EPA-650/2-74-057
June 1974
Environmental Protection Technology Series
-^eo ST^
*
$
55
V""1^
^> r^
^ PR0^G
ss
o
ol
O
^r
^;i;?^;^!!^!;^§^i;^^§i|!^
^^^^s^^^iiiSii^HiS
-------
EPA-650/2-74-057
REDUCTION
OF ATMOSPHERIC POLLUTION
BY THE APPLICATION
OF FLUIDIZED-BED COMBUSTION
by
G. J. Vogel, M. Haas, W. Swift,
J. Riha, C. B. Schoffstoll, J. Hepperly, and A. A. Jonke
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
Interagency Agreement No. EPA-IAG-0199CD)
ROAPNo. 21ADB-011
Program Element No. 1AB013
EPA Project Officer: D.B.Henschel
Control Systems Laboratory
National Environmental Research Cente*
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D. C. 20460
June 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.
11
-------
TABLE OF CONTENTS
Page
ABSTRACT 9
I. SUMMARY 10
II. INTRODUCTION 22
III. BENCH-SCALE, FLUIDIZED-BED, COMBUSTION-REGENERATION STUDIES. 23
A. Materials 23
1. Coal 23
2. Additive 24
3. Kerosene 25
B. Equipment 25
1. Pressurized, Fluidized-Bed Combustor 25
2. Fluidizing-Gas Supply System and 6-in.-dia Preheater 30
3. Fluidized-Bed Regenerator System 30
4. Analytical Equipment 33
C. Coal Combustion Experiments 35
1. Comparison of S0£ Retention in Old and New
Combustion Units 35
2. High-Pressure Operation 38
D. Regeneration of CaSOi* 41
1. Two-Step Regeneration Experiments 43
2. One-Step Regeneration of CaSO^, Batch Operation. . . 47
3. One-Step Regeneration of CaSO^, Continuous
Operation 53
4. One-Step Regeneration of Sulfated Dolomite,
Continuous Operation 56
5. Regeneration by the Reaction, CaS04 + CO -> CaC03
+ S02 59
6. Kerosene Combustion Products 62
IV. LABORATORY-SCALE REGENERATION STUDIES 63
A. Introduction and Review of Previous Work 63
B. Equipment 64
C. Materials. 64
D. Results and Discussion 64
1. Two-Step Process 64
2. One-Step Reductive Decomposition Process, Low
Temperature Studies 86
-------
TABLE OF CONTENTS (Cont'd.)
Page
3. Reaction of CaSO^ and C02/H2 at Temperatures of
1400°F and Less 88
4. Removal of H2S from Off-Gas by Reaction with
Potassium Permanganate ..... 90
V. TRACE-ELEMENT DISTRIBUTION STUDIES
A. Introduction
91
91
B. Applicability of Selected Methods for Analyzing Trace
Elements of Interest
1. Analysis of Coal, Limestone, and Coal Combustion
Residues by X-ray Spectrometric Analysis ...... 93
2. Application of Nondispersive X-ray (NDXR)
Spectrometry to Coal and Coal Combustion
Residues ...................... 94
C. Flue-Gas Sampling .................... 94
1. Description of Sampling Location .......... 95
2. Particulate Sampling ................ 97
3. Treatment of Particulate Samples .......... 100
4. Determination of Mercury in Particulate and Gaseous
Emissions ......................
5. Determination of Inorganic Fluorides in Gaseous and
Particulate Emissions ................
VI. ACKNOWLEDGMENTS ....................... 103
VII. REFERENCES .........................
APPENDIX A. English to Metric Unit Equivalents .........
APPENDIX B. Specific Analytical Procedures for Solids Samples
from Combustion and Regeneration Experiments .... 107
APPENDIX C. Operating Data and Results Obtained in Batch
Regeneration Experiments in the 1- and 2-inch-
Diameter Reactor ..................
APPENDIX D. Thermodynamic Considerations Relating to the
Conversion of Calcium Sulfate to Calcium
Carbonate ...................... ^27
-------
LIST OF FIGURES
No. Title Page
1. Pressurized Combustion-Regeneration Equipment Schematic . . 26
2. Six-in.-Dia, Pressurized, Fluidized-Bed Combustor 27
3. Lower Section of Combustor 28
4. Additive Hopper, Feeder, and Weigh Scale 29
5. Six-in.-Ola, Fluidizing-Gas Preheater 31
6. Three-in.-Dia, Fluidized-Bed Regenerator 32
7. Combustion and Regeneration Pilot Plants 34
8. Gas Sampling and Analysis System 35
9. Comparison of Sulfur Retention in Atmospheric Combustion
Experiments Conducted in Old Atmospheric and New
Pressurized Combustors 40
10. Conditions and Results for Experiment REDUC-8 51
11. Conditions and Results for Experiment REDUC-9 52
12. Kerosene Input and CHij Concentration in Process Off-Gas
for Experiment REDUC-10 54
13. Conditions and Results for Experiment REDUC-10 55
14. Sulfur Distribution in Regeneration Experiment REDUC-12 . . 58
15. Two-in.-Dia, Fluidized-Bed Reactor System 65
16. H2S Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-22 67
17. H2S Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-23 69
18. H2S Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-24 71
19. H2S Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-26 72
-------
LIST OF FIGURES (Cont'd.)
No. Title Page
20. Conceptual Two-Stage, Fluidized-Bed, Regeneration Reactor. 76
21. Infrared Spectrum of Material Taken from Cycle Two of
the Cyclic Sulfation- Regeneration Experiment: A. Material
from Sulfation Step; B. Material from Reduction Step;
C. Material from Regeneration Step ............ 79
22. Infrared Spectrum of Material Taken from Cycle Two of
Experiment CATS-22: A. Material after Reaction with HzS;
B. Material from Regeneration Step ............ 80
23. Electron Microprobe Scan of Particles Taken from the
Second-Cycle Sulfation Step of the Sulfation- Regeneration
Experiment ........................ 81
24. Electron Microprobe Scan of Particles Taken from the
Second-Cycle Reduction Step of the Sulfation-Regeneration
Experiment ................ ........ 82
25. Electron Microprobe Scan of Particles Taken from the
Second-Cycle Regeneration Step of the Sulfation-
Regeneration Experiment ................. 83
26. Equilibrium Concentrations of SC>2 for the Reaction
+ CO •* CaC03 + S02 ................. 87
27. Experimental Flow System for Flue-Gas Sampling ...... 96
28. Apparatus for Particulate Sampling of Flue Gas ...... 98
29. Single Impactor Stage ................ . . 99
30, Design of Impactor Collection Cups ............ 99
31. Apparatus for the Determination of Particulate and Gaseous
Emissions of Mercury in Flue Gas ............. 102
-------
LIST OF TABLES
No. Title Page
1. Chemical Characteristics of Coal .............. 23
2. Size Distribution of Coals ................. 24
3. Chemical Composition of Additives ............. 24
4. Size Distribution of As-Used Additives ........... 25
5. Chemical and Physical Characteristics of ANL Kerosene ... 25
6. Flue-Gas Analytical Equipment ............... 37
7. Operating Conditions and Results for Experiment SD-11 ... 39
8. Particle-Size Distribution of Feed and Product Materials
for Experiment ARK-3 .................... 41
9. Operating Conditions and Results for Experiment ARK-3 ... 42
lOa. Summary of Operating Conditions, Bench-Scale, Batch
Regeneration Experiments, Two-Step Process ......... 45
lOb. Summary of Operating Conditions and Results, Bench-Scale,
Batch Regeneration Experiments, Two-Step Process ...... 46
lla. Summary of Operating Conditions, Bench-Scale, Regeneration
Experiments, One-Step Process ............... 49
lib. Summary of Operating Conditions and Results, Bench-Scale
Regeneration Experiments, One-Step Process ......... 50
12. Sulfur Balance for Experiment REDUC-10 ........... 56
13. Comparative Particle-Size Distributions of Materials Fed
and Removed from Experiment REDUC-12 ............ 59
14. Chemical Composition of Solids in Exp. REDUC-12 ...... 59
15. Operating Conditions and Results for Experiment REDUC-13. . 61
16. Conditions and Results for I^S Addition and Regeneration
Steps of Experiment CATS-22 ................ 66
17. Screen Analysis of Initial and Final Bed Material and Fines
for Experiment CATS-22 ................... 68
-------
8
LIST OF TABLES (Cont'd.)
No. Title Page
18. Conditions and Results for H2S Addition and Regeneration
Steps of Experiment CATS-23 68
19. Conditions and Results for H2S Addition and Regeneration
Steps of Experiment CATS-24 70
20. Conditions and Results for H2S Addition and Regeneration
Steps of Experiment CATS-26 72
21. Process Conditions and Bed Compositions for Experiment
CATS-28 73
22. Composition of Effluent Gas Stream during Reaction of
CaSO^ with C0/C02 (Reduction Step) in Experiment CATS-27. . 77
23. Summary of X-ray Analyses of Particles from Fluidized-Bed
Experiments 84
24. Reaction Conditions and Analytical Results for Studies on
CaSO£|-H2-C02 Reaction . 89
25. Applicability of Selected Analytical Methods for Trace
Elements in Coal Combustion 92
26. Analysis of Coal and Coal Combustion Products with Non-
dispersive X-ray Spectrometer 95
27. Dimensions of Cascade Impactor Jets 100
-------
REDUCTION OF ATMOSPHERIC POLLUTION BY THE
APPLICATION OF FLUIDIZED-BED COMBUSTION
Annual Report
July 1972 - June 1973
by
G. J. Vogel, M. Haas, W. Swift, J. Riha,
C. B. Schoffstoll, J. Hepperly, and A. A. Jonke
ABSTRACT
The Argonne National Laboratory (ANL) program for
demonstrating the feasibility of fluidized-bed combustion
for possible use in power and steam plant applications is
divided into three studies: a) combustion of fossil fuels
in a pressurized combustor, b) regeneration of sulfated
additive for reuse in the combustor, and c) evaluation of
the type and level of trace-element pollutants in the
flue gas.
A fluidized-bed combustion pilot plant and a
fluidized-bed regenerator, both capable of operating
at 10-atm pressure, have been tested and operated. In
combustion of coal at 9-atm pressure and 1550°F (843°C)*,
high (>90%) retention of sulfur in the fluidized bed of
dolomite has been obtained with continuous addition of
fresh dolomite at a Ca/S ratio of 3. The NO level in
the flue gas has been low, <150 ppm. Two favored
methods for regenerating the sulfated dolomite have been
studied: a one-step reductive decomposition of the CaSOit
to CaO at VL950°F to release S02, and a two-step process
consisting of reduction of CaSOit to CaS at VL700°F
followed by reaction of CaS with H20/C02 to release
H2S. In results to date, complete release of sulfur
from the particles has not been achieved by either
regeneration method, and decrepitation of the particles
has been observed. Equipment has been constructed and
analytical methods have been selected for determining
the quantity and concentration of trace element pollutants
in solid or gaseous form in the flue gas.
*
Table of English to metric unit equivalents presented in
Appendix A.
-------
10
I. 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. The program consists of investigating
(1) the effects of operating variables in reducing pollutant
emissions when fossil fuels are combusted in a pressurized,
fluidized bed of sulfur-acceptor additive, (2) the regeneration
of this additive for reuse and the recovery of the sulfur, and
(3) the quantity and type of trace-element pollutants in the
flue gas.
In these studies, coal is completely combusted in a fluidized
bed of dolomite using an excess of oxygen. When sulfur contained
in the coal is released during combustion in the oxygen-excess
environment, CaSOt* is formed by reaction of the S02 with the calcium.
For regenerating CaSOi*, emphasis is being placed on two regeneration
methods. One is reductive decomposition of CaSO^ to CaO and S02
at ^2000°F:
CaSOi* + CO (or other reductant) -*• CaO + S02 + C02
A relatively concentrated stream of S02 is produced, which can be
processed in a sulfur recovery plant. The other method is a two-
step process in which CaSOii is reduced to CaS at ^1600°F and then
the CaS is reacted with C02/H20 at ^1100°F to release H2S for
sulfur recovery:
CaSOij + 4CO (or other reductant) ->• CaS + 4C02
CaS + H20 + C02 •»• H2S + CaC03
Bench-Scale, Fluidized-Bed, Combustion-Regeneration Studies
Materials. The coal used in the combustion experiments was a
bituminous coal from one of two sources: a Pittsburgh seam coal
from Consolidation Coal Co.'s Arkwright mine and an Illinois coal
from Peabody Coal Co.'s Christian County Mine. Coal was ground to
-14 mesh before use.
Two additives used in the studies were Tymochtee dolomite from
Ohio and No. 1359 limestone from Virginia. Additives were double
screened, using 14- and 100-mesh screens.
Kerosene and cylinder gases were obtained from ANL stock.
-------
11
Equipment. The ANL bench-scale equipment, designed for
operation to 10 atm, consists of a fluidized-bed combustor and an
additive regeneration reactor that have a common off-gas system
(cyclones, filters, gas-sampling equipment, pressure letdown valve,
and scrubber). The experimental system is thoroughly instrumented
and equipped with an automatic data-logging system.
The combustion unit consists of a 6-in.-dia pipe, 11 feet 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 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 maintained in the combustor by use of an internal overflow
pipe.
The 3-in.-dia regenerator was fabricated by casting a 2-1/2-in.
layer of refractory in an 8-in.-dia pipe. This assembly was enclosed
in a 12-in.-dia pressure shell. Cooling and heating coils extend
along the length of the 8-in.-dia pipe. As in the combustor, solids
are conveyed pneumatically to the fluidized bed and bed solids are
removed by use of an internal overflow pipe.
Coal Combustion Experiments. Construction of the 6-in.-dia
bench-scale equipment for fluidized-bed combustion was completed, and
operational testing of the plant was started at 1-atm pressure. At
this pressure, the data obtained on S02 retention and NO levels were
compared with those obtained in the previously installed atmospheric-
pressure combustor. The intent was to determine whether design
differences in the combustion zone would result in marked differences
in the S02 and NO levels in the flue gas. The principal design
differences are the following: (1) the presently installed combustor
has cooling tubes in the combustion zone whereas the previous one
did not, (2) coal and additive are injected vertically into the bed
(rather than horizontally, as in the previously installed unit),
and (3) bed material is removed continuously rather than batchwise.
During the atmospheric-pressure experiment with the new
combustor, Illinois coal and limestone No. 1359 were fed continuously
to the combustor. The operating conditions were a bed temperature of
1470°F, a Ca/S (calcium in the dolomite feed to sulfur in the coal
feed) mole ratio of 3.5, a superficial gas velocity of 2.8 ft/sec,
and 2.7 vol % 02 in the flue gas. Under these conditions, the flue-
gas S02 level was 420 ppm, corresponding to the retention of 89% of
the sulfur by the bed. This retention is similar to that obtained
in the previously installed combustor operated at the same conditions.
The flue-gas NO level of 375 ppm was within the range expected for
fluidized-bed coal-combustion experiments. Thus it was concluded
-------
12
that design differences in the two combustors did not markedly
affect the S02 and NO levels.
In subsequent experiments, system pressures were increased
gradually until the combustor was operating at design pressure.
Excellent sulfur retention (95%) in the bed and low levels of NO
(150-200 ppm) in the flue gas were demonstrated at 9-atm pressure.
In these experiments, Tymochtee dolomite and coal were fed continuous-
ly to give a Ca/S mole ratio of three. Gas velocity through the bed
was 2.5 ft/sec and the bed temperature was 1515°F.
Regeneration of CaSOu. During the past year, several regenera-
tion experiments were made using the newly completed 3-in.-dia
regeneration unit. Information regarding the operation of the unit
and the potential of the one- and two-step regeneration processes
was obtained.
The two-step regeneration process involves the reduction of
calcium sulfate to calcium sulfide with carbon monoxide in the
first step, then reacting the calcium sulfide with carbon dioxide
and steam to form CaCO$ in the second step. Both steps were
investigated in the 3-in.-dia, bench-scale regenerator. Four
reduction experiments (RED-1, -2, -3, and -4) were performed, and
with the exception of the first reduction experiment (RED-1), each
was followed by a C02/steam regeneration experiment (REGEN-2, -3,
and -4).
In a reduction run, 2.4 or 3.5 kg of calcium sulfate (Drierite)
was charged to the regenerator; the height of the settled bed was
2 and 3 feet, respectively. The bed was maintained at an operating
temperature of VL700°F by burning the carbon monoxide with oxygen.
Enough excess CO was fed to allow for the reduction of the calcium
sulfate. The pressure in the regenerator, initially near atmospheric,
was then increased to the operating pressure.
A regeneration run was begun immediately following each reduction
run without shutting down the equipment. To start the regeneration,
the temperature of the bed was dropped to ^1100°F from the reduction
process temperature of 1700°F. Carbon dioxide, diluted with nitrogen,
was passed through the bed and then steam was added to this mixture.
Experiments were conducted at pressures of 8, 30, 75, and 135
psig in this equipment-testing phase of the investigation.
Samples were taken of the final bed after the two steps (one
step only in the first experiment) had been completed and were
analyzed for sulfide and total sulfur content. In the first
experiment, no sulfide (<0.3%) was found in the sample which contained
18.0% total sulfur. The absence of reduction may have been caused by
an oxygen concentration in the bed that was too high. In RED-2 and
RED-3, oxygen concentration was carefully controlled (oxygen
-------
13
concentration in the exit gas was kept at a minimum), and sulfide
was formed in the bed. After the regeneration step in RED-REGEN-2,
37% of the sulfur was present as sulfide (Stotal = 20.2%;
S= = 7.5%) and in RED-REGEN-3, 44% (Stotal = 26.2% and S= = 11.6%).
If it is assumed that the sample of bed removed after the
regeneration step contained only (1) unreacted CaSO^, (2) CaS made
in the reduction step but not converted to CaSOtt by steam/C02, and
(3) CaC03 from the steam/C02 reaction with CaS, the final bed
of RED-REGEN-2 contained 54% CaSO^, 17% CaS, and 29% CaC03. The
CaCC-3 content suggests that a substantial portion of the CaS
reacted to form CaCOg and H2S. However, no H2S could be detected
in the reactor off-gas by the analytical instrument although a
strong odor of H2S was noted in the sample gas stream.
In RED-REGEN-3, the calculated values for the bed composition
were 62% CaS04, 26% CaS, and 12% CaC03. These data also suggest
that H2S should have been detected in the off-gas, although none was.
In RED-REGEN-4, the final bed composition was 77% CaSO^ and 23%
CaC03.
After the completion of these two-step regeneration process
studies, development studies in the 3-in.-dia unit were made on
the one-step process (CaSOi^ + CO •> CaO + S02 + C02). This process
required bed temperatures of ^1950°F. A new method (combustion of
kerosene plus CO, rather than pure CO) of reaching and maintaining
the desired bed temperature was successfully demonstrated in two
batch, reductive-decomposition experiments, REDUC-8 and -9. The
operating procedure was as follows:
1. The fluidized bed was preheated to 900°F. Kerosene was
then injected into the bed in a stream of carbon dioxide carrier
gas and combusted with oxygen to increase the bed temperature.
2. When the bed was at temperature, the oxygen concentration
in the off-gas was reduced to a value of 2%. Then carbon mon-
oxide from a cylinder supply was introduced and was combusted along
with kerosene to maintain the bed temperature and to provide the
necessary reducing gas for the sulfate decomposition. Maximum
allowable concentration of oxygen in the exit gas was 2%.
In REDUC-8, the bed temperature was maintained between
1500 and 1650°F to test the equipment, and in REDUC-9, the bed
temperature was increased until a maximum of 1950°F was reached.
In each experiment, the starting bed material was CaSO^, the
reactor pressure was ^2 atm, and the superficial gas velocity
through the bed was ^3 ft/sec.
-------
14
In both experiments, relatively large quantities of sulfur were
removed from the CaSOit bed. In REDUC-8, 77% of the sulfur was
removed in ^3-1/2 hours, which is somewhat higher than expected.
During the main part of the experiment, the CO and 02 concentrations
in the off-gas were ^12 and ^.5%, respectively. The S02
concentration in the off-gas varied considerably, the maximum
being 0.2%.
In REDUC-9, in which the bed temperature was gradually increased
from 800 to 1950°F in a 3-1/2 hour period, about 97% of the sulfur
was removed from the CaSO^. The CO and Q£ concentrations in the
exit gas were ^3 and M.%, respectively. The maximum S0£ concentration
in the exit gas during the experiment was 1.1%.
The experiments have demonstrated a satisfactory procedure for
increasing the bed temperature to the operating temperature and for
maintaining this temperature. They also demonstrated that some
release of sulfur from pure CaSO^ particles can be achieved by
the one-step regeneration technique.
An experiment (REDUC-12) was then made to regenerate, by the
one-step technique, sulfated dolomite (Tymochtee) produced in a
combustion experiment. The dolomite (6.1 wt % S) was regenerated
at 1.7-atm pressure and an average bed temperature of 1950°F. The
sulfated dolomite was fed to the fluidized bed throughout the
experiment at a rate of 11.6 Ibs/hr. Reductants for the reaction,
mainly CO, CH^, and H2, were produced by the incomplete combustion
of kerosene.(Pure CO was not added with the kerosene in this
experiment.)
The sulfur dioxide concentration in the process off-gas averaged
1.1%, with a maximum peak of 2.1%. The H2S content was M.% of the
S02 content in the process off-gas. The calculated sulfur balance
was 96%.
The particle-size distribution data for the feed material, the
primary cyclone product, and the combined final bed and overflow
products indicate that particle breakup occurred during the
regeneration step.
Partial combustion of kerosene has been the method for
obtaining the reducing gases required for most of the experiments
made in the 3-in.-dia unit. In order to characterize the nature of
the reducing gases formed, the concentration of reducing gases in
the off-gas was measured during kerosene combustion in an inert
fluidized bed at bed temperatures ranging from 1600-1900°F. The
data showed that the CO and also the CG^ concentrations in the off-
gas varied directly with the kerosene/oxygen feed ratio.
-------
15
Laboratory-Scale Regeneration Studies
Development studies on regeneration processes have continued
using the 2-in.-dia, batch, fluidized-bed reactor employed in
previous studies. Aspects of the two-step and one-step regeneration
processes have been investigated and a third scheme, a low-temper-
ature (<1400°F) reaction of CaSO^ with H2 and C02 (CaSO^ + 4H2 + C02
-»• CaCOs + 3H20 + H2S) , has been studied. Data have also been obtained
on the removal of H2S from process off-gas using a permanganate
solution.
