-------
Table 12 (Cont)
Particulate Removal Systems
Regenerator cyclones (1)
Regenerator air filter (1)
Kiln cylinders 1210 Am3/min
Kiln air filter 140 Am3/min
Off-gas bag filter (1)
Transport Systems
Cooled stone elevator 3000 kg/hr
Processed stone elevator 6000 kg/hr
(1) As in direct disposal
(2) Startup only - cost not included in comparisons
59
-------
Table 13
EQUIPMENT LIST FOR BRIQUETTING OPTION FOR PROCESSING
SPENT CAFB REGENERATOR STONE
Reactors
Pressure Vessels
Spent stone cooler
Design
Other Vessels
Fly ash silos
Curing chamber
Curing pond
Fans and Blowers
Regenerator air blower
Off-gas booster fan
Heat Transfer Systems
Waste heat boiler
Steam drum
BFW heater
Holoflite cooler
None
(1)
2-5mDx20mH
Concrete, precast
Double cone bottom
Bulk density - 640 kg/m (40 Ib/cf)
Cast iron rotary outlet valves
130 tnLx2mWx2mH
5 Conveyor belts,
1 m wide x 126 m L
17 m x 90 m x 2 m
(1)
(1)
(1)
(1)
(1)
(1)
60
-------
Table 13 (Cont)
Particulate Removal Systems
Regenerator Cyclones
Off-gas Bag filter
Regenerator air filter
Transport Systems
Air supply packages
Cooled stone elevator
Screw conveyor
Belt conveyors
Materials Processing
Hammer mill
Briquetter
Blender
Scalping screen
Fines screen
Briquetter feed hopper
Flyash surge hopper
Flyash weigh
Flyash feed hopper
(1)
(1)
(1)
1 - 3270 kg mix/hr
1 - 490 kg fly ash/hr
2 - 2780 kg stone/hr
3270 kg scrap/hr
1 m x 100 m
1 m x 90 m
1 m x 20 m
3270 kg/hr
3270 kg/hr
3270 kg/hr
3270 kg/hr
3270 kg/hr
2.0 m D x 2.3 m H
1 hr surge capacity
2.0 m D x 2.3 m H
4 hr surge capacity
0.5 m D x 1.3 m H
15 minute weigh cycle
0.5 m D x 2.3 m H
2 dumps from weigh hopper
(1) As in Direct Disposal Option
61
-------
Table 14
EQUIPMENT COSTS FOR DIRECT DISPOSAL OPTIONS FOR
PROCESSING SPENT CAFB REGENERATOR STONE
Vessels
Spent stone cooler $ 41,310
Blowers and Fans
Regenerator air blower 52,120
Off-gas booster fan 57,420
Heat Transfer Systems
Waste heat boiler 26,040
Steam drum 11,040
BFW heater 13,000
Holoflite cooler 19,990
Particulate Removal Systems
Regenerator cyclones 41,670
Off-gas bag filter 11,000
Regenerator air filter 700
Subtotal $274,290
Transport and Systems
Air supply package 12,000
Materials Handling
Storage silos 237,090
Total $523,380
62
-------
Table 15
COMPARISON OF INVESTMENTS REQUIRED FOR DEADBURNING AND SINTERING
OPTIONS FOR PROCESSING SPENT CAFB REGENERATOR STONE
Dead-burning Sintering
Basic items as in direct disposal (1) $254,300 (2) $146,840 (3)
Reactors
Rotary kiln 426,000 724,000
Vessels
Fines surge pot — 13,260
Kiln cyclone seal pot 13,260
Blower and Fans
Kiln gas booster fan 2,960
Off-gas booster fan (4) 62,430
Heat Transfer Systems
Holoflite cooler 5,400 39 930
Waste heat boiler (4) 2? 550
Air fin cooler 15,970
BFW heater (4) 14,040
Fuel oil system 15,560 26,360
Par calculate Removal Systems
Kiln cyclones 2,800 4,030
Kiln gas bag filter 2,060
Cooler cyclone 6,100 3,550
Off-gas bag filter (4) 11,900
63
-------
Table 15 (Cont)
Transport Systems
Cooled stone elevators
Materials Handling
Storage silos
Subtotal
Total equipment
(1) Common Items with Direct Disposal
Spent Stone Cooler
Regenerator air blower
Off-gas booster fan
Waste heat boiler
Steam drum
BFW heater
Regenerator cyclones
Holoflite cooler
Off-gas bag filter
Regenerator air filter
(2) 9 items
(3) 5 items
(4) Included in basic items
Dead-burning
$ 3,200
237,090
730,400
$984,700
41,310
52,120
57,420
26,040
11,040
13,000
41,670
11,000
700
$254,300
Sintering
2,600
205,330
1,134,980
$1,281,820
41,310
52,120
11,040
41,670
700
$146,840
64
-------
Table 16
COMPARISON OF INVESTMENTS REQUIRED FOR DRY SULFATION
OF SPENT CAFB REGENERATOR STONE
Dry Sulfation Option
Basic Items, as in Direct Disposal (1)
Reactors
Absorber
Rotary kiln
Vessels
Spent stone cooler
Fines surge pot
Pulverizer feed pot
Cyclone seal pot
Blowers and Fans
Regenerator air blower
Off-gas booster fan
Absorber air blower
Pulverizer air blower
Heat Transfer Systems
Waste heat boiler
Steam drum
BFW heater
Fuel oil system
Holoflite cooler
Absorber
463,520
Kiln
$ 700 (2) $ 155,570 (3)
3,200,000
46,360
13,260
13,260
13,260
62,290
15,750
26,840
29,950
45,060
22,080
13,520
10,000
69,570
—
—
(4)
15,360
—
—
(4)
(4)
(4)
10,000
45,800
65
-------
Table 16 (Cont)
Particulate Removal Systems
Regenerator cyclones
Absorber/kiln cyclones
Pulverizer cyclones
Off-gas bag filters
Absorber off-gas bag filter
Regenerator air filter
Absorber/kiln air filter
Pulverizer air filter
Transport Systems
Processed stone elevator
Cooled stone elevator
Materials Handling
Pulverizer
Storage silos
Subtotal
Total Equipment
(1) Common Items between Dry
Sulfation and Direct Disposal
Dry sulfation Option
Spent stone cooler
Regenerator air blower
Off-gas booster fan
Absorber
$ 42,890
36,240
7,220
11,430
19,960
(4)
320
350
2,900
Absorber
Kiln
(4)
51,250
(4)
(4)
420
2,900
1,200
73,740
325,660
1,305,810
$1,306,510
325,660
3,712,210
$3,867,780
Kiln
52,120
66
-------
Table 16 (Cont)
Waste heat boiler
Steam drum
BFW heater
Holoflite cooler
Regenerator cyclones
Off-gas bag filter
Regenerator air filter
(2) 1 item
(3) 7 items
(4) Included in basic items
Absorber
700
$700
Kiln
$ 26,040
11,040
13,000
41,670
11,000
700
$155,570
67
-------
Table 17
INVESTMENT REQUIRED FOR BRIQUETTING OPTIONS FOR PROCESSING
SPENT CAFB REGENERATOR STONE
Basic Items, as in Direct Disposal $ 274,290 (1)
Curing Chamber 260,270
Curing Pond 48,610
Transport Systems
Air supply packages 15,890
Cooled stone elvators 2,500
Screw conveyor 6,540
Belt conveyors 101,300
Materials Handling Systems
Storage silos 130,890
Hammer mill 2,810
Briquettes 145,000
Blender 8,710
Fines screen 5,580
Scalping screen 12,820
Briquetter feed hopper 12,750
Flyash hoppers 19,750
Additional Items 773,420
Total Equipment $1,047,710
(1) 10 items
68
-------
Table 16 (Cont)
Waste heat boiler
Steam drum
BFW heater
Holoflite cooler
Regenerator cyclones
Off-gas bag filter
Regenerator air filter
(2) 1 item
(3) 7 items
(4) Included in basic items
Absorber
700
$700
Kiln
$ 26,040
11,040
13,000
41,670
11,000
700
$155,570
67
-------
Table 17
INVESTMENT REQUIRED FOR BRIQUETTING OPTIONS FOR PROCESSING
SPENT CAFB REGENERATOR STONE
Basic Items, as in Direct Disposal $ 274,290 (I)
Curing Chamber 260,270
Curing Pond 48,610
Transport Systems
Air supply packages 15,890
Cooled stone elvators 2,500
Screw conveyor 6,540
Belt conveyors 101,300
Materials Handling Systems
Storage silos 130,890
Hammer mill 2,810
Briquettes 145,000
Blender 8,710
Fines screen 5,580
Scalping screen 12,820
Briquetter feed hopper 12,750
Flyash hoppers 19,750
Additional Items 773,420
Total Equipment $1,047,710
(1) 10 items
68
-------
Environmental Impact
Since all options would be designed to meet existing environmental
regulations and criteria, there should be no difference in SO emissions.
There could be differences in the environmental impact of solid residues,
however, the dry sulfation option would have a spent sorbent effluent
set by the Ca/S treat ratio of about 1:1. All the other options have
the theoretical possibility of a smaller spent sorbent rate because of
a lower Ca/S ratio. This could result from use of more attrition-
resistant limestones, modified main process conditions that degrade the
sorbent activity at a lower rate, or synthetic sorbents. The first of
these appears to have limited potential for reducing the generation rate
of residues, but the other two are receiving some attention at the
laboratory level.
In addition, all the options except dry sulfation would generate
residues from the sulfur recovery plant associated with them. While
different systems might be used, for the present we have concluded
that the stone processing options should be debited equally for this
aspect. These residues might be char, ash, and possibly chemical
wastes, as from the Stretford Process.
Conclusions about leachate characteristics have been reported in
Volume 3. Among them was the finding that the leachate quality from
processed spent sorbent was equal to or better than that from natural
gypsum. Trace elements are not expected to result in environmental
problems.
Effect of By-Product Credits
The effect of a by-product credit for sulfur recovery has proved
to be unimportant. The 200 MW plant at 3 percent sulfur will produce
about 10,770 Mg (10,600 long tons) sulfur per year at 90 percent overall
recovery. The price of domestic sulfur has fJuctuated roughly between
$17 and 28/Mg over the period 1950-1973.14 In J974, the price jumped to
about $30/Mg, and to $45/Mg the following year.J<3 it is currently
69
-------
39
about $64/Mg. Interestingly enough, while the price was increasing
sharply, demand declined only slightly, which confirms the view that
sulfur is an essential mineral for which industry has found no sub-
stitute. The rise in price can be attributed to the sharp increase
in energy costs, since most of the domestic sulfur is Frasch sulfur.
At $64/Mg, the gross realization of recovery is $689,000/yr which does
not even cover annual capital charges of 17 percent or $1,550,000.
This reconfirms an earlier assessment that sulfur recovery is uneconomi-
cal by available processes. Further, 70 percent of the domestic sulfur
is used in the southern states, with Florida taking 28 percent, the
latter presumably because of the phosphate industry. The demand for
electric power is nationwide, but the sulfur market is concentrated
in, perhaps, five states.
A word about possible by-product credits for briquets or lime:
at 92 Mg of briquets/day, the gross annual realization at $151,800
vs. charges on incremental capital of $268,600. Concrete block (8"
x 8" x 16"), however, is quoted in the Pittsburgh area at about $27/Mg.
If we assume that block made with spent sorbent could command the same
price, the gross realization would be $820,000. Thus, the case for
making cinder block appears more attractive than for making a coarse
aggregate. Different equipment, however, would be required.
Variations
After further data on the residence time for gas in the dry sulfa-
tion case are obtained, the size of the absorber and the amount of gas
recycled to the gasifier may prove uneconomical. This option may still
be made more attractive than any with a sulfur recovery plant, however,
by adding additional fresh lime either to the gasifier or to the
absorber. This technique would be contrary to current efforts to reduce
lime usage but is suggested as an alternative possibly more acceptable
to power plants than having essentially to operate a chemical plant in
the form of a sulfur recovery plant.
70
-------
If dry sulfation is rejected, another possibility would be to build
the CAFB plant next to an existing sulfuric acid plant and sell it the
regenerator off-gas. In this case the gas would have to be cooled from
1070°C, passed through a baghouse or a high-temperature electrostatic
precipitator, and boosted in pressure. The cost of these facilities
should be substantially less than those for a complete sulfur recovery
plant.
CONCLUSIONS
The technical and economic evaluations clearly show that a dominant
factor in the CAFB process is the investment required for a sulfur
recovery plant. Since this cost is eliminated in the dry sulfation
process, further development work should be carried out for this process.
In parallel with such work, development of a process such as
briquetting to make aggregate or direct disposal via block manufacturing
should continue:
• To preserve a back-up option to either direct disposal or dry
sulfation.
• To offer the possibility of by-product credits and resource
conservation.
In either case these options require confirmation of the acceptability
of the products.
71
-------
7. OCEAN DISPOSAL
BACKGROUND
As noted in Appendix S, Spent Limestone Disposition, of the 1975
3
Report, ultimate disposal of spent sorbent by ocean dumping was con-
sidered a possibility for locations in and near coastal areas although
Federal regulations already had severely reduced the probability of
obtaining a permit for a new source. Recognizing that ocean
disposal might be excluded as an option, we nevertheless considered tests
to determine the technical feasibility of ocean dumping of spent sorbent
to be constructive for providing a basis for evaluation. Of the five
classes of permits, at least three to be available. The special permit
was valid for three years, and the research permit held for only 18
months; both were renewable. Interim permits expired annually, could
be reapplied for, but would not be granted for new sources unless
Phase A of an implementation plan was completed. This plan would
either eliminate dumping or bring it within the requirements of Sec-
tion 227.3 of the Final Regulations. Utilization of spent limestone
sorbents as ocean reefs may provide an attractive option. This alterna-
•43
tive is under study by others for flue-gas desulfurization wastes.
EXPERIMENTAL WORK
The facilities of the Westinghouse Ocean Research Laboratory were
utilized to perform tests on three specimens of solid sorbents with
43
actual seawater. Some of the results were reported in March 1978.
The stone of main interest was regenerator stone from Run 9 of ERCA's
CAFB oil gasification pilot plant in Abingdon, England. The second
was a spent dolomite from Argonne National Laboratories (ANL). The
72
-------
third was a simulated spent dolomite made at the Westinghouse R&D
Center by successively calcining, sulfiding, and oxidizing Tymochtee
dolomite. Analyses of these specimens are shown in Appendix D.
The compositions shown were obtained by calculation from the chemi-
cal analyses for sulfate, sulfide, calcium, and magnesium. Carbon
dioxide was measured in the ANL stone but not on the other two, since
they are produced under conditions such that residual CO would be
very low. The dolomite from Run D-2 was selected because its composition
was closest to that expected for a commercial plant.
Initial Tests
Initial tests (Appendix I) measured the temperature rise and pH
changes in Maryland Bay water when contacted with the spent sorbents.
The sorbents are described in Table 1-1.
Table 1-2 shows that 90 percent of the temperature rise was pro-
duced within six hours. On the other hand, the pH showed an immediate
rise for all three specimens of 30 to 60 percent of the ultimate
increase, with nearly all of the total increase occurring within one hour.
Table 1-3 shows somewhat lower values for temperature rise and pH
for the ANL and Westinghouse stones at 100 g/£. We attribute this
difference to the considerably smaller degree of mechanical mixing in
Test II. Increasing the treat ratio increased the ultimate pH within
Test II, but the final values for ANL and Westinghouse stones were
about the same as for Test I. The CAFB stone showed significantly
greater pH at 400 g/£ treat ratio and about 10X the heat generation
of the other two stones.
Tests with Seawater
Preliminary tests of leaching were done with Maryland Bay water,
yielding 6-hour and 24-hour samples of filtrates and suspended floe,
plus the final solids. In general, the main tests, done with seawater
at an actual dump site, yielded four kinds of samples: filtrate and
73
-------
floe from the liquid, plus core and surface samples of the residual
solids. About two-thirds of the samples were inspected, as summarized
in Appendix J, by procedures described in Table J-l. Inspections
included the following:
1. Spectrochemical analysis for Cr, Cu, Ni, Pb, Sn and V
2. Atomic absorption (flameless) determination of Se and Hg on
three samples (IIA 400L, IIA 400CS, II SL1)
3. Weight-volume relationships of sediment in liquid samples
4. Total volume of solid-liquid samples submitted for analysis
5. Analysis of both solid and liquid fractions for the listed
impurities
6. Determination of fluorine (wet chemistry) on selected samples.
Results from Seawater Tests
The untreated sample of Maryland bay water contained levels of
selected trace elements relative to water standards as shown in Table 18.
Table 13
COMPARISON OF TRACE ELEMENTS IN MARYLAND BAY WATER
WITH ESTABLISHED DRINKING WATER STANDARDS
Element
Chromium
Copper
Lead
Nickel
Tin
Vanadium
US Public
Health Service
P
P
DNP
P
P
NS
Commonwealth
of Pennsylvania
NS
P
NS
NS
NS
NS
World Health
Organization
DNP
P
DNP
NS
NS
NS
P = Passes: level found is not more than the standard
NS = No standard
DNP = Does not pass, meaning the upper limit established by
spectrographic analysis is higher than the standard.
Actual value in the sample may be lower than standard.
-------
Except for lead, whose level is in the doubtful category, Maryland bay
water contained tolerable levels of the above six elements.
Effect of Stirring Time in Extraction of Trace Elements from Spent Sorbents
Test I included sampling, after 6 and after 24 hours of mechanical
stirring, of suspensions of seawater and spent sorbent. For all three
spent sorbents there was no difference in the upper limit found in the
filtrates for seven elements as a result of stirring longer than 6 hours.
Levels of these elements in the residual solids were 10 to 1000X
those in the filtrates, so apparently they are present in the spent sor-
bents in not readily leachable forms. As expected, the CAFB stone had
high levels of vanadium and nickel. The ANL stone showed 100 ppm of
chromium, while Tymochtee dolomite from the Westinghouse tests had
150 ppm of chromium and 200 ppm of nickel. These levels should be
checked on other samples of these stones. It nay be that chromium and
nickel are being picked up from the reactor systems.
Effect of Treat Ratio on Extraction of Trace Elements from Spent Sorbents
Treat ratios used ranged from 20 to 400 g/£ seawater. The only
agitation occurring was that due to the motion of the boat at the ocean
disposal test site. The results are for 24 hours of contact time. No
trend due to treat ratio was detected, either in the filtrates or in the
residual solids.
Extraction Rate of Trace Elements
In general, the levels of trace elements found in the filtrates
were no higher than background levels. This finding also supports the
conclusion that the trace elements of interest (Cr, Cu, Ni, Pb, Sn, V)
are apparently present in not readily leachable forms.
Spot checks were also made of the levels of three other elements, as
shown in Appendix K. Background levels of mercury and selenium in Mary-
land bay water (Specimen II SL 1) were below the limits of detection by
75
-------
flameless atomic absorption, while fluorine was found to be 0.68 ppm.
Although both the CAFB and the ANL stones showed fluorine contents of
the order of 40 ppm, their leachates showed values somewhat less than
the background bay water. Selenium, on the other hand, was not detected
in either the leachate or the residual solids from the ANL stones.
Mercury was found in the ANL stone to be less than 10 ppb, while the
leachate was not detectably higher than background.
Trace Element Material Balances
The accuracy of the spectrographic analysis is estimated to be
within a factor of 3. A result of, say, 12 ppm is to be read as having
a high probability of being in the range of 4 to 36 ppm. A spot check
of one sample, IIC20L, yields the following balance for vanadium:
Mg of V
Filtrate 1000 g @ 0.1 ppm 0.1
Floe 0.3 g @ 100 ppm 0.0084
Solids 20 g @ 2000 ppm 40.0
Total found 40.1
Feed Solids 20 g @ 1% 200.0 mg
Seawater 1000 g @ 0.1 ppm 0.1
200.1
V recovery = (40.1/200.1) 100 = 25%
The recovery is actually higher, since the vanadium content of the
residual solids was reported as more than 2000 ppm. For this initial
set of tests, we considered it unnecessary to pursue this disparity
further.
Comparison with Regulations
Table 19 shows a comparison of trace element concentrations arising
from spent sorbents with levels promulgated under federal regulations.
76
-------
Table 19
COMPARISON OF TRACE ELEMENT CONCENTRATION PRODUCED FROM SPENT
SORBENTS WITH ESTABLISHED CRITERIA, ppm
Mercury
In solids
In liquids
Selenium
In solids
In liquids
Fluorine
In solids
In liquids
Maximum
Permissible Level
Argonne Spent
Dolomite
CAFB Spent
Limestone
0.75*
1.5a: 0.0020*
0.01
2.41
<0.010
<0.001
34
0.23
46
0.41
Source: Reference 42.
Source: EPA Proposed Interim Primary Drinking Water Standards,
Federal Register, 40(51): 11989-98 (March 14, 1975).
The values shown for mercury under maximum permissible levels are for
solid and liquid phases of water, respectively, in the case of ocean
dumping. Where blanks are shown, no standards have been promulgated.
We conclude that mercury and fluorine do not constitute a problem,
while a conclusion on selenium is best deferred until additional
measurements are available.
ASSESSMENT
Overall, the data support the view that the leach rate of trace
elements such as chromium, copper, nickel, lead, tin, vanadium, mercury,
selenium, and iron are essentially zero in the first 24 hours from all
three samples tested. Experience with leaching of solids in the labora-
tory generally shows that the leach rates decrease with time from their
initial values.
77
-------
CONCLUSIONS
While these test results are certainly not definitive, they suggest
that ocean disposal of spent calcium-based sorbents from fluidized-bed
gasification or combustion of fossil fuels may not have a deleterious
impact on the ocean environment due to trace elements. The effect of
major constituents such as calcium, magnesium, and sulfate ions has
not been investigated here, but it is possible that additional calcium
may even be beneficial to aquatic life.
An opposing conclusion was obtained from EPA as recorded in Appen-
34
dix A of the 1978 report. The argument was based on several factors:
• The general policy of EPA to phase out ocean dumping by 1981
• The observation that the solid wastes under consideration here
may contain vanadium, mercury, and arsenic as well as being
highly alkaline
• The stringent criteria to be met by those seeking a permit
for ocean disposal
• The high cost.
This conclusion was directed at commercial-scale dumping and left open
the possibility of dumping for research purposes or on an interim basis.
The same report contained the conclusion from the Westinghouse
Ocean Research Laboratory that it may be possible to obtain interim
permits, noting that the Final Regulations specify a limiting permissible
concentration for selected elements in the receiving water, not in the
solid waste, and implying that dumping would meet criteria on trace
elements if these elements were fixed in the wastes in a nonleachable
form.
78
-------
8. REFERENCES
1. Archer, D. H. , D. L. Keairns, J. R. Hamm, R. A. Newby, W. C. Yang,
L. M. Handman, and L. Elikan, Evaluation of the Fluidized Bed
Combustion Process, Vols. I, II, and III. Report to EPA, Westing-
house Research and Development Center, Pittsburgh, PA, November 1971,
OAP Contract 70-9, NTIS PB 211-494, 212-916, and 213-152.
2. Keairns, D. L., D. H. Archer, R. A. Newby, E. P. O'Neill, E. J. Vidt,
Evaluation of the Fluidized-Bed Combustion Process, Vol. IV,
Fluidized-Bed Oil Gasification/Desulfurization. Report to EPA,
Westinghouse Research and Development Center, Pittsburgh, PA,
December 1973, EPA-650/2-73-048d, NTIS PB 233-101.
3. Keairns, D. L., R. A. Newby, E. J. Vidt, E. P. O'Neill, C. H.
Peterson, C. C. Sun, C. D. Buscaglia, and D. H. Archer, Fluidized
Bed Combustion Process Evaluation - Residual Oil Gasification/
Desulfurization Demonstration at Atmospheric Pressure. Report to
EPA, Westinghouse Research and Development Center, Pittsburgh, PA,
March 1975, EPA-650/2-75-027 a&b, NTIS PB 241-834 and PB 241-835.
4. Sun, C. C., Chemically Active Fluid Bed for SOX Control: Volume 3,
Sorbent Disposal. Report to EPA, Westinghouse Research and Develop-
ment Center, Pittsburgh, PA, July 1979, EPA-600/7-79-158c.
5. O'Neill, E. P., D. L. Keairns, and M. A. Alvin, Sorbent Selection
for the CAFB Residual Oil Gasification Demonstration Plant. Report
to EPA, Westinghouse Research and Development Center, Pittsburgh,
PA, March 1977, EPA-600/7-77-029, NTIS PB 266-827.
6. Bachovchin, D. M., P. R. Mulik, R. A. Newby, and D. L. Keairns,
Solids Transport between Adjacent CAFB Fluidized Beds. Report to
EPA, Westinghouse Research and Development Center, Pittsburgh, PA,
January 1979, EPA-600/7-79-021.
7. Keairns, D. L., W. G. Vaux, N. H. Ulerich, E. J. Vidt, and R. A.
Newby, Chemically Active Fluid Bed for SOX Control: Volume 1,
Process Evaluation Studies. Report to EPA, Westinghouse Research
and Development Center, Pittsburgh, PA, December 1979, EPA-600/7-
79-158a, to bt; issued.
79
-------
REFERENCES (Cont)
8. Keairns, D. L., C. H. Peterson, and C. C. Sun, Disposition of Spent
Calcium-Based Sorbents Used for Sulfur Removal in Fossil Fuel Gasi-
fication, Presented at the Solid Waste Management Session, 69th
Annual Meeting, AIChE, November 28 - December 2, 1976, Westinghouse
Scientific Paper 76-9E3-FBGAS-P1.
9. Craig, J. W. T., et al., Chemically Active Fluid Bed Process for
Sulfur Removal During Gasification of Heavy Fuel Oil - Second Phase.
Report to EPA, Esso Research Centre, Abingdon, UK, November 1974,
EPA-650/2-74-109, NTIS PB 240-632/AS.
10. Chemically Active Fluid Bed Process (CAFB). Monthly report to EPA,
Foster Wheeler Energy Corporation, Livingston, N. J. May 29 -
June 25, 1978, Contract 68-02-2106.
11. Minerals Yearbook 1975, Preprint on Stone, United States Department
of the Interior, Bureau of Mines, U.S. Government Printing Office,
Washington, DC
12.
Minerals Yearbook 1975, Preprint on Lime, United States Department
of the Interior, Bureau of Mines, U.S. Government Printing Office,
Washington, DC
13. Boynton, R. S., Chemistry and Technology of Lime and Limestone,
New York; Interscience Publishers; February 1967.
14. Statistical Abstract of the United States, 1975, U. S. Department
of Commerce, Bureau of the Census, Washington, DC
15. Agricultural Stabilization and Conservation Service, Current
Bulletins and Newsletters, Beaver, PA and Washington, PA.
16. Abernethy, R. F., M. J. Peterson, and F. H. Gibson, Spectrochemical
Analysis of Coal Ash for Trace Elements, R17281, Bureau of Mines,
U. S. Department of the Interior, July 1969.
17. Erickson, R. L., Coastal Abundance of Elements and Mineral Reserves
and Resources, United States Mineral Resources Geological Survey
Professional Paper 820, D. A. Brobst and W. P. Pratt, editors,
U.S. Department of the Interior, Washington, DC
18. Page, A. L. and A. C. Chang, Trace Element and Plant Nutrient
Constraints of Recycling Sewage Sludges on Agricultural Land,
The Second National Conference on Complete Water Use, 1975,
Chicago, ILL, AIChE and Environmental Protection Agency.
80
-------
REFERENCES (Cont)
19. Lovell, H. L, Appraisal of Neutralization Processes to Treat
Coal Mine Drainage, Pennsylvania State University, University Park, PA.,
EPA-670/2-73-093, November 1973.
20. Ford, C. T., J. F. Boyer, and R. A. Glenn, Studies of Limestone
Treatment of Acid Mine Drainage, Part II, Bituminous Coal
Research, Inc., Monroeville, PA., Water Pollution Control Research
Series Publication No. 14010 EIZ 10/71, Environmental Protection
Agency, December 1971.
21. Wilmoth, R. C., "Limestone and Limestone-Lime Neutralization of Acid
Mine Drainage", Industrial Waste Treatment Research Laboratory,
Rivesville, W. Va., EPA-670/2-74-051, June 1974
22. Moss, E. A., "Dewatering of Mine Drainage Sludge", Coal Research
Bureau, West Virginia University, Morganstown, West Virginia,
Water Pollution Control Research Series Publication No. 14010 FJX 12/71.
Environmental Protection Agency, December 1971.
23. Shreve, R. N., "The Chemical Process Industries, Ch. XI, Cements,
Calcium and Magnesium Compounds", McGraw-Hill Book Co., Inc.,
New York 1945.
24. "Energy Conservation Potential in the Cement Industry", FEA Conservation
Paper No. 26, 1975.
25. "Energy Consumption in Manufacturing", The Conference Board, Ballinger
Publishing Co., Cambridge, Mass, 1974.
26. Minerals Yearbook 1968, Vols. I-1I, U. S. Department of the Interior,
Bureau of Mines, U. S. Government Printing Office, Washington, D. C.
27. "New Cement Uses Fly-ash, Cost Less to Make", Chemical and
Engineering News, April 5, 1976.
28. Minerals Yearbook 1973, U. S. Department of the Interior, Bureau
of Mines, U. S. Government Printing Office, Washington, D. C.
29. Annual Book of ASTM Standards, 1973, Part 9, Cement, American
Society for Testing and Materials, Philadelphia, Pa.
30. Survey of Current Business, January 1976, Vol. 56, No. 1, Part 1,
U. S. Department of Commerce.
31. Annual Book of ASTM Standards, 1973, Part 10, Concrete and Mineral
Aggregates, American Society for Testing and Materials,
Philadelphia, Pa.
