l%-3%
of the sulfur in a fuel is emitted from the combustion system as
SO3 or acid. (In this paper SO3 and sulfuric acid are considered
synonymous, since the presence of water vapor and the reaction
of S03 with water vapor are so prevalent.) However, since such
acid can lead to various sulfates which might be a part of the
particulate emissions, it is important to consider these parti-
culates as well as S03 as part of the primary acid aerosol.
Further, as various combustion modifications (CM) become more
widely applied to control NOX emissions, one must be concerned
that these previously held postulations regarding SO3 and sulfate
emissions are valid.
In this paper we present a brief examination of how various
CM procedures might influence acid aerosol formation (3) and a
more detailed examination of the effect of staged combustion on
S03 formation (4).
COMBUSTION MODIFICATION
Acid aerosols, in the full sense of the term, include liquid
and solid particles containing sulfates, nitrates, and chlorides.
In essence, however, this paper is concerned with the sulfates,
i.e., SO3, H2S04, and sulfate salts. The principal reason for
this is that the great majority of emissions which may lead to
acid aerosols in the atmosphere are sulfur compounds, i.e., sul-
furic acid, S03, and sulfates, although it is recognized that
not all sulfates are acidic. Nitrates have not been observed,
nor are they expected thermodynamically in stack particles, but
a small amount of nitrate may be formed in the near plume. The
sparce information available on HC1 or chlorides is in general
agreement with basic thermodynamic considerations that the
chlorine in fuel will be emitted primarily as gaseous HC1 from
the stack (5). Evidence indicates that total primary sulfates
(i.e., those observed within the first half-mile) can be as high
as 20% of total sulfur emissions or as low as 2%.
100
-------
THERMODYNAMIC CONSIDERATIONS
If we consider the thermodynamic potential for producing acid
aerosols, we note that sulfates are by far the most likely acid
aerosols to be produced. Figures 1 and 2 illustrate this point.
In line with a major interest of this workshop, the characteriza-
tion of sulfurlc acid and sulfate particulate, it is of interest
to note in these two figures the prevalance of sulfuric acid
emissions from the combustion of oil relative to that from coal.
Figures 1 and 2 show this quite clearly. In effect the ash con-
tent of oils is too low to compete for the sulfuric acid; con-
versely the ash content of coals readily converts the acid to sul-
fate salts.
SPECIFIC EFFECTS OF COMBUSTION MODIFICATION ON ACID AEROSOL
Five combustion modification procedures were considered in
this study. These CM procedures are staged combustion, flue gas
recirculation, low excess air, low air preheat, and load reduc-
tion. Little pilot or field test data exist which directly
demonstrate that a particular combustion modification employed
to reduce NO and N02 will have an effect, good or bad, on primary
acid aerosol. The weight of the evidence is that anything which
tends to reduce super-equilibrium oxygen atom concentrations in
the flame zone will tend to reduce S03. On the other hand, if
the production of particulate, especially very small particles,
is increased, then the production of acid and sulfate solids
might be expected to increase through heterogeneous processes.
At this time, conclusions regarding the effect of a particular
combustion modification on specific equipment must be highly
speculative.
Staged Combustion
Archer, et al. (6), report on an investigation of a pilot
scale two-scale combustion of a high vanadium residual oil with
2.4% sulfur. Their results demonstrated that SO3 could be re-
duced essentially to zero when the first stage was slightly fuel
rich. They explained their results by noting from previous
work that carbonaceous particles inhibit S03 formation, react
with S03, and physically adsorb it.
Such changes do not mean S03 is completely eliminated from
the boiler, however. When air is added at the second stage to
complete combustion, SO3 can still be formed. As reported by
101
-------
439 710 980
&
3
CO
o
•ft
+-•
s
0)
I—«
o
10
2780
500
700
900
1100 1300
°K
1500 1700
1900
Figure 1. Equilibrium sulfur products for #6 oil combustion with
2% excess air, 2.80% sulfur.
102
-------
439 710 980
2780
10
H SO (g)VMgS04(s)
10
-5
500 700 900 1100 1300 1500 1700 1900
°K
Figure 2. Equilibrium sulfur products for coal combustion with 10%
excess air, 3.27% sulfur.
103
-------
Hedley (7), S03 may be produced even in excess of that produced
in a single-step combustion process. Medley's observations are
examined in detail in the second half of this paper. .
Flue Gas Recirculation
The effect of flue gas recirculation (FOR) on S03 formation
is not especially clear. Koizumi, et al. (8), studying the com-
bustion of a 2.5% sulfur heavy fuel oil, attempted to relate FOR
to S03 production and to flame length. Their observations on S03
formation are best illustrated by Figure 3. The effect of FGR
on acid dewpoint, i.e., 80s is inconclusive. One might venture
to say that there was a slight decrease in SO3 as percent FGR was
increased.
Low Excess Air
Thermodynamically it is recognized that the percent oxida-
o
level.
tion of S02 to SO3 is decreased as one reduces the excess air
Experience with oil-fired systems, where low excess air
operation is most practical at the present time, has demonstrated
that this mode of operation minimizes the formation of sulfates
in deposits in the high temperature portion of the boiler, reduces
the amount of sulfuric acid formed, and eliminates the emission
of acid smuts. Successful operation with low excess air requires
that the oxygen in the flue gas be maintained at levels below
0.2%. Such operation requires precise control of the fuel-air
ratio in all parts of the combustion system to prevent thermal
cracking of hydrocarbons and the emission of smoke. Consequently,
low excess air operation has been limited to oil-fired systems
because the technology for burning pulverized coal with minimal
oxygen does not exist.
Glaubitz (9) in Germany was probably one of the first people
to take advantage of this effect to control deposits and corro-
sion in oil-fired systems. By redesigning the oil burners and
exercising very close control on the fuel-air ratio, Glaubitz
was able to lower excess oxygen to 0.2% for routine operations.
Under these conditions, the sulfuric acid was reduced to such
104
-------
120
50
Excess air factor 1.1
Excess air factor 1.03
20 40
Recirculation ration %
60
Figure 3. Effect of exhaust gas recirculation ratio on acid
dewpoint (8).
105
-------
an extent that the dewpoint approached that of water. Glaubitz
stated that after 12,000 hours of operation, the boiler still did
not have to be shut down for cleaning, indicating that the
strongly bonded deposits, which build up as a result of the forma-
tion of large amounts of sulfates, had not developed in this boiler.
Low Air Preheat
Although there is considerable information regarding the
effect of lower air preheat on the SO3/SO2 ratio, Glebov (10)
points out, "data on the influence of flame temperature on process
of formation of S03 is very inconsistent." It has been firmly
established that in pulverized fuel-fired boilers, the content of
S03 in the gases decreases with increasing temperature in the fur-
nace. However, Crumley, et al. (11), on the basis of experi-
mental data they obtained..."using kerosene and distillate show
an increase in SO3 to a flame temperature of 1750°C (3182°F)
followed by a leveling off. The difference in the results from
the two fuels is considerably less than the difference in 2 percent
sulfur in the kerosene, and 3 percent in the distillate. At 70
percent excess air with kerosene, about 7 percent of the sulfur
was in the form of S03; at 28 percent excess air, about 5 percent."
Load Reduction
Based on very meager data, it appears that load reduction
has no significant effect on SO3 emissions. Glebov (10) found no
effect of load on S03 over a range of 20% to 80% design load in
his study of high sulfur, heavy oil in an experimental furnace.
In his theoretical computations he also found no change in going
from 100% to 70% load, assuming a catalytic activity of deposits
equivalent to that produced by Fe203, but some increase in SO3
with decreasing load, assuming catalysis by V205.
Table 1 summarizes the effects of CM procedures on SOs for-
mation. Examination of the literature suggests that to date
combustion modification procedures have not caused any pronounced
effects on acid aerosol formation in general or on S03 (sulfate)
formation in particular.
The work of Hedley (7), however, did imply that staged com-
bustion might lead to a potential deleterious effect, an increase
in S03, or, as we refer to it in the remainder of this paper,
enhanced S03. The second part of this paper discusses a labora-
tory investigation of staged combustion and its effect on SO3
formation.
106
-------
Table 1. Effects of CM on SO, Formation
CM
Effect
Staged combustion
Flue gas recirculation
Low excess air
Low air preheat
Load reduction
Enhanced (increased) S03 possible.
(Effect may be positive, negative,
or nil depending first stage stoi-
chiometry and temperature conditions
of second stage.)
Little direct effect
Decrease in S03
Inconsistent - decrease and increase
reported
No effect
STAGED COMBUSTION AND S03 FORMATION
The primary basis for examining staged combustion in detail
comes about not only from the Hedley study but also from three
other studies (12)(13)(14) which indicate that the high CO content
in the second-stage firing might readily "pump" oxygen atoms into
the second stage and thus promote the homogeneous oxidation of
S02 in the second stage.
Kinetically the formation of SO3 is an O-atom process best
described by the mechanism (12)(13)
S02+0+M
SO2+0
SO3+H
SO3+M
SO,
S02+OH
[1]
[2]
[3]
This mechanism readily accounts for the "excess" or "higher than
equilibrium" levels of SO3 observed by the authors (13), Hedley
(7), and others.
107
-------
Although the oxidation of CO occurs primarily via the OH
radical, CO+OH=CO2+H, as Gaydon points out, CO oxidation is
accompanied by high levels of oxygen atoms. The Semenov (15)
mechanism for the excess 0-atoms is
CO + 02 = C02* + 0 [4]
C02* + 02 = C02 + 20. [5]
Thus, if one couples the above-stated processes for S02 and for CO
oxidation, it becomes apparent that conditions may exist where
one might expect an increase in S03 formation in the second stage
of a two-stage combustion process.
Enhanced S03 Emissions
At the time of Medley's studies, no special attention was
directed to the use of staged combustion for NOX control. How-
ever, in explaining his evidence on the formation of S03 at
levels considerably above equilibrium, he describes a one-
dimensional controlled mixing experiment and, although this is
not presented as such in his paper, he states: "If combustion
took place under stoichiometric or fuel-rich conditions then no
trioxide formation took place. When less than stoichiometric air
was used, the unburnts in the gases consisted solely of carbon
monoxide with S02 but no SO3 . When the remaining excess air -was injected
into these gases, the maximum amount of S03 formed was greater than that formed
when this additional air was included with the initial combustion air, the overall
excess of air being the same in both cases. "*
This observation and its potential impact on the effects of
staging on SO3 formation then become closely tied to the effect
of CO oxidation kinetics in the second stage production of SO3.
The CO effect is best borne out in Figures 4 and 5 where one
notes the highest conversion of SO2 to S03 in sulfur-bearing CO
flames (12)(13). When one couples the observations in Figures 4 and
5 with Gaydon's observations of high concentrations of oxygen atoms
in CO flames and with the 0-atom mechanism for S03 formation in
combustion, Hedley's statements on the enhancement of SO3 in
staged firing appear quite consistent. Basically then, S03 for-
mation is an oxygen atom process, and the question to be
addressed is "what is the effect of staging on the oxygen atom
concentration?" and its corallary, "what is the effect of staging
on SO3 formation?"
*Italics signify authors' emphasis.
108
-------
eo
O
CQ
-------
CO
0)
CO
450
400
350
3 300
DO
•4-J
0)
I 250
o
6
O
en
200
SB 150
100
50
Flameholder
COS flame
(PT = 250 torr)
(P = 625 torr)
0 200 400 600 800 1000 1200
Distance Above Flameholder, mils
Figure 5. S03 profiles in H2S, COS, and CH3SH flames (13).
110
-------
Experimental
A quartz tube burner which allowed one to establish stable
methane-H?S flames within desired fuel-air was used. Two inlets
were provided above the flame (first stage) for adding air to
complete the combustion process (second stage). S03 was then
sampled at various positions downstream of the secondary air.
Burner System
The primary burner tube was constructed of quartz (13 mm
I.D.) and produced a laminar flow bunsen-type flame (1/D >60).
The quartz reaction chamber surrounding the burner tube was 19 mm
I.D. This chamber contained several temperature and sampling
ports spaced from 3-1/2 to 5-1/2 cm apart in the early postflame
zone and increased to about 12-1/2 cm apart in the far postflame
zone. These spacings provided appropriate time intervals for
collecting the S03. The reaction chamber was externally heated
(Chromel "A" wiring) to control second stage temperatures. Total
chamber length was approximately 76 cm and provided a maximum
gas residence time of about 250 msec.
Gas samples were removed at various locations above the flame
via a quartz sampling probe. S03 was analyzed colorimetrically
by the barium chloranilate procedure. CO, C02, 02, and SO2 were
also measured, mainly for purposes of confirming and comparing
postflame combustion conditions and sulfur oxide levels with cal-
culated cold gas compositions. Details of the apparatus and
analytical techniques are described in Reference 4.
Flame Conditions
Three flame compositions were probed in detail for S03 pro-
files in this study. The compositions were:
111
-------
Mole Fraction (Cold Cases)
Gas Single State 0S = 1.1
CH4 0.087
02 0.191
N2 0.720
H2S 0.0015
Two-Stage*
1st Stage 01 = 0.95* 1st Stage 0i = 0.90*
0.099 0.104
0.189 0.188
0.710 0.706
0.0017 0.0017
* In the second stage, 02= 1.1; mole fractions were the same as
in the single stage firing.
Equivalence ratios, defined as
= [Air/Fuel]/ [Air/Fuel] stoich,
are expressed as 0S ,
and
, i.e., 0S , single stage;
first stage; and 02 , second stage. In all experiments the second
stage firing introduced sufficient air that 2 was comparable to
0s-
Results and Discussions
The staged combustion experiments are summarized in Figures
6, 7, and 8. In essence the data show three distinct effects of
staged combustion on SO3 formation.
Figure 6 shows definite enhancement where -\ = 0.95. Curves
B and C show an absolute increase of some 9% in 863 formation
compared to Curve A, the single stage process.
Curves B and C in Figure 6 also show the effect of adding
the secondary air at two different positions in the postflame
gases. Comparison of these two curves shows that altering the
distance between the burner head and the introduction of secondary
air produced some changes in the shape of SOs curves, particularly
in the 40-150 msec range. Although maximum S03 levels are nearly
the same in either of the two-stage modes of firing, the deple-
tion of S03 appears to occur more rapidly when the addition of
secondary air is delayed several msec (Curve C) .
112
-------
o
o
co
O
CO
co
O
co
O
CO
_—-
CO
O
CO
O
CO
C
O
•O
•fl
O
2210 20501968 18181710 1635 1578
1515
I I I I I I
1865 1850 1835 1795 1735 1660
T
1505
I I \ I
17651585 1545 1520 1495 1445 1380
T°K (Curves A and E)
T°K (Curve B and D)
1260
T K (Curve C)
3.0-
20
40
80 100 120
Milliseconds
140
160
180
200
Figure 6. Oxidation of SCL in single- and staged-combustion, 01 =0.95.
-------
2.5 _
o
o
CO
o
CO
O
CO
O
2
c<
O
CQ
•8
I
•o
•fH
O
0\°
40 60
Milliseconds
100
120
Figure 7. Oxidation of SO^ in single- and staged-combustion, <£, = 0.99.
114
-------
60
Milliseconds
80
100
120
Figure 8. Oxidation of S02 in single- and staged-combustion,
= 0.90.
115
-------
Figure 7, where 0, = 0.99, barely sub-stoichiometric, the com-
parative differences of two-stage versus single stage combustion
are negligible.
Figure 8, where 0., = 0.90, shows that S03 formation can also
be decreased by staged combustion.
Enhancement—The data presented here confirm Medley's state-
ment that staged combustion enhances S03 formation. However, our
results indicate: (1) the effect may not be a pronounced
effect, (2) the enhancement is of short duration, the SO3 appearing
to approach steady state conditions about as rapidly as in single-
stage combustion, and (3) the enhancement effects and its duration
are dependent on the air/fuel ratio of each stage and the delay
interval in the addition of secondary air.
Time Delay Effects—It is obvious that delaying the addition
of secondary air to the point where the temperature is below that
required to produce favorable conditions (mainly 0-atoms) for S02
oxidation prevents further S03 formation. It follows that the
formation of SO3 would, therefore, decrease with decreasing tempera-
ture at the secondary air ports. Barrett, et al., have commented
to this effect in their examination of the formation of S03 in a
small combustor using single and two-stage firing modes (16).
They concluded from their studies that the addition of secondary
air at temperatures below about 950°C would likely produce little
or no additional SO3. In the present study we find that adding
the secondary air at about 850°C produced less S03 than when the
air was added at higher temperature, and no enhancement of S03
was observed at the lower temperature. Thus, the "nonreactive
temperature limit" may be somewhat lower than that observed by
Barrett, et al.
Air/Fuel Ratio Effects—In considering the ultimate effects
of different air/fuel ratios on S03 production in two-stage com-
bustion, one might expect an increase in SO3 with decreasing
air/fuel ratio in the first-stage firing. This is reasoned on
the basis of an increased CO level in the first stage, followed
by a greater enhancement of S03 from the CO oxidation chemistry.
However, temperature effects also influence the chemistry here.
As the air/fuel ratio is dropped well below stoichiometric, the
temperature of the first stage is decreased. As a consequence,
116
-------
a larger amount of secondary air is needed to restore the second
stage to the desired overall equivalence ratio. Further cooling
of the gases takes place with a resulting overall reduction in
the rate of CO oxidation, a lower 0-atora concentration, and hence
less SO3 formation, as observed in comparing the data of Figures
6 and 8.
On the other hand, approaching stoichiometric conditions in
the first stage increases the flame temperature to near a maxi-
mum, leaving less CO to be oxidized in the second stage. This
could, within limits, lead to less SO3 formation relative to a
richer first stage firing. Data from the present flame probings
do not show any enhancement in S03 formation in a two-stage
process at ^ = 0.99 (Figure 7).
S03 Fluctuation—The data in Figures 6, 7, and 8 show an
interesting, as yet unexplainable but repeatable, discontinuity
as SO3 approaches its maximum. The authors have observed similar
fluctuations in their microprobing of H2S flames, which they
attributed to the oxidation of SO (17). Interestingly, Hunter's
model for S03 formation shows a similar fluctuation (18).
Kinetics Analysis
The profiles of Figures 6, 7, and 8 provide a means for
analysis of the rate constants for K1, K2, and ratio k,/k2. The
kinetics analysis is summarized in Table 2. (Additional details
on these analysis are presented in Reference 4.)
Table 2. Kinetic Analysis
(T=1685K)
S02 + 0 + M SO3 + M
k, = 7.4 x 10 cm mole sec This study
k, = 1 x 1015 cm6 mole"2 sec"1 Reference 19
/k2 = 6.6 x 103 cm3 mole1 This study
/k2 = 104 cm3 mole~1 (est.) Reference 19
117
-------
SUMMARY
It does not appear that combustion modification procedures
will severely affect acid aerosols emissions or, more specifically,
S02 emissions. On the other hand under some conditions, dependent
on the fuel/air ratios of the first stage and the temperature of
the second stage process, staged combustion can lead to enhanced
S03, decreased SO3, or essentially the same S03 levels as observed
in normal, single stage combustion. This study points out that
one must consider the specific conditions existing in the combus-
tion chamber before comparing sulfate emissions from one boiler
to another.
This paper is based on two studies carried out at Battelle-
Columbus Laboratories under Environmental Protection Agency
support, namely Contract No. 68-02-1323 Task 49 and Grant No. E
805330010. The contents of these studies do not necessarily
reflect the views and policies of the Environmental Protection
Agency.
118
-------
REFERENCES
1. Altshuller, A. P. Regional Transport and Transformation of
Sulfur Dioxide to Sulfates in the U.S. J. Air Pollution
Control Assoc., 26:318-324, 1976.
2. Squires, A. M. Control of Emissions of Sulfuric Acid Vapor
and Mist in Air Quality and Stationary Source Emission
Control. National Academy of Sciences. Government Printing
Office, Serial No. 94-4, 1975. pp. 458-473.
3. Kircher, J. F., et al. A Survey of Sulfate, Nitrate, and
Acid Aerosol Emissions and Their Control. EPA-600/7-77-041
April 1977.
4. Merryman, E. L. , and A. Levy. Enhanced SO3 Emissions from
Staged Combustion. Seventeenth Symposium (International)
on Combustion, Leeds, August 1978.
5. lapalucci, T. L. , R. J. Demski, and Di Bienstock. Chlorine
in Coal Combustion. Bu. Mines, RI-7260, 1969.
6. Archer, J. S., P. D. Grout, and F. Eisenklam. Multistage
Combustion of Residual Fuel Oil. J. Inst. Fuel 43:397-404,
451-460, 1970.
7. Hedley, A. B. Factors Affecting the Formation of Sulphur
Trioxide in Flame Gases. J. Inst. Fuel, 40:142-151, 1966.
8. Koizumi, M. , H. Mizutani, Y. Takamura, and K. Nagata. High
Space Heat Release and Low Excess Air Combustion of Heavy
Fuel Oil Using Exhaust Gas Recirculation Methods. Bull.
J. Soc. Mech. Eng., 12:530-538, 1969.
9. Glaubitz, F. The Economic Combustion of Sulfur-Containing
Heating Oil, A Means of Avoiding Dewpoint Difficulties in
Boiler Operations. Combustion, 34:31-35, January 1963.
10. Glebov, V. P. Investigation of the Formation of S03 with
Combustion of Liquid High-Sulfur Content Fuel in a Cyclone
Chamber. Thermal Eng., 20:51-54, 1973.
119
-------
11. Crumley, P. H., and A. W. Fletcher. The Formation of Sulphur
Trioxide in Flue Gas. J. Inst. Fuel, 29:322-327, 1956.
12. Dooley, A., and G. Whittingham. The Oxidation of Sulfur
Dioxide in Gas Flames. Trans. Far. Soc., 42:354, 1946.
13. Merryman, E. L., and A. Levy. Sulfur Trioxide Flame Chemistry.
In: Proceedings of the Second International Air Pollution
Conference, Paper CP7A, 1970. 361 pp.
14. Gaydon, A. G. Continuous Spectra in Flames: The Role of
Atomic Oxygen in Combustion. Proc. Roy. Soc., A183:lll,
1944.
15. Semenov, N. Chemical Kinetics and Chain Reactions. Oxford
University Press, 1935.
16. Barrett, R. E., J. D. Hummell, and W. T. Reid. Formation
of 80s in a Noncatalytic Combustor. J. of Engineering for
Power, Trans. ASME, Series A, 88:165-172, 1966.
17. Levy, A., and E. L. Merryman. The Microstructure of Hydrogen
Sulfide Flames. Comb, and Flame, 9:229, 1965.
18. Hunter, S. C. , and P. K. Engel. Sulfur Oxides Emissions from
Boilers, Turbines, and Industrial Combustion Equipment.
Workshop on Measurement Technology and Characterization of
Primary Sulfur Oxides Emission from Combustion Sources,
Southern Pines, North Carolina, April 1978.
19. Cullis, C. F., and M. F. R. Mulcahy. The Kinetics of Combus-
tion of Gaseous Sulfur Compounds. Comb, and Flame, 18:225,
1972.
120
-------
Impact of Sulfuric Acid Emissions on Plume
Opacity
John S. Nader
William D. Conner
U.S. Environmental Protection Agency
ABSTRACT
Concurrent measurements were conducted on plume opacity,
stack gas opacity, sulfate, and gross particulate loading
in the emissions from coal- and oil-fired power plants.
Plume opacity was measured by EPA Method 9 and by Lidar,
stack gas opacity by an in-stack transmissometer, sul-
fates by modified EPA Method 6, and gross particulate
loading by EPA Method 5.
Results indicated a significant difference in plume opa-
city as compared with in-stack opacity in an oil-fired
combustion source. The plume opacity was always higher
and the difference increased with higher acid emissions
and with distance downstream from the stack exit. Plume
opacity data on coal-fired sources with particulate but
no gas controls were limited to Method 9, and the few
measurements available showed considerable scatter com-
pared with in-stack opacity data. No conclusions could
be drawn.
Preliminary measurements have been made on a coal-fired
power plant with both particulate and gas controls (FGD).
The data were not adequate to draw any conclusions. Fur-
ther studies are under way on high-sulfur coal-fired
plants with FGD systems and emissions with pronounced
and persistent visible plumes.
121
-------
INTRODUCTION
Emission standards for opacity and for mass concentration of
particulate matter have been established for new sources and are
applicable to fossil-fired combustion sources (1). Continuous
monitors for opacity (transmissometers) are required to be installed
on these sources to verify the maintenance and satisfactory opera-
tion of control systems used to meet the emission standards (2).
Reference Method 9 is the observer method of measuring the opacity
of the plume which forms as the particulate matter exits the stack
(1). The Lidar (JLight Detection and ranging) technique is an elec-
tro-optical instrumental technique to remotely measure the opacity
of the plume (3)(4). The transmissometers are installed in the
stack or in ducts leading to the stack and measure the opacity of
the gas stream in the stack or duct prior to its exiting the stack.
The Stationary Source Emissions Research Branch (SSERB) of the
Environmental Sciences Research Laboratory has had in its program
during the past three years tasks to generate a data base of con-
current measurements of in-stack gas opacity (Os) and plume opa-
city (Op) for emissions from various industries. The purpose of
these measurements was to identify those industries wherein the
plume and in-stack opacities do not agree. Measurements conducted
to date on combustion sources burning coal with sulfur _< 2% show
that Os and Op are comparable (5)(6). These results would imply
that no significant (observable effect on opacity) physical or
chemical transformation was occurring in the contents of the gas
stream as it was transported through the stack.
Similar opacity measurements were made on a power plant burning
oil with 2.4% S (sulfur) and 200 ppm to 600 ppm V (vanadium). In
contrast with opacity measurements made at sources burning coal or
oil of lower sulfur content, the plume opacity was found to be
significantly higher than the in-stack opacity. At this oil-fired
power plant a concurrent study was being conducted on sulfuric acid
emissions. The results of this acid study support the conclusion
that a physical transformation occurs as the gas stream exits the
stack and enters the atmosphere. The following phenomenon is indi-
cated: The sulfuric acid is above its dewpoint at stack tempera-
tures in excess of 150°C and does not affect the in-stack opacity.
When the gaseous sulfuric acid leaves the stack and is cooled to
ambient air temperatures which are below its dewpoint, it condenses
and the sulfuric acid droplets increase the plume opacity. Addi-
tional studies have been conducted and are ongoing to obtain more
data and understanding of the effect of sulfuric acid emissions on
plume opacity for various operating conditions, fuel composition,
and control systems for a number of fossil-fuel-fired utilities.
This paper presents and discusses the results of the above work
that SSERB has conducted thus far.
122
-------
The sulfur oxides potentially present in the stack gas stream
at temperatures above the sulfuric acid dewpoint are as shown in
Figure 1. The free H2SO4 and S02 are not sensed in the measure-
ments of the opacity of the stack gas stream. In the plume with
the temperature of the gas stream dropping below the acid dewpoint
and approaching ambient air temperature, the free H2S04 condenses
to form acid droplets. The condensed acid droplets and the acid
adsorbed on the fly ash add to the opacity of the plume. In our
studies, Os was measured by transmissometers, and Op was measured
by human observers or by a Lidar system; in some instances, 0
measurements were made by both of these methods.
COMBUSTION SOURCE FEATURES
Plume opacity measurements were conducted in conjunction with
emissions characterization studies at two oil-fired and three coal-
fired power plants. Table 1 summarizes the physical and operating
features of the plants.
There is a marked difference in composition of the fuel utilized
in the oil-fired sources in contrast to the coal-fired sources.
The ash content of oil was two orders of magnitude less than the
ash content of coal and the sulfur content of the coal was from
two to four times the sulfur content of the oil. In addition, very
high vanadium concentration (590 ppm) was found in the Venezuelan
oil. Excess boiler oxygen was typically in the 3% to 5% range
except for Plant A which operated at very low oxygen levels at
about 0.2%. Oil-fired sources had no emission controls; however,
fuel additives were used to minimize corrosion problems and did
provide some reduction in sulfate emissions (7). Coal-fired sources
had either particulate emission controls (electrostatic precipita-
tors, ESP) or both particulate and gaseous emissions controls (2
stage wet scrubbers).
SAMPLING LOCATIONS
Sampling locations for all in-stack measurements were at a
common location between the emission controls and the stack except
for Plants M and LC. At Plant M all in-stack measurements were at
a common location in the stack proper. At Plant LC in-stack opa-
cities were monitored in breechings leading into the stack, and all
other in-stack measurements were at the output of one of eight
scrubber modules that operated in parallel for the total boiler
output of 820 MW. Each module in effect handled about 100 MW of
power output. Except as noted, plume opacities were measured within
one stack diameter of the stack exit.
123
-------
Free H2SO4 (Gas Phase) " •_ .'*
Particles with Adsorbed H S04 • '''••'>-1 :•'.'•'•• ,~
• • • ' '''. Sulf ate Particles " *
Free S02 •
.* • " • Particles with Adsorbed SO
Stack Ducting
Figure 1. Sulfur oxides present in the stack gas stream.
124
-------
01
Table 1. Summary of Physical and Normal Operating Features
of Power Plants Studied
Fuel
Plant Burned
A Oil
M - Oil
P Coal
MC Coal
LC Coal
Height
(m)
150
60
88
118
213
Stack
Diameter
(m)
6.8
3.0
4.1
4.7
7.0
Fuel Content
Tempera- Ash
ture (°C) (%)
160 0.17
127 0.07
154 8.0
166 14.0
77 30.0
S V
( % ) ( ppm )
2.4 590
1.2 15
3.3 99
3.9 35
5.4 50
Excess8"
Boiler
02(%)
0.2
3.0
5.0
4.0
3.8
Emission
Controls
only fuel
additives
only fuel
additives
ESP
ESP
2-stage wet
scrubber
(particulate
plus FGD)
Power
Output
(MW)
525
190
100
330
820
As measured at economizer outlet.