Two-Step Process. Development studies of the two-step process
have included a) cyclic experiments in which dolomite was sulfided
with H2S and the CaS reconverted to CaCOs using C02 and steam to
determine if the absence of CaS(\ (which had been present in previous
cyclic experiments) affected the release of sulfur in the second
step, b) scouting experiments of the effect of higher regeneration
temperature on sulfur release, c) experiments using limestone
(instead of dolomite as in previous experiments) , d) effect of
presence of hydrogen during second step on sulfur release, e) use
of iron as a catalyst for reduction of CaSOi^ to CaS, f) a scouting
experiment on a concept in which the two steps of the regeneration
process are carried out at the same time in a single, staged
reactor, and g) examination of particles from some of the experiments.
Cyclic SutfidLng^Reqeneration Experiments. In previous attempts
to convert sulfated additive (CaSOi^) to "pure" additive (CaCOs) by
the two-step process, the product has had a sulfur content that was
too high for its reuse as an additive. * To determine whether
undesirable side reactions occur between CaSO^, CaS, and CaCOs
that inhibit the removal of sulfur, experiments were made in which
CaS was converted to CaCOs with the sulfate phase absent.
The experimental procedure consisted of partially calcining
dolomite No. 1337 (to form CaCOs and M8°)» reacting the CaCOs
fraction with H2S to form CaS, and converting the CaS to CaCOs by
reaction with C02/H20. In previous experiments, CaCOs ^n dolomite
had been reacted with S02 and the resulting CaSO^ reduced partially
to CaS before the C02/H20 reaction.
In experiment CATS-22, a bed of dolomite No. 1337 was half-
calcined at 1500° F with 40% C02-60% N2 for 18 hours. The bed of
CaC03-MgO was then reacted at 1500°F and 2-3 atm with C02-N2-H2S
for 30-60 minutes. The bed was reacted at 1100°F and 10 atm with
an equimolar C02/H20 mixture. This cycle was repeated three times.
Sulfide concentrations in the bed after the C02/H20 regeneration
step of cycles 1, 2, and 3 were 6.5, 13.7, and 13.2 wt %, respective-
ly. Before regeneration, the sulfide concentration was unknown for
cycle 1 and was 21.1 and 16.7 wt % for cycles 2 and 3, respectively.
Both the peak concentration of H2S in the effluent gas stream during
C02/H20 regeneration and the regeneration time decreased in subsequent
cycles. Apparently, the absence of sulfate does not improve the
-------
16
removal of sulfur because the same lower release of sulfur on cycling
was noted as when sulfate was present. A sieve analysis of initial
and final bed material and of fines collected in the filter chamber
during the three cycles indicated that substantial attrition of bed
particles occurred.
In experiment CATS-23, in which two cycles were completed, the
reaction conditions were the same as in CATS-22 except that less
sulfur was added to the bed in the H£S addition steps, and H2 was
added to the C02/H20 mixture during regeneration to prevent the
possible formation of calcium polysulfides, which could interfere
with sulfur release. The sulfur concentration in the bed material
decreased from 5.8 wt % after K2S addition to 2.9 wt % after
regeneration for cycle 1 and from 6.5 wt % to 5.1 wt % in cycle 2.
Approximately half of the sulfur present at the start was released
in cycle 1 and only a fourth in cycle 2. Peak concentrations of
H2S were 16.1 and 13.9 vol % in cycles 1 and 2, respectively.
Apparently, neither decreasing the quantity of sulfur in the bed
nor adding H2 to the regenerating gas mixture helped the sulfur
removal in the second cycle.
Effect of Higher Regeneration Temperature on Sulfur Release.
To determine if a higher temperature during the second or regeneration
step would increase the removal of sulfur, experiment CATS-24 was
performed at a regeneration temperature of 1250°F. The experimental
procedure consisted of half calcining dolomite No. 1337 (calcination
step) at 1500°F, reacting the CaCOa with H2S in a mixture to form
CaS (sulfiding step), and converting the CaS to CaCOs by reaction
with C02/H20 (regeneration step).
In experiment CATS-24, 1 kg of dolomite No. 1337 was calcined
at 1500°F with 40% C02-60% N2, forming a bed of CaC03-MgO. This
material was reacted at 1500°F with C02-N2-H2S-H2 until 25% of the
CaC03 was converted to CaS. The bed of CaS, CaCOa, MgO was reacted
at 1250°F and 10 atm with 45% C02-45% H20-10% H2 at a fluidizing-
gas velocity of 1.4 ft/sec. The sulfur contents of the bed before
and after the second or regeneration step were 6.2 and 2.5 wt %,
respectively. These results represent a moderate increase over the
highest sulfur removal achieved in the first cycle of the comparable
CATS-23 experiment, in which the regeneration temperature was 1100°F.
The peak concentration of H2S in the effluent gas stream during
regeneration was 5.9 vol % (dry basis), which may be compared with
an equilibrium H2S concentration of ^8 vol % (dry basis) at 1250°F.
Regeneration of Limestone. To determine if the composition of
the additive has an effect on removal of sulfur, experiment CATS-26
was performed with limestone additive instead of dolomite additive
(previously used). This experiment consisted of two cycles.
-------
17
In the sulfiding step of each cycle of the experiment, limestone
No. 1359 was reacted at 1500° F with C02-N2-H2S-H2 until breakthrough
of H2S (H2S addition step). In the regeneration step, the bed
(now consisting of CaS and CaC03) was reacted at 1250°F and 10 atm
with 453 C02-45/2 H20-103 H2 (cycle 1) and with 50% C02-50% H20 (cycle
2). The sulfur contents of the bed before and after regeneration
in cycle 1 were 6.7 wt % and 5.2 wt %. During the regeneration step
of cycle 2, the sulfur content of the bed stayed constant within
analytical error at 10.8 wt %.
The peak concentrations of H2S in the effluent gas stream
during regeneration were 5.3 vol % (dry basis) in cycle 1 and
5.8 vol % (dry basis) in cycle 2. Sulfur removal was less in the
cycles than that achieved with dolomite at identical experimental
conditions.
Effect of Presence of Hydrogen During Second Step on Sulfur
Release. The existence of possible intermediates such as poly-
sulfides may interfere with the regeneration of partially sulfated
additives. The presence of hydrogen during both stages of the
regeneration process might prevent polysulfide formation. A bed
of partially sulfated dolomite (15.4 wt % S) was reduced at 1600°F
with H2 as reductant for 5.5 hours. Conversion of CaSOit to CaS
was essentially complete. The bed now containing ^19% sulfide
was reacted with 41.5% C02-41.5% H20-17% H2 at 1200-1400°F for
52 minutes. The peak concentration of H2S in the effluent gas
stream at 1200°F was much less than the equilibrium concentration
expected at that temperature. At 1400 °F the equilibrium
concentration (1.5%, dry basis) of H2S was reached. The sulfide
concentration was 16.8%, indicating that the presence of hydrogen
does not markedly increase the conversion of CaS to
Reduction of CaSO^ Using CO and a Catalyst. The reduction of
calcium sulfate to calcium sulfide with CO as reductant is catalyzed
by the addition of Fe203 to the bed; a conversion of ^7% at 1250°F
after 45 minutes has been reported.2 In an attempt to duplicate
these results, a bed of partially sulfated dolomite was mixed
by tumbling with 10% by weight of Fe203 powder and reacted with an
equimolar mixture of CO-C02-N2 at 1250°F for 4.5 hours. The
conversion of CaSOij to CaS was 16.3% which is similar to results
obtained with 100% H2 at 1350°F.1 The Fe203 did not effectively
catalyze the conversion of CaSOi+ to CaS in dolomite at these
conditions ..
Tap-Stage Reactor Concept. A modified regeneration process
concept is discussed that utilizes a single-vessel, two-stage
fluidized-bed reactor utilizing the two-step process to produce an
off-gas containing H2S and S02 at a mole ratio of two; this mixture
would be suitable as feed to a Claus plant. The flow of solids and
gas would be cocurrent in the reactor. The lower or reduction stage
-------
18
would be operated at 1800°F with a CO/C02 gas mixture. In this stage,
S(>2 would be produced. Steam would be injected into the upper stage,
and H2S would be produced by reaction of CaS with H20/C02-
Examinat-lon of Bed Part-ides from SuT,fati.on-ReqeneTation
Experiments in Fluidized-Bed Reactors. Samples taken from several
experiments performed in fluidized beds have been examined by optical
microscopy, infrared spectroscopy, X-ray diffraction, and electron
microprobe techniques as part of an effort to clarify the mechanisms
associated with sulfation, reduction or sulfiding, and regeneration..
The results were obtained on samples taken from the combustion,
reduction, and regeneration steps of cycle two in a six-cycle,
sulfation-regeneration experiment1 and from CATS-22, a cyclic
calcination-sulfiding-regeneration experiment. The six-cycle tests
and CATS-22 were laboratory-scale experiments using dolomite 1337.
Material obtained from the second sulfation step in the six-
cycle sulfation-reduction-regeneration experiment contained
considerable sulfate, as indicated by X-ray and infrared spectrum
analyses. The infrared spectrum also indicated the presence of some
residual carbonate (003) in this material and the X-ray data
indicated the presence of some CaS. Careful examination of the
electron microprobe scan made on a similar material (taken from the
first sulfation step) showed that, in general, calcium and sulfur
occurred together throughout a cross section of the particle at
fairly uniform levels, whereas magnesium apparently was segregated
from the calcium and sulfur. This observation was confirmed by
X-ray data and indicated that sulfation took place throughout the
particle and was not limited to a surface reaction.
The presence of iron on the particle surface, inferred from
microscopic examination of the particles reported earlier,1 was
confirmed by the electron probe, which also showed some silicon
and calcium on the surface of particles from the six-cycle test.
The mineral form of this surface layer has not been identified.
Microscopic examination of this surface layer on a hot stage
indicated that its melting point was slightly less than 1100°C
(2000°F).
In the infrared spectrum of material taken from the second
reduction step of the sulfation-reduction-regeneration experiment,
the intensity of the sulfate bands was much lower than after the
sulfation step, as would be expected if calcium sulfate was indeed
reduced to calcium sulfide, but the reaction was clearly not complete.
The X-ray data indicated the same result. From the electron micro-
probe scan, it was not clear whether or not the conversion from
sulfate to sulfide had been limited to a surface layer.
-------
19
The most striking feature of the infrared spectrum and the X-ray
data for material taken from the second regeneration step in the six-
cycle sulfation-reduction experiment was the complete absence of
sulfate. The carbonate feature in the infrared spectrum of this
material was more prominent than in the previously discussed, samples,
and calcium carbonate could be detected in the X-ray pattern. The
electron microprobe scan for material from the second regeneration
step showed that calcium and sulfur appeared together uniformly
across the particle, and that the iron-rich layer was still evident.
Because all sulfate had disappeared at this stage of the process,
the iron-rich layer apparently was not a barrier to reaction within
the bulk of the particle.
Examination of the samples taken from CATS-22 showed that
conversion of calcium carbonate to calcium sulfide by reaction with
hydrogen sulfide did occur (although not all carbonate was converted
in the sample examined); however, the regeneration reaction caused
only partial conversion of sulfide to carbonate. Optical microscopy
did not reveal a surface layer on any sample particles; electron
microprobe scans were not made on the particles.
The results summarized above do not appear to explain directly
the difficulties encountered in the regeneration process. It seems
likely that an iron-containing layer forms on the particles during
the sulfation process (in a bed where combustion is occurring and
with sulfate or S0£ involved in the layer-forming reaction), but
that the layer is not directly responsible for the problems
encountered in subsequent steps. The presence of sulfate after
the reduction step and inability to effectively regenerate calcium
sulfide in the CATS experiments suggest that the thermodynamics of
the desired reactions may not be as favorable as previously believed,
and that existing thermodynamic data should be re-examined to determine
whether the present regeneration scheme is optimum.
One-Step Reductive Decomposition Process at Reduced Temperature.
Data obtained using the 3-in.-dia regenerator had shown that high
sulfur removals could be obtained even at temperatures as low as
1600°F. Previous work on the one-step process has been based on the
understanding that temperatures above 1900°F were required. A
program was initiated using the smaller 2-in.-dia unit to determine
the conditions that have an effect on the removal of sulfur in a
reduced—temperature, one-step process and to identify the vaporous
sulfur species in the off-gas. A bed consisting of CaS04 (Drierite)
was reacted at a bed temperature of 1200-1650°F with a gas composed
of mixtures of CO and one or more of the following gases: C02, H20,
02, and N2. The data show that:
1. Regeneration appeared to be most effective when the bed
temperature and C02 concentration were such that the
calcium product was CaCOs rather than CaO.
-------
20
2. Generally, high S02 concentrations were observed when the 02
concentration was high.
3. The maximum concentrations of S02 and maximum sulfur removals
were obtained with an inlet gas mixture consisting of 27%
CO, 19% C02, 6% 02, and the remainder nitrogen. The
maximum 862 concentration was approximately equal to the
equilibrium concentration expected for the reaction:
CaSOit + CO -*• CaCOs + S02
4. Elemental sulfur was present in the off-gas, but the
quantities detected were small.
5. Carbonyl sulfide was present in the off-gas. The concentration
was small, except when high concentrations of CO (>30%) were
used. The presence of carbonyl sulfide appeared to be
dependent on the presence of sulfur dioxide.
Reaction of CaSOu and CO?/H? at Temperatures of 1400°F and Less.
If the following reaction
CaSO^ + 4H2 + C02 '-* CaC03 + 3H20 + H2S
proceeds at low temperatures, <1400°F, the concentration of H2S in
the off-gas is reasonably high. Equilibrium constants at 1000,
1200,and 1400°F are 4.37 x 10 , 3.15 x 10 , and 1.51 x 107, respectively.
However, conversion of CaSOj^ to CaS (CaSOi+ + 4H2 -*• 4H20) must be
avoided because the equilibrium concentration of H2S for the reaction
of CaS with C02/H20 is low, less than 1500 ppm at 1400°F. A bed of
CaSOi, (Drierite) was reacted with 85% H2-15% C02 at 10 atm and 1000,
1200, and 1400°F for 75, 135, and 120 minutes, respectively. The
results indicate that:
1. The peak concentration of H2S in the off-gas increases with
temperature.
2. The sulfide content of the bed increases with temperature
indicating that reduction of CaSO^ to CaS is taking place, rather
than conversion to CaC03.
3. At 1400°F, the peak concentration of H2S (1160 ppm, dry
basis) was nearly equal to the equilibrium concentration of H2S
(1460 ppm, dry basis) expected for the reaction between CaS and C02/
H20, again indicating that the competing reaction in which CaS is
formed is dominating.
-------
21
Removal of H?S from Off-Gas by Reaction with Potassium Permangan-
ate. When CaS is converted to CaC03 by the reaction of CaS with
C02/H20, H2S will be formed. To avoid releasing H2S into the lab-
oratory exhaust system, a proposed scheme to remove the H2S from
the effluent gas stream is to pass the off-gas from the regenerator
through a scrubber system containing potassium permanganate solution.
The reaction expected to take place is
8KMn04 + 3H2S -*• 8Mn02 + 3K2S04 + 2KOH + 2H20
To determine if the above reaction occurs, as well as the contact
time necessary for reaction, a laboratory-scale experiment was
performed by passing a C02/H2S gas mixture through a bubbler
containing potassium permanganate solution and monitoring the
effluent gas stream for H2S with a quadrupole mass spectrometer.
The H2S was readily reacted. As a result of these tests, KMn(\
was used as the scrubber solution through which the regenerator
off-gas was passed during an experiment.
Trace-ElementJEnH ssion Studies
Studies have been initiated on the quantity and type of trace-
element emissions from fluidized-bed combustors. Emissions from
fluidized-bed combustors may be different than those from conventional
combustors, owing to differences in operating variables such as
combustion temperature and coal particle size. Comparisons with
available data from conventional plants will be made.
In the bench-scale combustion equipment, a sampling system has
been installed for collecting particulate and gaseous trace elements
in the flue gas. Methods of analysis for the primary elements of
interest (mercury, beryllium, lead, and fluorine) have been selected
and analysis of test specimens has begun. A Brink impactor has
been incorporated into the equipment so that the quantity and
distribution of trace metals in particulate solids smaller than three
microns can be determined.
-------
22
II. INTRODUCTION
An experimental program is being carried out at Argonne National
Laboratory (ANL) to develop advanced technology in pressurized,
fluidized-bed combustion of coal and to investigate the effects of
operating variables on combustion efficiency, sulfur retention
efficiency, and NOX and trace-element levels in the flue gas. Methods
for the regeneration of the sulfated additive are being studied so
that the additive can be reused in the combustor and the sulfur
values recovered.
The concept of pressurized, fluidized-bed coal combustion entails
the burning of coal at elevated pressure in a bed of additive
(commonly, limestone or dolomite) that sorbs the sulfur released
in the combustion. The capacity of the additive for sulfur retention
is reduced by sulfation and it must be removed and replaced with
fresh additive. However, it might be economically feasible to
regenerate the additive for reuse in the combustor. The low tempera-
ture of combustion, as compared with higher temperatures in
pulverized-coal boilers, may cause differences in the quantity and
type of~NOX and trace-element pollutants leaving via the flue gas;
hence, studies of both NOX and trace-element emissions are also being
made.
-------
23
III. BENCH-SCALE, FLUIDIZED-BED, COMBUSTION-REGENERATION STUDIES
A. Materials
1. Coal
Two coals have been combusted in the 6-in.-dia fluidized-
bed combustor: (1) an Illinois coal, obtained from the Peabody Coal
Co., Christian County, Illinois, which has been designated Shipment
No. 4; and (2) a bituminous coal obtained from Consolidation Coal
Co., Arkwright Mine. The chemical characteristics and the particle-
size distribution of these coals are listed in Tables 1 and 2,
respectively.
Table 1. Chemical Characteristics of Coals
Illinois Coal, Seam 6, Mine 10, Peabody Coal Co.,
Christian County, 111., Shipment No. 4
Arkwright Coal, Consolidation Coal Co., Arkwright Mine
Proximate Analysis (wt %)
As Received
Moisture
Volatile Matter
Fixed Carbon
Ash
Sulfur
Heating Value,
Btu/lb
Illinois
10.12
37.90
41.12
10.85
3.70
10,956
Arkwright
2.89
38.51
50.92
7.68
2.82
13,706
Dry
Illinois
42.17
45.75
12.08
4.14
12,163
Basis
Arkwright
39.66
52.43
7.91
2.90
14 ? 114
Ultimate Analysis (wt
Illinois Arkwright
Carbon
Hydrogen
Sulfur
Nitrogen
Chlorine
Ash
Oxygen (no difference)
66.96
4.84
4.35
1.33
13.50
9.02
77.14
5.23
2.90
1.66
0.19
7.91
4.97
-------
24
Table 2. Size Distribution of Coals
Illinois Coal. Seam 6, Mine 10, Peabody Coal Co.,
Christian County, 111., Shipment No. 4
Arkwright Coal, Consolidation Coal Co..
U.S. Sieve No.
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 + 170
- 170
Sieve Analysis
Arkwright Coal
8.0
14.2
12.3
24.7
17.9
23.0
Arkwright Mine
(wt %)
Illinois Coal
0.1
14.5
17.4
12.0
20.2
13.6
2-2.2
2. Additive
A limestone (identified as No. 1359), obtained from the
M.J. Grove Lime Co., Stephen City, Va., and a dolomite (Tymochtee),
obtained from C. E. Duff and Sons, Huntsville, Ohio, are characterized
in Tables 3 and 4. In operational testing of the regenerator,
commercial CaSO^ (Drierite) was used as a substitute for sulfated
additive. Particle-size distribution for this latter material is
presented in Table 4.
Table 3. Chemical Composition of Additives
Limestone No. 1359, M. J. Grove Lime Co., Stephen City, Va.
Tymochtee Dolomite, C. E. Duff aiid Sons, Huntsville, Ohio
Chemical Analysis (wt %)
Component
Ca
Mg
C02
H20
Derived
CaC03
MgC03
Limestone No.
37.90
0.27
NAa
1.56
94.80
0.95
1359 Tymochtee Dolomite
20.0 .
11.3
38.5
0.2
50.0
39.1
Available.
-------
25
Table 4. Size Distribution of As-Used Additives
Limestone No. 1359, M. J. Grove Co., Stephen City, Va.
Tymochtee Dolomite, C. E. Duff and Sons, Huntsville, Ohio
Commercial CaSOu (Drierite)
U.S. Sieve No.
-7 + 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 + 170
- 170
Sieve
No. 1359
Limestone
12.0
23.2
13.7
26.8
8.1
16.3
Analysis (wt %)
Tymochtee
Dolomite
0.1
26.5
21.4
12.2
18.3
9.8
11.8
CaSOu
70.5
26.4
1.1
2.2
3. Kerosene
Kerosene obtained from ANL stock was used in regeneration
experiments as an energy source and, by its incomplete combustion,
as a source of reductant gases (CO, H£, and CH^). The physical and
chemical characteristics of ANL kerosene are given in Table 5.
Table 5. Chemical and Physical Characteristics
of ANL Kerosene
Flash Point, Pensky-Martens, °F 119
Specific Gravity at 60/60°F 0.8086
Pounds per gal at 60°F 6.732
Specific Gravity, °API 43.5
Heating Value, Btu/lb 19,649
Btu per gal at 60°F 132,280
Sulfur, wt % 0.09
B. Equipment
Equipment has been installed for combusting coal at pressures
up to 10 atm absolute and for regenerating sulfated lime for reuse.
A simplified equipment schematic is shown in Fig. 1.
1. Pressurized, Fluidized-Bed Combustor
The pressurized, fluidized-bed combustor, shown schematically
-------
26
HIGH
REHEATER
PRESSURE
STEAM
COMPRESSOR
SURGE
TANK~
HOUSE
AIR
r»2-
02-
C02 !