81
-------
REFERENCES (Cont)
32. Orchard, D. F. , Concrete Technology, Vol. 3, Properties and
Testing of Aggregates, Third Edition, John Wiley & Sons.,
New York, N. Y.
33. Kunii, D. and 0. Levenspiel, Fluidization Engineering, Chapter 3,
John Wiley & Sons, New York, N. Y., 1969.
34. Perry, J. H., Chemical Engineer's Handbook, 4th Ed., New York;
McGraw-Hill Book Co.; 1963.
35. Kunii, D., and 0. Levenspiel, Op. cit., Chapter 7.
36. Guthrie, K. M., Capital Cost Estimating, Chemical Engineering,
March 24, 1969.
37. Pikulik, A., and H. E. Diaz, Cost Estimating for Major Process
Equipment, Chemical Engineering, October 10, 1972.
38. Merwin, R. W., Commodity Data Summaries, 1976 - Sulfur, Bureau
of Mines, U. S. Department of the Interior, U. S. Government
Printing Office, Washington, DC
39. A Growing Squeeze on Sulfur, Business Week, August 22, 1977.
40. Title I, Marine Protection, Reserve and Sanctuaries Act of 1972,
Public Law 92-532, 86 Stat. 1052 (33 U.S.C. 1411-1421).
41. Federal Water Pollution Control Act Amendment of 1972, Public Law
92-500, Section 403(c).
42. Title 40, Chapter I, Subchapter H - Ocean Dumping, Final Regulations
and Criteria, Federal Register, Vol. 38, No. 198, October 15, 1973.
43. Santhanam, C. J., R. R. Lunt, and C. B. Cooper, Current Alternatives
for Flue Gas Desulfurization (FGD) Waste Disposal—An Assessment,
Proceedings of the Symposium on Flue Gas Desulfurization, Vol. I,
Las Vegas, NV, March 1979, Washington, DC: Environmental Protection
Agency; 1979, EPA-600/7-79-167a.
44. Sun, C. C., C. H. Peterson, R. A. Newby, W. G. Vaux, and D. L.
Keairns, Disposal of Solid Residue from Fluidized Bed Combustion:
Engineering and Laboratory Studies. Report to EPA, Westinghouse
Research Laboratories, Pittsburgh, PA, March 1978, EPA-600/7-78-049.
82
-------
APPENDIX A
PRELIMINARY DEAD-BURNING/SINTERING STUDIES
This section covers tests conducted to determine whether CAFB spent
regenerator stone could be rendered environmentally inactive by subject-
ing it to high temperatures. In CAFB-9-DB-1* samples of spent stone
were tested at three temperature levels and two time intervals, and
changes in weight and BET surface area were observed.
Approximately 33 g of powder were used in each experiment. The
powder was placed on an alumina boat covered with platinum foil and heat
treated at temperatures of 1070, 1250, and 1550°C and for times of 2.5
and 24 hours at each temperature. The heat treatment was conducted in
air with a rate of rise of temperature of approximately 100°C/hr to the
test temperature. Following heat treatment the powder was furnace
cooled to room temperature.
WEIGHT CHANGES
Each sample was weighed before and after heat treatment. These
data are presented in Table A-l and Figure A-l. Since the finer frac-
tions of the powder sintered and stuck to the platinum foil, the powder
was weighed along with the alumina boat and the foil. The data showed
that, at the two higher temperatures, the weight gain ultimately changes
to a weight loss. The weight gain of approximately 4 to 4.5 percent was
complete in between 2 and 5 hours at 1070 and 1250°C, and in only 2 hours
.it 1550°C. Within the duration of the experiment, there was no loss in
weight at 1070°C. There was a rapid weight loss, however, relative to
*Label means: Sample No. 1 for dead-burning test; origin CAFB Run No. 9.
83
-------
Table A-l
CHANGE IN WEIGHT OF CAFB STONE WITH HEAT TREATMENT
Temperature, °C
Weight % Gain or Loss
2 hr
1070 +3. 70
1250 +3.19
1550 +4.57
5 hr
+3.54
+4.59
4.13
24 hr
+4.09
-1.95
-3.93
Oxidation of
CaS
Decomposition of
2 4 6 8 10 12 14 16 18 20 22 24 26 28
Time, hr
Figure A-l - Effect of Heat Treatment on CAF14 Stone in Air
84
-------
the initial weight of approximately 4 percent at 1550°C after 5 hours of
heat treatment, and this remained nearly constant during additional heat
treatment to 24 hours. The case at 1250°C was intermediate between 1070
and 1550°C.
The weight changes with temperature and time can be explained in
the following way. CAFB stone contains small amounts of CaS, CaSC>4, and
inert materils, such as iron oxide (FeO), silica, and so on. The main
constituent is CaO. When the stone is heated, the following reactions
can occur:
CaS + 200 -> CaSO. oxidation
2. 4
CaSO -> CaO + SO +1/2 0 decomposition
The weight gain may be attributed to the oxidation of CaS to CaSO and
the subsequent weight loss to the decomposition of CaSO, to CaO. As the
heat treating temperature and time are increased, the decomposition is
accelerated, as shown by the data at 1550°C. The predicted weight
changes due to this treatment depend on the accuracy of the chemical
analyses of the spent sorbent for sulfide and sulfate as well as on the
variability of the sulfur content on individual particles. The frac-
tional change in weight on oxidation, if one assumes no losses, should
be
f*= (136.14 - 72.14) ^^= 1.996 ^
where S is the weight fraction of sulfide sulfur in the spent sorbent
before heating. On decomposition, the fractional weight loss should
be
AW /72.14 - 56.08\ ^136.14 - 56.08\
W \ 32.06 / 1 V 96-06 / 2>
or AW/W = 0.501 $i + 0.833 S2, where S2 is the weight fraction of sulfate
sulfur in the spent sorbent expressed as SO,. Using sulfur analyses for
85
-------
CAFB stone (2.24 wt % S and 3.07 % SO,), we calculated the maximum weight
gain as 4.47 percent and the maximum weight loss as 3.68 percent. These
compare very well with the observed changes.
Examination of the sulfur contents of the 5-hour samples showed
sulfur losses with dead-burning as in Tables A-2 and A-3. A sulfate
content greater than that in the original stone was explained by
assuming that some of the CaS was oxidized to CaSO and the balance to
CaO. A smaller sulfate content meant that both the original CaS and
CaSO, decomposed to CaO. The calculations showed that about 15 percent
of the sulfide sulfur was lost at the lower temperatures, conceivably
applicable to dead-burning, but overall only 11 percent was lost. The
balance of the sulfide sulfur was oxidized to sulfate.
Table A-2
SULFUR RETENTION OF DEAD-BURNED CAFB-9 REGENERATOR STONE*
Temperature, °C
Time, hr
Composition, wt
Sulfide
1070 5 0.302
1250 5 0.021
1550 5 0.016
Sulfate
7.58
8.45
0.46
I
Calcium
60.00
62.88
67.84
*Sample was CAFB-9-DB1
Raising the temperature to 1250°C did result in 99 percent conver-
sion of the sulfide, with essentially the same overall sulfur loss.
Hence, if landfill is the end disposal method, 1250°C might be adequate,
At 1550°C the residual sulfide was about the same as at 1250°C, but the
overall sulfur loss (95 percent) approached completion.
86
-------
Table A-3
SULFUR LOSS IN DEAD-BURNING CAFB REGENERATOR STONE FOR 5 HOURS
Form of S Loss
CaS Conversion, %
Dead-Burning
Temperature
L070°C 1250°C
To CaO 15.7 14
To CaSO. 70.6 84
4
Unconverted 13.7 1
CaSO, Decomposition, %
To CaO
Unconverted
— _
.3
.7
.0
_
1550°C
99.3
0.7
85.6
14.4
Overall S Loss, % 10.7 9.3 95.0
We concluded that, if the CAFB stone was heated for not more than
2 hours at 1550°C or 5 hours at 1250°C, oxidation of the residual CaS
would be essentially complete. The stone may still not be dead-burned,
SURFACE AREA STUDY
The results of surface area estimates by BET are shown in Table A-4.
The results may be interpreted on the basis of simultaneous occurrence
of oxidation, decomposition, and sintering. The greatest surface area
was obtained after 2 hours of heat treatment at 1550°C, commensurate
with the time required for completion of oxidation. Also, at each tem-
perature level, there was an initial increase in BET area.
Both decoraposit ion and sintering should decrease BET area. Since
the CaSO lattice is larger than that of CaO or CaS, however, it is con-
ceivable that some dislocations are created on oxidation which, in
effect, exposes additional surface area. The BET results suggest that,
if dead-burning is indicated by minimum surface area, 1550°C for 2 to
5 hours or 1250°C for more than 24 hours is required.
87
-------
Table A-4
SURFACE AREA OF THE HEAT TREATED CAFB
STONE BY BET METHOD
Temperature, °C
o
Surface Area, m
2 hr
1070 0.58
1250 0.13
1550 1.75
5 hr
0.53
1.35
0.36
/g
24 hr
1.33
1.05
0.55
OTHER INDICATIONS
The leaching data reported in another section of this report
also shed some light on the dead-burning. The heat involved in producing
a suspension of Ca(OH) from solid CaO is above 65,300 J (15,600 cal)/g
mole. The 3 g samples, therefore, should release 3492 J (834 cal) if
hydration is complete. The 17°C rise in 20 ml water means 1424 J
(340 cal) were actually absorbed, which is only 41 percent of the theo-
retical maximum. This calculation neglects the heat of hydrolysis of
CaS, the thermal capacity of the flask, and the fact that no stirring
was used. The temperature rise observed, therefore, may be high because
of the first omission, low because of the second, and high because of
the third. With these reservations the stone as produced is about
40 percent dead-burned.
The 20 ml of water used is 23 times the theoretical needed for
hydration of the CaO present. The actual process, therefore, could
quench the CAFB spent stone in water, recycle any H S evolved, and dis-
pose of the slaked stone to a user such as a municipal sludge or an
acid mine drainage treatment plant. Alternatively, the stone could be
oxidized with air to retain the sulfur as CaSO . Previous leaching
tests have suggested that this may be done at ambient conditions.
88
-------
Dead-burning of CAFB-9 regenerator stone was examined further
by checking the effect of particle size on it. Two size fractions were
prepared by grinding CAFB stone to -88 + 63 and -44 ym.
Samples of about 10 g each were placed in alumina boats covered
with platinum foil and heat treated at 1250 and 1550°C for 2, 5, and
24 hours. The heating was done in air with a temperature rise of
50°C/hr to the test temperature. The samples were then furnace cooled
to room temperature.
Figure A-2 is a photograph of the dead-burned samples. All the
samples heated to 1550°C took on a yellowish color, whereas those at
1250°C remained an off-white color. Large aggregates were formed with
an increase in time and temperature, as expected. Simple heating, thus,
does result in sintering.
Weight changes are given in Table A-5. The gains are smaller
than those obtained previously, which can be explained by considering
that the decomposition of sulfate proceeded more readily with the finer
particles. The magnitude of the loss on continued heating, however, is
greater than expected. Some of the loss is possibly due to CO and
perhaps to moisture. The weight changes for the two particle sizes are
comparable.
Chemical analyses as summarized in Table A-6 show that 1250°C
for 2 hours is sufficient to essentially eliminate sulfide sulfur from
-44 urn particles in a static bed. The larger particles (-88 +63 pm)
may require times longer than 5 hours, even at 1550°C.
Sulfate sulfur can be reduced by about one order of magnitude to
the level of 0.5 to 1.0 wt % SO by heating at 1550°C for 2 hours. The
larger particles retain more sulfate sulfur (1% vs 0.5%) at 1550°C than
the smaller particles. Heating at 1250°C results in 0 to 10 percent
loss of sulfate sulfur, which may be offset by oxidation of sulfide
sulfur.
89
-------
Table A-5
WEIGHT CHANGES IN DEAD-BURNING OF GROUND CAFB-9 REGENERATOR STONE
Particle Size, um
Temperature, °C
-88 + 63
1250
1550
-44 + 0
1250
1550
Time, hr
2
5
24
Initial wt, g
Final wt, g
% change
Initial wt, g
Final wt, g
% change
Initial wt, g
Final wt, g
% change
8.466
8.633
+1.98
11.737
11.583
-1.30
10.196
10.054
-1.39
8.054
7.456
-8.05
7.154
6.615
-7.53
12.762
11.838
-7.24
10.636
10.854
+2.05
14.657
14.453
-1.39
14.408
14.278
-0.90
10.058
9.250
-7.99
8.892
8.166
-8.16
16.728
15.428
-7.77
Table A-6
EFFECT OF PARTICLE SIZE ON SULFUR RETENTION IN DEAD-BURNING
CAFB REGENERATOR STONE
Dead-burning Temperature, °C
Particle Size, um
1250
-88 + 63
-44
1550
-88 + 63
-44
Dead-burning Time, hr
2
5
24
2
5
24
Sulfide Sulfur
0.043
0.022
0.000
7.30
7.37
6.55
0.001
0.000
0.009
Sulfate Sulfur
7.08
8.64
8.52
0.006
0.019
0.000
(as SOp
0.96
1.03
0.98
0.000
0.000
0.000
0.48
0.60
0.48
90
-------
Figure A-2 - Dead-Burning of Ground CAFB-9 Regenerator Stone
-------
Overall, these results indicate that utilization of spent sorbent
from the CAFB regenerator as a high purity lime after heating it to
reduce sulfur appears unpromising. Heating to 1550°C from 1070°C
in a 1000 MW plant burning a 3 percent sulfur fuel oil and using a
limestone/sulfur molar makeup ratio of 1/1 would require about 5.6 bar-
rels of fuel oil/hr. At $10/bbl, assuming a 20-year project life,
16 percent capital charges, and a 50 percent tax rate, the incremental
alternative capital investment that could be made to avoid this fuel
cost is estimated at $1,700,000. This figure is even higher when the
other costs associated with installing dead-burning equipment are added
in but reduced by operating costs of the alternative facilities. The
energy penalty at 1550°C is at least 0.4 percent.
A further observation on dead-burning is possible through leaching
tests performed on the sample included in Table A-6. These are reported
in detail in another section of this report. In general, these showed
that calcium and sulfate ions could be leached from the dead-burned sam-
ples, presumably as Ca(OH)_ and CaSO,. Table A-7 presents the calcium
and sulfate results, which may be summarized as follows:
• Calcium
- At 1250°C reduction in leaching, if dead-burning time is
extended from 2 to 24 hours for both particle sizes, is
negligible.
- At 1550°C values are about two-thirds those at 1250°C,
but again effect of exposure time is negligible.
• Sulfate
- At 1250°C molar values are about one-half those of the
calcium levels; the effect of exposure time at both
particle sizes is negligible.
- At 1550°C values are about one-tenth those at 1250°C and
about one-twentieth of the corresponding calcium values;
the effect of exposure time is negligible.
92
-------
Table A-7
EFFECT OF PARTICLE SIZE ON LEACHATE COMPOSITION FROM
DEAD-BURNING CAFB REGENERATOR STONE
Dead-Burning Temperature, °C
Particle Size, ym
1250
-88 + 63
-44
1500
-83 +63
-44
Dead-burning Time, hr
Calcium, m moles/£
2
5
24
2
5
24
34.4
33.0
32.9
16.2
12.3
14.9
24.6
34.0
34.6
Sulfate,
10.8
15.4
15.5
20.6
21.0
21.1
m moles/ 8,
1.37
1.18
0.85
31.8
21.4
21.0
12.6
1.39
1.70
Saturation values for pure Ca(OH)2 and CaSO, at 20°C are nearly the
same - 22.0 m moles/£. Dead-burning appears to significantly affect
the solubility of CaSO but not that of Ca(OH)2> Since the regen-
erator stone is mainly a CaO, the leaching results appear to deny the
technical feasibility of dead-burning as a disposal method.
93
-------
APPENDIX B
FLUIDIZED-BED TEST FACILITY
A large portion of the data to evaluate the dry sulfation process
in this report was obtained on specimens processed in a 10-cm diameter
fluidized-bed test unit. Figures B-l and B-2 are process flow diagrams
for this unit. Basically, the unit is an apparatus for delivering mix-
tures of gases either from the laboratory supply or from commercial
cylinders through a replaceable distributor plate to a 10-cm x 45-cm
(4 in. x 18 in.) reaction chamber. The chamber is heated by external
electric wraparound heaters. The feed gas mixture may also be preheated
by a heating tape on the supply line.
Off-gas from the reactor is cooled by passing it through a bare
U-tube gas-air exchanger. Carry-over is removed in a cyclone, and fines
are caught in a sintered metal cartridge filter. When residual H S, H»,
or other combustibles are present, the off-gas is passed through a
burner. Methane may be added to the off-gas to ensure a combustible
mixture that will burn stably. The burner is an inverted, truncated,
square pyramid with three levels of wire gauze screens to aid further in
stabilizing flames from a wide range of gas flow rates.
A swing connection was inserted into the gas feed line to the reac-
tor to prevent accidentally mixing fuels with air or oxygen in the feed
system. This must be manually unfastened and reconnected into the
desired system, oxidant or fuel. A further safety measure was the addi-
tion of small vent valves on four of the rotameters as a partial protec-
tion against overpressure. Finally, a Plexiglas plate was installed
over the face of the control panel on which the rotameters were mounted.
-------
VO
M Shutoff Valve
A Regulating Valve
*J Check Valve
c£ Safety Valve
6 Watts Regulating Valve
Laboratory Air, 80psig Max
M
PI-9
Laboratory Nitrogen. 55psigMax yi
-•-- LJ
==^U_KS
~T i^ Connection
* * Fl-10
PI-10
2
DPI-1
Nitrogen
2490 psig
Hydrator
Swing
Connection
PI-2
Feed Gas to Reader
» Nitrogen to Monitor
» Methane to Off-Gas Burner
PI-1 PI-4 PI-£ PI-6] PI-7 f
ffi o4 of o4 cH
02 S02 C02 CH4 H2
2200 34 830 2265 2200 252
Cylinder Pressure, psig
To Sampling System
*• To $02 Monitor
— 10*> Sulfur Dioxide in
Nitrogen 350psig
Figure B-l - Process Flow Diagram for the Gas Supply Section of the
10-cm Fluidized-Bed Test Unit
-------
Dug I680820
10* S02J2£ Caltoration
Nitrogen for Calibration
Chemical
Seal
Trap
DP1-2
115V
Heating Tape y-
Feed Gas 2x576W .304"
Laboratory Nitrogen
Methane
Natural Gas
Pilot Flame
Figure B-2 - Process Flow Diagram for the Reactor Section of
the 10-cm Fluidized-Bed Test Unit
-------
Dug. 1701BI7
13.5"
1/8" Flexitallic Gasket
Not to Scale
1/8" Flexitallic Gasket
38-3/4"
0.9375"
Figure B-3 - Details of 10-cm Fluidized-Bed Reactor
97
-------
9. 170(818
Detail A, Typical Elwation
00
Align Holes In
Middle Plug to
Center between
Outer Plugs
Atountfng
Screw (4)
Detail B. Distributor Plate Center Plug
1/4"
D«tai1 C. Distributor Plate Edge Plugs
-J f-^64"
I—a 375"-4 I
-a 500"—H
Figure B-4 - Distributor Plate Assembly
-------
Owg. 1701616
n
Distributor Plate
1/2" Pipe x 3/8" Tube-
Conax Packing Gland
Assembly I Lava Glandl
SSTube
System Volume
31 ml
• 1/4" Pipe x 3/8" Tube
Gyrolock Fitting
1/2"x 3/8"Tubing Reducer
Whitey 1/2"SS Ball Valve
Figure 15-5 - Reactor Bed Sampling Connections
99
-------
Reactor details are given in Figure B-3, distributor plate details
are in Figure B-4, and bed sampling connections are in Figure B-5. The
reactor and distributor plate shown were used initially on Run L-10.
The hydrator was first used on Run L-9. The information available
indicated that the amount of water vapor required during dry sulfation
was low but not critical. A simple system was devised to split the air
feed, sending roughly half of it through the hydrator.
The bed sampling system was added on later runs. In operation this
system is emptied of solids and back-blown with nitrogen, permitting a
sample of bed material to be withdrawn at any desired time without the
need to purge excessive amounts of previous material.
100
-------
APPENDIX C
FLUIDIZATTON STUDIES
Figures C-l and C-2 show details of a 7.62-cm (3-in) Plexiglas
column in which observations were made on various samples of limestone
and spent sorbent in connection with the dry sulfation studies. Ini-
tially, a perforated plate distributor for the inlet air was used. The
steel wool exhaust filter shown was later replaced with a cartridge-type
filter.
As preparation for sulfating fine particles of CAFB regenerator
stone, fluidization observations were made first on limestone and then on
CAFB stone. This two-step procedure was necessitated by the limited
quantity of CAFB stone available. In one test the initial charge of
3
500 g made a bed 10.0 cm deep with an average bulk density of 1.32 g/cm .
Nitrogen flow was increased gradually to the maximum obtainable of
61 i/min at 15°C, which is equivalent to a superficial velocity of
27.3 cm/s. At 3.4 cm/s one rathole formed but with no visible movement
of the solids. At 4.0 cm/s additional ratholes formed, this time with a
continual ejection of solids from the holes. A crater formed and, as
the gas rate was increased, a large bubble would occasionally break
through and the bed would adjust to the new rate. The area covered by
the crater expanded until all of the surface was active at a gas flow of
10.6 cm/s. Further increases in gas flow resulted in an increasing frac-
tion of the bed becoming active, although even at the highest flow rate,
a portion of the solids at the wall near the distributor remained inac-
tive. Less than 1 g of solids was blown over to the filter on the efflu-
ent line, even though bed material was ejected from the bed to a height
of 17 to 34 cm above the distributor.
Fluidization was next observed with -149 + 74 ym limestone, by use of
a sintered metal distributor plate. The bulk density of a 500 g bed was
101
-------
e =
51
Pressure Tap
Steel Wool
Exhaust Filter
7 cm (3 in I I.D. Plexiglas
Fluidized Bed
V / / / ; = z
Air Inlet
Pressure Tap
- Brass Distributor Plate
^r- Pressure Tap
p Brass Air Chamber
Plexigtas
Fluidized Bed
Ruler
Ftuidizing *
Gas L_ Regulator
PA-Distributor Plate AP
PB - Bed AP
PF - Filter AP
Control Valve
Manometers
Figure C-l - 7.62-cm (3-in) Test Unit
Figure C-2 - Flow Diagram for the 7.62-cm (3-in)
Test Unit
-------
3
1.34 g/cm . At low flow rates the bed rose as an entity to about 21 cm
above the distributor before collapsing. As the flow rate increased, the
bed repeated this performance. Again, we found that even at maximum
flow rate some of the bed (bottom 2 to 8 cm) remained inactive. Adding
another 500 g of limestone yielded the same results. The distributor
pressure drop was a maximum of 10 cm HO.
Finally, fluidization was observed with a bimodal distribution, by
use of a bottom layer of. 250 g of -1680 + 1190 pm limestone overlaid with
500 g of -149 + 74 ym stone. No fluidization occurred at the maximum
flow when only the larger fraction was present. At the maximum flow of
22.0 cm/s, all but the bottom 6 cm were fluidized. There was no signifi-
cant elutriation.
Overall, we concluded that the presence of larger particles operated
to retain the fines even at flow rates considerably above the minimum
fluidization velocity of the fines. Proper design of the distributor
should permit all of the bed to be active.
As part of run CAFB-905, a series of tests was performed to investi-
gate the effect of particle size distribution on the fluidization charac-
teristics of CAFB-9 regenerator stone prior to sulfating the stone. The
objective was to determine the percentage of fine particles that can be
included in the bed material and still retain smooth fluidization charac-
teristics for the bed. The testing was performed using a 7.0-cm id by
91.4-cm (2.75 in. x 36 in.) Plexiglas tube. The tube was mounted on a
brass air plenum chamber, and a brass, orifice-type distributor plate with
thirteen 0,81-mra holes was used throughout the testing. Compressed air,
metered through standard laboratory rotameters, was used as the fluidiza-
tion gas. Bed pressure drop measurements were obtained by means of water
manometers connected to pressure taps along the 7.62 cm column. All test-
ing was done at ambient temperature. The basic particle size tested was
-420 + 125 \im (-40 + 120 mesh), to which various amounts of fines were
added, as noted in the following tables. Tamped packing was obtained by
gently tapping the Loaded column until no further compaction occurred.
Loose packing was produced by t'luidizing the bed and then shutting off
the air. The data are presented in Table C-l.
103
-------
Table C-l
BED COMPOSITION FOR FLUIDIZATION STUDIES
Case
Bed Weight,
g
Packing
Fines
Added,
8
Wt % Fines
I. No Fines Added
1-1
1-2
1-3
900
900
900
Tamped
Tamped
Loose
0
0
0
0
0
0
II. Fines Added, -44 + 0 ym
II-l
1 1-2
II-3
II-4
II-5
II-6
II-7
900
900
900
900
900
900
900
Loose
Loose
Loose
Loose
Loose
Loose
Loose
47.4
100.0
180.0
180.0
225.0
225.0
450.0
5.0
10.0
16.7
16.7
20.0
20.0
33.3
III. Fines Added, -63 + 44 pm
III-l
III-2
III-3
III-4
III-5
III-6
900
900
900
900
900
900
Loose
Loose
Loose
Loose
Loose
Loose
100.0
100.0
225.0
225.0
386.0
386.0
10.0
10.0
20.0
20.0
30.0
30.0
IV. Fines Added, 50-50 Mixture of -63 + 44 and -44 + 0 ym
IV-1
IV-2
IV-3
IV-4
IV-5
900
900
900
900
900
Loose
Loose
Loose
Loose
Loose
100.0
100.0
225.0
225.0
386.0
10.0
10.0
20.0
20.0
30.0
The data were reduced by a previously written computer program.
Typical curves are given in Figures C-3 through C-8. These show that
the addition of fines (1) increases the pressure energy needed to maintain
a fluidized bed, (2) increases the maximum bed pressure drop before
fluidization, and (3) produces a larger pressure drop at a lower super-
ficial gas velocity. Also, the transition from a packed bed to a fluidized
bed is less smooth and occurs over a wider range of gas velocities.
1.04
-------
Curve 716WS-"
"8
CO
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
T I I
0.0% Fines, Dense Packing
0 2 4 6 8 10 12 14 16 18 20
Gas Superficial Velocity, U, cm/sec
Figure C-3 - Fluidization of Ground CAFB-9 Regenerator Stone
Case 1-1
716504A
"8
00
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
i I I r
0.0% Fines, Loose Packing
1
t
1
8 10 12 14 16 18 20
Gas Velocity, U, cm/s
Figure C-4 - Fluidization of Ground CAFB-9 Regenerator Stone
Case 1-3
105
-------
o^
e
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
Curve 716SOH-A
I 1 1 1 \ 1
10.0% of-44 f 0 Mm Fines, Loose Packing -
j I
i i i
8 10 12 14 16 18 20
Gas Velocity, U. cm/s
Figure C-5 - Fluidization of Ground CAFB-9 Regenerator Stone
Case II-2
716509A
E
o
o
i i r
33.3 % of-44 fO urn Fines, Loose Packing
J_
! )0 12 14
Gas Velocity. U. cm/s
16
?0
Figure C-6 - Fluidization of Cround CAFB-9 Regenerator Stone
Case II-7
106
-------
Curve 71f517-A
o
k_
O
10.0* of -63 + 44 (jm Fines, Loose Packing
24 6 8 10 12 14 16 18 20
Gas Velocity, U, cm/s
Figure C-7 - Fluidization of Ground CAFB-9 Regenerator Stone
Case III-2
o_
"8
DD
30.0% of -63 f 44 um Fines, Loose Packing
2 4 6 8 10 12 14 16 18 20
Gas Superficial Velocity, U, cm/s
Figure C-8 - Fluidization of Ground CAFB-9 Regenerator Stone
Case III-6
107
-------
APPENDIX D
DRY SULFATION STUDIES
As in the fluidization studies, we needed to conserve the available
supply of CAFB regenerator stone, so we began the studies with limestone.
The ultimate objective was to determine the ranges of particle size, the
gas contact time, the solids residence time, and of any other factor
required to achieve satisfactory resulfation of the spent stone and
recapture of S0« in a fluidized bed. Data on the test runs are in
tables at the end of this section (Tables D-l through D-3).
INITIAL TESTS WITH LIMESTONE
The initial sulfation run, L-2, was made in the 10-cm unit using
1000 g of -149 + 74 urn limestone. This was calcined for three hours at
760 ± 10°C. The product weight was 955 g, showing less than 10 percent
calcination. The reactor was charged with 300 g of this material and
calcined at 900°C for 30 minutes, with a nitrogen flow of 27 £/min at
15°C. The unit was cooled overnight and a 15.2 g sample removed for
analysis.
The unit was brought to 870°C and the stone sulfated with a mixture
of 6.5 i/min of 10 percent SO and 21 £/min air for 30 minutes. Thn off-
gas was monitored with a Dynasciences S09 monitor. Feed gas was 2.4 per-
cent SO-, and the 0 supplied was about 13 times theoretical. The gas
superficial velocity at operating conditions was 23.3 cm/s. The gas
monitor indicated essentially no pickup of S0~. Chemical analysis of the
sulfated stone indicated 17.9 wt % CaSO,. When the reactor was opened
the bed appeared packed, and there was a single rathole in about the
middle of the surface. Thus, inadequate fluidization caused the SO to
substantially bypass the bed. Table D-l contains material balance Infor-
mation, Table D-2 has time-temperature data, and Table D-3 has chemical
analyses.
