SEPTEMBER
Prettdion Ttchnology Striti
v»viV**«*i^i*«v»V«V«V*VV*V»V«%v«**V«*«v.v«*i
*
-------
EPA-650/2-74-104
REDUCTION OF ATMOSPHERIC POLLUTION
BY THE APPLICATION
OF FLUIDIZED-BED COMBUSTION
AND REGENERATION
OF SULFUR-CONTAINING ADDITIVES
by
G. J. Vogel, W. M. Swift, J. F. Lenc, P. T. Cunningham,
W.I. Wilson, A. F. Panek, F. G. Teats, and A. A. Jonke
Argohne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
Interagency Agreement No. EPA-IAG-149(D)
ROAP No. 21ADB-011
Program Element No. 1AB013
EPA Project Officer: D.B.Henschel
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF COAL RESEARCH
DEPARTMENT OF THE INTERIOR
WASHINGTON , D. C. 20240
(OCR IAG 14-32-0001-1543)
and
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
September 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
TABLE OF CONTENTS
Page
ABSTRACT 10
SUMMARY 11
INTRODUCTION 22
BENCH-SCALE, PRESSURIZED, FLUIDIZED-BED COMBUSTION
EXPERIMENTS 25
Raw Materials 25
Coals 25
Additive 25
Inert Bed Material 25
Bench-Scale Equipment 26
Experimental Procedure 29
Combustion Experiments with Arkwright Coal 29
Sulfur Dioxide Retention 32
Analysis of Variance of S02 Data 32
Regression Analysis of S0£ Data 33
Kinetics of S02 Retention 34
Nitrogen Oxide Emissions 38
Combustion Efficiency 39
Combustible Carbon Hold-up in the Fluidized Bed. . . 42
Additive Utilization 42
Removal of Additive from Combustor by Elutriation. . 45
Terminal Velocities of Additive Particles ... 45
Experimental Elutriation Rates 45
Decrepitation of Tymochtee Dolomite 43
Heat-Transfer Coefficients at the Combustor Wall . . 49
Solids Loading in the Flue Gas 52
Carbon, Sulfur, and Calcium Material Balances. ... 55
Particle-Size Distribution of Fluidized-Bed and
Elutriated Solids 56
Combustion-Side Corrosion of Internal Cooling
Coils 58
Combustion of Low-Sulfur Subbituminous and Lignite Coals. 61
Operating Performance 61
Sulfur Dioxide Retention 61
Nitrogen Oxide Emissions 64
Combustion Efficiencies 64
Calcium Utilization 64
Material Balances and Screen Analyses 64
-------
TABLE OF CONTENTS (Cont'd.)
Effect of Combustor Pressure and Temperature on the
Concentration of NO and Other Gases in the Flue Gas. ... 65
Pressure Effect on NO Levels in the Flue Gas 66
Pressure Effect on the Concentration of Flue-Gas
Components Other Than NO 67
Effect of Temperature on the Concentration of Flue-
Gas Constituents 67
TRACE-ELEMENT DISTRIBUTION STUDIES 70
Solids Sampling 71
Flue-Gas Sampling 71
Analytical Methods 73
Results 73
Mercury Mass Balances 75
Iodine Monochloride Scrubbing Solution 77
Condensation of Mercury on Flue-Gas Ductwork . . 77
Efficiency of Mercury Sampling Apparatus .... 77
Retention of Mercury in Solid Products of Combustion. 78
Lead and Beryllium Mass Balances 78
Preferential Concentration of Lead and Beryllium in
the Finer Particulate Matter 81
Fluoride Mass Balances 84
Neutron Activation Analysis, Developmental Work ... 84
Sodium Concentrations in Flue-Gas Particulates. ... 86
Photomicrographs of Flue-Gas Particulates 86
KINETICS OF THE REACTION OF HALF-CALCINED DOLOMITE WITH SULFUR
DIOXIDE 89
Experimental 90
Apparatus 90
Materials 90
Procedure 91
Results and Discussion 92
ACKNOWLEDGMENTS 97
REFERENCES 98
-------
TABLE OF CONTENTS (Cont'd.)
Page
APPENDIX A. Characteristics of Raw Materials Used in Fluidized-
Bed Combustion Experiments 101
APPENDIX B. Carbon, Sulfur, and Calcium Material Balances for
Combustion Experiments 107
APPENDIX C. Screen Analysis Data 121
APPENDIX D. Concentration of Fine Particulate Matter in the Flue
Gas Exhausted from the Fluidized-Bed Combustion System. 126
APPENDIX E. Trace-Element Analytical Procedures 129
APPENDIX F. Conversion Factors, English to Metric Units 134
-------
LIST OF FIGURES
No. Title Page
1. Simplified Equipment Flowsheet of Bench-Scale Fluidized-
Bed Combustor and Associated Equipment 27
2. Detail Drawing of 6-in.-Dia, Pressurized Fluidized-Bed
Combustor 28
3. Effect of Ca/S Mole Ratio on S02 Level in Flue Gas. ... 35
4. Effect of Superficial Gas Velocity on S(>2 Level in Flue
Gas 36
5. Effect of Bed Temperature on S02 Level in JFlue Gas. ... 37
6. NO Concentration in Flue Gas as a Function of Ca/S Mole
Ratio 40
7. Combustion Efficiency as a Function of Bed Temperature. . 41
8. Additive Utilization of Final Bed Material as a Function
of Ca/S Mole Ratio 44
9. Terminal Gas Velocity of Solid Particles in a Gas Flow
as a Function of Particle Diameter for Various Physical
Situations 46
10. Entrainment of Sulfated Dolomite as a Function of
Superficial Gas Velocity 47
11. Various Size Fractions of Primary Cyclone Product
Recovered during Combustion Experiment VAR-6-2R 49
12. Factors Considered in the Calculation of Heat-Transfer
Coefficients at the Combustor Wall 51
13. The Effect of Velocity on Calculated Bed-to-Wall Heat-
Transfer Coefficients 53
14. Solids Loading of Flue Gas Leaving the Combustor as a
Function of Ca/S Mole Ratio and Gas Velocity 57
15. Mean Particle Diameters of Solid Samples from the "VAR"
Series of Combustion Experiments vs. Fluidizing-Gas
Velocity 58
16. Samples Number 3 and 11, Showing Intergranular-Type
Corrosion on the Bed-Side Wall of an Internal Cooling
Coil Immersed in the Fluidized Bed 59
-------
LIST OF FIGURES (Cont'd.)
NO. Title
17 . Photomicrograph and Electron Microprobe Scanning Images
of Intergranular-Type Corrosion Found in Process-Gas
Side of Sample Removed from an Internal Cooling Coil
in the Freeboard Area (All Images 40 x 50 ym) ...... 60
18. Effect of Pressure on Concentration of NO in Flue Gas. . 66
19. Effect of Pressure on Concentration of Flue-Gas
Components Other Than NO for Series NOX-1 ........ 68
20. Effect of Pressure on Concentration of Flue-Gas
Components Other Than NO for Series NOX-II ....... 69
21. Effect of Temperature on Concentration of Flue-Gas
Components ....................... 69
22. Experimental Flow System for Flue-Gas Sampling ..... 72
23. Flue-Gas Sampling Apparatus for Measuring Concentrations
of Hg, F~, and Entrained Particulate Matter ....... 72
24. Particle Size Distributions for Fly Ash Recovered by the
Primary and Secondary Cyclones in Experiment TRACE-3 . . 81
25. Photomicrographs of Arkwright Coal and Fly Ash Recovered
in Particulate Removal and Sampling Devices During
Combustion Experiment TRACE-3 .............. 88
26. Schematic Diagram of the TGA Apparatus ......... 91
27. Percent Conversion vs. Time for Various S02
Concentrations in the Reactant Gas at 750°C ....... 93
28. Initial Reaction Rate of S02 with 1337 Dolomite as a
Function of S02 Concentration in the Presence of H^O . . 93
29. Initial Reaction Rate of S02 with 1337 Dolomite as a
Function of S02 Concentration in the Absence of H^O. . . 94
30. Arrhenius Plot for the Reaction of S02 with Half-Calcined
1337 Dolomite ...................... 94
31. Percent Sulfation vs. Time for a) [CaO + MgO] ; b) [CaC03
+ MgO]; and c) MgO at 750°C ............... 95
-------
LIST OF TABLES
No. Title Page
1. Designed Operating Conditions and Matrix Representation
for the Eleven Experiments in the Latin-Square
Experimental Design 30
2. Operating Conditions and Flue-Gas Analyses for the
Eleven Experiments Investigating the Effects of
Independent Operating Variables 31
3. Analysis of Variance of S02 Flue-Gas Levels in the 3x3
Latin-Square Designed Series of "VAR" Experiments .... 32
4. Comparison Between Experimental and Calculated Values
for the S02 Flue-Gas Levels in the "VAR" Experiments. . . 34
5. Analysis of Variance of NO Flue-Gas Levels in the 3x3
Latin-Square Designed Series of "VAR" Experiments .... 38
6. Combustion Efficiencies for "VAR"-Series Experiments. . . 41
7. Combustible Carbon Content of Final Bed Material for
"VAR"-Series Experiments 43
8. Calcium Utilization in Solids Outlet Streams for the
Combustion of Arkwright Coal in a Fluidized Bed of
Tymochtee Dolomite 43
9. Inventory of +45 Mesh Additive Entering and Leaving the
Combustion System for the Eleven Experiments in the "VAR"
Series of Combustion Experiments 50
10. Heat-Transfer Coefficients for the "VAR" Series of
Statistically Designed Combustion Experiments 53
11. Distribution of Flue-Gas Particulates in "VAR"-Series
Experiments Expressed in Units of Grains/scf 54
12. Distribution of Flue-Gas Particulates in "VAR"-Series
Experiments Expressed in Units of lb/106 BTU 55
13. Carbon, Sulfur, and Calcium Material Balances for "VAR"
Series of Experiments 57
14. Actual Operating Conditions and Flue-Gas Analysis for
Combustion Experiments with Low-Sulfur Subbituminous
and Lignite Coals 62
-------
LIST OF TABLES (Cont'd.)
No, Title Page
15. Potential Effect of the Calcium Content of Coals Tested
on the Retention of Sulfur during Fluidized-Bed
Combustion. 63
16. Trace Elements of Interest and Their Order of Priority
for Experimental Investigation 70
17. Summary of Average Operating Conditions and Flue-Gas
Compositions for Trace-Element Experiments 74
18. Mercury Material Balances for Trace-Element Combustion
Experiments 76
19. Lead Material Balances for Trace-Element Combustion
Experiments 79
20. Beryllium Material Balances for Trace-Element Combustion
Experiments 80
21. Concentrations of Trace, Minor, and Major Elements in
Particulate Matter Recovered at Various Stages of Removal
from Flue Gas 83
22. Ratio of Lead to Beryllium in the Raw Materials and in
the Particulate Matter Removed from the Flue Gas during
the "TRACE" Experiments 83
23. Fluoride Material Balances for Trace-Element Combustion
Experiments 85
24. Neutron Activation Results on Samples from the TRACE-3
Trace-Element Experiment 86
25. Sodium Material Balances for Trace-Element Combustion
Experiments 87
-------
REDUCTION OF ATMOSPHERIC POLLUTION BY THE APPLICATION
OF FLUIDIZED-BED COMBUSTION AND REGENERATION
OF SULFUR CONTAINING ADDITIVES
Annual Report
July 1973-June 1974
by
G. J. Vogel, W. M. Swift, J. F. Lenc, P. T. Cunningham,
W. I. Wilson, A. F. Panek, F. G. Teats, and A. A. Jonke
ABSTRACT
The Argonne National Laboratory (ANL) program for
developing and demonstrating the feasibility of fluidized-
bed combustion for possible use in power and steam-plant
applications is divided into three studies: (a) the
combustion of coal in a pressurized combustor; (b) a
determination of the distribution of trace elements in
the combustion products; and (c) a fundamental investigation
of the kinetics of additive sulfation and regeneration
reactions.
A bench-scale, fluidized-bed combustion pilot plant
capable of operating at 10-atm pressure* was used to evaluate
the effects of operating variables on response variables
such as S02 and NO levels in the flue gas, combustion
efficiency, additive utilization, and heat-transfer
coefficients. High retentions of sulfur (>90%) and low
NO levels (<150 ppm) were achieved. The combustor was
also successfully tested using a variety of coals: a
highly caking, high-volatile bituminous coal, a high ash
subbituminous coal, and a low-heating-value lignite.
Data, in the form of material balances, are reported
for the four trace elements of primary concern: Hg, Pb,
Be, and F. Also reported are data on Na concentrations
in the particulate matter entrained in the flue gas from
the combustor.
The kinetics of the reaction of half-calcined dolomite
with S02 were studied at atmospheric pressure by thermogravi-
metric analysis. The reaction was found to be first order with
respect to the S02 concentration in the presence of H20 vapor
and approximately three-fourths order in the absence of
vapor.
*
For convenience, a table of English to metric unit conversion
factors is provided at the end of this report in Appendix F.
10
-------
SUMMARY
Argonne National Laboratory is investigating pollution control
aspects of pressurized, fluidized-bed combustion in a program funded
by the Control Systems Laboratory of the Environmental Protection
Agency and by the Office of Coal Research, Department of Interior.
The program consists of investigating (1) the effects of operating
variables at elevated pressures (up to 10 atm) on S02 removal, NO
suppression in the flue gas, combustion efficiency, additive
utilization, solids entrainment, and heat-transfer coefficients,
(2) the ability of the system to process a variety of coals includ-
ing bituminous, subbituminous, and lignite coals, (3) the quantity
and type of trace-element contaminants in the flue gas, and (4) the
kinetics of the reaction between S02 and sulfur-accepting additives.
In the combustion studies, coal is completely combusted in a
fluidized bed of dolomite using an excess of oxygen. Sulfur con-
tained in the coal is released during combustion as S02, which in
the excess-oxygen environment reacts with the CaO in the dolomite to
form CaSOjj.
The reaction kinetics between S02 and dolomite were studied at
atmospheric pressure using a thermal gravimetric analyzer (TGA) unit.
Bench~Scale, Pressurized, Fluidized-Bed Combustion Experiments
Materials. Combustion experiments were made using a variety of
coals of differing rank. The principal coal tested was a highly
caking, high-volatile bituminous, Pittsburg seam coal from the
Consolidation Coal Company's Arkwright Mine. Also tested, although
less extensively, were a subbituminous coal from the Western Coal
Company's San Juan Mine in New Mexico and a lignite from the
Consolidation Coal Company's Glenharold Mine in North Dakota. The
coals were fed to the combustor as received.
Tymochtee dolomite obtained from C. E. Duff and Sons, Huntsville,
Ohio, was used in the experiments reported here. The additive was
air-dried and screened to - 14 + 100 mesh.
Equipment. The ANL bench-scale equipment, designed for operation
to 10 atm pressure, consists of a flui'dized-bed combustor, a com-
pressor for supplying fluidizing-combustion air, a preheater for the
fluidizing-combustion air, coal and additive feeders, and an off-gas
system (cyclones, filters, gas-sampling equipment, and pressure let-
down valve). The system is thoroughly instrumented and equipped with
an automatic data logging system.
The combustion unit consists of a 6-in.-dia pipe, 11 ft long.
The exterior of the pipe is wrapped with electrical heaters (to heat
the bed from room temperature to the coal-ignition temperature at the
start of an experiment) and cooling coils (to cool the bed during coal
11
-------
combustion). Additional cooling is provided by hairpin coils immersed
in the fluidized bed. Coal and additive are conveyed pneumatically
and continuously to the fluidized bed from weighed hoppers. Each
hopper has a rotary valve at the bottom for metering the coal or
additive to the air-conveying stream. A constant bed level is main-
tained in the combustor by use of an internal overflow pipe.
Combustion Experiments with Arkwright Coal. A series of statis-
tically designed experiments were made to measure the effects of the
independent operating variables, temperature, fluidizing-gas velocity,
and Ca/S mole ratio (ratio of calcium content of dolomite feed to sul-
fur content of coal feed), on response variables such as 862 and NO
suppression in the flue gas, additive utilization, additive eritrainment
and decrepitation, combustion efficiency, and heat-transfer coefficients.
The three levels of the variables tested were (1) temperature at 1450,
1550, and 1650°F; (2) Ca/S mole ratio at 1, 2, and 3; and (3) gas vel-
ocity at 2.0, 3.5 and 5.0 ft/sec. All of the experiments were made at
a pressure of 8 atm absolute, a 3-ft fluidized-bed height, and 3% oxygen
in the flue gas.
Sulfur Dioxide Retention. Sulfur dioxide levels in the flue
gas ranged from 850 ppm at a Ca/S mole ratio of 1.0, gas velocity
of 4.9 ft/sec, and bed temperature of 1665°F to a low of 120 ppm
at a Ca/S mole ratio of 2.9, gas velocity of 2.1 ft/sec, and bed
temperature of 1445°F (S02 level at zero removal is ^2300 ppm).
These S02 levels correspond, respectively, to emission rates of 1.57
and 0.23 Ib S02/106 BTU as compared with the EPA emission standard of
1.2 Ib S02/106 BTU.
For Ca/S ratios above 2.0, the S02 removal is generally greater
than 90%. The level of S02 in the flue gas increases rapidly, how-
ever, with decreasing Ca/S ratio and with increasing gas velocity
at low Ca/S ratios. The combustion or bed temperature appears to
have very little effect over the range of conditions tested. The
observed variable effects on the S02 retention strongly support the
theory that the rate of S02 uptake by the sorbent is diffusion
limited.
The results indicate that for this coal and additive, and at a
gas velocity of less than 5 ft/sec and a bed temperature of less than
1650°F, it should be possible to operate close to a Ca/S mole ratio
of 1.0, and still meet the EPA emission standard of 1.2 Ib of S02/106
BTU.
Nitrogen Oxide Emissions. Nitrogen oxide levels in the flue gas
were extremely low over the range of operating conditions tested.
Values varied from 270 ppm at a Ca/S mole ratio of 3.2, temperature
of 1630°F, and gas velocity of 3.6 ft/sec to 120 ppm at a Ca/S mole
ratio of 1.0, temperature of 1665°F, and velocity of 4.9 ft/sec.
These NO concentrations correspond, respectively, to emissions of 0.40
and 0.15 Ib N02/106 BTU, as compared with the EPA emission standard of
0.70 Ib N02/106 BTU.
12
-------
The only variable that appeared to have a significant effect on
the NO levels was the Ca/S mole ratio, with the NO level increasing
with increasing Ca/S mole ratio.
Combustion Efficiency. Combustion efficiency varied directly
with the combustion temperature, ranging from -x-89% at 1450°F to ^97%
at 1650°F. These efficiencies are comparable to previously reported
efficiencies obtained in atmospheric combustion experiments.
Combustible Carbon Hold-Up in the Fluidized Bed. To provide a
measure of the combustible carbon hold-up in the fluidized bed during
combustion, samples of the final bed material from each experiment
were analyzed for total carbon and carbon present as carbonate ion.
The combustible carbon content was taken as the difference between
these two values. For the eleven experiments, the combustible carbon
in the samples ranged from <0.01 to 1.60% and was £1.0% in seven of
the experiments. These levels of carbon are sufficiently low to ensure
safe, stable operating performance,
Additive Utilisation. Utilization of additive in the final bed
samples was found to vary inversely with Ca/S ratio and was relatively
unaffected by fluidizing-gas velocity or bed temperature. Utilization
ranged from ^80% at a Ca/S ratio of 1 to ^35% at a Ca/S ratio of 3.
Additive utilization was consistently higher for the final bed
material than for the elutriated solids collected in the primary
cyclone. In addition, calcium utilization for elutriated solids
generally increased the farther downstream the particulate matter in
the flue gas was collected. The latter result is attributed to
increasing reactivity with decreasing particle size.
Additive Entvairment. Entrainment of additive from the combustor
was found to vary directly with the superficial gas velocity. Entrain-
ment increased from a low of 5% (of the additive being fed to the com-
bustor) at 2 ft/sec to as high as 80% at 5 ft/sec.
Theoretical consideration was given to the terminal velocities
associated with different sizes of dolomite particles. For a sphericity
factor of 0.6, 5 ft/sec corresponds to the terminal velocity of ^16 mesh
additive particles. Since 80-90% of the dolomite feed had a terminal
velocity below 5 ft/sec, it is understandable that 80% of the additive
fed was elutriated from the combustion at a gas velocity of 5 ft/sec.
It should be emphasized, however, that in a commercial unit, the problem
of additive elutriation would be reduced considerably by selection of a
suitable sorbent particle size.
A semi-quantitative attempt was also made to determine the contri-
bution of additive decrepitation followed by elutriation of the result-
ing fines to the high entrainment rates. The method chosen to provide
a measure of additive decrepitation was to inventory the +45 mesh addi-
tive entering and leaving the combustor.
13
-------
Average recoveries determined by the above procedure for each level
of the three independent operating variables ranged from 92 to 103%. The
average recovery for all the experiments was 96%. The results suggest
that very little decrepitation of the dolomite occurred over the range of
conditions tested.
Heat-Transfer Coefficients at the Combustor Wall. Values of heat-
transfer coefficients at the combustor wall were calculated for the elev-
en experiments in the study and were found to vary directly with the gas
velocity. The values ranged from ^40 BTU/(hr)(ft^)(°F) at a gas velocity
of 2 ft/sec to VL15 BTU/(hr)(ft2)(°F) at 5 ft/sec.
Solids Loading in the Flue Gas. Solids loading in the flue gas leav-
ing the combustor (before any particulate removal) varied directly with
both the fluidizing-gas velocity and the Ca/S mole ratio. Loadings ranged
from 4.1 grains/ftJ (7.3 lb/106 BTU) at a Ca/S mole ratio of 1 and veloc-
ity of 2 ft/sec to 23 grains/ft3 (39,1 lb/106 BTU) at a Ca/S mole ratio
of 3 and velocity of 5 ft/sec. At gas velocities of 2, 3.5, and 5 ft/sec,
increasing the Ca/S mole ratio from 1 to 3 increased the solids loading
by 60, 80, and 125%, respectively. While these results suggest the desir-
ability of maintaining low Ca/S ratios to minimize solids loading from
the additive, high Ca/S ratios could still be used at suitably selected
gas velocities and sorbent particle sizes. Thus, the loadings quoted here
are not representative of what could be achieved.
Combustion-Side Corrosion of Internal Cooling Coils. Corrosion on
the external (bed) side of an internal cooling coil was studied after
it had seen an estimated 500 hr of operating time. Samples taken from
the cooling coil where it extended below the surface of the bed exhibited
intergranular corrosion to a depth of ^30ym, sensitization, and sigma-
phase formation. A sample taken in the freeboard area of the combustor,
M I/2ft above the bed, from a cooling coil which had seen VLOOO hr of
operating time exhibited corrosion to a depth of 600ym. An electron-
microprobe examination of the sample detected Fe and Ni depletion and
the presence of sulfur along the grain boundaries to the depth of cor-
rosion.
Combustion of Low-Sulfur Subbituminous and Lignite Coals. Ex-
periments were made in the 6-in.-dia pressurized combustor to determine
whether any difficulties would be encountered in processing a San Juan
mine subbituminous coal with a high ash content of 17% and a Glenharold
mine lignite with a low heating value of 7,625 BTU/lb. The nominal op-
erating conditions for the two experiments were a bed temperature of
1550°F, gas velocity of 3.5 ft/sec, and ©2 concentration of 3% in the
dry flue gas (VL5% excess air). Due to the low sulfur content (<1 wt %)
of the western coals, a Ca/S mole ratio of 1 was used for the experiments.
Operating Performance. As with the high-volatile bituminous
Arkwright coal, the operating performance of the fluidized-bed combustor
in processing the western coals was excellent, thus demonstrating the
versatility of the fluidized-bed concept for processing coals of widely
varying rank and quality.
14
-------
Sulfur Dioxide Retention. The S02 levels of 250 and 120 ppm
observed for the combustion of the subbituminous and lignite coals,
respectively, correspond to emissions of 0.45 and 0.21 Ib S02/106
BTU. The combustion of Arkwright coal under similar operating con-
ditions would have a projected S02 emission of 610 ppm or 1.2 Ib
S02/106 BTU. The above emissions represent S02 retentions of
approximately 72, 72 and 85% for the bituminous, subbituminous, and
lignite coals, respectively.
The somewhat higher retention reported for the combustion experi-
ment with lignite suggests that the calcium in the coal may be an
active agent in helping to retain S02 during combustion.
Nitrogen Oxide Emissions. The NO levels of 150 ppm and 130
ppm, respectively, for the combustion experiments with the subbitu-
minous and lignite coals correspond to emissions of 0.19 and 0.18
Ib N02/106 BTU. The projected emission for the Arkwright coal
(140 ppm) under similar conditions is also 0.19 Ib N02/105 BTU.
Corribustion Efficiencies. Combustion efficiencies for the two
experiments were 94 and 97% for the San Juan and Glenharold coals,
respectively. These values agree quite well with the combustion
efficiency of 94% reported for the Arkwright coal at a combustion
temperature of 1550°F.
Dolomite Utilization. Calcium utilizations for the final-bed
material of 74 and 71% for the combustion experiments with the San
Juan and Glenharold coals, respectively, in Tymochtee dolomite at a
Ca/S mole ratio of 1.0, are in relatively good agreement with the
76 to 83% utilizations reported above for combustion experiments
with Arkwright coal using the same dolomite and the same mole ratio.
Effect of Combustor Pressure and Temperature on the Concen-
tration of NO and Other Gases in the Flue Gas. Two series of ex-
periments were conducted to specifically study the effect of com-
bustor operating pressure on concentrations of NO and other gases
in the flue gas. The NOX-I series involved the combustion of
Arkwright coal in a fluidized bed of Alundum. In the NOX-II series,
Arkwright coal was combusted in a fluidized bed of sulfated dolo-
mite while feeding dolomite at a Ca/S feed mole ratio of 1.5.
Nominal operating conditions in both series were a fluidized-bed
temperature of 1550°F, 3% 02 in the flue gas, and a fluidizing-gas
velocity of 3.5 ft/sec. Each of the two series of experiments
consisted of six parts, with the combustor pressure ranging from
1 to 8 atm. In a third series of experiments, the system pressure
was maintained at 4 atm, and the bed temperature was maintained at
1450, 1550, and 1650°F.
In addition to the concentrations of gases ordinarily monitored
in the flue gas (02, C02, S02, CO, NO, and CHt,), instrumentation
was installed for the continuous or intermittent determination of
15
-------
ammonia, hydrogen, and C2~C^ hydrocarbons.