CO
H,
-PREHEATER CYCLONES
COMBUSTOR
JI
FILTERS
TO GAS
ANALYSIS
REGENERATOR
TO
STACK
PRESSURED
LET-DOWN
VALVE
SCRUBBERS
ROOM
VENTILATION
AIR
Fig. 1. Pressurized Combustion-Regeneration Equipment Schematic
in Fig. 2, consists of a 6-in.-dia schedule 40 pipe (Type 316 SS),
approximately 11 ft long, contained within a 12-in.-dia schedule
10 pipe (Type 304 SS) over nearly the entire length. A bellows
expansion joint is incorporated into the outer shell to accommodate
differential thermal expansion between the pipes. The 6-in. pipe
is alternately wrapped with 3/8-in.-dia, 3000-watt, resistance-
type tubular heating elements and air-water cooling coils onto which
has been applied a layer of heat-conducting copper and stainless
steel spray. Sixteen heaters are attached to the unit, eight of
which are connected to Variacs and temperature indicator-controllers.
The controllers are activated by spring-loaded, coiled thermocouple
assemblies which pass through the outer jacket and contact the heated
pipe. Internal hairpin-shaped cooling coils of 3/8-in., schedule
40 pipe (Type 310 SS) extend down into the interior of the six-inch
vessel from the flanged top to provide additional heat-transfer area.
A water-air mixture is supplied to the internal and external cooling
coils. Coolant flow is adjusted on the basis of the temperatures of
the fluidized bed and the reactor wall.
A bubble-type gas distributor is flanged to the bottom of
the inner vessel. Fluidizing air, thermocouples, solids feed lines,
and solids take-off lines extend through the gas distributor (Fig. 3).
Coal and additive are pneumatically fed to the combustor by
two lO-in.-dia, periphery seal, rotary valve feeders (Fig. 4), both of
which are mounted on platform-type scales to indicate weight changes.
-------
27
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
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
SOLIDS FEED LINES
FLUIDIZING AIR
Fig. 2. Six-in.-Dia, Pressurized, Fluidized-Bed Combustor
-------
28
Fig. 3. Lower Section of Combustor
-------
29
Fig. A. Additive Hopper, Feeder, and Weigh Scale
-------
30
A constant bed height is maintained in the combust or by use of either
a 36-in. or 48-in. high overflow standpipe. Most of the fly ash and
additive that is elutriated from the bed is removed from the flue
gas by two cyclone separators in series . Essentially all of the
remaining fines are removed by two Pall Trinity filters (in series),
the first containing high temperature, epoxy-impregnated fiberglass
elements, the second containing sintered metal elements. Both
filters remove ^95 wt % of +5 ym particle-size material. The flue
gas is sampled downstream of the filters. An aliquot of the total
flow is partially dried by passage through a water condenser and
refrigerator and distributed to the gas analytical system where it is
analyzed for S02, NO, C02, CO, CHij, and 02-
2. Fluidizing-Gas Supply System and 6-in.-dia Preheater
Fluidizing air for the combustor is supplied by a 75-hp,
Gardner Denver, Electra-Screw compressor capable of delivering 100
cfm at 150 psig. The flow is metered by an orifice plate and flow
recorder-controller assembly. The fluidizing air can be heated
to approximately 1000°F by a 6-in.-dia preheater, shown schematically
in Fig. 5. The gas preheater was designed in accordance with ASME
code requirements and its design rating is 150 psig at 1500 °F.
Heat is supplied by eight 2700-watt, clam-shell-type heaters
contained within a 4-l/2-in.-dia center pipe.
3. Fluidized-Bed Regenerator System
A 3-in.-dia fluidi zed-bed regenerator (shown schematically
in Fig. 6), capable of operation at 150 psig and 2000°F, was
constructed by casting a 2-1/2-in. -thick lining of Plibrico castable
refractory in an 8-in.-dia, schedule 40, Type 316 stainless steel
pipe. This entire assembly was enclosed by a pressure shell made
of 12-in.-dia, schedule 20 carbon steel pipe. A bubble-cap-type
gas distributor is connected to the bottom of the inner vessel via
a slip fit, and held in place by retaining screws. Thermocouples,
solids overflow lines, kerosene feed line, and bed drain line pass
through the gas distributor and then through packing glands on the
bottom flange of the outer vessel. The solids inlet line passes
through the top flange of the unit and terminates above the bed.
A nitrogen purge is supplied to the annular space between the inner
and outer pipes and is maintained at a slightly higher pressure than
the reaction zone to keep reactant gases from leaving the reaction
zone.
The external wall of the inner vessel is wrapped alternately
with eight 3000-watt, 0.315-in.-dia tubular resistance heaters and
3/8-in. OD by l/16-in.-wall stainless steel tubing cooling coils
over its entire length. High-temperature <*550'C) heat-conducting
putty was applied to the heaters and cooling coils and this layer
was covered by an ^1/2-in. layer of Fiberfrax rope. The heaters
are controlled by temperature indicator-controllers and Variacs
-------
;i
HEATER LEADS
BYPASS
GAS FEED
CLAMSHELL
HEATERS
HEATER CONTROL
THERMOCOUPLES
TO REGENERATOR
TO
COMBUSTOR
Fig. 5. Six-in.-Dia, Fluidizing-Gas Preheater
-------
32
SOLIDS FEED
SPARK-PLUG IGNITER
PRESSURE TAP
THERMOWELL
HEATER
CONTROL
THERMOCOUPLES
ELECTRICAL HEATERS
LEADS
KEROSENE
INLET LINE
OFFGAS TO CYCLONE
SEPARATIONS AND
FILTERS
RUPTURE DISKS
12-irv JACKET
ELECTRICAL HEATERS
COOLING COILS
PLIBRICO CERAMIC LINER
THERMOWELL
BUBBLE CAP DISTRIBUTOR
THERMOCOUPLES
SOLIDS OVERFLOW LINE
24 inches
FLUIDIZING GAS
Fig. 6. Three-in.-Dia, Fluidized-Bed Regenerator
-------
33
actuated by thermocouples that are in spring-loaded contact with the
inner vessel wall of the regenerator. Cooling-water flow is manually
controlled.
Reactant gases are delivered from high-pressure gas
cylinders and metered by use of orifice-plate flow recorder-controller
assemblies. All gases, with the exception of oxygen and carbon mon-
oxide, are preheated in the 6-in.-dia preheater prior to distribution
through the bubble-cap distributor. Oxygen is introduced into this
heated gas stream at a point just prior to its passage through the
bottom flange of the regenerator. Carbon monoxide is introduced
directly into the bed via a separate line. Both the inner vessel
and annular space are protected by the use of rupture disks and
pressure relief valves.
Sulfated additive is pneumatically fed to the regenerator
via the solids inlet line by one of the two lO-in.-dia, periphery
seal, rotary valve feeders (see Fig. 4).
Off-gas and entrained solids from the regenerator are
processed by the cyclone-filter system common to the combustor,
which has been described above.
Fig. 7 shows an overall view of the pressurized combustion-
regeneration pilot plants.
4. Analytical Equipment
A schematic of the process-gas sampling and analysis system
associated with the fluidized-bed combustor is shown in Fig. 8.
Approximately 1% (1.0 scfm) of the total flue gas is withdrawn at
a point immediately after the secondary filter. Using system
pressure, this stream is passed through a water condenser, a sintered-
metal bayonet filter (to remove solids), and then through a
refrigerated condenser for additional water removal. Any residual
water, which could interfere with the measurement of flue-gas
constituents by an infrared technique, is removed by magnesium
perchlorate. (Approximately 3000 ppm moisture would give an infrared
analyzer reading corresponding to ^20 ppm NO. Carbon monoxide and
methane analyses would also be affected by moisture).
Continuous determinations of NO and S(>2 are carried out
using Beckman Model 315A infrared analyzers, and continuous analysis
of CHij and CO with Mine Safety Appliance (MSA) Model MRA infrared
analyzers. A Hays paramagnetic oxygen analyzer (Model 632) provides
continuous measurement of oxygen. Intermittent C02 analyses are
performed using a Hewlett-Packard Model 700 gas chromatograph.
-------
34
Pig. 7. Combustion and Regeneration Pilot Plants
-------
SECONDARY
CYCLONE
ROTAMETER
WATER
CONDENSER
PRIMARY SECONDARY
PRIMARY F|LTER
SINTERED S.S.
BAYONET FILTER
INSTRUMENT
GAS SUPPLY
MANIFOLD
REFRIGERATOR
EXHAUST
PRESSURE
CONTROL
BECKMAN 3I5A
INFRARED
ANALYZERS
(S02 AND NO)
HAYS
PARAMAGNETIC
OXYGEN ANALYZER
'MSA INFRARED
ANALYZERS
(CO AND CH4)
HEWLETT-PACKARD
GAS CHROMATOGRAPH
(C02)
Fig. 8. Gas Sampling and Analysis System
LO
Ln
-------
36
Each of these instruments is supplied with sample gas
through a pressurized manifold maintained at constant pressure.
Individual gas streams supply each instrument and are controlled by
flow valves in each line. Prior to and during each experiment, the
response of each analytical instrument is determined using standard
gas mixtures of flue-gas components in nitrogen. If required,
instrument zero and span adjustments are made. The analytical equip-
ment components are listed in Table 6.
Specific analytical procedures used on solid samples
obtained after an experiment are listed in Appendix B.
C. Coal Combustion Experiments
In addition to a series of shakedown experiments, which were
performed to gain familiarity with the newly installed pressurized
combustor and which are not reported, a coal combustion experiment
at atmospheric pressure was performed in the new unit using lime-
stone additive to compare results with those previously obtained
at'ANL in the atmospheric-pressure equipment. This was followed
by a high-pressure combustion experiment using Tymochtee dolomite
additive.
1. Comparison of SO? Retention in Old and New Combustion^Units^
The new combustor was operated at atmospheric pressure,
and the data on SC>2 retention and NO level obtained were compared
with data obtained in the previously installed atmospheric-
pressure combustor. The intent was to determine whether differently
designed combustion zones in the two combustors would contribute to
marked differences in SO2 and NO levels in the flue gas. Pertinent
differences in design are the following:
Previously Installed Newly Installed
Combustor Combustor
Coolant tubes in bed None Five hairpin tubes
occupy a maximum of
10% of the cross-
sectional area
Injection of coal and Perpendicular to the Parallel to the gas
additive into the fluidizing-gas flow flow and up into the
fluidized bed bed
Removal of sulfated Intermittent, through Continuous, through a
solids from the bed a line at the bottom vertical standpipe
of the bed whose top is 36 in.
above the gas-
distributor plate
-------
Table 6. Flue-Gas Analytical Equipment
Flue-Gas
Constituent
Sulfur Dioxide
Nitric Oxide
Oxygen
Methane
Carbon Monoxide
Carbon Dioxide
Method of
Analysis
Infrared
Infrared
Paramagnetism
Infrared
Infrared
Gas
Chromatography
Instrument
Model
Beckman 315A
Beck man 3 15 A
Hays 632
MSA LIRA 200
MSA LIBIA 300
Hewlett-
Packard 700
Output
Displayed On
Bristol, two-pen,
variable-range recorder
Bristol, two-pen,
variable-range recorder
Honeywell-Electronik 17,
2-pen, 0-10 mV range
recorder
Honeywell-Electronik 17,
2-pen, 0-10 mV range
recorder
Honeywe 1 1-B rown
Electronik,
variable range
Honeywell-Electronik 16,
variable chart speed,
reversed-drive recorder
Range
0-5000 ppm
0-10,000 ppm
0-500
0-1000 ppm
0-1 vol %
0-10 "Ol %
0-1000 ppm
0-5000 ppm
0-20 vol %
Accuracy
(% of range)
±1%
±1%
±2%
±1%
±1%
±5%
-------
38
In this experiment (SD-11), the operating procedure used in
the newly installed combustor was similar to that used in the previously
installed combustor. The partially sulfated limestone and alumina
bed was heated to a temperature at which coal would ignite. Follow-
ing ignition, coolant was supplied to the coolant tubes as necessary
to control the bed temperature. Illinois coal (see Table 1) and
No. 1359 limes tone were premixed, and the mixture was air-conveyed
to the bed. Premixing was necessary because the rates that would be
required for separate feeding were less than the minimum feed rates
of the newly installed solids feeders.
Operating conditions and steady-state flue gas concentrations
for Experiment SD-11 are presented in Table 7. The flue gas S02
level was 420 ppm, which corresponds to ^89% retention of sulfur in
the bed. This retention is similar to the sulfur retention expected
if the previously installed atmospheric combustor had been operated
at the same conditions (i,.e.t at a Ca/S mole ratio of 3.5, see Fig 9).
The flue gas NO level was within the range (300-600 ppm) expected
for coal combustion experiments conducted under similar conditions.
The conclusion is that the design differences in the combustors did
not markedly affect the results.
2. High-Pressure Operation
An experiment (ARK-3) has been performed in which Arkwright
coal was combusted in the 6-in.-dia pressurized combustor at 9.3 atm
absolute. The fluidized-bed temperature was maintained at vL515°F,
the superficial gas velocity was ^2.5 ft/sec, and Tymochtee dolomite
was fed at a Ca/S mole ratio of 3.
The as-received dolomite was double-screened prior to use
as the starting bed and feed material. All of the +14 mesh fraction
was removed to eliminate particles that might be difficult to air-
transport to the combustor and some of the -100 mesh particles were
removed because these would have elutriated rapidly from the bed.
The particle size distribution o£ the feed and product materials are
shown in Table 8.
In starting up the combustor, the bed in the combustor and
the fluidizing gas were preheated to ^850°F. Coal was then injected
into the bed and burned to bring the bed temperature up to VL500°F.
When this bed temperature had been established, the system pressure,
which had been 6 atm absolute, was increased to 9.3 atm absolute.
Table 9 lists the operating conditions and the results of
the experiment. The average S02 concentration in the flue gas was
100 ppm, indicating that over 95% of the sulfur in the coal was
retained by the additive. For comparison, S02 removals of 78 and
87% were achieved in two atmospheric pressure experiments (BC-2 and
-3) conducted in the previously installed combustor using Tymochtee
dolomite additive and an alumina starting bed.3
-------
Table 7. Operating Conditions and Results for Experiment SD-11
Equipment:
Coal:
Additive:
Starting Bed:
ANL 6-in.-dia, fluidized-bed, pressurized combustor
Illinois Seam No. 6, Peabody Coal Co., 3.7 wt % S (-14 mesh)
Limestone No. 1359, as received (94.80 wt % CaC03, 0.95 wt % MgC03)
Partially sulfated and calcined limestone No. 1359 (3.2 kg) plus
30-mesh Alundum (4.8 kg)
(^24-in. starting fluidized-bed depth)
Bed
Temp
(°F)
1470
System
Pressure
(atm)
Coal
Feed Rate
(Ib/hr)
5.0
Additive
Feed Rate
(Ib/hr)
2.0
Feed
Ca/S
Mole
Ratio
3.5
Superficial
Fluidizing-Air
Velocity 02
(ft/sec) (vol I
Steady-State
Flue-Gas Concentrations
2.8
2.7
C02
(vol 7.
16.8
CO
(ppm)
250
NO
(ppm)
375
S02
(ppm)
420
S02
Retention
89
vo
-------
40
100
90
80
70
*- 60
c
o
c
-------
41
Table 8. Particle-Size Distribution of Feed and
Product Materials for Experiment ARK-3
Sieve Analysis (wt %)
U.S. Sieve No.
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 + 170
- 170
Tymochtee
Dolomite,
As Fed
0.1
26.5
21.4
12.2
18.3
9.8
11.8
Arkwright
Coal,
As Fed
8.0
14.2
12.3
24.7
17.9
23.0
Sulfated
Bed Overflow
and Final Bed
0.5
46.0
32.4
15.3
5.8
Avg. Particle
Size, urn
531
323
772
Operating conditions for these latter experiments consisted of a
bed temperature of 1600°F, a gas velocity of 3 ft/sec, a bed depth
of 2 ft, and a nominal Ca/S mole ratio of 1.5. The NO concentration
in the flue gas during ARK-3 averaged 170 ppm, which is roughly
half of that found in the atmospheric-pressure combustion experiments.
Approximately 90 pounds of sulfated additive was obtained.
D. Regeneration of CaSO^
When coal is burned in a fluidized bed containing limestone or
dolomite additive, the product of the reaction between sulfur-contain-
ing gases and additive is calcium sulfate if combustion is carried
out under oxidizing conditions or calcium sulfide if combustion is
carried out with a deficiency of air. As the capacity of the bed
material to retain sulfur is approached, bed material must be removed
from the combustor and replaced with fresh material. Simply discarding
the sulfur-bearing material would be unattractive because of (1) the
cost of replacement material, (2) the cost of transporting fresh
material to the furnace and used materials from the furnace, (3) dis-
posal costs, and (4) pollution problems associated with the disposal
of used material (especially if it contained calcium sulfide, which
would slowly release hydrogen sulfide if exposed to air). A process
for regenerating used bed material so that the lime may be recycled
through the combustor is needed, both to circumvent these problems
and to recover the sulfur value as sulfuric acid or elemental sulfur.
Several regeneration processes have been considered. If the
product of the reaction between sulfur-containing gases and additive
in the combustor is calcium sulfate, the processes are
-------
to
Table 9. Operating Conditions and Results for Experiment ARK-3
Equipment:
Coal:
Additive:
Starting Bed:
ANL 6-in.-dia, fluidized-bed, pressurized combustor
Arkwright, Consolidation Coal Co., 2.82 wt % S (-14 mesh)
Tymochtee Dolomite, C. E. Duff and Sons, Huntsville, Ohio
(50.0 wt % CaC03, 39.1 wt % MgC03)
6.5 kg Tymochtee dolomite (-14 mesh)
Ov»15-in. starting, fluidized-bed height)
Fluidizing-Gas
Velocity: ^2.5 ft/sec
Bed Temp
(°F)
System
Pressure
(atm)
Coal Feed
Rate
(Ib/hr)
Feed
Additive Ca/S
Feed Rate Mole
(Ib/hr) Ratio
Average Flue-Gas
02
(vol %)
C02
(vol %)
CO
(ppm)
Concentrations
NO
(ppm)
CHi,
(ppm)
S02
(ppm)
1515
9.3
22.4
14.0
3.4
3.7
18.0
249
108
103
100
-------
43
1. Thermal decomposition of calcium sulfate to form CaO and 862 .
2. Reductive decomposition of calcium sulfate to form CaO and
S02.
3. Reductive decomposition of calcium sulfate to form CaS,
followed by process 1 or 2 below.
If the product of the reaction is calcium sulfide, the processes
are
1. Roasting of calcium sulfide in air or oxygen to form CaO and
S02.
2. Reaction of calcium sulfide with steam and carbon dioxide to
form CaCOs and H2S.
The sulfur dioxide or hydrogen sulfide formed would then be
converted to sulfur in a Claus plant or to sulfuric acid in a conven-
tional acid plant to recover the sulfur value.
During the past year, the bench-scale reactor (3- in. dia) has been
used to perform preliminary investigations of potential regeneration
methods. It has also been used in the initial stages of cyclic
combustion-regeneration experiments to determine the effects of recycle
under different operating conditions on the performance of the additive.
1. Two-Step Regeneration Experiments
A two-step regeneration process was investigated which
involved the reduction of calcium sulfate to calcium sulfide with
carbon monoxide in the first step and subsequent reaction of the
calcium sulfide with carbon dioxide and steam to form CaCOs ^n tne
second, or actual regeneration, step . The chemical equations for the
two reactions are as follows:
1/4 CaSdt + CO •*• 1/4 CaS + C02 (1)
CaS + C02 + H20 •*• CaC03 + H2S (2)
Both steps were investigated in the 3-in.-dia, bench-scale regenerator.
Four reduction experiments (RED-1, -2, -3, and -4) were performed,
and with the exception of the first reduction experiment, each was
followed by a C02/H20 regeneration experiment (REGEN-2, -3, and -4)
using the reduced material from the preceeding RED experiment. The
experiments were carried out during the initial testing of the new
equipment at which time the primary objective was to test the operation
of the regenerator at pressures up to 10 atm absolute.
Drierite (CaSO^) was the starting-bed material in these batch
experiments. Gases for reduction and regeneration were taken from
-------
44
either high-pressure cylinders or building supply sources. Arkwright
coal was burned in the bed during Experiment RED-4 to increase the
bed temperature.
In the reduction run, 2.4 or 3.5 kg of calcium sulfate
(Drierite) was charged to the regenerator, and the height of the
settled bed was 2 or 3 ft, respectively. Heated fluidizing air or
nitrogen was passed through the bed to increase the temperature to
approximately 1000°F. In RED-1, -2, and -3, a mixture of 5% C2H2 -
2% 02 in nitrogen was then fed to the reactor to increase the bed
temperature from 1000 to 1700°F. In RED-4, Arkwright coal was
injected and combusted in the bed in place of the acetylene. In all
four experiments, carbon monoxide was then substituted for the hydro-
carbon and the bed was maintained at ^1700°F by burning the carbon
monoxide with oxygen. Enough excess CO was fed to allow for the
reduction of the calcium sulfate. The pressure in the regenerator,
initially near atmospheric, was then increased to the operating
pressure. Run time for the four RED experiments was 3 to 5 hours.
To start a REGEN run, the CO combustion was stopped by
shutting off the CO and 02 feeds and the temperature of the bed was
allowed to drop to ^IIOOT from the reduction process temperature of
1700°F. Carbon dioxide, diluted with nitrogen, was passed through
the bed, and then steam was added to this mixture. Bed temperature
was maintained at 'vlOOO0? for 1 hr during the subsequent regeneration
step.
Experiments were conducted at pressures of 8, 30, 75, and
135 psig in this equipment-testing phase of the investigation. The
operating conditions and the outlet-gas concentrations for the
experiments are presented in Tables lOa and lOb.
Samples were taken of the final bed after the two steps had
been completed and were analyzed for sulfide and total sulfur content.
No sulfide (<0.3%) was found in the sample from RED-1, which contained
18.0% total sulfur. This may have been caused by the relatively
high oxygen concentration in the bed.
