ARGONNE NATIONAL LABORATORY
9700 South Cass Avenue
Argonne, Illinois 60439
REDUCTION OF ATMOSPHERIC POLLUTION
BY THE APPLICATION OF
FLUIDIZED-BED COMBUSTION
Annual Report
July 1970—June 1971
by
A. A. Jonka, G. J, Vogel, L. J. Anastasia,
R. L. Jarry, D, Ransaswami, M. Haas,
C. B. Schoffstoll, J, R, Pavlik,
G. N. Vargo, and R. Green
Chemical Engineering Division
Work performed under an agreement between the
U. S. Atomic Energy Commission
and
ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF AIR' PROGRAMS
DIVISION OF CONTROL SYSTEMS
EPA-IAG-Q020
Preceding Annual Reports
ANL/ES-CEN-10Q1 July 1968—June 1969
ANL/ES-CEN-1002 July 1969—June 1970
-------
OCLC Connexion
OCLC 1998655 Held by ESA - 2 other holdings
Page 1 of 1
Recstate Entered 19760218 Replaced 20190422
Type a ELvl I Srce d Audn Ctrl Lang eng
BLvl m Form Conf 0 Biog MRec Ctry ilu
Cont b GPub LitF 0 Indx 0
Desc i Ills a Fest 0 DtSt s Dates 1972 ,
040 DGU *b eng *c DGU *d OCLCQ *d OCLCF *d OCLCQ *d OCLCO *d ESA
050 4 TD885 *b ,R4 1971
088 EPA APTD-1128
099 EPA APTD-1128
049 ESAD
245 0 0 Reduction of atmospheric pollution by the application of fluidized-bed combustion : *b annual report, July 1970-
June 1971 / te by A.A. Jonke [and others].
260 Argonne, III. : *b Argonne National Laboratory : *b Distributed by National Technical Information Service, *c
[1972]
300 112 pages: *b illustrations; *c 28 cm
336 text +b txt +2 rdacontent
337 unmediated ^b n +2 rdamedia
338 volume *b nc +2 rdacarrier
500 At head ot title: ANL/ES-CEN-1004 Meteorology.
500 Work performed under an agreement between the U.S. Atomic Energy Commission and Environmental Protection
Agency, Office of Air Programs, Division of Control Systems, EPA-IAG-0020.
500 Includes abstract and summary.
504 Includes bibliographical references (page 112).
650 0 Air +X Purification.
650 0 Fluidization.
650 0 Pollution control equipment.
650 7 Air +x Purification. +2 fast +0 (OCoLC)fst00802185
650 7 Fluidization. *2 fast *0 (OCoLC)fst00928041
650 7 Pollution control equipment. +2 fast +0 (OCoLC)fst01070180
700 1 Jonke, A. A.
Delete Holdings- Export- Label- Submit- Replace-C Report Error- Update Holdings-C Validate-C
Workflow-In Process
about:blank
4/22/2019
-------
3
TABLE OF CONTENTS
Page
ABSTRACT 9
SUtiMARY 9
I. INTRODUCTION 18
II. BENCH-SCALE COMBUSTION EXPERIMENTS 20
¦A. Materials 20
1. Coal 20
2. Additives 20
3. Bed Material 21
B. Bench-Scale Equipment 21
C. Procedure . 23
D. Results and Discussion 24
1. Effects of Variables on Reduction in SO2 Emissions
during One-Stage Combustion 24
2. Effects of Variables on Reduction in Emissions of NO
during One-Stage Combustion 41
•3. 'E'ffects of Variables en "Chemical Ccir.pcGiticn and
Physical Properties of Bed and Elutriated Materials . 42
4. Material Balances 49
5. Exploratory Two-Stage Combustion Runs 50
III. MECHANISM OF THE LIME SULFATION REACTION 54
A. Introduction 54
1. Mechanism - Macro Effects 56
2. Mechanism - Micro Effects 57
B. Experimental 58
1. Microprobe Studies 59
2. Laboratory-Scale Experiments—Reduction of Sulfated
Limestone ..... 66
IV. REGENERATION OF CaO FROM CaSO. 71
4
A. Effect of Pressure ..... 72
B. Experimental 75
V. MODELLING STUDIES, S02 REMOVAL 77
A. Model and Assumptions 77
B. Model. Equations 78
C„ Effects of Parameters on SO^ Removal and Particle
Consumption 81
-------
A
TABLE OF CONTENTS
Page
D. Comparison of Model with Data 84
VI. MISCELLANEOUS 87
A. Control of Chloride Emissions 87
VII. FUTURE WORK 89
VIII. ACKNOWLEDGMENTS 90
APPENDIX A. MATERIALS 91
APPENDIX B. AUXILIARY EQUIPMENT AND INSTRUMENTATION 98
APPENDIX C. MATERIAL BALANCES 107
REFERENCES 112
-------
5
LIST OF FIGURES
No. Title Page
1 Simplified Equipment Flowsheet of Bench-Scale Fluidized-
Bed Combustor and Associated Equipment 22
2 Bottom Section of ANL 6-in.-dia Fluidized-Bed Combustor . . 23
3 Effect of Fluidiz'ecl-Bed Temperature on Sulfur Retention . . 25
4 Sulfur Retention as a Function of Ca/S Mole Racio 29
5 SC>2 Concentrations in Flue Gas for ANL and CRE Combustion
Tests with American and British Coal-Limestone Combinations. 32
6 Effect of Ca/S Ratio on Sulfur Retention for Three Series
of Runs 33
7 Effect of Superficial Gas Velocity on Sulfur Retention ... 36
8 Effect of Excess Oxygen in Combustion Air on Sulfur Retention
Experiments Amer-5C, -5D, and --5E 37
9 Concentration of NO in Flue Gas during Hurap-4 43
10 Laboratory-Scale "Reactors and "Associated Equipment 60
11 Electron Microprobe Traces Showing Relative Sulfur Levels
in Sulfated Particles, Experiment PB-3 63
12 Photomicrograph of Particle Examined in Scan 1, Experiment
PB-3 64
13 Microprobe Scans of Sulfated Particles, Experiment PB-7 . . 65
14 Reaction of Sulfated Lime with CO-CO2-H2O-N2 Mixtures,
Experiment Mech-1 .... 69
15 Reaction of Sulfated Lime with CO-CO2-H2O-N2 Mixtures
Followed by Reaction with Air, Experiment Mech-2 70
16 Equilibrium Mole Fraction SO^ at 1400-2000°F 74
17 Fluidized-Bed Regenerator 76
18 Schematic of Fluidized Bed, Illustrating Model Assumptions . 77
19 Normalized Generation Rate of S0? as a Function of Fractional
Bed Height 80
20 Typical Laboratory Kinetic Data and Interpretation 80
21 Calculated SO2 Removal and Particle Consumption, H = 10 . . 82
-------
6
LIST.OF FIGURES
No. Title Pap.c
22 Calculated SC^ Removal and Particle Consumption, 11 = 30 . . 82
23 Calculated SO^ Removal and Particle Consumption, H = 100 . . 83
24 Calculated SO^ Removal and Particle Consumption, r = 1 . . . 84
25 Calculated Equilibrium Partial Pressure of HC1 as a Function
of Temperature 88
B.l Heating and Temperature Sensing Arrangements for Combustor . 99
B.2 Combustor Cooling Circuit 100
B.3 Preheater 103
3.4 Fluidizing-Combustion Air Flow Control System 104
-------
7
LIST OF TABLES
No. Title Page
1 Average Operating Conditions and Flue Gas Compositions for
Bench-Scale Runs 26
2 Operating Conditions and Results for Sritish-Anerican
Experiments Performed at ANL ^1
3 Effect of Excess Oxygen in Combustion Air on Sulfur
Retention 38
4 Effect on Flue Gas Composition of Addition of Moisture to
the Fluidized Bed, Run Hump-4 39
5 Calcium Utilizations for the SA, BC, and AR Series 46
6 Effect of Cocurrent and Countercurrent Flow of Secondary Air
in Two-Stage Combustion 51
7 Operating Conditions for Experiments PB-3 and PB-7 in Bench-
Scale Combustor 62
8 Extent of Sulfation of Bed Particles as a Function of
..Reaction Time, Experiments PB-3 and PB-7 64
9 Equilibrium Composition for CO and CO2 during Regeneration
of CaSO. 73
10 Definition of Symbols 79
11 Values of Fitted Constants Obtained by Model Fitting .... 86
A.l. Size Distribution of Coals 91
A.2. Chemical Characteristics of Coals 92
A.3. Size Distribution of Illinois Coal and Coarse Limestone No.
1359 Used in Experiment AR-6 93
A.4. Chemical Characteristics and Particle Size Distribution for
Pittsburgh Coal 94
A.5. Chemical Composition of Limestones 95
A.6. Size Distribution of Limestones 96
A. 7. Characteristics of Welbeck Coal Ash 97
B.l. Analytical Methods and Procedures 101
B.2. Flue Gas Analytical Equipment 106
-------
8
LIST OF TABLES
No. Title Page
C.l. Material Balance for Sulfur, Carbon, and Calcium,
Experiments Araer-1, -3, and -4 108
C.2. Material Balance for Sulfur, Carbon, and Calcium,
Experiments Brit-1 to -3, Brit-Amer, and Amer-Brit 109
C.3. Material Balance for Sulfur, Carbon, Calcium, and Magnesium,
Experiments BC-6 to -10 110
C.4. Material Balances for Sulfur, Carbon, and Calcium,
Experiments AR-1, -2, -4, -5, and -6 Ill
-------
9
REDUCTION OF ATMOSPHERIC POLLUTION BY THE
APPLICATION OF FLUIDIZED-BED COMBUSTION
by
A. A. Jonke, G. J. Vogel, L. J. Anastasia, R. L. Jarry,
D. Raniaswami, M. Haas, C. B. Schoffstoll, J. R. Pavlik,
G. N. Vargo, and R. Green
ABSTRACT
Combustion of fossil fuels in a fluidized bed consisting of
partially reacted limestone is being studied to determine the
effect of operating variables on pollutant SO2 and N0X emission
in the flue gas. Coal has been burned with either an excess or
deficiency of oxygen in a 6-in.-dia bench-scale combustor operating
at atmospheric pressure with a bed of partially reacted limestone
(which is the acceptor for the sulfur compounds released during
combustion). The mechanism of sulfur capture has been studied.
Preliminary data have been obtained on conditions for regeneration
of the sulfur-containing acceptor for recycle to the combustor.
A mathematical model has been developed to predict the retention
of sulfur at different operating conditions when coal is burned
in a fluidized bed.
SUMMARY
The Office of Air Programs of the Environmental Protection Agency is
funding an investigation at Argonne National Laboratory of the effects of
variables on the removal of atmospheric pollutants (oxides of sulfur and
nitrogen) generated during the combustion of fossil fuels in a fluidized
bed. The concept involves burning fuel (coal, oil, or natural gas) in a
fluidized bed of particulate solids that react with gaseous sulfur compounds
(and possibly nitrogen compounds) released during combustion. A fluidized
bed is a highly efficient contacting medium for carrying out gas-solid
reactions and for removing heat generated by combustion.
The fluidized-bed combustion equipment at Argonne consists of a 6-in.-
dia, 6-ft-long bench-scale combustor operated in series with a 3-in.-dia,
7-ft-long gas preheater section. Ancillary equipment for operation of the
combustor includes'mechanical powder feeders for introducing coal and
additive into the combustor, gas manifolding for the air supply to the
combustor, cyclone separators and a final filter for removing particulates
from the flue gas, and a flue gas sampling and analytical system.
Bench-Scale Combustion Experiments
Experiments have been performed to investigate the effects of operating
variables on S0-7 and NO levels in die f.Lue gas. Variables affecting the
chemical composition and physical properties of the bed and elutriated
particles have been studied, and material balances have been obtained for
-------
sor.ie experiment?;. il:. ~opr j,: h f-.v runs in which the fluidized bed was
alumina or co;\i y.;,ri i; 1 , i\u: fluidized bed consisted of partially
reacted limestone, which reacts with the sulfur compounds. The combustor
was operated in the single-stage and two-stage modes.
In single- or one-stage combustion, more than the stoichiometric quan-
tity of oxygen (as air) is added to the fluidized-bed combustion zone to
burn the coal to CO2 and water. Typically, the flue gas coming from the bed
contains ^3% oxygen. In two-stage combustion, less than the stoichiometric
volume of air is added to the first stage. The oxygen concentration of the
gas leaving this stage is approximately zero or near zero and, because
relatively large amounts of CO and hydrocarbons are produced, the atmosphere
is highly reducing. As a process concept, the CO and hydrocarbons from this
stage could be burned in a second stage. The second stage could be physically
separate from the first stage but in the ANL combustor consists of the free-
board volume above the fluidized bed. Enough air is injected into the second
stage so that in most runs the oxygen concentration in the flue gas is ^3%.
Major findings in the one-stage combustion experiments were that reten-
tion of sulfur in the bed is strongly influenced by the following operating
variables over the range of the operating variables tested:
1. Fluidized-bed temperature. Maximum sulfur retention
was noted at different temperatures (e.g., 1450 and
1550°F) for different feed materials and operating
conditions.
2. Ca/S mole ratio (ratio of.moles of calcium'in the
additive to moles of sulfur in the coal). Sulfur
retention increases, rapidly db the Ca/S ratio is
increased to 3 and then increases less rapidly
as the Ca/S ratio is further increased.
3. Superficial gas velocity. Sulfur retention decreases
as the gas velocity is increased. At' a Ca/S ratio of
^4, sulfur retention decreases approximately 5% for
each 1 ft/sec increase in gas velocity in the range
2.0 to 7.6 ft/sec. At lower Ca/S ratios, the decrease
in sulfur retention is greater than 5% per ft/sec
increase in gas velocity; at higher ratios, sulfur
retention may be essentially independent of gas
velocity.
Variables that have a lesser effect on sulfur retention over the ranges
studied include the following:
1.
type of coal
2.
coal particle size'
3.
limestone type
4.
limestone particle size
5.
fluidized-bed height
6.
excess combustion air
7.
moisture
8.
solids feeding method
9.
temperature of the upper stage
of the combustor
-------
11
The effects of several operating variables on NO level in the flue gas
were also observed. When water was injected in the base of the bed at rates
representing a range of moisture contents of different coals, the NO concen-
tration in the flue gas decreased from 530 ppm with no water addition to
380 ppm with water added at a rate-corresponding to a coal moisture content
of 51%. It was also observed that NO concentration, normally 300-600 ppm
in one-stage runs, is reduced to less than 80 ppm (only a little above the
equilibrium level expected from the nitrogen-oxygen reaction) in two-stage
combustion runs.
Further details are given below for the experiments that provide the
basis for these observations.
Fluidized-Bed Temperature. In runs with Pittsburgh seam coal and
limestone No. 1359 at a Ca/S ratio of ^4.2 and with a gas velocity of ^2.6
ft/sec, sulfur retention (as a percentage of the sulfur in the feed streams)
increased from 78% to 96-99% as the fluidized-bed temperature was increased
from 1325 to 1450°F, and then decreased to 60% as the temperature was further
increased to 1650°F. Thus, ^1450°F was the optimum temperature for reducing
SO2 emissions when using this coal under the above operating conditions. In
contrast, an optimum temperature of 1550°F was observed in earlier work with
Illinois coal, a different batch of No. 1359 limestone, a lower Ca/S ratio
of 2.5, and a gas velocity of ^3 ft/sec (ANL/ES/CEN-1002, p. 35).
Ca/S Mole Ratio. The effect on sulfur retention of Ca/S mole ratio was
studied in two series of runs using different coals and different additive
particle sizes. Using Pittsburgh coal and -14 mesh limestone No. 135S at a
fluidized-bed temperature of 1450°F, sulfur retention increased from 46% at
a Ca/S ratio of ^1.0 to 96-99% at a Ca/S ratio of 4.2. With Illinois coal
and -14 mesh limestone No. 1359 at a temperature of 1550°F, the sulfur
retentions were 78, 95, and 94%-, respectively, at Ca/S stoichiometric ratios
of 2.5, 4.6, and 5.5.
Data on the effect of Ca/S mole ratio was obtained also in runs with
two 6-in.-dia combustors, one at Argonne National Laboratory and the other
at the National Coal Board's Coal Research Establishment Laboratories (CRE)
in England. In these experiments, Ca/S mole ratios of 1-3 were used at a
bed temperature of 1472°F. Various combinations of American and British
materials (Eritish Welbeck coal, American No. 1359 limestone, Illinois coal,
and British Stoke-on-Kent limestone) were tested at both installations. The
SO2 levels in the flue gas for ANL runs agreed reasonably well with those of
CRE ana showed again that increasing the Ca/S ratio increased the sulfur
retention.
Type of Coal. Data from experiments using limestone No. 1359 at
fluidized-bed temperatures of 1450-1470°F and Ca/S ratios of 1 to 4.3
indicated little difference in results when using either Illinois coal or
Pittsburgh coal at sulfur retentions of less than 90%. Additional data
would be required at other temperatures before a complete analysis can be
made.
Coal Particle Size. In experiments with -12 +50 mesh and -50 mesh
Illinois coaL (a -12 +50 mesh fraction ground to.-50 mesh), a Ca/S mole feed
-------
ratio of 2.4, and a fluidized-bed temperature of 1550°F, the sulfur retention
was 81% and 75% for the two size fractions. These sulfur retentions were
compared with a sulfur retention of 78% obtained in an experiment performed
"with -14 mesh coal at nearly identical operating conditions. The results
indicate a slight benefit from burning coarse coals or coal containing
small amounts of small particles.
Tvpe of Additive. Sulfur retentions for several types of additive were
compared in runs•performed at a Ca/S mole ratio of 2.5. With limestone
No. 1359 at 1550°F, sulfur retention was 78%, which is comparable to that
obtained with limestone No. 1360 and dolomite No. 1337 at combustion
temperatures of 1550-1600°F. This indicates that differences in these
additive types had minimal effects at the temperatures studied.
In runs with American and British coal-limestone combinations, the
reactivity of the British and American limestones was compared. Results
showed that the particular type of British limestone used removed a greater
fraction of the SC>2 from the combustion gases than did the American limestone,
but the effect was not large.
Additive Particle Size. Limestone No. 1359 of lOOO-yro average particle
diameter and limestones No. 1359 and 1360 and dolomite No. 1337 of 490- to
630-vim particle size were compared at a Ca/S mole ratio of 4.0 and a super-
ficial gas velocity of 3.0 ft/sec. (Some sulfur retention values were
calculated by extrapolation or interpolation of experimental values.) Sulfur
retention was ^87% for the larger-particle additive and 93% for the smaller-
particle additives, indicating that additive particle size, within the range
tested, has only a moderate effect on sulfur retention at a particular
fluidizing-gas velocity.
Fluidized-Bed Height. To determine the effect of fluidized-bed height
on sulfur retention, three one-stage combustion runs were made under similar
conditions but with fluidized-bed heights of 14, 24, and 46 in. Sulfur
retentions were 80, 81, and 85%, respectively, indicating only a minor
effect of bed height.
Superficial Gas Velocity. The effect of superficial gas velocity in
the combustor on sulfur retention was examined in several series of runs.
As an illustration, in one series of combustion experiments with Illinois
coal, coarse particles of limestone No. 1359 with diameters averaging 1010 um
were fed to the fluid bed to ensure their retention in the bed at the highest
gas velocity tested. The tests were made at 1550°F with a Ca/S mole ratio
in the feed of ^4 and at gas velocities of 3.5, 5.5, and 7.4 ft/sec. Sulfur
retentions were 83, 71, and 63%, illustrating an important effect of gas
velocity.
Excess Combustion Air. The effect on sulfur retention of oxygen concen-
tration in the flue gas (controlled by varying the volume of excess oxygen
in the feed) was investigated for oxygen concentrations in the flue gas of
^1 to ^5 vol % (5.5-32% excess, air). An experiment was performed at 1550"F,
using limestone No. 1359 additive, Illinois coal, and a Ca/S mole ratio of
^2.8. In this experiment, sulfur retention was 67% with 0.7 vol % oxygen
in the flue gas, 71% with 2.4 vol % oxygen in the flue gas, and 75% with
-------
13
5.6 vol % oxygen in the flue gas, showing a slight enhancement in sulfur
retention at higher excess air. .
Moisture Content of Coal. The effect of moisture content of the coal
w:-is 5X3-j.r.rii by .Lr.j<. -ting at different rates into the base of the
fluidized bed during combustion of Pittsburgh coal with limestone No. 1359
additive. (Pittsburgh coal contains 1.8% water and Illinois coal 10.1."
water.) The fluidized-bed temperature was 1450°F and the Ca/S ratio was 1.
Sulfur retention was not significantly influenced by moisture additions
equivalent to burning coal with moisture contents up to 51 wt %; the concen-
trations of NO, CH^, and CO in the flue gas decreased when moisture was
added.
Solids Feeding Method. The difference in solids feeding methods for the
British and American combustors was considered a possible reason for SO2
levels in the flue gas being slightly higher (by 100-200 ppm) in tests at
ANL than in tests at the Coal Research Establishment (CRE) with identical
coal and limestone. Premixed coal and limestone were fed through a single
line into the British combustor; coal and limestone were fed through separate
lines into the ANL combustor. When a mixture of American coal and limestone
was fed into the ANL combustor, sulfur retention did not differ significantly
from that obtained when coal and limestone were fed separately, indicating
that slight differences in ANL and CRE data cannot be attributed to the
feeding method.
T 6tr_p0 q f the Fr^ebo^rd. Tssts v.7 £ it s psirircrmsd L c determine if sriy
app til Ldb i.e traction of SO2 reaction with additive was occurring in the gas
space above the fluidized bed. In one experiment, the upper section of the
ANL combustor (above the fluidized-bed combustion zone) was insulated,
thereby increasing the temperature in that zone to 1340°F as compared with
1020°F in the absence of insulation; the temperature of the fluidized bed
was 1450°F. Insulation of the upper section had essentially no effect on
sulfur retention but resulted in lower concentrations of CH^ and NO in the
flue gas. These results suggest that sulfur-removal reactions occurred
essentially in the fluidized-bed zone, but that combustion of CH^ and
decomposition of NO increased when the upper combustor section was at 1340°F.
Insulation of the upper section of the combustor also slightly increased the
amount of CO burned.
Absence of Additive. Experiments were performed to determine how much
of the sulfur in coal is actually released as S09 during combustion, since
sorre sulfur can be retained by ash constituents." Illinois and Pittsburgh
coals were burned in a fluidized bed of refractory alumina with no additive
present and at a nearly constant coal feed rate (about 4 lb/hr). With
fluidized-bed temperatures ranging from 1325 to 1650°F, the concentrations
of S0.; in the flue gas were 1900 to 2200 ppm when Pittsburgh coal (2.4 wt %
S) was burned and 3750 to 4250 ppm when Illinois coal (3.7 wt % S) was burned.
The observed S09 levels when Pittsburgh coal was burned were 150-300 ppm
less than the levels calculated for complete release of all sulfur as SO2.
For Illinois coal, the observed S0? levels and the calculated levels for
coinpLete release of sulfur as S09 were approximately the same.
-------
NO Content of Flue. Gas. To determine the effect of additive on NO
content of the flue gas, fi-ve experiments were performed with Pittsburgh
coal at temperatures of 1325-lf>50°F, with no additive present, a coal feed
rate of 4 lb/hr, and fluidizin^ gas velocities of 2.4 to 2.8 ft/sec. NO
concentrations in the flue gas ranged from 450 to 630 ppm.
In experiments with Illinois coul, limestone No. 1359 additive, a
fluidized-bed temperature of 1550°F, a Ca/S mole ratio of -^4, and super-
ficial gas velocities of 3.5 to 7.4 ft/sec, nitric oxide concentration in
the flue gas ranged from 320 to 470 ppm. This is an indicated reduction in
NO emission of 20 to 45% as compared with NO emission in the absence of
additive.
The effect of moisture content of the coal on NO level in the flue gas
was studied by adding water to the fluidizing air at different rates during
the combustion of Pittsburgh coal. Limestone No. 1359 additive was used at
1450°F and a Ca/S ratio of 1. The concentration of NO decreased from 530 ppm
to 510 ppm upon adding 10 cc/min water (equivalent to 26 wt % water in the
coal) and to 380 ppm when the rate of water addition was increased to 30
cc/min (equivalent to 51 wt % water in the coal).
Chemical Composition and Physical Properties of Bed and Elutriated
Materials¦ To determine the extent of calcination (to CaO) of bed and
elutriated materials, samples from several series of bench-scale experiments
were analyzed for calcium, sulfur, and carbonate contents. Calcination of
the limestone in the fluidized-bed and final-filter solids was essentially-
complete (95% or greater); for solids in the cyclones, calcination ranged
from 60 to 84%.
The extent of calcium utilization (i.e., conversion of CaO to CaSO^) in
the bed and elutriated materials from several series of runs was determined
from analytical results. A higher degree of calcium utilization in the
fluid-bed solids was observed in a series of runs utilizing a small-particle-
size additive (25-103 pm) than in runs using larger particle size additive
(average particle sizes of 44-650 ym; 490 urn; 1640 um). The Ca/S mole ratios
in fluidized-bed and final-filter solids (0.9-3.4) were near the Ca/S mole
ratios of the input coal and limestone streams (1-3.6). The Ca/S mole ratio
for cyclone solids was as high as 11, indicating a low degree of sulfation
of this material (probably due to a shorter residence time in the comb us tor
reaction zone).
To measure any changes in particle size distributions during the runs,
particle-size analyses were performed on fluidized-bed solids from several
experiments. Results showed that particle size distributions of bed samples
taken during a run were related to the size distribution of the initial bed
material and were similar to the size distribution of.that fraction of the
fresh additive that was not capable of being elutriated at the gas velocity
existing in the combustor.
The decrepitation and attrition of several additives during coal
combustion has been estimated by comparing the quantity of small-particle
additive fed in a run to the quantity of additive elutriated. Results
indicated that the decrepitation rates of limestone No. 1359 and British
limestone are low. Limestones of this type are, therefore, desirable
-------
15
materials for use in a full-scale fluidized-bed combustor with regeneration
and recycle of additive. High carryover of additive was observed when
limestone No. 1360 and dolomite No. 1337 were used, indicating that these
may be less promising materials for regeneration and recycle.