Pressure Effect on NO Levels in the Flue Gas. In the absence of
dolomite, the NO level in the flue gas increased from less than 200 ppm
at 8 atm to "V1600 ppm at 1 atm. An NO level of 1600 ppm is unusually
high and may not be representative.
In the presence of dolomite (NOX-II series), an increase in NO
is also observed at reduced pressures, although it is considerably
less pronounced than in the first series of experiments. The level
of NO increased from ^200 ppm at 8 atm to around 400 ppm at 1 atm.
Pressure Effect on Concentration of Flue-Gas Components Other
Than NO. Analysis of the data from the NOX experiments can be
summarized as follows:
1. The S02 concentration is relatively unaffected by changes
in system pressure.
2. Methane levels in the flue gas tend to increase moderately
as the pressure is reduced from 8 to 6 atm and then increase rapidly
(up to 400 ppm) as the pressure is reduced to 1 atm.
3. In the presence of dolomite, CO levels increased from 170 ppm
at 8 atm to over 2000 ppm at 1 atm. In the absence of dolomite, the CO
level passed through a minimum (vLOO ppm) at about 4 atm pressure
and increased rapidly with either increasing or decreasing pres-
sure. This phenomenon remains unexplained.
4. The level of H2 (35 ppm) is relatively unaffected by changes
in pressure.
5. The levels of the other gases (NR^, C2Hlt, C2H6, 02^, C3Hs)
were below the detection limits of the analytical instrumentation.
Effect of Temperature on Flue-Gas Constituents. Under the con-
ditions at which the experiments were performed, temperature varia-
tions from 1450°F to 1650°F have little or no effect on the compo-
sition of the various flue-gas components tested.
Trace-Element Distribution Studies
The emissions of biologically toxic trace elements such as Hg and
Be from coal-burning power plants are receiving increasing attention as
potentially dangerous air pollutants. Relative to conventional boilers,
the lower combustion temperature of the fluidized-bed boiler and the
presence of additive for S02 retention may serve to reduce the emissions
of trace pollutants.
The elements of primary concern in the investigation are Hg,
Pb, Be, and F. The approach being taken is to make mass balances
16
-------
around the combustor for as many of the trace elements as possible with-
in the economic limitations of the study. This involves sampling and
analyzing all the solid materials charged to or recovered from the com-
bustor, including particulate matter entrained in the flue gas from the
unit. Analysis of the flue gas is also made for the more volatile trace-
element species, Hg and F~.
Four combustion experiments were made for the specific purpose of
investigating the behavior of trace elements during combustion. Experi-
ments TRACE-3 and TRACE-4B were made at a bed temperature of 1550°F, a
pressure of 10 atm, and 4% 02 in the off-gas. Experiments TRACE-5A and
TRACE-6 were made at 1650°F, 8 atm, and 3% 02 in the flue gas. In each
set of experiments, the first experiment involved the combustion of
Arkwright coal in a fluidized bed of alumina, whereas the second experi-
ment involved the combustion of Arkwright coal in a fluidized bed of
Tymochtee dolomite.
Mercury Balances. Mercury balances for experiments TRACE-3 and
TRACE-4B indicated Hg recoveries, expressed as the percentage of Hg
entering the combustor which can be accounted for in the combustion
products, of 56 and 43%, respectively. For experiments TRACE-5A and
TRACE-6, the respective recoveries were only 29 and 25%. The lower
recoveries for the latter two experiments resulted from the decreased
recovery of volatilized Hg in the flue gas which was down by a factor
of 6 (3 and 5% vs. 19 and 34%) from the prior experiments.
Analysis of the solid materials from the combustion indicate that
essentially no Hg was retained by the fluidized bed and that from 10
to 35% of the mercury which entered the combustion was retained by the
fly ash and elutriated additive removed from the flue gas. Data for
conventional coal-fired boilers indicate a Hg retention of VlO% by the
recovered ash products.
Attempts are being made to indentify the cause for the poor recov-
ery of Hg in the combustion products. Particular question areas that
are being examined include the following:
(1) The inability to detect mercury in the ICI scrubbing
solutions used in sampling the flue gas.
(2) The recovery efficiency of the mercury sampling equip-
ment.
(3) The possibility that mercury is plating out or condens-
ing on the walls of the flue-gas ductwork before reach-
ing the sampling zone.
Lead and Beryllium Mass Balances. Lead balances for the four
trace-element experiments range from 78 to 125% recovery of the lead
entering the combustor. Beryllium balances were somewhat lower, rang-
ing from 56 to 87%.
17
-------
The results indicate that lead is essentially retained by the
particulate matter in the combustion products. Emissions of Pb,
therefore, will be controlled by the efficiency of particulate re-
moval from the flue gas. However, because the concentration of Pb
in particulate matter increases with decreasing particle size, it
appears obvious that the efficiency of lead removal will be some-
what lower than the efficiency of particulate removal.
The beryllium mass balances, with the lower recoveries, suggest
the possible volatilization of beryllium (or compounds of beryllium)
during combustion.
Aside from material balance considerations, the results seem to
indicate that both Pb and Be concentrate preferentially in the finer
particulate matter leaving the combustor. The preferential concen-
tration of relatively nonvolatile trace elements in particles of
smaller diameter has been reported by others for samples taken from
ambient air and from the flue gas of coal-fired power plants. It
is suggested that the lower trace-element concentrations in the
larger particles are simply the result of dilution by unburned
carbonaceous materials.
Fluoride Mass Balances. The indicated recovery of fluoride for
the experiments at 1550°F are 123 and 110%, which are reasonably
acceptable values. The recoveries of 208 and 276% reported for the
experiments at 1650°F, however, are unaccountably high. The only
differences in sampling between the two sets of experiments were
the use of considerably larger flue gas samples and Na2C03 scrubbers
instead of NaHC03 scrubbers for the experiments at 1650°F.
Perhaps the most significant observation that can be made from
the F~ balances is that a significantly higher retention of the F~
appears to occur in the solid products of combustion when additive
is present. The recoveries of F~ in the solid samples were relatively
consistent, being 56 and 62% for combustion with additive and only
23 and 5% for combustion in an alumina bed. A possible explanation
for such a phenomenon could be the formation of CaF, which is a
relatively stable compound.
Neutron Activation Analysis, Developmental Work. The analytical
method of neutron activation analysis is being considered as a possible
instrumental method for expanding the trace-element study to include el-
ements of second- and third- priority interest (Cd, As, Ni, Zn, Cu, Ba,
Sn, P, Li, Mn, Cr, Se, and V). A preliminary testing of the method was
made by irradiating samples from the TRACE-3 experiment in the ANL CP-5
test reactor. After the irradiation, periodic y-ray counts were taken
using a Ge(Li) detector for the purpose of identifying activation pro-
ducts and their relative activity levels. Concentration data for seven
additional minor and trace elements are reported on the basis of the
preliminary results.
18
-------
Encouraged by the preliminary results, the remaining samples
from the trace-element combustion experiments are being analyzed
(including a reanalysis of the TRACE-3 samples) by this technique.
Additional developmental work (see Appendix D) is also being done
to increase both the number of elements detected and the accuracy.
Sodium Concentrations in Flue-Gas Particulates. Samples from
three of the TRACE experiments have been analyzed for Na because of
the considerable interest that has been expressed in Na as a source
of corrosion in a combined-cycle power system, particularly to the
blades of a high-temperature gas turbine. The results indicate
that Na is retained by the particulate matter during combustion,
and that the concentration of Na in the particulate matter in-
creases with decreasing particle size. Concentrations generally
varied from 0.5 wt % for material removed in the primary cyclone
to 1.5-2.0 wt % for material removed in the primary and secondary
filters.
Photomicrographs of Flue-Gas Particulates. Solid samples of
Arkwright coal and fly ash recovered from the cyclones and Brink
impactor at the conclusion of experiment TRACE-3, were examined
under a scanning electron microscope. Photomicrographs obtained
for several of the samples are presented. Unlike the generally
spherically shaped fly ash emitted from conventional coal-fired
combustors, the photomicrographs depict the fly ash from the fluid-
ized-bed combustor as a rather fragile and flake-like material.
Kinetics of the Reaction of Half-Calcined Dolomite with Sulfur
Dioxide
The work reported here represents the initial results obtained
by a small basic-chemistry support program associated with the
development of fluidized-bed combustion.
The active material used in this program is half-calcined
dolomite which reacts with S02 as indicated in equation (S-l)
[CaC03 + MgO] + S02 + 0.5 02 -»• [CaSO^ + MgO] + C02 (S-l)
For economic and environmental reasons, it is desirable to regener-
ate the reactive material from the product by some scheme such as
that indicated by equations (S^-2) and (S-3) .
[CaSOt, + MgO] + 4 H2 + [CaS + MgO] +4 H20 CS-2a)
[CaSO^ + MgO] + 4 CO -> [CaS + MgO] + 4 C02 (S-2b)
[CaS + MgO] + H20 + C02 -* [CaC03 + MgO] + H2S (S_3)
where the concentration of H2S resulting from equation (S-3) is
sufficiently high to allow sulfur recovery in a Claus plant.
19
-------
The goals of this program are i) to determine the detailed kinet-
ics of equations (S-l), (S-2), and (S-3); ii) to elucidate plausible
mechanisms for these reactions; and iii) to determine the conditions
that optimize each of these reactions. Kinetic results obtained at
atmospheric pressure for equation (S-l) are presented here.
Experimental Apparatus and Procedure. The experimental apparatus
consists of an Ainsworth thermal gravimetric analyzer (TGA) unit. An
automatic recording balance provides for continuous monitoring of the
sample weight with an accuracy of 0.1 mg. The sample is located in a
platinum basket that is suspended from one arm of the balance and
hangs in the heated region of a quartz reactor tube. The temperature
of the reaction zone is controlled to about + 2°C over the range from
ambient to 1000'C. Reaction gases are premixed and introduced into
the reactor from the bottom.
Commercial, research-grade cylinder gases were used to make up
the reactant-gas stream. The stone used in these experiments is
designated dolomite 1337 and is supplied by Charles Pfizer and Co.,
Gibsonburg, Ohio; the dolomite has an empirical formula Ca^ iifMgQ.86
(C03)2.
In a typical experiment, 200 mg of the dolomite-1337 particles
having diameters in a narrow range around 1.1 mm (-16, +18 U. S. Stan-
dard Screen) are placed in the apparatus under a flow of N2 and C02
and heated to 800°C at a rate of about 25°/min to half-calcine the
stone. Calcination is followed by observing the weight change of the
sample. When half-calcination of the stone is complete (usually after
about 45 min), sample temperature is adjusted to that selected for the
experiment, the sample is isolated under the N2 and C02 atmosphere,
and the reactant gas mixture adjusted to the appropriate composition
and flowrate for the experiment while bypassing the reaction tube.
At time zero, the reactant gas is diverted through the reaction tube
and the weight change of the sample observed as a function of time.
The reaction is followed until the rate of weight change is negligible.
Results. The reaction rate, and hence the extent of the reac-
tion at a given time, is a function of S02 concentration. A plot
was made of the log of the initial reaction rate (initial rates were
actually evaluated at t = 1 min because of scatter in the data near
t = 0) versus the log of S02 concentration. The straight line thus
obtained had a slope of 1.08 and indicates that the reaction is first
order with respect to S02 concentration in the reactant gas under the
reaction conditions tested.
It was noticed in several experiments that the concentration of
water vapor in the reactant gas appeared to influence the reaction
rate. A result from a series of experiments in which the concentra-
tion of water in the reactant gas was varied while the concentration
of other components was constant indicated, however, that the reaction
20
-------
was essentially zero order with respect to water concentration. An-
other series of experiments was made in which the reactant gas was dry
and the concentration of SC>2 was varied. Under these conditions, it
was observed that the rate varied with the 0.77 power of S02 concen-
tration. Thus it would appear that the rate determining step is dif-
ferent depending on whether or not water is present in the reactant
gas. With water present, the reaction is 0.22 order with respect to
oxygen concentration in the reactant-gas stream.
The temperature dependence of the reaction rate in experiments
with water present was examined over the range from 500 to 850°C (at
higher temperatures, the concentration of C02 necessary to prevent
decomposition of CaC03 could not be maintained with the present
apparatus). The initial reaction rate increased significantly with
temperature over this range. An Arrhenius plot of the data shows a
linear dependence of rate on 1/T and yields an apparent activation
energy of 7.3 kcal/mol. Such a value does not point conclusively
to a mechanism in which some chemical reaction is rate controlling
but is somewhat greater than one might expect if the reaction is
diffusion controlled.
Experiments were also made to assess the extent to which MgO
undergoes sulfation. It was found that the extent of sulfation of
MgO is not great and the reaction rate is much less than that for
the half-calcined stone.
21
-------
INTRODUCTION
The combustion of fossil fuels in a fluidized bed of calcined lime-
stone particles is a potentially efficient and economically attractive
process for the generation of steam for electric power production and
other uses. The process results in greatly reduced emissions of both
sulfur oxide and nitrogen oxide pollutants. Development work, evalua-
tions, and design studies are being carried out by various organizations
under contract with the U. S. Environmental Protection Agency (EPA) and
the Department of Interior's Office of Coal Research (OCR). The AEC's
Argonne National Laboratory (ANL) is one of the principal participants
in the development program under interagency agreements with EPA -and
OCR. Other participants in the program include the British National
Coal Board, Combustion Power Utility Company, Pope, Evans, and Robbins,
Inc., Exxon Research and Engineering Company, and Westinghouse Electric.
The fluidized-bed principle has been used extensively in the
petroleum and chemical-process industries. As applied to fossil-fuel
combustion, the fluidized bed consists of particles of partially
sulfated lime or dolomite that are held in suspension by the combustion
air which enters through a grid at the bottom of the combustor. Coal,
or any fossil fuel, is injected and burns in the bed, which is at a
temperature of 1600-1700°F. The heat of combustion is partially
removed by generation of steam in boiler tubes immersed in the bed.
The steam can be used as process steam or to generate electrical
power. The fluidized-bed heat-transfer rate from the hot bed particles
to the outside metal surfaces of boiler tubes in the bed is high.- The
sulfur dioxide released during the burning of the coal is reacted with
lime and, in the presence of excess oxygen in the bed, forms calcium
sulfate
CaC03 + S02 + 1/2 02 •> CaSO^ + C02 (1)
Fresh, crushed limestone (or dolomite) is injected continuously into the
bed and the sulfated lime is removed continuously to keep the level of
the bed constant. The sulfated lime that is removed can either be
regenerated to recover sulfur which is stored, and lime, which is reused
in the combustor, or the sulfated lime can be discarded with the ash
from the combustor.
While various regeneration schemes have been proposed, two processes
appear particularly promising. The first process is a one-step reductive
decomposition of CaSO^ at ^2000°F
CaSO^ + CO + CaO + S02 + C02 (2)
The second process is a two-step method involving the reduction of
to CaS at 1600-1700°F followed by reaction of the CaS with C02/H20 at
1000-1300°F to release H2S
+ 4 CO + CaS + 4 C02 (3a)
CaS + C02 + H20 + CaC03 + H2S (3b)
22
-------
Several variations of the fluidized-bed combustion concept are
under investigation for power-plant applications. First is a unit in
which coal is burned at atmospheric pressure. The combustion heat is
transferred to steam tubes immersed in the bed and to heat-recovery
units outside the bed. A second variation is a pressurized unit in
which the coal is burned at elevated pressure (e.g., 135 psig) in a
fluidized bed. Steam for operating a steam turbine is raised by
steam tubes immersed in the fluidized bed. In addition, the flue
gas is expanded through a gas turbine. Exit gas from the turbine is
passed through heat-recovery units. This variation is intended to
take advantage of the higher thermal efficiencies offered by a gas
turbine-steam turbine combined cycle. A third version is a fluidized-
bed combustor that is operated adiabatically at elevated pressure.
The heat of combustion is removed from the bed by providing a large
excess of air, ^150%. The flue gas is first expanded through a gas
turbine and is then exhausted to an unfired heat-recovery boiler.
No steam-raising tubes are present in-the bed.
Argonne National Laboratory has been conducting a basic research
and development program since 1968 to demonstrate the feasibility and
potential of fluidized-bed combustion for reducing sulfur and nitric
oxide emissions in accordance with environmental standards established
by state and federal regulatory agencies. Previous annual reports in
the series have presented comprehensive data on fluidized-bed combustion
and regeneration at atmospheric pressure. More recently, an experi-
mental program was initiated at ANL to develop advanced technology in
pressurized fluidized-bed combustion. Objectives of the program are
to optimize the combustion process with respect to sulfur dioxide
retention in the fluidized bed of additive, nitrogen oxide suppression
in the flue gas, and combustion efficiency; determine the behavior
and operability of the system over a wide range of operating conditions
and with a variety of coals including subbituminous and lignite;
characterize the distribution and emission of polluting trace elements
from fluidized-bed combustion systems and compare these emissions with
commercial, pulverized- fuel power plants; and elucidate basic
mechanisms of sulfation and regeneration reactions. The work reported
here has consisted of the following experimental investigations:
1. A statistically designed series of combustion experiments was
made to investigate the effects of the three independent operating
variables of Ca/S mole ratio (mole ratio of calcium in the additive
to sulfur in the coal), combustion temperature, and superficial
fluidizing-gas velocity on SC>2 and NO levels in the flue gas,
combustion efficiency, additive utilization, additive decrepitation
and entrainment, and heat-transfer coefficients. Three levels of
the independent variables xcere investigated: (1) Ca/S mole ratios
of 1, 2, and 3; (2) combustion temperatures of 1450, 1550, and 1650°F;
and (3) fluidizing-gas velocities of 2.0, 3.5, and 5.0 ft/sec. All
experiments were made in the ANL 6-in.-dia pressurized, fluidized-bed
combustor using high-volatile, bituminous Arkwright coal and Tymochtee
dolomite at a pressure of 8 atm, a 3-ft fluidized-bed height, and
excess combustion air (3% 02 in the dry flue gas).
23
-------
2. Two experiments were conducted in the ANL 6-in.-dia combustor
to evaluate the adaptability of pressurized, fluidized-bed combustion
technology for the combustion of low-rank coals. A San Juan mine sub-
bituminous coal containing 18% ash and a Glenharold mine lignite coal
with a heating value of only 7,400 BTU/lb were combusted at a Ca/S mole
ratio of 1.0, bed temperature of 1550°F, fluidizing-gas velocity of
3.5 ft/sec, combustor pressure of 8 atm abs, fluidized-bed height of
3 ft, and ^15% excess combustion air (3% 02 in the dry flue gas). The
experiments were evaluated for S02 and NO levels in the flue gas,
combustion efficiencies, and additive utilizations.
3. The effects of combustor pressure and temperature on the
concentration of S02, NO, CO, CH4, N02, NH3, H2 and C2-C^ hydro-
carbons were determined on the basis of three series of experiments
in the 6-in.-dia combustor. The first series of experiments involved
the combustion of Arkwright coal in a fluidized bed of inert solids
(Alundum). In the second series, Arkwright coal was combusted in a
fluidized bed of Tymochtee dolomite while feeding the dolomite at a
Ca/S mole ratio of 1.5. Operating conditions for both series were a
combustion temperature of 1550°F, a fluidizing-gas velocity of 3.5 ft/
sec and 3% 02 in the flue gas. Each of the two series of experiments
consisted of six parts with the combustor pressure ranging from 1 to 8
atm abs. In the third series of experiments, the system pressure was
maintained at ^4 atm abs and the bed temperature set at 1450, 1550, and
1650°F.
4. Two sets of experiments were completed in the 6-in.-dia
combustor to characterize the distribution and emission of trace
elements from fluidized-bed combustion systems. Data were obtained for
the biologically toxic trace elements, Hg, Pb, F, and Be, and for Na.
In each set of combustion experiments, Arkwright coal was first combusted
in a fluidized bed of inert solids (Alundum) and then in a fluidized
bed of Tymochtee dolomite under nominally similar operating conditions.
The first set of experiments were made at a bed temperature of 1550°F,
combustor pressure of 10 atm abs, and ^4% 02 in the flue gas. The
second set of experiments were made at 1650°F, 8 atm abs, and ^3% 02
in the flue gas.
5. Reaction kinetics were also investigated using thermogravimetric
analysis techniques. Results are presented on the effect of S02 concen-
tration on the rate of reaction between SQ2 (in the absence and presence
of H20) and half-calcined dolomite; on the effect of water vapor and
oxygen on the sulfation rate of half-calcined dolomite; on the temperature
dependence of the sulfation reaction with water vapor present; and on
the reaction rate between MgO and reaction gases containing S02, 02,
and H20.
24
-------
BENCH-SCALE, PRESSURIZED, FLUIDIZED-BED
COMBUSTION EXPERIMENTS
Experiments were performed in the ANL, 6-in.-dia, pressurized,
fluidized-bed combustor to demonstrate the feasibility and potential
of fluidized-bed combustion at elevated pressures (up to 10 atm).
Specific objectives of the experiments were to evaluate the effects
of combustor operating variables on S02 removal, NO suppression in
the flue gas, combustion efficiency, additive utilization, solids
entrainment, and heat-transfer coefficients. The results of these
experiments are reported below. Also reported are the results of
experiments made with low-rank coals and experiments made to measure
the distribution and concentration levels of trace elements entering
and leaving the combustion system.
Raw Materials
Coals. Combustion experiments were made using a variety of coals
of differing rank. The principal coal tested was a highly caking,
high-volatile bituminous, Pittsburgh seam coal obtained from the
Consolidation Coal Company's Arkwright mine. As received, the coal
contained 2.82 wt % S and 7.68 wt % ash, had a heating value of
13,700 BTU/lb, and had an average particle size of 323 urn. Also
tested, although less extensively, were a subbituminous coal obtained
from the Western Coal Company's San Juan mine in New Mexico (provided
by the Public Service Company of New Mexico) and a lignite obtained
from the Consolidation Coal Company's Glenharold mine in North Dakota.
Respectively, the coals, as received, contained 0.78 and 0.53 wt % S
and 17.0 and 6.11 wt % ash, had heating values of 9,620 and 7,620
BTU/lb, and had average particle sizes of 340 and 353 um. Complete
data on the chemical and physical characteristics of the three coals
are presented in Tables A-l to A-3 in Appendix A. The coals were fed
to the combustor as received.
Additive. A Tymochtee dolomite (^50 wt % CaC03 and ^30 wt %
MgC03> obtained from C. E. Duff and Sons, Huntsville, Ohio, was used
in all of the experiments requiring a sulfur-accepting additive. Prior
to its use in the combustor, the dolomite was air-dried and double
screened using 14- and 100-mesh screens. As fed to the combustor,
the dolomite had an average particle size of 750 ym. Table A-4 in
Appendix A summarizes the chemical characteristics and particle-size
distribution of the Tymochtee dolomite.
Inert Bed Material. A high-purity Type 38 Alundum grain
(electrically fused A1203> was obtained from the Norton Company for
use in combustion experiments that required a fluidized-bed of
chemically inert solids. The Alundum, which had a nominal grit size
of 30 mesh, is typically analyzed at 99.5 wt % A1203. A typical
chemical analysis and particle-size distribution for the Alundum
are given in Table A-5 of Appendix A.
25
-------
Bench-Scale Equipment
The experimental equipment and instrumentation consist of a
6-in.-dia, fluidized-bed combustor which can be operated at pressures
up to 10 atm, a compressor to provide fluidizing-combustion air, a
preheater for the fluidizing-combustion air, peripheral-sealed rotary
feeders for metering solids into an air stream entering the combustor,
two cyclone separators and two filters in series for solids removal from
the flue gas, associated heating and cooling arrangements and controls, and
temperature and pressure sensing and display devices. A simplified
schematic flowsheet of the combustion equipment is presented in
Fig. 1.
Details of the fluidized-bed combustor are illustrated in Fig. 2.
The reaction vessel consists of a 6-in.-dia, schedule 40 pipe (Type
316 SS), approximately 11 ft long. The reactor is centrally contained
inside a 9-ft section of 12-in.-dia schedule 10 pipe (Type 304 SS) . A
bellows expansion joint is incorporated into the outer shell to
accommodate differential thermal expansion between the pipes. A
bubble-type gas distributor is flanged to the bottom of the inner
vessel. Fluidizing-air inlets, thermocouples for monitoring bed
temperatures, solids feed lines, and solids removal lines are
accommodated by the distributor. The coal and additive feed lines extend
2 in. above the top surface of the distributor plate and are angled
20° from the vertical toward the longitudinal axis of the combustor.
A constant bed height is maintained in the combustor by use of either
a 36-in.- or 48-in.~high standpipe. The 6-in.-dia pipe is alternately
wrapped with resistance-type heating elements and cooling coils onto
which has been applied a layer of heat-conducting copper and then an
overlay of oxidation-resistant stainless steel. Additional cooling
capacity is provided by five internal, hairpin-shaped cooling coils
which extend down from the flanged top of the combustor to within
12 in. of the top surface of the distributor. The coolant is water
entrained in air.
Fluidizing-combustion air is supplied by a 75-hp, screw-type
compressor capable of delivering 100 cfm at 150 psig. The air can be
heated to approximately 1000°F in a 6-in.-dia, 10-ft-tall preheater
that has eight, 2700-watt, clam-she11-type heaters inside it to provide
heat.
Coal and dolomite additive are pneumatically fed from hoppers to
the combustor using two, lO-in.-dia rotary valve feeders. The feeders
and hoppers are mounted on platform-type scales.
The flue gas is sampled continuously and is analyzed routinely for
the components of primary importance. Continuous determinations of
nitrogen oxide, sulfur dioxide, methane, and carbon monoxide are made
using infrared analyzers. Oxygen is also monitored continuously using
a paramagnetic analyzer. Intermittent carbon dioxide analyses of the
flue gas are made by gas chromatography. Prior to and during each
26
-------
TO GAS ANALYSIS
SYSTEM
SCREW
COMPRESSOR
AIR
PREHEATER
9 VENTILATION
_<*, —
^g-~~*-L r^""? -'•'CrTj " "~ ---------
\J(J PRESSURE EXHAUST
FIBER SINTERED CONTROL
FILTER STAINLESS VALVE
STEEL
FILTER
SECONDARY
CYCLONE
PRIMARY
CYCLONE
Figure 1. Simplified Equipment Flowsheet of Bench-Scale
Fluidized-Bed Combustor and Associated Equipment
-------
60 AND 85 in. FREEBOARD
THERMOCOUPLES
INTERNAL COOLING
COIL LEADS
PURGE GAS OUTLET
HEATER CONTROL
THERMOCOUPLES
RUPTURE DISK
FLUE GAS TO
CYCLONE AND FILTERS
EXPANSION BELLOWS
RUPTURE DISK
12-in.JACKET
— r=- SOLIDS FEED LINES
SHELL PURGE
GAS INLET
EXTERNAL COOLING
COIL LEADS -
36 OR 48 in.
SOLIDS OVERFLOW
6, I2.AND 44 in.