-------
RESULTS
Emission data were obtained on particulate concentration,
S02, and sulfates under various operating conditions of the
boiler (excess 02) and of the control systems (cutting back on
electric fields of ESP). Opacity data, however, were obtained
concurrently for a limited number of operating conditions. The
Lidar system was inoperative and undergoing repairs during the
studies on the coal-fired sources. Plume opacity data in these
instances were limited to Method 9. Poor weather conditions (fog)
also restricted plume observations during the study at Plant LC.
Data from the various plant studies on emission concentrations
of gross particulate matter, S02, total water soluble sulfates
(SO^), plume opacity, and in-stack opacity were reviewed. As
much as possible, data were selected for those periods of time
when these measurements were made concurrently. Tables 2, 3, and
4 are a consolidation of these data. Table 2 summarizes the emis-
sion data for oil-fired power plants without any emission control
systems, Table 3, for coal-fired power plants with ESP controls,
and Table 4, for a coal-fired power plant with 2 stage wet scrubbers
(particulate control plus ^flue j£as ^.esulfurization, FGD) . Various
constraints, such as number of available sampling ports in a given
location, did not permit the desired measurements to be executed
concurrently at Plant LC, and this is reflected in the data in
Table 4. These data represent sampling visits to this source on
three different dates.
Figure 2 graphically portrays the data of Table 2 showing
the comparison of the plume opacity data with the in-stack opacity
data at the oil-fired power plants. Figure 2 also provides data
on the effect of additional condensation on plume opacity with
cooling of the plume downstream of the stack exit location.
DISCUSSION
At the oil-fired power plants M and A (Table 2), the most dis-
tinguishing features are the vanadium content of the fuel oil,
levels of excess O2, and their impact on plume opacity. Plant M
was burning domestic fuel oil with about 15 ppm V and 1.2% S com-
pared to Plant A which was burning Venezuelan fuel oil with 590
ppm V and 2.4% S. Both plants utilized fuel additives. A signi-
ficant difference between plume and stack opacity existed for both
fuels and for different levels of excess O2.
The plume opacity at Plant A at normal operation was 31% at
the stack exit and above the opacity emission standard of 20%.
126
-------
Table 2. Emission Characterization Data for Oil-Fired
Power Plants Without ESP
Date
Plant M
8/10/76
Plant A
8/19/76
8/19/76
8/17/76
8/19/76
8/19/76
Part. S02 SO4
Time (mg/Nm3) (mg/m3) (mg/m3)
1115-1215 27 1,600 68
1125-1145
1145-1200
1145-1205 250 3,800 340
0945-1000 390
1000-1015 390
Plume
S04/SO,, Opac. (%)
(%) Lidar
4.1 10+2
31+4
45+3b
8.2 23+3
23+2
54+3b
Obs.
6+1
30+1
52+2b
42+1
37+2
61+2b
In-Stacka
Opac. (%)
2-3
18-22
18-22
11-15
11-15
11-15
15
590
590
590
590
590
Remarks
ppm V;
ppm V;
ppm V;
ppm V;
ppm V;
ppm V;
4% 02
0
0
0
0
0
.2%
.2%
.4%
.6%
.6%
°7
oz
°2
°?
°2
Transmissometer measurements.
Measurements about 3 stack diameters (15 meters) downstream of stack exit,
-------
Table 3. EPA Multispectrometer XRF Analyzer Element
Sensitivities and Detection Limits (2)
IV)
09
Element
F
Na
Mg
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Sensitivity,
counts/ 100 sec
per
220
534
10280
8074
11614
13392
28013
25394
121286
87817
85635
18010
7484
17522
13300
Detection
limit
(100 sec, 3a)
ng/cm2
149
29
2
3
3
15
9
9
2
2
2
7
19
14
18
Element
Co
Ni
Cu
Zn
As
Se
Br
Cd
Sn
Sb
Ba
Pt
Au
Hg
Pb
Sensitivity,
counts/100 sec
per /zg/cm2
16540
14504
18880
21066
17125
22922
50340
17303
14800
31100
25000
6812
8498
5776
16583
Detection
limit
(100 sec,
ng/cm2
3
10
43
7
10
12
28
• 2
2
4
7
20
91
90
30
-------
CO
Table 4. Emissions Characterization Data for Coal-Fired
Power Plant LC with FGDa
Date
9/19/77
9/19/77
9/19/77
9/19/77
11/2/77
11/3/77
11/3/77
4/3/78
4/3/78
4/4/78
4/4/78
4/5/78
4/5/78
. , b ,
Part. S02 S04 S04/SOXD Plume
Time (mg/Nm3) (mg/m3) (mg/m3) (%) Opac. (%)
1014-1121
1201-1309
1390-1500
1528-1635
1529-1629
1345-1445
1530-1630
1245-1345
1515-1615
1145-1245
1345-1445
0930-1030
1130-1230
2,000 190 8.7
3,100 260 7.7
4,200 170 3.9
5,500 170 3.0
240
75
270
350 >90
320 >90
220 >90
280 >90
260 >90
280 >90
In-Stack
Opacity (%)
59-67
62-71
62-71
>90
>90
>90
>90
>90
>90
L30% Ash, 5.4% S, 50-ppm V, 820 MW.
Measurements made on one of 8 parallel FGD modules,
"Observer measurements.
Transmissometer measurements at stack breeching.
-------
p
G
CD
O
JH
0
a
•H
O
d
a
o
•a
•H
J
CD
s
50
40
30
20
10
A'
7
o
V
^
°
• 1%S. 15 ppm V, 4%O2
A 2.5%S, 590 ppm V, 0.2%CX
" 2.5%S, 590 ppm V, 0.6%O.
Lidar Measurement
Location above Stack Exit
Unprimed - 2 to 3 meters
Primed - 15 meters
\
10 20 30 40 50
In-Stack Transmissometer Opacity, percent
Figure 2. Concurrent plume and in-stack opacity data
for emissions from oil-fired power plants.
130
-------
The opacity increased markedly to 43% further downstream as more
condensation of sulfuric acid occurred with cooling of the plume.
The observers tended to read higher opacity values than the Lidar.
Both solid particulate matter and condensed sulfuric acid
(liquid droplets) affect opacity. Unfortunately, concurrent data
are not complete in Table 2, but qualitative data from acid emis-
sion measurements in other studies (8) (9)(10) reinforce some
observations which can be made from Table 2. The concentration
of solid particulate matter increased as excess 02 for combustion
was decreased to reduce acid formation. We attribute this to
unburned carbon soot particles which have been observed in a
related study (11). The increase of in-stack opacity with decrease
in excess 02 demonstrates this. Since the acid is in the gas
phase in the stack, a decrease in acid with a decrease in excess
02 does not affect the in-stack opacity data or the above inter-
pretation of the data. As a matter of fact, the possibility of
reduced sulfate salts exists with the decreased acid, and this
would tend to counteract the effect of the unburned carbon and
imply a greater impact of acid on stack opacity than was actually
observed.
In the plume, the acid (at temperatures below its dewpoint)
appears as liquid droplets after condensation. With increased
excess 02, the increased acid and sulfate salts tend to increase
the plume opacity, but a counteracting effect is the concurrent
reduction in unburned carbon resulting from more complete combus-
tion of the oil. Since the condensation of the acid is a function
of the plume temperature, one can infer on a semi-quantitative
basis the contribution of the condensed acid in the plume opacity
relative to that of the solid particulate (both unburned carbon
and sulfate salts) by a comparison of the plume opacity at the
exit to that downstream of the exit. At 0.2% 02, the plume
opacity increase was from 31% to 43%. At 0.6% 02, the increase
was from 23% to 54%, indicating the presence of more acid at
higher excess O2 levels. The impact of the acid may actually be
more than indicated because dilution of the plume downstream can
reduce the opacity and counteract the effect of the increase in
condensed acid.
It is of interest to note that there appears to be an increase
(56%) in particulate loading as determined by Method 5 with an
increase in excess O2 (Table 2, Plant A). One might expect a de-
crease because of a reduction in unburned carbon with more complete
combustion. There is the possibility of an increase in measured
particulate loading due to the collection of the gaseous acid and
sulfate salts by Method 5. Related studies in our laboratory
have shown the glass fiber filter to be a good collector of the
131
-------
gaseous acid (12). One can postulate that two overlapping functions
(one an increase in sulfate salts and acid, and another a decrease
in unburned carbon with increasing excess 02) contribute to the
particulate loading. The former will be a curve with a positive
slope, the latter a curve with a negative slope. The resulting
curve on particulate loading as a function of excess 02 would
have a positive or negative slope depending upon which function
has the steeper slope- The resulting curve would approach a
straight line (zero slope), as the two functions tend to exactly
counteract each other. Consequently, depending upon the amount of
excess 02, the particulate loading may be on either the rising or
declining slope of the curve, or it may be more or less constant.
The emission characterization data for high sulfur (>2% S)
coal-fired power plants with ESP controls show no significant
difference between plume and in-stack opacity under normal ESP
operation or with reduced electric fields in the ESP's. It is
possible that the high ash content of the coal and resulting high
particulate loading in the emission have a predominant effect on \
the in-stack and plume opacity. In-stack opacity at normal operal-
tion of the ESP was close to the opacity emission standard for ne*w
sources and_higher than that for the oil-fired power plants. Th^
ratio of 864 to particulate matter for the coal-fired emissions
is <1 and for the oil-fired emissions, >1.
The stack gas environment for the coal-fired power plant (LC)
with particulate and gas controls (2 stage wet scrubber) was unlike
that for plants with ESP controls. The stack gas temperatures
were below the sulfuric acid dewpoint, and the water vapor content
from the wet scrubbers was high. The result was that sulfuric
acid will appear in the gas stream as condensed acid droplets, and
these directly affect the in-stack opacity.
Concurrent emission data could not be obtained for Plant LC
(Table 4) in the same manner that it was for Plants P and MC (Table
3) because the required number of sampling ports was not available.
Nonetheless, the data obtained from the three visits to Plant LC
do permit some qualitative observations. The emission data on
the particulate loading appear to fall within a narrow range of
values indicating consistent plant operation. The plume and in-
stack opacity data show no significant difference, but the very
high opacity values are not consistent with the particulate
loading, size, and composition normally associated with fly ash
emissions. Optical transmittance measurements conducted at dif-
ferent wavelengths during the study gave data that varied with
wavelength in the visible portion of the spectrum (13). This
variation with wavelengths is indicative of submicron size distri-
bution. The submicron size was also substantiated with in-stack
132
-------
impactor measurements (14). The acid composition of the gas stream
was substantiated by the controlled condensation measurement data
(7). In this case, we attribute the high in-stack opacity levels
to fine particle concentration with a mass mean diameter in the sub-
micron size range and with a significant percentage of the composi-
tion consisting of condensed sulfuric acid.
There are a number of important questions raised by the data
obtained thus far. More studies are needed to adequately address
these questions and to determine the variation of these pollutant
emissions with operating parameters. The questions can be
briefly stated as follows:
• What is the quantitative distribution of H2S04 in the
gas stream between free acid and acid adsorbed on
particulate matter?
• What is the distribution of the sulfate ion (SO^)
between acid and salts?
• What is the size distribution of acid and salts?
There is need for more data on the physical properties of H2S04 in
both the stack and plume environments to support proper interpre-
tation of optical data.
SUMMARY
Emissions from oil-fired power plants without emission con-
trols and coal-fired power plants with ESP's and with FGD gystems
were characterized for plume and in-stack opacity, S02, S04 , and
mass concentration. Sulfuric acid content of the emissions from
the oil-fired power plant had a significant effect on the plume
opacity but no effect on the in-stack opacity. In the case of
the coal-fired power plants with ESP's, the in-stack and plume
opacities were essentially the same. This led to the conclusion
that the concentration of acid was low relative to the non-acid
particulate such that the acid did not contribute to any signifi-
cant degree to the opacity of the plume beyond that normally
associated with the fly ash. The in-stack and plume opacities of
the emissions for the high sulfur coal-fired power plant with the
2 stage wet scrubber system were comparable but significantly
high (70% to 90%). The high opacity was attributed mainly to
the sulfuric acid content of the emissions and to submicron size
of the particulate matter.
133
-------
REFERENCES
1. U.S. Environmental Protection Agency, Standards of Performance
for New Sources. Federal Register, 36, 1971. 24876-24895.
2. U.S. Environmental Protection Agency, Standards of Performance
for New Sources. Federal Register, 40, 1975. 43850-43854.
3. Cook, C. S., G. W. Bethke, and W. D. Conner. Remote Measure-
ment of Smoke Plume Transmittance Using Lidar. Appl. Opt.,
11:1742-1748, 1972.
4. Johnson, W. B., R. J. Allen, and W. E. Evans. Lidar Studies
of Stack Plume in Rural and Urban Environments. EPA 650/4-73-
002, U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina, 1973. 112 pp.
5. Peterson, C. M., and M. Tomaides. In-Stack Transmissometer
Techniques for Measuring Opacities of Particulate Emissions
from Stationary Sources. NTIS PB-212-741.
6. Herget, W. F., and W. D. Conner. Instrumental Sensing of
Stationary Source Emissions. Environ. Sci. and Technol.,
11:962-967, 1977.
7. Homolya, J. B. Unpublished data. U.S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina, 1978.
8. Homolya, J. B., and J. L. Cheney. An Assessment of Sulfuric
Acid and Sulfate Emissions from the Combustion of Fossil Fuels.
In: Proceedings of Workshop on Measurement Technology and
Characterization of Primary Sulfate Emissions from Combustion
Sources, J. S. Nader, ed., Southern Pines, North Carolina,
April 1978. EPA document (in press).
9. Dietz, R. N., and R. F. Wieser. Operating Parameters Affec-
ting Sulfate Emissions from an Oil-Fired Power Unit. In:
Proceedings of Workshop on Measurement Technology and Charac-
terization of Primary Sulfate Emissions from Combustion Sources,
J. S. Nader, ed., Southern Pines, North Carolina, April 1978,
EPA document (in press).
134
-------
10. Cheney, J. L. , and J. B. Homolya. Characterization of Combus-
tion Source Sulfate Emissions with a Selective Condensation
Sampling System. In: Proceedings of Workshop on Measurement
Technology and Characterization of Primary Sulfate Emissions
from Combustion Sources, J. S. Nader, ed., Southern Pines,
North Carolina, April 1978. EPA document (in press).
11. Bennett, R. L., and K. T. Knapp. Chemical Characterization
of Particulate Emissions from Oil-Fired Power Plants. In:
Proceedings of the Fourth National Conference on Energy and
the Environment, Cincinnati, Ohio, October 1976. pp. 501-506.
AICHE, Dayton, Ohio, 1976. 594 pp.
12. Cheney, J. L. Unpublished data. U.S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina, 1978.
13. Conner, W. D. Unpublished data. U.S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina, 1978.
14. Knapp, K. T. Unpublished data. U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1978.
135
-------
Query: Is There a Connection between the
Expansion of Areas of Acid Rain and a Shift
from Coal to Oil for Small-Scale Heat Needs?
Arthur M. Squires
Virginia Polytechnic Institute & State University
ABSTRACT
Oden's classic maps show the "explosion" of acid rain
over Northwest Europe between 1956 and 1966. Is there a
connection between Oden's data and the shift from coal to
oil for much of Europe's small-scale heat needs?
Since the oil in question had a much higher sulfur con-
tent than the coal it replaced, emissions of sulfur
dioxide sharply increased. Historic combustion data sug-
gest that a small oil furnace may emit appreciably more
sulfuric acid mist than a small coal furnace, even when
the fuel that is used contains low levels of sulfur.
Before the query can be answered, we need data on emis-
sions of sulfuric acid mist from real-world oil furnaces,
even those firing low-sulfur oil, as well as data on
atmospheric conversion of sulfur dioxide. The acid emis-
sions data should be for furnaces representing a wide
spectrum of age, type, history, condition of maintenance,
firing practice, etc. Needed data on emissions from small
coal furnaces of modern design will be acquired at VPI &
SU in a Coal Combustion Workshop emphasizing state-of-the-
art furnaces up to about 3 megawatts (thermal).
In the late spring, we will commission our first furnace
(an English fluidized-bed unit at 0.3 megawatt) and mea-
sure emissions of trace elements.
137
-------
Oden's classic maps of acid rain over Northwest Europe between
1956 and 1966 show the "explosion" of this phenomenon in those
years (1).
Those were the years, however, when Western Europe shifted
from coal to oil as its "growth fuel." It is evident from Table 1
that most planners chose oil for new sources of heat. (In Table 1,
Western Europe includes Greece and Yugoslavia.) Since coal for
steel and large-scale electricity generation continued to increase,
some replacement of coal by oil for heat needs on a smaller scale
must have occurred.
Table 1. Annual Consumption of Coal and Oil in
Western Europe, Expressed in Millions of Metric Tons
of Hard Coal Equivalent
Year Coal Oil
1950 464 86
1958 540 189
1966 486 516
Ockham's razor ("Pluralites non est ponenda sine necessitate"
= Multiplicity ought not be posited without necessity) is a use-
ful principle for the engineer, who must often blend engineering
science of limited applicability with imperfectly understood art
in order to make his design and to accomplish his job. Perhaps,
then, an engineer will be forgiven if he associates the spread of
acid rain in Western Europe with the shift to oil.
Further, it may be reasonable to consider the simplest explana-
tion of this association. Combustion engineers have long recog-
nized that oil furnaces tend to emit a good deal more sulfuric
acid mist, other factors being equal, than do coal furnaces.
Krause (2) reviewed the evidence prior to 1959.
We might, parenthetically, remember that in the 1960's engi-
neers developed the practice of limiting the excess air to large
utility boilers fired with oil, and this practice came into use only
late in that decade. It would no longer be true to say that oil
-•38
-------
furnaces inherently make more sulfur trioxide than coal furnaces,
since a carefully managed oil furnace can make less. Combustion
with low excess air, however, is a practice limited to relatively
large boilers and is not accessible to the operator of a small
furnace.
The simplest explanation of Oden's maps, then, is that the
growth of oil combustion produced a large increase in primary emis-
sions of sulfuric acid mist. Is this the correct explanation? It
would be foolhardy to do more than put it forward gingerly as a
hypothesis well worth examining.
What is missing, however, for a thorough test of the hypothe-
sis is an adequate data base on emissions of sulfuric acid mist from
oil furnaces, especially from those of smaller size. Few measure-
ments appear to have been made, for example, on furnaces in the
roughly 300 kilowatt to 6 megawatt thermal range of size (30 to
600 boiler horsepower; 1 to 20 million Btu's per hour), which accounts,
for example, for roughly one-half of the emissions of sulfur oxides
in New York City. No measurements have been made, as far as the
author is aware, that are representative of actual firing practices
over typical daily and annual operating cycles of such equipment.
Another curiosity, which the razor suggests might reasonably
be set alongside Oden's maps in our minds, is the persistence of
high levels of sulfate particulate matter in New York City after
emissions of sulfur dioxide have sharply declined. Seeking a simple
explanation, related to our hypothesis concerning acid rain, we
might consider the long-appreciated fact that emissions of sulfuric
acid mist from oil firing do not track the sulfur level of the oil.
Krause (2) reviewed the evidence. Unpublished data taken by KVB,
Inc. on five small furnaces (between 84 and 430 boiler horsepower)
burning low-sulfur oil (between 0.09% and 0.28% sulfur) tend to
confirm the historic evidence that conversions of fuel sulfur to
the trioxide in this low range can run many times greater than the
roughly 1% conversion to be expected for oil at 2.5% sulfur.
A subjective observation by a former resident of New York
City would be that something like one oil furnace out of twenty is
managed by an operator who daily puts out dense clouds of black,
sooty smoke for several fifteen-minute intervals, in summer as well
as winter. No doubt it will be difficult to develop a good figure
for input of sulfuric acid and sulfate particulate matter into New
York City's air, but the effort should include measurements for
poorly-run equipment. Obviously, measurements would need also to
be made for furnaces representing a wide spectrum of age, type,
history, and conditions of maintenance.
139
-------
Returning to Oden's maps, we should remember that the oils
burned in Western Europe in the 1950's and 1960's tended to have
higher sulfur levels than European coals, and so the data of Table 1
imply a large increase in emissions of sulfur dioxide. Perhaps, as
well as the increase in primary emissions of sulfuric acid, there
also may have been a significant increase in the synthesis of acid
from sulfur dioxide in the atmosphere.
Granted this probability, we may still regard it as curious
that acid rain did not draw attention before the sharp expansion
of oil firing. We may, therefore, still seek parsimonious hypo-
theses to test. There is a body of thought that associates con-
version of sulfur dioxide to trioxide in the atmosphere with pre-
sence of particulate matter. Perhaps a part of the explanation
of Oden's maps is associated with earlier deposition of sulfur tri-
oxide adsorbed upon larger particles from the stack of a coal-
fired furnace (even one fitted with a good precipitator). perhaps
oil combustion is placing ultra-fine particulate matter into the
atmosphere, laden with adsorbed sulfate, that travels long dis-
tances .
These suppositions would fit the recent evidence that sulfate
particulate matter in New York City tracks the level of such matter
at a rural sampling station upwind of the city (3).
There is also a body of thought that associates the persis-
tence of high levels of sulfates in urban atmospheres with conver-
sion of sulfur dioxide, even at the new low levels, upon particu-
late matter found in these atmospheres. Some light on this opinion
might be shed by study of selected cities that represent extremes
in the weight ratio of smoke to sulfur dioxide. This ratio is low,
for example, in London (about 0.2) and is astonishingly high in
Madrid (1.3 to 1.5) (4).
Perhaps Ockham's razor is a poor guide in this complex situa-
tion. It does, however, point to a serious gap in our information
concerning emissions of sulfuric acid mist from combustion, viz.,
from small oil furnaces.
Lest we be in for yet more surprises, we need better data for
emissions of all kinds from the variety of new coal furnaces that
may find increasing favor in the years ahead.
Virginia Polytechnic Institute & State University is setting
up a Coal Combustion Workshop that will emphasize state-of-the-
art furnaces up to about 3 megawatts thermal. We hope to commis-
sion our first furnace, an English fluidized-bed unit at 0.3 mega-
140
-------
watt, within the next few months. Our primary emphasis will be
upon extension instruction of the general public, but we will rou-
tinely study emissions from the dozen or so furnaces we hope to
acquire.
Let no one suppose that the hypotheses set forth here concern-
ing Oden's maps imply a defense of Europe's historic practices for
burning coal. The author himself experienced two of London's kill-
ing smogs, in late 1958 and late 1959. Therefore, he was much
impressed with the dramatic changes realized during the 1960's from
the implementation of Britain's Clean Air Act of 1956. Few Ameri-
cans, however, appreciate, as the English do, that small coal
furnaces can be clean. As a Nation, we remember the dirt and labor
of most small coal fires of the past. We remember the smoke and
soot of cities like St. Louis and Pittsburgh. The primary aim of
VPI & SU's Coal Combustion Workshop will be to demonstrate that
coal can be clean.
141
-------
REFERENCES
1. Oden, S. NederbOrdens fSrsurning-ett generellt hot mot
ekosystemem. I Mysterud (red.)- Forurensning og biologisk
miljovern, Universitetsforlaget, Oslo, 1971.
2. Krause, H. H. Oxides of Sulfur in Boilers and Gas Turbines.
In: Corrosion and Deposits in Boilers and Gas Turbines, ASMK
Research Committee on Corrosion and Deposits from Combustion
Gases, prepared by Battelle Memorial Institute. ASME, New York,
1959. pp. 44-77.
3. Leaderer, B. P. Summary of the New York Summer Aerosol Study
(NYSAS). J. Air Pollution Control Assoc., 28:321-327, 1978.
4. Stichting CONCAWE. Characteristics of Urban Air Pollution:
Sulphur Dioxide and Smoke Levels in Some European Cities.
Special Task Force: Characteristics of Urban Air Pollution.
Report Number 4/76, The Hague, The Netherlands, March 1976.
142
-------
Report of the Working Group on
Characterization of Gaseous Sulfur
Oxides Emissions
Arthur M. Squires, Reporter
This Working Group's objective was to review the status of
characterization data for gaseous sulfur oxides emissions. Its
conclusions and recommendations are as follows.
STATUS AND VALIDITY OF AVAILABLE DATA
Available sulfur dioxide data are generally reliable and use-
ful for judging sulfur dioxide emissions from combustion sources.
The Group noted, however, that measurements of sulfur dioxide
need to be accompanied by simultaneous measurements of oxygen level
and temperature at the point of measurement. When data are to be
used for emission comparisons, not only should the original data
be reported, but a sulfur dioxide level that is corrected to a stated
standard oxygen level should also be noted. (The level of 3% oxygen
has become standard for reporting nitrogen oxide emissions.)
Sulfur dioxide emission rate data are calculated from sulfur
dioxide and oxygen concentration measurements and an F-Factor.
Emission rate data can also be determined from measurements of
flue gas velocity and sulfur dioxide concentration.
The concentration of sulfuric acid vapor from an oil-fired
plant reflects a higher conversion of fuel sulfur, given prevalent
current firing practices, than the concentration of sulfuric acid
vapor from a coal-fired plant. Oil-fired sources tend to produce
particulates with high content of soluble sulfate matter. The
fraction of soluble particulate sulfates versus acid in coal-fired
emissions depends upon the amount and nature of the ash in the
coal and may vary widely.
143
-------
RECOMMENDATIONS
Sulfuric acid emissions data should be reported in parts per
million of H2SO4 by volume. If the data are reported as mg/Nm3,
or in other units, the conversion factor to ppm by volume should
be stated along with the data.
Water soluble sulfate particulates should be reported in mass
units of sulfate ion per gas volume (e.g., mg/Nm3).
Certain additional measurements need to be made in support of
sulfuric acid emissions data.
• Furnace oxygen level. This is difficult to measure
directly, and no good or convenient sampling technique
is available. The Working Group recognized development
of a sampling technique as a need for research.
• Temperature and oxygen level at the measuring point.
• Temperature of gases entering the convective pass of
a utility boiler, or temperature data generally indicating
the time-temperature history of the combustion gases.
In spite of some questions about measurement techniques and
conditions of measurement, the available sulfuric acid and soluble
sulfate particulate data do make a fairly consistent but incomplete
picture. More data are needed to support the picture available today
for large utility boilers firing oils of moderate to high sulfur
levels. Almost totally lacking are emissions data for furnaces of
all types burning low-sulfur oils, e.g., at 0.5% sulfur and below,
and emissions data for oil-fired industrial boilers and other smaller
furnaces, such as heating units for apartment buildings, business
establishments, and the like, most of which now burn low-sulfur oils.
Smaller boilers burning low-sulfur fuels may in combination
constitute significant emission sources of sulfuric acid and soluble
sulfate particulate matter for important air sheds, and their study
is needed. It costs as much to study a small unit as a large one,
and budgets will have to reflect this fact.
Workers who conduct studies of conversion of sulfur dioxide in
the atmosphere to secondary sulfate particulates (e.g., plume studies)
should be aware of the need to have good inputs of emissions data
for sulfuric acid and soluble sulfate particulates from the re-
levant sources. For example, there is need during a plume study to
check and record the operating conditions of the plant generating
the plume.
144
-------
For studies of plume opacity, there is need to acquire data
on the relative contribution of fly ash and acid aerosol to the
opacity of a plume, both in terms of in-stack measurements and of
data gathered in the open plume. There is need for data to allow
better understanding of the optical properties of plumes:
• Distribution of sulfate ion between acid and particulates,
• Size distribution of mist particles and of other
particulates.
• Role of solid particles as condensation nuclei for acid.
145
-------
Section 2
Paniculate Emissions
147
-------
Characterization of Fly Ash from Coal
Combustion
David F. S. Natusch
Colorado State University
ABSTRACT
Fly ash derived from coal combustion contains predominantly
spherical particles which consist of an insoluble aluminosi-
licate glass containing several mineral impurities. An
outer layer, 50 to 300 A thick, is rich in many potentially
toxic trace elements in the form of simple and complex sul-
fates. This layer, which is soluble in water, contains
essentially all of the particulate sulfur present in fly
ash in the form of sulfate. The actual mechanism(s) of
formation of particulate sulfate salts are ill-defined but
probably involve adsorption of condensation of gaseous sul-
fur species onto fly ash surfaces within the power plant
stack system.
INTRODUCTION
At the present time approximately 80% of the electric power
generated in the United States is derived from the combustion of fos-
sil fuels. Of this total, coal combustion accounts for approximately
70% with the balance made up by natural gas and oil. Furthermore,
it is now clear that increased coal utilization will be the primary
means of meeting the nation's electrical energy needs for the next
several decades at least.
It is well established (1) that sulfur present in coal is mo-
bilized, almost quantitatively, into the stack gas stream as a re-
sult of combustion in coal-fired power plants. In conventional com-
bustion operations, most of this sulfur is present as sulfur dioxide,
149
-------
together with small amounts in the form of sulfur trioxide, sulfuric
acid, and particulate sulfate salts. The latter are associated pri-
marily with fly ash particles whose physical and chemical character-
istics may play a controlling role in determining the environmental
impact of both emitted sulfate salts and of other sulfur species
(e.g., sulfur dioxide) which may interact with fly ash following
emission.
It is the purpose of this paper to summarize the available in-
formation about the physical and chemical characteristics of coal
fly ash and to assess the status of present knowledge about sulfur
species associated with fly ash. In both cases emphasis is placed
on what is known about the fundamental processes which control the
formation, transformation, and subsequent environmental behavior
of coal fly ash and its sulfur-containing constituents.