The bulk and tapped densities of samples of solids from the fluidized
bed, cyclones, and final filter were determined. Bulk densities of the
cyclone and final-filter solids were low, 0.15-0.77 g/cc, probably because
of the nature of the solids, which consisted of flyash and/or limestone fines.
The bulk and tapped densities of the fluidized-bed solids were very similar
to those of the precursor solids (limestone and coal ash).
Sulfur, carbon, and calcium material balances have been completed for a
large number of runs. Material balances ranged from 83 to 114%.
Two-Stage Combustion Experiments. In two-stage combustion, a sub-
stoichiometric volume of combustion air is fed at the base of the fluidized
bed, and a second volume of air is added at a point 6 in. above the bed to
burn CO and hydrocarbons in the gas leaving the fluidized bed. In the first
two exploratory runs, the secondary air was injected downward (toward the
bed and countercurrently to the fluidizing-gas flow); in two other exploratory
runs, air was injected upward (away from the bed and cocurrently with the
fluidizing-gas flow). These runs were made with Illinois coal and limestone
No. 1359 additive. The Ca/S mole ratio was ^3, and the fluidizing-gas velocity
was ^1.9 ft/sec. The starting bed was partially sulfated and calcined
limestone No= 1359,
The results for the two-stage runs with upward flow of secondary air
shoved a rri.ixir.u.- S0.; removal of 91% at a Ca/S ratio of 3.0. The NO level
in t:ie Hmz gas was 70 ppm, the lowest value recorded at ANL thus far.
Downward flow of secondary air resulted in poor SO2 removals. Possibly,
the downward-directed air contacted the bed solids releasing S0~ by reaction
with calcium sulfide. Sulfide has been qualitatively identified in bed
samples from these runs.
Mechanism of Lime Sulfation Reaction
Studies have been made on the lime-S02 reaction mechanism, specifically
on the mode of SO^ penetration into lime particles in fluidized beds. In
this laboratory, it is believed that the mechanism is essentially the shell-
formation model enhanced by localized reducing conditions in the dense phase*
of the fluid bed. Because of a large demand for oxygen by the carbon com-
bustion reaction near the bottom of the combustor, the emulsion phase in this
location is depleted in oxygen relative to the bubble phase. As a result,
¦k
The dense or emulsion phase is the portion of the bed where gas flows in
intimate contact with the solids. The remainder of the gas passes through
the bud in the form of bubbles, but as the bubbles rise through the bed,
gas circulates from the bubbles into the emulsion phase and back into the
bubbles. The gas in the bubbie phase is available for reaction only as
rapidly as it circulates into the emulsion phase.
-------
Che CO concentration is relatively high at some locations in the dense
phase where the SC^-lime reaction takes place.
ANL work, reported here on the mechanism comprises (1) electron micro-
probe examination of sulfated lime particles taken from elutriated solids
and from combustor beds and (2) laboratory-scale experiments with sulfated
lime to measure S07 evolution under reducing conditions. Microprobe
examination of particles taken from the combustor bed during coal-combustion
experiments showed that some particles had nearly uniform sulfur distributions
across the particle diameter and other particles had deep sulfur penetration
(but not all the way to the center).
Microprobe examination of particles taken during an experiment in which
calcined limestone was exposed to an SO^-air mixture (with no combustion
occurring) showed that this sulfur profile was radically different from the
sulfur profile of particles from a coal combustion experiment in which all
operating conditions other than the presence of combustion were similar.
Sulfur formed a very thin shell on the surface of the particles from the
former experiment. Furthermore, the thickness of this layer did not change
appreciably with increasing reaction time and was considerably thinner than
in lime particles from the coal combustion experiment. The conclusion from
these microprobe examinations is that sulfation reactions during the combustion
of carbon differ from sulfation reactions in the absence of combustion.
Although it cannot be said that the mechanisms of the reactions are different,
the extent of the sulfation reaction is considerably greater for the combus-
tion case. This lends support to the hypothesis that local reducing conditions
existing during combustion produce mobility of the SO2 for penetration into
the lime particles.
Laboratory-scale experiments to determine the effect of CO on the
sulfation reaction showed that SO2 can be released readily from sulfated
lime at about 1750°F, using CO as the reductant. At lower temperatures
(^1650°F), the reduction reaction apparently produces CaS rather than CaO
and SO2 since SO2 is released in high concentrations when air is passed
through a bed of sulfated lime that has been treated with CO at the lower
temperature. In experiments in which the temperature of the reaction was
1500°F, the rate of reaction to form CaS was nearly zero.
Regeneration of CaO from CaSO^
Since efficient SO2 removal from the gas phase in coal combustion will
probably require relatively large quantities of limestone, it will be
desirable to regenerate the partially sulfated lime for recycle and to
recover the sulfur value as sulfuric acid or elemental sulfur. Thermodynamic
equilibria for the regeneration of sulfated lime using CO have been analyzed.
Operation at 10-atm pressure is being considered for both fluidized-bed
combustion and regeneration. The equilibrium values for the reaction of
CaSO^ with carbon monoxide, derived from free energy data, were used to
calculate the SO2 partial pressure at 1- and 10-atm system pressures over
the temperature range 1400-2000°F. The S02 generation reactions are pressure
dependent, and increasing the system pressure reduces the SO2 concentration
proportionately.
-------
17
Laboratory-scale experimental work has been started to examine the
utility of various regeneration reaction schemes. Two preliminary batch
fluidized-bed experiments (in a 2-in.-dia reactor) were performed at 1950°F
wit.ii a superficial gas velocity of 2 ft/sec and an inlet gas stream contain-
ing 10 vol Z CO, 20 vol % C09, ana nitrogen (with oxygen added to burn CO
and maintain the desired c.or.i>ustor temperature) . One experiment was performed
at a pressure of 1 atm and the second experiment at 10 atin.
At 1-atm pressure, the maximum SC>2 concentration in the effluent gas
phase was about 3 vol %. Equilibrium calculations predicted a concentration
value of about 8 vol %. About 80% of the sulfur contained in the original
sample of partially sulfated limestone was evolved as SO2. Of the sulfur
"remaining in the particulate residue, about 90% was present as the sulfide,
CaS.
In the experiment performed at a total pressure of 10 atm, it proved
difficult to maintain a constant temperature, and a temperature excursion
resulted in partial caking of the bed material. The little SO2 evolved
during this experiment was released mainly during the heatup period.
Modelling Studies, SO., Removal
A mathematical model was devised for predicting, from the fluidized-bed
reaction parameters, sulfur retention in a fluidized bed by heterogeneous
reaction with additive limestone" particles. The model was fitted to experi-
mental. data and gave reasonably good agreement with observed sulfur retention.
Control of Chloride Emission
To examine the potential of fluidized-bed combustion for the'control of
chloride emission, thermodynamic equilibria were examined for several
reactions in which hydrogen chloride is reacted with solids. Hydrogen
chloride is produced (1) during combustion of coal (since most coals in the
United States contain chlorides) and (2) during the incineration of chloride-
bearing wastes such as polyvinyl chloride plastics. At a combustion tempera-
ture of 1500°F and with 10 vol % water present, reaction of HC1 with lime
does not reduce the hydrogen chloride concentration in the flue gas below
1000 ppm. At these conditions, BaO, KOH, and NaOH remove HC1 more effectively
than does lime, and MgO is less effective than lime. Small amounts of BaO,
KOH, or NaOH along with limestone possibly would remove HC1 from the gas
phase.
-------
18
I. INTRODUCTION
Methods are being developed for lowering the concentrations of noxious
pollutants emitted from power- and steam-producing plants to meet standards
set by state and federal governmental agencies. Progress is reported here
on a continuing study of the removal of pollutant (S0?, NO, particulate)
from the gas phase in the combustion of fossil fuels tsuch as coal, oil, and
natural gas). The concept investigated involves introducing fuel and a
sulfur-reactant additive material such as crushed limestone into a hot
fluidized bed of solids. Sulfur dioxide is generated when sulfur-containing
fuel burns in an excess oxygen atmosphere and reacts with calcined limestone,
forming solid CaSO^ particles.
During the first two years of this investigation, the effects of
independent variables on the reduction of SO2 and NO levels in the flue gas
have been studied at this site with a 6-in.-dia bench-scale combustor
operated at atmospheric pressure. A fine (<44 mesh) limestone additive was
generally used, and these data were supplemented by S07-additive reaction
rate data obtained in boat reactor experiments. Coal was the fuel burned
in most tests, but in a few, natural gas was combusted to demonstrate the
low NO levels achievable and to show that a source of the NO formed during
coal combustion is nitrogenous compounds in the coal. The combustor was
operated as a one-stage unit in this earlier work; in one-stage combustion,
all of the combustion air is added to the fluidized-bed region, where
essentially all combustion takes place in the presence of excess oxygen.
The ANL work reported here consisted of three major parts— (1) bench-
scale experiments to study the effects of variables and combustion"ftode on
sulfur retention in the fluidized bed, NO removal, and bed particle composi-
tion and size; (2) laboratory-scale investigations of the mechanism of lime
sulfation, and (3) the development of models to allow the extent of sulfur
retention to be predicted from the fluidized-bed reaction parameters.
Work at the Chemical Engineering Division of ANL during the past three
years has complemented studies of other contractors of the Environmental
Protection Agency's Office of Air Programs (OAP)—British National Coal
Board; Esso Research Center, Ltd.; Esso Research and Engineering; Pope,
Evans, and Robbins; Consolidation Coal; U.S. Bureau of Mines; and
Westinghouse. The program developed by OAP should result in lcwer pollutant
concentrations in the air and secondarily may result in a more economical
combustion process.
The objectives of the development program at ANL are:
1. To determine how sulfur retention and limestone utilization are
affected by independent fluidized-bed operating variables, such as bed
temperature, combustor pressure, gas velocity, oxygen concentration, bed
height, calcium-to-sulfur ratio, type of additive and coal, and additive
and coal particle size.
2. To determine the effect of operating vatiables on the level of NO
in the flue gas.
-------
19
3. To compare, as a function of the independent operating variables,
the characteristics and quantities of flyash generated in fluidized-bed
cjrnb tors and in conventional pulverized-fuel boilers.
4. To determine the effects of independent operating variables on the
regeneration of sulfated additive material and to investigate various methods
of regenerating this additive at elevated pressures.
5. To determine how effectively the regenerated additive removes SC^
from the flue gas.
6. To determine the chemical and physical mechanisms involved in
•a. -removal of SC^ by additive
b. NO formation
c. regeneration of additive
-------
20
II. BENCH-SCALE COMBUSTION EXPERIMENTS
The bench-scale comb us tor has been operated in the single-stage and
two-stage modes in experiments to investigate the effects of operating
variables on sulfur retention in the fluidi.;cd bed and NO removal from the
flue gas. Variables affecting the chemical composition and physical
properties of bed and elutriated particles have been studied, and material
balances have been obtained for some experiments.
Additional data have been obtained for the BC- and AR-series of experi-
ments discussed in the preceding annual report (ANL/ES/CEN-1002). New
experiments performed were the Amer, Brit, Hump, HP, and PBY series.
A. Materials
1. Coal
The coals used in the various series of experiments were Illinois
coal from Seam 6, Peabody Coal Co. Mine 10, Christian County, Illinois
(furnished by Commonwealth Edison); Pittsburgh Seam coal from Humphrey
Preparation Plant, Osage, West Virginia; and Welbeck coal (furnished by
the British National Coal Board).
The Illinois coal used in the experiments contained 3.77 wt %
sulfur and generally was crushed to pass a -14 mesh sieve. The sieve
analysis of this coal is given in Table A.1 of Appendix A of this report,
and its chemical characteristics in Table A.2. For experiments PB-5R and
-6R. coal from this batch was crushed and screened; the -12 +50 mesh fraction
. j -• —fj_j j * j r\ r¦on CD rpi ^
w ao ui vi uc u iiLLu twu ^uj. lxuiio • unc pui.cj.un woo j.ou .lu CA^cj-xuicub » mc
remaining portion was crushed and screened again to obtain a -50 mesh fraction,
which was fed in experiment PB-6R. The size distribution of the Illinois
coal used in experiment AR-6 is given in Table A.3.
Size distributions and chemical characteristics of the Welbeck
coal (containing 1.2 wt % sulfur) are given in Tables A.1 and A.2. The
proximate analysis, ultimate analysis, and particle size distribution of
the Pittsburgh coal which contained 2.4 wt % sulfur are given in Table A.4.
2. Additives
Limestone No.. 1359, used in many experiments performed during the
report period, was obtained from M. J. Grove Lime Co., Stephens City, Va.
and contained 95 wt % CaCO^ and 1 wt % MgCO^. The particle size distributions
of limestone No. 1359 before and after passage through the screw feeder are
given in Table A.3. Attrition of the limestone during passage through the
volumetric screw feeder resulted in the average particle diameter decreasing
from 1637 ym to approximately 1000 ym. The limestone used in experiment
AR-6 is that identified as "Limestone after Passage through Screw Feeder."
The chemical compositions and size distributions of the limestones
used in the Amer-Brit series experiments are given in Tables A.5 and A.6.
That supplied by the British National Coal Board (J. Gregory & Son, Stoke-on-
Kent, Staffordshire) contained, on an as-received basis, 95.6% CaC03 (average
of two analyses).
-------
21
3. Bed Material
In most of the bench-scale experiments performed during the report
period, the starting bed was partially calcined and sulfated limestone from
the preceding experiment. Some experiments utilized limestone No. 1359 as
the starting bed.
In the experiments employing British coal and limestone, the
starting bed material was ash from Welbeck coal. A chemical analysis and
a sieve analysis of this ash are listed in Table A.7. Comparison of the
chemical analysis for the ash (Table A.7) with that given for the mineral
content of Welbeck coal ash (Table A.2) shows reasonable agreement.
In experiments in which coal was burned with no additive present,
the starting bed consisted of fused alumina (Type T-61, 30-mesh, manufactured
by Alcoa).
B. Bench-Scale Equipment
The fluidized-bed combustion equipment consists of a vertical 6-in.-
dia, 6-ft-long fluid-bed combustor, a 3-in.-dia, 7-ft-long preheater for
the fluidizing-combustion air (Appendix B, Section 3), vibratory screw
feeders for introducing particulate coal and additive into transport air
streams that carry these solids to the combustor, gas manifolding for the
air supply to the combustor, associated heating and cooling arrangements
.and controls, the fIuidi-Zj.ng~cornbu3tiGri ai,r .supply (Appendix 3(, Section hj ,
temperature and pressure sensing and display devices, solids and gas sampling
•arrangements, a data logger (Appendix B, Section 7), and two cyclone
separators and a"glass fiber final filter for solids removal from the flue
gas. A sample stream of flue gas is routed to a gas analysis system
(Appendix B, Section 6). The gas analysis section provides in-line measure-
ment of the flue gas components SC^, NO, CH^, CO, and 0^ on a continuous
basis and of CO2 on an intermittent basis.
Figure 1 shows the components of the combustion system. The ANL
fluidized-bed combustor is described in detail below, and auxiliary equipment
in Appendix B. The gas analysis system is described in ANL/ES/CEN-1002,
pp. 21-22.
The 6-in.-dia combustor is 6 ft in length and consists of three sections.
The middle section (where combustion occurs) is 4 ft long, the top section
is 2 ft long, and the bottom section (which contains the gas distribution
place) aoou; c in. long. The combustor was constructed principally of type
304 stainless steel.
The 4-ft middle section is fitted with a flange at each end for connection
to the lower and upper sections. Four annular chambers 2 1/2 in. high and
8 in. in diameter are spaced 3 1/2 in. apart on the outer surface of the
middle section, with the bottom chamber 3 1/2 in. above the bottom flange.
These chambers are welded to the combustor section. Mr or an air-water
mixture is circulated through these annular chambers, permitting heat removal
fron specific ifones of the reactor. Resistance heaters installed between
the annular chambers assist in preheating the fluid bed at the start of an
-------
22
experiment and in maintaining the desired temperature during coirlnis tion.
The heating-cooling system is described in detail ia Appendix B, Section 1.
At intervals along the length of the combustor, access is provided to its
interior for thermocoupl.es, a solids sampling probe, pressure taps, and
solids injection probes. Solids injection and removal are discussed in
Appendix B, Section 2.
O2 ^2 AlR
COMBUSTOR
ADDITIVE fecder
COA
FE
J1
n
EDER Y7
1
J
T>
ri
irria
T
¦9
0
TO
CAS-ANALYSIS
SYSUM
I TO C-LtSS FIBER
r *" FI?JAL ( ILTCfl AND
tVENTILATION EXhAUST
1 K
\y N SECONDARY CVCUJNE
PRIMARY CYCLONE
Fig. 1. Simplified Equipment Flowsheet of Bench-Scale
Fiuidized-Bed Ccmbustor and Associated Equipment
The upper 2-ft-long section is flanged at each end; the upper flange
closure contains a sight glass mount. Surrounding most of this section is
an annular chamber 12 3/4 in. long and 8 3/8 in. in diameter through which
air or air-water mist can be circulated for temperature control. Heat
transfer from the gas inside the combustor to the metal wall is facilitated
by four fins (1 3/4 in. by 16 in. by 1/16 in.) welded to the interior
surface. From the upper section of the combustor, the particle-laden flue
gas flows to the cyclone separators and the glass fiber final filter for
solids removal. The flue gas filtration system is described in detail in
Appendix B, Section 5.
The third section (Fig. 2) of the 6-in.-dia combustor contains the'
gas distributor and is flanged for connection to the bottom of the 4-ft
middle section. On this blind flange are mounted up to 15 bubble-cap type
gas distributors, each with eight l/16-in.-dia holes. The bubble caps are
removable so that pressure drop across the gas distributor may be varied if
desired. The preheated fluidizing-comb ustion air enters the plenum below
the bubble-cap flange through a pipe at the left in Fig. 2. The welded pipe
nipple opposite the flanged pipe furnishes entry for a thermocouple and a
pressure tap. A 1 1/2-in. pipe used for withdrawing bed material from the
reactor passes through the bottom plate of the plenum and its top is at the
face of the bubble-cap plate.
-------
23
Fig. 2. Bottom Section of ANL 6-in.-dia
Fluidized-Bed Combustor
An air feed tube was incorporated into this combustor during some
experiments in which the two-stage combustion technique was explored. In
two-stage combustion, a deficiency of air is added to the fluid bed through
the gas distributor; enough air to complete combustion (i.e., secondary air)
is added in the region above the fluidized bed. Feeding of the secondary
air was through a 0.25-in.-ID tube installed through the top flange of the
combustor so that the tip of the tube was concentrically located 6 in.
above the top surface of the fluidized bed. (A 0.25-in.-ID tube was the
largest that could be installed without major modification of the combustor.)
C. Procedure
The startup procedure for one-stage coal combustion experiments at ANL
involves heating the fluidized bed in the bench-scale combustor to about
800°F by passing air preheated to ^900°F through the bed and simultaneously
employing resistance heaters located on the outer surface of the combustor.
When the bed temperature reaches 800°F, coal entrained in a transport air
stream is injected into the limestone bed and ignites. The burning coal
rapidly increases the bed temperature (within one-half hour) to the selected
experimental temperature. The selected combustion temperature is maintained
within 10°F of the set point (i.e., heat generation is balanced with heat
loss from the system) by heating with the external resistance heaters or
cooling with air passed through the annular chambers surrounding the middle
section of the combustor.
Injection of limestone additive is initiated soon after coal is introduced.
The limestone, as is the coal, is carried to the combustor entrained in a
transport air stream. Limestone is injected at a feed rate to give the
specified Ca/S mole ratio in the feed streams.
To keep the fluidized bed at tiie specified height, material is removed
from the bed, generally once per hour. The quantity of material rcr.oved
from the bed depends on (1) the quantities of limestone and coal fed and
-------
24
(2) the quantities of solids elutriated as flyash and removed as bed
samples.
As is stated above, the concentrations of Oy, SC^, NO, CO, and CH^ in
the flue gas are monitored and recorded continuously, and C09 concentration
is determined periodically. Attainment of steady-state conditions for SO2
removal from the flue gas is indicated by a relatively steady S09 concen-
tration.
At intervals during a steady-state SO2 removal period and at the end
of an experiment, the fluid bed is sampled and at the same time, the
elutriated solids collected in the primary and secondary cyclones are
sampled. The number of times the bed and cyclone solids are sampled depends
on the duration of an experiment.
The solids samples are generally analyzed for C, Ca, S, and CO3.
Material balances are made based on these analyses to determine carbon
burnup efficiency, degree of limestone calcination, limestone utilization,
and SO2 removal.
In two-stage combustion experiments (results are discussed in
Section II.D.5), startup is similar to startup in one-stage combustion
experiments. Initially, all air is fed at the base of the fluid bed, then
the coal feed rate is increased until the oxygen concentration in the flue
gas is approximately 2%. The flow rate of primary air fed at the base of
the fluid bed is reduced until the oxygen concentration in the flue gas
reaches zero, then reduced an additional amount so that a substoichiometric
volume of air is being fed to the fluid bed. Flew of secondary air through
a tube whose tip is in the freeboard space above the bed is then started,
and the oxygen concentration in the flue gas is increased to the desired
value.
D. Results and Discussion
The effects of operating variables on sulfur retention in the bed and
NO removal during one-stage and two-stage combustion have been studied in
experiments using the bench-scale combustor. The effects of variables on
the chemical composition and physical properties of bed and elutriated
materials have also been investigated. Material balances have been made
for several experiments.
Results of Argonne combustion runs are reported as "sulfur retained."
That is, the sulfur contained in the SO2 in the flue gas (measured
instrumentally) is compared with the quantity of SO2 that would be generated
from all of the sulfur in the coal fed. Some of the sulfur remains in
solid form during a run (e.g., the sulfur contained in fine coal particles
that are elutriated out of the bed), and this sulfur content is included in
"sulfur retained" values.
1. Effects of Variables on Reduction in SO2 Emissions during One-Stage
Co'.r.bus ti on
The variables investigated have been fluidized-bed temperature,
Ca/S mole ratio in the feed, combustor design, type and particle size of
-------
25
coal and additive, fluidized-bed height, superficial gas velocity, excess
air, moisture content of coal, solids feeding method, temperature of the
gas in the freeboard above the fluidized bed, and absence of additive.
Operating conditions and results for most of the experiments
completed during the past year have been tabulated as a computer readout
and are presented as Table 1.
a. Fluidized-Bed Temperature. Five runs (Hump-IA to -IE) were
performed with Pittsburgh coal and limestone No. 1359 (609-ym) to study the
effect of fluidized-bed temperature on the percentage of sulfur retained.
The temperature of the fluidized bed was varied between ^1325 and 1650°F,
the gas velocity was-2.5-2.9 ft/sec, and the Ca/S ratio was ^4.2. The
data on sulfur retention (as a percentage of the sulfur input) are plotted
in Fig. 3. The sulfur retention increased from ^78% to 96-99% as the
temperature was increased from 1325°F to 1450°F, then decreased to ^60% as
the temperature was further increased to 1650°F.
In contrast, with Illinois coal, a different batch of No.
1359 limestone (490-pm particle size), a gas velocity of ^3 ft/sec, and at
a lcwer Ca/S ratio of 2.5, the optimum temperature was 1550°F (ANL/ES/CEN-1002,
p. 35) as shown in the curve also plotted in Fig. 3. The shift in peak
temperature could be caused by the different characteristics of the coals,
Ca/S ratios, or an unknown factor.
100
90
BO
i 70
S 60
40,
30
O ILLINOIS COAL. Co/S-2 5
• PITTSBURGH COAL . Ca/S-4 0
20
1300
1400
1600
1700
TEMPERATURE. *F
Fig. 3* Effect of Fluidized-Bed Temperature on Sulfur Retention
Gas Velocity = ^3 ft/sec
Limestone size with Illinois coal: 490 urn
Limestone size with Pittsburgh coal: 609 ym ¦
-------
ro
a>
TABLE 1. Average Operating Conditions and Flue Gas Compositions for Bench-Scale Runs
TYPE OF CA/S STEADY STATE FLUE CAS CONC. I DRY)
FEEDRATES BED BED GAS
EXP. NO.
COAL¦ADD.
COAL .
ADD. ,
HEIGHT.
T EMP.
VEL .
02.
C 02 .
NO.
SO 2.
CH4.
CO.
® / HR
»/HK
IN.
F
FT/SEC
V/O
V/O
PPM
PPM
PPM
PPM
BC-7
2 1
5 . 4
1 . 8
2:30
24
1600
2 BO
0.0
0. 0
220
840
296
2 040
BC-8
2 1
4.3
1 . 6
2.50
24
1600
2 . 80
0.0
0. 0
320
800
72
1525
BC-9
2 6
4 . 1
1 . 7
2. 30
24
" 1600
2 . 70
0.0"
0.0
"4 00"""
1100
1 1 3
"1000"
ec-1oa
2 7
4 . i
2 . 6
2.2 0
24
1600
2 . 90
0. 0
0 . 0
250
80 0
82
8 70
BC-10B
2 7
4 . 3
2.6
2.20
24
1 480
2 . 80
0. ')
0 0
180
4 20
82
8 70
BC- IOC
2 7
4 .3
2 . 6
2. 2C
2 4
180 0
3 70
o.o"
0.0"
- 3 00 —
"36 5 0 "
" 82
~ " 870
AR- 1 A
2 1
"4.6
1 . 7
2.5 0
24
1400
2 . 50
0.')
0. 0
2 5-)
2570
2 30
4 200
AR-1B
2 I
4 . b
1 . 7
2.50
24
1 4b 0
2 . 70
0.0
0 0
280
1460
180
4 000
AR- 1 C
2 " 1
4 . 6
1 . 7
2.5 0
24
" 1 500
2 . 80
"" 0". 0
0. 0
420"
42 0
70"
"1 050
AR-1 0
2 1
4 . 6
1 . 7
2. 5 0
24
1550
2 . 9.)
u.0
0. J
4 2':
420
40
t 1 00
AH-1 E
2 1
4. b
1 . 7
2.5 0
2 4
1600
2 RO
0. )
0. 0
44 0
700
50
i 200
AR-1F
2 1
4". 6
1.7 -
2.50
2 4
" 1400
' 2 . 60 '
0. 0"
"0.0"
"260"
245 0""
"""2 30
200
AR-2 A
2 1
4 . i
1 . 6
2.60
24
1 550
2 80
0. .)