108
-------
Table D-l
MATERIAL BALANCE DATA FOR LIMESTONE SULFATION STUDIES
Batch
L-2
Calcination
Re calcination
Sulfation
L-5
Fluidization
Calcination
Sulfation
L-7
Calcination
Sulfation
L-8
Calcination
Fluidization
Recalcination
Sulfation
L-9
Fluidization
Calcination
Sulfation
Resulfation
L-10
Fluidization
Calcination
Sulfation
L-ll
Sulfation
L-12
Sulfation
Gas Flow Rates (£/min) at 15°C,
101.3 kPa Bat^h ^harg<%
N0 Air S09/N0 g
Z i. f. o
27 — — 1000
27 — — 300
21 6 . 5a
91 — -- 500
50 — — 1000
18
18 9.0 20. 8a
18 — — 1000
24.0 28. 7b
18 — ~ 1000
100
35 — — 810
18.3 29. 5b
126 — — 786
74 — _- X492
18
18.0 1.7C 856
47.0 4.1C
92 — 447
17-56 — 2116
18
17.9 1.65C
28.0-44.0 1.65C 1318
26.5 2.3C 384
a!0 percent S02 in nitrogen
5 percent S02 in nitrogen
C100 percent S02
109
-------
Table D-2
TIME/TEMPERATURE DATA FOR LIMESTONE SULFATION STUDIES
Batch
Particle
Size, ym
Heating,
min at °C
Reacting,
min at °C
Cooling,
min at °C
L-2 (-149 + 74 ym)
Calcination
Re calcination
Sulfation
L-5 (-149 + 74 ym)
Calcination
Sulfation
L-7 (-149 + 74 ym)
Calcination
Sulfation
L-8 (-177 + 44 urn)
Calcination
Recalcination
Sulfation
L-9 (-250 + 44 ym)
Calcination
Sulfation
Resulfation
L-10 (-250 + 63 ym)
Calcination
Sulfation
L-ll (-250 -I- 0 ym)
Sulfation
L-12 (-177 + 0 ym)
Sulfation
45/533-730
45/561-730
300/ 25*835
160/ 25-730
130/515-845
80/473-717
205/602-809
75/594-766
55/556-740
260/551-852
140/ 42-720
159/499-854
213/488-843
215/ 45-735
295/604-893
268/ 28-599
90/ 91-518
70/843-907
185/730-765
145/730-902
105/835-872
125/730-906
120/845-879
205/717-901
198/809-850
145/766-910
105/740-908
124/852-870
145/720-847
194/854-873
77/843-934
1123/735-832
108/832-872
192/893-872
334/599-866
190/518-843
60/890-846
60/765-579
65/902-533
35/872-629
55/906-602
50/879-589
95/901-543
23/850-705
195/910-285
230/908-151
116/870-374
20/847-623
122/873-437
100/848-319
100/690-371
23/872-849
10/846-740
78/846-720
110
-------
Batch L-5 was the second sulfation run. Prior to charging, the
reactor was fitted with a stainless steel ring to block off the middle
ring of holes in the distributor plate. Fluidization was observed by use
of nitrogen. When the gas flow increased, the bed became ratholed but
finally appeared completely fluidized at 91 £/min or 19.3 cm/s. Doubling
the bed depth by adding the balance of the charge of limestone resulted
in apparent fluidization at 50 £/min or 10.6 cm/s. The minimum fluidiza-
tion velocity calculated from the Ergun equation was 0.5 to 1.9 cm/s.
Elutriation was tolerably low. The bulk density of the slumped bed was
1.09 g/cc. The jet velocities through the 22-1.5 mm holes in the distri-
bution plate were 3900 cm/s (128.0 fps) and 2143 cm/s (70.3 fps),
respectively. One explanation for the disparity was that the apparent
fluidization was really multiple jet penetration of the bed for the half
charge, whereas the bed activity for the full charge was characterized by
bubble formation.
On completion of sulfation, the bed was observed to be crusted,
packed, and to have several channels extending the full depth of the bed.
The particle size distribution showed 11 percent +149 ym and 10 percent
-74 \im, neither of which was observed in the original charge. The carry-
over was 0.3 percent of the total reactor product. The level of sulfa-
tion achieved was higher than in L-2, as shown in Table D-l. The +149 ym
fraction contained 37 wt % CaSO, versus about half this amount for the
other fractions. Previous findings suggested that the reverse effect
of particle size should have been obtained. We concluded that the fines
were not effectively fluidized in this run. The nominal gas contact time
based on superficial velocity was 0.24 s.
Batch L-7 also used -149 +74 ym limestone but at a gas flow rate
20 percent higher than in Batch L-5 as an attempt to further improve
fluidization. A pressure build-up during the run to 6.3 psig was
observed. The gas rate was reduced to 10 percent over L-5, the reactor
pressure stabilized at 4.5 psig. When the run was completed, 18 ml of a
water/solids suspension was found in the filter, and the filter surface
111
-------
Table D-3
CHEMICAL ANALYSES OF STONE SAMPLES FROM LIMESTONE SULFATION STUDIES
Batch Ca
L-2
Calcined 39.6
Recalcined 41.9
Sulfated 44.2
L-5
Sulfated
+100
+140 43.9
+200
-200
L-7
Feed 39 . 8
Sulfated 40.9
L-8
Feed 39 . 2
Calcined 46.0
Recalcined
+80
+120 46.9
+170
+230
+325
-325
Sulfated
Composite
+80 41 . 5
+120 46.9
+170
+230
+325
-325
L-9
Feed 40.8
Sulfated 60.8
Resulfated
Composite 40.1
+60
+80 32.9
Component r wt %
C02 SO? Total Sulfur
41.8
36 . 2
24.7 8.9
27.37
14.00 4.26
13.58
14.51
43.4
21.3 — 3.33
43.4
32.2
3.00
22.4 0.62
0.51
0.48
0 . 39
0.99
14 . 5 14 . 3
22.4 23.2
26 . 7
26.1
23.7
23.5
13.1
3.41
51.51 16.53
52.8
1.56 50.3
112
-------
Table D-3 (Cont'd)
Batch Ca
L-9 (Cont'd)
+120
+170
+230
+325
-320
L-10
Sulfated
Top of bed
Edge of bed —
Center of bed
Bottom of bed
Cyclone —
Fitter
L-ll
Sulfated
+60 25.7
-325 31.1
L-12
Sulfated
+60 26.6
-325 27.1
Component, wt %
CO,, SOT Total Sulfur
2 4
0.14 51.6
40.5
27.4
27.5
28.8
36.4
14.1
42.7
43.5
6.4
11.6
1.30 56.0
10.5 32.7
6.10 35.7
1.51 34.0
was nearly all covered with a layer of wet powder. The origin of the
water was not identified. Extensive scaling corrosion was found in the
reactor and was attributed to the effects of the previous sulfidation
run.
In Batch L-8 the initial step was to observe fluidization of the
-177 + 44 urn material. Flow rates of nitrogen at ambient conditions were
used, at which either 1) the entire surface of the bed appeared active, or
2) the bed material was being thrown to the level of the outlet flange.
113
-------
Satisfactory fluidization without excessive elutriation was obtained at
100 fc/min.
The purpose of this run was to determine whether use of a wide par-
ticle size distribution would lead to 90 percent sulfation. We had also
intended to use a hydrator on the air supply, but it had not yet been
fabricated. The 975 g charge used was an arbitrary blend containing
equal weights of -177 + 125 ym, -125 + 88 ym, -88 + 63 ym, and -63 + 44 um
Limestone 1359. Following calcination at 900°C in a flow of 18 i N /min,
the reactor was cooled to ambient temperature and opened for inspection.
We observed that the bed was lightly packed and had a single 5 mm diameter
rathole. Carry-over to the cyclone was 0.2 g.
Nitrogen was used to observe the fluidization character of this
material. At 27 S,/min the rathole became an active crater; at 45 £/min
about 25 percent of the surface was active. We observed no further
change in the quality of fluidization up to the maximum nitrogen flow
available (137 £/min).
The flow of gas was shut off and the bed stirred thoroughly with a
metal rod. We saw no evidence of sintering, and the contents felt like
a bed of granular material. Fluidization was observed again. The first
active crater appeared at 18 £/min; at 100 Jl/min the whole surface was
active. A thin cloud of fines was elutriated, but the bed was still
visible through it.
The reactor contents weighed 835.8 g, corresponding to 34.2 percent
calcination, if one assumes 5 percent inerts. Actual chemical analysis
for calcium yielded 36.3 percent calcination and 3.4 percent inerts. The
inference was that the bed was not adequately fluidized at 18 £,/min. The
superficial velocity was 15.6 cm/s, which was theoretically enough to
fluidize 707 ym particles, whereas the largest ones used were only 177 ym.
The material was recalcined with a flow of 35 a N./min; allowing
for samples, the actual charge of partially calcined material was 809.7 g.
The weight of recalcined material was 738.5 g. After studying the data,
we concluded that the calcium analysis of the recalcined material was
114
-------
probably low by about 7 percent. From the bed weight and the inert
content estimated from the calcine charged, we calculated the extent of
calcination as 50.4 percent. Carry-over to the cyclone and to the filter
was negligible. The particle size distribution of the recalcined mate-
rial was as follows:
Size, urn Weight %
+177
+125
+88
+63
+44
-44
12.4
22.1
23.6
21.6
16.6
3.7
100.0
Fines production in 11.5 hours of fluidization amounted to 3.7 percent of
the product, corresponding to a rate of 0.0032 g/hr/g of final product.
It is of interest that all the fines were retained by the bed. Also,
some particle growth occurred: +177 pm formation was 0.011 g/hr/g final
product.
The reactor was recharged with 653 g of the recalcined stone and
sulfated at 870°C for two hours by using 46.8 £/min of 3.1 percent SO- in
nitrogen. Airflow was continued until the bed had cooled to 600°C.
The bed was cooled further with nitrogen to 400°C, then allowed to cool
overnight to ambient temperature at no flow. The product was a white
powder with no evidence of sintering or packing. Fluidization was
checked again using nitrogen. The bed fluidized readily; a flow of
86 i'/min was needed to keep all of the top surface active, somewhat less
than for the partially calcined stone, suggesting incomplete sulfation.
The rone Lor product was 706 g; carry-over was negligible. The percent
sulfation by calculation was 12.8 percent. Sulfate analyses by size
fraction showed more sulfation: 14 to 27 percent. This increase can
be accounted for by considering the C0_ analysis of the sulfated material
as high. The test showed that adequate fluidization alone was not enough
to achieve the high degree of resulfation desired.
115
-------
Batch L-9 was run next, using 1466 g of the following distribution:
Size, ym Weight %
-250
-177
-125
-88
-63
-44
36.6
30.6
10.4
11.3
11.1
0.0
100.0
Fluidization was observed at half charge and at full charge. The bulk
density of 786 g tapped to constant volume in a 1-liter graduate was
1.59 g/cc. To get the whole surface active with this amount of charge
required 126 £/min of nitrogen. For the full charge of 1466 g, only
74 £/min was needed. At 105 £/min, material was being thrown up to the
level of the outlet flange.
The stone was calcined in 18 £/min of nitrogen at 800 to 850°C.
Significant carry-over occurred:
Material Weight, g
Reactor product 856
Reactor piping 25
Cyclone catch 86
Filter catch 22
989
This corresponds to 78.0 percent calcined. The total carry-over of 133 g
occurred in four hours; the carry-over rate was thus 0.034 g/hr/g product.
The 856 g was charged to the reactor and stored under nitrogen until
the hydrator could be installed. This was pressure tested satisfactorily
with air at 15 psig.
The calcined stone was then sulfated at 870 to 880°C for 195 minutes
with a gas flow of 18 2,/min of air plus 1.7 £/min of pure SO . The con-
centration of S0» was thus 8.6 percent. At the end of this time, airflow
was continued until the bed temperature dropped to 437°C. When the system
116
-------
was inspected, carry-over was found to be negligible. The reactor prod-
uct was white, slightly packed, and showed one large rathole. Its fluidi-
zation behavior was again observed while still in the reactor, with
nitrogen as the fluidizing medium. At 102 £/min, all the surface was
active, but it was clear that one side of the bed below the surface was
inactive. The bed was stirred and the whole bed became active. The
pressure drop across the distributor plate and the bed dropped from 29.5
to 14.8 cm of H.O (11.6 to 5.8 in of 1^0). At 141 £/min, material was
thrown to the top of the 10-cm diameter section. The bed was still active
at a flow rate of 91 £/min, for which the pressure drop was 12.2 cm H?0
(4.8 in). A sample was taken before continuing with the sulfation.
Conditions used were 47 X. air/min plus 3.8 £ 100% SO /min for 25 min-
utes. The bed temperature was at 843°C when the SO was cut in and
2
rose rapidly to 980°C. Temperature thereafter was in the range of 805
to 850°C.
The reactor product yield was 928 g, while cyclone catch plus carry-
3
over in the piping was 112 g. The filter contained 11 cm of slurry with
about 1 cc solids. The reactor product size distribution and calcium
sulfate content were as followsi
Size, pm Weight % Weight %
+250 1.1 74.8
+177 45.5 71.2
+125 42.2 72.4
+88 5.0 57.4
+63 2.0 38.8
+44 2.0 39.0
-44 2.2 40.8
100.0
Thus, about half of the -125 + 88 ym fraction was retained in the
bed while about 80 percent of the fines fraction was elutriated. Sulfate
analyses were very encouraging, although again the large particles showed
significantly higher levels of sulfation. The higher degree of sulfation
in this run was attributed to the use of the hydrator. The moisture
content of the air corresponds to saturation at ambient temperature and
5 psig and is equivalent to 2.4 mol %.
117
-------
Dry sulfation test L-10 was carried out with a modified gas dis-
tributor to obtain improved fluidization. Test conditions were: bed
temperature 870°C, particle size -250 + 44 \im, reaction time 2 times
stoichiometric, and 8 percent SO- concentration. Approximately 1.4 vol %
water vapor was added to the gas.
The time to breakthrough was 53 minutes. The charge was 2117 g of
-250 + 63 ym limestone and contained 40 wt % calcium or 15.10 g moles.
The SO™ feed was 1.65 H/min or 0.0698 g moles/min. The sulfation was
thus 24.5 mol % at breakthrough. The S02 content of the off-gas rose
rapidly (within 20 minutes to the 7 percent level and then over a period
of 2-1/2 hours to the feed level of 8.5 percent). There was still a dis-
crepancy between the S0_ monitor indication and the gas concentration
derived from rotameter readings, one possible explanation for which was
a lag in monitor response.
These results were not encouraging. We therefore decided to return
to tests with smaller particle sizes, since TGA work had shown over
90 percent sulfation could be obtained below 74 um.
As preparation for Batch L-ll, fluidization observations were made
on various particle size distributions in the 10-cm unit. First, 225 g
of -88 + 0 urn Limestone 1359 was charged, producing a shallow bed about
2.3 cm deep. A fines return system was added so that, if elutriation did
occur, the contact time of the fines could be increased by recirculating
them from the cyclone to the reactor. When the system was tested to
the maximum air feed rate available, however, no circulation was observed.
The gas cooler and cyclone contained 21 g of fines, and the downstream
filter had 3.5 g. The carry-over was close to the 25 g of -44 + 0 pm in
the original charge.
Velocity conditions relative to U ,., the minimum fluidization veloc-
ity estimated by the Ergun equation, were as follows:
Mesh size 170 230 325
Particle diameter, ym 88 63 44
Umf 0.67 0.34 0.17
U0/Umf 37 73 149
118
-------
Since the theoretical terminal velocity for the fines was 80 - 93 x U f,
these observations are in accord with predictions.
The total solids were sieved, and, after adjusting the screen analy-
sis to a common basis, we made the following comparison:
Particle Size, ym Feed Product
+88
-88 + 63
-63 + 44
-44 + 0
Loss in sieving
0
106.5
85.2
21.3
213.0
5
70
113
18
206
7
TOTAL 213
The +88 ym fraction found was considered within the variability of sam-
pling and analysis. Some grinding apparently occurred in the -88 + 63 \im
fraction since there was a reduction of 36.5 g (34 percent) along with an
increase in the -63 +44 m fraction of 27.8 g (32 percent). This size
change occurred in the test period of about 30 minutes. An alternative
explanation, which is probably more reasonable, is that sieve analyses
are difficult to make for the small particle sizes, which tend to
stick to the larger particles.
A second try was made using 200 g of -88 + 44 um plus 615 g of
-44 + 0 um to test the hypothesis that there may not have been enough
fines present to permit the return system to become effective. The design
was based on picking up the cyclone catch in a jet of air and conveying
it back to the reactor. Fluidization at ambient temperature was carried
out. When high gas flow rates were reached, a cloud of fines began elu-
triating. The system was closed up and circulation of fines attempted.
Again no circulation was observed, even though the system was heated to
305°C, and the combined air and nitrogen flow was 145 &/min at 15°C, cor-
responding to a superficial velocity of 42.7 cm/s. The ratio U /U was
o ml
119
-------
255 for 44 ym particles. Assuming all the carry-over was -44 ym, only
4 percent of this fraction fed was elutriated. When the reactor was
opened, the bed was observed to have ratholed.
Fluidization was again observed at ambient conditions. At 142 £/min
gas flow, the pressure drop across the distributor and the bed was
(9.7 in) of oil (S.G. 2.95) or 72.7 cm of water. Deducting the bed
pressure drop, estimated at 8.4 cm at a bulk density of 1.3 g/cc leaves
a 61.8-cm water drop across the distributor. This yields an orifice
coefficient of 0.82, which appears reasonable for a submerged orifice.
The jet velocity from the distributor holes was 8250 cm/s (270 fps).
These observations led to the conclusion that a different particle
size distribution was needed to achieve fluidization at reasonably low
gas velocities while retaining fines in the bed. Accordingly, Batch 11
was prepared by adding 528 g of -250 + 88 ym stone to the reactor.
Fluidization was observed at ambient conditions. Fluidization was con-
sidered achieved at an airflow of 80 £/min, corresponding to a super-
ficial velocity of 17.8 cm/s. This is about 3 x U . for 250 ym particles
ml
and 107 x U f for 44 ym particles. No significant elutriation occurred.
Increasing the gas flow to 35 cm/s resulted in elutriation: this was at
209 x U f for 44 ym particles. The bed was thus acting as a sand filter
to retain fines at gas velocities twice the terminal velocities.
The stone was then sulfated at 630 to 865°C with a nominal 8-percent
S0? in air for 5 hours, 34 minutes. At the completion of the run, the
bed was found to be caked. The bed material was sieved. Lumps that could
not be broken easily by hand were included in the +250 ym fraction.
Mesh Size, ym Feed Reactor Product
+250
+177
+125
+88
+63
+44
-44
0
213
181
134
87
113
590
224
242
290
232
152
139
134
1318 1313
120
-------
There was no carry-over to any of the downstream equipment. Superficial
velocity at maximum bed temperatures was 18.0 cm/s, which may be com-
pared with the U f of 44 ym particles of 0.067 cm/s at 86.5°C. Chemical
analysis showed the +250 ym fraction contained 56.0 wt % sulfur as SO,,
while the -44 ym fraction had 52.7 percent. If one assumes pure lime-
stone and complete calcination of unsulfated stone, these figures corre-
spond to 79.4 and 74.7 mol % sulfation to CaSO,, respectively. Again,
the C0_ analyses seem discrepant, but sulfation appears to have been at
least 70 mol % when CaC03 was taken into account. The higher sulfation
compared to previous results was encouraging relative to lime utilization,
but the caking of the bed was a negative factor. Fluidizing velocities
would be maintained for a longer portion of the cooldown period in sub-
sequent runs.
To improve fluidization further, Batch 12 was run with a relatively
smaller content of -44 ym - namely, 7 percent. Even at 106 £/min, how-
ever, the bed was not completely fluidized, although fines were being
elutriated. Although we felt that heating to reaction temperature
would result in satisfactory fluidization, on the basis of observations
thus far, we arbitrarily adjusted the bed composition by adding 96 g
(25 percent) of -250 + 177 ym stone, bringing the total charge to 480 g.
The gas flow rates used corresponded to about 3 x U for the largest
particles in the bed (250 ym) and 95 x U .. for the 44 ym. After sulfa-
mr
tion the bed was again found to be packed, crusted, and ratholed. The
top flange was coated with 0.6 g of a greenish-yellow powder. This was
subsequently found to be insoluble in warm carbon disulfide (CS~) . Carry-
over was negligible, and the reactor product weight was 490 g.
Chemical analyses showed relatively low sulfate content with
essentially no difference between the +250 and -44 ym fractions. Calcium
was also low, however, so sulfation was again better than 73 mol %.
Before proceeding to tests with actual CAFB spent sorbent, fluidiza-
tion of this material was studied in more detail, as reported in another
section of this report (Appendix C).
121
-------
A summary of the test data with CAFB regenerator stone is shown in
Table D-4 and product analyses are in Tables D-5 and D-6. The first run
was CAFB-701 and used a feed with a normal distribution (mean 130 ym,
standard deviation 50 ym). This showed that the usable range of velocities
for the 10-cm reactor was a four-fold range from that needed to create
uniform fluidization as judged by the appearance of the bed to that at
which either elutriation was judged excessive or particles were being
ejected to the level of the outlet flange. At the 11.4 cm/s superficial
velocity required for minimum fluidization, the gas contact time for a
15-cm (5.9-in) bed at 50-percent voids was 0.8 seconds. To increase
this, a smaller mean particle size was required. As the particle size
decreases, however, the flow character of the CAFB stone changes. The
-177 + 125 yra fraction was observed to be free flowing, the -125 + 88 urn
fraction was somewhat sticky, and the -88 + 63 ym fraction was definitely
sticky. The -44 ym fraction readily compacted to a nonflowing mass. We
decided to modify the particle size distribution before proceeding with
the sulfation.
CAFB-702 was a modified blend of particle sizes to increase the
amount of +125 ym from 53 to 64 percent, reducing the -125 + 44 ym from
42 to 31 percent and retaining the -44 ym at 5 percent. The distribution
was still normal, with a mean of 147 ym and a standard deviation of 62 ym.
This change reduced the U , by about 10 percent, but elutriation was
mi
noticeably greater at the maximum flow rate. The flow rates increased
by about 10 percent. Using a typical analysis for CAFB-7 regenerator
stone, the degree of sulfation was calculated from the total sulfur con-
tent as 44 to 68 mol %, with a trend toward higher sulfation for the
smaller particle sizes.
An incidental result was obtained toward the end of the run. Drop-
lets of liquid were observed leaving the final vent on the apparatus.
The liquid was viscous and acid, as evidenced by the corrosion of the
aluminum countertop where some of the droplets had deposited. Inspection
122
-------
Table D4
TEST CONDITIONS FOR CAFB REGENERATOR STONE RUNS
D«q. 168^809
!
1 Run
I Number
i
Charge
Weight, g I Size, urn
Gas Flow Rates (i/min at 15°C, 1 dtm)d
N2
Air
so2
Total
Superficial
Gas Velocity,
Reactor Time/Temperature Conditions
Heating
Reacting
Cooling
NJ
U)
CAFB-701 Fluidization observations only
CAFB-702
CAFB-903
Phase I
Phase II
Phase III
CAFB-904
Phase I
Phase II
Phase III
Phase IV
CAFB 905
Phase I
Phase II
Phase III
CAFB 906
Phase I
Phase II
250
500
500
-420+0
-420+0
-420+0
20 1.74 21.74
19.7
44 mins at 63-404°C 542 mins at 404-904°C 52 mins at 616-577°C
1975 -354+0
1000 -500+0
32
30
20
6.6
21
21
27.5
27.5
27.5
20.1
44
17.1
1L4
2.53
L10
L10
L45
1.54
L54
1.08
2.3
0.9
0.6
1.5
1.0
41.13
22.10
22.10
28.95
29.04
29.04
21.18
46.3
18.0
12.0
31.5
21.0
23.0
18.3
W.3
25.6
24.1
25.7
18.7
25.9
14.2
10.4
22.5
17.4
65 min
144 min
150 min
109 min
177 min
291 min
53 min
184 min
148 min
190 min
145 min
187 min
at 260
at 457
at 190
at 221
at 118
at 518
at 516
at 105
at 380
at 370
at 308
at 397
-404°C
-738°C
-699CC
-654°C
-640°C
-813°C
-627°C
-452°C
- 752°C
-846°C
- 655°C
-795°C
165 min
91 min
94 min
339 min
290 min
171 min
280 min
243 min
244 min
221 min
277 min
109 min
at 404
at 738
at 699
at 654
at 640
at 813
at 627
at 452
at 742
at 845
at 643
at 791
-460°C
-854°C
-827°C
-871°C
-832»C
-895°C
-899°C
-457°C
-758°C
-853°C
- 655°C
-808°C
Over night to 190° C
224 min at827-127°C
30 min at846-638°C
34 min at 832-651°C
Overnight to 516°C
70 min at899-570°C
37 min at450-340°C
88 min at 749-365°C
83 min at850-485°C
84 min at 647-352°C
145 min at803-222°C
a 1 atm = 101.32501 kPa
-------
Own. 1705887
Table D-5
CHEMICAL ANALYSES OF STONE SAMPLES FROM CAFB SULFATION STUDIES
Composition, wt%
Batch
CAFB -702
Feed
Sulfated Product, urn
1-420
1-250
1-177
1-125
1-88
1-63
1-44
-44
Filter liquid
CAFB -703
Feed
Sulfated
CAFB -904
Feed
Sulfated I
Sulfated II -IV, urn
1-420
1-250
1-177
1-125
1-88
1-63
f44
-44
S= S04=
—
—
45.1
-
—
—
-
-
—
1092gH2S04/*
—
40.2
-
39.7
63.5
49.5
60.0
65.9
68.8
68.2
58.5
58.1
54.2
Total Sulfur
__
15.4
15.9
16.2
16.5
18.4
19.7
18.3
18.5
—
—
—
—
—
—
—
—
—
—
—
—
—
124
-------
of this liquid showed that it contained 1092 g H2SO,/2, and had a specific
gravity of 1.62, corresponding to 67 percent acid. This finding was
explored further in the subsequent runs.
In CAFB-903, the amount of -44 ym particles was doubled to 10 per-
cent. Ths distribution had the same mean, but the standard deviation was
increased to 80 ym. Compared to CAFB-702, the charge had fewer particles
in the midrange and more at both ends. Elutriation was found to be
appreciable at a considerably lower velocity than in the previous run and
the range of operable flow rates in the test reactor appreciably reduced.
Sulfation was carried out in two phases, first at 450°C and then at
700 to 850°C. No significant pickup of S0_ was indicated by the monitor
at 450°C. Although this temperature is favorable for the conversion of
S0? to S0_, whatever mechanism was responsible for acid production in
CAFB-702 was clearly not operating in this run. Actual sulfation finally
achieved was 57.0 wt % CaSO,, appreciably lower than in the previous run.
The reactor product was granular, although a few small lumps were present.
The difference in degree of sulfation was attributed to the difference in
reaction time: 542 minutes in CAFB-702 versus 350 in CAFB-903.
CAFB-904 was a variation on CAFB-903. We hypothesized that the
sulfated limestone was displaying catalytic activity in converting SO- to
SO . The revised approach was to sulfate the stone partially at 870°C
and then explore absorption/conversion of S0_ at 450°C. Particle size
distribution was the same as in CAFB-903. The reactor product from
Phase I contained 56.3 wt % CaSO,. There was evidence that gas bypassing
may have occurred during part of the Phase I test.