In RED-1, as well as the three succeeding RED experiments,
some sulfur was released from CaSOij as S02 by the one-step reductive
decomposition reaction:
CaSOit + CO -»• CaO + S02 + C02 (3)
Bed temperatures were high enough at times for this to occur. In RED-2
and RED-3, oxygen concentration was carefully controlled (oxygen
concentration in the exit gas was kept at a minimum), and sulfide
was formed in the bed. After RED-REGEN-2, 37% of the sulfur was
present as sulfide (Stotai = 20.2% and S= = 7.5%) and in the third
experiment, 44% of the sulfur was present as sulfide (S^ota! = 26.2%
-------
Table lOa. Summary of Operating Conditions, Bench-Scale, Batch Regeneration Experiments,
Two-Step Process
Experiment
RED-1
RED-2
REGEN-2
RED- 3
REGEN-3
RED-4
REGEN-4
Starting Bed
Material
2.4 kg CaS04
(-7+25 mesh)
2,4 kg CaSOi,
(-7+25 mesh)
Final Bed
from RED-2
3.5 kg CaSOi,
(-7+25 mesh)
Final Bed
from RED- 3
2.4 kg CaSOi,
(-7+25 mesh)
Final Bed
from RED-4
Feed
Material
None
None
None
None
None
None
None
Temp
(°F)
1700
1560-
1800
1060-
1140
1630-
1890
1000-
1090
1310-
1950
900-
930
System
Pressure
(psia)
22.7
28.7
36.7
75
75
150
127
Gas
Velocity
(ft/sec)
-x.3.3
%3.5
Vl.3
^3.0
-v-2.3
Varied
•vO.9
02
(scfm)
0.45-
0.67
0.28-
0.48
0.45-
1.1
Varied
Input Flow Rates
CO H20 C02
(scfm) (scfm) (scfm)
1.33
0.56-
1.8
0.7 2.2
2.8-
3.3
0.52 5.5-
6.4
1.9-
2.5
0.63 2.2
N2
(scfm)
1.6
2.4-
2.8
5.9-
8.2
5.0
5.3
5.3
-------
Table lOb. Summary of Operating Conditions and Results, Bench-Scale, Batch Regeneration Experiments,
Two-Step Process
Experiment
RED-1
RED-2
REGEN-2
RED- 3
REGEN-3
RED-4
RE GEN- 4
02
(vol %)
0.6-
0.9
0.1-
0.15
0.1-
0.15
0.3-
0.7
0-
0.15
0-
0.05
0.05
C02
(vol %)
30
22. 5a
33a
18-
23
18.2-
46.6
<10
ND
Outlet-Gas
CO
(vol %)
2-13
16.3-
>26
1.8-
2.6
10-
19.6
0.01-
0.11
2.4-
18
ND
Concentrations
S02
(ppm)
0-
730
375-
2730
325-
430
420-
2250
25-
80
25-
65
80
NO
(ppm)
0-
710
35-
45
0-5
20-
33
10-
30
20
ND
Analysis of
Final Bed
Solids or Overflow
Sulfur Total
CH4 Removal S Sulfur
(ppm) (%) (wt %) (wt %) Purpose
75 ^29 <0.3 18 A
/
300- NDb ND ND
400
250- %31 7.5 20.2
300
\
150- ND ND ND Two-Step
200 Process
Studies
20- >vl3 11.6 26.2
50
200- ND ND ND
300
200 ND ND ND \
(
a.
One determination.
Not determined.
-------
47
and S= = 11.6%).
Analysis showed that the starting CaSOit bed material in these
experiments was relatively pure. Based upon Ca and S analyses, the
final bed from RED-REGEN-2 contained 54% CaSO^, 17% CaS, and 29% CaC03,
if it is assumed that the sample of bed contained only (1) unreacted
CaSOij, (2) CaS made in the reduction step, but not converted to CaSOt,
by steam/C02, and (3) CaC03 from the steam/C02 reaction with CaS. The
CaC03 content suggests that a substantial portion of the CaS was
reacted to form CaC03 and H2S in the second step. However, no H2S
could be detected in the reactor off-gas using the quadrupole mass
spectrometer analysis method, even though a strong odor of H2S was
noted in the sample gas stream. The instrument readings during this
and the other experiments are therefore considered suspect.
In RED-REGEN-3, the calculated values for the bed composition
were62% CaSOi^, 26% CaS, and 12% CaC03. These data also suggest that
H2S should have been detected in the off-gas during REGEN-3, but none
was detected.
In RED-REGEN-4, the final bed composition was calculated to be
77% CaSOit and 23% CaC03. As in previous experiments, H2S should have
been detected during REGEN-4, but again, none was detected.
From the data obtained in these experiments, it was concluded
that the experimental system will be operable at design conditions.
Certain modifications were made to the equipment, however, in an
attempt to provide better temperature control and reduce the oxygen
concentrations in the gas phase to very low levels. The modifications
include (1) adding a spark ignition source* inside the regenerator
unit and substituting kerosene for acetylene and (2) relocating the
oxygen inlet line so that, in the reduction step, the CO and 02 mix
before they enter the bed. In the above investigation, CO was added
directly to the bed and the oxygen added to the gas stream upstream
from the preheater.
2. One-Step Regeneration of CaSOu, Batch Operation
After the completion of the two-step regeneration studies, the
emphasis on regeneration methods shifted to the one-step process, which
requires higher temperatures than the reduction step of the two-step
process. The one-step process is based on the following chemical
reaction:
CaSOij + CO + CaO + S02 + C02
Two experiments involving the batch reduction of calcium sulfate
(Drierite) were completed to test a procedure for bringing the
fluidized bed to process temperature and maintaining that temperature
by combusting kerosene and CO.
-------
48
In the first experiment, REDUC-8, which was a test experiment,
a 1650°F bed temperature was maintained rather than a higher temperature
required for reductive decomposition. In the second experiment, REDUC-9,
a temperature of 1950°F (required for the one-step process)was attained.
The operating procedure consisted of the following steps:
1. The fluidized bed was heated to 900°F. Kerosene was then
injected into the bed in a stream of C02 carrier gas and combusted in
the preheated, fluidizing-air stream to increase the bed temperature to
1600°F.
2. When the bed was at temperature, the oxygen flowrate was
reduced until its concentration in the off-gas was less than 2%. CO
from cylinder sources was then added. Bed temperature was then adjusted
to the temperature required in the experiment.
The starting-bed material was CaSO^ (Drier! f-e) in each of the
two batch experiments conducted at 2 atm absolute pressure. No
CaSOtt was added to the bed during the experiments and kerosene
combustion was continued throughout the expetiments.
In REDUC-9 the CO concentration was purposely kept low (as was
the C0/C02 ratio) to maximize the S02 concentration in the off-gas.
Operating data for these experiments are shown in Tables lla and lib.
The bed temperature, the inlet gas and kerosene flowrates, and
the measured outlet-gas concentrations in REDUC-8 and REDUC-9 are
plotted in Figs. 10 and 11, respectively. Superficial gas velocities
in the regenerator averaged 2.5 and 3.3 ft/sec for REDUC-8 and -9,
respectively.
Analyses of samples of material removed from the final bed of
each experiment showed that sulfur removal was high. The weight of
sulfur in the CaSOtf starting bed for REDUC-8 was 535 g. Of this
weight, 67 g (12.5%) of the sulfur remained in the final bed.
Calculated composition of the final bed was 12.2% CaS, 5.2% CaSO^,
82.6% CaC03 (Stotal " 5.40%; S= = 4.19%). Thus, 87.5% of the initial
sulfur in the CaSOij was removed in about 3-1/2 hr at a 1500-1600°F
bed temperature. The average S02 concentration in the off-gas during
this period was 0.1%. Only 42 g (7.8%) of the initial sulfur left
the system as S02. The remaining 426 g (79.7%) of the initial sulfur
was unaccounted for. This latter material might have left the system
in the form of COS, H2S, or elemental sulfur in the off-gas. The
concentrations of these off-gas constituents were not determined. The
high level of sulfur removal that resulted at the relatively low
operating temperature was unexpected.
The weight of sulfur in the CaSO^ starting bed for REDUC-9
was also 535 g. Of this weight, 5.7 g (1.1%) of the sulfur remained
in the final bed. Calculated composition of the final bed was 0.3%
-------
Table Ha. Summary of Operating Conditions, Bench-Scale Regeneration Experiments,
One-Step Process
Feed System
Experiment Type
REDUC-8 Batch
REDUC-9 Batch
REDUC-10 Cont.
REDUC-12 Cont.
REDUC-13 Cont.
Starting Bed Feed
Material Material
CaSO^ None
2.27 kg
CaSOi, None
2.27 kg
CaS04 Same
2.27 kg
(-7+25 mesh)
Sulfated Same
Tymochtee
Dolomite
2.72 kg
(6.1% S)
CaS04 Same
2.40 kg
(-7+25 mesh)
Rate Temp
(Ih/hr) (°F)
1300-
1650
850-
1950
4.0 1720-
1920
11.6 1950
6.0 1724
1500
1455
Pressure
(psia)
29.4
24.7
29.7
25
41
Gas
Velocity
(ft /sec)
^2.3
•x.3.3
5-6
3.0
2.4
2.7
2.4
02
(scfm)
0.2-
1.1
0.3-
0.6
0.6
1.7
0.6
0.4
0.4
Input
Kerosene
(cc/min)
6.8-
22.0
0-
22.0
23
31.6
23.3
29.3
23.3
Flow Rates
CO
(scfm)
0-
1.4
0-
0.1
0-
1.4
C02
(scfm)
0.7-
2.9
0.5-
1.6
0.1-
0.3
0.1
3.9
5.0
5.0
N2
(scfm)
1.4-
2.1
1.4-
1.6
2.2-
3.5
2.1
1.0
1.0
0.8
VO
-------
Table lib. Summary of Operating Conditions and Results, Bench-Scale Regeneration Experiments,
One-Step Process
Outlet-Gas Concentrations
Experiment Type
REDUC-8 Batch
REDUC-9 Batch
REDUC-10 Cont.
REDUC-12 Cont.
REDUC-13 Cont.
02
(vol Z)
0-
11.5
0-
3.2
0.02-
4.5
1.3
1.8
1.5
1.4
C02
(vol %)
41-
51
47tt
7.2-
19.3
28.4
82
86
86
CO
(vol %)
2.5-
15.0
1.5-
6.0
1.0-
6.8
4.1
5.2
4.3
4.3
S02
(ppm)
100-
1900
0-
10000
0-
42400
11000
1400
1500
1100
NO
(ppm)
30-
440
50-
785
25-
820
166
274
610
620
CHi,
(ppm)
40-
9900
2000-
10000
0-
6800
7900
14000
20000
20000
Approx.
Solids
Sulfur
Removal
(%)
77
97
85
59
12
Analysis of
Final Bed
or Overflow
Total
S Sulfur
(wt %) (wt %) Purpose Comments
4.2 5.4 .— |
0.1 0.6
JTest of 1-Step
Process
Conditions
2.5 7.5 One-Step Continuous
Process Feeding Test
Studies
0.06 3.9
(Overflow)
0.07 21.5
(Overflow)
Regeneration
Step, Cycle 1
of Cyclic Study
Effect of High
C02 Cone.
V
One determination.
-------
51
3.2
2.4
O)
2
§ 1.6
C3
0.8
Time, hr
Fig. 10. Conditions and Results for Experiment REDUC-8
(Pressure: 2 atm absolute)
-------
52
u
o o
o >
CM •
O <"
O O
01
oo
c
•r-
E
c o
01 u
O
S.
-------
53
CaS, 2.0% CaSOit, and 97.7% CaCQ3* (Stotai - 0.62%; S= - 0.14%). Thus,
98.9% of the initial sulfur in the CaSO^ was removed in about 3-1/2 hr
at a 1600-1900 °F bed temperature. In this experiment, the bed temper-
ature was higher but the CO concentration was lower than in REDUC-8.
The average 862 concentration in the off-gas was 0.24%. Only 75 g
(14.0%) of the initial sulfur left the system as S02. The remaining
454 g (84.9%) of the initial sulfur was unaccounted for. The form of
this latter material is unknown. Data from the quadrupole mass-analysis
unit showed a slight blip where the H2S peak would be expected, but the
H2S concentration could not be measured because the peak was too small.
No COS peak was noted. It is possible that elemental sulfur vapor was
produced and condensed in the off-gas stream.
3. One-Step Regeneration of CaSOq, Continuous Operation
A single one-step regeneration experiment (REDUC-10) was made
in which the CaSO^ was fed continuously to the regenerator. In this
experiment, the starting bed material and the solids fed to the
regenerator bed during the experiment were CaSO^ (Drierite) .
In the startup procedure, the fluidizing gas was preheated
and the solids in the f luidized bed were heated by electrical heaters
until a bed temperature was reached sufficient to combust kerosene
(MJOO °F) . By burning kerosene, the bed temperature was then increased
to the experimental operating temperature -of M.900°F. After about one
hour (with the bed at 1900°F), CaSO^ was fed into the bed at a rate
of 4 Ib/hr. Pressure in the system was 15 psig, and the fluidizing-
gas velocity was 5-6 ft/sec. The bed temperature, the inlet gas
and kerosene flow rates, and the concentrations of the gases exiting
from the bed^are shown in Figs. 12 and 13 (see also Tables lla and
lib).
Both S02 and I^S were present in the process off-gas. The
S02 concentration generally ranged between 1 and 4% while CaSOi^ was
being fed. The maximum 50% concentration recorded was ^6%. The
concentration was not measured continuously, as was the S02
CaCOs assumed rather than CaO because the bed was cooled in a C02
gas stream after the experiment.
The concentration of the off-gas constituents exiting the bed are
calculated values. In regeneration experiments with continuous
additive feed, gas leaving the fluidized bed is diluted with additive
transport gas and the concentrations of the off-gas constituents
are measured after the combined gases leave the regenerator. The
concentration of a particular off-gas constituent above the bed is
then calculated by multiplying its concentrations in the diluted
stream by the ratio of the flow rates in the diluted and undiluted
streams.
-------
54
O) O
tO U
e
27
26
25
24
23
ex
Q.
Ol
CO
O)
O
JQ
O
c
O
O
O
O
7000 I—
6000 -
5000
4000
3000
2000 -
1000 -
Time, hr
Fig. 12. Kerosene Input and CHi* Concentration in
Process Off-Gas for Experiment REDUC-10
-------
55
o
VI
to
CJ
I
•U
3.2
2.4
r 1.6
« 0.8
CaS04 Feed to Bed Started
Time, hr
Fig. 13. Conditions and Results for Experiment REDUC-10
(Pressure: 2 atm absolute)
-------
56
concentration. The total I^S content of the process off-gas was
obtained by analyzing the quantity of CuS residue in a copper sulfate
solution through which a proportionate part of the process off-gas
was passed. Based on this analysis, the sulfur, as I^S, was approxi-
mately 20% of the sulfur found in the process off-gas.
The sulfur material balance, presented in Table 12, shows that
a little over 60% of the sulfur was accounted for. One source of the
unaccounted sulfur is the very fine particulate matter in the off-gas
that was not collected in the cyclone separators. Other unrecovered
sulfur might have been in the form of elemental sulfur that deposited
in process lines by the H£S-S02 reaction. Thorough cleanup of the
off-gas piping system would recover this sulfur.
Table 12. Sulfur Balance for Experiment REDUC-10
IN; Sulfur, kg
CaSOi*, Starting Bed
CaSOij, Fed during Experiment
1.54
Final Bed
Cyclone Solids
Off-Gas: As S02
As H2S
TOTAL OUT: 0.95
DEFICIT: 0.59 or 38%
The solids removed from the regenerator after the experiment
contained 73% CaO, 21% CaSOtt, and 6% CaS. No product (i.e., no over-
flow from the bed) was collected in the product container during
the experiment. A considerable quantity of solids (about 20 wt % of
the starting bed plus fed solids) was collected in the cyclone receivers.
These solids consisted of about 80 wt % CaS04 and 20 wt % CaO. Lower
(less than 5-6 ft/sec) fluidizing-gas velocities would result in less
of the solids being elutriated from the bed.
4. One-Step Regeneration of Sulfated Dolomite. Continuous Operation
An experiment (REDUC-12) was completed in which the sulfated
Tymochtee dolomite (6.1 wt % S) produced in combustion experiment
AKK-3 was used as the feed. The regeneration was carried out at 1.7
atm absolute pressure and an average bed temperature of 1950°F.
-------
57
Sulfated dolomite was fed to the fluidized bed throughout the experi-
ment at a rate of 11.6 Ibs/hr, and the resultant solids overflow from
the bed was collected by use of the internal 24-in.-high standpipe.
Reductants for the reaction, mainly CO, CH^, and H2> were produced
entirely by the incomplete combustion of kerosene within the bed.
No separate addition of CO was made as had been done in past experiments.
Kerosene also was the energy source for elevating the bed temperature
from the 700°F preheat temperature to the operating temperature of
1950°F.
Operating conditions and results for this experiment are shown
in Tables lla and lib. The oxygen flow to the bed was controlled so
that the oxygen concentration in the off-gas was maintained at less
than 1.5%, which resulted in the presence of reducing gases in the ^
off-gas. The CO and CH^ concentrations in the off-gas above the bed
averaged 4.1 and 0.8%, respectively. Hydrogen concentration was not
measured. The S02 concentration in the process off-gas above the bed
averaged 1.1%, with a maximum peak of 2.1%. The H2S content was VL%
of the S02 content in the process off-gas. The sulfur material
balance, Fig. 14, has been calculated to be 96.1%.
One objective of cyclic combustion-regeneration experiments is
to determine the extent of particle breakup (and elutriation from the
fluidized bed) during regeneration. Breakup could result from
attrition occurring in the fluidized bed or from chemical conversion
of the compounds. During regeneration, calcium carbonate is decar-
bonated and CaSOit is converted to CaO (or CaS).
The particle-size distribution data (shown in Table 13) for
the sulfated Tymochtee dolomite feed material, the primary cyclone
product, and the combined final bed and overflow products indicate
that decrepitation occurred during the regeneration step. The feed
material, with a mean particle diameter of 772 ym, had less than 6%
of -45 mesh solids. Material elutriated from the fluidized bed and
collected in the cyclone had a lower mean particle diameter of 292 ym
and considerably more of -45 mesh solids (58.5% of the total).
The distribution data for the primary-cyclone solids show that
about 40% of the solids collected in the cyclone were in the -25 to +45
mesh range. Particles in this size ranee would not be expected to
*
The concentration of the off-gas constituents exiting the bed are
calculated values. In regeneration experiments with continuous
additive feed, gas leaving the fluidized bed is diluted with additive
transport gas and the concentrations of the off-gas constituents
are measured after the combined gases leave the regenerator. The
concentration of a particular off-gas constituent above the bed is
then calculated by multiplying its concentrations in the diluted
stream by the ratio of the flow rates in the diluted and undiluted
streams.
-------
58
S02 in Off-Gas 524 g
Final Beds
37 g SULFUR (1.6%)
2,259 g SOLID
H2S in Off-Gas £ g
_ To Filters
Negngible Solids Collected
I
Secondary Cyclone
Negligible Solids Collected
Primary Cyclone 1.252 g SULFUR (6.9%)
18,252 g SOLID
Bed Overflow 391 g SULFUR (3.9%)
10,080 g SOLID
Y
Starting Beds +
Solids Fed
2.289 g SULFUR (6.1%)
37,531 g SOLID
Sulfur in Kerosene ]Q g
SULFUR BALANCE:
In 2,299 g
Accounted For 2.210 g
BALANCE -89 g
Fig. 14. Sulfur Distribution in Regeneration Experiment REDUC-12
elutriate at the 3.0 ft/sec gas velocity in the bed. However, the gas
velocity is higher than 3.0 ft/sec in the upper zone of the regenerator.
This higher gas velocity results from the fact that the gas which
pneumatically conveys the additive solids to the regenerator (through
a line that enters the top of the regenerator and terminates about a
foot above the bed) and the fluidizing-gas streams meet in the upper
zone of the regenerator and exit together. The gas velocity of the
combined streams above this meeting point may be as high as 6 ft/sec.
Solids that may slug from the bed into the upper zone of the regenerator
could be carried out of the regenerator by the higher gas velocity. A
better system, either terminating the carrier gas line at the top of the
regenerator or removing the feed solids in the carrier gas by a
cyclone external to the regenerator and then feeding, the solids to
the bed via a dipleg, is being considered.
-------
59
Table 13. Comparative Particle-Size Distributions of Materials
Fed and Removed from Experiment REDUC-12
U. S. Sieve
Sizes
+ 14
-14 + 25
-25 + 35
-35 + 45
-45 + 80
-80 + 170
- 170
Starting Bed and
Feed Material
(wt %)
0.54
45.77
32.65
15.22
5.82
0.00
0.00
Primary
Cyclone
(wt %)
0.00
0.00
10.22
31.56
28.44
8.87
20.92
Composite
Final Bed and
Overflow
(wt %)
0.65
33.23
41.79
16.99
6.22
0.84
0.29
Mean Particle Dia
(urn)
772
292
706
Because of weight changes that occur in the solids through the
chemical reactions, the percentage of feed solids elutriated from the
regenerator cannot be determined from weight data for solids input
and output. However, calcium material balance data indicate that
about 55% of the solids fed to the regenerator were elutriated from the
regenerator in this experiment.
The chemical composition of the feed to the regenerator, the
product solids removed from the regenerator, and the solids removed
from the first cyclone are shown in Table 14. Very little sulfide
(0.06 wt %) was found in the regenerator product. The regenerator
feed and product streams contained 25.9 and 13.3 wt % CaSOtf, respec-
tively.
Table 14. Chemical Composition of Solids in Exp. REDUC-12
Feed to Regenerator
Product from Regenerator
First Cyclone
CaSOif
(wt %)
25.9
13.3
27.6
CaS
(wt %)
0.06
1.20
CaO
(wt %)
32.7
25.9
5. Regeneration by the Reaction, CaSOu + CO •*• CaCOg + S02
The one-step reductive regeneration process is generally con-
ducted at a high temperature (1900-2000°F) to reduce the sulfated
additive to CaO and liberate S02» However, the results of experiment
REDUC-8 led to consideration of conducting the reductive regeneration
-------
60
process at lower temperatures so as to yield CaCO$ (rather than CaO)
and S02. One potential advantage of a lower-temperature one-step
reductive regeneration process is less rapid deactivation of the sorbent
.than that which occurs in the higher- temperature one-step process.
The CaSOjf resulting from the sulfation of dolomite or limestone
in the comb us tor could, if the kinetics are favorable, be regenerated
by the reaction
CaSOit + CO ^ CaCOa + S02
Equilibrium S02 concentrations are relatively high at low temperatures.
For example, with a 5% CO concentration and a temperature of 1700°F,
the equilibrium S02 concentration is M>.5%. The ^4% peak concentrations
at 1600-1700°F obtained in the 2-in.-dia laboratory-scale regeneration
experiments described below may have resulted because conditions were
suitable for this reaction to occur (see data in Appendix C) .