BED THERMOCOUPLE WELL
EXTERNAL COOLING COIL
(WRAPPED ON 6-in.DIA WALL)
CALROD TYPE HEATER
INTERNAL COOLING COILS
BUBBLE CAP DISTRIBUTOR
ELECTRICAL HEATER
LEADS
PLIBRICO FILLED VOLUME
FLUIDIZING AIR
Figure 2. Detail Drawing of 6-in.-Dia,
Pressurized Fluidized-Bed Combustor
28
-------
experiment, the response of each analytical instrument is checked using
standard gas mixtures of flue-gas components in nitrogen. Batch samples
of the flue gas can be taken and analyzed for components of secondary
importance.
The combustion system is equipped with a Hewlett-Packard 2010C
data acquisition system to monitor and record temperature, pressure,
gas flow, and flue-gas concentration data for subsequent data handling
and analysis.
Experimental Procedure
The experimental procedure, although subject to minor variations,
was basically as follows. A preweighed amount (VL5 kg) of either
partially sulfated additive from a previous experiment or tabular
alumina was charged to the reactor to provide an initial bed of
material. The starting bed was then raised to about 800°F by passing
fluidizing air, preheated to between 800 and 900°F, through the com-
bustor and simultaneously employing the resistance heaters on the
combustor wall. Once the bed temperature reached 800°F, the system
was brought to the desired operating pressure and coal entrained in
a transport air stream was injected into the bed. To prevent the
possibility of carbon accumulation in the fluidized bed during start-
up, coal was initially injected in small, intermittent amounts until
ignition and sustained combustion was confirmed by a rapidly increasing
temperature and a changing flue-gas composition. Continuous injection
of coal was then initiated and the bed temperature was raised to a
selected combustion temperature. The bed temperature was maintained
by use of the external and internal cooling coils.
Injection of the sulfur-accepting additive was initiated when the
bed reached operating temperature. The air, coal, and dolomite feed
rates were adjusted to give a specified mole ratio of calcium in the
additive to sulfur in the coal, a specified superficial gas velocity,
and a specified level of oxygen in the flue gas leaving the combustor.
Sulfated additive was removed from the combustor by means of a stand-
pipe to maintain a constant fluidized-bed level.
Combustion Experiments with Arkwright Coal
A series of nine experiments in a 3 x 3 latin-square experimental
design (1/3 replicate of a 33 factorial design) plus 2 replicate
experiments were made to measure the effects of the independent operating
variables, temperature, fluidizing-gas velocity, and Ca/S mole ratio,
on response variables such as sulfur dioxide capture efficiency,
nitrogen oxide levels in the flue gas, additive utilization, additive
entrainment and decrepitation, combustion efficiency, and heat-transfer
coefficients. The three levels of the variables tested were (1) temper-
ature at 1450, 1550, and 1650°F; (2) Ca/S mole ratio at 1, 2, and 3;
and (3) gas velocity at 2.0, 3.5, and 5.0 ft/sec. All of the experiments
were made at a pressure of 8 atm absolute, a 3-ft fluidized-bed height,
29
-------
and 3% oxygen in the flue gas. Table 1 summarizes the designed operating
conditions and shows the matrix representation for the eleven experi-
ments in the experimental design. The actual operating conditions
and flue-gas compositions for the eleven experiments are given in
Table 2.
Table 1. Designed Operating Conditions and Matrix Representation
for the Eleven Experiments in the Latin-Square Experimental Design
Equipment:
Coal:
Additive:
Pressure:
Fluidized-Bed Height:
Excess Air:
6-in.-dia pressurized fluidized-bed combustor
Arkwright (Consolidation Coal Company)
Tymochtee dolomite (C. E. Duff and Sons)
8 atm abs
3 ft
M.5% (3% 02 in the dry flue gas)
Levels of Independent Variables
(T)
(R)
(V)
Variable
Temperature, °F
Ca/S Mole Ratio
Gas Velocity, ft/sec
Level
1450
1.
2.
1
0
0
Level
1550
2
3
2
.0
.5
Level
1650
3
5
3
.0
.0
Matrix Representation of Experimental Combinations
V2 a
(VAR-5)3
V3
(VAR-7)
Vi
(VAR-1)
Vi
(VAR-3)
V2
(VAR-6)
V3
(VAR-2)
V3
(VAR-9)
Vl
(VAR-4)
V2
(VAR-8)
Experiment number.
Two replicate experiments made of this experimental combination;
VAR-6-R, VAR-6-2R.
30
-------
Table 2. Operating Conditions and Flue-Gas Analyses for the Eleven Experiments
Investigating the Effects of Independent Operating Variables
Coal:
Additive:
Fluidized-Bed Height:
System Pressure:
Experiment
VAR-1
VAR-2
VAR-3
VAR-4
VAR-5
VAR-6
VAR-6-R
VAR-6-2R
VAR-7
VAR-8
VAR-9
Combustion
Temp, °F
1445
1565
1575
1650
1460
1565
1560
1550
1460
1630
1665
Ca/S Mole
Ratio
2.9
2.9
1.1
1.9
1.0
2.0
2.1
2.0
2.2
3.2
1.0
Gas
Velocity,
ft/sec
2.1
4.8
2.1
2.3
3.4
3.6
3.6
3.5
4.2
3.6
4.9
: Arkwright (2.82% sulfur)
: Tymochtee Dolomite
: 3 ft
: 8 atm abs
Flue-Gas Analysis
02,%
2.7
3.0
3.0
3.0
3.0
3.0
2.9
3.0
2.7
3.0
2.9
C02,%
17
17
15
16
15
16
15
16
16
17
16
S02,PPm
120
130
350
120
680
170
210
190
260
130
850
N0,ppm
190
210
140
180
150
190
180
160
150
270
120
C0,ppm
100
30
30
50
760
30
30
30
90
40
40
Dry basis,
-------
Sulfur Dioxide Retention. Sulfur dioxide levels in the flue gas
ranged from 850 ppm at a Ca/S mole of 1.0, gas velocity of 4.9 ft/sec,
and bed temperature of 1665°F to a low of 120 ppm at a Ca/S mole ratio
of 2.9, gas velocity of 2.1 ft/sec, and bed temperature of 1445°F (S02
level at zero removal is about 2300 ppm). These sulfur dioxide levels
correspond, respectively, to emission rates of 1.57 and 0.23 Ib S02/106
BTU as compared with the EPA emission standard of 1.2 Ib S02/106 BTU.
Analysis of Variance of SO? Data. An analysis of variance was
made on the S02 levels in the flue gas and is presented in Table 3.
According to the F-test for significance at the a = 0.1 level, the
Ca/S mole ratio is the only significant source of variation in the ob-
served experimental S02 levels. However, the validity of the analysis
of variance for the latin square experimental design (essentially a
one-third replicate of a 33 factorial design which does not confound
main effects with each other) is based on the assumption that the con-
trolled variables act independently of one another. This is expressed
mathematically by the model equation for the latin square experimental
design as follows:
= y + a.
.
where X^., denotes the observed response,
possible^responses, a-, 3.,, and
'k ijk
y equals the true mean of all
equal the true treatment effects,
and Gijk equals the residual or error between the observed response
and the expected response.
Table 3. Analysis of Variance of S02 Flue-Gas Levels
in the 3x3 Latin-Square Designed Series of "VAR" Experiments
Source of
Variation
Ca/S Ratio
Velocity
Temperature
Error
Sum
of
Squares
4
7
3
2
.434
.120
.780
.80
x
x
X
X
105
10"
10"
10"
Degrees of
Freedom
2
2
2
2
Mean
Square
2
3
1
1
.217
.56
.89
.40
x 105
x 10"
x 104
x 10
F
15
2
1
a
e
.8
.5
.4
Fo.i
9.00
Total
5.804 x 105
F = ratio of variable mean square to the error mean square.
32
-------
In practice, the assumption of independence is generally taken to
mean that the interactions are not large compared to the main effects.
When interactions are large, in addition to inflating the experimental
error mean square, they confuse the effects required to be estimated
and can give misleading F-tests for significance. Evidence that the
error mean square has been inflated in the analysis of variance for
the S02 data is apparent in comparing the error mean square in the
analysis of variance table (1.4 x 104) to the S02 variance determined
from the replicated experiments (4.0 x 102). Since both are estimates
for o , the variance of the response data, the difference between the
two values seems large. The large error mean square in the analysis
of variance table seems to indicate that interactions may be relatively
large and that the model equation for the latin square analysis of
variance does not hold for the SC>2 data.
Regression Analysis of SO? Data. Using nonlinear regression
methods, the 862 levels in the flue gas were correlated using the
following power-curve-type equation:
vl/2
In S02 = A + B (|^ )
Where SC>2 = level of S02 observed in the flue gas, ppm
A,B = constants
V = superficial fluidizing-gas velocity, ft/sec
R = Ca/S mole ratio
T = Bed temperature, °R x 10 3
The values of the coefficients A and B were determined by the
method of least squares to be 4.06 + 0.06 and 2.53 + 0.10, respectively.
The nominal or designed levels of the independent operating variables
were used for the correlation.
In Table 4, the calculated and experimental values for the eleven
experiments in the "VAR" series are presented for comparison.
Examination of the residuals indicates an excellent correlation of
the data with the empirical equation. It should be emphasized that
the correlation equation is not represented as a functional explanation
for the dependence of the S02 flue-gas concentration on the operating
parameters.
Figures 3, 4, and 5 illustrate, respectively, the effect of Ca/S
mole ratio, superficial fluidizing-gas velocity, and bed temperature
on the S02 flue-gas concentration. The solid lines in the figures are
based on the empirical correlation of the data given above. The
experimental data, corrected where necessary for the effect of the
third independent variable, are plotted along with the correlation
curves. For Ca/S ratios above 2.0, the S02 removal is generally
greater than 90%. The level of S02 in the flue gas increases rapidly,
however, with decreasing Ca/S ratio and with increasing gas velocity
at low Ca/S ratios. The combustion temperature appears to have very
little effect over the range of conditions investigated. This is
33
-------
considerably different from the temperature effect observed in
atmospheric-pressure combustion experiments which showed a maximum
sulfur dioxide removal (hence a minimum sulfur dioxide level) at
a combustion temperature of 1550°F.1
Table 4. Comparison Between Experimental and Calculated Values
for the S02 Flue-Gas Levels in the "VAR" Experiments
Levels of SO2 in
the Flue Gas, ppm
Experiment Experimental Calculated Residual
VAR-1
VAR- 2
VAR- 3
VAR- 4
VAR- 5
VAR- 6
VAR- 7
VAR- 8
VAR- 9
VAR-6-R
VAR-6-2R
118
129
351
122
681
169
262
131
846
209
190
110
145
340
135
690
185
255
120
845
185
185
-8
16
-11
13
9
16
-7
-11
-1
-24
-5
The results indicate that for this coal and additive it should
be possible to operate close to a Ca/S mole ratio of 1.0 and still
meet the EPA emissions limitation of 1.2 Ib of sulfur dioxide per
106 BTU. It should be emphasized, however, that the effects reported
here are based on a minimum number of observations. Since the sulfur
dioxide level in the flue gas is so highly responsive between a Ca/S
mole ratio of 1.0 and 2.0, additional data in this range are needed
to provide a greater degree of reliability to the effects observed
above.
Kinetics of SO? Retention. The pronounced effect of the Ca/S
mole ratio on the S0£ flue-gas concentration (Fig. 3) suggests that
the rate of 862 uptake by the additive is diffusion limited. At
high Ca/S ratios (;>2.0), sufficient additive surface area is available
for high S02 removal (^90% or better). As the Ca/S ratio approaches
unity, however, diffusion of reactants and products to and from the
internal surfaces of the additive particle becomes increasingly
important for high removal rates. The relatively poor removal rates
obtained at a Ca/S mole ratio of 1.0 suggest, therefore, a diffusion-
limited process.
Examination of Fig. 4 also supports the theory that the S02-
dolomite reaction is diffusion limited. In a diffusion-controlled
34
-------
TEMPERATURE: 1650 °F
EXCESS AIR :~ 15% (3% 02 IN FLUE GAS)
FLUIDIZED-BED HEIGHT : 3ft
O VELOCITY £ 5.0 ft/sec
Q VELOCITY £ 3.5 ft/sec
O VELOCITY £ 2. Oft/sec
DENOTES DATA POINT
CORRECTED FOR TEMPERATURE
S02 LEVEL CORRESPONDING TO EPA
LIMIT OF 1.2 lbsS02/IO" Btu
V= 5.Oft/sec
V= 3.5 ft/sec
V= 2.0 ft/sec
1.5 2.0 2.5
Co/S FEED MOLE RATIO
Figure 3. Effect of Ca/S Mole Ratio
on S02 Level in Flue Gas
35
-------
100
E
ex
CL
1000
900
800
cr
UJ
o
•z.
o
o
CO
<
CD
I
UJ
ID
TEMPERATURE. 1650 °F
EXCESS AIR:~ 15% (3% 02 IN FLUE GAS)
FLUIDIZED-BED HEIGHT: 3ft
O Ca/S £ I
D Ca/S £2
O Ca/S £ 3
-X- DENOTES DATA POINT
CORRECTED FOR TEMPERATURE
2 700
600
500
400
LIMIT OF 1.2 Ibs S02/I06/Btu
-------
E
o.
Q.
Z
O
(T
I-
Z
UJ
o
z
o
o
CO
UJ
ID
co
100
1000
900
800
700
600
500
400
300
200
100
GAS VELOCITY: 2 ft/sec
EXCESS AIR:- 15%(3% 02 IN FLUE GAS)
FLUIOIZED-BED HEIGHT: 3ft
O Ca/S S I
D Co/S S2
O Ca/S £ 3
* DENOTES A DATA POINT
CORRECTED FOR VELOCITY
S02 LEVEL CORRESPONDING TO EPA
LIMIT OF 1.2 lbsS02/!06 Btu
1
1
1
1400 1450
1500 1550 1600 1650
BED TEMPERATURE, °F
1700
Figure 5. Effect of Bed Temperature
on S02 Level in Flue Gas
37
-------
situation, the smaller particles are more reactive than the larger
particles because of their significantly greater external surface area
per unit mass. Increasing the superficial, fluidizing-gas velocity
through the bed in effect reduces the reactor residence time of the
smaller, more reactive particles by increasing the rate of their
elutriation from the reactor. At the high Ca/S ratios, loss of the
finer particles is less significant because of the relatively large
quantity of dolomite available to react with the sulfur. At the
lower Ca/S ratio of 1.0, however, the loss of the finer, more reactive
particles becomes more noticeable.
The relatively small effect of temperature on the S02 removal, as
shown in Fig. 5, again suggests a diffusion-controlled process.
Reaction-controlled processes should show a strong temperature
dependence.
Nitrogen Oxide Emissions. Nitrogen oxide levels in the flue gas
were extremely low over the entire range of operating conditions
tested. Values ranged from 270 ppm at a Ca/S mole ratio of 3.2,
combustion temperature of 1630°F, and fluidizing-gas velocity of 3.6
ft/sec to 120 ppm at a Ca/S mole ratio of 1.0, temperature of 1665°F,
and velocity of 4.9 ft/sec. These nitrogen oxide concentrations
correspond, respectively, to emissions of 0.40 and 0.15 Ib N02/105
BTU, as compared with the EPA emission standard of 0.70 Ib N02/106
BTU.
An analysis of variance was made for the NO data as for the S02
data, and is presented in Table 5. The only apparently significant
effect on the NO level in the flue gas was the Ca/S ratio. It would
appear from the analysis of variance that the problem of interaction
between the independent operating variables was not as serious as in
the analysis of variance for the S02 data. The error mean square of
412 compares well with the variance of 272 determined for the NO data
from the three replicate experiments.
Table 5. Analysis of Variance of NO Flue-Gas Levels in the
3x3 Latin-Square Designed Series of "VAR" Experiments
Source of
Variation
Ca/S Ratio
Temperature
Velocity
Error
Sum of
Squares
11,494
1,425
2,627
823
Degrees of
Freedom
2
2
2
2
Mean
Square
5,747
712
1,314
412
F
e
13.9
1.7
3.2
F0.1
9.00
Total
16,369
38
-------
In Fig. 6, the experimental values of the nitrogen oxide levels in
the flue gas are plotted as a function of the Ca/S mole ratio. The
broken lines in Fig. 6, which connect data from experiments performed
under nominally similar combustion temperatures, suggest a possible
temperature dependence, but the results are inconclusive on that
point. The observed correlation with Ca/S mole ratio is consistent
with the reported correlation of NO with SC>2 levels during fluidized-
bed combustion. Reduced mole ratios of Ca/S correspond to high levels
of S02 in the fluidized bed of partially sulfated additive, which in
turn has been shown to correspond with low NO levels in the flue-
gas. 2>3
The nitrogen oxide levels reported here for combustion at 8 atm
are considerably below the 300 to 550 ppm values previously obtained
during atmospheric combustion studies.1 The pressure effect on
nitrogen oxide emissions has been observed by other investigators.**
Combustion Efficiency. Combustion efficiencies, expressed as
the percentage of the total combustible carbon fed to the combustor
that was converted to carbon dioxide, have been determined for the
eleven experiments in the combustor variable study. Results of the
calculations are listed in Table 6.
Combustion efficiencies for the eleven experiments ran from 88
to 97%. In the analysis of variance, combustion temperature was the
only variable indicated to have a significant effect at the a = 0.1
level. At a given gas velocity, the data indicate that the combustion
efficiency increases with increasing temperature. The only anomaly
in the combustion efficiency data is the low value of 91% for
Experiment VAR-9. A value nearer 95% or 96% for this particular
experiment would have been more consistent with the other values
reported.
The data also suggest that, at a given temperature, the combustion
efficiency may decrease slightly with increasing gas velocity. This
can also be seen in Fig. 7, which graphically illustrates the temper-
ature dependence of combustion efficiency. The spread of the data
indicates that increasing the velocity may result in slightly lower
combustion efficiencies at a given temperature, owing perhaps to
slightly shorter hold-up times for the combustible material in the
fluidized bed. If the effect is real, however, it was too small to
be indicated as significant by the analysis of variance.
The combustion efficiencies reported here are comparable with
previously reported efficiencies obtained for atmospheric combustion
experiments.1 In the previous experiments, an Illinois No. 6 coal
containing 4.5 wt % sulfur was burned in a limestone bed at combustion
temperatures between 1550 and 1650°F, a fluidizing-gas velocity of
3 ft/sec, excess oxygen in the flue gas between 2.0 and 3.0%, and
Ca/S mole ratios between 0 and 4.2. Measured combustion efficiencies
for those experiments ranged from 94 to 97%. Whereas 10 to 20% of the
39
-------
300
e
Q.
CL
CO
<
o
UJ
CE
\-
2
UJ
O
•z.
o
o
o
250
COMBUSTION TEMPERATURE
O I450°F
D I550°F
A I650°F
200
150
100
REGRESSION
CURVE
50
FLUIDIZING GAS VELOCITY: 2-5ft/sec
EXCESS AIR :~I5%(3%02 IN FLUE GAS)
FLUIDIZED-BED HEIGHT: 3ft
0
0
23
MOLE RATIO, Ca/S
Figure 6. NO Concentration in Flue Gas
as a Function of Ca/S Mole Ratio
40
-------
Table 6. Combustion Efficiencies for "VAR"-Series Experiments
Nominal
Experiment
Bed Temp,
°F
Ca/S
Mole Ratio
Combustion
Efficiency,*
Gas velocity, 2.0 ft/sec
VAR-1 1450
VAR-3 1550
VAR-4 1650
3.0
1.0
2.0
90
96
97
Gas velocity, 3.5 ft/sec
VAR-5 1450
VAR-6 1550
VAR-6-R 1550
VAR-6-2R 1550
VAR-8 1650
1.0
2.0
2.0
2.0
3.0
89
94
94
94
96
Gas velocity, 5.0 ft/sec
VAR-7
VAR-2
VAR-9
1450
1550
1650
2.0
3.0
1.0
88
94
91
Defined as the percentage of the total combustible carbon fed
that was completely burned to C02.
100
98
96
Ca/S RATIO: 1-3
EXCESS AIR : ~ 15 % (3% 0, IN FLUE GAS)
— FLUIDIZED-BED HEIGHT: 3ft
GAS VELOCITY, ft/sec:
O ~ 2.0
0 ~ 3.5 o
A ~ 5.0
94 -
* 92-
CD
§ 90
86
86
O
_L
_L
_L
Figure 7. Combustion Efficiency
as a Function of Bed Temperature
1400 1450
1500 1550 1600
TEMPERATURE. °F
1650
1700
41
-------
unhurried carbon was in the form of carbon monoxide and methane in the
flue gas for the atmospheric pressure experiments, only 0.2 to 1.3%
of the unburned carbon was present as carbon monoxide and methane in
the flue gas for the experiments at 8 atm. The major carbon loss
in the tests at 8 atm occurred as the result of the elutriation of
fine, unburned carbon particles from the fluidized bed.
In an application of the fluidized-bed combustion concept, the
fly ash containing the unburned carbon will be fed to a second
fluidized bed to complete the combustion. Earlier work has shown
that combustion efficiencies of over 99% are then readily achieved.5
Combustible Carbon Hold-up in the Fluidized Bed. Related to
combustion efficiency is the combustible carbon hold-up in the
fluidized bed. High carbon inventories in the fluidized bed could
present highly unstable and hazardous operating conditions, particularly
during start-up and shut-down of a unit.
To provide a measure of the combustible carbon hold-up in the
fluidized bed during combustion, samples of the final bed material
from the experiments in the variable study were analyzed for total
carbon and for carbon present as carbonate ion. The combustible
carbon content was taken as the difference between these two values.
The results of these calculations are presented in Table 7.
For the eleven experiments, the combustible carbon in the final
bed material ranged from <0.01 to 1.60% and was £l.0% in seven of the
experiments. These levels of carbon are sufficiently low to ensure
safe, stable operating performance. The analysis of variance indicates
no significant effects of the operating conditions on the combustible
carbon levels in the final bed samples.
Additive Utilization. Of considerable economic importance is the
level of additive utilization that can be achieved for a given sulfur
removal. Additive utilization, or degree of sulfation, is defined as
the percentage of the total moles of calcium in the additive converted
to calcium sulfate. As the ratio of calcium in the additive to the
sulfur in the coal decreases, the additive utilization increases.
This is accompanied, however, by a net decrease in the amount of
sulfur retained by the additive and results in higher sulfur dioxide
emissions. For the Arkwright coal, for example, the EPA emission
standard requires approximately 71% sulfur removal. It should be
theoretically possible, therefore, to meet the requirements using a
Ca/S ratio of 0.71, assuming 100% utilization of the additive.
Table 8 lists the calculated calcium utilization values for the
solid samples recovered from the bed, two cyclones, and a filter at
the conclusion of each experiment. The entrained solid samples
represent solids recovered from the flue gas during the entire
experiment, including startup. The startup time was relatively short
compared with the total time of the experiment, and run times were of
42
-------
Table 7. Combustible Carbon Content of Final Bed
Material for "VAR"-Series Experiments
Nominal
Experiment
Bed temp, 1450°F
VAR-1
VAR-5
VAR-7
Bed temp, 1550°F
VAR-3
VAR-6
VAR-6-R
VAR-6-2R
VAR-2
Bed temp, 1650°F
VAR-4
VAR-8
VAR-9
Gas Velocity,
ft/sec
2.0
3.5
5.0
2.0
3.5
3.5
3.5
5.0
2.0
3.5
5.0
Ca/S, mole
ratio
3.0
1.0
2.0
1.0
2.0
2.0
2.0
3.0
2.0
3.0
1.0
Final Bed Com-
bustible Carbon,
wt %
0.60
1.27
<0.01
0.68
1.60
<0.01
1.00
1.30
0.85
0.36
0.42
Estimated accuracy, ±10%.
Table 8. Calcium Utilization in Solids Outlet Streams
for the Combustion of Arkwright Coal in a Fluidized Bed
of Tymochtee Dolomite
Calcium Utilization
Experiment
VAR-1
VAR-8
VAR-2
VAR-4
VAR-6
VAR-6-R
VAR-6-2R
VAR-7
VAR-3
VAR-5
VAR-9
Ca/S
Mole
Ratiob
2.9
3.2
2.9
1.9
2.0
2.0
2.0
2.2
1.1
1.0
1.0
Fluidization
Velocity,
ft/sec
2.1
3.6
4.8
2.3
3.6
3.6
3.5
4.2
2.1
3.4
4.9
Final
Bed,
%
37.2
32.3
34.1
49.8
41.3
49.7
48.5
47.3
77.4
76.5
82.7
Primary
Cyclone,
%
34.9
25.5
30.1
35.1
37.9
41.7
41.1
39.0
55.5
56.9
38.4
Secondary
Cyclone,
%
52.6
37.4
47.9
32.6
56.5
62.5
59.5
37.5
66.3
65.2
54.8
Filter,
%
93.7
27.3
50.0
39.8
75.0
82.0
77.8
46.5
79.2
77.8
88.9
Llefined as the percentage of total moles of calcium present as
In coal and additive feed streams.
43
-------
sufficient duration to allow for ^80% bed replacement, assuming
perfectly backmixed flow of additive through the combustor. The
final bed samples, therefore, are fairly representative of steady-
state conditions.
Examination of the data in Table 8 indicates that utilization of
additive in the final bed material was affected inversely by the Ca/S
ratio and relatively unaffected by the fluidizing-gas velocity or bed
termperature. This is graphically illustrated in Fig. 8, which depicts
the sulfation of the final bed material as a function of the Ca/S mole
ratio. According to the analysis of variance, the effect of Ca/S mole
ratio on final bed utilization is highly significant at the a = 0.005
level.
It is interesting to note, however, that the additive utilization
was consistently higher for the final bed material than for the
elutriated solids collected in the primary cyclone. In addition,
calcium utilization for elutriated solids generally increased the
farther downstream the particulate matter was collected from the flue
gas. This latter result is attributed to increasing reactivity with
decreasing particle size of the particulate matter removed sequentially
from the primary cyclone, the secondary cyclone, and the filter.