0 0
390
470
282
? 0 50
AR- 2 R
2 1
4. 4
1 . 7
2.60
24
1550
2 . 50
0. 0
0 . 0
35 )
730
2 82
^ 050
AR-2 C
2 ' 1
4. 5
1.7
2.60
24
1550
"' "2 .<>0
""0.0
""""0.0 ""
3 00"
"12 5 0""
282
"TO 50"
AR-2 0
2 1
4 . 3
1 . 8
2. 60
24
1550
3.10
0. '1
0. 0
4 30
85 0
28 2
2 050
AR-4
3 1
4 . 2
1 . 6
2.80
24
1550
2 80
0.0
0.0
310
75 0
150
2 000
AR- 5 A
3 1
3.6
1 . 6
3.00
24
" 1550
2: 80
0. ¦')
""0.0"-
37 0
1100 "
0
" 6"
AR- 5 8
3 1
4 . 4
3. 1
5.5 0
24
1550
2 . 80
0. .)
0 . 0
3 50
200
1 00
1 200
AR-S C
3 1
4 . 1
2.4
4.6 0
24
1 550
2 . 80
0. 0
0 . 0
440
160
1 50
1 600
AR-5 D
3 1
4. 1
1 . 3
2.5 0
24
1 55 0
_ 2 . 80"
""""0 . T"
... Q-y-
"3 3"1""
"72 0 ""
1 70
"16 00-:
AR-b C
3 1
5 . 6
3 - 5
4.20
24
1 55 0
7 . <.0
0. 0
0. 0
4/0
150 0
100
1 700
AR-6D
3 1
10.2
5 . 3
4.10
24
1 550
3 50
0.0
0. 0
34 0
77 0
¦ 500
3 000
AR - 6 E
3 1
13.8
7.4
3 .80
24
1550
5 .50
"""0.0
""""0." 0 "
320 "
i 2 5 0 "
"""700
" 4 100
AHER- I
3 3
4 . 5
0. 5
1. 00
24
1470
2 . 50
0. 0
0. 0
240
2630
300
3 750
AMt R - 3
3 3
4 . 6
1 . 6
2.8 0
2 4
1470
2 SO
0 . 0
0 . 0
26 0
1120
7 0 0
3 000
AMfcR-4
3 3
4 . 6
1.0
1.90
2 4
1 470"
" 2 60
""""0.0"
— 0T0 ~
~~2 2'":
"14 60"-
"3 00 0""
"" " "~0~
BR1T- 1
A 4
5 .0
0. 4
2.10
24
1 4 f 0
2 . 50
0 . ij
0 . 0
3 5a
320
600
3 2 00
OKIT-2
4 4
4 . 9
0 . 7
3(. 5 0
24
1 4 7 0
2 . 60
0.0
0 0 ¦
310
25 0
20 0
0
BR 1 T-3
4 4
5 . 2
"0-2
l: 10
24
1470
2 . 60
0. 0
0.0"
26 5 "
"66 0
"" 300
" 4 000
AMER-BK i T
3 4
4.6
¦1.0
1.80
24
1470
2 . 60
0.
0. 0
25')
1 30 0
200
3 1 50
BR 11 -AMEH
4 3
5. 2
0 . 4
1 .80
2 4
1470
2 hi)
0. 0
0. 0
265
500
0
4 800
tutu5C
3 3
"5.2
1 . 8
- 7. 99
24
1552
2 . 64
"0.8
17.6""""
214 "
1516""
37 9 8
"5 000"
AMtR-SD
3 3
5.4
1.8
2.94
24
.1544
2 . 64
2 . 6
1 9. j
2 64
1195
1 6 74
i 006
A ME R - 3 J
3 3
4.5
1. 5
2.86
24
1476
2 .63
2 . 5
16.3
252
8 8 8
1371
i 2 00
A Mt K-8 A
3 3
5 . u
2.3
3:99
14
1 54 5
2 .85
2.8
16. 2
348
89 1"
37 0
2 00 "
AHER-8B
3 3
4 .8
2 . 4
4.28
24
1560
2 . 98
2 . 6
16.9
279
75 1
49 7
720
-------
TABLE 1 (Contd.)
fTpe uf ca/s
FfEDRATFS 8EO BED
EXP. NO. COAL ADO. COAL. ADD., HEIGHT. TL;MP.
» / HR
»/HK
IN.
F
AMEK 8-C
3
3
5 . 0
2. 3
3.98
46
1553
MUMP- 1 A
S
3
4 . 1
1 . 3
4.18
24
1 44 1
HUMP- IB
5
3
4 . 2
1 . 4
4.30
24
1 548
HUMP - 1C
S
3
4 . 1
1 . 4
4.4 5
24
1650
HUMP- ID
5
3
3 .8
1 .3
4.58
24
1 455
HUMP - 10
5
3
4 . 8
1 . 5
4.20
24
1 459
HUMP-1e
5
3
4 . 0
1 . 3
4.44
24
1 395
HUMP- 1£
5
3
4 . 0
1 . 3
4.3 2
24
1319
HUMP-?AI
5
3
4 . 0
0.8
2.6?
24
1456
HUMP-2U3
5
3
4 . 1
0. 3
1. 00
24
1443
HUMP 3
5
3
4.0
0 . 3
1. 10
24
1452
HUMP-3-2
5
3
4.0'
0 . 4
1.28
24
1446
HUMP-4-1
5
3
4 . 0
0 . 6
2. 00
24
1 46 4
HUMP-4- I
5
3
4". U
0 . 3
0.9 4
24
1 44 1
HUMP-4-3
5
3
4 . 1
0 . 3
1. 00
24
1443
HUMP-4-4
5
3
4 . 0
0. 4
1. 46
24
1448
HP-5-A
5
0
4 . 0
0.0
0.0 0
2 4
1 325
HP-5-B
5
0
3 . 9
0. 0
0. 00
24
1 450
HP-b-C
5
0
3.6
0. 0
0. 00
24
15 38
HP-5-0
5
0
4 . 2
0. 0
0.0 0
2 4
1 60 5
HP-5-E
5
0
4 . 3
0.0
0. 0 0
24
1325
HP-6-A
5
0
4 . 4
0.0
0. 00
24
1 328
HP-6-B
5
0
4 . 0
0 . 0
0. 00
24
1 439
HP-fa-C
5
0
4 . 4
0.0
0. 00
24
1 544
HP-6-D
5
0
5. 0
0 . 0
0. 00
24
164 2
PBY-2-A
3
0
4 . 4
0. 0
0. 00
24
1 3 26
PBY-2-B
3
0
4 . 1
0. 0
0. 00
24
1 450
PBY-2-C
3
0
4 . 2
0. 0
0. 00
24
155 1
PBY-2-0
3
0
4 . 3
0.0
0. 00
"24
1 64 4*
PBY-2-E
3
0
4 . 2
0.0
0. 00
24
1 450
PEABY-4
3
3
4 . 0
2.1
4.54
24
1 479
PEABY-5
3
3
3 . 9
1.2
2.58
24
1547
PEABODY-5K
3
3
3 . 9
1.1
2. 43
24
1550
PEAB Y-6
3
3
4 . 0
2.89
24
1551
PEAB0DY-6R
3
3
3.9
2.43
24
1550
AMEK"33 3
3
3
4. 3
3.50
24
1 47 1
AHER 333-3
3
3
4 . f
2.75
12
1 468
AHER 33 3-4
3
3
5 . U
1 - 6
3.25
12
1477
sreaov STATfc FLU0 CAS CONC. lUHYi
GAS
(12.
C02 .
MO.
SO 2 •
CH4,
' Cu.
/S^L
V / 0
V/0
PPM
PPH
PPH
PPM
3 . 15
3.0
16.4
352
570
422
200
"2 . 60
3. 0
0. 0
46 4
2 5
180
300
2 . 7/
0.0
0.0
529
380
283
200
2.83
0.0
0.0
610
98 0
192
250
2 7h2
"" "37.i"
"0 0 '
5 29""
64
" 2 0 0""
500
2.63
3 . 1
16.3
493
8 5
344
500
2 -42
3. 3
16.4
447
276
449
1 200
2.50
3. 1
15.5
338
48 0
701
1 200
2.61
2.2
16. 2
336
56 4
549
1 000
2 57
3.7
15.6
461
1310
250
800
""2". 5 3"
2 .8 '
15.6"
4 8 6 "
15 00"
2 59
l"0OO
2 . 52
3. ')
15. 3
534
15 26
449
15 00
2 . 55
10.4
16. 5
531
1571
773
2 900
2.51
3.1
15 .9 "
"506"
1 4R 0
" 53 8 "
2100
2 . 50
3.3
15.2
433
141 3
550
I 000
2.52
3.4
14.8
396
1 30 6
69 7
2 200
' 2 . 40
0.0 "
1 6. 7
4 62"
1910
"7 2 0"
"4 400
2 . 79
0.0
0.0
6 09
1911
527
i 00 0
2 . 90
0. 0
16.5
626'
205 7
880
4 500
2 . 99
0.0
16. 7
~" 6 01
22 3 1
84 2"
3 400
2 . 70
0. "
0.0
585
198 7
800
5 3 00
2 . 59
2 . 0
16.3
6 00
2282
754
0
2 . 73
"2.0 '"
16. 7
" 684 ""
2119
577
0
2 . 87
1 . 4
16.7
684
2289
1 003
0
3 . 04
1. 2
16.7
642
245 2
1319
0
2 . 40
2.1
"15.5"
5 34 "
'390 3
4 60
" 4 400
2 . 60
3. 2
15.4
646
3677
165
1 5 00
2 . 70
2.9
0.0
654
3759
174
1 5 00
2 . 80
"2.7
15 . 7
6 49
4095
400
""2 5 00
2 . 60
0. 0
0. 0
672
3733
60
1 000
2 - 61
3. 3
15.1
5
463 7
98
60 0
2 . 70"
3.2 "
15. 4"
"" 318
45 2"
"349
400
2 • 69
3.4
15.0
294
64 9
372
500
2 68
3. 1
15.6
388
1 169
624
1 500
3 . 0"1 "
0.0
"15 .5
2 36
84 5
20 0
63
2.79
.2-3
16 7
121
660
786
135
2 .59
2.2
17.2
49 4
145 9
22
3 5fc9
2 56
2.7
17.8
"2 0 3
117 2
266
3 610
-------
ro
00
TABLE 1 (Contd.)
NOTE THAT A *01)00" OR ft "-10" OK ft "O.O" (IN 02 'AND C02 CONC.) INDIC A IE TH AT DATA EITHER
HAVE NOT BE EM TABULATED OR ARE NOT AVAILABLE " " ' "
Additional Notes:
1, Only data from combustion runs are tabulated. If the data logger was not in operation during a run, no
data from that run are listed.
2. Some numbers in this table differ from those reported in the text. This sometimes is caused by a
.rounding-of f procedure but is more often caused by a malfunction of the data logger. These malfunctions
have gradually been corrected as the system has been used. ~~ " — -
An example of a malfunction has been the printing of "l's" instead of "rs". Numbers are obtained every
five minutes during a run and are averaged to obtain, for example, the average bed temperature during a
run. If the numbers averaged include incorrect numbers, an error results. However, the plotted dat'a-
for each run have been examined, and the correct average number has been used in the text.
3,. In the .M.coal"_cplumn,_ "2" represents Illinois coal, shipment 2; "3" represents Illinois coal, shipment 3;
"4" represents British coal; and "5,f represents Pittsburgh coal. In the "additive" column, "1" represents"
limestone No. 1359, shipment 1; ,,2H represents limestone No. 1359, shipment 2; "4" represents British
limestone; ,,6" represents, limestone No. 1360; and "7" represents Dolomite No. 1337.
-------
zy
b. Ca/S Mole Ratio in the Feed. Additional, experiments have
been done to investigate the effect on sulfur retention of Ca/S mole ratio
in the solids feeds. The additive types,¦coals, and operating conditions
used differed from those studied in earlier work. The effect of Ca/S ratio
is shown (Fig. 4) for the Pittsburgh and the Illinois coals at the fluidized-
bed temperature.where sulfur retention had been maximized—1450°F for the
Pittsburgh coal and 1550°F (some 1600°F data are included also) for Illinois
coal. Additional data on the effect of Ca/S ratio on sulfur retention were
obtained during combustion experiments in which data from the ANL 6-in.-dia
combustor were compared with data from the British 6-in.-dia combustor
(these experiments are discussed below).
iUO
1 1
1 IO- 1
/O A
1
90
80
—
—
70
—
Illinois I
/,
cy
/
¦? Pittsburgh
Coal
—
Coal / /
/ /
//
//
//
bO"
Symbol Run No.
Coal Type
W Ad^ve
Q Hunp-LA,
-1D,-2A,-2B
Pittsburgh
1450 1359
50
/
A AR-4,-5
Illinois
1550 1359
/
1
A AR-1E
Illinois
1600 1359
40
O BC-8
O BC"9
Illinois
Illinois
1600 1359
1600 1360
30
—
© BC-10
Illinois
1600 1337
20
— /
Average particle
490-630 ym
Gas velocity in
size range
combustor:
for additive:
2.6-2.8 ft/sec
10
—
0
I 1
l I
1
1
0 1 2 3 4 5 6
Ca/S Mole Ratio
Fig. 4. Sulfur Retention as a Function of
Ca/S Mole Ratio in the Feed
-------
30
Four runs (IIump-lA, -ID, -2A, and -2B) were performed with
Pittsburgh coai and 609-pm limestone No. 1359, a 1450°F fluidized-bed
temperature, a gas velocity of 2.2-3.7 ft/sec, and Ca/S ratios oE l.Q to
4.6 in the feed. The sulfur retention increased from r^46% at a Ca/S ratio
of 1 to 75% at a Ca/S ratio of 2.7, and to 96% at"a Ca/S ratio of 4.2-4.6.
Experiments AR-5B to -5D were completed at 1550°F with 490-pm
limestone No. 1359 and Illinois coal containing 3.7% S (dry basis) to study
the effects on sulfur retention of Ca/S mole ratios in the feed streams of
2.5, 4.6, and 5.5. Sulfur retentions were 78, 95, and 94%, respectively,
as shown in Fig. 4.
The effect of Ca/S mole ratio was also examined in experiments
carried out in the 6-in.-dia ML combustor and a 6-in.-dia combustor at the
Coal Research Establishment (CRE) laboratories (at England) to determine if
SOn removals would differ significantly when the two combustors were operated
under similar conditions. Three limestone stoichiometric addition levels
were used in experiments with (1) American coal and limestone fed separately,
(2) premixed American coal and limestone, (3) British coal and American
limestone, (4) British coal and limestone, and (5) American coal and British
limestone (Table 2). The British coal and limestone were Welbeck coal and
a limestone furnished by J. Gregory & Son. The American coal and limestone
were Peabody Mine 10 coal and limestone No. 1359. The sulfur retention data
in Fig. 5 (presented as ppm SO2 in the flue gas) indicate that sulfur retention
increased at higher Ca/S mole ratios for British and American coal-limestone
combinations in both combustors.
c. Type of Coal. The effect of type of coal on sulfur retention
could not be evaluated because of insufficient data. However, at bed
temperatures of 145Q-1470°F and retentions of less than 90%, the sulfur
retentions are not much different for Illinois and Pittsburgh coals (Fig. c).
Data are not available for comparing sulfur retentions at a bed temperature
of 1550°F. No data on sulfur retentions above 90% are available for Illinois
coal at 1450°F.
Sulfur retentions obtained in experiments Hump-IA, -ID, -2A,
and -2B, in which Pittsburgh coal was burned in the ANL combustor with
limestone No. 1359 additive, are plotted in Fig. 6 along with sulfur
retentions for earlier ANL work with Illinois coal and limestone No. 1359
additive (Amer-6, -3, -33, -4, and -1). Also plotted are sulfur retention
data for British experiments^ at the Coal Research Establishment laboratories,
in which Illinois coal was burned in the CRE 6-in.-dia fluidized-bed
combustor with limestone No. 1359 additive. The fluidized-bed temperature
was nearly the same (1450 to 1472°F) in all of these runs. Below 90%
retention, the data from the various experiments fell along a smooth curve,
indicating that at ^1450°F with limestone No. 1359 additive, sulfur retention
was independent of the source of coal. Zielke et al.^ had reported sulfur
retention to be independent of the source of coal when Disco char (3.7 wt %
S) , Cresap char (1.2 wt % S) , and Ireland coal (1.3 \'t TI S) var; burned in
a fluidized bed of Tymochtee dolomite at I800°F.
-------
TABLE 2. Operating Conditions and Results for British-American Experiments Performed at ANL
Equipment:
British Coal:
American Coal:
British Additive:
American Additive:
British Starting Fluidized Bed:
American Starting Fluidized Bed:
Tempo rature:
ANL 6-in. dia Fluldized-Bed Combustor
Welbeck, 1.2 wt % S
Illinois, 3.7 wt % S
British limestone, Stoke-on-Kent, 94.6 wt % CaCO,
No. 1359 limestone, 94.8 wt % CaCO^
Welbeck Coal Ash (from British combustor)
Calcined and partially sulfated No. 1359 limestone
14 70°F
Utiliz.
Superficial
of CaO in
Run
Ca/S Mole
Gas
Flue
Gas Composition
Sulfur
Additive
Time
Coal
Addi tive
Ratio3
Velocity
°2
so2
NO
CO+CH
C0?
Retention
Feed
Expt.
(hr)
(lb/hr)
(lb/hr)
A B
(ft/sec)
(vol %)
(ppm)
(ppm)
(ppm)
(vol^%)
(%)
(%)
British
Coal
- British
Additive
Brit-1
14
5.0
0.42
2.1 2.3
2.5
2.9
320
350
4200
16.0
78
37
Brit-2
6.5
4.9
0.69
3.5 3.8
2.6
3.0
250
310
b
16.3
82
23
Brit-3
8
5.2
0.23
1.1 1.3
2.6
2.4
660
265
>5500
16.2
55
50
American
Coal
- American Additive
Araer-1
39
4.5
0.53
1.0 1.1
2.6
2.7
2480
240
4050
16.0
38
38
Amer-3
21
4.6
1.6
2.8 3.0
2.6
2.5
870
260
3700
16.5
78
23
Amer-4
13
4.6
1.05
1.9 2.0
2.6
3.1
1460
215
3000
16.2
63
33
Ame r- 3 3
6.5
4.6
1.5
2.7 2.8
2.6
2.5
840
240
3600
16.8
79
29
American
Coal
- British
Additive
Amer-Brit
8
4.6
1.0
1.8 2.0
2.6
2.6
1300
250
2400
16.2
68
38
British Coal -
• American.
Additive
Bri t-Amer
4
5.2
0.38
1.8 2.0
2.6
2.4
^500
265
4800
16.4
-v66
37
3Column A includes only the calcium in the limestone feed; column B includes the calcium in the
limestone feed and the coal.
b
CO concentration not known for this experiment, CH^ = 200 ppm.
u>.
-------
4000
3500
3000
2500
2000
1500
1000
500
0
Fig
CRE Data
(No Data Points)
ANL Data
Illinois coal- —
American limestone No. 1359
Amer-6, Premixed Illinois
coal-
limestone Mo. 1359
A Welbeck coal-Illinois
American Limestone
Welbeck coal-
British limestone
Illinois coal-
British limestone
Welbeck Coal
-O
1.0 2.0 3.0
Ca/S Mole Ratio (including calcium content of coal)
4.0
5. SC>2 Concentrations in Flue Gas for ANL and CRE Combustion Tests
with American and British Coal-Limestone Combinations
Gas Velocity: ^2.6 ft/sec
Fluidized Bed Temp.: 1470°F
Bed Height: Ik in.
-------
33
100
C
o
Illinois coal. Limestone No.
1359, T » 1472°F (ANL Data)
e
0)
4J
V
a:
A Illinois coal. Limestone No.
1359, T = 1470°F (British Data)
u
3
Pittsburgh coal, Limestone No.
O 1359, T - 1450°F (ANL Data)
0
1
2
3
5
6
Ca/S Ratio
Fig. 6. Effect of Ca/S Ratio on Sulfur Retention
for Three Series of Runs (Combustion of
two coals in ANL combustor, one coal in
British conibustor)
-------
d. Coal Fa;' ir i _> r j . To te:muv.i i:«;j ~r of particle
size of the coal feed on sui: r-niion, two iks (PB-5K and PB-6R)
were completed with -12 +50 mesh and -50 mesh Illinois coal. To ensure
uniformity of all of the properties except size, the coal for both experi-
ments was from the same batch of coal. The batch was crushed and screened,
and the -12 +50 mesh fraction was divided into two portions. One portion
was fed during PB-5R and the remaining portion was crushed to obtain the
-50 mesh coal fed in PB-6R. Other experimental conditions were selected
(a Ca/S mole feed ratio of 2.4 and a fluidized-bed temperature of 1550°F)
to give moderate retention of sulfur so that the effect of particle size,
if any, would be clearly indicated.
It was expected that most of the -12 +50 mesh particles would
be burned within the fluidized bed and that some of the -50 mesh coal would
burn in the zone above the fluidized bed (where SO^ absorption by limestone
would be less efficient). In earlier work, it has been observed that
stratification of particles can occur in the bed to some extent for particle
size distributions similar to that used in experiment PB-5R. The larger-
sized particles of coal are probably burned close to the bottom of the bed
and the SO2 released is sorbed by CaO at a bed level above the burning site.
Sulfur retentions by limestone additive were 81% and 75%, respectively, in
PB-5R and -6R.
These sulfur retentions were compared with that obtained in
experiment AR-5D, in which Illinois -14 mesh coal was burned under operating
conditions nearly identical to those used in the above experiments. The
particle sizes for the -14 mesh coal included most of the particle sizes
present in the other two fractions. Sulfur retention in experiment AR-5D
was 78%. Thus, similar sulfur retentions were obtained by burning coal of
three particle size distributions. Nevertheless, the results indicate a
small benefit from burning coarse coals or coals containing small amounts
of fine particles.
e. Additive Type. The SO^ removals for runs performed at a
Ca/S mole ratio of 2.5 and fluidized-bed temperatures of 1550 or 1600oF with
several types of additives may be compared in Fig. 4. ~ With limestone No.
1359 and dolomite No. 1337, sulfur retention was 75-85Z indicating that
differences in these additive types had only minimal effects at these
temperatures.
Experiments performed at the Coal Research Establishment and
ANL allowed comparison of the relative reactivities of British and American
limestones for SO2 removal. Illinois coal and British Welbeck coal were
each fed with 609-ym limestone No. 1359 and 555-pm British limestone.
Figure 5 shows that in the tests with Illinois coal, the British limestone
apparently was more effective for removing SO2 from the flue gas than was
limestone-No. 1359. In tests with Welbeck coal, the SO^ levels in the flue
gas were about the same for limestone No. 1359 and British limestone—
probably within the experimental error.
f. Additive Particle Size. Sulfur retention in AR-6 (which used
1000-ym limestone No. 1359 and a Ca/S ratio of 4) was compared with sulfur
retention in experiments AR-5, BC-8, BC-9, and BC-10 (which used limestones
No. 1359 and 1360 and dolomite No._ 1337 of 490-630 pm average diameter, a
-------
35
superficial gas velocity of 3 ft/sec, and a Ca/S of 2.2-5.5). Sulfur
retention in AR-6 was measured at gas velocities of 3.5-7.4 ft/sec, and
the sulfur retention at 4 ft/sec was based" on extrapolation of experimental
values. Sulfur retention was calculated to be ^87% for the larger particle
additive and 93% for the smaller particle at a Ca/S ratio of 4.0. This
suggests that additive particle size, at least in the region of high sulfur
retention, has only a moderate effect on sulfur retention at a particular
fluidizing-gas velocity.
g. Fluidized-Bed Height. The effect of the height of a fluidized
bed on the percentage of sulfur retained in the bed was studied in one-stage
combustion runs. In Amer-8A, -8B, and -8C performed under similar conditions
except 'that bed heights were 14, 24, and 46 in. [length to diameter (L/D)
ratios of 2.3, 4.0, and 7.7], the sulfur retentions were 78, 80, and 83
(Table 1), respectively, indicating a minor effect of bed height. At the
highest L/D ratios, solids slugging was severe. This did not affect sulfur
retention, but operation of the combustor was difficult—heat transfer was
poor, and high feed rates of coal were noted several times that caused
variations in the SO2 concentration in the flue gas.
h. Superficial Gas Velocity. Increased sulfur retention with
decreased superficial gas velocity (in the range of 3.5 to 7.4 ft/sec) was
observed in experiment AR-6, in which the operating conditions were a coal
combustion temperature of 1550°F, a starting bed of partially sulfated
limestone No. 1359, and addition of limestone No. 1359 (>1000 pm particle
.size) ana Illinois .coal ,at a Ca/'.S .mole .feed ratio of ^4. .The relatively
coarse additive was selected to ensure that additive particles would be
retained in the fluidized bed at high gas velocities. At gas velocities
of 3.5, 5.5, and 7.4 ft/sec, the average SO2 concentrations in the flue gas
were 770, 1250, and 1500 ppm, corresponding to retentions of 83, 73, and 66%
of the sulfur fed to the reactor (Table 1). These data are also shown in
the semilog plot of Fig. 7 and may be correlated with the equation:
R = 101.79 e-0*0625 v (!)
where R = SO2 removal, %
v = superficial gas velocity, ft/sec
Previous ANL data (SACC-8 and -9, ANL/ES/CEN-1002, p. 26) for
25-um No. 1359 limestone additive, coal combustion at 1600°F, and alumina
fluidized beds do not show the same correlation between sulfur retention
and gas velocity for gas velocities of 2.7 and 8.6 ft/sec. Similarly, in
tests carried out by Pope, Evans, and Robbins using <44 pm No. 1359 limestone,
a Ca/S ratio of '^2, a sintered coal-ash fluidized bed, and a combustion
3
temperature of 1525 to 1600°F, gas velocity in the range 6.0 to 12.8 ft/sec
had essentially no effect on sulfur retention.