A competing reaction
1/4 CaSOij + CO -»• 1/4 CaS + C02
must be avoided by operating at a high C02/C0 ratio and a low temper-
ature and with 02 present to reconvert CaS to CaSO^. To avoid
decomposing CaC03, the partial pressure of C02 must be 0.8 to 1.8 atm
at 1600°F and 1700°F, respectively.
Operating conditions and results for a single experiment
(REDUC-13) completed in the 3-in.-dia regenerator are presented in
Table 15 (see also Tables lla and lib). The experiment was made at a
system pressure of 2.8 atm absolute. A C02 partial pressure of 2.4
atm was maintained, sufficient to keep the carbonate from decomposing
at a temperature of ^1730 °F. The CaSO^ (Drierite) was fed continuously
into the bed at a rate of 6 Ib/hr. The starting bed was five pounds
of CaSO^. The temperature of the fluidized bed varied during the
experiment and the resulting data ate only partially indicative of the •
results that might be obtained if steady state had been achieved.
A stable bed temperature of 1650°F was desired but lower temperatures
mainly prevailed.
The measured S02 concentrations in the process off-gas were
considerably less than the equilibrium concentrations. Concentrations
of S0£ never exceeded 0.2% during the experiment, whereas equilibrium
concentrations at the process conditions range from 3 to 6% S02.
About 15% of the sulfur content in the off-gas was H2S. Very little
sulfur as sulf ide was present in samples removed from the product
(which continuously overflowed from the bed) and from the final bed.
The final bed contained 20.5 wt % total sulfur and 0.25 wt % sulf ide
and the overflow product contained 21.5 wt % total sulfur and 0.07 wt 7,
sulf ide. Only ^70% of the sulfur was accounted for and VL2% total
regeneration of the CaSO^ was achieved.
-------
Table 15. Operating Conditions and Results for Experiment REDUC-13
Equipment: ANL 3-in.-dia, fluidized-bed regenerator
Feed Material: CaSOi, (Drierite)
Feed Rate: 6 Ib/hr
Starting Bed: CaSOi,. (Drierite)
System Pressure: 2.8 atm absolute
Bed Temp
(°F)
1724
1500
1455
Gas
Velocity
(ft /sec)
2.4
2.7
2.4
Input Flow Ratesa
02
(cfm)
0.6
0.4
0.4
Kerosene
(cc/min)
23.3
29.3
23.3
N2
(cfm)
1.0
1.0
0.8
C02
(cfm)
3.9
5.0
5.0
02
(vol %)
1.8
1.5
1.4
Off-Gas Concentrations
CO
(vol I)
5.2
4.3
4.3
NO
(ppm)
274
610
620
CH4
(vol %)
1.4
2.0
2.0
Avg. S02
Cone.
(vol %)
0.14
0.15
0.11
3A11 gas flows at 70°F and 1 atm.
-------
62
6. Kerosene Combustion Products
In the final two regeneration experiments conducted in the
3-in.-dia regenerator, the partial combustion of kerosene was the sole
source of the reducing gases used for the reduction of CaSOj*. To
obtain data on the species and concentrations of gases at different
temperatures and kerosene/oxygen ratios, kerosene was partially
combusted in a fluidized bed of alumina (A^Os). Nitrogen at 2.5
cfm (70°F, 1 atm) was the diluent gas. Gas velocity ranged from 2.1
to 2.8 ft/sec, and the system pressure was ^25 psla. Bed temperatures
ranged from 16QO°F to 1900°F and kerosene/oxygen ratios from 16 to 27
cc kerosene/ft oxygen.
Data were obtained during 8 steady-state periods and showed
that the CO and CR^ concentrations in the off-gas varied directly
with the kerosene/oxygen ratios. Analysis of components other than
CO and CHtt in an off-gas sample, taken when the temperature of the
fluidized bed was 1890°F and the kerosene/oxygen ratio was 25.9 cc
kerosene/cfm 02,showed that
a) the ethylene concentration was 0.15%,
b) the ethane concentration was less than 0.05%,
c) the hydrogen concentration was *\>9%.
An additional and similar experiment will be made to obtain
more data so that CHi^ and CO concentrations in the off-gas can be
correlated with dependent variables of bed temperature and kerosene/
oxygen ratio.
-------
63
IV. LABORATORY-SCALE REGENERATION STUDIES
A. Introduction and Review of Previous Work
Data have been obtained in the 2-in.-dia, batch, fluidized-bed
reactor on two regeneration processes under consideration. These
processes are as follows:
1. Reductive decomposition of calcium sulfate.
2. Reduction of calcium sulfate to calcium sulfide followed
by reaction of calcium sulfide with carbon dioxide and
steam.
A preliminary investigation has also been made of a third scheme
in which CaSOi^ is reacted with H2 and C02 at a temperature below
1400°F. In any practical application of these processes, the
sulfur-containing gas produced in the reaction would be treated
further to make elemental sulfur.
During the previous year, the thermodynamics of the above
processes were studied.1 The yields of sulfur-containing gases and
the compositions of solid phases were calculated. The variations
of these yields and compositions with temperature, pressure, and gas
composition for a system at equilibrium were also calculated.
Also, in previous studies of the reductive decomposition process,
experiments were performed in a static system to test the accuracy
of the equilibrium compositions calculated for the reduction of
CaSOif with CO/C02 mixtures. In most cases, good agreement was
obtained between experimental and calculated values for S02 levels
when CaSO^ was reacted with CO/C02 over a range of ratios.
Previous developments in the two-step process have shown that:
1. Complete reduction of CaSO^ to CaS can be achieved at
1600°F with H2 or CO as reductant.
2. Equilibrium concentrations of H2S can be obtained when CaS
is reacted with C02/H20 gas mixtures at 1000-1200°F, but
the sulfur removal is poor, generally 50% or less.
3. When cycling solids through several sulfation-reduction-
regeneration steps (the product of the second step was
resulfated with S02~containing gas), the peak concentrations
of HgS decrease and less of the total sulfur is released
as the cycling continues.
-------
64
B. Equipment
The experiments were performed in batch, fluidized-bed reactors:
CATS-34 and CATS-35 in a l-in.-dia alumina tube reactor and all other
experiments in a 2-in.-dia stainless steel reactor. A schematic of
the 2-in. reactor system is presented in Fig. 15. Except for the
reactor, the system is the same for the l-in.-dia reactor. The
pressure of the system is controlled by a pressure transmitter,
pneumatic controller, and a pneumatic valve. Temperature in the
reactor is controlled by adjusting the amperage in the resistance
heaters around the reactor. Analysis of the effluent gas is performed
with infrared analyzers, a gas chromatograph, and a quadrupole mass
spectrometer.
C. Materials
The bed material consists of either CaSOij (Drierite) or
partially sulfated dolomite or limestone obtained from the combustion
of coal in the additive. Nitrogen and carbon dioxide are obtained
from the evaporation of the liquids or from normal cylinder sources.
The other reactant gases were obtained from normal cylinder sources.
D. Results and Discussion
1. Two-Step Process
Development studies of the two-step process have included
(1) cyclic experiments in which CaO was sulfided with H2S, and the
CaS reconverted to CaCO$ with H20/C02 to determine if the absence
of CaSOit (which had been present in previous cyclic experiments)
affected the release of sulfur, (2) scouting experiments on the
effect of higher regeneration temperature on sulfur release, the
use of limestone (instead of dolomite), and the catalytic reduction
of CaSOit by iron, (3) the assessment of two-stage, two-step regenera-
tion process, and (4) examination of particles from the cyclic
experiments.
a. Cyclic Sulfiding-Regeneration Experiments. To determine
whether the removal of sulfur is inhibited by undesirable side
reactions between the components of the system ( CaSC^, CaS, and
CaCOs ),experiments were performed to convert CaS to CaCOs with
sulfate phase absent. Data obtained previously in the cyclic experi-
ments including sulfation had shown that much sulfur was not being
released from the particle during the regeneration step and that
sorbent activity fell substantially upon cycling.
Experiment CATS-22 consisted of three cycles using
dolomite. Each cycle comprised calcination of the MgCOa fraction
of the dolomite to MgO, conversion of CaCOs to CaS with H2S, and
regeneration of CaS to CaC03 with C02/H20. Data for the inlet gas
composition, temperature, reaction time, and final bed composition
-------
SOLIDS
REMOVAL
LINE-
CONDENSER
RECEIVER
TO SAMPLING
INSTRUMENTS
PI PURGE ROTAMETER
M
T-^
REACTOR
SOLIDS BALLAST VACUUM PUMP
RECEIVER
TAYLOR PRESSURE
TRANSMITTER
FILTER
CHAMBER
PREHEATER
PNEUMATIC
CONTROLLER
PNEUMATIC VALVE
Fig. 15. Two-in.-Dia, Fluidized-Bed Reactor System
Ui
-------
66
are presented in Table 16. Data for the H2S concentration in the
effluent gas stream during regeneration are presented in Fig. 15.
The regeneration step was terminated when the H2S release appeared
to stop.
The data indicate that after the sulfiding steps
the sulfur content of the bed was essentially in the form of sulfide;
therefore, any effect of sulfate interfering with the removal of
sulfur may be neglected. After the regeneration step in cycles 1,
2, and 3, the sulfide concentrations in the bed were 6.5, 13.7, and
13.2 wt %, respectively. Less sulfur was removed as H2S in succeeding
cycles.
Table 16. Conditions and Results for H2S Addition and
Regeneration Steps of Experiment CATS-22
Initial Bed: 1 kg dolomite No. 1337
(Bed half-calcined at 1500°F with 40Z C02-60% N2 before each cycle.)
H2S Addition
Inlet-Gas3 Flow Rates
Cycle
No.
1
2
3
C02
(cfm)
0.27
0.39
0.35
N2
(cfm)
0.62
0.33
0.41
H2S
(cfm)
3.3
2.6
2.2
Temp
1500
1500
1500
Reaction
Time Pressure
(min) (atm)
60 2.8
50 2.7
32 1.9
Sulfur
in Final
Total S
(wt Z)
-
21.3
17.4
Cone.
Bed
(wt %)
-
21.1
16.7
Regeneration
Inlet-Gas3
Cycle
No.
1
2
3
Flow
C02
(cfm)
0.88
0.88
0.88
Rates
H20
(cfm)
0.88
0.88
0,88
Gas
Velocity
(ft/sec)
1.3
1.3
1.3
Approx.
Temp
1100
1100
1100
Reaction
Time Pressure
(min) (atm)
60 10
35 10
25 10
Sulfur
in Final
Total S
(wt Z)
7.2
14.3
14.1
Cone.
Bed
s-
(wt Z)
6.5
13.7
13.2
All Gas flows at 1 atm and 70 "F.
-------
67
The peak H2S concentrations in the effluent gas steam
during regeneration of CaS to CaCO$ in cycles 1, 2, and 3 were
26.5, 14, and 12.7 vol % (dry basis) (see Fig. 16). The calculated
equilibrium concentration at 1100°F is -v21 vol % (dry basis). The
duration of the regeneration reaction decreased from 58 min for
the first regeneration to 25 min for the third regeneration. These
results indicate that the amount of sulfur removal decreased with
each succeeding regeneration as it had when sulfate was present in
earlier cyclic experiments.1 It appears that the problem of sulfur
removal cannot be attributed to interference due to the presence
of CaSOit in the bed.
A particle-size analysis (Table 17) was performed
on initial and final bed material and on fines elutriated into the
filter chamber. The results indicate that substantial attrition
of bed particles occurred during the cycles.
30
in
N
10
CYCLE *l
10
30
TIME, min
80
Fig, 16. H2S Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-22
-------
68
Table 17. Screen Analysis of Initial and Final Bed
Material and Fines for Experiment CATS-22
Sieve Analysis (wt %)
U.S. Sieve No.
+ 25
-25 + 40
-40 + 60
-60 + 80
-80 +100
-100 +230
-230
Initial Bed
Dolomite No. 1337
6.7
49.3
20.2
16.6
4.9
1.8
0.4
Total wt, g
S content, wt %
Weight of S, g
Final Bed
6.0
31.3
26.6
15.6
9.6
8.3
2.6
384
14.1
53.3
Fines
0.5
0.5
0.9
4.3
58.3
35.5
222
6.2
13.8
Experiment CATS-23 was performed to determine whether
a low conversion of CaCOa to CaS and the addition of a reductant
(H2) to the regenerating gas mixture would affect the removal of
sulfur from the bed. Two cycles were performed at the operating
conditions and with the results shown in Table 18.
table 18. Conditions and Results for H2S Addition and
Regeneration Steps of Experiment CATS-23
Initial Bed: 1 kg dolomite No. 1337
(Bed half-calcined before each cycle at 1500'F with 40% C02-60% N2.)
H2S Addition
Inlet-Gas Flow
Cycle
No.
1
2
C02
(cfm)
0.19
0.19
N2
(cfm)
0.26
0.26
Rates
H2S
(cfh)
1.3
1.3
Temp
(°F)
1500
1500
Reaction
Time
(min)
40
40
Pressure
(atm)
1.3
1.3
Sulfur Cone.
in Final Bed
(wt Z)
5.8
6.5
Regeneration
Inlet-Gas3 Flow
Cycle
No.
1
2
C02
(cfm)
0.88
0.88
H20
(cfm)
0.88
0.88
Rates
H2
(cfm)
0.18
0.18
Temp
(°F)
1100
1100
Reaction
Time
(min)
78
68
Pressure
(atm)
10
10
Sulfur Cone.
in Final Bed
(wt %)
2.9
5.1
All gas flows at 1 atm and 70°F.
-------
69
The sulfur content of the bed decreased from 5.8 wt %
after the first H2S addition to 2.9 wt % after the first regeneration
step, and from 6.5 wt % to 5.1 wt % in the second cycle. The peak
H2S concentrations in the effluent gas (Fig. 17) were 16.1 vol %
and 13.9 vol % (dry basis) for cycles 1 and 2, respectively. Much
less sulfur was released in the second cycle, which has been typical
in all of the cyclic experiments, and it was concluded that neither
low conversion of CaCQ$ to CaS nor the presence of hydrogen would
significantly help the problem of sulfur release during cycling.
10
20 30 40
TIME, min
50
6O
Fig. 17. H2§ Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-23
-------
70
b. Effect of Higher Regeneration Temperatures on Sulfur
Release from the Particle. Earlier attempts to convert CaSO^ to
yielded a product that probably contained too much sulfur for
reuse as additive. In this earlier work, conversion of CaS to CaCO$
was performed at 1000 °F or 1100 °F to take advantage of high equili-
brium concentrations of H2S (37 and 20 vol %, dry basis, respectively)
in the gas phase.
To determine if reaction of CaS and C02/H20 at a higher
temperature would increase the removal of sulfur, a laboratory- scale
experiment was performed in which CaS was converted to CaC03 at
1250°F (at this temperature the equilibrium concentration of H2S
is ^8 vol %, dry basis). Gas compositions, temperatures, reaction
time, and bed compositions for this experiment (CATS-24) are presented
in Table 19.
Table 19. Conditions and Results for H2S Addition and
Regeneration Steps of Experiment CATS-24
Initial Bed: 1 kg dolomite No. 1337
(Bed half-calcined at 1500°F with 40% C02-60% N2 at 1 atra.)
H2S Addition
Inlet- Gas3
C02
(cfm)
0.24
N2
(cfm)
0.33
Flow Rates
H2S
(cfm)
0.027
H2 Temp
(cfm) (°F)
0.04 1500
Reaction
Time
(min)
40
Pressure
(atm)
2.1
Sulfur Cone, in Bed
at End of Step
(wt %)
6.16
Regeneration
Inlet-Gas a Flow
C02
(cfm)
0.88
H20
(cfm)
0.88
Rates
H2
(cfm)
0.18
Temp
CD
1250
Reaction
Time
(min)
40
Pressure
(atm)
10
Sulfur Cone, in
Final Bed
(wt %)
2.45
Gas flow rates at 70°F and 1 atm.
Data for the E^S content during the run are presented in
Fig. 18. The peak concentration of H2S in the effluent gas stream
was 5.9 vol % (dry basis). As was previously noted, the equilibrium
concentration of H2S expected at 12501>F is ^8 vol % (dry basis}.
After completion of the reaction, a sample of the bed was
obtained and analyzed for sulfur content. The data indicate that
the sulfur content of the bed decreased from 6.2 wt % at the end of
the H2S addition step to 2.5 wt % at the end of the regeneration
step. In the first cycle of an earlier experiment (CATS-23, reported
above) that was identical to CATS-24, except that the regeneration
step was performed at 1100°F instead of 1250°F, the sulfur concentrations
-------
71
1 3
1 i r
10
20 30
TIME, min
40
5O
Fig. 18. H2S Concentration in Effluent-Gas Stream
during Regeneration Step of Experiment CATS-24
in the bed before and after the regeneration step for cycle 1 were
5.8 wt % and 2.9 wt %, respectively. These data indicate that sulfur
removal can be moderately increased by increasing the regeneration
temperature to 1250 °F.
c. Regeneration of Limestone. To determine the effect of
additive composition on the removal of sulfur, a laboratory-scale
experiment (CATS- 26) was performed in which limestone instead of
dolomite was used as the additive. Inlet-gas compositions, temperatures,
reaction times, and bed compositions for the two cycles of the
experiment are presented in Table 20.
In the sulfiding step of cycle 1, breakthrough of
(i.e.3 when H2S was first detected in the off-gas) occurred after
30 min of reaction time. Conversion of 12.5% of the CaC03 to CaS
was obtained. In cycle 2, H2S breakthrough was indicated after 45 min
of reaction time, and 25% of the CaC03 remaining in the bed after
cycle 1 was converted to CaS. Peak concentrations during regeneration
of 5.3 vol % (dry basis) and 5.8 vol % (dry basis) H2S were obtained
in cycles 1 and 2, respectively (see Fig. 19). The equilibrium I^S
concentration expected at 1250°F for the reaction is ^8 vol %
(dry basis) .
In cycle 1, the sulfur content of the bed decreased from
6.7 wt % before regeneration to 5.2 wt % after the conversion. The
sulfur content, 10.8 wt %, did not change during the regeneration
step of cycle 2. In a previous experiment (CATS-24) with reaction
-------
72
Table 20. Conditions and Results for H2S Addition and
Regeneration Steps of Experiment CATS-26
Initial Bed: 1 kg limestone No. 1359
Equipment: 2-in.-dia, fluidized-bed reactor
H2S Addition
Cycle
No.
1
2
Inlet-Gasa Flow Rates
C02 N2 H2S H2 Temp
(cfm) (cfm) (cfm) (cfm) (°F)
0.23 0.28 0.03 0.027 1500
0.23 0.28 0.03 0.027 1500
Reaction
Time
(min)
30
45
Pressure
(atm)
2
2
Sulfur Cone.
Bed at End of
(wt %)
6.7
10.8
in
Step
Regeneration
Cycle
No.
1
2
Inlet-Gas3 Flow Rates
C02 H20 H2 Temp
(cfm) (cfm) (cfm) (°F)
0.26 0.26 0.18 1250
0.26 0.26 0.00 1250
Reaction
Time
(min)
30
40
Pressure
(atm)
10
10
Sulfur Cone.
Final Bed
(wt %)
5.2
10.8
in
a
Gas Flow Rates at 70°F and 1 atm.
20 30
TIME, min
Fig. 19. H2S Concentration in Effluent-Gas Stream during
Regeneration Step of Experiment CATS-26
-------
73
conditions identical to cycle 1 except that partially calcined
dolomite was the bed material, the sulfur content decreased from 6.2
wt % to 2.4 wt 7, upon regeneration in cycle 1. These data indicate
that,at these experimental conditions, the composition of the additive
has an effect on the removal of sulfur and that dolomite is a better
choice as an additive than is limestone. It should be noted that
limestone is not decarbonated (calcined) at the experimental conditions
whereas the dolomite is half calcined. Calcination probably increased
the particle surface area of the dolomite.
d. Effect of Presence of Hydrogen during Second Step on
Sulfur Release. Experiment CATS-28 was performed to determine whether
the presence of hydrogen during both steps of the process prevents
the suspected formation of intermediates such as calcium polysulfides
(which may interfere with the regeneration of sulfated dolomite).
The bed material, consisting of 0.5 kg of partially sulfated dolomite
(containing 15.4 wt % S), was reduced at 1600°F with 25 vol % H2-37.5
vol % C02-37.5 vol % N2 at 10 atm for 5.5 hr to convert the CaSOit to
CaS. Process conditions for the reduction step are presented in
Table 21. The sulfur and sulfide concentrations of the bed material
after reduction were 18.9 wt % and 19.1 wt %, respectively, indicating
that conversion of sulfate to sulfide was complete. The bed, now
consisting of CaS, CaC03, and MgO, was reacted at 1200-1400°F with 17
vol % H2-41.5 vol % C02-41.5 vol % H20 at 10 atm for 52 min. The
process conditions for the regeneration step are also presented in
Table 21.
Table 21. Process Conditions and Bed Compositions for Experiment CATS-28
Bed Material: 0.5 kg partially sulfated dolomite (15.4 wt % S)
Reduction Step
Inlet-Gas Flow Rates
Temp
Pressure
(atm)
Time
(hr)
C02
(cfm)
N2
(cfm)
H2
(cfm)
Sulfur Cone, in
Bed at End of Step
Total S S=
(wt %) (wt %)
1600
10
5.5
0.7
18.9
19.1
Regeneration Step
Inlet-Gas" Flow
Temp
(°F)
1200
1400
1200
Pressure
(atm)
10
10
10
Time
(min)
0-27
27-42
42-52
C02
(cfm)
1.3
1.3
1.3
H20
(cfm)
1.3
1.3
1.3
Rates
H2
(cfm)
0.5
0.5
0.5
Sulfur Cone, in
Final Bed
Total S S=
(wt %) (wt %)
16.3 16.8
Gas Flow Rates at 70°F and 1 atm.
-------
74
A portion of the effluent gas stream was monitored for
content. The H2S content of the effluent gas stream was ^.3 vol %
(dry basis) during the first 27 min at 1200°F. When the temperature
was increased to 1400°F, the H£S content of the effluent increased to
1.2 vol % (dry basis) in 6 min, then decreased slowly to 0.8 vol % (dry
basis) after a total of 15 min at 1400°F. The temperature was lowered
to 1200°F, and the H2S content of the effluent gas decreased to r»0.3
vol % (dry basis) and remained constant at that concentration for 10 min.