<
LLl
K
<
S
a
LJ
CD
o
100
90
80
70
N 60
P
£ 50
a
§ 40
30
I I \
COMBUSTION TEMPERATURE: I450-I650°F
EXCESS AIR:~I5%(3% Oj, IN FLUE GAS)
FLUIDIZED-BED HEIGHT: 3ft
FLUIDIZING GAS VELOCITY:
O S 2.0 ft/sec
D S 3.5 ft/sec
A = 5.0 ft/sec
Ca/S RATIO
Figure 8. Additive Utilization of Final Bed Material
as a Function of Ca/S Mole Ratio
44
-------
Removal of Additive from Combustor by Elutriation. The results
of the preceeding section on utilization indicate a definite advantage
in minimizing the amount of additive leaving the combustor entrained
in the flue gas. Not only is the utilization consistently higher for
the additive in the fluidized bed (and hence for the additive leaving
the combustor by overflow through the standpipe) than for the bulk of
the entrained additive removed from the flue gas by the primary cyclone,
but high additive entrainments may also present additional problems in
flue-gas cleaning to meet the standards for particulate emissions.
The two factors most strongly affecting elutriation of additive from
the combustor are the' gas velocity relative to the particle-size
distribution of additive in the feed (terminal velocity effects) and
the amount of decrepitation of additive that occurs during fluidization.
Terminal Velocities of Additive Particles. A theoretical consid-
eration was made of the terminal velocities associated with different
sized particles of additive. By means of a terminal-velocity correlation
reported by Kunii and Levenspiel,6 a series of curves was derived to
illustrate the effect of system pressure (i.e,3 as it affects the
density of the fluidizing gas), particle density, and particle
sphericity on the relationship of particle size to its terminal
velocity. The results of these calculations are presented in Fig. 9-
The top curve relates the terminal gas velocity as a function of
particle size for spheres having a particle density of 2.24 g/cm3 (an
approximate value for half-calcined dolomite). The fluidizing gas was
considered to be at a temperature of 1800°F and a pressure of 1 atm.
At a fluidizing-gas velocity of 5 ft/sec, no entrainment of +45 mesh
solids would be expected. The lower curves, however, illustrate the
effect of including a factor for the nonspherical nature or lack of
sphericity of real particulate matter. By assuming a sphericity of
1.0 (true spheres), a typical flue-gas composition of the major
components, a temperature of 1450°F, and a pressure of 8 atm, a set
of curves was calculated for particle densities of 2.85 and 2.24 g/cm3
(estimated range of density for half-calcined, partially sulfated
dolomite). Again, little if any entrainment of +45 mesh additive
would be expected at fluidizing velocities of only 5 ft/sec. A
repetition of the calculations for sphericities of 0.8 and 0.6
produced the lower two sets of curves. A sphericity of 0.6 or lower
is common for crushed particulate materials.6 At these lower values
of sphericity, a fluidizing-gas velocity of 5 ft/sec is well above
the terminal velocity required to entrain additive in the +45 mesh
range. The two data points in Fig. 9 correspond to the maximum
particle size of additive recovered from the primary cyclone for "VAR"
experiments made at 2.1 and 4.9 ft/sec. It would appear, therefore,
that 80-90% of the dolomite used in the combustion experiments had a
terminal velocity below 5 ft/sec.
Experimental Elutriation Rates. On the basis of calcium material
balances made for the experiments in the variable study, it was
possible to calculate values for additive entrainment expressed as a
percentage of the additive feed rate. In these experiments, the
45
-------
80
70
60
50
40
30
o
§ 20
to
3 10
TOP CURVE: FLUIDIZING GAS; AIR
TEMP: 1800 °F
PRESSURE: lotm
PARTICLE DENSITY: 2.24 g/cm3
LOWER SETS OF CURVES:
FUimm GAS.' 3%02; 7%H20, 1
TEMP: 1450 °F
PRESSURE; 8 atm
PARTICLE DENSITY: 2.89 (SOLID LINE)
2.24 (BROKEN LINE)
<£8-SPHERICnT OF PARTICLES
o- DATA POINTS "VAR" EXPERIMENTS
CURVES DERIVED FROM TERMINAL
VELOCITY CORRELATIONS IN "FLUIDIZATION
ENGINEERING" BY KUNII AND LEVENSPIEL
MESH SIZE
45 40 35 30 25 20 18 16 14 12 10 8
i i i i ii i i i i i i
400 800 1200
PARTICLE DIAMETER,
2400
I I
1.5 2.0 2.5 3 45678
PARTICLE DIAMETER, in. xlO2
Figure 9. Terminal Gas Velocity of Solid Particles
in a Gas Flow as a Function of Particle Diameter
for Various Physical Situations
46
-------
particle-size distribution of the dolomite fed to the combustor was
held constant.
The results of the calculations are presented in Fig. 10, which
depicts the relative rate of additive entrainment from the combustor
as a function of the superficial-gas velocity. Entrainment of the
additive increases rapidly from ^5% at 2 ft/sec to as much as
80% entrainment at a velocity of 5 ft/sec. The high elutriation rates
at 5 ft/sec fluidization velocity are certainly in keeping with the
theoretical terminal velocities of the dolomite particles.
o
o
100
90
80
70
60
50
40
30
20
10
T
T
T
T
T
T
TEMPERATURE : I450-I650'F
EXCESS AIR :~I5% (3%02 IN FLUE GAS!
FLUIOIZED-BEO HEIGHT: 3ft
REGRESSION LINE
O Co/S = I
Q Co/S = 2
A Co/S = 3
I
J_
_L
I
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
SUPERFICIAL GAS VELOCITY , ft/sec
5.5
Figure 10. Entrainment of Sulfated Dolomite
as a Function of Superficial Gas Velocity
47
-------
The data suggest that the inverse effect of velocity on sulfur
removal is a result of the increased entrainment of additive at the
higher velocities. We previously observed that velocity had no
significant effect on the utilization of additive in the fluidized
bed. It may be possible, therefore, to offset the inverse effect of
velocity on sulfur retention by using a larger additive particle size
at the higher velocities to reduce entrainment.
Decrepitation of Tymochtee Dolomite. A semi-quantitative attempt
was made to determine to what extent the high entrainments of additive
reported above were the result of decrepitation of additive followed
by elutriation of the resulting fines. The method chosen to provide
a measure of additive decrepitation during a combustion experiment was
to inventory the +45 mesh additive material (particle diameter >354 ym)
entering and leaving the combustion system. A more exacting or
detailed analysis of decrepitation was hindered by the presence of
fly ash in the entrained solids removed from the flue gas in the
cyclone separators.
This is illustrated by a visual analysis of the different size
cuts (+45, -100 to +170, and -325 mesh) of a typical primary cyclone
product shown in the photograph of Fig. 11. The +45 mesh fraction of
the product appears to be essentially free from fly ash and unburned
carbonaceous materials. The two finer size fractions, however,
obviously contain considerable amounts of interfering ash and carbon.
Chemical analysis of the +45 mesh material in Fig. 11 indicates a
total carbon (from unburned carbon and CaC03) of only 3.9%. Chemical
analysis of the unsplit material indicates a level of 4.1% carbon as
carbonate in the additive. Thus, the +45 mesh fraction does not contain
any significant levels of carbon or ash derived from the coal.
Conversely, the -100 +170 and -325 mesh fractions illustrated in
Fig. 11 contain 36 and 20 wt % carbon, respectively, which implies a
considerable carbon and ash content derived from the coal.
The screen-analyses data for the "VAR" experiments (tabulated in
Appendix C) were used to inventory the +45 mesh additive entering and
leaving the combustor. Weights of half-calcined and partially sulfated
additive were corrected on the basis of calcium-utilization data in
Table 8 to the equivalent weight of fresh unreacted stone represented
by the sample. An assumption made in the analysis was that the size
of a particle does not change appreciably during calcination or
sulfation. Minor changes in size would only affect the very small
percentage of additive with a particle diameter at or near the 354 ym
cut point. The +45 mesh fraction, however, accounts for approximately
87 wt % of the unreacted additive fed to the reactor.
The results of the calculations for the eleven experiments in
the variable study are presented in Table 9. Additive recovery values
determined by the above procedure ranged from 78 to 108% for the
individual experiments. Average recoveries, however, calculated for
each level of the three independent operating variables only ranged
48
-------
from 92 to 103% as compared with the average recovery of 96% for all
the experiments. The data indicate that very little decrepitation
of the additive occurs over the range of conditions tested.
+45
m
-too
+ 170
-325
Figure 11. Various Size Fractions of Primary Cyclone Product
Recovered during Combustion Experiment VAR-6-2R
Heat-Transfer Coefficients at the Combustor Wall. Values of heat-
transfer coefficients at the combustor wall have been determined for
the eleven experiments in the statistically designed series of
combustion experiments. Figure 12 illustrates the factors considered
and the method used in the calculation of the coefficients. Using the
notation given in Fig. 12, the amount of heat transferred through the
reactor wall was determined by means of the equation.
=
-------
Table 9. Inventory of +45 Mesh Additive Entering and Leaving the
Combustion System for the Eleven Experiments in the "VAR" Series
of Combustion Experiments
Experiment
+45 Mesh Additive
a
Entering System,
kg
+45 Mesh Additive
Leaving System,
kg
Recovery of +45
Mesh Additive,
VAR-1C
VAR-2
VAR-3
VAR-4
VAR-5C
VAR- 6
VAR-6-R
VAR-6-2R
VAR-7C
VAR-8C
VAR- 9
32.8
41.5
44.6
45.6
53.6
44.8
43.4
43.3
41.6
50.3
44.2
34.7
40.9
34.6
42.5
58.1
45.0
40.7
42.0
39.4
45.0
45.5
105
98
78
93
108
100
94
97
95
89
103
Variable
Nominal
Operating Levels
Average Recovery Value of
+45 Mesh Additive, %
Velocity, ft/sec
Temperature, °F
Ca/S Mole Ratio
2.0
3.5
5.0
1450
1550
1650
1.0
2.0
3.0
92
98
99
103
93
95
96
96
97
Includes initial bed and dolomite fed to the combustor.
Includes final bed, overflow, primary cyclone, and secondary
cyclone products.
•»
'Required an estimate of the size distribution of the initial
bed based on the previous history of the additive material.
Average recovery for all "VAR" experiments = 96%.
50
-------
SENSIBLE HEAT
REMOVED BY
INTERNAL COOLING
COiLS
WALL
TEMPERATURE-
SENSIBLE HEAT REMOVED IN
GASEOUS AND ELUTRIATED
SOLID PRODUCTS
FLUIDIZED
BED
/> HEAT ^2
/OF REACTION}
BED
TEMPERATURE,
SENSIBLE HEAT
REMOVED THROUGH
REACTOR WALL TO
EXTERNAL COOLING
COILS
SENSIBLE HEAT ADDED
IN REACTANTS
SENSIBLE HEAT REMOVED
IN OVERFLOW PRODUCTS
Figure 12. Factors Considered in the Calculation
of Heat-Transfer Coefficients at the Combustor Wall
Assumptions inherent in Eq. 4 include: (1) heat transfer by
radiation from the surface of the bed to metal surfaces above the
level of the bed is negligible; (2) heat losses by conduction through
the gas distributor supporting the fluidized bed and by conduction in
the axial direction in the reactor walls are negligible; (3) combustion
products (gaseous and solid) leave the fluidized bed at the bed
temperature; and (4) the internal cooling coils remove heat only from
the fluidized bed (i.e., not from the gas above the bed). The first
two assumptions tend to bias qw toward higher than actual values,
whereas the last two assumptions tend to bias qw toward lower than
actual values. Thus the omitted terms in Eq. 4, all estimated to be
relatively minor compared to the amount of heat transferred between
the fluidized bed and the vessel wall, tend to balance out the
possibility of a large net cumulative bias of qw in either direction.
After obtaining a value for the heat transferred through the
reactor wall, qw, by Eq. 4, the bed-to-wall heat-transfer coefficient,
Uj[, was calculated by means of the equation
Ui =
AT
(5)
51
-------
where Aj_ is the inside surface area of the combustor wall between the
distributor surface and the top of the standpipe regulating the height
of the fluidized bed, and AT is the temperature difference between
the bed temperature, T^, and the inside wall surface temperature, T^.
As TJ was not measured directly, it was determined by an iterative
solution of the following equations
T. = T -
i w
m
< V
(6)
' ko * C(T» + V (7)
m T
where kj equals the thermal conductivity of the reactor wall at the
average wall temperature, AL is the log mean wall area for heat
transfer, Xw represents the wall thickness, TW is the experimentally
determined outside wall temperature, and ko and C are constants.
The values of bed-to-wall heat-transfer coefficients determined
for the eleven experiments in the "VAR" series are presented in
Table 10 and Fig. 13 as a function of velocity. Values of the
coefficients range from about 40 BTU/(hr)(ft2)(°F) at a gas velocity
of about 2 ft/sec to more than 135 BTU/(hr)(ft2)(°F) at 5 ft/sec.
The high degree of scatter that the data exhibit may be attributable
to large uncertainties regarding (1) the effective height of the
fluidized bed, because this determines the wall surface area actually
being affected by the fluidized bed for heat transfer, and (2) the
particle density of the fluidized bed, because this directly affects
both the rate of heat transferred by conduction and convection through
the lamina or sublayer of gas at the combustor wall. Bed temperature
may also be a significant factor affecting the rate of radiant heat
transfer between the bed and vessel wall. The effect of temperature
is not obvious, however, from the data shown in Fig. 13.
The values reported here for bed-to-wall coefficients appear to
be slightly higher than the 50 to 100 BTU/(hr)(ft2)(°F) values reported
for other fluidized-bed combustion systems as heat-transfer coefficients
between the bed and tube surfaces immersed in the bed.7
Solids Loading in the Flue Gas. Data on the distribution of
flue-gas particulates in the "VAR" series of experiments are
presented in Tables 11 and 12. The flue-gas particulate concentration
is expressed as grains/scf in Table 11, because these units are
generally used in determining the suitability of the pressurized flue
gas for driving a gas turbine. In Table 12, the distribution of flue-
gas particulates is expressed as lb/10& BTU. The latter units are
52
-------
Table 10. Heat-Transfer Coefficients for the "VAR" Series
of Statistically Designed Combustion Experiments
Experiments
Fluidizing-Gas
Velocity, ft/sec
Bed-to-Wall Heat-Transfer
Coefficient3, BTU/(hr)(ft2)(°F)
VAR-1
VAR- 3
VAR- 4
VAR- 5
VAR- 6
VAR-6-2R
VAR-6-R
VAR- 8
VAR- 7
VAR- 9
VAR- 2
2.1
2.1
2.3
3.4
3.5
3.5
3.6
3.6
4.2
4.9
4.8
65
42
78
65
84
101
96
115
100
96
136
Based on inside area of combustor wall.
140
•z.
UJ
120 -
o
o
cc ,-
UJ u.
100
!
cc. •-
X 80
tr 60
"a
40
o
i-
Q
I
T
D
Co/S RATIO: 1-3
EXCESS AIR :~I5% (3%02 IN FLUE GAS)
. FLUIDIZEO-BED HEIGHT: 3ft
TEMPERATURE, "F:
O ~ 1450
D ~ 1550
A ~ 1650 a /O
I
I
234
VELOCITY, ft/sec
Figure 13. The Effect of Velocity on Calculated
Bed-to-Wall Heat-Transfer Coefficients
53
-------
Table 11. Distribution of Flue-Gas Particulates in "VAR"-Series Experiments
Expressed in Units of Grains/scf
Nominal
Experiment
Gas velocity, 2.0 ft/sec
VAR-3
VAR-4
VAR-1
Gas velocity^ 3.5 ft/sec
VAR-5
VAR-6
VAR-6-R
VAR-6- 2R
VAR-8
Gas velocity, 5.0 ft/sec
VAR-9
VAR-7
VAR-2
Ca/S,
Mole Ratio
1.0
2.0
3.0
1.0
2.0
2.0
2.0
3.0
1.0
2.0
3.0
Bed Temp,
°F
1550
1650
1450
1450
1550
1550
1550
1650
1650
1450
1550
Q
Flue-Gas Particulate Concentration, grains/scf
After
Combustor
4.1
4.8
6.3
7.5
8.3
10.8
10.0
13.5
10.4
17.8
23.4
After 1st
Cyclone
0.7
0.5
0.8
1.3
0.6
0.6
0.8
0.9
0.6
4.7
1.2
After 2nd
Cyclone
0.3
0.3
0.3
0.5
0.2
0.3
0.4
0.3
<0.3
2.1
0.7
After
Filter
b
b
b
b
b
b
b
b
<0.1
<0.1
b
Estimated accuracy, +10%.
No data obtained.
-------
Ui
Table 12. Distribution of Flue-Gas Participates in "VAR"-Series Experiments
Expressed in Units of lb/106 BTU
Nominal
Experiment
Gas velocity, 2.0 ft/sec
VAR-3
VAR-4
VAR-1
Gas velocity, 3.5 ft/sec
VAR-5
VAR-6
VAR-6-R
VAR-6-2R
VAR-8
Gas velocity, 5.0 ft/sec
VAR-9
VAR-7
VAR-2
Ca/S,
Mole Ratio
1.0
2.0
3.0
1.0
2.0
2.0
2.0
3.0
1.0
2.0
3.0
Bed Temp,
°F
1550
1650
1450
1450
1550
1550
1550
1650
1650
1450
1550
Flue-Gas Particulate Level, lb/106 BTUa
After
Combustor
7.3
8.8
10.5
12.3
13.8
18.7
16.7
24.0
17.1
27.9
39.1
After 1st
Cyclone
1.2
0.8
1.3
2.2
0.9
1.0
1.3
1.6
1.0
7.4
2.0
After 2nd
Cyclone
0.5
0.5
0.5
0.9
0.4
0.5
0.7
0.6
<0.5
3.3
1.1
After
Filter
b
b
b
b
b
b
b
b
<0.1
0.1
b
Estimated accuracy, +10%.
No data obtained.
-------
used by EPA for control of particulate emission from a fossil-fuel-
fired steam-generating plant. Such a plant has a maximum particulate
emission standard of 0.1 lb/106 BTU heat input.
The solids loading in the flue gas as it left the combustor is
graphically presented in Fig. 14 as a function of Ca/S mole ratio at the
three levels of gas velocity studied. Loadings ranged from a low of 4,1
grains/ft3 at a Ca/S ratio of one and velocity of 2 ft/sec to a high of
23 grains/ft3 at a Ca/S ratio of three and velocity of 5 ft/sec. Extrapo-
lation of the curves to zero Ca/S ratio would roughly indicate the
contribution to each curve made by the fly ash from the coal. At gas
velocities of 2, 3.5, 5 ft/sec, increasing the Ca/S mole ratio from one
to three increased the solids loading in the flue gas by 60, 80, and 125%,
respectively. While these results suggest the desirability of maintaining
low Ca/S ratios to minimize solids loading from the additive, loadings at
high Ca/S ratios could be reduced significantly at suitably selected gas
velocities and sorbent particle sizes. Thus, the loadings quoted here
should not be considered representative of what could be acheived under
optimized conditions.
The data presented in Tables 11 and 12 on the concentration of
flue-gas particulates after the two cyclones and the filters are for
informational purposes only and should not be construed as representing
the limit of particulate retention. Design of the cyclones and filters
used in the ANL combustion pilot plant was not optimized for particulate
collection.
Carbon, Sulfur, and Calcium Material Balances. Carbon, sulfur,
and calcium material balances have been calculated for the "VAR"
experiments. The calculations were based on chemical analyses of
riffled samples of the various solids streams added to, or removed
from, the combustor and on analyses of the flue gas for C02, CO, CH^,
and SC>2. The results are summarized in Table 13. More detailed
material-balance data for each of the eleven experiments in the "VAR"
series are presented in Appendix B of this report, Tables B-l through
B-ll, inclusive.
As shown in Table 13, the material balances ranged from 103 to
116% for carbon, from 80 to 99% for sulfur, and from 79 to 107% for
calcium. In light of the numerous possibilities for inaccuracies in
weighing, sampling, and analysis of the various streams, material
balances of 100 + 10% are considered acceptable. Of the 33 values
shown in Table 13, 25 (or 76%) of the values are within this range.
Particle-Size Distribution of Fluidized-Bed and Elutriated Solids.
Screen analysis data for the "VAR" experiments are presented in Table C-l
of Appendix C and summarized in Fig. 15, which is a pilot of the mass-mean
particle diameters of the indicated samples against gas velocity. The
data reflect the increased elutriation of additive with increased velocity
as shown in Fig. 10. As the gas velocity increases, progressively larger
bed particles are elutriated from the bed and retained in the cyclones.
Thus, the mean diameters of all the samples tend to increase with increas-
ing gas velocity.
56
-------
28
26
24
O
^ 22
'5
w 20
I le
5% sulfur, analytical accuracy, 45% relative for
solid samples. For <5% sulfur, analytical accuracy,
+_10% relative for solid samples.
"Analytical accuracy, +2% relative.
57
-------
UJ
UJ
2
5
UJ
_j
o
t-
cc
2
<
UJ
CO
to
1000
900
800
700
600
500
400
300
200
100
0
o DOLOMITE FEED
SAMPLE:
o FINAL BED
D PRIMARY CYCLONE
O SECONDARY CYCLONE
' PRIMARY
CYCLONE
SECONDARY
CYCLONE
01 23456
FLUIDIZ1NG GAS VELOCITY, ft/sec
Figure 15. Mean Particle Diameters
of Solid Samples from the "VAR"
Series of Combustion Experiments
vs. Fluidizing-Gas Velocity
The results of limited experience gained in the use of a Brink iro-
pactor to measure the solids loading and particle-size distribution in
the flue gas being vented from the system are presented in Appendix D.
Combustion-Side Corrosion of Internal Cooling Coils. Corrosion on
the external (or bed) side of an internal cooling coil was studied after
an estimated 500 hr of operating time. Samples No. 3 and 11, taken from
the cooling coil where it extended below the surface of the fluidized
bed, exhibited intergranular corrosion to a depth of ^30 pm (see Fig. 16).
Both samples exhibited sensitization and sigma phase formation. Sensi-
tization of type 310 SS occurs at 800 to 1500°F and sigma phase forma-
tion at 1200°F. It should be noted that tube temperatures were not care-
fully controlled and may have occasionally approached the temperature of
the fluidized bed. In a commercial unit with optimized tube materials
and carefully controlled metal temperatures, it should be possible to
keep corrosion within acceptable limits.
Sample No. 12, shown in Fig. 17, was taken from the cooling coil
in the freeboard area of the combustor, "^4 1/2 ft above the bed. Very
little internal, or coolant-side, corrosion was observed, but external,
or bed-side, corrosion was to a depth of 600 ym. The sample was exam-
ined using an electron microprobe. Photographic images show some deple-
tion of Fe and Ni along the grain boundaries where corrosion took place.
Sulfur is seen in these grain boundaries to the depth of corrosion. Elec-
trolytic etching of the sample showed that the metal was sensitized.
This particular sample had seen vLOOO hr of operating time.
58
-------
As-polished Photomicrograph of
Sample No. 3 (275X)
b. As-polished Photomicrograph of
Sample No. 11 (275X)
Figure 16. Sample No. 3 and 11, Showing Intergranular -
Type Corrosion on the Bed-Side Wall of an
Internal Cooling Coil Immersed in the
Fluidized Bed
59
-------
a. As-polished Photomicrograph
of Sample No. 12 (116X)
b. Specimen Current
c. Iron Phase
d. Nickel Phase
e. Chromium Phase
f. Sulfur Phase
Figure 17. Photomicrograph and Electron Microprobe Scanning Images of Intergranular-Type Corrosion
Found in Process-Gas Side of Sample Removed from an Internal Cooling Coil in the Free-
board Area (All Images 40x50
-------
Combustion of Low-Sulfur, Subbituminous and Lignite Coals
Combustion experiments were made in the 6-in.-dia pressurized
combustor to determine whether any difficulties would be encountered
in processing a San Juan mine, subbituminous coal having a high coal-
ash content of 17% and a Glenharold mine lignite having a low heating
value of 7,625 BTU/lb. The nominal operating conditions for the two
experiments were a bed temperature of 1550°F, fluidizing-gas velocity
of 3.5 ft/sec, and an 02 concentration of 3% in the dry flue gas. Due
to the low sulfur content (<1 wt %) of the two coals, a Ca/S mole ratio
of 1.0 was selected. In Table 14 are summarized the actual operating
conditions and flue-gas analyses for the two experiments. Included
in this table are estimated SC-2 and NO flue-gas compositions for
Arkwright coal combusted under nominally similar operating conditions.
The estimates are based on correlations reported for the results of
combustion experiments with Arkwright coal.
Operating Performance. The operating performance of the fluidized-
bed combustor during the combustion of both the subbituminous and
lignite coals was excellent. These results, coupled with the performance
of the combustor in processing the highly-caking, high-volatile-
bituminous Arkwright coal, successfully demonstrate the versatility
of the fluidized-bed concept for processing coals of widely varying
rank and quality.
Sulfur Dioxide Retention. The 862 levels of 250 and 120 ppm for
the combustion of the subbituminous and lignite coals, respectively,
correspond to emissions of 0.45 and 0.21 Ib S02/106 BTU. By way of
comparison, combustion of the Arkwright bituminous coal under similar
conditions would have a projected 862 emission of 610 ppm or 1.15 Ib
S02/106 BTU. Correspondingly, the above S02 emissions represent
sulfur retentions of approximately 72, 72, and 85% for the bituminous,
subbituminous, and lignite coals, respectively.
The somewhat higher retention reported for the combustion
experiment with lignite may be a result of the relatively high calcium
content of the coal itself. Data in Table 15 show that Ca/S ratios
calculated just on the basis of calcium and sulfur contents for the
respective coals range from as low as 0.08 for the bituminous coal up
to 0.96 and 1.94 for the subbituminous and lignite coals, respectively.
Including the coal calcium with the calcium in the additive results in
potentially effective Ca/S ratios of 1.1, 2.1, and 3.0 for the three
combustion experiments as indicated in Table 15.
The potential effect of the coal calcium on sulfur retention
during combustion is indicated by the percentage of the sulfur in the
coal which was reportedly retained by the coal during ashing for
mineral analyses. For the bituminous coal, where the coal Ca/S ratio
is only 0.08, the amount of sulfur retained during ashing was only
3.7%. Sulfur retentions during ashing for the subbituminous and
lignite coals, however, increased to as high as 42 and 99%, respectively.
61
-------
Table 14. Actual Operating Conditions and Flue-Gas Analysis for Combustion Experiments
with Low-Sulfur Subbituminous and Lignite Coals
Experiment
Operating Conditions
Coal : Mine
Rank
Total time, hr
Time at steady operating
conditions, hr
Combustor pressure, atm
Bed temperature, °F
Ca/S mole ratio in feed
Gas velocity, ft/sec
Coal feed rate, Ibs/hr
Dolomite feed rate, Ibs/hr
Air feed rate, cfm at 70°F
and 1 atm
b
Flue-Gas Analysis
02, %
S02, ppm
NO , ppm
CO , ppm
C02, %
SJ-1A
San Juan
Subbituminous
21.9
19.0
8.0
1565
1.1
3.6
45.3
2.5
78
2.9
250
150
50
17
LIG-1
Glenharold
Lignite
27.2
24.7
8.0
1560
1.1
3.5
54.4
2.0
78
3.2
120
130
80
17
"VAR"-Typea
Arkwright
Bituminous
—
—
8.0
1550
1.0
3.5
29.4
5.2
77
3.0
610C
140C
—
—
Values obtained in this column are for comparison purposes only. Experiment was not
made at these conditions.
