United States Industrial Environmental Research EPA-600/7-78-164c
Environmental Protection Laboratory August 1978
Agency Research Triangle Park NC 27711
v>EPA Environmental
Assessment of
Coal- and Oil-firing
in a Controlled
Industrial Boiler;
Volume III.
Comprehensive
Assessment and
Appendices
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect the
views and policies of the Government, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-164c
August 1978
Environmental Assessment of Coal-
and Oil-firing in a Controlled
Industrial Boiler;
Volume III. Comprehensive
Assessment and Appendices
by
C. Leavitt, K. Arledge, C. Shih,
R. Orsini, W. Hamersma, R. Maddalone,
R. Beimer, G. Richard, and M. Vamada
TRW, Inc.
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-2613
Task No. 8
Program Element No. EHE624A
EPA Project Officer: Wade H. Ponder
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory, U. S. Environmental Protection Agency, and approved for publi-
cation. Approval does not signify that the contents necessarily reflect
the views and policies of the U. S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorsement
or recommendation for use.
-------
ABSTRACT
The report gives results of a comparative multimedia assessment of
coal versus oil firing in a controlled industrial boiler. Relative environ-
mental, energy, economic, and societal impacts were identified. Compre-
hensive sampling and analyses of gaseous, liquid, and solid emissions from
the boiler and its control equipment were conducted to identify criteria
pollutants and other species. Major conclusions include: (1) while the
quantity of particulates from oil firing is considerably less than from
coal firing, the particles are generally smaller and more difficult to
remove, and the concentration of particulates in the treated flue gas from
oil firing exceeded that from coal firing. (2) N0X and CO emissions
during coal firing were about triple those during oil firing. (3) Sulfate
emissions from the boiler during coal firing were about triple those
during oil firing; however, at the outlet of the control equipment, sulfate
concentrations were essentially identical. (4) Most trace element emis-
sions (except vanadium, cadmium, lead, cobalt, nickel, and copper) were
higher during coal firing. (5) Oil firing produces cadmium burdens in veg-
etation approaching levels which are injurious to man; coal firing may pro-
duce molybdenum levels which are injurious to cattle. (6) The assessment
generally supports the national energy plan for increased use of coal by
projecting that the environmental insult from controlled coal firing is not
significantly different from that from oil firing.
This report was submitted in fulfillment of Contract Number 68-02-2613,
Task 8 by TRW Environmental Engineering Division under the sponsorship of
the U. S. Environmental Protection Agency. This report covers a period
from October 24, 1977 to May 5, 1978, and work was completed as of May 5,
1978.
iii
-------
CONTENTS
Abstract iii
Abbreviations . v
Acknowledgments vi
1. Introduction 1-1
2. Summary and Conclusions 2-1
3. Plant Description 3-1
4. Test Description 4-1
5. Comprehensive Assessment of Coal Firing Case for an Industrial
Boiler 5-1
6. Comprehensive Assessment of Oil Firing Case for an Industrial
Boiler 6-1
Appendices
A. Simplified Air Quality Model A-l
B. Organic Analysis Methods B-l
C. Inorganic Analysis Methods C-l
iv
-------
LIST OF ABBREVIATIONS
acm/mi n
— Actual Cubic Meters Per Minute
ACFM
-- Actual Cubic Feet Per Minute
DSCM
-- Dry Standard Cubic Meters
ESCA
-- Electron Spectroscopy for Chemical Analyses
FGD
— Flue Gas Desulfurization
ICPOES
-- Inductively Coupled Plasma Optical Emission Spectroscopy
MATE
-- Minimum Acute Toxicity Effluent
NAAQS
— National Ambient Air Quality Standards
NSPS
-- New Source Performance Standards
SSMS
~ Spark Source Mass Spectrometry
TSP
~ Total Suspended Particulate
v
-------
ACKNOWLEDGMENTS
The cooperation of the Firestone Tire and Rubber Company and FMC is
gratefully acknowledged. We are particularly indebted to Gary Wansley of
Firestone and Carl Legatski of FMC, without whose cooperation this assess-
ment could not have been completed.
vi
-------
SECTION I
INTRODUCTION
A comparative multimedia assessment of coal-f1r1ng and o1l-f1r1ng 1n
an Industrial boiler was conducted. Extensive sampling and analysis of
all major gaseous, liquid, and solid emissions and effluents was done.
The test boiler was a dual fuel 10 MW equivalent unit that 1s capable of
burning both coal and oil. During the tests conducted for this study the
boiler burned either exclusively coal or oil, although 1t can burn both
fuels simultaneously 1f required. The boiler 1s equipped with a pilot
double-alkali flue gas desulfurlzatlon (F6D) unit designed to treat
approximately 30% of the total flue gas, approximately 3 MW equivalent.
During the tests, however, the FGD was processing only about 10-15% (13%
average) of the total when coal was fired and 23-30% (25% average) when
oil was burned. This corresponds to approximately 1.5 MW and 2.5 MW,
respectively. Because the FGD was operating at less than design capacity,
there 1s some question about the typlcalness of the test results. That 1s,
the pilot unit may have been performing better than the full size commercial
version will.
The assessment consists of several parts. First, comprehensive
emissions assessments of each fuel were conducted. These assessments
consist of detailed examinations of gaseous, liquid and solid emissions
and effluents considering both pre- and post-scrubber emissions for each
fuel. That 1s, the emissions to all three media were determined for the
case where no FGD was present and for the case where an FGD capable of
scrubbing 100% of the flue gas was present. The comprehensive emissions
assessment also consisted of an examination of the effects and
efficiencies of the scrubber for both fuels.
These comprehensive emissions assessments were used to develop
a comparative emissions assessment. This assessment examines the
differences 1n the quantities and character of the emissions resulting
from the combustion of each fuel. In this case the primary emphasis was
on the emissions and effluents resulting from the burning of each fuel
where 100% scrubbing capacity was available. This part of the project
was concerned with determining the emissions resulting from each fuel
1-1
-------
with emission controls 1n place. Of special concern was the cross-media
impacts of each fuel.
On the basis of the emissions assessments a comparative environmental,
societal and energy impact assessment was developed. The emphasis here was
on the relative impacts of each fuel assuming full flue gas desulfurtzation.
The uncertainties inherent in these types of analyses dictated that the
differences rather than the absolute magnitude of these Impacts be
considered.
The report consists of three volumes. Volume I is an executive
summary which summarizes the major results and conclusions of the study.
Volume II presents the comparative assessments. Included are a descrip-
tion of the test setting, the comparative emissions assessment, and the
comparative environmental, societal and energy Impact assessment. This
volume, Volume III, contains the comprehensive assessments and appendices
and includes detailed descriptions of the test site, the test protocol
and a detailed presentation of data.
1-2
-------
SECTION 2
SUMMARY AND CONCLUSIONS
A comparative assessment of coal and oil firing 1n a controlled
Industrial boiler was conducted. The comprehensive emissions assessments
for each fuel were used to develop a comparative emissions assessment.
On the basis of the emissions assessment a comparative environmental
assessment was developed.
The following is a 11st of the major conclusions resulting from the
comprehensive coal and comprehensive oil assessments;
COMPREHENSIVE COAL ASSESSMENT
• Uncontrolled emissions of criteria pollutants generally
corresponded well with values reported in AP-42. Although
NOx emissions were slightly higher than the average AP-42
value, they appear to be within the normal range for similar
Industrial units.
• NOx reductions varying from approximately 0 to 24 percent
were measured across the scrubber. However, the magnitude
of NOx reductions could not be correlated to changes in
variables monitored during the test period (I.e., tempera-
ture, gas flow rate, liqu1d/gas ratio, boiler load, etc.).
For this reason, 1t 1s considered feasible that observed
NOx reductions are a sampling phenomenon, perhaps related
to leaks 1n the sample train.
• Sulfur dioxide removal data Indicated an average scrubber
efficiency of 97 percent. Controlled S02 emissions were
36.3 ng/J (0.08 pounds/MM Btu) which 1s less than either
existing or proposed NSPS limitations for utility boilers.
• Mass balance data Indicate that the multlclone unit up-
stream of the scrubber was removing little or no fly ash
during the test period. The scrubber was found to remove
99.4 percent of the Inlet particulate removal.
• Although the removal efficiency for total particulates 1s
high, there appears to be a net Increase 1n emission rates
across the scrubber for particulates less than 3jim 1n size.
This net increase can be attributed to the poor removal
efficiency of the scrubber for fine particulates, and to the
sodium blsulfate (NaHS04) and calcium sulfite hemihydrate
(CaS03•!/2H2O) particulates generated by the scrubber. Both
NaHS04 and CaS03»l/2H20 have been Identified at the scrubber
outlet but not at the Inlet.
2-1
-------
• The relatively poor removal efficiency (approximately 30%)
for SO3 across the scrubber is an indication that SO3 is
either present as very fine aerosols in the scrubber inlet,
or is converted to very fine aerosols as the flue gas
stream is rapidly cooled inside the scrubber.
• The overall removal efficiency for trace elements across
the scrubber is 99.5 percent. Of the 22 major trace ele-
ments, 18 exceed their MATE values at the scrubber inlet
and four at the scrubber outlet. The four trace elements
in the scrubber flue gas that pose a potential hazard are
arsenic, chromium, iron, and nickel. In addition, the
emission concentration of beryllium at the scrubber outlet
ts equal to its MATE value.
• The relative removal efficiency for trace elements across
the scrubber can be explained by enrichment theory. In
general, trace elements that occur as element vapors or
form volatile compounds at furnace temperatures are more
concentrated in the smaller particulates, as a result of
subsequent condensation and surface adsorption. These are
the same trace elements that are removed less efficiently
by the scrubber.
• ESCA analysis has shown that while there may be higher
surface concentration of sulfur containing compounds in
the particulates emitted from the scrubber, most of the
sulfur containing compounds are probably present as solid
sulfates and sulfites. Thus, it 1s conceivable that
sulfuric acid vapor 1s condensed and deposited on the
particulates emitted, whereas sodium blsulfate and calcium
sulfite hemlhydrate are emitted as fine, solid particulates.
• The overall sulfur balance indicates that over 92 percent
of the fuel sulfur is emitted as S02> less than 1 percent
of the fuel sulfur is emitted as SO3, and approximately
3 percent of the fuel sulfur Is emitted as S04~.
• Total organic emissions were generally less than 6 ng/J
(0.01 pound/MM Btu) and these emissions appear to be
primarily Ci to C6 hydrocarbons and hydrocarbons heavier
than C-jg. While uncontrolled emission rates for C7 to C-jc
and higher hydrocarbons are low, emissions of these organics
were further reduced by 90 to 100 percent in the scrubber
unit.
• No polycycllc organic material (POM) were detected 1n
either scrubber inlet or outlet samples drawn during the
test period. Total organic emissions were sufficiently
low that attempts to identify specific compounds resulted
only 1n Identification of substances normally associated
with laboratory analyses and sampling equipment.
2-2
-------
• The combined wastewater stream generated from the boiler
operation apparently does not pose an environmental hazard
since the discharge concentrations of Inorganics and organlcs
are all well below their MATE values.
• The scrubber cake produced during coal firing contained
29% coal fly ash. With the exception of boron, trace
element concentrations in the scrubber cake far exceeded
their MATE values. Because the trace elements may leach
from the disposed scrubber cake, these solid wastes must
be disposed of in specially designed landfills.
• Mass balance closure for most of the trace elements have
been found to be 1n the 75 to 107 percent range. This
closure Instills confidence on the validity of the sampling
and analysis data for trace elements.
COMPREHENSIVE OIL ASSESSMENT
• Uncontrolled emissions of criteria pollutants do not
generally correspond with emission factors from AP-42. N0X
emissions were nearly 23% lower than the AP-42 emission factor,
although they appear to be within the normal range for similar
industrial units. CO emissions were nearly 63% lower than
the AP-42 emission factor. SO? and total hydrocarbons corres-
ponded well with their respective AP-42 emission factors.
Particulate emissions, 1n the absence of coal ash contamina-
tion, are approximately twice the value tabulated 1n AP-42.
t Sulfur dioxide removal data indicated an average scrubber
efficiency of 97%. Controlled S02 emissions were 26.8 ng/J
90.06 lb/MM Btu) which 1s less than either existing or pro-
posed NSPS limitations for utility boilers.
0 Particulate removal data Indicate that, on the average, scrubber
efficiency was 84% during the test period. However, based on
particulate catches essentially free of coal ash contamination,
the scrubber efficiency was approximately 75% for o1l-f1r1ng
particulates.
• Organic emissions were generally less than 5 ng/J (0.01 lb/MM
Btu) and appear to be composed primarily of C"| to C« hydro-
carbons and organlcs heavier than C-jg. Approximately 88 and
83% of the C7 to C]g and higher than C15 organlcs, respectively,
were removed by the scrubber.
• The organic compounds Identified 1n the gas samples were
generally not representative of combustion-generated organic
materials, but were compounds associated with materials used
1n the sampling equipment and 1n various analytical procedures.
This again confirms the low level of organic emissions.
2-3
-------
• Polycyclic organic material (POM) was riot found in the scrubber
inlet or outlet at detection limits of 0.3 yg/m3. MATE values
for most POM's are greater than this detection limit. However,
since the MATE values for at least two POM compounds - benzo(a)-
pyrene and d1benz(a,h)anthracene - are less than 0.3 yg/m3,
additional GC/MS analyzes at higher sensitivity would be
required to conclusively preclude the presence of all POM's
at MATE levels.
• When emissions are uncontrolled, 90 to 923 of the sulfur 1n
the fuel feed is emitted as S02» less than 1% as SO3, and
1.5% as SO^58.
• SOg is efficiently removed by the scrubber (97 to 98? effi-
ciency). The SO3 removal efficiency (28 to 29%) suggests that
SO3 1s associated with fine particulates or aerosols. S04a 1s
about 60% removed by the scrubber, and so is probably associated
with the larger particulates.
t Of the 22 major trace elements analyzed 1n the flue gas stream,
11 exceeded their MATE values at the scrubber inlet while only
5 exceeded MATE values at the scrubber outlet. These 5 elements
are arsenic, cadmium, chromium, nickel and vanadium. With the
exception of chromium, elements exceeding their MATE values at
the scrubber outlet were removed from the flue gas stream with
efficiencies lower than the overall average removal efficiency
of 87%.
• Beryllium emissions were 0.001 mg/m^ after scrubbing this
corresponds to half the MATE value for this element. At this
emission concentration, the National Standard for Hazardous Air
Pollutants limitation of 10 grams beryllium per day would only'
be exceeded by boilers of 100 MW capacity or greater.
• The combined wastewater stream from the boiler operation
apparently does not pose an environmental hazard, since the
discharge concentrations of organics are all well below their
MATE values.
• Mass balance closure for 10 of the 20 trace elements analyzed
is between 50 and 136%. Poorer mass balance closure
was obtained for the remainder of trace elements due to the
extremely low element concentrations and/or contamination of
the scrubber recycle solution by coal firing components.
/
t The scrubber cake produced contains about 1% oil flyash. With
the exception of antimony, boron, molybdenum and zinc, trace
element concentrations In the scrubber cake exceeded their
health based MATE values. All ecology based MATE values were
exceeded by trace element concentration. Because the trace
elements may leach from the disposed scrubber cake, these
solid wastes must be disposed of in specially designed land-
fills.
2-4
-------
SECTION 3
PLANT DESCRIPTION
The host for this assessment was the Pottstown, Pennsylvania plant
of the Firestone T1re and Rubber Company. Boiler number 4, one of four
used to supply process and heating steam to the plant, was used 1n the
assessment. The boiler burns either coal or oil and has a pilot FMC
double alkali flue gas desulfurizatlon system designed to treat approxi-
mately one-third of the boiler flue gas.
The excellent cooperation and assistance by the Firestone T1re and
Rubber Company and FMC was Invaluable 1n performing this assessment.
GENERAL SETTING
Pottstown 1s situated 64 kilometers (40 miles) northwest of Philadel-
phia along the Schuylkill River. (See Figure 3-1.) The Firestone T1re and
Rubber Company plant layout 1s shown 1n Figure 3-2.
Steam Plant
The steam plant consists of four separate boilers which supply process
and heating steam to the entire facility. Boiler Numbers 3, 4, and 5,
operate at a fairly constant rate of 45,000 kg/hr (100,000 lb/hr) of steam.
Process steam demand 1s relatively steady, since the plant operates 24
hours per day, seven days per week. Fluctuations 1n heating load are
satisfied by either boosting steam generation rates on these boilers or by
operating Boiler No. 1. The steam generation rate of Boiler No. 1 varies
from zero to approximately 22,700 kg/hr (50,000 lb/hr) of steam.
Boiler Number 4 was chosen for this assessment because 1t has a pilot
scale flue gas desulfurizatlon system and a dual fuel capability.
BOILER DESCRIPTION
Boiler No. 4 1s a Type P-22 EL, Integral furnace, once through Bab-
cock and Wilcox unit. (See Table 3-1 for boiler specification data and
Figure 3-3 for a schematic of the boiler and associated equipment.) When
1t was Installed 1n 1958, the boiler was designed as a coal-fired unit but
was converted to fire either coal or oil 1n 1967. The changeover from one
fuel to the other can be accomplished 1n less than thirty minutes.
3-1
-------
Fagleysville
Douglatville
stow. POTTSTOWN
Sanatoga
Limerick
CHESTER
CO
Parker
Ford/
Stamfoi
Jersey City
Easton
Bethlehem
Reading
POTTSTOWN
Harrisburg
Trenton
Philadelphia
Lancaster
Camden
Wilmington
Delaware
Bay
\/
Layfield
Boyertown*^^ /
S xjGilbertsville
Morytville
Little
BERKS CO
' New ' ,3J
Hanover )Obell«k
MONTGOMERY
CO I
Scrantonj
Wilkes-4
Barre
Atlantic
Ocean
'Atlantic
City
Figure 3-1. Location of Pottstown, Pennsylvania
3-2
-------
LANDFILL
MANUFACTURING
AND WAREHOUSE
ENGINEERING
STEAM
PLANT
VISITOR
PARKING
LOT
CHENICAL DIVISION
MANUFACTURING
Figure 3-2. Industrial site plant layout.
TABLE 3-1. BOILER NUMBER 4 DESIGN DATA
Boiler Type:
Manufacturer:
Type of burner:
Number of burners:
Burner arrangement:
A1r Preheater:
Fuel:
Design steam rate:
Ofl/PuHverized coal;
face fired;
Integral furnace;
dry bottom.
Babcock and Wilcox, Type P-22 EL
Circular conical
3
Triangular, one face
yes
Number 6 fuel o11;
High volatile bituminous coal,
Class II, Group 2, of ASTM D388
45,000 kg/hr (100,000 lb/hr);
1.4 mPa (190 ps1);
at approximately 193°C (380°F).
JiiSJL
Process steam.
3-3
-------
SCRUBBER
FEED SOLIDS
EXHAUST
GAS TO STACK
EXHAUST GAS
TO STACK
FGD
LANDFILL
CAKE
SCRUBBER ^
MAKEUP WATER
FLYASH
STORAGE
FLYASH
MULTICLONES
EXHAUST
GAS
FEEDWATER FROM
PRETREATMENT UNIT
r-EAi
iRUI
STEAM
BOILER
OIL
TO MUNICIPAL
SEWAGE
TREATMENT
PIT
Figure 3-3. Boiler system schematic.
3-4
-------
The two fuels are usually not burned simultaneously except when con-
verting from oil to coal firing. The coal 1s Ignited by continuing oil
firing until a stable coal flame is obtained. 011 1s fired simultaneously
with coal to maintain acceptable steam generation rates when coal with a
low heat content is burned.
The boiler was designed to burn either Number 6 fuel oil or a high
volatile eastern bituminous coal. The boiler does not presently have fuel
oil specifications. Therefore, sulfur and heat content vary with supplier.
Analysis data for the fuel oil burned during testing can be found in
Section 6. The coal burned is required to meet Class II, Group 2 of
ASTM D388. Normally Pennsylvania coal is used. However, coal purchased
from a mine 1n Kentucky was burned during the 1978 coal strike. Analysis
data for the coal burned during testing can be found in Section 5. The
fuel handling systems are shown in Figures 3-4 and 3-5 for coal and oil,
respectlvely.
An air preheater 1s located in the flue gas plenum directly downstream
of the boiler. This gas to gas heat exchange recovers approximately 4.2
gigajoules/hr (4 million BTU/hr) when the boiler is operating at full load.
This represents a potential annual savings of 907,000 kg (1000 tons) of coal,
or 5500 barrels of oil.
Tables 3-2 and 3-3 present data on the monthly fuel consumption and
steam generation during 1976 and 1977, respectively.
The boiler is scheduled to operate 45 weeks per year on a 24-hour-per
day, 7-day-per-week basis.
The following maintenance is scheduled:
• Continuous surface blow-down of the steam drum to control alkalinity
arid sol Ids 1s maintained;
t Once per shift each end of the mud drum 1s blown down;
• Soot 1s blown once per shift;
t Bottom ash Is removed from the unit once per day.
Table 3-4 presents data on the number of days that boiler No. 4
was off-line during 1976 and 1977. Coal use, oil use, and the off-line
schedule are compared for 1976 and 1977 1n Figures 3-6 and 3-7. An over-
haul in July 1977 accounts for part of this off-line schedule.
3-5
-------
BUCKET ELEVATOR
CAR
ABOVE \
GROUND
COAL STORAGE
COAL RECEIVING PIT
PRIMARY AIR FAN
Figure 3-4. Coal handling flow diagram.
TRUCK
CAR
BOILER M
B&W
400 kg/hr
UNDER GROUND
STORAGE
4 VESSELS
190,000 LITERS EACH
2 VESSELS
1> 060,000 LITERS EACH
ABOVE GROUND
STORAC-E
Figure 3-5. Fuel oil distribution to boiler number 4.
3-6
-------
TABLE 3-2. 1976 STEAM GENERATION AND FUEL CONSUMPTION DATA FOR
BOILER NUMBER 4
Month
Total
steam
generated
by unit
Steam
generated
by coal
firing
Steam
generated
by oil
firing
Coat
Used to
fire
unit
011
used to
fire
unit
Pounds
Pounds
Pounds
Pounds/Month
Pounds/Month
JANUARY
63,191,120
MOKE
63,191,120
NONE
479,630
FEBRUARY
56,731,849
NONE
56.731.349
NONE
426,064
MARCH
59.284.34S
NONE
59.204,3<16
NONE
437.432
APRIL
65,082,129
NONE
65,032,129
NONE
481,852
MAY
54,663,247
NOME
54.663.247
NONE
408,942
JUNE
24,615,374
NOME
24,615,374
NONE
381,786
JULY
64.379.436
NONE
64.379,436
NONE
478,111
AUGUST
32,185,973
NONE
32,185,973
NCJ.'E
407,381
SEPTEMBER
17,249,137
NONE
17,249,137
NONE
125,047
OCTOBER
67,981,950
45,367,432
22,614,518
3.809.400 "
165,614
NOVEMBER
61,173,241
58,149,635
3,023,606
5,183,440
23,822
DECEMBER
70,591,100
67,407,612
3,183,488
5,733,500
23,408
TABLE
3-3. 1977 STEAM GENERATION AND FUEL CONSUMPTION DATA FOR
BOILER NUMBER 4
Month
Total
steam
generated
by unit
Steam
generated
by coal
firing
Steam
generated
by oil
firing
Coal
used to
fire
unit
011
used to
fire
unit
Pounds
Pounds
Pounds
Pounds/Month
Pounds/Mo nth
JANUARY
64,436,650
54,930,366
9,506,484
4,831,100
70,554
FEBRUARY
63,707,600
55,303,646
8,403,854
4,092,600
17,053
MARCH
61.963,231
54,338,243
7,624,988
4,886,200
12,165
APRIL
61.350,362
58,033,817
3,316,545
4,875,820
24,567
HAY
69,071,439
53,378,917
15,692,522
4,710,700
118,251
JUNE
67,044,394
60,757,611
6,286,703
5,038,420
46,492
JULY
12,301,200
10,634,135
1,667,065
610,040
11,703
AUGUST
48,932,444
41,596,830
7,335,614
3,458,880
59,448
SEPTEMBER
65,526,263
60,504,449
5,021,614
5.408,900
37,750
OCTOBER
72,127,226
50,656,817
21,470,409
4,176,380
160,074
NOVEMBER
54,799,650
43,500,403
11,299,247
3,690,060
84,153
~DECEMBER
-
-
-
-
-
*At the tine that this table was developed no Information was available for December 1977.
3-7
-------
TABLE 3-4. NUMBER OF DAYS THAT BOILER MAS OFF-LINE FROM 1 JANUARY 1976
TO NOVEMBER 1977
Month
No. of days off-line
1976
iyy/
JANUARY
—
—
FEBRUARY
--
—
MARC 15
*
3
APRIL
2
MAY
9
--
JUNE
19
—
JULY
1
23
AUGUST
18
8
SEPTEMBER
20
1
OCTOBER
—
—
NOVEN3ER
1
6
DECEMBER
—
*
Total 68 45
Exhaust gas cleaning
The flue gases are treated by an air pollution system which consists of
multlclone units and a pilot FGD unit. The multlclones are the primary
particulate control device. All of the flue gas passes through the multi-
clones after which the steam 1s split and two-thirds of the flue gas 1s
ducted to the stack. The other one-third 1s ducted to the pilot FGD
system which removes S02 and additional particulates. There are no NO
controls on the system.
Multiclone Unit
The collection efficiency of the multlclone varies as a function of
the particle size distribution and grain loading. Typically, multlclones
remove 90 percent of those particles with diameters of 10/1 and greater, and
50 to 80 percent of those particles with diameters of 3/uand greater. The
3-8
-------
• PERCENT OF CAPACITY SUPPLIED BY OIL FIRING
+ PERCENT OF CAPACITY SUPPLIED BY COAL FIRING
^ TOTAL PERCENT CAPACITY IN RELATION TO DESIGN
STEAM GENERATION."
100
90
80
70'
60 ¦
50'
40 <
30 ¦
20
~ ~
<2>
° o +
+
~
—» +
T 30
. . 20
•10
F M A M J J A S
MONTHS IN YEAR (1976)
x ¦<
r~
•< -•
'ASSUMING 100,000 LBS/hR AS 100* STEAM GENERATION
Figure 3-6. Comparison of fuel use and off-line schedule for 1976.
• PERCENT OF CAPACITY SUPPLIED BY OIL FIRING
4 PERCENT OF CAPACITY SUPPLIED BY COAL FIRING
^ TOTAL PERCENT CAPACITY IN RELATION TO
DESIGN STEAM GENERATION*
100
90'
80- .
70
60
50' •
40- •
30
.. XL
0 <3> O O
+ *
ILiX
o
o +
1- • n
«>10
t/> r
o at
T1 O
-n •
r o
T1
z
m a
3 3
X •<
r- o
-< H
m
3
JFMAMJJASON
MONTHS IN YEAR (1977)
'ASSUMING 100,000 LBS/HR AS 100X STEAM GENERATION
Figure 3-7. Comparison of fuel use and off-line schedule for 1977.
3-9
-------
collection efficiency of multiclones drops off rapidly for particles less
than 3m diameter.
The fly-ash is periodically collected and transported to an on-site
landfill for final disposal.
Flue Gas Desulfurization System
The flue gas desulfurization (FGD) system was designed and manufactured
by FMC Corporation. The FGD system is a pilot unit designed to handle 280
acm/min (10,000 ACFM) of flue gas, which is approximately one-third of the
volume of the flue gas from the boiler. The pilot plant was placed on-Hne
in January of 1975.
Figure 3-8 is the basic flow diagram of the FMC FGD system as it is
applied at this site.
The flue gas (stream 1) 1s withdrawn downstream of the boiler or the
exit side of the multiclone dust collectors. Fly-ash loading at the scrub-
ber inlet is substantially higher during coal-firing than during o1l-f1r1ng.
To accommodate the wide variation in fly-ash loading the FGD system was
designed to operate with or without fly-ash, and can be operated without
any mechanical changes on either fuel.
Upon entering the FGD unit the flue gases are contacted with a slightly
acidic scrubbing solution (stream 4) which removes SOg and particulates.
The SOg and particulates are removed at the scrubber throat and carried
away in the scrubbing solution. The process utilizes a sodium sulfite-
sodium bisulfite solution as the absorbent. The basic reaction for SOg
removal is:
Na2S03 + S02 + H20 2NaHS03
A bleed stream (stream 5) of the scrubbing solution 1s removed from
the system at a rate which keeps the pH of the solution 1n an acceptable
range. The bleed stream 1s reacted with calcium hydroxide in a short reten-
tion time, agitated vessel to regenerate the sodium sulfite. The basic
chemistry of sodium sulfite regeneration process 1s:
3-10
-------
CYCLONIC MIST
MESH MIST ELIMINATOR
w
i
VENTURI
SCRUB3ER
SLURRY
THICKENER
ELIMINATOR
STORAGE
ROTARY DRUf
FILTER
LEGEND:
1. BOILER FLUE GAS TO SCRUBBER
2. SCRUBBER OUTLET TO ATMOSPHERE
3. SOLID WASTE TO LANDFILL
4. ABSORBENT SOLUTIONTO SCRUBBER
5. ABSORBENT SOLUTION TO REGENERATION
6. SODIUM CARBONATE MAKEUP
7. REGENERATION SOLUTION
8. REGENERATED SCRUBBER SOLUTION
9. CONCENTRATED SLURRY
10. RETURNED SCRUBBER SOLUTION
Figure 3-8. FMC unit at the industrial facility.
-------
2NaHS03 + Ca(CH)2
CaS03 • 2H20 + 1-1/2HZ0 + Na2S03
The slurry of precipitated sulfur compounds (stream 8) 1s concentrated
and pumped to a rotary drum filter where the essentially clear liquid and
solid waste products are separated. The clear liquid (stream 10) 1s returned
to the system for further utilization. The solid wastes, 1n the form of
filter cake containing 40 percent (by weight) water, (stream 3) are removed
from the rotary drum filter and conveyed to a storage bin to await trans-
portation to the dump site. Because of the heavy particulate loading, more
filter cake is produced during coal firing than during oil firing.
The on-site landfill, which 1s the final disposal facility for all of
the solid waste generated at the facility, has several test wells from which
samples are collected every three months and sent to an Independent labor-
atory for analysis. Monthly tests are conducted by plant personnel to
monitor sodium and specific conductivity. With permission of the Pennsyl-
vania Department of Environmental Resources, this site 1s being used as an
experimental disposal area for the filter cake from the FMC unit.
Boiler Water Pretreatment
Water is brought 1n from the nearby Schuylkill River and treated by
a pretreatment facility to make 1t acceptable as boiler feedwater. The
water treatment process 1s made up of three stages; a precipitator stage,
a hot softening stage, and a filtration stage.
On exiting the filtration stage, the treated water 1s piped to a storage
vessel where it is mixed with condensate returned from the boiler.
Liquid wastes are generated by the water pretreatment plant at a
relatively constant rate Independent of the fuel being used 1n the boiler.
The major constituents of the effluent are compounds of calcium and
magnesium. The effluent 1s sent to a collection pit, mixed with acid water
from the resin plant and blow-down from the boilers, and then disposed of
directly into the municipal sewage system.
3-12
-------
Boiler Blowdown
There are two blowdown sources 1n the boiler system, the steam drum
and the mud drum. There 1s a continuous blowdown from the steam drum
which keeps the level of suspended sol Ids 1n the boiler feedwater within
an acceptable range. Tests of the steam drum blowdown effluent are made
every four hours and adjustments to blowdown rate are made accordingly.
The mud drum 1s blown once per shift.
The effluent from both blowdowns 1s sent to the same pit that collects
effluent from the water pretreatment unit.
Cooling Water
Approximately 150-190 l/m1n (40 to 50 gpm) of water from the boiler
feedwater storage tank is used in a once-through cooling system. This
water is composed of approximately 50 percent condensate return and
50 percent make-up water. After use, 1t 1s also sent directly to the
municipal sewage system»
3-13
-------
3-14
-------
SECTION 4
TEST DESCRIPTION
Multi-media emission tests were conducted on Boiler No. 4 of the
Firestone plant from 27 September through 8 October, 1977. Gaseous,
liquid and solid emissions were sampled during coal and oil firing to
obtain data for the assessment. Flue gas sampling was conducted before
and after the scrubber to determine which pollutants are removed or
modified by the control device.
Emissions were characterized using EPA's phased approach. This
approach utilizes two levels of sampling and analysis (Level 1 and Level
2). Level 1 procedures are accurate within a factor of about 3. They
provide preliminary assessment data and identify problem areas and Infor-
mation gaps. Based on these data a site specific Level 2 sampling and
analysis plan 1s developed. Level 2 provides more accurate and detailed
Information to confirm and expand on the Information gathered 1n Level 1.
The methods and procedures used for Level 1 are documented 1n the manual,
"Combustion Source Assessment Methods and Procedures Manual for Sampling and
Analysis", September 1977. The Level 2 methods and procedures include
"state-of-the-art" techniques adapted to the needs of this site. They are
described 1n Volume III Appendices B and C.
Normally all Level 1 samples are analyzed and evaluated before moving
to Level 2. Because of the program time constraints, the Level 1 and Level
2 samples were obtained during the same test period. However, analysis of
the samples did proceed in a phased manner except where sample degradation
was of concern. In that case Level 2 analysis was performed on the samples
prior to Level 1 completion.
TESTS AND FIELD ANALYSES
The industrial boiler assessment tests were conducted on the solid,
liquid and gaseous effluent streams and the fuel. Tests were conducted
during both coal and oil firing. Figure 4-1, the system schematic for
boiler No. 4, and desulfurlzatlon unit, shows the sampling locations.
Parameters sampled during coal and oil firing at each location are
summarized 1n Table 4-1. The table also identifies the sampling and
analysis methods used.
4-1
-------
EXHAUST
GAS TO STACK
EXHAUST GAS
TO STACK
SCRUBBER
SOLIDS
EXHAUST
LANDFILL
CAKE
SCRUBBER
MAKEUP
WATER
FLY ASH
PLY ash
MULTICLONES
STORAGE
EXHAUST
GAS
STEAM
COAL
BOILER
FEEDWATER
FROM
PRETREATMENT
UNIT
TO MUNICIPAL
. SEWAGE
TREATMENT
LEGEND
1 - FUEL
2 - BLOWDOWN
3 - FLYASI-
4 - EXHAUST GAS
FGD INLET
5 — EXHAUST GAS
FGD OUTLET
6 - SCRUBBER CAKE
7 - MAKEUPWATER
8 - SCRUBBER FEED SOLIDS
Figure 4-1. Boiler system schematic and sampling locations.
Gaseous Effluents
The boiler flue gas was sampled at the inlet and the outlet of the
pilot flue gas desulfurization unit. Integrated bag samples were taken *
at both points during each test. On-site analyses of C02» 02» N2 and C^
Cg organics were conducted. Continuous monitors were used to analyze
CO, N0/N0x, S02 and total hydrocarbons (as CH4). Figure 4-2 1s a
schematic of the continuous monitor setup. A Thermal Electron Corporation
(TECO) gas conditioner was used to remove condensate and particulate from
the gas sample. The instruments used are specified 1n Table 4-1.
Isokinetic sampling was performed at each location during all tests
using four different sampling trains (gaseous streams only).
The Source Assessment Sampling System (SASS) was used to collect Level
1 gaseous and particulate emission samples at the scrubber Inlet and out-
let. The SASS train 1s Illustrated 1n Figure 4-3. The train consists of
a heated probe, three cyclones and a filter in a heated oven. The cyclones
were used only during the coal Inlet tests. During the other tests the
4-2
-------
TABLE 4-1. PARAMETERS SAMPLED FOR COAL AND OIL FIRING
Location
Parameter
Sampling Method
Analysis
1
FUEL (coal & oil)
C, H, N, S, ash,
moisture,
heating value
Inorganics
Grab
Ultimate (lab)
Level II (lab)
4&5
COMBINED BLOWDOWN
Alkal1n1ty/ac1d1ty
PH
conductivity
hardness
TSS
nitrate
sulfate
sul f1 te
phosphate
ammonia
nitrogen
organlcs
FLYASH
Inorganics
organlcs
Composite dipper On-site HACH kit
Composite grab
FLUE GAS (Inlet & outlet)
CO
CO,
N0/N0,/N0
Ng.Og
so2
SOo/SO,
H2S0-,I1C1, HF,
particulate sulfate,
total hydrocarbons
(as CH )
Cj - Cjj organlcs
particulate & vapor
particulate sizing
SCRUBBER CAKE
Inorganics
organlcs
Level 1 & 2 (lab)
Level 1 & 2
Level 1 & 2
(lab)
(lab)
Continuous, Beckman
Model 865
Grab (bag)
Grab (bag)
Continuous, TECO
Model 10A
Grab (bag)
Continuous, TECO
Model 41
Goksoyr-Ross
Continuous, Beckman
Model 400
Grab (bag)
SASS
Method 5
Anderson Impactor
SASS
Composite grab
Direct reading
GC (TCD) on site
GC (TCD) on site
Direct reading
GC (TCD) on site
Direct reading
Level 2 (lab)
Direct reading
GC (FID) on site
Level 1 (lab)
Level 2 (lab)
Level 1
Level 2
Level
Level
1 & 2
1 & 2
llfil
BOILER & SCRUBBER
MAKEUP WATER
organlcs
1norganlcs
SCRUBBER MAKEUP
SOLIDS
Top grab
Grab
Level 1 (lab)
Not required
Not required
4-3
-------
EXHAUST
GAS OUTLET
EXHAUST
GAS INLET
[HEAT TRACED SAMPLE LINES
GAS CONDITIONER
NO/NOX
ANALYZER
THC
ANALYZER
CO
ANALYZER
S02
ANALYZER
FGD UNIT
Figure 4-2. Flue gas continuous monitor setup.
HEATER
CON-
TROLLER
STACK T.C.
GAS COOIJR
SS PROBE
10 >1
GAS
TEMPERATURE
T.C. /
S-TYPE PITOT
CONVECTION
OVEN
1
OVEN
T.C.v
9~
XAD-2 /
CARTRIDGE
V~~4
INGE*
T.C
CONDENSATE
COLLECTOR
DRY GAS METER ORIFICE METER
CENTRALIZED TEMPERATURE
AND PRESSURE READOUT
10 CFM VACUUM PUMPS
CONTROL MODULE
Figure 4-3. Source assessment sampling system (SASS) schematic.
4-4
-------
particulate loadings were too low for the cyclones to work effectively.
The remainder of the system consists of a gas conditioning system, an XAD-
2 polymer absorbent trap and a series of Impingers. The polymer traps
gaseous organics and some Inorganics and the Impingers collect the remain-
ing Inorganics. All sample contact surfaces are Type 316 stainless steel,
teflon, or glass. The train was run for 6 to 8 hours until a minimum
of 30 cubic meters of gas had been collected.
Previous sampling and analysis experience had indicated that SASS
train materials may contaminate certain organic and Inorganic samples.
The combination 1s of concern only when the pollutant 1s present at a
concentration that 1s near the detection limit of the Level 2 methods.
To avoid that possibility all-glass sampling trains were used to collect
Level 2 samples. Method 5 sampling trains were modified as shown in
Figure 4-4 for organics and Figure 4-5 for Inorganics. Both trains
sampled approximately 10 cubic meters of flue gas during a 6 to 8 hour
test run.
A controlled condensate train (Goksoyr-Ross), as shown 1n Figure 4-6,
was used at each location to obtain samples for S02» SOg (as HgSO^),
particulate sulfate, HC1 and HF.
During Level 2 test runs, Anderson cascade Impactors were used to
obtain particulate samples by particle size fraction. A pre-separating
10n cyclone was used up-stream of the Impactor on the Inlet side.
Liquid Effluents
The combined boiler blowdown was sampled using the composite dipper
method and boiler and scrubber makeup water were sampled by the top grab
method. The samples were analyzed as shown 1n Table 4-1. In addition,
each sample was extracted with methylene chloride and the extracts returned
to the lab for further analysis.
Solid Effluents
Composite samples of the flyash and scrubber filter cake were collected
per Level 1 procedures and returned to the lab for analysis. Grab samples
of the scrubber feed solIds were also obtained.
4-5
-------
S0R8ENT MODULE
FILTER
^ ir T
CYCLONE
U GLASS LINED SS PROBE
^ PITOT TUBE
ORIFICE
MANOMETERS
Figure 4-4. Organic sampling train.
s
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
/
GREENBURG-S.YITH
IMP INGERS
{c=A frl\ J]
t?) <¥ <¥ <¥ &
FILTER
CYCLONE
GLASS LINED SS PROBE
PITOT TUBE
lj m
ORIFICE
LJ
MANOMETERS
Figure 4-5. Inorganic sampling train.
4-6
-------
RUIBER VACUUM
HOSE
ADAPTER FOR CONNECTING HOSE
TC WELtv
VACUUM
(GAUGE
askstos cloth
INSULATION
GLASS-COL
HEATING
MANTLE
STACK
QUARTZ
FILTER
HOLDER
fHREE WAY
VALVE
SILICA GEL
THERMOMETER
Figure 4-6. Controlled condensation train.
LABORATORY ANALYSES
The samples Identified 1n Table 4-2 were returned to the laboratory
for analysis. The Level 1 analysis scheme for particulates and gases from
the SASS train 1s shown 1n Figure 4-7. The Level 1 analysis scheme for
solids, slurries and liquids 1s shown 1n Figure 4-8. Detailed analysis
procedures can be found 1n the manual "Combustion Source Assessment
Methods and Procedures Manual for Sampling and Analysis," September 1977.
The Level 2 Inorganic liquid sample analysis scheme 1s shown 1n
figure 4-9 with the Inorganic solids analysis scheme shown 1n Figure 4-10.
The analysis scheme for the controlled condensate train 1s presented in
Figure 4-11. A discussion of the specific procedures used for this program
and their results can be found 1n Volume III Appendix C.
The Level 2 analysis scheme for organic SASS and modified Method 5
components 1s shown 1n Figure 4-12. The analysis scheme for organic
aqueous samples 1s shown 1n Figure 4-13. A discussion of the specific
procedures and results for this program can be found 1n Volume III
Appendix C.
4-7
-------
TABLE 4-2. SUMMARY OF SAMPLES
SASS Train
Probe rinse
Cyclone catch
Filter
Resin
Resin Condensate
Implnger solutions
Controlled Condensation
Probe rinse
Filter
Impinger solutions
' TO LABORATORY FOR ANALYSIS
Inorganic and Organic Trains
Probe rinse
Cyclone catch
Filter
Implnger solution
Liquid Samples
Neat
Extracts
Solid Samples
As found
PARTICULATE
MATTER
so^/so3
OPACITY
(STACKS)
•WEIGH
INDIVIDUAL
CATCHES
"If INORGANICS
ARE GREATER THAN
10°* OF TOTAL CATCH.
PROBE AND
CYCLONE
RINSES
SASS TRAIN GAS
CONDITIONER
CONDENSATE
ORGANICS
EXTRACT
SASS TRAIN
IMPINGERS
INORGANICS
PHYSICAL SEPARATION
INTO FRACTIONS,
LR/IR/MS
ELEMENTS (SSMS) AND
SELECTED ANIONS
1ST
2ND
W
INORGANICS ELEMENTS AND
5t, At, 5b, Hg SELECTED ANIONS
>10M *
3-10
EXTRACTION
inorganics
ELEMENTS (SSMS) AND
SELECTED ANIONS
PHYSICAL SEPARATION
INTO FRACTIONS
IC/IR/MS
1-3» '
<)U *
FILTER
SAME AS ABOVE
INORGANIC
(GRAB)
ORGANIC
MATERIAL Ca
ONE-SITE GAS
CHROMATOGRAPHY
XAD-2
ABSORBER, -
J INORGANICS CEMENTS (SSMS) AND
1INUROANICS SELECTED ANIONS**
PHYSICAL
SEPARATION
INTO FRACTIONS,
LR/IR/MS
INORGANICS
ELEMENTS (SSMS) AND
SELECTED ANIONS
ORGANIC
MATERIAL C. - C,
MODULE RINSE
ON-SITE GAS
CHROMATOGRAPHY
EXTRACTION
PHYSICAL SEPARATION
INTO 8CLASSES
ORGANICS
>Cti
ALIQUOT FOR GAS
CHROMATOGRAPHIC
ANALYSIS
PH/SICAl SEPARATION
INTO FRACTIONS
LC/IR/MS
Figure 4-7. Basic level 1 sampling and analytical scheme for
particulates and gases.