The equilibrium concentrations of H2S expected for this reaction at
1200°F and 1400°F are ^5 vol % (dry basis) and 1.5 vol % (dry basis).
These results indicate that at low temperatures, where the equilibrium
concentration of H2S is favorable, the kinetics of the reaction are not
rapid enough for equilibrium to be established in the system.
The sulfur and sulfide concentrations of the bed material
after regeneration were 16.3 and 16.8 wt %, respectively, indicating
that 12% of the sulfur had been removed. These results indicate that
the presence of hydrogen during regeneration does not yield satisfactory
sulfur removal.
e. Reduction of CaSOu Using CO and a Catalyst. It has been
reported2 that the reduction of calcium sulfate to calcium sulfide
with CO as reductant is catalyzed by the addition of Fe£03 to the bed,
and that in one experiment a conversion of 97% occurred at 1250°F after
45 min. Experiment CATS-29 was made in an effort to confirm these
results. In CATS-29, a bed of partially sulfated dolomite (9 wt % S)
was mixed by tumbling with 10% by weight of Fe£03 powder and reacted
with an equimolar mixture of CO-C02-N2 at 1250°F and 1 atm for 4.5 hours.
The conversion of CaSOi* to CaS was only 16.3% which is considerably
less than the 97% reported conversion. It appeared that the Fe203 did
not effectively catalyze the conversion of CaSO^ to CaS in the presence
of sulfated dolomite at these conditions.
f. Two-Stage Reactor Concept. Results of experiments on the
first step of the two-step regeneration process indicate that sulfur
compounds are likely to be evolved in small amounts in the reduction
step. Loss of these sulfur compounds would be undesirable. When the
reduction is carried out at 1600°F with CO, a small quantity of S02 is
evolved. When the reduction is performed with 100% H2, small amounts
of H2S are evolved.
To preclude release of these gases to the atmosphere, it
appears desirable to pass both the gas and the reduced solids co-
currently from the reduction reactor to the regeneration reactor.
Fortunately, the effluent gas from the reductor can be utilized as a
reactant in the regenerator. When CO is used as the reductant, a
reaction product is C02, which is a component of the regeneration
gas. Similarly, when H2 is used as reductant, H20, which is another
component of the regeneration gas, is produced. Reductant in moderate
excess is not likely to interfere with the subsequent process step.
-------
75
For cocurrent flow of gas and solids between the two stages,
it may be convenient to carry out the two steps in a single vessel having
two stages. Such a vessel is -illustrated in Fig. 20. Solids are fed to
the lower stage, where reduction occurs. Then gas and solids flow
upward, through the restriction, to the upper stage where regeneration
occurs. Water reactant is fed to the upper stage. Some or all of this
water could be in liquid form to help reduce the bed temperature to
the desired level. The gas velocities in the two stages would be
slightly different, but this should have no important consequences.
A further modification and simplification of the overall
regeneration, sulfur-recovery process may be possible. If the temperature
of the reduction stage is increased, say to 1800°F, greater amounts of
S02 will be present in the gas leaving the reduction stage. By proper
adjustment of the stage temperatures, it may be possible to produce a
regenerator off-gas containing a mole ratio of H2S/S02 of two. This
gas would then be cooled and fed to a Glaus reactor for direct conversion
to sulfur.
To simulate the reduction stage of the envisioned reactor,
sulfated dolomite from a combustion experiment was reacted at 1850 °F
with 2% CO in C02 for 80 min, and with 0.25% CO in C02 for 40 additional
minutes. After this treatment, the bed was reacted with an equimolar
mixture of C02/H20 at 1250°F for 17 min to simulate the second stage.
The results for the first-stage tests (Table 22) indicate that during
the reduction step, the ratio of CO to C02 in the effluent gas stream
was greater than the ratio in the inlet gas, possibly because carbon
in the bed was reacting with C02 to produce CO. As the effluent gas
CO/C02 ratio increased or moved to the CO-rich side of the coexistence
line for CaSO^ and CaS, the S02 content of the effluent gas stream was
greater than would be expected from equilibrium calculations, but was
less than would be expected at the CaS/CaSO^ coexistence line. (Pressure
of S02 in equilibrium with CO/C02 mixtures and the CaS/CaSOif coexistence
line as a function of temperature have been presented earlier. ) When
the ratio of CO to C02 was close to the coexistence line, the S02 content
of the effluent gas stream was as expected from equilibrium calculations.*
For example, at 1710°F the S02 pressure in the effluent gas stream was
0.011 atm at a C0/C02 ratio of 0.023. For comparison, the predicted S02
pressure at a C0/C02 ratio of 0.023 and 1710°F is 0.015 atm. At the
coexistence line, the S02 pressure would be M3.140 atm.
During the following step to convert CaS to CaCOa, no H2S
was found in the effluent gas stream; possibly no CaS was present after
the reaction with C0/C02. Previous successful reductions of CaSO^ to
CaS using CO as reductant were performed at 1500°F and 100% CO.1 It
is possible that in this experiment (CATS-27) the CO content was too
low to effect much reduction of CaSO^ to CaS.
-------
76
RFGFKIFRATinW QTAftF
(I200°F)
/CaS + H20 + C02\
\** CaC03 + H2S /
SULFATED <=
ADDITIVE e^r
IJ3JJJ&
$$$$
am
?%.&'0c^
w
?l£$vt&<
Q*O O^1^- i.'^S'V^1"
iS^w
1 ,
Kx
\^ SOLIDS REMOVAL
REDUCTION STAGE
(I800°F)
CGaS04 + 4CO
— CaS + 4C02
CaS04 + CO
— » CaO + COo +
-------
77
Table 22. Composition of Effluent Gas Stream during Reaction
of CaS04 with C0/C02 (Reduction Step) in Experiment CATS-27
Inlet-gas composition:
2% C0-C02, 0-80 min
0.25% C0-C02, 80-120 min
Time
(min)
10
20
30
40
50
60
70
80
90
100
110
120
a
Temp
(°F)
1850
1840
1840
1840
1840
1835
1825
1810
1730
1710
1685
1665
Effluent
Ratio
C0/C02
0.004
0.036
0.048
0.059
0.063
0.065
0.052
0.032
0.025
0.023
0.022
0.022
Gas Stream
Pressure
of S02
(atm)
0.018
0.069
0.075
0.061
0.052
0.043
0.038
0.027
0.015
0.011
0.009
0.008
Equilibrium
Pressure of S02
at Exit C0/C02
Ratio and Bed
Temperature
(atm)
•»—
0.015
0.006
0.004
0.002
0.002
0.004
0.015
0.015
0.015
0.014
0.011
Approximate
S02 Pressure
on CaS/CaSO^
Coexistence Line
(atm)
0.225
0.215
0.215
0.215
0.215
0.210
0.205
0.195
0.145
0.140
0.125
0.115
Temperature at middle of the bed.
Samples from different parts of the bed following the
second step were analyzed for sulfur content. The bed had caked near
the middle. The sulfur content in the upper third of the bed decreased
from 3.7 wt % to 0.9 wt % during the reduction and regeneration steps.
These results represent a sulfur removal which is better than the
maximum sulfur removals of ^50% obtained in previous two-step
regeneration experiments.
g. Examination of Bed Particles from Sulfation-Regeneration
Experiments in Fluidized-Bed Reactors. In earlier work, individual
particles from the 2-in.-dia regenerator unit were examined by optical
microscopy as part of an effort to clarify the mechanisms associated
with sulfation, sulfiding, and regeneration.1 In that work, a surface
layer about 20 ym thick was observed on presumably sulfated dolomite
particles in the 1-mm-dia size range.
The bed particles examined in current work were from (1) an
earlier sulfation-regeneration experiment, in the 2-in.-dia unit, that
consisted of six cycles with each cycle including sulfation and two-
stage regeneration1 and (2) sulfiding-regeneration experiments (CATS-22,
-------
78
-23, -24, and -26) in which conversion of CaS to CaC03 was studied under
various conditions. In the latter experiments, CaS was formed by
reacting CaCOs with H2S, rather than by the reduction of
The techniques used to examine bed particles were optical
microscopy, infrared spectroscopy, X-ray diffraction, and electron
microprobe analysis. Initial examination was done with an optical
microscope. Material obtained from the sulfation-regeneration experiment
had a surface layer that was darker, in color (brown-black) than the
interior of the particles and was magnetic. Particles from experiments
in which CaSO^ was not formed (i.e., sulfiding-regeneration experiments)
did not exhibit this surface layer. Starting material for the sulfation-
regeneration experiments had been obtained from a combustion experiment
carried out in the 6-in.-dia combustor. In all samples, the bulk
structure of the particles appeared to consist of very small (0.3 ym)
crystallites, as might be expected in calcined material. Because the
size of the crystallites was below the resolving power of an optical
microscope, reliable identification of the minerals and compounds
present was not possible.
Therefore, a number of samples were examined by infrared
spectroscopy to determine the presence of characteristic anions (e.g.,
carbonate and sulfate). Satisfactory results were obtained using both
Nujor mulls and KBr pellets as matrix material.
X-ray diffraction patterns were obtained for a number of
samples in order to identify the major minerals present. The X-ray
results were in agreement with the infrared results.
The electron microprobe was used to examine a limited
number of typical particles in order to establish the distribution of
elements in the particles. Results of these examinations confirmed the
presence of the surface layer that had been observed optically.
The IR spectroscopy, X-ray diffraction, and electron micro-
probe results were obtained on samples taken from the sulfation,
reduction, and regeneration steps of cycle two in the sulfation-
regeneration cyclic experiment, and from Experiment CATS-22. The infrared
spectra are presented in Figs. 21 and 22, the microprobe data in
Figs. 23 to 25, and the X-ray data in Table 23.
Material obtained, from the second cycle sulfation step of
the sulfation-regeneration experiment can be clearly seen to contain
considerable sulfate (S0=) by examination of Fig. 21a and Table 23.
The infrared spectrum in Fig. 21a also indicates the presence of some
residual carbonate (COf ) . The carbonate, however, was not detected
by X-ray examination (Table 23). On the other hand, the X-ray data
indicate some calcium sulfide (CaS) that was not identified by infrared
(Figs. 21 and 22). Careful examination of the electron microprobe scan
made on this material (Fig. 21) and on similar material (taken from the
first sulfation step) shows that, in general, calcium and sulfur occur
-------
79
v>
{£.
I-
UJ
o
(£
LU
O.
I
WAVE NUMBER
Fig. 21. Infrared Spectrum of Material Taken from Cycle Two of the
Cyclic Sulfation-Regeneration Experiment: A. Material
from Sulfation Step; B. Material from Reduction Step;
C. Material from Regeneration Step
(x = Sulfate and y = Carbonate)
-------
00
o
o
CO
CO
CO
<
oc.
UJ
o
a:
UJ
a.
WAVE NUMBER
Fig. 22. Infrared Spectrum of Material Taken from Cycle Two of
Experiment CATS-22: A. Material after Reaction with H2S;
B. Material after Regeneration Step
(y • Carbonate)
-------
Si
Fe
Ca
Mg
0
__ ^
W»*J«H™**^
J
i
V.
PARTICLE
PARTICLE-
Fig. 23. Electron Microprobe Scan of Particles Taken from the Second-
Cycle Sulfation Step of the Sulfation-Regeneration Experiment
oo
-------
00
t-J
SI
WLJL*A«*aJj mL**>L J»..w ^\uju^J/UW^u
Mg
PARTICLE
Fig. 24. Electron Microprobe Scan of Particles Taken from the Second-Cycle
Reduction Step of the Sulfation-Regeneration Experiment
-------
SI
Fe
I a
Mg
MJ'WUt^^
PARTICLE
PARTICLE-
k-
Fig. 25. Electron Microprobe Scan of Particles Taken from the
Second-Cycle Regeneration Step of the Sulfation-
Regeneration Experiment
'
-------
00
JS
Table 23. Summary of X-ray Analyses of Particles from Fluidized-Bed Experiments
Source of Solids Chemical Compounds Found
Sulfation-regeneration experiment Major: CaSO^ (Anhydrite), MgO (Periclase),
cycle 2 (sulfation) ' Minor: CaS (Oldhamite)
Sulfation-regeneration experiment Major: CaS (Oldhamite), CaS04 (Anhydrite),
cycle 2 (reduction) MgO (Periclase)
Possible Very Minor: Si02 (Alpha-Quartz)
Sulfation-regeneration experiment Major: CaS (Oldhamite), MgO (Periclase)
cycle 2 (regeneration) Possible Very Minor: CaC03 (Calcite)
CATS 22 - Product of second reaction Major: CaS (Oldhamite), MgO (Periclase)
with H2S Minor: CaC03 (Calcite)
CATS 22 - Product after second Major: CaC03 (Calcite), CaS (Oldhamite),
regeneration MgO (Periclase)
uetails of this experiment have been reported earlier .
-------
85
together throughout the particle at fairly uniform levels, whereas
magnesium appears to be segregated from the calcium and sulfur. This
indicates that sulfation takes place throughout the particle and is
not limited to a surface reaction. The presence of iron on the
particle surface, which had been inferred from microscopic examination
of the particles, is confirmed by the electron probe scan (Fig. 23),
which also shows some silicon and calcium concentrated at the surface.
The mineral form of this surface layer is not identified, although a
number of iron-calcium-silicon-oxygen (and/or sulfur) compounds exist.
Microscopic examination of this surface layer on a hot stage indicates
that its melting point is slightly less than 1100°C (2000°F).
The infrared spectrum of material taken from the second
reduction step in the sulfation-regeneration experiment is shown in
Fig. 21b. The intensity of the SO™ bands is much reduced in comparison
with Fig. 21a (sulfation step), as would be expected if the calcium
sulfate was indeed reduced to calcium sulfide, but the reaction is
clearly not complete. The X-ray data (Table 23) indicate that the
concentration of CaS increased, but that CaSO^ was still present in
appreciable amounts. From the electron microprobe scan in Fig. 24,
it is not clear whether or not the conversion from sulfate to sulfide
was limited to a surface layer; the general features of the scan appear
to be the same as those in Fig. 23.
Examination of the data for material taken from CATS-22
(Figs. 22a and 22b) shows that conversion of calcium carbonate to
calcium sulfide by reaction with hydrogen sulfide did proceed (although
not all carbonate was converted in this sample, see Fig. 22a). However,
the regeneration reaction caused only partial conversion of sulfide to
carbonate. Optical microscopy did not indicate that a surface layer
formed on any of these samples, and no electron microprobe scans of
the sample were made.
The most striking feature of Fig. 21 and of X-ray data
for material taken from the second regeneration step in the sulfation-
regeneration experiment is the complete absence of sulfate. The carbonate
feature in the infrared spectrum of this material (Fig. 21) is more
prominent than in the previously discussed samples, and calcium
carbonate can be detected in the X-ray pattern (see Table 23). Again,
the electron microprobe scan (Fig. 24) shows that calcium and sulfur
appear together uniformly across the particle and that the iron-rich
layer is still evident. The fact that all sulfate has disappeared
at this stage of the process seems to indicate that the layer is not
a barrier to reaction in the bulk of a particle.
All of the results appear to be consistent, but they do
not appear to explain directly the difficulties encountered in the
regeneration process. It seems likely that an iron-containing layer
forms on the particles during the sulfation process (i.e., that sulfate
or S02 is involved in the layer-forming reaction), but that the layer
is not directly responsible for the problems encountered in subsequent
-------
86
steps. The presence of sulfate after the reduction, step and the
inability to effectively regenerate carbonate from CaS in both the
cyclic and the CATS experiments suggest that the thermodynamics of
the desired reactions may not be as favorable as previously believed.
2. One-Step Reductive Decomposition Process, Low Temperature Studies
In the continuous, one-step, reductive decomposition experiments
made in the 3-in.-dia regenerator, the bed temperature is typically
maintained at above 1900°F. At above 1900°F, the calculated S02
equilibrium pressure (presented earlier for the CaSOi,.-CaS-CaO-CO-C02
and CaS04-CaS-CaC03-CO-C02 systems)1 in the gas is sufficiently high
so that the gas can be processed in a Glaus plant to recover the sulfur
content of the gas. Lower temperatures, although desirable from an
equipment corrosion standpoint, would not produce a gas with a high
enough S02 concentration for Glaus plant processing.
In experiment REDUC-8 made in the 3-in.-dia unit a high sulfur
removal from CaSO^ was demonstrated at a bed temperature of only 1600°F.
In the 3-in. unit, oxygen is present in addition to CaSOif-CaS-CaO-CO-CC^
to combust kerosene which provides heat to maintain the bed temperature.
The oxygen can react with CaS and the system would then consist of
CaSOtt-CaO-C02-CO for which the SC>2 equilibrium pressure is relatively
high, as shown in Fig. 26, at a temperature of 1600°F. Therefore, a
laboratory-scale scouting program (Experiments CATS-34 through -45) was
initiated using the 1- and 2-in.-dia, batch, fluidized-bed regenerators
to determine conditions which have an effect on the removal of sulfur
and to identify the sulfur species in the off-gas. A bed consisting of
CaSOij (Drierite) was reacted at a bed temperature range of 1200-1650°F
with a gas composed of mixtures of two or more of the following gases:
CO, C02, H20, and N2- The inlet gas compositions, operating conditions,
and SC>2 and 02 concentrations in the off-gas are presented in
Appendix C, Figs. C-l to C-12 and Tables C-l to C-5. Conclusive
results were not obtained in these experiments but some observations
were that:
1. Regeneration appeared to be most effective when
the bed temperature and C02 concentration were
such that the calcium product was CaCOa rather
than CaO.
2. Generally, high S02 concentrations were observed
when the 02 concentration was high.
3. The maximum concentrations of S02 and maximum
sulfur removals were obtained with an inlet
gas mixture consisting of 27% CO, 19% C02,
6% 02, and the remainder nitrogen. The maximum
S02 concentration was approximately equal to the
equilibrium concentration expected for the reaction:
+ CO •»• CaCOs + S02
-------
87
0 O.I 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fig. 26. Equilibrium Concentrations of S02 for the Reaction
CaSOit + CO •*• CaC03 + S02
-------
88
4. Elemental sulfur was present in the off-gas, but
the quantities detected were small.
5. Carbonyl sulfide was present in the off-gas. The
concentration was small, except when high concen-
trations of CO (>30%) were used. The presence of
carbonyl sulfide appeared to be dependent on the
presence of sulfur dioxide.
6. The sulfur material balances for the experiments
were generally >97%.
3. Reaction of CaSOq and COg/H? at Temperatures of 1400°F and Less
If the following one-step reaction
CaS04 + 4 H2 + C02 •*• CaC03 + 3 H20 + H2S
proceeds at low temperatures (<1400°F), the concentration of H2S in the
off-gas would be reasonably high. Equilibrium constants at 1000, 1200,
and 1400°F are 4.37 x 108, 3.15 x 10% and 1.51 x 107, respectively.
However, conversion of CaSOt* to CaS must be avoided because the equili-
brium concentration of H2S for the reaction of CaS with C02/H20
(CaS + C02 + H20 -*• CaC03 + H2S) is low, less than 1500 ppm.
A bed of CaS04 (Drierite) was reacted with 85% H2-C02 at 10 atm
and 1000, 1200, and 1400°F for 75, 135, and 120 min, respectively.
Experimental data are shown in Table 24. The results indicate that:
1. The peak concentration of H2S in the off-gas
increases with temperature.
2. The sulfide content of the bed increases with
temperature, thereby indicating that reduction
of CaSO^ to CaS is taking place, rather than
conversion to CaC03.
3. At 1400°F, the peak concentration of H2S (1160 ppm,
dry basis) was nearly equal to the equilibrium
concentration of H2S (1460 ppm, dry basis) expected
for the reaction between CaS and C02/H20.
A further analysis of the competing reactions has been made
and the information is presented in Appendix D. The conclusions were
that the addition of water to the reactant mixture would have suppressed
the CaS producing reaction, but that the H2S concentration would still
be too low for further processing of the gas in a Claus plant.
-------
Table 24. Reaction Conditions and Analytical Results for Studies on CaS04-H2-C02 Reaction
Pressure: 10 atm
Initial Bed: 700 g CaSOi, (Drierite, 22.6 wt % S)
Inlet-Gas Flow Rates: 0.85 cfm H2 at 70°F, 1 atm
0.15 cfm C02 at 70°F, 1 atm
Experiment
No.
CATS- 33
CATS-32
CATS-30
Bed Reaction
Temp Time
(°F) (min)
1400 0-25
25-50
50-75
75-120
1200 0-30
30-60
60-90
90-135
1000 75
Weight of Sulfur
Removed as
H2S
(g)
0.71
1.14
1.29
1.82
Total 4.96
0.21
0.59
0.67
1.04
Total 2.51
low
Avg. Cone.
of H2S
in Off-Gas
(ppm)
650
1030
1160
910
160
440
500
520
low
Weight of
Final Bed
(g)
544
666
677
Total
Sulfur in
Final Bed
(wt %)
27.6
22.6
X
22.2
Sulfide in
Final Bed
(wt %)
12
1.2
Not
Determined
Calculated Composition
of Final Bed
CaSOi,
(wt %)
66
(8.
96
(7.
Ca3 CaC03
(wt %) (wt %)
27
2 g S released)
3
7 g S released)
7
1
00
-------
90
4. Removal^ of H?S from Off-Gas by Reaction with Potassium
Permanganate
When CaS is converted to CaCOa by the reaction of CaS with the
C02-H20 mixture, large quantities of H2S will be formed. To avoid
releasing H2S into the laboratory exhaust system, a proposed scheme
to remove the H2S from the effluent gas stream is to pass the off-gas
from the regenerator through a scrubber system containing potassium
permanganate solution. The expected reaction is as follows:
8 KMnOi^ + 3 H2S -v 8 Mn02 + 3 K2SOit + 2 KOH + 2 H20
To determine if the above reaction occurs, as well as the contact
time necessary for reaction, a laboratory-scale experiment was performed.
The procedure consisted of passing a C02-H2S gas mixture through a
bubbler containing potassium permanganate solution and monitoring the
effluent gas stream for H2S with the quadrupole mass spectrometer.