Phases II through IV were concerned with maximizing the degree of
sulfation at 650 to 900°C. Total exposure time to 5 percent S02 was
IS hours for the run. The S02 monitor showed some pickup of SO through-
out most of this exposure. Overall, the reactor product contained
90.0 wt % CaSO,. Maximum sulfation was in the -177 + 125 urn and the
-i:3 + 88 urn fractions (97.6 and 96.6 wt % CaSO,, respectively). The low-
est sulfation was in the +420 urn fraction (70.1 percent) and in the -44 pm
125
-------
Dwo. 1705688
Table D-6
CHEMICAL '\ALYSES OF STONE SAMPLES FROM CAFB SDLFATIOTS STUDIES
CAFB -905
Measured
S=
S04=
Calculated
CaS
CaSO,
BET
Surface Area,
mZ/g
Feed
-500 4- 177
-171 4-88
-88
Sulfation I
450°/0 min
-500 4- 127
-1774-88
-88
450"/15 min
-500 4- 177
-177+88
-88
450°/75 min
-500 4- 177
-177+88
-88
450°/135 min
-500 4- 177
-177 4-88
-88
450e/255 min
-500 f 177
-177 4-88
-88
Cyclone Catcti
-500 4- 177
-177+88
-88
Filter Catch
-500 +• 177
-1774-88
-88
1.94
0.31
1.85
1.97
1.90
1.95
1.09
1.59
0.81
1.49
1.14
0.78
0.97
1.09
0.70
1.45
1.13
1.51
0.52
0.64
0.98
0.21
0.69
0.62
4.08
7.15
4.23
3.79
3.64
3.94
5.16
4.98
6.96
6.07
6.07
6.36
6.47
6.89
8.29
7.77
7.47
11.35
4.36
0.70
4.16
4.43
4.27
4.30
2.45
3.58
1.82
3.35
2.56
1.76
2.18
2.45
1.58
3.27
2.54
3.40
1.18
1.43
2.21
0.47
1.55
1.40
5.78
10.13
6.00
5.37
5.16
5.58
7.31
7.06
9.86
8.60
8.60
9.01
9.17
9.76
11.74
11.01
10.59
16.08
2.79
2.71
10.71
3.80
6.63
7.98
1.98
3.99
8.39
2.11
3.00
5.78
1.60
1.90
4.93
0.943
1.354
4.97
9.96
12.43
7.52
126
-------
Dwo. 1705B89
Table 0-6 (Cont'd)
CAFB-905
Measured
S=
so;
Calculated
CaS
CaS04
BET
Surface Area,
m2/g
Sulfation 11
750% min
-500 * 177
-177 f 88
-88
750 °/ 15 min
-500 * 177
-177*88
-88
750°/75 min
-500 * 177
-177*88
-88
750°/135 min
-500 * 177
-177*88
-88
750°/255 min
-500 f 177
-177*88
-88
Cyclone Catch
-500 * 177
-177*88
-88
Filter Catch
-500 * 177
-177*88
-88
Sulfation III
850°/0 min
-500 * 177
-177 *88
-88
1.52
1.65
2.26
1.43
1.54
1.39
1.28
1.33
NS*
1.30
1.19
NS
0.97
1.14
NS
0.79
1.80
1.81
1.06
1.12
NS
10.13
9.89
15.57
13.86
14.43
19.14
21.17
21.80
25.31
25.06
25.49
30.88
29.65
32.29
10.44
31.51
21.45
11.77
17.44
15.29
29.08
30.36
42.38
3.42
3.71
5.08
3.22
3.46
3.13
2.88
2.99
NS
2.92
2.68
NS
2.18
2.56
NS
1.78
4.05
4.07
2.38
2.52
NS
14.36
14.02
22.07
19.64
20.45
27.12
30.00
30.90
35.87
35.02
36.12
43.76
42.02
45.76
14.80
44.66
30.40
16.68
24.72
21.67
41.21
43.03
60.06
1.75
2.01
0.996
1.227
NS
NS
6.96
NS
11.73
6.30
1.017
2.830
•NS = Not Sufficient Sample Quantity
127
-------
Dw". 1705B90
Table D-6 (Cont'd)
CAFB 905
Measured
S=
so4=
Calculated
CaS
CaS04
BET
Surface Area.
m2/g
850°/15 min
-500 + 177
-177+88
-88
850°/60 min
-500 + 177
-177 +88
-88
850°/120 min
-500 1 177
-177 +88
-88
850°/239 min
-500 + 177
-177+88
-88
Reactor Product
+ 500
-500 + 354
-354 + 177
-177+88
-88
Cyclone Catch
-500 + 177
-177+88
-88
Filter Catch
-500 + 177
-177+88
-88
CAFB 906
Feed
Sulfation I
6508C/280 min
-500+354
Reactor Product
0.95
1.09
NS
1.03
1.04
NS
0.95
1.06
NS
0.88
1.00
NS
0.60
1.13
0.88
0.%
0.73
0.46
0.57
0.71
-
-
—
30.80
32.75
37.11
31.31
—
39.37
32.87
35.75
NS
35.46
39.75
NS
52.50
37.14
34.72
39.78
31.52
61.94
42.42
13.40
—
-
—
2.14
2.45
NS
2.32
2.34
NS
2.14
2.38
—
1.98
2.25
—
1.35
2.54
1.98
2.16
1.58
1.04
1.28
1.60
—
-
—
43.65
46.41
52.59
44.37
—
55.80
46.58
50.67
NS
50.26
56.34
—
74.40
52.64
49.21
56.38
44.67
87.78
60.12
18.99
—
-
—
_
—
-
0.955
0.961
-
—
-
-
0.505
4.608
—
1.70
1.41
1.06
0.90
1.98
17.88
17.40
8.52
31.94
27.00
~
Measured
S=
3.28
3.73
so;
-
-
Total
Sulfur
7.64
9.44
Ca
58.67
53.87
co2
8.03
10.3
120
-------
fraction (76.8 percent). This run was taken as evidence that when fluidi-
zation is achieved, the calcium can be sulfated to essentially 100 percent
by contact times of the order of 20 hours. The same result was achieved
in a 2.5-cm fixed bed.
CAFB-905 was intended to collect additional information to clarify
the effect of particle size distribution in fluidization and to correlate
the progress of sulfation in a fluidized bed with surface area and pore
volume measurements made on the same fractions. As a preliminary, exten-
sive observations were made in an existing atmospheric pressure 7.6-cm
(3-in) Plexiglas column on the behavior of -420 + 125 urn spent CAFB
sorbent to which were added various proportions of -63 + 0 ym powder.
These showed that the addition of fines
• Increased the pressure energy needed to maintain a fluidized
bed
• Increased the maximum bed pressure drop before fluidization
was achieved
• Produced a larger pressure drop at a lower superficial gas
velocity.
The transition from a fixed bed to a fluidized bed was also less smooth
and occurred over a wider range of gas velocities. With a bed containing
one-third of its weight as -44 + 0 ym particles, for example, the transi-
tion occurred over the range of 2.5 to 11.0 cm/s versus 7.5 to 8.0 cm/s
for a bed without these fines.
Sulfation was carried out in three phases corresponding to the reac-
tor temperatures of 450, 750, and 850°C. Samples were collected at five
reaction times and were analyzed for sulfur content, both sulfide and
sulfate, BET surface area, and pore volume as a function of particle
size. Observations were made on sorbent carry-over. A summary of sor-
bent distribution is given in Table D-7. Figures D-l through D-3 show
the time-temperature curves.
In the meantime, the significance for dry sulfation is that the
design previously worked out for the absorber provides more than enough
129
-------
Table D-7
MATERIAL BALANCE DATA TOR TEST CAFB-905
OKI. niCflZ
Particle Size Distribution, g
Sample
m 9
Phase 1 Initial Conditions
Reactor Charge 25. 1
Phase? SulfationatWC
Sample Tube Cleanings 25.1
Initial Sample 25.1
1$ Minutes 26.5
75 Minutes 26.7
1)5 Minutes 31.7
251 Minutes 29.3
Sample Purges 36. 8
Cyclone Catch 25.8
Filter Catch O 18-6
Phase} Su««lonat750BC
Initial Sample 30.7
IS Minutes 3317
75 Minutes 34.1
135 Minutes 445
259 Minutes 46.8
Sample Purges 40.0
Cyclone Catch 24.4
Filter Catch 3.4
Phase 4 SulMoniteO*C
India) Sample 420
15 Minutes «3
M Minutes 46.8
120 Minutes «'
ZWMInutei 44.8
Sample Purgej 78.7
Cyclone Catch 216.9
Filter Catch 2.1
Tubing Catch 10.2
Reactor Product »B. 5
Total Sample Weight Recovered
0. 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
13
0.0
0.0
0.0
0.0
0.0
0.0
0.0
a?
0.0
0.0
0.0
°-°
0.0
0.0
9.9
ftl
*•*
68.0
-35*45 -45+80 -80 + 170
1.1 20.0 1.8
as
1.1
0.9
0.2
0.9
0.7
16.7
19.1
20.6
21.7
2S.6
24,6
3.8
it
It
3.0
3.6
31
This sample was not sieved
0.030
0.1
0.9
1.0
0.9
1.1
1.1
1.1
0.021
0.027
1.0
1.1
1.5
1.3
2.3
2.9
1.6
0.1
as
45.0
0.1
0.1
26.8
29.2
34.Z
38.8
40.9
34,3
0.064
0.045
36.6
35.1
418
37.6
38.9
68.2
16
0.2
1.4
97ZO
2.8
4.3
29
3.3
4.0
46
48
41
0.2
0.9
44
4,1
45
4.2
3.6
7.6
42
0.8
2.1
147.5
-i70+a
2.2
3.8
2.4
it
1.8
16
0.7
tt.1
lit
0.1
0.2
0.042
0.014
0.022
0.5
24.2
t2
0.021
0.0)1
0.012
0.004
0.012
a 026
1916
a9
il
86.0
Total
25.1
245.6
262.6
18314
2371.7
Notes
1.
2.
U. S. sieve mesh sizes
Includes about 3g of material found in inlet line lo filter
130
-------
500
£ 400
i
I 300
£ 200
100
0
SO.on
Reaction Time, hr
Figure D-l - Dry Sulfation of CAFB-9 Regenerator Stone at 450°C
800
700
& 600
E
o>
o 500
400
300
Phase II
Cumulative Contact
Time »ith S0?, min. -
0
15
75
135
255
26
28 30 32
Reaction Time, hr
36
Figure D-2 - Dry Sulfation of CAFB-9 Regenerator Stone at 750°C
ai
E
900
800
700
600
500
400
300
n
Phase III
Cumulative Contact -
Time with S02, min.
0
15
60
120
239
76
78 80
Reaction Time, hr
82
84
86
Figure D-3 - Dry Sulfation of CAFB-9 Regenerator Stone at 850°C
131
-------
residence time (VI00 hours versus 18) for the solid. The conclusion sug-
gests that the regenerator stone may not have to be ground, thus permit-
ting elimination of the jet pulverizer and the associated air compression
cost.
One further run was made, CAFB-906, to explore the reactivity of
CAFB stone to dry S02 versus S02 in moist nitrogen. Previous TG results
had shown that fresh limestone would pick up substantial amounts of S02>
The reactor was charged with 1000 g of -500 um CAFB-9 stone obtained
by a combination of sieving out the -500 ym fraction from the stone as
received and grinding +500 urn to accumulate the desired charge.
Fluidization behavior was observed at atmospheric pressure using
air, as shown in Figures D-4 and D-5. Sulfitation was then carried out
in two phases. In Phase I, the charge was heated to 650°C and then
treated with 5 percent S02 in nitrogen, with about half the nitrogen
being fed through the hydrator. Samples were taken at time intervals as
shown on Figure D-6. The reactor was cooled to below 300°C overnight.
In Phase II the reactor temperature was 800°C; sampling was at similar
intervals as shown on Figure D-7. After we had cooled the sulfited
product to ambient temperature, we observed its fluidization behavior,
using dry nitrogen as shown in Figures D-8 and D-9.
Table D-8 gives material balance information by particle size.
Table D-9 shows that a slight change in distribution occurred as a result
of attrition and sulfidation. Some particle growth occurred in the
+500 ym fraction, and a reduction in the percent of -500 + 354 um can be
noted.
Chemical analysis of the reactor showed only 9.4 wt % total sulfur,
essentially the sulfite sulfur, and 3.73 wt % of sulfide sulfur. Allow-
ing for the disproportionation of calcium sulfite into sulfate and sulfide,
we can calculate a theoretical S02 pickup. The analytical results, how-
ever, were not sufficiently consistent to permit this. If we use a typi-
cal analysis for the regenerator stone, as in the 1975 Annual Report,
the 9.4-percent sulfur works out to an 18-percent conversion of the
132
-------
Curve 716512-fl
O
•
a 4
CO
<§)
o
o 2
CO
o>
.c
o_
<
0
III II II
Low Flow Range
1 Active Crater
30
o
CVJ
20o
E
10
0
Q.
<
24 6 8 10 12 14 16 18 20 22 24
Airflow Rate, ^/min @760 mm Hg, 15°C
Figure D-4 - Fluidization Behavior of CAFB-9 Regenerator Stone,
Rim HAFR-906. prior to Sulfitation
O
*
O
a.
00
<§)
9
8
7
6
5
4
3
to
CD
1 2
1
0
T I I 1 I
High Flow Range
2 Active Craters
tiii
j L
60
50
o
CM
40 ±
O
30 g
20
10
0
10 20 30 40 50 60 70 80 100 120
Airflow Rate, l/min@760 mm Hg, 15°C
Figure D-5 - Fluidization Behavior of CAFB-9 Regenerator Stone,
Run CAFB-906, prior to Sulfitation
133
-------
Curve 716514-A
Reactor Temperature, °C
1 1 1 1 1 I 1 1
Phase I
Sample
Number
0
1
2
3
S02on 4
O^^^-k-O-rt ^ -f^
Cumulative Contact
Time with S02, min.
0
30
60
120
280 S09off
r •
v /FTtr^ i\ -
/ 0 1 2 3 4 \ -
^ J\ \ \ 1 1 II 1
16 17 18
19 20 21 22
Reaction Time, hr
23 24
Figure D-6 - Absorption of S02 on CAFB-9 Stone in Absence of Air
O
o
0>
(
E
o>
o
TO
0>
cc.
CumulativeContac
Time with $03, mia
42 43 44 45
Reaction Time, hr
47
Figure D-7 - Absorption of S02 on CAFB-9 Stone in Absence of Air
134
-------
Curve 716513-A
O
o
a
to
CVJ
<§)
O
*o
o>
JZ.
o
c
a.
<
3
2
I
0
i \
Low Flow Range
30
20
10
o
E
o
2 4 6 8 10 12 14 16 18 20 22 24
Nitrogen Flow Rate, ^/min@760 mm Hg, 15 °C
0
Figure D-8 - Fluidization Behavior of CAFB-9 Regenerator Stone,
Run CA^B-POS, after Sulfitation
O
•
o
8.
00
cv
©
O
"o
«/»
CD
JI.
U
c
Q_
<
8
7
6
5
4
3
2
1
0
High Flow Range
1 Active Crater
I
J I
i I I I
I I
70
60
50
40 i
o
£
o
30
10 20 30 40 50 60 70 80 90 100 120
Nitrogen Flow Rate, ^/min@760 mm Hg, 15°C
20 <
10
0
Figure D-9 - Fluidization Behavior of CAFB-9 Regnerator Stone,
Run CAFB-906, after Sulfitation
135
-------
Table D-8
MATERIAL BALANCE DATA FOR TEST CAFB-906
Sampl
Phase I Sulfitation at 650°C
Initial sample 27.
30 minutes 28.
60 minutes 38 .
120 minutes 45.
280 minutes 37.
Subtotal 178
Phase II Sulfitation at 800°C
Initial sample 28
30 minutes 38
60 minutes 36
120 minutes 34
Subtotal 137
Combined Sample Purges 73
Reactor Product Sample 25
Reactor Product 745
Cyclone Catch 1
Filter Catch 0
Line Plugging 1
Total output 1163
e Particle Size
+35 +45 j +60 | +80
0 1.1 15.8 5.7 2.6
9 1.0 16.6 6.1 3.0
7 1.6 22.2 8.4 4.0
9 1.8 25.4 10.0 5.1
5 1.9 20.4 7.7 4.3
0
.5 1.7 15.8 5.8 3.0
.6 2.1 21.3 8.1 4.1
.0 1.1 20.3 7.7 4.0
.8 1.9 19.5 7.2 3.7
.9
.6 2.9 44.2 16.1 6.7
.0 Not sieved
.6 23.1 379.9 147.5 81.3
.9 Not sieved
.0
.6 Not sieved
.6
Distribution, ga
+120
1.3
1.6
2.0
2.7
2.3
1.6
2.2
2.2
1.9
2.9
67.2
+170
0.4
0.5
0.4
0.8
0.8
0.5
0.6
0.6
0.5
0.7
39.1
+230 1 +325
<0.1 0
<0.1 0
<0.1 0
<0.1 0
<0.1 0
<0.1 0
<0.1 <0.1
<0.1 0
<0.1 0
<0.1 0
6.0 1.0
-325
0
0
0
0
0
0
0
0
0
0
0.5
U. S. sieve mesh sizes
-------
Table D-9
EFFECT OF SULFITATION ON PARTICLE SIZE DISTRIBUTION
OF CAFB-9 REGENERATOR STONE
u. s.
Screen
Size
Reactor Charge
Weight. R I Wt %
Phase I
Initial Sample
Weight, el Wt %
Reactor Product
Weight, e 1 Wt %
25 + 35
35 + 45
45 + 60
•60 + 80
•80 + 120
•120 + 170
•170 + 230
230 + 325
•325
0.0
587.8
212.4
103.6
75.4
38.1
6.1
1.2
0.4
0.00
57.34
20.72
10.11
7.35
3.72
0.60
0.12
0.04
1.1
15.8
5.7
2.6
1.3
0.4
<0.1
0.0
0.0
4.07
58.52
21.11
9.63
4.82
1.48
0.37
0.00
0.00
23.1
379.9
147.5
81.3
67.2
39.1
6.0
1.0
0.5
3.10
50.95
19.78
10.90
9.01
5.25
0.81
0.13
0.07
1025.0
100.00
27.0
100.00
745.6
100.00
available CaO. As a check, the sulfide sulfur was 2.9 percent versus the
3.73 percent found. For comparison, in the same contact time of about
six hours, sulfitation in air (sulfation, i.e.) proceeded to a level at
least twice as high. Apparently, the presence of oxygen enhances the
ability of SO- to penetrate the spent stone particles.
In contrast earlier work had shown that much higher sulfur burdens
could be achieved with fresh limestone, an indicator of the reduced chem-
ical reactivity of CAFB regenerator stone but also suggesting that in
sulfation the oxygen present may have more than a stoichiometric role.
A sample of CAFB-9 stone, therefore, was tested in the TG apparatus
for reaction of CaO with SO- in the absence of oxygen. In 0.5 percent
S02/N2, the reaction was sluggish and corresponded to 9.7 percent of
stoichiometric for formation of CaS'3CaSO,, in 70 minutes at 800°C.
Raising the temperature to 825"C had little effect on the rate of reaction.
Previous TG tests had shown that fresh CaO reacts strongly with SO in
the absence of air.
137
-------
These test results overall support the conclusion that dry sulfation
is technically feasible. An optimum temperature for the dry sulfation
of small particles (<74 ym), however, has not been shown. The model pre-
sented in the March 1975 report seems to be applicable to both fixed and
fluidized beds in the case of small particles but may not be applicable
to a fluidized bed in the case of large particles.
ANALYSIS OF DATA FROM CAFB RUNS
Figures D-10 through D-12 present the basic information collected in
CAFB-905 in terms of weight percent CaSO^ versus run time for three size
fractions of spent stone. Some sulfation, 5 to 10 wt %, occurred even at
450°C, but most of it occurred at 750°C (about 30 wt %). A further incre-
ment was obtained on raising the temperature to 850°C (5 to 10 percent).
The maximum sulfation was obtained in the +500 urn fraction (about 75 wt %),
while the -88 urn fraction showed only 45 percent sulfation. The middle
fraction, -500 + 88 ym, contained 49 to 56 percent CaS04.
Figure D-13 shows the changes in BET surface area obtained for three
size fractions during sulfation. The BET areas for the two larger frac-
tions (-500 + 177 and -177 + 88 ym) were about the same, whereas the area
for the -88 ym fraction was, at least initially, larger by a factor of
more than 2. It is likely that the area declines gradually over the
period covered by Phase II. It is also possible that, for the -88 ym
fraction, the BET area in Phase I declines to about the level of 2 m /g,
as for the other two fractions. The sulfation obtained in Phase I,
which was no more than about 5 percent of the calcium, was associated
with a three-fold reduction in BET area. Despite the substantial sulfa-
tion that occurred in Phase II, there did not appear to be a comparably
large reduction in BET area. Continued sulfation as in Phase III appears
to be associated with a resumption of the decline in BET area.
For comparison with previous work done by others, the March 1975
Annual Report, Vol. II, page 280,D1 showed data on the sulfation of CAFB-7
stone ground to less than 148 ym (100 mesh). In 5 percent S02, TG
138
-------
100
80 -
60
re
O
20
II
10
20
_L
J_
30
40 50
Run Time, hrs
III
70 80
90
Figure D-10 - Dry Sulfation of -500 + 177 pm Fraction of
CAFB-9 Regenerator Stone
100
80
Curve 69?6i'3-A
•2
"5
60
* 40
2D
II
J_
_L
III
10
20
30
40 50
Run Time, hr
60 70
80
Figure D-ll - Dry Sulfation of -177 + 88 \im Fraction of
CAFB-9 Regenerator Stone
139
-------
JUU
80
1
" 60
e
3
S
o
5 40
°oj
20
^
C
.0. 8
1 6
fc
CD
2
0
, , , , , , ,.,..T_...T, , ., ^ , ( n ( ( 1 —
F
MT
i i , i , i , i , |
10 20 30 40 50 60 70 80
Run Time, hr
Figure D-12 - Dry Sulfation of -88 urn Fraction of CAFB-9
Regenerator Stone
i.urvp >.^.'t ;;-.
"II I 1 II! 1 1 1 II 1 1 1 1 I
\
•\_U ^^ ^L_ -
\\ U. S. Sieve Size
\ \S -170
\
'X XX HO + 170 \^
^'^V. — \V""^--*^~"~"* — — nCT x . on \^ ~
^^•^-— ^— n\ — ^^^^ — -* KVi —35+80 ^s^
^v, — " V \\ .^^^
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
_
90
0 2 4 6 8 10 24 26 28 30 32 34 70 72 74 76 78 80 82
Run Time, hr
Figure D-13 - Effect of Dry Sulfation on BET Surface
Area of CAFB-9 Regenerator Stone
140
-------
sulfation experiments showed two hours was sufficient to achieve 95 per-
cent sulfation for -149 + 125 pm particles at 920°C or for -74 pm par-
ticles at 825°C. No more than about 5 minutes was required for 35 percent
sulfation (56 wt % CaSO,). In contrast, the results from the 10-cm
fluidized unit show the increase in CaSO, from 14 wt % to about 46 wt %
required 255 minutes at 750°C for -177 + 88 pm particles. This corre-
sponds to a conversion of CaO from 6.5 to 26.0 mol %. Thus, there is
a discrepancy of a factor of 50 in reaction rate to be rationalized.
First, the 10-cm unit results do not appear to have been limited by
stoichiometry. At 5 percent S0? in air, the oxygen available was 8 times
theoretical. The total SO- supplied was 1.27 times theoretical. About
18 percent of the supplied SO. was absorbed. This amount was determined
by noting that the CaSO, content of the final product was about 53 wt %
more than initial value.
Next, the 1 to 5 percent conversion of calcium at 450°C might have
been responsible for the low reaction rate. The BET data, strictly
speaking, show that the number of sites capable of adsorbing nitrogen
was greatly reduced in Phase I. Although this interpretation is consis-
tent with the inference that chemical reactivity for SO /O. was also
reduced, the BET area was essentially constant in Phase II, when most of
the sulfation occurred. Thus, one sequence is initial reduction of avail-
able pores, either by pore blockage or by lining the internal pore area
with a less reactive CaSO layer. Calculations show that a very small
amount of conversion is sufficient to create a monolayer of CaSO,, so the
first phase of sulfation at any temperature can be seen as converting all
surface CaO to CaSO,. Thereafter, SO- and oxygen must diffuse through
this and subsequent layers of CaSO to achieve the demonstrated levels of
conversion. The constancy of the BET values in Phase II supports this
view and thus implies that the reaction with subsurface calcium leaves
the surface area essentially constant over a considerable portion of the
reaction. We consider it unlikely that the observed effects can be
explained by blockage of pore mouths except, perhaps, in the later stages,
as in Phase IV.
14.1
-------
Another possibility is inadequate fluidization or, more basically,
inadequate contact between the fines and the gas. Such failure of con-
tact could happen if segregation occurred in the bed so that the bulk of
the fines was out of the gas path or if fines were elutriated to reduce
their concentration in the bed. The final product had 6.5 percent of
-88 ym material, so at least half of the original fines charged were
elutriated. Also, the fines content of the samples withdrawn during the
course of the test began at about 10 percent of -88 pm and dropped to
less than 0.1 percent. This also resulted in very small samples for
chemical analysis, although we do not believe that the sulfate contents
found are unrepresentative. Therefore, despite the evidence that fluidi-
zation occurred, it does appear likely that the combination of elutria-
tion and segregation could be responsible for the low sulfation of the
fines. The low sulfation of the coarse particles is in line with the
previous work and the literature.
Superficial gas velocities in each of the three phases are summar-
ized in Table D-10. We had found previously that in some cases velocities
to 200 times minimum fluidization velocity were needed to elutriate
fines. Only the -44 Mm particles at 450°C and 750°C met or surpassed
this criterion. More than 90 percent of the +88 pm material was retained
in the bed. Thus, to retain -44 pm fines in a bed of -354 + 125 pm par-
ticles requires the superficial gas velocity at 750°C to be less than
13.6 cm/s, whereas to fluidize -354 pm particles would require it to be
4.5 cm/s.
Further analysis of the data was accomplished by use of the shrinking
core model for reaction of a gas with a nonporous solid developed in the
literature.2 Several assumptions are required.
1. Particles remain spherical during the reaction.
2. Reaction occurs only at the boundary between the reaction
product and the unreacted core.
3. Temperature is uniform.
4. Density of solid remains constant.
342
-------
5. There is no gaseous region between the reaction product and
the unreacted core.
6. The diffusion rate of the gaseous reactant is large relative
to the rate of decrease of unreacted core radius.
Table D-10
GAS VELOCITIES IN THE DRY SULFATION OF REGENERATOR
STONE IN RUN CAFB-905, cm/s
Phase
I II
III
Temperature Level, °C 450
U , Superficial Velocity 24.66
U ,., Minimum Fluidization Velocity
Particle Size, ym
44 (325 mesh) 0.0896
88 (170 mesh) 0.358
354 (45 mesh) 5.77
500 (35 mesh) 11.40
750
13.56
0.0701
0.286
4.54
9.02
850
9.94
0.0661
0.265
4.19
8.53
o' mf
44 ym
88 ym
354 ym
500 ym
275
69
4.3
2.2
193
47
3.0
1.5
150
38
2.4
1.2
Where the diffusion rate of the reacting gas through the product
layer is controlling, the equation developed is expressible as:
rg[l + (1 -
- 2(1 -
A + B
.[1 - (1 - Xg)1/3]
where
r = particle radius, cm
s
X = (initial mass - mass of unreacted core)/ initial mass
t = reaction time, minutes
143
-------
A = constant = -6 D /k, cm
e
D = effective diffusivity of reacting gas through the product
e 2
layer, cm /min
k = first-order reaction rate constant, cm/min
2
B = constant = 6 Dfi b Mg (cA)b/PB» cm /min
M., = molecular weight of solid reactant
(C ) = concentration of reactant gas in the bulk of the gas,
A b o
moles/cm
b = moles of solid reacting per mole of reactant gas
3
p = density of solid reactant, g/cm .
Chemical analysis of the sorbent samples yields y, the weight fraction
of CaSO, in the sample, which is related to Xg as follows:
_ (y " yo)/(1 " yo}
^ 1 + (1 - y:
where
y = weight fraction CaSC^ at the beginning of sulfation
and
y = weight fraction CaS04 at time t.
This model neglects the small corrections due to CaS and impurities.
For y0 = 0 and y = 0.50, XB = 0.2917; for y = 0.90, XB = 0.7876. For
y0 = 0.50 and y = 0.75, XB = 0.3685; for y = 0.95, XB = 0.8400.
Data from Run CAFB-905 for the dry sulfation of CAFB-9 regenerator
stone are summarized in Tables D-ll through D-13 and are plotted in Fig-
ures D-14 through D-16. The correlating lines are least-squares repre-
sentations of the form Y = A + BX. The constants A and B are given in
Table D-14. To extract values of DB and k, use is made of pore volume
data. The pore volume distributions are condensed on Figures D-17
through D-19, plotted to show the cumulative percent of pore volume
above 6 x 10~3 pm pore diameter. The values of pore volumes shown on
the figures are estimates of intraparticle pore volume, the balance of
cumulative volume above pore diameters of 8.8 um being attributed to
interparticle voids.
144
-------
Table D-ll
DRY SULFATION OF CAFB-9 REGENERATOR STONE, PHASE I - 450°C
Time, min
0
15
75
135
255
0
15
75
135
255
Wt %
CaSO^
-35 +
5.37
7.31
8.60
9.17
11.01
-80 +
5.16
7.06
8.60
9.76
10.59
*B
80 U. S. Sieve
0.000000
0.008824
0.014809
0.017484
0.026251
X x 10~6
Size
^
0.301
0.894
1.362
1.709
' Y x 104
_
1.493
2.508
2.961
4.451
170 U. S. Sieve Size
0.000000
0.008610
0.015737
0.021196
0.025151
_
0.787
2.147
2.864
4.552
_
0.571
1.044
1.408
1.671
-170 U. S. Sieve Size
0
15
75
135
255
5.58
9.86
9.01
11.74
16.08
0.000000
0.019822
0.015801
0.028867
0.050593
—
1.025
6.438
6.315
6.756
_
0.437
0.348
0.637
1.119
145
-------
Table D-12
DRY SULFATION OF CAFB-9 REGENERATOR STONE, PHASE II
- 750°C
Time , min
0
15
75
135
255
0
15
75
135
255
Wt %
CaS04
-35
14.36
19.64
30.00
35.52
42.02
-80
14.02
20.45
30.90
36.12
45.76
KB
+ 80 U. S. Sieve
0.000000
0.028713
0.091343
0.128144
0.176712
X x 10~5
Size
_
0.917
1.410
1.785
2.401
Y x 103
0.488
1.562
2.201
3.053
+170 U. S. Sieve Size
0.000000
0.035017
0.098831
0.134437
0.208053
_
1.917
3.321
4.337
5.145
0.233
0.662
0.905
1.413
-170 U. S. Sieve Size
0
15
75
135
255
22.07
27.12
35.87
43.76
14.80
0.000000
0.031759
0.092446
0.154378
^
—
6.372
10 . 715
11.288
"
0.070
0.206
0.334
—
146
-------
Table D-13
DRY SULFATION OF CAFB-9 REGENERATOR STONE, PHASE III - 850°C
Time, rain
0
15
60
120
239
0
15
60
120
239
Wt %
CaSO^
-35 +
41.21
43.65
44.37
46.58
50.26
-80 +
43.03
46.41
NAa
50.67
56.34
*B
X x 10~5
Y x 104
80 U. S. Sieve Size
0.000000
0.023001
0.029958
0.051822
0.090017
_
1.147
3.514
4.033
4.562
—
3.903
5.807
8.822
15.392
170 U. S. Sieve Size
0.000000
0.033613
-
0.078690
0.143925
^
1.998
-
6.721
7.146
_
2.235
-
5.260
9.695
-170 U. S. Sieve Size
0
15
60
120
239
60.06
52.59
55.80
NA
56.34
^
0.000000
0.041513
-
0.048726
_
-
19.434
-
65.788
.