4-8
-------
SOURCE
SOLIDS
SLURRIES
LIQUIDS
LEACHABLE
MATERIALS
ft
u
ORGANICS
INORGANICS
PHYSICAL SEPARATION
INTO FRACTIONS LC/IR/MS
ELEMENTS (SSMS) AND
SELECTED ANIONS
¦* INORGANICS
ORGANICS
SUSPENDED
SOLIDS
ELEMENTS (SSMS) AND
SELECTION ANIONS
PHYSICAL SEPARATION
INTO FRACTIONS,
LC/IR/MS
INORGANICS
ORGANICS
ELEMENTS (SSMS) AND
SELECTED ANIONS
PHYSICAL SEPARATION
INTO FRACTIONS
LC/IR/MS
INORGANIC^ ELEMENTS (SSMS) AND
INQK0A1NIV.J ANIONS
SELECTED
WATER
TESTS
ORGANIC
EXTRACTION
OR DIRECT
ANALYSIS
ORGANICS
ORGANICS
PHYSICAL SEPARATION
INTO FRACTIONS,
LC/WMS
ALIQUOT FOR GAS
CHROMATOGRAPHIC
ANALYSIS
Figure 4-8.
Basic level 1 sampling and analytical scheme for
solids, slurries and liquids.
ANIONS ASSf JSIO
CATION/ANION )
CALCULATED AND
MQUILf LIST OF
MAtfGSPfOfS ,
identified and }
V QUANT I Fit D /
II A MASS \
lALANCt
ACCOMPLISHED^
CXACT CATION
ANO ANION
RATIO* IN THIS
SAMFU? /
AND ANIONS
HAVI NOT ICfN
ASSIGNED TO .
f»TI CONCENTtATIONSl
EVALUATED ANO
tlNSULTS ESTABLISHEDI
(KNOWLEDGE OF
PMYVCAl
v nommiL,
/koiahi stamK
SOLID SPECIES?
ICOMfOUNOS AND.
\ VALANCE /
\ STATES) /
CSTAKISHCO ANO
SAMPLE EVAPOftATED
AND HANDLED AS I
[CATIONS AND ANIONS
\ WITH ESTAIUSHfD ,
\ CONCENTRATIONS >
WET TESTS.SPECIFIC
ION ELECTIOOE,
CHROMATOGRAPHY
ON SPECIFIC
CATIONS FROM
Figure 4-9. Level 2 Inorganic liquid sample analysis scheme.
4-9
-------
Figure 4-10. Level 2 inorganic solids analysis scheme
-------
ANALYSIS
TRAIN COMPONENT
PREPARATION
SPECIES POUND
EXTRAC1
SOLIDS
EXTRACT
SOLIDS
OI8CARO
SAME AS
PROSE PARTICULATE
MATTER
SO
CI". F . SO
SOLUBLE SO,
CI", F~
WATER
INSOLUBLE
SOA F-
SOr HCI.
HP
WEIGHT
AND
RECORD
RINSE
PROBE RINSE
BOIL TO
DESTROY
H202
PROBE
PARTICULATE
MATTER
HOT 10%
HCI EXTRACTION
HOT H20
EXTRACTION
FILTER
(AND BLANK)
CONTROLLED
CONDENSATION
COIL RIN8E
H,0, IMPIN6ER
DILUTE TO
100 ML
EMPTY IMPINQBR
H,0, IMPINQER
SILICA QEL
WEIGH
AND
RECORO
COMBINE
WITH PROBE
PARTICULATE
MATTER ANO
EXTRACT
ADO ENOUGH
HIOH PURITY
NtjCOg TO
MAKE ALKALINE
PH IB-SI
Figure 4-11. Analysis scheme for controlled condensate train components.
4-11
-------
SASS TRAIN
RETAINED SAMPLES
I
IN>
OV-101
WEIGH
COMPARE WITH
IC FRACTION
GC/MS OATA
PROtE WASH
SAMPLE
GC/MS ON AUQUOT
OV-17 COLUMN
WEIGH
DEXIL 300
1 M
CYCLONE ~ FILTER
SAMPLE
CONCENTRATE
(AIR DRY)
WEIGH
FRACTIONATE
BY LC
Zn AND 10*
CYCLONES
SAMPLE
COMPARE WITH
LC FRACTION
GC/MS DATA
WE! GH
ON ALIQUOT
OV-17
GC/MS
WEIGH
CONCENTRATE
(AIR DRY)
WEIGH
OV-17
XAD-Z
MODULE
FRACTIONATE
PENTANE
BY LC
DATA SIMPLE
WELL RESOLVED
AND EASY TO
INTERPRET
VN EXTRACT
GC/MS
OV-17
AUQUOT
CONCENTRATE
AIR DRY
WEIGH
WEIGH
FRACTIONATE
BY LC
WEIGH
x ALL
STOPj COMPOUNDS
IDENTIFIED
2ND
TIME
AROUND
WEIGH
GC/MS
TENAX GC
OR
h3po4/car»owax
ALIQUOT
DATA SIMPLE
WELL RESOLVED
AND EASY TO
INTERPRET
ACIO EXTRACT
PH <2
Ms Q
CONCENTRATE
AIR DRY
WEIGH
FRACTIONATE
BY LC
WEIGH
2ND \ m /—S.
TIME \i i fsTOM
OUND/ ^ J COMPOUNDS
IDENTIFIED
CONCENTRATE
XI00
GC/MS
OV-17
FFAP
KOH-CARBOWAX 20 M
5
6
10
8d
8o,b
12
11
13
13®
23 (LOW BOILING)
8c
18
19
20
23
0 a MAY K USEFUL TO EASE
SPECTRAL INTERPRETATION
(D SULFUR COMPOUNDS ARE REACTIVE AND
MAY SQUIRE SPECIAL CARE
(3) NITROSOAM]NCS AM EXPCC1ED ONLY
AT LOW CONCENTRATIONS, IF PRESENT
SPECIAL PREPARATION WILL K REQUIRED
FOLLOWED BY GC/MS ANALYSIS USING
CARiOWAX 20 M AND MULTIPLE ION
DETECTION (MID)
FFAP OR
SP-216-PS
DCttVUIZE
GC/MS
TMS
OR
—
ov-toi
8
8
METHYLATE
u
Figure 4-12.
General logic flow chart for level 2 organic SASS component analysis.
-------
mmcg compounds
lOfNIRJO
SWCTUl INTIIftfTATlON
FOtCOtUMN TCfcMINATORS WHIN
i|OP I MMtG COMPOUNDS
StOPj MMEC COMPOUNDS
CONCINTUT! FOC
IC KPAIATION
A5 fO« SA5S SAMPLES
), I*, 14 LOW lOllMS ONIV)
», [1]), 14, II, l«, W.flJ)
m
01HCT AOUIOUi INJICtlON GC Mi
ANALVSI) FO# HIGH CONCINllATlON
I < * mi ¦ ujtryt
lb. 11,17, (13). 15,14.17, J 1, *MIN|S If PWStNT
CAHGOIItS 10,12
Figure 4-13. General logic flow chart for level 2 organic aqueous samples.
4-13
-------
4-14
-------
SECTION 5
COMPREHENSIVE ASSESSMENT OF COAL FIRING
CASE FOR AN INDUSTRIAL BOILER
This section provides a comprehensive multimedia assessment of
emissions/effluents associated with a coal-fired industrial boiler equipped
with an FGD system. Data from Level I/Level II sampling and analyses are
utilized to quantitatively determine emissions in gas, solid and liquid
waste streams and to evaluate performance of pollution control equipment
in use daring coal firing. Waste stream pollutant concentrations are
compared with Minimum Acute Toxicity Effluent (MATE) values, where appro-
priate, to provide an indication of risk to public health and ecology.
Simplified air quality models are used to determine the relative ground
level air quality resulting from both uncontrolled and controlled emissions.
TEST CONDITIONS
Five tests were performed with a coal-fired industrial boiler gener-
ating from 34,000 to 44,200 kg steam per hour (75,000 to 97,500 pounds per
hour) which corresponds to between 75 and 97% of full load operation for
the unit tested. Specific test conditions are summarized in Table 5-1.
Tabulated fuel feed rates are the nominal feed rates maintained during
each test. Assuming a constant thermal efficiency of 90%, steam produc-
tion data and coal analyses indicate that nominal fuel feed rates are
accurate to within approximately 13%. Oxygen concentrations provided in
Table 5-1 were measured in flue gas samples drawn from the inlet of the
system's wet scrubber unit. Due to air leakage into upstream ducting
operating at sub-atmospheric pressure and possible air leakage into the flue
gas bag sampling system, tabulated data are not representative of furnace
gas concentrations. During normal operation, the concentration of oxygen
in the furnace after combustion is between 3 and 4% which corresponds to
an excess air input of 16 to 23% during coal firing. Excess air estimates
presented in Table 5-1 were computed assuming a mean oxygen concentration
of 3.5% in the furnace and utilizing fuel analyses data.
Test data relating to scrubber throughput and loading, and total flue
gas generation rates are presented in Table 5-2. Measured flow rates
5-1
-------
TABLE 5-1. SUMMARY OF TEST CONDITIONS - COAL FIRING
r..t WMm Production Rate * of Nominal Estimated*
La IbT- Maximum Coal Feed Scrubber Excess A1r
steam/hr steam/hr Load Rate, Inlet to Furnacet
kg/hr
200
39,700
87,500
87.5
3629
7.8
20
201-1
44,200
97,500
97.5
3629
8.2
20
201-2
43,100
95,000
95.0
3629
8.4
20
201-3
34,000
75,000
75.0
3175
8.3
20
201-4
40,800
90,000
90.0
3629
6.7
20
Due to air leaks in ducting upstream of the scrubber inlet, tabulated Og
values are not representative of combustion zone O2 concentrations.
Combustion zone O2 concentrations normally range from 3 to 4% for this unit
02 - CO/2
% excess air is estimated to be 100 x q ^ ^—(Oo - CO/21
where: ' 2 " 2 )
O2 was assumed to be 3.5% and other species concentrations are computed
from fuel analyses.
TABLE 5-2. FRACTION OF FLUE GAS PROCESSED BY THE SCRUBBER
DURING COAL COMBUSTION
Test No.
Flow Rate
at Scrubber
Inlet, *
dscm/min
% of Design
Load
Total
Flue Gas
Flow Rate,
dscm/m1n*
Fraction of
Total Flue Gas
Processed by
the Scrubber
200
99
56
754
0.13
201-1
91
51
761
0.12
201-2
89
50
798
0.11
201-3
98
55
684
0.14
201-4
102
58
706
0.14
Average
96
54
741
0.13
* Dry standard cubic meters per minute (dscm/min).
5-2
-------
through the scrubber inlet expressed as dry standard cubic meters per
minute (dscm/min) and percentage of design scrubber loading are presented
in the first two columns. For discussion in this report, standard temper-
ature and pressure are defined as 20°C and one atmosphere pressure,
respectively. Although the scrubber was designed for a capacity of
approximately 177 dscm/min (6,250 dscfm), it is a manually controlled
variable venturi type unit capable of processing a substantial range of
loadings. A loading of 90 to 102 dscm/min (51 to 57% of design loading)
was maintained during testing, rather than full loading, because failure
of the multiclone particulate removal unit upstream of the scrubber
resulted in high solids loading at the scrubber and unacceptably high
scrubber cake production rates. The scrubber is a pilot unit and, as such,
was not sized to process the total flue gas output of the furnace. Typi-
cal inlet and outlet gas temperatures for the scrubber unit were 300°F and
125°F, respectively. Total flue gas flow rates presented in the table
were computed from coal analyses, coal feed rate data and flue gas analy-
ses utilizing the following expression:
4.762 (nc + ns) + .9405 nH - 3.762 nQ^
A - ¦
r . 1 - 4.762 (02/100)
where: npg = gm moles of dry effluent/gm of fuel.
n. = gm moles of element j in fuel per gm of fuel.
J
02 = volumetric 02 concentration in percent.
As Indicated in Table 5-2, the flue gas slip stream drawn for scrubber
processing ranged from 11 to 14% of the total flue gas generated.
A high volatile bituminous feed coal was utilized for all coal-fired
tests. Ultimate analyses of feed coal samples obtained during each of the
five tests performed are presented in Table 5-3. These data indicate an
essentially constant feed coal composition during the five days of testing.
Average coal moisture and ash contents were 7.15 and 9.90 % w/w, respec-
tively. Mean coal sulfur, nitrogen and chlorine contents were 1.64, 0.92,
and 0.12 % w/w, respectively. An average coal heat content of 29,485 kJ/kg
(12,683 Btu/pound) was determined. Additional analyses were performed on
feed coal samples from test 201-1 to determine concentrations of 17 trace
5-3
-------
TABLE 5-3. SUMMARY OF ULTIMATE COAL ANALYSES
Test Number
Weight % 200 201-1 201-2 201-3 201-4 Average a*
Moisture 8.44 7.37 6.54 7.18 6.23 7.15 0.86
Carbon 71.69 70.62 73.30 71.94 72.97 72.10 1.07
Hydrogen 4.33 4.18 4.31 4.27 4.30 4.28 0.06
Nitrogen 0.86 1.00 0.88 0.87 0.98 0.92 0.07
Chlorine 0.10 0.10 0.12 0.13 0.15 0.12 0.02
Sulfur 1.64 2.00 1.38 1.68 1.50 1.64 0.23
Ash 8.94 11.19 9.41 9.85 10.09 9.90 0.85
Oxygen 4.00 3.54 4.06 4.08 3.78 3.89 0.23
kJ/kg 29,263 28,872 29,997 29,419 29,874 29,485 459
~
a - One standard deviation.
elements (Ca, Mg, Sb, As, Cd, Cr, Co, Cu, Pb, Mn, Mo, Ni, V, Zn, Se, Sr,
and Zr) and 2 minor elements (Fe and Al). The method employed for analysis
of these elements was inductively coupled plasma optical emission spec-
troscopy (ICPOES) which is generally considered to be more accurate than
spark source mass spectrometry (SSMS). However, a single feed coal sample
from test 200 was analyzed for the trace elements boron and beryllium by
SSMS, and for mercury by cold vapor analysis. These data are presented
in Table 5-4. Considering the uniformity of feed Coal ultimate analyses
obtained during the test period, it appears reasonable to assume that
tabulated trace and minor element analyses are typical of the coal fired
during the five day test period. Although analyses of other coal samples
from the same source (mine or cleaning plant) are not available for
direct comparison, analyses of most trace and minor elements presented
in Table 5-4 appear to be consistent with concentration limits typifying
Appalachian and Eastern Interior Basin coals. No coal strontium analyses
were found for comparison. Trace elements present in somewhat higher con-
centrations than are indicated to be typical by the limited published data
are antimony, arsenic, cobalt, copper, molybdenum, and zirconium. However,
5-4
-------
TABLE 5-4. CONCENTRATION OF MAJOR TRACE ELEMENTS IN COAL-TEST 201-1
Element pg/q Coal Reference
Sample A Sample B Average Typical Ranget
Ca
820
720
770
0
- 1600
2,4
Mg
400
300
350
0
- 959
2,4
Sb
80
90
85
0.2
- 8.9
2
As
140
133
137
0.5
- 93
1,2
B
-
-
~
2.4
4
- 115
1,3
Cd
3
4
3.5
0.1
- 65
2
Cr
49
47
48
4
- 144
1,3
Co
102
152
127
0.5
- 43
1,2
Cu
98
46
72
3
- 61
1,2
Fe
10,750
13,750
12,250
0.3
- 40,000
2,4
Pb
77
95
85
4
- 218
1,2
Mn
13
11
12
6
- 181
2
Mo
238
348
293
0.4
- 30
1,2
Ni
23
51
37
2
- 80
1,2
V
49
44
47
2
- 147
1,2,3
Zn
48
64
56
6
- 5,350
2
Se
68
77
73
0.4
- 74
2,3
Sr
72
64
68
NO
DATA
A1
15,900
12,100
14,000
0.4
- 40,700
2,4
Zr
325
215
270
8
- 133
2
Be
-
-
2.3*
0.6
- 4.1
1
Hg
-
-
0.14+
0.07
- 0.49
1
Boron and beryllium analysis were performed by SSMS on a feed coal
sample from test 200.
+ Typical range for Appalachian and Eastern Interior Basin coals.
+ Mercury was determined by cold vapor analysis of a coal sample
from test 200.
5-5
-------
the significance of these higher concentrations is not apparent due to
the limited quantity of published data and the complete absence of source
specific data.
STACK EMISSIONS
As discussed previously, the wet scrubber unit processed from 11 to
14% of the total generated flue gas. Flue gas analyses were performed on
samples drawn from the scrubber inlet and outlet. Results of these
analyses were utilized to estimate total boiler emissions on the basis of
100% of the flue gas being treated by a scrubber. That is, it was assumed
that additional scrubber modules could be added to the system such that
the total flue gas output would be processed with a mean scrubber effi-
ciency identical to that obtained using the pilot scrubber. All emissions
data presented in the following sections are based on this assumption.
Criteria Pollutants
Federal New Source Performance Standards (NSPS) currently in effect
define allowable N0X (as N02), S02, and total particulate emission rates
from utility boilers having 25 MW or greater output. Existing NSPS
limitations on N0X emissions from coal fired units are 300 ng/J (0.70 lb/MM
Btu) although a proposed standard of 200 ng/J (0.50 lb/MM Btu) maximum
emissions and 65% reduction of uncontrolled emissions is under considera-
tion. Proposed N0X standards are based on boiler emission levels achiev-
able with proper excess air control and staged combustion. The NSPS
limitation on SO2 emissions for coal fired units is 520 ng/J (1.20 lb/MM
Btu). Potential standards based on S02 emission levels achievable through
flue gas desulfurization (FGD) would impose the further requirement for
90% reduction of uncontrolled emissions from sources producing more than
85 ng/J (0.20 lb/MM Btu). The NSPS limitations for total particulates is
currently 43 ng/J (0.10 lb/MM Btu). A proposed particulate standard of
13 ng/J (0.03 lb/MM Btu) maximum emissions and 99% reduction of uncontrolled
emissions is being considered. Proposed particulate standards are based
on emission levels achievable with electrostatic precipitators (ESP) and
baghouses. Federal NSPS do not currently address either CO or total
hydrocarbons.
5-6
-------
Similar standards relating to industrial boilers have not been promul-
gated to date. Therefore, criteria pollutant emission data presented in
this section will be discussed in the context of the existing and proposed
Federal NSPS for utility boilers.
As mentioned previously, 5 coal-fired tests were performed on the
Firestone industrial boiler. Criteria pollutant concentrations were
generally measured at frequent intervals during each test and averaged to
obtain the mean concentration for the test. The 5-test averages of criteria
pollutant emissions data are presented in Tables 5-5 through 5-8. Average
emissions data from the individual tests are presented in Tables 5-9 and
5-10. Additionally, the 5-test averages of scrubber Inlet data are pre-
sented in Table 5-11 for comparison with the EPA AP-42 emission factors
for uncontrolled sources. The data are discussed by specific compound in
the ensuing sub-sections.
Nitrogen Oxides
Mean NO emissions measured during the testing period were 421 ng/J
J\
(0.98 lb/MM Btu) prior to FGD contacting. Full load N0X emissions were
approximately 491 ng/J (1.14 lb/MM Btu) at the scrubber inlet. This emis-
sion rate is somewhat higher than the average uncontrolled emission rate
of 343 ng/J (0.80 lb/MM Btu) tabulated in AP-42 for bituminous coal fired,
dry bottom, 10 MW industrial boilers. However, examination of published
industrial boiler data indicates that N0X emissions measured during the test
period are not at variance with typical ranges of N0X emissions (Reference 5).
The average measured N0X emission rate after FGD was 372 ng/J (0.87 lb/MM
Btu) which is nearly 25% higher than the current NSPS limitation of 300 ng/J.
As indicated by data presented in Tables 5-1 and 5-9, N0X data varied with
boiler load, as expected. Test 201-3 was performed at the lowest boiler
load (75% of full load) and, therefore, resulted in the lowest N0X emis-
sions measured. Indeed, the average N0X emission rate measured at the
scrubber outlet during test 201-3 was 259 ng/J (0.60 lb/MM Btu) which is
lower than the current NSPS limitation.
N0y data generally indicate a reduction of N0X emissions across the
scrubber during the test period. Measured N0X data obtained during test
201-2 are presented in Figure 5-1 to display typical fluctuations in N0X
5-7
-------
TABLE 5-5 CRITERIA POLLUTANT EMISSIONS FOR A COAL-FIRED INDUSTRIAL
BOILER IN CONCENTRATION UNITS (5 TEST AVERAGE)
Pollutant mg/Nm (Grain/SCF)
Before Scrubber After Scrubber
NOx (as N02) 992 (0.43) 880 (0.38)
CO 37 (0.02) 34 (0.01)
S02 2600 (1.14) 90 (0.04)
Organics (as CH^) 14 (0.01) 15 (0.01)
Total Particulates 7100 (3.10) 45 (0.02)
TABLE 5-6. CRITERIA POLLUTANT EMISSIONS FOR A COAL-FIRED INDUSTRIAL
BOILER IN TEMPORAL UNITS (5 TEST AVERAGE)
Pollutant kg/hr (lb/hr)
Before Scrubber After Scrubber
NOx (as N02) 44.3 (97.7) 39.2 (86.4)
CO 1.66 (3.66) 1.50 (3.31)
S02 117 (258) 3.80 (8.38)
Organics (as CH^) 0.59 (1.30) 0.65 (1.43)
Total Particulates 315 (694) 1.99 (4.39)
5-8
-------
TABLE 5-7. CRITERIA. POLLUTANT EMISSIONS FOR A COAL-FIRED INDUSTRIAL
BOILER IN THERMAL UNITS (5 TEST AVERAGE)
Pollutant
nq/J (lb/MM Btu)
Before Scrubber
After Scrubber
NOx (as N02)
421 (0.98)
372 (0.87)
CO
15.9 (0.04)
14.3 (0.03)
so2
1112 (2.59)
36.3 (0.08)
Organlcs (as CH^)
5.79 (0.01)
6.29 (0.01)
Total Particulates
2951 (6.86)
18.6 (0.04)
TABLE 5-8. CRITERIA POLLUTANT EMISSIONS FOR A COAL-FIRED INDUSTRIAL
BOILER IN PRODUCTION RATE UNITS (5 TEST AVERAGE)
Pollutant
mg/kg steam (lb/1000 lb steam)
Before Scrubber
After Scrubber
NOx (as N02)
1089 (1.09)
962 (0.96)
CO
41.2 (0.04)
37.1 (0.04)
so2
2883 (2.88)
93.8 (0.09)
Organlcs (as CH^)
15.4 (0.01)
16.7 (0.02)
Total Particulates
7570 (7.57)
47.6 (0.05)
5-9
-------
TABLE 5-9. SUMMARY OF CRITERIA POLLUTANT EMISSIONS-COAL FIRING
Test No. ng/J (Ib/MM Btu)
NO CO SO? HC* C-i - Cfi C7 - C.- Organics Total
(as CH^) Higher Particulates
(as NOp) Organics Organics Than
Cl6
200 Inlet
417
20.7
1009
3.79
<5.49
0.34
2.28
2361
(0.97)
(0.05)
(2.35)
(0.01)
(<0.01)
(0.00)
(0.01)
(5.49)
200 Outlet
367
18.5
25.4
4.22
<5.49
0.27
0.33
14.3
(0.85)
(0.04)
(0.06)
(0.01)
(<0.01)
(0.00)
(0.00)
(0.03)
201-1 Inlet
491
16.5
1284
4.35
<5.65
3122
(1.14)
(0.04)
(2.99)
(0.01)
(<0.01)
(7.26)
201-1 Outlet
457
15.7
39.0
5.22
<5.65
20.9
(1.06)
(0.04)
(0.09)
(0.01)
(<0.01)
(0.05)
201-2 Inlet
455
17.7
1295
0.88
<5.75
(1.06)
(0.04)
(3.01)
(0.00)
(<0.01)
201-2 Outlet
358
15.9
35.5
1.33
<5.75
(8.33)
(0.04)
(0.08)
(0.00)
(<0.01)
201-3 Inlet
330
16.2
1028
10.9
<5.69
(0.77)
(0.04)
(2.39)
(0.03)
(<0.01)
201-3 Outlet
258
12.7
31.8
10.9
<5.69
(0.60)
(0.03)
(0.07)
(0.03)
(<0.01)
201-4 Inlet
409
8.56
942
8.95
<5.06
3370
(0.95)
(0.02)
(2.19)
(0.02)
(<0.01)
(7.84)
201-4 Outlet
420
8.95
49.7
9.73
<5.06
20.6
(0.98)
(0.02)
(0.12)
(0.02)
(<0.01)
(0.05)
Average Inlet
Average Outlet
421
(0.98)
372
(0.87)
15.9
0.04)
14.3
0.03)
1112
(2.59)
36.3
(0.08)
5.79
(0.01)
6.29
(0.01)
<5.53
(<0.01)
<5.53
(<0.01)
0.34
(0.00)
0.27
(0.00)
2.28
(0.01)
0.33
(0.00)
2951
(6.86)
18.6
(0.04)
* Total hydrocarbons as determined by flame ionization detection (FID) analysis.
-------
TABLE 5-10. SUMMARY OF CRITERIA POLLUTANT EMISSIONS - COAL FIRING
Test No. kg/hr
NOx
(as N02)
CO
so2
HC
(as CH4)
C1 " C6
Organics
C7 ' C16
Organics
High
Molecular
Weight
Organics
Total
Particulates
200 Inlet
45
2.2
106
0.40
<0.59
0.04
0.25
248
200 Outlet
39
2.2
3.0
0.56
<0.59
0.001
0.004
1.7
201-1 Inlet
51
1.7
134
0.45
<0.60
--
—
324
201-1 Outlet
46
1.6
3.9
0.53
<0.60
—
—
2.1
201-2 Inlet
49
1.9
139
0.09
<0.62
—
—
--
201-2 Outlet
40
1.8
3.9
0.15
<0.62
—
—
—
201-3 Inlet
30
1.5
97
0.97
<0.53
--
--
—
201-3 Outlet
23
1 .1
2.7
1.01
<0.53
—
—
—
201-4 Inlet
44
0.9
TOO
0.96
<0.55
—
—
362
201-4 Outlet
46
0.9
5.2
1.00
<0.55
—
—
2.1
Average Inlet
44
1.6
115
0.58
<0.57
0.04
0.25
311
Average Outlet
39
1.5
3.7
0.64
<0.57
0.001
0.004
2.0
-------
TABLE 5-11. COMPARISON OF CRITERIA POLLUTANT EMISSIONS
WITH EMISSION FACTORS FOR UNCONTROLLED BOILERS
Pollutant nq/J (Ib/MM Btu)
Test Data Average Emission Factors^
Before Scrubber for Uncontrolled Sources
N0X (as NOg at full load)
491
(1.14)
343
(0.80)
CO
15.9
(0.04)
19
(0.04)
(/>
O
ro
1112
(2.59)
1189
(2.77)
Organics
5.79
(0.01)
5.72
(0.01)
Total Particulates
2951
(6.86)
3212
(7.47)
Factors are computed from AP-42 values using the national average
bituminous coal higher heating value of 11,263 Btu/lb (Reference 6).
levels observed during an eight hour test. Average N0X removals were
computed for each test by integrating the area between plots of scrubber
inlet and outlet concentrations versus time and averaging the total area
with respect to time. Time integrated N0X removal data are presented in
Table 5-12. The slight increase in N0X emissions measured during test
201-4 results from a single outlying scrubber outlet data point in a group
of data which otherwise corresponds well with inlet data. Thus, the appar-
ent increase is considered not to be significant. Otherwise, these data
indicate that average N0X removals across the scrubber ranged from 6 to 24%.
Although N0X removal in both wet and dry F6D systems has been reported
in the literature, no information is available regarding the chemistry of such
occurrences (References 7 & 8). It is feasible that some degree of N0X removal
may be effected by dissolution of N0£ in the slightly acidic scrubber
solution. However, it should be noted that data presented in Table 5-12 do
not indicate a correlation between the extent of N0X removal and scrubber
inlet N0X concentration. And, as mentioned previously, inlet and outlet
gas temperatures and flow rates in the scrubber unit were essentially iden-
tical for all tests. Thus, the extent of N0X removal cannot be correlated
to variables monitored during testing. Further, as indicated by data in
Table 5-9, N0X removal trends are paralleled by CO removal trends. These
5-12
-------
cn
i
u>
600
i
Q_
*
z
0
1 500
i—
z
UJ
U
z
8 400
X
O
z
§ 300
z>
to
<
UJ
2
200
(
i
» INLET
L OUTLET
<
~ • —j
/v
V.
\
\
a
V
f
8
10
11
12 13
TIME, HOURS
14
15
16
17
Figure 5-1.
N0X Concentration Profiles Measured
at Scrubber Inlet and Outlet During
Test 201-2.
-------
TABLE 5-12
. SUMMARY OF N0X
REDUCTION DATA
Test
No.
Average NO As NO,
A C.
ng/J (ppm)
Inlet Outlet
Average NO
Integration
ng/J (ppm)
Removal by
Method*
%
200
417
(515)
367
(456)
37
(46)
8.9
201-1
491
(591)
457
(551)
28
(34)
5.7
201-2
455
(539)
358
(424)
93
(110)
20.3
201-3
330
(395)
258
(309)
79
(95)
24.0
201-4
409
(547)
420
(564)
Not Significant
The area between plots of inlet and outlet N0X concentrations vs time
was integrated and averaged with respect to time to yield average
removals.
observations may indicate that measured N0X (and CO) removal is actually
a sampling phenomenon, perhaps related to small air leaks in the sampling
train and subsequent dilution of the sample gas prior to analysis.
Carbon Monoxide
Carbon monoxide emission rates in the flue gas stream were measured
to be 15.9 ng/J (0.04 lb/MM Btu). Measured CO emission rates correspond
well with the value reported in AP-42, namely 19 ng/J (0.04 lb/MM Btu). A
slight reduction in CO emissions was observed after scrubbing in most tests
with an average reduction of 10% being measured. However, the significance
of measured CO emission reductions at such low concentrations is question-
able due to low analytical sensitivity (approximately 7% at these CO levels)
and, as discussed for NO data, data trends and lack of correlation may
indicate a sampling phenomenon.
5-14
-------
Sulfur Dioxide
Average SO2 emission rates prior to scrubbing were measured to be
1112 ng/J (2.59 lb/MM Btu). Measured uncontrolled SO2 emission rates
compare favorably with the value of 1189 ng/J (2.77 lb/MM Btu) reported
in AP-42. Average SO2 emission rates measured after scrubbing were
36.3 ng/J (0.08 lb/MM Btu) which corresponds to a mean scrubber efficiency
of 96.7%. Thus, SO2 emissions after FGD were substantially lower than
either existing or proposed NSPS limitations.
Hydrocarbons
Emissions of organic material measured as methane were found to be
5.79 ng/J (0.01 lb/MM Btu), on the average, prior to scrubbing. Measured
total organic emissions compare well with the tabulated AP-42 value of
5.72 ng/J (0.01 lb/MM Btu). It should be noted that, during these tests,
the flue gas analyzed was processed by a gas conditioner. Therefore,
higher molecular weight organics may have been condensed or scrubbed from
the flue gas prior to flame ionization detection (FID) analysis. As will
be discussed subsequently, total hydrocarbon emissions measured by FID
analysis of the scrubber inlet gas may be low by approximately 30%. The
slight increase in hydrocarbons across the scrubber, although statistically
insignificant with respect to mean hydrocarbon values due to concentration
fluctuations during testing, has statistical significance with respect to
available real time inlet and outlet data pairs. That is, based on avail-
able inlet and outlet data pairs (measured not more than 30 minutes apart),
outlet samples are biased high with respect to inlet samples. The magni-
tude of the bias is approximately 5 ppm (^1.5 ng/J) at the excess oxygen
levels measured. The cause of the observed bias is not known at the
present time. However, data from recent tests will be available to deter-
mine whether data bias is related to gas sampling techniques. The possi-
bility of bias arising from moisture interference, improper FID calibration
or variable sample gas flow rate has been evaluated and subsequently
discarded.
In addition to FID analyses, gas chromatograph analyses were performed
on limited bag samples of flue gas and sample catches from the Level I
(SASS) train. Gravimetric analyses were also performed on Level I samples
5-15
-------
to quantify high molecular weight organics. Bag samples were collected
over a 30 to 45 minute period, one sample per test, and utilized to deter-
mine C-j to Cg organics. The SASS train contacts approximately 30 cubic
meters of flue gas which were drawn isokinetically during the test.
Analysis of SASS train samples provides quantitative measurement of to
organics by GC and organics higher than C-jg by gravimetric methods.
Analytical results for scrubber outlet SASS train XAD-2 resin samples
were not available due to sample handling problems. However, data from
coal-fired utility boilers were utilized to obtain an average ratio of
resin adsorbed organics to all other organics collected by the SASS train.
These ratios were used in conjunction with data from the scrubber outlet
probe rinse, resin module rinse, particulate organics and other organic
catches from the SASS train to estimate the resin catch. Calculated out-
let organic data from the C7 to C-jg and higher than organic fractions
are considered to be accurate to within a factor of three to four.
Data from FID, GC and gravimetric analysis indicate 28 to 60% of the
scrubber inlet organics are heavier than C-jg with the balance being com-
posed primarily of C-j to Cg hydrocarbons. Scrubber outlet hydrocarbons
appear to consist almost entirely of C^ to Cg hydrocarbons due to removal
of organics higher than C7 (21% removal of C^ to C-jg fraction and 85%
removal of the >C-|g fraction). It is interesting to note that while
removal of organics higher than C^ is apparently 2.02 ng/J, no decrease in
total hydrocarbons by FID analysis was observed. This may indicate that
the FID was analyzing only the C-| to Cg fraction and that heavier frac-
tions were removed by the gas conditioner. Under the assumption that FID
data reflect the C-j to Cg fraction only, total hydrocarbon emissions would
be approximately 8.4 ng/J (0.02 lb/MM Btu) or approximately 45% higher
than indicated by FID. Scrubber inlet hydrocarbons would then consist of
11% organics higher than Clg with the balance being the C-j to Cg fraction.
Again, scrubber outlet hydrocarbons would consist primarily of the C-j to
Cg fraction.
Total Particulates
Average emission rates of total particulates prior to scrubbing were
2951 ng/J (6.86 lb/MM Btu). A mass balance of the coal ash indicates that
5-16
-------
approximately 75%* of the total ash was present in the flue gas at the
scrubber inlet. These data indicated that the multiclone unit located
upstream of the scrubber was removing little or no particulate material.
This observation was subsequently verified by site operators who noted
that the multiclone unit was subject to mechanical failure during the test
period. Hence, particulate loadings measured at the scrubber inlet appear
to be representative of uncontrolled emissions. Mean particulate loadings
measured at the scrubber inlet are approximately 8% lower than the value
of 3212 ng/J (7.47 Ib/MM Btu) presented in AP-42. Particulate emissions
after scrubbing were 18.6 ng/J (0.04 Ib/MM Btu) which corresponds to 99.4%
particulate removal efficiency in the scrubber. It is evident from these
data that the scrubber is an effective particulate control device which
does not appear to require auxiliary mechanical collectors. Controlled
particulate emissions are well below the existing NSPS limitation of
43 ng/J (0.10 lb/MM Btu) although they are slightly higher than the pro-
posed limitation of 13 ng/J (0.03 lb/MM Btu).
Particulate Size Distribution
The particulate size distributions for the scrubber inlet and outlet
were determined using two different methods. For the high particulate
loading at the scrubber inlet, the fractions of particulates in each size
range were determined optically using polarized light microscopy (PLM)
and are expressed in terms of number percent (i.e., the number of particles
in each size range). For the lower particulate loadings at the scrubber
outlet, the fractions of particulates in each size range were determined
using an Anderson cascade impactor. The cascade impactor determines weight
percent in each size range. Different sizing methods were used because
the particulate concentration at the scrubber inlet was too high and caused
problems in the interpretation of the cascade impactor data.
An estimate of the scrubber inlet particulate size distribution by
number is presented in Table 5-13. However, PLM size data are based on
particle diameter and a number distribution, and cannot be directly
* As shown in Table 5-21, approximately 10-20$ of the particulates at the
scrubber inlet is oil soot which is a system residual from previous oil
firing.
5-17
-------
TABLE 5-13. APPROXIMATE SCRUBBER INLET PARTICULATE SIZE
DISTRIBUTION FOR COAL FIRING (PLM) -
TEST 201-1
Particle Diameter
Approximate Numerical
Size Range, Microns
%
<1
6
1 - 3
26
3 - 10
40
>10
28
compared with the cascade impactor size data, which are based on aerody-
namic diameter and a weight distribution. The PLM particulate size data
have, therefore, been converted to aerodynamic diameter and weight distri-
bution basis, by assuming that particulates of different size have the
same density (refer to Table 5-14). This is a reasonable assumption
because the major components of the particulates generated from coal com-
bustion, the aluminosilicates and the iron oxides, are known to partition
equally among small and large particulates. With the constant density
assumption, the weight distribution in each size range would be proportional
to the product of the number distribution and the particulate volume
representing the size range. The particulate volume was calculated based
on the geometric mean diameter for the size range.
In Table 5-14, the scrubber inlet particulate size distribution by
weight, as estimated from the PLM size distribution by number, is presented
along with the scrubber outlet particulate size distribution determined
using the Anderson cascade impactor. The data presented show a significant
change in the particulate size distribution before and after scrubbing, as
the larger particulates are effectively removed in the venturi scrubber.
The emission rates of the particulates in each size fraction before and
after scrubbing are presented in Table 5-15. These again show that the
removal efficiency for particulates in the 3 - 10 ym range is 97.9%, and
the removal efficiency for particulates larger than 10 ym is more than
99.9%. It is also interesting to note that for particulates less than 3 ym
5-18
-------
TABLE 5-14. SCRUBBER INLET AND OUTLET PARTICULATE
SIZE DISTRIBUTION BY WEIGHT FOR COAL
FIRING - TEST 201-1
Aerodynamic Diameter
Size, Range, Microns
Scrubber
Inlet
Weight %
Scrubber
Outlet
< 1
0.0017
62
1-3
0.041
30
3-10
2.24
7
>10
97.7
1
TABLE 5-15. EMISSION RATES OF PARTICULATES FOR A
COAL-FIRED BOILER IN TEMPORAL UNITS -
TEST 201-1
Aerodynamic Diameter
Size Range,
Microns
Scrubber
Inlet
kg/hr
Scrubber
Outlet
Removal
Efficiency
<1
0.0055
1.30
<0
1-3
0.13
0.63
<0
3-10
7.3
0.15
97.9
>10
316.5
0.021
>99.9
Total
324.0
2.10
99.3
5-19
-------
in size, there is a net increase in emission rates across the scrubber.
This net increase indicates that the venturi scrubber is probably not
effective in removing the fine particulates present in the flue gas, and
that fine particulates may be generated within the scrubber. Based on the
analysis of SO3 and SO^ " emission data, it has been estimated that up to
40% of the fine particulate emissions at the scrubber outlet could be
contributed by scrubber generated NaHSO^. The remaining portion of the
net increase in fine particulates across the scrubber can probably be
attributed to the uncertainties associated with the assumptions used in
converting PLM number size distribution data to weight size distribution,
and to calcium sulfite hemihydrate (CaS03-1/2 H20) particulates generated
by the scrubber.
Sulfur Compounds: SOp, SO3, SO^
The Goksoyr-Ross controlled condensation system was used to determine
SO3 emissions for tests 201-1 and 201-4. As shown in Table 5-16, less
than 1% of the fuel sulfur is emitted as SO3. The overall sulfur balance
indicates that over 94% of the sulfur input in the fuel is emitted as SO2,
SO3 or sulfate (SO^-) in the flue gas. The remainder of the sulfur is
associated with the bottom ash or may be unaccounted for. The scrubber,
in addition to removing 95 to 97% of the S02, also removes 32 to 33% of the So
and 88% of the SO^ . The relatively poor removal efficiency for SO3 is an
indication that SO3 is either present as very fine aerosols in the scrubber
inlet, or is converted to very fine aerosols as the flue gas stream is
rapidly cooled inside the scrubber. The higher removal efficiency for
S04~ is an indication that most of the S04~ in the scrubber inlet is
present as metallic sulfates in particulates larger than 3 nm in size.
The 3 to 5% of the SO2 and 67 to 68% of the SO3 that are not removed pass
through the scrubber. For SO^", however, a more complex process occurs.
At the scrubber inlet 24% of the sulfates is water soluble and 76% is acid
soluble (15% HNO3). At the outlet the values are 97% and 3% respectively
(These data are summarized in Table 5-17). This is an indication that the
combustion generated sulfates are not simply passing through the scrubber.
5-20
-------
TABLE 5-16.
SO21 SOj
AND S04"
EMISSIONS FROM
COAL FIRING
Pollutant
Concen-
Mass
Thermal
Produc-
% of
Removal *
tration
emission
emission
tion
fuel
efficiency
,3
rate
rate
emission
sulfur
%
mg/m
g/hr
ng/J
rate
found 1n
mg/kg steam
flue gas
so2
201-1 Inlet
2,970
134,000
1,280
3,030
92
97
201-1 Outlet
90.0
3,950
37.8
89.3
2.7
201-4 Inlet
2,420
102,000
937
2,490
94
95
201-4 Outlet
130
5,180
47.8
127
4.8
201-1
Inlet
16.9
764
7.3
17.3
0.42
201-1
Outlet
11.7
513
4.9
11.6
0.28
201-4
Inlet
12.6
527
4.9
13.0
0.39
0
CM
Outlet
9.0
359
3.3
8.8
0.26
S04
201-1 Inlet 154.1 6,950 66.3 157 3.2
201-1 Outlet 19.1 838 8.00 19.0 0.39
Total
201-1 Inlet - - - - 95 +
201-1 Outlet - 3.4+
201-4 Inlet - - - - 94 +
201-4 Outlet - 5.1f
*
Calculated from mass emission rate.
+
As all three sulfur species.
^This removal is actually a net removal since the scrubber both removes and
generates sulfates.
5-21
-------
TABLE 5-17. SUMMARY OF SULFATE EMISSIONS DURING
COAL FIRING - TEST 201-1
mg/m3
Inlet Outlet
Water soluble 36.7 (24*) 18.6 (97%)
Acid soluble 117.4 (76*) 0.5 (3 36)
Total 154.1 19.1
Because of the possibility that the S04~ emissions from coal combustion
may be modified by the scrubbing process, an analysis effort to determine
the nature of SO^ emissions was initiated. Both the Fourier Transform IR
(FTIR) analysis and the X-Ray Diffraction (XRD) analysis have confirmed
the presence of sodium bisulfate (NahSO^) in the scrubber outlet, but not
In the scrubber inlet. This is positive proof that sulfates, as the result
of oxidation of sodium bisulfite (NaHSO^) and sodium sulfite (Na^SO^), are
generated within the scrubber and emitted in the scrubber effluent gas.
Also, tests on boilers with flue gas concentrations of 400 to 8,000 ppm SO
have shown that there is no correlation between initial S02 concentration
and the net sulfate formation rate (Ref. 9). This Implies that the scrubber
has a minimum sulfate emission rate that is virtually unaffected by inlet
S02 concentration.
Based on the above findings, it is believed that NaHSO, emissions
3 ^
from the scrubber are on the order of 5 mg/m (the difference between total
outlet sulfate and emissions was determined as S03). Furthermore, if
one assumes that only a small fraction of the SO^ was collected on the
filter as f^SO^ because of the high filter temperature (175°C), then the
scrubber contribution could be as high as 19 mg/mor 0.8 kg/hr for test
201-1. As discussed in the previous section, this amount of scrubber
generated NaHSO^ could account for 40% of the fine particulate (3 urn)
emissions at the scrubber outlet.