The bubbler contained 200 ml of 0.4 M KMnO^ (0.0825 mole
A 90% C02-10% H2S gas mixture was passed through the solution at ambient
temperature and a flow rate of 0.013 cfm. The gas composition was
controlled by feeding with previously calibrated rotameters. The
reaction was rapid, as evidenced by immediate formation of a brown
precipitate (Mn02) and the absence of H2S in the effluent gas stream
at the start of the experiment. After 20 min of reaction, the presence
of H2S in the effluent gas stream was indicated by the quadrupole mass
spectrometer. The H2S flow was discontinued at that time. The total
quantity of H2S used during the reaction was 0.033 mole. The ratio
of moles of KMnO^ to moles of H2S was ^8:3, as would be expected from
the above reaction. The products of the reaction were a clear solution
and a brown precipitate of fine particles. A permanganate-containing
solution was used in the process gas scrubber as a result of these
tests.
-------
91
V. TRACE-ELEMENT DISTRIBUTION STUDIES
A. Introduction
The potential of a fluidized-bed combustion system for retaining,
in a solid constituent, trace elements released by the combustion
of coal will be investigated. Coal will be combusted at 10 atm in a
partially sulfated dolomite bed and also in an inert bed, such as
alumina. Data on the distribution of trace elements in experiments
with the two bed materials will be compared with each other and also
with data published on trace-element emissions from conventional boiler
plants.
B. Applicability of Selected Methods for Analyzing Trace Elements
of Interest
The elements of interest, in order of priority as pollution agents,
are listed in Table 25. Also listed in Table 25 are procedures
believed to be the most applicable for analyzing these elements at
trace concentration levels.
»
Ideally, a method of analyzing trace elements in coal combustion
products should (1) determine a large number of elements of interest
simultaneoulsy, (2) require relatively little sample preparation,
(3) be capable of automation, (4) produce an output compatible with
computerized data processing, and (5) be rapid.
Of the methods listed in Table 25, emission spectroscopy is
capable of measuring with suitable sensitivity all elements, except
fluorine and phosphorus. However, if this method is to meet the
criteria of rapid output and computerized data processing, a rather
sophisticated (and expensive) instrument equipped with direct-reading
detectors would be necessary. Calibration of the instrument using
materials of composition very close to those being analyzed (a time-
consuming effort) also would be necessary. The resulting system would
be highly specific to coal, ash, fly ash, etc., would be very rapid,
and would have an accuracy of perhaps 10-100%, varying for different
elements. The resulting system would be analogous to those used in
steel mills for monitoring the quality of routinely produced alloys.
X-ray spectrometry is another method capable of simultaneously
measuring most of the elements specified. It is capable of measuring
all elements of atomic number greater than about 15 (the atomic number
of phosphorus), but is is more sensitive for elements with higher
atomic numbers. Generally, X-ray spectrometry is not sufficiently
sensitive to be applicable at trace levels, particularly if sample
sizes are small. The X-ray spectrometer has the merit of internal
standarization which neutralizes matrix effects and makes for
relatively high accuracy. For a production system, the X-ray
-------
92
TABLE 25. Applicability of Selected Analytical Methods3
for Trace Elements in Coal Combustion
Priority
1
1
1
1
2
2
2
3
3
3
3
3
3
3
3
3
3
Element
or Ion
Fluoride
Lead
Mercury
Beryllium
Cadmium
Arsenic
Nickel
Copper
Zinc
Barium
Tin
Phosphorus
Lithium
Vanadium
Manganese
Chromium
Selenium
XRS
X
X
X
X
X
X
X
X
X
X
X
X
X
' ES
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
SIE AA/FP FLUOR COL
Xb
X
X
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
aXRS - X-ray Spectrometry.
ES - Emission Spectroscopy.
SIE - Specific Ion Electrode Method.
AA/FP - Atomic Absorption Spectrometry/Flame Photometry.
FLUOR - Fluorimetry.
COL - Colorimetric Method.
bX - Probably applicable for analysis at trace concentration levels,
spectrometer should be coupled to a computer.
An improved nondispersive X-ray spectrometer (NDXR) is
particularly suited for coupling to a computer because analysis of
the fluorescent X-ray spectrum from the sample is done by a solid-
state detector/pulse-height analyzer that provides digital data
for computer processing. Analysis by X-ray Spectrometry is discussed
further in a following section of this report.
Atomic absorption Spectrometry is a method with high sensitivity,
particularly since new atomization procedures have been developed,
for example, graphite furnaces and plasma units. The method is suited
for accurate analysis of specific elements because (unlike emission
-------
93
and X-ray spectroscopy) it does not provide information for more than
two elements simultaneously. Atomic absorption is particularly suited
to the determination of mercury and would also be appropriate for
lithium. It is most applicable to metals because the sensitivity
of the method declines with increasing nonmetallic character of the
element. The method is also less sensitive to elements that form
refractory oxides (and hence atomize with greater difficulty). This
method should probably be viewed as a supplement to emission or X-ray
spectrometry.
The specific ion electrode method is preferred for measuring trace
quantities of fluoride in aqueous solution. Generally, it is used
with a pyrohydrolysis technique for liberating the fluoride as HF and
collecting it in aqueous solution.
Fluorimetry is a very powerful method of analyzing very small
quantities of certain elements. For example, it is the method of
choice for uranium and beryllium. It is claimed that concentrations
of beryllium as low as 10" % can be determined in air dust without
chemical separations.
A variety of ultrasensitive colorimetric methods are available
for many elements that are suited for trace-element analysis of coal
and coal combustion residues. We have chosen to list only phosphorus
under the "Colorimetric" heading because other methods are more
appropriate for the other elements. The colorimetric method is
capable of measuring phosphorus in microgram quantities.
1. Analysis of Coal, Limestone, and Coal Combustion Residues by
X-ray jjpectrometric Analysis
X-ray spectrometry is a particularly attractive method for
analyzing coal and coal combustion residues (CCCR) because (1) in
principle, little sample preparation is required, (2) a number of elements
are determined simultaneously, (3)application is to all elements of
atomic number greater than 15 (the atomic number of phosphorus),
and (4) by the use of internal standards, quantitative information of
relatively good accuracy can be acquired.
In X-ray spectrometry, as usually practiced, a sample of the
material to be analyzed is irradiated with X-rays from a tungsten
(or possibly molybdenum) target X-ray tube. Elements in the sample
are excited and emit fluorescent X-rays whose energies are characteristic
of the elements in the sample. The X-ray spectrum so produced is
analyzed by a properly oriented crystal that disperses a beam of X-rays
according to Bragg1s law of diffraction. The intensities of the
dispersed X-rays are measured by a radiation detector that physically
traverses the area upon which they impinge. By suitable calibration,
the intensities are converted to concentrations.
-------
94
The X-ray spectrometer being developed at ANL differs from
the conventional instrument in two respects: (1) the exciting radiation
is itself a characteristic X-ray formed when a secondary target is
irradiated with X-rays from a conventional X-ray tube, and (2) the
detector is a solid-state radiation detector/pulse-height analyzer
that electronically sorts the fluorescent X-rays from the sample
according to energy and measures the intensities. The advantages of
the new system are simultaneous measurement of all X-rays and markedly
increased sensitivity. As little as tenths of a microgram of some
elements can be detected. By adding known quantities of elements
having about the same X-ray fluorescence energy as the element to be
detected, the quantities of the elements present can be accurately
estimated by direct comparison of the relative intensities.
2. Application of Nondispersive X-ray (NDXR) Spectrometry to
Coal and Coal Combustion Residues
The capability of NDXR spectrometry for quantitative simul-
taneous measurement of components at trace levels has been demonstrated
in a number of applications. Recently, the concentrations of U, Au,
Zn, and Fe in a sample were estimated to be 10, 0.1, 8-10, and 5 yg/ml
of solution, respectively; the sample analyzed consisted of less than
1 ml of solution with an internal standard. Detection of Cl, K, Ca,
Fe, Cu, Zn, Fb, and Br in bile has also been demonstrated.
Because this instrument may be applicable to the analysis
of CCCR, preliminary analyses were performed on several samples
retained from earlier coal combustion experiments. Approximately
0.010 g of sample was used in each analysis and the data obtained
were intended to be only qualitative.
The results are reported in Table 26. These are preliminary
data obtained with a far-from-optimized system. Constituents
identified as major are believed, to be present at concentrations
greater than 1% and minor constituents, at less than 1%. The column
headings "Principal," "Trace," and "Questionable" represent three
levels of concentration derived by visual comparison of peak heights
on an oscillogram. No internal standards were used to quantify the
amounts present.
If analysis for a particular element (or elements) were of
interest, the sensitivity could be optimized by using a secondary
target that would produce X-rays of an energy corresponding to the X-ray
absorption edge of the element of interest. (This would enhance
absorption and increase fluorescence X-ray intensity and, hence,
sensitivity.)
C. Flue-Gas Sampling
To determine the behavior of trace elements in coal during
fluidized-bed combustion, it will be necessary to analyze the final
-------
Table 26. Analysis of Coal and Coal Combustion Products with Nondispersive X-ray Spectrometer
Material
Commonwealth Edison Coal,
Shipment #4
Primary Cyclone Product from
Experiment AR-6
Secondary Cyclone Product from
Experiment AR-6
Final Bed from Experiment AR-6
Final Filter from Experiment AMER-33
Limestone No. 1359
Major Constituents3
Fe
Fe, Ca
Fe, Ca
Ca, Fe
Fe, Ca
Ca
Minor Constituents
Principal
Sr, Se
Sr
Sr
Si
Sr, Br
Fe
Trace
Cu,
Zn,
Pb,
Zn,
Cu
Zn,
Mo,
Se,
Ti»'
Mo,
Sr,
Ti
Ca
Mo
Ti
Ti
Pb
Zn
Cu
Se
K,
, Mn
, Ba
, Mo
, Cu
, Mo
, Zn,
Ba
Questionable
Br, Rb
Rb
Rb, Se
—
.
Rb
Believed to be present at >1 wt % cone.
Believed to be present at <1 wt % cone.
-------
96
flue gas for particulate and gaseous trace-element compositions, as
well as the various solids streams associated with the process. Flue-
gas sampling procedures have been determined and the assembly of the
necessary sampling equipment has been completed.
1. Description of Sampling Location
A very important consideration in trace-element sampling is
the location of the sampling point. Since the Brink Impactor to be
used in the particulate sampling is designed to handle particles in the
range from 0.3 to 3 ym in diameter, the sampling will be done after the
flue gas has passed through gas-solid separators which remove particles
larger than about 3 ym in diameter and through the pressure letdown
valve. This has the advantage of allowing the sampling to be
accomplished at near-atmospheric pressure, even though the combustor
and the ancillary equipment are operating at pressures of up to 10 atm.
The material removed in the separators can also be recovered and
analyzed, if so desired.
The modifications made in the flue-gas line to accommodate
the sampling process are indicated in Fig. 27. During sampling
CONTROL VALVE
FOR MAINTAINING
SYSTEM PRESSURE
SAMPLE LOOP
BY-PASS VALVE
WELL-TYPE
MANOMETER
Fig. 27. Experimental Flow System for Flue-Gas Sampling
-------
97
periods, the sample-loop isolation valves are opened and the bypass valve
is closed to direct the flue gas through the sampling zone. 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 increased cross section of
the sampling zone is intended to reduce the average gas velocity past
the sample probe to between 6 and 12 ft/sec, depending on the operating
conditions in the combustor.
To allow for isokinetic sampling of the flue gas, the total
gas flow through the sample loop will be determined using the orifice
meter. Based on the relative conditions at the meter and the sampling
location, the linear gas-flow rate past the sample probe can be
determined by assuming fully developed flow at the sampling point.
This is a very reasonable assumption, because the sample probe is
located approximately 12 pipe-diameters downstream and 3 pipe-diameters
upstream from any flow disturbances.
2. Particulate Sampling
The apparatus for the sampling of flue-gas particles is
illustrated in Fig. 28. The sample probe is constructed from 1/4-in.
Type 304 stainless steel tubing. A tapered nozzle with a 5/32-in.
opening is affixed to the end of the probe to provide for streamlined
sampling. The probe can be inserted to any depth along the diameter
of the flue to allow for sampling at several traverse points. The
probe, which normally faces upstream, can also be pointed downstream
when samples are not being taken to prevent possible accidental
plugging of the nozzle opening.
The particles in the gas sample are collected in the Brink
Cascade Impactor and the follow-up,glass-fiber filter (Gelman Type A).
The impactor has five in-line stages, each of which has a jet that
utilizes a collection cup as an impaction plate. A spring holds each
collection cup in place. A single stage of the impactor is illustrated
in Fig. 29.
The particles suspended in the gases pass through a jet.
Particles with sufficient inertia impact against a cup, and the remainder
pass through annular slots located around the cup. Figure 30 illustrates
the cup design.
Each collection cup has annular slots with a total cross-
sectional area 30 times the area of the largest jet, thereby reducing
turbulent effects to a negligible level. The dimensions of the impactor
jets are given in Table 27. The impactor, which is machined from
Type 316 stainless steel, collects particles in the range from 0.3
to 3.0 urn.
The glass-fiber filter at the outlet of the impactor collects
the particles that pass the final stage; in general, these are particles
-------
98
TO VENT
ROTOMETER
DRYING
1 THERMOCOUPLE
TUBE
WELL-TYPE
MANOMETER
DRY-GAS
METER
THERMOCOUPLE
ICE BATH
SAMPLING
ZONE
SAMPLE
PROBE
BRINKS 5-STAGE
IMPACTOR
I
CONTROLY
VALVE
a-
GLASS FIBER
FILTER
CONDENSATE
RECEIVER 1
A
U-TYPE
MANOMETER
Fig. 28. Apparatus for Particulate Sampling of Flue Gas
smaller than 3 ym.
The control valve is used to regulate the volumetric flow
through the impactor from 0.9 to 1.2 cfm, so that the velocity at the
nozzle inlet of the sample probe equals the linear velocity of the
flue gas flowing past the probe. This is done to promote representative
-------
:
COLLECTION
CUP
SPRING
JET SPINDLE
GASKET
Fig. 29. Single Impactor Stage
sampling of the particles in the flue gas by isokinetic sampling. It
may be necessary to heat the probe, impactor, and filter to prevent
condensation of flue-gas water.
The remainder of the sample train is designed to accurately
determine the sampling rate and the total volume of the gas sample
taken, so that the particulate loading of the main flue-gas stream can
be calculated.
-3 SLOTS
Fig. 30. Design of Impactor Collection Cups
-------
100
TABLE 27. Dimensions of Cascade Impactor Jets
Dimensions, (cm)
Spacing of
Jet No. Jet Dia Jet Opening3
1
2
3
4
5
0.2490
0.1775
0.1396
0.0946
0.0731
0.747
0.533
0.419
0.282
0.220
From collection-cup surface.
3. Treatment of Particulate Samples
It is anticipated that, in addition to obtaining total parti-
culate and trace-element concentrations in the flue gas, it will also
be possible to obtain trace-element data as a function of particle size.
It will be necessary, therefore, to obtain complete analytical and
particle-size-distribution data for the material impacted on each cup
or stage of the impactor. Since it is not uncommon to collect from 40
to 80 percent of the particulate matter on one stage of the impactor,
the sample time will be limited by re-entrainment from this cup. The
max!mum sample time must be used, therefore, to insure the collection
of reasonable quantities of material on the other cups.
To obtain particle-size-distribution data, samples of the
material from each stage and the final filter will be examined using
a scanning electron microscope (SEM). Experience with the instrument
available indicates that reasonably good photomicrographs can be
obtained,at magnifications as high as 20.000X. It should be possible,
therefore, to obtain from such photographs reasonably accurate
particle-size information.
Consideration is also being given to the possible application
of a scanning computer system developed by C. B. Shelman and Donald
Hodges of the Applied Mathematics Division of Argonne National
Laboratory to the counting and size analysis of the particles. The
system, called ALICE, is capable of searching photographs for predeter-
mined details and then analyzing these details and recording the
resulting data. In the examination of nerve fibers, for example,
ALICE counted each fiber in the nerve samples, calculated the area
of each fiber, and remembered its location, so that it could provide
information on the distribution of fibers of any particular size. A
great deal of attention will be given to preparing the photomicrographs
in a way that will ensure their adequacy for computer analysis.
To obtain the necessary elemental information, the material
from each stage of the impactor will also be subjected to chemical
-------
101
analysis. The particular elements of interest and the analytical
methods being considered for their determination have been described
above.
4. Determination of Mercury in Particulate and Gaseous Emissions
Supplemental to the primary determination of particulate
emissions in the flue gas is the determination of possible trace-
element gaseous emissions. Of particular concern in this area is
mercury. Billings et al. have reported that up to 90% of the mercury
in the coal is released during combustion and appears as vapor
discharged in the stack gas, while only 10% remains in the residual
ash. It would seem essential, therefore, that in the case of mercury,
a method for sampling both particulate and gaseous emissions should
be employed.
The method being considered is essentially the one proposed
by the Environmental Protection Agency.5 The apparatus to be used
in the sampling train is illustrated in Fig. 31. In principle, the
particulate and gaseous emissions are isokinetically sampled from the
flue gas and collected in the acidic iodine monochloride solution.
The mercury collected (in the mercuric state) is reduced to elemental
mercury in basic solution by hydroxylamine sulfate. Mercury is
subsequently vaporized from the solution by a mercury-free air stream
and analyzed using an atomic absorption spectrophotometer in the
flameless mode.
The entire sampling train up to the check valve will be
constructed of Pyrex glass. The probe dimensions, which have not yet
been determined, will be somewhat larger than the stainless steel probe
to accommodate the larger volumetric sampling flow rates of 0.5 to
1.0 cfm.
The second impinger in the sampling train is of the Greenburg-
Smith design with the standard tip. The remainder of the impingers
are gas-washing bottles modified by replacing the standard tip with
a l/2-in.-ID glass tube, which extends to 1/2 in. from the bottom of
the flask. Each flask has a capacity of 500 ml. The first, second,
and third impingers contain the mercury scrubbing solution. The
fourth impinger contains soda lime to neutralize acid vapors, while
the fifth impinger contains silica gel to remove water vapor.
The filter holder will also be made of Pyrex and is designed
to accept 50-mm glass-fiber filter paper. The filter provides for
the removal of particulates and liquid mist entrained in the flowing
gas from the first three impingers.
-------
102
SAMPLING
ZONE
THERMOCOUPLE
SAMPLE
GLASS FIBER
FILTER
PROBE
ICE
BATH
IMPINGERS
TO VENT
-X*-
CONTROL
VALVE
ROTOMETER
1
M
1
1
f
1
1
/
i>
1
1
1
GAS DRYING
BOTTLES
THERMOCOUPLE
CHECK
VALVE
DRY-GAS
METER
WELL-TYPE
MANOMETER
PUMP
Fig. 31. Apparatus for the Determination of Farticulate
and Gaseous Emissions of Mercury in Flue Gas
Assembly of the sampling equipment has been completed and the
collecting of samples from the flue gas will be initiated in the near
future.
5. Determination of Inorganic Fluorides in Gaseous and Particulate
Emissions
It is also possible that some fluorine may be present in the
gaseous phase of the flue gas from the combustor. It has been demon-
strated by Pack et at. that gaseous and particulate inorganic fluorides
can be sampled from the atmosphere using a water-filled impinger.6
It will be possible, therefore, to use the same sampling train for both
the mercury determination and the fluoride determination. Analytical
methods for trace quantities of fluoride are currently being researched.
When an appropriate analytical method is selected, sampling for fluoride
will also be initiated.
-------
103
VI. ACKNOWLEDGMENTS
We gratefully acknowledge the help given by Dr. R. C. Vogel,
Mr. L. Burris, Mr. D. S. Webster, Dr. S. Lawroski, and Mr. L. Link in
directing and reviewing the program, to Mr. D. E. Grosvenor,
Mr. J. F. Lenc, and Mr. J. J. Stockbar for their many contributions
and efforts during their temporary appointment to the group, and to
our secretary, Mrs. Vita Sniffer.
We are indebted to Dr. P. Cunningham, Dr. V. Maroni, Dr. C. Johnson,
Mr. S. Johnson, Mr. R. Schablaske, and Mr. W. Shinn for their analysis
of bed particle samples, and to Dr. P. Cunningham for preparing the
section of this report relative to the results of their analysis of the
bed particles. Also, to Dr. P. Cunningham, Dr. M. Blander, and
Dr. R. Yang for their study of the thermodynamic considerations relating
to the conversion of CaSOi^ to CaCO%.
-------
104
VII. REFERENCES
1. A. A. Jonke et al.3 Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion and Regeneration of
Sulfur-Containing Additives, Annual Report, July 1971-June 1972,
EPA-R2-73-253. (Available from National Technical Information
Service as PB 227-058/AS)
2. T. W. Zadick, R. Zavalita, and F. P. McCandless, Department of
Chemical Engineering, Montana State University, private
communication.
3. A. A. Jonke et ai.3 Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion, Annual Report, July 1969-
June 1970, ANL/ES-CEN-1002.
4. . C. E. Billings, A. M. Sacco, W. R. Matson, R. M. Griffin,
W. R. Coniglio, and R. A. Barley, Mercury Balance on a Large
Pulverized Coal-Fired Furnace, Preprint of paper presented at the
65th Annual Meeting of the Air Pollution Control Association,
Miami Beach, Florida, June 1972.
5. Environmental Protection Agency, National Emission Standards for
Hazardous Air Pollutants, Federal Register 36: No. 234, pp. 23239-
23256, Dec. 7, 1971.
6. M. R. Pack, A. C. Hill, M. D. Thomas, and L. G. Transtrum,
Determination of Gaseous and Particulate Inorganic Fluorides in
the Atmosphere, American Society for Testing Materials, Special
Technical Publication No. 281, pp. 27-44, 1960.
-------
105
APPENDIX A. ENGLISH TO METRIC UNIT EQUIVALENTS
-------
106
Table A-l. English to Metric Unit Equivalents
English System
Metric Equivalent
Length
Area
Volume
Mass
Pressure
Temperature
Energy
Power
in.
ft
in.2
ft2
in.3
ft3
oz
Ib
lb/in.2
in. H20
atm
°R
BTU
BTU/min
2.54 cm
0.305 m
6.45 cm2
0.093 m2
16.39 cm3
28.32 1 (liter)
28.35 g
453.6 g
51.70 mm Hg
1.865 mm Hg
760 mm Hg
1.8 (°C) + 32
1.8 (°K)
252 cal
252 cal/min
-------
107
APPENDIX B. SPECIFIC ANALYTICAL PROCEDURES
FOR SOLIDS SAMPLES FROM COMBUSTION AND REGENERATION EXPERIMENTS
-------
108
Table B-l. Specific Analytical Procedures
A. Elemental and Functional Groups;
1) Calcium. Determined by means of an EDTA volumetric titration.