NATURE OF COAL FLY ASH
A number of workers have undertaken detailed physical and
chemical studies of coal fly ash (2-5) , and its general character-
istics are now quite well known. It is, however, important to note
that fly ash derived from different power plants may exhibit con-
siderable variability due, primarily, to differences between coal
types and the nature of the combustion conditions. In this regard,
combustion temperature is a very important factor insofar as it
determines whether or not the fly ash matrix is molten at any stage
and whether potentially volatile species actually experience a vapor
phase history. It is also extremely important to recognize that
most studies of fly ash are conducted on samples which are retained
by particle control devices so do not truly represent material emitted
to the atmosphere.
Morphology and Matrix Composition
Derived from mineral impurities present in the coal, coal fly
ash particles are primarily inorganic in nature. Consequently, the
amount of mineral matter present in a given coal strongly influences
the particle emission factor for that coal. The major elemental
constituents of coal fly ash are Si, Al and Fe, together with minor
amounts of Ca, Mg, K, Na, Ti, and S. Some typical concentration ranges
of these elements in U.S. coal fly ashes are presented in Table 1 (6).
In general, fly ashes derived from western U.S. coals have a higher
ash content and exhibit higher alkali and alkaline earth metal con-
tents than do those from eastern coals, which are typically higher
in sulfur.
150
-------
Table 1. Typical Matrix Element Composition Ranges
of Some U.S. Coal Fly Ashes Expressed as
Weight Percentages of the Oxides
Matrix element composition,
wt-%of oxide
Major constituent
A12O3 14-30
SiO2 22-60
Fe203 3-21
K20 0.2-3.5
CaO 0.5-31.0
Minor constituents
Li20 0.01-0.07
Na20 0.2-2.3
MgO 0.7-12.7
TiO2 0.6-2.6
P205 0.1-1.1
S04* 0.1-2.2
*
Soluble sulfate
During combustion in a modern coal-fired power plant, the min-
eral impurities in coal melt and form small, mostly spherical, par-
ticles. The extent to which these molten particles coalesce or
disintegrate into even smaller droplets is determined in part by
the geometry, flow characteristics, and combustion conditions within
the plant. Consequently, the size distribution of the particles
produced may vary significantly between different plants. In a few
plants of obsolete design (e.g., chain grate stoked), as well as in
modern fluidized bed plants, combustion temperatures are not suf-
ficiently high to melt the fly ash matrix, so that irregularly shaped
particles are formed. Since these cannot readily disintegrate, their
size distribution is generally centered around larger, median values
than those encountered with spherical fly ash particles.
The aerodynamic equivalent mass median diameters of coal fly
ashes in the absence of particle control devices typically lie in
the range 8 fim to 30 pm (7), and the mass is reasonably approximated
by a log-normal distribution. The mass median diameters of fly ashes
emitted from control devices are considerably smaller than indicated
above and depend largely on the collection efficiency of the control
devices. In the case of electrostatic precipitators, mass removal
151
-------
efficiencies in excess of 98% are often achieved, and aerodynamic
mass median diameters of emitted fly ash are typically in the range
0.5 fzm to 2 fim.
While a number of distinct morphological forms of fly ash can
be distinguished (10), only three are highly abundant. The first
involves solid, or slightly voided, spheres and accounts for most
fly ash particles having physical diameters less than about 5 fim.
The second morphological form consists of hollow spheres whose in-
terior voids are filled with carbon dioxide at a pressure of about
0.2 atm (4). These particles predominate in the physical diameter
range 10 ptm to 60 jzm. Finally, and most intriguing, are hollow
particles filled with large numbers (10 to 200) of small solid par-
ticles. This encapsulation phenomenon is encountered primarily for
particles in the physical diameter range 20 fim to 60 fim (8) . The
phenomenon of particle encapsulation in fly ash is not fully under-
stood; however, there is good evidence to show that encapsulated
particles are actually formed inside their hosts so are not available
to interact with external vapors and gases such as S02 (9).
As a result of the widespread occurrence of hollow and encap-
sulating fly ash particles, the measured density of coal fly ash is
essentially unrelated to the density of the matrix material. X-ray
and electron diffraction studies of fly ash indicate that the matrix
consists, for the most part, of an aluminosilicate glass together
with small amounts of the minerals a quartz (SiC>2), mullite
(3Al203.2Si02) , hematite (Fe2O3), and magnetite (Fe304). Fly ashes
derived from western U.S. coals also have some crystalline gypsum
(CaS04 .2H2O) and lime (CaO) . It is apparent, therefore, that coal
fly ash is highly heterogeneous in nature and is likely to exhibit
low aqueous solubility.
Trace Element Distribution
The specific concentrations (^tg/g) of individual trace elements
in coal fly ashes depend primarily on the trace element content of
the original coal. In general, a fly ash which contains high con-
centrations of one trace element will also have high concentrations
of most others as well. However, the relative elemental concentra-
tions encountered in fly ash may differ markedly from those in the
original coal due to the different partitioning characteristics of
individual trace elements between bottom ash and fly ash. Table 2
lists some typical specific concentration ranges for a number of
trace elements encountered in coal fly ash and compares them with
values for oil fly ash.
152
-------
Table 2. Specific Concentrations of Elements
in Coal and Oil Fly Ashes
Element
Al
As
Au
B
Ba
Be
Br
Ca
Cd
Ce
Cl
Co
Cr
Cs
Cu
Fe
Ga
Hf
Hg
I
In
K
La
Lu
Mg
Mn
Mo
Na
Ni
Pb
Rb
Sb
Sc
Se
Sm
Sn
Sr
Ta
Th
Ti
Tl
ri
u
V
W
Yh
JL L/
Zn
Coal Fly Ash
Specific concentration,
Mg/g
70,000-140,000
2-500
0.004-0.1
10-600
500-7000
1-10
0.3-20
6000-180,000
0.1-50
100-300
10-500
5-100
50-300
1-20
50-650
25,000-300,000
10-250
5-10
0.02-0.4
0.5-7
0.1-0.3
1500-35,000
35-100
0.5-2
11,000-60,000
50-500
5-40
1200-18,000
5-100
5-1000
40-300
1-15
10-40
. 1-20
10-20
30-30
50-4000
0.5-1.5
15-70
3500-8500
2-30
5-20
100-500
3-10
3-7
50-5000
Oil Fly Ash
Specific concentration,
Mg/g
100-5000
30
__
—
500-10,000
—
—
10-1000
—
—
—
90
66
—
50-2000
10,000-100,000
—
—
—
—
—
1000
—
—
500-5000
1-100
—
2000-50,000
—
200-2000
—
5
—
5
__
—
—
—
— —
—
—
— —
100-200,000
— —
— —
200-3500
153
-------
It is now well established (2)(3)(8) that a number of elements,
including As, Cd, Cu, Ga, Mo, Pb, S, Sb, Se, Tl, and Zn, tend to
increase in specific concentration with decreasing particle size.
This is attributed to a mechanism whereby certain elements, or their
compounds, are volatilized during combustion and then condense back
onto the surfaces of co-entrained fly ash particles as the tempera-
ture falls to the dewpoint of each vapor species. A great deal of
evidence has been presented in support of this mechanism (2)(3)(10);
however, it is becoming increasingly apparent that several additional
factors may also operate. For example, recent work (8) suggests
that the physical and chemical behavior of individual elements dur-
ing combustion can be correlated with their geochemical classifica-
tion. Thus, the chalcophile, lithophile, and siderophile elements
exhibit different partitioning behavior which determines their
distribution in coal fly ash. In addition, a distinct influence of
individual particle matrix composition and specific surface area
is observed.
Undoubtedly the most important consequence of the volatilization-
condensation phenomenon exhibited by some trace elements is their
pronounced enrichment in the region of individual particle surfaces.
An example of such enrichment is presented in Figure 1 in which the
concentration dependence of lead on depth below the particle sur-
face is illustrated. The importance of this surface enrichment
lies in several factors, viz.
(1) It is the particle surface which comes in direct contact
with the external environment so that the highest concentrations of
potentially toxic and reactive trace elements are mostly readily
accessible. A rough comparison of estimated surface and bulk con-
centrations of several elements in a coal fly ash is presented for
illustration in Table 3.
(2) Material present in the region of surface enrichment is
readily soluble in aqueous media (Figure 1), thereby rendering the
most environmentally undesirable trace elements mobile and available
for interaction with the external environment. In this regard it
should be recognized that only about 2%-3% of the total mass of
coal fly ash is soluble in water.
(3) Conventional analyses of bulk fly ash grossly underesti-
mate the effective concentrations of most trace elements which are
actually available for interaction with the external environment
(Table 3).
154
-------
APPROXIMATE DEPTH (A)
210
3 160 320 480 640
I 1 I 1 I I i ! I
800 960
I 1 1
en
t/5
Z
UJ
IU
40000
CO
30000 8:
20000
10000
LU
o
X
o
Of
a.
a.
40 80 120
TIME (SECS)
Figure 1. Dependence of the elemental concentration of
lead on depth below the surface of a coal fly
ash particle before (A) and after leaching
with water ( • ) and dimethyl sulfoxide ( • )
as determined by secondary ion mass spectro-
metry.
160
200
240
-------
Table 3. Estimated Surface Concentrations
of Elements in Coal Fly Ash
Estimated surface
Bulk concentrated in
concentration, 300 A layer,
Element Mg/g Mg/g
As
Cd
Co
Cr
Pb
S
V
600
24
65
400
620
7,100
380
1,500
700
440
1,400
2,700
252,000
760
When one takes account of the fact that condensation of trace
metals onto fly ash particle surfaces almost certainly takes place
at much higher temperatures than does condensation of S03, or adsorp-
tion of SO2, it will be recognized that much of the interaction of
sulfur species with fly ash is likely to be with trace metal species
rather than with the particle matrix material. This may be extremely
important in determining both the nature of the interactions and
their resulting products.
SULFUR IN COAL FLY ASH
Current information about the chemical and physical status of
sulfur present in fly ash is fragmentary. Nevertheless, a useful
picture of its probable behavior can be assembled. For this pur-
pose, it is helpful to consider the inter- and intra-particle dis-
tribution of sulfur, its chemical forms, and the probable mechanisms
of formation of particulate sulfate salts.
Distribution of Sulfur
As pointed out earlier, sulfur is one of those elements which
increase in specific concentration with decreasing particle size
(2). However, unlike most of the trace metals, which exhibit such
156
-------
size dependence, a linear correlation between specific concentra-
tions of sulfur and particle surface area is difficult to establish
(2)(3). Indeed, separation of the dependences of concentration on
physical size, density, and feromagnetism, as illustrated in Table 4,
indicates a rather complicated dependence on both particle size
and density. The reasons for these dependences are not clear.
Table 4. Distribution of Sulfur Concentration (% by wt)
as a Function of Physical Size, Density, and
Ferromagnetic Character in Coal Fly Ash
Particle Size
Density (g/cm3)
2.1-2.5
2.5-2.9
>2.9
Nonmagnetic
<20
20-44
44-74
>74
0.24
0.11
0.21
0.31
0.40
0.48
0.37
0.12
0,22
0.82
1.26
0.48
—
0.43
1.02
0.71
Magnetic
<20
20-44
44-74 0.10 0.21
>74 — 0.43
0.16
0.45
0.34
0.20
0.19
0.09
0.28
0.14
Analyses of individual particles and groups of particles by
means of ion microprobe mass spectrometry and Auger electron spec-
trometry (10) establishes beyond reasonable doubt that the sulfur
associated with coal fly ash is present in a layer of the order of
50 I thick at the particle surfaces (Figure 2). Furthermore, this
layer is sufficiently soluble to enable almost quantitative removal
of all sulfur species by continued washing with water or mineral
acids. An example of such removal is presented in Figure 3 which
illustrates the dependence of sulfate concentration on time in
individual washings during Soxhlet extraction of coal fly ash with
water at 25°C. (This technique is later referred to as Time Resolved
Solvent Leaching, TRSL.)
157
-------
30
60
APPROXIMATE DEPTH (A)
90 120
150
180
210
i
X
S(LI¥1: 152 eV)
LU
O1
00
GO
LU
>
LU
cc:
20
^^
"V
TIME
Figure 2. Dependence of the elemental concentration of
sulfur on depth below the surface of a coal
fly ash particle as determined by Auger elec-
tron spectrometry.
-------
en
CO
T.R.S.E. PROFILE OF FLY ASH
(WATER)
F"
cr
S04=X2(10"2)
20 24 28
TIME (HR.)
32 36 40 44
48
Figure 3. Time resolved leaching profile for sulfate, chloride
and fluoride anion extraction from coal fly ash.
-------
Analyses of fly ash which has been exhaustively leached with
water indicate that very little, if any, sulfur remains, even though
only 2%-5% of the fly ash mass actually dissolves. It is apparent,
therefore, that sulfur is, at most, only a trace constituent in the
fly ash matrix even though it is a major component of the particle
surface layers.
Chemical Forms of Sulfur
Studies of fly ashes derived from the oxidative combustion of
coal and oil using Electron Spectrometry for Chemical Analysis (ESCA)
show that sulfur is present in the +6 oxidation state (10). Parti-
culates derived from coal conversion processes, which involve reduc-
ing conditions, contain sulfur in the -2 oxidation state, however
(11). Neither result is unexpected. Time resolved solvent leaching
studies of coal fly ash, in which analyses of soluble anions are per-
formed by means of ion chromatography, indicate that sulfate is the
only sulfur-containing anion leached by water.
It is probable, therefore, that the sulfur species present in
the surface layer of coal fly ash is, at least predominantly, and
probably exclusively, in the form of sulfate.
Some evidence is available regarding the cations which are
associated with sulfate species in coal fly ash. Thus, X-ray powder
diffraction patterns of some fly ashes indicate the presence of
either anhydrite (CaS04) or gypsum (CaS042H20). These species are
present most commonly in fly ashes derived from western U.S. coals
which contain especially high levels of calcium. The two forms
result, apparently, from exposure of the highly hygroscopic anhy-
drite to moisture. In a sense, therefore, the occurrence of gypsum
is probably artefactual.
Quite strong indications have also been obtained for the exis-
tence of several trace metal sulfates in coal fly ash. Thus, both
Fourier Transform Infra Red Spectroscopy and Time Resolved Solvent
Leaching (TRSL) provide evidence for the presence of Cd, Co, Cr,
Mo, and Ni sulfates in coal fly ash. The alkali metals Ba, Cu, and Ca
are also present, at least partly, in the form of sulfates. Even
stronger evidence is available (12) for the existence of Al and Fe
as sulfates in the surface layer of fly ash.
While definitive evidence is lacking, present indications are
that essentially all of the elements present in the so-called sur-
face layer of coal fly ash exist in the form of sulfates. Two points
160
-------
must, however, be recognized. First, the actual sulfate compounds
are probably not simple but may consist of mineral forms which may
include double salts. For example, the existance of alkali iron
tri-sulfates has been suggested (10). Secondly, it is clear (at
least in the case of the minor elements such as Ba, Ca, Mg, K,
and Na) that a given metal may be present in more than one chemical
form. No evidence has been found for the presence of free H2S04
in fly ash particles.
Association of Sulfur with Fly Ash
The fact that sulfur present in coal fly ash is present almost
entirely in the so-called particle surface layer provides very
strong support for the proposition that sulfur-containing gases or
vapors interact with the surfaces of co-entrained fly ash particles
in a power plant stack. What is not clear is whether the inter-
action is via condensation, adsorption, or chemical reaction, or
whether sulfur dioxide, sulfur trioxide, or sulfuric acid is the
primary reactant.
Simple vapor pressure calculations indicate that condensation
of 863 and H2SO4 is unlikely to occur at the temperatures encountered
in a coal-fired power plant. Yet fly ash with well-formed sulfate
surface layers is routinely collected at such temperatures (e.g.,
from electrostatic precipitators). One is inclined, therefore,
to rule out condensation processes as being responsible for sur-
face deposition of sulfates unless direct condensation of a metal-
sulfate from the vapor phase occurs. As far as we are aware, there
is no evidence whatsoever to support such an idea.
By default, therefore, one is left with the process of adsorp-
tion of S02, S03, or H2S04 as being responsible for formation of
particulate sulfate salts. In this regard it should be noted that
adsorption of SO2 would require fairly rapid (possibly catalytic)
oxidation to the sulfate species.
It is apparent from the foregoing remarks that further research
into the mechanism(s) of formation of particulate sulfate salts is
required. In this regard, it is stressed that the toxicological
implications of particulate sulfate salts make such research far
from academic insofar as knowledge of formation mechanisms may well
provide information necessary for development of effective control
strategies.
161
-------
CONCLUSIONS
Overall, it appears that the physical and chemical character-
istics of coal fly ash are quite well defined. Thus, the material
is in the form of spherical particles which consist primarily of
an alumino-silicate glass containing several effectively insoluble,
mineral forms. On the surface of this insoluble substrate, however,
there exists a thin layer (50-300 1) of readily soluble material
which is rich in trace metals and which contains essentially all of
the particulate sulfur in the form of metal sulfates.
It seems highly probable that the soluble sulfate layer pre-
sent on the surface of coal fly ash particles is formed by gas-to-
particle conversion of sulfur species involving adsorption and/or
condensation processes. Certainly, the necessary increase in speci-
fic concentration of sulfur with decreasing particle size is observed,
although agreement with theoretically predicted size dependences is
poor. Essentially nothing is known about the actual species which
are involved in gas-to-particle conversion.
Due to their potential toxicity it is important to identify and
quantitate the mechanism(s) of formation of particulate sulfate salts.
ACKNOWLEDGMENTS
Aspects of the work referred to herein were supported in part
by grants ERT-74-24276, MPS-74-05745 and DMR-73-03026 from the United
States National Science Foundation; by grant R-803950-03 from the
United States Environmental Protection Agency, Environmental Research
Laboratory-Duluth; and by grant EE-77-S-02-4347 from the United States
Department of Energy.
162
-------
REFERENCES
1. Levy, A., E. L. Merryman, and W. T. Reid. Environ. Sci. Tech-
nol. 4:653, 1970.
2. Davison, R. L., D. P. S. Natusch, J. R. Wallace, and C. A. Evans,
Jr. Environ. Sci. Technol., 8:1107, 1974.
3. Kaarkinen, J. W., R. M. Jorden, M. H. Lawasani, and R. E. West.
Environ. Sci. Technol., 9:862, 1975.
4. Raask, E. J. Inst. Fuel, 41:339, 1968.
5. McCrone, W. C., and J. G. Delly. The Particle Atlas. Ann Arbor
Science Publishers, Ann Arbor, Michigan, 1973.
6. Bickelhaupt, R. E. J. Air Pollution Control Assoc., 25:18,
1975.
7. Natusch, D. F. S. Proc. 2nd Federal Conference on the Great
Lakes, Public Information Office of the Great Lakes Basin Com-
mission, Ann Arbor, Michigan, 1976. 114 pp.
8. Natusch, D. F. S., C. F. Bauer, H. Matuslewicz, C. A. Evans,
Jr., J. Baker, A. Loh, R. W. Linton, and P. K. Hopke. Proc.
International Conference on Heavy Metals in the Environment,
Toronto, Ontario, Canada, Vol. II, Part 2, 1975. 553 pp.
9. Natusch, D. F. S., and C. F. Bauer. Unpublished results,
1978.
10. Linton, R. W., P. Williams, C. A. Evans, Jr., and D. F. S.
Natusch. Anal. Chem., 49:1514, 1977.
11. Keyser, T. R., D. F. S. Natusch, C. A. Evans, Jr., and R. W.
Linton. Environ. Sci. Technol., in press, 1978.
12. Natusch, D. F. S., and M. A. Tompkins. Unpublished results,
1978.
163
-------
Sulfur and Trace Metal Paniculate Emissions
from Combustion Sources
Roy L. Bennett
Kenneth T. Knapp
U.S. Environmental Protection Agency
ABSTRACT
Oil-fired and coal-fired power plant particulate emis-
sions have been physically and chemically character-
ized. The concentrations of sulfur and 27 other ele-
ments in the samples were determined by high resolution
wavelength dispersive X-ray fluorescence analysis (XRF),
Particle size distributions were measured with in-stack
and extractive cascade impactor arrangements. Elemen-
tal particle size distributions were also determined
by XRF analysis of sized emissions collected on the
impactor stages.
Oil-fired emissions were collected for analysis during
eight separate field studies. Three of the studies
were conducted at the same plant at three different
times. The plant was burning high sulfur oil but with
different vanadium contents. The studies conducted at
the other five plants included a range in boiler size,
in sulfur content in fuel, in trace metal content in
fuel, and in level of excess oxygen in the boiler.
Emission samples were collected from five coal-fired
power plants, one of which was equipped with flue gas
desulfurization (FGD) facilities. Two of the plants
were consuming eastern interior coal of high sulfur
(3%-5%) content. Two plants were burning a fairly high
ash, subbituminous western coal with less than 7.0% sul-
fur. Emissions from the fifth plant, which was
165
-------
equipped with FGD units, were collected from one of the
scrubber modules. Sulfur was the most abundant ele-
ment determined in the particulate emission from the
scrubber; reported as sulfate, it was 58%-70% of the
total analyzed mass. Analysis of all significant ex-
pected companion cations accounted for only about 75%
of the total sulfate, which suggests the remainder was
sulfuric acid.
INTRODUCTION
The primary objective of a series of investigations conducted
by the Particulate Emissions Research Section has been the chemi-
cal and physical characterization of particulate emission from
oil-fired and coal-fired power plants. Additional objectives of
the field studies at these combustion sources have been to eval-
uate new monitoring instruments and to evaluate sampling tech-
niques used for stationary source testing. These investigations
have included field studies at six oil-fired power plants and
five coal-fired plants. Emphasis is placed on coal-fired plants
in this paper since portions of the oil-fired studies have been
previously described (1)(2).
MEASUREMENT PROCEDURES
At each site particulate samples were collected on various
filter substrates for chemical analysis of the emissions. Par-
ticle sizing samples were collected with cascade impactors and
small cyclone samplers. Total particulate mass emission tests
were made by Method 5 type procedures. Details of these collec-
tion procedures and the analytical methods used for chemical
characterization of the samples are presented elsewhere in these
proceedings (3). Most of the chemical analyses of the emission
samples were by X-ray fluorescence (XRF) spectrometry, so only
the elemental composition, not the chemical compounds present, is
determined. For example, results are reported for particulate
sulfur only as the amount of elemental sulfur. If the sulfur is
present as sulfate, the mass of the sulfate would be three times
the reported values.
OIL-FIRED POWER PLANTS
Site Descriptions
Characterization studies were conducted at six oil-fired
166
-------
power plants varying in generating capacity, burning fuel with a
wide range of sulfur and trace metal contents, operating at
different levels of excess oxygen in the boiler, and utilizing
fuel additives at various levels. The detailed description of
these plants and the operating conditions during the sample
collection periods have been previously reported (1). Table 1
is a summary of the important operating parameters of the oil-
fired plants and shows the wide variety of conditions which were
examined. All sites were operating without emission control
devices.
Three sampling trips (Al, A2, and A3) were made to one modern
plant that was designed to operate at extremely low excess oxygen
levels in the boiler. The fuel being burned had high sulfur con-
tent (2.4%) on each occasion, but the vanadium content was
different (192 ppm, 593 ppm, and 292 ppm).
Table 1. Operating Conditions at Oil-Fired Power Plants
During Testing Periods
Power load
Site (MW)
Al
A2
A3
B3
J
M
W
N
300-560
525
525
127,127,240
50-60
150-190
600
95
Excess Oxygen
(vol. %)
0
0
0
1
1
3
0
1
.2-1
.1-0
.2-1
.7,2
.5-1
.0-6
.8-2
.5-2
.2
.7
.1
.0,0.6
.7
.0
.3
.8
Fuel Content
V (ppm)
192
593
292
192
111
15
447
113
Ni (ppm) S (%)
16
68
50
16
14
21
62
67
2
2
2
2
1
1
2
1
.43
.40
.40
.43
.42
.23
.15
.74
Q
Three units, each tested twice.
167
-------
Summary of Results of Oil-Fired Power Plants
Details of some of the earlier results obtained at oil-fired
power plants have been previously reported (1)(2); therefore,
only the more important general observations are reported here.
Of the elements analyzed by XRF , sulfur, vanadium, and nickel
were the most predominant. The emission concentrations of vana-
dium and nickel were directly related to the fuel content of these
metals. The particulate sulfur emission levels were more complex,
as the concentrations were found to be related to the fuel sulfur
content, boiler excess oxygen levels, and the fuel vanadium con-
tent. At sites where sampling was made at several excess boiler
oxygen levels, the particulate sulfur increased with increased
excess oxygen. This is presumably due to greater oxidation of the
sulfur to sulfate forms. Carbon content increased as the excess
oxygen decreased. At Site Al carbon was as high as 70% of the
total mass at excess oxygen levels of 0.2%. Particle size deter-
minations with both in-stack and extractive cascade impactor
arrangements indicated that at all sites most of the particles
are less than 0.4 /urn in diameter. At Site Al , during operation
at low excess oxygen, a second mode around 10 /um was present.
These larger particles were highly carbonaceous. Elemental anal-
ysis of sized fractions revealed that sulfur was present in the
larger size fractions to a greater extent than nickel or vanadium.
Most of the nickel and vanadium were associated principally with
smaller particles (mass median diameters less than 0.4 //m) . Iron,
which often originated from corrosion to a greater extent than
from the fuel, was associated with the larger particles (mass
median diameter about 3
COAL-FIRED POWER PLANTS
Site Descriptions
Sampling was conducted at five coal-fired power plants that
were consuming fuel of widely varying sulfur and mineral concen-
trations. Four plants had electrostatic precipitator (ESP) con-
trol systems. Three of these were sampled from ports in the
breeching between the ESP units and the stack; the fourth was
sampled from ports in the stack. The control system at the fifth
plant consisted of two scrubbers in series, a Venturi scrubber and
a flue gas desulfurization (FGD) scrubber. The emissions samples
were collected after a FGD unit at this plant. Some of the plants
were sampled at various operating conditions, including different
excess oxygen levels in the boiler and a variation in the number
of ESP fields operating in order to change the particulate emis-
sion loading.
168
-------
Brief descriptions of the five coal-fired plants and operat-
ing conditions during the sampling periods follow. Table 2 lists
the important operating conditions at each site.
Table 2. Operating Conditions at Coal-Fired Power Plants
During Testing Periods
Site
P
L
SC
CB
K
Power
(Ml)
100
330
520
88
720a
Coal
Type
Eastern Interior
Eastern Interior
Wyoming
Wyoming
Western Interior
% S
3.3
3.9
0.7
0.7
4.8
% Ash
8
14
12
12
29
Control
System
ESP
ESP
ESP
ESP
Venturi Scrul
Sub-bituminous
followed by FGD
Scrubber
Emissions were sampled from the FGD unit operating at 115 MW.
Site P—This facility was an old, small coal-fired power
plant that was burning an eastern interior, high-sulfur coal.
The unit sampled had a normal load rating of 100 MW. Sampling
ports were located on the outlet breeching of the ESP serving the
boiler. The particulate collection efficiency of the ESP was
lowered by reducing the number of ESP fields operating.
Site L—This plant had a normal load of 300 MW, was burning
a high-sulfur eastern interior coal, and had an ESP that was
operated at several efficiency levels during the tests. Samples
were taken from ports in the breeching between the ESP and the
stack.
Site SC—The unit sampled at this plant had a nominal gross
generating output of 520 MW. A very low-sulfur coal (0.7%) that
was mined in Wyoming was burned. Particulate emissions were con-
trolled by 32 ESP units. Loading was varied at times by cutting
out 8 or 12 units. Sampling ports were located on the stack.
169
-------
Site CB—This was a small unit, 88 MW nominal output, that
was controlled by an ESP. It was burning a Wyoming-mined coal
with a nominal sulfur content of less than 0.7%. Particulate
loadings were reduced by cutting off some of the ESP fields.
Sampling ports were located between the ESP and the stack.
Site K—This plant operated at loads from 300 to 720 MW
during the testing periods. Particulate emissions were controlled
by eight parallel dual units, a Venturi scrubber and a FGD scrub-
ber in series. Sampling was conducted after the newest and most
modern of the FGD units. The unit was operating at a 115 MW load
during all of the tests. The boiler was burning a very high-
sulfur (4.8%), high-ash western interior sub-bituminous coal.
Particulate Emissions
Total particulate emissions concentrations (milligrams per
normal cubic meter, mg/Nm3), as determined by Method 5 procedures,
are listed for Sites P, L, SC, CB, and K in Tables 3 through 7.
For the first four sites, which had ESP control systems, the emis-
sion values are given for full ESP operation and conditions with
some ESP units off. Although cutting off the first one or two
ESP fields often did not greatly change the particulate loadings,
sufficient units were cut off to obtain significant increases in
loading at all but Site CB. The elemental data show drastic com-
positional changes as the particulate loading is changed by
reducing the electrostatic precipitation efficiency.
Elemental Composition
The composition of the particulate emissions from the two
high-sulfur ESP controlled plants were analyzed for 28 elements
by XRF. The results are listed in Table 8 for Site P and in
Table 9 for Site L. These results are presented in percentage of
the total analyzed material which does not include elements with
atomic numbers less than that of fluorine (9). The composition
changes drastically with electrostatic precipitation. For
example, at Site P the sulfur was about 70% of the total at nor-
mal (full ESP) conditions but only about 5% with two ESP units
out. Iron was 8%-12% with normal ESP control but increased to
45%-55% with two units out. The same extreme changes were ob-
served at Site L. Sulfur dropped from 75% of the total analyzed
material in samples which passed the fully operating ESP units to
less than 6% when ESP units were cut out. Again, the percentage
of iron was less in samples that were obtained with the ESP units
operating at normal levels.