5-22
-------
Inorganics
The emission concentrations for 22 major trace elements at the scrubber
inlet and outlet are presented in Table 5-18. To assess the hazard poten-
tial of these emissions, the emission concentrations are compared with the
Minimum Acute Toxicity Effluent (MATE) values. The MATE values are
emission level goals developed under the direction of EPA, and can be
considered as concentrations of pollutants in undiluted emission streams
that will not adversely affect those persons or ecological systems exposed
for short periods of time (Ref. 10). MATE values for air derived from human
health considerations are used as the basis for comparison here.
As shown in Table 5-18, of the trace elements presented, 18 exceed
their MATE values at the scrubber inlet and 4 at the scrubber outlet. The
four trace elements in the scrubber effluent that pose a potential hazard
are arsenic, chromium, iron and nickel. The MATE value for arsenic is
extremely low because arsenic 1s a cumulative poison producing long-term
chronic effects 1n humans, and the MATE values for chromium and nickel are
extremely low due to considerations for potential human carcinogenicity.
The established Threshold Limit Values (TLV's) for arsenic, chromium and
3
nickel are 0.5, 0.5 and 0.1 mg/m , respectively. If the TLV's are used as
the basis for comparison, the emission concentrations for arsenic, chromium
and nickel are all less than their respective TLV's and would be considered
less hazardous. Additionally, 1t may be noted that the emission concen-
tration of beryllium at the scrubber outlet 1s equal to Its MATE value.
At this emission concentration, the total beryllium emissions from boilers
greater than 50 MW In capacity would amount to more than 10 grams per day
and exceed the National Emission Standard for Hazardous Air Pollutants.
In Table 5-19, the emission factors and the mass emission rates for
the 22 major trace elements at the scrubber Inlet and outlet are presented.
The mass emission rates were used to calculate the removal efficiency for
these trace elements by the scrubber. The overall removal efficiency for
these trace elements 1s approximately 99.5%. As indicated In Table 5-19,
however, some of the trace elements are not being removed as effectively
as others.
5-23
-------
TABLE 5-18. EMISSION CONCENTRATIONS OF TRACE ELEMENTS
DURING COAL-FIRING TEST 201-1
Trace
Element
Scrubber
Inlet
mg/m3
Scrubber
Outlet
mg/m3
MATE
Value
mg/m3
Degree of Hazardf
Scrubber Scrubber
Inlet Outlet
*
Be
0.1
0.002
0.002
50
1.0
Hg*
0.011
0.005
0.05
0.22
0.10
Ca
74
0.036
16
4.6
0.002
Mg
19
0.011
6.0
3.2
0.002
Sb
3.7
0.025
0.050
74
0.5
As
7.8
0.22
0.002
3900
110
*
B
0.2
0.03
3.1
0.07
0.01
Cd
0.47
0.0010
0.010
47
0.1
Cr
2.6
0.13
0.001
2600
130
Co
3.6
0.012
0.050
72
0.24
Cu
9.6
0.020
0.20
48
0.10
Fe
450
2.4
1 .0
450
2.4
Pb
8.5
0.021
0.15
57
0.14
Mn
0.78
0.015
5.0
0.16
0.003
Mo
10
0.027
5.0
2.0
0.005
Ni
1.4
0.063
0.015
93
4.2
V
3.1
0.058
0.50
6.2
0.12
Zn
2.3
0.048
4.0
0.58
0.012
Se
3.2
0.099
0.200
16
0.50
Sr
11
0.058
3.1
3.5
0.019
A1
480
2.6
5.2
92
0.5
Zr
1.6
0.018
5.0
0.32
0.004
Total
1100
6.2
Approximate values as determined by Spark Source Mass Spectrometry
(SSMS). The other values presented are determined by Inductively
Coupled Plasma Optical Emission Spectroscopy (ICPOES) analysis.
+ Degree of hazard is defined as the ratio of the discharge concentration
to the MATE valve.
f Mercury was determined by cold vapor analysis of SASS train samples
taken during test 200.
5-24
-------
TABLE 5-19. EMISSION FACTORS AND MASS EMISSION RATES OF
TRACE ELEMENTS DURING COAL-FIRING TEST 201-1
Trace
Emission
Factor,nq/J
Emission
Rate, q/hr
Removal
Enrich-
Element
Scrubber
Scrubber
Scrubber
Scrubber
Efficiency
ment
Inlet
Outlet
Inlet
Outlet
%
Factor
*
Be
0.04
0.001
5
0.09
98
3.7
Hgt
0.08
0.037
0.50
0.23
55
84
Ca
32
0.015
3300
1.6
99
0.09
Mg
8.2
0.0046
860
0.48
99
0.11
Sb
1.6
0.010
170
1.1
99
1.2
As
3.4
0.092
350
9.7
97
5.3
B
0.1
0.01
10
1.2
88
2.1
Cd
0.20
0.00042
21
0.044
99
0.4
Cr
1.1
0.054
120
5.7
95
9.5
Co
1.6
0.0050
160
0.53
99
0.6
Cu
4.1
0.0084
430
0.88
99
0.4
Fe
190
1.0
20,000
110
99
0.99
Pb
3.7
0.0088
380
0.92
99
0.5
Mn
0.34
0.0063
35
0.68
98
3.4
Mo
4.3
0.026
450
1.2
99
0.5
N1
0.60
0.026
61
2.8
95
8.6
V
1.3
0.024
140
2.5
98
3.6
Zn
0.99
0.020
100
2.1
98
3.9
Se
1.4
0.041
140
4.3
97
5.8
Sr
4.7
0.024
500
2.5
99
0.9
A1
210
1.1
22,000
110
99
1.0
Zr
0.69
0.0075
72
0.79
99
2.1
Total
470
2.6
50,000
270
99
*
Approximate values as determined by SSMS. The other values were
determined by ICPOES analysis.
+ Mercury was determined by cold vapor analysis of SASS train samples
taken during test 200.
5-25
-------
To better understand the removal efficiency of the individual trace
elements, the enrichment factor for each trace element across the scrubber
has been computed. The enrichment factor is defined here as the ratio of
the concentrations of trace element to aluminum in the scrubber outlet,
divided by the corresponding ratio in the scrubber inlet. Aluminum is
selected as the reference material because it has been known to partition
*
equally among particulates of different size . The enrichment factors
presented in Table 5-19 show that beryllium, mercury, antimony, arsenic,
boron, chromium, manganese, nickel, vanadium, zinc, selenium and zirconium
are enriched across the scrubber. The enrichment observed 1s due primarily
to the partitioning of trace elements as a function of particulate size,
and the greater collection efficiency of the scrubber for the large size
particulates. It may also be noted that many of the trace elements that
show an enrichment trend, such as mercury, selenium and arsenic, either
occur as element vapors or form volatile oxides and halides at furnance
temperatures. Condensation and surface adsorption of the more volatile
elements or their oxides and halides downstream of the furnace could,
therefore, result in higher concentrations of these trace elements on
smaller particulates.
To gain better insight into the nature of the particulates generated
from coal combustion, the Electron Spectroscopy for Chemical Analysis (ESCA)
technique was used to determine the surface and subsurface concentrations
of elements in the particulates. The main use of ESCA, however, was In
the study of the sulfur depth profile of sulfur containing compounds 1n
the particulates. The ESCA results, expressed as normalized atom percent,
are presented in Table 5-20. For particulates at the scrubber Inlet,
comparison of the 201-1 cyclone with the 201-1 filter results shows similar
sulfur surface content. The sharp drop in S atom percent after etching of
the scrubber Inlet filter sample, on the other hand, Indicates that some
surface coating of the particulates occurred.
Assuming that the bulk of the particulate is homogeneous with respect
to aluminosilicates, the thickness of the surface coating can be most
Silicon, Iron, and scandium have also been used by other Investigators
as the reference element In the computation of enrichment factors.
Notice that Iron has no enrichment 1n this study while silicon and
scandium were not measured.
5-26
-------
TABLE 5-20. DEPTH PROFILE ANALYSIS OF COAL PARTICULATE WITH CONCENTRATIONS
EXPRESSED AS NORMALIZED ATOM PERCENT* - TEST 201-1
Na
Si
A1 Fe CI
Ca
Inlet Level II cyclone catch
Level II filter catch
Level II filter catch; 76 A
57 4 11 5 7 2 2 1
56 2.4 11.5 14.1 8 1.4 1.5 1 .2 2
54.2 2.1 4.0 17.9 12.6 2.1 1.3 2.2 1.0 2.7
Level
II
filter catch
45.7
5.5
13.2
7.1
2.2
1.2
1.1
0.7
1.3
14.9
7.1
Level
II
O
filter catch; 75 A
48
7.4
11.5
9.0
5.2
1.2
1.1
1.1
1.1
12.1
1 .8
Level
II
filter catch; 150 A
48.3
9.1
10.0
10.1
7.7
1.2
1.2
1.2
10.0
1.2
Level
II
filter catch; 300 A
48.3
8.8
8.0
10.6
9.6
2.1
1.0
1.2
0.5
8.0
1.8
Level
II
filter catch; 500 A
47.9
7.3
6.7
10.8
13
1.7
1.2
1.1
1 .2
8.1
1.1
Level
II
filter catch; 700 A
47
7.3
6.1
11.6
11.0
2.3
0.9
1.5
0.5
1.0
8.8
1.5
The atom percent of the 12 elements presented here adds up to 100 percent. Other elements present in
the cyclone and filter catches were not studied in ESCA. Hence, the atom percents in this table are
normalized atom percents and not absolute atom percents.
-------
readily estimated by considering the concentration of coating elements
relative to the concentration of aluminum or silicon. The relative concen-
tration may be plotted as a function of penetration to enable graphical
estimation of coating thickness. Aluminum is selected as the reference
element here rather than silicon because the ESCA analyses were run directly
on the filter and the silicon content of the filter would interfere with
the interpretation of data.
Depth profiles for six major elements in the scrubber outlet partic-
ulates are depicted graphically in Figure 5-2. These profiles show that
sulfur and carbon are more concentrated near the surface than deeper inside
the particles. Both the sulfur and the carbon curves, however, level off
o
after ^275 A. Hence, the thickness of the deposited layer of sulfur and
O
carbon appears to be approximately 275 A. For sulfur, these data indicate
that while there may be higher surface concentration of sulfur containing
compounds in the particulates emitted from the scrubber, some of these
compounds are probably deposited on particles composed of solid sulfate or
sulfite. This would be the case if sulfuric acid condensed on the sodium
bisulfate (NaHSO^) or calcium sulfite hemihydrate (CaS0^*l/2 H^O) particles
that have been found to be present. For carbon, the depth profile data
indicate that a fraction of the carbon, either oil soot*, or as a carbon
containing compound (such as carbonate or bicarbonate}, could be deposited
on the surface of the fine particulates emitted. The leveling off of the
carbon curve, however, indicates that another fraction could be emitted as
solid oil soot particles or solid carbonate/bicarbonate particles. The
depth profile for iron, vanadium, chlorine, and calcium is reasonably flat
and indicates that the relative concentrations of these elements remain
approximately constant.
The approximate composition of the particulates at the scrubber inlet
and outlet has also been investigated using PLM analysis. The estimated
weight percentages of the major components of the particulates are
* Oil soot deposited in the duct-work downstream of the boiler during
oil-firing.
5-28
-------
<
*
X
&
uu
O
(O m CO CM T-
NOUVH1N30NOO WnNIKimV Afl Q3QIAia N0I1VUJ.N30N0D 1N3W313
Figure 5-2. Depth profile analysis of outlet coal particulate - Test 201-1.
5-29
-------
presented in Table 5-21. It may be noted that a substantial fraction of
the particulates at both the scrubber inlet and outlet is composed of oil
soot. The increase in the weight percent of oil soot in the particulates
across the scrubber is consistent with the results of ESCA analysis, as
discussed in the previous section. Both the PLM analysis and the ESCA
analysis have shown that oil soot could be emitted as fine, solid partic-
ulates. The calcium sulfite hemihydrate (CaS03«l/2 H20) and the unknown
sulfate identified at the scrubber outlet are mostly generated by the
scrubbert. As discussed previously, the unknown sulfates may be composed
primarily of sulfuric acid and sodium bisulfate.
TABLE 5-21. MAJOR COAL PARTICULATE COMPONENTS - TEST 201-1
Approximate weight %
Component Inlet Outlet
Ash*
Fused 15-30
Unfused 50-65
Minerals
Fe203 1-5 15-25
Feo04 (magnetite) 10-15
Si O2 <2
CaC03 - 1-5
Oil soot+ 10-20 25-40
Coke <2
CaS03.1/2 H20 and - 50-65
Unknown sulfate
*
Iron-aluminum silicates.
+ See text for discussion.
t Strictly speaking, CaS03*l/2 H2O is formed in the regeneration of
Na2S03 and a fraction of it is carried over with the regenerated
Na2S03 into the scrubber.
5-30
-------
Chloride, Fluoride, and Nitrate Emissions
Emissions data for chloride (CI"), fluoride (F~), and nitrate emis-
sions (NO3") are presented in Table 5-22. Chlorides and fluorides are
removed with high degrees of efficiency, at greater.than 99% and greater
than 85%, respectively. These are to be expected because the overall
removal efficiency of the trace element cations with which these anions
may be associated is greater than 99%. The removal efficienty for nitrates
cannot be determined because the nitrate emissions at both the scrubber
t
inlet and outlet are expressed as "less than" values.
Organics
Four methods of analyses were utilized in determining flue gas organic
loadings. Continuous FID analyses were performed to determine total
organic concentrations assuming all carbon to be present as methane. A
field chromatograph was employed to analyze concentrations of organics 1n
the range of C-| to Cg hydrocarbons and a laboratory chromatograph was used
in determining hydrocarbons 1n the C7 - C16 boiling range. Gas chromato-
graphy was applied to bag samples of gas collected over a 30-45 minute
period. Higher molecular weight organics were determined gravlmetrlcally
from residues of solvent rinses and extracts.
For Cj - C^g organics, the subscripted carbon number refers to a
boiling range rather than a specific molecular structure. For example,
the Cg designation refers to compounds boiling in the range of approximate-
ly 110 to 140°C while normal octane bolls at 125.7°C. The approximate
boiling ranges corresponding to each carbon are presented 1n Table 5-23.
TABLE 5-22. CHLORIDE, FLUORIDE, AND NITRATE EMISSIONS FROM COAL
FIRING - TEST 201-1
Inlet* Outlet*
Removal
mg/m ng/J mg/m ng/J Efficiency
10.9 4.7 <0.009 <0.004 >99
0.51 0.22 <0.076 <0.03 >86
<1.1 <0.48 <0.58 <0.25
Mass as the ion.
Chloride (CI")
Fluoride (F")
Nitrate (N03")
5-31
-------
TABLE 5-23. APPROXIMATE BOILING RANGES CORRESPONDING
TO EACH CARBON NUMBER
Carbon
Approximate
Carbon
Approximate
Number
Boiling Range
Number
Boiling Range
Cl
-160 to
-100°C
C9
140 to 160°C
C2
-100 to
- 50°C
C10
160 to 180°C
c3
- 50 to
0°C
Cll
180 to 200°C
c4
0 to
30°C
C12
200 to 220OC
C5
30 to
60°C
c13
220 to 240OC
c6
60 to
90°C
c14
240 to 260OC
C7
90 to
110°C
C15
260 to 280°C
C8
110 to
140°C
C16
280 to 300°C
Results of organic determinations made during test 200 are summarized
in Table 5-24. Total organics determined as methane by FID show a slight
increase across the scrubber. However, as mentioned previously, this
increase can not be explained with available data. Total hydrocarbon
analyses appear to correspond well with GC and gravimetric hydrocarbon
analyses. These data indicate that, although uncontrolled total organic
emissions are low, organics which are produced consist primarily of light
hydrocarbons in the C^ to Cg range and hydrocarbons heavier than C^. Also
concentrations of hydrocarbons higher than C^g appear to decrease by more
than 80% across the scrubber. It should be noted that analytical results
for the outlet resin samples from the SASS train were not available, so
these were calculated using data from ten coal fired boiler tests which
used similar sampling trains, and assuming the proportion of organic
material in the resin sample to the rest of the organics trapped in the
sampling train to be constant. These calculated values are good to a
factor of 3 to 4.
A quantitative analysis of the to C^g organic fraction from test
200 indicates that organic material was present for all boiling ranges
with the exception of the and fractions which were present at con-
centrations lower than 0.4 yg/m3 (limit of detection). Measured concentra-
3
tions at the scrubber inlet ranged from 21 to 215 ug/m . The highest
5-32
-------
TABLE 5-24. COMPARISON OF ORGANIC MEASUREMENT METHODS
DURING COAL FIRING - TEST 200
Organic Concentration
Method
mg/m3
Inlet
Outlet
% Change
Total as CH^ (FID)
9
11
Not
Significant
Cj - Cg (GC)
512.7*
512.7*
-
C7 - C16 (GC)
0.81
0.652
-20
C-jg (Gravimetric)
5.4
0.788
-85
~These values represent the detection limit of the instrument used.
concentrations were found in the Cg, C12> and C1g fractions which had
215, 130, 128 and 121 ^g/m3, respectively.
Organic materials collected by the XAD-2 resin during sampling of the
test 200 scrubber inlet gas were separated by liquid chromatography (LC)
and the fractions were subsequently analyzed by infrared spectroscopy (IR)
per standard Level 1 procedures. Only resin extracts and resin module
rinse samples were submitted for LC/IR analyses since the organic contents
of other samples were too low for such treatment. Results of sample
analyses (not corrected for resin blank species contents) and blank
analyses are presented in Table 5-25. Many of the compound categories
found in the sample are also present in the blank, although in most
instances blank contributions appear to be minor. The presence of sili-
cones and esters in the sample is most probably characteristics of the
sampling system (i.e., lubricants and resin interference) rather than the
source.
An attempt was made to identify the specific compounds that made up
the total organic emissions. Table 5-26 summarizes the organics that were
identified. None of the identified compounds is directly associated with
combustion. They are, however, representative of the types of compounds
that are used in the manufacture of the sample bags, XAD-2 resin and the
5-33
-------
TABLE 5-25. SUMMARY OF THE INFRARED ANALYSIS OF ORGANICS PRODUCED
DURING COAL COMBUSTION - TEST 200 INLET (NOT CORRECTED FOR BLANK)
Total Organics,
mg/m^
LCI*
Sample/Blank
0.14 0.02
LC2
Sample/Blank
LC3
Sample/Blank
0.06 0.03
LC4
Sample/Blank
0.12 0.01
LC5
Sample/Blank
0.37 0.02
LC6
Sample/Blank
0.95 0.10
LC7
SamDle/Blank
0.06
Total
Sample/Blank
1.70 0.18
Category Intensity*
Aliphatic Hydrocarbons
Aromatic Hydrocarbons
Chlorinated Hydrocarbons
Silicones
Heterocyclic Sulfur
Compounds
XM XM
XM XM
XM XM
XM XM
Km
Xn
XM
Xm
Xm
Xm
Xm Xm
Xm
Xm
Om
Xm
Xm
Thiocarbonyl Compounds
Nitro Compounds
Ethers
Esthers
Amides
Xm
Xm
XM
xw
XM
Xm Om
XM XM
XM XM
Xm
Om
XM XM
Om Om
Alcohols
Glycols
Phenols
Carboxylic Acids
Silicates
Xm
Xm
Xm
Xm
XM
XM XM
Xm
XM
OM
XM XM
t
*
LC fractionation was performed per standard level 1 analytical procedures.
* D = At least one species suspected present
X = At least one species present
M = Material type accounts for 50% of the sample
m = Material type accounts for 10X of the sample
-------
TABLE 5-26. ORGANIC COMPOUNDS IDENTIFIED IN THE
FLUE GAS DURING COAL FIRING*
vg/rc3
Compound Name
201-2
In
Out
201-3
In Out
Propionaldehyde
170
380
54
Ethyl-n-butyl ether
2000
1500
910 1500
Hydrocarbon
-
-
380 '
Chloropropanol
-
-
3.9
Unidentified Alcohol
-
4.0
-
Ketone (MW 140)
3.9
-
6.9
Methyl substituted aromatic
1.7
-
-
* As discussed 1n the text, these compounds are not considered to be
directly asoociated with combustion.
solvents used in the analysis. This finding is consistent with both GC
and FID analyses indicating very low organic emissions during testing.
Polycyclic Organic Material
Polycyclic organic material (POM) was not found in either the scrubber
3
inlet or outlet samples at detection limits of 0.3 yg/m . This observa-
tion is consistent with the findings to date from the EPA sponsored project
"Emissions Assessment of Conventional Combustion Sources". However, two
POM compounds for which MATE values are below 0.3 yg/m are benzo(a)pyrene
and dibenz(a,h)anthracene. The MATE values for benzo(a)pyrene and
3 ?
dibenz(a»h)anthracene are 0.02 yg/m and 0.09 yg/m , respectively. While
available data indicate that many POM compounds are not present at signifi-
cant concentrations during coal firing, additional analyses at higher
GC/MS sensitivity would be required to conclusively preclude the presence
of all POMs at significant concentrations.
5-35
-------
Scrubber Efficiency
Flue gas analyses indicate that scrubber processing removes a signif-
icant percentage of input sulfur oxides (SOg. SOg and particulate S0^~),
total particulates and organics of the C7 class and higher. Scrubber
removal efficiency data for these flue gas components are presented in
Table 5-27. As discussed in the criteria pollutants section, the signifi-
cance of data indicating N0X and CO removal appears to be somewhat question-
able. Therefore, these components are not included in this discussion.
Average removal efficiencies have been discussed previously. However,
it is Important to note that it 1s the Cy hydrocarbons and higher hydro-
carbons which are removed with 77% efficiency. These fractions comprise
32 to 69% of the total generated organics. Hence, based on the total
generated organics, a removal efficiency of 25 to 53% was obtained.
TABLE 5-27. SCRUBBER EFFICIENCY DURING COAL FIRING
Test Number
% Removal
so2
so3
~
so4=
Total
Particulates
C- and higher^"
Organics
200
97
--
99
77
201-1
97
33
88
99
--
201-2
97
—
—
—
—
201-3
97
—
—
—
--
201-4
95
32
—
99
--
Average
97
32
00
00
-+
99
77+
*
This removal rate is actually a net change rate. As described earlier
the scrubber both removes and generates sulfates.
* One data point only.
It is not known for certain by what process organics are removed with
such a high degree of efficiency. There are, however, at least three
.possible mechanisms:
5-36
-------
• Dissolution - Some organics are partially water soluble.
These compounds could be removed by dissolving in the
slurry.
• Condensation - High boiling organics could condense and
be removed as "particulate".
• Sorption - Some organics could adsorb on particulates.
One or any combination of these mechanisms may account for the high removal
efficiencies.
LIQUID WASTE
As discussed 1n Section 3, only one significant waste water stream 1s
produced. The stream is a combination of water treatment waste, boiler
blowdown, and acid waste water from elsewhere 1n the plant. The quality of
this combined stream is such that 1t is acceptable for disposal Into the
municipal sewer system. Liquid streams from the scrubber operation are
passed to the thickener and recirculated to the scrubber after the filtration
step. There is no direct wastewater discharge from the scrubber operation,
as the process 1s designed to dispose of all of the water that enters its
system through evaporation and moisture entrained 1n the scrubber cake.
Because several streams are mixed together, it 1s not possible to
accurately determine what part of the effluent 1s attributable to the boiler.
However, the flow rate of the combined stream 1s approximately 10,000
I1ters/hr (40 gallons/m1n).
Water Quality Parameters
Table 5-28 summarizes the waste water parameters for the combined
waste water stream. Note that these values do not represent water pro-
duced solely by the boiler but also include process waste.
Inorganics - Combined Waste Water Stream
Table 5-29 shows discharges of the major inorganics in the combined
stream as determined by SSMS analyses (note that these are accurate only
to within a factor of about 3). None of these elements exceeds its MATE
value based on human health considerations. However, based on the
uncertainty in SSMS analysis, cobolt, cadmium, nickel and copper may
exceed their MATE values based on ecological considerations.
5-37
-------
TABLE 5-28. COMBINED STREAM WASTE WATER PARAMETERS - COAL FIRING
Parameter
Test
200
Test
201-1
Test
201-2
Test
201-3
Test
201-4
Average
PH
7.9
7.5
8.2
8.0
7.3
7.8
Hardness, ma/1
210
158
135
145
100
150
(as CaC03)
Alkalinity, mg/1
115
125
130
125
145
128
(as CaC03)
Cyanide, mg/1
0
0
0
0
0
0
TABLE 5-29. WASTE WATER INORGANICS FOR COAL FIRING
Trace
mg/1
g/hr*
MATE Value
, mg/1
Degree of Hazard
Element
Health
Ecology
Health
Ecology
Be
<0.001
<0.01
0.030
0.055
<0.033
<0.018
F
0.8
8
38
—
0.021
—
V
0.003
0.03
2.5
0.15
0.0012
0.02
Cr
0.002
0.02
0.25
0.25
0.008
0.008
Co
0.1
1
0.75
0.25
0.13
0.40
HI
0.005
0.05
0.23
0.010
0.022
0.50
Cu
0.02
0.2
5.0
0.050
0.004
0.40
Sr
0.5
5
46
—
0.011
—
Cd
<0.001
<0.01
0.050
0.001
<0.02
<1.0
Sb
<0.001
<0.01
7.5
0.20
<0.0001
<0.005
Pb
0.01
0.1
0.250
0.050
0.04
0.20
~
Flow rate of 10,000 liters per hour, combined waste water.
5-38
-------
Organlcs - Combined Waste Mater Stream
Table 5-30 summarizes the concentrations of C7 - C^g organlcs 1n the
combined stream. High molecular weight organlcs (>C^6) were present at a
concentration of 0.21 mg/Hter, but are probably attributable to process
wastes generated at the manufacturing site. As a basis for comparison,
the water MATE values for alkanes, alkenes, and alkynes are 1n the 500 to
14,000 mg/Hter range based on human health considerations, and 1n the 1.0
to 100 mg/Hter range based on ecological considerations. The discharge
concentrations of organics are all well within these MATE values.
TABLE 5-30. SUMMARY OF C7 - Clfi ORGANICS IN THE WASTE WATER -
COAL FIRING
Carbon Number
mg/1
Carbon Number
mg/1
C7
ND
C13
ND
C8
ND
C14
0.1
C9
ND
C15
0.1
C10
0.1
C16
ND
C11
ND
C12
0.1
Total Cy — C-|6
0.4
ND means none detected.
SOLID WASTE
Three solid waste streams are produced by the system:
• Bottom ash;
• Fly ash;
• Scrubber cake.
Table 5-31 shows the approximate quantities of bottom ash and scrubber
cake that were produced. Only small quantities of fly ash were produced
during the test period due to the malfunctioning of the multlclone.
5-39
-------
TABLE 5-31. SOLID WASTE PRODUCTION RATES - COAL FIRING
Bottom Ash Scrubber Cake+
es kg/hr ug/J kg/hr ug/O
200 ^80 ^0.75 1100 10.2
201-1 ^80 ^0.76 1100 10.5
201-2 -v80 ^0.73 1200 11.0
201-3 ^80 ^0.86 850 9.1
201-4 ^80 ^0.74 840 7.8
•k
15 tons a week
+ Scaled up to represent scrubbing of 100% of the flue gas for boiler
number 4.
The scrubber cake produced after filtration has the appearance of a
clayey silt. Its grain size is quite uniform and characteristic of sllty
soils, but Its behavior closely resembles a clay in many respects. As
obtained from the vacuum filter, the scrubber cake consists of small lumps
and appears to be relatively dry; in actuality, however, the water content
generally ranges from about 30 to 50%.
If it is assumed that calcium sulfite hemihydrate (CaSO.j.1/2 HgO) 1s
formed from SOg scrubbing and ^SOg regeneration, then the mass balance
in Table 5-32 shows that the scrubber cake is composed of 28.5% coal fly
ash and 23.8% CaSO^.1/2 HgO. However, if the multlclone had been func-
tioning properly during the test period, more fly ash would be removed
upstream of the scrubber and the fly ash content of the scrubber cake
would be lowered proportionally. The amount of scrubber cake produced
could be reduced to 600-750 kg/hr on wet basis, assuming approximately
60 to 80% multiclone efficiency.
5-40
-------
TABLE 5-32. SCRUBBER CAKE MASS BALANCE FOR COAL FIRING - TEST 201-1
Component
Contribution to Scrubber Cake
kg/hr Weight %
Fly Ash Removed by Scrubber
CaS03*l/2 H2O Formed from SO2
Scrubbing and Na2S03
Regeneration
CaSO^t CaC0o» NaoSO^, Ca(0H)2»
NaHSOj ana Na2§03 Losses
(estimated)
Water
Total
324
262
10-85
429-504
1100
29.5
23.8
0.9-7.7
39-46
100
Although the scrubber cake material 1s composed predominately of
relatively Insoluble solids (calcium sulfite, calcium sulfate, and some
calcium carbonate), the intersltitlal water does contain soluble residues
of lime, sulfate, sulfite and chloride salts. Trace elements in the fly
ash may also leach from the disposed scrubber cake and are of special con-
cern. The concentrations of 20 trace elements 1n the scrubber cake are
presented In Table 5-33. Here 1t may be noted that with the exception of
boron, the trace element concentrations 1n the scrubber cake have far
exceeded the MATE values for solids. This 1s the result of reducing a
large volume of low concentration pollutants to a smaller volume of con-
centrated pollutants. The degree of hazard for the trace elements 1n
the scrubber cake is sufficiently high to warrent the disposal of these
solid wastes in specially designed landfills.
5-41
-------
TABLE 5-33. INORGANIC CONTENT OF SCRUBBER CAKE FROM
COAL-FIRING (DRY BASIS) - TEST 201-1
Element
Concentration
pg/g
MATE Value
, wg/g
Degree of Hazard
Health
Ecology
Health
Ecology
Ca
60,715
480
32
126
1,897
Mg
1 ,458
180
174
8.1
8.4
Sb
315
15
0.4
21
788
As
532
0.5
0.1
1,064
5,320
B
88
93
50
0.9
1.8
Cd
13
0.1
0.002
130
6,500
Cr
141
0.5
0.5
282
282
Co
424
1.5
0.5
283
848
Cu
112
10
0.1
11
1,120
Fe
47,241
3.0
0.5
15,738
94,482
Pb
297
0.5
0.1
594
2,970
Mn
51
0.5
0.2
102
255
Mo
1 ,117
150
14
7.4
80
Ni
114
0.45
0.02
253
5,700
V
195
5.0
0.3
39
650
Zn
282
50
0.2
5.6
1,410
Se
256
0.10
0.05
2,560
5,120
Sr
642
92
--
7.0
—
A1
45,310
160
2.0
283
22,655
Zr
106
15
--
7.1
—
Total
159,409
5-42
-------
The concentrations of 20 trace elements present 1n fly ash are pre-
sented in Table 5-34. Again, in almost every case, the trace element
concentration in the fly ash has far exceeded its MATE value for solids.
Trace element concentrations in the bottom ash would be similar to those
of the fly ash, except that the more volatile elements and the elements
that form volatile compounds would be more enriched in the fly ash. Thus,
the concentrations of arsenic, antimony, boron, chromium, manganese,
nickel, vanadium, zinc, selenium and zirconium would all be lower in the
bottom ash.
An overall mass balance for the 20 trace elements has been performed
and the results are summarized in Table 5-35. The percent of trace element
1n the coal feed that could be located in the effluent streams (scrubber
cake, scrubber effluent gas, bottom ash and fly ash) 1s used as a measure
of mass balance closure. With the exception of boron, copper, strontium
and zirconium, the closure of mass balance for the trace elements has been
found to be good. This Is an Indication of the reliability and accuracy
of sampling and analysis data for trace element and flow rate measurements.
The scrubber cake was also analyzed for organlcs but none were de-
tected. This is to be expected since the concentration of organlcs In
the flue gas stream was extremely low.
ANNUAL EMISSIONS
Table 5-36 presents an estimate of the annual emissions of the major
pollutants for the controlled and uncontrolled case. It was assumed that
the boiler operates at 100% load, 87% of the year (7560 hours/years), and
that coal is the only fuel burned.
5-43
-------
TABLE 5-34. INORGANIC CONTENT OF FLY ASH FROM
COAL-FIRING - TEST 201-1
Element Concentration MATE Value, ug/g Degree of Hazard
ug/g HealthEcology Health Ecology
Ca
378
480
32
0.8
12
Mg
2,478
180
174
14
14
Sb
438
15
0.4
29
1,095
As
1,015
0.5
0.1
2,030
10,150
B
20
93
50
0.2
0.
Cd
18
0.1
0.002
180
9,000
Cr
434
0.5
0.5
868
868
Co
408
1.5
0.5
272
816
Cu
320
10
0.1
32
3,200
Fe
129,330
3.0
0.5
43,110
258,660
Pb
438
0.5
0.1
876
4,380
Mn
121
0.5
0.2
242
605
Mo
1,288
150
14
9
92
N1
165
0.45
0.02
367
8,250
V
376
5.0
0.3
75
1,253
Zn
179
5.0
0.2
36
895
Se
378
0.10
0.05
3,780
7,560
Sr
728
92
—
8
—
A1
109,450
160
2.0
684
54,725
Zr
187
15
—
12
--
Total
248,149
5-44
-------
TABLE 5-35. MASS BALANCE ON TRACE ELEMENTS - TEST 201-1
Element
Coal Feed
g/hr
Scrubber
Cake
g/hr
Scrubber
Effluent Gas
g/hr
Bottom and
Fly Ash*
g/hr
Percent.
Recovery
Ca
2,794
40,072
1.6
30
*
Mg
1,270
962
0.5
198
91
Sb
308
208
1.1
35
79
As
497
351
9.7
81
89
B
8.7
58
1.2
1.6
700
Cd
12.7
8.6
0.04
1.4
79
Cr
174
93
5.7
35
77
Co
461
280
0.53
33
68
Cu
261
74
0.88
26
39
Fe
44,455
31,179
110
10,346
94
Pb
308
196
0.92
35
75
Mn
44
34
0.68
9.7
100
Mo
1,063
737
1.2
103
79
N1
134
75
2.8
13
68
V
171
151
2.5
30
107
Zn
203
186
2.1
14
100
Se
265
169
4.3
30
77
Sr
247
424
2.5
58
196
A1
50,806
29,905
110
8,756
76
Zr
980
70
0.79
15
9
For mass balance calculations, bottom ash has been assumed to have the
same trace element concentrations as fly ash. This 1s an approximate
assumption, as some trace elements are enriched 1n the fly ash.
+ Percent recovery is defined as 100 times the ratio of the sum of the
emissions for a trace element to the trace element in the coal feed.
+ Percent recovery for calcium 1s not calculated because most of the
calcium in the scrubber cake is from the Hme slurry.
5-45
-------
TABLE 5-36. ANNUAL EMISSIONS - COAL FIRING*
kq/year
0/
to
Pollutant
Scrubber Inlet
Scrubber Outlet
Difference
Gaseous N0X (as NO2)
500,810
442,520
-12
S02
1,127,300
36,800
-97
S03
6,184
4,157
-33
SO4
67,214
8,110
-88
CO
16,119
14,497
-10
Organics (as CH4)
5,870
6,377
+ 9
Ci - Cg Organics
<5,606+
<5,606+
—
C7 - Cifi Organics
345
274
-21
High Molecular Weight Organics
2,311
335
-86
Total Particulates
2,991,700
18,856
-99
0 - ly
11,691
— '
1 - 3 y
5,657
3 - lOy
1,320
—
>10y
188
M^/year
Liquid Blowdown/wastewater
^76,000
<76,000
0
Cooling water
^86>000
^86,000
0
kg/year
Solid Bottom ash
~778,600
^ 778,600
0
Fly ash (mechanical separator)*
~U80Q,000
*1>8QQ ,000
0
Scrubber cake t
0
8,054,100
—
* Assuming 100% load, 45 weeks/year (7,560 hours/year),
f These values represent the detection limit of the instrument used.
+ Assuming that the mechanical separator is operating properly.
-------
AIR QUALITY ASSESSMENT - COAL FIRING
Simplified air quality models were used to determine the relative
ground level air quality resulting from uncontrolled and controlled emis-
sions. The ambient air quality values are approximate only. The emphasis
should be placed on the relative values for each case as opposed to their
absolute values.
Worst case weather conditions and typical weather conditions were
considered. The worst case was assumed to be plume trapping. An equation
proposed by Blerly and Hewson [11] was used with the following assumptions:
inversion height 100 meters, wind speed 1.0 meter/second, D class stability
(neutrally stable) in the mixing layer, and effective stack height of 50 m
(164 ft). The typical case was assumed to correspond to the standard
Gaussian convective diffusion equation, [12]. The following conditions were
used: wind speed 4.0 meters/second and D class stability. These conditions
could reasonably be expected to occur almost anywhere in the country.
Typical does not mean average. It was assumed that all species were inert.
No photochemical reactions were considered. (See Appendix A for details.)
Figures 5-3 through 5-10 present plots of approximate ground level
ambient air quality as a function of distance directly downwind from a
single 10 MW equivalent source. Data for N0X, CO, SOg and particulates
are presented. The purpose of these figures is not to attempt to accurately
predict air quality but to compare the effects of controlled and uncon-
trolled emissions under an arbitrary but realistic set of meteorological
conditions. It 1s Implicit 1n this approach that each set of meteorologi-
cal conditions remains constant for a sufficient length of time for the
ambient air quality to reach steady state conditions at each distance.
Table 5-38 presents a summary of the federal ambient air quality
standards for each pollutant. The standards are also shown on each plot.
Keeping in mind the caveats mentioned above* several observations can
be made:
• The N0x standard 1s exceeded under both weather conditions.
Since the scrubber does not remove significant amounts of
N0x, there 1s no substantial difference between the controlled
and uncontrolled cases. (The boiler has no N0X controls.)
5-47
-------
700
i*>
I 500
5
i
z
400
TEST 201-1
»-
z
INLET
—
1 300
OUTLET
PRIMARY AND 8ECONDARY
STANDARD: ANNUAL
ARITHMETIC MEAN
100
DISTANCE FROM STACK, km
Figure 5-3. Relative N0X air quality under worst case
weather conditions - coal firing.
5-48
-------
INLET
140
\
-• 120
§
«C
K
S
8 TOO-
PRIMARY AND SECONDARY
STANDARD: ANNUAL
ARITHMETIC MEAN
TEST 202-1
H
2
OUTLET
DISTANCE FROM STACK, km
Figure 5-4. Relative N0X air quality under typical weather
conditions - coal firing.
5-49
-------
THE MAXIMUM EXPECTED AMBIENT CO
CONCENTRATIONS ARE WELL BELOW THE
MOST RESTRICTIVE STANDARD OF 10 mg/m3
(8-HOUR AVERAGE)
INLET
TEST 201-1
OUTLETX.^
DISTANCE FROM STACK, km
Figure 5-5. Relative CO air quality under worst case weather
conditions - coal firing.
5-50
-------
7
6- -
THE MAXIMUM EXPECTED AMBIENT CO
CONCENTRATIONS ARE WELL BELOW THE
MOST RESTRICTIVE STANDARD OF 10 mg/m3
(8-HOUR AVERAGE)
3 S - -
4- -
orT
\\
i\
i\
TEST 201-1
3- -
2 - -
A
OUTLET^
• •
DISTANCE FROM STACK, I
10
12
14
Figure 5-6, Relative CO air quality under typical weather conditions
coal firing.
5-51
-------
1,<00
INLET
SECONDARY STANDARD:
MAXIMUM 3-HOUR AVERAGE
1.000
TEST 202-1
800
PRIMARY STANDARD:
MAXIMUM 24-HOUR
AVERAGE
400
200
PRIMARY STANDARD:
ANNUAL ARITHMETIC
MEAN
LET
DISTANCE FROM STACK, km
Figure 5-7. Relative SO2 air quality under worst case weather
conditions - coal firing.
5-52
-------
800
r*
M
400 + I \
I I
I \
rr t
INLET
PRIMARY STANDARD:
MAXIMUM 24-HOUR AVERAGE
| 300-.
5
I
TEST 202-1
ui
200- -
\
100- -
\
\
\
\
PRIMARY STANDARD:
ANNUAL ARITHMETIC MEAN
OUTLET
+
t
14
DISTANCE PROM STACK, km
Figure 5-8, Relative SOg air quality under typical weather
conditions - coal firing.
5-53
-------
8,000
4,000- -
I
I
3,000-
r\
' \ INLET
' \
\
\
\
I
THE MAXIMUM AMBIENT PARTICULATE
CONCENTRATION RESULTING FROM
OUTLET EMISSIONS WOULD BE ~27Alg /M3,
WELL BELOW ALL AIR QUALITY STANDARDS.
<
K
TEST 201-1
3
H
Z
141
2,000- -
\
1,000- -
\
\
\
+
PRIMARY STANDARD:
24-HOUR MAXIMUM
OUTLET
T
2
1 1
6 8
DISTANCE FROM STACK, km
10
12
Figure 5-9, Relative particulate air quality under worst case weather
conditions - coal firing.
5-54
-------
1,000- .
900
•00 - -
700- -
000- -
800- -
.. I \
300- -
200- -
100- -
11
I ¦ INLET
1
I *
TEST 202-1
\
-+
\
J
¦f
1 OUTLET
\
PRIMARY STANDARD:
ANNUAL GEOMETRIC MEAN
PRIMARY STANDARD:
MAXIMUM 24-HOUR AVERAGE
SECONDARY STANDARD:
MAXIMUM 24-HOUR AVERAGE
SECONDARY STANDARD:
' ANNUAL GEOMETRIC MEAN
1
2
10
12
14
DISTANCE FROM STACK, km
Figure 5-10. Relative particulate air quality under typical
weather conditions-coal firing.
5-55
-------
t CO standards are not exceeded under any conditions. The
most restrictive standard is 10 mg/m3 (10,000 >ig/m3) and
the maximum predicted level is only about 0.2% of this value.
As with N0X there is no substantial difference between the
cases based on scrubber inlet and outlet data.
t All primary SO2 standards are exceeded under both weather
conditions for uncontrolled emissions. For controlled
emissions no standard is exceeded under either weather
condition.
• All particuate standards are exceeded under both weather
conditions for uncontrolled emissions. For controlled
emissions no standard is exceeded.
TABLE 5-37. NATIONAL AMBIENT AIR QUALITY STANDARDS
FOR CRITERIA POLLUTANTS
Pollutant standard
Pollutant Primary* Secondaryt
Nitrogen dioxide
Carbon monoxide
Sulfur dioxide
Total suspended
Particulate
100>ng/mJ(0.05 ppm) Same as primary
annual arithmetic mean.
10 mg/m3 (9 ppm)
maximum 8-hour average;
40 mg/m3 (35 ppm)
maximum 1-hour average.
Same as primary
80/jg/m3(0.03 ppm) 1300/ug/m3 (0.05 ppm)
annual arithmetic mean: maximum 3-hour average
365 jug/m3 (0.14 ppm)
maximum 24-hr average.
75 jug/m3 annual geo-
metric mean: 260 jjg/m
maximum 24-hr average.
60 jug/m3 annual geo-
metric mean: 150jug/m3
maximum 24-hr average.
*
Primary, necessary to protect the public health.
Secondary, necessary to protect the public welfare.
5-56
-------
CONCLUSIONS - COAL FIRING IN A 10 MW INDUSTRIAL BOILER WITH FGD
1) Uncontrolled emissions of criteria pollutants generally
corresponded well with values reported in AP-42. Although
NO emissions were slightly higher than the average AP-42
value, they appear to be within the normal range for similar
industrial units.
2) N0X reductions varying from approximately 0 to 24 percent
were measured across the scrubber. However, the magnitude
of N0X reductions could not be correlated to changes 1n
variables monitored during the test period (I.e., temperature,
gas flow rate, boiler load, etc.). For this reason, it
is considered feasible that observed NO reductions are
a sampling phenomenon, perhaps related to leaks in the
sample train.