Other American and British data for series of coal combustion
runs per formed at several gas velocities and various Ca/S mole ratios have
been examined to further elucidate the relationship between sulfur retention
and superficial gas velocity. Results of recent British experiments using
Welbeck coal, 440-um Brinish limestone, Ca/S ratios of 1 and 2, and a coal-
ash bed show that sulfur retention is greater at a gas velocity of 2 ft/sec
-------
36
than at 3 ft/sec (Fig. 7). The equations fitted to the two experimental
points at each stoichiometric Ca/S ratio are:
R = 100.4 e~°*071 V Ca/S = 2 (2)
and R = 102.1 e"°'266 V Ca/S = 1 (3)
To allow further comparison, Fig. 7 includes a datum point from a curve
(Fig. 4) for limestone fluidized beds at 1550°F, a Ca/S ratio of 4, limestone
and dolomite additives, and a gas velocity of ^3 ft/sec.
The slopes represented by equations 1, 2, and 3 decrease as
the Ca/S ratios increase; thus, at sufficiently high Ca/S ratios, sulfur
retention may be essentially independent of superficial gas velocity.
100
90
80
70
60
^•5
c 50
I «
4>
CC
u
3
^ 30
20
10
CRE Data (1470°F)
0" "81 Welbeck Coal, British Limestone (AhO um)
©—-e)
AN'L Data (1550°F)
© © Run AR-6, Illinois Coal, Limestone No. 1359 (^1000 um)
A Illinois Coal, point taken from curve (Fig. A) for
limestones and dolomite (^630 nn), Ca/S ° 4
Ca/S ¦ U
Ca/S =» 1
_L
I
_L
I
_L
2 4 6
Superficial Cas Velocity, ft/sec
Fig. 7. Effect of Superficial Gas Velocity on Sulfur Retention
-------
37
i. Excess Air. Runs to determine the effects of excess air
(or oxygen) in the combustion gas on sulfur retention utilized Illinois coal,
limestone No. 1359 additive, and a starting bed of partially calcined and
sulfated limestone. Experiments Amer-5C, -5D, and -5E (at 1550°F) were
performed at three excess air levels. The oxygen level in the off-gas was
varied by adding pure oxygen at several rates to the fluidizing air before
it entered the preheater. The purpose of adding oxygen in preference to air
was to maintain the fluidizing-gas velocity essentially constant, since this
variable affects sulfur retention.
Data from experiment Amer-5 are summarized in Table 3 and
Fig. 8. In experiments Amer-5C, -5D, and -5E, the SO2 level in the flue gas
decreased from 1540 ppm with 0.7 vol % C>2 in the flue gas to 1350 ppm with
2.4 vol % C^, and to 1170 ppm with 5.6 vol % O2. Sulfur retentions were
67, 71, and 75%, respectively. Apparently, oxygen concentration slightly
affects the reaction of SO2 with limestone, and sulfur retention can be
expected to increase when oxygen concentration in the flue gas is increased.
. 60 —
a; 50
Fig. 8. Effect of Excess
Oxygen in Combustion Air
on Sulfur Retention,
Experiments Amer-5C, -5D,
and -5E
Ca/S Mole Ratio: ^2.8
Fluidized-Bed Temp.: 1550'
Superficial Gas Velocity:
2.6 ft/sec
j 4
Oxygen Concentration, vol X
-------
U)
00
XAB^E 3. Effect of Excess Oxygun in Combustion Air on Sulfur Retention
AIJL 6-in.-dia fluidized-bed combustor
Equipment:
Coal:
Additive:
Starting Fluidized Bed:
Fluidized-Bed Temperature: 1550°F (Amer-5)
ILlinois Seam No. 6 (Peabody-Commonwealth Edison),
3.7 wt % S (-14 mesh)
LlDiestone No. 1359, as received
(94.8 wt % CaC03, 0.9 wt % MgCC>3)
18.0 lb partially sulfated and calcined limestone No. 1359
(609 vim) (^24-in. fluidized-bed depth)
Expt.
Amer-
Time at
Equilibrium
(hr)
Feed
Rates
Ca/Sa
Mole
Ratio
Superficial
Gas
Velocity
(ft/sec)
Sulfur
Retention
(%)
CaO
Utilization
(%)
Flue
Gas Composition^
Coal
(lb/hr)
Additive
(lb/hr)
°2
(vol %)
so2
(P pin)
CO
(vol %)
5C
7.5
5.2
1.8
2.8
2.6
67
24
0.7
1540
17.8
5D
4.5
5.2
1.8
2.8
2.6
71
25
2.4
1350
>19.3
5E
5.5
5.0
1.8
2.9
2.6
75
26
5.6
1170
19.6
Values do not include calcium content of coal, which would have increased Ca/S by ^.1.
^The concentrations of the components in the flue gas; Ln Runs 5D and 5E should not be compared with similar
data from other runs made to date. In these two runs the excess oxygen requirements were supplied by oxygen
rather than air, resulting in a smaller volume of gas than if air had been used. Consequently, the measured
flue gas concentrations are proportionately higher.
-------
39
j. Moisture Content of Coal. In a run to determine the effect
of moisture content of the coal on sulfur retention, Pittsburgh coal was
fed at 4 lb/hr, limestone No. 1359 was fed at 0.3 lb/hr, and water was added
at different rates to the fluidizing air (at the base of the ANL 6-in.-dia
combustor). The fluidized-bed temperature was maintained at 1450°F, the
fluidizing gas velocity at 2.6 ft/sec, and the Ca/S ratio at 1. The starting
bed for this run (Hump-4) consisted of partially calcined and sulfated
limestone No. 1359. The rate of water addition and the concentrations of
flue gas components are listed in Table 4. The water injection rates were
equivalent to burning coal with moisture contents up to 51 wt %.
The concentration of SO2 in the flue gas remained the same
(^1400 ,ppm) .throughout the run. The concentrations of CH^ and CO in the flue
gas decreased upon the addition of moisture, but did not decrease further
with increased rate of addition. The NO concentration decreased as well,
as discussed in Section II.D.2.
TABLE 4. Effect on Flue Gas Composition of Addition
of Moisture to the Fluidized Bed, Run Hump-4
Coal: Humphrey (Pittsburgh) coal,
Morgantown, W. Va., 2.4 wt % S
Additive: Limestone No. 1359
Fluidized-Bed Temperature: 1450°F
....Ca/S-.Mole..Ratio.: , 1..0
Gas Velocity: 2.6'ft/sec
Water
Addition
Concentration
in Flue
Gas3
Run
Duration
(hr)
Volume
Rate
(cc/min)
Equivalent
wt % water
in coal
°2
(vol %)
co2
(vol %)
CO
(ppm)
ch4
(ppm)
NO
(ppm)
S02
(ppm)
1.25
0
1.8
2.2
16.5
2900
850
530
1400b
2.00
10
26.2
2.7
16.0
2100
580
510
1400b
2.00
20
40.8
3.3
15.0
2100
580
440
1400b
2.00
30
50.6
3.4
15.0
2100
580
380
1400b
aMoisture-frce basis.
^Reduction in SC^ emission
42%.
-------
k. Solids Feeding Mot hod. An experiment (Arier-6) was made in
which coal and limestone were preniixed before being fed to the combustor
rather than being fed separately. This change had little effect on the
sulfur retention, which was similar for runs without prefixing (see Fig. 5).
Operating conditions were a temperature of 1470°F at a Ca/S ratio of 2.6,
.and a gas velocity of 2.6 ft/sec.
1. Temperature of Gas in the Freeboard. During the first half
of Hump-3 with -14 mesh Pittsburgh coal and 609-vim limestone No. 1359 , the
upper section of the ANL fluidized-bed combustor was insulated; during the
remainder of this run (Hump-3-2, Table 1), the upper .section was not insulated
(absence of insulation is normal for the ANL combustor) and was cooled by
natural circulation of the ambient air. The temperature of the fluidized
bed, where most of the coal was burned, was 1450°F for the entire run. The
temperature of the upper section of the combustor was 1340°F during the
first half and 1020°F during the remainder of the run. The Ca/S mole ratio
was'l, and the gas velocity was ^2.6 ft/sec.
The concentration of SC>2 in the flue gas was ^1500 ppm during
both periods, indicating that sulfur-removal reactions occurred essentially
in the fluidized-bed zone. Upon removal of insulation of the upper section,
the concentrations of CH, in the flue gas increased from 260 ppm to 450 ppm,
that for NO from 490 to 540 ppm, and that for CO from 1000 to 1500 ppm. This
indicates that gaseous combustion of CO, CH^, and possibly additional reaction
of CO with NO increased when the upper section of the combustor was insulated
and the gas temperature reached 1340°F.
m. Absence of Additive. Currently, the results of combustion
experiments are given as percent sulfur retained—the percentage of sulfur
in the feed that does not appear as SO2 in the flue gas. However, previous
studies on combustion efficiency (ANL/ES/CEN-1002, Appendix D) indicate
that some unburnt coal particles are elutriated with the ash. The unburnt
carbon in the ash may contain some sulfur. Therefore, not all sulfur retained
has been released as SO2 and reacted with additive. To determine how much
SO2 and NO is actually released during combustion, experiments Hump-5A to
-5E (Table 1) and PB-2A to -2E were performed with no additive present.
Five experiments (Hump-5A to -5E) were performed with -14 mesh
Pittsburgh coal (2.4 wt % S), with no limestone injected, coal feed rates of
about 4 lb/hr, and the fluidized-bed temperature ranging from 1325 to 1650°F.
The duration of operation at each temperature was 1 to 1.5 hr. The fluidized
bed contained refractory alumina as the major constituent and unelutriated
ash as the minor constituent. The fluidized-bed depth was about 24 in.,
and the fluidizing-air velocity was 2.4 f.o 2.8 ft/sec. The observed concen-
trations of SO2 were 150 to 300 ppm less than the values calculated by
assuming that all sulfur in the coal had been converted to SO2, as shewn
in the following tabulation.
-------
41
S02(ppm)
Calc. S0„(ppm)
Bed Temp (°F)
Hump-5A
1900
2290
1325
-5B
1900
2210
1450
-5C
2040
2300
1550
-5D
2200
2360
1650
-5E
1960
2250
1325
When Illinois coal. (3.7 -w-c- & was burned under similar
condiciori i- five experiments {PB—2-A :£<>. ^2E) , che' concentrations o£ SO9
were higher, as expected due to the.higher concentration of sulfur in this
coal. The observed and calculated S0« concentrations were about the same
in three cases and-in the fourth and fifth runs the calculated concentrations
were 17% less and 7% higher than the measured concentrations, as shown:
S0„(ppm)
Calc. S0„(ppm)
Bed Temp (°F)
PB-2A
3950
3900
1325
-2B
3750
3750
1450
-2C
3750
3700
1550
-2D
4250
3550
1650
_ O T? ,
4. J-J
3650
lAAft
1450
2. Effects of Variables on Reduction in Emissions of NO During
One-Stage Combustion
Nitrogen oxides, principally nitric oxide (NO), are formed during
the combustion of fossil fuels. At 1600°F (a common temperature for fluid-
bed combustion), the equilibrium concentration of NO as a result of the
nitrogen fixation reaction, 1/2 N2 + 1/2 O2 J NO, varies from 50 to 200 ppm,
depending on the oxygen concentration. However, during combustion of coal
in the bench-scale combustor, nitric oxide levels of 400-800 ppm have been
measured. A possible explanation for these high NO concentrations is that
the nitrogenous content of the coal (1-1.5 wt % in U.S. coals) is the source
of NO. This amount is sufficient to form several thousand ppm. The residence
time in the combustor is not sufficient for decomposition to equilibrium
levels.
In recent investigations, NO release during coal combustion in the
absence of additive was studied, as well as the effect of moisture content
of the coal on NO removal.
In five experiments with Pittsburgh coal (Hump-5A to -5E) and five
with Illinois coal (PB-2A to -2E), the quantity of NO released when no
additive was present in the bed was determined. Coal was burned in an
alumina bed at different temperatures (1325-1650°F). The NO concentration
in the flue gas ranged from 450 to 630 ppm:
-------
42
Fl.-Iiod Tcr.p (°F) NO in Flue Gas (ppn)
Hump-5A 1325 450
-5B 1450 610
-5C 1550 630
-5D 1650 600
-5E 1325 570
The lowest NO concentration was obtained at the lowest bed temperature.
With Illinois coal, the NO concentrations ranged from 530 to 680 ppm,
and again the lowest NO concentration was obtained at the lowest temperature.
Fl.-Bed Temp (°F) NO in Flue Gas (ppm)
PB-2A 1325 530
-2B 1450 650
-2C 1550 660
-2D 1650 650
-2E 1450 680
However, with either coal, no good temperature correlation could be made.
The NO levels are higher than those expected if additive had been present,
indicating that sulfated additive has a beneficial effect, reducing NO
levels in the flue gas up to approximately 50%.
The effect of moisture content of the coal on NO level was studied
in Hump-4, in which water was added at different rates to the fluidizing
air, at the base of the combustor (Table 4). Pittsburgh coal and limestone
No. 1359 were used. The fluidized-bed temperature was maintained at 145D°F
and the Ca/S ratio at 1. A Calcomp plot of NO concentration in the flue gas
during all of Hump-4 is shown in Fig. 9. The concentration of NO decreased
from 530 ppm to 510 ppm upon adding 10 cc/min water (equivalent to 26 wt %
water in the coal), and to 380 ppm when the rate of water addition was further
increased to 30 cc/min (equivalent to 51 wt % water in the coal). Reductions
in the NO concentration may be due to a catalytic effect of H2O.
3. Effects of Variables on Chemical Composition and Physical Properties
of Bed and Elutriated Materials
Samples from the fluidized bed and from elutriated material collected
in the primary and secondary cyclone separators and the final filter were
examined and analyzed. An electron microprobe was used to examine sulfur
distribution in additive particles. The extent of limestone calcination and
calcium utilizations were calculated from analyses of feed and product streams.
The extent of hydration of partially sulfated limestone upon exposure to the
atmosphere was measured. Elutriated material was analyzed for the presence
of calcium sulfite, and particle size distributions of bed samples taken at
intervals during a run were measured. Decrepitation cf additive particles
in the bed was calculated from calcium levels and size distributions of the
-------
43
fresh additive and of elutriated solids. Bulk and tapped densities of
solids from the bed, cyclones, and final filter were also determined.
8
13
CD
Z
a
•.Bl
*x X
X >
X
XXX
X V * X*
X x
X XX XX ,
x XX ;
X
«x*
v- yV f
XX X
X*
* x„x„xKx *
-) I 1 h
VS*
+-
U
-+-
9.00 9. BO 1C.B0 ll.ua
12.20 13.00
TIME IN HOURS
13.B0 1U.E0 15.W 16.20 17.00
Fig. 9. Concentration of NO in Flue Gas during Hump-4
(Calcomp plot)
a. Effect of Type of Fuel on Sulfur Distribution in Additive
Particles. The distribution of sulfur in bed and elutriated particles from
bench-scale combustion experiments was determined with an electron microprobe,
which offers nondestructive, in situ elemental analysis in micron-scale
areas.
^Two samples of elutriated material from the primary cyclone
were examined. One sample was from a coal combustion experiment, and the
other was from a natural gas combustion experiment with SO2 added from a
separate source. Elutriated particles from the" two samples showed differences
in particle composition. For particles from the coal combustion experiment,
the sulfur concentration usually was lower at a greater distance from the
surface of the particle. For particles in samples from the natural gas
experiment, sulfur concentrations appeared to be more uniform throughout
•k
Examination performed by K.. Natesan.
-------
44
the particles. Few particles exposed to combustion of either fuel showed
the calcium and sulfur levels that would be observed if CaO had been complete
converted to CaSO/, Some particles from the natural gas combustion experi-
ment had the calcium concentration of CaCO^, indicating that little calci-
nation or sulfation had occurred. For these elutriated particles the
residence time in the bed was relatively short as compared to particles
remaining in the bed itself.
In samples removed from the bed, the distribution of sulfur
across the cross section of most limestone and dolomite particles from coal
combustion experiments was rather uniform. The possible relationship of
this sulfur distribution pattern to the reaction mechanism was considered.
The most probable mechanism for the CaO + SC^ reaction, when it occurs under
oxidizing conditions, is one in which the reaction initially proceeds in
the surface layer of the particle, after which the SC>2 diffuses through the
initially sulfated layer. Some of the particles examined did show higher
sulfur concentrations at the surface, which would be expected for the shell-
diffusion reaction mechanism. Although most of the bed particles came from
experiments in which oxidizing conditions prevailed, both uniformly distributed
and surface-concentrated sulfur were present. More detailed results are
presented in Section IV of this report, Mechanism of Lime Sulfation Reaction.
b. Limestone Calcination. In some experiments analyses for
calcium, sulfur, and occasionally CO3 were performed on samples from the
fluidized bed and from elutriated material. From the analyses, the extent
of calcination and occasionally the Ca/S ratios for these materials were
determined.
The extent of calcination of limestone material was determined
for fluid-bed samples from experiments AR-6C, -6D, and -6E, performed to
explore the possible effect on SO2 removal at 1550°F of various superficial
gas velocities. Illinois coal ana 490~um limestone No. 1359 were fed to a
starting bed of partially calcined and sulfated lime to give a Ca/S mole
ratio in the feed of ^4.0. The average values of the extent of limestone
calcination were 95.7, 94.5, and 97.3% for superficial gas velocities of
3.5, 5.5, and 7.4 ft/sec. These results show no direct relationship of
the extent of calcination to the superficial gas velocity" but indicate
that calcination is rapid enough at 1550°F to be essentially independent
of gas velocity with material of this particle size.
Calcination was essentially complete (90-100%) for fluidized-
bed and final-filter solids and ranged from 60 to 84% for solids elutriated
to the cyclones in the Amer and Brit experiments. The high calcination
level of the fluidized-bed solids is accounted for by their long residence
time in the combustor; the high calcination level of the limestone component
of the final-filter material is probably a result of rapid reaction in the
combustor of this very finely divided material. The lower extent of calci-
nation for particles collected in the cyclones is probably related to their
relatively short residence times in the reaction zone due to their inter-
mediate size.
-------
45
c. Calcium Utilization. The extent of conversion of CaO to
CaSO^ for bed and elutriated materials was calculated from the Ca/S mole
ratios for materials from several series of experiments. The Ca/S mole
ratio in a solids stream was calculated from calcium and sulfur concen-
trations determined by wet chemical analysis.
Average values and ranges of calcium utilizations have been
obtained for solids streams from the SA-, BC-, and AR-series experiments
(Table 5). The major difference observed is the higher degree of utilization
of calcium in the fluid-bed solids for the SA-series than for the BC- or
AR-series, due most probably to the use of a smaller particle size of
limestone additive in the SA series.
The Ca/S mole ratios for solids streams of Amer and Brit
experiments also were calculated from the calcium and sulfur contents of
the solids. In most runs, the Ca/S mole ratios in the fluidized-bed and
final-filter solids were 0.9-3.4, and are close to the Ca/S mole ratios
(1-3.6) of the coal and limestone feed "streams (Table 1). The calcium
utilizations for the primary and secondary cyclone solids were low,
corresponding to Ca/S ratios of 3.3 to 10.9. Probably, the lower extent
of sulfation of the solids in the cyclones was due to their shorter
residence times in the combustor reaction zone. The higher calcium
utilization for the final-filter solids was probably a result of their small
size (evidenced by their passage through the cyclone separators) and greater
reactivity.
d. "Hydration o"f"Partlal'ly Sulfated Limestone. "Because samples
from the cyclones and fluidized bed contain partially sulfated limestone
that is highly calcined (99% or greater), the weight of a sample submitted
for analysis could increase upon exposure to the atmosphere since hydration
of lime (CaO) to form the hydroxide, Ca(0H)2, occurs readily. An increase
in the weight of a sample because of hydration of compounds in the sample
could affect the value reported as the percentage of a component—sulfur,
calcium, or carbon—in the sample. Therefore, experiments were performed
to determine the rate and extent of hydration of partially sulfated fluidized-
bed and cycLore samples by exposing samples to the atmosphere and weighing
them at intervals.
From analyses of the exposed fluidized-bed samples for CaO,
the hydration rate was found to be initially about 2%/hr and to drop rapidly
to about 0.5%/hr. Some samples continued to gain weight beyond the point
at which hydration was theoretically complete. Weight gain continued at
a very slew rate, about 0.1%/hr. This additional pickup of moisture may be
a result of water being absorbed on an active limestone surface. Since
fluidized-bed material is exposed only briefly to the atmosphere during
analysis, excessive hydration is not expected.
e. Calcium Sulfite Formation. The possibility that calcium
sulfite formation and calcination are related through a common mechanism
was suggested by the results of an earlier BMI coal combustion study^ in
which sulfite was found in the reacted limestone. The fraction of sulfur
in the form of sulfite was related to the degree of calcination. Although
there is considerable scatter in the data presented, an inverse relationship
-------
TABLE 5. Calcium Utilizations for the SA, BC, and AR Series
Experimental Conditions
Expt.
Series
Type of
Additive
Size of
Additive
Temp
CD
Gas
Velocity
(ft/sec)
Ca/Sc
Starting
Bed
Calcium Utilization (%)
Primary Secondary Fluid Final
Cyclone Cyclone Bed Filter
SA No. 1359
BC BC-6,
-7,-8 No. 1359
BC-9 No. 1360
BC-10 No. 1337
25-103 1550-1650
3.0
Varied
0-4.0
44-650 1480-1800 2.7-3.2 2.2-2.5
Alumina
Calcined-
Sulfated
Lime
Average
Range
Average
Range
20
14-24
35
21-47
40
38-45
25
21-29
70
67-71
44
32-54
70
62-100
84
56-100
AR-1,
-2,-4
AR-5
AR-6
No. 1359
No. 1359
No. 1359
490
490
1640
1400-1600 2.5-2.1
1550
1550
^2.8
2.5-2.8
2.5-5.5
3.5-7.4 3.8-4.2
Calcined-
Sulfated
Lime
Calcined-
Sulfated
Lime
Calcined-
Sulfated
Lime
Average
Range
25
21-29
13-17
29
26-31
27-28
41
40-42
20-30
15-20
67
56-77
2 3
45
Calculated ratio of calcium in additive fed to combustor to sulfur in coal fed to combustor.
-------
47
could possibly be postulated. For example, according to BMI, the molar
percent of sulfur present as sulfite was in the 55-65% range at ^25% calci-
nation, in the 10-30% range at 80% calcination, and in the 5-25% range at
95% calcination. The suggested reactions include the following:
CaC03 CaO + C02 (4)
CaO(s) + S02(g) -> CaS03(s) (5)
CaS03(s) + 1/2 02(g) -> CaS04(s) (6)
4 CaS03(s) -+ CaS(s) + 3 CaS04(s) (7)
The presence of calcium sulfite in the products from limestone-
SCL reaction is being studied at Argonne. The sulfite levels of fluidized-
bea samples were expected to differ from sulfite levels in samples of
elutriated solids. Fluidized-bed particles have relatively long residence
times and are heated to the combustion temperature (in excess of 1400°F,
where CaSO^ is thermodynamically unstable), and such particles would be
expected to contain little sulfite. In contrast, sulfite may be detected
in the elutriated solids collected in the primary and secondary cyclones,
which may not have reached combustion temperatures (the calcination reaction
is endothermic) because of shorter residence times in the combustor.
At ANL, elutriated material from each cyclone separator in
-experiments BC-2->to -,5 ^(.performed..at .i6GG°F_; .see ANL/ES/CEN-1G02 , p. 47)
was analyzed for sulfite. The particle sizes of the Tymochtee dolomite
additive were 650 pm for BC-2 and -3 and <44 pm for BC-4 and -5. Tfie extent
of calcination for samples from the primary cyclone ranged from 76 to 81%,
and that for the secondary cyclone was 87 to 96%.
Sulfite was not detected in solids from BC-2; however, sulfite
was detected in the three other experiments up to a maximum of 0.23 wt %.
A sulfite concentration of 0.2 3% in the cyclone samples corresponds to 4.2%
of the total sulfur fed. In the BMI study, 10 to 30% of the sulfur was in
the form of sulfite at the same extent of calcination. It is thought that
the samples in which no sulfite was detected had been heated to 1400°F or
higher. The fact that sulfite was not found does not preclude it as an
intermediate as suggested by Battelle.
f. Particle Size of Fluidized-Bed Material. The particle size
distributions of input and output solids streams for experiments Amer-1,
Amer-3, Brit-1, Brit-3, Brit-Amer, and Amer-Brit (Table 2) were determined
to identify any changes in particle size distribution in the bed as a result
of passage through the screw feeder, elutriation, additive feed rate, and
type of starting bed. In these runs, the gas velocity was ^2.6 ft/sec.
The particle size distributions for limestone No. 1359 (used
in Amer and Brit-Amer experiments) before and after it had passed through
the screw feeder were measured. Passage through the screw feeder resulted
in the +25 mesh fraction decreasing from 40.4 to 33.1 wt % and the -325
mesh fraction increasing from 8.2 to 11.1 wt %.
-------
AS
In the Arner experiments, Illinois coal was burned. It" w;ls
assumed in interpreting the results of particle size analyses that the ash
generated was elutriated overhead and did not accumulate to an appreciable
extent in the bed.
In experiment Air.er-1, the limestone feed rate was 0.53 lb/hr.
The starting bed material was the final bed (partially sulfated and calcined
limestone) from experiment AR-5. A major change was observed in the particle
size distribution of the bed material during this 39-hr run—the percentage
of +25 mesh particles increased from 13.6 to 21.5. Also, the -170 mesh
particles in the starting bed apparently were removed by elutriation early
in the experiment, and the percentage of -80 +170 mesh particles dropped
from 7.2 to 2.1.
In experiment Amer-3, the limestone feed rate was three times
as great as in Amer-1, about 1.6 lb/hr, and the original bed material was
the final bed from Amer-1. At this higher feed rate, the particle size
distribution in the bed changed significantly. The percentage of +25 mesh
particles in the bed after 6 hr of operation was 40.6 wt %, as compared with
21.5 wt % after 39 hr of Amer-1. The percentage of +25 mesh particles in
the final fluidized-bed sample for this 21-hr experiment also was high, 41.4%.
Bed samples taken during Amer-3 contained no -170 mesh particles, which
apparently had elutriated. When the particle-size distribution for limestone
additive was recalculated with the -J.7.0 mesh fraction omitted, the calculated
distribution was similar to the particle size distribution found in Amer-3
bed material after 21 hr.