170
-------
Table 3. Total Mass and Elemental Particulate Emission
Concentrations - Site P
Concentrations
(mg/Nm3)
ESP Fields Out Total Mass
0 63
72
174
147
108
105
Average: 111
2 569
422
424
484
Average: 475
3 1,830
S
6.1
7.7
7.1
8.4
13.
6.9
3.5
8.3
11.0
2.8
5.5
6.5
2.2
3.0
3.8
3.2
5.3
9.3
5.4
V
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.35
0.19
0.16
0.70
0.13
0.07
0.39
0.22
0.09
0.35
0.58
0.76
0.29
Fe
(7.7)*
0.83
0.59
0.89
0.99
0.82
64.
39.
33.
46.
27.
22.
67.
42.
27.
64.
117.
160.
59.
*Sample was probably contaminated with corrosion products and the
value was not included in the average.
171
-------
Table 4. Total Mass and Elemental Particulate Emission
Concentrations - Site L
Concentrations
(mg/Nm3 )
ESP Fields Out Total Mass
0 107
291
461
Average: 286
1 313
Average: 313
2 365
461
900
950
Average: 740
3 3,400
2,100
S
11
14
15
13
16
16
16
13
18
15
5.2
12
17
14
—
—
V
.45
.22
—
.22
.93
.64
.79
.80
1.8
1.0
1.2
.82
1.0
1.1
—
—
Fe
2.9
1.2
1.2
1.8
50.
37.
44.
63.
186.
82.
133.
71.
81.
103.
—
—
Average: 2,700
172
-------
Table 5. Total Mass and Elemental Particulate Emission
Concentrations - Site CB
Concentrations
(mg/Nm3)
ESP Fields Out Total Mass S
0 176 0.65
0.96
0.72
98 0.63
0.56
0.62
0.58
66 0.64
0.76
0.69
78 0.69
0.73
0.67
Average: 105 0.68
1 83 -0.93
0.92
1.07
89 0.82
0.87
0.85
V
0.03
0.04
0.02
0.02
0.02
0.02
0.03
0.02
0.03
0.02
0.01
0.01
0.01
0.02
0.02
0.01
0.01
0.01
Fe
2.2
3.1
2.0
2.5
2.3
2.0
2.0
3.6
2.8
2.8
2.9
2.5
2.8
2.6
2.7
2.8
2.9
3.0
2.6
2.2
Average: 86 0.91 0.01 2.7
173
-------
Table 6. Total Mass and Elemental Particulate Emission
Concentrations - Site SC
Concentrations
(mg/Nm3)
ESP Fields Out Total Mass S
0 15.3 0.04
0.05
0.07
22.0 0.25
0.21
0.23
Average: 18.7 0.14
8 23.9 0.33
0.37
0.34
0.34
Average: 23.9 0.34
12 88.1 0.74
0.98
1.17
0.90
1.03
1.18
1.24
1.49
Average: 88.1 1.09
V
—
0.008
0.010
—
0.012
0.011
0.010
0.013
0.011
0.047
0.046
0.057
0.057
0.056
0.059
0.055
0.057
0.054
Fe
0.06
0.06
0.09
0.58
0.48
0.36
0.27
0.67
0.70
0.52
0.52
0.60
3.2
3.4
4.0
3.8
3.8
4.0
4.0
5.4
3.9
174
-------
Table 7. Total Mass and Elemental Particulate Emission
Concentrations - Site K FGD Scrubber
Concentrations
(mg/Nm3)
Total Mass
137
123
89
193
175
Average: 143
S
21.1
23.4
19.6
32.3
30.2
25.3
Zn
8.1
5.8
4.2
9.4
12.9
8.1
Fe
10.2
7.7
8.9
13.8
13.7
10.9
No. of Tests
8
6
3
3
3
175
-------
Table 8. Elemental Composition of Particulate Emissions
Site P
(Percent of Total Amount Analytically Determined)
ESPs Out:
Load :
Excess 0 2 :
Element*
F
Na
Mg
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Cd
Sn
Sb
Ba
Hg
Pb
0
Full
Normal
0.22
5.0
5.8
0.03
70.
0.25
0.49
0.56
4.3
1.4
8.2
0.18
0.17
0.42
0.06
2.6
0
Full
High
0.24
6.5
7.8
69.
0.62
0.39
0.76
2.0
0.06
12.
0.09
0.09
0.49
2
Full
Normal
0.07
0.40
14.
16.
0.12
5.7
3.0
2.0
2.7
0.28
1.3
0.15
55.
0.16
0.10
0.04
0.45
0.10
0.01
0.01
0.26
0.45
2
Half
Normal
0.52
19.
22.
0.13
6.7
3.2
1.5
2.4
0.17
1.2
0.14
45.
0.09
0.08
0.29
0.07
0.01
0.18
0.39
Samples were analyzed for all elements; blank indicates content
was below detection limits (3).
176
-------
Table 9. Elemental Composition of Particulate Emissions
Site L
(Percent of Total Amount Analytically Determined)
ESP Units Out:
Element*
F
Na
Mg
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Cd
Sn
Sb
Ba
Hg
Pb
5.6
8.4
0.33
75.
0.84
2.6
0.61
6.2
0.11
0.11
0.07
0.57
8.2
11.
0.25
5.7
3.1
5.0
1.4
0.34
0.61
0.19
18.
0.03
0.03
0.06
0.35
0.01
0.07
0.13
0.31
17.
24.
0.23
5.9
6.1
8.1
2.2
0.38
1.5
0.26
31.
0.06
0.05
0.08
0.52
0.01
0.10
0.19
0.46
*Samples were analyzed for all elements; blank indicates content
was below detection limits (3).
177
-------
Elemental Emission Concentrations
The particulate concentration (mg/Nm3) of the elements sulfur,
vanadium, and iron in the emissions from plants controlled by ESP
are listed in Tables 3 through 6. The concentrations of sulfur,
iron, and zinc in the particulate emission from the FGD scrubber at
Site K are given in Table 7. For the two plants burning high-sulfur
coal and controlled by ESP units, Plants P and L, the sulfur con-
centration values did not exhibit the extreme increase that was
observed with iron and vanadium when the ESP units were cut off.
For example, the sulfur concentration at Site L (Table 4) was
unchanged, within experimental error, whereas the iron concentra-
tion was about 60-fold lower at full ESP operation compared to two
units out. Similar results were observed at Site P (Table 3). The
sulfur concentrations at Sites SC and CB were very low, principally
due to the low sulfur coal being burned. In contrast to the results
at the plants burning high-sulfur coal, a significant reduction in
the particulate sulfur concentration with an increase in the number
of ESP units in operation was observed at Site SC (Table 6). The
sulfur concentration emitted from the FGD scrubber at Site K, about
20-30 mg/Nm3, was the highest of the five sites tested. In the
samples from this site, insufficient cations were found to account
for all the sulfur when it is assumed to be all sulfate. The
results indicate that when all cations are considered to be as
sulfate, which is not necessarily the case, 25% of the sulfate is
unaccounted for and is probably present as free sulfuric acid. This
suggests at least 15 mg/Nm3 sulfuric acid was being emitted.
Particle Size Distribution
The effect of electrostatic precipitation on the particle size
distribution at Sites P and L is shown by the in-stack cascade
impactor data in Tables 10 and 11. At Site P the apparent mass
median diameter (MMD) when all precipitators were operating was
very small. The data from the two in-stack impactor runs indicate
the MMD was less than 0.3 ^m, as more than 50% of the mass was on
the back-up filters. The data from the extractive impactor indicate
values of about 0.45 and 0.9 /um for MMD of two runs. This small
particle size does not agree with the size distribution data pre-
viously reported for other ESP controlled coal-fired plants or with
data based on concurrent transmissometer measurements (slope of
extinction coefficient versus particulate mass concentration) which
indicate that the particles were much larger (4). With two ESP
units out at Site P, the MMD was large, about 5-10 jum. At Site L
the very fine particles with full ESP operation were not observed.
178
-------
Table 10. Effect of Electrostatic Precipitation on
Particle Size Distribution
Site P
Stage
1
2
3
4
5
6
7
Filter
No. of Tests: 1
Conditions:
ESPs Out: 0
Load: Full
Excess 02 : Normal
1
0
Full
High
5
2
Full
Normal
DSO
(/um^ (milligrams/normal cubic
23 0.3
10 0.3
4.7 0.8
1.9 0.8
1.0 6.7
0.52 6.4
0.27 7.6
13.0
1.4
1.3
3.0
4.3
1.2
0.9
0.9
12.7
1770.
1250.
770.
167.
56.
37.1
23.4
134.5
1
2
Half
Normal
meter)
61.8
31.1
98.6
29.6
8.8
1.3
0.6
5.4
179
-------
Table 11. Effect of Electrostatic Precipitation on
Particle Size Distribution
Site L
No. of Tests: 111 11
Conditions:
ESPs Out: 012 23
Excess 02: Normal Normal Normal High Normal
Stage
1
2
3
4
5
6
7
Filter
(/urn)
23
10
4.7
1.9
1.0
0.52
0.27
(milligrams/normal cubic meter)
1.1
1.5
4.4
5.9
3.2
1.2
0.9
9.3
7.2
5.3
18.6
11.7
5.0
1.4
1.0
27.1
280.
295.
84.
40.4
11.1
2.5
1.7
15.8
302.
148.
89.
36.
13.5
3.4
1.3
12.8
2820.
1340-
600.
220.
44.2
12.6
15.1
109.
180
-------
The particles in the emissions from the FGD unit at Site K had
a MMD less than 0.4
Summary of Results of Coal-Fired Power Plants
A summary of the total mass and the sulfur, iron, vanadium, and
zinc elemental concentrations for the five coal-fired power plants
at normally controlled conditions is presented in Table 12. The
sites (P and L) burning high-sulfur coal emitted fairly high parti-
culate sulfate concentrations, while Sites SC and CB, which were
consuming low-sulfur coal, as expected had low particulate sulfur
values. At the plants burning high-sulfur coal, the sulfur con-
centration, in contrast to that of the trace metals such as iron
and vanadium, did not change greatly with changes in the number of
ESP fields that were operating. These results indicate that the
ESP controlled emissions have a higher percent particulate sulfur
than uncontrolled emissions. Most of the particulate sulfur emis-
sions from these two plants are apparently in a form that is not
efficiently removed by the ESP units. These forms might be fine
particles or vapor of sulfuric acid which subsequently condenses or
is adsorbed after the material passes through the ESP units.
The highest sulfur concentrations were found at Site K which
was sampled after the FGD scrubber. Iron and zinc concentration
were also high at this site.
ACKNOWLEDGMENTS
The authors acknowledge the contributions of Ray Steward
and Robert Griffin, U.S. EPA, and the sampling team of Engineering
Sciences for the characterization sample collection; Robert Kellogg
and John Lang, Northrop Services, Incorporated, for X-ray fluores-
cence analyses; and William Henry, Battelle-Columbus Laboratories,
for fuel analyses.
181
-------
Table 12. Summary of Emission Concentrations at
Coal-Fired Power Plants
Site
P
L
SC
CB
K
Power
(MW)
100
330
520
88
115
Concentrations
(mg/Nm3)
Total Mass
111
195
19
104
143
S
6.9
13.0
0.14
0.68
25.3
V
<0.01
0.33
0.03
0.23
L
Fe
0.8
1.8
2.7
2.6
10.9
Zn
—
—
—
8.1
182
-------
REFERENCES
1. Bennett, R. L., and K. T. Knapp. Chemical Characterization of
Particulate Emissions from Oil-Fired Power Plants. In: Proceed-
ings of the Fourth National Conference on Energy and the Environ-
ment, AICHE, Dayton, Ohio, 1976. pp. 501-506.
2. Knapp, K. T., W. D. Conner, and R. L. Bennett. Physical Charac-
terization of Particulate Emissions from Oil-Fired Power Plants.
In: Proceedings of the Fourth National Conference on Energy
and the Environment, AICHE, Dayton, Ohio, 1976. pp. 495-500.
3. Knapp, K. T., R. L. Bennett, R. J. Griffin, and R. C. Steward.
Collection Methods for Particulate Sulfur and Other Chemical
Determination. In: Measurement Technology and Characterization
of Primary Sulfur Oxides Emission from Combustion Sources,
Southern Pines, North Carolina, 1978.
4. Conner, W. D- ESRL, EPA, Personal Communication, April 1, 1978.
183
-------
Inorganic Compounds Present in Fossil Fuel Fly
Ash Emissions
William M. Henry
Ralph I. Mitchell
Battelle-Columbus Laboratories
Kenneth T. Knapp
U. S. Environmental Protection Agency
ABSTRACT
Vast tonnages of fly ash are emitted from coal-fired and
oil-fired power plants. Based on elemental analyses,
these emissions contain hazardous substances, but full
assessments of the health hazards they produce require
knowledge of the chemical forms of the largely inorganic
emissions.
X-ray diffraction, (subtractive) Fourier Transform In-
frared, chemical phase analyses, coupled with extensive
elemental determinations and limited equilibrium thermo-
dynamic calculations were applied to a group of samples
collected from oil-fired and coal-fired power plant stacks
to provide information on the chemical structure of fly
ash emissions.
Investigations produced considerable data important in
toxicity evaluations. Notably, oil-fired fly ashes con-
tain a highly water-soluble phase, with sulfate as the
principal anion component. Vanadium is present largely
as a water-soluble vanadium (IV) oxysulfate. Valence
state analyses show direct correlations of V (IV) with
water solubilities.
Due to the presence of a high concentration insoluble
iron aluminum silicate glass phase, water-soluble sul-
fate compounds are less prominent in coal-fired fly ashes,
with a range of up to 30%. Metal oxides also are present
in both types of fly ash, but they are not the principal
constituents on which toxicity evaluations should be based
185
-------
INTRODUCTION
The literature abounds with analytical methodology descriptions
and applications to inorganic particulate pollutant analyses (1-13).
The great majority of these publications and references describe
and/or are applied to the elemental and anionic contents of pol-
lutant samples.
Comparatively, methods applicable to inorganic compound or
chemical form identification and analysis are few, and descriptions
of these applied to pollutant samples are quite limited in the lit-
erature and in ongoing research and development activities. This
lack of attention given to inorganic compound identification in pol-
lutants is unusual in view of frequently declared needs for such
information in health and toxicity assessment studies, control meas-
ures, and disposal means. Several reasons can be cited for this
anomaly, but a principal cause appears to be the relative difficulty
of inorganic compound identification of samples as complex and heter-
ogeneous, as are pollutant emission particulates. The commonly and
readily used techniques for analysis of inorganic constituents con-
sist of initially breaking samples down to their ionic forms and/or
utilizing the atomic characteristics of the samples' constituents
and then isolating individual elements, cations, or anions, chemical-
ly or spectrally for identification and quantification. This is in
contrast with the more commonly used organic species analysis methods
such as infrared and mass spectrometry which are based largely on
identifications of molecular fragments and thus are relatable more
directly to elucidation of organic compound constituents. These,
of course, are generalizations; since with selected sample dissolu-
tion the valence state of certain elements can be retained and quan-
tified, and certain inorganic species such as alpha quartz, asbestos,
etc., have distinctive molecular spectral characteristics and/or
specific crystalline forms. However, the use of compound specific
techniques for inorganic species identification has not been exploited
to any great degree on complex pollutant emission samples. Inorganic
compound identification and analyses of pollutant emission samples,
what little has been done, has relied mostly on XRD techniques plus
morphological characterization of sample by component recognition,
using the microscopy-instrumented tools of SEM, STEM, and EMP where-
in microscopic viewing can be aided by elemental analyses of the
viewed particle or particle groupings. More recently the surface
identification techniques of ESCA, Auger, SIMS, etc., have been
applied to pollutant particulates, but these techniques are diffi-
cult to standardize and to interpret derived data.
More surprising than the dearth of information on the inorganic
compound structures of particulates emitted from sources using
186
-------
fossil fuels are the comparatively sparse data available even on
the elemental and anionic contents of oil-fired fly ash emissions.
Oil-fired power plants still are a major source of electrical energy,
especially in the eastern and southern regions of the U.S. Most
frequently these plants are operated with little or no emission
control equipment. As a result, although fuel oil thermal ash con-
tents range only from 0.05% to 0.2%, the particulate emission rates
from oil-fired plants can be high even as compared to rates from
the much higher ash content coal-fired sources which are operated
with more rigorous control measures. However, literature sources
reveal very little concerning even the elemental compositions of
particulate emissions from these sources.
Due to the high ash contents of coal fuel sources and the con-
sequent need to find disposal means and/or alternate usages for the
high tonnages of collected emission particulates, and to concern
over the potential health hazards consequences of these measures,
the chemical and physical natures and elemental contents of fly
ashes from coal-fired sources have been studied in more detail than
have those from oil-fired sources. However, until recently even
these studies have not focused primarily on methodologies to deter-
mine the inorganic compound forms present in fossil fuel particu-
late emissions. It should be pointed out that both oil-fired and
coal-fired fly ashes contain <0.1% organics, as based on MeCl2 ex-
tractions, so the particulates emitted as fly ashes are primarily
inorganic in form.
In summary, in view of the vast tonnages of fly ashes emitted
from oil-fired and coal-fired power plants and other sources using
or processing fossil fuels and the increased concerns of the poten-
tial health hazards of these due to the use of additional and dirtier
fossil fuel sources, there exists a need for methodologies to iden-
tify and quantify the chemical structure of the emissions which are
largely inorganic. The work described herein was performed to help
fill the information gap caused primarily by lack of existing method-
ology. The work is ongoing and has been directed mostly toward
oil-fired fuel emissions since these are the least known chemically
and may pose the greatest health hazards.
EXPERIMENTAL
The techniques considered most applicable to complex pollutant
particulate samples are listed in Table 1.
The first three techniques listed in Table 1, coupled with de-
tailed cation and anion determinations and limited equilibrium ther-
187
-------
modynamic evaluations of the chemical data, provided most of the
data presently obtained on the inorganic constitution of oil-fired
and coal-fired fly ash samples. ;
Table 1. Analytical Techniques Applicable to
Inorganic Compound Identification
Techniques
Comments
X-ray Diffraction
Chemical Phase
Fourier Transform Infrared
Thermal
Microscopy - optical, petro-
graphic, chemical, scanning
electron, scanning electron
transmission, and electron
microprobe
Surface Methods
Mass Spectrometry
X-ray Emission Spectrometry
Components must have crystalline
structures.
Valence state measurements and
component separations and analyses.
Application to inorganic compounds
is a new development.
TGA, DTA, DSCA, calorimetry
Except for the petrographic and
chemical methods, the identifica-
tion of compounds is by elemental
association within particles.
ESCA, Auger, SIMS, IMA
Knudsen cell and possibly high
resolution mass spectrography.
Wavelength peak shifts due to
chemical bonds show various oxi-
dation states.
The experimental efforts were directed principally toward the
investigation and use of techniques better established in their
development and application to inorganic compound identification
and quantitation than are techniques such as ESCA, SIMS, and IMA.
This selection was guided by the need to fill the large data gap
existing as to the total chemical constitution of fossil fuel par-
ticulate emissions. An exception to this technique selection pro-
cess was the exploration and use of FT-IR for inorganic compound
identification.
188
-------
Field Sample Collections
Samples of oil-fired and coal-fired fly ashes were collected
from several power plant sites which burn fossil fuels of various
origins with the objective of obtaining a range of fly ash sample
compositions which would be representative of present power produc-
tion processes. Sampling was performed at the port holes in the
stacks beyond any emission control process operation. The fly ash
samples were obtained by simply inserting a 2-cm-diameter glass-
lined probe into the center of the stack perpendicular to the stack
stream flow and, with a 1-hp blower, drawing a portion of the flow
into a fine mesh Teflon bag. A 24-hour sampling time period usually
provided 50 to 75 grams of stack emission particulates. At the con-
clusion of the sampling period the Teflon bag was removed from the
Hi-Vol container, sealed in a polyethylene bag, and returned to the
laboratory for analyses of the collected particulates. No efforts
were made to relate the samples with combustion conditions, to
sample isokinetically, or to separate aqueous phase emissions from
the particulates. The objective of the sampling was to obtain
relatively large masses of samples for purposes of methodology devel-
opment with the expectation that the developed methodologies can be
applied to more carefully planned sampling efforts at a later time.
Sample pretreatment was considered in carrying out the analyses
of fossil fuel particulate emissions samples since unknown alter-
ations of their chemical forms must be avoided. Samples collected
in the way described from stack exit flues at temperatures of about
150°C do contain a lot of moisture, and pretreatments such as desic-
cation and heating can alter the sample weight and chemical forms.
From structural, crystallographic and/or optical—XRD, IR, petro-
graphy—analytical aspects, it is desirable to work with samples in
a stable, moisture-free condition, since the presence of loose and
even bound forms of water complexes the identification efforts. A
common practice of drying samples at 105°C before bottling, weighing,
and analysis is not applicable to the wet particulate emissions since
for many samples there is no point where loose, unbound, capillary
water only is removed by heating in air atmosphere. This is illus-
trated by the data given in Table 2 for samples collected at the
stack exit ports of coal-fired and oil-fired power plants. Thermo-
grams of a composite of four oil fly ash samples (equal amounts of
each mixed together) heated slowly at 1°C per minute in air and in
argon are shown in Figure 1. The thermogram for the oil-fired fly
ash composite heated in air shows a continuous weight loss over a
15-hour heating increase at a 1°C per minute change. The sequence
of weight losses, as shown by individual sample TGA and DTA plots
in air, indicates capillary or unbound water, hydrated or bound
water, carbon, and the partial S04 losses.
189
-------
CD
O
0
0
tOO 200 300
400 500
Temperature, °~
600 7QO 80° 9''0
-------
Table 2. Weight Losses of Fly Ash Samples on
Slow Heating in Air (in Percent)
105°C
200°C
400°C
750°C
Oil Fly Ash No. 1
Oil Fly Ash No. 2
Oil Fly Ash No. 4
Oil Fly Ash No. 5
Coal Fly Ash NBS
Coal Fly Ash No. 1
Coal Fly Ash No. 2
Coal Fly Ash No. 3
2.4
3.0
4.5
5.05
0.25
1.0
4.0
4.0
4.5
4.8
12.5
10.6
0.55
1.8
5.4
6.5
18.0
69.5
28.0
36.9
1.1
2.6
13.0
9.0
22.5
74.0
57.0
45.5
4.1
4.7
19.2
24.2
Thermograms based on heating the samples under argon show minor
incremental weight changes between 200°C and 400°C, as illustrated
by the composite sample in Figure 1, indicating probable loss of most
unbound water contents. IR and XRD spectral and pattern images ob-
tained on the samples after heating under argon are much improved,
as are the microscopic appearances of viewed sample particle fields.
Based on these findings, heating the samples under argon appears
to be a reasonably satisfactory mode of removing the unbound water
without altering otherwise the integrity of the sample structure.
Based on individual thermograms for each sample, heating samples at
350°C under argon was adapted as the preparation mode for IR, XRD,
and microscopic examinations. Other determinations were carried out
on air-dried samples.
Six oil-fired and four coal-fired fly ash samples have been
used to date for the methodology development work. The NBS Standard
Reference Material Coal Fly Ash is actually a group of precipitator
and mechanically collected ashes which have been sieved and blended
prior to standardization. It is planned to add two Western coal-
fired fly ashes to the experimental samples. The analyses of the
fuels being burned at the times of sample collections are given in
Table 3.
ELEMENTAL ANALYSES
Inorganic compound identification can be aided considerably by
knowledge of the elemental constituents of samples so complete
analyses of the work samples were obtained as shown in Tables 4 and
5. Significant data to note in Table 4 are:
191
-------
Table 3. Analyses of Fuels Used During Collections of
the Fly Ash Samples3 - Results in PPM
Except Where Percent Is Given
Fuel Oils
S
V
Ni
Fe
Mg
Al
Si
Ca
Na
K
Ash at 550°C
No. 2
2.5%
540
69
5
139
2
5
10
20
4
0.18%
No. 4
2.15%
446
62
45
6
2
3
5
10
4
0.10%
No. 5
2.65%
292
50
17
114
1
4
7
159
7
0.14%
No. 6
1.56%
40
20
15
1490
60
10
5
8
6
0.15%
Coals
No. 2
3.87%
—
—
1.0
0.02%
1.3%
1.0%
0.2%
—
—
8.2%
No. 3
3.62
—
—
0.9
0.1
1.0
1.7
0.3
—
—
14.4%
Q
No fuel oil sample was available for the No. 1 fly ash. Fly ash
No. 1 was on hand from a 1973 research program taken from a power
source purportedly using a domestic origin No. 6 fuel oil.
The No. 3 fuel oil and fly ash samples were taken about 1 month
later from the same power source as the No. 2 and are nearly
identical with those of the No. 2.
No coal sample was available for the No. 1 coal fly ash - the
fly ash was on hand from a previous program.
192
-------
Table 4. Oil-Fired and Coal-Fired Fly Ash Compositions - Major Constituents (Percent)
8
C H N N03
NO" NH+ S04 SO-
S~ Cl P Si
Oil-Fired Fly Ashes
No. 1
No. 2
No. 4
1
No. 5
No. 6
Total Sample Content 12.4 0.9 0.1 0.005
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 69.0 0.7 0.9 0.013
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 21.5 1.0 0.9 0.02
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 1,5 1.2 0.1 0.02
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 14.5 2.4 6.5 <0.01
Water-Soluble Content
Water-Insoluble Content
<0.01 0.012 36.9 <0.01
36.0
0.9
0.005 0.13 12.0
12.0
0.15
0.01 0.81. 41.2
41.1
0.1
<0.01 0.16 57.6
58.6
0
0.03 7.3 49.2
48.4 "
0.8
<0.01 0.05 0.008 0.31
<0.01
0.31
0.02 0.002 0.2
<0.01
" 0.2
0.02 0.004 0.2
<0.01
0.2
0.05 0.001 0.05
" <0.01
_ " 0.05
0.06 0.05 0.22
<0.01
0.22
Coal-Fired Fly Ashes
NBS
No. 1
No. 2
No. 3
SRM 1633 Total Sample 3.3 0.1 <0.1 <0.01
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 1.7 0.3 <0.1 <0.01
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 7.0 0.5 0.1 <0.01
Water-Soluble Content
Water-Insoluble Content
Total Sample Content 0.5 0.7 <0.1 0.02
Water-Soluble Content
Water-Insoluble Content
<0.01 <0.01 0.98
0.60
-CO .01
-------
Table 4. (Continued)
Al
Fe
Ni
V
Mg
Ca
Na
K
Total
Organics
Water
Solubility
H2°
H2S04
pH
CD
Oil-Fired Fly Ashes
No. 1 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
No. 2 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
No. 4 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
No. 5 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
No. 6 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
Coal-Fired Fly Ashes
NBS SRM 1633 Total Sample
Water-Soluble Content
Water-Insoluble Content
No. 1 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
No. 2 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
No. 3 Total Sample Content
Water-Soluble Content
Water-Insoluble Content
1.25
0.5
0.75
0.05
0.02
0.03
0.40
0.23
0.17
0.01
<0.01
0.01
1.42
0.27
1.15
0.61
0.30
0.31
0.40
0.25
0.15
0.41
0.20
0.21
0.48
0.49
0
0.40
0.43
0
1.66
1.0
0.66
0.85
0.60
0.25
1.29
1.06
0.23
2.28
2.31
0
0.35
0.30
0.05
2.27
0.50
1.77
6.68
2.23
4.45
10.2
8.98
1.22
12.85
12.9
0
1.10
0.78
0.32
18.4
4.71
13.7
3.41
1.15
2.26
5.94
5.0
0.94
2.60
2.65
0
2.4
2.4
0
1.0
0.6
0.4
0.31
0.15
0.16
0.1
0.07
0.03
0.20
0.19
0.01
0.32
0.16
0.16
3.91
3.90
0.01
0.30
0.30
0
0.50
0.51
0
2.02
2.0
0
0.20
0.21
0
0.13
0.13
0
0.1
0.1
0
0.10
0.12
0
0.10
0.09
0
0.12
0.11
0
<0.1
0.053
<0.1
<0.1
12.7
6.5
0.01 0.02
2.0
4.2
11.3 12.6
10.9 14.1
0.63 0.56
10.27 13.54
8.79 7.90
1.63 1.94
7.16 5.96
0.06
0.02
0.52 1.5
0.30
0.60
1.75
1.54
0.02 0.03
0.01 0.01
0.01 0.02
0.06 0.06
0.04 0.04
0.02 0.02
0.2 0.40 0.05 1.0
0.01 0.18 0.03 0.5
0.01 0.22 0.02 0.5
0.6 3.0 0.08 1.1
0.2 1.8 0.06 0.6
0.4 1.2 0.02 0.5
<0.1
0.04
0.072
0.11
58.0
23.3
72.0
98.5
83.0
3.5
5.3
13.0
34.0
7.0
5.0
4.5
5.5
2.1
0.3
1.0
4.0
5.0
<0.1 3.9
0.2 2.7
0.04 2.42
1.0 2.15
1.5 2.22
<0.1 11.35
<0.1 4.50
2.0 3.17
2.1 2.73
-------
Table 5. Oil-Fired and Coal-Fired Fly Ash Compositions - Trace Constituents
Element
Li
Be
B
F
Sc
Ti
Cr
Mn
Co
Cu
Zn
Ga
Ge
As
Se
Br
Rb
Sr
Y
Zr
Nb
Mo
Ru
Rh
Pd
Ag
Cd
In
Sn
Sb
Te
I
Cs
Ba
La
Ce
59-71
Hf
Ta
W
Re
Os
Ir
Pt
Au
Hg
Tl
Pb
Bi
Th
U
Oil-Flred Fly Ashes
No. 1
0.5
0.05
30.