-
0.918
-
1.078
NA
Not available.
147
-------
Curve 693266-A
U. S. Sieve Size
a - 35 + 80
- 80 +170
0-170
Xn)-2
-------
Curve 693263-A
Y = r ll + (l-Xcr -2(1-XC)
S L D D
U. S. Sieve Size
o -35 + 80
A -80 + 170
8
X x l(f *
Dry Sulfation of CAFB-9 Regenerator Stone in a
Fluidized Bed at 850°C
Figure D-17 shows that the -177 + 88 ym fraction had a broader dis-
tribution of pore diameters than did the other two fractions inspected.
From Table D-15 the total pore volume in the fresh feed can be related to
the arithmetic average particle diameter by the equation:
Vp = 2.63 D~1/2
3
where V is in cm /g and D is in ym. This suggests that smaller parti-
cles as obtained by grinding are more porous, and presumably their
calcium content can be more fully utilized.
With respect to the distribution of the intraparticle pore volume,
the 50 percent points are at 0.50, 0.16, and 0.85 ym for the three frac-
tions in decreasing order of particle size. Thus, the intermediate size
fraction (-177 + 88 |im) had fewer of the larger pores than did the other
two sizes and might, therefore, be expected to be less reactive because
of a less open structure. Opposed to this is the fact that its total
149
-------
Table D-14
CONSTANTS IN THE LINEARIZED CORRELATION OF CONVERSION VS
REACTION TIME IN THE DRY SULFATION OF CAFB-9
REGENERATOR STONE
Constant
Phase I - 450 °C
A
B
Phase II - 750°C
A
B
Phase III - 850°C
A
B
•——"»— ~rv r «.«. "'" J — i
-35 + 80
7.898E-05
1.935E-10
-9.691E-04
1.717E-08
-5.124E-05
2.660E-09
U. S. Sieve Size
-80 +170
4.071E-05
2.962E-11
-4 . 723E-04
3.466E-09
-1.612E-05
1.114E-09
-170
3.568E-05
5.429E-10
-2.239E-04
4.515E-10
NAa
NA
aNA = Not available.
pore volume was twice that of the -500 + 177 ym fraction. Table D-13
shows that a somewhat higher degree of sulfation was obtained with the~
-177 + 88 ym fraction than with the -500 +177 ym, but there is a sug-
gestion that even more sulfation was obtained with the -88 ym fraction.
Thus, the initial distribution of pore volumes does not appear to be a
predictor of the degree of sulfation.
Figure D-18 shows the effect of dry sulfation on the -500 +177 wm
fraction. Phase I yielded a product with about 10 percent less pore vol-
ume. This was reduced by 50 percent in Phase II, but further sulfation
in Phase III from 40 to 50 wt % CaSO^ was not accompanied by a further
change in pore volume. Table D-16 shows how the distribution of pore
volume was affected by sulfation. Pores larger than 0.4 ym in the
-500 + 177 ym fraction were essentially eliminated, while volume due to
smaller pores was roughly unchanged.
Figure D-19, however, shows Phase I sulfation reduced pore volume
for the -177 + 88 ym fraction by 50 percent in going from 5 to 10 wt %
150
-------
Table D-15
POROSITY AND PARTICLE DENSITY OF SAMPLES FROM CAFB-905
Sample
No.
Identification
U. S. Sieve
Size
Pore Volume,
cm3/g
Wt %
CaSO^
Calculated
Density,
g/cm3
176B
176C
176D
186B
186C
I860
196B
196C
I960
200B
200C
200D
204B
204C
204D
Reactor Feed -500 -
-177 H
-88
Product I -500 H
-177 H
-88
Product II -500 -
-177 H
-88
Feed III -500 H
-177 H
-88
Product III -500 -
-177 H
-88
1- 177 ;im
h 88
1- 177 ym
1- 88
1- 177 ym
1- 88
1- 177 ym
1- 88
h 177 ym
h 88
0.1358
0.2443
0.3887
0.1203
0.1316
NAa
0.0688
0.0717
NA
0.0571
0.0502
NA
0.0689
0.0413
NA
5.37
5.16
5.58
11.01
10.59
16.08
42.02
45.76
NA
41.21
43.03
60.06
50.26
56.34
NA
2.278
1.827
1.445
2.350
2.890
NA
2.595
2.566
NA
2.678
2.724
NA
2.576
2.754
NA
NA = Not available.
. Phase II reduced pore volume by another 50 percent as for the
-500 + 111 urn fraction. Phase III reduced it again in half, contrary to
the behavior of the -500 +177 urn fraction. Table D-13 shows pore vol-
ume was reduced by 80 to 85 percent in all pore diameter ranges.
Unfortunately, sample sizes available for the -88 urn fraction were too
small to permit pore volume measurements. These results show that in
large particles (-500 + 177 pm) the large pores are nearly completely
utilized in preference to small pores, whereas in small particles all
pores are utilized to about the same degree. The manner in which the
BET surface area changes during sulfation appears inconsistent
with loss of pore volume. This point needs further study.
151
-------
Table D-16
EFFECT OF DRY SULFATION ON THE DISTRIBUTION OF PORE VOLUME, cc/g
Pore Diameter,
ym
Feed
Phase I
Products
Phase II
Phase III
-500 + 177 ym Fraction
Above 0.4
0.04 to 0.4
Below 0.04
0.08498
0.02852
0.02235
Total 0.13585
-177 + 88 Mm Fraction
Above 0.4
0.04 to 0.4
Below 0.04
Total
0.06856
0.08099
0.09479
0.24434
0.08131
0.02302
0.01598
0.12031
0.08958
0.01917
0.02287
0.13162
0.03389
0.01160
0.02330
0.06879
0.00962
0.03248
0.02964
0.07174
0.00253
0.03132
0.03481
0.06866
0.01261
0.01151
0.01718
0.04130
Table D-17
AVERAGE PARTICLE DENSITY OF SAMPLES FROM CAFB-905
Particle
Average
Particle
Phase
Phase
Phase
particle radius, ym
density
I - 450°C
II - 750°C
III - 850°C
U. S. Sieve Size
-35 + 80
169.25
2.314
2.472
2.622
-80 + 170 | -170
66.25
2.058
2.428
2.739
22.00
1.842
2.388
2.762
Particle densities were calculated from the pore volume data sum-
marized in Table D-15 and plotted on Figure D-20. Finally, average
particle densities for each phase of the experiment were calculated as
in Table D-17. For these calculations the density of CaSO. and CaO were
taken as 2.96 and 3.32, respectively.
152
-------
Curve 693138-A
Product
Sample
Phase I
Phase II
Phase III
40 60
Cumulative Pore Volume,
Figure D-17 -
Distribution of Pore
Volumes in CAFB-905
Sample Phase I Feed
10 20 30
Cumulative Pore Volume, *
Figure D-18 - Distribution of Pore
Volumes in CAFB-905
Pro due t Samp1es,
-35 + 80 U. S.
Sieve Size Fraction
10 20 30
Cumulative Pore Volume, *
Figure D-19 -
Distribution of Pore
Volumes in CAFB-905
Product Samples,
-80 + 177 U.S.
Sieve Size Fraction
-------
2.8
2.6
2.4
2.2
e
u
T 2.0
2-
tS
£
I L8
U
2 1.6
1.4
1.2
1.0
I I
Phase III Feed
Phase II Product
Phase III Product
Phase I Product
Phase I Feed
50
100 150 200
Average Particle Radius, M
250
Fig. D-20 - Particle Density of Samples from Dry Sulfation of
CAFB-9 Regenerator Stone
i
Diffusion coefficients were calculated for Phases II and III by using
VL = 56.08 and (C ) =5 percent, corrected to temperature. Comparable
values for Phase I could not be calculated because the graphs had posi-
tive instead of negative Y-intercepts. This might mean that diffusion
was not controlling at low conversions. The results are in Table D-18
from which the following statements with respect to the effect of temper-
ature may be made:
• Increasing the temperature from 750°C to 850°C increases k,
the reaction rate constant, for -500 + 177 \m particles by
a factor of greater than 3 versus nearly 12 for the
-177 + 88 ym fraction.
• Increasing the temperature as above reduces De, the diffusion
coefficient per particle, by a factor of 5.5 for the
-son + 177 urn versus 2.5 for the -177 + 88 urn fraction.
Thus, increasing temperature to at least 850°C is favorable in that the
reaction rate is increased, but the effect of temperature on the diffusion
154
-------
Table D-18
PARAMETERS FOR DRY SULFATION OF CAFB-9
Parameters
Phase II
9 V»
D , cm /min/particle
k, cm/mln/particle
Phase III
O
D , cm /min/particle
k, cm/min/particle
Activation Energy
QD, cal/g mole
AH, cal/g mole
-35 + 80
2.118E-04
1.311
3.819E-05
4.472
-39,200
28,000
U. S. Sieve Size
-80 + 170 |
4.199E-05
0.533
1.671E-05
6.223
-21,100
56,100
-170
5.380E-06
0.144
NAa
NA
NA
NA
3NA = Not available.
b
The dimensions used are intended to remind the reader that the theory
was derived for a single particle, while the observations were made on
a bed of particles, thus requiring the assumption that every particle
was exposed to the same average conditions.
coefficient may be obscured by the different sulfation levels in Phases II
and III.
Additional insight can be obtained by calculating activation ener-
gies: Q for D and AH for k in accord with the equations
and
In
2 AH 1 I
~ =~R
K
Qt
R~
(D
(2)
The latter is derived from Barrer and includes consideration of entropy
of activation. As shown in Table D-13, the activation energies for k
155
-------
Table D-19
CALCULATION OF BULK VALUES OF PARAMETERS
Size Fraction
-35 + 80 | -80 +170 | -170
Size Range, Pm
Average Radios, Mm
Particle Density, g/cnf
-500 + 177 -177 + 88 -88
169.25 66.25 22.00
Phase II
Phase III
N, Particles/g
Phase II
Phase III
Reaction Rate, kN, cm/min/g
Phase II
Phase III
2
Diffusion Rate, D N, cm /min/g
Phase II e
Phase III
2.472
2.622
1.9919E+04
1.8780E404
2.611E+04
8.398E+04
4.218
0.717
2.428
2.739
3.3815E+05
2.9975E+05
1.802E-I-05
1.865E+06
14.199
5.009
2.388
2.762
9.3888E+06
8.1174E+06
1.352E+06
NAa
50.512
NA
NA = Not applicable.
appear reasonable but suggest that grinding the spent sorbent is not
desirable since the smaller particles have a higher activation energy.
This higher energy may also aid in the unfavorable changes, such as modi-
fication of pore structure. The Q_ values are negative, reflecting the
decrease in D with increase in temperature, possibly coupled with the
increase in degree of sulfation.
The role of particle size is further clarified in Table D-19, in
which the parameters are calculated per gram of sorbent. Grinding to
smaller sizes increases the reaction rate per gram of sorbent nearly two
orders of magnitude in both phases. This also increases the diffusion
coefficient. We conclude that grinding to at least -177 um is
advantageous.
156
-------
REFERENCES
1. Keairns, D. L., et al., Fluidized Bed Combustion Process
Evaluation, Phase 1 - Residual Oil Gasification/Desulfurization
Demonstration at Atmospheric Pressure, Vols. I and II, Report
to EPA, Westinghouse Research Laboratories, Pittsburgh, PA,
March 1975, EPA 650/2-75-027 a and b, NTIS PB 241-834 and 241-835,
2. Smith, J. M., Chemical Engineering Kinetics, New York: McGraw-
Hill Book Co.; Ch. 2, 14.
3. Jost, W., Diffusion in Solids, Liquids and Gases, New York:
Academic Press, Inc.; 1952, Ch. VII.
157
-------
APPENDIX E
PORE VOLUME STUDIES
In support of the dry sulfation and the dead-burning studies, paral-
lel work on pore volumes and surface areas of Limestone 1359 and CAFB
regenerator stone was carried out. The upper limit on processing tem-
perature for the CAFB stone might be taken as that at which thermal
decomposition of the CaSO, it contains occurs. If the S02 released can
be recycled to the gasifier or to a sulfur recovery plant, then the upper
limit might be set higher by other considerations, such as process eco-
nomics. The lower limit is set by the rate of reaction, either for
S02/02 pickup or for migration of the constituents of the crystal lat-
tices leading to inactivation. Either limit may be modified from that
for pure CaSO, by the presence of other elements in the matrix.
Initial studies used three size fractions of Limestone 1359:
-3360 + 2000, -595 + 420, and -105 + 74 ym. Approximately 10 g of each
fraction was placed in an alumina boat covered with platinum foil. The
boat was inserted in a furnace heated to 1070°C in air. The samples
were held at that temperature for 1, 2, 9, 17, and 32 hours, then quickly
cooled, and placed in a desiccator. Surface areas were measured by BET
using nitrogen as the absorbed gas. Similar measurements were made on
the as-received powder. The data are presented in Tables E-l and E-2
and Figure E-l along with weight loss measurements.
The following observations can be made from the table and the figure.
• The calcination is accompanied by an increase in surface area
as the decrease in weight proceeds to the theoretical value of
44 percent.
• For each size fraction, calcination is complete in about one
to two hours as determined by weight loss data. The large
particles take nearly twice as long for complete calcination
as do the small particles, which is reflected both in weight
loss and BET surface area data.
158
-------
• A peak surface area appears at the completion of calcination,
followed by a sharp decrease in surface area on continued
calcination. This change is essentially complete, irre-
spective of the size, when the calcination is continued
for about four hours.
• The effect of particle size is noticeable only at the early
stage of calcination prior to the completion of decomposition.
The results indicate two competing mechanisms during calcination.
A loss of C0? increases the area while continued exposure at high tem-
perature decreases it. Surface area can be increased by increasing the
number of particles/g or by creating more pores or larger pores within
each particle.
Table E-l
WEIGHT LOSS OF LIMESTONE 1359 DURING CALCINATION AT 1070°C
11.
s.
Mesh
Size
Particle
Diameter,
pm
Weight
After 1
Loss (%) at 1070
hr
After
O
2
C
hr
-6 + 10
-30 + 40
-140 + 200
-3360 + 2000
-595 + 420
-105 + 74
12.7
43.3
42.9
43.5
42.8
43.5
Theoretical weight loss = 44 percent.
Table E-2
BET SURFACE AREA OF LIMESTONE 1359, m /g
Particle
Diameter, pm
3360 - 2000
595 - 420
105 - 74
Calcination
0
0.09
0.23
0.81
1
2
5
6
1
.29
.19
.47
2
5.91
5.07
4.42
Time at
1 *
1.64
1.38
1.58
1070°C,
1 9
1.07
1.12
1.16
hr
1 17
0.80
0.87
0.98
32
0.79
0.88
0.92
159
-------
?u
•MO
§30
J20
lio
n
ill 11
-it f
1 I L Theoretical toss 43. 97* for pure CaCO,
£\
& v -30 * 40 and -140 f 200 Mesh
7 \
f- — 6 + 10 Mesh —
11 .11
o -6 ••• 10 Mesh
A -30* 40 Mesh
a -140 + 200 Mesh
I
10
30
35
15 20 25
Calcination Time, hr
Figure E-l - Effect of Calcination Time at 1070°C on the
Surface Area of Limestone 1359
Curve 69725S-A
10.0
(VI
S
at
I
CO
1.0
0.1
IHour
0 Hours
Parameter: Calcination Time
0.01
0.02 0.1 1 10
Particle Size Range, mm
Figure E-2 - BET Surface Area of Limestone 1359 Calcined at 1070°C
160
-------
The absolute values of surface area are uncertain due to possible
reabsorption of moisture and CO^ after calcination, yet the data seem
to support the view that some process resulting in reduction of surface
area is occurring along with the calcination. This is probably the well
known dead-burning process, but what is also of interest here is how
long do the particles continue to have a reasonable surface area, and
does this reflect directly on the capacity of the stone to absorb S0~?
The calcination is extended to 32 hours because preliminary design work
on the dry sulfation system indicated that a retention time in the
absorber of at least 20 hours would be required.
Figure E-2, plotted from Table E-2, shows that the surface area for
the uncalcined stone is nearly a straight line with a slope of -0.693.
To evaluate the reasonableness of this result, consider dividing a
spherical particle of diameter dQ into n spherical particles, all of
diameter d.. . The number of particles is the ratio of volumes:
n = 6 *dQ3/6 *dl °T dO/dl * (1)
The ratio of the external spherical surfaces is
= d . (2)
The total BET area found may be taken as the sum of the external spheri-
cal surface (Sn or S.) and the internal surface of the pores in the par-
ticles: (Si)Q or (Si).^ Thus,
BET - Sx + (Si)1 . (3)
Now, if S-, » (Si),, then the BET will vary inversely as d, and the log-
log plot will have a slope of -1. If (Si) » S. , then a slope of -1
means Si varies the same way as S,: inversely with d.. . Since the slope
obtained is less than 1 (in absolute value), the internal surface area
must vary as a smaller power of particle diameter, regardless of the
relative magnitudes of S and (Si).
161
-------
After 17 hours of calcination, the exponent on diameter dropped to
-0.0608, meaning that the residual BET area was then nearly independent
of particle diameter.
After one hour of calcination, the BET area increased at all three
particle diameters. For small particles, the BET area had already
attained the limiting condition of independence of diameter.
PORE VOLUME STUDIES ON LIMESTONE
We sought further information by measuring pore volumes of four of
the calcined limestone samples. The data we obtained are summarized in
Tables E-3 and E-4. The apparatus used was a Micromeritics Instrument
Corporation Model 900 Series Mercury Penetration Porosimeter. We deter-
mined pore volumes by forcing liquid mercury at ambient temperature into
the sample at pressures in the range 0.9 to 30,000 psia. Pore diameters
are calculated from the surface tension formula if one assumes the sur-
face tension of mercury to be 474 dynes/cm and the contact angle to be
130°.
An existing computer program was modified to produce Figures E-3
through E-10. One modification normalized the pore volume data by
dividing each value by the largest value found. A second generated dif-
ferential normalized volume curves by plotting the incremental volume
between successive measurements against the arithmetic average of the
corresponding pore diameters.
Both size fractions showed significant pore volumes in the range
0.02 to 0.2 ym: 0.0568 (50% of 0.1136) versus 0.1746 (19% of 0.9191)
cc/g for the 1-hour samples 92-1 and 94-1, respectively. Peak volumes
in these ranges occurred at 0.08 and 0.15 ym, respectively. At 32 hours,
pores in this diameter range are absent.
Samples 92-1 and 93-1 have lower total pore volumes than the other
two samples. This difference is in line with BET surface area measure-
ments reported previously, showing that calcination of large particles of
limestone was not completed under the conditions used in one hour.
All of the samples show that a significant fraction of the total
jlume (10 to 15 percent) occurs in the average diameter range of 8 to
162
-------
Table E-3
PORE VOLUME DATA ON LIMESTONE SAMPLES CALCINED AT 1070°C
Sample No.
U. S.
Mesh Size
Calcination
Time , hr
Total Pore Volume,
cc/g
Pore Volume
below 8 pm, cc/ga
92-1
93-1
94-1
95-1
-6 + 10
-6 + 10
-140 + 200
-140 + 200
1
32
1
32
0.1136
0.3336
0.9191
0.8283
0.0772
0.2669
0.4044
0.3893
Discussed later in text.
Table E-4
DISTRIBUTION OF SIGNIFICANT CONTRIBUTIONS TO INTRAPARTICLE PORE VOLUME
Sample
No.
Contributions to Intraparticle
Pore Volume
Average Pore
Diameter Range,
ym
% of Total
Volume in
This Rangea
Location o£ Significant
Contributions
% of
Total
Volume^
Pore
Diameter,
ym
92-1
93-1
94-1
95-1
0.02
0.3
0.02
1.0
0.3
1.8
- 0.2
- 2.0
- 0.2
- 8.0
- 1.8
- 6
50
73
19
14
24
11
8.0
6.0
6.0
2.9
2.0
2.0
1.9
1.8
7.0
4.0
3.0
0.09
1.1
0.15
7.0
4.0
2.2
1.5
1.2
8.0
3.0
5.0
aTotal volume is cumulative normalized volume, including interparticle
voids
bRefers to the ordinate on a plot of differential normalized volume versus
average pore diameter.
163
-------
TOTflL VOLUME =0-1136CC/3
T3TRL VOLUME = 0 .11 36 CC/G
Figure E-3 - Cumulative Pore Volume for Calcined
Limestone 1359 (See ordinate for
calcination conditions)
T8TRL V9LUME = 0.3336 CC/G
*10* 2 5 Iff* Z
Figure E-5 - Differential Pore Volume for Cal-
cined Limestone 1359 (See ordi-
nate for calcination conditions)
TBTflL VBLUnE =0.3336 CC/G
-
i iff
Figure E-4 - Differential Pore Volume for Calcined
Limestone 1359 (See ordinate for
calcination conditions)
Figure E-6 - Differential Pore Volume for Calcined
Limestone 1359 (See ordinate for
calcination conditions)
-------
= C.9191CC/G
T?Tflu VOLUME =0.919ICC/G
.
i K
i\
! xi
:ff 2 s if 2 s
cir' - t :?' 2 5 nr z s :ff 2
oiancTtn. me
Figure E-7 - Cumulative Pore Volume for Calcined
Limestone 1359 (See ordinate for
calcination conditions)
3TflL VBUmE =O.B283CC/G
1 ;
;
: ! i
; ; 1
• • • \
i
j :
. ' : i
i • i
i
, : i : I i i
i
!
1
i
1
1
., i
,
\
\
^
^
T
S \tf 2 S Itf
HMttfu. n;c»>«
Figure E-8 - Differential Pore Volume for Calcined
Limestone 1359 (See ordinate for
calcination conditions)
-------
20 ym. This is considered to be volume between the particles. The
value of 8 ym as the largest intraparticle pore is obtained by further
analyses described below.
Samples with smaller particle sizes (-105 + 74 ym) have higher
total pore volumes and higher intraparticle pore volumes than the sam-
ples with larger particle sizes (-3360 + 2000 Mm). This finding is in
contrast with the nearly equal surface area recorded for these samples,
allowing for the time lag for the larger particles.
The data support the conclusion that short-time calcination develops
significant pore volume in the 0.02 to 0.2 ym pore diameter range,
whereas exposure to 1070°C for more than 2 to 4 hours essentially elim-
inates these pores. Pore volume also develops in the 0.5 to 2.0 ym size
range. It is likely that these have consolidated from the smaller pores
found initially.
ANALYSIS OF PORE DATA BY MODEL
To examine these data more intensively, a model was developed from
which the number distribution of pores, the pore surface area, the pore
mouth area, and the pore volume could be estimated. Development of the
model is given at the end of this section in an addendum.
Pores are present in the sorbents in a variety of sizes and config-
urations. The purpose of a model is to provide a simplified way of
describing the essential relationships among the variables involved.
El
Klinkenberg proposed straight cylindrical, not interconnected, capil-
laries but concluded that pore sizes in sandstone calculated from
miscible displacement data were smaller than those obtained by mercury
E2
porosimetry. Dullien and Azzam proposed modifying this model
to a network of cylindrical capillaries with step changes in diameters.
Their model would require information on the sum of the lengths of the
necks and bulges in the capillaries, an effective length of the capil-
laries, and the radii of the necks and bulges.
Instead of such a model we took a more direct approach. The pores
are visualized as right circular cones with their apices at the center
166
-------
of spherical particles and their mouths on the spherical surfaces of the
particles. Such a structure might permit application of a single factor,
derived from experimental data, to yield an effective surface area.
Three physical characteristics of the pores are of interest: pore
volume, pore surface area, and pore mouth area. The primary data
required are the particle diameter, the porosimetry data, and the particle
density. The porosimetry data are in the form of cumulative volume of
mercury added to the sample versus pressure, from which a distribution
of pore diameters versus pore volume can be calculated. The model
developed permits estimation of the number distribution of pores by pore
diameter, permitting, in turn, calculation of the three quantities of
interest as distributions by pore diameter.
Number of Pores/Particle
The number of pores per particle in the ith pore diameter interval
is given by
2
where
4f1(pVT)
Id \' 1
d
1 P/i
2"
- ( --E I
id y.
1/2
f. = the fraction of the pore volume in the i pore diameter
interval
3
p = the particle density, g/cm
/ = the total measured pore vi
d = the particle diameter, Mm
3
V = the total measured pore volume, cm /g
(d ). = the average pore diameter for the i interval, pm.
The data for the fine powder (-105 + 74 ym) were used to calculate (N )..
The results are plotted in Figure E-ll. We found that the number of
pores in each pore diameter interval increased at an accelerating rate
as the pore diameter decreased. This behavior continued down to the
smallest diameter measurable, corresponding to the maximum pressure of
206.8 MPa (30,000 psi) attainable in the porosimeter. Consequently, the
absolute total number of pores cannot be inferred from these data.
167
-------
10*
- ,_o
I
fie'1
10
,-z
10
,-3
Calcination of-140+ 200 Mesh Limestone
!Hour«1070cC
32Hour$SW70°C
(Cu rves exclude pores smaller than 0. OObu m i
10 20 »
Cumulative Number ol Pores/Particle, millions
Figure E-ll - Effect of Calcination Time on Number of Pores/Particle
There is a further problem in interpreting the results. The
porosimeter measures all voids as pores, whether they are inside the
particles or between them. The former are the pores of interest in this
investigation and are referred to as internal pores or simply as pores.
The latter are called interparticle pores where it is necessary to dis-
cuss them.
To determine how much of the "pore" volume obtained was in internal
pores, the data were cumulated from the lower end of the pore diameter
range, as shown in Figure E-ll. Calcination for 32 hours reduces the
total number of "pores" from 32.6 million to 18.7 million/particle. It
is evident that above a certain pore diameter, there is no significant
increase in the number of pores in either case. For both samples, about
85 percent of the total "pores" are below about 0.01 ym in diameter.
The distribution of pores, however, is apparently different from the two
168
-------
samples, since when the pore diameter is increased the 99 percent level
is reached at 0.025 \im for 32 hours of calcination but not until 0.2 ym
for 1 hour of calcination.
Pore Mouth Area/Particle
If the area of the pore mouth for each diameter interval is multi-
plied by the number of pores in that interval and the results cumulated,
again from the lower end of the diameter range, the total will be limited
by the spherical surface area of the particle. Figures E-12 and E-13
show the results of this calculation. First, note that the end points
_o
are very nearly the same: 5.02 x 10 for calcination for 1 hour versus
o o
5.35 x 10 m for 32 hours of calcination. Second, the curves do not
differ much until one considers diameters smaller than about 1 pm. In
the range of 0.15 to 0.5 pm, as is more clearly shown in Figure E-13, the
pore mouth area is nearly an order of magnitude higher for particles
calcined for 1 hour compared to those calcined for 32 hours. Extended
calcination apparently results in a loss of pore mouth area in this range
of diameters. This decreases the probability that a gas molecule strik-
ing the particle will find itself at a pore mouth and, therefore, in a
position to continue into the particle.
The limits shown on Figure E-12 are from Table E-5, giving the par-
ticle characteristics. It may be a coincidence, but the curves of Fig-
ure E-12 cross at the theoretical value for spherical surface/particle
_Q O
of 2.38 x 10 m . Since the pore diameter at this point is about 6 ym,
larger pores are concluded to be outside the particles and are, there-
fore, interparticle pores. Compare this conclusion with deductions
from the volume fraction of pores as summarized in Table E-6. Both
approaches yield about the same limit. These calculations say in effect
that, if the pore volume/particle approached the particle volume estimated
on the basis of spherical particles, the maximum pore diameter to be
considered as an internal pore corresponds to that for which the cumulative
pore volume equals the particle volume. This pore diameter is identified
169
-------
10
CM
e
a>
o
«-8
O
a.
J2
i
3
O
10
Spherical Area of Mean Particle Size
Calcination of -140 +200 Mesh Limestone
O 1 Hour (<* 1070°C
A 32 Hour>ifn>l0700C
10
,-3
10
,-2
Mf1 10°
Average Pore Diameter, \im
101
10
Figure E-12 - Effect of Calcination Time on Pore Mouth Area/Particle
©
10.0
O
Q.
I
i—«
<§>
"5
O
a.
1.0
0.1
Calcination of -140 + 200 Mesh Limestone @ 1070°C
-3
10
Figure E-13 -
10
^" L
n-1
1 ' , —' 1 1—>-l
* 101
10 ' 10V
Average Pore Diameter, pm
Effect of Calcination Time on Pore Mouth Area Ratio
170
-------
Table E-5
SUMMARY OF CHARACTERISTICS OF -140 + 200 MESH SPHERICAL PARTICLES
Particle diameter, pm
Range 74 to 105
Harmonic mean 86.8
Surface area/particle, m2 2.36695 x 10~8
o _7
Volume/particle, cm 3.42419 x 10
o
Particle density, g/cm 2.6
3 -1
Particle volume, cm /g 3.84615 x 10
Mass of one particle, g 8.90289 x 10~7
Number of particles/g 1.12323 x 10
Table E-6
ESTIMATE OF MAXIMUM AVERAGE INTERNAL PORE DIAMETER FOR
CALCINED LIMESTONE 1359
Sample Identification
Number 94-1 95-1
Mesh size -140 + 200 -140 + 200
Calcination time at 1070°C 1 hour 32 hours
Measured total "pore" volume, cm3/g 0.9191 0.8283
Calculated no. of spherical particles/g 1.123 x 10 1.123 x 10
"Pore" volume/particle, cm3 8.133 x 10~7 7.374 x 10~7
Volume of spherical particle, cm3 3.424 x 10~7 3.424 x 10~7
Particle volume/"pore" volume 0.4185 0.4643
Average pore diameter, ym 7.69 9.59
from the original data as that corresponding to the fraction in Table E-6
labeled particle volume/pore volume. This approach yields about the same
limits as Figure E-12.
In summary, the pores may be considered intraparticle pores if
their average diameters are less than 8 .1: 2 vim, based on the criteria
that the pore mouth area does not exceed the spherical surface area of
171
-------
the particle and the pore volume does not exceed the particle volume.
The number of pores, however, is essentially negligible down to about
0.2 urn for even short-term calcination.
Surface Area/Gram
The calculated pore surface areas are shown in Figure E-14. After
2
1 hour of calcination, the particles show a total area of 30.4 m /g.
2
Calcining for 32 hours reduces this to 13.7 m /g. Of the total area, in
each case 84 percent is in pores smaller than about 0.13 pm. An addi-
tional 12 percent occurs in the range of 0.4 to 2.0 urn for the 32-hour
calcination, while the 1-hour case shows a continuous increase in area
amounting to an additional 15 percent in the interval from 0.13 to 2.0 pm.
The surface areas calculated from the model may be compared with
measured values (Table E-7). Values derived from porosimetry are also
given. The surface area changes inferred from mercury porosimetry are
10'
I
01 — 1
s-io '
I
10
Calcination ol -HO-f ?40Mesh Jjmeslpne
o" 1 Hour »~1070*C"~"
A 32 Hours *»1070°C
10
Figure E-14 -
Surface Area. m?/cj
Effect of Calcination Time on Total Surface Area/Gram
172
-------
Table E-7
SURFACE AREAS OF LIMESTONE 1359
Sample Number 94-1 95-1
Calcination time at 1070°C 1 hour 32 hours
2
Calculated area, m /g
Below 8 pm 30.3 13.7
Below 0.2 ym 28.7 11.6
50% point 0.0105 pm 0.00802 um
2
Measured area, m /g
Total by BET 6.47 0.92
Below 8.8 ym by porosimetry 19.8 8.1
Reduction in area
By BET 85.8%
By porosimetry 59.1%
By model 54.8%
in good agreement with those predicted by the model. The BET values,
however, show a much larger reduction. These values may be low due to a
decrease in the absorptive power of CaO for nitrogen after calcination
Another possibility is that the model overestimates the surface area
since not all of the pores extend deeply into the particles.
If C ±s the fraction of altitude of the conical pore measured from
2
the apex, then the surface area for a given pore varies as C , and the
3
volume of the pore varies as C . If C is 25 percent of the cone altitude,
meaning the pore extends from the surface of the particle to 75 percent
of a particle radius of the center, then truncating this much of the cone
reduces the surface area by only 6.25 percent and the volume by 1.56 per-
cent. Clearly, the reduction in calculated area would have to come from
a substantial adjustment in the model near the surface of the particle.
If the pores were more compact than a cone, the calculated surface area
for a given pore volume would be decreased. If, for example, the volume
173
-------
attributed to one conical pore were assigned to a cylindrical pore, the
ratio of surface areas would be
s
s~
p
2
~ 3
/d \2"
'-W
1/2 d
+ <*
which for 0.1 urn pores in 100 ym particles is about 2/3. One may con-
clude that a significant number of the pores are more cylindrical than
conical, but elaboration of the model to reflect this finding is outside
the scope of the present work.
Relative Distribution of Pores before and after Calcination
The progress of the calcination can be further illuminated by com-
paring the number of pores left after 32 hours of calcination for each
interval of pore diameter with the number produced by 1 hour of cal-
cination. This comparison is shown in Figure E-15 which reveals three
important consequences of extended calcination:
• Pores in the diameter range of 0.088 to 0.36 um are essentially
eliminated (97+%).
• Pores in the interval from 0.36 urn to the upper end of the
diameter range show either retentions of essentially 100 per-
cent or increases of up to 700 percent.
• About half the pores with diameters smaller than 0.04 ym sur-
vive 32 hours at calcination of 1070°C.
Thus, there is a critical pore diameter range of, say, 0.04 to 0.4 pm in
which pores essentially disappear on extended calcination. Above this
range larger diameter pores are generated, and below it about half the
pores are somehow able to survive calcination. In Figure E-14 the criti-
cal diameter range accounts for about 30 percent of the surface area
present at 1 hour of calcination and only 3 percent of that at 32 hours.
The increase in the number of pores in the large diameter range is
considered part of the same process responsible for eliminating pores
in the middle range of diameters. Similar results have been observed
F3 E4
in sintering other oxides like MgO, ZnO, and UO?.J ' This process is
174
-------
believed to depend on the free surface energy which is minimized by
reducing the area in the approach to equilibrium.
The survival of very small pores suggests an additional mechanism
is operative. One possibility is that the gas present in the fine pores
somehow stabilizes their dimensions, a theory that could be tested by
sintering in a vacuum.
Figure E-15 shows two discontinuities. The one between 6.4 and
-3
7.2 x 10 ym is because no incremental pore volume was measured at this
diameter interval for the 1-hour calcination. No significance is attached
to this fact; another sample would probably have had a nonzero increment
_2
here. The other discontinuity between 1.7 and 2.1 x 10 ym is produced
by a zero increment to pore volume for the 32-hour sample. This informa-
tion may be significant, but no explanation is apparent for a zero at
this diameter followed by a 40 to 60 percent retention of pores in the
interval of 2.4 to 3.2 x 10~2 ym.
rtl
c.
o
£
c
o
o
ro
10
10
Average Pore Diameter, pm
Figure E-15 - Relative Distribution of Pores in Limestone
Calcined at 1070°C
175
-------
One additional aspect of the data merits comment. The BET areas
for large, and small particles over the 32 hours of calcination are
similar, agreeing with the model. The pore volume of the large particles,
however, is considerably smaller than that for small particles. Longer
calcination time did not result in comparable pore volumes. The model
predicts that the pore volume should be independent of the particle
diameter for the case where the particles have the maximum pore volume.
The 1-hour sample of the -3360 + 2000 pm fraction was not fully calcined,
but the 32-hour sample, as noted in Table E-3, showed only 0.27 cc/g
versus about 0.40 for the -105 -f 74 pm fraction. It may be that the
lattice structure is mechanically more rigid in large particles than in
small, and the C0_ evolved on calcination has time to diffuse through
the lattice into the pores formed initially before additional pores can
develop.
Active Pores
From the preceding analysis one sees that pores in a midrange of
diameters essentially disappear on continued calcination (sintering).
The width of this range depends on the criterion used. If one uses pore
mouth area, as in Figure E-13, this range is 0.008 to 0.8 ym, correspond-
ing to a factor of at least 1.5 between the area at 1 hour and that at
32 hours. In the narrower range of 0.12 to 0.62 pm, the factor is at
least 5.
If one uses relative survival of pores, however, as shown in Fig-
ure E-15, the range is 0.04 to 0.4 ^m, over which less than 10 percent
of the pores survive extended calcination. Table E-8 summarizes the cal-
culated pore distribution by pore diameter range. Large pores (those
over 0.4 \tm in diameter) increase in number by a factor of 3.5. About
60 percent of the small pores (those smaller than 0.04 |im in diameter)
survive calcination and after 32 hours are about 500 times as numerous
as the pores in the midrange of 0.04 to 0.4 pm.
176
-------
Table E-8
NUMBER OF PORES PER GRAM OF LIMESTONE IN
SELECTED PORE DIAMETER RANGE
Diameter range, ym
8 - 0.4
0.4 - 0.04
0.04 - 0.005
Sample Number
94-1
q
8.20 x 10
1.17 x 1012
3.54 x 1013
95-1
1 0
2.86 x 10
4.49 x 1010
2.09 x 1013
Ratio
3.49
0.0385
0.591
Similar results are obtained by comparing total surface area as in
Table E-9., pore mouth area as in Table E-10, or pore volume as in
Table E-ll. This table shows that the small pores contribute less than
4 percent of the total pore volume after 1 hour of calcination and less
than 2 percent after 32 hours.
One may speculate that the disappearance of pores in the range of
0.04 to 0.4 ym and the conversion of the limestone to a dead-burned or
inactive stone are related. Pores in this diameter range are, therefore,
called active pores. The increase in the number of large pores is insuf-
ficient to maintain activity, so these are termed inactive pores. The
very small pores apparently contribute little to activity.
Table E-9
SURFACE AREA, m /g OF LIMESTONE,
IN SELECTED PORE DIAMETER RANGES
Diameter
8
0.4 -
0.04 -
range, |im
0 . !\
0.004
0.005
Sample
i 94-1
0.584
8.89
20.85
Number
95-1
1.783
0.369
11.44
Ratio
3.05
0.0415
0.549
177
-------
Table E-10
PORE MOUTH AREA, m2/g OF LIMESTONE,
IN SELECTED PORE DIAMETER RANGES
Sample
Diameter range, pm 94-1
8 - 0.4 1.181 x 10~2
0.4 - 0.04 1.412 x 10~2
0.04 - 0.005 2.259 x 10~3
Number '
95-1 Ratio
3.04 x 10~2 2.57
8.53 x 10~4 0.0604
1.246 x 10~3 0.487
Table E-ll
PORE VOLUME, cm3/g LIMESTONE, IN SELECTED
PORE DIAMETER RANGES
Diameter Range, urn
Sample Number
94-1
Volume
Fraction
Volume
95-1
Volume
Fraction
Volume
Ratio
8
0.4
0.04
- 0.4
- 0.04
- 0.005
Total
Bed voidage
0.1608
0.2223
0.0354
0.4185
0.5815
0.1478
0.2043
0.03249
0.3661
0.0102
0.0192
0.3955
0.6045
0.3032
0.00849
0.01594
2.051
0.04156
0.4908
Further insight is obtained by noting that the active pores account for
about 22 percent of the "pore" volume after 1 hour of calcination and
only 1 percent after 32 hours. If one takes into account the previous
deductions about internal pores and interparticle pores, the former
amount to 41.85 percent and 39.55 percent of the "pore" volume measured,
respectively.
Tt is interesting that the interparticle pore volume fractions of
0.582 and 0.604 agree well with general estimates of voidage in packed
beds. The agreement is an indirect support of the choice of 8 \im as the
largest diameter pore to be considered as an internal pore.
178
-------
Why this diameter range (0.04-0.4 ym) should be critical is uncer-
tain, but an explanation can be offered in terms of the mean free path
of gas molecules estimated from Maxwell's equation:
L = 3/4 TO2N , (5)
where
a = molecular diameter
and
N = number of molecules per unit volume.
o
Gases like S09, C09, and N~ are about 4A in diameter. At standard condi-
^ ^ f- o O
tions the mean free path is 560A, versus 400 to 4000 A for the critical
o
diameter range. At 870°C the mean free path will be about 2300 A. A
given molecule of gas will collide with up to two other molecules before
striking the pore wall. For larger pores the frequency of collision
with other gas molecules increases the probability that a given molecule
will be deflected back from the pore mouth into the bulk gas. For smaller
pores the entrance to the pore will quickly become saturated or spent
by adsorption or chemical reaction. If we consider only mechanical
transport, a gas molecule would have a lower probability of penetrating
the full depth of a narrow pore.
Sulfation of Limestone
These observations on surface area and pore volume are helpful in
understanding the sulfation of limestone. After an initial period during
which essentially 100 percent of the SO™ passing through a fluidized bed
of limestone particles is absorbed, the percentage of absorption falls
off, but not immediately, to zero. Rather, there is a period during
which absorption continues but at a significantly reduced rate. The
literature frequently mentions that sulfating limestone either directly
or by oxidizing CaS produces a layer of sulfate that effectively prevents
the reaction from going to completion. This notion can be examined in the
light of the present work.
!79
-------
In studies of calcite ' the high degree of reactivity of the
oxide following calcination has been shown to be due to the difference
in structure of the original calcite and the CaO. The change from cal-
cite structure to cubic CaO has been noted as accompanied by a linear
shrinkage of 23.7 percent, meaning that either an unstable CaO lattice
is formed, in which the calcium and oxygen atoms retain their original
position, or pores are formed between CaO units.
pc go
Studies conducted by several workers indicate that rearrange-
ment of oxygen atoms takes place immediately following removal of CO
E9
and that the decomposition begins at the surface.
Calcium oxide has a face-centered cubic structure with a lattice
o
spacing of 4.81 A, while CaSO, is orthorhombic, with spacings as shown
in Table E-12. If one visualizes a monolayer of CaSO. on the surface of
a spherical particle of calcined limestone, one may estimate the area
o o
-15 2
occupied by one molecule of CaSO, as 3.5 A x 7.0 A, or 2.45 x 10 cm .
If one assumes a particle diameter of 100 ym, one may make the following
calculations:
Particle weight, g 1.361 x 10~6
Number of particles/g 7.346 x 10
2 -4
Particle surface area, cm 3.14 x 10
Number of molecules of CaSO./particle 1.28 x 10
-13
Number of g-mols of CaSO /particle 2.13 x 10
Weight of CaSO,, g/particle 2.90 x lO"11
-3
Wt % CaSO, 2.13 x 10 , or 21.3 ppmw
This amount of CaSO is very much lower than that attainable, so more
than the spherical surface of the particles must become suI fated.
If the total surface area available because of the pore structure
is taken into account, a higher percent of sulfation can be explained.
2
BET measurements have shown areas of the order of 5 m /g. This is
6.807 x 10 nT/particle, or a factor of 217 times the spherical surface-
of the particle. Hence, the sulfate in a monolayer is 0.461 wt %, which
is still low.
180
-------
a
o
4.96
4.81
5.6948
6.238
b
0
7.97
4.81
5.6948
6.991
c
0
5.74
4.81
5.6948
6.996
Table E-12
CRYSTAL LATTICE PARAMETERS
CaCO- Orthorhombic
CaO Cubic
CaS Cubic
CaSO. Orthorhombic
4 o
Note: All dimensions are in Angstroms.
o
The calcium ion is relatively small compared to SO., or S0»: 2 A
o J Z
versus at least 5 A. In addition, the crystal lattice of CaSO, is large
compared to other calcium compounds, as shown in Table E-12. Sulfur in
any of its forms would thus appear to have difficulty diffusing through
even the first layer of CaSO,.
During calcination, the lattice of the carbonate must undergo shrink-
age in all three directions. Sulfidation of the oxide then requires an
expansion of the lattice in all three directions, whereas if sulfidation
of the carbonate can occur directly, it should proceed more readily
because two of the dimensions are larger than those of the sulfide lat-
tice. Sulfation, either by oxidation of the sulfide or by direct sulfa-
tion of the oxide, requires expansion of all three lattice dimensions.
Direct sulfation of the carbonate should proceed somewhat more readily
because one of the dimensions is larger than those of the sulfate.
These comments cast some light upon the nature of the sulfation
problem. Unless the SO,, and 0., molecules can reach the calcium lattice
before it shrinks from the dimensions of the carbonate to those of the
oxide, the sulfation rate should be reduced. Sufficient energy must be
available to permit calcination and yet not permit so much ion mobility
as to result in lattice shrinkage. The sulfation itself is exothermic
and, conceivably, before the energy of reaction can be dissipated, some
of it will be utilized in permitting lattice shrinkage of the calcined
but unsulfated calcium.
181
-------
To achieve 100 percent sulfation for the assumed 100 pm particles,
5 ° ° 4
the SO- has to penetrate 5 x 10 A/7.97 A or 6.27 x 10 layers of lattice
structure. An extensive pore structure clearly is helpful in exposing
surface, but it would appear that a substantial portion of the sulfation
must occur by penetration of the lattice structure. Hence, any condi-
tions that tend to maximize the size of the lattice will aid sulfation.
The foregoing analysis suggests that to use limestone as a sorbent
for a gas at 1070°C, the absorption should be carried out simultaneously
with the calcination. A critical question is the rate of pore loss
versus the rate at which the gas can be brought to the site of the pore.
One technique would preheat the limestone to just below the calcination
temperature desired. As a variation, using fine particles, the rate at
which the particle temperature equilibrates to the gas temperature may
be sufficiently fast as to make short contact times feasible. Hence,
with fine particles, it may not be necessary to preheat the stone.
There is a flux of CO-, however, from the interior of the particles that
tends to sweep out any gas diffusing into the pores. An improved version
of the process would, thus, appear to involve calcination at an optimum
combination of time and temperature so as to maximize the reactivity of
the stone. The latter might be directly proportional to some portion of
the pore volume, as for example, that which lies in the diameter range
of 0.15 to 0.50 ym. The precalcined stone would then be used either in
a fluidized bed or in an entrained flow reactor, according to the contact
time needed, with the temperature level similarly optimized to achieve
maximum utilization of the sorbent.
STUDIES WITH SPENT CAFB REGENERATOR STONE
The preceding work was extended to cover sulfation of spent CAFB
regenerator stone as in the dry sulfation process. A detailed test was
made in the 10-cm laboratory fluidized bed, as described in Appendix D.
Samples were collected at various times during the reaction and inspec-
tion for sulfur content, BET area, and pore volume distribution.
182
-------
Figures E-16 through E-28 show the primary data obtained by porosi-
metry for selected samples from Run CAFB-905. The last four figures show
that the apparatus was yielding reproducible results over the several
weeks required to process the samples.
Figure E-29 shows the cumulative pore volume distribution in three
size fractions of spent CAFB regenerator stone from Run 9. Figures E-30
through E-32 show the effect of dry sulfation. For simplicity in making
comparisons, the -500 +177 pm fraction will be referred to as the
B fraction, the -177 + 88 pm fraction as the C fraction, and the -88 pm
fraction as the D fraction. Table E-13 contains comparative data. We
have assumed that pores larger than 8.8 pm are interparticle voids. (The
slight change from the value of 8 pm used earlier in this report was a
matter of convenience in reading data from the computer printouts.) Since
the curves appear to break at about 2 pm, the pore volume below 1.8 urn
was also noted, as was the diameter range in which the major fraction of
this volume was found. The lower limit of this diameter range was 0.1
to 0.4 pm, so a second break point at 0.18 pm was selected. The follow-
ing observations are obtained from Table E-13.
The internal pore volume increases as particle size
decreases
(B:C:D = 1:1.8:2.9) .
Small pores (<0.18 pm) account for 17 and 18 percent of
the intraparticle pores in the B and D fractions, respec
tively, but 52 percent in the C fraction.
Most of the pores are smaller than 1.8 ym.
The total pore volume remaining is about the same for the
B and C fractions; no data were available for the
D fraction.
The C fraction lost 83 percent of its pore volume versus
49 percent for the B fraction.
The middle range of pore diameters (0.18 to 1.8 urn) expe-
rienced a loss of 84 and 94 percent for the B and C
fractions, respectively.
The small pores volume increased 2.3-fold for the B frac-
tion but decreased 81 percent for the C fraction.
183
-------
- Although the value of pore volume for the large pores
(1.8 to 8.8 urn) may be uncertain because they are small
in magnitude, they were nearly eliminated in the B frac-
tion but increased by a factor of 4.3 in the C fraction.
Table E-13
EFFECT OF DRY SULFATION ON THE PORE VOLUME DISTRIBUTION IN
SPENT CAFB REGENERATOR STONE
Fraction
U. S. Mesh Size
Particle Size, ym
Total Pore Volume, cc/g
Distribution, cc/g
- 8.8 + 1.8 ym
- 1.8 + 0.18 ym
- 0.18 ym
Total internal volume
Percent of feed
Distribution, %
- 8.8 + 1.8 ym
- 1.8 + 0.18 ym
- 0.18 ym
B
-35 + 80
-500 4- 177
Feed
0.3055
0.0070
0.1060
0.0228
0.1358
5.2
78.0
16.8
Sulfated
Product
0.2416
0.0003
0.0168
0.0517
0.0688
50.7
0.4
24.4
75.2
C
-80 +170
-177 + 88
Feed
0.2508
0.0023
0.1155
0.1265
0.2443
0.9
47.3
51.8
Sulfated
Product
0.1926
0.0099
0.0069
0.0245
0.0413
16.9
24.0
16.7
59.3
D
-177
-88
Feed
0.5329
0.0526
0.2740
0.0712
0.3978
13.2
68.9
17.9
These observations support the view that more pore volume is present in
the -177 + 88 ym fraction than in the -500 + 177 ym fraction, and it appar-
ently is more available for whatever processes occur during dry sulfation.
These include sulfation, pore coalescence, and pore formation. Since the
C fraction was more highly sulfated, it appears advantageous to grind
spent regenerator stone at least to -80 mesh (177 ym).
184
-------
Figure E16a
Figure E16b
Differencial Pure Volume for Sulfated CAFB Regenerator Stone
(see ordlnate fur sulfatlun condition!)
TQTRL VOLUME =0.3055CC/G
TQTflL V0LUME =0.3055CC/G
S ID- 2 E or' 2 _. 5 __
5 Iff 2
nrTcit. nictO
j
5 0*2 S 10* 2 S tf
Figure E17a
Figure E17b
Differential Pure Vulume fur Sulfnted CAKIl KoKl-fi,Taioi Stuno
(•«« urdlnete fur tulfatlun cundlth.nsi
TQTRL VOLUME =0.2508CC/G
T3THL VOLUME = 0-Z5D8CC/U
w
o 5
\l
\
5 Kf2*"s""n'"z"*&
s 10* 10"? s itr*2 s itr'z 5 icf/ s iff z s
5 icf/ s
UIHHCTCK. HICRONS
itf z 5 tf
185
-------
Figure E18a
Figure E18b
DitfirmtKl for. folgn for l.irlt*f CAH btnuritor SIOM
(••• ortlniK* for iiiKtclo* eofiltlnm)
CunuUtlvr Pur* Voliw for SulfitaJ CAfB Regenerator Stone
)
T0TRL VOLUME =0.53Z9CC/0
Figure E19a
Figure E19b
etlve Pore VoltMC for Sulfateil CAFB Rrxcn
dee ordlnate for mlfetlon condlt limit)
Differential Pure Vulune for Svlfeted CAfl legeoeretor Sto«e
(«f* urdlnet* fur MU)fat Ion cuAdltlun.)
!£
:r
T8TRL VOLUME =0.7801CC/G
I I
T0TRL VOLUME =0.2801CC/D
0 i
(D n^
* u
10 5
i "e
. o
" :is
186
-------
Figure E20a
Figure E20b
Z x
• s= -1
!"> 5s
cr ^.j
(K
•
01
u E
« ?
tf
tr i
4i
fMV
cr'
**>•*
i^-j.
»
A
n1
^
L.
i
i
J
K
i
i
/
^-
*
. — . .
i
187
-------
Figure E22a
Figure E22b
Cuojulaclvo Per* VO}OM for Sulf«to4 CAT! MgtiMrator stoi
(•« ordl»u for «ilf«tlo» euBdltloni)
irontlal Pur* VoluM for BulficW CAF1 l«fMMr«t
-------
Figure E24a
>. Volume fur Sutfjted CAKB Regenerator Stui
:ir.*lt fur miration rondltluni)
Figure E24b
eg -
T- JBLUME .- 0.2254 CC/B
cr f s ir* ? s itr1 ?
^__
0* Z 5 Iff * 5
« «
I ^ "'
U t
O •
TOTHL VOLUME =0.22S4CC/G
»?«RO€ 0
IRnETFR. rlCR6MS
Figure E25a
Figure E25b
in^t* fur lulfiitiun »u«J 111 UI
3TM.. VOLUMt = D.241R CC/C.
TOTHL VOLIirtL = 0.2416 CC/G
'§»
cr' ? s irr
OIWICTCI. ni
"iff* ? P i
? '"' 5
it. nicttoMa
„„ _
S 10" ! S Itf
189
-------
Figure E26a
Figure £26b
Differential For* Voliaw for Sulfated CAFB JU-generiUi.r S[»
(•«• ordinal* for aulfatlun conJUfimM)
Cumulative Pure Vultmo fur Suiraled LAfB IU(«
(IBM ordinal*) for •ullallun 'undlllulu)
TQTHL VOLUME = 0.1927CC/G
T0THL VBLUMC - 0.I92TCC/G
°- 5*
»»?•*
icul
S Iff I S
Figure E27a
Figure E27b
CiMul«tlve Porf Vulune fur Check Samj.Ii- at Start »l
.il !'.».• Vul.t** lor Hi*, k SMpI* «t Start of SttrU*
L VBLUMF - n.279r>rr/rj
8«
<*i
I
!$.
' 5 l(f
oiwilrrn. ni
190
-------
Figure E28a
Figure E28b
THTIIL viiuinr - n.2544rr/u
The data were examined further with the aid of the pore volume model
developed earlier. Figures E-33 through E-35 show the cumulative distri-
bution of pore diameters in terms of numbers of pores per particle. As
in the case of calcination of limestone without sulfation, most of the
pores are smaller than 0.01 pm in diameter. Sulfation increased the num-
ber of pores/particle from 288 to 432 million for the B fraction but
reduced the number for the C fraction from 180 to 27 million/particle.
Figures E-36 and E-37 show the incremental number of pores/particle
versus pore diameter. These clarify the meaning of the pore volume
changes for diameters less than 0.1 tt |im. For the C fraction, the number
of pores/particle is less for all diameters shown after sulfation,
whereas for the B fraction, at nearly every diameter interval, the num-
ber of pores is increased. Prior to sulfation the sorbent particles
may he considered as having a distribution of pores covering the exter-
nal surface of the particles. If sullation proceeds at an initially
constant rate, one may postulate that a monolayer of CaSO is laid down
on the exposed surface of the particles and the pores. Since the crystal
Yff ? " ' 5 " Iff
191
-------
10°
Curve 61749; «
10
£
£
.1
5
3
=>
O
10
,-2
10
,-3
10
176 D-170 Mesh
..
176 B-35+ 80 Mesh
Hf1 10°
Average Pore Diameter, pm
10'
Figure E-29 - Pore Volume Distribution in Spent CAFB
Regenerator Stone
10
t; 10
-1
I10"'
10
,-3
176 B
-35 + » Mesh
Curve 6974B9-A
10
-2
10-10
Average Pore Diameter, pm
10'
Figure E-30 - Effect of Dry Sulfation on Pore Volume Distribution of
Spent CAFB Regenerator Stone
192
-------
Curve i--j'49:-fl
10
10
,-2
10 ' 10"
Average Pore Diameter,
Figure E-31 - Effect of Dry Sulfation on Pore Volume Distribution of
Spent CAFB Regenerator Stone
Curve- S
-------
10'
10
i 10
-i
10
,-2
10
,-3
176 B
204 B
Curve 712953-A
10'
10° h
lie-
5
|
10
,-z
0 100 200 300 400 500 600
Cumulative Number of Pores/Particle, millions
10
,-3
204C
176 C
I
I
E
10'
10'
10
10
-2
0 100 200 300
Cumulative Number of Pores/Particle, millions
10
-3
176 D
0 10 20 30
Cumulative Number of Pores/Particle, millions
Figure E-33 - Effect of Dry Sul-
fation on Pore Diameter Dis-
tribution of -35 + 80 Fraction
of Spent CAFB Regenerator
Stone
Figure E-34 - Effect of Dry Sul-
fation on Pore Diameter Dis-
tribution of -80 + 170 Frac-
tion on Spent CAFB Regenerator
Stone
Figure E-35 - Pore Diameter Dis-
tribution in -170 Fraction of
Spent CAFB-9 Regenerator Stone
-------
Curve 712919-A
Ul
108r
o
3
g 106
105
2MB - Sulfated Product
176B - Feed
10
,-2
10
Pore Diameter. \im
,-1
10"
10
S 10
S.
s
IV
B, 10
10'
204C Sulfated Stone
-— 176C Feed
1
I L
10
,-2
-1
10
Pore Diameter, pm
10
Figure E-36 - Effect of Dry Sulfation on Pore Diam-
eter Distribution in -35 + 80 Frac-
tion of Spent CAFB Stone
Figure E-37 - Effect of Dry Sulfation on Pore
Diameter Distribution in -80
+170 Fraction of Spent CAFB
Stone
-------
lattice of CaSO, is larger than that of CaO, and since one may assume
that exposure of the sorbent to 1070°C in the CAFB regenerator permitted
the lattice structure to shrink from the dimensions of CaCO and CaS to
those of CaO, sulfation therefore forces a stretching of the atomic bond
lengths. The CaSO, formed does not necessarily detach from the lattice,
but it does tend to fill up the pore volume. It also can lead to lat-
tice dislocations, which means creation of pores. The results with the
C fraction can be explained by saying that pores at the 0.18 urn end of
the range shown on the figures are reduced in diameter and that pores
at the lower end of the range are rapidly closed. The results with the
B fraction can be explained by saying either that pores larger than
0.18 um are reduced to pores smaller than 0.18 urn or that pores are
created. It is likely that both effects occur, but the creation of pores
is the more important mechanism.
Calculations from the model as reported in Table E-14 show the
total number of pores in the various diameter ranges. Regarding the
B fraction, although there is a loss of 575 pores/particle in the
-8.8 + 1.8 um range, if these were merely reduced through sulfation,
for example, they would represent an insignificant increment to the num-
ber of pores in any lower size range. Hence, there is a real loss of
pores corresponding to a major loss of pore volume in the intermediate
range of -1.8 + 0.18 pm. There is, however, a 50 to 100-percent increase
in the number of pores in the two smaller ranges that is considered to
indicate creation of pores.