To determine any changes iri particle size distribution during
Brit-series experiments, particle size distributions of input solids streams
(Welbeck coal and CRE limestone) and fluidized-bed samples for experiment
Brit-1 were measured, as well as those for coal ash (the starting bed). The
limestone addition rate was only 0.4 lb/hr, and the coal ash accumulation
in the bed (assuming no elutriation) may have been several times the limestone
addition rate. The data shew that the particle size distribution of the
fluidized bed at the end of the experiment differed little from the size
distribution of the original coal ash.
During experiment Brit-3, with a limestone feed rate of
0.23 lb/hr, the particle size distribution of the bed remained relatively
constant. The +25 mesh fraction, which constituted 27% of the bed material
at the end of Brit-1, had decreased to 19% at the end of experiment Brit-3.
Particle-size distributions were determined for the Brit-
Amer experiment, in which British coal and American limestone No. 1359 were
fed. In this short duration experiment (4 hr) with a low limestone feed
rate (0.38 lb/hr), particle size distribution in the bed did not change
significantly. In the 8-hr Amer-Brit experiment, which utilized American
coal and British limestone, the particle size distribution did not differ
significantly from those for the Brit-Amer and Brit experiments. In general
the particle size distribution of the bed particles does not change
significantly during an experiment.
-------
49
g. Decrepitation of Additive Particles. Decrepitation and
attrition of several additives during coal combustion experiments has been
estimated by comparing the calcium content of elutriated material with the
calcium content of fines in the bed. (The starting fluidized bed also
contained calcium.) In most experiments a gas velocity of ^2.6 ft/sec was
used; at this velocity, it is expected that all of the flyash and additive
particles having diameters of <177 urn will be elutriated from the fluidized
bed and that their calcium content would be found in the elutriated material.
Because the particulate nvitter thac is elutriated during the combustion of
coal in a rluidized bed is a mixture of solids of different origins and
co-positier.s , the fraction of additive carried over can only be estimated.
The expected elutriation of additive for Amer-, BC-, AR-, and
Brit-series experiments was calculated and compared with the actual
elutriation determined from calcium material balances. (An allowance was
made for the calcium content of the flyash, which was also elutriated.) The
difference between actual and expected elutriation was the estimated
decrepitation of large additive particles.
It is estimated that decrepitation of No. 1359 limestone was
^8%. No decrepitation of British limestone was evident. Decrepitation
of limestone No. 1360 and dolomite No. 1337 particles was more severe—
40 and 85%, respectively. These results indicate that limestone No. 1359
and British limestone are desirable materials for use-in a full-scale
fluidized-bed combustor if the additive is.to.be regenerated and recycled.
The higher decrepitation rates for Nc. 1360 and 1337 additives may make
these' materials less promising for regeneration and recycle.
h. Bulk and Tapped Densities of Solids from the Bed, Cyclones,
and Final Filter. Bulk and tapped densities of fluidized-bed material and
of solids collected in the cyclones and final filter in Amer and Brit experi-
ments were determined. The bulk densities of the cyclone and final-filter
solids ranged from 0.15 to 0.77 g/cc, and the tapped densities from 0.23 to
1.1 g/cc. These low densities were due most probably to the nature of the
solids, which consist of coal ash and/or limestone fines. The bulk and
tapped densities of the fluidized-bed solids were very similar to those'' of
the precursor solids. In the Amer experiments, the precursors were limestone
No. 1359 and possibly small amounts of Illinois coal ash; in the Brit
experiments, the precursors were Welbeck coal ash and small amounts of
British limestone.
4. Material Balances
Material balances for sulfur, carbon, and calcium i-n several
series of runs are tabulated below.
Amer-1,-3,-4
Brit-1 to -3, Amer-Brit,
B rit-Amer
BC-6 to -10
BC-10 (dolomite additive)
AR-1,-2,-4,-5,-6
Sulfur
113%
98%
81%
82%
Carbon
9 7%
114%
97%
9 7%
Calcium Magnesium
92%
92%
87%
99%
82%
-------
The experimental conditions.for these experiments are given in Table 1 of
this report and in ANL/ES/CEN-1002, Table 8. The procedure for calculating
material balances in these AHL runs is presented in Appendix C, along with
details on the weights of sulfur, carbon, and calcium in each material
entering and leaving the cor.ibustor.
5. Exploratory Two-Stage Combustion Runs
Four exploratory runs (Amer-7A to -7D) using two-stage combustion
(see Section III.C. for procedure) were the first two-stage runs made at ANL
to determine the effect of this mode of combustion on SC^ and N0X concen-
trations in the flue gas. In two-stage combustion, a substoichiometric volume
of combustion air is fed at the base of the fluidized bed, and a second
volume of air is fed into the space above the bed to burn CO and hydrocarbons
in the gas leaving the fluidized bed. The hot gases from the second stage
would be directed to a gas turbine.
In each part of Amer-7, the secondary air was injected 6 in. above
the fluidized bed. In Amer-7A and -7B, the secondary air was injected down-
ward (toward the bed and countercurrently to the fluidizing-gas flow); in
Amer-7C and -7D, air was injected upward (away from the bed and cocurrently
with the fluidizing-gas flow). These runs were made with Illinois coal and
limestone No. 1359 additive at fluidized-bed temperatures of 1450 and 1550°F.
The Ca/S mole ratio was ^3_, and the fluidizing-gas velocity was VL.9 ft/sec.
The starting bed was partially sulfated and calcined limestone No. 1359.
Operating conditions and results are given in Table 6.
^ M 1_ "F i_l T" Rot"anfi qti Hnrn no A fn or_7 A 0 -PI 3 3 d—bed tC Hip C IT w "U 2T
was maintained at 1550°F. The concentrations of SO2 in the flue gas at steady
state combustion was ^3000 ppm. The temperature 16 in. above the secondary
air inlet point (probably in the CO combustion zone) was only ^1020°F, showing
that the gas cools rapidly after leaving the bed (which was at ^1550°F).
When secondary air injection was stopped in Amer-7A, the
concentration of SO2 in the flue gas decreased from 3000 ppm to about 800 ppm.
Upon the resumption of air injection, the S0? concentration returned to the
original value of 3000 ppm. This suggests that secondary air may have beer,
reacting with a compound in the bed to form SO2.
In Amer-7B, operating conditions were maintained essentially
the same as in Amer-7A except that a lower fluidized-bed temperature, 1450°F,
was used. At steady-state combustion, the SO2 concentration in the flue gas
was 1100 ppm. Interruption of secondary gas flow caused the SO2 concentration
to decrease to 400 ppm, and resumption of air injection returned the SO2
concentration to 1100 ppm. The different SO2 levels at 1450 and 1550°F
indicate that SO2 emission was affected by combustion temperature.
Secondary air was injected cocurrently into the fluidizing
gas during Amer-7C; all other operating conditions were essentially the
same as in Amer-7A. Concentrations of SO2 in the flue gas after steady
state was attained were 3000 and 500 ppm, respectively, for Amer-7A and
-7C.
-------
TABLE 6. Effect of Cocurrent and Countercurrent Flow of Secondary Air
in Two-Stage Combustion
Equipment: ANL 6-in.-dia fluidized-bed combustor
Coal: Illinois (3.7 wt % S; -14 mesh)
Additive: Limestone No. 1359, as received
Starting Bed: Partially calcined and sulfated limestone No. 1359
Fluidized-Bed
Total
Run
Feed
Rates
Ca/S
Fluidizing-Air
Velaci ty
(ft/sec)
Secondary
Air
Injection
Rate
Flue Gas Concentrations
Sulfur
Expt.
A tic r-
Temperature
en
Time
(he)
Coal
(lb/hr)
Additive
(lb/hr)
Mole
Ratio
(% of Total
Air Feed)
°2
(vol %)
C0„
(vol X)
CO
(ppm)
NO
(ppm)
so2
(ppm)
Retention
7Aa
1550
6.5
5.0
1.7
2.9
li9
20
0.9
18.5
6000
100
3000
43
7Ba
1450
6.5
5.1
1.8
3.0
1.-8
20
1.0
17.2
6400
100
1100
80
7Cb
1550
7.0
5.2
1.8
3.0
lift
16
1.0
15.9
7800
70
500
91
7Db'C
1550
11.0
5.2
1.8
3.0
2:1
13
1.0
17.5
6500
70
1000
81
SDirection of secondary air flow was down, toward the fluidized bed.
^Direction of secondary air flow was up, away from the fluidized bed.
Cl)pper section of combustor (where the secondary air was reacting with CO and hydrocarbons) was insulaced.
Ui
-------
The higher SC>2 concent ration levels during An>er-7A and -7B,
in comparison to SO2 levels in Amer-7C, and the decrease in SO, levels upon
interruption of air injection from the probe in Amur-7A and -7S indicate
that possibly the following reactions occur. In the lower part of the
combustor, where too little oxygen is available for complete combustion,
one or both of the following reactions are postulated:
CaSO. + 4 CO i CaS + 4 C0o (8)
4 I
CaO + H2S + CaS + H20 (9)
When air is injected countercurrently into the off-gas from the fluidized
bed, a portion of the bed solids may be exposed to the air jet and the
following reaction may occur:
CaS + 1 1/2 02 t CaO + S02 (10)
Sulfide has been found in samples removed from the bed.
The sulfur retention corresponding to the 500 ppm S02 concen-
tration in the flue gas in Amer-7C is 91% (Table 6). This may be compared
with the sulfur retention of 80% obtained in one-stage combustion run Amer-8B.
In Amer-8B, the coal, limestone, and bed temperatures were the same as in
Amer-7C, but the Ca/S ratio was ^4 rather than 3.0 used in Amer-7C. It
would be expected that S02 removal would be less "than 80% in Amer-7C since
this run was performed at a Ca/S ratio of 3.0 and with other Amer-8B
operating conditions.
In run Amer-7D, an attempt was made to increase the gas
temperature above the fluid bed by insulating the top section of the
combustor. Thereby, the amount of CO and hydrocarbons combusted raight be
increased. Temperature in the second stage, which was 980°F in Amer-7C,
was 1275°F in run Amer-7D. The SO9 concentration in the flue gas in run
Amer-7D reached 1000 ppm, compared with 500 ppm observed in run Amer-7C
(which was made at the same conditions but with the second stage uninsulated).
To account for the increased S02 level at the higher second-stage temperature,
it is suggested that CaS in the solids being elutriated from the fluidized
bed reacted with oxygen, releasing S02 (reaction 10) and that the reaction
rate was greater at the higher temperature. This preliminary result also
suggests that provision should be made to remove most of the solids from
the flue gas prior to second-stage combustion.
b. Emission of NO. The concentrations of NO in the flue gas in
these two-stage runs ranged from 70 ppm during Amer-7C to 100 ppm during
Amer-7A and -7B. In contrast, NO levels in one-stage combustion have ranged
from 180 to 500 ppm. The 70 ppm level corresponds approximately to the
equilibrium NO concentration expected at this temperature and oxygen concen-
tration. The lew NO level experienced in two-stage combustion indicates
that with this mode of operation, nitrogen is not oxidized to NO during the
combustion process because of the high CO concentrations present in the bed.
c. CO Concentration in the Flue Gas. The concentrations of CO
in the flue gas were 6000, 6400, 7800, and 6500 ppm, respectively, for
Amer-7A, -7B, -7C, and -7D. During Amer-7A and -7B, interruption of secondary
-------
53
air injection through the probe resulted in higher levels of CO concen-
trations in the flue gas. Since oxygen-deficient conditions in the fluidized
bed produced high concentrations of CO and the low temperatures prevalent
at the downstream side of the fluidized bed did not favor the combustion
of CO to CO2, only about one-fourth of the CO reacted. It was also observed
that when tne temperature of an upper section of the combustor was increased
from 977°F (525°C) in Amer-7C to 1275°F (690°C) in Amer-7D, the CO concen-
tration decreased (see Table 6), indicating that the CO levels In the flue
gas night be further lowered by increasing the temperature of the off-gas
stream.
-------
III. MECHANISM OF THE LIME SULFATION REACTION
A. Introduction
The mechanism by which the reaction between lime and SO2 occurs to
form CaSO^ has been speculated upon and investigated by various groups6~10
outside this laboratory. In general, the two main hypotheses for the reaction
can be called the continuous reaction and the shell-formation models. In
the continuous model, the reaction is considered to occur more or less
uniformly throughout the lime particle, with the sulfating gas passing into
the interior of the particle through pores. In the shell model, the initial
reaction occurs on the surface of the particlej • the' continuing reaction
proceeds by passage ,of "the sulfating'gas through this initial sulfate layer,
and the shell thickness increases with time. The controlling factors in
these two reaction models would be gas diffusion through the pore structure
in one case and through the CaSO^ shell in the other.
Experimental work reported by Borgwardt^ of OAP supports the continuous-
reaction model, while work reported by Hatfield and Kim? at the TVA Muscle
Shoals site supports the shell-reaction model. Both modes of reaction were
supported by workers at Battelle Memorial Institute (BMI)^—the continuous
model for an oxygen-deficient reaction condition and the shell model for the
oxygen-sufficient reaction condition. In a study employing X-ray diffraction
and electron microscopy, McClellan and coworkers' came to the conclusion that
the sulfation reaction follows the shell model. Of interest in the results
reported by McClellan and coworkers is that CaSO^ crystallite size was
dependent on whether the calcination and sulfation reactions occurred'
simultaneousnr v° /~>v"ystcL^ wcis smul 1cit when
reactions occurred consecutively.
The reaction for sulfation under oxygen-rich conditions can be written
as follows,
CaO + S02 + 1/2 02 = CaS04 (11)
which can, in turn, be represented by the following pairs of reactions
either CaO + S02 = CaS03 (12)
CaS03 + 1/2 02 = CaS04 (13)
or S02 + 1/2 02 = S03 (14)
CaO + SO- = CaSO, (15)
J H
It has not been shown which reaction route prevails at the nominal fluidized-
bed temperature of about 1600°F. The presence of neither CaSO^ nor SO^ is
favored at this temperature; the S02 dissociation pressure from the
equilibrium for equation 12 is about 0.04 atm, while the SO3 partial pressure
for equation 14 is about 3 x 10-^ atm at equilibrium concentrations of SO2
and O2 of 0.0004 and 0.03 atm, respectively, which would be the nominal
concentration of these effluents from a fluidized-bed combustor. Although
these equilibrium data favor the reaction path employing the reaction of
-------
55
SO2 with lime, no kinetic data are available and the reaction path cannot
be stipulated at this time.
In work being performed in England on fluidized-bed gasification, the
reaction route through SO^ is advocated.^ Workers at BMlH have shown,
in experiments simulating limestone injection into flue gas for SO2 pickup,
that the short-time reaction produces mainly sulfite, while longer-term
reactions produce sulfate.
It must be kept in mind that most of the studies of the limestone-SC^
reaction to date have been directed at. conditions existing for limestone
injection into conventional pulverized-coal furnaces. This type of operation
involves shorter reaction times and higher reaction temperatures than in a
fluid-bed combustor where the lime particles would be retained for extended
periods of time. Therefore, data from the above-described investigations
can only be used to guide the fluid-bed combustion program.
In the ANL fluidized-bed program, the mechanism of the lime-SC^ reaction
has been a continuing area of interest. Early in the program, it was evident
that knowledge of the mechanism of the reaction would be important in devising
the most efficient reaction conditions for reducing SOo emission. Ample
evidence exists from the ANL work that the mechanism of sulfur oxide suppression
in the fluidized-bed combustor is much more complex than simple consecutive
reactions involving CaO, SO^, and C^. Many anomalous results have been
observed that cannot be explained by the reaction pair, eq. 12 and 13, or the
alternative reaction pair, eq. 14 and 15. It is reasonable to suppose that
the 'above reactions cio occur in the cc::3us to.", but tr.at these reactions alone
are not sufficient to account for all of the observed results. Some of the
evidence that supports a more complex'mechanism is the following: *'
1. When natural gas is burned in a partially sulfated bed of lime
(without fresh limestone addition), SO2 is released from the bed at
temperatures as low as 1650°F; yet in the absence of combustion, decomposition
of CaSO^ is known to occur only at much higher temperatures (>2000°F) .
2. The optimum temperature range for sulfur retention in the ANL
combustor has been observed to be 1450 to 1550°F. This temperature range
is appreciably lower than the optimum for SC^ removal from flue gases outside
the combustor (>1600°F). In the fluidized-bed combustor, sulfur retention
efficiency rapidly decreases at temperatures above 1600°F.
3. In the fluidized-bed combustor, the optimum temperature for sulfur
retention possibly varies with the Ca/S ratio, being lower at high ratios of
Ca/S. This is contrary to the expectation that the CaO-SC^-C^ reaction
should exhibit a fixed optimum temperature for a particular type of lime.
4. Recycle of flyash containing partially sulfated lime to the
fluidized-bed combustor at ANL does not increase SO2 removal to an observable
extent; yet this same flyash material is found to have considerable capacity
to react with SO-; when a sample in a boat is exposed to dilute SC^-air
mixtures in a furnace.
-------
56
5. When less-th an-stoichioir.et ric air is fed to the bottom of the
combustor and a stream of secondary air is directed toward the top of the
bed, sulfur retention by the lime bed is very poor; yet when the secondary
air stream is directed cocurrently with the off-gas and does not contact
bed material, sulfur retention is good.
The above observations suggest that the combustion process itself
influences sulfur oxide reaction with lime. During fluidized-bed combustion,
the sulfur appears to have an exceptional mobility for penetration into lime
particles.. The observation that SO2 can be released from bed material at
rather low temperatures (1650°F) during combustion of natural gas suggests
that chemical reduction of CaSO, occurs despite the introduction of excess
air to the bottom of the bed. The need to employ a relatively low temperature
(1450-1550°F) for optimum sulfur retention by the lime bed may be influenced
by the rate of the reduction reaction.
1. Mechanism - Macro Effects
Because not all needed information is on hand, our current concept
of the mechanism is necessarily speculative. Nevertheless, a summary of
our current thoughts about the mechanism might help illustrate its
complexity.
It is necessary, first, to consider the flow behavior of gases
in the fluidized bed. When air enters the bed at the gas distributor, a
portion of the gas flows through the bed in intimate contact with the solids .A
This mixture of gas and solids is called the emulsion phase. The remaining
Dortion of the air f.1 ows through the bed in the form of gas bubbles. As
the bubbles rise through the bed, gas from the bubbles continually circulates
into the emulsion phase and back into the bubbles. The gas in the bubble
phase reacts only as it circulates through the emulsion phase. On the basis
of combustion efficiency data for the fluid-bed combustor, it is known that
circulation of air from the bubbles to the emulsion phase is essentially
complete by the time the bubbles leave the bed. Nevertheless, it is reasonable
to assume that in the lower portion of the fluidized bed, not all of the air
from the bubble phase has circulated into the emulsion phase and that the
emulsion phase contains a deficiency of air. Thus the emulsion phase in the
lower part of the bed would contain a reducing atmosphere, and the emulsion
phase in the upper part of the bed would contain an oxidizing atmosphere.
The behavior of sulfur under the above conditions can be illustrated
by an extreme example from experimental work. In one part of a two-stage
coal combustion experiment (Amer-7A), the reducing condition in the bed was
accentuated by feeding less-than-stoichioinetric air at the bottom of the bed.
The sulfur concentration in the flue gas was 800 ppm. In another part of
the same experiment, additional secondary air was introduced through a tube
in such a way that the air impinged on the top of the bed and the sulfur
concentration in the flue gas increased to 3000 ppm. These results suggest
that sulfur was accepted by the lime in the lower part of the bed, forming
CaS, but that reaction of the sulfite with excess air in the upper part of
the bed released SC^. Of course, in a run made under typical single-stage
combustion conditions, extreme effects such as those in this example would
not be encountered.
-------
57
It-is well known that CaSO^ can be reduced rapidly to CaS by
reducing gases such as CO or hydrogen at temperatures of about 1750°F.
At lower temperatures, the rate decreases, and not much is known quantita-
tively about the kinetics at lower temperatures. The reaction of CaS with
air to produce CaO and SO2 is also well known, but again little is known
about the kinetics at the temperatures commonly used in the fluid-bed
comb us tor.
It is reasonable to assume that at some fluidized-bed temperatures,
a small portion of the CaSO^ in the bed can be reconverted to CaO and SO2
by a mechanism involving circulation of the particles into localized reducing
atmospheres resulting from fuel combustion. This was borne out in an
experiment in which evolution of SO2 was observed during combustion of
natural gas in a partially sulfated bed at 16'50°F. If oxidation of small
amounts of CaS is a part of the mechanism, some SO2 may be generated near
the top of the bed, limiting the overall sulfur removal.
It appears that the reaction in which CaSO^ is formed from lime
and SO2 is continually being reversed by exposure of the CaSO^ to localized
reducing conditions in the lower portion of the bed. The importance of the
reverse reaction would depend upon the conditions in the combustion bed;
lower temperatures would minimize the rate of the reverse reaction, high
concentrations of reductants would tend to promote the reaction. Thus, the
net SOj retention could be influenced by combustion conditions and by the
distribution of fuel (or the distribution of reducing agents) in the bed.
In a large conbustor, for example, the number of coal feed points might
.influence S0o removal. .Since-di'Mer-ent -coals-burn^-at different rates, the
local concentrations of gaseous reductants in the bed may vary for different
types of coal. ' Thus, the type of fuel burned may influence "sulfur removal,
and the optimum combustion temperature may vary slightly for different types
of fuel. A number of other factors could be important; for example, in a
large combustor with a deep bed that is heavily baffled by steam tubes, the
rate of solids circulation would be reduced, thus decreasing the reverse
reaction. Such factors could account for the variability of results achieved
in various types and sizes of combustors using various types of coal.
2. Mechanism - Micro Effects
If the circulating bed particles are continually cycling into and
out of a reducing atmosphere and SO2 can thereby be regenerated from the
CaSO/, this would have an effect on the sulfation of individual additive
<4
particles. When a fresh lime (or limestone) particle is fed to the bed,
reaction with SO2 occurs rapidly and CaSO^ forms, mostly on the outer
surfaces and in the pores of the particle. A condition is soon reached
wherein the probability that an SO2 molecule will be released from the
particle by the reduction mechanism is almost as great as the probability
that an S09 molecule will be accepted from the bulk gas phase. The particle
would then appear to have a low reactivity for SO2, although it would still
have a considerable capacity for accepting SO2 at a low rate.
The average residence time of a particle in the fluidized bed is
several hours. During-this long residence period, the particle would be
subjected to many cycles of alternate reducing and oxidizing conditions.
-------
58
Any SO^ released by Che reduction mechanism would have an opportunity either
to penetrate deeper into the particle or to escape into the gas phase.
Ultimately, a condition would be readied --.-herein the sulfur would be
uniformly distributed within the particle at a concentration determined by
a pseudoequilibrium set up by the sulfur acceptance and rejection cycle.
This sulfur concentration would be appreciably less than that represented
by total conversion of CaSO^ since at this level of sulfation, any additional
pickup of SO2 would be balanced by release of SC^ and the sulfur level would
remain constant.
The above-conjectured mechanism would limit total sulfur pickup
by the lime particles. On the other hand, the sulfur would have a greater
mobility for penetration into particlesy and there would be no sulfate
shells (which would seriously reduce the reactivity and capacity). As
mentioned previously, preliminary experimental data indicate that the net
effect is that fluidized lime beds in which combustion is occurring have a
greater reactivity and capacity for SO2 removal than do similar beds in
which combustion is not occurring.
Let us now consider the fine lime particles contained in recycled
flyash. These particles, which are only lightly converted to sulfate, have
considerable capacity for accepting additional SO2, as is observed when
they are exposed to SO^-containing gases outside the combustor. Yet when
the elutriated material is recycled to the combustor, the additional lime
has little apparent effect on S0„ retention. This could be explained by
assuming that the outer sulfate-5earing portion of the particles is already
in sulfur-equilibrium with the fluidized-bed coisbustion medium. Because
these fine particles have a very short residence time in the fluidized bed,
further sulfur penetration into the particles is insignificant. Thus, the
recycled particles pass through the combustor bed with little apparent
reaction.
Next considered is feeding (to a coarse-particle bed of partially
sulfated lime) of coal and limestone at rates corresponding to very high
ratios of li~:e to sulfur. As the Ca/S ratio is increased, the average
residence time of the additive particles decreases proportionately. At high
Ca/S ratios, the probability that an SO2 molecule will be picked up near
the outer surface of a relatively fresh particle rather than penetrating
into an aged particle increases. Thus, at high Ca/S ratios, a sulfur gradient
within a particle is more likely. Under these circumstances, the relative
effect of the "reverse" reaction (reduction of CaSO^) becomes greater, and
the temperature for optimum removal is slightly lower. This hypothesis
conforms to the experimental evidence that the optimum combustion temperature
is lower at a Ca/S ratio of four (1450°F) than at a Ca/S ratio of two (1550°F).
B. Experimental
The mechanism of the sulfation reaction has been investigated by two
approaches: (1) microprobe examination of sulfated lime particles to
determine the sulfur distribution and (2) laboratory-scale reduction of
sulfated limestone under moderate temperature conditions. An objective of
the latter experiments was to evaluate the hypothesis that the sulfation reaction,
proceeds via alternate sulfation and reduction by studying the evolution of
SO2 from sulfated limestone. Microprobe examination would reveal differences
-------
59
in the distributions of sulfur in lime particles that in turn might indicate
how the mechanism of sulfation during coal combustion differs from the
mechanism when limestone is exposed to a SC^-air mixture during natural
gas combustion.
A Model EMX-SM electron microprobe manufactured by Applied Research
Laboratory was used to examine sulfated lime particles. The electron probe
analyzer offers nondestructive, in situ elemental analysis of areas as small
as one square micrometer. It can quantitatively determine all elements having
atomic numbers higher than that of sodium in samples as small as a few cubic
micrometers in volume, or on surfaces as small as 1 ym in diameter. Accuracy
is 1-2%. The probe uses the sample as a target for a magnetically focused
beam .of .highly .accelerated electrons. Upon striking the sample, the electrons
excite the emission of X-rays, whose wavelengths characterize the elements
present and whose intensities yield concentrations. When the probe is
adapted for scanning instead of point analysis, it provides a clear map of
element distribution over the sample surface. The equipment and instrument
components of the bench-scale fluidized-bed comb us tor, in which the solids
examined with the microprobe were generated, are described in detail in
Section III.B of this report. The partially sulfated particles sampled for
electron probe examination were cold-mounted in Polylite, polished through
1-um diamond paste, then either carbon-^ or gold-coated by vacuum sputtering.