Dry basis.
r*
Predicted values based on correlations derived from data of "VAR"-series experiments.
-------
Table 15. Potential Effect of the Calcium Content of Coals Tested
on the Retention of Sulfur during Fluidized-Bed Combustion
Experiment
Coal : Mine
Rank
Coal sulfur content, wt %
Coal calcium content, wt %
Coal Ca/S ratio
Percent sulfur retained by
coal during ashing for
mineral analysis3
Ratio of calcium in additive
to sulfur in coal for
experiment
Total potentially effective
Ca/S ratio
Sulfur retention, %
"VAR"-Type
Arkwright
Bituminous
2.82
0.27
0.08
3.7
1.0
1.1
72b
SJ-1A
San- Juan
Subbituminous
0.78
0.75
0.96
42
1.1
2.1
72
LIG-1
Glenharold
Lignite
0.53
1.03
1.94
99
1.1
3.0
85
a
Chicago, Illinois.
Estimated from results of combustion experiments with Arkwright coal.
-------
A further indication of the activity of coal calcium in sulfur
retention was observed during a combustion experiment in which Arkwright
coal was combusted in a fluidized bed of Alundum. Flue-gas analyses
for the particular experiment indicated a sulfur retention in the bed
of approximately 3% as compared with 3.7% retained by the coal during
ashing.
Nitrogen Oxide Emissions. The NO levels of 150 ppm and 130 ppm,
respectively, for the combustion experiments with the subbituminous
and lignite coals correspond to emissions of 0.19 and 0.18 Ib NC^/IO^
BTU. The projected emission for the Arkwright coal (140 ppm) under
similar conditions is 0.19 Ib N02/106 BTU.
Combustion Efficiencies. Combustion efficiencies for the two
experiments were 94 and 97% for the San Juan and Glenharold coals,
respectively. These values agree well with the combustion efficiency
of 94% reported for the Arkwright coal at a combustion temperature of
1550°F (see Fig. 7).
Calcium Utilization. Calcium utilizations for the final bed
material of 74 and 71% for the combustion experiments with the San Juan
and Glenharold coals, respectively, in Tymochtee dolomite at a Ca/S
mole ratio of 1.0, are in relatively good agreement with the 76 to
83% utilizations reported above for combustion experiments with Arkwright
coal using the same dolomite and the same mole ratio. The slightly
lower utilizations for the lower rank coals is undoubtedly a reflection
of the high calcium content of the coals themselves. The relatively
high calcium ashes of the low rank coals, even at low ash concentrations
in the final bed material, may give somewhat misleading and depressed
additive utilizations (assuming coal Ca is not as fully utilized as
additive Ca).
Material Balances and Screen Analyses. Overall carbon, sulfur,
and calcium material balances were calculated for the combustion
experiments with the low-sulfur coals. The material balances are
presented in Tables B-12 and B-13 of Appendix B for experiments SJ-1A
and L1G-1, respectively. Except for the low Ca balance (75% recovery)
for experiment SJ-1A and the high sulfur balance (116% recovery) for
experiment L1G-1, the material balances are similar to those reported
for the "VAR"-series experiments.
Particle-size data for the starting bed, final bed, bed overflow,
and elutriated solids collected in the primary cyclone for experiments
SJ-1A and LlG-1 are shown in Table C-2 of Appendix C. No particle-
size data were obtained for the elutriated solids collected in the
secondary cyclone or the primary and secondary filters because the
material collected in these units "balled" when screening was attempted.
64
-------
Effect of Combustor Pressure and Temperature on the Concentration of
NO and Other Gases in the Flue Gas
The concentration of NO in the flue gas is significantly lower
in coal combustion experiments made at eight atmospheres pressure than
at one atmosphere. In the "VAR" series of experiments at eight
atmospheres pressure, NO concentrations ranged from 120 to 270 ppm
(see Table 2). In atmospheric tests made with a variety of coals and
conditions in the previously installed atmospheric-pressure combustor,
the NO concentration was generally above 300 ppm (see ANL/ES/CEN-1004).
Two series of experiments were conducted for the specific purpose of
studying the effect of combustor operating pressure on concentrations
of NO and other gases in the flue gas. The NOX-I series involved the
combustion of Arkwright coal in a fluidized bed of Alundum. In the
NOX-II series, Arkwright coal was combusted in a fluidized bed of
sulfated dolomite while feeding dolomite to the bed at a Ca/S feed
mole ratio of 1.5. Nominal operating conditions in both series were
a bed temperature of 1550°F, 3% 02 in the off gas, and a fluidizing-
gas velocity of 3.5 ft/sec. Each of the two series of experiments
consisted of six parts with the combustor pressure ranging from 1 to
8 atm. In a third series of experiments, the system pressure was
maintained at 4 atm (60 psia), and the bed temperature was changed
from 1450 to 1550 and to 1650°F.
In addition to the concentrations of gases ordinarily monitored
in the flue gas (02, C02, S02, CO, NO, and CH^), instrumentation was
installed for the continuous or intermittent determination of ammonia,
hydrogen, and C2-C4 hydrocarbons. An EAI Quad 1200 residual gas
analyzer (quadrupole mass analyzer) was used to determine the NH3
concentration in the flue-gas stream. The H2 and C2-Cit hydrocarbons
were analyzed by gas chromatography. Hydrogen was also analyzed by
a thermal conductivity detector. Ethylene, ethane, acetylene, propane,
iso-propane, and butane were analyzed by a flame ionization detector.
An attempt was also made to monitor N02 continuously using a
Mast Model 724-11 analyzer, preceded by sodium carbonate scrubbers to
remove potential S02 interference. (Separate laboratory experiments
indicated that a gas stream containing up to 2500 ppm S02 can be
scrubbed to less than 1 ppm using this procedure.) However, the Mast
analyzer, which is actually designed for ambient air monitoring, gave
results that appeared to be biased by other flue-gas components, such
as strong oxidizing agents. The pressure and temperature effects of
N02 flue-gas levels will, therefore, be reserved for a future
investigation.
The N02 component has generally been considered to be a relatively
minor flue-gas component (<50 ppm), although only limited data are
available. Information that has been published was obtained by British
workers, who, in combustion experiments made at 5 to 6 atm, measured
NO and N02 levels on the order of 150 ppm and 5 to 10 ppm, respectively.
65
-------
Pressure Effect on NO Levels in the Flu6 Gas. The NO data gathered
from the NOX-I and NOX-II experiments are presented in Fig. 18, which
illustrates the effect of pressure on the concentration of NO in the
flue gas. In the absence of dolomite (NOX-I series), a rapid increase
in NO concentration is noted as the combustor pressure is reduced,
going from less than 200 ppm at 8 atm to around 1600 ppm at 1 atm.
The levels of NO reported for this experiment were higher than those
previously observed during combustion of coal in Alundum beds. No
explanation is available for this result.
In the presence of dolomite (NOX-II series), an increase in NO
is also observed at reduced pressures, although it is considerably
less pronounced than in the first series of experiments. The level
of NO increased from ^200 ppm at 8 atm to ^400 ppm at 1 atm. The NO
levels observed for this series are consistent with previously
reported values.
2400-
J L
s.
4
NOX-I SERIES:
ARKWRIGHT COAL
ALUNDUM BED
1550 °F BED TEMP
3.5 ft/sec GAS VELOCITY
NOX-II SERIES:
ARKWRIGHT COAL
TYMOCHTEE DOLOMITE
Ca/S RATIO = 1.5
1550 °F BED TEMP
3.5 ft/sec GAS VELOCITY
^NOX-n SERIES
SERIES
_L
15 30 45 60 90 120
PRESSURE, psia
Figure 18. Effect of Pressure on Concentration
of NO in Flue Gas
66
-------
Pressure Effect on the Concentration of Flue-Gas Components Other
Than NO. The observed concentrations of S02, CO, H2, and CH^ in the
flue gas of the NOX-I and NOX-II series of experiments are presented
in Figs. 19 and 20, respectively. After an analysis of the data,
the following conclusions were drawn:
1. The 862 concentration is not affected greatly by changes in
system pressure. During the NOX-I experiments, the S02 level was
essentially constant at ^2300 ppm between 1 and 8 atm of pressure.
In the NOX-II experiments, the S02 level ranged from 400 to 600 ppm
over the same pressure range.
2. Methane levels in the flue gas increase as the operating
pressure is reduced below about 6 atm. In both series of experiments,
the CHi^ level was a fairly constant 25 ppm at 6 and 8 atm. The level
of CH.4. then increased rapidly with decreasing pressure to around
400 ppm at 1 -atm.
3. In the presence of dolomite, CO levels in the flue gas
increased from 170 ppm at 8 atm to over 2000 ppm at 1 atm. In the
absence of dolomite, the CO level passed through a minimum (/x/lOO ppm)
at around 4 atm pressure, and increased rapidly with either increasing
or decreasing pressures. This phenomenon remains unexplained.
4. The level of H2 is relatively unaffected by changes in
pressure from 8 atm to 2 atm, remaining at about 35 ppm in both
series of experiments. In the NOX-II experiments with dolomite,
the level of H2 increased to 330 ppm at 1 atm.
5. The levels of the other gases were below the detection limits
of the instrument used for their measurement. Ammonia was found to be
<100 ppm for both series of experiments. The flue-gas ethylene, ethane,
acetylene, propane, and iso-propane concentrations were <10 ppm, and
the butane, <20 ppm.
Effect of Temperature on the Concentration of Flue-Gas Constituents.
A series of three experiments was. also completed to determine the effect
of system temperature on concentrations of flue-gas constituents. The
data from these experiments are presented in Fig. 21. Under the
conditions at which the experiments were performed, temperature
variations from 1450°F to 1650°F had little or no effect on the
composition of the flue-gas components investigated.
67
-------
SO 2
NOX-I SERIES:
ARKWRIGHT COAL
ALUNDUM BED
1550 °F BED TEMP
3.5 ft/sec GAS VELOCITY
CO
0 15 30 45 60 90 120
PRESSURE, psia
Figure 19. Effect of Pressure on Concentration
of Flue-Gas Components Other Than NO for Series NOX-I
68
-------
2000
E
^ 1800
c/>
ul 1600
O
| 1400
O
O
en 1200
cp
UJ
u.
O
O
55
cr.
UJ
O
O
O
1000
900
800
700
600
500
400
300
200
100
0
NOX-E SERIES:
ARKWRIGHT COAL
TYMOCHTEE DOLOMITE
Co/S RATIO = 1.5
1550 °F BED TEMP
3.5 ft/sec GAS VELOCITY
t
g
T
0 15 30 45 60 90 120
PRESSURE, psia
Figure 20. Effect of Pressure on Concentration of
Flue-Gas Components Other Than NO for Series NOX-II
ARKWRIGHT COAL
TYMOCHTEE DOLOMITE
Ca/S RATIO = 1.5
3.5 ft/sec GAS VELOCITY
E
a.
CL
CO
h-
—y
^_
U. lit
0^
zg
2s
HO
£<"
^•<
5
s H
^ s
i T
i
-
____^^_^. ^^^_
- :^nr H^F
1 — 4 !s
0 1450 1550
TEMPERATURE, °F
CO
rO
til
X
O
i
Cf)o
,NO
— ,co
-^^CH4
i i
1650
Figure 21. Effect of Temperature on Concentration
of Flue-Gas Components
69
-------
TRACE-ELEMENT DISTRIBUTION STUDIES
A study is being made of the distribution and potential emissions
of biologically toxic trace elements during the fluidized-bed combustion
of coal. Interest in the emissions of such elements as Hg, Be, Pb, and
F has increased considerably in recent years. Although these elements
are present at very low levels in fossil fuels, the large consumption of
these fuels in the United States represents an annual potential emission
release of up to several thousand tons for some of the elements.
Since most of the trace elements in coal are likely to be found in
the sorbent used for S02 removal, there is some interest in comparing
the behavior of the trace elements in a fluidized-bed combustor with con-
ventional, coal-fired power stations. Relative to conventional boilers,
the lower combustion temperatures of the fluidized-bed boiler and the
presence of additive for SC>2 retention may serve to retain potentially
volatile trace elements in particulate form. The elements of interest
are listed in Table 16 in order of their priority as pollution agents as
determined by the EPA.
Four combustion experiments have been completed to study the distri-
bution of trace elements in the combustion system. The approach being
taken is to make mass balances around the combustion system for as many
of the trace elements as possible within the economic limitations of the
study. This involves sampling and analyzing all the solid materials
charged to or recovered from the combustion system, including particu-
late matter entrained in the flue gas. The flue gas is also analyzed
for the more volatile trace-element species, Hg and F~.
Table 16. Trace Elements of Interest and Their Order of Priority
for Experimental Investigation
First Priority Third Priority
Beryllium Copper
Mercury Zinc
Fluorine Barium
Lead Tin
Phosphorus
Second Priority Lithium
Vanadium
Cadmium Manganese
Arsenic Chromium
Nickel Selenium
70
-------
Solids Sampling
The mass balances are made over the duration of the combustion
experiments and not just over a specific interval of time. Inputs
to the mass balance, therefore, consisted of the initial bed charge
(either Alundum or partially sulfated additive from a previous
combustion experiment), the coal fed to the combustor and, for
experiments with a bed of additive, the sorbent fed to the combustor.
Outputs from the system consisted of the final bed, overflow material
removed through the standpipe, and entrained solids from the combustor
recovered in the primary cyclone, secondary cyclone, primary filter,
and secondary filter.
With the exception of the coal and additive feeds, samples
submitted for analysis were obtained by splitting or core sampling
the bulk materials. Because of the large volume of coal and additive
used during a single experiment, samples of these materials consisted
of randomly taken grab samples, which were combined into a single
sample for analysis. All samples were ground to a powder using a
mortar and pestle to promote homogeneity before further dividing the
samples for the individual analyses.
Flue-Gas Sampling
Whereas the solids were sampled prior to and at the conclusion
of the experiment, the flue gas was sampled over an interval of time
during the course of the experiment. The flue gas was sampled after
it had passed through the principal gas-solid separators, which
remove nearly all of the particulate matter above 5 pm in diameter,
and through the pressure let-down valve. This had the advantage of
allowing the sampling to be accomplished at atmospheric pressure,
even though the combustor and the ancillary equipment were operating
at pressures of up to 10 atm.
The modifications made in the flue-gas line to accommodate the
sampling process are indicated in Fig. 22. The sampling zone is a
5-ft length of 4-in.-dia, Type L copper tube with a nominal inside
diameter of 3.905 in. The sample probe is located approximately
12-pipe diameters downstream and 3-pipe diameters upstream from any
flow disturbances.
Sampling of the flue gas for particulate matter and gas phase
Hg and F~ was made using the apparatus illustrated in Fig. 23. The
sample probe was constructed from 1/4-in., Type 304 stainless steel
tubing. A tapered nozzle with a 5/32-in. opening was affixed to the
end of the probe to provide for streamlined sampling. The probe could
be inserted to any depth along the diameter of the flue to allow for
sampling at several traverse points. However, because of the relatively
small diameter of the flue, the highly turbulent gas flow, and the
location of the probe away from flow disturbances, sampling was done
only at the centerline of the flue. Thus the assumption made in
sampling was that the concentrations of interest were not a significant
function of radial position.
71
-------
CONTROL VALVE
FOR MAINTAINING
SYSTEM PRESSURE
Figure 22. Experimental Flow System for Flue-Gas Sampling
CLASS FIBER
THERMOCOUPLE
THERMOCOUPLE
SAMPLING
ZONE
IELL-ITPE if]
NANOMETER
ORT-CAS
METER
PUMP
Figure 23. Flue-Gas Sampling Apparatus for Measuring Concentrations
of Hg, F~, and Entrained Particulate Matter
72
-------
The particulate matter in the gas sample was removed in the Brink
Cascade Impactor (in-line cyclone + 5 stages) and the follow-up, glass-
fiber filter (Gelman Type A). The impactor removed and separated the
particulate matter into five size ranges, which covered the range
from 'vO.S to 3.0 pm in diameter. The sample probe, impactor, and
filter housing were heated to over 250°F to prevent condensation of
the Hg and F~ on the metal surfaces.
The sample of flue gas, after removal of the particulate matter,
passed through a series of scrubbers in an ice bath to remove the
volatile F~ and Hg compounds. The first scrubber contained ^400 ml
of either a 1 wt % NaHC03 or saturated Na2CC>3 solution to remove the
F~. The second and third scrubbers each contained V100 ml of a 0.1 M
iodine monochloride (IC1) solution for removing the gas-phase Hg. In
two of the four combustion experiments, the first IC1 scrubber was
replaced by a straight section of glass tubing containing two frits
of compactly wound fine gold wire. Because gold is an excellent
amalgamator of Hg, it was anticipated that the Hg in the flue-gas
sample passing through the first scrubber would be trapped by the
gold. The second IC1 scrubber was retained to serve as a back-up
for the gold frits.
The volumetric flow rate of the gas sample through the impactor
and scrubbing solutions ranged from 0.1 to 0.15 cfm.
Analytical Methods
Mercury and lead are being determined by atomic absorption
spectroscopy, fluoride by specific ion electrode, and beryllium by
fluorimetry. Neutron activation analysis methods are also being
developed as an instrumental method for analyzing as many of the
second and third priority elements as possible. Detailed analytical
procedures currently being used for the various trace elements and
types of samples (solutions, coal, ash, and additive) are presented
in detail in Appendix E,
Results
As indicated above, four combustion experiments were made for
the specific purpose of investigating the behavior of trace elements
during combustion. A summary of the average operating conditions and
flue-gas compositions for the trace-element experiments are presented
in Table 17. Nominal conditions for experiments TRACE-3 and TRACE-4B
were a bed temperature of 1550°F, 10 atm pressure, and 4% 02 in the
off-gas as compared with values of 1650°F, 8 atm, and 3% 02 for
experiments TRACE-5A and TRACE-6. At each set of conditions, one
experiment involved the combustion of Arkwright coal in a fluidized
bed of Alundum, whereas the other experiment involved the combustion
of Arkwright coal in a fluidized bed of Tymochtee dolomite.
73
-------
Table 17. Summary of Average Operating Conditions and Flue-Gas
Compositions for Trace-Element Experiments3
Expt.
Run
Time,
hr
Bed
Temp,
°F
(Experiments
TRACE-5A
TRACE- 3
6.0
4.0
1670
1560
(Experiments
TRACE-6
TRACE-4B
7.5
4.75
1660
1550
Sys
Feed Rates
teui
Pressure, Coal, Dolomite, Air,
atm Ib/hr Ib/hr
in a
8
10
in a
8
10
scfm
Ca/S
Ratio
Gas
Velocity,
ft/sec
Flue-Gas Analysis:
Dry Basis
02>
%
C02,
%
S02>
ppm
NO, CO,
ppm ppm
Fluidized Bed of Alumina)
29.4 0.0
24.9 0.0
67.1
63.0
0.0
0.0
3.3
2.4
3.0
3.9
16.0
16.1
2180
1660b
110 70
180 120
Fluidized Bed of Tymochtee Dolomite)
29.7 6.1
28.2 14.2
67.9
75.5
1.2
2.9
3.4
3.0
2.9
3.8
16.5
19.0
440
140
150 39
200 35
All experiments made with as-received Arkwright coal.
Low value due to malfunction of S0£ IR analyzer.
-------
Mercury Mass Balances. Mercury balances for the four experiments
are presented in Table 18. Experiments TRACE-3 and TRACE-4B, which
were completed first (with two IC1 scrubbers in the gas-sampling
apparatus), exhibited Hg recoveries, expressed as the percentage of
Hg entering the combustor which can be accounted for in the combustion
products, of 56 and 43%, respectively. For experiments TRACE-5A and
TRACE-6 (with two gold frits and one Id scrubber in the gas sampling
apparatus) the respective recoveries were only 29 and 25%. The lower
total Hg recovery values for the latter experiments apparently resulted
from the decreased recovery of Hg in the volatile form, which was
lower by a factor of 6 as compared with the earlier experiments (3 and
5% VS. 19 and 34%). The amount of Hg recovered in the solid-phase
combustion products ranged from 9 to 37% with an average recovery of 23%
for the four experiments.
Measured concentrations of Hg in the flue gas for experiments
TRACE-5A and TRACE-6 (made at 1650°F) were only 0.03 and 0.05 ppb
as compared with 0.32 and 0.66 ppb for experiments TRACE-3 and
TRACE-4B (made at 1550°F). It would seem more logical to expect an
increase in the level of Hg in the flue gas at the higher combustion
temperatures. The lower measured concentrations did not, however,
appear to result from the modifications in the flue-gas sampling
equipment discussed above. In the experiments with the modified
apparatus, traces of Hg were found in the carbonate solution
(considerably less than in the first two experiments, however), on
the gold frits, and on the filter following the IC1 scrubbing solution.
As with the first two experiments, no Hg was detected in the IC1
solution.
Several attempts have been made to determine the cause for the
poor recovery of Hg in the combustion products. The particular
question areas that have been or are being examined include the
following:
(1) The inability to detect mercury in the IC1 scrubbing
solutions used in sampling the flue gas.
(2) The recovery efficiency of the mercury sampling
equipment.
(3) The possibility that mercury is plating out or
condensing on the walls of the flue-gas ductwork
before reaching the sampling zone.
Efforts that have been expended to define the problems are discussed
below.
75
-------
Table 18. Mercury Material Balances for Trace-Element Combustion Experiments
Component
IN:
OUT:
Coal
Additive
Initial Bed
TOTAL Hg In
Final Bed
Overflow
Prim. Cyclone
Sec . Cyclone
Prim. Filter
Sec. Filter
Flue Gas
TOTAL Hg Out
Combustion in
TRACE-3
(1550°F)a
Cone . , Wt ,
ppm ug
Alundum Bed
TRACE-5A
(1650°F)
Cone. ,
ppm
0.15 5,900 0.07
No Additive Used
<0.005 Neg.C 0.028
5,900
<0.005 Neg. <0.005
0.46 2,100 0.15
0.46 130
No Sample
No Sample
0.32 ppb 1,100
3,300
0.39
0.45
0.52
0.03 ppb
Wt,
U8
5,680
640
6,300
Neg.
1,280
130
150
10
180
1,800
Combustion in
TRACE-4B
(1550°F)
Cone.
ppm
0.15
0.017
<0.01
<0.01
<0.01
0.035
0.28
Wt,
vg
9,100
510
Neg.
9,600
Neg.
Neg.
520
170
0.15 90
No Sample
0.66 ppb 3,300
4,100
Dolomite Bed
TRACE- 6
(1650°F)
Cone.
ppm
0.07
0.025
0.005
<0.005
<0.005
0.06
0.35
Wt,
vg
7,050
530
70
7,600
Neg.
Neg.
990
420
0.28 110
No Sample —
0.05 ppb 390
1,900
Recovery, %:
TOTAL
In Flue Gas
In Solid Form
56
19
37
29
3
26
43
34
9
25
5
20
a
Bed temperature
Estimated.
°Considered negligible
d.
Recovery = (Hg out/Hg in) x 100%.
-------
Iodine Monochloride Scrubbing Solution. Sulfur dioxide has been
reported as interfering with the detection of Hg in gas streams when
using IC1 or K^MnO^ scrubbing solutions.8 To test this, S02 was
bubbled through an Id solution containing a known concentration of
Hg. Samples were taken of the IC1 solution at various times and
analyzed for Hg. The Hg concentration remained constant, thereby
indicating no interference from the SC^. However, excessive exposure
to SC>2 can cause I2 or MnC>2 precipitation, which then makes analysis
of the solutions for Hg impossible. This, however, explains the
inability to detect Hg in the IC1 scrubber in only the TRACE-3
experiment.
In subsequent tests, it was observed that if the IC1 solution is
first exposed to flue gas, then trace quantities of Hg subsequently added
to the solution cannot be detected. Recovery is generally less than 50%.
Trace amounts of Hg added to ICl solutions unexposed to flue gas, however,
exhibit 100% recovery of the Hg by analysis. This strongly suggests that
components in the flue gas other than S02 may be interfering with the Hg
analysis in the ICl solutions.
Condensation of Mercury on Flue-Gas Ductwork. Several tests
have been made and additional tests are in progress to determine the
possibility that Hg condenses on the walls of the ductwork before it
reaches the sampling zone. Copper and stainless-steel tabs were
suspended in the flue gas just upstream from the sampling probe during
a typical combustion experiment. When analyzed for mercury at the end
of the experiment, the concentrations of Hg on the copper and stainless
steel tabs, respectively, were 0.7 and 0.2 ppm. Copper and stainless
steel tabs which had not been suspended in the flue gas exhibited Hg
concentrations of ^2.5 and 0.2 ppm, respectively, indicating that the
Hg on the test pieces was not Hg which had condensed out of the flue
gas.
The fiber filters used in cleaning the flue gas of particulate
matter were also examined. Indications were that any mercury on the
filters was associated with the particulate matter trapped on the
filters.
Scrapings were also taken from the walls of the flue-gas piping
and analyzed to determine whether abnormally high levels of Hg were
present there. The samples did have Hg concentrations approximately
an order of magnitude higher than the combustion products recovered
in the cyclones and filters. It is doubtful, however, that enough
Hg accumulates in the piping to explain the poor recoveries of Hg
reported for the TRACE experiments.
Efficiency of Mercury Sampling Apparatus. Plans are to inject
a known amount of Hg into the flue-gas sampling line during a future
combustion experiment at a point just upstream of the sampling apparatus
and determine the mercury material balance. This should help to
determine whether the problems are with the sampling of the flue gas or
whether the Hg is condensing out in the system before reaching the
sampling location.
77
-------
Retention of Mercury in Solid Products of Combustion. The recovery
of mercury in the solid combustion products is in keeping with results
reported by other investigators. Billings et al, made a mercury
balance around a 660 MWe coal-fired power station using a coal con-
taining 21% ash (<1% S) and having a heating value of ^9000 BTU/lb.9
They reported that ^10% of the mercury remained with the furnace
residual ash, and 90% was emitted in the vapor phase. The average
concentrations of Hg in the coal and ash products were 0.3 ppm and
0.2 ppm, respectively.