3) Sulfur dioxide removal data indicated an average scrubber
efficiency of 97 percent. Controlled SO2 emissions were
36.3 ng/J (0.08 pounds/MM Btu) which is less than either
existing or proposed NSPS limitations for utility boilers.
4) Mass balance data indicate that the multiclone unit up-
stream of the scrubber was removing little or no fly ash
during the test period. Under the resulting near full
fly ash loading, the scrubber removed 99.4 percent of
the inlet particulates.
5) Although the removal efficiency for total particulates 1s
high, there appears to be a net Increase 1n emission rates
across the scrubber for particulates less than 3 ym 1n size.
This net increase can be attributed to the poor removal
efficiency of the scrubber for fine particulates, and
to the sodium bisulfate (NaHSO^) and calcium sulfite
hemihydrate (CaSOg-1/2H20) particulates generated by the
scrubber. Both NaHSO^ and CaS03*l/2H20 have been identi-
fied at the scrubber outlet but not at the inlet.
5-57
-------
The relatively poor removal efficiency (approximately 30%)
for S03 across the scrubber is an indication that SOj 1s
either present as very fine aerosols in the scrubber inlet,
or 1s converted to very fine aerosols as the flue gas
stream is rapidly cooled inside the scrubber.
The overall removal efficiency for trace elements across
the scrubber is 99.5 percent. Of the 22 major trace ele-
ments, 18 exceed their MATE values at the scrubber inlet
and four at the scrubber outlet. The four trace elements
in the scrubber effluent that pose a potential hazard are
arsenic, chromium, iron, and nickel. In addition, the
emission concentration of beryllium at the scrubber outlet
is equal to its MATE value. The National Emission Standard
for Hazardous Air Pollutants limitation of 10 grams beryl!ii
per day would not be exceeded except by boilers of 50 MW
capacity or greater at this concentration.
The relative removal efficiency for trace elements across
the scrubber can be explained by enrichment theory. In
general, trace elements that occur as element vapors or
form volatile compounds at furnace temperatures are more
concentrated in the smaller particulates, as a result of
subsequent condensation and surface adsorption. These are
the same trace elements that are removed less efficiently
by the scrubber.
ESCA analysis has shown that while there may be higher
surface concentration of sulfur containing compounds in
the particulates emitted from the scrubber, most of the
sulfur containing compounds are probably present as solid
sulfates and sulfites. Thus, it is conceivable that
sulfuric acid vapor Is condensed and deposited on the
particulates emitted, whereas sodium bisulfate and
calcium sulfite hemihydrate are emitted as fine, solid
particulates.
5-58
-------
The overall sulfur balance indicates that over 92 percent
of the fuel sulfur is emitted as SC^, less than 1 percent
of the fuel sulfur is emitted as SOg, and approximately
3 percent of the fuel sulfur is emitted as S0^~.
Total organic emissions were generally less than 9 ng/J
(0.02 pound/MM Btu) and these emissions appear to be
primarily C-| to Cg hydrocarbons and hydrocarbons heavier
than C-jg. While uncontrolled emission rates for to
C^g and higher hydrocarbons are low, emissions of these
organics were further reduced by 21% and 85%, respectively,
in the scrubber unit.
Polycyclic organic material (POM) was not found in the
3
scrubber inlet or outlet at detection limits of 0.3 pg/m .
MATE valves for most POM's are greater than this detection
limit. However, since the MATE values for at least two
POM compounds - benzo(a)pyrene and dibenz(a,h)anthracene -
3
are less than 0.3 yg/m , additional GC/MS analyses at
higher sensitivity would be required to conclusively pre-
clude the presence of all POM's at MATE levels.
The combined wastewater stream generated from the boiler
operation apparently does not pose an environmental hazard,
since the discharge concentrations of most inorganics and
organics are all well below their MATE values. However,
based on the uncertainty in SSMS analyses, cobalt, cadmium,
nickel and copper may exceed their MATE values based on
ecological considerations.
The scrubber cake produced contains a significant amount of
coal fly ash. With the exception of boron, trace element
concentrations in the scrubber cake exceeded their health
MATE values. All ecology MATE values were exceeded. Because
the trace elements may leach from the disposed scrubber cake,
5-59
-------
these solid wastes must be disposed of in specially
designed landfills.
Mass balance closure for most of the trace elements have
been found to be in the 75 to 107 percent range. This
closure instills confidence on the validity of the sampling
and analysis data for trace elements.
5-60
-------
REFERENCES FOR SECTION 5
1. Magee, E.M., H.J. Hall and G.M. Varga, Jr. Potential Pollutants In
Fossil Fuels. Report prepared by ESSO Research and Engineering Co.
for EPA under Contract No. 68-02-0029. June 1973.
2. Ruch, R.R., H.J. Gluskoter and N.F. Shimp. Occurrence and Distribution
of Potentially Volatile Trace Elements In Coal: A Final Report.
Illinois State Geological Survey Environmental Geology Notes, Number
72. August 1974.
3. Hamersma, J.W. and M.L. Kraft. Applicability of the Meyers Process
for Chemical Desulfurization of Caol: Survey of Thirty-Five Coals.
Report prepared by TRW Systems Group for EPA under Contract No.
68-02-0647. September 1975.
4. Koutsoukos, E.P., M.L. Kraft, R.A. Orsini, R.A. Meyers, M.J. Santy and
L.J. Van Nice. Meyers Process Development for Chemical Desulfurization
of Coal, Vol. I. Report prepared by TRW Systems Group for EPA under
Contract No. 68-02-1336. May 1976.
5. Cato, G.A., L. J. Muzio and D.E. Shore. Field Testing: Application
of Combustion Modifications to Control Pollutant Emissions From
Industrial Boilers—Phase II. Report prepared by KVB for EPA under
Contract No. 68-02-1704. April 1976.
6. Steam-Electric Plant Air and Water Quality Control Data. Federal Power
Commission. March 1975.
7. Kircher, J.F., A.A. Putnam, D.A. Ball, H.H. Krause, R.W. Coutant, J.O.L.
Wendt and A. Levy. A Survey of Sulfate, Nitrate and Acid Aerosol Emis-
sions and Their Control. Report prepared by Battelle - Columbus
Laboratories for EPA under Contract No. 68-02-1323. April 1977.
8. Proceedings: Symposium On Flue Gas Desulfurization - Hollywood, F1.,
November 1977.
<). Capabilities Statement. Sulfur Dioxide Control Systems. Technical
Report 100. FMC Corporation, Environmental Equipment Division, Hasca,
111., November 1977.
10. Cleland, J.G. and G.L. Kingsbury. Multimedia Environmental Goals for
Environmental Assessment, Vols. I and II, EPA-600/7-77-136a, November
1977.
11. Bierly, E.W. and E.W. Hewson. Some Restrictive Meteorological Condi-
tions to be Considered in the Design of Stacks. J. Appl. Metero.,
1, 3 pages 383-390, 1962.
12. Turner, D.B. Workbook of Atmospheric Dispersion Estimates. U.S. Dept
of HEW, 1969.
5-61
-------
5-62
-------
SECTION 6
COMPREHENSIVE ASSESSMENT OF OIL FIRING
CASE FOR AN INDUSTRIAL BOILER
This section provides a comprehensive multimedia assessment of
emissions/effluents associated with an oil-fired industrial boiler equipped
with an FGD system. Data from Level I/Level II sampling and analyses are
utilized to quantitatively determine emissions in gas, solid and liquid
waste streams and to evaluate performance of pollution control equipment
in use during oil firing. Waste stream pollutant concentrations are
compared with Minimum Acute Toxicity Effluent (MATE) values, where
appropriate, to provide an indication of risk to public health and ecology.
Simplified air quality models are used to determine the relative ground
level air quality resulting from both uncontrolled and controlled emissions.
TEST CONDITIONS
Five oil-fired tests were performed on the industrial boiler at the
Firestone facility. Unit loadings ranged from 31,800 to 45,400 kg steam
per hour (70,000 to 100,000 pounds per hour) which corresponds to between
70 and 100% of full load operation for this boiler. Specific test con-
ditions are presented in Table 6-1. Tabulated fuel feed rates are nominal,
although oil analysis and steam production data Indicate that they are
accurate to within 3% 1f a 90% thermal efficiency is assumed. Oxygen
concentrations presented in Table 6-1 were measured 1n flue gas samples
drawn from the inlet of the scrubber unit. Due to air leaks into upstream
ducting operating at sub-atmospheric pressure and the possibility of air
leaks in the sampling train, tabulated oxygen concentrations are not
generally representative of concentrations at the furnace outlet. During
normal operation, oxygen concentrations in the furnace after combustion
range from 3 to 4% which corresponds to an excess air input of 17 to 25%
for oil firing. Estimated excess air levels presented 1n Table 6-1 were
computed assuming an average oxygen concentration of 3.5% in the furnace
and utilizing fuel analyses data.
Test data relating to flue gas flow rates and scrubber loading are
summarized in Table 6-2. Flue gas flow rates through the scrubber were
6-1
-------
TABLE 6-1. SUMMARY OF TEST CONDITIONS - OIL FIRING
Test
Steam Production Rate
% of
Nominal
% O2 at
Estimated
3» Excess
A1r to
Furnace"1"
No.
kg
Steam/hr
lb
Steam/hr
Maximum
Load
011 Feed
Rate,
Gal/hr
Scrubber
Inlet
202-1
45,400
100,000
100
900
5.8
21
202-2
45,400
100,000
100
900
6.3
21
202-3
44,200
97,500
97.5
880
6.1
21
202-4
42,200
93,000
93.0
805
4.0
21
203
31,800
70,000
70.0
600
Not
Measured
21
Due to air leaks 1n ducting upstream of the scrubber inlet, tabulated 03
values are not representative of combustion zone O2 concentrations. Combus
tlon zone O2 concentrations normally range from 3 to 4? for this unit.
+ 02 - CO/2
% excess air 1s estimated to be 100 x 0 ^4 _ (q2 . co/2)' where °2 Was
assumed to be 3.5% and other species concentrations were computed from fuel
analyses.
Test
No.
TABLE 6-2. FRACTION OF FLUE GAS PROCESSED BY SCRUBBER
DURING OIL COMBUSTION
Flow Rate
at Scrubber
Inlet, *
dscm/m1n
% of
Design
Load
Total Flue
Gas Flow
Rate, *
dscm/m1n
Fraction
Processed
by the
Scrubber
202-1
193
109.0
824
0.24
202-2
192
108.5
850
0.23
202-3
189
107.8
818
0.23
202-4
155
87.6
657
0.24
203
171
96.6
537
0.32
Average
180
101.9
737
0.25
Dry standard cubic meters per minute (dscm/mln),
6-2
-------
155 to 193 dry standard cubic meters per minute (dscm/m1n) which corres-
ponds to between 88 and 109% of full design loading. As discussed In
Section 5, the scrubber is a pilot unit which does not process the entire
flue gas output of the furnace. From 23 to 32% of the total flue gas was
processed through the scrubber during oil-firing tests.
Ultimate analyses of fuel oil samples for the five oil fired tests
are presented in Table 6-3. The fuel composition was essentially constant
during testing. Additional analyses were performed on fuel samples from
test 202-4 to determine concentrations of 22 trace elements (Ca, Mg, Sb,
As, B, Cd, Cr, Co, Cu, Pb, Mn, Mo, Ni, V, Zn, Se, Sr, Fe, Al, Be, Hg and Zr).
These data are presented in Table 6-4. The method employed for analysis
of most of these elements was Inductively Coupled Plasma Optical Emission
Spectroscopy (ICPOES) which is generally considered to be more accurate
than Spark Source Mass Spectrometry (SSMS).
TABLE 6-3. SUMMARY OF ULTIMATE OIL ANALYSES
Weight %
202-1
202-2
202-3
202-4
203
Average
*
a
Carbon
85.96
85.83
86.54
86.77
86.29
86.28
0.39
Hydrogen
10.95
10.88
10.95
10.92
10.89
10.92
0.03
Nitrogen
0.31
0.34
0.42
0.30
0.42
0.36
0.06
Sulfur
2.08
1.95
1.97
1.97
1.85
1.96
0.08
Ash
0.02
0.02
0.02
0.01
0.02
0.02
0.004
Oxygen
0.68
0.98
0.10
0.03
0.53
0.46
0.40
kJ/kg
-
-
-
-
-
40,741+
-
*o ¦ One standard deviation.
+The heat content of the oil burned 1s nearly constant at this value;
Individual values were not available.
6-3
-------
TABLE 6-4. CONCENTRATION OF MAJOR TRACE ELEMENTS IN OIL
(TEST 202-4)
Element
ppm 1n
Fuel 011
Typical
Range*
Reference
Ca
5.50
No Data
Mg
<0.4
No Data
Sb
0.03+
0.002-0.8
1
As
2.0*
0.0006-1.1
1
B
<0.15
No Data
Cd
<3.5
No Data
Cr
2.2
0.002-0.02
1
Co
<1.25
No Data
Cu
1.40
No Data
Fe
12.3
0.003-14
!»2
Pb
2.6*
No Data
Mn
0.4f
0.001-6
1
Mo
2.9+
<0.1-1.5
1
N1
16.0
14-68
3
V
36.5
15-590
3
Zn
3.0
No Data
Se
0.7+
0.03-1
1
Sr
0.23
No Data
A1
3.5
No Data
Zr
<0.05
No Data
Be
<0.05
No Data
Hg
0.09++
0.02-30
1
*Except for V and N1, these ranges are for U.S. and foreign crude oils.
Ranges of V and Ni concentrations are for fuel oils.
+Values were calculated from concentrations at the scrubber Inlet when
ICPOES analysis provided upper limit data only.
+Arsen1c concentration calculated from concentration at the scrubber
inlet (see text).
Performed by SSMS on a feed oil sample from test 203.
++Performed by cold vapor analysis on a feed oil sample from test 203.
6-4
-------
However, oil beryllium was analyzed by SSMS on a sample from test 203, and
mercury was analyzed using cold vapor analysis on the same sample. Several
of the oil trace elements were below the ICPOES detection limit. Approximate
values for these elements were calculated from the concentrations found at
the scrubber inlet, assuming that essentially all of the oil trace elements
reach the scrubber. As no bottom ash was generated during oil firing, this
assumption should be valid. The value for arsenic presented in the table,
2 ppm, was also calculated because the value obtained from ICPOES analysis,
45 ppm, appeared to be unreasonable compared to both the SSMS value (0.1 ppm)
and the typical ranges found 1n the literature.
Considering the uniformity of fuel oil ultimate analyses obtained
during the test period, 1t appears reasonable to assume that tabulated
trace element analyses are typical of the oil-fired during the five day test
period. Listed for comparison are typical concentration ranges for nlckle and
vanadium 1n fuel oils, and other trace element concentrations 1n U.S. and foreign
crude oils. The crude oil values should be used for rough comparison.
Analyses of most trace elements for which typical values are available
appear to be consistent with the crude oil values. Arsenic and molybdenum
values are on the high side, and chromium 1s quite a bit higher than the
typical crude oil values. However, the significance of these higher concen-
trations 1s not apparent due to the limited quantity of published data and
the complete absence of source specific data.
STACK EMISSIONS
As discussed previously, the scrubber unit processed 23 to 3235 of the
total flue gas generated during oil firing. However, under the assumptions
specified 1n Section 5, mass emission rate data presented in the ensuing
sections are reflective of full stream loadings at the scrubber inlet and
outlet.
6-5
-------
Criteria Pollutants
Current NSPS define allowable emission rates of NO^ (as NOg), SOg and
total particulates from utility boilers having 25 MW or greater output.
Existing NSPS for limitation of N0X emissions from oil-fired units 1s
129 ng/J (0.30 Ib/MM Btu) and proposed standards would Impose the further
requirement for 65% reduction of uncontrolled emissions. The current NSPS
limitation on SO2 emissions from oil-fired boilers 1s 344 ng/J (0.80 lb/MM
Btu). Potential standards could impose further requirements for up to 90%
reduction of uncontrolled S02 emissions. The existing NSPS limitation on
particulate emissions Is 43 ng/J (0.10 lb/MM Btu) and proposed standards
would reduce this limit to 13 ng/J (0.03 lb/MM Btu). Current federal NSPS
do not address either CO or total hydrocarbon emissions.
Because similar standards for Industrial boilers have not been pro-
mulgated to date, criteria pollutant emissions data presented In this
section will be discussed 1n the context of existing and proposed federal
NSPS for utility boilers.
As mentioned previously, 5 o1l-f1red tests were performed on the
Firestone Industrial boiler. Criteria pollutant concentrations were generally
measured at frequent Intervals during each test and averaged to obtain the
mean concentration for the test. The 5-test averages of criteria pollutant
emissions data are presented in Tables 6-5 through 6-8. Average emissions
data from the individual tests are presented 1n Tables 6-9 and 6-10.
Additionally, the 5-test averages of scrubber Inlet data are presented in
Table 6-11 for comparison with the EPA AP-42 emission factors for uncon-
trolled sources. The data are discussed by specific compound 1n the ensuing
sub-sections.
6-6
-------
TABLE 6-5. CRITERIA POLLUTANT EMISSIONS FOR AN OIL-FIRED
INDUSTRIAL BOILER IN CONCENTRATION UNITS (5 TEST AVERAGE)
mq/Nm^ (Graln/SCF)
Pollutant Before Scrubber After Scrubber
N0X (as N02) 467 (0.20) 450 (0.20)
CO 15.3 (0.01) 14.9 (0.01)
S02 2763 (1.21) 74.3 (0.03)
Organlcs (as CH4) 6.81 (0.00) 7.47 (0.00)
Total Particulates 317 (0.14) 50.7 (0.02)
TABLE 6-6. CRITERIA POLLUTANT EMISSIONS FOR AN OIL-FIRED
INDUSTRIAL BOILER IN TEMPORAL UNITS (5 TEST AVERAGE)
kg/hr
(1b/hr^
Pollutant Before ScrubberAfter Scrubber
N0X (as N02) 22.2 (48.9) 21.4 (47.2)
CO 0.71 (1.58) 0.69 (1.52)
S02 130.7 (288) 3.55 (7.83)
Organlcs (as CH4) 0.34 (0.75) 0.36 (0.79)
Total Particulates 14.8 (32.5) 2.28 (5.03)
6-7
-------
TABLE 6-7. CRITERIA POLLUTANT EMISSIONS FOR AN OIL-FIRED
INDUSTRIAL BOILER IN THERMAL UNITS (5 TEST AVERAGE)
no/J (Ib/MM Btu)
Pollutant Before Scrubber After Scrubber
NOx (as N02) 168 (0.39) 161 (0.37)
CO 5.47 (0.01) 5.31 (0.01)
S02 993 (2.31) 26.8 (0.06)
Organlcs (as CH4) 2.49 (0.01) 2.74 (0.01)
Total Particulates 113 (0.26) 17.6 (0.04)
TABLE 6-8. CRITERIA POLLUTANT EMISSIONS FOR AN OIL-FIRED
INDUSTRIAL BOILER IN PRODUCTION RATE UNITS (5 TEST AVERAGE)
mg/Kg Steam (lb/1000 lb. steam)
Pollutant Before Scrubber After Scrubber
N0X (as N02)
499
(0.50)
482 (0.48)
CO
16.1
(0.02)
15.5 (0.01)
so2
2945
(2.95)
79.7 (0.08)
Organlcs (as CH4)
7.62 (0.01)
7.99 (0.01)
Total Particulates
331
(0.33)
51.6 (0.05)
6-8
-------
TABLE 6-9. SUMMARY OF CRITERIA POLLUTANT EMISSIONS - OIL FIRING
ng/J (Ib/MM Btu)
Test No.
NOx
CO
so2
HC+
(as CH4)
c - c
1 L6
Organics
C7 " C16
Organics
Organics
Higher
Than C]g
Total
Particula
202-1 Inlet
175
(0.41)
5.08
(0.01)
938
(2.18)
2.84
(0.01)
<4.63
(< 0.01)
—
—
166
(0.39)
202-1 Outlet
166
(0.39)
4.70
(0.01)
32.1
(0.07)
3.20
(0.01)
<4.63
(< 0.01)
—
—
20.2
(0.05)
202-2 Inlet
175
(0.41)
5.30
(0.01)
1075
(2.50)
4.61
(0.01)
<4.79
(< 0.01)
—
—
—
202-2 Outlet
165
(0.38)
5.03
(0.01)
29.2
(0.07)
5.01
(0.01)
<4.79
(< 0.01)
—
—
—
202-3 Inlet
181
(0.42)
6.22
(0.01)
1085
(2.52)
1.71
(0.00)
<4.73
(< 0.01)
—
—
—
202-3 Outlet
177
(0.41)
5.92
(0.01)
26.7
(0.06)
1.98
(0.01)
<4.73
(< 0.01)
—
—
—
202-4 Inlet
141
(0.33)
5.30
(0.01)
874
(2.03)
0.82
(0.00)
<4.14
(< 0.01)
—
—
59.0
(0.14)
202-4 Outlet
138
(0.32)
5.60
(0.01)
19.2
(0.04)
0.76
(0.00)
<4.14
(< 0.01)
—
—
15.0
(0.03)
203 Inlet*
--
—
—
--
0.17
(0.00)
2.61
(0.01)
—
203 Outlet*
—
—
--
0.02
0.43
m m,
(O.OO) (0.00)
Average 168 5.47 993 2.49 <4.57 0.17 2.61 113
Inlet (0.39) (0.01) (2.31) (0.01) (< 0.01) (0.00) (0.01) (0.26)
Average 161 5.31 26.8 2.74 <4.57 0.02 0.43 17 6
Outlet (0.37) (0.01) (0.06) (0.01) (< 0.01) (0.00) (0.00) (0.04)
*Emiss1on factors were computed assuming an Og concentration of 5.5%, the average concentration for test 202.
+Total hydrocarbons as determined by FID.
-------
TABLE 6-10. SUMMARY OF CRITERIA POLLUTANT EMISSIONS - OIL FIRING
kg/hr
Test No.
NOx
CO
so2
HC
(as
ch4)
C1 "C6
Organlcs
r _ r
7 16
Organlcs
Organlcs
Higher
Than Cjg
Total
Particulates
202-1 Inlet
202-1 Outlet
23.7
22.7
0.69
0.64
128
4.35
0.39
0.45
< 0.63
< 0.63
—
22.4-
2.74
202-2 Inlet
202-2 Outlet
24.0
22.4
0.71
0.66
146
3.98
0.61
0.66
< 0.64
< 0.64
—
—
202-3 Inlet
202-3 Outlet
24.1
23.6
0.83
0.78
143
3.53
0.25
0.25
< 0.62
< 0.62
-•
202-4 Inlet
202-4 Outlet
16.9
16.9
0.63
0.67
106
2.33
0.12
0.08
< 0.50
< 0.50
7.13
1.82
203 Inlet
203 Outlet
~
—
—
—
--
0.02
0.00
0.25
0.04
—
Average
Inlet
22.2
0.71
131
0.34
< 0.60
0.02
0.25
14.8
Average
Outlet
21.4
0.69
3.55
0.36
< 0.60
0.00
0.04
2.28
-------
TABLE 6-11. COMPARISON OF CRITERIA POLLUTANT EMISSIONS
WITH EMISSION FACTORS FOR UNCONTROLLED BOILERS
ng/J (lb/MM Btu)
Test Data Average Emission Factors*
Before Scrubber for Uncontrolled Sources
N0X (as N02 at full load)
175 (0.41)
217
(0.51)
CO
5.47 (0.01)
14.7
(0.03)
so2
993 (2.31)
904
(2.10)
Organlcs (as CH^)
2.49 (0.01)
2.94
(0.01)
Total Particulates
113 (0.26)
29.4
(0.07)
*Factors are computed from EPA AP-42 values using the national average
oil heating value of 146,285 Btu/gal (Reference 4).
Nitrogen Oxides
Average N0X emissions measured during o1l-f1r1ng were 168 ng/J (0.39
lb/MM Btu) prior to F6D processing. Full load N0X emissions were 175 ng/J
(0.41 lb/MM Btu) prior to FGD processing. This emission rate 1s approximately
19% lower than the tabulated AP-42 emission factor of 217 ng/J (0.51 lb/MM Btu)
for 10 MW oil-fired industrial boilers. However, available published data
indicate that the measured uncontrolled N0X emission rate 1s well within
the range typical of oil-fired Industrial boilers (Reference 5). Average
N0X emissions after FGD were 161 ng/J (0.37 lb/MM Btu) which 1s 23% higher
than the current NSPS limitation of 129 ng/J (0.30 lb/MM Btu). As Indicated
in Table 6-9, N0x emissions decreased upon decreasing the boiler load from
100 to 93%, as expected. However, comparison of full load N0X emission
rates with 97.5% loading data (175 ng/J at full load versus 181 ng/J at
97.5% loading) Indicates slight data scatter which appears to be on the
order of 10 ng/J.
6-11
-------
Data presented in Table 6-9 indicate N0X reduction across the scrubber
for all oil-fired tests; the magnitude of measured average reduction is
from 2 to 6%. However, the larger N0X reductions of 5 and 6% measured during
tests 202-1 and 202-2 are reflective, at least in part, of a small air leak
which was discovered in the sampling line to the scrubber outlet. This
leakage problem, associated with a faulty coupling, was rectified prior to
tests 202-3 and 202-4. Data from the latter tests indicate N0X reductions
of approximately 2%. Hence, N0X removal during this test period appears to
be approximately 2% as indicated by tests 202-3 and 202-4 rather than the
somewhat higher value indicated by tests 202-1 and 202-2.
Carbon Monoxide
Emission rates of CO prior to FGD were 5.47 ng/J (0.01 Ib/MM Btu).
This measured average CO emission rate is nearly 63% lower than the AP-42
value of 14.7 ng/J (0.03 Ib/MM Btu) for uncontrolled oil-fired industrial
boilers. Average CO emissions after FGD processing were 5.31 ng/J (0.01
lb/MM Btu). Although average CO emission data indicate a 3% reduction
across the scrubber, data from the tests which were not subject to known
problems with the sampling train (tests 202-3 and 202-4) indicate a 6%
Increase to a 5% reduction across the scrubber. Further, at these low CO
concentrations, analytical sensitivity 1s approximately 15% of the measured
value. Therefore, the slight reduction in CO measured across the scrubber
is considered to be of little significance.
Sulfur Dioxide
Average S02 emissions prior to scrubbing were 993 ng/J (2.31 lb/MM Btu)
Measured SOg emissions compare well with the AP-42 value of 904 ng/J (2.1o
lb/MM Btu) for uncontrolled oil-fired industrial boilers. Average SOg
emissions after scrubbing were 26.8 ng/J (0.06 lb/MM Btu) which corresponds
to 97% FGD efficiency. The controlled S02 emission rate of 26.8 ng/J is
substantially lower than either the existing NSPS limitation of 344 ng/J
(0.80 lb/MM Btu) or potential standards requiring up to 90% reduction.
6-12
-------
Hydrocarbons
Emissions of hydrocarbons measured as methane were, on the average,
2.49 ng/J (0.01 lb/MM Btu) prior to scrubbing. The measured uncontrolled
hydrocarbon emission appears to compare well with the AP-42 average value
of 2.94 ng/J (0.01 lb/MM Btu) for uncontrolled oil-fired Industrial boilers.
However, it should be noted that, during these, tests, flue gas samples
were processed through a gas conditioner prior to FID analysis. Therefore,
higher molecular weight organics may have been scrubbed or condensed from
the sample gas slipstream prior to analysis. As will be subsequently
discussed, total hydrocarbon emissions measured by FID analysis of the
scrubber inlet gas may be low by a factor of 2 to 3. Emission data obtained
after flue gas scrubbing generally indicate a slight Increase of hydrocarbons
across the scrubber unit; the magnitude of this increase 1s 0.25 ng/J or
approximately 10%, on the average. However, examination of all real time
scrubber Inlet and outlet data pairs (obtained not more than 30 minutes
apart) measured during the test period indicates that the difference between
scrubber inlet and outlet hydrocarbon analyses is statistically Insignificant.
Similar treatment of the individual tests Indicated that a statistically
significant bias between scrubber Inlet and outlet hydrocarbon analyses
existed for only one of the four tests performed, namely test 202-2. There-
? fore, these data appear to Indicate that hydrocarbon analyses measured by
FID at the scrubber Inlet and outlet do not differ significantly.
In addition to FID hydrocarbon analyses, gas chromatograph analyses
were performed on limited bag samples of the flue gas and sample catches
from the Level I sampling (SASS) train. Gravimetric analyses were also
performed on Level I samples to quantify high molecular weight organics.
Each bag sample was collected over an interval of 30 to 45 minutes, with a
single sample being collected per test. These samples were utilized to
measure to Cg hydrocarbons. The SASS train contacts approximately 30
cubic meters of flue gas which were drawn 1sok1net1cally during the test.
Analysis of SASS train samples provides quantitative measurement of organic
compounds higher than Cy.
6-13
-------
Analytical results for scrubber outlet SASS train XAD-2 resin samples
were not available due to sample handling problems. However, data from
oil-fired utility boilers were utilized to obtain an average ratio of resin
adsorbed organlcs to all other organics collected by the SASS train for
each fuel. These ratios were used in conjunction with data from the
scrubber outlet probe rinse, resin module rinse, particulate organics and
other organic catches from the SASS train to estimate the resin catch.
Calculated outlet organic data from the Cy to and higher than C-^
organic fractions are considered to be accurate to within a factor of two
to three. Average chromatograph and gravimetric hydrocarbon analyses
indicate that 35 to 94% of the scrubber inlet hydrocarbons are higher than
C16 with the balance being composed primarily of C-| to Cg organics. Organics
In the scrubber outlet consist of 9 to 16% hydrocarbons higher than C^g
and, again, the balance is primarily C-| to Cg organics. These data indicate
that 88% of the Cy to C-jg organics and 83% of the organics higher than
are removed by the scrubber. This corresponds to removal of 2.33 ng/o
organic material which is in conflict with FID total organic data indicating
no organic removal. These data may indicate that the FID was analyzing only
the C-j to Cg fraction and that heavier fractions were removed by the gas
conditioner prior to analysis. Under the assumption that FID data reflect
the C1 to Cg fraction only, the data would indicate that 49% of the scrubber
inlet hydrocarbons are higher than C^ with the balance being composed
primarily of C1 to Cg organics. The scrubber outlet organics would consist
primarily of C-| to Cg organics with 13% organics higher than C-jg. Further
the total hydrocarbon emissions, although still low, would increase by a
factor of 2 to 3.
Total Particulates
Average emission rates of total particulates at the scrubber inlet
were 113 ng/J (0.26 lb/MM Btu). As discussed in Section 5, these emissions
approximate uncontrolled emissions owing to multlclone malfunction during
the test period. This uncontrolled emission rate is nearly a factor of
four greater than the AP-42 value of 29.4 ng/J (0.07 lb/MM Btu). Data
presented 1n Table 6-9 Indicate that particulate emissions prior to
6-14
-------
scrubbing were substantially greater during test 202-1 than during test
202-4, indicating that coal ash from previous tests may have been emitted
during early oil-firing tests. If particulate data from test 202-4 are
assumed to be representative of oil-firing emissions, the particulate
emission rate of 59 ng/J (0.14 lb/MM Btu) measured at the scrubber Inlet
exceeds the AP-42 value by a factor of two. This assumption appears to be
valid since PLM data indicate that particulates from Test 202-4 are composed
primarily of oil soot and sulfate compounds (refer to the Inorganic sub-
section for analysis).
Total particulate emissions after scrubbing were 17.6 ng/J (0.04 lb/
MM Btu), on the average. This controlled emission rate corresponds to a
scrubber particulate removal efficiency of 84%. However, based on the
particulate catch known to be free of coal ash contamination, the scrubber
efficiency appears to be approximately 75% for oil-firing particulates.
Particulate emissions after scrubbing are well below existing NSPS standards
of 43 ng/J (0.10 lb/MM Btu) although they are slightly higher than the
proposed limitation of 13 ng/J (0.03 lb/MM Btu).
Particulate Size Distribution
Size distributions of oil-firing particulates at the scrubber inlet
and outlet were determined with an Anderson cascade impactor. Attempts at
additional size distribution determinations by PLM proved unsuccessful due
to the agglomerating nature of oil soot, a major component of oil-fired
particulates. As discussed previously, particulates from early o1l-f1r1ng
tests (such as Test 202-1) collected prior to scrubbing appear to contain
residual particulate from coal fired tests. Since coal particulate 1s
generally coarser than oil particulate, contaminated oil particulate would
probably appear to be somewhat coarser than pure oil particulate samples.
Impactor data for test 202-1 are summarized in Table 6-12. The substantial
difference between scrubber inlet and outlet size distributions Indicates
that larger particulates are removed with greater efficiency than small
particulates. Emission rate data and removal efficiencies are presented 1n
Table 6-13. These data show that particulates larger than 3 microns 1n
diameter are removed with greater efficiency than smaller particles.
6-15
-------
TABLE 6-12. PARTICULATE SIZE DISTRIBUTION
BY WEIGHT - TEST 202-1
Aerodynamic Diameter Weight %
Size Range, Microns Scrubber Inlet Scrubber Outlet
<1
20
83
1 -
3
1
12
3 -
10
74
5
>10
5
0
TABLE 6-13. EMISSION RATES OF PARTICULATES
FOR AN OIL-FIRED BOILER IN
TEMPORAL UNITS - TEST 202-1
Aerodynamic Diameter kg/hr Removal
Size Range, Microns Scrubber Inlet Scrubber Outlet Efficiency
<1
4.48
2.27
49.2
1 - 3
0.22
0.33
<0
3 - 10
16.6
0.14
97.4
>10
1.12
0.00
100
Total
22.4
2.74
87.8
6-16
-------
Sulfur Compounds: S02» SOg, and SO^
S02 was monitored continuously using a pulsed fluorescent analyzer.
SOg was determined as using the Goksoyr-Ross controlled condensation
system, and SO^" was determined by anion analysis of the particulate
extracts from the Method 5 sampling train. A summary of these analytical
results is presented in Table 6-14. As can be seen from the sulfur balance,
90 to 92% of the input sulfur is emitted as SO2 when emissions are un-
controlled. The removal efficiency for S02 is high: 97 to 98%. 28 to 29%
of the SO3 was also removed by the scrubber. This low efficiency shows
that SO^ 1s not associated with larger particulates, which are efficiently
scrubbed, but probably as fine liquid aerosols. Using similar reasoning,
the higher S04~ removal efficiency indicates that S0^~ 1s more likely than
SOg to be associated with larger particulates.
Table 6-15 shows the breakdown of the sulfate emissions into the water
and acid soluble fractions before and after the scrubber. While both types
of sulfates were removed by the scrubber, the fraction of water soluble
sulfates increased from 56 to 88%. One explanation is that the acid soluble
fraction is more efficiently removed than the water soluble fraction. This
cannot be checked by comparing the removal efficiencies of the major element
cations (Table 6-17) as a function of the solubility of their sulfates
because the types of sulfates present are not known. A second explanation
may be that the water and acid soluble fractions are removed with comparable
efficiency, and that the scrubber contributes small quantities of water
soluble sulfate to the gas stream passing through. Because of this possi-
bility, an analysis effort utilizing Fourier Transform Infrared Analysis
(FTIR) and X-ray diffraction analysis was initiated to determine the nature
of SO^ emissions. For the dual alkali system, possible sulfate species
would be CaSO^ and CafHSO^^ from the scrubber regeneration step, NaHSO^ and
from oxidation of NaHSOg and ^SOg. Calcium sulfate and b1 sulfate
are ruled out because of the low calcium concentration at the outlet (70
O
yg/m ; see Table 6-16). Fourier transform Infrared analysis showed the
presence of NaHSO^ at both the inlet and outlet of the scrubber, but x-ray
diffraction results Indicated that NaHSO^ made up less than 1% of the
6-17
-------
TABLE 6-14.
SO2 » SOg »
AND S04 EMISSIONS FROM
OIL FIRING
Pollutant
Concen-
tration
mg/m3
Mass
Emission
Rate
g/hr
Thermal
Emission
Rate
ng/J
Production
Emission
Rate
mg/kg steam
% of Fuel
Sulfur
Found in
Flue Gas
Removal
Efficiency
%
so2
202-1 Inlet
2582
127,500
940
2811
92
202-1 Outlet
88.5
4,371
32.2
96.4
3
97
202-4 Inlet
2689.2
106,060
874
2514
90
202-4 Outlet
59.2
2,335
19.2
55.4
2
98
so3
202-1 Inlet
20.6
1,017
7.50
22 ^ 4
0.59
202-1 Outlet
14.6
721
5.31
15.9
0.42
29
202-4 Inlet
25.8
1,018
8.39
24.1
0.69
202-4 Outlet
18.6
734
6.05
17.4
0.50
28
so4a
202-4 Inlet
70.4
2,776
22.9
65.8
1.58
202-4 Outlet
28.0
1,140
9.10
26.2
0.63
60
Total
202-1 Inlet
93
202-1 Outlet
3.6
202-4 Inlet
93
202-4 Outlet
3.1
6-18
-------
TABLE 6-15.
SUMMARY OF SULFATE EMISSIONS *
DURING OIL-FIRING - TEST 202-4
/ 3
mg/m
Inlet
Outlet
Water Soluble
39.7 (56%)
24.5 (88%)
Acid Soluble
30.7 (44%)
3.5 (12%)
Total
70.4
28.0
*
See Appendix C, Table 15, for details.
particulate matter 1n the gas stream, whereas sulfates accounted for 40%
of the particulate matter at the Inlet and 60% at the outlet. It 1s
possible that scrubber generated Na,,S04 caused the change in sulfate
solubilities, but this has not been confirmed.
Inorganics
Concentrations of 22 major trace elements present 1n the flue gas at
the scrubber Inlet and outlet are presented In Table 6-16. MATE values [6]
for these elements are also presented for comparison. As discussed 1n
Section 5, tabulated MATE values represent air stream concentrations which
are not considered to adversely affect human health upon short term exposure.
Of the 22 elements analyzed, 11 exceed their respective MATE values at
the scrubber Inlet and 5 exceed their MATE values at the scrubber outlet.
The 5 elements exceeding their MATE values at the scrubber outlet are
arsenic, cadmium, chromium, nickel and vanadium. The MATE value for arsenic
is extremely low because arsenic 1s a cumulative poison producing chronic
effects 1n humans. MATE values for nickel and chromium are extremely low
due to considerations for potential human carcinogenicity. Similarly, the
low MATE value for cadmium results from considerations of potential carci-
nogenic, oncogenic and teratogenic effects upon humans. The MATE value for
6-19
-------
TABLE 6-16.
EMISSION CONCENTRATIONS
DURING OIL-FIRING TEST
OF TRACE ELEMENTS
202-4
Element
Scrubber
Inlet
mg/m3
Scrubber
Outlet
mg/m3
MATE
Value
mg/m3
Degree of Hazard*
scruDDer ScruBB^F
Inlet Outlet
CD
!
i
i
i
<0.001
0.001
0.002
<0.50
0.50
Hg+
0.0016
0.0002
0.05
0.032
0.004
Ca
0.41
0.070
16
0.026
0.004
Mg
0.31
0.030
6.0
0.052
0.005
Sb
0.062
0.006
0.050
1.24
0.120
As
0.15
0.030
0.002
75.0
15.0
B
0.53
0.039
3.1
0.171
0.013
Cd
0.28
0.066
0.010
28.0
6.60
Cr
0.17
0.018
0.001
170
18.0
Co
0.10
0.012
0.050
2.0
0.24
Cu
0.54
0.007
0.20
2.70
0.035
Fe
4.8
0.28
1.0
4.8
0.28
Pb
0.20
0.013
0.15
1.333
0.087
Mn
0.03
0.004
5.0
0.006
0.001
Mo
0.22
0.025
5.0
0.044
0.005
N1
1.1
0.20
0.015
73.3
13.33
V
2.7
0.82
0.50
5.40
1.640
Zn
0.61
0.065
4.0
0.153
0.016
Se
0.050
0.006
0.200
0.25
0.03
Sr
0.043
0.001
3.1
0.014
0.0003
A1
5.7
0.48
5.2
1.096
0.092
Zr
0.015
0.001
5.0
0.003
0.0002
Total
18
2.5
__ ^
Beryllium was determined by Spark Source Mass Spectrometry (SSMS). The
other values, with the exception of mercury, are determined by Inductive!v
Coupled Plasma Optical Emission Spectroscopy (ICPOES) analysis. y
+Mercury was determined by cold vapor analysis of SASS train samples taken
during test 203.
^Degree of hazard 1s defined as the ratio of the discharge concentration to
the MATE value.
6-20
-------
vanadium 1s comparatively higher since vanadium has been associated with
eye and bronchial Irritation without Indication of chronic effects. Emis-
sions of arsenic and chromium after scrubbing are below their TLV's which
3
are each 0.5 mg/m . Hence, if the TLV's were used as the basis for compar-
ison, emissions of arsenic and chromium would be considered less hazardous.
Emissions of cadmium, nickel and vanadium after scrubbing exceed their TLV
values of 0.05, 0.1 and 0.5 mg/m , respectively, in addition to exceeding
their respective MATE values.
3
Beryllium emissions were measured to be 0.001 mg/m after scrubbing;
this corresponds to half the MATE value for this element. At this emission
concentration, the National Standard for Hazardous Air Pollutants limitation
of 10 grams beryllium per day would only be exceeded by boilers of 100 MW
capacity or greater.
Emission factors and mass emission rates for the 22 elements analyzed
are presented 1n Table 6-17. Also presented in the table are scrubber
removal efficiencies for each element. An overall removal efficiency of
87% was obtained for these elements although several elements were removed
with less efficiency, namely calcium, arsenic, cadmium, nickel, and
vanadium. It is interesting to note that, with the exception of
chromium, all elements which exceeded their MATE values at the scrubber
outlet were removed with lower than average efficiency during scrubbing.
Enrichment factors across the scrubber have been computed for each
element and are presented 1n the last column of Table 6-17. The enrichment
factor 1s defined as the ratio of the concentrations of trace element to
aluminum at the scrubber outlet divided by the corresponding ratio at the
scrubber Inlet. As discussed in Section 5, aluminum is selected as the
reference material because 1t has been known to partition equally among
particulates of different sizes. Tabulated enrichment factors Indicate
that all elements analyzed are enriched across the scrubber with the
exception of boron, copper, Iron, lead, strontium and zirconium. The
observed enrichment 1s principally due to the partitioning of trace
6-21
-------
TABLE 6-17. EMISSION FACTORS AND MASS EMISSION RATES OF
TRACE ELEMENTS DURING OIL-FIRING TEST 202-4
Element
Emission Factor,
nq/J
Emission
9/hr
Rate,
Removal
Efficiency
%
Enrichment
Factor
Scrubber
Inlet
Scrubber
Outlet
Scrubber
Inlet
Scrubber
Outlet
*
Be
<0.0003
0.0003
<0.04
0.04
Unknown
>11.9
Hg+
0.0006
0.0001
0.05
0.006
87
1.48
Ca
0.13
0.022
16
2.7
83
2.03
Mg
0.10
0.0094
12
1.1
91
1.15
Sb
0.02
0.0019
2.5
0.23
91
1.15
As
0.049
0.0094
5.9
1.1
81
2.37
B
0.17
0.012
21
1.5
93
0.87
Cd
0.091
0.021
11
2.5
77
2.80
Cr
0.055
0.0057
6.7
0.69
90
1.26
Co
0.033
0.0038
3.9
0.45
89
1.43
Cu
0.18
0.002
21
0.27
99
0.15
Fe
1.6
0.088
190
11
95
0.69
Pb
0.065
0.0041
7.9
0.50
94
0.77
Mn
0.010
0.0013
1.2
0.15
87
1.58
Mo
0.072
0.0079
8.7
0.95
89
1.35
N1
0.36
0.063
43
7.7
83
2.16
V
0.88
0.26
110
31
71
3.61
In
0.20
0.02
24
2.5
90
1.27
Se
0.016
0.002
2.0
0.23
87
1.43
Sr
0.014
0.0003
1.7
0.038
98
0.28
A1
1.9
0.15
220
18
92
1.0
Zr
0.0049
0.0003
0.59
0.038
94
0.79
Total
6.0
0.78
710
96
87
Beryllium was determined by SSMS. The other elements, except mercury,
determined by ICPOES.