2.
1.
500.
100.
200.
500.
500.
40.
5.
10.
30.
7-
10.
5.
500.
50.
50.
10.
100.
<1.
<1.
<1.
<1.
1.
<0.5
20.
3.
<0.5
<0.5
<1.
1000.
40.
50.
120.
<0.5
<1.
<1.
<0.2
<0.4
<0.2
0.3
<0.5
<1.
10.
3000.
<0.1
10.
10.
No. 2
0.5
0.3
0.5
<1.
. <1.
300.
500.
200.
50.
100.
40.
5.
1.
20.
5.
3.
1.
100.
3.
•5.
2.
50.
<1.
<1.
<1.
<1.
1.5
<0.5
3.
5.
<0.5
<0.5
<1.
200.
10.
5.
15.
<0.5
<1.
<1.
<0.2
<0.4
<0.2
0.3
<0.5
<1.
<0.2
200.
<0.1
5.
5.
NO. 4
0.2
0.2
5.
3.
5.
400.
1000.
200.
200.
200.
200.
50.
10.
30.
10.
25.
5.
300.
5.
20.
1.
100.
<1.
<1.
<1.
<1.
4.
<0.5
5.
10.
<0.5
<0.5
<2.
200.
50.
25.
70.
<0.5
<1.
2.
<0.2
<0.4
<0.2
<0.3
<0.5
<1.
0.5
400.
<0.1
6.
2.
No. 5
3.
0.3
3.
5.
3.
400.
500.
500.
300.
400.
400.
60.
10.
30.
3.
25.
4.
300.
10.
20.
1.
100.
<1.
<1.
<1.
<1.
3.
<0.5
5.
10.
<0.5
<0.5
3.
1000.
50.
30.
60.
<0.5
<1.
2.
<0.2
<0.4
<0.2
<0.3
<0.5
<1.
1.
400.
0.3
4.
10.
No. 6
100.
0.1
1.
<1.
<1.
700.
450.
300.
300.
400.
200.
40.
7.
20.
7.
2.
7-
200.
10.
10.
2.
150.
<1.
<1.
<1.
<1.
4.
<0.5
3.
150.
<0.5
<0.5
1.
1000.
150.
100.
90.
<0.5
1.
2.
<0.2
<0.4
<0.2
<0.4
<0.5
<1.
0.2
300.
<0.1
2.
2.
Coal-Fired
NBS
300.
5.
100.
10.
20.
6000.
130.
500.
50.
120.
200.
50.
20.
60.
10.
10.
150.
1500
30.
200.
7.
20.
<0.5
<0.5
<1.
<0.5
15.
<0.5
3.
7.
<0.5
<0.5
10.
2500.
70.
125.
90.
10.
2.
5.
<0.2
<0.4
<0.2
0.4
<0.5
0.1
2.
80.
0.7
20.
15.
No. 1
0.1
0.5
100.
20.
10.
2000.
100.
50.
5.
30.
20.
10.
10.
20.
5.
15.
5.
2000.
50.
100.
5.
10.
<0.5
<1.
<1.
<1.
2.
<0.5
3.
1.
<0.5
<0.5
1.
1000.
40.
75.
40.
3.
<1.
<1.
<0.2
<0.4
<0.2
0.3
<0.5
<1.
2.
200.
<0.1
10.
10.
Flv Ashes
No. 2
200.
1.5
300.
30.
30.
100.
150.
300.
70.
200.
800.
100.
70.
100.
20.
5.
400.
200.
100.
200.
10.
40.
<0.5
<0.5
<1.
0.7
8.
<0.5
10.
10.
<0.5
<0.5
20.
1000.
60.
100.
130.
2.
1.
4.
<0.2
<0.4
<0.2
0.4
<0.5
<1.
15.
150.
0.7
40.
40.
No. 3
200.
1.5
200.
60.
40.
3000.
1000.
200.
30.
300.
1200.
60.
70.
100.
20.
20.
700.
150.
30.
50.
7.
70.
<0.5
<0.5
<1.
0.7
10.
<0.5
10.
15.
<0.5
<0.5
40.
500.
30.
30.
60.
2.
1.
4.
<0.2
<0.4
<0.2
0.4
<0.5
1.
30.
100.
1.
20.
30.
195
-------
(a) There are high concentrations of SO4 in the samples.
(b) The SO^ is nearly entirely water soluble.
(c) The SO^ is essentially the only anion in the water-
soluble phase.
(d) There are very high water solubilities of the oil-
fired emissions and, to a lesser extent, the coal-
fired emissions.
As can be seen, the coal-fired fly ash samples are much less
water soluble than are the oil-fired fly ashes, but as is discussed
later, the coal fly ash particulates predominantly are amorphous
aluminum-iron-silicate glasses which of course, are insoluble. The
chemical inertness of these glasses may have favorable health aspects
since they contain portions of trace heavy elements which generally
are regarded as hazardous. However, the SO^ contents of the coal-
fired fly ashes are nearly entirely water-soluble. Extensive deter-
minations made on these samples show the sulfur contents to be
nearly entirely in the SO4 form.
Water-Phase Separation
The above findings suggest a ready, simple mode of fractionating
fossil fuel particulate emissions into water-soluble metal (and
ammonium) sulfates and water-insoluble metal oxides (and silicates)
plus inert carbon. Any free H2SO4 acid, of course, also is contained
in the water-soluble phase of the samples, but H2SO4 acid has not
been found to be present in large percentages, although the method
used for its determination has given erratic results.
The separation of samples into water-soluble/insoluble phases
has proved useful for structural identifications of specific metal
sulfate forms, principally by FT-IR, and of oxide forms by XRD.
The water solubility separation is simply achieved by stirring
a 2-gram sample in 150 ml of water at room temperature for one hour
using a mechanical ("Mag-Mix") stirrer, filtering, washing and drying
the insoluble phase, and gently taking to dryness an aliquot of the
soluble phase. After drying, the insoluble residue is weighed to
give the percent insoluble fraction with the percent soluble obtained
by difference. As stated earlier, the only anion of any significant
concentration in the concentration in the soluble phase is the SOT
and, in fact, the soluble phase contains nearly all of the 864"
present in the total unfractionated oil-fired fly ash samples.
196
-------
In the oil-fired fly ash work samples used in this program,
the water-soluble phase represents from 66% to nearly 100% of the
sample components exclusive of the inert, soot-like carbon. The
soluble, phase components of the oil fly ashes are primarily metal
and NH4 sulfates plus any H2S04 acid, while the insoluble phase
components are carbon, oxides, and minor amounts of insoluble sul-
fates.
The coal-fired fly ash samples also contain a water-soluble
sulfate phase. These are much lower in percentage contents due to
the high concentrations of insoluble inert iron-aluminum silicates
and lesser amounts of insoluble crystalline minerals such as quartz,
hematite, and magnetite in the coal fly ashes.
Oil-Fired Fly Ash Compound Identification
Chemical Valence of Vanadium—In conjunction with the water
solubility studies, it was determined that vanadium in the oil fly
ash samples is present, principally in a water-soluble form. It
was noted that the water-soluble solutions had a greenish to green-
ish-blue color nearly proportional to the concentrations of vanadium
determined present. Valence state measurements of vanadium in the
oil fly ash samples were made using an adaptation of an extraction-
photometric method described by Shcherbakova, et al. (14), for the
determinations of Vv and VIV (actually reduced vanadium) in vanadium
catalyst samples. Vv was determined in the presence of VIV at an
acidity of 0.2 N since it was determined that at pH >_1 vanadium (IV)
is oxidized to vanadium (V) by atmospheric oxygen. ,
Following the procedure described by Shcherbakova, et al., 0.1
gram of oil fly ash sample was dissolved in 10 ml of 0.2 N HC1, and
the insoluble portion was filtered and washed with 0.2 N HC1 and
diluted to a volume of 100 ml. Extraction was carried out on a 2-ml
aliquot using 5 ml of 10~2 PMBP solution (l-phenyl-3 methyl-4 benzoyl-
pyrazolone-5), 1 ml pentanol, 4 ml chloroform, and 8 ml 0.2 N HC1.
Vv in the presence of any reduced vanadium was read spectrometrically
at 500 nm. Total V in the sample was determined by oxidizing another
aliquot of the above sample solution to Vv and repeating the extrac-
tion-photometric procedure. Reduced vanadium was found by the dif-
ference between the total vanadium determination result, and the Vv
value was determined in the presence of reduced vanadium. Total
vanadium in the sample and in the water-soluble phase also was
determined by atomic absorption analyses with better precision and
accuracy than obtained by use of the extraction-photometric proce-
dure.
197
-------
The results obtained on the oil fly ash samples by use of the
above methods are given in Table 6. As can be seen in the table,
the reduced vanadium values (Column 6) coincide closely with total
vanadium contents of the water-soluble fraction (Column 5). Since
V11 and V111 vanadium states are very unstable, it is highly prob-
able that the water-soluble vanadium is in the VIVstate.
Table 6. V in the Presence of Reduced Vanadium and Total
Vanadium Determinations (Oil-Fired Fly Ash)
Sample
No.
1
2
4
5
6
Extraction-Photometric Atomic Absorption
v
vv
1.70
4.50
0.90
0.14
0.35
yTotalc
2.25
5.7
10.7
11.75
1.1
vTotal
2.27
6.68
10.2
12.85
1.10
vWater-Soluble
0.50
2.23
8.98
12.90
0.78
yRedueed
0.57
2.18
9.3
12.71
0.75
9
Results in percent.
b V
V in presence of reduced vanadium.
c V
V after oxidation of reduced vanadium.
d V
Difference between Column 1 (V in presence of reduced vanadium)
and Column 4 (total V determined by AAS) results.
The water solubilities of two reference vanadium compounds (ICN
Pharmaceuticals vanadium sulfate and Alfa vanadium oxysulfate) were
compared with oil fly ash samples Nos. 2, 4, and 5 before and after
heating under argon at 350°C. The vanadium sulfate was found to be
very water insoluble both before and after heating. The VOS04« 5H20
was found to be highly water soluble before heating, exhibiting a
deep greenish-blue color but was only very slightly water soluble
after heating. (Anhydrous VOS04 is reported as insoluble in the
literature.) The oil fly ash samples behaved similarly with the un-
heated samples giving deep greenish coloration in the water solutions
and the heated samples imparting no color. Semiquantitative analy-
ses of the two reference vanadium compounds and fly ash samples
showed no vanadium (<0.1%) was dissolved in water after the samples
had been heated.
198
-------
Based on the valence state determinations, the water solubility
color tests, and chemical assays, it appears that the oil fly ash
samples contain water-soluble VIV OS04- 5H20 and water-insoluble V2O5.
Using the elemental analyses from Tables 4 and 5, the data
given in Table 7 for oil-fired fly ash samples were calculated
based on the assumptions that:
(1) The cation concentrations contained in the soluble
fractions probably were sulfate forms, since no
other anions of any significant concentrations were
present.
(2) The cation concentrations contained in the insoluble
fractions were oxide forms primarily, plus limited
concentrations of insoluble sulfates.
(3) The carbon, of course, would be present as an insoluble
form.
For example, considering the Mg in Sample No. 1, of the 18.1%
present in the total sample, 4.71% is contained in the soluble phase
and the remaining 13.7% in the insoluble phase. Using the gravi-
metric factor for Mg-*-MgS04 of 4.95, the MgS04 content would be
4.95 x 4.71% = 23.3% MgS04.
Similarly, using the gravimetric factor for Mg-»-MgO of 1.66 x 13.7%,
the calculated insoluble MgO would be 22.8%.
These assumptions are not at odds with equilibrium thermodynamic
calculations. As can be seen in Table 7, the possible calculated
combinations total close to 100% for the Nos. 1, 2, 4j. and 5 samples.
The No. 6 combinations total only ~90%. The total 80$ contents of
the calculated compounds given at the bottom of Table 6 check reason-
ably well with determined concentrations givgn in Table 3 except
for the No. 5 sample where the calculated S04 totals 46.5% versus
the determined value of 57.6%, and for the No. 6 sample where the
calculated S04 totals 55.2% versus the determined value of ~49%.
X-ray Diffraction of Oil-Fired Fly Ashes—Samples were pre-
pared for X-ray diffraction analyses by heating in an argon atmos-
phere at 350°C for two hours to drive off loosely bound and capillary
water, mixing in a mechanical shaker to break up agglomerated par-
ticles, and storing in a desiccator prior to X-ray analysis. The
total samples, water-soluble and water-insoluble fractions, were
199
-------
Table 7. Possible Compound Compositions of Oil-Fired Fly Ash
Samples Based on Chemical Analyses of Soluble and
Insoluble Phases
Calculated Species
C as C
H20*
H2S04*
NH4 as (NH4)HS04
Mg as MgO
Mg as MgS04
V as V205
V as VOSO4 • 5H2 0
Fe as Fe203
Fe as FeSO4
Ni as NiO
Ni as NiS04
Al as A1203
Al as A12(S04)3
Si as Si02
Na as Na2S04
K as K2S04
Ca as CaO
Ca as CaS04
Other Elements as
Oxides/ Sul fates
Totals of Above
Sulfates**
No. 1
12.4
7.0
0.1
0.08
22.8
23.3
3.2
2.5
0.45
0.8
0.85
2.65
1.4
3.2
0.55
12.1
0.3
0.55
2.0
1.3
97.5
35.15
No. 2
63.7
5.0
0.2
0.83
3.8
5.7
7.9
11.1
0.2
0.7
0.3
1.6
0.06
0.15
0.4
0.95
0.2
0.2
0.5
0.35
104.2
12.3
No. 4
21.5
4.5
0.05
5.18
1.6
24.8
2.2
44.5
0.3
0.5
0.3
2.8
0.3
1.45
0.45
1.55
0.2
0.04
0.25
0.75
113.1
46.0
No. 5
1.5
5.5
1.0
1.03
0
12.9
0
64.0
0
1.3
0
6.0
0.02
0
0.05
6.25
0.2
0.01
0.65
1.0
101.3
46.0
No. 6
14.5
2.0
1.5
46.6
0
11.9
0.6
3.9
0
1.1
0.06
0.8
2.2
1.7
0.5
0.6
0.25
0.2
0.55
1.0
90.1
55.2
*H20 and H2S04 values are those determined as given in Table 4
rather than calculated values based on H.
**S04 contents of the calculated species.
200
-------
run separately. The prepared sample specimens were placed in cavity
mounts and analyzed using a Norelco-Phillips X-ray unit, with CuK
radiation and a graphite monochrometer. Strip chart recordings were
obtained over a 2 grange of 10°C to 70°C. Structural interpreta-
tions were made by use of ASTM reference data plus synthetic stan-
dards and reference compounds. Peak heights were read from the
strip chart recordings for intercomparisons of sample and sample
fraction peak heights with each other and with reference intensities.
As anticipated, the X-ray patterns were very complex due to
the numerous phases present and their variances in contained waters
of hydration. Interpretations were assisted materially by use of
reference patterns obtained on chemical forms postulated as present
based on the chemical analyses of the total samples and of their
water-soluble and insoluble fractions.
These calculated chemical forms proved useful for both the XRD
identifications and the subtractive FT-IR work described later. MgO,
V205, and carbon patterns were readily identified in the total sam-
ples and in the insoluble fractions of these in, semiquantitatively,
the concentration ranges given in Table 7. No V205 or MgO patterns
were obtained in the No. 5 and No. 6 oil-fired fly ash samples.
XRD patterns were obtained on the vanadium sulfate and oxysul-
fate reference salts after baking at 350°C under argon and compared
with patterns obtained on the total samples and soluble phases of
oil fly ashes. Diffraction peaks were obtained on the total samples
and water-soluble phases at the d-spacings obtained for the vanadium
oxysulfate reference but not the vanadium sulfate. Assays made of
the vanadium oxysulfate reference salt indicated a VOSO4*1H20 com-
position after baking and VOS04«5H20 in the untreated salt. The
XRD work is continuing to identify other sulfate and oxide phases
present in the samples using an internal standard (NaCl) to quantify
the data.
Fourier Transform Infrared (FT-IR)—Based on present investi-
gations, it appears that Fourier Transform infrared spectrometry
offers greater potential for inorganic compound identification in
complex mixtures than does XRD. While in the past infrared spectro-
scopy has been used mostly for organic compound identifications, by
using the sensitivity of the Fourier Transform infrared system and
the data handling capability of a dedicated computer, it has been
found that inorganic compounds can be identified as well. The
dedicated computer permits the storage of infrared reference com-
pound spectra, identified via elemental—anion/cation—analyses as
possibly present on the samples to be examined, and the subtraction
of these spectra from unknown sample spectra.
201
-------
The details of this new application of infrared to oil-fired
and coal-fired fly ash samples are given in another paper (15) in
this program.
Coal-Fired Fly Ashes
Lesser efforts have been expended to date on coal-fired fly
ash inorganic compound identifications. Petrographic examinations
and X-ray diffraction analyses have been made which confirm the
conclusions of previous investigators (16-20) that coal-fired fly
ashes are composed principally of an amorphous glass structure.
Water extractions have shown the presence of a water-soluble phase
composed principally of sulfates. However, due to the presence of
the high concentration glass phase in coal-fired fly ashes, the
water-soluble phase is less than in oil-fired fly ashes.
To identify the glassy phase constituents, two synthetic oxide
mixtures, G-l and G-2, were prepared for use as XRD and FT-IR refer-
ence materials. Aluminum, iron and silicon oxides were mixed and
ground together in the proportions given below:
A1203 - 51% A12O3 - 40%
Fe2O3 - 20% Fe2O3 - 15%
Si02 - 29% Si02 - 45%
Portions of these mixes were then fired to obtain liquid melts, and
after quenching and solidification, the melts were ground to ~300
mesh. XRD patterns were obtained on the synthetic oxide mixes, the
oxide melts, the coal-fired fly ash total samples, and their water-
insoluble fractions. The resultant XRD patterns of the synthetic
oxide mix, before melting, showed very strong A1203, Fe203, and Si02
structures. The oxide melts showed only a weak a quartz, Fe203,
and very weak Fe304 and A1203 patterns. X-ray pattern structures
found in the coal-fired fly ash samples and in the insoluble phases
thereof resembled those of the oxide melts in respect to the strengths
of the a SiO2, Fe203, Fe304, and A1203 patterns. Some additional
but weak patterns also were found present in the total fly ash
samples but not in the water-insoluble phases.
Only limited additional XRD work is planned on the coal-fired
fly ashes since, except for the few metal sulfate and weak oxide
patterns which can be seen, the coal-fired fly ashes exhibit little
XRD pattern structure. Thus the presence of a large amorphous
glass phase is indicated only indirectly by XRD by the absence of
strong diffraction patterns.
202
-------
FT-IR spectra of the glassy or amorphous phase, while strikingly
different from spectra obtained of crystalline oxides, appear to
be sufficiently unique and definitive for direct identification
using appropriately prepared reference materials and spectral sub-
tractions. The water-soluble phases of the coal fly ash samples—
sulfates—can be determined by FT-IR similarly to the oil-fired
fly ash identifications. At this point, pending further method
development to ascertain sensitivity and accuracy, the FT-IR tech-
nique appears superior to XRD in identifying the glass phase and
metal sulfate constituents of coal-fired fly ashes.
Distributionof Selected Trace Metals in Fly Ash Samples —
Since health effect aspects of fly ash emissions are of key interest,
and certain trace elements are suspected to be hazardous, the water-
soluble and water-insoluble phases of the Nos. 2 and 3 coal-fired
fly ashes were analyzed using spark source mass spectrography. If
the trace elements were found to be tied up in the glass phases of
the fly ash emissions, it might be presumed that their toxicity
effects would be less than if present in a water-soluble form. The
results obtained are shown in Table 8.
Table 8. Ratios of Water-Soluble to Water-Insoluble Trace
Element Concentrations in Coal Fly Ash
Element No. 2 Coal Fly Ash No. 3 Coal Fly Ash
Be
V
Cr
Mn
Co
Ni
Cu
Zn
As
Se
Mo
Cd
Sn
Sb
Tl
Pb
U
1-1
1-2
1-1
1-1
1-2
1-1
1-1
2-1
1-1
1-4
1-1
5-1
1-5
1-5
1-1
1-1
1-2
4-1
2-1
1-1
1-1
1-1
2-1
5-1
5-1
3-1
4-1
4-1
5-1
1-2
1-2
10-1
5-1
1-1
203
-------
In viewing these data, it should be kept in mind that the spark
source determinations are semiquantitative and that the water-insol-
uble phases also contain metal oxides and some insoluble sulfates,
so the insoluble phase represents more than the glassy portion of
the ashes.
No trend is apparent in the data, but this work will be con-
tinued on additional coal-fired fly ash samples. It is important to
learn about the distribution of these trace quantities of hazardous
elements since it appears virtually impossible to pin down their
specific compound forms except perhaps by tedious microscopic
searches using a STEM instrument to show cation/anion associations.
SUMMARY
Methodology Work
(1) Separations of fly ash samples into water-soluble and
insoluble phases simplifies the analytical efforts.
(2) Complete cation-anion determinations are useful in
guiding and quantitating the structural analytical
techniques.
(3) A new application of infrared spectrometry which uses
the capacity and capability of Fourier Transform to
store reference compounds and to subtract spectra has
been found to offer great potential for the identifi-
cation of inorganic compounds.
Composition
(1) Oil-fired fly ash emissions are highly water soluble
and the water-soluble fractions consist primarily of
sulfates. Thus health effects evaluations of the tox-
icities of metal contents should consider this form
in addition to the more commonly tested oxide form.
(2) Coal-fired fly ash emissions also contain a signifi-
cant water-soluble sulfate phase.
(3) In addition, coal-fired fly ash emissions contain a
major glass phase which contains some of the heavy
metals deemed hazardous. This glass phase may be less
hazardous to health than the water-soluble sulfate phase.
204
-------
RECOMMENDATIONS
Coal-fired fly ash samples from the Western and Great Plains
areas and from petroleum refinery operations should be obtained
and analyzed in order to provide a greater representation on which
to apply and test the developed methodologies so as to increase
the limited analytical data bank on fossil fuel derived particulate
emissions.
Consideration should be given to expanding the program scope
to examining partlculates emitted from nonconventional fossil fuel
combustion sources.
A more complete library of reference spectra should be pre-
pared for the Fourier Transform infrared spectrometer work. A
planned replacement of the presently used FTS-14, which has limited
storage (~20 low resolution files), by an FTS-10 of increased
storage capacity will permit permanent cataloging for storage and
retrieval of the needed metal sulfate and oxide reference spectra
to facilitate identifications in the samples.
Additional studies should be carried out at a microscopic level
to examine single particles for compositions as functions of their
surface and depth concentrations in order to ascertain the chemical
forms of trace constituents in the particulate emissions.
205
-------
REFERENCES
1. Altshuller, A. P. Anal. Chem., Annual Review, 41:5, 1969.
2. Mueller, P. K., et al. Anal. Chem., Annual Review, 43:5, 1971.
3. Coleman, R. F. Comparison of Analytical Techniques for Inorganic
Pollutants. Annal. Chem., 48(8):640A-653A, 1976.
4. Manual of Methods for Chemical Analysis of Water and Wastes.
EPA-625/6-74-003.
5. Potential Pollutants in Fossil Fuels. PB-225-039, June 1973.
6. Compendium of Analytical Methods. PB-228-425, April 1973.
7. Evaluation of Selected Methods for Chemical and Biological
Testing of Industrial Particulate Emissions. EPA-600/2-76-137,
May 1976.
8. Technical Manual for Process Measurement of Trace Inorganic
Materials. EPA Report Pending, July 1975, Contract No. 68-02-
1393.
9. U.S. Federal Register, Federal Register List of Approved Test
Procedures, 40(111), June 9, 1975. pp. 24535.
10. ASTM Annual Book of Standards, Part 23, 1976.
11. Pollutant Analysis Cost Survey, EPA-650/2-74-125, December
1974.
12. Analytical Guide, Am. Indust. Hyg. Assoc. J., August 1975.
pp. 642-645.
13. Anderson, P. L. Free Silica Analysis of Environmental Samples -
A Critical Literature Review. Am. Ind. Hyg. Assoc. J., pp. 767-
778, September 1975.
14. Shcherbakova, N. A., N. V. Mel'chakova, and V. M. Peshkova.
Determination of Vanadium (V) and Vanadium (IV) in Each Other's
Presence. J. of Analyt. Chem. of the USSR, Eng. Ed., 31(2),
Part 2, February 1976.
206
-------
15. Jakobsen, R. J. , R. M. Gendreau, W. M. Henry, and K. T. Knapp.
Inorganic Compound Identification by Fourier Transform Infrared
Spectroscopy. Paper presented at conference on Measurement
Technology and Characterization of Primary Sulfur Oxides
Emission from Combustion Sources, Southern Pines, North Car-
olina, April 24-26, 1978.
16. O'Gorman, J. V., and P. L. Walker. Mineral Matter and Trace
Elements in U.S. Coals. R&D Report 61, Interim 2, Office of
Coal Research, Department of the Interior, July 1972.
17. Watt, J. D. , and D. J. Thome. Composition and Pozzolanic Prop-
erties of Pulverized Fuel Ashes, I and II. J. Appl. Chem.,
15:585-604, 1965.
18. Minnick, L. J. Fundamental Characteristics of Pulverized Coal
Fly Ashes. 62nd Annual ASTM Meeting, June 1959.
19. Simons, H. S., and J. W. Jeffery. An X-ray Study of Pulverized
Fuel Ash. J. Appl. Chem., 10:328-336, 1960.
20. Bickelhaupt, R. E. Volume Resistivity Fly Ash Composition Rela-
tionship. Environ. Sci. & Tech., 9(4):336-342, 1975.
207
-------
Investigation of Participate Sulfur by ESCA
Arthur S. Werner
GCA/Technology Division
ABSTRACT
Sulfur species emitted by combustion systems play roles
of varying significance in atmospheric chemistry and
have been linked to health and ecological hazards.
Gaseous and particulate sulfur emissions include a
range of oxidized and reduced compounds. We have been
investigating particulate sulfur forms emitted by oil-
fired and coal-fired combustion sources using X-ray
photoelectron spectroscopy (XPS) or ESCA.
In ESCA, a soft X-ray beam strikes a sample, ejecting
inner shell electrons from atoms on or near the sample
surface. By measuring the kinetic energy spectrum of
the photo-ejected electrons, the elemental abundances
of particulate surfaces can be determined in concentra-
tions as low as 0.1%. High resolution spectra of
individual core electrons reveal a "chemical shift,"
which is a function of the oxidation state of the pre-
cursor element.
As part of this study, a particulate emitted by combus-
tion sources was collected on filters and on impactor
substrates. ESCA analyses were performed directly on
the filters and substrates with no prior sample prepara-
tion. Sulfur species identified include sulfate,
sulfite, sulfide, and organic sulfur compounds.
209
-------
INTRODUCTION
Environmental effects of inorganic species are functions not
only of elemental composition but also of compound form and physi-
cal state. A number of analytical techniques are commonly
employed to determine elemental composition and anion content of
particulate emitted by combustion systems. These techniques in-
clude spark source mass spectrometry, X-ray fluorescence, atomic
absorption spectroscopy, and wet chemistry. Particulate sulfate
is usually determined as the soluble anion by precipitation and
turbidity or by acid-base titration.
We report here on an investigation of particulate cations and
anions (including sulfates) using X-ray photoelectron spectro-
scopy, or ESCA, which was undertaken as part of an environmental
assessment of the Chemically Active Fluid Bed (CAFB) process.
This study, carried out in 1975, was an evaluation of the emis-
sions from the oil-fired CAFB. We are presently engaged in a
follow-up program to assess this process for lignite firing.
THE CHEMICALLY ACTIVE FLUID BED (CAFB) PROCESS
The Chemically Active Fluid Bed (CAFB) process was developed
by the Esso Research Centre, Abingdon (ERCA), England, as a means
to generate electrical energy from high-sulfur, high-metal heavy
fuel oil. Fuel oil is fed continuously into a fluidized bed of
limestone maintained at 870°C (1600°F) by preheated, substoichio-
metric air. The fuel oil entering the gasifier is vaporized,
oxidized, cracked, and reduced to produce a low-Btu, low-sulfur
gas which is then burned in a conventional gas-fired boiler.
Sulfur contained in the oil initially forms various gaseous com-
pounds which then react with the bed lime to yield calcium sul-
fide. The sulfided lime is cycled to a regeneration unit where
it is oxidized to produce CaO, which is returned to the gasifier,
and SO2, which is sent to a sulfur recovery unit. An additional
feature of the CAFB process is that the gasifier bed material
adsorbs vanadium, nickel, and sodium contained in the fuel oil,
thus limiting air emissions of these trace elements.
At present, the only existing CAFB unit is a 2.93 MW pilot
plant at the ERCA facility. Foster-Wheeler Energy Corporation
(FW) is in the final construction stages of a 10 MW retrofit
demonstration plant to be constructed in San Benito, Texas, at
the La Palma Power Station of the Central Power and Light Company.
The ERCA pilot plant is the facility at which all sampling dis-
cussed in this paper took place.