For the C fraction, again there is an almost complete loss of pores
in the intermediate range. In contrast to the B fraction, there is also
a loss of pores in the smaller diameter ranges. The pores in the C frac-
tion, therefore, appear more reactive than those in the B fraction, and
grinding the spent regenerator stone to at least -177 um appears
advantageous.
Two other characteristics may be inspected: pore mouth area and
pore surface area. Table E-15 summarizes certain geometric properties
196
-------
Table E-14
EFFECT OF DRY SULFATION ON THE DISTRIBUTION OF PORE DIAMETERS
IN SPENT CAFB REGENERATOR STONE
Fraction
U. S. Mesh Size
Particle Size, urn
Quantity
Total Pores/Particle
Size Range, pm
+ 8.8
-8.8 +1.8
-1.8 + 0.18
-0.18 + 0.018
-0.018
B
-35 + 80
-500 + 177
Feed
287.73+06
89+00
590+00
504+03
6.17+06
281.06+06
Sulfated
Product
432.24+06
110+00
15+00
282+03
13.01+06
418.96+06
C
-80 +170
-177 + 88
Feed
179.49+06
-
44+00
154+03
6.06+06
171.58+06
Sulfated
Product
26.94+06
21+00
64+00
11+03
1.13+06
25 . 79+06
D
-170
-88
Feed
138.10+06
6+00
85+00
17+03
0 . 26+06
135.31+06
Note: Exponential notation is used: 278.73 +06 means 278.73 x 10 .
of spherical particles. Note that the "average" diameter of a particle
has several values according to the definition used. For examination of
the results of the area calculations, the surface area mean was used.
Figures E-38 through E-40 show the pore mouth area per particle for the
three fractions of unsulfated stone and, in some cases, limits on diameter
and area taken from Table E-15. Only for the -88 VJHI (C) fraction were
these limits helpful in distinguishing between intraparticle pores and
interparticle pores. Fortunately, the curves level off in the range of
1 to 10 urn so a choice of 10 jjm as the largest internal pore does not
appear unreasonable.
Table E-16 shows the distribution of pore mouth areas. Before sul-
fation most of the pore mouth area for intraparticle pores is in the diam-
eter range 0.18 to 1.8 |im for the B and D fractions. For the C fraction,
the pore mouth area is more evenly distributed by pore diameters. This
is interpreted in terms of the probability of an S0? (and an 09) molecule
£- £.
striking and entering a pore. For the C fraction, the probability is
relatively independent of pore diameter below 1.8 vim, but it is much
197
-------
!
1"
_ Wwk*1 ATM oMrimmttk MMfl^rlkM Sin
Avtragr Part Wwwter. y m
Figure E-38 - Effect of Dry Sulfation on the Pore Mouth Area in -35 + 50
Fraction of Spent CAFB Regenerator Stone
Am of ArllhiMlk (Mm PMkk Sl»
Figure E-39 - Effect of Dry Sulfation on the Pore Mouth Area in -80 +170
Fraction of Spent CAFB Regenerator Stone
SHitfkjl Amol Arlm»mtcM«in Urtlclt Slit
i
\"
10-' »-' 10° »' J
AMD* Hn MiMbr. i>"
Figure E-40 - Distribution of Pore Mouth Area in -170 Fraction of
Spent CAFB Regenerator Stone
198
-------
Table E-15
SUMMARY OF CHARACTERISTICS OF SPHERICAL PARTICLES
Quantity
Particle Diameter, ym
Range
Harmonic mean
Arithmetic mean
Surface area mean3
-35
-500
261
338
375
U.
+ 80 j
+ 177
.45
.5
.0
~S". Mesh Size
-50 +170
-177 + 88
117.56
132.5
139.8
-170
-88
—
44
62.2
Particle Density, g/cm 2.278
Spherical Surface Area, 3.600 x 10
m2/particlec
Volume, cm3/particle 2.031 x 10
Mass, g/particle 4.626 x 10
1.827
1.445
-7
-5
-5
5.515 x 10~8 6.082 x 10 9
1.218 x 10~6 4.460 x 10"8
2.225 x 10~6 6.445 x 10~8
Number of particles/g
2
Surface area, m /g
Volume, cm /g
2.162 x 10
o
7.782 x 10
0.4390
4.494 x 10 1.552 x 10
-2 -?
2.478 x 10 9.437 x 10
0.5473 0.6920
SIT!
Table D-27, Appendix D
Based on arithmetic mean
diameter.
higher for the 0.18 to 1.8 urn range for the B and C fractions. After
sulfation, the distributions for both B and C fractions appear more even,
implying that those pores that are available by reason of size or fre-
quency are sulfated sooner.
Figure E-41 shows the ratio of pore mouth area after sulfation to
that before sulfation for the B and C fractions. This more clearly shows
the loss of pore mouth area for the C fraction at all pore diameters con-
trasted with an increase in pore mouth area at nearly all diameter inter-
vals smaller than 0.14 urn.
199
-------
Table E-16
EFFECT OF DRY SULFATION ON THE DISTRIBUTION OF PORE MOUTH AREA
IN SPENT CAFB REGENERATOR STONE
Fraction
U. S. Mesh Size
Quantity
Total pore mouth
area, 10~9 m2/
particle
Size Range, ym
+ 8.8
-8.8 + 1.8
-1.8 + 0.18
-0.18 + 0.018
-0.018
B
-35 + 80
Feed
278.46
151.33
6.60
92.60
10.46
17.47
Sulfated
Product
213.40
148.98
0.30
15.40
23.10
25.62
C
-80 +170
Feed
35.04
NA
0.34
16.56
7.65
10.49
Sulfated
Product
29.39
23.47
1.43
0.99
1.60
1.90
D
-170
Feed
7.49
1.14
0.90
4.33
0.40
0.72
i
10.0
„
0.1
0.01
— — 1 — 1—
i [
-J
n
j
-
-
-
i
1 1 1 — i— | 1 T — r r | . i i T | i , , ,
n r n
1 1
n rj !
! 1 L J -35 + 80 Mesh
j I--,
i
i
i ,
j
| (-' 1 -80+170 Mesh
LJ —
u
J L | i 1 1 1 ll 1 1 1 - — 1 1 1— 1 1 1 1 L-
O'2 10"1 10° 101 10
Pore Diameter, \tm
Figure E-41 - Ratio of Pore Mouth Area/Particle after Dry
Sulfation to Initial Area
200
-------
Finally, Figure E-42 shows the cumulative distribution of the total
2
surface area in m /g before and after sulfation. Table E-17 shows the
distribution of pore diameter interval. On totals, the B fraction showed
2
a 50 percent increase from 14 to 21 m /g, while the C fraction showed an
2
84 percent decrease from 57 to 9.5 m /g. The B fraction also had more
surface area initially than the other two fractions. Incrementally, both
the B and the C fractions after sulfation had most of their surface area
in pores smaller than 0.018 pm.
Table E-17
EFFECT OF DRY SULFATION ON THE DISTRIBUTION OF SURFACE AREA
IN SPENT CAFB REGENERATOR STONE
Fraction
U. S. Mesh Size
Quantity
B
-35 + 80
Feed
Sorbent
Product
C
-80 +170
Feed
Sorbent
Product
D
-170
Feed
Total Surface Area, m /g 14.456 20.946 56.610 9.445 37.399
Distribution, pm
+ 8.8
-8.8 +1.8
-1.8 + 0.18
-0.18 + 0.018
-0.018
0.015
0.012
1.183
1.362
11.884
0.017
0.000
0.407
2.914
17.608
NA
9.004
1.762
7.359
47.485
0.024
0.011
0.126
1.494
7.790
0.036
0.072
2.325
3.480
31.986
Overall, while we do not claim that the pore model used is an accu-
rate description of actual pores, we feel that the model is helpful in
thinking about what aspects of pore geometry are important for chemical
reaction. First, gas molecules must have access to the interior of the
sorbent particles; hence, pore mouth area is important. Once the react-
ing molecules have entered the pores, large surface area should enhance
reactivity. Finally, since in this case the reaction product has a
larger molecular volume than the solid sorbent, large pore volume is
desirable. We conclude that all of these criteria are best met by grind-
ing the sorbent from the CAFB process to about -177 \im. Whether the fines
201
-------
10 -
Parameter: Particle Size. U.S. M«sh
Feed Stone
— Sullated Product
X 40
Surface Area, mtyg
Figure E-42 - Cumulative Distribution of Total Surface Area in
CAFB-9 Regenerator Stone
(-88 ym) should be handled differently is not clear from the current data
but, for the present, processing this fraction along with the -177 + 88 pm
appears satisfactory.
REFERENCES
1. Klinkenberg, L. J., Pore Size Distribution of Porous Media and
Displacement Experiments with Miscible Liquids, American Insti-
tube of Mining, Metallurgical and Petroleum Engineers Transac-
tions, 210: 366-69; 1957.
2. Dullien, F. A., and M. J. S. Azzam, Comparison of Pore Size as
Determined by Mercury Porosimetry and by Miscible Displacement
Experiment , Industrial and Engineering Chemistry Fundamentals
15 (2): 147; 1976.
3. Gupta, T. K., J. Mat. Sc., 6 (25); 1971
202
-------
4. Coble, R. L., and T. K. Gupta, On Sintering and Related Phenomena,
ed. G. C. Kuczynski et al, New York: Gordon and Breach; 1967,
p. 423.
5. Fischer, H. C., J. Am. Ceram. Soc. 38 (7): 245; 1955, and 38 (8):
284; 1955.
6. Hartman, M., and R. W. Coughlin, Ind. Eng. Chem., 13 (3): 248;
1974.
7. Farnsworth, M., Ind. Eng. Chem., 19 (5): 583; 1927.
8. Clark, G. L., W. F. Bradley and V. J. Azbe, Ind. Eng. Chem., 32:
9; 1940.
9. Norton, F. H., Fine Ceramics, New York: McGraw-Hill; 1970.
203
-------
ADDENDUM TO APPENDIX E
PORE VOLUME MODEL
NUMBER OF PORES PER PARTICLE
Studies involving absorption of gases on solid particulate sorbents
often include attempts to interpret the data in terms of pores and diffu-
sion through these pores to the interior of the particles. This paper
looks at a simple model of such a pore structure and compares the results
with experimental measurements of surface area and pore volume.
The volume of a spherical particle is
Vx = 7rd3/6 • (AE-1)
where
d = the particle diameter
V = the particle volume.
The number of particles per gram of sorbent is
N = 1/V.^ , (AE-2)
where p is the particle density.
Note that this is not the number of particles in a bed of such particles,
for this would depend on the bulk density, not the particle density. The
spherical surface area of one particle is
Trd2 , (AE-3)
and the area of each pore mouth is
2
A •= ird /4 , (AE-4)
P P
where d is the pore diameter.
P
204
-------
Visualize a pore structure consisting of only conical pores with the
mouths of the pores all opening on the outer surface of the spherical
particles and with the apex of the cone at the center of the particle.
A cross-section through a diametral plane is shown in Figure AE-1. The
mouths of the pores are assumed to be on the equivalent of an equilateral
triangular spacing with some finite ligaments between the adjacent pores.
From spherical trigonometry are obtained the following formulae.
If a, b, and c are the lengths of the sides of a spherical triangle in
radians, the area of the spherical triangle is given by
A = SL E/4TV (AE-5)
where E is the excess spherical angle over TT radians in the triangle.
E in turn is given by
tan2(E/4) = tan (s/2) tan ((s - a)/2) tan ((s - b)/2)
tan ((s - c)/2) (AE-6)
in which
s = (a + b + c)/2 . (AE-7)
If the angles between the sides of the triangle are known,
E=A + B + C-ir . (AE-8)
To illustrate the use of these formulae to calculate tht. area of one
octant of a sphere, set a, b, and c each equal to ir/2. The value of s
is then 3n/4 and
2 3
tan (E/4) = tan (3i;/8) tan (ir/8) (AE-9)
E = 1.57079637 (AE-10)
= TT/2 •
The area of an octant is then S.(7r/2) (l/4ir) or S.^/8.
From Figure AE-1, the length of the arc connecting the centers of
two adjacent pores is d . This subtends an a-.igle of (u + a) radians.
s p
The angle « is given by
a = w/(d/2) , (AE-11)
205
-------
Dwg. 6445A46
Figure AE-1 - Diametral Plane Section Showing Ttoo Adjacent Pores
Figure AE-2 - Intersection of a Right Circular Cone and a Sphere
206
-------
where w is the width of the ligament between adjacent pores as measured
along the spherical surface.
The pore included angle is
6p = 2 arcsin (dp/d) . (AE-12)
Applying the above formulae:
a = 9 + a
P
s = (3/2)(6p + a) (AE-13)
s - a = (9 + a)/2
P
tan2 (E/4) = tan ((3/A)(6 + a)) tan3 (6 + ct)/4) • (AE-14)
P P
Consider a 100 ym particle with all of its pores 0.1 ym in diameter.
The width of the ligament may be taken for purposes of illustration as
0.01 urn. Then a is 0.01/50 or 0.0002 radian and the pore included angle
is
6=2 arcsin (0.1/100) or 0.002000 radians .
P
Both these angles are very small so we may take advantage of an
approximation:
sin x = x = tan x . (AE-15)
Equation (AE-12) reduces to
= 2 dp/d . (AE-16)
Equation (AE-14) becomes
(E/4)2 = (3/4)(l/4)3(6p + a)4 (AE-17)
E = (/3/4)(Op + a)2 . (AE-18)
Substituting from equations (AE-16) and AE-11)
E = /3(dp/d)2(l + (w/dp))2 . (AE-19)
207
-------
Finally,
A/S1 = inf (dp/d)2(l + (w/dp))2 . (AE-20)
The number of pores per particle is
Np = (1/2)(S1/A) (AE-21)
because each of the S../A equilateral spherical triangles has half of a
pore. Also the number of pores per gram is
(d
P
(AE-22)
SURFACE AREA PER PARTICLE
Next, expressions may be written for the surface area per particle.
Since each pore is visualized as a right circular cone, the lateral area
of one such cone is
s = irrl (AE-23)
where
r = the radius of the pore mouth
and
1 = the slant height of the pore
s = ir(d d/4) . (AE-24)
To this is added the residual surface area of the particle represented
by the ligaments between the pores. To calculate this quantity, certain
auxiliary relations are needed.
The portion of the spherical surface subtended by one pore depends
on the solid angle u that can be related to the included angle 6 . In
P
Figure AE-2, a right circular core is shown positioned with its apex at
the origin and its altitude coincident with the y-axis. The portion of
208 ,
-------
the spherical surface cut out by the intersection of the core with the
sphere can be generated by rotating the arc AB about the y-axis
B
S = 2ir I rds
A
S = 2ir
R
/
R cos (9 /2)
P
1/2
dy
(AE-25)
R = y + z
R
S = 2?r
(z) - dy = 2TrR2(l - cos
z
R cos (0 /2)
P
(AE-26)
(AE-27)
and the fraction of the spherical surface subtended by one pore is
e
f = 1 - cos
The solid angle u) is defined by
S ,. . S
to = (4Tr) = —
(AE-28)
(AE-29)
2 (1 - COS
(AE-30)
The residual surface area of one particle having N pores is
s = Trd2(l - N f)
r P
1 - COS -r
1 - -^ -L N
(AE-31)
209
-------
The total surface area per particle is
Sp = sp + sr (AE-32)
N /d N N
We may now calculate illustrative numerical values for these various
quantities, using a 100 ym particle with 0.1 ym pores. From equa-
tion (AE-20), A/SL is 1.66777 x 10~7 and N is 2.99802 x 106 pores per
particle. From equation (AE-33), the total surface area per particle is
S /S, = 749.505 + 1 - 0.749502
P 1
= 749.758 .
_Q O
Since the surface area of the sphere is S = 3.14159 x 10 m , S is
-52 p
2.35719 x 10 D m .
These values show that we can drop the last two terms in equa-
tion (AE-33); combining with equations (AE-20) and (AE-21),
S /S, =
Ir(T}-~^—2 •
This says that for a given pore diameter and w/d ratio, the surface area
due to porosity varies as the cube of the particle diameter. This is for
one particle. Combining with equations (AE-1) and (AE-2)
NS = surface area/gram of sorbent
P
210'
-------
pd
(AE-35)
Thus, the surface area/gram varies inversely as the pore diameter, and
is nearly independent of the particle diameter.
PORE VOLUME PER PARTICLE
Turning next to pore volume, the volume of one pore is
P 3 4
and the pore volume per particle is
,1/2
1-
V = N v
P P P
(AE-36)
(AE-37)
The fraction of the volume of each particle occupied by pores is
V
_
v
N /d
_E _
4 \d
Substituting from equations (AE-20) and (AE-21)
ll/2
I /dY
v
_E
V,
Since d « d, a close approximation is
P
V
_£
Vl
_ ^ 1
2/3
r i2
l + r
P
(AE-38)
(AE-39)
(AE-40)
211
-------
o
This says the pore volume/particle varies as d for a given w/d ratio
P
Combining equations (AE-1), (AE-2), and (AE-39)
NV = Pore volume/gram of sorbent
P
1 -
1 7T 1
V ^^ 2/3
w "
1 * d~
P
2
1/2
IT 1
2/3p
1 + d
P
2
- -B.
(AE-41)
Thus, the pore volume/gram is independent of the size of the particles
and only somewhat dependent on the pore diameter. This means that if the
maximum number of pores of the type visualized in the model is present,
then it does not matter whether the particles are large or small. The
total pore volume will be the same per gram of sorbent.
Using the numerical values above,
V IV, = 0.74950 .
P 1
One more relationship is obtained by combining equations (AE-40) and
(AE-34):
S /S .
//V~ = d~ * (AE-42)
pi P
Pore Mouth Area Per Particle
An expression for the pore mouth area is obtained by combining
equations (AE-20), (AE-21), and (AE-4)
2/3
(AE-4 3)
212
-------
Thus, the pore mouth area/particle varies as the square of the particle
diameter. Combining with equation (AE-2)
NEA
jr/3
Pd
(AE-44)
Hence the pore mouth area/gram of sorbent increases as the particle size
is reduced.
MODIFICATIONS FOR DISTRIBUTION OF PORE DIAMETERS
These relationships apply for the case of uniform diameter spheres
with pores of equal diameter. For the practical case, a given spherical
particle will have a distribution of pore sizes. By porosimetry the
fraction of pore volume within a given diameter range can be obtained by
measurement and then the number of such pores estimated from equation (AE-38)
. th
as applied to the i pore diameter interval:
V
-------
Similarly, from equation (AE-33), by dropping the last terms
4 \d
and
VV 1 i /d \ •»•
-------
i
.-p'i /d Y i
4 4IiPVT d
\P4
irflPvTd2
2
/d r
^d 4
• - (tf
v i
1/2
1/2 '
(AE-55)
Per gram of sorbent, the pore mouth area is
6Vn
NZ(Ap). (Np).
1/2
(AE-56)
So, as for uniform pore diameter, the total pore mouth area/gram of sor-
bent increases as the average particle diameter decreases.
These relations are summarized in Table AE-1 to show how the various
particle parameters per gram of sorbent vary with particle dimensions.
For the case where d is much less than the width of the ligament between
the pores, if grinding results only in size reduction and does not
create pores, then pore mouth area is increased but surface area and pore
volume are not. This would make the interior of the particles more
accessible to a reactant and would favor increased utilization. It
would increase reaction rate only if transport rate to reacting surface
of the sorbent rather than reaction rate were controlling.
215
-------
Table AE-1
DEPENDENCE OF PARTICLE PARAMETERS ON PARTICLE DIMENSIONS3
Functional
Dependence
d
P
Practical Dependence
» w
d * w
P
d « w
P
Number of Pores/g
Surface Area/g
d(d + w)'
P
(d + w)'
d2d
P
4d2d
P
_
4d
dw
d
2
w
Pore Volume/g
Pore Mouth Area/g
(dp + w)'
d(d + w)'
P
1.
A
4d
w
a
Basis:
1.
2.
Spherical particles of diameter d
Maximum number of conical pores of pore mouth diameter d «
separated by ligaments of width w. p
216
-------
NOMENCLATURE
A Area of spherical triangle
A Area of mouth of one pore
P
d Diameter of a particle
d Diameter of a pore
P
d Arc length between adjacent pore centers
s
f. Fraction of pore volume in i diameter interval
N Number of particles/gram
N Number of pores/particle
P
s Lateral area of one pore
P
s Residual spherical surface corresponding to the ligaments
between pores
S Spherical surface area of one particle
S Total surface area in one particle
P
V Volume of one pore
P
V Volume of one particle
VT Total measured pore volume
w Arc length of ligament between adjacent pore
a Projected angle subtended by the ligament between adjacent pores
0 Projected angle between pore walls
P
p Particle density
217
-------
APPENDIX F
LOW-TEMPERATURE FLY ASH BLENDING
Exploration of fly ash blending began with bench-scale feasibility
tests using oxidized sulfated limestone from Batch L-l as a simulated
spent sorbent. Mix details are in Table F-l. This stone contained
34.0 wt % CaSO, composition. Details are given in Appendix I.
Table F-l
EFFECT OF ADDITION OF SIMULATED SPENT SORBENT ON THE COMPRESSIVE
STRENGTH OF 2-INCH PORTLAND CEMENT CUBES
Mix Composition, g
Type I Portland cement 250.0
White sand 687.5
Simulated spent sorbent 30.0
967.5
Water 121.2
Ratios
Water/cement 0.485
Sand/cement 2.75
Sorbent/cement 0.12
Compressive Strength, MPA (psi)
7 days 14-1 (2050)
26 days 25.4 (3690
The mix was placed in 2-in. cube molds, well vibrated and tamped.
The cubes were stripped 24 hours after casting and cured in wet paper.
The compressive strengths obtained at 7 and 26 days are in the range
for normal structural concrete, 13.79 to 42.37 MPa (2000 to 6000 psi).
Figure F-l shows the strength development curve compared to that of
218
-------
Curve 690461-A
Psi
6000
5000
c.
•&
c
I 4000
I
"*/l
I 3000
o
o
2000
1000
Test mix contains 12g of simulated spent sorbent
per lOOg cement.
Water/cement = 0.49 for both curves
Type I Portland Cement
MPa
40
30
20
10
10 15 20
Curing Time, days
25
Figure F-l - Effect of Spent Sorbent on Compressive Strength of
Type I Portland Cement
normal Type I cement. The use of the spent sorbent apparently led to a
substantial reduction (30-50%) in compressive strength, not attributable
solely to the presence of 3.6 percent CaSO, in the spent sorbent/cement
mix.
The effect of CaSO was explored further in mixes containing reagent
grade gypsum. The test specimens were cylinders 7.6 cm diameter and
15.2 cm high (3 in. x 6 in.), and the mixes contained 10 percent of either
gypsum or simulated spent sorbent (oxidized sulfided limestone). To test
the oxidized sulfided limestone addition, we first placed the material in
water, and when the exothermic slaking reaction was complete, we removed
the solids and used them in the mortar. The 7- and 14-day compressive
strength data (see Table F-2) indicate that the addition of pure gypsum
resulted in a significant strength reduction at 7 days, whereas the addi-
tion of the oxidized sulfided limestone, which had been reacted with water
prior to its utilization in the mortar mix, did not result in any initial
219
-------
loss of strength. Further, there appears to be a gain in strength between
7 and 14 days. At 44 days, strengths for all three mixes reached a common
value of about 55 MPa (about 8800 psi), as shown in Figure F-2. The
effect of gypsum is, thus, to delay development of compressive strength,
while CaSO, in the form and amount present in the simulated spent sorbent
was without noticeable effect.
Table F-2
EFFECT OF CALCIUM SULFATE ADDITION ON THE COMPRESSIVE
STRENGTH OF CEMENT MORTARS
Addition
7
Age Days
1 14 44
None 43.0 (6250) - 53.6 (7780)
Oxidized Sulfided
Limestone 42.9 (6240) 49.6 (7210) 54.3 (7860)
Gypsum 30.9 (4490) 28.2 (4100) 57.6 (8370)
Notes:
1. Test specimens were cylinders 7.6 cm diameter x 15.2 cm high
(3 in. x 6 in.)
2. The oxidized sulfided limestone was slaked with water before
blending with mortar.
3. The basic mortar was a sand-Type I Portland cement mortar.
Additives were 10 percent by weight of the cement.
4. Compressive strengths are in MPa (psi).
The cylinder tests were extended to higher gypsum/cement ratios. In
these, 7.6 cm by 15 cm (3 in. by 6 in.) cylinders were cast, moist cured
for 7 days, cured in air for 14 days, and checked for axial compressive
strength. The sand/cement ratio was as per ASTM 109-73 "Compressive
Strength of Hydraulic Cement Mortars." The compressive strength
increased linearly with the gypsum/cement ratio. The actual values of
compressive strength were about half those of the previous test. The
220 ..
-------
difference was attributed to the method of curing (air versus immersion)
These data support the view that the gypsum/cement ratio can be as high
as 0.5, as shown in Figure F-3.
Table F-3
EFFECT OF GYPSUM/CEMENT RATIO ON COMPRESSIVE STRENGTH
Mix
Type I Portland Cement
Reagent Grade Gypsum,
White Sand, g
Water, ml
Compressive Strength,
» g
g
psi
MPa
1
1000
100
2750
500
3305
22.8
2
1000
300
2750
500
3735
25.8
3
1000
500
2750
500
4190
28.9
Other tests at even higher gypsum/cement ratios are summarized in
Table F-4
Table F-4
GYPSUM/CEMENT MIXES USED TO COMPARE COMPRESSIVE STRENGTHS
Mix
Type I Portland Cement, g
Gypsum, g
White Sand
Water, ml
1 | 2 | 3
1000 50
0 50
500 500
125 125
30
70
500
125
4
10
90
500
125
Although all of the mixes set completely within 48 hours, mixes 3 and 4
were rather weak at the end of 10 days. Mix 4 could be crushed by hand.
Except for mix 1, these were extremely lean mixes relative to normal con-
crete compositions.
221
-------
r-o
70
-o Plain Mortar
Plain Mortar 4- Spent Sorbent
a Plain Mortar + Gypsum
50 -
S. 40
£ 30
8 20
o.
o
o
Curve 716809-A
20.7MPa
(3000psi)
10 20 30 40
Age of Mix, Days
50
60
Figure F-2 - Effect of Simulated Spent Sorbent
on the Compressive Strength of
Cement Mortars
10
0
0.0 0.2 0.4 0.6 0.8 LO
Gypsum/Cement Ratio
Figure F-3 - Effect of Aging on the Compressive
Strength of Gypsum Mortars
-------
A comparison was made of the compressive strengths of 7.6 cm by 15 cm
(3 in. by 6 in.) cylinders made with and without spent stone. The mix
composition was as follows:
Type I Portland cement, g 1000
White sand, g 2750
Fly ash, g 200
Water, ml 700
To one batch of the above mix was added 100 g of spent stone (oxidized
sulfided limestone from test L-l) after it had been slaked in 200 ml of
v.'atcir. The total water content of the two batches of mix was the same
at the beginning of the test. The average compressive strength of three
cylinders containing spent stone was 18.1 MPa (2627 psi) versus 16.0 MPa
(2323 psi) for the cylinders of plain mix. This is an increase of
11 percent, which shows that fly ash/spent sorbent compacts are possible.
This test may be regarded as the first of the tests of utilizing spent
sorbent as part of the aggregate in normal concrete.
Subsequent tests to explore the effect of mix composition on com-
pressive strength are summarized in Tables F-5 and F-6. Mixes including
CAFB stone were made by first slaking the stone and allowing the product
to cool to ambient temperature. Slaking was vigorously exothermic,
reaching temperatures of 105°C. The test specimens were 5.08 cm (2 in.)
cubes. Figure F-4 shows the relative locations of the mixes on a tri-
angular diagram from which the sand has been omitted inasmuch as it was
held at a constant ratio to the cement.
All of the mixes showed a general trend toward increased compressive
strength as they aged. This is shown in Figure F-5. Mix A, the normal
cement mortar mix, developed a 28-day strength of 45 MPa (6530 psi).
Adding fly ash to the level of 31.2 wt % on cement while keeping other
ratios constant resulted in an increase in 28-day strength to 50 MPa
(7250 psi). However, adding spent stone to the level of 66.7 percent
on cement while keeping the sand/cement rates constant at 2.75 resulted
223
-------
in a sharp drop in 3-day compressive strength. The trend on aging indi-
cated a 28-day strength of 14 MPa (2030 psi). The drop may be due to
the sharp increase in water content as discussed below.
Table F-5
COMPRESSIVE STRENGTH OF 5.08 cm (2 in.) CUBES MADE WITH THE
MIXES IN TABLE F-6
Age at Test,
days
Compressive Strengths, kPa (psi)
Mix
A
Mix
B
Mix
C
Mix
D
Mix
E
Mix
F
Mix
Fl
Mix
F2
1
3
7
11
14
27
28
22850
(3314)
26890
(3900)
35420
(5137)
44200
(6410)
41070
(5957)
-
27850
(4040)
36040
(5227)
37440
(5430)
48310
(7007)
49900
(7238)
7690
(1115)
-
11340
(1645)
12060
(1750)
-
-
1520
(220)
5760
(835)
8030
(1165)
9240
(1340)
-
-
12340
(1790)
790
(115)
930
(135)
1620
(235)
-
2520
(365)
-
3520
(510)
240
(35)
210
(30)
930
(135)
-
1100
(160)
-
1450
(210)
"
-
2390
(346)
-
2965
(430)
-
3170
(460)
Mixes D to F, in which the cement/fly ash ratio was held constant
at 3.2 and the sand/cement ratio at 2.75, showed a continued drop in
strength with the increase in the relative amount of spent stone. Mix D
was observed at 90 days to have developed a strength of 22 MPa (3190 psi)
The ultimate strength of the other mixes appeared to be no more than
4 MPa (580 psi).
224
-------
Table F-6
o,.