2) Sulfur. Leco combustion method employing an automatic
volumetric titration of an IC^-I-starch system.
3) Carbon. Leco combustion method using a thermal-conductivity
detector employing the C<2/C02 relationship.
4) Carbonate. Determined by measuring with a sensitive gas
chromatograph the C(>2 evolved by acidification of the material.
5) Sulfate, S0u.°. A gravimetric method is employed. The sulfate
present is converted to barium sulfate which is treated to
900°C to remove any organic material and weighed.
6) Combined Nitrogen. A micro-Kjeldahl method is used. The
nitrogen is released as ammonia, absorbed, and titrated.
B. Coal and Coal Ash
These analyses are performed by a commercial laboratory (Commercial
Testing and Engineering Co., South Holland, 111.) employing ASTM
designation D271-64 methods. The analyses made are proximate,
ultimate, mineral (ash), fusibility of ash, free swelling index,
and calorific value.
-------
109
APPENDIX C. OPERATING DATA AND RESULTS OBTAINED IN
BATCH REGENERATION EXPERIMENTS IN THE 1- AND 2-INCH-DIAMETER REACTOR
Regarding the Reaction
+ CO J CaC03 + S02
(CATS-34 through -45)
-------
110
CL
CVJ
o
1700
1600
1500
1400
1300
1200
30
20
10
0
4
3
2
1
0
50
40
30
20
0
10
20
_L
30
40 50
Time, min
60
70
80
Fig. C-l. Operating Conditions for Experiment CATS-34
-------
Ill
Q.
O)
•o
OJ
CO
in
10
CD
I
O.—
o
c >
CSJ
o
O i—
O
C >
CSI
CD
tJ
O
>
CM
O
1700
1600
1500
1400
1300
1200
30
20
10
0
5
4
3
2
1
0
60
50
40
30
12.5
10.0
7.5
5.0
2.
0
0
inlet
cone, in Outlet Gas:
0.3% max for first _
35 min; <0.1 rest
of time
1
10 20
30 40 50
Time, min
60
70
80
Fig. C-2. Operating Conditions for Experiment CATS-35
-------
112
Q.
E
OJ
O)
CO
1700
1600
1600
1400
1300
1200
30
20
10
0
to
5 0.3
*«- -o
o^I 0.2
0.1
0
O i—
o
c >
o
o
CM
o
to
re
CD
o
>
o
40
30
20
10
0
I
10 20 30 40 50
Time, min
60
70
30
Fig. C-3. Operating Conditions for Experiment CATS-36
-------
113
o.
01
•o
o>
CD
I/)
ID
C3
I
O •—
O
cva
o
10
in
»o
c\j
O
o
o
>
1500
1400
1300
1200
1100
30
20
10
0
4
3
2
1
0
50
40
30
20
7
6
5
4
3
2
1
0
inlet
20 30
Time, min
40
50
Fig. C-4. Operating Conditions for Experiment
CATS-37
-------
114
o
a.
1700
1600
1500
1400
<£ 1300
1200
30
o^- 20
o
c >
10
0
5
4
3
2
1
0
60
50
O)
+j **
o-—
o
c >
3 &«
O
Co 40
•?- >
CM
O
30
20
«J
*J
to a«
t/l r—
n> o
ts >
c
•^
6
4
2
0
in
III
X
Is — i
10
30 40 50
Time, rain
Fig. C-5. Operating Conditions for Experiment CATS-38
-------
115
E
01
CQ
1600
1500
1400
1300
1200
10
5
0
0.2
•".- 0.1
CM+» O
O 3 >
COO
50 ,_
8-4
40
o
o
>
C\J
o
1
20
30 40 50
Time, min
60
70
Fig. C-6. Operating Conditions for Experiment CATS-39
-------
116
o
CL
o t
«_> 3
o
at
O •—
O
c >
C\J
o
CO
1300
30
20
10
0
0.8
0.6
0.4
0.2
0
ai
O r—
O
c >
CM
O
o
>
C\J
o
30
20
10
0
4
3
2
1
0
inlet
I
Outlet <0.1% throughout
I I I
10 20 30 40 50
Time, nin
60
70
80
Fig. C-7. Operating Conditions for Experiment CATS-40
-------
117
(1)
CO
01
+J
o
>
CM
O
O
a*
r—
O
CNJ
o
1600
1500
1400
1300
1200
40
30
20
1
0.8
0.6
: 0.4
0.2
0
90
80
70
60
50
40
5
4
3
2
1
0
Above Detection Lirat
inlet
outlet
10 20 30 40
Time, min
50
60
Fig. C-8. Operating Conditions for Experiment CATS-41
-------
118
Q.
01
1600
1500
-o 1400
03
1300
30
•»
C -M &S
'-^^ 20
O -M O
O 3 >
10
0.5
. 0.4
^3S* 0.3
0.2
O
o- 0.1
0
60
CVHJ *O
O 3 >
°° 40
OJ
O
wl
4
3
2
1
n
—
—
J
-
__ inlet
IX
^iii^——++^i^^~~ outlet
- — •f" 1 **l 1
—
L—
10 20 30 40
Time, min
50
60
Fig. C-9. Operating Conditions for Experiment CATS-42
-------
119
O
*
ex
•o
OJ
CO
C ••-> 65
r>
O
1700
1600
1500
1400
1300
15
10
5
0
0.5
0.4
0.3
5 0.2
CM
O
OO
•r- +J
OJ
o -i-» o
o
CvJ
O
0.1
0
90
80
5
4
3
2
1
0
r
1
inlet
outlet
10
20
30
40
50
60
70
80
90
Fig. C-10. Operating Conditions for Experiment CATS-43
-------
120
o
>
CNJ
O
0 10 20 30 40 50 60 70 80 90 100 110 120
Time, min
Fig. Oil. Operating Conditions for Experiment CATS-44
-------
121
1700
1600
1500
1400
1300
C •!->
O +J O
LJ 3 >
O
c: o
•r- >
PJ
O
O)
C O
CM
O
o
9f
10 20 30 40 50 60 70 80 90 100 110 120
Time, min
Fig. C-12. Operating Conditions for Experiment CATS-45
-------
10
Table C-l. Operating Conditions and Results of Experiments CATS-34 to CATS-37
System pressure: 2 atm Superficial gas velocity: 2 ft/sec at 1600°F
Experiment
Operating Conditions
Reactor
Temperature, °F
Reaction time, min
Inlet-Gas Flow Rates
CO, cfm at 70"F, 1 atm
C02, cfm at 70"F, 1 atm
02, cfm at 70°F, 1 atm
H20, cfm at 70°F, 1 atm
N2, cfm at 70°F, 1 atm
Total
Bed Data
Weight of initial bed, g CaS04 (Drierite)
Weight of sulfur in initial bed, g
Weight of final bed, g
Concentration of sulfur in final bed, wt %
Weight of sulfur in final bed, g
Concentration of sulfide in final bed, wt %
Weight of sulfide in final bed, g
Final Bed Composition
CaS, wt %
CaS04, wt %
CaC03-CaO, wt % (by difference)
Sulfur Balance
Weight of sulfur as CaS, g
Weight of sulfur as CaS04, g
Weight of sulfur as S02, g
Weight of sulfur as H2S, g
Total, g
Percent material balance
CATS-34
alumina
1210-1670
80
0.09
0.07
0.17
0.33
140
31.7
72
32
23
20.9
15.1
/47.1
46.6
6.3
15.1
7.9
8.9
31. 9a
100.5
CATS-35
alumina
1220-1620
80
0.09
0.07
0.02-0.04
0.14
0.33
140
31.7
77.5
14.8
11.5
0.02
0.015
0.05
62.8
37.2
0.02
11.5
17.2
28. 7a
90.5
CATS -36
SS
1200-1620
80
0.37
0.32
0.02
0.71
1.42
570
128.8
492
23.3
114.5
3.6
17.7
8.1
83.5
8.4
17.7
96.7
1.4
4.8
120. 6a
93.7
CATS-37
SS
1175-1460
45
0.37
0.32
0.03-0.09
0.61
1.36
570
128.8
473
21.4
101.3
2.4
11.5
5.5
80.8
13.7
11.5
89.8
15.7
117.03
91
Does not include weight of sulfur as S or COS.
-------
Table C-2. Operating Conditions and Results of Experiments CATS-38 to CATS-41
System pressure: 2 atm Superficial gas velocity: 2 ft/sec at 1600°F
Experiment
Operating Conditions
Reactor
Temperature, "F
Reaction time, min
Inlet-Gas Flow Rates
CO, cfm at 70° F, 1 atm
C02, cfm at 70°F, 1 atm
02, cfm at 70° F, 1 atm
H20, cfm at 70°F, 1 atm
N2 , cfm at 70° F, 1 atm
Total
Bed Data
Weight of initial bed, g CaSOit (Drierite)
Weight of sulfur in initial bed, g
Weight of final bed, g
Concentration of sulfur in final bed, wt %
Weight of sulfur in final bed, g
Concentration of sulfide in final bed, wt %
Weight of sulfide in final bed, g
Final Bed Composition
CaS, wt %
CaS04, wt %
CaC03-CaO, wt % (by difference)
Sulfur Balance
Weight of sulfur as CaS, g
Weight of sulfur as CaSOi, , g
Weight of sulfur as S02 , g
Total, g
Percent material balance
CATS-38
SS
1250-1610
80
0.37
0.26
0-0.08
0.71
1.38
570
128.8
450
20.8
93.5
1.12
5.0
2.5
83.5
13.9
5.0
88.5
33.3
126. 8a
98.5
CATS-39
SS
1220-1500
80
0.1
0.5
0.02-0.05
0.71
1.34
570
128.8
566 ,
22.8
129
0.02
0.1
0.04
97.2
2.9
0.1
128.9
3
132 .Oa
102.5
CATS-40
SS
1330-1650
80
0.37
0.02-0.04
0.97
1.36
570
128.8
389
30.6
119.2
21.7
84.5
49
38
13.1
84.5
34.7
7.6
126. 8a
98.5
CATS-41
SS
1240-1510
60
0.76
0.62
0.02-0.07
1.42
570
128.8
432
24.7
106.8
10.9
47.2
24.6
58.6
16.9
47.2
59.6
8.6
115. 4a
89.8
T)oes not include weight of sulfur as S or COS.
ro
Oo
-------
N3
Table C-3. Operating Conditions and Results of Experiments CATS-42 and CATS-43
System pressure: 2 atm Superficial gas velocity: 2 ft/sec at 1600°F
Experiment
Operating Conditions
Reactor
Temperature, °F
Reaction time, min
Inlet-Gas Flow Rates
CO, cfm at 70° F, 1 atm
C02, cfm at 70T, 1 atm
02, cfm at 70°F, 1 atm
H20, cfm at 70°F, 1 atm
N,, cfm at 70° F, 1 atm
£' '
Total
Bed Data
Weight of initial bed, g CaSOi, (Drierite)
Weight of sulfur in initial bed, g
Weight of final bed, g
Concentration of sulfur in final bed, wt %
Weight of sulfur in final bed, g
Concentration of sulfide in final bed, wt %
Weight of sulfide in final bed, g
Final Bed Composition
CaS, wt %
CaSOin wt %
CaC03-CaO, wt % (by difference)
Sulfur Balance
Weight of sulfur as CaS, g
Weight of sulfur as CaSOi>, g
Weight of sulfur as S02, g
Total, g
Percent material balance
CATS-42
SS
1370-1540
53
0.40
0.49
0.02-0.06
0.43
1.36
570
128.8
507
22.2
112.3
4.4
22.3
9.9
75.5
14.6
22.3
90.0
4.2
116. 5a
90.6
CATS-43
SS
1370-1680
90
0.14
0.97
0.02-0.07
0.14
1.29
570
128.8
567
21.8
123.4
0
0
0
92.5
7.5
0
123.4
6.1
129. 5a.
100.5
not include weight of sulfur as S or COS.
X
-------
Table C-4. Operating Conditions and Results of Experiment CATS-44
System pressure: 2 atm Superficial gas velocity: 2 ft/sec at 1600°F
Operating Conditions
Reactor
Temperature, °F
Reaction tine, min
Time Period, min
Inlet-Gas Flow Rates
CO, cfm at 70°F, 1 atm
C02, cfm at 70° F, 1 atm
02, cfm at 70° F, 1 ata
H20, cfm at 70°F, 1 atm
N2, cfm at 70°F, 1 atm
Total
Bed Data
Weight of initial bed, g CaS04 (Drierite)
Weight of sulfur in initial bed, g
Weight of final bed, g
Concentration of sulfur in final bed, wt %
Weight of sulfur in final bed, g
Concentration of sulfide in final bed, wt %
Weight of sulfide in final bed, g
Final Bed Composition
CaS, wt %
CaS04, wt %
CaC03-CaO, wt % (by difference)
Sulfur Balance
Weight of sulfur as CaS, g
Weight of sulfur as CaSOi,, g
Weight of sulfur as S02 , g
Weight of sulfur as H2S, g
Total, g
Percent material balance
SS
1350-1590
120
0-72 72-82 82-86 86-95 95-99
0,28 0.28 0.28 0.14 0.2
1.03 0.69 1.03 1.17 1.17
0.01-0.05 0.04 0.04 0.04-0.05 0.04
0.34
1.34 1.35 1.35 1.35 1.41
570
128.8
542
21.3
115.4
1.4
7.6
3.2
84.5
12.3
7.6
107.8
10.3
125. 7a
97.6
99-120
0.1
1.2
0.04
1.34
Does not include weight of sulfur as S or COS.
-------
Table C-5. Operating Conditions and Results of Experiment CATS-45
System pressure: 2 atm Superficial gas velocity: 2 ft/sec at 1600°F
Operating Conditions
Reactor
Temperature, °F
Reaction time, min
Tine Period, min 0-35
Inlet-Gas Flow Rates
CO, cfm at 70"F, 1 atm 0.41
C02, cfm at 70° F, 1 atm 0.69
02, cfm at 70°F, 1 atm 0.0-0.04
H20, cfm at 70°F, 1 atm
N2, cfm at 70"Ff 1 atm 0.24
Total 1.36
Bed Data
Weight of initial bed, g CaSOi, (Drierite)
Weight of sulfur in initial bed, g
Weight of final bed, g
Concentration of sulfur in final bed, wt %
Weight of sulfur in final bed, g
Concentration of sulfide in final bed, wt %
Weight of sulfide in final bed, g
Final Bed Composition
CaS, wt %
CaS04, wt %
CaC03-CaO, wt % (by difference)
Sulfur Balance
Weight of sulfur as CaS , g
Weight of sulfur as CaSOi,, g
Weight of sulfur as S02 , g
Weight of sulfur as H2S, g
Total, g
Percent material balance
SS
1370-1620
120
35-44 44-54 54-79 79-89
0.41 0.41 0.41 0.17
0.93 0.69 0.93 0.93
0.04 0.04 0.04 0.04
0.24 0.24
1.38 1.38 1.38 1.38
570
128.8
474
21.1
100
8
37.9
18
55.7
26.4
37.9
62.1
18.3
118. 3a
92
89-98 98-120
0.06 0.01
0.99 1.27
0.04 0.04
0.28
1.38 1.32
Does not include weight of sulfur as S or COS.
-------
127
APPENDIX D. THERMODYNAMIC CONSIDERATIONS
RELATING TO THE CONVERSION OF CALCIUM SULFATE TO CALCIUM CARBONATE
-------
128
Thermodynamic Considerations Relating to the Conversion of Calcium
Sulfate to Calcium Carbonate
The present process being considered for the conversion of
to CaC03 involves a two-step process in which CaSOi^ is converted to CaS
and CaS in turn is converted to CaC03. We have attempted to reexamine
the thermodynamics of .these processes and suggest possible alternative
schemes to accomplish conversion of CaSO^ to CaC03.
Among the many possible reactions arising in the presence of CaSO^,
H2, C02, H20, CaS, CaC03, H2S, predominant ones are:
CaS04 + 4 H2 ->• CaS + 4 H20 (1)
CaS + H20 + C02 -*• CaC03 + H2S (2)
4 H2 + C02 -»• CaC03 + 3 H20 + H2S (3)
H20 + CO -> C02 + H2 (4)
CaS04 + 4 CO -»- CaS + 4 C02 (5)
+ 3 H2 + CO -*• CaC03 + H2S + 2 H20 (6)
CaS04 + CO •* CaC03 + S02 (7)
S02 + 3 H2 •*• H2S + 2 H20 (8)
The equilibrium constants for these reactions at selected
temperatures are tabulated in Table D-l.
In the presence of CaSOt^, H£, and COz,reaction (1) may proceed to
a much greater extent than reaction (3). However, when PH20/PH2 £ ^Kj,
where KI is the equilibrium constant for reaction (1), reaction (1) is
suppressed and reaction (3) proceeds.
Thermodynamic considerations of these and relevant reactions follow:
-------
Table D-l. Equilibrium Constants for Predominant Reactions in the Conversion
of Calcium Sulfate to Calcium Carbonate at Selected Temperatures
1000 °F
Reactions In K K
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
CaS04
CaS +
CaS +
CaCOa
CaSO,
CaC03
H20 +
C02 +
CaSO,,
CaS +
CaSO^
CaC03
CaS04
CaC03
S02 +
H2S +
+ 4H2 •»•
21.768 2.818 x 109
4H20
H20 + C02 ->•
-1.862 0.153
+ H2S
-1- 4H2 + C02 -*•
19.896 4.37 x 108
+ 3H20 + H2S
CO •>
1.373 3.949
H2
+ 4CO -»
27.25 6.853 x 1011
4C02
•*- 3H2 + CO ->
21.269 1.728 x 109
+ H2S + 2H20
+ CO ->•
-1.442 0.236
+ S02
3H2 -»•
£ in
22.713 7.316 x 1010
2H20
Temperature
1200 °F 1400 "F
In K K In K K
21.025 1.352 x 109 20.329 6.74 x 108
-3.759 0.023 -5.199 5.52 x HT3
17.265 3.148 x 107 15.129 3.719 x 106
0.710 2.033 0.235 1.265
'
23.863 2.31 x 10 10 21.268 1.724 x 109
17.976 6.407 x 107 15.364 4.704 x 106
-0.787 0.455 -0.270 0.763
o c
18.763 1.408 x 108 15.634 6.164 x 106
10
\o
-------
130
(a) At 1000°F from reaction (1) and (3)
£ 2.82 x 109 •*• PH20/PH2 % 230
(PH20)3 (PH2S)
£ 4.37 x 108
(PH2)tf(PC02)
Let PH20/PH2 > 230, reaction (1) does not substantially proceed and
the following equilibrium condition holds for (3):
PH2S
< 36
(PH2)(PC02)
For reaction (4),
(PC02)(PH2) PC02
"V q qc __ > QAQ
(PH20)(PCO) * J'y:> °r PCO yUy
For reaction (5) to reach equilibrium,
(PC02)'t PC02
— £ 6.85 x 1011 or
(PCO)*
Therefore, the amount of CO produced by reaction (4) is just
enough to initiate reaction (5). To prevent this, PH20/PH2 should be
greater than 230. Reaction (6) would be in equilibrium when (3) and
(4) are in equilibrium.
If the feed gas contains: 0.04 atm H2, 9.8 atm H20, 0.2 atm C02.
PH20/PH2 = 245
PH2S
30
(PH2) (PC02)
Equilibrium PH2S =0.28 atm
Stoichiometric PH2S = 0.01 atm [from (4)]
Therefore, total consumption of H2 is expected.
-------
131
By making PH20/PH2 high enough, reaction (3) would be predominant.
Similar calculations for T = 1400°F indicate that the actual equilibrium
constant, PH2S/(PH2) (PCOa), at lower temperature is much greater than at
higher temperature. Thus, lower temperature is preferrable. The PH20/PH2
ratio should not be made too great because the actual equilibrium
constant, PH2S/(PH2)(PC02), is proportional to [1/(PH20/PH2)3]. At too
low a temperature,PH2 is very low and the partial pressure of H2S will
be correspondingly low.
It is recognized that the above reaction produces H2S at partial
pressures too low to permit subsequent sulfur recovery reactions.
Nevertheless, it is recommended that the conditions described as being
favorable for reaction (3) be experimentally verified. The results of
such an experiment can be expected to provide greater insight into the
problems.
-------
132
TECHNICAL REPORT DATA
(Please read JuiJfuctions on llie reverse before c
1. REPORT NO.
EPA-650/2-74-057
2.
4. TITLE AND SUBTITLE
Reduction of Atmospheric Pollution by the Application
of Fluidized-Bed Combustion
3. RECIPIENT'S ACCbSSION-NO.
5. REPORT DATE
June 1974
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(s)G.J.Vogel, M.Haas, W.Swift, J.Riha,
C.B.Schoffstoll, J.Hepperly, and A.A.Jonke
8. PERFORMING ORGANIZATION REPORT NO.
ANL/ES-CEN-1006
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-IAG-0199(D)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Annual; 7/72-6/73
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The report gives results of a bench-scale and laboratory-scale experi-
mental investigation of the feasibility of applying fluidized-bed combustion (FBC) of
coal to power generation. The FBC concept would combust coal at elevated pressure
(to 10 atm) in a fluidized bed of dolomite, which reacts with the SO2 formed during
combustion. The partially sulfated dolomite is then regenerated using one of the two
alternative techniques: a one-step reductive decomposition technique (converting
CaSO4 to CaO and SO2); or a two-step technique involving the reduction of CaSO4 to
CaS, followed by reaction of the CaS with H2O and CO2 to form CaCO3 and H2S.
Results of the pressurized FBC tests indicate favorable air pollutant emission control
with 90-95% reduction in SO2 emission, and with NO levels of about 150 ppm, well
below EPA's New Source Performance Standards for large coal boilers. Initial
sorbent regeneration studies showed significant deactivation and decrepitation of the
dolomite; further investigation is necessary in the area of regeneration.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Fluidized-Bed
Processing
Coal
Combustion
Electric Power
Generation
Regeneration
(Engineering)
Sulfur Dioxide
Air Pollution Control
Stationary Sources
Pressurized Fluidized
Bed Combustion
13B
10A
13H, 07A
21D, 08G
21B, 07B
Dolomite (Mineral) Nitrogen Oxide (NO)
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Reportj
Unclassified
21. NO. OF PAGES
132
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |