&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

-------