210
-------
SAMPLING
Sampling of the oil-fired CAFB in late 1975 was carried out
in accordance with procedures for environmental assessments as
they were specified at that time. Particulate sampling was
accomplished using a standard RAC train constructed according to
the.procedures outlined in EPA Method 5. Due to the positions of
the installed ports, eight point traverses were taken on two
diameters 120° apart. The train was modified slightly to allow
for sampling of gaseous organic species. Particulate size dis-
tribution measurements were taken with a University of Washington
eight-stage in-stack impactor using ungreased substrates. A sin-
gle point was sampled isokinetically for sufficient time (15 to
30 minutes) to collect a weighable quantity on each stage. In
addition to particulate, flue gas was sampled for NOX by Method 7,
S02/SO3 by Method 8, and H2S by Method 11. An Orsat analyzer
was used to measure CO, C02, and 02. In addition to collecting
stack-emitted particulate, samples of stack and internal cyclone
fines and bed material were acquired.
ANALYSIS
Particulate emissions, cyclone fines, bed material, and fuel
were analyzed for trace elements, surface species, and inorganic
anions and cations. The principal technique used for elemental
determinations was spark source mass spectrometry (SSMS). Atomic
absorption spectroscopy (AAS) and wet chemical techniques were
employed to supplement the SSMS measurements.
Particulate and solid samples were investigated for surface
elements and inorganic compounds using X-ray photoelectron
spectroscopy (XPS), also known as electron spectroscopy for
chemical analysis (ESCA). In ESCA, a high energy X-ray beam (for
the analyses reported here, the MgKa line having an energy of
1253.6 eV was used) impinges on a solid, knocking out core elec-
trons from atoms on the solid surface. The resulting electrons
pass through an energy analyzer and are pulse-counted by a parti-
cle multiplier. The binding energy of the electrons is then
calculated from the energy of the incident X-ray, the spectrometer
work function, and the measured electron kinetic energy. Binding
energy ranges can be uniquely associated with specific precursor
elements. In fact, ESCA is sensitive to all elements in the
periodic table. An additional feature of ESCA spectra is that
the precise electron binding energy in a known range is a function
211
-------
of the valence state of the atom of interest. For example, sulfur
combined as sulfate can be differentiated from sulfur as sulfide.
In addition, because core electron ejection cross sections are
relatively independent of valence state, the ratio of the areas
under the peaks corresponding to sulfate and sulfide is a measure
of the sulfate-sulfide concentration ratio.
All samples analyzed by ESCA in this study were first scanned
over the entire electron binding energy range (broadband scan) to
identify those elements present in concentrations greater than
0.1% to 1% (the sensitivity of ESCA to any one element is a func-
tion of the photoionization cross section of the most intense core
electron emission of that element). These broadband spectra were
then analyzed to yield surface concentrations of all identifiable
elements. To supplement the bulk SSMS analyses, high energy
argon ions were used to etch away surface layers exposing strata
20 to 100 A deep. The exposed sample layers were then rescanned
over the entire binding energy, and the resultant elemental con-
centrations were compared with the surface and bulk values.
In this study, ESCA was used to analyze impactor substrates,
RAC filters, gasifier and regenerator bed solids, and cyclone
fines. The principal elements of interest were sulfur and vana-
dium. The discussion below primarily concentrates on sulfur.
Our interest in examining the particulate and solid matter
was to determine the fate of vanadium (550 ppm) and sulfur (3.5%)
contained in the fuel oil. The total particulate emission rate
for this run was 0.1 lb/106 Btu; SO2 was 1.60 lb/106 Btu (828
ppm); and S03 was 0.02 lb/106 Btu (11.1 ppm). Thus, sulfur emis-
sions were distributed as 98.5% S02, 1% as SO3, and 0.5% as
particulate sulfur.
Figure 1 is a broadband ESCA scan of the surface of the
stack cyclone particulate. The elements observable on the surface
(and thus present at concentrations greater than 0.1%) are oxygen,
vanadium, calcium, carbon, sodium, and sulfur. Figure 2 is a
scan of the 2p electron of sulfur. Sulfate and sulfide are pre-
sent at a rate of roughly 3 to 1. Figures 3 and 4 are similar
ESCA scans of particulate collected on the RAC filter. Comparison
of Figures 1 and 3 indicates that the filter particulate has
less surface carbon and more surface sulfur than particulate cap-
tured by the cyclone. Figure 4 shows that essentially all
particulate sulfur on the RAC filter is sulfate.
212
-------
ro
co
c
3
o
UJ
o
O
600
•"is
480
360 240
BINDING ENERGY, eV
SC9
120
Figure 1. Broadband ESCA scan of stack cyclone particulate.
-------
in
'c
3
h_
o
O
O
SC9
I
I
I
175
171
167 163
BINDING ENERGY, eV
159
155
Figure 2. ESCA scan of 2p electron of sulfur of stack cyclone particulate.
-------
N>
cn
c
3
o
uT
!5
tr
600
480
360 240
BINDING ENERGV «V
120
0
Figure 3. Broadband ESCA scan of sampling train filter particulate.
-------
JO
hw
o
uT
5
cc
o
o
175
171
167 163
BINDING ENERGY, eV
159
155
Figure 4.
ESCA scan of 2p electron of sulfur of sampling train filter
particulate.
-------
To characterize further the compositions of the particulate
emissions, each impactor substrate was analyzed by ESCA. Table 1
summarizes the results of broadband scans of each particulate
fraction. Fractions are denoted UW91 through UW98 in descending
order of size. Columns labeled "surface" refer to scans of un-
modified samples;o"subsurface" indicates scans taken after etching
away roughly 100 A of surface. Codes SC9 and FS9 refer to stack
cyclone and RAG filter particulate, respectively.
It is apparent from Table 1 that the material captured in the
stack cyclone is not representative of particulate emissions from
the stack. Material captured by the cyclone contains a good deal
of unburned carbon and unoxidized sulfur. Most of this material
is likely large particulate which passed through the boiler un-
affected by the highly oxidizing atmosphere.
The surface/subsurface sulfur ratio of impactor fractions
UW91, 93, 95, and 97 indicates that sulfur is clearly enriched on
particulate surfaces. This phenomenon is due to two factors:
adsorption of sulfur on lime particles in the bed, and condensa-
tion and adsorption of gaseous SO2 and S03 on particulate when
the flue gas leaves the boiler for the cooler stack region. The
latter explanation has been invoked to explain enhanced concen-
trations of certain trace metals in small ambient particulate.
The analogous enrichment on surface versus subsurface of carbon
is also partially due to condensation of organics in the stack.
However, this should not be overinterpreted because enhanced
surface carbon is almost always found by ESCA on samples exposed
to ambient air. The surface adsorption explanation is supported
by the appearance of calcium in the subsurface scans of impactor
particulate but not in the surface scans. The bulk of the parti-
culate is lime from the gasifier bed.
The ongoing lignite study is concentrating on particulate
sulfur abundance as a function of depth from the surface. Sulfur
to calcium ratios are being measured after etching the surface for
various time periods. Preliminary data indicate that sulfur con-
centrations decrease by a factor of about five from the surface to
a depth of 500 A.
ACKNOWLEDGMENTS
This work was supported by the U.S. Environmental Protection
Agency, Industrial Environmental Research Laboratory, Research
Triangle Park, under contract numbers 68-02-1316, task order
number 14, and 68-02-2632.
217
-------
ro
cx>
Table 1. Surface and Subsurface Concentrations of Stack Particulate
Collected During Oil Gasification
Sample, % abundance
Element
0
V
N
C
Na
S
Ca
SC9
Surface
12.8
1.1
-
80.8
0.8
3.1
1.5
FS9
Surface
34.6
2.4
-
49.8
2.9
7.7
2.6
UW91
Surface
34.2
0.7
3.3
52.4
1.4
7.9
-
Sub-
Surface
61.0
1.0
-
31.5
0.9
4.2
1.4
UW92
Surface
28.8
0.4
2.7
60.7
1.0
6.4
-
UW93
Surface
37.4
0.5
2.3
50.1
1.3
8.3 ,
-
Sub-
Surface
67.9
0.9
-
26.1
0.9
2.9
1.2
UW94
Surface
32.1
0.6
3.1
56.2
1.5
6.5
-
CW95
Surface
34.4
0.4
2.2
55.1
1.1
7.0
-
Sub-
Surface
65.9
0.9
-
28.1
0.8
2.8
0.7
UW96
Surface
32.1
0.4
3.2
55.6
1.6
7.1
-
UW97
Surface
35.4
0.5
2.7
52.9
1.6
6.9
-
Sub-
Surface
63.8
1.3
-
30.5
1.3
2.4
0.7
UW98
Surface
14.1
0.3
-
82.4
0.8
2.5
-
Sub-
Surface
10.5
1.9
-
83.9
0.8
2.8
-
See text for explanation of colunn headings.
-------
Sulfur Emissions Sampling and Analysis
Ray F. Maddalone
TRW, DSSG
ABSTRACT
Sampling and analysis for S02, H2S04, and particulate
sulfate were spurred by the concern over sulfate emissions
from flue gas desulfurization (FGD) units.
It has been demonstrated that FGD units remove S02; how-
ever, it was postulated that they emit potentially more
toxic sulfate aerosols. Not enough data are available'at
the current time to support or reject that supposition.
Over the past three years, TRW has been involved in the
sampling and analysis of sulfate particulate and sulfuric
acid. This presentation will review the particle size
and chemical speciation data collected at several combus-
tion sites employing FGD units.
In particular, this presentation will describe the
modifications, the laboratory testing, and the use of
the controlled condensation system (CCS) for sulfuric
acid collection at combustion sources. Sulfuric acid
data collected using the CCS will be presented. The data
were gathered during a 60-day test program conducted at
the EPA/TVA Shawnee test facility, at a coke oven, and at
a full-scale industrial boiler. Sulfate particulate
sampling and analysis were undertaken concurrently at the
utility boiler. This program included the analysis of
particulate matter using XRD, FTIR, and ESCA. The results
of these tests will be discussed, and possible analysis
approaches will be outlined.
219
-------
INTRODUCTION
Sampling and analysis for S02, H2S04, and particulate sulfate
have been spurred by the concern over sulfur oxide emissions from flue
gas desulfurization (FGD) units. While it has been demonstrated that
the FGD units remove SO2 efficiently, it has been postulated that
they emit potentially more toxic sulfate aerosols.
With this concern over the sulfur oxide output of the FGD, con-
certed efforts to develop new or improved sampling methods have been
conducted by the EPA and its contractors over the last five years.
These programs sought to define the problems and improve current sul-
fate sampling methodology. TRW has been involved in the research
and development of methods as well as in the field measurement of
sulfates and H2S04 from the outlet of FGD's. This paper will discuss:
• Considerations for sampling oxidizable and volatile
sulfur emissions from a wet scrubber, to effectively
collect and preserve species.
• The development of H2SO4 and sulfate sampling procedures
using the controlled condensation system.
• The results of H2SO4 and sulfate sampling tests at the
EPA/TVA Shawnee FGD test facility and at a coke oven
controlled by a charged droplet scrubber using the
controlled condensation system.
• Sulfate size distribution at a utility boiler equipped
with a soda ash wet scrubber.
SULFITE/SULFATE STABILITY STUDIES
A series of experiments were devised and run to obtain infor-
mation on the stability of typical sulfate and sulfite compounds
which might be expected to be emitted from an FGD. As an initial
survey procedure, a Thermal Gravimetric Analysis (TGA) was run. If
the TGA results indicated any instability in the 100°-150°C (typical
flue gas sampling temperature) range, an isothermal TGA was run which
would simulate the thermal conditions that a particle would experience
sitting on filter in a standard flue gas particulate matter sampling
train. Table 1 summarizes these stability studies.
220
-------
Table 1. Sul C i to/Sulla Lu Stability Studies
Compound
Thermal Gravimetric
Analysis
Isothermal Gravimetric
Analysis
Simulated ParlieuluU
Sampling
Na2S03 1. Dry Air
-No reaction up to 500°C (932°F)
-Weight increase of 2% between
500°-625°C (932°-1157°F)
150°C (302°F) -
3 hours with water
in the bubbler;
99% of the expected
sulfite activity
recovered
1. Dry Air - 100°C (212°F) 1. 150°C (302°F) -
-No reaction up to 150°C (302°F)
-Weight loss of 38% between
150°-235°C (302°-455"F)
Literature reports that Na2SO4
is stable up to 1200°C (2192°F),
Therefore, it was not tested.
-2% weight loss
. in a one hour and 20
minute period
2. Dry Air - 150°C (302°F)
-38% weight loss
in a one hour and 10
minute period
3. Dry Air - 100°C (212°F)
-1 part ferrous ferric
oxide mixed with NaHS03.
2% weight loss in a
one hour and 20
minute period
3 hours with water
in the impinger;
36% of the expected
sulfite activity
recovered
NaHS04 1. Dry Air
-1% weight loss from 20°-70°C
(68°-158°F)
-Stable from 70°-170°C
(158°-338°F)
-14% weight loss from 170°-600°C
(338°-1112°F)
(NH4)2S04 1. Dry N2
-No weight loss to 225°C
(437°F)
-Weight loss of 89%
from 225°-425°C (437°-797"F)
NH4HSO4 1. Dry N2
-15% weight loss from 20°-350°C
(68°-662°F)
-50% weight loss from 350°-400°C
(662°-752°F)
1. Dry N2 - 125°C (257°F)
-1.8% weight loss
in a 3 hour period
. Dry N2 - 125°C (2.r)7
-1.8* weight loss
i n a II hour period
221
-------
In two cases, Na? S03 and NaHSO3, a simulated sampling test
was run. This experiment closely approximated the conditions ex-
pected in sampling a wet scrubber using the EPA Method 5. A
sample train (Figure 1), consisting of an impinger, a Jtube packed
with glass wool, and a 47 mm filter holder, was placed in an oven
set at 150°C (302°F). A known weight of powdered Na2S03 or NaHS03
was placed on a 5.0 /u Mitex Teflon filter and inserted in the
filter holder. After allowing the system to equilibrate, labora-
tory air was bubbled through the impinger with 250 ml of H20 and
was passed through the glass wool and the filter holder. After
three hours the filter containing the Na2S03 or NaHS03 was placed
in 10 ml of 0.10 N I2, which was back titrated with 0.10 N Na2S203.
The results in Table 1 indicate that at 150°C (302°F) Na2S03
is stable even under conditions stimulating a heated filter in a
water-saturated air stream. On the other hand, NaHS03 decomposes
or volatilizes (exhibits weight loss) at 150°C. It is expected
that (NH4)2S03 and NH4HS03 will exhibit similar instabilities at
even lower temperatures. Ammonium sulfite decomposes at 60°-70°C
(140°-158°F) and sublimes at 150°C (302°F), while ammonium bisul-
fite is delinquescent and sublimes at 150°C (302°F) .
In most cases, sampling at an FGD will involve the collection
of liquid aerosols possibly containing dissolved scrubber materials.
A series of experiments were devised and run to obtain background
information on the stability of sulfite compounds under this sampl-
ing condition. These experiments were devoted to the study of the
stability of sulfite compounds during the collection on a heated
filter, where the sulfite aerosol (solution) would be converted
to a sulfite particle (solid) on the surface of the filter. As
seen above, sulfite compounds have remarkable thermal stability
under dry conditions. The objective of the following experiments
was to measure the sulfite activity after sulfite solutions were
dried. The results of these experiments will provide an indication
of the compounds' ability to resist oxidation or decomposition
going from a wet to dry state. The procedure employed is described
below:
1. Solutions of NH4HS03 and Na2S03 were made up to
contain approximately 15-20 jug SO3/10yul.
2. A 15 mm (diameter) by 80 mm vial was placed on a
hotplate and allowed to thermally equilibrate.
222
-------
OVEN ~150°C
GLASS
WOOL
FILTER
HOLDER
TO PUMP
Figure 1. Sulfite stability apparatus.
223
-------
3. Ten fil of each sulfite solution was dropped on the
heated bottom of the vial and allowed to dry for a
specified amount of time at 125°C (257°F) and at
150°C (302°F).
4. As soon as the liquid was evaporated, 10 ml of 0.1
M tetrachloromercurate (II) (TCM) was added to the
vial to stabilize the sulfite.
5. Using the West-Gaeke procedure, the sulfite content
in the vials was measured and compared to 10 fil
standards placed in a separate vial containing
10 ml of 0.1 M TCM.
The percentage sulfite activity recovered is summarized
in Table 2.
Table 2. Wet to Dry Sulfite Stabilities
Compound
NH4HS03
NH4HS03
NH4 HS03
NH4HS03
Na2S03
Na2S03
Na2S03
Na2S03
Temp.
°C
125
125
150
150
125
125
150
150
Drying Time
(Min.)
4.0
4.0
1.5
1.5
4.0
4.0
1.5
1.5
Recovery Average Rec.
54'2 48.4
42.6
2°'° 15.0
10.0
4'6 3.3
1.9
16-3 14.3
12.3
As can be seen in Table 2, both Na2S03 and NH4HS03 were
not stable while being heated to dryness. The relatively high
stability of NH4HS03 at 125°C is surprising compared to the
low stability of Na2S03 at the same temperature. The higher
stability of Na2S03 at 150°C temperature is probably the
result of the shorter time (1.5 versus 4.0 minutes) required
to evaporate the water; the Na2S03 spent less time in the
224
-------
water being heated, and, consequently, there was less time
available for any reaction in the liquid. The same reasoning
did not apply to NH4HS03, since by nature it is less thermally
stable than Na2S03.
The conclusion resulting from these experiments is that
a percentage of the sulfite aerosols sampled on a dry (heated)
filter will be oxidized or otherwise changed during the course
of sampling a wet scrubber. If sulfite species are to be
collected, an approach must be developed that allows for the
sulfite species to be collected efficiently under wet conditions
and immediately stabilized to ensure that the sample collected
reflects the in situ concentration.
SULFURIC ACID/SULFATE SAMPLING PROCEDURES
The systems used to quantify H2SO4 are based on selective
absorption or controlled condensation. A series of workers
(1)(2)(3) have refined the selective IPA absorption method.
This approach uses an impinger with 80% isopropyl alcohol to
collect the SOa and to pass the SO2• The SOa is collected in
a back-up impinger of 3% H2O2• This method is currently the
basis of the EPA compliance test (4) for sulfuric acid mist.
The major problem with this procedure is the lack of a pre-
fllter to effectively prevent particulate matter from reaching
the IPA impinger. The particulate matter in the impinger can
act either as a direct interferent by contributing S04 from
sulfate salts, or as an indirect inferent by catalyzing the SO^
to S04 oxidation in the liquid phase through action of trace
elements like Fe, Cu, or V.
The controlled condensation approach was first proposed
by Knol (5) and has been further developed by Goksoyr and Ross (6)
The Goksoyr-Ross system is the basis of an ASTM procedure for
SOX (7). In the controlled condensation approach, SO3 is
separated from the gas stream by cooling the temperature of
the flue gas below the dewpoint for S03 but above the dew-
point of H2O. The resulting aerosol is either collected on the
walls of the cooling coil or on a back-up frit. Investigators
(8)(9) studying controlled condensation in the laboratory
have found the precision and accuracy to be +_ 6% in synthetic
gas streams. However, these researchers addressed neither the
problem of particulate matter removal nor the possible
neutralization of H2S04 by alkaline particulate matter in a
filtration system.
225
-------
A sulfuric acid generator based on the design of Lisle and
Sensenbaugh (8) was modified and used to provide test gas streams
of varying H2S04, H20, S02, 02, and N2. The generator was a
12 mm O.D. quartz tube with a side arm injection port to
introduce liquids onto a heated course quartz frit (Figure 2).
The evaporator was wrapped with heating tape to vaporize
solutions of H2S04 that were metered into the evaporator using
a syringe pump. Gas outlet temperatures ranged from 300°-350°C.
Adjusting the H2S04 solution strength and flow produced a range
of H2S04 and water concentrations. The rest of the test system
(Figure 2) consisted of a heated quartz filter section and a
Pyrex Controlled Condensation Coil (CCC). The CCC was a
modified Graham Condenser which had a 60 mm medium frit added
to one end of the cooling coil.
The preliminary evaluator tests showed that the CCC collected
H2S04 efficiently over the ranges of moisture (4%-8%), H2S04
(10 ppm-20 ppm), and CCC temperature (35°-60°C) with a coefficient
of variance of +_ 7%. Before the system was taken to a coal-
fired combustor, an effort was made to determine the effect of
fly ash on H2S04 recovery. A series of experiments were run with
varying amounts of coal fly ash from TVA Shawnee Power Plant
placed on the filter prior to the start of the run. This approach
represented a worse case evaluation, since under normal field
conditions the H2S04 would see a slowly increasing amount of fly
ash. The fly ash was titrated after the run to determine if
there was any decrease in the alkalinity of the fly ash. By
expressing the decrease in the fly ash alkalinity in milli-
equivalents, it could be added to the H2S04 recovered from the CCC
to determine if an acid mass balance was retained. The results
of these experiments are summarized in Table 3.
Table 3. Summary of Fly Ash H2S04 Recovery Tests
Equivalent
Fly Ash
(g/m3)
ppm %
H2S04 02
ppm
S02
Average
Percentage H2S04 Found
Filter CCC Total
1.3
1.3
1.3
0.13
9
12
11
11
8
8
0
650
5300
700
15
14
11
0
81
86
87
89
96
100
98
89
226
-------
BACK PRESSURE GAUGE
SYRINGE PUMP
ro
TEFLON TUBE
COARSE QUARTZ FRIT
EVAPORATOR
HEATING MANTLE
HEATING
TAPE
L V-
TO IMPINGERS
AND PUMP
CONTROLLED
CONDENSATION
COIL
Figure 2. H~SO, test apparatus,
-------
These data indicate that the fly ash reacted with a portion
of the H2S04. Within this limited test series, there appears to
be a slight improvement in the H2S04 recovery with increasing
SOa concentrations perhaps because the SO2 competes with the
H2S04 for the "alkaline sites on the fly ash. With data available
at the present time, it is only possible to estimate that the
fly ash caused a 12%-14% reduction in the amount of H2SO4
collected by CCC.
SULFATE/SULFURIC ACID FIELD TESTS WITH CCS
Shawnee Test Facility
Extensive field tests with the Controlled Condensation System
(CCS) were conducted at a pilot plant located at a TVA coal-fired
power plant in Paducah, Kentucky, using the apparatus shown in
Figure 3. The H2S04 tests were designed to support an evaluation
program of FGD operating parameters being conducted at the Shawnee
site. This pilot facility had two prototype FGD units utilizing
wet lime or limestone S02 scrubbing chemistry. The inlet flue
gas mass loadings were approximately 11.4 or 0.17 g/m3 depending
on whether the gas for the prototype FGD's was obtained directly
from the boiler or from the outlet of the ESP. Sulfur dioxide
concentrations varied from 2,000 ppm to 4,000 ppm at the inlet and
from approximately 400 ppm to 800 ppm at the outlet. Gas tempera-
tures and moisture percent varied from 165° to 121°C and 8% to 17%,
respectively, across the FGD unit.
Data were obtained from simultaneous inlet/outlet H2S04
measurements across the FGD unit taken over a period of approximately
30 days. The average inlet H2S04 value was 8.3 ppm (ranged from
0.4 to 24.8), while the average outlet value was 3.1 ppm (ranged
from 0.0 to 13.9). The average H2S04 removals by the FGD were
on the order of 60%, well below the SO2 removal efficiency seen
at Shawnee which typically ranged from 75% to 95%. The fact that
the H2SO4 removals did not parallel the S02 removal indicates that
the H2S04 does not exist as a true vapor in the FGD. In fact,
under the conditions that exist in the FGD (high humidity and
particulate matter for condensation sites), it seems reasonable
that S03 would exist as a liquid aerosol of H2SO4. The size of
these aerosols can be inferred from the aerodynamic sizing tests
(10) (11) conducted at Shawnee. These tests showed a mass removal
by the FGD of ~60% for particles in the range of ~0.5 p. Further
sizing/sulfate analysis studies will be necessary to determine
the exact size of H2SO4 aerosols.
Since the inlet H2S04 coefficient of variance (CVT) represents
228
-------
ADAPTER FOR CONNECTING HOSE
TC WELL
STACK
N>
ro
(O
RECIRCULATOR
THERMOMETER
RUBBER VACUUM
HOSE
ASBESTOS CLOTH
INSULATION
GLASS-COL
HEATING
MANTLE
DRY TEST
METER
THREE WAY
VALVE
SILICA GEL
co
Figure 3. Controlled condensation system field set-up.
-------
7 2 1 / 2
the sum, (CVT) = [(CVS) + (CVM) J , of source fluctuation
(CVS) and method error (CVM), the source fluctuation can be
estimated by assigning a coefficient of variance to the field
CCS. This coefficient of variance reflects errors in reading
temperature and pressures, accuracy of leak rates, as well as
rinsing recovery or titration errors which might affect the
ultimate calculation of the H2S04 concentrations. It is estimated
that the field accuracy of the CCS is _+ 11%, and thus the coefficient
of variance of the source is + 65%. Samples taken in the morning
and evening yield a variance of ± 32.4%. Using these values,
Figure 4 was generated. The use of this type of graph will
permit the most effective design of a test program to meet
accuracy requirements within the time and funding limitations.
Coke Oven Measurements
The same CCS was used to monitor the S02/H2S04 content of a
waste gas stream from a bank of coke ovens. In this application
a three point simultaneous test was performed across a charged
droplet scrubber (CDS) used to control coke oven particle
emissions. Two separate tests were performed under upset (during
coke oven charging, high mass loading) and non-upset (steady
state coking period, low mass loading) conditions. The goal of
these tests was to determine whether H2S04 was a major contributor
to a corrosion problem in the CDS. Consequently, sampling
positions were chosen at points where H2SO4 could be condensed.
Figure 5 shows the gas phase locations: at the inlet, after a
pre-queneh spray, and after the electrode system (50 feet down-
steam). Table 4 contains the results of the tests.
Table 4. S02 and H2S04 for Coke Oven/CDS Tests
H2S04
Condition Position
Upset Inlet
Pre-cooler
Exit
Non-upset Inlet
Pre-cooler
Exit
SO2
154 ppm
163 ppm
186 ppm
210 ppm
199 ppm
217 ppm
Based on
SOj
13
6
0
18
13
5
ppm
.4
.5
.1
.2
ppm
ppm
ppm
ppm
Based on
H+
12
17
6
18
13
4
ppm
ppm
.4
.2
.3
.9
ppm
ppm
ppm
ppm
230
-------
o
to
cs
X
O
LU
<
Q
LU
1/1
LU
LL.
O
u
u
LU
D_
X
234
SAMPLES PER DAY
Figure 4. Expected coefficient of variance (CV) of the H^S
measurement based on the number of samples taken.
231
-------
K)
CO
no
Spray *~C
Quench Sample
Coolant •-£ N v2
Inlet
Sample
?
9
J
9?
•iv .<»
?
U~
Quench Water
^
T .
Liquor
58
Outlet
I
Sample
Figure 5. Sampling positions at CDS unit.
-------
In the non-upset condition a steady decline in the H2S04
concentration across the CDS is seen, indicating a removal
efficiency of 72% in this low (<10% opacity) mass loading case.
The data for the upset condition are not as clear-cut, since
the SO4 and H"1" titration of the coil rinse did not agree. It
appears that another acidic species was condensed in the coil
along with H2S04. Visual observations in the field indicated
the coil rinse to be discolored instead of clear as in the non-
upset conditions. The plume during the upset condition was black
with coal and hydrocarbon aerosol by-products. The filter system,
which was heated to 200°C, probably allowed low vapor pressure
material to pass to the coil. While there was not enough solution
for further confirmatory analysis, it is possible that an organic
acid or phenol caused a positive interference with the Hf analysis.
Using the S04 titration values as the true H2S04 concentration,
the CDS was 100% effective in removing H2S04 during the upset
condition. The reason for this improved efficiency was probably
due to the presence of the large quantity of particles. The
H2S04, in the high humidity of the CDS, probably condensed on
the solid particles which grew in size because of the hydroscopic
nature of sulfuric acid. Once again the H2S04 appears to behave
more as a particle than as a gas. In this case, S02 (gas) was
not removed, while 60%-100% of the H2S04 was removed by the CDS.
SULFATE SIZE DISTRIBUTION TESTS
To gain an insight into the emission of sulfate aerosols from
an FGD, TRW conducted an aerosol sizing test at the outlet of a
utility boiler controlled with a soda ash FGD. These tests
consisted of two runs with a Meteorology Research, Inc. impactor
maintained at 121°C. This impactor has seven stages and a filter
providing a particle size distribution of ~0.5 to 30 \i. The runs
were conducted during times when the boiler was under full and
half load, so that a comparison of the size distributions could
be made. Figures 6 and 7 show the results of gravimetric and
sulfate analysis of each stage.
It is readily apparent that the bulk of the sulfate particles
collected were <1 p. Close inspection of the figures shows that
the particle size distribution was skewed to the smaller particles
when the boiler was under full load. However, from only these
two runs, it is not possible to determine whether boiler combustor
233
-------
10
35.0 17.5
6.2 2.8 1.7 0.8
AERODYNAMIC DIAMETER, f
0.54 <0.54
Figure 6. Size distribution of sulfate particles with boiler under
full load.
234
-------
N>
35.0 17.5 6.2 2.8 1.7 0.8 0.54 <0.54
AERODYNAMIC DIAMETER, M
Figure 7. Size distribution of sulfate particles with boiler under half load.
-------
conditions or FGD operating conditions caused the change in the
particle size distribution.
SUMMARY
The sampling of sulfate/sulfuric acid from combustion
processes poses several problems:
• Stability of bisulfate, sulfite, and bisulfite
compounds under conventional sampling approaches is not
adequate to maintain their identity.