+Mercury was determined by cold vapor analysis of SASS train samples taken
during test 203.
6-22
-------
elements as a function of particulate size, and the greater collection
efficiency of the scrubber for the large size particulates. It may also
be noted that many of the trace elements that show enrichment, such as
selenium, arsenic and mercury, either occur as elemental vapors or form
volatile compounds at furnace temperatures. Condensation and surface
adsorption of the more volatile elements or their compounds downstream of
the furnace could, therefore, result in higher concentratlonf of these
elements on smaller particulates.
ESCA analyses were performed on particulates from oil firing to deter-
mine surface and subsurface elemental concentrations. The ESCA results are
presented 1n Table 6-18 as normalized atom percentages for each particulate
catch and penetration depth. Scrubber inlet data Indicate that coarser
particulate matter collected by the cyclone differs somewhat from finer
particulates collected on the filter. The principal difference 1s the
higher carbon and lower silica contents of the coarser particulate. It 1s
Interesting to note that the filter catch particulates from the scrubber
inlet and outlet yielded very similar ESCA analyses both at the particle
0
surface and at 76 A penetration. Elements which appear to be enriched on
the surface Include sulfur, phosphorous and carbon. On the other hand,
vanadium and the typical ash components aluminum and silicon appear to be
more concentrated on the bulk of the particulate matter.
The composition of particulates at the scrubber Inlet and outlet
during Test 202-4 has also been determined by PLM analysis. Estimated
weight percentages of particulate components are presented 1n Table 6-19.
Particulates at the scrubber Inlet are composed primarily of oil soot,
various sulfate/sulfite compounds and fused ash while outlet particulate
1s composed largely of sulfates and sulfites. Sulfate data presented
previously Indicate that scrubber inlet particulates contain approximately
4035 sulfate 1on. Hence, the PLM estimate for CaS0g*l/2 H^Q and unknown sulfates
weight percentages of 20 to 3935 appears to be somewhat low. Further, the
CaS03*l/2 H20 would appear to be present only as a minor constituent. This
may Indicate that the tabulated Inlet weight percentage of oil soot, the most
6-23
-------
TABLE 6-18. DEPTH PROFILE ANALYSIS OF OIL PARTICULATE WITH CONCENTRATIONS
EXPRESSED AS NORMALIZED ATOM PERCENT* - TEST 202-4
0
Na
S
Si
A1
Fe
CI
P
V
Ca C K
Inlet
Level
II Cyclone Catch
38.5
3.2
12.9
2.6
2.3
1.4
0.6
1 .2
37.4
Level
II Filter Catch
48.6
4.2
10.7
11.8
3.1
1.3
1.5
1 .7
17.2
Level
II Filter Catch;
76A
46.3
4.7
6.5
17.1
3.3
0.8
1.4
1.1
3.3
0.9 12.9 1.7
Outlet
Level
II Filter Catch
45.5
5.7
9.7
14.6
3.4
1.3
2.3
1.6
1.7
14.1
Level
II Filter Catch;
76A
53.5
3.2
6.8
22.0
4.0
1.3
1.0
2.0
6.2
The atom percent of the twelve elements presented here adds up to 100 percent. Other elements
present in the cyclone and filter catches were not studied in ESCA. Hence, the atom percents in
this table are normalized atom percents and not absolute atom percents.
-------
TABLE 6-19. MAJOR COAL PARTICULATE COMPONENTS - TEST 202-4
Approximate Weight %
Component Inlet Outlet
Ash
Fused 13-23
Unfused 1 - 4
Minerals
Fe203 <1 8-16
Fe304 (magnetite)
< 1
CaC03 S3
Oil Soot 43-57 2-8
CaS03«l/2 H?0 and 20 - 39 80 - 90
Unknown Sulfate
*Iron-a1um1num silicates.
difficult particulate component to quantify, is high with respect to the
true particulate composition. Similarly, sulfate data indicate that outlet
particulates are approximately 632! sulfate ion. Assuming that sulfate ions
are combined with calcium or sodium, outlet particulates may consist of
up to 92% sulfate compounds. Hence, CaS03*l/2 H20, If present in either
the scrubber inlet or outlet, appears to be a minor constituent of parti-
culates generated during oil-firing.
Chloride, Fluoride, and Nitrate Emissions
The results of anion analysis on extracts of particulate matter
collected at the inlet and outlet of the scrubber are presented 1n Table
6-20. The fluoride removal efficiency Is high, 89%, as would be expected,
since the overall removal efficiency of the trace element cations 1s also
high, 87%. The lower removal efficiencies for chlorides and nitrates,
52% and 57%, respectively, suggest that these anions may be associated with
the finer particulate matter which is not efficiently removed by the scrubber.
6-25
-------
TABLE
6-20. CHLORIDE, FLUORIDE, AND NITRATE EMISSIONS
FROM OIL FIRING - TEST 202-4
Inlet
~
Outlet
Removal
mg/m3
ng/J
3
mg/m ng/J
Efficiency,
%
Chloride (CI")
0.46-0.47
0.15
0.22 -0.23 0.072-0.075
52
Fluoride (F~)
0.054
0.017
0.006-0.008 0.002-0.003
89
Nitrate (NO^-)
0.24
0.076
0.102 0.033
57
~
Mass as the 1on.
Organlcs
Organlcs In the gaseous effluent stream were analyzed using both Level
1 (Test 203) and Level 2 (Tests 202-1 through 202-4) procedures. Samples
were taken using the SASS train 1n Test 203 and were analyzed by gas
chromatography for total chromatographable organlcs (GC/TCO), by gravlmetry
by Infrared absorption spectroscopy (IR), and by low resolution mass
spectrometry (LRMS). GC/TCO analysis yields values for organlcs 1n the
C7 to Clg range, a value for the total concentration of high molecular
weight organlcs is determined gravlmetrlcally, IR analysis of the residues
from the gravimetric determination Identifies the classes of compounds
present 1n the high molecular weight fraction, and LRMS on sample fractions
separated by liquid chromatography (LC) provided further Identification
information. The subscripted carbon numbers refer to boiling ranges of
organic mixtures; these are presented 1n Table 5-23.
In Tests 202-1 through 202-4, three types of samples were taken. The
gas stream was continuously monitored with a flame Ionization detector (FID)
for total organlcs, expressed as CH4; gas bag samples were taken for
analysis of the C-j to Cg organlcs; and samples from a modified Method 5
6-26
-------
sampling train (with an XAD-2 resin module) and from gas bags were analyzed
for Identification of volatHes using gas chromatography/mass spectrometry
(GC/MS).
Table 6-21 summarizes the organic concentrations by category. As
mentioned previously, the apparent Increase 1n total organlcs across the
scrubber 1s not statistically significant. Any C-j to Cg organlcs were
present at sufficiently low concentrations that they were not detected.
The C7 to C16 fraction was removed with 85% efficiency, and higher molecular
weight organlcs with 83% efficiency. In the range of C7 and above, the
concentration of higher molecular weight organlcs was much greater than the
concentration of volatiles.
TABLE 6-21. COMPARISON OF ORGANIC MEASUREMENT METHODS
Test
Method
Organic Concentration
mq/m*
% Change
Inlet
Outlet
202-1
Total as CH^
(continuous FID)
8
9
Not
Significant
202-1
C"| - Cg (GC on
bag samples)
*12.7
*12.7
—
203
C7 " ^16 on
SASS train samples)
0.469
0.07*
85%
203
*c16
(Gravimetric on
SASS train samples)
7.27
*
1.2
83%
*The values for the XAD-2 resin fraction were calculated using data
obtained from similar sampling trains on four oil-fired boilers.
6-27
-------
Results of IR analysis on the fractionated scrubber Inlet XAD-2 resin
extract from SASS train samples are presented 1n Table 6-22. Other SASS
train samples contained Insufficient organic material to warrant IR analysis.
It 1s apparent that the blank contained many organic materials released by
the XAD-2 resin. This was a source of difficulty 1n Interpreting analytical
results, especially because the amount of resin-generated material seemed
to vary from sample to sample. The two classes of compounds analyzed by ir
which were not also found 1n the blank are heterocyclic sulfur compounds
and sulfonic acids. LRMS Identified the heterocyclic sulfur compound as
dlbenzothlophene, which 1s probably a reaction product of S0£ and organic
material from combustion or from the resin. Likewise the sulfonic acid may
be an S02 reaction product, or it may be contamination from a cleaning
agent.
Table 6-23 presents results of the GC/MS analysis. Only those com-
pounds are listed which were not present 1n blank samples. None of these
compounds would be directly associated with combustion. Several are, however
associated with lubricants, material used 1n the manufacture of gas bags,
and solvents used for cleaning test apparatus. These results agree with
the GC/TCO, gravimetric, and FID data indicating low organic emissions.
Polycycllc Organic Material
Polycycllc organic material (POM) was not found in either the scrubber
Inlet or outlet samples at detection limits of 0.3 yg/m3. This observation
1s consistent with the findings to date from the EPA sponsored project
"Emissions Assessment of Conventional Combustion Sources". However, two
POM compounds for which MATE values are below 0.3 yg/m3 are benzo(a)pyrene
and d1benz(a,h)anthracene. The MATE values for benzo(a)pyrene and
3 3
d1benz(a,h)anthracene are 0.02 yg/m and 0.09 yg/m , respectively. While
available data indicate that many POM compounds are not present at
concentrations greater than or equal to their MATE value during oil firing
additional analyses at higher GC/MS sensitivity would be required to con-
clusively preclude the presence of all POMs at concentrations above their
MATE values.
6-28
-------
TABLE 6-22. SUMMARY OF THE INFRARED ANALYSIS OF ORGANICS
FROM OIL COMBUSTION - TEST 203, INLET
LC Fraction
Total Organles, LCI LC2* LC3 LC4 LC5 LC6 LC7 Total
Sample/Blank Sample/Blank Sample/Blank Sample/Blank* Sample/Blank Sample/Blank Sample/Blank Sample/Blank
mg/w3 0.09 0.02 0.05 0.03 0.03 O.D1 0.04 0.02 0.2B 0.10 0.02 0.51 0.18
Category Intensity
Aliphatic Hydrocarbons
XM
XM
XM
XM
Xm
Xm
Xm
Xm
Aromatic Hydrocarbons
Xm
Xm
XM
XM
Xm
Xm
Xm
Xm
Xm
Silicones
Xm
Xm
Xm
Heterocyclic Sulfur
Compounds
Xm
Xm
Xm
N1tro Compounds
0m
0m'
Om
Om
Esters
Xm
XM
XM
XM
XM
XM
XM
Xm
Amides
OM
Om
Glycols
XM
XM
XM
XM
Carboxylie Acids
XM
XM
XM
XM
Sulfonic Acids
XM
Silicates XM
*Not analyzed. 0 = At least one species suspected present M = Major component
X = At least one species present m * Minor component
-------
TABLE 6-23. ORGANIC COMPOUNDS IDENTIFIED BY
GC/MS IN FLUE GAS DURING OIL FIRING
Compound
Proplonaldehyde
Nitromethane
Ethyl-n-butyl Ether
Ethyl Acetate
4-raethyl-3-pentene-2-one
Octanol
Ketone (MW 138)
Ketone (MW 140)
Phthallc Anhydride
Amyl Benzoate
Glycerol Triacetate
Concentration, uq/m
Inlet
200
1700
14
Outlet
570
1.6
WF,
Inlet Outlet
142
63
1000
350
45
20
29
42
2300
28
6-30
-------
Scrubber Efficiency
Flue gas analyses indicate that scrubber processing removes significant
percentages of flue gas sulfur oxides (SOg, SOg and SO^ ), total particulates
and organlcs of the Cy class and higher. Scrubber removal efficiency data
for these flue gas components are summarized in Table 6-24. The average
removal efficiencies have been discussed in previous subsections. However,
1t 1s Important to note that it 1s the Cy and higher hydrocarbons which are
removed with 84% efficiency. These fractions comprise 38 to 96% (the most
accurate estimate appears to be 53%) of the organics measured at the
scrubber inlet. Hence, based on the total organics generated, a 32
to 84% removal efficiency was obtained.
TABLE 6-24. SCRUBBER EFFICIENCY DURING OIL FIRING
Test % Removal
Number I* Total and Higher
SO2 SO3 S04~ Particulates Organics
202-1 97 29 -- 88
202-2 97
202-3 97
202-4 98 28 60 75
203 - - ~ — 84
Average 97 29 60+ 82 84+
if
This removal rate 1s actually a net change rate. As described earlier, the
scrubber removes and, possibly, generates sulfates.
+0nly one data point.
6-31
-------
It 1s not known for certain by what process organlcs are removed with
such a high degree of efficiency. There are, however, at least three
possible mechanisms:
• Dissolution - Some organlcs are partially water soluble.
These compounds could be removed by dissolving 1n the
slurry.
• Condensation - High boiling organlcs could condense and
be removed as "particulate".
• Sorption - Some organlcs could adsorb on particulates.
One or any combination of these mechanisms may account for the high removal
efficiencies.
LIQUID WASTE
As discussed previously, only one significant waste water stream is
produced. The stream 1s a combination of water treatment waste, boiler
blowdown, and acid waste water from elsewhere In the manufacturing plant.
The quality of this combined stream 1s such that 1t 1s acceptable for
disposal Into the municipal sewer system. Liquid streams from the scrubber
operation are passed to the thickener and recirculated to the scrubber
after the filtration step. There 1s no direct wastewater discharge from
the scrubber operation, as the process 1s designed to dispose of all of the
water that enters Its system through evaporation and moisture entrained 1n
the scrubber cake.
Because several streams are mixed together, It 1s not possible to
accurately determine what part of the effluent 1s attributable to the boiler
However, the flow rate of the combined stream 1s approximately 10,000
Uters/hr (40 gallons/m1n).
Water Quality Parameters
Table 6-25 summarizes the waste water parameters for the combined waste
water stream. Note that these values do not represent water produced solely
by the boiler but also Include process waste.
6-32
-------
TABLE 6-25. COMBINED STREAM WASTE WATER PARAMETERS - OIL FIRING
Test Test Test Test Test
Parameter 202-1 202-2 202-3 202-4 203 Average
PH 7.5 6.5 7.5 6.5 6.9 7.0 ± 0.5
105 140 155 150 110 132 ± 23
"5 65 UO 50 140 102 ± «
Cyanide, mg/1 0 0 0 0 0 0
Inorganics - Combined Waste Water Stream
Analyses of major Inorganic cations 1n the combined waste water
stream are presented 1n Table 6-26. These data are based on the SSMS
technique which 1s accurate to within a factor of approximately 3. Of the
eleven elements analyzed, none exceeds Its MATE value based on health
considerations while only nickel and copper may exceed their respective
HATE values based on ecological considerations.
rtrqanlcs - Combined Waste Water
•
Concentrations of C^ to C16 organlcs measured 1n the combined waste
water stream are summarized 1n Table 6-27. The only fractions analyzed to
contain even low concentrations of organlcs were the C-jq, C^ and C^g
fractions, and these were each below 0.1 mg/Hter. Higher molecular weight
organlcs were not detected. Thus, the total quantity of organlcs measured
in the combined waste water stream was less than 0.3 mg/l1ter. As a basis
for comparison, the water MATE values for alkanes, alkenes and alkynes are
fn the 500 to 14,000 mg/1iter range based on human health considerations and
in the 1.0 to 100 mg/1 range based on ecological considerations. Discharge
concentrations of organlcs 1n the combined waste water stream are well
*1th1n these MATE values.
6-33
-------
TABLE 6-26. WASTE WATER INORGANICS - OIL FIRING
~
MATE Value, mq/1
Degree of Hazard
Element
mg/1
g/hr
Health
Ecology
Healtn
Leo logy
Be
<0.001
<0.01
0.030
0.055
0.033
0.018
F
4
40
38.0
—
0.1
--
V
0.002
0.02
2.50
0.150
0.001
0.013
Cr
0.02
0.02
0.250
0.250
0.08
0.08
Co
0.007
0.07
0.750
0.250
0.009
0.028
N1
0.02
0.2
0.230
0.010
0.087
2
Cu
0.02
0.2
5.0
0.050
0.004
0.4
Sr
0.3
3
46.0
—
0.007
--
Cd
<0.001
<0.01
0.050
0.010
<0.02
<0.004
Sb
0.001
0.01
7.50
0.200
0.0001
0.005
Pb
0.006
0.06
0.250
0.050
0.024
0.005
*Flow rate of 10,000 liters per hour.
TABLE 6-27. SUMMARY OF C7 - Ci6 ORGANIC IN
THE WASTE WATER - OIL-FIRING
Carbon Number mg/1 Carbon Number mg/1
~
C7
ND
C13
ND
C8
ND
C14
<0.1
C9
ND
C15
<0.1
C10
<0.1
C16
ND
C11
ND
C12
ND
Total C^ — C^g
<0.3
*ND means none detected.
6-34
-------
SOLID WASTE
Three solid waste streams are produced by the system:
• Bottom ash;
• Fly ash;
• Scrubber cake.
Table 6-28 shows the approximate quantities of bottom ash and scrubber cake
that were produced. Only small quantities of fly ash were collected during
the test period due to the inefficiency of multlclones in capturing fine
particulates generated during oil firing.
The scrubber cake produced after filtration has the appearance of a
clayey silt. Its grain size 1s quite uniform and characteristic of sllty
soils, but Its behavior closely resembles a clay in many respects. As
obtained from the vacuum filter, the scrubber cake consists of small lumps
and appears to be relatively dry; in actuality, however, the water content
generally ranges from about 30 to 50%.
Assuming that calcium sulfite hemihydrate (CaS03*l/2 H20) 1s the
principal product from S02 scrubbing and NagSOj regeneration, data presented
In Table 6-29 represent the estimated composition of scrubber cake produced
during o1l-f1r1ng. Although the scrubber cake production rate was not
measured for Test 202-4, 1t has been estimated as the average of production
rates determined for other oil-fired tests performed. Data presented 1n
Table 6-29 Indicate that the cake 1s composed of 44 to 50% unbound water
and at least 47% calcium sulfite hemihydrate. These data reflect the low
particulate emissions which are characteristic of oil firing. Only 1% of
the scrubber cake 1s estimated to be particulate.
Although the scrubber cake material 1s composed predominately of
relatively Insoluble solids (calcium sulfite, calcium sulfate, and some
calcium carbonate), the 1ntersit1t1al water does contain soluble residues
of Hme, sulfate, sulfite and chloride salts. Trace elements In the fly
ash may also contribute to the leachate from the disposed scrubber cake and
are of special concern. Concentrations of 20 trace elements 1n the scrubber
6-35
-------
TABLE 6-28. SOLID WASTE PRODUCTION RATES - OIL FIRING
Bottom Ash Scrubber Cake*
Test
kg/hr
ng/J
kg/hr yg/j
202-1
<1
< 7.4
400 2.9
202-2
<1
< 7.4
550 4.1
202-3
<1
< 7.6
380 2.9
202-4
<1
CO
00
V
Not Measured
203
<1
<11.1
Not Measured
¦k
Scaled up to represent 100% of the flue gas for boiler No. 4.
TABLE 6-29. ESTIMATED SCRUBBER CAKE MASS BALANCE
FOR OIL FIRING - TEST 202-4
Contribution to Scrubber Cake
Component kg/hr Weight %
Fly Ash Removed by Scrubber
5
1
CaS03*l/2 H2O Formed from SO2
Scrubbing and NagS03
Regeneration
210
47
CaS04, CaC03, NagSO^ Ca(0H)2
NaHS03 an" Na2S03 Losses
(Estimated)
6 - 35
1 - 8
Water
193 - 222
44 - 50
Total
443
100
Total cake production rate was estimated from the average of tests
202-1, 202-2 and 202-3.
6-36
-------
cake are presented in Table 6-30. With the exception of antimony, boron,
molybdenum and zinc, all trace elements were found to exceed human health
based MATE values for sol Ids. Similarly, with the exception of boron,
all trace elements were found to exceed ecology based MATE values for
solids. These results are a consequence of reducing a high volume of low
concentration wastes to a low volume of concentrated wastes. The high
degree of hazard for most elements appears to warrant disposal of these
solid wastes in specially designed landfills.
An overall mass balance for the 20 trace elements is presented 1n
Table 6-31. The percent of the trace element in the oil feed that could be
located 1n the effluent streams (scrubber cake and scrubber effluent gas)
1s used as a measure of mass balance closure. Elements having concentrations
1n the oil and scrubber cake which were less than the ICPOES detection limits
have been estimated as Indicated 1n the table. Where ICPOES analysis of
oil has yielded only upper limit data, elemental concentrations have been
estimated from ICPOES analysis of the scrubber Inlet gas stream assuming
that no losses were Incurred prior to scrubbing. Where ICPOES analysis of
the scrubber cake yielded only upper limit data, SSMS analysis were utilized.
Good mass balance closure was obtained for arsenic, boron, chromium, cobalt,
copper, molybdenum, nickel, vanadium, zinc and selenium. However, as would
be expected due to these extremely low elemental concentrations, mass balance
closure for some elements 1s poor. Instances 1n which the effluent flow
rate of an element substantially exceeded the input feed rates, such as
with iron and aluminum, may be the result of the extremely high elemental
concentrations attained during coal firing and subsequent contamination of
the recycle scrubber solution.
The scrubber cake was also analyzed for organlcs but none were detected.
This may be expected since the concentrations of organlcs 1n the flue gas
stream was very low.
6-37
-------
TABLE 6-30. INORGANIC CONTENT OF SCRUBBER CAKE
FROM OIL FIRING (DP.Y BASIS) - TEST 202-4
Element Concentration MATE Value, uo/g Degree of Hazard
ua/a Health Ecology Health Ecology
Ca
200,000
480
32
417
6,250
Mg
3,799
180
174
21
22
Sb
*
3
15
0.4
0.2
7.
As
*
15
0.5
0.1
30
150
B
40
93
50
0.4
0.
Cd
*
1
0.1
0.002
10
500
Cr
15
0.5
0.5
30
30
Co
*
19
1.5
0.5
13
38
Cu
16
10
0.1
2
160
Fe
2,164
3.0
0.5
721
4,328
Pb
*
6
0.5
0.1
12
60
Mn
6
0.5
0.2
32
80
Mo
*
14
150
14
0.1
1
N1
132
0.45
0.02
293
6,600
V
203
5.0
0.3
41
677
Zn
36
50
0.2
0.7
180
Se
*
9
•o.io
0.05
90
180
Sr
239
92
—
2.6
--
A1
1 ,684
160
2.0
11
842
Zr
37
15
mi •*
2.5
"• •»
Total
208,450
*SSKS analyses were utilized where 1CP0ES analysis provided upper limit
data only.
6-38
-------
TABLE 6-31. MASS BALANCE OF TRACE ELEMENTS - TEST 202-4
Element
Oil Feed
Scrubber
Scrubber
Percent
g/mln
Cake
Outlet
Recovery
g/m1n
g/m1n
Ca
16.4
50,000
2.8
**
Mg
(12.2)*
950
1.2
>1,000
Sb
( 2.4)
0.8*
0.2
42
As
( 5.9)
~
3.9
1.2
68
B
(20.9)
10.0
1.5
55
Cd
(11.0)
~
0.2
2.6
25
Cr
3.6
3.9
0.7
125
Co
( 3.9)
~
4.7
0.5
133
Cu
4.2
4.1
0.3
105
Fe
36.7
541
11.0
>1,000
Pb
( 7.9)
~
1.5
0.5
25
Mn
( 1.2)
4.0
0.2
350
Mo
< 8.7)
3.5*
1.0
52
N1
47.7
33.0
7.9
91
V
108.8
50.7
32.3
76
Zn
8.9
9.1
2.7
133
Se
( 2.0)
it
2.4
0.2
136
Sr
0.7
59.8
0.04
>1,000
A1
10.4
421
18.9
>1,000
Zr
( 0.6)
9.2
0.04
>1,000
*
SSMS data were utilized where ICPOES analysis provided upper limit
data only.
* ICPOES data from the analysis of scrubber Inlet particulates were
utilized when fuel analysis provided upper limit data only.
* Percent recovery of a trace element is 100 times the ratio of its total
emission rate (scrubber cake plus scrubber outlet) to Its feed rate.
iir*
Percent recovery for calcium 1s not calculated because most of the
calcium in the scrubber cake 1s from the lime slurry.
6-39
-------
ANNUAL EMISSIONS
Table 6-32 presents an estimate of the annual emissions of the major
pollutants for the controlled and uncontrolled case. It was assumed that
the boiler operates at 100% load, 87% of the year (7560 hours/year), and
that oil 1s the only fuel burned.
AIR QUALITY ASSESSMENT - OIL FIRING
Simplified air quality models were used to determine relative air
quality resulting from uncontrolled and controlled emissions. The ambient
air quality values are approximate only. The emphasis should be placed on
the relative values for each case as opposed to their absolute values.
Worst case weather conditions and typical weather conditions were
considered. The worst case was assumed to be plume trapping. An equation
proposed by Bierly and Henson [Ref. 7] was used with the following
assumptions: Inversion height 100 meters, wind speed 1.0 meter/second,
D class stability (neutrally stable) in the mixing layer and an effective
stack height of 50 m (164 ft). The typical case was assumed to correspond
to the standard Gaussian convective diffusion equation, [Ref. 8]. The
following conditions were used: wind speed 4.0 meters/second and D class
stability. These represent conditions that could reasonably be expected
to occur almost anywhere in the country. Typical does not mean average.
It was assumed that all species were inert. No photochemical reactions
were considered. (See Appendix A for details.)
Figures 6-1 through 6-8 present plots of approximate ground level
ambient air quality as a function of distance directly downwind from a
single 10 MW equivalent source. Data for N0X, CO, SOg and particulates
are presented. The purpose of these figures 1s not to attempt to accurately
predict air quality but rather to compare the effects of controlled and
uncontrolled emissions under an arbitrary but realistic set of meteo-
rological Gondltions. It 1s Implicit in this approach that each set of
meteorological conditions remains constant for a sufficient length of time
6-40
-------
TABLE 6-32. ANNUAL EMISSIONS - OIL FIRING*
Pollutant
Scrubber
Scrubber
% Difference
Inlet
Outlet
Gaseous
N0X (as N02)
164,230
157,390
- 4
CSJ
o
906,020
24,453
- 97
so3
7,249
5,183
- 28
II
sr
O
CO
20,894
8,303
- 60
CO
4,991
4,845
- 3
Organlcs (as CH^)
2,272
2,500
+ 16
C1 " C6
<4,164*
<4,164*
—
C7 " C16
155
18
- 88
>C16
2,381
392
- 83
Total particulates
53,832
13,686
- 75
10w
0
"
m
^/year
Liquid
Blowdown/waste water
<\.76,000
^76,000
0
Cooling water
<^86,000
-------
200
180
\ INLET
140
120
TEST 202-1
100
PRIMARY AND SECONDARY
STANDARD: ANNUAL
ARITHMETIC MEAN
80
OUTLETX
14
DISTANCE FROM STACK, km
Figure 6-1. Relative N0X quality under worst case
weather conditions - oil firing
6-42
-------
THE MAXIMUM EXPECTED AMBIENT NO*
CONCENTRATIONS ARE WELL BELOW
THE PRIMARY STANDARD OF 100 ^ g/m3
INLET
<
£
z
%
8
*-
Z
UJ
S
TEST 202-1
20- -
1
OUTLET
DISTANCE FROM STACK, km
Figure 6-2. Relative NOv air quality under typical
weather conditions - on firing
6-43
-------
CO
i
3.
z
o
i-
<
c
~-
z
1U
o
z
o
u
H
z
ui
THE MAXIMUM EXPECTED AMBIENT CO
CONCENTRATIONS ARE WELL BELOW THE
MOST RESTRICTIVE STANDARD OF
10 mg/m3 (MAXIMUM 8-HOUR AVERAGE).
y INLET
TEST 201-1
OUTLET
DISTANCE FROM STACK, km
Figure 6-3. Relative CO air quality under worst case
weather conditions - oil firing
6-44
-------
1.4
THE MAXIMUM EXPECTED AMBIENT CO
CONCENTRATIONS ARE WELL BELOW THE
MOST RESTRICTIVE STANDARD OF
10 ms/i»3 (MAXIMUM 8-HOUR AVERAGE).
INLET
TEST 201-1
OUTLET
DISTANCE FROM STACK, km
Figure 6-4. Relative CO air quality under typical
weather conditions - oil firing
6-45
-------
900
INLET
700
•00
TEST 202-1
tsoo
PRIMARY STANDARD:
MAXIMUM 24-HOUR
AVERAGE
400
300
200
PRIMARY STANDARD:
ANNUAL ARITHMETIC MEAN
100
OUTLET
DISTANCE FROM STACK, km
Figure 6-5. Relative SO2 air quality under worst case
conditions - oil firing
6-46
-------
260
226
INLET
200
176'
100
PRIMARY STANDARD:
ANNUAL ARITHMETIC
MEAN
50- -
OUTLET
DISTANCE FROM STACK, km
Figure 6-6. Relative SO? air quality under typical
weather conditions - oil firing
6-47
-------
200
INLET
SECONDARY STANDARD:
MAXIMUM 24-HOUR AVERAGE
100
140
Z 120
TEST 202-1
100
PRIMARY STANDARD:
ANNUAL GEOMETRIC
MEAN
80
SECONDARY STANDARD:
ANNUAL GEOMETRIC MEAN
OUTLET
DISTANCE FROM STACK, km
Figure 6-7. Relative particulate air quality under
worst case weather conditions - oil firing
6-48
-------
50
40- -
\
i
p
<
p
z
z
III
20
10- K
\
INLET
THE MAXIMUM EXPECTED AMBIENT
PARTICULATE CONCENTRATIONS ARE
WELL BELOW THE MOST RESTRICTIVE
STANDARD OF 60^ g/f3
(ANNUAL GEOMETRIC MEAN).
\
TEST 202-1
I
\
\
\
\
\
\
OUTLET
• •
DISTANCE FROM STACK, km
10
12
14
Figure 6-8. Relative particulate air quality under typical
weather conditions - oil firing
6-49
-------
for the ambient air quality to reach steady state conditions at each
distance. Note also that the plots represent a single line extending
directly downwind from the source.
Table 6-33 presents a summary of the ambient air quality standards
for each pollutant. The standards are also shown on each plot.
Keeping in the caveats mentioned above, several observations can be
made:
• The N0X standard is exceeded under worst case weather
conditions but not under typical conditions. Since the
scrubber does not remove significant amounts of N0X,
there is no substantial difference between the controlled
and uncontrolled cases. (The boiler has no N0X controls.)
• CO standards are not exceeded under any conditions. The
most restrictive standard is 10 mg/m^ (10,000 yg/m3)
while the maximum predicted level is only about 0.06% of
that value.
• Under worst case conditions uncontrolled SO2 emissions
result in both primary standards being exceeded. For
controlled emissions no standards are exceeded. Under
typical conditions the only annual primary standard is
exceeded by uncontrolled emissions. For controlled
emissions no standards are exceeded.
• One primary and both secondary standards are exceeded
by uncontrolled particulate emissions under worst case
weather conditions. Controlled emissions do not result
1n the violation of any standard. No standards are
violated under typical weather conditions.
6-50
-------
TABLE 6-33. NATIONAL AMBIENT AIR QUALITY STANDARDS
FOR CRITERIA POLLUTANTS
Pollutant Standard
Pollutant
it
Primary
Secondary+
Nitrogen dioxide
Carbon monoxide
Sulfur dioxide
Total suspended
particulates
100 yg/m (0.05 ppm)
annual arithmetic mean.
3
10 mg/m (0 ppm) maximum
8-hour average; 40 mg/m
(35 ppm) maximum 1-hour
average.
3
80 yg/m (0.03 ppm) annual
arithmetic mean; 365 yg/m3
(0.14 ppm) maximum 24-hour
average.
3
75 yg/m annual geometric
mean; 260 yg/m maximum
24-hour average
Same as primary
Same as primary
1300 yg/m (0.5 ppm)
maximum 3-hour average.
j
60 yg/m annual geometric
mean; 150 yg/m3 maximum
24-hour average.
*
Primary, necessary to protect the public health.
^Secondary, necessary to protect the public welfare.
-------
CONCLUSIONS - OIL FIRING IN A 10 MW INDUSTRIAL BOILER WITH FGD
1) Uncontrolled emissions of criteria pollutants do not generally
correspond with emission factors from AP-42. N0X emissions were
nearly 23% lower than the AP-42 emission factor, although they
appear to be within the normal range for similar Industrial units.
CO emissions were nearly 63% lower than the AP-42 emission factor.
SOg and total hydrocarbons corresponded well with their respective
AP-42 emission factors. Particulate emissions, in the absence of
coal ash contamination, are approximately twice the value tabulated
In AP-42.
2) Sulfur dioxide removal data Indicated an average scrubber efficiency
of 97%. Controlled SO2 emissions were 26.8 ng/J (0.06 lb/MM Btu)
which is less than either existing or proposed NSPS limitations
for utility boilers.
3) Particulate removal data Indicate that, on the average, scrubber
efficiency was 84% during the test period. However, based on
particulate catches essentially free of coal ash contamination, the
scrubber efficiency was approximately 75% for oil firing particulates
4) Organic emissions determined by FID analysis were generally less than
5 ng/J (0.01 lb/MM Btu) and appear to be composed primarily of C-j to
Cg hydrocarbons and organlcs heavier than C^g. However, gas chroma-
tograph and gravimetric data Indicate that FID values may be low by a
factor of 2 to 3. Approximately 88 and 83% of the C^ to C^g and
higher than C^g organlcs, respectively, were removed by the scrubber
5) The organic compounds Identified in the gas samples were generally
not representative of combustion-generated organic materials, but
were compounds associated with materials used in the sampling equip-
ment and 1n various analytical procedures. This again confirms the
low level of organic emissions.
6) When emissions are uncontrolled, over 90% of the sulfur In the fuel
feed 1s emitted as S02» less than 1% as S0g, and 1.5% as S0^~.
6-52
-------
7) S02 1s efficiently removed by the scrubber (97 to 98% efficiency).
The S03 removal efficiency (28 to 29%) suggests that S03 is present
as fine liquid aerosols. SO^* 1s about 60% removed by the
scrubber, and so 1s probably associated with the larger particulates.
8) Of the 22 major trace elements analyzed 1n the flue gas stream, 11
exceeded their MATE values at the scrubber Inlet while only 5 exceeded
MATE values at the scrubber outlet. These 5 elements are arsenic,
cadmium, chromium, nickel and vanadium. With the exception of
chromium, elements exceeding their MATE values at the scrubber outlet
were removed from the flue gas stream with efficiencies lower than
the overall average removal efficiency of 87%.
9
9) Beryllium emissions were 0.001 mg/m after scrubbing; this corresponds
to half the MATE value for this element. At this emission concen-
tration, the National Standard for Hazardous A1r Pollutants limitation
of 10 grams beryllium per day would only be exceeded by boilers of
100 MW capacity or greater.
10) The combined wastewater stream from the boiler operation may not
pose an environmental hazard in terms of organic materials since the
discharge concentrations of organics are all well below their MATE
values. A similar conclusion may be drawn with respect to inorganic
materials since inorganics, with the exception of nickel and copper,
did not exceed their MATE values for liquid streams. Owing to
uncertainty associated with SSMS analysis, nickel and copper may
exceed their MATE values although this 1s not necessarily the case.
11) Polycycllc organic material (POM) was not found 1n the scrubber Inlet
3
or outlet streams at detection limits of 0.3 yg/m . MATE values for
most POMs are greater than this detection limit. However, since the
MATE values for at least two POM compounds - benzo(a)pyrene and
d1benz(a,h)anthracene - are less than 0.3 ug/m , additional GC/MS
analysis at higher sensitivity would be required to conclusively pre-
clude the presence of all POMs at MATE levels.
6-53
-------
Mass balance closure for 10 of the 20 trace elements analyzed
is between 50 and 136 percent. Poorer mass balance closure
was obtained for the remainder of trace elements due to the
extremely low element concentrations and/or contamination of
the scrubber recycle solution by coal firing components.
The scrubber cake produced contains about 1% oil fly ash.
With the exception of antimony, boron, molybdenum and zinc,
trace element concentrations in the scrubber cake exceeded
their health based MATE values. All ecology based MATE values
were exceeded by the trace element concentrations. Because the
trace elements may leach from the disposed scrubber cake,
these solid wastes must be disposed of in specially designed
landfills.
6-54
-------
REFERENCES FOR SECTION 6
1. Magee, E.M., H.J. Hall and G.M. Varga, Jr. Potential Pollutants In
Fossil Fuels. Report prepared by ESSO Research and Engineering Co.
for EPA under Contract No. 68-02-0029. June 1973.
2. R.A. Woodle and W.B. Chandler, Jr. "Mechanism of Occurrence of Metals
in Petroleum Distillates." Industrial and Engineering Chemistry 44:
2591, November 1952.
3. R.L. Bennett and K.T. Knapp. "Particulate Sulfur and Trace Metal
Emissions From Oil-Fired Power Plants." Presented at AIChE meeting.
June 1977.
4. Steam-Electric Plant Air and Water Quality Control Data. Federal
Power Commission. March 1975.
5. Cato, G.A., L.J. Muzio and D.E. Shore. Field Testing: Application
of Combustion Modifications to Control Pollutant Emissions From
Industrial Boilers—Phase II. Report prepared by KVB for EPA under
Contract No. 68-02-1704. April 1976.
6. Cleland, J.6. and G.L. Kingsbury. Multimedia Environmental Goals for
Environmental Assessment, Vols. I and II. EPA-600/7-77-136a, November
1977.
7. Bierly, E.W. and E.W. Hewson. Some Restrictive Meteorological Condi-
tions to be Considered in the Design of Stacks. J. Appl. Metero.,
1, 3 pages 383-390, 1962.
8. Turner, D.B. Workbook of Atmospheric Dispersion Estimates. U.S. Dept
of HEW, 1969.
6-55
-------
6-56
-------
APPENDIX A
SIMPLIFIED AIR QUALITY MODEL
-------
APPENDIX A
SIMPLIFIED AIR QUALITY MODEL
Simplified air quality models were used to determine relative air
quality resulting from uncontrolled and controlled emissions. The ambient
air quality values are approximate only. The emphasis should be placed on
the relative values for each case as opposed to their absolute values.
Worst case weather conditions and typical weather conditions were
considered. The worst case was assumed to be plume trapping. An equation
proposed by Blerly and Henson was used with the following assumptions:
inversion height 100 meters, wind speed 1.0 meter/second, D class stability
(neutrally stable) 1n the mixing layer and one-hour averaging time. The
typical case was assumed to correspond to the standard Gaussian convectlve
diffusion equation. The following conditions were used: wind speed 4,0
meters/second, D class stability, and one-hour averaging time. These
represent conditions that could reasonably be expected to occur almost
anywhere in the country but are not specific to the area of the plant.
Typical does not mean average. It was assumed that all species were
Inert. No photochemical reactions were considered.
There are several meteorological conditions which can produce high
ground level pollutant concentrations. These conditions can result 1n
plume coning, looping, fumigation, and trapping, all of which can cause
high ambient concentrations. In the case of coning, high levels occur
along the plume centerline. Looping causes high ground level concentra-
tions at points where the plume impacts the ground. Fumigation causes
high ground level concentrations which are generally lower than those from
plume trapping.
For this study 1t was assumed that plume trapping constituted the
worst case. That 1s, the case that would result 1n the highest ground
level concentrations.
Plume trapping occurs when the plume 1s trapped between the ground
surface and a stable layer aloft. Blerly and Hewson [1], have suggested
the use of the following equation to account for the multiple eddy
reflections from both the ground and the stable layer.
A-l
-------
X (x, 0> z; H) =. o„..?
2uuo a 1eXP
y z 1
-1/2
(^f] + exP [-i/2(^Ji)2]
+ Z [exp - ]/2(Z'H-2NL) + exp -1/2 (ZfH'2NL) 2 (1)
N=1 L az az
+ exp -1/2 (¦
Z-H+2NL
) + exp -1/2
(Z+H+2NL)2j
Where: X (x,y,z;H) = Concentration at point (x,y,z) assuming an
effective stack height at H;
H = Effective stack height, meters;
Q = Pollutant emission rate, kg/hr;
jU = Mean wind speed, M/S
O" = Concentration distribution within the plume
1n the horizontal (cry) and vertical (trz)
directions, unltless;
z = Height above the ground, meters;
J ¦ Wind speed class Index, unltless
N = Wind speed class Index, unltless;
L = Height of the stable layer, meters.
At ground level (z = 0) and at the plume center line (y * 0)
equation (1) reduces to:
X (x,0,0; H) = —^
l-Htf ] * §
exp -0.5
'H+2JL
+ exp -0.5
'H -2JL1
1
(2)
A-2
-------
For the typical case, ground level concentrations were calculated
using the standard Gaussian solution to the convectlve diffusion equation.
This was obtained by neglecting the "Inversion height" terms 1n equation
(1), that 1s by setting z=0. The solution becomes:
In equation 3, H 1s defined by:
H = Hs + AH
Where H = physical height of the stack and AH « plume rise, both
expressed In meters. There are more than 30 plume-rise formulas 1n the
literature. All of which require empirical determination of one or more
constants. For the purpose of this study, the Brlggs plume rise formula
was chosen to calculate the final plume rise 1n stable conditions.
X (x, 0, 0; H) 9 exp
(3)
Where: |i ¦ Wind speed, m/s;
s = Stability parameter, unltless;
F « Buoyancy flux.
The stability parameter, s, 1s defined as:
2
Where: g - Gravitational constant, m/s ;
T = Absolute ambient air temperature, °K.
Where: 9 0/ aZ ¦ (8T/0Z) + 9.8°C/km, the potential temperature gradient.
The buoyancy flux, F, 1s defined as:
-------
Where: AT - Stack temperature minus the ambient air temperature, K° ;
Ts = Stack temperature, °K ;
g = Gravitational constant, m/s^ ;
w = Stack exit velocity, m/s ;
r = Inside radius of the stack, m.
Table 1 shows the input data used in the plume rise calculations.
TABLE 1. INPUT DATA USED IN PLUME HEIGHT CALCULATIONS
Coal
Oil
Stack Temperature (°C)
51.7
55.6
Ambient Temperature (°C)
20.0
20.0
Stack Velocity (m/s)
16.44
16.44
Stack Area (m^)
0.245
0.245
Stack Height (m)
10.0
10.0
Wind Speed (m/s)
4.0
4.0
With the above input data effective stack heights of 56.0 m and 60.0m
were calculated for coal and oil firing, respectively. The difference
was due entirely to variations 1n the flue gas temperature.
As shown in both equations (2) and (3) concentrations vary Inversely
with wind speed. As y approaches zero, ground level concentrations become
infinite. This means that at zero wind speed the Gaussian solution is no
longer applicable. (The lower wind limit is about 1.0 m/s.) For plume
trapping to occur and for the ground layer to become well mixed, the
most stable ground layer must be chosen so that y and z 1n equation (2)
remains as small as possible. D stability 1s such a case.
For the typical case, D stability occurs most often (see Table 2).
The table shows D stability to be the most common of those listed.
A-4
-------
TABLE 2. ANNUAL PERCENT FREQUENCY OF PASQUILL STABILITY
CATEGORIES FOR ALL WIND DIRECTIONS AND SPEEDS, [4]
Pasquill Stability Category
A
B
C
D
E
F
Birmingham, Alabama
1
7
12
44
36*
Tucson, Arizona
2
10
14
33
41*
Los Angeles, California
0
4
15
48
13
19
Miami, Florida
0
5
14
42
39*
Chicago, Illinois
1
5
11
55
12
17
New York, New York
0
3
10
67
13
6
Philadelphia, Pennsylvania
0
5
11
51
14
18
indicates E and F categories combined.