Microphotographs were then made of selected particles or areas of the mounts
to guide the microprobe work.
The equipment (Fig. 10) for laboratory-scale study of the evolution of
S09 from sulfated Time 'in a reducing atmosphere :wS3 a ".edified porticn of
the laboratory-scale apparatus constructed for high-pressure chemical reaction
studies. It comprised an electrically heated 2-in.-dia stainless steel
vertical fixed-bed reactor, associated manifolding for supplying and mixing
pure gases from cylinders, and provisions for water saturation of the feed
gas stream at the inlet to the reactor. The off-gas from the reactor was
dried and filtered prior to continuous analysis by the analytical system
described in ANL/ES/CEN-1002, pp. 21-22.
1. Microorobe Studies
a. Elutriated Particles. Samples of material elutriated to the
primary cyclone in two runs were examined; one sample was taken at the end
of a coal co:?bus cion experiment and the other at the end of a natural gas
combustion experiment. Fresh limestone No. 1359 additive employed in these
experiments had an average particle size of 25 ym and contained 97% CaO
and 1.1% MgO on a calcined basis. In both experiments, the starting fluid-
bed material was refractory alumina.
Electron probe scanning traces (not shown) for cross sections
of four typical elutriated particles from each experiment showed that particle
composition apparently is related to the type of fuel combusted. With coal
as the fuel, sulfur concentration is usually lower at increasing distance
from the surface of the particle; on the other hand, sulfur concentrations
in particles collected during the combustion of natural gas appear to be
more uniform over a range of distances from the particle surface. Only one
trace out of four displays the calcium and sulfur concentrations expected for
-------
CaSO/. Traces for soir.e particles appear to indicate that their composition
was CaCO^, suggesting that these fine limestone particles were not calcined
when passing through the bed.
EXHAUST To
Hood plenum
00}
TO Art^LmCML
rt^r/.'COLO
-sitJTteeo
M£T#l Filter
nO,
I—tl
i—tr
Fgom crurJoE/zs
BM 8/1LL SALVE.
px3 rJeeoie /&wa
exj GATE VP.L/E
Q Pop off f/ujfe.
' +/> OH,-, DC
PUZMCC n//TH
tZEACTOK ¦
_ '— t-L/i^rirlCc" >-»'/ 7%
WATER m/Mii-i*. ZLZnCTCZ.
STEAM T3AP i
ZEoaarts sr/tneuj
Fig. 10. Laboratory-Scale Reactors and Associated Equipment
b. Fluid-Bed Particles. One set of electron probe analyses
was performed on fluid-bed samples from coal combustion experiments BC-6,
AR-1, and AR-2, in which the starting fluid-bed material was partially
calcined and sulfated limestone or dolomite. (This dolomite contains 57%
CaO and 42.2% MgO in the calcined state.) Sulfur, calcium, and magnesium
levels over cross sections of particles of limestone No. 1359 (BC-6 and
AR-2) and dolomite No. 1337 (AR-1) were measured.
Since experiments BC-6 and AR-1.were both performed under
ordinary oxidizing conditions (^3 vol % oxygen in the off-gas), the
distribution of sulfur across a cross section of a particle was expected
to conform to that' for a shell diffusion reaction mechanism model, i.e.,
a high sulfur concentration at the edges and a drastically lewer concen-
tration at the center of the particle. The basis for this expectation was
work reported by Battelle Memorial Institute,^ in which microprobe analysis
showed a shell structure when reaction occurred under .oxidizing conditions,
but a uniform sulfur distribution in the particle when reaction occurred unde
-------
61
The sulfur traces for experiments BC-6 and AR-1 indicate that
the sulfur was distributed almost uniformly over the cross sections of the
300- to 400-ym-dia particles. The average sulfur concentrations in the
limestone and limestone-dolomite beds from which these particles were taken
were 7.2 and 9.5%, respectively. The sulfur content for a completely sulfated
limestone is calculated to be 23.5% and that for a completely sulfated
dolomite 18.2%.
Scans were also performed for bed particles from coal
conbustion experiment AR-2 (Table 1) . Part of this run was performed under
relatively mild oxidizing conditions with only ^1 vol % oxygen in the flue
gas. Examples were found of both the shell model and the uniform concen-
tration mod^i for S07 reaction with limestone. Either model (the shell
model or continuous penetration model) or both could and in fact appear
to be involved.
At the time particle samples were taken, the fluid bed had
been used for a total of about 70 hr in seven experiments (BC-6 to -10,
AR-1, AR-2) and fresh limestone and dolomite had been added during these
experiments. Because any one particle could have been in the bed for a
long time before samples were taken, individual particle histories could
not be determined. Moreover, the sulfur concentration profile for a particle
having a total possible bed residence time of 11 hr is similar to that for
another particle having a total possible bed residence time of 70 hr. This
suggests that a long residence time in the fluid bed is not an important
factor in determining the extent or shape of the sulfur concentration profile.
c. Limestone Bed Particles from Control Experiments. To eliminate
the uncertainty about the reaction history of individual limestone particles,
two control experiments (Table 7) were performed in the bench-scale combustor,
each with a starting bed of fresh limestone and with no additive fed during
the experiment. In one experiment (PB-3), the limestone No. 1359 was sulfated
by SO2 generated by the combustion of coal. In the second experiment (PB-7) ,
the bed was electrically heated to reaction temperature, then the limestone
No. 1359 was sulfated by S09 in a mixture with air in the absence of combustion
but with heat supplied by electric heating units. Bed samples were taken
frequently during the experiments.
(1) Coal Combustion Experiment PB-3. At the start of
experiment PB-3, SO2 concentration in the flue gas increased with time as
the lime bed was sulfated, as expected. The SO2 concentration finally
reached a level near that predicted for the absence of absorption of SO2
by the lime bed. Samples taken during the course of this experiment were
examined with the tnicroprobe. Seven typical sulfur traces from the micro-
probe analysis are shown in Fig. 11. A photomicrograph of the particle
examined in scan 1 (Fig. 11) is shown in Fig. 12. Sulfur distributions
shown in scans 1, 2, 4, and 6 of Fig. 11 support the idea that sulfation
proceeds throughout the body of the particle. However, scans 3, 5, and 7
shew sulfur concentrated in shells on the particles. The latter particles
might be intermediates in the sulfation cycle, representing the condition
in which sulfation is proceeding inward with time. In bed samples taken
more than 2 hr after the start of a run, approximately half of the particles
examined showed uniform sulfur distribution and the other half showed shells
-------
of sulfate. The presence of both uniformly sulfated particles and particle
having shells of sulfate indicates that sulfation proceeds by diffusion
through a layer of sulfate.
TABLE 7. Operating Conditions for Experiments PB-3 and PB-7
in Bench-Scale Corabustor
Starting Bed: Fresh Limestone No. 1359
Fluidized-Bed Temperature: 1500°F
Gas Velocity: ^2.7 ft/sec
PB-3 PB-7
Elapsed time, hr 34 8
Source of SC^ Combustion of -14 mesh Feed of 0.6 vol X SO^
Illinois coal in air, equivalent to
combustion of 6.4 lb/hr
coal
Coal combustion 4.5
rate, lb/hr
Oxygen concentration 2.3 3.0
(dry basis) in the flue gas,
vol %
It should be noted that the maximum sulfur concen-
tration level at any point in any of the particles examined corresponds to
a count rate of about 500 cps, which is estimated to represent about 50%
conversion. This maximum sulfur level is reached at the outer surfaces of
the particles within 6 hr residence time. Thereafter, the sulfur level at
the outer surfaces of the particles remains constant, and all additional
sulfation occurs within the particles. This is also suggested by the
observation that SC^ concentration in the flue gas increases rapidly after
about 6 hr.
The percentage conversion to CaSO^ during the course
of experiment PB-3 (Table 8) was determined by chemical analyses of bed
samples. These data show that ^30% of the lime was converted to sulfate
in 14 hr and that the extent of lime sulfation essentially stayed at that
level for the remainder of the experiment.
(2) Noncombustion Experiment PB-7. The SO2 concentrations
in the effluent during experiment PB-7 (in which the gas phase fed was a
mixture of about 0.6 vol % SO2 in air) were very low during the first half-
hour, increased rapidly during the next hour, and increased at a moderate
rate during the following 6 1/2 hr. The sulfation rate decreased much more
rapidly in experiment FB-7 than in PB-3.
-------
63
li:li
-i+hwh?
-l-l-l-j ! H-!
|_ -
L 1L
I--
1
j—
life
i—
- -t- -
.d ; Li » iL I
mVl mW hri
w
vWW
_ _Li_ i.
-H-l-1-
\*vJa
Ji==!:
U)
Cu
U -l_i
Scan 1 for particle sampled at 2 hr; scan length, ^1100 ym
PTFI
O
O
o
w
p.
o
o
o
o
-j-
¦l i V.
beanL
ri; ri| rUl
i-fc nnyil MWW l I ¦! I t_: ^ 1 > I iJi i I. I. I J TtlBM II II I LJ.-J ¦ » ¦¦ . 1- I
can i roc ^arcicie sanpiea at o uit; scan length, ^601) (iiii
|i'i.i ii , i i tj--p-| i . i. i i ir~ ~ Jl,^r'7"|-xzr7|ri^i
r^~JTrTHTT^=~rf:
izl: iT.r;.j-L,r;:i:r M-TT-h-; lit
ik-fibti
: l-i-KdH-F
" 'tf-
-M-r
_J—l_t ,
- r i.. it
FF
-i-
L±
7W1
zkh
Scan 3 for particle sampled at 6 hr; scan length, ^1000 ym
-1 -f ~ r- f- r r-f 4-4 - r U
:i leiiscn, "OUU um
i° n'frTTJTI
Cfl
a.
CJ
o
o
o
Scan 5 for particle sampled at 10 hr; scan length, A^600 ym
; i ! "I i_]Ti ]':~?_ !Tn f-r-'-t-f- -t r-pf~)
'v V ^ I
'1
-1
; : " ! ; i 1 . ~'"j ! i i'iTi !
I < ;
i . . . : ; i -! ; j ; i
; i , . I | i h .
"LiL'1"!1 ^ ror ')~1 r c a:- J" " L ' sc.'.a ' umgdu,' '^iu'OtJ
-P
¦*\!7
r! T
5
inn
i
W
d.
U
O
o
o
r* - 1 » v- i- ! : : 5 : \ ' *
| I - : ¦ ' : ! 1 ; ; ¦ ' ! j
i ' 1 !
ji ¦i i 'j jJJ | i!1 [_
Scan 7 foe particle sampled at 14 hr; scan length, ^1000 um
Fig. 11. Electro.! Micron rube Traces Showing Relative Sulfur Levels
in Sulfated Particles, Experiment P13-3
-------
64
SCAN 1
x50
Fig. 12. Photomicrograph of Particle Examined in Scan 1,
Experiment FB-3
Table 8. Extent of Sulfation of Bed Particles as a Function
of Reaction Time, Experiments PB-3 and PB-7
Elapsed Micro-
Time probe S Ca Ca/S
(hr) Scan (wt %) (wt %) Mole Ratio
/ Sulfated
Experiment PB-3
2 Yes
6 Yes
10 Yes
1A Yes
31 No
3.1
8.4
11.5
12.7
11.2
62.8
48.6
41.8
37.9
38.7
16
4.6
2.9
2.4
2.8
5.5
8.5
29.3
33
28
Experiment PB-7
2 Yes
8 Yes
2.8
4.0
59.5
57.5
17
12
4.5
7.5
-------
65
Microprobe scans of particles from bed samples removed
at 2, 4, and 8 hr are shown in Fig. 13. The microprobe scans for experiment
PB-7'are distinctly different from those for experiment PB-3. Depth of
¦sulfation of the particles was very limited, and increasing reaction time
did not produce greater penetration. The level of sulfur concentration at
the outer surface of the particles from PB-7 was about twice that in
experiment PB-3, and is estimated to represent 100% conversion to CaSO,.
The total sulfation, however, was substantially less, as shown by the aata
in Table 8. The highest level of conversion to CaSO, in PB-7 was only about
8%.
s. in in.?-:!!!-
u
h rrh-rH-^H
r rn- i . i~t
Scan 1 for particle sampled at 2 hr; scan length, ^1200
um
Scan 2 for particle sampled at 2 hr; scan length, ^1000
um
! IT! i
Scan 3 for particle sampled at 4 hr; scan length, ^700 pm
:Hr
CA
C-
u
o
o
o
Scan 4 for particle sampled at 8 hr; scan length, VL200 pm
::]¦>}.!r[
I . ,
—! -i '
-"i • r
-! / '! ! ' ' i L '
1 I —'/ .!
ucrmrru:i i .t-| j
1 • r '' ; i
i
-4
:.j L\ T | ;-f-^
- i-rpi
4-
¦ . 1 ! ! !
_ .Lj_L
i-
_L_
—
-/M
—1—i—t—j—
-I--
4:
is
i -
\ - -
..j ij-i:L
-' ! ! !
—r -I * i -I -1 -
T
m
Scan 5 for particle sampled at 8 hr; scan length, ^800 pm
Fig. 13. Microprobe Scans of Sulfated Particles,
Experiment PB-7
-------
66
These results suggest that the mode of sulfur acceptance
differed in the two experiments. In experiment PB-7, a nearly impenetrable
sulfate shell built up that limited further sulfur pickup by the lime. In
experiment PB-3, the sulfate shell was much more easily penetrated by SO2.
Since the principal differences between the two experi-
ments were the source of the S02 and the reaction environment (generation of
S09 by combustion in experiment PB-3 and the absence of combustion in experi-
ment PB-7), the presence of the combustion reaction apparently is important.
These results strongly support the hypothesis that partial reducing conditions
in the emulsion phase of the fluid bed can act to release bound SO2, allowing
it to be transported deeper into the lime particles.
2. Laboratory-Scale Experiments—Reduction of Sulfat^d Iir.estone
As was stated earlier (ANL/ES/CEN-1002, pp. 50-52), it has been
observed that when natural gas is burned in a partially sulfated bed of lima
(without fresh limestone addition), SOp is released from the bed at temperatures
as low as 1650°F; yet in the absence of combustion, decomposition of CaSO^
is known to occur only at much higher temperatures (>2000°F). A chemical
reaction that would release SO2 from sulfated lime is the following:
CaS04 + CO = CaO + C02 + SC>2 (16)
Two intermediate reactions control the rate of this reaction:
1/4 CaS04 + CO = 1/4 CaS + C02 (17)
3/4 CaS04 + 1/4 CaS = CaO + S02 (18)
The equilibria for these reactions are discussed in Section IV of this report.
Laboratory-scale experiments were performed to determine the effect
of CO on the sulfation reaction and to investigate the use of CO in a
regeneration step to remove S02 from sulfated lime (discussed further in
Section IV). The procedure consisted of exposing a mixture of partially
sulfated lime and particulate alumina in a vertical fixed-bed reactor to a
mixture of gases that included CO and measuring the S02 content of the
effluent gas. The partially sulfated lime was bed material from a bench-
scale fluidized-bed coal combustion experiment (DUO-1) and typically would
have a CaO conversion to CaSO, of about 30%. In each experiment, 200 g of
the bed material was dispersed in an equal volume of refractory alumina
spheres having an average diameter of somewhat less than 1/8 in. The
mixture of lime bed material and alumina balls was supported on a 3- to 4-in.
layer of alumina balls of the same size. A similar quantity of alumina
balls was placed on top of the section containing the bed material.
In an experiment, the reactor was heated to the selected temperature
with only nitrogen flowing through the reactor. When the selected reaction
temperature was reached, a gas mixture containing CO, C02, H20, and N„ was
admitted to the reactor at a flow rate of about 4 liters/min. Water t>3 vol %)
was added to the C0-C02-N2 mixture by passing it through a bubbler at room
temperature. The effluent gas phase was analyzed for CO, C02, and SO2; at
-------
67
the end of the experiment, the solid phase was submitted for analysis for
total sulfur, SO^, and S~ content.
Four experiments (Mech-1 to -4) were performed. The initial
stage of each experiment was the treatment with CO-COj-^O mixture in
nitrogen, described above. Additionally, as the final stages of Mech-2
and -3, air was passed through the reactor to react with any CaS present
in the bed.
The SC>2 and CO concentrations in the effluent gas during Mech-1
are shown in the plots of Fig. 14. During the initial period when only
nitrogen was passed through the reactor, SO2 was evolved. The concentration
of-S0„ -peaked at about 4200 .ppm ,¦within 1/2 hr as the temperature was increased
from 1500 to 1700°F, then fell to about 200 ppm after an additional 1/2 hr
at 1700°F, staying at that level until the CO-containing gas was admitted
several hours later. A possible explanation for the large evolution of SOn
during the nitrogen flow period is decomposition of calcium sulfite (CaSO^),
which might have been an intermediate reaction product on the surface of the
lime particles. The sulfite has a much lower decomposition temperature than
does the sulfate. Another possibility is that reducing agents such as
carbon or hydrocarbons in the bed may have reduced the CaSO^.
When CO-containing gas was admitted at about 1750°F, SO2 was
evolved immediately. Its concentration in the gas phase rose to about
6000 ppm. However, dependence of the SO2 concentration on the CO concen-
trstiop., prsdictsd by sc^uiiibrium snd phels0 ruls consi.dsrsti.ons (sss
Section IV), was not observed. The presence or absence of water vapor in
the reacting gas did affect SO2 evolution; the SO2 concentration dropped to
about 3000 ppm when feeding of ^0 in the reaction gas mixture was stopped.
The results of gas phase analysis for the second experiment,
Mech-2, are shown in Fig. 15. In this experiment, the initial reaction
temperature was 1650°F. During the period when only nitrogen was passing
through the reactor, SO2 evolution was noted but the initial peak was much
lower than .in Mech-1 at 1750°F. When the reducing gas (10% CO, 15% CO2,
^3% II >0, aad }'>-,) vis 2 inictui to the reactor, a spike was noted in the SO2
evolution trace, hue cheii the SO2 concentration rapidly dropped to about
500 ppm. Increasing the reaction temperature to 1750°F did not greatly
change SO2 evolution. On the chance that reaction 17 was dominating the
process and CaS was being formed, the reducing gas mixture was replaced with
air to convert the CaS to CaO and SO2 according to the reaction
CaS + 3/2 02 = CaO + S02. (19)
This change brought about an immediate increase in the SO2 concentration to
somewhat more than 4 vol %. The high rate of SO2 evolution continued for
about 10 min, then the SO2 level dropped gradually to several thousand ppm,
indicating CaS was present.
In Mech-3, the portion of the experiment in which reducing gas
was fed was performed at 1500°F to determine if the formation of CaS would
be favored. This reaction produced little SO2 during either the 1500°F
reducing step or the air oxidation step at temperatures up to 1800°F. This
-------
68
result indicates that a temperature of 1500°F is too low for reaction 17.
In experiment Mech-4, with reducing gas fed at 1600°F, similar SC^ data
were obtained as at 1750°F in experiment Mech-2.
In general, these laboratory-scale experiments show that SC^ can
be released under reducing conditions. When the temperature is about 1600°F,
CaS is formed instead of SC^. At 1300°F, no reaction apparently occurs.
The results lend support to the hypothesis that in the combustor, CaSO^ is
cyclically reduced to CaS which is then oxidized to generate SO2. This
reaction could contribute to SO^ penetration into the lime particle and could
limit SCU removal at high combustion temperatures. The reduction mechanism
may still be operative at a fluidized-bed temperature of 1500°F, since
particle temperatures are known to exceed the measured bed temperatures.
In addition, these results suggest that recovery of SCL by a two-
step cycle employing reactions 17 and 19 is feasible at relatively moderate
temperature. These reactions are also favorable in that they are not
repressed by high-pressure operation. In fact, reaction 19 would be enhanced
by operation at high pressure.
-------
1700]
No
present
10% CO
N' only
5%
CO
10*
CO
IZ
CO
CO
n
e.
oouo
X
ft
500>
c
o
Kf f ] uent SO,
-a
c
r-j
KffIuent CO
1000
Reaction Time, hr
Fig. 14. Reaction of Sulfated Lime with CO-CO^-H^O-N^ Mixtures,
Experiment Mech-1
o*
SO
-------
2004
— 1800
increase Co
1 7 50° K ^
1600
Inlet Cas
N_ only
1200
107. C0-15X CO
Ail-
only
mixture
— 1000
effluent CO
000
5 —
SO
.000
4 —
0
3000
3 —
000
000
e f f 1 uent SO
30
Reaction Time, min
Reaction Tlrae, lir
Fig. 15. Reaction of Sulf ited LLine with CO-CC^-^O-^ Mixtures
Followed by Read, ion V7ith Air, Experiment Mech-2
-------
71
IV. REGENERATION OF CaO FROM CaSO^*
Regeneration of partially sulfated lime would be desirable (1) for
.recycle of the lime, thereby reducing the solids waste burden imposed by
the sulfur-removal process and (2) to recover a sulfur value as sulfuric
acid or elemental sulfur.
Listed below are chemical reactions for a process to regenerate CaO
from CaSO^ using Co reductant (H^ is an alternative reductant):
CaS04(s) + CO(g) ? CaO(s) + C02(g) + S02(g) (20)
*1 X1 X 1
1/4 CaSO (s) + C0(g) J 1/4 CaS(s) + CO (g) (21)
1/4 x2 x2
3/4 CaS04(s) + 1/4 CaS(s) £ CaO(s) + S02(g) (22)
Reactions 20 and 22 are endothermic, while reaction 21 is exothermic. Since
this system is considered to be in equilibrium, the phase rule can be applied:
P + F = C + 2 (23)
where P = number of phases at equilibrium
F = number of degrees of freedom
C = number of components
There are our components and four phases (CaSO^, CAO, CaS, and the gas
phase), giving a total of two degrees of freedom. However, there is a
restriction caused by the fixed ratio between calcium and sulfur in the
initial material (CaSO^). This cancels one degree of freedom. This could
be temperature, or pressure, or one mole fraction in the gas phase. Obviously
this univariant system exists only under very limited conditions. Under
other conditions other reactions will occur. The range of conditions will
be expanded in further work.
Only two of the reactions 20, 21, and 22 are independent since any
two reactions can be combined to give the third. Therefore, reactions 20
and 21 were arbitrarily chosen to make the equilibrium calculations.
Let x^ = moles of each product formed by
reaction 20 at equilibrium
x2 = moles of each product formed by
reaction 21 at equilibrium
y^ = mole fraction CO in feed gas
y2 = mole fraction C02 in feed gas
y^ = mole fraction inert gases (N2) in
feed gas.
~
We are indebted to Professor Scott Wood of the Illinois Institute of
Technology for his advice and assistance in these equilibrium calculations,
-------
For one mole of entering gas, the equilibrium composition for each
compound is given in Table 9. From this,
CO,
SO,
y2 + X1 + X2
1 + x,
,1 + x.
K_
(24)
CO
yl ~ X1 + X2
1 + x.
CO,
and
P
2 CO
y2 + *1 * "2
1 + X^
yl - xl - x2
1 + x.
(25)
(y + x + x )(x ) 71
which reduces to K_ = n r (24a)
1 (yl " X1 " x2 Xl^
y2 + x + x
and K_ = — (25a)
2 yl ' X1 " X2
TT (x )
and finally to K = K ——: r (26)
T ?2 (1 -r
where it = total pressure
Equation 26 is consistent with the phase rule since the fixing of two
degrees of freedom (temperature, pressure, or mole fraction SO2) provides
an invariant system at equilibrium.
A. Effect of Pressure
An important variable in the regeneration process is the total pressure
of the system. It may be desirable that the regeneration cycle operate at
the same pressure as the combustor since this would facilitate coupling the
regenerator to the combustor for a continuous mode of -operation. Therefore,
a high-pressure process is envisioned for both fluidized-bed combustion and
regeneration. With a high-pressure process either the throughput with smaller
equipment will be equivalent to that obtained at 1 atm with large equipment,
or the throughput will be increased in equipment having the same size as
atmospheric units. The effect of pressure—especially on the concentration
of SO2 in the gas phase—is being evaluated.
-------
TABLE 9. Equilibrium Composition for CO and C02
during Regeneration of CaSO^
Basis: 1 mole of feed gas
Inlet Composition
(mole fraction)
Equilibrium
Conversion
(moles)
Total
Moles at
Equilibrium
CO
yl
" X1 " X2
yl " X1 " X2
CO 2
y2
+ + x2
y2 + xJi + x2
so2
0
+ X1
+ X1
N2
y3
0
y3
1.00
+ X1
1.00 + x
Equation 26 has been used to calculate the equilibrium pressure of S02
at system pressures of 1 and 10 atm over the temperature range 1400 to 2000°F
(Fig. 16). It is important to note that the SO2 mole fraction is inversely
proportional to the total system pressure at any fixed temperature and that
on the basis of equilibrium considerations, pressurized regeneration of
CaSO^ is not attractive from this standpoint. The equilibrium calculations
indicate that the expected SO2 yield at 10 atm and 1950°F would be no greater
than 4 vol %, a relatively low concentration if the gases are to be fed.to
a Claus plant for recovery of sulfur.
12
A reaction scheme proposed by Squires and Graff is under consideration
but has not yet been studied experimentally. In their reaction scheme,
partial reduction of CaSO^ to CaS with hydrogen is followed by the reaction
of CaS with CO2 and ^0 to form CaCO^ and I^S as shown in the following
reaction:
CaS + H20 + C02 = CaC03 + H2S
This reaction is reported to occur at 1100°F and to be favored by high
pressure.