• Extractive methods for sulfuric acid analysis encounter
neutralization problems during the separation of
particulate matter and the H2S04.
• Based on removal efficiencies, H2S04 acts more like
a small (<0.5 pi) particle than a gas in FGD units.
• Sulfate aerosol size distribution can vary significantly
with process conditions.
In order to learn more about the composition of sulfur species,
new techniques need to be developed. Only where in situ sulfur
species measurements can be made will an accurate picture of sulfur
chemistry be attained.
ACKNOWLEDGMENT S
The sulfate/sulfite stability and sulfate aerosol sampling
tests were conducted under EPA contract 68-02-21-1412 under the
direction of Dr. Robert Statnick. The sulfuric acid sampling
studies, Shawnee field tests, and the preparation of this paper
were supported by EPA contract 68-02-2165 under the direction of
Mr. Frank Briden. Both contracts were from the Process Measurements
Branch of IERL/RTP.
236
-------
REFERENCES
1. Corbett, P. F. A Photo-turbidimetric Method for Estimation
of S03 in the Presence of S02. J. Soc. Chem. Ind., 67:227,
1948.
2. Flint, D. Determination of Small Concentrations of S03 in
the Presence of Larger Concentrations of S02. J. Soc. Chem.
Ind., 67:2, 1948.
3. Fielder, R. S., P. J. Jackson, and E. Raask. Determination
of S02 and S03 in Flue Gases. J. Inst., Fuel, 33:84, 1960.
4. Environmental Protection Agency. Determination of Sulfuric
Acid Mist and Sulfur Dioxide Emissions from Stationary
Sources. Federal Register 41, (111), 1976. 23087.
5. Knol, B. P. Improvements in Determination of S03 and SO2 in
Combustion Gases. Riv. Combustible, 4:542, 1960.
6. Goksoyr, H., and K. Ross. The Determination of Sulfuric Trioxide
in Flue Gases. J. Inst., Fuel, 35:177, 1962.
7. American Society of Testing Materials. Part 26, ASTM
Method D3226-73T, 1974.
8. Lisle, E. S., and J. D. Sensenbaugh. The Determination of Sulfur
Trioxide and Acid Dew Point in Flue Gases. Combustion,
36:12, 1965.
9. Driscol, J. N., and A. W. Berger. Improved Chemical Methods
for Sampling and Analysis of Gaseous Pollutants from Combustion
of Fossil Fuels. Volume I, Sulfur Oxides. Walden Research
Corporation, PB 209-267, June 1971.
10. Maddalone, R. F., A. Grant, D. Luciani, and C. Zee. Procedures
for Aerosol Sizing and S03 Vapor Measurement at TVA Shawnee
Test Facility. EPA Contract 68-02-2165, Task 2, TRW DSSG,
January 1977.
11. Rhudy, D., and H. Head. Results of Flue Gas Characterization
Testing at the EPA Alkali Wet-Scrubbing Test Facility. Presented
at: Second EPA Fine Particle Scrubber Symposium, EPA 600/2-77-
193, Sept. 1977.
237
-------
Operating Parameters Affecting Sulfate
Emissions from an Oil-Fired Power Unit
Russell N. Dietz
Robert F. Wieser
Leonard Newman
Brookhaven National Laboratory
ABSTRACT
Any voluntary or legislated action taken to control
sulfates should be based at least in part on a thorough
knowledge of the character of primary sulfates (i.e.,
H2SO4 and water-soluble sulfate salts) and on an
understanding of the principal variables that govern
the magnitude of those emissions.
A system comprising isokinetic flue gas sampling for
particulate sulfate on an in situ quartz fiber filter
assembly, followed by controlled condensation for H2S04
collection, was used primarily at a high (2.5%) sulfur
content oil-fired power unit. Speciation during collec-
tion demonstrated that the ESP, depending on efficiency,
reduced particulate sulfate generally by 50% to 90%.
Particulate sulfates at a low (0.3%) sulfur content
oil-fired unit decreased almost in proportion to the
decrease in sulfur content of the oil. Sulfuric acid
concentrations were found to correlate well with excess
furnace 02 over the range investigated (0 to ~2% 02).
A good correlation was also found for particulate
sulfate and H2SO4 at the ESP inlet, with an indication
of sulfate formation controlled both in the flame
region and in the high temperature heat transfer region.
Elemental and carbon analyses indicated that the principal
metals in the soluble fraction (Mg and V) nearly accounted
for the total measured soluble sulfates. The insoluble
fraction was composed primarily of MgO and carbon.
239
-------
INTRODUCTION
Any voluntary or legislated action taken to control sulfates
should be based, at least in part, on a thorough knowledge of
the character of primary sulfates, their emission rates from
power plants, institutional and industrial boilers, and apartment
and home heating units, and an understanding of the principal
variables which govern the magnitude of those emissions. Both
for consideration of potential health effects and for determina-
tion of the mechanisms and parameters which affect the magnitude
and distribution of such emissions, the character of primary
sulfates (i.e., H2SO4 and water soluble sulfate salts) must be
determined with flue gas sampling methods that differentiate
between the acid form and the less nocuous sulfates.
A reliable sampling method, utilizing a Brookhaven-designed
nozzle and filter assembly (1) for collection, in situ, of flue
gas particulates (including water-soluble metal sulfates), fol-
lowed by a version of the Goksoyr-Ross (2) condenser coil for
separate collection of the flue gas sulfuric acid, has been lab-
oratory and field tested at three different power plant units
(1)(3). Validation of the methodology has been described (Ib) and
will be reported in detail elsewhere (4).
The range of emissions of sulfuric acid and particulate
metal sulfates at three commercial power plant units will be
presented in this paper. The principal plant operating para-
meters controlling the magnitude of their emission rate, including
furnace oxygen, sulfur, and vanadium content of the fuel, and
electrostatic precipitator efficiency, have been correlated with
emission concentration. Details on the elemental distribution
of carbon and the principal metals between the soluble and in-
soluble particulate fractions will be presented.
With more than one year of sampling experience at the Long
Island Lighting Company's (LILCO) Northport Power Station, Unit
3, it was concluded that the full-scale (365 MW) unit, burning
fuel oil with an average of 2.4% sulfur and 350 ppm vanadium,
was capable of continuous operation with less than 1 ppm of
H2SO4 and less than 2 ppm of particulate metal sulfates in the
flue gas emitted to the stack. Thus, with the proper utilization
of furnace and emissions controls, the results in this paper show
that even potentially high sulfate emitting oil-fired sources
can be readily controlled to less than 0.2% of the sulfur in the
fuel emitted as H2S04 and metal sulfates.
240
-------
EXPERIMENTAL
The Sampling Apparatus
Flue gas sampling was performed primarily with the Brookhaven
controlled condensation system (CCS) with some qualified measure-
ments using the Brookhaven version (5) of the Modified (6) EPA
Method 6. Basically the CCS (la) consisted of an in-situ filter
for particulates located directly behind the isokinetically sized
nozzle, a partially heated glass probe terminating in a 17-turn
6 mm glass coil maintained at 140°F for collection of the H2S04
aerosol, a back-up Pyrex wool plug, a 10-turn coil and receiver
vessel maintained at ice water temperature for condensing most
of the water vapor, two impingers containing peroxide for collec-
tion of S02, and finally, a dryer, pump, and dry test meter. A
critical orifice and pre-evacuated 1 liter bottle was located
between the pump and dry test meter for subsequent chromato-
graphic determination of the flue gas oxygen and carbon monoxide
levels at the sampling location.
Analytical Procedures
Specific details on the work-up of the collected material
can be found elsewhere (1). The total particulate load on the
filter was determined gravimetrically. Then, any free H2SO4
was recovered from the filter with a 100% isopropyl alcohol (IPA)
wash, and the soluble metal sulfates were recovered with a water
wash (5% IPA); negligible sulfur remained on the filter, which
was weighed to determine the total mass of the insoluble parti-
culates. The water wash solution was analyzed initially in the
program by the Autoanalyzer turbidimetric procedure for total
sulfate and later by the more sensitive ion chromatographic
approach. A portion of about one-third of the solutions were
also analyzed for the principal metals by atomic absorption spec-
troscopy. The insoluble fraction remaining on the quartz filter
was analyzed for carbon and the principal metals by grinding
uniformly in a mortar with pestle prior to determination.
The only sulfate fraction to pass the filter, i.e., sulfuric
acid, was recovered separately by washing each main section of
the sampling system—namely, the probe, the acid condensation
coil, and the final filter plug. Generally more than 90% of the
H2SO4, determined by titration with 0.02 N NaOH with some con-
firmatory analyses by ion chromatography, was contained on the
combination of the probe and the acid, condenser coil.
241
-------
Leaks in the flue ducting were taken into account by com-
paring the oxygen content at the flue gas sampling location with
that in the furnace. The chromatographically determined CO con-
tent was used to infer the relative level of furnace oxygen.
Sampling Locations
From November 1976 through March 1978, 81 flue gas sampling
runs were performed, 60 of which were performed at LILCO North-
port Unit 3. The fuel for the 365 MW oil-fired unit typically
contained 2.4% sulfur and 350 ppm of vanadium; the latter was
previously shown (7) to be responsible, in part, for the magni-
tude of the sulfuric acid emissions.
The two sampling locations at Northport Unit 3 were just
prior to the electrostatic precipitator (ESP) and just after the
induced draft (I.D.) fan on the outlet side of the ESP (Figure 1).
Unit 2 at Northport was essentially the same as Unit 3 with
the exception that there was no ESP. (A precipitator is cur-
rently being installed, and the unit is projected to be back on-
stream with the ESP some time after June.) Sampling was per-
formed just at the point where the duct turned into the stack.
The third unit sampled, the Barrett Station Unit 1 at Island
Park, New York, was also a tangentially fired boiler, but the
fuel was of low sulfur (~0.3% S), low vanadium (~15 ppm V)
type. Sampling was also performed at the exit of the induced
draft fan. The unit had an ESP, but it was not in operation.
The Barrett unit had at one time burned coal but is now exclu-
sively fired with low sulfur oil as mandated by New York City.
All three units had Liqui-Mag (MgO) added for corrosion protec-
tion.
FLUE GAS SAMPLING RESULTS
Because sampling was performed at three different units, the
results will first be presented according to the unit sampled.
Subsequently, comparisons will be made of the statistical distri-
bution or range of emissions between each unit.
242
-------
CO
TEST PORT
NORTHPORT
TEST PORT UNIT 3
TEST PORT ARRANGEMENT
SECONDARY AIR DUCT
ELEVATION VIEW
Figure 1. Schematic of the sampling port arrangement at Northport Unit 3.
-------
NORTHPORT UNIT 3
The furnace oxygen levels, estimated from consideration of
plant measurements as well as the bottle CO levels, varied from
0.0% to 1.1% with an average of 0.25%—indicative of the very
close control of combustion air to the stoichiometric require-
ments. The fuel sulfur level of about 2.4% gave expected S02
levels of about 1460 ppm; the measured levels were generally
within 3% of the calculated amount. Similarly the measured
amount of flue gas water vapor was usually within 5% of the ex-
pected amount. The successful attempt at a mass balance for
those two compounds placed a certain degree of confidence on the
measured levels of sulfuric acid, metal sulfates, and total
particulates which, of course, cannot as yet be predicted by any
reliable means.
Sulfuric Acid Emissions(Effect of Furnace Oxygen)
Furnace 'oxygen was shown to play a key role with the level
of H2S04 found in the flue gas as shown in Figure 2. A least
mean square fit of the data gave
[H2SO4] = 0.5 + 10.8 [02] [1]
where [H2S04] = sulfuric acid concentration, vol. ppm
[02] = furnace oxygen, percent
with a coefficient of determination (i.e., the correlation co-
efficient, squared) of 0.92, indicating a reasonably good corre-
lation of acid with oxygen level in the furnace. There appar-
ently was no effect whether the sampling was done at the inlet
or outlet of the ESP.
Particulate Metal Sulfates (Effect of ESP)
As anticipated, although the ESP had no effect on H2S04,
there was a significant effect on the amount of emissions of
total particulates, including the water-soluble metal sulfate
fraction, as shown in Table 1. In April, with only half the
modules in operation at a total power level of 68 KW, the mea-
sured total particle removal efficiency was 63% based on the
measured outlet and inlet particle level.
244
-------
o ESP OUTLET
• ESP INLET
O.I 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
FURNACE 02, %
Figure 2. Effect of furnace oxygen on sulfuric acid levels at
Northport Unit 3.
245
-------
Table 1. Effect of Precipltator
(Average Sampling Results at Northport Unit 3)"
ro
-p*
CD
Date
4/20/77
4/21/77
5/25/77
5/24-AM
5/24-PM
6/24/77
6/23/77
7/20/77
7/18/77
7/19/77
10/7/77
10/4/77
10/5/77
12/14/77
12/15/77
12/13/77
Location
to ESP
Inlet
Outlet
Inlet
Outlet
Outlet
Inlet
Outlet
Inlet
Outlet
Outlet
Inlet
Outlet
Outlet
Inlet
Inlet
Outlet
Estimated
Furnace H2S04
02 % vol . ppm
0.6
0.1
0.0
0.1
0.0
0.6
0.8
0.2
0.2
0.2
0.0
0.2
0.1
0.1
0.3
0.1
6.2
1.8
<0.02
1.4
0.3
7.0
9.6
1.8
3.5
3.6
0.1
4.5
1.4
0.3
3.8
0.5
Soluble Part. S04
vol. ppm
15.8
5.2
9.7
3.1
1.1
17.2
1.1
12.1
9.2
8.8
7.0
4.6
2.6
6.0
13.9
4.1
mg/m3
64.0
21.1
39.2
12.5
4.5
69.7
4.6
49.1
37.3
35.8
28.5
18.8
10.6
24.5
56.3
16.7
Total S04
% of fuel S
1.46
0.49 ,
0.66
0.31
0.10
1.67
0.77
1.01
0.84
0.91
0.48
0.62
0.25
0.42
1.21
0.30
Total
Part. ,
mg/m3
157
59
157
23
14
112
6
149
107
73
243
97
35
c
c
c
Electrostatic Precipitator
Efficiency,
or
63
86
91
95
28
51
60
86
--
Modules
on %
50
100
100
88
62
62
62
75
75
Operating
Power , KW
68
117
123
118
b
b
72
81
b
.Each value is generally the average of 3 or more runs.
Data not available.
Gravimetric particulate weights were inappropriately measured.
-------
In May about three weeks after the unit had been thoroughly
cleaned, all modules were operating at a total power level of 117
KW to 123 KW, and efficiency had increased to 86%-91% removal of
total particulate. A month later, even though operating with one
module off but at the same total power level, efficiency had climbed
to 95%. The higher efficiency, in spite of the lower inlet partic-
ulate load which should probably have reduced efficiency somewhat,
was probably due to the presence of a larger concentration of H2S04.
Experiments have shown that the addition of SO3 or H2S04 to flue
gas streams with normally low acid levels has resulted in improve-
ments in the operation of electrostatic precipitators (8)(9). How-
ever, the improved operation of the precipitator with concurrent
reduction in sulfate particulate emissions did not off-set the
increased emission of H2S04. Thus, total sulfate emissions, i.e.,
H2S04 plus particulate sulfate, were higher in June.
Only five of the eight precipitator modules were operating
during the July runs, total particulate removal efficiency was the
lowest of all experiments, and particulate metal sulfate emissions
were the highest. In October the ESP was operating only slightly
better than in July.
Particulate metal sulfate correlated well with the measured
ESP efficiency as shown in Figure 3. From the ordinate on the right
it can be seen that the metal sulfate level was reduced to less
than 1.5 vol. ppm when the ESP efficiency exceeded about 90%. Thus,
with the ESP operating as it was designed, particulate metal sulfate
emissions from even a moderately high sulfur-containing fuel can
be kept below an emission level of 0.1% of the sulfur in the fuel.
All of the precipitator inlet data from April through July
(Data for October and December had not been available yet) were
used to plot the ESP inlet particulate metal sulfates versus the
measured sulfuric acid level as shown in Figure 4. The slope of
nearly unity represented a one-to-one relationship between changes
in sulfuric acid levels and corresponding changes in particulate
metal sulfates from the boiler. Thus there was apparently a region
of the boiler in which the formation of sulfuric acid caused a
corresponding increase in metal sulfates. The intercept at 10.2
ppm of metal sulfates at zero concentration of H2S04 implied the
existence of another boiler region responsible for the formation
of particulate metal sulfates independent of H2SO4 levels—i.e.,
independent of furnace oxygen levels. These observations were only
possible because of the utilization of the Brookhaven controlled
condensation system which reliably separated the H2SO4 from the
metal sulfates during the sampling procedure. Further discussion
will be given in a later section.
247
-------
100
10 20 30 40 50 60 70 80 90 100
MEASURED ELECTROSTATIC PRECIPITATOR EFFICIENCY, %
Figure 3. Effect of electrostatic precipitator on the emission
of particulate metal sulfates.
248
-------
0.94(H2S04)
r2=0.87
1
1
I 2345678
PRECIPITATOR INLET H2S04,vol. ppm
Figure 4. Correlation between particulate metal sulfates and
sulfuric acid at the ESP inlet of Northport Unit 3.
249
-------
NORTHPORT UNIT 2
Because this unit did not as yet have an electrostatic precipi-
tator, which is currently being installed, it was necessary to
operate generally with furnace oxygen in excess of 1.5% in order to
stay below the opacity emissions standard. As will be seen, this
condition resulted in order of magnitude higher H2SO4 emissions.
Sulfuric Acid Emissions (Effect of Furnace Oxygen)
Seven sampling runs at this unit were performed with furnace
oxygen variable from 1.2% to 2.7%, resulting in H2SO4 concentra-
tions ranging from 20 ppm to 40 ppm as shown in Figure 5. Nine
sampling runs were performed in April 1978 during which the furnace
oxygen varied from 0.4% to 1.0%. Although not shown, those lower
sulfuric acid levels indicated the curve went through the origin.
The data for Unit 3 were shown for comparison. There was a
higher slope (by about a factor of 1.8) for the data at Unit 2
compared to Unit 3, and the similarities between the units precluded
all but one possible explanation. Since the fuel vanadium levels
were identical, that key ingredient could not be the answer.
Effect of Fly Ash Recirculation—The most probable cause for
the lesser dependence on oxygen may be related to the recirculation
of the ESP-collected fly ash back into the furnace in the case of
Unit 3. The region suspected of being responsible for the sulfuric
acid dependence on furnace oxygen levels was the surfaces of the
superheater and reheater tubes in the boiler which became contin-
uously coated with catalytically active (3)(10) vanadium-containing
deposits from the oil ash. Based on typical ash determinations,
considering that about half the oil ash was deposited as bottom ash
(11), and assuming that typically 100 mg/m3 of fly ash were recir-
culated back to the furnace, the ash coating on the tube surfaces
in the case of recirculation (Unit 3) was probably comprised to a
greater extent of less chemically active material than in the case
of a once-through (Unit 2) system.
Future measurements at Unit 2 after the ESP has been installed
may confirm this hypothesis.
250
-------
50
Q.
O.
CJ
X
(f)
<
CD
LJ
40
30
20
10
0
0
NORTHPORT
UNIT 2
NORTHPORT
UNIT 3 (ESP inlet)
FURNACE 02, %
Figure 5. Effect of furnace oxygen on sulfuric acid levels at
Northport Unit 2.
251
-------
Particulate Metal Sulfates (No ESP)
Because there was no electrostatic precipitator and because
there was generally an order of magnitude higher H2804 level in
the flue gas of Unit 2 compared to Unit 3, much higher particulate
metal sulfates might have been expected. However, the total amount
of metals present, including oil ash and MgO additive, limited the
metal sulfates to only a 50% increase above that typically at Unit
3.
BARRETT UNIT 1
Although the furnace oxygen levels when sampling at Barrett
were generally in the range of 0.5% to 1.0%, the low sulfur content
of the fuel, coupled with the low vanadium content of the oil ash,
resulted in H2S04 levels less than 0.2 ppm. Particulate metal
sulfates ranged from about 2 ppm to 4 ppm.
Because the levels were of such a reduced magnitude compared
to those at Northport, it was not possible to deduce the effect of
furnace oxygen.
OVERALL RESULTS AT ALL UNITS
Since a large number of measurements had been made, an easier
way to view the general range of emissions from each of the three
units was in the form of histograms or frequency-distribution diagrams
as shown in Figures 6 through 9.
Northport Unit 3 Distribution
Figure 6 shows that the range of sulfuric acid flue gas con-
centrations at Northport Unit 3 was from 1 ppm to 11 ppm at the
electrostatic precipitator outlet and from 1 ppm to 9 ppm at the ESP
inlet, that is, independent of the ESP. Similarly, the average at
the outlet (3.3 + 2.6 ppm) was essentially identical to that at the
inlet (3.2 + 2.8 ppm). Thus, as expected, the ESP had no effect on
the flue gas vapor phase sulfuric acid content.
On the other hand, the particulate metal sulfate did reflect
the presence of the precipitator. At the ESP inlet (lower histogram
in Figure 7), the range of metal sulfates was from 5 ppm to 19 ppm
with an average of 11.9 + 4.2 ppm. At the outlet (upper histogram),
252
-------
\c.
10
8
6
M-
<*_*
v> 4
iii
i&j
O
•y
LJ o
o: ^
tr
Z>
8 o
o
U_
o 10
8
6
4
2
r*
_
/
—
1
H9SO^ =3.3 DDm
i
C- -T • •
er = 2.6 ppm
I
i i i i i
1 3 57 9 II 13 15
H2S04,vol.ppm (ESP OUTLET)
-
_
—
—
—
__l
H2S04 = 3.2 ppm
o- = 2.8 ppm
_l 1
l_
_J !_., 111.-
3 579 II 13 15
H2S04,vol.ppm (ESP INLET)
*•
ESP inlet.
253
-------
Jo. OF OCCURRENCES (f)
^ O ro & CD oo o
—
-
I
1
1
PART SO^ = 4.6 ppm
-------
UJ
o
z
UJ
o:
•D
o
o
O
O
6
Z
10
8
6
TOTAL 30= =0.56%
a- =0.26%
Ql I I I I I I I I I I I
O.I 0.3 0.5 0.7 0.9 I.I 1.3 1.5 1.7 1.9 2.1
TOTAL SULFATE, % OF FUEL S (ESP OUTLET)
8
6
4
2
n
TOTAL SO^ =1.03%
a- =0.50%
-
1
1
1
i
1
1
1
1
1
i
O.I 0.3 0.5 0.7 0.9 I.I 1.3 1.5 1.7 1.9 2.1
TOTAL SULFATE, % OF FUEL S (ESP INLET)
Figure 8. Histograms of total sulfate emissions at Northport
Unit 3: (upper) at the ESP outlet; (lower) at the
ESF inlet.
255
-------
99S
No. OF OCCURENCES (f)
D — rv> GJ -£> ui O • — ro GJ .£. 01
Ho
-
1
5
so4
-------
the range was less—from 1 ppm to 11 ppm—as was the average 4.6 +
2.9 ppm. Thus, on the average, the ESP reduced metal sulfates by ~
61%.
Total sulfates, i.e., the sum of the H2S04 and the metal
sulfates, expressed as a percentage of the sulfur in the fuel, are
plotted in Figure 8. At the ESP inlet (lower curve), the range was
much broader than at the ESP outlet. On the average, the total
sulfate emissions to the stack were 0.56 + 0.26% of the sulfur in
the fuel.
Northport Unit 2 Distribution
The upper portion of Figure 9 shows the histograms for the
concentrations of H2SO4 and metal sulfates in the flue gas at
Northport Unit 2 as well as the total sulfate emissions as a per-
centage of the fuel sulfur. Comparing the average H2S04 emissions
at Unit 2 (31 _+ 7 ppm) with those at Unit 3 (3.3 +• 2.6 ppm), nearly
an order of magnitude higher acid emission was present at Unit 2.
The metal sulfate concentration (19 +_ 5 ppm) was only about 50%
higher than that at the ESP inlet of Unit 3 but 4 times higher than
the metal sulfate entering the stack at Unit 3.
Role of the Northport ESP—Both the much higher acid levels
and the somewhat higher metal sulfate levels at Unit 2 compared to
Unit 3 indicated that the ESP played a very important role—a some-
what direct role with respect to the metal sulfate emissions and an
indirect role with respect to the acid. Burning the fuel oil at a
furnace oxygen level of 0.2 + 0.3%, as was the case for Unit 3,
resulted in a very heavy soot load when nearly the same oxygen level
was used at Unit 2 (April 1978 tests not yet available). The resul-
ting black plume, which was at the threshold of opacity emission
standards, was not especially aesthetically appealing. In order to
reduce the visibly dark plume, the unit without the precipitator
normally resorted to combustion at the moderate level of furnace
oxygen of about 1.0% to 1.5% 02 . Thus the amount of oxygen avail-
able for increasing the rate of catalytic formation of H2SO4 was
generally from 4 to 6 times higher at the uncontrolled unit compared
to that with the ESP. In addition, recirculating the fly ash removed
by the ESP back into the furnace apparently reduced the overall
catalytic activity of the surfaces by about a factor of two.
Thus the ESP allowed the emissions of total sulfate, on the
average, to be reduced from 4.1% to 0.56% of the sulfur in the fuel—
a 7-fold reduction.
257
-------
Barrett Unit 1 Distribution
The sulfuric acid emissions at Barrett, as shown in the lower
portion of Figure 8, were remarkably lower than those at Northport
Unit 2—0.10 + 0.07 ppm compared with 31+7 ppm, respectively—that
is, over 300-fold lower. Yet, the particulate metal sulfates were
only about 7 1/2-fold lower (2.5 +_ 0.8 ppm) than at Unit 2 (19 +_ 5
ppm).
Role of Fuel Sulfur and Vanadium—Assuming that flue gas sul-
furic acid was primarily controlled by and directly dependent on
the sulfur content of the fuel, the vanadium content of the fuel,
and the furnace oxygen level, then the ratio of the product of those
three variables to the average H2S04 concentration should be about
the same for each unit. As shown in Table 2, the aforementioned
ratios were, indeed, very nearly identical, giving strong weight
to the direct dependence of H2S04 on sulfur and vanadium content.
Table 2. Effect of Fuel Sulfur and Vanadium
Average Fuel Oil Content Average Average Ratio of
Sulfur,Vanadium, 02, H2S04, Product
% Product9- ppm to H,S04
Unit
a
ppm
Northport 2.1
Unit 2
Barrett 0.3
Unit 1
390 2.0 1638 31 53
15 1.0 4.5 0.10 45
Product of sulfur content times vanadium content times
furnace oxygen.
Future studies should be conducted at one unit for which the
sulfur, vanadium, and furnace oxygen could be varied in a systematic
fashion.
ANALYSIS OF COLLECTED PARTICULATES
To obtain a better understanding of the mechanisms and
variables affecting the emissions of H2S04 and particulate metal
sulfate, a selected number of the water wash solutions from the
258
-------
nozzle filters, as well as the remaining participates on the fil-
ter, were analyzed for the principal metal content. Material
balances based on the stoichiometry for the metal sulfates and
oxides were performed, and the information was used to do a
metals inventory on the power plant. Examples of those results
are shown in the following section. Complete details are avail-
able in other reports (la)(3a).
Particulate Material Balances (Water Soluble)
Determinations of some of the water washings containing dis-
solved metal sulfates were made for the principal metals—
typically Mg, V, and Na with smaller amounts of Ni and Ca and
traces of iron—as shown in Table 3. Assuming that the stoichio-
metric amount of sulfate with each element corresponded to that
for the compounds shown in the table, the sum of those amounts
was shown to differ from the total sulfate measurement by gen-
erally no more than +_ 10%.
By summing the individual compound concentrations, com-
parison was also made with the total water-soluble particulate
concentration determined by difference between gravimetrically
determined total particulate and insoluble particulate. The
agreement in the last two columns was very good.
The choice of compound formulations for each of the sulfate
salts was unambiguous except for nickel and iron—the lower (-ous)
valencies were assumed. For vanadium, the principal sulfate has
been shown to be vanadyl sulfate, VOSO4 (12). The hydrate formu-
lations were assumed to be the highest hydrate stable at the
conditioning humidity of the filter assemblies using data available
in the literature (13)(14). The hydrates of MgSO4 were the most
stable with the presence of MgS04. GH^O having been demonstrated
by other workers using X-ray diffraction (15).
Particulate Material Balances (Water Insoluble)
As shown in Table 4, the principal insoluble metals also
involved Mg and V with smaller amounts of Ni and Fe and traces
of Na and Ca. However, carbon contributed substantially to the
insoluble fraction, generally the largest of any element present.
The sum of the stoichiometrically determined metal oxides plus
carbon differed from the gravimetrically determined total insolu-
ble particulate by generally less than + 30%; the sum was
probably the more accurate number.