Presented 1n Table 3 are the wind speed distribution and D stability
for Baltimore, Maryland, for the period June - August, 1968. Winds 1n the
range 3.6 to 5.2 m/s (7 to 10 knots) occur most often.
-------
TABLE 3. FREQUENCY (PERCENT) OF PASQUILL STABILITY CATEGORY D;
BALTIMORE, MARYLAND; BASED ON ALL REGULAR 3-HOURLY
WEATHER OBSERVATIONS, JUNE - AUGUST, 1968; 1 KNOT =
0.515 M/S (HOLZWORTH, 1974)
Direction
Wind Speed (knots)
0-3
4-6
7-10
11-16
17-21
>21
Total
N
0.2
0.3
0.3
0.0
0.0
0.0
0.7
NNE
*
0.5
0.1
0.0
0.0
0.0
0.7
NE
~
0.3
0.1
0.0
0.0
0.0
0.4
ENE
~
0.4
0.1
0.0
0.0
0.0
0.6
E
0.8
1.5
0.1
0.0
0.0
2.5
ESE
~
0.4
0.4
0.0
0.0
0.0
0.8
SE
0.3
0.3
0.5
0.0
0.0
0.0
1.1
SSE
0.1
1.1
0.8
0.0
0.0
0.0
2.0
S
0.2
0.9
1.4
0.4
0.0
0.0
2.9
SSW
0.1
0.1
0.7
0.0
0.0
0.0
1.0
sw
~
0.4
0.4
0.1
0.0
0.0
1.0
wsw
0.2
0.8
0.4
0.4
0.0
0.0
1.8
w
0.1
1.1
1.6
0.9
0.0
0.0
3.7
WNW
~
0.3
0.5
0.8
0.0
0.0
1.6
NW
~
0.1
1.1
0.3
0.1
0.0
1.6
NNW
0.0
0.0
0.4
0.3
0.0
0.0
0.7
Total frequency of D stability = 23.1%
Frequency of calms distributed with D stability = 0.4%
~Indicates <0.05,
A-6
-------
REFERENCES FOR APPENDIX A
1. Bierly, E.W., and E. W. Hewson. Some Restrictive Meteorological
Conditions to be Considered in the Design of Stacks, J. Appl.
Meteor, 1,3, 383-390, 1962.
2. Turner, D.B. Workbook of Atmospheric Dispersion Estimates, U.S.
Dept. of Health, Eductation and Welfare, 1969.
3. Slade, D.H. Meteorology and Atomic Energy, U.S. Atomic Energy
Commission, 1968.
4. Holzworth, G.C. Chronological Aspects of the Composition and
Pollution of the Atmosphere, World Meteorological Organization,
Technical Note No. 139.
A-7
-------
A-8
-------
APPENDIX B
ORGANIC ANALYSIS METHODS
-------
APPENDIX B
ORGANIC ANALYSIS METHODS
This section summarizes the philosophy and methods used for Level 2
organic analysis of the Industrial samples. The purpose here 1s to present
a summary of the results which have been previously reported and to present
the methods and approaches used for the analysis. The methods are explained
and the figures show how various samples from the Level 2 organic train as
well as process and gas bag samples were analyzed using primarily GC/MS.
LEVEL 2 ORGANIC ANALYSIS PLAN
The Level 2 organic analyses have been designed to Identify and quan-
tify organic species. The analysis plan provides for the following type of
Information, listed according to the priority levels that have been
assigned.
t Identification of compounds present at significant levels
• Identification of all compounds
• Quantification of all compounds
To achieve this prioritized listing, the steps outlined In the Level 2
priority flow diagram (Figure 1} have been followed. Primary analytical
decisions" are based on Level 1 analysis data and established minimum levels
for compounds on the MEG and priority pollutant lists. Subsequent analy-
tical decisions are based on cost effectiveness and a prescreenlng process.
The prescreenlng 1s accomplished by analyzing all samples to the point of
LC separation as shown fn Figure 2. Once the preliminary analysis Is
complete, a determination Is made as to whether 1t 1s cost effective to
continue the analysis. At a minimum, volatile organic matter (C£ - C-|g)
and polynuclear organic material (POM) were Identified and quantified 1n
all samples. The samples were paired for analysis as pre- and post-scrubber
couples. Any determination made on one of the pair was automatically made
on the other. The purpose of this effort was to maximize the amount of
Information related to scrubber efficiency and scrubber operation on
various organic compounds. The central technique for Level 2 organic
analysis was gas chromatography/mass spectrometry. The procedures
outlined are the starting point for the overall analysis. Information
B-l
-------
gained in the initial phase was to be used to direct subsequent analysis if
it was found necessary to continue beyond the prescreening phase.
Level 2 Analyses Performed
All samples including the process samples were analyzed through the
prescreening phase using the procedures discussed. It was determined that
further sample work up would not result in additional information based on
the lack of organic compounds detected. The GC/MS was used on all samples
for the prescreening process. The GC columns used together with sample
preparation steps taken are presented in this report.
Conclusions
Table 1 summarizes the findings of the Level 2 organic analyses.
Table 2 summarizes the samples analyzed by GC/MS. The lack of significant
levels of organic materials suggests a very clean source as related to organic
emissions. No polynuclear aromatic hydrocarbons were detected. In fact, no
organic species which could be directly associated with combustion could be
identified. Calculations of sample volume and sample concentration have been
performed and confirm the limit of detectabi1ity of the GC/MS techniques as
related to the effluent streams. There is no reason at this time to suspect
anything but a very clean combustion source with extremely low level organic
emissions.
LEVEL 2 ORGANIC ANALYSIS PHILOSOPHY
This Level 2 organic analysis plan is based on Level 1 analysis data
and is intended for use on process, gas bag, and sampling train collected
samples. The plan assumes that Level 1 analysis has been completed and the
information from Level 1 analysis is available. The techniques discussed
are implemented by a skilled mass spectrometrist, since at several points in
the analysis, judgements and even modifications are necessary depending on
sample source or what compounds are identified during the course of the
analysis. B_2
-------
START
LEVEL %.
DOES LEVEL 1 DATA
INDICATE ANY POTEN.
TIALLY HAZARDOUS
MATERIAL IN EXCESS
OF ITS ESTABLISHED
MINIMUM?
YES
IDENTIFY C
USING ORG
ANALYSIS f
OMPOUNDS
ANIC
lOWCHART
ESTIMATE QUANTITY
OF EACH COMPOUND
IS IT COST EFFECTIVE
OR A PROGRAM REQUIRE
MENT TO ANALYZE THIS
SAMPLE FURTHER?
IS THERE SUFFICIENT
MATERIAL FOR
ANALYSIS?
DO ANY POTENTIALLY
HAZARDOUS COMPOUNOS
EXCEED OR APPROXIMATE
MINIMUM ESTABLISHED
LEVELS?
QUANTIFY POTENTIALLY
HAZARDOUS COMPOUNDS
QUANTIFY THOSE
COMPOUNDS FOR
WHICH IT IS COST
EFFECTIVE TO DO SO
IS FURTHER
QUANTITATION
COST EFFECTIVE?
^ REPORT
LEVEL i
ANALYSIS
COMPLETE
Figure 1. Logic flow chart for Level 2 organic analysis.
B-3
-------
UVfl J
I IAIN SAMMS
GC/MS
OV-17
Ft
m
<
4=»
Data sivrvc. ^
wiu rtscxvio > w
AND EASY to
iNKupwr?
cc/Vi
cw-i»
AllOOOI
rros; w>jh
SAMPlf
rllTli
s*«ni
J* )
SECOND
llMf
AtOUNO
?
fKl«ACl
M.CI
CONaNlMi
xico
All COMPOUNDS
iDENtlFIED
ixltACT
M.CI
XaO-1
MS IN
fUCnONAH
IV LC
ac/M
ov-i;
• /N (KIXACT
M > U
m,ci
CONWNVOt
gc/ms (D
ov-u
ffA t
KOH-CAMO-
WAX ?0M
GC/»/!
JIKAX GC
CM
HjKJ4/ca*»Owa*
Aiiouor
DATA 51 Mm,
WIU MSOIVEO
AND EASY tO
iNittmi i
acio exiract
Prt < I
M.CJ
GC/MS
FFA» OH
sr-ju-fs
ocovinzc
GC/MJ
SECOND
tIMi
AROUND
7
OV-W
All COMFOUNDS
IDENtlFIED
0 a MAY U USEFU. tO CASE
SttCtKAl INTlimtATION
(5) NU«OSOAM!Nf S A«£ (XKCtlD ONIV
At ICWCCNCfNjtATIONS, if PMSENT
SICIAl MCAIuriON Will I! IfOUIItO
fOUOWED IY GC/MS ANALYSIS USING
CaMOwaX ;o M AND SELECTED ION
MONIIOtlNG (SIM)
:ONCENIUtf
<100
Figure 2. Analysis of Level 2 samples from glass sampling train.
-------
S LWAIUSAMW
CD
f
Uf
©
OMict Aouout iHjcctioN gc/ms
ANAlYilS fCM MtCM CONdNUAIION
AND HIGHLY VCXAIIU MAICUALS
knak cc faumo
fCX»« COMPOUNDS NON-rouui COMTOUNOS
- MOST CAT! GOUCS •
StNUTlVIIY 100 NG/rl INJCCTID
-0
- HIGH CO'CINHAIION
«o»l *N0 CHI'* VOLATlU
' MMfG COMKMNOS
I0€NTIH€»
m*a a no im» cqncshtiahon
FOt WAUIIMMISIIU
VCXAIIU IHMVO OIGANICS
©
GC/MS
o j% camowax tseo
ONCAMOTAKC
CATCGOtlCS
I. I, U. U (LOW KXUtt ONLY)
WATEi IMM4SSIIU
MMCG COMPOUNDS
lOfNimco
KtUN SAMfti USING
CAhLlAlY COLUMN OC/MS
MCTAi O* GLASS COLUMN
OV-17 O* CAWOwax 10 M
SAME DATA OiTAiNCD AS
kfou loan iMnovto
CHIOMATOG*Aft«C ((SOLUTION
WAttl IMMISSIIU
MMCCOMfOUNOS
IMNTlfUD
¦Ait 'mut«al fXTiACiiON
m >11
GC'MS®®
OV-17
CAKGOMiS®
CMOMASOU H»
1
fl
—
i.i. j.*. a.*). 7,»
n,n.u.as)
may it uscrui ro*
AMINfS If rUUNT
CATIGOUtS 10, u
0 CMAVMUWfU IOIA4
Maw INIIIWIATION
(D m MjK>4 cuts wooi nuci
FO* COLUMN KWMNAtOH WMfN
ubNG columns roi Acme snots
(3> ( | INOtCATiS CATfGOMS WMCM
MA* f»«» IN KM r(ACTIONS 0U
to acuity (H.)
9 ML COLUMNS GLASS UNUSS
OMtCMM HOWI.
ACOtPr f xiiaci
M <)
CtXTtACT WITH
M. dj
CONaNtlAH
GC/MS©®
TINAX GC _
oi
-------
CP
t
Ol
NO
DATA SIMPLE, >
WELL RESOLVED
AND EASY TO
V INTERPRET?/
DATA SIMPLE, ^
WELL RESOLVED
AND EASY TO >
\ INTERPRET/
NO
YES
YES
y° /SECOND TIME
\ AROUND ?
NO/ SECOND TIME
\ AROUND ?
YES
YES
ALL COMPOUNDS
IDENTIFIED
CONCENTRATE
X100
BAG
SAMPLE
05 1)
CONCENTRATE
X100
GC/MS
ALIQUOT ON
PORAPAK Q
GC/MS ALIQUOT
ON MORE
APPROPRIATE
COLUMN
Figure 4.
Lecel 2 analysis of gas bag samples.
-------
TABLE 1. ORGANIC ANALYSIS RESULTS
Oil
Coal
!-3
COMPOUND NAME
FRACTION
201 -2
201
-3
202-2
202
IN
OUT
IN
OUT
IN
OUT
IN
OUT
Propionaldehyde
GB
170
380
54
200
—
142
-
Nitromethane
GB
-
¦ -
-
-
-
-
63
-
Ethyl-n-butyl ether
GB
2000
1500
910
1500
1700
-
1000
-
Ethyl acetate
GB
-
-
-
-
-
-
-
2300
Hydrocarbon (C6H14)
GB
-
-
380
-
-
-
-
-
™ Chloropropanol
PR
-
-
-
3.9
-
-
-
-
4-methyl-3-pentene-2-one* PR
-
-
-
-
-
570
350
28
Unidentified alcohol
PR
-
4.0
-
-
-
-
-
-
Octanol
PR
-
-
-
-
-
-
45
-
Ketone (MM 138)*
PR
-
-
-
-
-
-
20
-
Ketone (HW 140)
PR
3.9
-
6.9
-
14
-
-
-
Phthallc anhydride
PR
-
-
-
-
-
-
29
-
Aaiyl benzoate
PR
—
-
-
-
-
1.6
-
-
Glycerol triacetate
PF
-
-
-
-
-
-
42
-
Methyl sub aromatic
XR
1.7
-
-
-
-
-
-
-
GB - Grab bag samples
PR - Probe rinse
-------
TABLE 2. PROCESS SAMPLES ANALYZED BY GC/MS
CO
I
OO
SAMPLE NUMBER
201-2-1-S-KD
201-3-1-S-KD
202-2-1-S-KD
202-3-1-S-KJ
202-2-1-S
202-3-1-S
201-2-1-B-S-KD
201-3-1-B-S-KD
202-2-1-B-S-KD
202-3-1-B-S-KD
201-2-1B-3-S-KD
201-2-1-B-S
201-2-1-B-S
SAMPLE DESCRIPTION
Coal-fired, Test #2, concentrated field extraction of 4 liters
of scrubber water.
Coal-fired, Test #3, concentrated field extraction of 4 liters
of scrubber v/ater.
Oil-fired, Test #2, concentrated field extraction of 4 liters
of scrubber water.
Oil-fired, Test #3, concentrated field extraction of 4 liters
of scrubber v;ater.
Oil-fired, Test K2, unconcentrated field extraction of 4 liters
of scrubber water.
Oil-fired, Test #3, unconcentrated field extraction of 4 liters
of scrubber water.
Coal-fired, Test #2, concentrated lab base/neutral extraction
of 2 liters of field extracted scrubber water.
Coal-fired, Test #3, concentrated lab base/neutral extraction
of 2 liters of field extracted scrubber water.
Oil-fired, Test §2, concentrated lab base/neutral extraction
of 2 liters of field extracted scrubber water.
Oil-fired, Test #3, concentrated lab base/neutral extraction
of 2 liters of field extracted scrubber water.
Coal-fired, Test #2, concentrated lab base/neutral extraction blank.
Coal-fired, Test #2, unconcentrated lab base/neutral extraction
of 4 liters of field extracted scrubber water.
Coal-fired, Test 83, unconcentrated lab base/neutral extraction
-------
TABLE 2. (Continued)
09
I
to
202-2-1-B-S
202-3-1-B-S
201-2-1B-B-S
201-2-1 -A-S-KD
201-3-1-A-S-KD
202-2-1-A-S-KD
202-3-1-A-S-KD
201-2-1-1B-A-S-KD
201-2-1-A-S
201-3-1-A-S
202-2-1-A-S
202-3-1-A-S
201-2-1B-A-S
201-2-SC-S-KD
201-3-SC-S-K0
202-2-SC-S-KD
202-3-SC-S-KD
Oil-fired, Test #2, unconcentrated lab base/neutral extraction
of 2 liters of field extracted scrubber water.
Oil-fired, Test #3, unconcentrated lab base/neutral extraction
of 2 liters of field extracted scrubber water.
Coal-fired, Test #2, unconcentrated lab base/neutral extraction blank.
Coal-fired, Test #2, concentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Coal-fired, Test #3, concentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Oil-fired, Test #2, concentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Oil-fired, Test S3, concentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Coal-fired, Test #2, concentrated lab acid extraction blank.
Coal-fired, Test #2, unconcentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Coal-fired, Test #3, unconcentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Oil-fired, Test #2, unconcentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Coal-fired, Test #2, unconcentrated lab acid extraction of 2 liters
of field extracted scrubber water.
Coal-fired, Test #2, unconcentrated lab acid extraction blank.
Coal-fired, Test #2, concentrated lab extraction of scrubber cake.
Coal-fired, Test #3, concentrated lab extraction of scrubber cake.
Oil-fired, Test #2, concentrated lab extraction of scrubber cake.
Oil-fired, Test #3, concentrated lab extraction of scrubber cake.
-------
TABLE 2. (Continued)
^01-3-FA-S-XD
Coal-fired. Test
#3
201-2-SCB-S-KD
Coal-fired, Test
#2
201-2-SC-S
Coal-fired, Test
n
201-3-SC-S
Coal-fired, Test
$3
202-2-SC-S
Oil-fired, Test
#2
202-3-SC-S
Oil-fired, Test
#3
201-3-FA-S
Coal-fired, Test
#3
201-2-SCB-S
Coal-fired, Test
#2
OS
I
o
concentrated lab extraction of fly ash.
concentrated lab extraction blank,
unconcentrated lab extraction of scrubber cake,
unconcentrated lab extraction of scrubber cake,
unconcentrated lab extraction of scrubber cake,
unconcentrated lab extraction of scrubber cake,
unconcentrated lab extraction of fly ash.
unconcentrated lab extraction blank.
-------
The combination of gas chromatography and mass spectrometry is the
central technique to this analysis plan. 6C/MS combines the separation power
of the gas chromatograph with the unexcelled identification potential of the
mass spectrometer. The incorporation of a computer based data handling system
with the GC/MS provides the most powerful compound identification technique
available to the analyst. The technique is highly cost effective but requires
an experienced spectrometrist to apply the technique and analyze the data
generated. Judgments as to sample size, depending on Instrument sensitivity,,
and mass range to be scanned, depending on Instrument resolution, as well as
selection of an alternate GC column for a specific sample, are at the dis-
cretion of the analyst. General direction can be given, however, a total
analysis requires some modification of the procedures.
The most cost effective Level 2 analysis scheme is specific analysis
based on compound category data obtained from Level 1. This information
would provide data for GC column selection and would generally simplify the
overall analysis. The analysis scheme outlined here is for all categories
of compounds with the exception of those compounds which are volatile and
are analyzed by the field GC technique,and those compounds which are reactive
and chemically modified by sampling or standing.
The Level 2 analysis plan incorporates wet chemical separations,
including sample extractions and liquid chromatography, and instrumental
analysis using primarily GC/MS. The other techniques discussed.have been
applied in special cases but require further research into their appli-
cation. These include high resolution mass spectrometry (HRMS), Chemical
ionization mass spectrometry (CIMS), gas chromatography with selective
detectors, and capillary column GC/MS. The proposed analysis plan is pat-
terned after Level 1 which will provide information to ease the total sample
burden imposed by the identification of a large number of organic compounds.
A variety of specific GC columns is described together with appropriate
conditions for their use on specific categories of compounds. If a category
is known to be absent in a specific sample, based on information from Level
1, it is expected that this knowledge will be used to modify the analysis.
Typical sensitivities for various analysis steps are given as a part of the
specific method discussions. It is important that the analyst implementing
the Level 2 plan, have a working knowledge of Level 1 organic analysis.
B-11
-------
HARDWARE REQUIREMENTS AND OPTIONS FOR LEVEL 2 ANALYSIS
The primary tool for Level 2 analysis is a high sensitivity GC/MS
instrument. A discussion of corollary GC/MS techniques, useful during the
course of analysis, are also given together with their appropriate application,
GC/MS
In order to apply GC/MS and obtain reliable data it is necessary to
have a spectrometer which is capable of high speed scanning (i.e. recording
a full mass spectrum in 3 seconds or less) with resolution that will allow
separation nominal mass peaks to at least mass 600. The gas chromatograph
should be capable of using glass columns since many materials to be analyzed
are sensitive to metal surfaces. The mass spectrometer should be capable
of chemical ionization with a variety of reagent gasses such as methane or
isobutane. The ability to use capillary columns is useful in many of the
GC/MS analysis. The incorporation of a computer based data handling system
eases the labor involved in acquiring mass spectral data, and reduces the
time for data reduction and interpretation. The computer does not eliminate
the need for an experienced mass spectrometrist, it merely provides a more
cost effective means of handling large volumes of mass spectral data.
Chemical Ionization (CI) Mass Spectrometry
Normal mass spectrometry is done by bombarding the sample with
electrons at an energy level of 7Q electron volts (eV). The ionization
process produces a spectrum which contains characteristic fragment Ions
from the molecule under study. In many cases, a molecular ion is produced
(I.e., the ion representative of molecular weight) and its identification
is unambiguous, however, in some cases, no molecular 1on is produced or 1t
is present at such a low level that it cannot be Identified. The most
important peak in any mass spectrum 1s the molecular ion since a knowledge
of molecular weight reduces the total number of organic compound possi-
bilities substantially. Electron Ionization does provide a great deal of
compound structure Information, but when the molecular ion is absent much
information 1s lost, making spectral interpretation difficult or
impossible.
B-12
-------
Chemical ionization incorporates a reagent gas to perform the ioniza-
tion process. The use of methane or isobutane for the chemical ionization
process are most common. When these reagent gases are used, the energy of
ionization is reduced from 70 eV to about 7 eV. The result is ionization of
a sample without sufficient excess energy to cause significant fragmentation
and in most cases the "pseudo" molecular ion dominates the spectrum yielding
molecular weight information. The CI process involves a transfer of a proton
from the reagent gas to the sample when ionization occurs. The resulting
spectrum is a "pseudo" molecular ion at 1 mass unit higher than the molecular
weight of the compound. Chemical ionization should always be used in con-
junction with electron ionization for spectral interpretation. As is true
with most analytical techniques, chemical ionization is not without its dif-
ficulties. The ionization of many materials such as alcohols, causes a
protonation of the hydroxy1 group followed by a loss of water from the
"pseudo" molecular ion by a thermal process. An example of this type of
ionization is given below.
Cli3
HC-OH + CHc+ (Reagent Gas)
/ 0
ch3
HC-OH
The loss of water from the "pseudo" molecular ion is primarily dependent
on source temperature, and is increased with higher temperatures. This
fragmentation process may not take place when electron Ionization is used
and in many cases causes confusion in the determination of molecular weight.
Similar occurrances take place when primary amines are being studied, showing
a loss of ammonia from the "pseudo" molecular ion, and to a lesser extend
acids, ethers, esters, and halogenated compounds. Hydrocarbon samples are
typically not sensitive to chemical ionization. This is especially true of
straight chain hydrocarbons. Under CI conditions, straight chain hydrocarbons
often show a loss of 1 from the molecular 1on rather than an addition, together
with a significant reduction in overall sensitivity. Materials which con-
tain heteroatoms such as nitrogen, oxygen, sulfur, etc. generally show an
B-13
-------
increase in sensitivity relative to their electron ionization spectra. This
variation in sensitivity is useful in identifying heteroatom structures in
complex hydrocarbon samples using chemical ionization. Other reagent gases
are available, (e.g. ammonia, nitrous oxide, and hydrogen), however, less
work has been done with these reagents gases.
High Resolution Mass Spectrometry (HRMS)
The techniques discussed to this point require that the compound of
interest be amenable to gas chromatography. Many materials, of course,
cannot be chromatographed and therefore do not lend themselves to GC/MS.
High resolution mass spectrometry is a technique by which one can analyze
low volatility residual materials. Total compound identification may
not be possible in all cases depending on mixture complexity, however,
functional groups and heteroatoms can generally be identified unambiguously.
The technique as discussed employs the direct insertion probe which is used
to introduce the sample into the ion source of the mass spectrometer. The
use of a high resolution data system together with the high resolution mass
spectrometer is important in obtaining useful data in a reasonable amount of
time. Full spectra should be recorded and the computer used to reduce the
data to element maps for selected mass peaks. The element maps will give
the elemental composition for mass peaks and an experienced mass spectro-
metrist can use this information to determine the compound types in a
sample. The sophistication of a high resolution mass spectrometer is much
greater than GC/MS and the sophistication of the operator must also be
greater. The technique should be limited and applied only when deemed
necessary since the availability of such equipment is not wide spread.
Infrared Spectroscopy
Infrared spectroscopy is very useful in determining compound function
ality. The technique is not applied directly in this Level 2 plan since it
is assumed that IR spectra have previously been recorded for Level 1.
It is also assumed that this information is available and is used by
the analyst to select appropriate GC columns and to insure that he has
analyzed all materials by this Level 2 plan which were present in Level 1.
3-14
-------
Capillary GC/MS
In some cases, packed columns do riot provide the chromatographic
resolution necessary to obtain good mass spectral data. In these cases the
use of high resolution capillary columns is recommended. This is especially
true in the direct analysis of the extracts prior to concentration, or liquid
chromatographic fractionation. There are a wide variety of capillary
columns available to the analyst. These columns include standard open tublar
wall coated columns, SCOT columns, and micro packed columns. For the begin-
ner, the use of a SCOT column is recommended, since it is more tolerant of
temperature and sample size, while providing increased resolution over its
packed column counterpart. The liquid phase chosen for a capillary column is
generally based on information obtained using packed columns. A wide variety
of liquid phases are available, however, due to their expense, only a selected
few columns are expected to be used routinely. For general application in
Level 2 analysis, it is recommended that a OV-17 SCOT column and a Carbowax
20M SCOT column, which are between 50 and 100 ft. in length be used. These
two columns will satisfy 90% of the requirements for capillary column GC.
SAMPLE PREPARATION AND EXTRATION PROCEDURES
The preparation and extraction procedures described in this section
are similar to those used in Level 1 analysis. For those analysts familiar
with Level 1 analysis, the only modification is in the extraction of the
condensate of the XAD-2 sorbent trap.
Probe Wash, Cyclones, and Filter from Level 2 Train Samples
The probe wash, cyclones, and filter samples are prepared for analysis
by extraction with methylene chloride. The extractions should be conducted
using a Soxhlet apparatus for 24 hours. The Soxhlet cup should have been
previously extracted to remove contamination which would lead to erroneous
results following the established procedures outlined for Level 1 analysis.
XAD-2 Sorbent Trap
The XAD-2 resin from the sorbent trap should be extracted with
methylene chloride using a large Soxhlet apparatus. The procedure for this
extraction as well, as resin preparation are,identical to Level 1 procedures.
B-15
-------
Extraction of the Condensate
The condensate from the sorbent trap should first be extracted with
methylene chloride after the pH has been adjusted to 11 or greater with 6N
sodium hydroxide. This base/neutral extract should be set aside for subse-
quent analysis. The solution pH is then adjusted to less than 2 using 6N
hydrochloric acid and the extraction with methylene chloride repeated. This
division of the condensate sample into two extracts may eliminate the need
for an LC separation step, making the overall analysis less expensive and time
consuming.
PRESCREENING ANALYSIS OF THE EXTRACTS FOR ORGANIC COMPOUNDS
Concentration of extracts prior to analysis causes the loss of most
materials with boiling points below about Cg (220°C). To obtain data on
low boiling extracted compounds from the sampling train samples, GC/MS analy-
sis is run on the sample prior to concentration. A 5 cc aliquot of the
extract should be saved for this analysis. One GC/MS run on each sample is
made using a general purpose column. If specific classes of compounds are
found to be present from the Level 1 LC fraction data for a given extract,
repeat analysis of the uncondensed extract may be necessary to determine if
more volatile materials in the same compound class are present. Column
selection and the rerun of a sample should be based on the categories identi-
fied from the Level 1 analysis of the LC fractions.
Prescreeninq GC/MS Analysis of the Probe Wash, Cyclones, and Filter Extrartg
The GC/MS analysis of the probe wash, cyclones and filter extracts
should be run using the chromatographic conditions givpn below.
Liquid phase-OV-17
Liquid loading-3%
Solid Support-Chromasorb W, AW, DMCS
Column type-glass
Column size - 2 mm ID x 2 meters long
Temperature program-50°C for 5 minutes,50-280°C at 6°C per minute
Hold at maximum until all peaks elute.
Injector temperature-280°C
Detector and transfer line to separator temperature-280°C
Flow rate of helium-30cc per minute
B-16
-------
Sample size - 5 ul
Sensitivity 3 ug/m3
This set of chromatographic conditions is general and is designed to
separate and quantify most organic compounds suspected to be present in a
sample. Specific categories and concentrations determined from the LC
fractionation step may dictate the use of an alternate column and/or modifi-
cation of the conditions for this column. This judgement can be made only
by the analyst based on his ability to interpret the GC/MS data. The com-
plexity of this extract is expected to vary widely depending on the source.
When a sample 1s highly complex, the use of chemical ionization mass
spectrometry is recommended. Chemical Ionization may aid 1n the interpre-
tation of individual mass spectra,especially if no molecular ion is present
1n the EI spectrum.
When problems of chromatographic resolution are present due to sample
complexity, the use of capillary GC/MS may aid compound Identification.
Liquid phase selection should be made based on data from the Level 1 data.
A good starting column would be a 50 foot 0V-17 SCOT column. The appli-
cation of capillary GC/MS is discussed elsewhere in this section.
Prescreening GC/MS Analysis of the XAD-2 Resin Extract
The procedure outlined for the probe, cyclones, and filter extract,
1s adequate for the methylene chloride extract of the XAD-2 sorbent
material. No special precautions other than those discussed above are
necessary. The sensitivity of this analysis Is also expected to be 3 ug/m3.
Prescreening GC/MS Analysis of the Condensate Extract
The condensate extract consists of two parts, a base/neutral fraction
and an acid fraction. The GC/MS analysis of these fractions 1s based on
the polarity of compounds expected to be present. The separation of the
condensate into two parts is to possibly eliminate the need for an LC
fractionation step on this sample and to assure total organic removal from
the condensate. The base/neutral fraction may be somewhat complex but the
acid fraction should be relatively clean. If the chromatograms are not too
B-17
-------
complex, it is advisable to concentrate the samples 100 fold and repeat
this analysis to increase the overall sensitivity without having to do LC
fractionation. A probe HRMS run on the residue of the sample will provide
information on the compounds which are not amenable to GC/MS. If both
fractions are complex, the samples should be blended prior to LC fractiona-
tion, however, if only one fraction is complex, only that fraction need be
submitted for further workup,
GC/MS Analysis of the Base/Neutral Fraction of the Condensate
The same procedure outlined for the probe, cyclones, and filter extracts
is applied to the base/neutral fraction of the condensate. Sensitivity of
this analysis is expected to be 3 ug/m^,
GC/MS Analysis of the Acid Fraction of the Condensate
Due to the polarity and the acidic nature of the acid fraction, a polar
column is used for this analysis. For prescreening of the acid fraction a
phosphoric acid treated carbowax. 2QM column is recommended. Other columns
such as tenax and FFAP will also work. When using the carbowax column,
phosphoric acid treated glass wool should be used to plug the column ends.
This will minimize adsorption of acidic species. The gas chromatographic
procedure for the carbowax column is given below:
Liquid phase - 3% phosphoric acid and 10% carbowax 20M
Solid support - Chromasorb W-AW
Column type - glass
Column size - 2 MM ID x 2 meter long
Temperature program - 50°C for 5 minutes; 50-180° at 4°C per
minute. Hold at maximum until- all peaks elute.
Injector temperature - T9Q°C
Detector and transfer line to separator temperature - 190°C
Flow rate of helium - 30cc per minute
Sample size - 5 ul
Sensitivity - 5 yg/m3
3-13
-------
Ultimate Organic Compound Identification
Once the prescreening GC/MS work has been completed on the extracted
samples, a general idea of compound type or class is available. If no
significant organic compound presence is found the Level 2 analysis can be
stopped at this point. V materials are detected and have not been satis-
factorily identified, the analysis should continue on those samples using
the following procedures.
The next step is to separate the various extracts, after they have
been condensed,by liquid chromatography to permit compound Identification.
The purpose of this LC procedure is to separate the samples Into approxi-
mate classes based on polarity using a gradient elution technique. The
detailed procedure for the LC fractionation is given 1n the Level 1 manual.
The LC separation is not a high resolution technique, therefore overlap in
the compound classes in many of the fractions is common. The procedure for
Level 1 is followed, even though several of the fractions are blended after
separation prior to analysis. The blending of fractions is due to compound
class similarity and allows a more cost effective GC/MS analysis. Table
3 gives the blending of the fractions following the LC separation using
the solvent gradient outlined in Table 4 . (Unblended aliquots can be
analyzed if the analyst decides complexity warrants).
Table 3. LC FRACTION BLENDING
LC Fraction
Blend
A
3
4
C
5
7
B-19
-------
TABLE 4. SOLVENTS USED IN LIQUID CHROMATOGRAPHIC SEPARATIONS
Fraction No.
Solvent Composition
1
Pentane
2
20% Methylene Chloride in Pentane
3
50% Methylene Chloride in Pentane
4
Methylene Chloride
5
5% Methanol in-Methylene Chloride
6
20% Methanol in Methylene Chloride
7
50% Methanol in Methylene Chloride
GC/MS ANALYSIS OF LC FRACTIONS
Each of the blended LC fractions are concentrated to a volumn of less
than 10 cc using an air drying technique. Once this has been achieved, 1
of an internal standard is added and diluted to exactly 10 cc using methylene
chloride in a volumetric flask. In specific cases, where sensitivity is very
important a smaller volume may be used as long as it is known exactly. The
GC analysis of the individual fraction blends are discussed below.
Fraction A (1)
Fraction A contains the compounds that generally fall in categories
1 and 2 of the MEG list. These include aliphatic hydrocarbons and alkyl-
halides. These are the least polar compounds to be analyzed and are well
suited to low polarity silicon liquid phase GC columns. The column condi-
tions given below provide complete analysis of this fraction.
Liquid Phase OV-17
Liquid Loading - 3%
Solid support - Chromasorb W-AW-DMCS
Column type - glass
B-20
-------
Column size - 2 mm ID x 2 meters long
Temperature program - 50°C for 5 minutes; 50-280°C at 6°
per minute. Hold at maximum until all peaks elute.
Injector temperature - 290°C
Detector and transfer line to separator temperature - 290°C
Flow rate of helium - 30 cc per minute
Sample size - 5 ul
Sensitivity - 3 yg/m3
Because of the nature of the compounds found in fraction A, the use of
chemical ionization is not recommended to improve sample identification.
The loss in sensitivity and the confusion resulting from a mixed ionization
process suggests that electron ionization is the method of choice.
Fraction B (2 & 3)
Fraction B has been blended and contains categories 2, 15, 16, 21, and
22 of the MEG list. These compounds can normally be classified
as unsaturated hydrocarbons and halogenated species. In general these
classes produce strong molecular ions in the electron ionization mode of
operation,resulting in spectra which are easy to interpret.
The GC separation is best conducted on a high temperature nonpolar
chromatographic column such as Dexil 300. The conditions for a typical
analysis are given below.
Liquid phase - Dexil 300
Liquid loading - 3%
Solid support Chromasorb W-AW-DMCS
Column type - glass
Column size 2 mm ID x 2 meters long
Temperature program 50°C for 5 minutes; 50-300°C at 6° per
minute. Hold at maximum until all peaks elute.
Injector temperature - 300°C
Detector and transfer line to separator temperature - 290°C
Flow rate of helium - 30cc per minute
a—21
-------
Sample size - 5 yl
3
Sensitivity - 3 yg/m
The compounds generally found in LC fractions 2 and 3 also not
amenable to chemical ionization. Electron ionization spectra of these
materials should be sufficient for compound identification.
GC/MS Analysis Fraction C (4 & 5)
Fraction C represents classes of compounds with increased polarity
over the previous fractions. Several intermediate polarity nitrogen,
sulfur, and oxygen,containing compounds elute in these fractions. Analy-
sis of this material is best suited to an intermediate polarity silicon
column of which there are several to choose from. The chromatographic
conditions given below represent a compromise for this class of materials.
Liquid Phase - OV-17
Liquid loading - 3%
Solid support - chromasorb W
Column type - glass
Column size - 2 mm ID x 2M long
Temperature program - 50°C for 5 minutes; 50-290°C at 6°C per
minute. Hold at maximum until all peaks elute.
Injector temperature - 290°C
Detector and transfer lines to separator temperature - 290°C
Flow rate of helium - 30 cc per minute
Sample size - 5 ul
Sensitivity - 3 ug/m3
Due to the nature of these classes of compounds and the fact that
they generally contain heteroatoms, chemical ionization is recommended as
a supplemental technique to aid in the interpretation of the mass spectral
data.
B-22
-------
GC/MS Analysis of Fraction D (6 & 7)
LC fractions 6 and 7 represent a complex mixture of compounds which are
rather polar and have widely varying acidities. In these two fractions both
basic and acidic compounds elute together, and such mixtures are not amenable
to a single gas chromatographic column. Without previous information as to
the nature of compounds present, it is necessary to run this fraction on at
least three different GC columns in order to insure that all materials in
the sample have been identified.
The columns selected for these analyses (given below) include an inter-
mediate polarity silicon column, a column designed to elute free fatty acids
and glycols, and another to elute free amines. A class of compounds known as
nitrosoamines elute in this fraction. These materials are very toxic even
at low concentrations. An attempt to analyze for nitrosoamines in this
mixture, without special care would be virtually impossible. If nitrosoamines
are expected, special precautions should be taken. Specifically designed
clean up steps should be used followed by chromatographic analysis with a
column such as carbowax 20M,which is especially good for nitrosoamines at
low concentration.
Liquid Phase OV-17
Liquid loading - 3%
Solid support - Chromasorb W
Column type - glass
Column size - 2 mm ID by 2 meters long
Temperature program - 50°C for 5 minutes; 50-300°C at 6° per
minute. Hold at temperature maximum until all peaks elute.
Injector temperature - 290°C
Detector and transfer line to separator temperature - 290°C
Flow rate of helium - 30 cc per minute
Sample size - 5 microliters
Sensitivity - 3 yg/nf*
This column is designed to elute those compounds with intermediate polar-
ity such as esters, ketones, and nitrogen heterocycles. The more polar ma-
terials are better suited to an FFAP column described below.
8-23
-------
Liquid Phase - FFAP
Liquid loading - 10%
Solid Support - Chromasorb W-AW
Column type - glass
Column size - 2 mm ID by 2 meters long
Temperature programs - 50°C for 5 minutes; 50-230°C at 6°
per minute. Hold at temperature maximum until all peaks elute.
Injector temperature - 240°C
Detector and transfer line to separator temperature - 250°c
Flow rate of helium - 30 cc per minute
Sample size - 5 yl
Sensitivity - 10 yg/m3
The basic compounds such as amines are better suited to columns specific
for basic materials. The following set of conditions will provide good chro-
matographic separation for basic compounds.
Liquid Phase - 10% carbowax 20M - 3% KOH
Solid Support - Chromasorb W
Column type - glass
Column size - 2 mm ID by 2 meters long
Temperature program - 50°C for 5 minutes 50-180 at 6° per
minute. Hold at temperature maximum until all peaks elute.
Injector temperature - 180°C
Detector and transfer line to separator temperature - 190°c
Flow rate of helium - 30 cc per minute
Sample size - 5 ul
Sensitivity - 10 yg/m3
The use of these three columns should provide compound identification
on fractions 6 and 7. Alternate columns may be used if information from the
GC/MS analysis of the original extracted material shows specific categories
present.
B-24
-------
LEVEL II ANALYSIS OF PROCESS WATER SAMPLES
This Level 2 plan for analysis of water samples is taken from the
Sampling and Analysis Procedures for the Survey of Industrial Effluents
for Priority Pollutants, published by the Environmental Protection Agency,
Cincinnati, Ohio. The schematic plan is outlined in Figure 3-3. The
analysis is divided into three parts. The first is a direct Injection of
the aqueous sample for the determination of very high concentrations of
organic materials and those compounds which are not amenable to the
Bellar purge and trap technique. The second step is the purge and trap
technique where an aqueous sample is purged with an inert gas,and the water
immiscible volatile organic compounds are trapped on a Tenax solid sorbent
prior to GC/MS analysis. Finally.the sample is extracted, first at an alka-
line pH followed by an acidic pH extraction to separate the higher boiling
and water miscible organics which may be either neutral, basic, or acidic.
Direct Aqueous Injection GC/MS
When impurities in the water exits at very high concentration, they
can be most easily determined,both qualitatively and quantitatively by direct
aqueous injection of the water sample. Typically.a 5 microliter sample
of the water is injected onto an appropriate GC column such as Tenax for
polar compounds and Porapak Q for non-polar compounds using the conditions
given below. The direct injection technique is also useful for the analy-
sis of extremely volatile impurities which cannot be determined by the
purge and trap technique.
Tenax GC
Column type - Tenax GC
Column material - glass
Column size - 2 mm ID x 2 meter long
Temperature program - 50°C for 5 minutes; 50-300°C at 6°
min; Hold at maximum until all peaks elute.
Injector temperature - 280°C
Detector and transfer line to separator temperature - 280°C
Flow rate of helium - 30 cc per minute
Sample size - 5 pi
Sensitivity - 100 uq/1
B-25
-------
Tenax GC is a gas-solid chromatographic material. It does not contain
a liquid phase and has very good temperature stability. It tends to elute
polar materials with ease, however, non polar compounds are highly retained
on the column. The ultimate sensitivity achieved with this column is some-
what lower than many others due to its adsorbtive character.
Porapak Q
Porapak Q is a porous polymer which is also a gas solid absorbent, and
will elute most nonpolar compounds with good resolution.
Column type - Porapak Q
Column length - 4 mm ID by 2M long
Temperature program - Room temperature to 240°C at 6°C
per minute. Hold at maximum until all peaks elute.
Sample size - 5 ul
Sensitivity - 100 ug/l
Purge and Trap Concentration Technique
The purge and trap technique is designed to concentrate those organic
compounds from water which are immiscible and have a boiling range up to
about 130°C, very low boiling immiscible materials are not trapped by this
technique. The apparatus used for this analysis consists of a purging
chamber in which the sample is placed. The chamber is purged with an
inert gas such as helium at a flow rate of 40 cc per minute. The purge
time is approximately 12 minutes. The organic vapors are trapped on a a
Tenax and silica gel column which is subsequently heated. The desorbed
gases are injected into a gas chromatograph, followed by separation on a
carbowax 1500 column.
Liquid phase - 0.2% Carbowax 1500
Solid support - Carbopak C
Column type - glass
Column size - 2 mm ID by 3 meters long proceeded by a short
column of 3% Carbowax 1500 on Chromasorb W
Helium flowrate - 30 cc per minute
Temperature program - room temperature during trap desorption followed
by rapid heating to 60°C hold for 4 minutes then program at 8°C per minute
to 170°C and hold for 12 minutes or until all compounds have eluted.
3-26
-------
Sensitivity - variable depending on trapping efficiency, must be
determined daily when analysis technique is used. Internal standard must
be used for quantitation.
The column used in this analysis has very high resolution for nonpolar
materials which are low boiling. These include categories 1, 2, 15, and 16
of the MEG list. If the sample is highly contaminated and chromatographic
resolution is insufficient for compound identification, the use of a capil-
lary column, either OV-17 or Carbowax 20M, may be used as a substitute in
this analysis.
When using the purge and trap technique it is necessary to run blank
water samples between each analysis sample. It is also necessary to bake
the trap during the course of the GC run to remove all possible inter-
ferring organic substituents which may cross over from one sample to the
next due to insufficient trap heating.
Extraction of Water Samples for GC/MS Analysis
The extraction of water samples for analysis by GC/MS is identical
to the procedure outlined for the condensate sample of the sampling train,
iIf the chromatographic analysis of the extracts are complex
and incomplete compound identification results, the LC fractionation step
should be Implemented as outlined 1n Figure 2-2. The sensitivity of this
technique 1s 7 g/1.
ANALYSIS OG GAS BAG SAMPLES BY GC/MS
The gas bag samples are analyzed by expulslng a known volume of the gas
and trapping on a tenax support. The tenax trap 1s then heated and the de-
sorbed compounds are condensed 1n a LN2 trap before injection onto a Porapak
Q GC column. The sample 1s chromatographed and spectra recorded for each
peak which elutes. Blank bag samples must run to correct for normal bag
background.