It is interesting to note that the fly-ash samples in the study
by Billings et al. were less concentrated in Hg than the coal being
combusted, whereas in the TRACE experiments the elutriated solids were
consistently higher in Hg concentration than the coal (except for pri-
mary cyclone samples highly diluted with sulfated dolomite). This is
reflected in the higher retention of Hg by the ash products (37 and
26%) reported for the TRACE experiments made with an Alundum bed. This
is an indication, at least, that fluidized-bed combustion may offer the
potential for increased retention of mercury by the solid products of
combustion.
Diehl et al. have also studied the fate of Hg during the
combustion of coal in combustion units of various sizes.10 In a
100-g/hr test unit, ^60% of the mercury was retained in the fly ash
(combustion at 2100°F) when burning a coal containing 0.15 ppm Hg
(^0.9 ppm in the ash). Approximately, 35% of the mercury was retained
in the fly ash when burning a coal containing 0.24 ppm Hg (M3.35 ppm
in the ash). In a 500-lb/hr combustor burning pulverized coal
containing 0.18 ppm Hg, ash removed in cyclones at ^360°? contained
0.22 ppm Hg, which accounted for vL2% of the Hg in the coal. They
also sampled fly ash from 250 MWe and 175 MWe power plants, and
reported mercury contents in the ash of 0.10 and 0.26 ppm, respectively,
for estimated Hg retentions of 7 and 19% in the fly ash.
Lead and Beryllium Mass Balances. Lead balances for the four
trace-element experiments are presented in Table 19. The material
balances range from 78 to 125% recovery of the lead entering the
combustor. Beryllium balances for the trace-element experiments are
presented in Table 20. Recoveries of Be were somewhat lower than
for the Pb, ranging from 56 to 87%.
The results strongly indicate that the lead, with the observed
high recoveries, is essentially retained by the particulate matter
in the combustion products. Emissions of Pb, therefore, will be
controlled by the efficiency of particulate removal from the flue
gas. However, because the concentration of Pb in particulate matter
increases with decreasing particle size, the efficiency of lead
removal will be somewhat lower than the overall efficiency of parti-
culate removal.
The beryllium mass balances, with the lower recoveries, suggest
the possible volatilization of beryllium (or compounds of beryllium)
during combustion. This seems an unlikely possibility, however, for
the bed temperatures used in these combustion experiments.
78
-------
Table 19. Lead Material Balances for Trace-Element Combustion Experiments
Combustion in Alundum Bed
TRACE- 3
(1550°F)a
Component
IN: Coal
Additive
Initial Bed
TOTAL Pb In
OUT: Final Bed
Overflow
Prim. Cyclone
Sec. Cyclone
Prim. Filter
Sec. Filter
TOTAL Pb Out
Recovery, %: TOTAL
Cone.
ppm
29
1
51
95
260
- No
- No
Wt,
mg
TRACE- 5A
(1650°F)
Cone. ,
ppm
1,100 1.6
.1 14 1.1
1,100
670 2.8
No Overflow
440 15
70 13
Sample - 22
Sample - 46
1,200
109
wt,
mg
130
25
160
60
130
4.3
7.4
0.6
200
125
Combustion in Dolomite Bed
TRA.CE-4B
(1550°F)
Cone . ,
ppm
29 1
6
21
2
16
15
70 1
180
300
1
wt,
mg
,800
180
290
,300
260
220
,000
110
170
,800
78
TRACE- 6
(1650°F;
Cone. ,
ppm
1.6
12
11
11
12
16
27
96
)
wt,
mg
160
310
150
620
150
100
270
33
38
590
95
Bed temperature.
-------
Table 20. Beryllium Material Balances for Trace-Element Combustion Experiments
co
o
Combustion
TRACE- 3
(1550°F)a
in Alundum Bed
TRACE- 5 A
(1650°F)
Cone . , Wt , Cone . ,
Component
IN: Coal
Additive
Initial Bed
TOTAL Be In
OUT: Final Bed
Overflow
Prim. Cyclone
Sec. Cyclone
Prim. Filter
Sec. Filter
TOTAL Be Out
Recovery, %: TOTAL
ppm mg
0.7 28
ppm
0.66
Wt,
mg
54
iiw flX
0.83 10
38
0.76 10
NT
JNO
2.65 12
5.95 2
- No Sample
- No Sample
24
63
IU-LU-LVG UOCU
0.83
0.79
ru s* f^
2.29
6.62
6.75
8.05
19
73
17
19
2.2
2.3
0.1
41
56
Combustion in
TRACE-4B
(1550°F)
Cone . ,
ppm
0.7
0.67
0.78
0.75
0.73
1.55
5.20
6.77
Wt,
mg
42
20
11
73
12
10
23
3
4
— \Tr»
52
71
Dolomite Bed
TRACE-6
(1650°F)
Cone . ,
ppm
0.66
0.75
0.66
0.80
2.44
2.24
5.63
7.70
O 1 *i
Wt,
mg
66
16
8.9
91
11
21
37
6.8
3.0
79
87
Bed temperature.
-------
Preferential Concentration of Lead and Beryllium in the Finer
Particulate Matter. Aside from material-balance considerations, the
results indicate that both lead and beryllium concentrate preferentially
in the finer particles leaving the combustor. Figure 24 shows Coulter
Counter particle-size analysis curves for the primary and secondary
cyclone materials as recovered from the TRACE-3 experiment. The
combustion gases leaving the combustor pass through a primary cyclone,
a secondary cyclone, and a porous filter, in that order. A 50% cutoff
value of approximately 10 urn was obtained from Fig. 24 for the "mean"
diameter of the primary cyclone material. A "mean" value of 1.5 um
was found for the diameter of the secondary cyclone material. For the
coarser primary-cyclone ash product, the lead and beryllium concen-
trations were 95 and 2.65 ppm, respectively. In the finer, secondary-
cyclone ash, the lead and beryllium concentrations increased to 255
and 5.94 ppm, respectively.
Although particle-size data are not available for the primary and
secondary cyclone and filter solids from the other trace-element experi-
ments, we can infer from the concentration data for Pb and Be in Tables
19 and 20 a definite trend of increasing trace-element concentration in
the finer fractions of the material recovered from the flue gas.
o PRIMARY CYCLONE
x SECONDARY CYCLONE
0
O.I 0.2
0.5 I 2 5 10 20 50 100
DIAMETER ,/y.m
Figure 24. Particle-Size Distributions for Fly Ash
Recovered by the Primary and Secondary Cyclones
in Experiment TRACE-3
81
-------
The preferential concentration of relatively nonvolatile trace
elements in particles of smaller diameter has been demonstrated in
samples taken from ambient air11'12 and more recently in fly-ash
particles retained in the precipitation systems and in the airborne
fly ash leaving eight different coal-fired power plants in the United
States.1^ it has been proposed that preferential concentration occurs
by volatilization of the elements (or one of their compounds) during
combustion, followed by condensation or adsorption onto the larger
surface area, per unit mass, of the smaller particles.13 It was noted
that the normal temperature in the combustion zones of the conventional
plants tested was between 1300° and 1600°C, a temperature similar to
or above the boiling points of the elements investigated (Cd, As, Ni,
Pb, Cr, and Zn).
Although our results appear to confirm the preferential concen-
tration of Pb and Be in smaller particles, there is little evidence
to support the idea of volatilization and subsequent condensation or
absorption as the mechanism by which this concentration occurs. In
a report by the Illinois Geological Survey, it was shown that there
was no significant loss of beryllium in coal ash prepared at 700°C.1I+
As the TRACE-3 and TRACE-4B experiments were carried out at 10-atm
pressure and a bed temperature of only 850°C, it does not appear
likely that volatilization would be a significant factor in the
apparent concentration of the beryllium in the finer particles. A
somewhat simpler explanation may be that the trace elements in the
larger particles are simply diluted by unburned carbonaceous materials.
This is evidenced in Table 21, which tabulates the concentrations of
trace, minor, and major elements in particulate matter recovered
during the TRACE experiments at various stages of removal from the
flue gas. As experiments TRACE-3 and TRACE-5A were made using an
Alundum bed, the changes in concentration with subsequent stages of
solids recovery from the flue gas are not biased by the presence of
large amounts of additive. In the TRACE-3 experiment, the level of
carbon decreases from 51.3 to 29.2% as the concentrations of the
trace and minor elements in the coal increase in the ash. In
experiment TRACE-5A, the concentration of carbon decreases from 34.1
to 6.6% as the levels of the other elements show concentration
increases.
An interesting observation can also be made by comparing the
ratios of lead to beryllium in the coal, additive, and in each of
the materials removed at various stages of the gas cleanup system.
These ratios are presented in Table 22.
The columns of particular interest are those under experiments
TRACE-3 and TRACE-5A, in which additive does not appear as a factor
in the calculated ratios. While there does appear to be some
variation in the ratio of Pb to Be, these variations could well fall
within the accuracies of the analytical methods. If volatilization
and condensation are the mechanisms by which preferential concen-
tration occurs, it seems likely that the Pb-to-Be ratio would change
significantly over the range of samples analyzed.
82
-------
Table 21. Concentrations of Trace, Minor, and Major Elements
in Particulate Matter Recovered at Various Stages
of Removal from Flue Gas
Concentration
Pb,
Source of Solids ppm
Primary Cyclone 95
Secondary Cyclone 260
Primary Cyclone 70
Secondary Cyclone 180
Filter 300
Primary Cyclone 15
Secondary Cyclone 13
Primary Filter 22
Secondary Filter 46
Primary Cyclone 16
Secondary Cyclone 27
Primary Filter 96
Be, C, S,
ppm wt % wt %
(Experiment TRACE-3)
2.65 51.3 0.86
5.95 29.2 1.80
(Experiment TRACE-4B)
1.55 2.3 4.5
5.20 13.1 2.6
6.77 8.4 3.1
(Experiment TRACE- 5A)
2.29 34.1 1.8
6.62 16.8 2.5
6.75 6.6 4.9
8.05
(Experiment TRACE-6)
2.24 19.6 4.8
5.63 11.3 3.1
7.70 5.2 4.3
Ca,
wt %
1.29
1.80
15.4
5.58
6.65
2.11
2.02
2.89
8.11
4.89
5.44
Table 22. Ratio of Lead to Beryllium in the Raw Materials and in the
Particulate Matter Removed from the Flue Gas during the
"TRACE" Experiments
Experiment
Material TRACE-3
Coal 41
Additive not used
Primary Cyclone 36
Secondary
Cyclone 43
Primary Filter no sample
Secondary
Filter no sample
TRACE-4B TRACE- 5A
41 2.4
9 not used
45 6.5
34 2.0
44 3.3
no sample 5.7
TRACE-6
2.4
16
7.1
4.8
12
no sample
83
-------
Fluoride Mass Balances. Fluoride material balances for the trace-
element experiments are given in Table 23. The indicated recovery of
fluoride for the experiments at 1550°F are 123 and 110%, which are
reasonably acceptable values. The recoveries of 208 and 276% reported
for the experiments at 1650°F, however, are unaccountably high. The
only differences in sampling between the two sets of experiments were
the use of considerably larger flue-gas samples and the use of Na2C03
scrubbers instead of NaHC03 scrubbers for the experiments at 1650°F.
Perhaps the most significant observation that can be made from
the F~ balances is that the retention of the F~ in the solid phases
leaving the combustor appeared to be significantly higher when additive
was used in the experiment. The reported recoveries of F~ in the solid
samples were 56 and 62% for combustion with additive present and only
23 and 5% for combustion in an Alundum bed. A possible explanation
for such a phenomenon could be the formation of CaF, which is a
relatively stable compound.
Neutron Activation Analysis, Developmental Work. Neutron
activation analysis is being considered as a possible instrumental
method for expanding the trace-element study to include elements of
second- and third-priority interest (Cd, As, Ni, Zn, Cu, Ba, Sn, P,
Li, Mn, Cr, Se, and V). A preliminary testing of the method was made
by taking 100-mg samples of coal, TRACE-3 primary-cyclone ash, and
TRACE-3 secondary-cyclone ash and irradiating them in the ANL CP-5
test reactor for 2.5 hr. Flux monitors irradiated with the samples
revealed that neutron exposure of the samples was considerably
lower than had been anticipated. Subsequent inquiries indicated that
an error may have occurred in positioning the samples in the reactor;
this would account for the low measured neutron flux.
After the irradiation, periodic Y~raY counts were taken using
a Ge(Li) detector on all three samples for the purpose of identification
of activation products and relative activity levels. Results for
seven of the elements detected are presented in Table 24. Additional
elements were detected, but the data on these elements were lost when
the magnetic tape on which the data were stored was accidentally
erased. With the exception of the material balance for iron, the
remaining four elements balance as well as, or better than, the values
obtained for mercury, lead, beryllium, and fluoride by direct chemical
analysis.
Encouraged by the preliminary results, the remaining samples from
the trace-element combustion experiments are being analyzed (including
a reanalysis of the TRACE-3 samples) by this technique. Additional
developmental work (see Appendix D) is also being done to gain the
capability of detecting a larger number of elements and to improve
the accuracy of the method.
84
-------
Table 23. Fluoride Material Balances for Trace-Element Combustion Experiments
00
Ui
Combustion in
TRACE-3
(1550°F)a
Cone.
Component ppm
IN:
OUT:
Coal 25
AQQH.LJ.ve
Initial Bed 100
Total F~ In
Final Bed - Not
Overflow
Prim. Cyclone 20
Sec. Cyclone 10
Prim. Filter - No
Sec. Filter - No
Flue Gas
Total F~ Out
Recovery, %: TOTAL
In Flue Gas
In Solids
Wt,
g
1.0
M/-> i^i^-i
1NO /\QQ1
1.2
2.2
Detected -
— Ttfrt
0.1
Neg.
Sample -
Sample -
2.6
2.7
123
118
5
Alundum
Bed
TRACE- 5 A
(1650°F)
Cone.
ppm
29
tive Use
100
36
Overflow
36
12
5
Wt,
g
2.4
2.3
4.7
0.8
0.3
Neg.
Neg.
Insufficient
Sample
7.4
8.5
180
157
23
Combustion in Dolomite
TRACE-4B
(1550°F)
Cone.
ppm
25
14
350
38
86
150
47
115
- No
Wt,
g
1.5
0.4
4.9
6.8
0.6
1.2
2.3
Neg.
Neg.
Sample -
3.3
7.4
110
48
62
Bed
TRACE-6
(1650°F)
Cone . ,
ppm
29
14
130
86
41
71
37
29
- No
Wt,
g
2.9
0.3
1.8
5.0
1.2
0.4
1.2
Neg.
Neg.
Sample -
9.2
12.0
240
184
56
Bed temperature.
-------
Table 24. Neutron Activation Results on Samples from
the TRACE-3 Trace-Element Experiment
Concentration, ppm
Element
?e
Cr
Sc
Na
K
Zn
Cu
Coal
11,300
—
2.4
960
460
710
33
Primary
Cyclone
33,000
250
18
11,000
2,800
1,600
330
Secondary Material Balance,
Cyclone % Recovery
39,000
330
23
10,000
2,800
—
410
37
—
96
143
76
—
127
Sodium Concentrations in Flue-Gas Particulates. Samples from
several of the TRACE experiments have been analyzed for Na because of
the considerable interest that has been expressed in Na as a source
of corrosion in a combined-cycle power system, particularly corrosion
of the blades of a high-temperature, gas turbine. The results of the
analyses that have been completed thus far are presented in Table 25.
It would appear from the material balances that the Na is retained by
the particulate matter during combustion. As with Pb and Be, the
concentration of Na in the particulate matter increases with decreasing
particle size. Concentrations generally vary from 0.5 wt % for
material removed in the primary cyclone to 1.5-2.0 wt % for material
removed in the primary and secondary filters. It should be emphasized
however, that the flue gas is cooled to approximately 600-800°F during
the solids removal process. Because the flue gas entering the turbine
will be at considerably higher temperatures (VL600°F), considerably
less Na may be retained by the particulates during their removal at
these higher temperatures.
Photomicrographs of Flue--Gas Particulates. Solid samples of
Arkwright coal and fly ash recovered from the cyclones and Brink
impactor at the conclusion of experiment TRACE-3 were examined under
a scanning election microscope. Photomicrographs obtained for several
of the samples are pictured in Fig. 25. Unlike the generally
spherically shaped fly ash emitted from conventional coal-fired
combustors, the photomicrographs depict the fly ash from the fluidized-
bed combustor as a rather fragile and flake-like material.
86
-------
Table 25. Sodium Material Balances for Trace-Element Combustion Experiments
oo
Component
IN : Coal
Additive
Initial Bed
Total Na In
OUT: Final Bed
Overflow
Prim. Cyclone
Sec. Cyclone
Prim. Filter
Sec. Filter
Total Na Out
Recovery, % TOTAL
Combustion in
TRACE-3
Cone., a Wt,
wt % g
0.10° 40
— M/-* A A A -i fr- -i
0.11 14
54
Alundum Bed Combustion in Dolomite
TRACE- 5 A
Cone. ,
wt %
0.05b
ve Used -
0.11
0.12 16 0.06
— No Overflow —
1.1° 51 0.47
1.0C 2.9 0.88
No Sample 1.97
No Sample 1.65
70
130
Wt,
g
41
25
66
13
40
2.7
6.6
0.2
62
94
TRACE- 6
Cone . ,
wt %
0.05
0.04
0.06
0.07
0.06
0.35
0.95
1.51
No
Wt,
g
50
8.4
8.1
66
10
5.2
57
11
6
Sample
89
135
Determined by atomic absorption spectroscopy.
Commercial Testing Company, Chicago, Illinois reported a value of 0.07 wt
in Arkwright coal.
•»
"Determined by neutron activation analysis.
% Na
-------
a. Arkwright Coal
b. Primary Cyclone Fly Ash
Secondary Cyclone Fly Ash
e.
Brink Impactor
Stage 3 Fly Ash
d. Brink Impactor
Stage 1 Fly Ash
f. Brink Impactor
Stage 5 Fly Ash
Figure 25.
Photomicrographs of Arkwright Coal and Fly Ash Recovered
in Particulate Removal and Sampling Devices during
Combustion Experiment TRACE-3
-------
KINETICS OF THE REACTION OF HALF-CALCINED
DOLOMITE WITH SULFUR DIOXIDE
The kinetics of the reaction of half-calcined dolomite with
sulfur dioxide is being studied to gain further insight into the
sulfation process. Preliminary results of the study which is being
carried out by a small basic chemistry support group, are presented
below. Besides their application to the current f luidized-bed
studies, it is hoped that these studies may also be of value in other
processes for S02 control, e.g. panel-bed filters.
The active material used in this program is half-calcined
dolomite, which reacts with S02 as indicated in Eq. (8)
[CaC03 + MgO] + S02 + 0.5 02 -»• [CaS04 + MgO] + C02 (8)
For economic and environmental reasons, it is desirable to regenerate
the reactive material from the product by some scheme such as that
indicated by Eqs. (9) and (10)
+ MgO] + 4 H2 + [CaS + MgO] + 4 H20 (9a)
+ MgO] + 4 CO •* [CaS + MgO] + 4 C02 (9b)
[CaS + MgO] + H20 + C02 -»• [CaC03 + MgO] + H2S (10)
where the concentration of H2S resulting from Eq. (10) is sufficiently
high to permit sulfur recovery in a Glaus plant.
A brief mention of some earlier work by others on this and other
related systems seems appropriate. A considerable amount of work,
including detailed kinetic studies,15'16 has been reported on the
reaction of calcined limestones with S02 according to Eq. (11)
CaO + S02 + 0.5 02 ->• CaS04 (11)
These studies were done in connection with both the f luidized-bed
combustion and dry limestone injection processes. The reactions of
fully- and half-calcined limestones with H2S [see Eqs. (12) and (13)]
[CaO + MgO] + H2S + [CaS + MgO] + H20 (12)
[CaC03 + MgO] + H2S ->• [CaS + MgO] + H20 + C02 (13)
have been studied in some detail in connection with other desulfurization
schemes. 16~20 Some results have also been reported on the reactions
of interest in work discussed here.21"23 Exploratory experiments on
the application of the system considered here to panel-bed filtration
have indicated its feasibility.23 Conclusions that can be drawn from
this prior work include:
89
-------
(1) Dolomitic limestones (fully- or half-calcined) are more
effective reagents than calcite and show considerable
promise in sulfur emission control.
(2) The rate of the reduction reaction of the sulfation
product to the sulfide [reaction (9)] appears to be
satisfactory.
(3) The active reagent can be more readily regenerated
[by reaction (10)] from sulfide produced by direct
sulfidation with H2S than from sulfide formed by
reduction of the sulfated product resulting from
reaction with S02-
In the light of this prior work, the goals of our program are
(a) determine the detailed kinetics of reactions (8), (9), and (10);
(b) elucidate plausible mechanisms for these reactions; and (c) de-
termine those conditions that optimize each of these reactions. Ki-
netic results for Eq. (8) at atmospheric pressure are presented here.
Experimental
Apparatus. The experimental apparatus is similar to that used
by other workers^-7>19>22 and is schematically depicted in Fig. 26.
The reactant gas mixture, which is prepared by controlling the flow
of each constituent by means of a diaphragm-type regulator and
calibrated rotameters, flows upward through the heated reaction tube,
past the sample, and exits through a condenser and a series of
scrubbers. Total flow can be controlled in the range from 200 to 400
cc/min with an accuracy for the total flow and that of each constituent
of about +2%. The water content of the reactant gas is controlled by
a thermostated humidifier. Sulfur dioxide is added to the stream
after humidification. The sample is suspended from one arm of a
recording balance and is contained in a platinum basket. The balance,
which provides continuous weight data over the range from 0.2 to 1.0 g
with an accuracy of + 0.1 mg, is protected from corrosive gases by
a purge flow of nitrogen. Temperature in the reaction zone is
controlled by a Marshall furnace with an accuracy of +5°C up to about
950°C and is recorded along with sample weight on a recorder.
Materials. The apparatus is fabricated from glass and type 304
stainless steel. Commercial research-grade cylinder gases are used
to make up the reactant gas stream. The stone used in these
experiments was dolomite obtained from Charles Pfizer and Co., Gibsonburg,
Ohio, and has an empirical formula Ca^ .ii+MgQ .86(^03)2. Chemical analysis
and petrographic characteristics of 1337 dolomite stone have been reported
by Harvey.24
90
-------
RECORDING
MICROBALANCE
PURGE
ROTOMETERS
VENT
PURGE
GAS IN-
REACTANT
GAS OUT"K=
REACTANT
-GAS
FURNACE -
REACTANT
*—*- GAS
A A
SCRUBBERS
REACTANT
GAS IN—*:
-BAFFLES
Q'
-SAMPLE
GAS HANDLING
SYSTEM
REACTOR
SYSTEM
Figure 26. Schematic Diagram of the TGA Apparatus
Procedure. In a typical experiment, 200 mg of 1337 dolomite
particles having diameters in a narrow range around 1.1 mm (-16 +18
U.S. mesh, U.S. Sieve Series) are placed in the apparatus under a flow
of N2 and C02 and heated to 800°C at a rate of about 25°/min to half-
calcine the stone. Calcination is followed by observing the weight
change of the sample. When half-calcination of the stone is complete
(usually after about 45 min), sample temperature is adjusted to that
selected for the experiment, the sample is isolated under the N2 and
C02 atmosphere, and the reactant gas mixture adjusted to the
appropriate composition and flowrate for the experiment while by-
passing the reaction tube. At time zero, the reactant gas is
diverted through the reaction tube and the weight change of the
sample observed as a function of time. The reaction is followed
until the rate of weight change is negligible (typically about two
hours for the experiments reported here).
To study the reaction of MgO, a sample of reagent-grade MgC03
was pressed into a dense pellet, which was then broken up to obtain
a sample containing particles of the size used in the other experiments.
Calcination and subsequent procedures were identical to those for
other samples.
91
-------
The reaction rate, r, at any time during the reaction was
calculated from the equation
where w is the total weight of the sample after calcining, n is the
amount of sulfate, as moles of 863, reacted and t is time in seconds.
Results and Discussion
Typical experimental results are shown in Fig. 27, where the
fraction of the stone reacted [according to equation (8)] is plotted
against time for several different SC>2 concentrations. The reaction
temperature and reactant-gas composition are given in the figure
caption. It is evident that the reaction rate, and hence the extent
of the reaction at a given time, is a function of S02 concentration.
Figure 28 shows this same data in a plot of the log of the initial
reaction rate (initial rates were actually evaluated at t = 1 rain,
because of scatter in the data near t = 0) versus the log S02
concentration. The straight line thus obtained has a slope of 1.08
and indicates that the reaction is first order with respect to S02
concentration in the reactant gas under these reaction conditions.
It should be noted that similar first-order dependence for the
reaction of fully-calcined 1337 dolomite has been reported by
Borgwardt^5 and, in fact, the reaction rates reported by Borgwardt
are very similar to those observed here.
It was noticed in several early experiments that the concen-
tration of water in the reactant gas appeared to influence the reaction
rate. Results from a series of experiments in which the concentration
of water in the reactant gas was varied while the concentration of other
components was constant indicated, however, that the reaction was
essentially zero order in water concentration. Another series of
experiments in which the reactant gas was dry and the concentration
of SC>2 was varied gave the results shown in Fig. 29. Here the slope
is 0.77, indicating that the rate varies with roughly the three-
fourths root of S02 concentration. Thus, it would appear that the rate-
determining step is different depending on whether or not water is
present in the reactant gas. With water present, the reaction is
0.22 order with respect to oxygen concentration in the reactant gas
stream.
The temperature dependence of the reaction rate with water present
was examined over the range from 550 to 850°C (at higher temperatures,
the concentration of C02 necessary to prevent decomposition of CaCO^^
could not be maintained with the present apparatus) . The initial
reaction rate increased significantly with temperature over this
range. An Arrhenius plot of the data, shown in Fig. 30, indicates a
linear dependence of rate on 1/T and yields an apparent activation
energy of 7.3 kcal/mol. Such a value does not point conclusively to
a mechanism in which some chemical reaction is rate controlling but
is somewhat greater than one might expect if the reaction is diffusion
controlled.