-------
so,,
600
SO,
ppfi
100
10
Temp
1400
2000
1900
1800
1700
1600
1500
0.01
T Oy x 10
Fig. 16. Equilibrium Mole Fraction SC^ at 1400-2000°F
CaSO. + CO = CaO + C0o + SCL
4 2 2
1/4 CaS04 + CO = 1/4 CaS + C02
3/4 CaSO, + 1/4 CaS = CaO + SO„
4 I
-------
75
B. Experimental
Initial laboratory-scale experimental work to evaluate the effect of
pressure has been on the reduction of CaSO^ in a fluid bed, using CO as the
reductant. Two experiments were performed at 1950°C and pressures of 1 and
10 atm. The gas velocity was 2 ft/sec, sufficient to fluidize the partially
sulfated and calcined lime beds. The first of the regeneration experiments
was performed at 1-atm pressure (Regen-5) and the second at 10-atm pressure
(Regen-9).
The laboratory-scale regeneration reactor used in these experiments
is shown diagrammatically in Fig. 17. The 2-in.-dia vertical tube reactor
.is. constr.uc.texl .of t.y.pe .316 .s.tainless .s.teel. .It is about 5 ft .long, with
the lower 2 ft enclosed by electric furnaces. The upper part of the
reactor is insulated to reduce heat loss.
The starting fluid-bed material (from previous experiments) consisted
of 628 g of sulfated limestone containing 10.7 wt % sulfur in Regen-5 and
consisted of 1169 g of sulfated limestone containing 10.0 wt % sulfur in
Regen-9. Nitrogen flowed through the reactor at incipient fluidization
velocity as the reactor was heated by electric furnaces to about 1700°F.
To increase the temperature to 1950°F, CO was burned. At this temperature,
a gas mixture consisting of 10 vol % CO-20 vol % CO2 and the remainder
nitrogen and oxygen (when required) was admitted to the reactor. The reactor
effluent gas stream was passed through a filter chamber to remove entrained
solids, through a pressure—reducing stage if necessary, and then at nearly
atmospheric pressure to the gas-sampling manifold for SO2, CO, CO^, and 0£
analysis. The solid residue in the reactor was analyzed for total sulfur
and for sulfur present as sulfide.
During the experiment at 1 atm (Regen-5), the SO2 concentration in the
reactor effluent rose to 3 to 4 vol % within the first minute of reaction
and then decreased during a 20-min period to essentially zero. The CO
concentration in the effluent was low, less than 5000 ppm, during the period
of peak S09 evolution and then rose rapidly as SO2 evolution subsided. The
solid residue confined 3.2 wt % sulfur, of which 2.8 wt % was sulfide (CaS).
The results sncv; tnat about 80% of the sulfur contained in the original
sample was converted to S0„. Nearly 90% of the sulfur in the residue was
sulfide, indicating that the reaction had stopped short of complete conversion
to S09 because of an unfavorable CaS to CaSO^ ratio.
During experiment Regen-9 (performed at a total pressure of 10 atm),
control of the temperature in the fluidized bed was difficult and sporadic
at best, apparently because of partial caking of the bed soon after the
CO-CO2-O2-N2 mixture was fed.
-------
76
high pressure
Blovback
solenoid
Oxygen
To depressurisat1 on
^ and analytical
system
Sintered metal
-bayonet filter
Internal thermocouple
Internal thermocouple
Marshal furnaces
Internal thermocouple
Internal thermocouple
3/£" Alumina ball
gas distributor
From gas s upply
manifold and
p reheater
(Not to scale)
Fig. 17. Fluidized-Bed Regenerator (2-in. dia)
-------
77
V. MODELLING STUDIES, SO REMOVAL
(Lowell B. Koppel;
The process of reducing the emissions of SO2 by heterogeneous reaction
with additive limestone particles in a fluidized-bed corobustor is mathemat-
ically complex. The objective of this study is to devise a mathematical
model which is convenient for use and which yields reasonably accurate
predictions of removal level and predicts how SO2 removal is affected by
typical operating variables. To achieve this, several assumptions regarding
the process are made.
A. Model and Assumptions
Figure 18 illustrates certain of the model assumptions. In accord
with current theory on flow through fluidized beds, the bed is assumed to be
divided into two phases—a particulate phase through which gas flows at the
minimum fluidizing velocity, and a bubble phase which contains all the gas
flow in excess of that required for incipient fluidization of the bed. The
gas in the particulate phase is in plug flow, having uniform SO2 concentration
across any horizontal plane. Additive particles in the particulate phase
are perfectly mixed so that the entire particulate phase is at a uniform
temperature. As the additive particles in the bed absorb SO2, their average
reactivity (expressed in terms of first-order reaction velocity constant)
is assumed to decrease linearly with the extent of SO2 absorption. Gas is
continuously exchanged between the bubble and particulate phases.
F (contains unremoved SO^ at
concentration c)
Solids
withdrawal
Combustion
/ , so2 /
Generation
Bubble Phase
\_/(Gas only, No SO- generation)
Particulate /
Phase
(incipiently
fluidized)
Gas
Transfer
Column Uall
Column
Wall
Additive Feed, v
Coal Feed
v. Distributor
Gas Feed, F
Fig. 18. Schematic of Fluidized Bed, Illustrating Model Assumptions
*
L. B. Koppei, Purdue University, consultant to the Chemical Engineering
Division, Argonne National Laboratory.
-------
With these assumptions, and with a group of curves (to be discussed
below) to describe the SO^ generation pattern, equations can be derived to
predict SO2 removal R, as well as particle consumption C. Numerical study
of these equations showed that the effects of gas bubbles bypassing the bed
could be omitted. Hence, the equations for SO2 removal R and particle
consumption C are reproduced here only for the case of no bubble flow (3=0).
, -a -H(l - -).
R = 1 - a[e _a~6 R r ] (27)
(1 - e )[H(l - p - a]
C = - (28)
r
Symbols are defined in Table 10, and details of the derivation have been
presented in another document.13 The values of r and H are obtained from
operating data and supporting kinetic data; a is a shape parameter describing
the SO2 generation pattern. Given values of the parameters, ot, r, and H,
Equation 27 is solved iteratively for SO2 removal, R. Equation 28 then
yields the particle consumption C.
Figure 19, previously published in ANL/ES/CEN-1002, is printed here to
show the relationship of a to the generation pattern. At higher values of
ct, relatively more SO2 is generated near the bed entrance. The equation
which generates the curves of Fig. 19 is
-ax/h
exs
Relative local rate of SC^ generation =
Postive values of a imply that most SC^ is generated near the entrance;
negative values imply that most SO2 is generated near the exit. Uniform
SC>2 generation corresponds to a = 0; generation of all SC^ at the bed
entrance corresponds to a + ®, Figure 19 illustrates the diversity of
generation patterns which can be obtained simply by changing the single
parameter a; it is this factor that leads to selection of this particular
family of generation patterns for inclusion in the model.
To calculate H, it is necessary to know K, the reaction velocity
constant for fresh additive particles. Typical laboratory kinetic data
might be in the form of Fig. 20.6 ^ batch of additive CaO is exposed to
SO2 at a known concentration cQ and at the temperature and pressure
anticipated during combustion conditions. The moles of SO2 absorbed per
pound of CaO are plotted against time. The parameter K is obtained from
the initial slope of the curve.
_ Initial Slope
-------
79
TABLE 10. Definition of Symbols
b, .b„,b_,b. ,bc = model fitting constants
1 2 3 4 5
C B particle consumption, y/yg
c = exit concentration of SC^, moles/volume
c = concentration of SO., used in laboratory test
F = total gas flow, volume/time
,f = ratio of .actual to ideal absorption capacity, y&lyQ
H = bed inventory parameter, KW/F
h = fully expanded bed height
K = value of reaction velocity constant for fresh additive,
volume/(time)(weight of additive)
M = Ca/S mole ratio fed to fluidized bed
R = fractional removal of SC^, 1 - Fc/S
r = effective feed ratio of additive to SO^, wyg/s
S = total rate of generation of SC^, moles/time
T = operating temperature of fluidized bed, °F
W = total weight (reacted + unreacted) of additive in
fluidized bed
w = feed rate of fresh additive, mass/time. Also equals
removal rate of exhausted additive.
x = distance above bed entrance.
y = average S09 absorbed by additive particles in bed,
moles SC^/mass additive
y = ultimate S09 absorption capability of additive
particles, moles SC^/mass additive
y = stoichiometric SO2 absorption capability of additive
particles, moles SC^/mass additive
z = x/h
a = shape parameter describing the S0? generation pattern;
higher values of a correspond to relatively more SC>2
generation near the bed entrance.
3 = fraction of total gas feed going to bubble phase
-------
80
Fig. 19. Normalized Generation Rate
of SCL as a Function of
Fractional Bed Height
0 0.1 0.2 0.3 0 4 0.5 0.6 0.7 0.8 0.9 1.0
-.FRACTIONAL DISTANCE FROM ENTRANCE
Fig. 20. Typical Laboratory
Kinetic Data and
Interpretation
/ S\ Initial Slope = Kc
time
-------
In addition, y£ is obtained as the asymptotic value of y, as shown.
Example:
To illustrate the model, some typical calculations will be made.
Consider the following operating data for a fluidized-bed combustor:
w = 0.8 lb CaO/hr
W = 16 lb CaO
S = 0.006 mole S02/hr
*F = 1000 cfh
and the following kinetic data:
Initial Slope = 0.04 mole S02/(lb Ca0)(hr)
y = 0.013 mole S02/lb CaO
c = 4 x 10"^ mole S02/ft
o
The calculations then proceed as follows:
^ = (0.8) (0.013) =
S (0.006)
3
Initial Slope _ 0.04 _ .4 ft
~ 4 X 10~^ (lb CaO)(hr,
a.Si. „ 160
F 103
These values are used in Eqs. 27 and 28 to compute predicted values of
R and C. With an H of 100 and values of a of -10, -5, -2, 0, 2, and 5, SO2
removals are 88, 93, 97, 99, 99, and ^100%. Particle consumptions may be
calculated by dividing the removals by 1.7, the value of r.
Note from this example that r is the effective feed ratio of additive
to SO2. Thus, based on the stoichiometry of one mole of SO2 to one mole of
CaO, the theoretical ultimate value of y is 0.018, and the theoretical
stoichiometric feed rate is (0.8)(0.018)/(0.006) = 2.4. This figure is
reduced to an effective value of 1.7 because CaO does not absorb the
stoichiometric quantity of SO2.
C. Effects of Parameters on S0„, Removal arid Particle Cons uTiD t ion
Figures 21-23 present illustrative values, calculated as described in
the preceding section, of S0? removal and particle consumption for various
values of r, a, H. In each "Figure, the normalized bed inventory variable H
is constant. Removal, R, and consumption, C, are plotted against additive-
to-S02 feed ratio r at various values of the S0? generation pattern shape
parameter a.
-------
82
WAJIMUM pObSiBlF NAV-WuM POSr.lHl.L
PLM^VAL \ ! rOfr.UVPtlO'J 0_i®
MAXIMUM POSSIBLE
SO? REMOVAL
MAXIMUM POSSIBLE —
\V, WRTICLE CONSUMPTION
\\\
\i
v&
005
C. PARTICLE
CONSUMPTION
r,so2 removal
Fig. 21. Calculated SO2 Removal and Particle Consumption, H = 10
J J 1 11IIX
\*
MAXIMUM POSSIBLE
SO? REMOVAL
\\V MAXIMUM POSSIBLE
\\
I
\\v PARTICLE CONSUMPTION
ml 1 < 1 1 1111I 1 1 >1 mil 1 1 rTTrrr
Fig. 22. Calculated SC^ Removal and Particle ' Consumption, H = 30
-------
MAXIMUM POSSIBLE
PARTICLE CONSUMPTION
MAXIMUM POSSIBLE-
SO^REMOVAL
1.0
0.01
0.1
I o
10
100
Fig. 23. Calculated 'SO^ "Removal 'ancT'T'art'i'cle 'Ccnsump'ti'cn, -H- = 100
As r is increased beyond approximately 10, removal does not significantly
increase with r in a given graph. This is because the bed reactivity H is
constant. At r = 10, the bed essentially contains only highly reactive,
unconsumed particles. This represents the maximum possible bed reactivity
and further increases in r would not be expected to improve S0£ removal.
Obviously, consumption is so low at these higher values of r that even the
asymptotic value cannot be economically attained.
For an additive-poor bed, H = 10, the effect of SO2 generation pattern
is quite strong. However, as the bed inventory is increased, a becomes less
important.
Also sham' in each of Figs. 21-23 are two curves, one representing
maximum possible values of particle consumption and the other maximum
possible SO7 removals. Since r is the effective additive to SO2 feed ratio,
maximum possible removal occurs at complete consumption (C = 1) if the value
of r is less than unity, and maximum possible removal is unity if r exceeds
unity. Similarly, maximum possible consumption occurs at complete removal
(R = 1) for values of r greater than unity, and is unity for values of r
less than unicy. Thus,
R
r and C = 1 if r < 1
max
max
R
1 and C = — if r.> 1
max
max r
-------
The'curves in Figs. 21-23 show how closely ideal conditions are approached
at different values of a.'
Figure 24 shows the effect of bed inventory on S0o removal and particle
consumption. The additivc-to-SO^ feed ratio r is held~fixcd at unity, and
R and C (which are equal when r = 1) are plotted against II for various values
of a. Clearly, additive-rich beds are desirable and a bed as rich as H = 1000
(or an equivalent series of smaller beds) should give virtually complete
SC>2 removal and particle consumption, regardless of the SO2 generation
pattern.
This analysis shews why knowledge of the kinetic parameters K and yg
is needed for rational design. Reasonable SO2 removals apparently require
operations with r > 1 and H > 100. Values of r and H depend not only on
the corresponding operating variables—additive feed rate and auditive
inventory—but also on the kinetic parameters.
0.9
08
07
0.6
R or
05
OA
OS—
02
0 I
100
1000
10
H—»
Fig. 24. Calculated SC^ Removal and Particle Consumption, r = 1
D. Comparison of Model with Data
Further attention will be restricted to two specific SO2 generation
patterns. If all SO2 is generated at the entrance to the bed, a -*¦ 00 and
Eq. 27 reduces to _
„ . ! - e-11'1 " 7' (29)
-------
85
If the SO2 is generated uniformly throughout the bed, ct = 0 and Eq. 27 reduces
to n
. -H(l - ~)
R = 1 - r—(30)
H(1 - f
The parameters H and r are related to the physical parameters by
H = ^ (31)
F
wy
r =
e
(32)
However, if M is the Ca/S mole ratio fed to the column, then r can also be
expressed as
r = Mf (33)
where the fraction f is defined by
y
f = (34)
ys
and is the ultimate S09 absorption capability of additive particles relative
~to -the i-deal'-stoichi'omet-ric -absorption ¦ cap-abi-1-ity of-• the- addifci ve particles.
The reaction velocity constant K.' and the fraction f may depend on the
temperature T at which the bed is operated. Two forms of this dependence
were chosen for investigation:
K. = bl + b2 (T + 460) + b3 (T + 460)2 (35)
and
f = b4
K = bx + b2 (T + 460)
f = b4 + b5 (T + 460)
(36)
Thus, four models were investigated for fitting the experimental data.
Model 1 — Eqs. 29 and 35
Model 2 — Eqs. 29 and 36
Model 3 — Eqs. 30 and 35
Model 4 — Eqs. 30 and 36
A computer program was devised to find values of the fitted constants b^,
b7, bp b^, b,- such that best (in the least-squares sense) agreement was
oG taLned between values of R calculated from the model and those observed
experimentally. A total of 45 runs were used as data to be fitted to the
model. The ranges of experimental values covered were:
-------
86
T 1325 to 1800°F
W 18 lb
F 458 to 1412 cfh
M 1.0 to 5.5
The results (Table 11) show that the standard error for each model is
similar, in the range 10-11%. This standard error appears to be reasonable
on the basis of inspection which showed some lack of reproducibility in the
data. Thus, the fitting results do not allow determination of which SO2
generation pattern is more realistic.
TABLE 11. Values of Fitced Constants Obtained by Model Fitting
Standard
Error of
Model b, b, b, b. bc Prediction
1 2 J 4 5
1 -5153.4870 5.3025818 -0.00133601 0.49891591 - 10.3
2 -532.00413 0.32007391 - 0.34173206 -0.001462656 11.3
3 -13830.389 14.32897 -0.00362929 0.49500000 - 10.4
4 -1964.5366 1.1602147 - 3.0735411 -0.0013070000 11.3
-------
87
VI. MISCELLANEOUS
A. Control of Chloride Emissions
Most U.S. coals contain a small amount of chloride that is emitted as
hydrogen chloride when the coal is burned. Since lime is capable of reacting
readily with hydrogen chloride, it was decided to examine, theoretically,
the potential of several oxides and hydroxides for control of chloride
emissions at the concentrations encountered in the fluidized-bed combustion
process. This information might also be of value for application of fluidized-
bed combustion to the incineration of chloride-bearing wastes such as poly-
vinyl chloride plastics.
The capability of a lime bed for removing hydrogen chloride from
combustion gases would depend on the thermodynamic equilibrium of the
reversible reaction
CaO + 2 HC1 £ CaCl2 + H20
Calculated values of the equilibrium partial pressure of HC1 with several
oxides and hydroxides as a function of temperature at a partial pressure of
water of 0.1 atm (approximately that in flue gas) are shown in Fig. 25. At
a combustion temperature of 1550°F (1090°K), the equilibrium partial pressure
of HC1 with CaO is approximately 10atm or 1000 ppm. Since HC1 concen-
trations in flue gases from the combustion of coal are far lower than 1000
ppm, no removal of....h,ydr.o,gen ..chloride ,by «lime.»would..be ^expected..
Consideration of the calculated equilibrium data for the reaction of
HC1 with other possible additives (MgO, BaO, NaOH, and KOH) indicates that
of these materials, BaO and the alkali hydroxides are favorable for accepting
HC1. Possibly, the addition of small amounts of one of these agents would
remove HC1. Experimental data would be needed to verify the suitability
of such materials and to determine the quantities of additives needed for
HC1 emission control. No work in this area is planned until other more
pressing work is completed.
-------
88
-10
n
u
X
Reactions
a.
MO + 2 HCl(g) - MCI +
MOH + HCl(g) - MCI + H,
10
400
500
600
1100 1000 900 800
700
Temp, °K
Fig. 25. Calculated Equilibrium Partial Pressure of HC1
as a Function of Temperature (P = 0.1 atm)
2
-------
89
VII. FUTURE WORK
The future experimental program at ANL will encompass the following:
1. Completion of study of the effects of variables, such as coal
particle size, for one-stage .combustion at a pressuTe of 1 atm. Additional
experiments at 1-atm pressure will be required to aid in establishing the
mechanisms involved in pollution control. Studies having a lcwer priority
include (a) measurement of SC^ and NO levels during the combustion of oil
in a fluidized bed under both oxidizing and reducing conditions and (b)
experiments to determine pollutant removal with coals and limestones not
yet tested.
2. Determination of the effect of independent variables on sulfur
retention, NO level, and particulate emission, using the atmospheric-pressure
bench-scale combustor in the two-stage mode of operation.
3. Construction of a multipurpose, bench-scale plant for operation
at up to 10 atm to study the effects of the most important variables in
a. one-stage, high-pressure combustion,
b. two-stage, high-pressure combustion,
c. regeneration of the partially sulfated
and sulfided additives from one-stage
and two-stage operation.
4. Continuation of'stu'die's to eluc'ldat'e "the "sech'an'rsms involved in
one-stage and two-stage combustion and the regeneration reactions.- In one-
stage combustion, a sulfate is produced that is regenerated to oxide with a
reducing gas; in two-stage combustion, a sulfide is produced that is regen-
erated by an oxidizing gas.
-------
VIII. ACKNOWLEDGMENTS
We gratefully acknowledge the help given by Dr. R. C. Vogel,
Mr. D. S. Webster and Dr. S. Lawroski in directing and reviewing the
program, the assistance and advice of Mr. K. Natesan (Materials Science
Division) in mounting the limestone particles and for performing some of
the microprobe examinations, the assistance of Mr. A. Sanders and
Mr. A. Martin in mounting samples and of Mr. C. Johnson, Mr. K. Anderson,
and Mr. C. Seils in performing microprobe examinations and of the analysis
team of Mr. M. Homa, Mrs. C. Blogg, Miss F. Ferry, Miss J. Williams, and
Mr. Z. Tomczuk directed by Dr. R. Larsen and Mr. E. Kucera, and of our
typist Miss P. Wood.
-------
91
APPENDIX A. MATERIALS
TABLE A.l. Size Distribution of Coaxs
Illinois. Seam 6, Peabody Coal Co.,
Mine 10, Christian County, Illinois
Welbeck. East Midlands Field (British National Coal Board)
Sieve Analysis (wt %)
As Charged After Feeding
to Hopper of Through
Screw Feeder Screw Feeder
U. S. Sieve No. Illinois Welbeck Illinois Welbeck
+14
-
-
N.A.b
-14 +25
19.4
11.3
14.0
-25 +35
13.5
13.9
15.4
-35 +45
16.9
23.1
21.0
-45 +80
18.9
27.8
22.0
-80 +170
13.2
9.9
11.1
-170 +325
7.6
7.0
7.4
-325
10.4
7.0
9.1
; Particle Diameter,
pmc
350
396
412
Analyses for Welbeck coal are for as-received material; the
Illinois coal was ground to pass a U.S. No. 14 sieve before
being fed.
^N.A. - not available,
c
Calculated by summing products of weight fractions and average
screen openings for the sieve size range.
-------
Chemical Characteristics of Coals
Seam 6, Mine 10, Peabody Coal Co.,
Christian County, Illinois
East Midlands Field (British National Coal Board)
Proximate Analysis (wt %)
As-Received Dry Basis
Illinois Welbeck Illinois Uelbeck
Moisture
10.12
5.08
-
-
Volatile Matter
37.90
31.12
42.17
32.79
Fixed Carbon
41.12
46.65
45.75
49.15
Ash
10.85
17.16
12.08
18.07
Sulfur
3.7
1.23
4.14
1.30
Heating Value, Btu/lb
10956
11206
12163
11807
Ultimate Analysis (wt %)
Illinois Welbeck
Carbon 68.45
Hydrogen 4.97
Sulfur 4.14
Nitrogen 1.18
£
Mineral Content of Ash (wt %)
Illinois Welbeck
P2°5
0.12
0.19
sio2
38.67
58.65
Fe2°3
22.49
7.48
A12°3
16.31
21.67
Ti02
0.80
1.08
CaO
8.48
2.56
MgO
0.88
1.40
so3
8.59
2.13
k2o
1.43
2.85
Na20
1.66
1.65
Undetermined
0.57
0.34
TABLE A. 2.
Illinois.
Welbeck.
67 .Cb
4.50
1 1Q
1.21
Average of all analyses.
-------
TABLE A.3. Size Distribution of Illinois Coal
and Coarse Limestone No. 1359 Used in Experiment AR-6
Limestone after
Illinois Limestone Charged to Passage through
U.S. Coal Screw-Feeder Hopper Screw Feeder
Mesh (wt %) (wt %) (wt %)
+7
0
0
0
»—1
+
r-*
1
35.6
7.3
-10 +14
25.3
18.1
-14 +25
27.3
27.5
37.2
-25 +35
11.3
5.9
10.1
-35 +45
13.2
3.3
8.4
-45 +80
16.7
1.5
6.5
-80
0.9
12.5
-80 +170
14.6
-170 +325
7.8
-325
9.0
Av. particle
dia, um
490
1637
1010
-------
TABLE A. 4. Chemical Characteristics and Particle Size Distribution
for Pittsburgh Coal
Sieve Analysis
Mineral. Content
of Ash (wt %)
U.S. Mesh
As Received
(wt %)
As Used in
Current Runs
(wt %)
P2°5
0.26
+4
33.4
0
Si02
48.51
-4 +14
41.0
0
Fe2°3
15.17
-14 +25
16.0
22.0
A1203
22.54
-25 +45
5.8
30.0
Ti02
1.01
-45 +80
1.4
21.1
CnU
3.93
-80 +170
0.9
13.1
MfjO
0.80
-170 +200|
4.0
)
1.5
S03
4.54
-200 J
9.8
K 0
1.73
0
CM
0.66
Undetermined
0.86
aAs received.
-------
95
TABLE A.5. Chemical Composition of Limestones
Limestone No. 1359, M. J. Grove Lime Co., Stephens City, Va.
British Limestone, J. Gregory & Son, Stok.e-on-K.ent, Staffordshire
Chemical Analysis (wt %)
British Limestone
Component Limestone No. 1359 ANL Analysis CRE Analysis
Ca
37.9
37.9
39.1
Mg
0.27
0.12
0.12
C0„
t.
NAb
49
42
H20
1.56..
0.8
NA
Si02
NA
NA
1.6
Derived
CaC03
94.80
94.6
96.5
MgC03
0.95
0.4
0.4
For as-received material.
^NA - not available.
-------
TABLE A.6. Size Distribution of Limestones
Limestone No. 1359, M. J. Grove Co., Stephens City, Va.
British Limestone, J. Gregory & Son, Stoke-on-Kent, Staffordshire
Sieve Analysis (wt %)
As Charged After Feeding
to Hopper of Through
Screw Feeder Screw Feeder
U.S. Sieve No.
1359
British3
1359
British
+12
-
-
-
-
-12 +14
-
2.1
-
-
-14 +25
39.3
29.8
33.1
22.6
-25 +35
14.0
13.7
14.2
12.8
-35 +45
13.8
13.7
14.6
13.6
-45 +80
12.5
13.3.
13.8
-14.2
-80 +170
¦7.9
9.7
9.0
11.6
-170 +325
3.5
5.7
4.2
9.0
-325
9.0
11.3
11.1
16.1
Average Particle
Dia, pm
609
555
555
440
For as-rece_Lvic r.at
erial.