259
-------
Table 3. Detailed Analvses of Soluble Particulates
CT>
o
Flue
Gas,
Run° ra3
PR-31-0 0.3484
34-0 0.2929
36-0 0.2973
37-1 0.2246
Mg
(S04)C
[MgSCUT
4.74
(18.7)
[37.5]
4.85
[19.2]
[38.4]
4.49
(17.8)
[35.5]
7.06
(27.9)
[55.8]
V
(S04)
[voscv
1
(1
[3
2
(4
[6
3
(6
[11
5
(9
[16
.02
.9)
.3]
.10
.0)
.7]
.53
.7)
.3]
.12
.7)
.4]
Na
(S04)
1 [Na2S04]
2
(5
[7
3
(6
[10
2
(5
[8
3
(7
[11
.38
.0)
•4]
.23
.8)
.0]
.64
.5)
-2]
.74
.8)
.6]
Ni
(S04)
[NiS04 ]
0
(0
[1
0
(0
[1
1
(1
[3
0
(1
[2
.52
.9)
.4]
.53
• 9)
.4]
.13
.8)
.0]
.85
.4)
.2]
Fe
(S04)
[FeS04]
0
(0
[0
0
(0
[0
0
(0
[0
0
(0
[0
.04
•1)
.2]
.03
•1)
.1]
.04
•1)
.2]
.03
.1)
.1]
Ca
(S04)
(CaS04)
0.86
(2.1)
[2.9]
1.11
(2.7)
[3.8]
_
-
-
_
-
—
Total Sulfate Total Soluble
2S Meas.f 2g Meas.h
(28.7) (32.0)
[52.7] [53 .5 ]
(33.7) (35.7)
[60.4] [62.1]
(31.9) (30.0)
[58.2] [56.2]
(46.9) (45.0)
[86.1] [73.8]
All concentrations in mg/m of flue gas determined from elemental atomic adsorption
nO after run no. represents sampling performed at Xorthport electrostatic precipitator outlet;
^Calculated stoichiometric amount of sulfate
Calculated as MgS04-4H20; all other sulfate salts were anhydrous at desiccator humidity
^3ur. of the stoichioraetric sulfates
From ion chromatograph total soluble sulfate determinations
^Sur. of the stoichiometric metal sulfate salts
"Difference between gravimetrically determined total particulate and insoluble particulate
I, inlet
-------
Table 4. Detailed Analyses of Insoluble Particulates
O)
Flue
b Gfs-
Run m
PR-31-0 . 0.3484
34-0 0.2929
36-0 0.2973
37-1 0.2246
Total Insoluble
Mg „
(MgO)C
4.84
(8.0)
3.65
(6.0)
3.33
(5.5)
11.57
(19.2)
V
(V02)
5.10
(8.3)
4.73
(7.7)
3.00
(4.9)
10.83
(17.6)
Na Ni
(?) (NiO)
0.36
(0.5)
0.39
(0.5)
0.31
(0.4)
2.00 1.75
(3.0) (2.2)
Fe
(Fe203)
4.93
(7.0)
4.14
(5.9)
1.39
(2.0)
2.85
(4.1)
Ca
(Ca02)
0.19
(0.3)
0.15
(0.3)
0.27
(0.5)
1.05
(1.9)
vd
Carbon *•
(22.5) (46)
(12.5) (33)
(13.4) (28)
(32.5) (80)
Meas.6
(68)
(41)
(8)
(64)
.All concentrations in mg/m3 of flue gas determined fron elemental atomic absorption (carbon by IR-CO,)
"See Table 3A
^Calculated stoichiometric amount of metal oxides
Sum of the stoichiometric metal oxides plus carbon
Gravimetrically determined insoluble particulate
-------
Power Plant Metals Inventory
Based on the fuel oil ash analyses and the Liqui-Mag (MgO)
additive rate, it was possible to calculate the expected furnace
gas concentration of the principal metals as shown in Table 5.
At the Northport Unit 3, the soluble and insoluble metal deter-
minations were added together and averaged at both the inlet
and outlet of the ESP as shown for the July and October 1977
runs.
For the July runs, a major fraction (71%) of the MgO was re-
moved by the furnace system (as bottom ash and as fly ash tube
deposits) since the Mg decreased from 65.2 to 18.6 mg/m3. Of
this latter amount entering the ESP, about 54% was removed (i.e.,
another 15% of the furnace MgO) even though the ESP was func-
tioning only at 40% efficiency. Thus the total amount of the
injected MgO retained in the unit was 87%. As stated earlier,
Liqui-Mag (MgO) was added to the fuel primarily to inhibit high
temperature wastage and low temperature corrosion, but, at North-
port, also aided substantially in the recovery of generally over
80% of the salable vanadium from the oil (11).
Although the boiler portion of the system retained 72% of
the Mg and the ESP another 15% (for a total of 87%), the boiler
only retained 34% of the vanadium and the ESP another 39% for a
total of 73% retained. In October, when the ESP was found to be
much more efficient, the portion of Mg retained by the boiler was
the same (70%), but that by the ESP climbed to 26% for a total of
96% retained in the unit; with respect to Mg, the ESP was 86%
efficient. However, the vanadium retained by the boiler and ESP
combined was not much higher than in July, primarily because the
efficiency of the ESP with respect to vanadium was only 64%.
This difference in ESP efficiency for Mg and V particulates, 86%
and 64%, respectively, implied that the V particles must have
been substantially smaller in size.
Further details on the distribution of the elements will be
presented elsewhere (3a).
262
-------
Table 5. Power Plant Metals Inventory
rv>
Power Plant
Northport
(Unit 3)
July 1977
Northport
(Unit 3)
Oct. 1977
Barrett
(Unit 1)
Aug. 1977
Flue Gas
Region
Q
Furnace
ESP Inlet
ESP Outlet
(% Retained)
Furnacea
ESP Inlet
ESP Outlet
(% Retained)
Furnacea
Stack
(% Retained)
0
Gas Concentration, mg/m
Mg
65.2
18.6
8.6
(87)
35.5
10.8
1.5
(96)
7.5
1.2
(84)
V
24.4
16.0
6.5
(73)
30.0
17.2
6.2
(79)
0.32
0.43
(0)
Na
9.4
5.7
2.8
(70)
6.3
5.0
1.9
(70)
3.3
1.6
(51)
Ni
3.2
2.6
1.1
(66)
(3.2)
2.1
0.5
(86)
1.4
0.3
(79)
Ca
1.15
1.1
1.2
(0)
(1.2)
0.8
0.3
(71) ~
1.2
— — —
Efficiency
Esp, %
40
73
none
Furnace gas concentration of metals was calculated from oil ash analysis and MgO
additive rate.
-------
PARAMETERS AFFECTING SULFATE EMISSIONS
In the preceding sections, the principal variables affecting
either sulfuric acid or sulfate formation and emission levels
were shown to be
1. Furnace 02 H2S04 (|) and SO4 (>f)
2. ESP S04 (f) only
3. Fly ash recirculation H2S04 (|) and S04 (?)
4. Sulfur in oil H2S04 (|) and S04 (?)
5. Vanadium in oil H2S04 (|) and S04
6. Other metals in oil H2S04 (?) and S04
For example, increasing furnace 02 had direct effects on in-
creasing sulfuric acid and metal sulfates. The electrostatic
precipitator directly affected metal sulfates only; the higher
the efficiency of the ESP, the lower the metal sulfate emissions.
Indirectly, the ESP at Northport appeared to be responsible
for a reduction in H2S04 emission because of reduced catalytic
activity of the recirculated fly ash. Increasing vanadium
apparently had a direct effect on increasing catalytic activity
and thus increasing H2S04 and metal sulfates. The level of
total metals in all probability controlled the level of total
metal sulfates.
Some additional factors which were not yet investigated but
are to be considered in upcoming measurements (3a) were the
Liqui-Mag additive rate, the power level, and the burner tilt in
the furnace.
Since increasing the MgO content would increase the total
metal content in the flue gas, quite possibly the metal sulfates
might increase. Counteracting that, however, would be the
diluting effect of the MgO on the concentration of vanadium
available for catalytic formation of H2SO4 and metal sulfates.
Thus, the net effect of increasing MgO would probably be to
reduce H2S04 and metal sulfate emissions; such results have
been noted by others (3b).
All else being equal, reducing power level, or load, has
been shown to have a disproportionately higher reduction in total
sulfate emission (10)(11). The precise effect on H2SO4 and metal
sulfates will be assessed in future measurements.
264
-------
Finally, reducing burner tilt has been shown to have a
slight but perceptible decrease on total sulfate emissions (16)
Again, there was no distinction made between sulfuric acid and
metal sulfates.
COMPARATIVE STUDIES
Although there have been numerous studies conducted over the
years on the magnitude of the emissions of metal sulfates and
sulfuric acid from oil-fired power plants, there are two results
of interest on the determination of metal sulfates that will be
reported here for comparison with the Brookhaven measurements.
The EPA has been conducting studies (17) at several oil-
fired power plants using the EPA Method 5 (18) for the deter-
mination of metal sulfates. Their results were plotted in Figure
10 as total metal sulfate concentration versus furnace oxygen as
well as all the Brookhaven data for the Northport Unit 3 mea-
surements at the ESP inlet and that obtained at Barrett. The
Brookhaven results at Northport were in qualitative agreement
with those determined by EPA at plants Al, A2, and W. Similarly,
the Brookhaven results at Barrett were in excellent agreement
with the EPA results at a very similar plant M.
The results depicted in Figure 10 confirm the existence of
an intercept for particulate metal sulfate formation at zero
furnace oxygen, i.e., in the absence of sulfuric acid. The
figure also demonstrated the dependence of metal sulfates on fur-
nace oxygen levels as well as the sulfur and vanadium content of
the fuel. The magnitude of the metal sulfate level at Barrett
and plant M was lower than that at the other plants because the
fuel sulfur content was less. The dependence of the metal sul-
fate concentration on furnace oxygen level, that is, the slope
of the line, was much less than the others because of the much
lower vanadium content.
A study was performed by LILCO (19) at the Barrett Unit 2
about nine months prior to the Brookhaven measurements at Barrett
Unit 1. LILCO found a total sulfate level of 1.9 +_ 0.3 ppm,
quite in line with the Brookhaven value of 2.5 + 0.8 ppm, espe-
cially when consideration was given to the fact that Unit 1 had
no operating particulate controls, but Unit 2 had an operating
cyclone where all the collected particulates were re-injected
back into the furnace. All other conditions were the same.
265
-------
24-
E
Q.
Q.
O
LJ
LJ
=>
O
h-
o:
O
LJ
O
cr
h-
z
O
O
w
600 MW
2.2%S
450 ppm V
NORTHPORT
UNIT 3
350 MW
2.4%S
rA2
525 MW
2.4% S
600 ppm V
Al
300-560 MW
2.4% S
200 ppm V
BARRETT
UNIT I
>^70-180 MW
M
70 MW
I.2%S
30 ppm V
1.2
FURNACE
Figure 10. Effect of furnace oxygen and fuel oil sulfur and vanadium
content on metal sulfate emissions at several oil-fired units,
266
-------
CONCLUSIONS
The field utilization of the Brookhaven controlled conden-
sation system at several oil-fired power plant units demonstra-
ted the capability of a reasonably simple but quite reliable
approach to the sampling of flue gas for the specific consti-
tuents, H2S04 and total particulates; the latter were subse-
quently separated into a water-soluble and insoluble fraction.
The soluble fraction was shown to be entirely composed of water
soluble metal sulfates—principally of Mg, V, and Na. Carbon
was the main element in the insoluble fraction which also con-
tained metal oxides primarily of Mg, V, and iron.
Furthermore, it was conclusively shown that unless the flue
gas sampling method used to study the effect of operating para-
meters on the emission of total sulfates was specific for H2S04
and metal sulfates, the exact nature of the mechanisms respon-
sible for the variability of sulfate emissions would have been
more difficult to ascertain.
Those parameters shown to have the main effects on sulfuric
acid emissions were furnace oxygen (H2S04 increased with in-
creased O2), fly ash recirculation (H2S04 decreased with re-
circulation) , and the fuel oil sulfur and vanadium content (an
increase in either constituent increased the H2S04). Metal
sulfates (MS04) were primarily governed by furnace oxygen (MS04
increased with increasing 02 but to an extent proportional to
the vanadium content of the fuel—i.e., with little or no vana-
dium in the fuel, MS04 was constant and independent of furnace
02), the ESP (MS04 decreased with increasing precipitator effi-
ciency) , and the vanadium and other metal content of the fuel
(MSO4 increased with increasing vanadium and generally also with
other metals).
From the results to date it appeared that there were two
regions responsible for the formation of H2S04 and metal sul-
fates. At one, the post-flame region, it is postulated that a
portion of the flame-induced super-equilibrium S03 (3b) caused the
formation of metal sulfates somewhat independent of furnace 02
levels. Those metal sulfates corresponded to the intercept values
in Figure 10. The other region, the superheater and reheater
tubes of the boiler, was responsible for the catalytic formation
of sulfuric acid with a concomitant amount of metal sulfates,
dependent on the vanadium and furnace oxygen levels.
Finally it was demonstrated that one direction that can be
taken to reduce H2S04 and metal sulfate emissions was to burn
267
-------
fuel oil containing low amounts of sulfur (<0.3% S) with little
vanadium (<30 ppm V)—that is, the experience at Barrett. How-
ever, an equally acceptable and probably less expensive approach
has been demonstrated at Northport Unit 3—a high sulfur (2.4%),
high vanadium (350 ppm) oil-fired unit. By maintaining furnace
oxygen at or below 0.1%, sulfuric acid was held to about 1 ppm,
and by maintaining the ESP at 90% efficiency or better, metal
sulfates were held to less' than 2 ppm. Thus, at a properly con-
trolled high sulfur, high vanadium oil-fired unit, the absolute
emissions of total sulfates (~3 ppm) were held to the same
level as at a low sulfur, low vanadium oil-fired unit (~2 ppm).
ACKNOWLEDGMENT
We would like to thank Bob Gergley and Bob Wilson for their
help in performing the field experiments as well as Lance Warren
and Fred Glaser at Northport and Ted Kempf and Ken Abrams at
Barrett for helping with arrangements and supplying field data,
Harold Cowherd and Fred Lipfert of LILCO for special arrange-
ments and B.T. Hagewood of LILCO for fuel analyses. A special
appreciation goes to the BNL Analytical Group for the sulfate,
carbon, and elemental analyses, and to Irv Meyer of the BNL
glass shop for fabrication of components. Several discussions
with Jim Homolya and John Nader of EPA and with Jack O'Neal
of LILCO were very helpful.
268
-------
REFERENCES
1. Dietz, R. N., and R. F. Wieser. Sulfate Emissions from Fossil
Fueled Combustion Sources, a. Progress Report No. 6,
March 1978; b. Progress Report No. 5, September 1977; c.
Progress Report No. 4, February 1977. Brookhaven National
Laboratory.
2. Goksoyr, H. , and K. Ross. The Determination of Sulfur
Trioxide in Flue Gases. J. Inst. Fuel, 35:177-179, 1962.
3. Dietz, R. N. , and R. W. Garber. Power Plant Flue Gas and Plume
Sampling Studies. a. Progress Report No. 2, in prepara-
tion; b. Progress Report No. 1, November 1977. Brookhaven
National Laboratory.
4. Dietz, R. N. , and R. F. Wieser. Sampling Power Plant Flue Gas:
Separate Collection of Suspended Particulates and Sulfuric
Acid, in preparation. Brookhaven National Laboratory.
5. Dietz, R. N., R. F. Wieser, and L. Newman. An Evaluation of
the Modified EPA Method 6 Flue Gas Sampling Procedure.
Presented at Workshop on Measurement Technology and Charac-
terization of Primary Sulfur Oxides Emission from Combustion
Sources, Southern Pines, North Carolina, April 1978.
6. Cheney, J. L. , W. T. Winberry, and J. B. Homolya. A Sampling
and Analytical Method for the Measurement of Primary Sul-
fate Emission. J. Environ. Sci. Health A12, 10:549-66,
1977.
7. Homolya, J. B., H. M. Barnes, and C. R. Fortune. A Characteri-
zation of the Gaseous Sulfur Emissions from Coal and Oil-
Fired Boilers. Fourth National Conference on Energy and
Environment, Cincinnati, October 1976.
8. Sparks, L. E. Effect of a Fly Ash Conditioning Agent on
Power Plant Emissions. EPA-600/7-76-027, October 1976.
9 Kircher J. F., et al. A Survey of Sulfate, Nitrate, and
Acid Aerosol Emissions and Their Control. EPA-600/7-77-041,
April 1977.
269
-------
10. Plumley, A. L., J. Jonakin, and R. E. Vuia. A Review Study of
Fireside Corrosion in Utility and Industrial Boilers. Pre-
sented at Corrosion Seminar, McMaster University and En-
gineering Institute of Canada, Hamilton, Ontario, May 1966.
11. O'Neal, A. J., Jr. The Nature and Cost of Residual Fuel Oil
Problems and the Profits Realized from Proper Corrective
Action. Combustion, 37-45, Dec. 1977.
12. Kera, Y., and K. Kuwata. The Formation of VOS04 on the Sur-
face of V205 in the Oxidation of S02 as Studied by ESR.
Bull. Chem. Soc. Japan, 50:2438-2441, 1977.
13. Ephraim, F. Decomposition Pressures of Hydrates. In:
International Critical Tables Vol. VII, National Academy of
Sciences, 1930. pp. 224-313.
14. Barnett, E. de P., and C. L. Wilson. Inorganic Chemistry,
Longmans, Green, and Co., 1957. p. 181.
15. Cavanaugh, L. A., et al. Particulate Sampling Program for
the Encina Power Plant. SRI Project 6747-1, SRI Interna-
tional, January 1978.
16. O'Neal, A. J., Jr. Flue Gas Studies. Long Island Lighting
Company Internal Report, January 1978.
17. Bennett, R. L., and K. T. Knapp. Particulate Sulfur and Trace
Metal Emissions from Oil-Fired Power Plants. AICHE Meeting,
New York City, November 1977.
18. U.S. Environmental Protection Agency. Standards of Per-
formance for Stationary Sources. Federal Register 41 (111),
June 1976. 23076-83.
19. O'Neal, A. J., Jr. Total Particulates and Sulfate Parti-
culates at Barrett #2. Long Island Lighting Company Inter-
nal Report, February 1978.
270
-------
Report of the Working Group on
Characterization of Paniculate Sulfur
Oxides Emissions
Ray F. Maddalone, Reporter
This Working Group had the objective of reviewing the status
of characterization data for particulate sulfur oxides emissions.
Its conclusions and recommendations are as follows.
STATUS AND VALIDITY OF AVAILABLE DATA
The group felt that the data presented during the workshop,
while certainly very extensive, did not represent all that was
known about particulate sulfur species. It further noted that
additional information, especially that pertaining to emissions
from coal combustion, was likely to be forthcoming from sources
outside the United States, due to the greater length of time
during which a number of countries (notably in Europe) have been
utilizing coal as a fuel.
The foregoing statement notwithstanding, it was felt that suf-
ficient data were available to provide considerable insight into
both qualitative and quantitative aspects of the formation, trans-
formation, and emission of particulate sulfur-containing species.
In assessing the validity of these data the following points were
noted:
1. Available sampling and measurement technology are quite
adequate to describe the general behavior of primary
sulfur oxide emissions and, in several cases, are suf-
ficiently reliable to enable rather precise and accurate
measurement of individual species.
• Measurement methods for sulfur dioxide are quite
adequate for present needs.
271
-------
• Determination of particulate sulfate salts is also
adequate, although it is noted that most of the avail-
able procedures only measure those sulfate salts
which are readily soluble in water under specified
conditions.
• Measurement methods for determining sulfur trioxide
and sulfuric acid definitely need to be improved and
at present probably only indicate lower limits of
these species.
• At present, methods for the identification of in-
dividual metal sulfates are in their infancy and
need to be improved.
2. Processes which result in the formation and transformation
of particulate sulfur-containing species and which are
known to occur both in combustion sources and in
emitted plumes include:
• Condensation of sulfur trioxide/sulfuric acid.
• Adsorption of sulfur trioxide/sulfuric acid onto
fly ash particles.
• Chemical reaction of sulfur trioxide/sulfuric acid
with chemical constituents of fly ash particles.
• Direct interaction of sulfur dioxide with fly ash
particles. While definitive information is sparse,
it is probable that this also takes place.
3. The parameters which control the formation and subsequent
emission of particulate sulfate salts are generally rather
poorly defined; however, it is clear that the following
factors undoubtedly influence these processes:
• The temperature profiles within the combustion zone,
stack system and plume.
• The amount of oxygen present.
• The characteristics of the fuel.
272
-------
RECOMMENDATIONS
The group recognizes that particulate sulfur species generally
account for only a minor fraction of the total primary sulfur oxides
derived from conventional combustion processes. On the other hand,
it is these particulate species which often play a controlling
role in determining the environmental impact of emitted sulfur
oxides. Consequently, the following recommendations are directed
towards elucidation of the characteristics of combustion particu-
lates and of the processes of formation of particulate sulfur oxides,
1. More detailed information is required concerning the
condensation of vapor phase sulfur trioxide/sulfuric
acid so that prediction of the temperature dependence
of this process can be made more precise.
• More precise definition of the controlling
parameters, especially the role of nucleating
particles, is required.
• The size distribution of condensed droplets or
particles must be established more precisely, and
the factors controlling this size distribution
must be identified.
2. Information is required concerning the rate, extent, and
mechanism of formation of sulfate salts from sulfur
trioxide and/or sulfuric acid.
• The influence of the chemical composition of fly ash
particles and the temperatures encountered must be
established.
* The particle size distribution of sulfate salts and
the factors controlling this size distribution must
be determined.
3. It is necessary to establish the relative importance of
homogeneous and heterogeneous oxidation of sulfur dioxide
to sulfur trioxide. In particular, there is a need to
determine the role of fly ash particles in catalyzing this
oxidation. In this regard it is noted that such catalysis
may be due to one or more of the trace metal species
associated with fly ash or to the presence of carbonaceous
soot particles.
273
-------
4. It is also necessary to establish the importance of direct
sulfur dioxide-to-sulfate salt conversion. Relatively
little is known about this potentially important pathway;
however, it seems likely that its mechanism involves
catalytic promotion, the nature and extent of which should
be clearly established.
5. Studies of the physical and chemical characteristics of
combustion particulates should be continued.
• Compositional versus size characteristics should
be firmly established.
• Increased knowledge of the physical and chemical
composition of particle surfaces should be sought.
• Recognition should be given to the fact that the
characteristics of emitted material may differ from
those of material retained in particle control
devices.
6. It is necessary to extend present studies of conventional
combustion systems to emerging technologies involving
fuel combustion or conversion and pollutant emission
control. It is apparent that both the nature and amounts
of sulfur emissions associated with several emerging
processes differ significantly from those encountered
in conventional combustion.
7. Finally, yet perhaps most important, it is recommended
that an international working group or task force be
established to gather information and establish a com-
prehensive data base dealing with primary sulfur oxide
emissions from combustion sources. It is suggested that
such a group be established through a United Nations
Agency (e.g., WHO), since the problems addressed are of
considerable international concern.
274
-------
Appendix
PARTICIPANTS AND OBSERVERS
Jeffrey W. Adams
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02140
617/864-5770 x.3036
Aubrey P. Altshuller
Director
Environmental Sciences
Research Laboratory
Environmental Protection Agency
Environmental Research Center
MD/59
Research Triangle Park
North Carolina 27711
919/541-2191
John Bachmann
Environmental Protection Agency
Environmental Research Center
MD/12
Research Triangle Park
North Carolina 27711
919/541-5231
Elizabeth M. Bailey
Division of Environmental
Planning
Tennessee Valley Authority
Muscle Shoals, Alabama 35660
205/383-4631 x.2788
Roy L. Bennett
Research Chemist
Environmental Sciences
Research Laboratory
Environmental Protection Agency
Environmental Research Center
MD/46
Research Triangle Park
North Carolina 27711
919/541-3173
Richard K. Chang
Department of Engineering and
Applied Science
Yale University
New Haven, Connecticut 06520
203/432-4470
James L. Cheney
Environmental Protection Agency
Environmental Research Center
MD/46
Research Triangle Park
North Carolina 27711
919/541-3172
Harold Cowherd
Environmental Engineering
Long Island Lighting Company
175 East Old Country Road
Hicksville, New York 11801
516/733-4700
Kenneth M. Gushing
Research Physicist
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
205/323-6592
Daryl DeAngelis
Research Engineer
Monsanto Research Corporation
1515 Nicholas Road
Dayton, Ohio 45407
513/268-3411
Russell N. Dietz
Chemical Engineer
Brookhaven National Laboratory
Building 426
Upton, New York 11973
516/345-3059
275
-------
James Dorsey
Industrial Environmental
Research Laboratory
Environmental Protection Agency
Environmental Research Center
MD/62
Research Triangle Park
North Carolina 27711
919/541-2557
Brian Doyle
Principal Engineer
KVB, Inc.
246 North Central Avenue
Hartsdale, New York 10530
914/949-6200
Edgar S. Etz
Research Chemist
Center for Analytical Chemistry
National Bureau of Standards
Chemistry Building
Room A-121
Washington, D.C. 20234
301/921-2862
Richard C. Flagan
California Institute of
Technology
MS 138-78
Pasadena, California 91125
213/795-6811 x.1383
William M. Henry
Projects Manager
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
614/424-5210
James B. Homolya
Environmental Protection Agency
Environmental Research Center
MD/46
Research Triangle Park
North Carolina 27711
919/541-3085
James E. Howes, Jr.
Senior Researcher
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
614/424-5269
Skillman C. Hunter
KVB, Inc.
17332 Irvine Boulevard
Tustin, California 92680
714/832-9020
Peter Jackson
Central Electric Generating
Board
Marchwood Engineering
Laboratories
Marchwood Southampton
England SO44ZB
Ashok K. Jain
Research Engineer
NCASI
Box 14483
Gainesville, Florida 32604
904/377-4708
Robert J. Jakobsen
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
614/424-5617
Kenneth T. Knapp
Chief, Particulate Emissions
Research Section
Environmental Sciences
Research Laboratory
Environmental Protection Agency
Environmental Research Center
MD/46
Research Triangle Park
North Carolina 27711
919/541-3085
276
-------
Arthur Levy
Manager
Combustion Systems Technology
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
614/424-4827
Dale Lundgren
Environmental Engineering
Sciences
University of Florida
Gainesville, Florida 32611
904/392-0846
Ray F. Maddalone
Section Head
TRW Defense and Space
Systems Group
One Space Park 01/2020
Redondo Beach, California 90278
213/535-1458
Richard E. Marland
Office of the Assistant
Administrator for Research
and Development
Environmental Protection Agency
RD 672
401 M Street, S.W.
Washington, D.C. 20460
202/755-2532
William R. McCurley
Research Engineer
Monsanto Research Corporation
1515 Nicholas Road
Dayton, Ohio 45418
513/268-3411
John Nader
Chief
Stationary Source Emissions
Research Branch
Environmental Protection Agency
Environmental Research Center
MD/46
Research Triangle Park
North Carolina 27711
919/541-3085
David F. S. Natusch
Professor
Department of Chemistry
Colorado State University
Fort Collins, Colorado 80523
303/491-5391
A. Jack O'Neal, Jr.
Chief Chemist
Electric Production Department
Long Island Lighting Company
P.O. Box 426
Glenwood Landing, New York 11547
516/671-6783
Richard Rhudy
Project Manager
Electric Power Research Institute
Box 10412
Palo Alto, California 94303
415/855-2421
Roosevelt Rollins
Environmental Protection Agency
Environmental Research Center
MD/46
Research Triangle Park
North Carolina 27711
919/541-3171
Arthur M. Squires
Department of Chemical
Engineering
Virginia Polytechnic Institute
Blacksburg, Virginia 24061
703/951-5972
Paul Urone
National Environmental
Investigation Center
Denver Federal Center
Building 53 - Box 25227
Denver, Colorado 80225
303/234-4661
277
-------
Jack Wagraan Arthur S. Werner
Environmental Protection Agency Manager
Environmental Research Center Analytical Laboratory
MD/46 GCA/Technology Division
Research Triangle Park Burlington Road
North Carolina 27711 Bedford, Massachusetts 01730
919/541-3009 617/275-9000
278
-------
1. REPORT NO.
EPA-600/9-78-Q2Qb
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before competing)
4. TITLE AND SUBTITLE
WORKSHOP PROCEEDINGS ON PRIMARY SULFATE EMISSIONS FROM
COMBUSTION SOURCES
Volume 2. Characterization
'6. PERFORMING ORGANIZATION COD:
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS"
Kappa Systems, Inc.
1501 Wilson Boulevard
Arlington, Virginia
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
15. SUPPLEMENTARY NOTES
3. RECIPIENT'S ACCESSION NO."
5. REPORT DATE
.Au_giist_iaZ8
10. PROGRAM ELEMENT NO.
1AD712 BC-52 (FY-78)
11. CONTRACT/GRANT NO.
68-02-2435
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/09
16. ABSTRACT
Technical papers on the characterization of primary sulfate emissions from combustion
sources, presented at a workshop sponsored by the U.S. Environmental Protection Agency,
are compiled in Volume 2 of a proceedings.
The objectives of the workshop were to review and discuss current measurement methods
and problem areas for sulfur oxides emission with attention focused on sulfuric acid,
sulfates, and sulfur-bearing particulate matter; to review and discuss emission data
from various combustion sources operating under different conditions which include
various pollutant controls, fuel composition, excess boiler oxygen, etc.; and to
delineate and recommend areas in need of research and development effort.
Scientists were invited to present the result of their studies on primary sulfate
emissions. The 3-day workshop devoted one day to measurement technology, a second to
characterization, and a third to critical assessment of the presented papers and
development of summary working group reports on each half-day session of the initial
2 days. Thirty-one papers were presented by 29 participants on measurements and
characterization. Four working group reports were developed and summarized in the
last day.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
* Air pollution
* Sulfates
* Emission
* Combustion products
* Chemical analysis
* Physical properties
I).IDENTIFIERS/OPEN ENDED TERMS
.'. COSATI Held/Croup
13B
07B
21B
07D
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report}
UNCLASSIFIED
>1. NO. OF P,
287
20. SECURITY CLASS (This page)
22. PRICE
:LASSIFIED
EPA Form 2220—1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
279
------- |