92
-------
100
REACTANT-GAS
COMPOSITION, %
AS LABELED
15.0
5.0
2.9
BALANCE
20 40 60 80 100 120 140
TIME.min
Figure 27. Percent Conversion vs. Time
for Various S02 Concentrations in the Reactant Gas at 750°C
REACTION TEMP = 750 "C
REACTANT-GAS
COMPOSITION. %
: VARIABLE
C02:I5.0
02: 5.0
H20: 2.9
N2: BALANCE
l _ I _ I
.2 .4 .6 .8 I 2 4 6 8 10
S02 CONCENTRATION X I07. mol/cc
Figure 28. Initial Reaction Rate of S02 with 1337 Dolomite
as a Function of SC>2 Concentration in the Presence of
93
-------
1
REACTION TEMP = 750 °C
REACTANT-GAS
COMPOSITION^
S02: VARIABLE
C02:I5.0
02:5.0
N2: BALANCE
' [CoC03 + MgO] + S02 + £
[CoSO,+MgO]
.2 .4 .6 .8 I 2 4 6 8 10
S02 CONCENTRATION X I0?, mol/cc
Figure 29. Initial Reaction Rate of S02 with 1337 Dolomite
as a Function of S02 Concentration in the Absence of H20
8
6
°
a- 4
;= .4
[CoCOjt MgO] + S02 + 5 02 —
.9 1.0 I.I
IO3/T, -K'1
1.2 1.3
Figure 30. Arrhenius Plot for the Reaction of S02
with Half-Calcined 1337 Dolomite
In some experimental runs, a greater than theoretical conversion
to sulfate, based on Eq. (8) was observed. It has been reported that
MgO undergoes sulfation;22 therefore, the sulfation of calcined MgC03
was examined to assess the extent to which this reaction might interfere
with detailed analysis of our results. Typical results are shown on
Fig. 31, together with results for the half-calcined stone. The extent
of sulfation of MgO is not great and the reaction rate is much less
than that for the half-calcined stone; nevertheless, the effect is
too great to be ignored in any more detailed treatment of the half-
calcined results.
94
-------
[CaO+MgO]to[CaS04
[CaC03+MgO]to[CaS04+MgO]
REACTION TEMPERATURE=750°C
0 20 40 60 80 100 120
TIME,min
Figure 31. Percent Sulfation Vs. Time
for a) [CaO + MgO]; b) [CaC03 + MgO]; and c) MgO at 750°C
Figure 31 also shows results obtained for the sulfation of fully-
calcined stone reacted under the same conditions as half-calcined
except that C02 was absent in the reactant gas. The results for the
half-calcined and fully-calcined dolomite are very similar as one
might expect in view of the generally close correlation between our
work and that of Borgwardt.15
To date, several attempts to fit our results with a simple
kinetic model such as that based on a shrinking unreacted core or on
reaction of a porous solid have not been satisfactory. A possible
explanation for the apparent role of water in the mechanism might
be that the S02 is oxidized by the water to 803 on the surface of the
reactant stone, followed by S03 diffusing through a product layer to
the unreacted core of the solid for reaction with CaCC^. This concept
would be consistent with the similar kinetics observed for half- and
fully-calcined stone since the rate-determining step would presumably
be the same in either case. This view is supported by the observation
that reactivity in a fluidized bed drops off somewhat above about
850°C, where the thermodynamics of S02 oxidation become less favorable.
On the other hand, Borgwardt's observations with fully-calcined stonellf
suggest that the decreased reactivity is due to hard-burning of the
stone.
95
-------
Further work is needed to understand more fully the role of
water in the sulfation mechanism and to extend the kinetic studies to
the reduction and regeneration reactions outlined above. The potential
advantages of a process utilizing dolomite in a closed cycle for SC>2
control are sufficiently great to warrant continued effort.
96
-------
ACKNOWLEDGMENTS
We gratefully acknowledge the help given by Mr. L. Burris,
Mr. D. S. Webster, Dr. S. Lawroski, and Mr. L. Link in directing and
reviewing the program.
We are indebted to C. Schoffstoll and J. Stockbar for their
technical assistance in the operation and maintenance of the bench-
scale fluidized-bed combustor.
We would also like to express appreciation to Mr. E. Kucera,
Mr. M. Homa, Mrs. Florence Williams, Mrs. Jackie Williams, Mr. R.
Telford, and Mr. K. Jensen for general analytical services, to Mr.
R. Bane for trace-element analyses, to Mr. R. Heinrich for neutron
activation analyses, and to Mr. J. Paris for spectrochemical
analyses.
97
-------
REFERENCES
1. A. A. Jonke et al., Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion, Annual Report,
July 1971-June 1972, ANL/ES/CEN-1004 (1972).
2. G. A. Hammons and A. Skopp, A Regenerative Limestone Process
for Fluidized-Bed Coal Combustion and Desulfurization, Final
report prepared under contract No. CPA-70-19 for Process Control
Engineering Program, Air Pollution Control Office: Esso Research
and Engineering Company, Government Research Laboratory, Linden
New Jersey (Feb. 28, 1971).
3. A. Skopp, M. S. Nutkis, G. A. Hammons, and R. R. Bertrand,
Studies of the Fluidized Lime-Bed Coal Combustion Desulfurization
System, Final report prepared under contract No. CPA-70-19 for
the Control Systems Division, Office of Air Programs, U. S.
Environmental Protection Agency: Esso Research and Engineering
Company, Government Research Laboratory, Linden, New Jersey
(Dec. 31, 1971).
A. Proceedings of Third International Conference on Fluidized-Bed
Combustion, U. S. Environmental Protection Agency Report,
EPA-650/2-73-053 (December 1973).
5. S. Ehrlich, E. B. Robison, J. S. Gordon, and J. W. Bishop,
Development of a Fluidized-Bed Boiler, AIChE Symposium Series
No. 126, Vol. 68 (1972).
6. D. Kunii and 0. Levenspiel, Fluidization Engineering, John Wiley
and Sons, Inc., New York, N. Y., 76-80 (1969).
7. E. L. Carls, Review of British Program on Fluidized-Bed
Combustion; Report of U. S. Team Visit to England,
February 17-28, 1969, Argonne National Laboratory Report,
ANL/ES/CEN-1000 (1969).
8. R. M. Statnik, D. K. Oestreich, and R. Steiber, Sampling and
Analysis of Mercury Vapor in Industrial Streams Containing
Sulfur Dioxide, Preprint of paper presented at the 1973 American
Chemical Society Annual Meeting, August 26-31, 1973.
9. C. E. Billings, A. M. Sacco, W. R. Matson, R. M. Griffin,
W. R. Coniglio, and R. A. Harley, Mercury Balance on a Large
Pulverized Coal-Fired Furnace, Preprint -of paper presented at
the 65th Annual Meeting of the Air Pollution Control Association,
June 18-22, 1972.
10. R. C. Diehl, E. A. Hattman, H. Schultz, and R. J. Haren, Fate of
Trace Mercury in the Combustion of Coal, Bureau of Mines Managing
Coal Wastes and Pollution Program, Technical Progress Report
No. 54 (1972).
98
-------
11. R. E. Lee, Jr., S. S. Goranson, R. E. Enrione, and G. B. Morgan,
National Air Surveillance Cascade Impactor Network II. Size
Distribution Measurements of Trace Metal Components, Environ.
Sci. and Technol. 6^, 1025 (1972).
12. M. Kertesz-Saringer, E. Meszaros, and T. Varkoni, Atmospheric
Environment .5, 429 (1971).
13. D. F. S. Natusch, J. R. Wallace, and C. A. Evans, Jr., Toxic
Trace Elements: Preferential Concentration in Respirable
Particles, Science 183, 202 (Jan. 18, 1974).
14. R. R. Ruch, H. J. Gluskoter, and N. F. Shimp, Occurrence and
Distribution of Potentially Volatile Trace Elements in Coal,
Illinois State Geological Survey, Environmental Geology Notes
No. 61 (1973).
15. R. H. Borgwardt, Environ. Sci. and Technol. 4^ 59 (1970).
16. E. P. O'Neill, D. L. Keairns, and W. F. Kettle, Paper presented
at 3rd International Conference on Fluidized-Bed Combustion,
Hueston Woods, Ohio, Oct. 29-Nov. 1, 1972.
17. M. Pell, Ph.D. Thesis, City University of New York (1970).
18. M. Pell, R. A. Graff, and A. M. Squires, Sulfur and SO?
Developments, AIChE (1971).
19. A. M. Squires, R. A. Graff, and M. Pell, Chemical Engineering
Progress Symposium Series 67, pp. 23-24 (1971).
20. L. A. Ruth, A. M. Squires, and R. A. Graff, Environ. Sci. and
Technol. 6:12, 1009 (1972).
21. R. R. Bertrand, A. C. Frost, and A. Skopp, Fluid Bed Studies of
the Limestone Based Flue Gas Desulfurization Process, Interim
Report, Contract No. PH 86-67-103, for National Air Pollution
Control Admin., Esso Research and Engineering Company, Linden,
N. J. (1968).
22. S. G. Narayanan, M.S. Thesis, City University of New York (1971).
23. A. M. Squires and R. A. Graff, Paper presented at 63r.d Annual
Meeting of the Air Pollution Control Association, St. Louis, Mo.,
June 14-18, 1970.
24. R. D. Harvey, Petrographic and Mineralogical Characteristics of
Carbonate Rocks Related to Sulfur Dioxide Sorption in Flue Gases
Final report, contract No. CPA 22-69-65 for U.S. Environmental
Protection Agency, Illinois State Geological Survey, Urbana,
Illinois (July 15, 1971).
99
-------
25. K. J. Hill and E. R. S. Winter, J. Phys. Chem. j>0, 1361 (1956).
26. A. A. Jonke et al,, Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion and Regeneration of Sulfur-
Containing Additives, Annual Report, July 1972-June 1973,
ANL/ES-CEN-1006 (in press).
100
-------
APPENDIX A. CHARACTERISTICS OF RAW MATERIALS
USED IN FLUIDIZED-BED COMBUSTION EXPERIMENTS
101
-------
Table A-l. Particle-Size Distribution and Chemical
Characteristics of Arkwright Coal
Sieve Analysis
U.S. Sieve No.
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 +170
-170
% on Sieve
0.0
8.0
14.2
,3
,7
12,
24.
17.9
23.0
Mean Particle Dia: 323 ym
Moisture
Volatile Matter
Fixed Carbon
Ash
Sulfur, wt %
Heating value,
BTU/lb
Proximate Analysis (wt %)
As Received Dry Basis
2.89
38.51
50.92
7.68
100.00
2.82
13,706
39.66
52.43
7.91
100.00
2.90
14,114
Carbon
Hydrogen
Sulfur
Nitrogen
Chlorine
Ash
Oxygen (by difference)
Ultimate Analysis (wt
77.14
5.23
2.90
1.66
0.19
7.91
4.97
102
-------
Table A-2. Particle-Size Distribution and Chemical
Characteristics of San Juan Subbituminous Coal
Sieve Analysis
U.S. Sieve No.
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 +170
-170
% on Sieve
0.00
11.16
10.11
15.12
25.77
19.27
18.57
Mean Particle Dia: 340 urn
Proximate Analysis, wt %
As Received Dry Basis
Moisture
Ash
Volatile Matter
Fixed Carbon
Sulfur, wt %
Heating value,
BTU/lb
9.28
16.96
33.28
40.48
100.00
0.78
9,621
._
18.70
36.68
44.62
100.00
0.86
10,605
Ultimate Analysis, wt
As Received
Dry Basis
Moisture
Carbon
Hydrogen
Nitrogen
Chlorine
Sulfur
Ash
Oxygen
(by difference)
9.28
55.82
3.96
1.14
0.10
0.78
16.96
11.96
100.00
__
61.53
4.36
1.26
0.11
0.86
18.70
13.18
100.00
Average of two samples, as received.
103
-------
Table A-3. Particle-Size Distribution and Chemical
Characteristics of Glenharold Lignite
Sieve Analysis
U.S. Sieve No.
on Sieve
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 +170
-170
Mean Particle Dia:
0.00
9.88
11.19
19.46
28.90
18.15
12.42
353 \im
Proximate Analysis, wt
As Received
Dry Basis
Moisture
Ash
Volatile Matter
Fixed Carbon
Sulfur, wt %
Heating value,
BTU/lb
30.90
6.11
30.00
32.99
100.00
0.53
7,625
__
8.84
43.42
47.74
100.00
0.77
11,035
Ultimate Analysis, wt %
As Received Dry Basis
Moisture
Carbon
Hydrogen
Nitrogen
Chlorine
Sulfur
Ash
Oxygen
(by difference)
30.90
46.04
3.03
0.72
0.04
0.53
6.11
12.63
100.00
66.63
4.38
1.04
0.06
0.77
8.84
18.28
100.00
Average of two samples, as received.
104
-------
Table A-4. Particle-Size Distribution and Chemical
Characteristics of Tymochtee Dolomite
Sieve Analysis
U.S. Sieve No. % on Sieve
+ 14 0.4
-14 +25 48.6
-25 + 35 19.9
-35 + 45 18.8
-45 +80 11.7
-80 +170 0.4
-170 0.4
Average Particle Size: 750 ym
Component Chemical Analysis (wt %)
Ca 20.0
Mg 11.3
C02 38.5
H2 0.2
Derived Composition
CaC03 50.0
MgC03 39.1
105
-------
Table A-5. Particle-Size Distribution and Chemical
Characteristics of Type-38 Alundum Grain
Obtained from the Norton Company
Sieve Analysis
U.S. Sieve No. Wt % on Sieve
+14 0.0
-14 +25 15.7
-25 +35 73.3
-35 +45 11.0
-45 +80 0.0
Total 100.0
Typical Chemical Analysis
Component Wt %
A1203 99.49
Si02 0.05
Fe203 0.10
Ti02 0.01
Na20 0.35
106
-------
APPENDIX B. CARBON, SULFUR, AND CALCIUM MATERIAL BALANCES
FOR COMBUSTION EXPERIMENTS
107
-------
O
00
Table B-l. Carbon, Sulfur, and Calcium Material Balances for VAR-1 Experiment
Combustor Pressure, psia = 120 Gas Velocity, ft/sec =2.1
= 1445 Flue-Gas 02, % =2.74
= 2.9
Bed Temperature, °F
Ca/S Mole Ratio
Source of Material
Material in
*3
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gask
% Balance
Weight,
g
14,000
39,917
19,958
15,730
16,264
4,998
454
278
—
Carbon,
wt %
1.9
74.91
10.5
3.8
4.4
51.4
30.8
10.0
—
Carbon,
g
266
29,902
2,096
32,264
598
716
2,569
140
28
3,324C
37,375
116
Sulfur,
wt %
10.9
2.82
0.3
7.0
5.8
1.0
1.8
3.4
Sulfur,
g
1,526
1,126
60
2,712
1,101
943
50
8
9^
61d
2,172
80
Calcium,
wt %
23.9
0.44
20.0
23.5
25.1
3.58
4.2
4.4
—
Calcium,
g
3,346
176
3,992
7,514
3,697
4,082
179
19
12
—
7,989
106
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 47.5.
'Calculated for avg. CO cone, of 98 ppm plus avg. C02 cone, of 172,000 ppm.
Calculated for avg. S02 cone, of 118 ppm.
-------
o
VO
Table B-2. Carbon, Sulfur, and Calcium Material Balances for VAR-2 Experiment
Combustor Pressure, psia = 120 Gas Velocity, ft/sec = 4.8
Bed Temperature, °F = 1566 Flue-Gas 02, ft =3.05
Ca/S Mole Ratio = 2.9
Source of Material
Material in
a
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas6
% Balance
Weight,
g
11,353
63,731
34,474
15,000
3,018
32,391
787
961
—
Carbon,
wt %
1.8
74.91
10.5
4.2
3.7
11.1
11.7
4.7
—
Carbon,
g
204
47,741
3,620
51,565
630
112
3,595
92
45
54,109°
58,583
114
Sulfur,
wt %
8.8
2.82
0.3
6.3
6.5
4.5
3.6
4.2
—
Sulfur,
g
999
1,797
103
2,899
945
196
1,458
28
4°,»
107d
2,774
96
Calcium,
wt %
22.1
0.44
20.0
23.1
23.3
18.7
9.3
10.4
—
Calcium
g
2,509
280
6,895
9,684
3,465
703
6,057
73
100
—
10,398
107
Partially sulfated dolomite.
3 Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 102.9.
"Calculated for avg. CO cone, of 31 ppm plus avg. C02 cone, of 174,000 ppm.
Calculated for avg. S02 cone, of 129 ppm.
-------
Table B-3. Carbon, Sulfur, and Calcium Material Balances for VAR-3 Experiment
Combustor Pressure, psia = 120
Bed Temperature, °F = 1574
Ca/S Mole Ratio = 1.1
Gas Velocity, ft/sec = 2.1
Flue-Gas 02, % =2.95
Source of Material
Material in
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gasb
% Balance
Weight,
g
12,603
187,790
34,927
9,978
30,997
15,714
1,742
1,254
—
Carbon,
wt %
2.4
74.91
10.5
1.04
1.7
31.1
21.0
6.1
—
Carbon,
g
302
140,673
3,667
144,642
104
527
4,887
366
76
143,621°
149,581
103
Sulfur,
wt %
8.5
2.82
0.3
12.2
10.2
2.0
2.0
3.0
Sulfur,
g
1,071
5,296
105
6,472
1,217
3,162
314
35
38^
914d
5,680
88
Calcium,
wt %
21.4
0.44
20.0
19.7
20.1
4.5
3.8
4.8
Calcium
g
2,697
826
6,985
10,508
1,966
6,230
707
66
60
9,029
86
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 46.0.
Calculated for avg. CO cone, of 31 ppm plus avg. C02 cone, of 147,000 ppm.
Calculated for avg. S02 cone, of 351 ppm.
-------
Table B-4. Carbon, Sulfur, and Calcium Material Balances for VAR-4 Experiment
Combustor Pressure, psia = 120 Gas Velocity, ft/sec = 2.3
Bed Temperature, °F = 1648 Flue-Gas 02, % =2.97
Ca/S Mole Ratio = 1.9
Source of Material
Material in
Starting beda
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas°
% Balance
Weight,
g
14,000
98,204
33,793
12,228
25,871
10,725
457
667
—
Carbon,
wt %
3.3
74.91
10.5
1.8
2.4
17.7
9.2
5.7
—
Carbon,
g
462
73,565
3,548
77,575
220
621
1,898
42
38
84,642C
87,461
113
Sulfur,
wt %
6.2
2.82
0.3
8.8
7.6
2.3
2.7
3.1
—
Sulfur,
8
868
2,769
101
3,738
1,076
1,966
247
12
21J
170d
3,492
93
Calcium,
wt %
24.0
0.44
20.0
22.1
23.2
8.2
10.1
9.9
Calcium,
g
3,360
432
6,759
10,551
2,702
6,002
879
46
66
—
9,695
92
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 47.9.
"Calculated for avg. CO cone, of 48 ppm plus avg.
Calculated for avg. S02 cone, of 122 ppm.
cone, of 162,000 ppm.
-------
Table B-5. Carbon, Sulfur, and Calcium Material Balances for VAR-5 Experiment
Combustor Pressure, psia = 120 Gas Velocity, ft/sec = 3.4
Bed Temperature,
Ca/S Mole Ratio
Source of Material
o
Material in
Starting bed
Arkwright coal
Dolomite additive
o
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gasc
% Balance
Weight ,
g
28,310
195,182
33,113
31,890
20,467
27,108
3,452
2,344
°F = 1461
= 1.0
Carbon,
wt %
2.74
74.91
10.5
2.04
2.61
48.83
37.40
23.85
Carbon,
g
111
146,210
3,477
150,464
649
535
13,236
1,291
559
144,572d
160,842
107
Flue-Gas 02
Sulfur ,
wt %
9.37
2.82
0.3
10.51
9.98
1.59
1.74
2.60
, %
Sulfur,
g
2,653
5,504
99
8,256
3,352
2,043
431
60
61
l,696e
7,643
93
3.00
Calcium,
wt %
18.93
0.44
20.0
17.18
17.38
3.49
3.33
4.18
Calcium,
g
5,360
859
6,622
12,841
5,478
3,557
947
115
98
10,195
79
values shown represent composite data for two segments of the experiment.
Partially sulfated dolomite.
p
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 76.6.
Calculated for avg. CO cone, of 755 ppm plus avg. C02 cone, of 154,000 ppm.
£
Calculated for avg. S02 cone, of 681 ppm.
-------
U5
Table B-6. Carbon, Sulfur, and Calcium Material Balances for VAR-6 Experiment
Combustor Pressure, psia = 120 Gas Velocity, ft/sec =3.6
Bed Temperature, °F = 1565 Flue-Gas 02, % =3.00
Ca/S Mole Ratio = 2.0
Source of Material
Material in
•a
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas"
% Balance
Weight ,
g
14,266
91,854
32,432
16,210
17,047
16,219
618
492
—
Carbon,
wt %
4.2
74.91
10.5
4.5
3.8
25.1
19.6
5.7
—
Carbon,
8
599
68,808
3,405
72,812
729
648
4,071
133
28
70,601°
76,210
105
Sulfur,
wt %
6.3
2.82
0.3
7.0
7.0
4.0
2.8
4.2
—
Sulfur,
g
899
2,590
97
3,586
1,135
1,193
649
19
21J
201d
3,218
90
Calcium,
wt %
23.1
0.44
20.0
21.2
22.0
13.2
6.1
7.2
—
Calcium,
g
3,295
404
6,486
10,185
3,437
3,750
2,141
42
35
9,405
92
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 77.5.
"Calculated for avg. CO cone, of 32 ppm plus avg. C02 cone, of 158,000 ppm.
Calculated for avg. S02 cone, of 169 ppm.
-------
H
Table B-7. Carbon, Sulfur, and Calcium Material Balances for VAR-6-R Experiment
Combustor Pressure, psia = 120 Gas Velocity, ft/sec = 3.6
Bed Temperature, °F = 1558 Flue-Gas 02, % =2.86
Ca/S Mole Ratio = 2.0
Source of Material
Material in
Starting bed3
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas"
% Balance
Weight ,
g
15,385
84,1^3
31,298
13,665
15,194
20,317
642
559
Carbon,
wt %
1.6
74.91
10.5
2.4
2.3
18.8
18.3
9.0
Carbon,
g
246
63,032
3,286
66,564
328
349
3,820
117
50
64,982
69,646
105
Sulfur,
wt %
10.8
2.82
0.3
8.5
8.9
4.6
2.5
3.8
Sulfur,
g
1,662
2,373
94
4,129
1,162
1,352
935
16
21J
235d
3,721
90
Calcium,
wt %
18.0
0.44
20.0
21.4
21.6
13.8
5.0
5.8
Calcium,
g
2,769
370
6,260
9,399
2,924
3,282
2,804
32
32
—
9,074
97
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 76.9.
Calculated for avg. CO cone, of 33 ppm plus avg. C02 cone, of 154,000 ppm.
Calculated for avg. S02 cone, of 209 ppm.
-------
H1
M
Ui
Table B-8. Carbon, Sulfur, and Calcium Material Balances for VAR-6-2R Experiment
Combustor Pressure, psia = 122 Gas Velocity, ft/sec = 3.5
Bed Temperature, °F = 1549 Flue-Gas 02, % =2.95
Ca/S Mole Ratio = 2.0
Source of Material
Material in
rt
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas°
% Balance
Weight,
g
15,386
86,864
29,938
16,152
12,059
18,335
752
813
Carbon,
wt %
2.4
74.91
10.5
3.3
2.8
20.8
18.9
7.4
—
Carbon,
g
369
65,070
3,143
68,582
533
338
3,814
142
60
67,266°
72,153
105
Sulfur,
wt %
7.8
2.82
0.3
8.3
8.0
4.4
2.6
4.0
—
Sulfur,
g
1,200
2,450
90
3,740
1,341
965
807
20
33,
214d
3,380
90
Calcium,
wt %
20.6
0.44
20.0
21.4
24.0
13.4
5.6
6.5
Calcium,
g
3,169
382
5,988
9,539
3,456
2,894
2,457
42
53
8,902
93
rartially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 77.1.
Calculated for avg. CO cone, of 33 ppm plus avg. C02 cone, of 159,000 ppm.
Calculated for avg. S02 cone, of 190 ppm.
-------
Table B-9. Carbon, Sulfur, and Calcium Material Balances for VAR-7 Experiment
Combustor Pressure, psia = 127
Bed Temperature, °F = 1458
Ca/S Mole Ratio = 2.2
Gas Velocity, ft/sec = 4.2
Flue-Gas 02> % =2.7
Source of Material
Material in
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas"
% Balance
Weight ,
g
11,740
83,916
32,659
15,777
7,070
23,586
4,718
3,739
—
Carbon ,
wt %
2.6
74.91
10.5
2.4
2.4
23.0
38.2
17.4
Carbon ,
8
305
62,861
3,429
66,595
379
170
5,425
1,802
651
60,662C
69,089
104
Sulfur,
wt %
7.9
2.82
0.3
7.8
8.2
4.4
2.3
3.6
Sulfur ,
g
927
2,366
98
3,391
1,231
580
1,038
109
135,
266d
3,359
99
Calcium,
wt %
21.0
0.44
20.0
20.6
23.5
14.1
7.7
9.7
Calcium,
g
2,465
369
6,532
9,366
3,250
1,661
3,326
363
363
8,963
96
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 100.4.
"Calculated for avg. CO cone, of 86 ppm plus avg. C02 cone, of 159,000 ppm.
Calculated for avg. S02 cone, of 262 ppm.
-------
Table B-10. Carbon, Sulfur, and Calcium Material Balances for VAR-8 Experiment
Combustor Pressure, psia = 120
Bed Temperature, °F = 1630
Ca/S Mole Ratio = 3.2
Gas Velocity, ft/sec =3.6
Flue-Gas 02, % =3.05
Source of Material
Material in
Starting bed
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas
% Balance
Weight ,
g
15,565
65,318
36,742
15,787
12,928
20,054
874
545
—
Carbon,
wt %
4.0
74.91
10.5
3.3
3.4
11.6
8.5
4.6
—
Carbon,
g
623
48,930
3,858
53,411
521
439
2,326
74
25
58,771C
62,156
116
Sulfur,
wt %
5.8
2.82
0.3
6.2
5.7
4.1
3.0
3.5
—
Sulfur,
g
903
1,842
110
2,855
979
737
822
26
19,
118d
2,701
95
Calcium,
wt %
24.4
0.44
20.0
24.0
24.1
20.1
9.9
9.9
—
Calcium,
g
3,798
287
7,348
11,433
3,789
3,116
4,031
87
54
11,077
97
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 76.5.
"Calculated for avg. CO cone, of 45 ppm plus avg. C02 cone, of 174,000 ppm.
Calculated for avg. S02 cone, of 131 ppm.