Calculated by the procedure
described in
footnote
c, Table
-------
97
TABLE A. 7. Characteristics of Welbeck Coal Ash
Constituent
Mineral Analysis,
Ignited Basis
(wt.%)
U.S. Mesh
Sieve Analysis
(As Received)
(wt %)
P2°5
0.06
+25
24.5
Si02
69.74
-25 +35
20.2
Fe2°3
4.76
-35 +45
30.6
Ai203
20.34
-45 +80
22.1
Ti02
0.81
-80 +170
1.5
CaO
3.70
-170 +325
0.5
MgO
2.05
-325
0.5
S°3
2.44
K2°
2.98
Na20
1.73
Undetermined
0.39
-------
APPENDIX B. AUXILIARY EQUIPMENT AND INSTRUMENTATION
1. Heating-Cooling System
Two 2.7-kW, 230-V resistance heaters are the external souice of
heat for the lower portion of the 4-ft combustion section and a 2.0-kW,
230-V heater for the bubble-cap section. These heaters assist in bringing
the fluid bed to the coal ignition temperature of ^800°F and in maintaining
the heat balance in the reaction system. The locations of the heating
circuits and of the controlling and recording thermocouples on and in the
combustor are shown in Fig. B.l. Stainless steel-sheathed Chromel-Aluniel
thermocouples are used. The temperature indicator-controllers are Honeywell-
Brown Pyro Vanes with indicating-controlling ranges of either O-20OO°F or
0-2400°F. Thermocouple output is displayed on a Honeywell-Brown Electronik
12-point, 0-2400°F range recorder.
Removal of heat from the combustor is accomplished by passing air
or an air-water mist through the annular chambers surrounding the lower
portion of the combustor section. When it is necessary to remove only
relatively small quantities of heat, air alone is circulated. The flow of
air (metered with a rotameter) is adjusted to obtain the desired output of
a thermocouple immersed in the fluid bed. When a greater quantity of heat
must be removed, air-water mist is injected into the annular chambers. A
schematic of one of four systems for the injection of air-water mist is
shown in Fig. B.2. The other three systems are similar. The system consists
of a 5-gal distilled water reservoir, an air supply, and a temperature
indicator-controller (TIC) - thermocouple - solenoid valve combination.
The thermocouple indorsed in the fluiuizeu beu acts through the TIC to
energize a solenoid valve which, in turn, allows water to be injected into
a carrier air flow and hence to the annular chamber of the combustor. The
air-water mist leaving the combustor is vented through a heat-exchanger
to remove the water.
2. Solids Injection and Removal
Coal and limestone are transported to the fluidized-bed reactor
as a dilute phase in an air stream that is part of the total fluidizing-
air flow. The solids are delivered into the air stream by Vibra Screw
Line bin feeders. The standard units have solids bins of 3-cu ft capacity,
solids contact surfaces of Type 304 stainless steel, and mechanical variable-
speed transmissions with a 10 to 1 turndown ratio. They can utilize screws
of various diameters (in the range 1/4 to 5/8 in.), depending on the feed
rate requirements. Solids feed rates of 0.2 to 17 lb/hr are employed. To
maintain a continuous check of the weight of the contents of Vibra Screw
feeders, the feeders are mounted on Toledo "One Spot" platform scales, each
with a capacity of 500 lb; the smallest scale division on these scales is
1/2 lb.
The dilute-phase air-solids mixture i's carried to the cutr.L us tor
in 3/8-in. stainless steel lines. Coal and limestone are injected sc-parateiy
into the combustor through a fitting welded to the combustor at a point
just above the gas distributor plate.
-------
99
Flue Gas
To Cyclone
Separators
TR2
TI
TR ¦ Thermocouple Co
Recorder
TI e Thermocouple to
Indicator or
Indicator-controller
TR7
TIC s Temperature
Indicator Controller
voltage source
TRIO
TR9
TR5
TR8
Combuscion
Air —
In let
Fig. B.l. Heating and Temperature Sensing Arrangements for Combustor
-------
o
o
To Steam
Condenser & Drain
Vent to
Drain
i *1.
nc
0
¦0-
0
Temperature
Indicator-Controller
Pressure Gage I
fl
Regulator
Pop-Off Valve
Valve
Control Valve
Rotameter
Ek
Water Reservoir
TL
Air Supply —£3-
K
Wa ter
Supply
Fig. B.2. Combustor Cooling Circuit
-------
101
Solids leave the combustor (1) as bed samples, (2) as excess bed
material, and (3) as flyash-limestone fines. Samples are removed from the
combustor bed through 1/2-in. stainless steel pipes welded into the combustor
wall-at points about 2 and 20 in. above the gas distributor plate. The
solids sampling system consists of a manifold that provides access to a
vacuum pump, high-pressure (35-psig) air, and the sampling ports on the
combustor. Solids are drawn into an evacuated sampling vessel by opening
a valve connecting the vessel to a combustor port for a sufficient time to
withdraw ^50-100 g of the bed material. The sample is then allowed to
cool, the sampling vessel is brought to atmospheric pressure, and a bottom
valve on the sampling vessel is opened to allow the bed sample to flow out.
Details on procedures for elemental analysis and coal analysis are presented
in Table B.l.
TABLE B.l. Analytical Methods and Procedures
A. Methods for Elemental Analysis
1) Calcium. EDTA volumetric titration.
2) Sulfur. Leco combustion method, employing automatic volumetric
titration of an,IO3- -I-starch system.
3) Carbon. Leco combustion method, using a thermal-conductivity
detector and employing the 0^/002 relationship.
4) Carbonate. Measurement, with a sensitive gas chromatograph,
of the CO2 evolved by acidifying the material.
5) Sulfite, S0^ . Turbidimetric method. The sulfur value released
by acid treatment is converted to sulfate and titrated with a
barium solution.
6) Sulfide, S . Turbidimetric method. A sample is treated with HC1;
the evolved gases are collected in an inert atmosphere and then
treated with an acidic acid-bismuth nitrate solution. Bismuth
sulfide precipitates and is measured.
7) Combined Nitrogen. Micro-Kjeldahl method. The nitrogen is released
as ammonia, absorbed, and titrated.
B. Methods for Analysis of Coal and Coal Ash
Analyses of coal and coal ash are performed by a commercial laboratory
(Commercial Testing and Engineering Co., South Holland, 111.) employing
ASTM designation D271-64 methods. The analyses made are as follows:
proximate, ultimate, mineral (ash), fusibility of ash, free swelling index,
and calorific value.
-------
102
Bed material is removed intermittently during a run to maintain
a constant bed height; after a run, the entire bed is removed. ^Bed material
is withdrawn through a 1-in. pipe. The top of the pipe is at the face of the
gas distributor, and the pipe passes through the plenum chamber and its
bottom plate. When a plug valve at the bottom of the pipe is opened, bed
material flows out of the reactor.
Most of the fine flyash and limestone carried in the flue gas
stream are removed by two cyclone separators in series. Essentially all
of the remaining fines are removed with a glass wool final filter.
3. Air Preheater
The preheater is constructed of a 7-ft length of 3-in. dia schedule
AO, type 304 stainless steel pipe. Figure B.3 is a schematic of the unit.
The preheater is packed with 6.5-ft lengths of 3/8-in. dia type 304 stainless
steel tubing to increase the heat exchange capacity.
Heating is provided by four Hevi-Duty clam-shell"furnaces which
cover the entire length of the preheater. The top furnace is rated at
1.6 kW and the other three are each rated at 3.4 kW; all are operated at a
230-V input. Temperature control is accomplished with thermocouple,
temperature-indicator-controller (TIC) circuits as shown in Fig. B.3.. These
units are identical to those used for the combustor.
4. Fluidizing-Combustion Air Supply
Fluidizing-combustion air is taken from the laboratory high-
pressure (35 psig) air supply. After passing through a pressure reducer
where its pressure is reduced to 15 psig, the filtered air is metered
through a flow control system to the preheater. Figure B.4 is a diagram of
the flow control system. This system consists of a Taylor pressure trans-
mitter, a flow recording and control device, and two calibrated orifice-
flow control valve arrangements. The ranges of the two orifice-valve
combinations are 0-15 and 0-32 scfm.
An example of air flow rates in a run is as follows: To achieve
superficial gas velocity of 3 ft/sec in the combustor at 1472°F, a total
flow rate of 10 scfm at 70°F and 1 atm is needed. This total flow is divided
as follows: 6.5 scfm through the orifice-valve arrangement to the preheater,
1.5 scfm through each of the solids feeders, and 0.5 scfm through an
auxiliary feed port.
5. Flue Gas Filtration
The approximately 10 scfm (70°F, 1 atm) of flue gas leaving the
combustor is passed through a particle-removal system before it is routed
through the building exhaust system. The particle-removal equipment
comprises two cyclone separators and a glass wool final filter.
-------
103
Air
1.6 kW
230 V
TIC
..J
\_L
TIC
230 V
TR1
3.4 kVJ
230 V
230 V
To combustor
= Pressure
TIC ® Temperature Indicator-Controller
v = Power Source
TR = Thermocouple to Recorder
Thermocouple to TIC
Fig. B.3. Preheater
The cyclones were manufactured by Universal Oil Products (UOP)
Air Correction Division and are designated sizes 2 and 3 (see UOP Tech.
Bull. 604A). The calculated flow handling capacities for a gas at 500°F
and a ZiP of 6-in. water gauge are 49 and 111 cfm for size 2 and size 3
cyclones, respectively.
The final process filter is a 2-sq ft, 1-in.-thick glass fiber
commercial unit that constitutes one side of a filter box. The flue gas is
passed through this filter and into the room exhaust system. In the latter
system, the gas (after a manyfold dilution) passes through a set of high-
efficiency filters before it is exhausted to the atmosphere.
-------
I—1
o
->
Building High-Pressure
Air
FRC-1
Flow recording controller
PT-1
Pressure transmitter
Valve
To
-O* Preheater
Regulator
Pressure Gauge
Flow Control Valve
||| . Taylor Calibrated Orifice Assembly
(7) Range, 0-32 scfm
© Range, 0-15 scfm
Fig. B.4. Fluidizing-Combustion Air Flow Control System
-------
105
6. Gas Sampling and Analysis System
A sample of approximately one-twentieth of the total flue gas
(0.5 scfm) is taken at a point midway between the primary and secondary
cyclones. This stream is pumped through a 12-in. sintered-nickel bayonet
filter to remove solids. The gas is then passed through a water condenser
and a refrigerator to remove water. Any residual water, which could inter-
fere with the measurement of flue-gas constituents by an infrared technique,
is removed by magnesium perchlorate. (About 3000 ppm moisture in the gas
gives a reading corresponding to ^30 ppm SO2 and ^20 ppm NO. Moisture does
not affect oxygen analysis.)
Continuous analyses .for NO..and S0? are carried out using Beckrnan
315A infrared analyzers, and continuous analyses for Ctt^ and CO with Mine
Safety Appliance (MSA) MRA infrared analyzers. Continuous measurement of
oxygen is accomplished with a Hays paramagnetic oxygen analyzer. Intermittent
analyses for CO2 are performed by gas chromatography. The instrument supply
manifold is maintained at a constant pressure by a pressure-control device.
Aliquot samples are removed from the manifold to supply each instrument with
the required gas flow. Prior to and during an experiment, the response of
each analyzer to nitrogen gas and to a mixture of the analyzed component
in nitrogen are determined and the instrument zero and span are adjusted,
if necessary. The analytical equipment components are listed in Table B.2.
The temperature of the gas is determined at several parts of the sampling
system, and values of the gas-residence time are calculated. It was estimated
that the total residence time of flue gas in the sampling system .is abouc
'1? "rC'0 n ^
7. Data Logger
The combustor pilot plant has recently been equipped with a
Honeywell data logger capable of handling up to 200 variables: 175 thermo-
couple signals, 19 pneumatic signals, and 6 millivolt signals. During a
log cycle, a record of the 200 variables (in groups of 25) with time and
calibration data (system check) is recordid by a Flexowriter typewriter.
Simultaneously, the same information is stored on punched tape. Information
ultimately is stored on data cells for use in the IBM 360-50/75 computer
and Calcomp plotter systems.
The heart of the logger is the digitizing servo, which is connected
sequentially to each of the 200 channels to receive the input voltage,
measure it, and establish an analog shaft rotation position. This shaft
position is encoded in cyclic binary form by a Giannini encoder for later
translation into the decimal form used by the typewriter.
-------
o
o
TABLE B.2. Flue Gas Analytical Equipment
Flue Ga9
Cons tituent
Method of
Analynis
Instrument Model
Output Displayed On
Range
Accuracy
(% of range)
Sulfur Dioxide
Infrared
Beckman 315A
Bristol, two-pen,
variable-range recorder
0-5000 ppm
0-10,000 ppm
+1%
Nitric Oxide
Infrared
Beckman 315A
Bristol, two-pen, 0-500
variable-range recorder 0-1000 ppm
+1%
Oxygen
Paramagnetism
Hays 632
Honeywell-Electronik 16, 0-1 vol %
2-pen, 0-10 raV range 0-10 vol %
recorder
Methane
Infrared
MSA LIRA 200
Honeywell-Electronik. 16,
2-pen, 0-10 mV range
recorder
0-1000 ppm
+1%
Carbon Monoxide Infrared
MSA LIRA 300
Honeywell-Brown Electronik 0-5000 ppm
variable range
+1%
Carbon Dioxide Gas Chromatography Hewlett Packard 700 Honeywell-Electronik 16, 0-20 vol 7. +5X
variable chart speed,
reversed-drive recorder
-------
107
APPENDIX C. MATERIAL BALANCES
Material balances were calculated from the sulfur, carbon, and calcium
contents of (1) the input streams (coal, limestone additive, and initial
fluid bed) and (2) the output streams (cyclone solids, final bed, bed
samples, and SO2, CC>2, CO, and CH^ contents of the flue gas).
The weights of material in the solids input streams (coal and limestone
additive) were obtained from feed hopper scale readings taken during the
runs. The accuracy of these weights was checked by direct weighing of the
quantities of coal and limestone charged to the solids feeders and of the
quantities remaining in the solids feeders at the end of an experiment.
The .quantities ..of ~s.ulf.ur., carbon., and calcium in these input streams were
calculated from available analytical data. For the solids output streams
(material on the cyclones, bed samples, and final bed), the quantity
accumulated or removed was weighed, then sampled and analyzed for sulfur,
carbon, and calcium.
The quantities of sulfur and carbon in the flue gas were estimated
from data supplied by continuous monitoring for SO2, CO, and CH^ by IR
analyzers and from periodic analyses for CO2 by gas chromatography. The
flue gas composition data were integrated to determine the total quantity
of either sulfur or carbon exhausted in the flue gas over the full course
of an experiment.
Material balances for sulfur, carbon, and calcium in Amer-1, -3, anu -4
are Jgiven in Table C.l and those for'Brit-1 to -3, Brit-Amer, and Amer-Brit
in Table C.2. The results of the material balance calculations for the
five final BC-series experiments are listed in Table C.3 and those for the
AR-series experiments in Table C.4.
-------
TABLE C.l. Material Balance for Sulfur, Carbon, and Calcium,
Experiments Amer-1, -3, and -4 (weights in pounds)
Weight
Amer-1
Ca Height
Amer-3
Ca
Weight
Ame r-4
Ca
Material In
Coal
326.0
12.13
200.56
2.15
108.5
4.04
72.17
0.72
71.2
2.65
43.81
0.47
Additive
59.1
-
6. 74
22.40
34.5
-
3.93
13.08
13.7
-
1.56
5.19
Bed
16.9
1. 79
0.17
6.52
16.1
1.80
0.16
5.59
14.3
1.53
0.14
5.62
Z
-
13.92
207.5
31.1
-
5.84
76.3
19.4
-
4.2
45.5
11. 3 '
Material Out
Flue Gas
-
6.36
195.45
-
--
1.23
63.96
-
-
1.09
40.28
-
Primary Cyclone
58.9b
2. 78
12. 30
7.68
16,8
0.55
2.74
3.40
9.83
0.24
2.41
1.26
Secondary Cyclone
-
0.07
0.88
0.26
2.9
0.07
0.64
0.44
1.09
0.02
0.29
0.15
Final Filter
c
-
-
-
c
-
-
-
1.46
-
-
-
Fluidized Bed
52.3
7.80
0.31
17.74
37.0
4.54
0. 36
18.13
21.7
2. 35
0. 21
7.65
Z
-
17.01
208.94
25.68
-
6.39
67.70
21.97
-
3.7
43.2
9 .1
% Balance
_
122
101
83
-
109
89
113
88
95
82
All
experiments,
balance =
Total in x
100
113
97
92
Total out
Weight, for example, of coal fed in specified run(s).
'This is the total weight of the material removed from the primary and secondary cyclones in this run.
"Not available.
-------
TABLE C.2. Material Balance for Sulfur, Carbon, and Calcium,
Experiments Brit-J to -3, Brit-Amer, and Amer-Brit (weights in pounds)
Brit-1
Brit-2
Brit-3
Brit-Amer
Amer-Brit
a r r 1 1 In
Con 1
AddiL ive
Bod
Materia] Out
We i eht
141. 3
10.9
18.0
Ca
1.74 65.89 0.44
1.34 4.20
0.18
1.92 67.23
0.44
5.08
Weight
60.0
10.2
16.1
0. 74
0.47
1.21
Ca
27.99 0.19
1.25 3.93
0.16 1.50
29.40 5.62
Weight S
Ca Vleiftht
96.6
4.6
1.19 45.06 0.30
0.57 1.77
1.19 45.63 2.07
30.9
1.5
Fiue Gas
-
0.69
75.73
-
-
0.15
31. 77
-
-
0.56
51.46
-
-
Prtcnry Cyclone
29. 3
0.15
10.00
1.61
12.9
0.09
4.01
1.10
18.2
0.13
8.54
0.86
7.4
Secondary Cyclone
2.6
0.02
1.28
0.11
1.1
0.01
0.46
0.07'
1.6
0.02
1.08
0.07
0.7
Final Filter
4.5
0.18
0.62
0.20
2.1
0.02
0.24
0.07
1.7
0.02
0.53
0.07
d
Fluidized Bed°
34.4
1.02
0.43
2.28
7.4
0.29
0.08
1.25
11.4
0.59
0.13
2.40
2.6
2.06
88.06
4.20
-
0.56
36.56
2.49
-
1. 32
61.74
3.40
-
J Balance
10 7
131
83
height, for example, of coal fed in a specified run.
^Sunrr.ation of Brit-2, -3, Brit-Amer, and Amer-Brit. Material balance is for total,
included in Brit-3.
Combined with Brit-3.
g
Includes the fluid bed In experiments Brit-1 and Amer-Brit; includes only the fluid-bed samples and the excess
bed in other experiments.
0.38
14.41
0.17
Ca
0.10
0.56
0.38 14.58 0.66
All Experiments. Balance
WeiRhta
S
C
Ca
54.7
2.05
33.96
0.36
11.6
-
1.43
*.49
2.05
35.39
4.B5
Ib
4.83
125.00
13.20
_
0.65
30.32
-
10.2
0.31
1.74
2.05
1.1
0.02
0.22
0.13
0.5
0.02
0.07
0.04
21.5
1.66
0.20
4.54
-
2.66
32.55
6.76
4.54
130.85
12.65
94
105
96
98
114
92
O
VC
-------
TABLE C.3. Material Balance for Sulfur, Carbon, Calcium, and Magnesium,
Experiments BC-6 to -10 (weights in pounds)
BC-6
BC-7
BC-8
BC-9
BC-10
S
C
Ca
S
C
Ca
S
C
Ca
S
C
Ca
S
C
Ca
MR
Material In
Coal
4.23
60.47
0.47
5.30
74.70
0.58
6.41
90.29
0.70
5.59
78.83
0.61
5. 70
80. 37
0.62
-
Additive3
-
0.67
2.23
-
4.08
13.40
-
4.54
14.89
-
4.45
11.10
-
9.20
15.08
9.41
Bedb
1.50c
-
11.43
3.42
1.65
8.43
2.02
1.14
7.02
1.47
0.79
7.33
L. 99
1.39
6.40
_
I
5.7
61.1
14.1
8.7
80.4
22.4
8.4
95.9
22.6
7.1
84.1
19.0
7.7
91.0
22.1
9.4
Material Out
Flue Gas
0. 81d
56.21e
-
0.99d
64.04®
-
0.96d
82.97*
i
1.63d
76.67e
-
0.78d
74.89e
-
Primary Cyclone
0.42
3.29
1.64
0.87
7.04
2.81
0.49
6.31
1.63
0.94
4.24
6.61
2.22
4.94
13.35
7.57
Secondary Cyclone
0.05
0.45
0.21
0.44
5.88
0.22
0.02
6.37
0.12
0.09
0.61
0.65
0.04
0.25
0.03
0.09
Final Filter
0.09
0. L2
0.11
0.06
0.13
0.09
0.03
0.08
0.07
0.11
0.16
0.17
0.11
0.17
0.14
0.08
Fluidized Bed
Samples
3.83
21
9.45
4.73
2.77
15.79
4.08
2.42
17.56
3.43
2.36
9.47
3.30
0.38
7 .16
1.55
£
5.2
61 . 8
11.4
7.1
79.9
18.9
5.6
92.2
19.4
6.2
84.0
16.9
6.4
80.6
20.7
9.3
% Balance
91
.101
81
81
99
84
66
96
86
88
100
89
84
89
94
99
All Experiments, Balance
83
97
87
c
This quantity of sulfur was added as S0_ from a gas cylinder during initial bed sulfation.
Calculated from time-averaged concentration of SC^ in the flue gas.
a
'Calculated from time-averaged concentrations of CC>2, CO, and CH^ in the flue gas.
Limestone No. 1359 (97.8% CaCO~, 1.3% MgCOj) used in experiments BC-6, -7, and -8; Limestone No. 1360 (69.8% CaCO^,
19.2% MgC03) used in BC-9; Dol.omite No. 1337 (53.4% CaC03> 46.. 5% MgCO-j) used in BC-10.
'New starter bed of limestone No. 1359 was sulfated in experiment BC-6. This bed or part of it was then carried over
in each following experiment.
-------
TABLE C.4. Material Balances for Sulfur, Carbon, and Calcium,
Experiments AR-1, -2, -4, -5, and -6 (weights in pounds)
AR-1
AR-2
AR-4
AR-5
AR-6
S
C
Ca
S
_C_
Ca
_S_
C
Ca
S
C
Ca
C
Ca
Material In
Coa 1
5.48
77.28
0.60
5.61
79.02
0.61
8.67
126.85
1.36
11.06
182.96
1.96
16.67
275.60
2.95
Additive
-
5.12
16.81
-
5.59
18.38
-
4.86
15.95
-
17.68
58.10
-
27.66
90.87
Bed
2.27
0.06
4.66
1.78
1.03
6.32
-
3.14
10. 32
1.57
0.19
7.54
1.73
0.36
16.5 7
Z
7.8
82.5
22.1
7.4
85.3
24.9
8.7
134.9
27.6
12.6
200.8
67.6
18.4
30 3.6
110.4
Material Out
Flue Gas
1. 72
65.58
-
0.42
76.62
-
1.04
119.64
-
0.63
198.02
-
2.09
262.29
-
Primary Cyclone
1. 73
5.75
7.57
0.64
3.99
3.70
0.67
8.36
4.65
0.58
3.99
3.58
3.44
29.12
27.67
Secondary Cyclone
0.04
0.53
0.16
0.03
0.56
0.15
0.04
0.83
0.25
0.04
0.60
0.20
0.20
1.17
0.93
Final FiIter
0.09
0.18
0.14
0.09
0.23
0.21
0.11
0.71
0.18
0.02
0.24
0.12
0.08
0.33
0.22
Fluidized Bed
3.84
1.96
14.10
4.99
1.64
16.40
.4.84
0.48
17.09
8.76
1.05
42.17
9.03
1.64
67.23
and Samples
Z
7.4
74.0
22.0
6.2
83.6
20.5
6.7
129.5
22.2
10.0
203.9
46.1
14.8
294.5
96.1
X Balance
96
90
100
83
97
82
77
96
80
79
102
68
81
97
87
All
Experiments, Balance
82
97
82
-------
112
REFERENCES
1. Fifth Monthly Progress Latter on Reducing Emission of Sulphur, Nitrogen
Oxides and Particulates by Using Fluidised Combustion of Coal, October
19 7 J , DHB-101170, Prepared for Environmental Protection AgencyOffice
of Air Programs by National Coal Board, London.
2. C. W. Zielke, H. E. Lebowitz, R. T. Struck, and E. Gorin, J. Air
Pollution Control Assoc. _20_(3), 164-169 (1970).
3. Pope, Evans, and Robbins, Progress Report No. 22 (August 1969).
4. First three monthly reports on Reducing Emission of Sulphur, Nitrogen
Oxides and Particulates by Using Fluidised Combustion of Coal, August
1970, DHB-200870, Prepared for Environmental Protection Agency, Office
of Air Programs by National Coal Board, London.
5. R. W. Coutant et al., Investigation of the Reactivity of Limestone and
Dolomite for Capturing SO^ from Flue Gas, Summary Report, Battelle
Memorial Institute, Contract No. PH86-67-115 (June 27, 1969).
6. R. H. Borgwardt, Kinetics of Reaction of SO2 with Calcined Limestone,
paper presented at U.S. Public Health Service Symposium on Limestone-
SC'2 Reaction Kinetics and Mechanism, NAPCA, Cincinnati, Ohio (1969).
7. J. D. Hatfield and Y. K. Kim, Limestone Calcination and Sulfation
Investigations, ibid.
8. R. W. Coutant et al., Investigation of the Reactivity of Limestone and
Dolomite for Capturing SO9 from Flue Gas, Summary Report, Battelle
Memorial Institute, p. 40 (Aug. 30, 1968).
9. G. H. McClellan, S. R. Hunter, and R. M. Scheib, X-ray and Electron
Microscope Studies of Calcined and Sulfated Limestones, The Reaction
Parameters of Lime, ASTM Special Technical Publication 472, Philadelphia
(June 1970).
10. G. Moss, The Fluidised Bed Desulphurising Gasifier, paper presented at
Session II-, Second International Conference on Fluidized-Bed Combustion,
Hueston Woods, Ohio, October 4-7, 1970.
11. R. W. Coutant, R. E, Barrett, and E. H. Laugher, SO^ Pickup by Limestone
and Dolomite Particles in Flue Gas, paper 69-WA/APC-l, presented at
ASME Winter Annual Meeting, Los Angeles, Calif., November 16-20, 1969.
12. A. M. Squires and R. A. Graff, Panel Bed Filters for Simultaneous
Removal of Flyash and SOp. IV. Reaction of SOq with Half-Calcined
Dolomite, paper presented at 63rd Annual Meeting of the Air Pollution
Control Assoc., St. Louis, Missouri, June 14-18, 19 70.
13. L. B. Koppel, Purdue University, private communication (November 1970).
------- |