ANALYSIS OF BULK SOLID SAMPLES
Bulk solid samples such as scrubber cake and fly ash are extracted with
methylene chloride using a Soxhlet apparatus. The sample 1s then analyzed by
GC/MS using the same procedures as discussed earlier. The sensitivity for
this technique, based on a 5 g sample, 1s 4 g/kg. If more sensitivity 1s
required a larger sample must be used.
B-27
-------
B-28
-------
APPENDIX C
INORGANIC ANALYSIS METHODS
-------
APPENDIX C
INORGANIC ANALYSIS METHODS
The comparative assessment tests conducted at the Firestone boiler
were designed to study the effect of emission control devices on the flue
gas composition. As part of this program, Level 1 and comprehensive
Level 2 sampling and analysis procedures were used to study the inorganic
compounds in the flue gas streams. The Level 2 sampling consisted of
using modified method 5 train for particulate matter in the flue gas and
the controlled condensation system for the HgSO^ content of the flue gas.
These trains are shown in Figures 1 and 2. In this approach, the Level 1
SSMS elemental data was to be the focus of Identification efforts on specific
elements. The criteria used for the evaluation was MEG compounds and MATE
values^ developed from the Industrial Environmental Research Laboratory
of the EPA at Research Triangle Park.
After the Level 1 data was reviewed, specific analytical techniques
were used to look for compounds in the sample. These methods included:
• Thermogravimetric Analysis (TGA) — Used to determine drying
temperatures or stability data.
• Polarized Light Microscopy (PLM) — Used visually to identify
materials present in the sample.
• Inductively Coupled Plasma Optical Emission Spectroscopy (ICPOES) -
Used to determine accurate inlet/outlet concentration of elements.
• Fourier Transform IR (FTIR) - Used to identify inorganic compounds
from specific IR band correlations.
§ X-Ray Diffraction (XRD) — Used to directly identify crystalline
material in the solid samples.
• Electron Spectroscopy for Chemical Analysis (ESCA) — Used to
study the surface and sub-surface sulfur concentrations and
oxidation state of bulk samples.
• Secondary Ion Mass Spectrometry (SIMS) — Used to study the sur-
face and sub-surface composition of bulk samples.
• Scanning Electron Microscope with Energy Dispersive X-Ray
Fluorescence (SEM-EDX) - Used to obtain high resolution photo-
graphs and elemental composition of single particles.
• Transmission Electron Microscope with Selective Area Electron
Diffraction (TEM-SAED) - Used to identify individual particles
by their electron diffraction pattern.
C -l
-------
In addition to these instrument methods, specific anion analyses for
Cl~, F~, NO^'j and SO^" were run on all the samples.
The following sections will discuss the results from each of the tech-
niques. Specific correlation found in the data will be presented in the
discussion section.
RESULTS OF ANALYSES
Complete sets of samples were available for Test 201-1 (coal) and 202-4
(oil). Only these sample sets were analyzed, since the boiler operated at
85 percent or better load through all of the tests.
The following sections contain the data from each of the methods
employed for each sample analyzed. In some cases, two or more methods
will be discussed together for comparison purposes.
Evaluation of Level 1 SSMS Data
The SSMS data was reviewed and compared to the air MATE values for the
most toxic MEG compound of each element. The comparison consists of deter-
mining the ratio of the value found in each sample to the appropriate MATE
value. It has been decided, in view of the relative inaccuracy (.factor of
2-3) of SSMS, that any ratio of sample to MATE air value exceeding 0.5 would
require further research be focused on those elements. Table 1 shows an
example of how*the Level 1 data was reduced for the Site 203 outlet sample
Both the coal (200) and the oil (203) sites outlet SSMS data were evaluated
in this fashion, and the results exceeding 0.5 percent of their MATE values
are tabulated in Table 2. Though the oil samples had a lower inlet partic-
ulate loading, the number of elements exceeding their MATE values was high
for the oil outlet emission values. The following analyses seek to find
more quantitative information about these elements, and determine the exact
species present.
TGA Test Results
TGAs were run on all available loose particle samples. These
included: 201-1-1 cyclone and filter, 201-1-scrubber cake, 202-4-cyclone
and 202-4 scrubber cake. The analyses were performed using a duPont 950
instrument in a ^ atmosphere at up to 600aC; the spectra may be found In
Appendix A. Both 201-1-1 cyclone and filter material showed little weight
loss up to 500°C. After that point, a steady weight loss was noted. The
C -2
-------
TABLE 1. EXAMPLE OF LEVEL 1 SSMS DATA REDUCTION FOR 203 OUTLET
CATCGOAV
COMPOUNO
NATE
NATE
HATER
1.9/1
H£ALTM
HATE
MATES
1.9/I
Ecotocr
HATE
LAKO
i«9/l
HEALTH
MATE
LAM
M9/1
ECOLOGY
SAMPLE
»9/¦
(w*>
*9/1
RATIO
SAWLE
LEVEL 2
KQUIDEO
».TES
N.NO
mtt
27. LITHIUM
Li
2.2 « 10
3.3 a 10*
3.6 a 102
7.0 a 10-'
7.5 x lO"1
*.1
o.oy
/V
Li*
2.2 « 10
3.) « 102
3.8 a 102
Lif (as LI)
2.2 x 10
1.) < to?
3.8 a 102
L*2C03 (as Li)
2.2 » 10
3.3 « to?
3.8 a 102
L1H
2.S « 10
3.a « 102
*
28. SODIUM
Na*
S3 x 10*
8.0 a I0&
n
1.6 a 103
N
NaOH
2.0 x 103
3.0 a 10*
N
6.0 a 10
N
U.I
V
29. POTASSIUM
KOH
2.0 x 103
3.0 a 10*
N
6.0 a 10
N
OIL
AJ
K
H
*
N
N
t* (as K)
N
N
2.3 a 10*
8
4.6 x 10
30. RUBIDIUM
1.21 x 10s
1.62 x 10®
N
3.64 a 10?
N
31. CESIUM
CS*1
8.19 x 10*
1.23 a 10*
N
2.46 a 103
N
32. BERYLLIUM
Be
2.0
3.0 a 10
5.5 a 10
(.0 a 10-2
1.1 a 10-1
1
OS
y
le+*
2.0
3.0 a 10
5.5 a 10
6.0 a 10-2
1.1 a 10-1
BeO (as B*)
BcOAItOj-SIO;
(•« Bel
2.0
3.0 a 10
5.5 a 10
6.0 a 10"*
1.1 a 10-1
33. MAGNESIUM
Nagnesiua, Ng
6.0 x I03
9.0 a 10*
8.7 a 104
1.8 a 102
1.7 a 102
%*fC
0.79
V
thgnesfui Ioa, Ng**
6.01 x 103
9.0 a 10*
8.7 a 10*
1.8 a 10?
1.7 a 102
M>9nes1ia Oxide.
M
1.01 x 10*
1.5 a IOS
1.0 a 10*
3.0 a 102
2.0 a 102
Magnesiu* fluoride,
1*9*2 (•» Ng)
6.0 x 103
9.0 a 10*
8.7 a 10*
1.8 a 102
1.7 a 102
Ni9Msiua Sulfate,
HgSOf (as Ng)
6.0 x 103
9.0 a 10*
8.7 a 10*
1.8 a 102
1.7 a 102
Hagneslte, M9CO3
(« Mg)
6.0 « 103
9.0 a 10*
8.7 a 10*
1.8 a 102
1.7 a 102
Ooloaite, N9CO3-
CaCOi {as Ng)
6.0 x 1Q3
9.0 a 10*
8.7 a 10*
1.8 a 102
1.7 a 102
Asbestos (as Ng)
6.0 x 103
9.0 a I04
8.7 a 10*
1.8 a 10*
1.7 a 102
34. CALCIUM
Calciu* Ion, U++
1.6 x 10*
2.4 a 10*
1.6 a 10*
4.8 a 102
3.2 a 10
a
-------
TABLE 2. SAMPLE TO MATE RATIOS *0.5
200-0utlet
203-0utlet
Li
Na Sb
Be
Be S
Si
Mg V
As
Ba Cr
S
Si Fe
Cr
Pb Co
Fe
P Ni
Ni
As Cd
202-4-I-cyclone sample showed a completely different thermal weight loss
profile. That sample showed an immediate loss of 1 percent up to 5Q°C,
1 percent between 50 and 100°C, 2 percent from 100 to 425°C and 26 percent
from 425 to 600°C. The first two weight losses probably represent water
of hydration and the final weight loss was probably due to a decomposition
or volatilization.
If the composition of a sample is known, TGA can be used to quantitate
a compound based on a specific weight loss due to waters of hydration or
decomposition patterns. In this case, CaSO^'l/SHgO should be present in
the scrubber cake, and at ^100°C all of the HgO molecules are lost. Based
on the weight fraction of H20 in CaSOj-l/^h^O and the percentage weight
loss between 25 to 125°C, an upper limit to the amount of CaSO^ present in
the scrubber cake can be set. The 201-1-scrubber cake (Attachment A) shows
an initial weight loss probably due to moisture in the sample. The second
plateau between 50 and 125°C could be due to the loss of the water hydra-
tion. This weight change indicates that as much as 27 percent CaS03*l/2 was
present in the 201-1-scrubber cake. No similar plateau was found in the
202-4-scrubber cake.
PLM and SEM-EDX
PLM analysis of samples from the coal (201) and oil (202) tests was
completed on the following samples:
• 201-1-flyash
• 201-1-I-cyclone
C -4
-------
• 201-1-I-filter
• 201-1-scrubber cake
• 201-1-0-filter
• 202-4-I-cyclone
• 202-4-I-filter
• 202-4-scrubber cake
• 202-4-0-filter
PLM analysis consists of viewing the samples under a microscope to
study: size, shape, color, and color fracture index of the particles.
Using these and other physical properties, an analyst can identify individ-
ual particles with diameters as small as ^0.5y. Below that level SEM must
be used to provide an image of the particles for morphological study.
The EDX attachment provides elemental composition information for
selected areas of a particle. Thus, SEM-EDX can be used to identify an
unknown particle from its elemental composition as well as its general
morphology. In this section selected SEM-EDX photographs will be used to
illustrate and support the discussion of the PLM results. The entire set
of PLM photographs may be found in Attachment B.
A particle size distribution of each powder sample was determined using
optical microscopy with manual sizing. The results of this analysis along
with the calculated arithmetic mean diameter of the particles in each sample
are presented in Table 3.
Table 4 displays the results of the polarized light microscope
analysis. Estimated weight percentages, estimated modal size, and size range
for each component in each sample are summarized in this table.
Three of the nine samples (201-1-0-Filter, 202-4-1-Fi1ter and 202-4-
0-Filter) were impacted on glass fiber filters. These were so heavily
loaded that particles could not be distinguished from one another. An
attempt was made to remove particles by "teasing" them off with a needle.
This was unsuccessful because the particles firmly adhere to one another
and the filter. For this reason, no overall particle size distribution
was performed on these samples, nor was SEM-EDX analysis possible.
-------
TABLE 3. OVERALL SIZE DISTRIBUTION
Size Ranges
201-1-
Flyash
201-1-1-
Cyclone
202-4-1-
Cyclone
<3.2ym
20.8%
39.6%
21.2%
3.2-6.4
21.3%
26.9%
28.1%
6.4-12.8
19.9%
17.0%
26.6%
12.8-19.2
17.6%
9.9%
11.3%
19.2-32.0
12.0%
5.2%
8.9%
32.0-48.0
5.1%
0.9%
2.5%
I 48.0-64.0
I
2.3%
0.5%
0.5%
64.0-96.0
0.5%
0.0
1.0%
96.0-128
0.5%
0.0
0.0
>128um
i
0.0
0.0
0.0
l
j Arithmetic Mean
! Diameter
I
|
13.4um
6.63pm
10.4ym
Size Ranges
201-1-1-
Filter
201-1 -
Scrubber Cake
¦
202-4-
Scrubber Cake
| <1.Opm
19.6%
33.0%
37.7%
1.0-2.0
22.5%
20.9%
29.1 %
j 2.0-3.0
16.7%
13.1%
11.2%
3.0-4.0
10.5%
10.2%
9.9%
4.0-6.0
12.0%
10.7%
7.2%
i 6.0-8.0
j
8.6%
5.3%
2.2%
8.0-10.0
4.3%
2.4%
0.4%
10.0-14.0
2.9%
2.4%
1.3%
14.0-20.0
1.0%
1.5%
2.2%
>20.Oym
0.0
0.5%
0.9%
Arithmetic Mean
] Diameter
3.32um
2.99ym
2.78pm
C -6
-------
TABLE 4. PLM ANALYSES RESULTS
COAL FIRED LOILES SAMPLES
201-
1-Flyash
201-1
-I-Cyclone
201
-1
-I-Filter
201
-1-Scrubber
Cake
201-1-0-
Filter
Components
A
B
C
A
B
C
A
B
C
A
B
C
A
ParLially fused flyash
65-80*
30
5-140
55-7%
25
5-65
40-552
12
5-40
55-70%
20
4-60
Flyash
l-5i
4
1-20
10-20%
4
1-16
35-502
2
<1-13
10-20%
5
1-15
Oil soot
5-1 i%
25
2-100
10-20%
15
2-80
10-25%
8
1-40
10-20%
15
1-60
25-40^
f-Sa ^net i te
10-20%
12
3-45
10-20%
12
3-25
<2%
5
2-14
1-5%
10
5-45
Iron oxide
2-52
15
1-40
1-5%
7
<1-50
<2%
5
1-1B
<2%
3
1-21
15-252
Coke
1-52
40
5-160
<23!
60
6-100
Quartz
<2%
12
5-60
<2%
20
5-40
Calcite
<1%
6
1-10
1-52
CaS03-l/2H20
10-20%
6
(length)
3-21
(length)
50-65%
Urkno^n sulfate
__
OIL FlkEl) BOILER SAMPLES
201-4
-I-Cyclone
202-
4
-I-Filter
202-4-Scrubber Cake
202-4-0-
Filter
Components
A
B
C
A
A
B
C
A
Partially fused flyash
20-35%
20
6-40
f Flyufh
1-52
3
<1-15
Oil Sl,o(:
£0-752
20
<1-140
10-20%
1-5%
30
1-80
Hd'ir.uL i ti-
<2%
12
3-50
lrr,n o::i £i<-'S
<1%
8
3-35
15-30%
5-10i
Unknown sulfate
1-5*
45-60%
50-65%
iaSO.,-l/2ii..O
95%+
30
2-80
Calcite
1-5*2
Water clropl jts
15-30%
30-45%
K^y: A - Estimated weitjltt percent
B - Estimated medal diameter (tim)
C - ratine-
-------
Particles found in most samples were flyash, partially fused flyash,
oil soot, and iron oxide (hematite and magnetite). Traces of quartz
shards, coke, and calcite were found in many samples. Both scrubber cake
samples contained calcium sulfite hemi-hydrate (CaS03-1/2H20), That hydrate
was the principal component of sample 202-4-scrubber cake.
Regardless of whether coal (series 201) or oil (series 202) was used
to fire the boiler, oil soot was found in the samples collected. There
was, however, a difference in the appearance of the oil soot frojn the two
series. Figure 1 shows oil soot in sample 202-4-I-cyclone collected
during oil combustion. The oil soot is largely in the form of complete
cenospheres with smooth, unbroken walls. Oil soot from sample 201-l-fiyaSf,
which is representative of the oil soot in all the 201 series samples is
shown in Figure 2. It appears to be broken, abraded, and has a grainy sur-
face texture. The more worn appearance of the 201 series oil soot indi-
cates that it was probably oil soot retained in the ducts from some earlier
oil combustion. It was probably not freshly produced oil soot from oil
combustion which occured during the sampling period.
Descriptions of the individual samples follow.
*
Coal Fired Samples —
Sample 201-1 -Flyash —
Between 65-80 percent of the sample mass is partially fused flyash.
These are the large white particles seen in Figure B-3. The opaque white
appearance is due to reflection off air bubbles entrapped in the trans-
parent, glassy, colorless material which comprises the partially fused
flyash. In this sample these particles are very large, roughly spherical
up to 14u in diameter. Figure 3 shows a SEM photograph of a partially
fused flyash particle with its elemented analysis.
Magnetite comprises 10-20 percent of the sample mass. These shiny
black spheres range in diameter from 3 to 40u, with an average size of
about 12y. They are easily detectable by moving a magnet near the sample
and watching for a corresponding motion in the sample. Figure 4 shows a
SEM photo of an iron oxide particle.
C -8
-------
t
Figure 3. SEM-EDX of partially fused flyash from 201-1-Flyash. Ele-
ments present at position marked: Si, Fe (strong); K, Ti
(medium) and Cu (trace).
* •• 8
Figure 4. SEM-EDX of iron oxide particle from 201-1-Flyash. Approx-
imately 36y diameter. Elements present: Fe (strong);
Si (weak). _ ^
-------
On soot was also present, primarily as fragments of cenospheres.
There are many intact cenospheres in this sample, however, they looked worn
and abrated as previously described. Oil soot represented 5-15 percent of
the sample mass.
The flyash ranged in color from clear, colorless spheres to golden,
brown, or red spheres. All are transparent and around 10 percent contain
air bubbles. The average diameter was estimated to be 4y and the largest
flyash sphere seen was 20ym in diameter. Flyash was a minor sample compo-
nent, contributing less than 5 percent of the total mass.
Up to 5 percent of this sample was iron oxides. Most of these were hem-
atite: nonmagnetic, red or orange red, roughly spherical, birifringent
particles. Trace components were quartz shards and a few large (>100u in
diameter) pieces of coke.
Sample 201-1-Cyclone —
As in the 201-1-flyash sample, the glassy masses of partially fused
flyash were the largest sample components (55-70 percent of the sample
mass). In general, this component was morpholically similar to that in the
201-1-flyash sample but smaller in size (see Figure B-4). The largest par-
ticle of partially fused flyash found in this sample had a diameter of
only 65y as compared with 140y in the 201-1-flyash sample. Most of the
partially fused flyash particles in this sample were between 20 and 30u m
in diameter.
Oil soot was a major component at 10-20 percent of the sample mass.
Few intact spheres were present, most were large portions of the spheres and
fragments. As in the other 201 series samples, the oil soot appeared more
abraded than in the 202 series samples. The average diameter of an oil
soot particle was about 15y, but they range from 2y up to 80y.
Magnetic particles also constituted about 10-20 percent of the sample
mass. These black, opaque spheres average about 12y in diameter and
spheres up to 25y in diameter were seen in this sample.
About 10-20 percent of this sample was flyash (Figure 5). These
transparent spheres ranged from colorless to dark brown but most were
yellow-tan in color. A small portion of these spheres contain air bubbles
C -10
-------
Inside ¦-*
4~" Skin
i t
Inside Skin
Figure 5. SEM photo of a cenosphere from 201-1-I-Cyclone.
Elemental composition by area:
1. Inside: Si, Fe (strong), Ti, Ca, K (medium)
2. Skin: Si (strong), Ca (medium), Ti, Fe (weak).
and some had an orange peel surface texture. Their average diameter is
4ym and almost all were between lpm and 12pm. The largest sphere seen had
a diameter of 16ym. Figure B-5 is an example of the type of cenospheres
found in this sample.
Quartz shards, coke, and iron oxides are trace sample components.
The quartz shards are sharp and look freshly broken. Most of the iron
oxide is in the form of hematite.
Sample 201-1-I-FiIter —
It should be noted in the particle size distribution tabulation
(Table 3) there were not as many large particles (>20ym) in this sample
C -11
-------
(Figure 6) as in the other powder samples analyzed (Figure 3 or 4, for
example). This sample appeared to be composed primarily of partially fused
flyash and flyash.
The partially fused flyash was similar to that in the samples already
discussed except that it was considerably smaller. The average diameter is
only 12ym (maximum diameter: 40ym) as compared with a 30ym average diameter
in sample 201-1-flyash and a 25ym average diameter in sample 201-1-I-cyclone
Partially fused flyash is 40-55 percent of the sample mass.
Glassy flyash spheres contribute 35-50 percent of the sample mass.
As with the partially fused flyash, the glassy flyash was morphologically
similar to that in the other samples (Figure 7) but generally smaller. The
flyash had an average diameter of 2ym. Both the 201-1-flyash and 201-1-1-
cyclone contained flyash with an average diameter of about 4pm.
2J
Ji
Figure 6. SEM photo of CaS03-l/2H20 laths impacted on flyash
cenosphere. Elements present are:
1. Fe, S (strong); Si, Ca (medium)
2. Fe (strong); Si, Ca, S (weak).
C -12
-------
*J
tl
Figure 7. SEM photo of typical aggregates found in 201-scrubber cake-
Elements present:
1. Si, Fe (strong); K (weak)
2. Ca, S, Si (strong); Fe (weak).
Fragments of oil soot cenospheres composed 10-25 percent of the sample
mass. The oil soot appeared crushed and eaten away and no intact spheres
were seen. Magnetite and iron oxides were trace sample components. Iron
oxides were present primarily as hematite.
Sample 201-1-Scrubber Cake --
Partially fused flyash was again the largest sample component, 55-70
percent of the sample mass. It was similar in morphology (Figure 8) to
that seen in previously discussed 201 series samples. The average diameter
was 20y and the range was from 4 to 60p.
Several flat, thin particles (Figure 9) (laths) were identified as
CaSO^-l^f^O. This form of CaSO^-l/^^O comprised the largest portion of
the CaSO^-l^^O present. Spherulites of CaSO^-l^H^O were a-|so present.
The spherulites in this sample were covered with other particles such as
C -13
-------
Figure 8. 1000X SEM photo of sphere!ites found in 204-4-scrubber
cake. Elemental analysis: Ca, S (strong).
Figure 9. 600X SEM photo of "ball of twine" in 202-4-scrubber cake.
Elemental analysis: Ca, S (strong).
C -14
-------
oil soot fragments arid flyash (Figure B-1Q) in them. Figures 6 and 7 show
examples of the large amounts of flyash present in the 201 series scrubber
cake. In Figure 6 a crystal (area 1) impacted on the cenosphere was ana-
lyzed by EDX. The elemental composition identifies it as CaSO^, but the
strong Fe signal from the cenosphere surface indicates the thinness of the
crystals. Analysis of an aggregate shown in Figure 7 by area showed the
presence of CaSO^ in area 2 and flyash in area 1. The CaSO^-l^b^O is
10-20 percent of this sample as compared with 80-95 percent of the 202-4-
scrubber cake sample.
Ten-twenty percent of this sample was oil soot, almost all of which
was in the form of broken fragments with an average diameter of 15pm and
a size range of 1 to 60ym in diameter.
The flyash particles in this sample were similar to that in other
samples. Their color was primarily tan although they ranged from color-
less to brown with an occasional red sphere. All were transparent, a few
contained air bubbles and some had an orange peal surface texture. All
forms of flyash were responsible for 10-20 percent of the sample mass. The
average flyash diameter was about 5y.
Sample 201-1-0-Filter —
This sample appeared to be composed primarily of fine carbonaceous
particles on a filter saturated with unknown sulfate crystals (Figure B-ll).
The bright birifringence (speckled color) of the filter examined under
polarized light may be due to the unknown within the filter.
The fine carbonaceous particles were probably oil soot. This hypothe-
sis is based on the appearance of these particles with oil soot being the
major carbonaceous component in all of the other 201 series samples. An
analysis for vanadium (associated with oil soot) would confirm or refute
this supposition.
Low temperature ashing of a small portion of this sample removed the
carbonaceous material, revealing iron oxides (Figure B-12). Both hydrated
(yellow) and unhydrated (red) iron oxides were present. Overall, iron
oxides probably accounted for 15-25 percent of the sample mass. Calcium
carbonate (calcite) particles were also seen after ashing, however, they
did not contribute more than 5 percent of the total mass.
C -15
-------
011. F i r ed Bo i1er Samp1es
Sample 202-4-I-Cyclone —
Oil soot, both complete cenospheres arid fragments, was the largest
sample component, comprising 60-75 percent of the sample mass (Figure B-13).
They ranged in diameter from
-------
Traces of birifringent calcium carbonate particles can be seen after
ashing. These, however, did not compose more than 5 percent of the sample
mass.
Sample 202-4-Scrubber Cake --
This sample was almost entirely composed of CaSO^-1/2^0 spherulites
(see Figure B-17). Figure B-18 shows the spherulites with the compensator in
place. The SEM photo in (Figure 8) clearly shows the spherulites in the
202-4-scrubber cake. Note also the lack of flyash in the sample. The SEM
photo (Figure 9) shows a "ball of twine" which turns out to be composed of
Ca and S indicating CaSO^. The spherulites ranged for 15 to 80um in dia-
meter with an average diameter of about 30 pm. Thin, plate-like sheets
(seen in Figure B-19) were also present, probably being CaS03•1/2H2O. This
form accounted for only a small part of the sample mass.
Oil soot fragments and calcium carbonate particles were minor sample
components. Neither contributed more than 5 percent of the sample mass.
Sample 202-4-0 --
This filter has an overall blue-green color (Figure B-20). The color
was probably due to some cation dissolved in water droplets which covered
and were absorbed into the glass fiber filter. About 30-45 percent of the
sample appeared to be water. Because of this, there was a continuous
change in the crystal forms found on that filter resulting from dissolution
of chemicals and recrystallization. Two of the crystal forms found ini-
tially on this filter are shown in Figures B-20 and B-21. Figures B-22 and B-23
show recrystallized forms after the filter remained in the immersion oil
for several weeks. Manipulation of the filter caused redissolution of
those crystals as shown in Figure B-23. Crystals similar in morphology to
those in Figure B-23 were seen forming in water droplets which had been
moved off of the filter. The identity of the crystals is not known at
this time, however, it was at least partially calcium sulfite hemi-hydrate
and the rest was an unknown sulfate. Later analyses will seek to deter-
mine their identity more precisely.
Low temperature ashing of this sample revealed that hydrated iron
oxides constitute 5 to 10 percent of the sample mass. The oxides were fine,
grainy yellow particulates. Figure 1-24 shows the filter after ashing.
C -17
-------
Artion and Elemental Data
The modified Method 5 train was operated at the inlet and outlet of
the F6D. The particulate matter collected in the probe, cyclone, filter
and impinger system was dissolved and then analyzed for 20 elements by the
EPA Corvallis Laboratory using ICPOES. Particulate aliquots from these
samples were extracted with hot water and then HNO3, The extraction solu-
tions were analyzed for CI", F~, NO3 (except HNG3 extract) and SO4 by spe-
cific ion electrodes or standard wet chemical procedures. The results of
these analyses are detailed in Tables 5, 6, and 7.
The anion analysis (Table 5) showed that SO4 was the predominant anion
in all the samples. Chloride was the ion found in the second greatest con-
centration (< 1 percent of the sulfate value).
Performing both water and acid extractions of the sample, provides
additional information with respect to the type of species present. For
example, 201-1-1 SO4 was primarily HNO3 soluble, whereas at the outlet only
3 percent of the sulfate was found in the NHO3 extract. Clearly the nature
of the sulfur species was altered by the sulfur species scrubber process
(see Section 3).
The elemental data shown in Tables 6 and 7 show trace element concen-
trations in gm/m3. The 20 elements listed represent those elements which
could be analyzed by the specific ICPOES instrument. Earlier SSMS analysis
of sample 200-0 showed 8 elements exceeding their MATE values by a fac-
tor >0.5. Of those eight elements, three (Ni, Fe,and Cr) were analyzed by
ICPOES in 201-1-0 and they were found to exceed their MATE values. The
fact that all three elements again exceeded their MATE values by a fac-
tor >0.5 indicates that the SASS train did not cause artificially high Ni,
Cr or Fe values through contamination, since all Level 2 samples were
obtained using a glass train. In the oil fired case Ni and Cr concentra-
tion ratios to MATE values were >0.5, and Fe was not. From the SSMS data,
sixteen elements exceeded their MATE ratios. Ten of these elements were
analyzed by ICPOES. Only 5 elements (Ar, V, Cr, Ni, Cd) of this set
exceeded the MATE ratio .
C -18
-------
TABLE 5. ANION CONCENTRATION INLET & OUTLET OF FGO, mg/m
,3
CI
F~
NO"
SO
4
Site
Hot Water
Extract
hno3
Extract
Hot Water
Extract
hno3
Extract
Hot Water
Extract
Hot Water
Extract
hno3
Extract
201-1-Inlet
1.605
9.262
0.503
0.012
<1.14
36.7
117.4
201-1-0 Inlet
<0.005
<0.004
<0.070
<0.006
<0.585
18.6
0.5
202-4-Inlet
0.463
<0.006
0.045
0.009
0.235
39.7
30.7
202-4-0utlet
0.223
<0.006
0.006
<0.002
0.102
24.5
3.5
-------
TABLE 6. TEST 202-4 (OIL) TRACE ELEMENT SAMPLING RESULTS
Element
Trace Element Composition, mg/m^
Inlet
Outlet
Ca
0.411
0.070
Mg
0.309
0.030
Sb
0.062
0.006
As
0.146
0.030
[ B
0.531
0.039
I Cd
i
0.280
0.066
!
Cr
0.166
0.018
Co
0.104
0.012
i Cu
1
0.539
0.007
Fe
4.847
0.277
j Pb
0.197
0.013
! Mn
1
0.031
0.004
1
| Mo
0.217
0.025
Ni
1.076
0.204
V
2.690
0.822
Zn
0.609
0.065
Se
0.050
0.006
Sr
i
0.043
0.001
AT
5.678
0.484
Zr
0.015
0.001
C -20
-------
TABLE 7. TEST 201-1 (COAL) TRACE ELEMENT SAMPLING RESULTS
Element
Trace Element Composition, mg/m3
Inlet
Outlet
Ca
73.80
0.036
Mg
19.OQ
0.011
Sb
3.74
0.025
As
7.83
0.224
B
5.81
0.334
Cd
0.47
0.001
Ca
2.60
0.132
Co
3.55
0.012
Cu
9.56
0.020
Fe
454.00
2.400
Pb
8.48
0.021
Mn
0.78
0.015
Mo
9.97
0.027
Ni
1.37
0.063
V
3.05
0.058
Zn
2.28
0.048
Se
3.19
0.099
Sr
10.50
0.058
A1
480.00
2.570
Zr
1.62
0.018
C-21
-------
XRD and TEM-SAED Results
Both XRD and TEM-SAED have the capability to identify compounds present
in a sample. XRD is normally used on bulk samples and has a sensitivity of
^0.5 percent in most cases, but under the best conditions can see 0.05 per-
cent. TEM-SAED on the other hand is used for single particle identifi-
cation. Consequently thick particles or large aggregates produce indis-
tinct diffraction patterns.
Only the loose particulate matter could be analyzed by TEM-SAED, and
few of the TEM-SAED photographs provided much more detail than the SEM
photos. Difficulty in breaking up aggregates and the size of most of the
particles reduced the number of diffraction lines obtained. The informa-
tion that was derived from TEM-SAED tended to confirm the XRD results, so
only two examples of TEM-SAED results will be shown.
All of the particulate samples were analyzed by XRD. The three glass
fiber filter samples were run directly and were subsequently subjected to
an ultrasonic Freon bath for fifteen minutes to extract the particles from
the filter. This process did not yield significant amounts of particles.
Therefore, the filters were subjected to an ethyl alcohol, ultrasonic bath.
The alcohol, the suspended particles and some glass fibers were then fil-
tered through 200 mesh glass fiber and the resultant liquid evaporated.
The particulate left in the evaporation dish were then placed in a diffrac-
tometer and compounds were identified by comparing the diffraction patterns
to standard listings from the International Centre for Diffraction Data
(JCPDS). The following sections discuss the individual results from the
samples.
Coal Fired Samples —
These samples were: 201-1 Fly Ash, 201-1-1 Cyclone, 201-1-1 Filter
(free powder), 201-1-Scrubber Cake and 201-1-0(1) Filter. An alcohol
extraction of the last sample was also performed and the results included.
C-22
-------
Four crystalline compounds were definitely identified in at least one
of each of these samples. These were as follows.
Symbol Name Formula JCPDS No.
Q Alpha Quartz SiOg 5-490
M Mullite A15Si2°13 15-776
H Hematite Fe2°3 13-534
S Calcium Sulfite CaSCL-1/2 H?0 4-0588
Hemi-Hydrate
Three crystalline compounds were determined as possibly present in at least
one of each of the samples. These were:
Symbol
C
Cr
Tr
On a comparative basis these crystalline species varied as follows in
the five as-received samples:
Q - Fly ash > Cyclone = Filter (Inlet)
> Scrubber Cake > Filter (Outlet)
M - Fly ash - Cyclone < Filter (Inlet)
> Scrubber Cake > Filter
H - Fly ash - Cyclone - Filter (Inlet)
> Scrubber Cake > Filter (Outlet)
S - Positively identified in Scrubber Cake and possibly in Filter.
Large amount in former, trace in the latter.
Name Formula JCPDS No,
Calcium Aluminum CaO-Al?0~»2SiCL 5-528
Ortho Silicate
Cristobal ite Si02 4-359
Tridymite Si02 18-1170
C- 23
-------
C - (Not confirmed positively in any sample)
Fly ash * Cyclone - Filter (Inlet)
> Scrubber Cake > Filter (Outlet).
Figures A and B show examples of TEM studies of the 201-1-Fly Ash and
201-I-Inlet Filter. Details are obscured by the thickness of the samples
and size of the aggregates formed during the sampling. Figure A does show
a "haystack" quartz fiber and Figure B indicates that Fe-jO^ might also be
present.
Oil Fired XRD Results —
Four samples (202-4-I-Cyclone, 202-4-1-Fi1ter, 202-4-Scrubber Cake
and 202-4-0-Filter) were run direct and alcohol extractions of the filter
samples were obtained and run.
Symbol
Q
M
S
Name
Alpha Quartz
Mullite
Calcium Sulfite
Hemi-Hydrate
Formula
Si02
A15Si2°13
CaS03-l/2 H20
JCPDS No,
5-490
15-776
4-0588
No comparative determination of amount of species present in each
sample could be made for the oil fired burner samples; however, the
following comments should be noted about each sample.
Test Sample 202-4-I-Cyclone —
Cristobalite and tridymite (forms of quartz) possibly present and
may contribute to the very strong 4.11 peak; however, their presence cannot
fully account for its very high intensity. This same peak also appears in
the scrubber cake, but not on the filter samples.
Test Sample 202-4-I-FiIter ~
The many unidentified peaks at large d spacings from this sample seem
to be indicative of vanadyl sulfate in various degrees of hydration; however,
no listed compound could be positively identified. Vanadyl sulfate is very
C-24
-------
Figure A. TEM photo of 201-1 Fly ash showing quartz studies.
Materials present: aQuartz Kaolinite.
Figure B. TEM photo of 201-1-I-Fi1ter.
Material present: aQuartz, Fe304-
C-25
-------
soluble in water and the peaks shown on this chart may indicate a spectrum
of hydrated states, many of which are not yet listed in the JCPDS cards.
The alcohol extraction specimen seems to confirm this since
23-723, was positively identified. However, this identification must be
considered in the light of chemical reactions occurring during the extrac-
tion. In particular V2O3 could have reacted with an acidic sulfate and
formed Vg (504)3.
Test Sample 202-4-Scrubber Cake —
Calcium sulfite hemi hydrate, CaS03-l/2 H20 (4-588) was identified as
present. There were a number of peaks which could not be identified, most
notable was that at 4.11 d spacing.
Test Sample 202-4-Q-Filter --
The same comments are applicable for this sample and its extract as
for 202-4-I-Filter, except that there were fewer peaks, indicating fewer
varieties of crystalline specimens.
There was not sufficient similarity between these four samples from
the oil burner to compare the relative amount of crystalline species
between them.
Summary —
Table 8 summarizes the results from the XRD and TEM studies.
FTIR Analysis Results
Infrared analysis of samples was performed using a Nicolete Fourier
Transform IR Spectrometer. The loose particle samples were run in the
region of 1900 cm"1 to 400 cm"1 as a mull of mineral oil between AgCl
windows to prevent ionic exchange variations that could occur with KB
pelleting procedures. The data obtained is in a digital form and thus
can be manipulated by the instrument in a variety of ways. In particular
spectra can be subtracted from each other to remove impurities or as a
means of identifying unknown materials. Samples of the 201-1 fly ash was
extracted in HNO3 and H20 and used as a reference spectra for removal of
the mineral background. When the subtraction was performed, the resulting
spectra proved to be too difficult to interpret. In general the use of
C-26
-------
TABLE 8. XRD and TEM-SAED SAMPLE COMPOSITION ANALYSES
Samples
Definite
Possible
Flyash
Q; M; H; S a-quartz,
mullite, FegO^
Ca0-Al203-2Si02
—
CaS03-l/2H20
1
1
Cyclone
ct-quartz, mullite, Fe203
CaO'Al203-2Si02
o
CSJ
-a
0)
S-
Filter
(inlet)
Na2Al2Si20g.H20
a-quartz, mullite
CaO*A1203«2Si02
•r-»
<+-
i
o
o
Filter
(outlet)
NaHSO^ a-quartz, mullite
Fe203, CaS03-l/2H20
Scrubber
Cake
CaS03«l/2H20 a-quartz,
mullite, Feo0,
2 3
NagSO^, Ca0-Al203
•2Si02
Cyclone
a-quartz
Filter
(inlet)
FeSi205; NaAl(S03);
Fe3S4 V0(S04)
(202-4-
Filter
(inlet)
extracted
1/2 (S04)3
-o
~-
1
O
Filter
(outlet)
Filter
(outlet)
extracted
1/2 (S04)3
Na2Si20g-, KFeSi30g
C-27
-------
FTIR's subtraction capability was not possible due to the complexity of the
spectra obtained, difficulty in preventing band shifts due to hydration,
particle size effects and the lack of having the correct compounds in the
computer library. As it will be shown in the discussion section, FTIR,
even with these limitations, provided significant information in the iden-
tification of NaHSO^ in the outlet samples. For the next test site samples
an effort will be made to:
• Dry and store all samples to the same degree of moisture
content
• Establish grinding procedures for the samples
• Add key compounds to the computer library for subtraction
The following sections briefly describe the FTIR results. All the
spectra obtained are found in Attachment C.
Test Sample 201-1-Cyclone —
No correlation of the spectrum with any specific materials was pos-
sible. It is believed that most of the spectrum was due to either silicates
in the fly ash or a cell interference caused by placing insufficient sample
between the windows.
Test Sample 201-1-Filter —
The main spectral bands for this sample are shown below with SiC^ for
comparison:
work.
Sample (cm"1)
Si02 (cm"1)
1460 (sp)
1375 (sp)
1075 (vb)
1075 B
797 w
800 m
780 w
555 w
570 w
C-28
-------
The 1375 and 1460 bands could be due to the mineral oil mull, which could
not be totally subtracted. The band at 1075 is very broad and is not of
great utility in identification. The 800 and 570 bands for SiO2 match up
reasonably well, but the 797 band disappeared after HNO3/H2O extraction,
which should not happen if was present.
Test Sample 201-1-Scrubber Cake ~
This spectrum had little detail, as though several weak lines could
have been assigned to the CaS03, CaSO^ and Na2S04 known to be present in
the sample. The low level of detail probably was due to the high moisture
content of the samples.
Test Sample 201-1-0-Filter --
The spectrum from this sample was attained directly from the filter
using Attenuated Total Reflectance (ATR) techniques. ATR was used because
the particles were embedded in the filter. The spectrum was reasonably
rich in bands and a possible assignment (NaHSO^HgO) was made based on the
following comparison:
Sample (cm"1)
NaHSQfl'HgO
(Hz0)
Broad
1660 (m)
1235 (m)
1160 (m)
1120 (w)
1095 (w)
1080 (m)
1038 (s)
857 (s)
773 (w)
865 (s)
775 (m)
Broad
Broad
653 (w, sh)
620 (m)
650 (w, sh)
605 (m)
C- 29
-------
As a test to see whether the inlet filter contained any materials
similar to those on the outlet filter, an electronic subtraction of the
spectrum of the 201-1-I-FiIter from the 201-1-Q-FiIter was performed (see
Appendix C). No change was seen in the 201-1-0-FiIter spectrum, indicating
the NaHSO^ was unique to the outlet.
Sample 202-4-I-Cyclone and 202-4-I-Filter —
The cyclone spectrum exhibited many medium to weak peaks in the 1200-
800 cm'1 region. None of these bands could be attributed to sulfate or any
other material. The inlet filter was scanned using ATR and the background
filter spectrum was electronically subtracted. In general it had inany
characteristics of 201-1-0-FiIter, with additional fine structure super-
imposed on the strong bands. A comparison between the filter and NaHSO^*
HgO is shown below:
-1
Sample, cm
NaHS0d-H20, .
1645 (s)
1660 (s)
1155 (s)
1175 (s)
1047 (s)
1045 - 1075
863 (s)
865 (s)
745 (s)
775 (db, w)
The bands are shifted slightly, but that could be due to the degree of
hydration. No oxides such as Si02 or Fe304 can be seen.
Test Sample 202-4 Scrubber Cake —
The scrubber cake spectrum showed many of the bands seen in the cyclone
material except bands in the 621-645 cm"1 region. The spectrum details and
possible compounds are shown below:
Sample, cm"1 CaS03-2H20, cm-1 Fe304, cm"1
1625 (b) "1625
1305 1325
1150
C-30
-------
Sample, cm"1 CaSQ3-2H2Q, cm"1 Fe3°4' cm_1
1112 1100
955 (b) 955 (b)
721 721
621 653
585 570
542
500
460
455 452
The shifts in the CaS03 spectrum are probably due to the moisture
content of the samples.
Test Sample 202-4-0-Filter —
The outlet filter, while similar to 201-1-0 has fewer details and
seems to consist of bands entirely assignable to NaHSO^:
Sample, cm"1 NaHSO^-HgO, cm"1
1645 (s) 1660 (s)
1155 (s) 1175 (w)
1047 (s) 1060 (b, s)
863 (s) 865 (s)
745 (w) 775 (m)
725 (w, db)
ESCA AND SIMS ANALYSIS RESULTS
Both ESCA and SIMS are surface analysis techniques. In ESCA, X-rays
bombard the sample and knock out inner electrons. Depending on the sample
material and parent element, the electron escape depth can be as much as
9 0
50 A, but is normally closer to 25 A. The electrons which do escape the
sample, are energy classified in a electron spectrometer. Knowing the .
X-ray energy, the binding energy of the electrons can be calculated. Since
the binding energy can be influenced by the electron density at the ele-
ment, changes in an elemental binding energy for a given electron can be
C-31
-------
correlated with the oxidation state of the parent element. This approach
has been used with great success in determining sulfur oxidation states.
Depth profile information is obtained by etching the sample with an Ar+
beam and repeating the ESCA scan.
SIMS uses a beam of Ar+ (or 0) to remove and ionize layers of elements
from a sample. The ions are then identified by a mass spectrometer and the
ion output can be plotted. The following sections will discuss the results
of these tests.
ESCA Results
The original ESCA survey scans are found in Attachment D. All nine of
the solid samples were analyzed and most were re-analyzed after a 75 A Ar+
etching The results of these tests are summarized in Table 9 as surface
atom percent concentrations. While ESCA's sensitivity for elements on the
surface is excellent, bulk sensitivity is ^0.1 percent. With this in mind,
the atom percent data shows no undue surface concentrations of heavy or
volatile trace elements. Overall the coal samples contained more surface
K, Ca, Fe, and A1 than the oil samples, while significant differences
between samples were seen for C. The Inlet 202-4 cyclone contained a large
amount of C and probably was due to the large soot.particles seen earlier
in the PLM studies.
The main use of ESCA was in the study of the sulfur depth profile in
the samples. Comparing 201-1-cyclone to 201-1-Filter, the sulfur surface
content appeared similar. The sharp drop 1n S atom percent after etching of
the inlet filter sample indicates some surface coating of the particles
occurred. One technique to study surface concentrations versus bulk compo-
sition is to normalize the data to a bulk concentration element. In this
study A1 rather than Si was chosen since 201-1-0-F, 202-4-I-F, and 202-4-
0-F were run directly on the filter and the Si content of the filter would
mask the true changes. Figure C is a plot of the S/Al ratio for 201-1-0-F
Ar+ etched to a depth of 750 A. Note that there is some surface dependence
of S, but that the curve levels off after -275 A. These data indicate that
while there may be some surface concentration of S in the 201-1-0-Filter,
it probably is on a particle composed of solid sulfate. This would be the
case if H2S04 condensed on the NaHSO,. particles that appear to be present.