-------
Table B-ll. Carbon, Sulfur, and Calcium Material Balances for VAR-9 Experiment
oo
Combustor Pressure, psia = 121
Bed Temperature, °F = 1665
Ca/S Mole Ratio = 1.0
Gas Velocity, ft/sec = 4.9
Flue-Gas 02, % =2.93
Source of Material
Material in
Starting bed3
Arkwright coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Secondary filter
Flue gas"
% Balance
Weight,
g
11,761
206,615
35,608
14,279
9,338
45,430
1,531
1,370
Carbon,
wt %
3.3
74.91
10.5
0.5
0.9
30.0
14.9
5.3
Carbon,
g
388
154,775
3,739
158,902
71
84
13,629
228
73
157,114°
171,199
108
Sulfur,
wt %
8.3
2.82
0.3
11.7
10.5
3.5
2.8
4.7
Sulfur,
g
976
5,826
107
6,909
1,671
980
1,590
43
64^
2,229
6,577
95
Calcium,
wt %
21.4
0.44
20.0
17.7
20.4
11.4
6.4
6.6
Calcium,
g
2,517
909
7,122
10,548
2,527
1,905
5,179
98
90
9,799
93
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 100.3
'Calculated for avg. CO cone, of 38 ppm plus avg. C02 cone, of 159,000 ppm.
Calculated for avg. S02 cone, of 846 ppm.
-------
Table B-12. Carbon, Sulfur, and Calcium Material Balances for Experiment LIG-1
120 Gas Velocity, ft/sec = 3.51
Combustor Pressure, psia
Bed Temperature, °F = 1562
Ca/S Mole ratio = 1.07
(excluding Ca in coal)
Flue-Gas 02, %
= 3.2
Source of Material
Material in
Starting bed3
Glenharold Lignite
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Filters
Flue gas
% Balance
Weight,
8
13,240
669,060
25,628
13,422
17,731
50,769"^
982 >
875J
Carbon,
wt %
1.5
46.04
10.5
1.7
2.2
16.9
Carbon,
g
199
308,035
2,691
310,925
288
390
8,894
272,345°
281,857
91
Sulfur,
wt %
9.4
0.53
0.3
10.6
10.6
2.9
Sulfur,
g
1,245
3,546
77
4,868
1,423
1,879
1,526
818d
5,646
116
Calcium,
wt %
15.9
1.03
20.0
18.6
17.6
13.5
Calcium,
g
2,105
6,891
5,126
14,122
2,496
3,121
7,104
12,721
90
Partially sulfated dolomite.
Avg. flue-gas flowrate (dry basis), cfm at 70°F and 1 atm = 75.9.
•>
"Calculated for avg. CO cone, of 123 ppm plus avg. C02 cone, of 172,000 ppm.
Calculated for avg. S02 cone, of 194 ppm.
-------
K3
O
Table B-13. Carbon, Sulfur, and Calcium Material Balances for Experiment SJ-1A
Combustor Pressure, psia = 120 Gas Velocity, ft/sec = 3.56
Bed Temperature, °F = 1565 Flue-Gas 02, % =2.91
Ca/S Mole Ratio = 1.1
(excluding Ca in coal)
Source of Material
Material in
Starting bed
San Juan coal
Dolomite additive
Material out
Final bed
Bed overflow
Primary cyclone
Secondary cyclone
Filters
Flue gas
% Balance
Weight,
g
16,356
449,971
30,164
13,240
21,851
87,109"^
2,956 \
2,187 J
Carbon,
wt %
1.1
55.82
10.5
1.5
1.2
16.4
Carbon,
g
180
251,174
3,167
254,521
199
262
15,129
235,217°
250,807
98
Sulfur,
wt %
11.0
0.78
0.3
9.4
10.5
1.2
Sulfur ,
g
1,799
3,510
90
5,399
1,245
2,294
1,107
1,213d
5,859
108
Calcium,
wt %
16.3
0.75
20.0
15.9
16.4
3.6
Calcium,
g
2,666
3,375
6,033
12,074
2,105
3,584
3,321
9,010
75
Partially sulfated dolomite.
"*Avg. flue-gas flowrate (dry basis) at 70°F and 1 atm = 76.5.
•»
"Calculated for avg. CO cone, of 46 ppm plus avg. C02 cone, of 166,000 ppm.
Calculated for avg. S02 cone, of 321 ppm.
-------
APPENDIX C. SCREEN ANALYSIS DATA
121
-------
Table C-l.
M
fO
Sieve Analysis of Bed and Elutriated Solids from "VAR" Experiments
RUN NO.
Temp, °F
Ca/S, Mole Ratio
Gas Velocity, ft/sec
FINAL BED
% On Screen Size:
+14
-14, +25
-25, +35
-35, +45
-45, +80
-80 » +170
-170
Mean Particle Dia, ym
OVERFLOW
Wt % On Screen Size:
+14
-14, +25
-25, +35
-35, +45
-45, +80
-80, +170
-170
Mean Particle Dia, ym
2
1550
3
4.8
0.5
71.3
26.6
1.4
0.1
0.1
0.0
930
0.3
76.9
18.2
2.6
1.6
0.2
0.2
945
7
1450
2
4.2
0.8
51.2
31.5
15.0
1.6
0.0
0.0
813
0.0
45.3
23.4
7.1
24.1
0.1
0.1
714
9
1650
1
4.9
0.7
52.4
43.8
2.9
0.2
0.0
0.0
846
0.7
66.9
29.4
2.7
0.2
0.2
0.0
909
5
1450
1
3.4
0.4
49.2
35.7
12.4
2.3
0.0
0.0
803
1.0
60.7
28.3
8.6
1.4
0.0
0.0
874
6
1550
2
3.6
0.6
64.7
29.8
5.0
0.0
o.o
0.0
897
1.0
73.7
22.6
2.8
0.0
0.0
0.0
947
6-R
1550
2
3.6
0.5
53.7
34.6
10.7
0.5
0.0
0.0
833
1.2
67.7
25.5
5.6
0.1
0.0
0.0
919
6-2R
1550
2
3.5
0.5
48.5
39.8
10.0
1.2
0.0
0.0
808
0.5
61.4
30.4
6.9
0.8
0.0
0.0
874
8
1650
3
3.6
0.0
45.9
42.2
10.4
1.4
0.0
0.0
787
0.0
58.6
34.4
6.5
0.5
0.0
0.0
856
1
1450
3
2.1
0.6
44.4
33.4
15.0
6.7
0.0
0.0
765
3.7
22.3
50.5
17.2
4.2
0.8
1.2
704
3
1550
1
2.1
0.0
31.8
28.7
19.0
20.3
0.2
0.0
643
0.0
26.7
29.4
20.2
23.1
0.5
0.2
607
4
1650
2
2.3
0.0
37.3
35.1
17.3
10.2
0.0
0.0
706
0,0
69.4
23.3
5.5
1.8
0.0
0.0
902
-------
Table C-l (Cont'd.). Sieve Analysis of Bed and Elutriated Solids from "VAR" Experiments
OJ
RUN NO.
Temp, °F
Ca/S, Mole Ratio
Gas Velocity, ft/sec
PRIMARY CYCLONE
Wt % on Screen Size:
+45
-45, +80
-80, +100
-100, +170
-170, +230
-230, +325
-325
Mean Particle Dia, ym
SECONDARY CYCLONE
Wt % on Screen Size:
+45
-45, +80
-80, +100
-100, +170
-170, +230
-230, +325
-325
Mean Particle Dia, ym
2
1550
3
4.8
51.9
23.1
1.9
7.7
3.9
3.9
7.7
257
16.4
6.0
6.0
7.5
3.0
3.0
58.2
109
7
1450
2
4.2
47.2
24.5
1.9
7.6
3.8
3.8
11.3
244
51.2
17.1
2.4
7.3
4.9
4.9
12.2
241
9
1650
1
4.9
38.2
23.5
2.9
8.8
5.9
5.9
14.7
208
6.7
6.7
2.2
8.9
6.7
8.9
60.0
61
5
1450
1
3.4
8.4
23.0
9.5
16.0
7.9
9.1
26.2
142
5.9
17.7
0.0
17.7
11.8
11.8
35.3
74
6
1550
2
3.6
34.6
23.1
3.9
11.5
3.9
7.7
15.4
193
9.1
9.9
4.6
9.1
4.6
9.1
54.6
69
6-R
1550
2
3.6
36.9
24.1
4.8
7.3
3.6
4.2
19.1
229
4.4
4.4
0.0
4.4
4.4
4.4
78.3
43
6-2R
1550
2
3.5
48.2
25.0
1.8
7.1
3.6
3.6
10.7
249
4.9
12.2
9.8
14.6
9.8
9.8
39.0
91
8
1650
3
3.6
47.7
23.1
1.5
6.2
3.1
4.6
13.9
244
20.0
6.7
2.2
4.4
6.7
4.4
55.6
107
1
1450
3
2.1
4.0
12.0
4.0
16.0
12.0
16.0
36.0
61
15.8
10.5
0.0
10.5
10.5
15.8
36.8
79
3
1550
1
2.1
0.0
19.7
14.1
32.4
15.5
18.3
0.0
124
3.7
3.7
0.0
7.4
7.4
14.8
63.0
36
4
1650
2
2.3
5.8
15.4
3.9
13.5
7.7
9.6
44.2
97
11.1
4.4
2.2
4.4
4.4
4.4
68.9
71
-------
Table C-l (Cont'd.)- Sieve Analysis of Bed and Elutriated Solids from "VAR" Experiments
RUN NO. 27956
Temp, °F 1550 1450 1650 1450 1550
Ca/S, Mole Ratio 32112
Gas Velocity, ft/sec 5 5 5 3.5 3.5
FILTER
Wt % on Screen Size:
+45 MATERIAL NOT SCREENABLE
-45, +80
-80, +100
-100, +170
-170, +230
-230, +325
-325
Mean Particle Dia, ym
6-R 6-2R
1550 1550
2 2
3.5 3.5
28.0
28.0
12.0
16.0
8.0
4.0
4.0
193
8134
1650 1450 1550 1650
3312
3.5 2 2 2
MATERIAL NOT SCREENABLE
-------
Table C-2. Particle-Size Distribution of
Bed and Elutriated Solids for SJ-1A and LIG-1 Experiments
Experiment
Temp, °F
Ca/S, Mole Ratio
Gas Velocity, ft/sec
Starting Bed
Wt % on Screen Size
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 +170
-170
Mean Particle Dia, ym
Final Bed
Wt % on Screen Size
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 +170
-170
Mean Particle Dia, um
Overflow
Wt % on Screen Size
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 +170
-170
Mean Particle Dia, urn
Primary Cyclone
Wt % on Screen Size
+ 45
-45 + 80
-80 +100
-100 +170
-170 +230
-230 +325
-325
Mean Particle Dia, pm
SJ-1A
1565
1.10
3.56
0.00
38.70
40.16
20.58
0.56
0.00
0.00
734
0.00
34.70
32.76
30.81
1.62
0.11
0.00
693
0.23
41.21
33.41
23.75
1.28
0.12
0.00
742
13.02
26.93
6.84
18.54
8.83
14.13
11.70
145
LIG-1
1562
1.07
3.51
0.00
34.70
32.76
30.81
1.62
0.11
0.00
693
0.00
41.61
36.77
20.97
0.32
0.32
0.00
737
0.00
47.47
32.26
19.59
0.46
0.23
0.00
772
12.15
21.55
3.87
13.26
9.39
15.47
24.31
112
125
-------
APPENDIX D. CONCENTRATION OF FINE PARTICULATE MATTER
IN THE FLUE GAS EXHAUSTED FROM THE FLUIDIZED-BED
COMBUSTION SYSTEM
126
-------
APPENDIX D. CONCENTRATION OF FINE PARTICULATE MATTER
IN THE FLUE GAS EXHAUSTED FROM THE FLUIDIZED-BED
COMBUSTION SYSTEM
A Brink, five-stage inertial impactor, capable of sampling and
classifying gas-entrained particulate matter into five size ranges from
^0.3 to 3,0 um in diameter, was an integral component of the mercury
and fluoride sampling train used in the trace-element distribution
studies. As illustrated in Figs. 22 and 23 of the text, sampling of the
flue gas was accomplished as the gas was being vented from the system,
i.e., after the flue gas had passed through the particulate removal de-
vices (cyclones and fitters) and the system-pressure control valve.
The impactor, together with an in-line cyclone and glass fiber filter,
removed the fine particulate emissions from the sample of flue gas be-
fore the sample passed through the fluoride and mercury scrubbing solu-
tions (see Fig. 23). Details concerning the impactor and the sampling
apparatus are included in both this report (pp.71-73) and in a previous
annual report.26
With the Brink impactor, it was possible during the operation of
the mercury and fluoride sampling equipment to obtain a limited amount
of data on the loading and size distribution of the fine particulate
emissions (^0.3 to 3.0 vim particles) in the flue gas being vented from
the combustion system.
It should be emphasized, that the data reported here are not neces-
sarily representative of the loadings and particle-size distributions
that would be obtained in a commercial unit with an optimally designed
system for the high-temperature removal of fine particulate matter from
the flue gas. The data are presented here as a reporting of what was
observed and as a possible basis for making future comparisons and
evaluations.
With this understanding, Table D-l summarizes the data obtained
with the Brink impactor. The first three experiments in the table were
performed with two cyclone stages and a single filter stage in the
off-gas clean-up system. The loadings are still quite high, ranging
from 0.051 to 0.089 grains/ft3. The last three experiments in the
table, made with two cyclone stages and two filter stages in the off-gas
clean-up system, exhibited concentration levels which should be compat-
ible with the requirements for use with a gas turbine.
127
-------
TABLE D-l. Fine Particulate Emissions for Selected Combustion Experiments
oo
Experiment
TRACE-3b
TRACE-4Bb
VAR-7b
VAR-9C
TRACE-5AC
TRACE-6C
Operating
Conditions
Summarized
in Text
Table No.
17
17
2
2
17
17
Particulate Matter Recovered on
In-Line
Cyclone
15.2
10.5
9.3
Td
T
T
Stage
1
,3.6-*
M.O J
20.9
13.6
11.8
0.7
T
0.1
Stage
2
(2.0)
26.5
12.4
12.2
0.8
T
T
Stage
3
(1.2)
13.8
10.3
15.6
0.4
T
T
Stage
4
(0.8)
6.3
7.7
6.7
T
0.1
T
Each Stage, mg
Stage
5
(0.4)
4.4
7.0
7.3
T
0.1
T
Wall
Losses
12.7
1.7
3.4
T
T
T
Filter
9.4
18.9
13.0
0.6
0.6
0.6
Emissions,
grains/ft3
at 70°F, 1 atm
(dry basis)
0.089
0.064
0.051
0.001
<0.001
<0.001
Geometric mean particle diameter in microns. Determined by means of a computer program furnished by
Mr, Bruce Harris, Research Branch, Control Systems Laboratory, National Environmental Research Center,
Research Triangle Park, North Carolina and substantiated by analysis of photomicrographs of material
impacted on the various stages.
Experiments performed with 2 cyclone stages and 1 filter stage in off-gas clean-up system.
r»
Experiments performed with 2 cyclone stages and 2 filter stages in off-gas clean-up system.
Trace (<0.1 mg).
-------
APPENDIX E. TRACE-ELEMENT ANALYTICAL PROCEDURES
129
-------
APPENDIX E. TRACE-ELEMENT ANALYTICAL PROCEDURES
First-Priority Elements
Mercury, Solids. For solid samples (coal, fly ash, additive,
etc.) approximately 0.1 to 1 g of well-ground material is used. The
sample is placed in a quartz boat in a quartz tube and is combusted
at 600-700°C in a pure oxygen stream flowing at a rate of 280 ml/min.
The combustion gases containing the mercury vapor are passed through
an absorbent train consisting of two tubes, each of which contains
12.5 ml of a 4% w/v solution of KMnOit, 14 ml of 1:4 E^SC^, and 23.5 ml
of water. After combustion is complete (15-20 min), the absorbent
solution is decolorized with 10% hydroxylamine hydrochloride dropwise
(no excess) and made up to a final volume with water. This volume
may be as large as 1 liter depending on the anticipated mercury
concentration. An aliquot is then transferred to an aeration bottle
and 2 ml of stannous chloride is added just prior to the analysis to
reduce mercury to the metallic state. (The stannous chloride solution
is prepared by dissolving 20 g of SnCl2 in 40 ml HC1 and diluting this
solution to 250 ml with water.) The solution is then aerated, and
the air stream containing the mercury is passed through a 10-cm
absorption cell where the mercury is measured by cold vapor atomic
absorption at the 253.7 nM wavelength. The instrument can be used
to measure mercury in the range of 0.05 to 0.4 ng. Analyses are
done at least in triplicate because experience has indicated that
the distribution of mercury may be somewhat heterogeneous.
Mercury, Gold Coil. Mercury which has been amalgamated on the
gold coil is released by heating in a quartz tube to 700°C within an
02 or inert gas purge stream. The mercury vapor is then treated as
indicated in the above procedure.
Mercury, Solutions. For iodine monochloride solutions a suitable
aliquot is transferred to a small beaker, 5 ml of 1 M NaOH is added,
and the solution is mixed well. This is then transferred to the
10-cm cold-vapor cell of the atomic absorption unit and diluted to
45 ml with water. Five milliliters of reducing agent (12 g of
hydroxylamine sulfate and 12 g of sodium chloride diluted to 100 ml
with water) is then added and the analysis is performed in the same
way as indicated for solid samples.
For sodium carbonate solutions, 10 ml of solution is transferred
to a small beaker, and KMnO^ is added until a pink color is retained.
The solution is then diluted with water at least twofold and made
acidic (to pH <6) with sulfuric acid prior to taking 50 ml and
transferring it to the atomic absorption unit. The instrumental
procedure used is the same as that outlined above.
Fluoride, Coal. To determine fluoride in coal, 0.8 to 1 g of
sample and 5 ml of 1 M NaOH are placed in a Paar Bomb. Combustion
requires ^1/2 hr in pure oxygen. After cooling, the contacted bomb
130
-------
surfaces are rinsed with water and the solution is quantitatively
transferred, to a beaker and diluted to ^40 ml with water. The pH is
then adjusted to 5.2 with dilute (0.25 M) H2SOi,. and heated over a hot-
water bath for approximately 10 min. On cooling, 5 ml of acetate
buffer (294 g sodium citrate and 20.2 g KN03 adjusted to pH 6 with
citric acid) is added and the sample diluted to 50 ml. Instrumental
analyses are performed using the fluoride specific ion electrode
(S.I.E.) and standards prepared in,acetate buffer solutions.
Fluoride, Non-Coal Solid. The determination of fluoride in
samples from the combustor bed, cyclone, and filter is being done
currently by a two-step leaching procedure followed by analysis of
solutions with the S.I.E. In this procedure, VLOO ml of a 25% by
volume solution of TISAB (Orion total ionic strength adjustment buffer)
in water is added to 1 to 2 g of sample. After the residue is allowed
to settle, it is separated from the supernatant liquid, washed
thoroughly, and transferred to a platinum crucible, where it is air-
dried. The dried residue is mixed with 2 g of Eschk's mixture (2:1
MgO/Na2C03), covered with an additional 1 g of mixture, and heated
at 900°C for 35 min. When cool, the mixture is leached with fresh
TISAB solution. Fluoride analyses of all solutions, either combined
or on an individual basis, are done using the S.I.E.
Another procedure which will be explored involves the separation
of the fluoride from samples by a pyrohydrolysis technique, followed
by determination in aqueous medium using the S.I.E. In this procedure,
a sample of known weight would be mixed with a suitable chemical
accelerator and heated to ^900°C in steam and oxygen. The volatilized
fluoride silicates would be collected in an aqueous catch solution
for S.I.E. analysis.
Fluoride, Solutions. Fluoride analysis on sodium carbonate
solutions are performed by first diluting with the Orion TISAB solution
followed by a S.I.E. determination. Depending on the concentration of
fluoride anticipated, 75 ml of sample or sample plus water is taken
to a 100-ml final volume with TISAB. The S.I.E. determination on this
solution is compared with standards prepared in the TISAB solution.
Lead. A well-ground coal sample of ^1 g is placed in a platinum
crucible and ashed overnight at 450°C. (Non-coal samples are not ashed.)
The ash or starting sample is then treated with 4 ml of 50% HN03 (or
10 ml for samples of additive), evaporated to approximately one-half
the original volume, and filtered through Whatman #42 filter paper.
An atomic absorption determination is performed at 217 nM and compared
with standards in a similar matrix. The optimum concentration for
atomization is 0.1 to 1 ug/ml.
Beryllium. Coal samples are ashed at 450°C overnight and then
transferred to a 250 ml Erlenmeyer flask. (Non-coal samples are not
ashed.) Heat is applied with caution to be sure to retain any aerosols
formed while S03= fumes are allowed to escape. One to two ml of hydrogen
131
-------
peroxide is added to oxidize organic materials present and heating is
continued until reaction ceases. (This is apparent by the cessation
of bubbling.) Three grams of anhydrous sodium sulfate is added and
heating is continued on a hot plate for approximately 15 min to
volatilize excess acid and ensure the formation of pyrosulfates. The
solution is then heated over a Meeker burner (by hand) until the
vigorous reaction subsides (3-4 min) and no additional fumes form.
The flask is allowed to cool, 20 ml of water is added, and heat is
applied to dissolve the solid melt formed. Silicates are then removed
by filtration, washed with hot water, and all solutions combined and
taken to a 50 ml final volume. A pH adjustment is made and the Be
concentration of the solution is determined by fluorescence spectro-
scopy using a Beckman DU spectrophotometer.
Second- and Third-Priority Elements
Elements of second and third priority are determined by neutron
activation analysis, a nuclear technique in which the elements to be
analyzed are made radioactive by irradiation with neutrons. The gamma
rays produced by the induced radioactive species are then measured and
identified. The radioactivity of a particular neutron activation
product is directly proportional to the concentration of its parent
isotope. The basic equation for neutron activation is
AN = na = neutron flux, in n/cm2-sec
t = irradiation time, in sec
By using a suitable flux monitor and a detector of known counting
efficiency, the concentration of an element in a sample can be calculated
by means of the above equation. Fundamentally, the steps in a typical
analysis are as follows:
1. Irradiate weighed quantities of the sample (M.OO mg) and a
flux monitor in suitable containers for a sufficient time to produce
adequate radioactivity for the element of interest.
2. After irradiation, the analysis is facilitated by gamma
counting the samples and flux monitor on an absolutely calibrated
Ge(Li) detector at optimum times to directly measure radioactive
species with short, intermediate, and long half-lives. To ensure
proper identification of the species being measured, the half-life,
and, where possible, secondary associated gamma-rays are checked at
convenient time intervals.
132
-------
The samples and flux monitors are irradiated in Argonne's "CP-5"
reactor in either the "isotope tray" or the "rabbit" facility where
the thermal neutron flux available is about 1.5 x 1012 to 5 x 1013
n/(cm2)(sec). The samples are irradiated to a fluence of about
1 x 1017 n/cm2. The Ge(Li) detector used for counting is coupled to
a 4096-channel analyzer system. The detector is a coaxial-type Ge(Li)
having a full width half maximum (FWHM) of 2.0 keV for the 1.33-MeV
gamma ray of 60Co and a peak-to-Compton ratio of about 30:1. The data
readout from the analyzer system is in the form of magnetic tape
compatible with the IBM 370/195 computer system which is used for
analysis of the pulse-height spectra.
The neutron activation procedure is currently being developed.
133
-------
APPENDIX F, CONVERSION FACTORS. ENGLISH TO METRIC UNITS
134
-------
TABLE F-l. TABLE OF CONVERSION FACTORS, ENGLISH TO METRIC UNITS
Length
Area
Volume
Mass
Pressure
Temperature
Energy
Velocity
Flow
Heat-Transfer
Coefficient
English Unit
in
ft
in2
ft*
in3
ft3
Ib
atm
psi
°F
°R
BTU
ft/sec
cfm
Btu/(hr)(ft2)(°F)
Metric Equivalent
2.54 cm
0.305 m
6.45 cm2
0.0929 m2
16.4 cm3
.0283 m3
453.6 g
760.0 mmHg
51.7 mm Hg
1.8 (°C) +32
1.8 °K
252 Cal
30.48 cm/sec
0.472 I/sec
5.68 W/(n?)fC)
135
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing}
1. REPORT NO.
EPA-650/2-74-104
2.
3. RECIPIENT'S ACCESSIOf+NO.
4. T.TLE AND SUBT.TLE Reduction of Atmospheric Pollution
by the Application of Fluidized-Bed Combustion and
Regeneration of Sulfur-Containing Additives
5. REPOBT DATE
September 1974
6. PERFORMING ORGANIZATION CODE
7-AUTHOR(s)G.J.Vogel, W.M.Swift, J.F.Lenc, P. T.
Cunningham, W.I.Wilson, A. F.Panek, F.G. Teats,
and A. A. Jonke
8. PERFORMING ORGANIZATION REPORT NO.
ANL/ES-CEN-1007
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21ADB-011
11. CONTRACT/GRANT NO.
(EPA) EPA-IAG-149(D)
(OCR) IAG 14-32-0001-1543
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, ORD, NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711, and'
Office of Coal Research, Department of the Interior
Washington, DC 20240
13. TYPE OF REPORT AND PERIOD COVERED
Annual; 7/73-6/74
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT The report gjves results of using a bench-scale, fluidized-bed combustion
plant (capable of operating at 10-atm pressure) to evaluate the effects of combustor
operating variables on the following major response variables: SO2 and NO levels in
the flue gas, combustion efficiency, additive utilization, and heat-transfer coeffic-
ients. It also gives results of combustor testing, using a variety of coals: a highly
caking, high-volatile bituminous coal, a high ash subbituminous coal, and a low-
heating-value lignite. It reports material balance data for four trace elements--Hg,
Pb, Be, and F--and for Na concentrations in the particulate matter entrained in the
flue gas from the combustor. Kinetics of the reaction of half-calcined dolomite with
SO2 was found to be first order with respect to the SO2 concentration in the presence
of H2O vapor and approximately three-fourths order in the absence of H2O vapor.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Fluidized-Bed
Processing
Sulfur Oxides
Nitrogen Oxides
Dolomite (Rock)
Combustion
Coal
Regeneration (Engin
eering)
Additives
Stoichiometry
Kinetics
Transfer
Air Pollution Control
Stationary Sources
Fluidized-Bed Com-
bustion
13B, 21D
13H, 07A
07B
07D
08G, 20K
21R, 20M
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
136
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
136
-------
ENVIRONMENTAL PROTECTION AGENCY
Technical Publications Branch
Office of Administration
R«March Triangle Park. N.C. 27711
POSTAGE AND f£ES PAID
ENVIRONMENTAL PROTECTION AGENC»
EPA - 335
OFFICIAL BUSINESS
AN EQUAL OPPORTUNITY EMPLOYER
Return this sheet if you do NOT wish 10 receive this material Q,
or if change of address is needed [~~1. (Indicate change, including
ZIP code.)
PUBLICATION NO. EPA-650/2-74-104
------- |