C-32
-------
TABLE 9. SURFACE ATOM PERCENT FROM ESCA DATA
o
I
L*J
o>
Sample*
0
Na
S
Si
A1
Fe
CI
P
V
Ca
C
K
201-1-1 Cyclone
57
4
11
6
7
2
2
1
201-1-1 Filter
56
2.4
11.5
14.1
8
1.4
1.5
1.2
2
201-1-1 Filter 76 A SP
54.2
2.1
4.0
17.9
12.6
2.1
1.3
2.2
1.0
2-7
201-1-0 Filter
45.7
5.5
13.2
7.1
2.2
1.2
1.1
0.7
1.3
14.9
7.1
201-1-0 75 A SP
48
7.4
11.5
9.0
5.2
1.2
1.1
1.1
1.1
12.1
1.8
201-1-0 150 A SP
48.3
9.1
10.0
10.1
7.7
1.2
1.2
1.2
10.0
1.2
201-1-0 300 A SP
Q
48.3
8.8
8.0
10.6
9.6
2.1
1.0
1.2
0.5
8.0
1.8
201-1-0 500 A SP
47.9
7.3
6.7
10.8
13
1.7
1.2
1.1
1.2
8.1
1.1
201-1-0 750 A SP
47
7.3
6.1
11.6
11.0
2.3
0.9
1.5
0.5
1.0
8.8
1.6
201-1 Flyash
44.6
2.6
7.6
13.9
7.5
1.2
1.2
1.2
0.4
18.1
1.6
201-1 Scrubber Cake
47
10.0
14.8
3.2
3.7
0.6
1.5
0.9
5.8
9.9
2.6
202-4-1 Cyclone
38.5
3.2
12.9
2.6
2.3
1.4
0.6
1.2
37.4
202-4-1 Filter
48.6
4.2
10.7
11.8
3.1
1.3
1.5
1.7
17.2
202-4-1 Filter 76 A SP
46.3
4.7
6.5
17.1
3.3
0.8
1.4
1.1
3.3
0.9
12.9
1.7
202-4-0 Filter
46.6
5.8
9.9
14.9
3.5
1.3
2.4
1.7
1.8
14.5
202-4-0 Filter 76 A SP
53.5
3.2
6.8
22.0
4.0
1.3
1.0
2.0
6.2
202-4-SC
41
9.8
15.3
2.1
2.3
1.7
2.1
1.1
1.4
6.5
12.7
3.6
Filter Blank
38.4
5.8
7.5
24.7
1.8
1.1
1.6
1.1
0.9
17.2
Filter Blank 75 A SP
47.3
8
2
28.5
3.2
0.9
2.6
0.6
1.2
4.4
1.3
*A SP refers to Depth of Ar+ sputtering in A.
-------
4.0
2.4
CM
230 330 400 50J c3j 730 C30
ETCHJ NG DEPTH (A)
Figure C. Sulfur dependence with etching depth for outlet
coal fired filter sample.
-------
The 201-1-1-Cyclone and filter samples show a sharp decline to S/Al values
0
for below the 201-1-0-Filter even at 750 A, indicating a high surface con-
centration, perhaps from SOg absorption.
SIMS Results
SIMS analysis up to mass 150 was performed on all solid samples. The
profiling rate was determined by etching a known thickness substrate of
Si02 over Ga. It was determined that the Ar+ beam removed approximately
0.78 A/sec. The output from each etching was stored as uncorrected (for
abundance and ionization probability) ion counts. This raw data was
reduced and drawn as a plot of intensity of the element ratioed to Si
versus time in seconds (depth). The entire set of those plots are found
in Appendix E. Elements that have a surface dependence would show a decay
curve with increasing depth. Only 14 elements were plotted because of
space limitation, but they are representative of the elements found in the
sample. In the 201-1 series no clear-cut surface dependence is seen. The
cyclone and inlet filter plots look very similar, with the exceptionof some
anomolies for the Li ¦* Ca plot. The outlet filter samples show a large
difference in shape and element distribution compared to the inlet filter.
One significance of this differenece is not clear since results from a
free particle are being compared to an impacted particle on a filter.
The 202-4 gas particulate samples show several interesting charac-
teristics:
• Li, Na, Mg and to some0extent A1 and Fe show a maxima
near 400 seconds (280 A)
§ The inlet filter shows little surface dependence for
Ti -+• Ba, but the outlet curves show a much slower decay
curve and an apparent enrichment.
More precise information about compositional changes in the sample are shown in
Table 10. There the atom concentration shown represents isotope abundance
and ionization probability corrected data. Comparisons can be made across
the FGD by looking at inlets versus the outlets. Care should be exercised,
however, since the 201-1-0-Filter has impacted particles in a filter com-
pared to the loose particle samples for the 201-1-I-Filter/Cyc1 one. This
same problem exists for the 202-4-I-Cyclone versus the 202-4-0-Filter.
C-35
-------
TABLE 10. ATOM CONCENTRATION AT A DEPTH OF 1400 A FOR GAS PARTICULATE SAMPLES
0
1
CO
o\
Element
201-1-I-Cyclone
201-1-I-Filter
201-1-0-Filter
202-4-I-Cyclone
202-4-I-Filter
202-4-0-Filter*
Li
922 ppm
0.10%
770 ppm
0.11%
291 ppm
37 ppm
Na
1.05%
1.33%
22.84%
2.27%
9.57%
10.36%
Mg
0.92%
0.77%
1.34%
1.06%
4.51%
2.44%
A1
18.48%
19.28%
9.17%
11.7%
8.1%
5.04%
Si
18.74%
17.88%
10.42%
22.44%
12.3%
23.13%
K
3.23%
3.62%
2.55%
2.75%
3.32%
2.68%
Ca
1.59%
1.91%
2.17%
2.62%
3.75%
3.55%
Ti
0.88%
0.89%
0.15%
0.42%
0.26%
890 ppm
V
294 ppm
310 ppm
187 ppm
0.97%
3.06%
1.22%
Cr
322 ppm
279 ppm
141 ppm
760 ppm
670 ppm
203 ppm
Mi)
304 ppm
314 ppm
140 ppm
251 ppm
759 ppm
298 ppm
Fe
4.82%
4.02%
1.21%
5.44%
4.87%
1.42%
Sr
752 ppm
784 ppm
119 ppm
758 ppm
445 ppm
99 ppm
Ba
232 ppm
268 ppm
42 ppm
449 ppm
516 ppm
42 ppm
0
Concentrations at 1200 A depth.
-------
Changes can be seen in some of the elements with particle size {cyclone to
filter) and across the FGD (inlets versus outlets). In particular a large
enrichment is seen in the Na content of the outlet 201-1 filter sample.
DISCUSSION
The results of the test program were described in the preceding sec-
tions and to some extent discussed. The following sections will present
two major conclusions from the test program.
Trace Element Enrichment Across the FGD
As it has been shown in Section 2.4.3 the overall emission rate of
trace elements was reduced by the FGD. However this data only tells part
of the story. Tables 11 and 12 show the mass loading, percentage removal
and enrichment trend. The enrichment trend was calculated by determining
the ratio of the mass loading of an element to the mass loading of Al. The
ratio normalizes the data and allows comparison of the results.
The average percentage removal varied between the coal and oil tests.
It is believed this is more a mathematical result of low input values for
oil compared to a baseline scrubber output. In the coal case a much greater
input is compared to the same baseline output, which would result in an
artificially greater removal efficiency.
On the other hand, the enrichment trend data shows that while lower
mass outputs are attained, the composition of output is drastically differ-
ent from the input. Summarized below in Table 13 are the elements which
show a position enrichment.
The data can be differentiated into those elements showing between
0-30 percent, 31-60 percent and >61 percent either positive or negative
enrichment trends. In the coal case for positive trends 31 percent showed
a 0-30 percent, 31 percent between 31-60 percent increase and 38 percent
showed >61 percent increase. The same comparison for oil was: 22 percent,
0 percent, and 82 percent, respectively.
Several researchers 2, 3, 4, 5 have found elements such as Sb, As, Cd, Cr,
V, Zn, and Se inhibit concentration increases with decreasing particle size.
C-37
-------
TABLE 11, TEST 201-1 (COAL) TRACE ELEMENT SAMPLING RESULTS
Trace Element
Composition, mg/m^
Percentage
Removal
Enrichment Trend
Element
Inlet
Outlet
> 30%
30 - 60%
C 60%
Ca
73.8
0.036
>99.9
-
Mg
19.0
0.011
99.9
-
Sb
3.74
0.025
99.3
+
As
7.83
0.224
97.1
+
B
5.81
0.334
94.3
+
Cd
0.47
0.001
99.8
-
Cr
2.60
0.132
94.9
+
Co
3.55
0.012
99.7
-
Cu
9.56
0.020
99.8
-
Fe
454
2.40
99.5
-
Pb
8.48
0.021
99.8
-
Mn
0.78
0.015
98.1
+
Mo
9.97
0.027
99.7
-
Ni
1.37
0.063
95.4
+
V
3.05
0.058
99.6
+
In
2.28
0.048
97.9
+
Se
3.19
0.099
96.9
+
Sr
10.5
0.048
99.4
+
A1
480
2.57
99.5
* 1
1.62
0.018
98.9 |
Average 98.5 ± 1.8 1
+
1
-------
TABLE 12. TEST 202-4 (OIL) TRACE ELEMENT SAMPLING RESULTS
Trace Element
Composition, mg/ni3
Percentage
Removal
Enrichment Trend
Element
Inlet
Outlet
> 30%
30 - 60%
<60%
Ca
0.411
0.070
83.0
+
Mg
0.309
0.030
90.3
+
Sb
0.062
0.006
90.3
+
As
0.146
0.030
79.5
+
B
0.531
0.039
92.7
-
Cd
0.280
0.066
76.4
+
Cr
0.166
0.018
89.2
+
Co
0.104
0.012
88.5
+
Cu
0.539
0.007
98.1
-
Fe
4.847
0.277
94.3
-
Pb
0.197
0.013
93.4
-
Mn
0.031
0.004
87.1
+
Mo
0.217
0.025
88.5
+
Ni
1.076
0.204
81.0
+
V
2.690
0.822
69.4
+
Zn
0.609
0.065
89.3
+
Se
0.050
0.006
88.0
+
Sr
0.043
0.001
97.7
-
A1
5.678
0.484
91.5
Zr
0.015
0.001
93.3
Average 88.1 ± 7.1
-------
TABLE 13. ELEMENTS SHOWING POSITIVE ENRICHMENT TRENDS
Coal
Oil
Sb, As,
Ca, Mg,
B, Cr,
Sb, As,
Mn, Ni,
Cd, Cr,
V, Zn,
Co, Mn,
Se, Zr,
Mo, Ni,
V, Zn, Se
The importance in this bulk data is that it indicates either:
1. The FGD selectively removes certain elements, concentrating
the rest, or, more likely,
2. The mist eleminator system on the FGD has low removal effi-
ciency for small particles which contain the trace elements.
Further tests are needed in which bulk and size dependent data is taken
at the inlet and outlet of the FGD. However, the data shown should signal
a concern on the part of FGD designers that fine particle removal is as
important as SO2 removal to the environment.
Scrubber Generated Particulate Matter
During this test, a CCS and a modified Method 5 train were operated at
both the inlet and outlet of the FGD. The modified Method 5 train used a
cyclone and filter at the inlet but only a filter at the outlet of the FGD.
Cyclone/Filter oven temperatures were maintained near 175°C. Because both
oil and coal were burned during the tests at this industrial boiler, a
direct comparison of scrubber outputs was possible. Table 14 shows the SOg
concentration, mass loading, and average gas flow across the scrubber during
the test program. Even though the inlet (I) mass loading for coal versus
oil differed by a factor of 25, the outlet (0) mass loadings were within
10 percent of each other. These data were the first indication that the
scrubber output might be independent of input conditions, and that a base-
line emission rate for the FGD existed. In order to determine the validity
of this statement an analysis effort following the IERL-RTP Level 2 inor-
ganic analysis procedures developed by TRW was initiated.
C-40
-------
TABLE 14. FLUE GAS CONDITIONS ACROSS THE SCRUBBER
S02
Concentration,
ppm
Mass
Loading,
mg/m3
Average Gas
Flow Across
the Scrubber,
dscm/mi n
Coa\
201-1-1
1114
7180
89.3
O
1
1
o
CVJ
33.8
48
201-4-1
913
8660
201-4-0
48.2
53
Oil
202-1-1
2969
452
202-1-0
33.2
55
202-4-T
1009
182
152.7
202-4-0
22.2
46
In Table 15 the sulfate extraction values for the 201-1 (coal) and the
202-4 (oil) samples are tabulated. The sulfate values for water and acid
soluble sulfates are shown as well as the total inlet (I) and outlet (0)
concentration. While the coal showed a greater inlet concentration for
sulfate than the oil, both the coal and oil outlet concentrations are rela-
tively close. Surprisingly, the oil has a higher outlet concentration.
The higher flue gas velocity in the scrubber during the oil tests could
have reduced the efficiency of the mist eliminator system, reducing the
particle removing capability of the scrubber.
Studying Table 15 further, there is an apparent change in the chemical
form of the outlet sulfate versus the inlet sulfate. This change is clearly
shown from the difference in the solubility between the water extractable
and the acid extractable portions of the inlet and outlet 201-1 filters.
Because the sulfate found represented 57 percent and 40 percent of the total
outlet particle catch for oil and coal respectively, an effort was made to
identify the chemical species inlet and outlet to the FGD. In this way any
modification of the particle matter by the FGD could be traced.
-------
TABLE 15. TOTAL SULFATE VALUES FROM THE MODIFIED METHOD 5 TRAIN
Sulfate Extraction Values (mg/m3)
Sample
Hot H20 Extraction
HNQ3 Extraction
Total
?ni-l Coal Inlet
1
Probe
j Cyclone (>3m)
I Filter
1
<100
36.7
0.2
107.0
10.2
107.0
46.9
Inlet Total
36.7
117.4
154.1
201-1 Coal Outlet
Probe
j Filter
18.6
0.5
19.1
i
I Outlet Total
i
18.6
0.5
19.1
202-4 Oil Inlet
Probe
Cyclone (>3y)
Filter
10.4
29.3
29.0
0.7
1.0
29.0
11.1
30.3
Inlet Total
39.7
30.7
70.4
!
202-4 Oil Outlet
i
Probe
Filter
24.5
3.5
<0.33
3.5
i 24.5
1 ...
j Outlet Total
24.5
3.5
28.0
C -42
-------
ESCA analysis of the inlet and outlet filter material showed only
S(VI) present even after Ar+ etching was employed to remove 150 A of
material. A series of Fourier Transform IR (FTIR) scans were made directly
from the inlet and outlet filters using attenuated total reflectance tech-
niques. In the coal case, the inlet filter showed only a few weak lines
providing little anion information. The IR bands for the coal outlet fil-
ter are shown in Table 16. A search of typical sulfate compounds showed
little correlation with these bands. However, Na-HSC^-HgO had several bands
(Table 17) that matched the spectra. The size of the water band at 1660 cm"1
indicated a high degree of hydration, and coula explain the differences in
the sample spectra and the reference NaHSO^. The fact that the filter
material was deliquescent is in line with the nature of a bisulfate.
TABLE 16. FTIR BANDS, cm
Coal Filter
NaHS04 • H20
Oil Filters
Outlet
Inlet
Outlet
653 (w, sh)
650 (sh)
773 (w)
775 (m)
857 (s)
865 (s)
863 (s)
872 (s)
1038-1160 (s, b)
1040-1080 (b)
1047 (s)
1040 (s)
1235 (m)
1150-1350 (vb)
1155 (s)
1160 (s)
1660 (b)
1645 (b)
1645 (b)
XRD analysis was able to confirm the presence of NaHS04 on the outlet, but
not the inlet filter. The implication from these results was that the
NaHS04 could have resulted from a scrubber generated emission of NaHSOj,
which was oxidized during collection and storage.
The FTIR and XRD analyses of the oil filters were not as conclusive.
The FTIR spectra of the inlet and outlet filters was similar and thus in
both samples NaHSO^ was indicated. X-Ray Diffraction could not confirm the
presence of NaHSO^ in either sample, because of its low concentration
(<1 percent) or the high degree of moisture in the samples.
C -43
-------
TABLE 17. H2S04 CONCENTRATION, mg/dscm (ppm)
Run
Inlet
Outlet
Scrubber Efficiency (%)
201-1
20.7* (5.09)*
14.3 (3.51)
31
201-4
15.5* (3.80)*
11.0 (2.69)
29
Average 201
18.1
12.7
30
202-1
25.2 (6.19)
17.9 (4.39)
29
202-4
31.6 (7.74)
22.7 (5.58)
28
Average 202
28.4
20.3
29
*Values corrected (+12%) for loss of H2SO4 to fly ash on filter.
Because the HSQ4 was f°uncl on the inlet and outlet oil sample filters,
it could have been an artifact from H2SO4 collecting on the filters.
Table 16 shows the results of the H2SO4 sampling tests. Interestingly, the
oil fired case had much higher H2SQ4 values. This is possibly due to the
high V content found in the flue gas. However, even these H2SO4 levels
were not high enough to account for all of the SO4 seen on the outlet fil-
ters. In the coal and oil samples an excess of 6.4 and 7.7 mg/m3 of SO4
respectively was unaccounted for. Once again outlet values for sulfate are
quite close, even though the inlets differed greatly.
One final test was run on the outlet filters using Secondary Ion Mass
Spectrometry. In this method, layers of a particle are removed, ionized,
and analyzed by the mass spectrometer to produce an elemental depth profile
of the sample. At a depth of 1400 A the mole percent of Na suddenly
increased from 1.33 to 22.34% for the inlet to outlet coal filters. Even
correcting for a 1% contribution from the filter background for the outlet
sample, there was a distinct change in the bulk particle composition between
the inlet and outlet. The mole percent from Na for the oil filter data
(7.2 vs 10.4%) showed an increase of roughly 50%, but since both inlet and
outlet oil filter samples were analyzed directly on the filter, the large
background Na of the filter probably masked any change in the sodium
concentration.
C -44
-------
In summary the following statements can be made:
• Outlet mass loadings for coal and oil combustion in the
industrial boiler were within 10 percent of each other.
• Sulfate outlet values for oil and coal combustion are
relatively similar considering the difference between
inlet conditions.
• Outlet sulfate species are more water soluble than the
inlet species.
• FTIR analysis confirmed the presence of NaHS04 in the coal
combustion outlet samples.
• XRD analysis confirmed the presence of NaHS04 in the coal
combustion outlet samples.
• Outlet H2SO4 concentrations were not adequate to account
for all of the sulfate collected.
Based on these findings, it is believed that the baseline scrubber
emission of NaHS04 is on the order of 6 to 7.7 mg/m^ (the difference of total
sulfate minus H2SO4 concentration). If one assumes that only 50 percent
of the H2SO4 was collected because of the high filter temperature (175°C),
then the scrubber contribution could be as high as 12 to 18 mg/m^.
It should also be considered that the reason for the clear-cut enrich-
ment of Na in the outlet 201-1 case was a result of operating conditions.
Either an operational problem or error could have caused Na loss during
the coal run. Further investigation of FGD operation during 201-1 tests
should be made to isolate any engineering problems.
C-45
-------
REFERENCES FOR APPENDIX B
1. Cleland, J.G.; G.L. Kingsbury. Multimedia Environmental Goals for
Environmental Assessment, Volume I and II. EPA-600/7-77-136a,
(1977).
2. Lee, R.E. et al. Environ. Science and Technical, £ (7), 643, (1975).
3. Davison, R.L. et al. Environ. Science and Technical, 8 (13),
1107 (1974). ~
4. Lee, R.E. and von Lehmden, D.J. JAPCA, 23 (10), 853, (1973).
5. Toca, F.M. Diss. Abstr. Int», 33, 3156B (1973).
6. Maddalone, R.F., S.F. Newton, R.G. Rhudy, R.M. Statnick. "Laboratory
ad Field Evaluation of the Controlled Condensation System (Goksoyr/
Ross) for SO3 Measurements in Flue Gas Streams." National APCA
Meeting, (1977).
C-46
-------
APPENDIX C
ATTACHMENT A
TGA SCANS
C -47
-------
. »\ \; f* _. i VOOUo
\ bl>AX:S
Ijsi- \ »r -1*. hs?*.
=>J i£C
!pr
-S.
-------
¦¦¦! iO. vV;C'.13
:! 2k-hT ! -axis
HS2<5M 13CAi_E. -*c
. ..CAT-J'
> Cfi 13CAi_E.
__ SCALE, rr.g/in.
.2.
SUPPRESSION, mg
WEIGHT, nr.g
T.'.VJE CONST, sec J
dY. £ ma/rr.in J/in
TMA
SCALE. r-i:s/in_
(VI DDE
SAV.PLE SIZE
Load, g „
jaY. [1QX ), [rniis/rr.nVin .
I
-Pk
-------
t*A.V7 MO. 9TCOD3
jp.u.M no date^^It?
^OPE^ator
SAMPL E:
IaI- ^'1 -C^tcone
ATM.
PLOW RATE
T-AXIS
SCALE. 'C/in SO
PROG. RATE. *C/min_l0_
HEAT. _ COOI ISO.
SHIFT, in 0
DTA-DSC
SCALE. "C/in .
(mcal/ce
WEIGHT, mg
:3/in_
SCALE, me/in.
JL
SUPPRESSION, mg
WEIGHT, ma _
TIME CONST., sec i_
dY, (nng/min J/in .
TMA
SCALE, mils/in.
MODE
SAMPLE SIZE .
LOAD. Q _
dY, CIOX). tmils/minVin
<
in
o
TEMPERATURE. *C £ CHROMEL/ALUMEL}
-------
NO.
SKlJN mo DATE
^C—ER-^TOR 1
-------
PART S?0088
i
in
ro
¦Wilt-ft;
;3LjN; \3 DATE
iO^EPATOR _A< SSiH5rj\H
iSAViPLE:
j T-ol.-V
CM-<*
! AT!/, fn
l=LOW RATE N.v
T-AXiS
SCALE. °C/in 50_
PBOG. BATE. "C/min
HEAT COOI ISO.
DTA-DSC
SCALE. "C/in .
Cmcal/sec)/in_
WEIGHT, rng
REFERENCE
TGA
mg/in.
SUPPRESSION, mg
WEIGHT. mo_13-j?.51_
TIME CONST., sec L
dY. tm^minj/in .
TMA
SCALE, mils/in.
MODE
SAMPLE SIZE.
LOAD, g
dY. C10XJ. Cmils/min >/in
• -rr
TEMPERATURE. 'C t CHROMEL/ALUMEL )
-------
APPENDIX C
ATTACHMENT B
PLM PHOTOGRAPHS
C -53
-------
FIGURE B-l. 202-4-I-Cyclone, showing oil soot collected during oil
combustion partially uncrossed polars (pup) 51X
-------
FIGURE B-2. 201-1-Flyash, showing oil soot typical of that
collected during coal combustion; PUP, 131X
-------
FIGURE B-3. 201-Flyash, showing partially fused flyash, oil soot, and ion
oxide; transmitted and reflected light (R&L), PUP, 51X
-------
FIGURE B-4. 201-1-I-Cyclone, showing partially fused
flyash and oil soot; R&L, PUP, 51X
-------
0
1
cn
CD
FIGURE B-5. 201-1-I-Cyclone, showing the flyash (some
with air bubbles); PUP, 131X
-------
FIGURE B-6. 201-1-I-Filter, showing partially fused flyash and oil soot,
also, overall smallness of particle size; R&L, PUP, 51X
-------
c~>
I
cr»
o
FIGURE B-7. 201-1-I-Filter, showing flyash; PUP, 131X
-------
FIGURE B-8. 201-1-Scrubber cake, showing partially fused
flyash and oil soot; R&L, PUP, 51X
-------
r->
I
cr»
ro
^ »• v. "
-* > • dttsM "• ^
• .' ^ ' m-^^rn V 4
4 3? v* - .-JS* ' *- *t ^ - *
¦•¦¦ * ¦" tfi -AT f ^
"B &| _*» *
* ' 5 ¦ Jfr'** JjfV' • •
Kt£> '* -V. >.->*,
„ . ^ . j« •}
FIGURE B-9. 201-1-Scrubber cake, showing CaSo3. 1/2 H20 laths; PUP, 131X
-------
I
•aft
«
FIGURE B-10.
201-1-Scrubber cake, showing foreign particulates
incorporated in spherilites; PUP 131X
-------
FIGURE B-ll. 201-1-0, showing carbonaceous coating on material
embedded in the filter.
-------
FIGURE B-12. 201-1-0, after LTA revealing minerals; R&L, PUP, 51X
-------
FIGURE B-14. 202-4-I-Cyclone, showing crystals of an unknown
sulfate oil soot; PUP, 131X
-------
0
1
CD
O0
0
FIGURE B-15.
202-4-1, showing carbonaceous particles on briefing
on the background; R&L, PUP, 51X
-------
FIGURE B-16. 202-4-1, after LTA, revealing mineral background;
R&L, PUP, 51X
-------
FIGURE B-17. 202-4-scrubber cake, showing spherilities;
PUP, 131X
-------
FIGURE B-18. 202-4-scrubber cake, spherilite with first order red
plate compensator; completely crossed polars, 131X
-------
FIGURE B-19. 202-4-scrubber cake showing spherilities, laths
and plates of CaSC^. 1/2 ^0; PUP, 131X
-------
FIGURE B-20. 202-4-0, showing sulfate crystals on filter,
as well as the blue droplets; PUP, 51X
-------
FIGURE B-21. 202-4-0, showing another crystal form on filters
PUP, 51X
-------
FIGURE B-22. 202-4-0, showing recrystal1ization on filter,
after sitting; PUP, 51X
-------
FIGURE B-23. 202-4-0, showing another type of recrystal1ization
on filter; PUP, 51X
-------
FIGURE B-24. 202-4-0, after LTA, revealing minerals; R&T, PUP, 51X
-------
C-78
-------
APPENDIX C
ATTACHMENT C
FTIR SPECTRA
C -79
-------
TRW-201-1 -1 -CYCLONE-MIN. 01L FCR. 0974
MDRC-HB CHEM tflB
o
O ijj
I
00
o
itbo lsbo i2bo lioo lobo
WRVENUMBERS
TS00 1800
900
700
800
BOO
500
-------
o
CD 1
O
O
a)
a
a
o
uJ
a
a
in
a
a
jf
LT
LJ
^ O
So
*
o
i— —.
!- IP
1
«—•
oo
r
01 a
z a
L_ '
a: to
¦— j-
n
n
I/1
a
n
o
a
03
n
a
a
Cfl.
*1
"1"
T"
1 -FlILTIIR
cui 7 i ~r :il i !-n
CORRECTED FORiMINEROL OIL
-------
I
00
ro
SiiDO 1300 1200
1100 1000
WnVENUMBERS
loi-j-i-OMtiiNuq blank] hinds'
RIR
LE
SUBtRRqrim RESULT
'^01 ~
II - I i (MljNUS, MINERAL. OIL)
.00
00
.28
K SRMPL
~9 * RE
E MINUS
iFERENCIj:
:~1
600
..L-
-------
TRW ; ¦
fcOl-Jl-FlYRBH
pORlfECTED FOR M1NERRL OIL
NICOLE! 7199 FT
o
o
CJ
7B0
s5o
900
WRVENUM3ERS
1000
BOO
600
-------
oo
a
o
NTCQLElj ¦ tlM ffT
'V
iobo
WRVENUMBERS
t "f" "i r~ v
I ! I i !
U i ! i ! ' i
201rl-SCRUBBER CRKE
CORRECTED FOR MINERAL OIL
-------
a
in
cn _
m
a
in
MDRC-HB CHEM LRB
in
o
iT.
ci „
O
in
.
a
" y .
i if)
00 ; **
01 ^r,
t. m
CO •
z (n
rz ¦*"
tz
•"O
in
o
in
a
in
o
in
•
(n
o
in
in
n
*\
i8bo i7bo TbBo isbo 11 bo isoo
II r
.t.
: i
t-
TRU-202-%-1-CYCLONE - M1N.0IL FCB.05225 SMOOTHED 9 PTS
1200 llbO 1000
WRVENUMBERS
sBo *bo
-------
NICQLET 7199 FT-IF
Ttbo
i ibo lofio gbo ado
WflVENUMBERS
! ;
TRW
SAMPLE FLYflSH 3LPNK
CORRECTED . FOR M i N..P.HL OIL
-------
o
CO
I
00
MDflC-HB
CHEM LRB
1800 I7b0 16b0 1500 1100 1300
TRW-202-%-SCRB• CRKE - MIN.OIL FCR.0<»177
-------
12D0 1100 1OdO
900
i COLE: I; 7199
[TBI
202
'LEiSUqTRflCTIlJlN RESULT
-H-O-Flii-TER (NINJS R-flNK I
¦tINyS 202-4-1 I iMiNUS BLQNK
CHrl E ONB NUlj-I EQ
66*
JO 22 bO 2000 1900 1800 1700 1600 1500
03 3*00 3200 3000 2800 2600
~OVENUMOERS
-------
NICOLET 7199 FT-IR
URVENUHBERS
SUBTRACTION F.CSdLT
aox-'t-n-FJi.Trn i:njs pii
O.aObQ » S!3M«-!_F :i;nU3
. I .QOOO « RtTeriEKLE
/DO
oOO
1300 tzbo
bo
-------
iuIDTR!
t~
39 ff T|- I
r
ioo i zoo i loo jobo
DO 2ZDO 20001900 1800 1700 1600 1500
go:
900
UPVENUMBERS
-------
NICGLET 7199 FT-IR
TT5o 3203 3cb0 2903 26b0 2
-------
C -92
-------
APPENDIX C
ATTACHMENT D
ESCA SPECTRA
C -93
-------
® ESCfl SUR
PE= 100
KG. 10.0KV. 40-OUR
rae* cc
12/13/77
SF= 1.465
DRT= 5.00
tot i I cr
loo.oa
CM >00
CM. SO
900.00
tOS.00
BINDING ENERGY» EV
-------
$ CSCft suit
PCs 100
no. io.okv. 40.our
riu» ca
12/13/77
5F= 1.465
DflT= 5.00
6IN0IN0 ENER&T, CV
-------
0 ESCfi SUR
PE= 100
no. 10.0KV. 40.OUR
riui cs«
12/14/77
SF= 1.9S0
DflT= 7.SO
SM.00
BINDING ENERGY. EV
-------
<£ CSCA SUR 12/13/77
P€r 100 SFs 2.926
no* 10.0KV. 40.QUA OflT= s.oo
riu* cii ni it mrai
o
«o
•M
m
•M
»
BINDING ENERGY. EV
-------
® ESCft SUR
PE= 100
tlG. 10.0KV, 40.00ft
FSLCt CM
12/14/77
SF= 2.925
DRT= 5.00
hi 11 » « »
Xji
MO.M
100.M
7O0.M
BINDING ENERGY, CV
-------
$ ESCft SUP
PE* 100
HG» 10-0KV, 40.OMR
riui cv
12/14/77
SF= 2.925
DATs 5.00
Ml I f ICO A Sf
1M.M
BINDING ENERGY» EV
-------
® ESCR SUR
PE= 100
HG. 10.0KV, 40.0ft«
riu* cm
SF= 2.925
DRT= 5.00
Nl i I m R tf
i j
12/14/77
I
BINDING ENERGY* EV
-------
® escr sur
PCs ioo
NG. 10.0KV, 40.ONA
rzut cm
12/14/77
SFs 2*925
DRT= S.00
CM I • CM ft 1
If-
BINDING ENERGY" » EV
-------
$ ESCA SUR
PEa 100
KG. 10.0KV, 40-OUR
POe* C44
12/14/77
SF= 2.S2S
DAT* 6.00
cot i • no a tt
IM.M
BINDING ENERGY» EV
-------
$
ESCft SUR
PC= 100
HG. 10.0KV, 40.onA
rtui cm
12/14/77
SF= .366
ORTs 13.34
\
BINDING ENERGY # CV
-------
$ ESCfl SUR
PE= 100
HG. 10.0KV, 40.OUR
fXLC. CC
12/13/77
SF= .975
DflT= 6.00
Ml 1 rl-T
<00.00
0 4M.M
BINDING ENERGY » EV
-------
$ CSCA SUR
PCs 100
HG# 10.0KV. 40.0HA
nut c«
12/13/77
SFs 1-465
DATs S.00
Ml I SC
/
BINDING ENERGY. EV
-------
($ ESCft SUR
PE= 100
Mr. 10.0KV.
met ci
40*ann
0
1
o
I at
0\Lt>
7M.M
12/13/77
SFs 1.46S
DflT= 6.00
tot « i cr
W~J
y
BINDING ENERGY. EV
-------
$ ESCA SUR
PEs 100
HG. 10.0KV, 40.OUR
rn£« ca
12/13/77
SFs 1.465
ORT= 6.00
m «i ram
BIHOING ENERGY. EV
-------
(J ESCfl SUR 12/1 i/77
PE= 100 SF= 2.340
10.0KV. 40.OUR DRT= 6.28
FtLC* at tot 4 I FIITC* 7$ ft *
t&l
09
loa.ao
.00
MO.M
000.00
BINDING ENERGY» EV
-------
® CSCA MJL CLs C 1 12/13/77
PEs 50 SF= .048
HO. 10.0KV, 40.0HR DATs 10.10
ID
tat
.M
t/S
BINDING ENERGY. EV
-------
J) ESCfl SUR
PE= 100
HG. 10.0KV, 40.0f1fl
FIUi Ctt
12/13/77
Sf= 1.466
DflT= 5.00
cot « o FILTER
i i
WO. 00
100.00
*00.00
BINDING ENERGY, EV
-------
$ ESCfl SUR 12/14/77
PCs 100 Sf= 2.926
nc. io.okv. 40.ona ort= io.oo
FtLCi on
4 *7t « «r /-//fc*.
r»
.M
BINOINC CNCRGY• EV
-------
{£ ESCfl SUR 12/14/77
PC= 100 SF= 1.462
no, lo-OKV. 40.onn dat= io.oo
HUE» CM tOt 4 • 7« ft ST F.dUh
¦ —i * . -»——.— -i -1 »
BINDING CNCRGY. €V
-------
$ CSCA SUR
PEr 100
HG. 10.0KV, 40.OMR
riui c*
12/13/77
SF= 1.465
DATs 6.00
CM 4 *C
IN.00
BINDING ENERGY* EV
-------
$ ESCfi SUR
PE= 100
flG. 10.0KV, 40.OUR
FKCt CO
12/13/77
SF= 1.465
DflT= 5.00
fit TCI tUWK
> I
100.00
•00.00
400.00
000.00
700.00
BINDING ENERGY» EV
-------
$ ESC* SUR
PCs 100
nc. 10-0KV, 40•ona
P1LC* CM
12/14/77
SF= 3.900
DATs 3.75
FZVTCM «UWR 7S A
BINDING ENERGY• EV
-------
c -116
-------
APPENDIX C
ATTACHMENT E
SIMS DATA DEPTH PROFILE PLOTS
C -117
-------
201-1-1- CYCLONE
00
OJ
O
h-
a
u
o
H
i—
a:
oc
H
CO
u
h—
M
w h a n a
ida—'—
nsr
TIME IN SECONDS
2300
o - 7 / 28 - LITHIUM X 468
* - 23 / 28 - SODIUM X 26
0 « 2f / 28 - MAGNESIUM X 106
+ - 27 / 28 « ALUMINUM X 5
~ - 39 / 28 - POTASSIUM X 12
v - 40 / 28 - CALCIUM X 29
PILE NAME 1-IC2
C -118
-------
201-1-1- CYCLONE
00
CNJ
O
I—
a
u
O
H
h-
CE
a:
05
z
u
H-
21
H
2000
TIME IN SECONDS
o ¦ H8 / 28 ¦ TITANIUM X 95
* « 51 / 28 - VANADIUM X 1411
0 - 52 / 28 ¦ CHROMIUM X 2295
+ - 55 / 28 - MANGANESE X 2678
a - 56 / 28 - IRON X 25
v - 88 / 28 ¦ STRONTIUM X 675
a » 138 / 28 • BARIUM X 2936
FILE NAME 1-IC2
C -119
-------
201-1"I LOOSE FILTER
a,
• > '
% «
O
h-
a
ui
o
H
I—
CE
C£
y-
h-
H
CO
U
H
§09
WS
Z33
5000
TIME IN SECONDS
o « 7 / 28 « LITHIUM X 59i
* « 23 / 28 - SODIUM X 30
0 « 2* / 5.8 - MAGNESIUM X 122
+ « 27 / 28 - ALUMINUM X f
~ - 39 / 28 - POTASSIUM X It
* - *0 / 28 - CALCIUM X 21
FILE NAME 1-IF3
C -12Q
-------
201-1-1 LOOSE FILTER
m
time in second:-:
o - *8 / 28 ¦
* - 5; / ze ¦
O - 52 / 28 »
+ - 55 / 28 ¦
a ¦ 56 / 28 B
~ ¦ 88 / 28 ¦
* ¦ 138 / 28
TITANIUM X 88
VANADIUM X 1116
CHROMIUM X 183*
MAN6AQE8E X 2395
IRON X 29
STRONTIUM X 61*
' BARIUM X 2190
FILE NAME 1-IF3
C -121
-------
201-1 O-PF
00
CM
O
h-
Q
U
O
H
I-
CE
CC
H
CO
z
UJ
100
TIME IN SECONDS
o - 7 / 28 » LITHIUM X 520
« « 2? / 28 - SODIUM X 5
0 . 25 / 28 - MAGNESIUM X 511
+ - 27 / 28 - ALUMINUM X 58
o « 39 / 28 - POTASSIUM X 15
* - 50 / 28 - CALCIUM X 53
FILE NAME TJW13
C -122
-------
201 1-0 PF
00
CM
a
u
o
H
i—
-------
202-^-i-cyCLONE
00
CM
o
a
u
o
H
h-
tr
C£
>-
h-
M
CO
zz.
u
M
W1 % Q Qi
m
m
Tm
TIME IN SECONDS
o * ? / 28 « LITHIUM X 11?
* - 23 / 28 - SODIUM X *
* - 2* / 28 - MAGNESIUM X 27
+ - 27 / 28 - ALUMINUM X 3
a - 39 / 28 « POTASSIUM X 2
* » 10 / 28 - CALCIUM X 2
FILE NAME 202CY
C -U4
-------
20 2- '-f-I--CY CLONE
c&
CM
O
LJ
o
H
cn
IX'
>-
i*-
M
'/•
Z
u
h~
z
H
TIME IN SECONDS
o « H8 / 7.8 - TITANIUM X cB
* - 5; / 28 - VANADIUM X S
0 « 52 / 28 « CHROMIUM X 77
+ * 55 / 28 - MANGANESE X 210
o - 56 / 28 ¦ IRON X *»
* ¦ 88 / 28 * STRONTIUM X 80
4 - 138 / 28 ¦ BARIUM X 153
FILE NAME 202CV
-------
20 7- 't-I-PRPER 5
oo
CM
O
h-
a
u
o
H
I—
a:
ee
H
CO
u
H
»¦ M
m
m
m
imr ~Ti
TIME IN SECONDS
2000
o « 7 / 28 - LITHIUM X 1330
* » 23 / 28 - SODIUM X 3
0 - 24 / 28 - MAGNESIUM X 14
+ - 27 / 28 - ALUMINUM X 7
a - 39 / 28 - POTASSIUM X 4
v - ^0 / 28 - CALCIUM X 8
FILE NAME 24FP
C -126
-------
20/:-^-I-PftPER 5
CO
CM
O
I—
q
uJ
a
H
cE
oc
>-
\—
H
CO
•z.
u
H
1200 1
TIME IN SECONDS
OS
5&
o - 48 / 28 ¦ TITANIUM X
* - 51 / 28 - VANADIUM X 1
0 - 52 / 28 - CHROMIUM X 135
+ ¦ 55 / 28 - MAN6ANE8E X 411
a - 54 / 28 - IRON X «
~ - 88 / 28 - STRONTIUM X 270
* - 198 / 28 - BARIUM X 251
FILE NAME 21FP
C -127
-------
202-4-0 FILTER
00
CM
O
a
u
o
H
I—
a:
en
H
CO
u
H
Sib
da
T*00
2000
TIME IN SECONDS
o
*
+
~
7 / 28 "
23/28
24/28
/
/
/
27
39
40
28
28
28
LITHIUM X 21110
i SODIUM X 6
' MAGNESIUM X 50
' ALUMINUM X 23
> POTASSIUM X 15
' CALCIUM X 1*
FILE NAME TRV9
C -128
-------
202-^-0 FILTER
CO
OJ
O
a
uj
o
H
h-
CC
CC
H
Cf>
z
UJ
M
TIME IN SECONDS
o « 18 / 28 - TITANIUM X 577
* « 51 / 28 - VANADIUM X 20
* • 52 / 28 - CHROMIUM X 1903
+ - 55 / 28 - MANGANESE X 1838
o « 56 / 28 - IRON X 82
~ • 88 / 28 - STRONTIUM X 3907
* • 138 / 28 » BARIUM X 12138
FILE NAME TRW9
C -129
-------
201--1--I
!* U
CO
c\j
o
h-
Q
U
O
M
h-
TIME IN SECONDS
o « 7 / 28 - LITHIUM X U54
* - 23 / 28 - SODIUM X
0 • 2H / 28 « MAGNESIUM X 29?.
+ - 27 / 28 ¦ ALUMINUM X 3?
n « 39 / 28 - POTASSIUM X •»
v « 40 / 28 - CALCIUM X 5
FILE NAME 1-IFC
C -13Q
-------
r/> <
-1 -I-F
CO
OJ
o
ui
o
H
£
c£
>-
h-
H
CO
Z
UJ
H
ran
vn
800
TIME IN SECONDS
o . 48 / 28 « TITANIUM X *4
* - 51 / 28 - VANADIUM X 514
0 - 52 / 28 - CHROMIUM X 924
+ - 55 / 28 ¦ MANGANESE X 98
a - 56 / 28 ¦ IRON X 3
v - 88 / 28 « STRONTIUM X 198
* ¦ 138 / 28 ¦ BARIUM X 1097
PILE NAME 1-IFC
C -131
-------
202-H SCRUBBER CRKE
100
00
CNf
O
a
ui
o
M
h-
a:
cc
M
CO
21
u
H
"530 1200" 1600
TIME IN SECONDS
2000
0
m
*8
/
28 -
TITANIUM X 70
*
m
51
/
28 -
VANADIUM X *80
m
32
/
28 -
CHROMIUM X 365H
+
-
55
/
28 -
MANGANESE X 320
a
¦1
5*
/
28 -
IRON X 2
v
«
88
/
28 -
STRONTIUM X 158
A
a
138 / 28 - BARIUM X 7697
FILE NAME TRV12)
C -132
-------
202-* SCRUBBER CfiKE
a?
CM
a
u
o
H
i—
a:
DC
M
if)
Z
u
z
M
1000
800 -
600 -
*00 -
TO
52»
m
xm
m*
TIME IN SECONDS
o - 7 / 28 - LITHIUM X 735
t - 23 / 28 - SODIUM X 1
0 - 2* / 28 - NA6NESIUN X 82
~ - 27 / 28 « ALUMINUM X 118
a - 39 / 28 « POTASSIUM X 17
v « 40 / 28 - CALCIUM X 3
FILE NAME TRV10
C-133
-------
C-134
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/7-78-164C
3. RECIPIENT'S ACCESSION NO.
•*.title anosubtitle Environmental Assessment of Coal-
and Oil-firtng in a Controlled Industrial Boiler;
Volume m. Comprehensive Assessment and Appendices
6. REPORT DATE
August 1978
6. PERFORMING ORGANIZATION CODE
7.author
------- |