-------
Dwg. 6358A31
X=Type I Portland Cement
Y = Spent Stone (CAFB
Regenerator Stone)
Z=Flyash
x Mix A
22850
MixB
7850
MixC
7690
Notes: 1. Compositions are in weight % of primary mix (X +Y +Z)
2. Numbers on the graph are 3-day compressive strengths of
2" cubes in kilopascals
3. Sand/ cement ratio is 2.75
Figure F-4 - Effect of Mix Design on 3-day Compressive Strength
226
-------
Curve 716808-A
Mix Composition, wt %
A B C D E
-o DO A •> <]
100 76 60 50 36 19 19
0 0 40 34 53 75 75
0 24 0 16 11 6 6
Symbol
Type I Portland Cement
Spent Sorbent
Fly Ash
10
40 50 60
Age of Mix, Days
Figure F-5 - Effect of Pilot Plant Spent Sorbent on the
Compressive Strength of Fly Ash/Cement
Mortars
Specimens of composition F, after soaking in water three days,
cracked into several pieces, and the color of the water became yellowish.
A repeat batch was made and, instead of being cured in water, this batch
was kept in the mold and covered with wet cloth. When the strength was
found still to be low (0.210 MPa or 30 psi) at 3 days, observations on
mix F were terminated in favor of mixes with a lower total water content.
The specimens were also cured differently from mix F. Instead of ponding
the specimens in water, specimens of mix F and F~ were air cured after
their removal from the mold. Air curing enhanced the compressive strength
at the reduced water content.
Normally, hydration of cement paste proceeds best when the capillary
spaces in the paste are water filled. Curing by ponding works well for
227
-------
the high cement content mixes, but for very low cement content the
hydraulic pressure created in the capillaries exceeds the early mechanical
strength of the specimens, causing them to fall apart.
Table F-5 shows that mix F_ developed a compressive strength 18 per-
cent greater at 14 days than did mix E, even though the stone/cement ratio
was increased from 1.5:1 to 4:1. This result was obtained by simple
changes in process technique: lower total water/cement ratio plus air
curing rather than water curing. Table F-6 shows the distribution of the
water added between slaking water and free water. Normally, a total of
0.5 part of water/part of cement is used. Some water, however, is needed
for plasticity. With the low cement/stone ratio, the amount of water
needed to achieve a plastic slurry greatly exceeds that needed for hydra-
tion of the cement. Even so, the apparently slight reduction from 5.2
for mix F to 3.7 for mix F» in free water/cement may be credited with at
least some of the strength gain. Mixes E and F_, incidentally, developed
about the same 28-day strength, 3.5 versus 3.2 MPa (510 versus 460 psi),
suggesting that increasing spent stone in the total of stone 4- cement
+ fly ash from 53 to 75 percent did not change the inherent strength of
the mix.
The next step was to explore the other end of the composition range,
'•.amely, zero cement content.
Mixes of CAFB stone and coal fly ash were made as per the propor-
tions shown below:
Mix 1 Mix 2 Mix 3
CAFB stone, g 100 100 100
Fly ash, g 100 50 25
Water, ml 120 115 105
The mixes were air cured and were found not only to attain final set but
also to possess sufficient handling strength when the specimens had dried.
Qualitatively, mix 3 was adjudged the least permeable to water. Com-
pressive strengths on these specimens were not determined because they
228
-------
were too irregular. The effect of using stone ground to 63 um (-230 mesh)
was tested in three 5.08-cm (2-in.) cubes of the following compositions
on a weight basis:
Mix 4A 100 Stone + 23 Fly ash
Mix 4B 100 Stone + 10 Fly ash
Mix 4C 100 Stone + 36 Fly ash
Mix 4A was based on calculations from the typical composition of fly ash
and that of CAFB-9 stone. The object was to have the free CaO content the
same as that of the total calcium in Portland cement. This put the Al-O-
also in the correct range but left the SiO^ much lower than in cement.
Mix 4B was arbitrarily set for half this proportion of fly ash while 4C
was 50 percent greater. The stone was first slaked in water, using a
water/total solids ratio of 1.125. After cooling, the slaked stone was
blended thoroughly with the desired amount of fly ash. Water content
was increased to 1.14 to improve plasticity. The specimens were cured
in a humidifier at 95 percent R.H. and 21°C. Compressive strengths are
in Table F-7.
Table F-7
COMPRESSIVE STRENGTH OF CAFB-9 REGENERATOR
STONE/FLY ASH MIXES, kPa (psi)
Mix
4A
4B
4C
7
758 (110)
1131 (164)
648 (94)
Age,
Days
14
889
1413
869
(129)
(205)
(126)
I 28
1144 (166)
1689 (245)
1034 (156)
These data show that mix 4B, containing 10 percent fly ash, continued to
develop higher compressive strength on aging than did mixes containing
more or less fly ash.
229
-------
The values were still low but suitable for landfill. These same
samples were subjected to leaching tests in chunk and in powder form.
Sulfate levels in the leachate from 10 g in 100 ml after 256 hours,
which were in the range 2000 to 3400 ppm, approached the saturation value
for CaSO,. Use of Student's t test on the means for the chunk data ver-
sus the powder data show that the probability of finding as large a dif-
ference in means (1944 versus 1747 ppm) by chance was about 5 percent,
supporting the view that processing the spent sorbent into large blocks
should reduce the leach rate of sulfate ions.
The next experiment was a detailed investigation of the effect of
particle size using narrow cut fractions. The mix composition was 765 g
stone, 85 g fly ash, and 1000 g HO, except for the -60 + 80 mixes, which
used 950 g. The amount of water used was based on the plasticity of the
mix. Duplicate measurements were made to permit statistical analysis of
the variance of the results. The data are in Table F-8, the analysis of
variance in Table F-9.
Table F-8
COMPRESSIVE STRENGTH OF 5.08 cm (2 in.) CUBES MADE WITH CAFB-9
REGENERATOR STONE AND COAL FLY ASH
Size Range, pm
-250 +177
-177 + 175
-125 + 88
-88
Curing Time, Days
7
100; 11 3
45;43
20; 25
70; 40
14
2 34; 260
170;168
220;258
160;148
21
188;
205;
210;
125;
161
313
210
148
66
193;213
340; 370
398;405
208; 2 35
Notes:
1. Coal fly ash is from Duquesne Light Company's Phillips Plant.
2. Mix composition is 765 g stone, 85 g fly ash, 1000 g water,
except for -250 + 177 urn mixes, where only 950 g water was used.
3. Compressive strengths in psi.
230
-------
Table F-9
ANALYSIS OF VARIANCE FOR COMPRESSIVE STRENGTH DATA ON CAFB
STONE/FLY ASH CUBES
Main Effects
Size
Age
Interaction
Size & Age
Remainder
TOTAL
Sum of Squares
27522.8
231210.3
67926.5
9018.5
335678.1
Degrees of
Freedom
3
3
9
16
31
Mean
Square
9174.3
77040.1
7547.4
563.6
F. Ratio
16.28
136.73
13.39
Table F-8 presents the data on the effect of particle size range on
the compressive strength of compacts made from CAFB regenerator stone and
coal fly ash. Table F-9 contains a statistical analysis of these data.
Curing time was, as expected, highly significant, but the effect of par-
tical size range and the interaction of curing time and particle size
were also significant at better than the 99-percent level.
Closer examination of these findings revealed that the contribution
of the effect of replication to the total sum of squares of deviations
from the grand mean is relatively small. This is a formal way of stating
that the replicates in general were in close agreement (within 0 to 20 per-
cent), and this agreement prevailed over a wide range of compressive
strengths. In only two of the 16 pairs of results was the difference
greater than 20 percent, nor was there any trend to greater differences
with a decrease in particle size or curing time.
The grand mean of the compressive strengths was 1.290 MPa (187.2 psi),
Both early and late strength data deviated widely from the grand mean and,
therefore, made large contributions to the total sum of squares. The
statistical analysis was repeated after deleting the 7-day data to deter-
mine whether they were responsible for the effect of particle size range.
231
-------
In commercial use the early strength of concrete determines how soon
forms can be removed and load placed on the concrete. The strengths
obtained at 7 days were in all cases far below those for normal weight
concrete: 5.52 to 14.48 MPa (800 to 2100 psi). All three effects were
still found to be significant at better than the 95 percent level.
The original data support the view that intermediate particle sizes
-170 + 88 um gave higher compressive strengths, of the order of 2.76 MPa
(400 psi), about twice the strength of compacts made with -250 + 177 um
material. The gain is not sufficient, however, to justify processing
spent sorbent to -177 + 88 um, but grinding to at least 100 percent
less than 177 ym appears worthwhile.
It is also interesting that all four size ranges showed a drop in
strength at an intermediate age, followed by a recovery to generally
higher levels.
Since relatively low 60-day compressive strengths were obtained,
use of narrow particle size ranges of CAFB spent sorbent in blends with
coal fly ash appears unattractive. Other variations, however, need to be
explored. The large amount of water needed for plasticity is considered
a major contributor to the low strengths. Two methods available for
reducing water content are the use of a surfactant and the use of iso-
static pressing. We plan to employ these techniques in later tests.
Following the thought of producing synthetic aggregate from spent
sorbent, cubes previously made from -250 +177 and -125 + 88 ym CAFB
stone were crushed to -1.27 cm (-1/2 in.) for testing as a synthetic
aggregate. The mix composition was 100 g of Type I Portland cement, 175 g
sand, and 100 g crushed CAFB composite. Compressive strengths of single
specimens at 14 days were 8.69 MPa (1260 psi) and 5.38 MPa (780 psi),
respectively. These values are 2 to 5 times those for the original
stone/fly ash compacts but still below the normal concrete range.
232
-------
APPENDIX G
HIGH-TEMPERATURE FLY ASH BLENDING
Feasibility experiments were conducted whose object was to determine
whether stable solid compacts could be made by sintering mixes of sul-
fated limestone with fly ash and clay. The initial tests used -595
+ 420 pm oxidized sulfided limestone from Batch L-l made in the 10-cm
laboratory fluidized bed. Data for these tests are in Table G-l.
Three levels of additive — 20, 40, and 60 wt % of the blend — and
four levels of sintering temperature — 800, 900, 1000, and 1200°C —
were used. The mixes were heated for two hours in a stream of 3.3 £
nitrogen/min (7 cfh).
In the case of the fly ash additive, CF-3, the composition contain-
ing 40 percent OSL* + 60 percent fly ash, yielded a clinkerlike product
when sintered at 1000°C. With other compositions and sintering tempera-
tures the product was either a powder or a melt.
With clay additives all the compositions sintered to a solid mass
at all temperatures, but when these products were aged for three to
five days in the laboratory, they crumbled to powder without exception.
The results indicate that, within the experimental range studied,
compositions containing fly ash need to be heated to at least 1000°C
before sintering occurs.
Chemical analyses for sulfide and sulfate sulfur are also in
Table G-l and presented graphically in Figures 0-1 and G-2. In all cases
most of the sulfate sulfur was lost. An unexpected result was the
increase in sulficle sulfur with increased fly ash content. In contrast,
the residual sulfide content did not vary with ball clay content.
*OSL is oxidized sulfided limestone.
233
-------
Curve 712951-A
Curve 712952-A
N3
U>
30
in
E 20
10
a Before Sintering at 1000°C
for 2 hours in Nitrogen
O After Sintering
Calcium Sulfate Content
30
-o-
JL
20 40
Wt % Fly Ash
10
A Before Sintering
V After Sintering
Calcium SulfMe Content
I
20 40
Wt % Fly Ash
60
3
s
30
a Before Sintering §1200 °C
for 2 Hours in Nitrogen
O After Sintering
- Calcium Sulfate Content 30
20
10
_CL.
20 40
Wt * Clay
10
A Before Sintering
V After Sintering
- Calcium SulfMe Content
60
I
20 40
Wt % Clay
Figure G-l - Sulfur Retention in Sintered
Spent Sorbent/Fly Ash Mixes
Figure G-2 - Sulfur Retention in Sintered
Spent Sorbent/Clay Mixes
-------
The sulfur retention was examined, as shown in Table G-2. The cal-
culations include an allowance for the SCL content of the fly ash used,
which was 0.00874 g moles SO.,/100 g fly ash. The average sulfate reten-
tion was 6.3 percent. This is high by an amount unknown because the
yield of product per 100 g charge was not measured. The maximum weight
loss if all the sulfur were evolved as S0_ and no other changes occurred,
however, would work out to 21.6 g/100 g sorbent. For mix CF-2 with
60 percent sorbent, the maximum loss possible was 13.0 percent plus
0.3 percent from the SO,, in the fly ash.
Table G-2
SULFUR RETENTION IN SINTERED SPENT SORBENT/ADDITIVE MIXES
Specimen
CF-1
CF-2
CF-3
CF-4
CF-5
CF-6
Wt % Sorbent 80
Input, moles/100 g mix
60
40
80
60
Total SO,
Total S~
Total S
0.2015
0.0822
0.2837
0.1533
0.0616
0.2149
0.1051
0.0411
0.1462
0.1998
0.0822
0.2820
Cutout, moles/100 g product
Total S0,= 0.0090 0.0116 0.0087 0.0005
Total S= 0.0683 0.1444 0.1852 0.0224
Total S 0.0773 0.1560 0.1939 0.0319
Ratios, Output/Input
Total SO,
Total S~
Total S
0.0448
0.832
0.273
0.0760
2.344
0.726
0.0824
4.507
1.326
0.0478
0.273
0.114
40
0.1498 0.0999
0.0616 0.0411
0.2115 0.1410
0.0140 0.0037
0.0274 0.0224
0.0414 0.0261
0.0931 0.0368
0.445 0.546
0.196 0.185
Calculations show further that the sulfur levels found in CF-2 can be
interpreted as follows:
• The overall weight loss was 10.4 percent (which is within the
maximum above).
235
-------
• There was a 54 percent conversion of CaSO, to CaS.
• There was a 25 percent loss of sulfur from the original
CaS present.
• There was a 39 percent decomposition of CaSO, to CaO.
The identity of the reductant was a matter of speculation earlier. It
would appear that something in the fly ash was responsible, and at this
point, although the earlier thought of FeO involvement might still be
a contributor, it appeared more likely that the carbon content of the
fly ash was the source. The NEES fly ash showed a 7.6 percent LOT,
and 3.5 percent CCL. The difference of 4.1 percent may be interpreted
as carbon. Mix CF-2, therefore, contained only 82 percent of the car-
bon needed for the CaSO, reduction calculated.
A second set of sintering experiments was conducted on mixtures of
fly ash and gypsum and fly ash and CaS. These two calcium compounds
are potential end products of fuel desulfurization processes. The tests
were aimed at narrowing down the effective field of investigation of
blends of fly ash with actual spent stones.
Five compositions, three sintering temperatures, one to three sin-
tering times, and two kinds of ambient atmospheres were used. Discs
were pressed in a 1-in diameter die at about 2100 MPa (3000 psi) and
then sintered. Data are presented in Table G-3 for the gypsum mixes
and Table G-4 for the CaS mixes.
Colors developed in the specimens showed a gradation from light to
dark (gray, yellow, brown, black) with increasing temperature. The
specimens were characterized as fragile, coherent, or strong in accord-
ance with their behavior when handled. There was a trend toward greater
cohesiveness as the sintering temperature increased.
The density data did not preaent a clear picture graphically so a
statistical analysis was made. The data matrix is unbalanced (i.e.,
incomplete) because of our desire to minimize the number of tests. The
complete matrix would have required 360 tests (2 sorbent types x 5 fly
ash contents x 3 sintering times x 3 sintering temperatures x 2 atmosphere
236
-------
Table G-3
Dwg.2«2C70
EFFECT OF FLY ASH CONTENT ON SULFUR RETENTION IN SINTERED CaSO, MIXES3
Composition, wt%
Sample
No.
75-CF-7
75-CF-8
75-CF-9
75-CF-10
75-CF-ll
Gypsum
80
70
60
50
40
Fly Ash
20
30
40
50
60
Temp.,
°C
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
Time,
hrs
1
2
4
2
4
2
2
4
4
2
1
2
4
2
4
2
2
4
4
2
1
2
4
2
4
2
2
4
4
2
1
2
4
2
4
2
2
4
4
2
1
2
4
2
4
2
2
4
4
2
Nitrogen
Flow,
scfh
4
0
4
0
4
0
4
0
4
0
Product
Character
Coherent
Coherent
Coherent
Coherent
Coherent
Strong
Coherent
Fragile
Fragi le
Strong
Coherent
Coherent
Coherent
Coherent
Coherent
Strong
Coherent
Powdered
Powdered
Coherent
Coherent
Coherent
Coherent
Coherent
Coherent
Strong
Fragile
Powdered
Powdered
Fragile
Fragile
Fragile
Fragile
Coherent
Fragile
Strong
Fragile
Powdered
Powdered
Strong
Fragile
Fragile
Fragile
Coherent
Fragi le
Strong
Fragi le
Fragile
Coherent
Strong
Color
White
White
White
Light gray
Light gray
Yellow white
Gray
Yellow gray
Light brown
Dark gray
White
White
White
Light gray
Light gray
Brown
Gray
Yellow white
Light brown
White
Light gray
Light gray
Light gray
Yellow gray
Dark brown
Gray
Yellow gray
Light brown
Black
Light gray
Light gray
Light gray
Light gray
Yellow gray
Yellow brown
Dark gray
Yellow gray
Light brown
Dark gray
Light gray
Light gray
Light gray
Light gray
Yellow gray
Dark brown
Dark gray
Dark gray
Light brown
Black
Density.
g/cm
1.761
1.770
1.400
2.118
2.094
2.091
1.489
1.393
1.644
1.103
1.735
1.741
1.226
1.992
1.835
1.783
1.364
1.265
1.254
0.978
1.552
1.509
1.044
1.797
1.345
2.318
1.128
1.036
0.998
0.917
1.386
1.248
0.977
1.667
1.151
1.572
0.996
1.018
0.962
1.664
1.202
1.184
0.999
1.449
1.142
Melted
1.016
0.978
1.045
Sintem
Composition, wt%
CaSO.
4
77.99
71.85
58.42
61.65
59.20
2.38
71.64
60.76
57.52
23.48
-
-
-
-
0.85
22.85
1.70
3.54
0.34
CaS
1.35
0.36
0.45
26.24
0.45
1.44
1.26
0.27
1.26
1.44
-
-
-
0.36
0.63
0.72
0.54
0.45
a Fly Ash was trom the New tnyland Electric System.
237
-------
Table G-4
EFFECT OF FLY ASH CONTENTS ON SULFUR RETENTION IN SINTERED CaS MIXES
Composition, wt%
Simple
No.
75-CF-U
75-CF-U
7S-CF-14
7VCF-15
7S-CF-16
CIS
Calcium
Sulflde
«
70
to
SO
40
Fly Ash
20
»
40
50
60
Temp.,
•c
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
1000
1100
1200
Time,
hrs
2
4
2
4
2
2
4
4
2
2
4
' 2
4
2
2
4
4
2
2
4
2
4
2
2
4
4
2
2
4
2
4
2
2
4
4
2
2
4
2
4
2
2
4
4
2
Nitrogen
Flow.
scfh
4
0
4
0
4
0
4
0
4
0
Product
Character
Strong
Coherent
Strong
Strong
Strong
Strong
Strong
Strong
Strong
Strong
Strong
Coherent
Coherent
Coherent
Coherent
Strong
Coherent
Strong
Coherent
Coherent
Strong
Strong
Coherent
Coherent
Coherent
Coherent
Strong
Coherent
Coherent
Strong
Coherent
Strong
Coherent
Coherent
Coherent
Coherent
Coherent
Coherent
Coherent
Coherent
Coherent
Fragile
Strong
Coherent
Coherent
Fragile
Coherent
Coherent
Coherent
Fragile
Fragile
Fragile
Fragile
Strong
Coherent
Color
Gray
Gray
Gray
Gray
Gray, tinge of yellow
Gray
Gray
Gray, tinge ol yellow
Light brown
Dark gray
Gray
Gray
Gray
Dark gray
Gray, yellow tinge
Gray
Gray
Gray, yellow tinge
Brown
Dark gray
Gray
Gray
Dark gray
Dark gray
Gray.l yellow tinge)
Dark gray
Gray
Gray, (yellow tinge I
Brown
Black
Gray
Gray
Dark gray
Dark gray
Gray, tinge of yellow
Dark gray
Dark gray
Dark gray
Brown
Black
Dark gray
Gray
Dark gray
Dark gray
Dark gray, tinge ol
yellow
Dark gray
Dark gray
Gray,
Brown
Dark gray
Density.
g/cm
1:575
1.555
1.612
1.578
1.799
1.652
1.626
1.534
1.576
1.603
1.834
1.534
1.485
1.772
1.521
1.554
1.502
1.537
1.475
1.526
1.554
1.566
1.499
1.449
1.513
1.429
1.365
1.260
1.467
1.473
1.306
1.499
1.021
1.461
1.386
1.446
1.412
1.394
1.359
1.402
1.4»
1.159
1.409
1.087
1.423
1.343
1.413
1.354
1.300
1.214
1.3*4
1.390
1.088
1.357
0.984
Composition, wt*
CaS04
1.19
1.53
0.85
1.02
1.02
11.06
1.02
2.55
0.85
1.36
6.13
2.72
-
-
-
-
2.04
1.11
1.70
l'.^4
11. W
US
66.32
70.86
44.96
75.80
71.04
47.34
53.89
63.23
62.60
62.69
41.67
30.W
-
~
-
-
34.63
34.10
39. K
ti.V>
U.36
Notts
1. Fly ash MslramtrMNM Enffcnd {Metric Sytltn.
238
-------
types x 2 replicates). The actual number of density determinations
available was 103. A series of F-tests were made by extracting bal-
anced submatrices from Tables G-3 and G-4. The level of significance
chosen was 5 percent, meaning that if the ratio of the variance of the
effect under test to the error variance exceeded the applicable standard
F-value, the effect would be termed significant since this result would
be obtained by chance less than 5 percent of the time. All such results
are marked with an asterisk in the ANOVA (analysis of variance) tables.
The first step in the analysis was to obtain an estimate of the
error variance. This is best obtained where tests were replicated. One
such set was available, as shown in Tables G-5a and G-5b, covering CaS/
fly ash mixes sintered in static nitrogen for four hours. Here, fly
ash content was shown to be significant while temperature was not.
Since the fly ash and temperature interaction was also insignificant
it can be combined with the error variance to get a revised error esti-
mate. The value obtained was 0.00878 with 14 degrees of freedom (d.f.).
This does not change the conclusions on significance since the numerical
values for the F-ratios are changed only slightly. For fly ash content,
F becomes 6.18 versus a table value of 3.11.
Table G-5a
EFFECT OF SINTERING TEMPERATURE AND FLY ASH CONTENT ON THE DENSITY
OF CaS/FLY ASH MIXES SINTERED IN STATIC NITROGEN FOR FOUR HOURS
Weight %
Fly Ash
20
30
40
50
60
Mean Density
Temperature, °C
1000
1.626; 1.534
1.537; 1.475
1.467; 1.473
1.402; 1.430
1.383; 1.390
1.472
1100
1.576; 1.603
1.526; 1.554
1.306; 1.499
1.159; 1.409
1.088; 1.357
1.408
Mean
Density
1.585
1.523
1.436
1.350
1.304
1.440
-------
Table G-5b
ANOVA FOR DATA IN TABLE G-5a
Source of
Variation
Fly Ash Content, W
Temperature, T
W x T
Error
Total
Sum of
Squares
0.2173
0.0205
0.0296
0.0934
0.3608
Degrees of
Freedom
4
1
4
10
19
Mean
Square
0.0543
0.0205
0.0074
0.00934
Calculated
F-Ratio
5.82*
2.19
0.79
F-Ratio
-------
Table G-6a
EFFECT OF MIX COMPOSITION, TEMPERATURE, AND TYPE OF SINTERING
ATMOSPHERE ON THE DENSITIES OF MIXES SINTERED FOR FOUR HOURS
Weight %
Fly Ash
20
30
40
50
60
Gypsum
Flowing '
Nitrogen
1000°C
1.400
1.226
1.044
0.977
0.999
1100 °C
2.094
1.835
1.345
1.151
1.142
Static
Nitrogen
1000°C
1.393
1.265
1.036
1.018
0.978
1100 °C
1.644
1.254
0.998
0.962
1.045
Calcium Sulfide
Flowing
Nitrogen
1000 °C
1.555
1.485
1.449
1.386
1.343
1100 °C
1.578
1.521
1.429
1.412
1.354
Static j
Nitrogen3 !
1000°C
1.580
1.506
1.470
1.416
1.387
1100 °C
1.590
1.540
1.403
1.284
1.223
NOTE;
aThe values for calcium sulfide-static nitrogen are averages of
the repeats shown in Table G-5.
Tables G-8a and G-8b explore the significance of sintering time
for the CaS mixes. Neither time nor temperature was significant, and
the effect of fly ash content was shown to be limited to a linear
relationship.
Compositing the four interaction terms led to a smaller estimate
of the error variance: 0.0026 with 13 d.f. Use of this value made the
effect of time significant at the 1 percent level: 10.4 versus 9.07.
This result means further tests are needed to define the contribution
of sintering time.
One further effect on the case of the CaS mixes can be checked -
the effect of atmosphere - using the data for four hours of sintering
time. Tables G-9a and G-9b present this analysis. We concluded that
only the weight % fly ash was significant.
241
-------
Table G-6b
ANOVA FOR DATA IN TABLE G-6a
Source of
Variation
Temperature, T
Atmosphere, A
Mixture, M3
Weight %, Wb
T x A
T x M
A x M
T x W
A x W
M x W
T x A x M
T x A x W
T x M x W
A x M x W
T x A x M x W
Total
Error
D.F.
1
1
1
4
1
1
1
4
4
4
1
4
4
4
4
39
10
Sum of
Squares
0.0894
0.0751
0.4213
1.050
0.1106
0.1413
0.0568
0.0902
0.0075
0.2383
0.0431
0.0091
0.0356
0.0318
0.0403
2.4404
0.0934
Mean
Square
0.0894
0.0751
0.4213
0.2624
0.1106
0.1413
0.0568
0.0226
0.0019
0.0596
0.0431
0.0023
0.0089
0.0080
0.0101
0.0093
Calculated
F- Ratio
9.6*
8.1*
45.3*
28.0*
11.9*
15.2*
6.1*
2.4
0.2
6.4*
4.6
0.2
0.96
0.86
1.1
F-Ratio
(§5% Level
4.96
4.96
4.96
3.48
4.96
4.96
4.96
3.48
3.48
3.48
4.96
3.48
3.48
3.48
3.48
aRefers to type of sorbent.
Refers to fly ash content.
Compositing interaction terms led to 0.0022 as the internal estimate
of error, with 13 d.f. Use of this smaller error did not change any of
the conclusions on significance.
Turning next to the gypsum results, Tables G-lOa and G-lOb contain
an analysis of the effect of sintering temperature and weight of fly ash
on density. In contrast to the CaS results, and, as suspected from the
analyses of Tables G-5b and G-6b, temperature had a significant effect
242 '
-------
Table G-7a
EFFECT OF SINTERING TEMPERATURE AND FLY ASH CONTENT
ON THE DENSITY OF CaS/FLY ASH MIXES SINTERED
IN FLOWING NITROGEN FOR TWO HOURS
Weight Percent
Fly Ash
Mean Density
20
30
40
50
60
1000
1.575
1.534
1.499
1.461
1.423
1.498
Temperature, °C
1100
1.612
1.772
1.513
1.446
1.413
1.551
1200
1.799
1.554
1.365
1.394
1.300
1.482
Mean Density
1.662
1.620
1.459
1.434
1.579
1.510
Table G-7b
ANOVA FOR THE DATA OF TABLE G-7a
Source of
Variation
Temperature, T
Weight % Fly Ash, W
T x W
Total
Error
Degrees of
Freedom
2
4
8
14
10
Sum of
Squares
0.0130
0.1826
0.0759
0.2715
0.0934
Mean
Square
0.0065
0.0457
0.0095
0.0093
Calculated
F-Ratio
0.7
4.91*
1.02
F-Ratio
@ 5% Level
4.10
3.48
3.07
on the density of gypsum/fly ash sinters. The temperature x weight %
fly ash interaction was also significant. The actual effect of fly ash
content on density is shown in Table G-lOa.
Examination of the data in Table G-lOa reveals that the data point
at 1200°C for 40 percent fly ash is out of line with the rest of the
table. At each level of fly ash, density at 1100'C is greater than at
243
-------
Table G-8a
EFFECT OF SINTERING TIME ON THE DENSITY OF CaS/FLY ASH
MIXES SINTERED IN FLOWING NITROGEN
Weight %
Fly Ash
20
30
40
50
60
Mean density
Sintering Time, hr
2
Temperature, °C
1000
1.575
1.534
1.499
1.461
1.423
1.498
1100
1.612
1.772
1.513
1.446
1.413
1.551
4
Temperature, °C
1000
1.555
1.485
1.449
1.386
1.343
1.444
1100
1.578
1.521
1.429
1.412
1.354
1.459
Mean Density
1.580
1.578
1.472
1.426
1.383
1.488
Table G-8b
ANOVA FOR DATA IN TABLE G-8a
Source of Variation
Time, t
Temperature, T
Weight % Fly Ash, W
Linear
Remainder
t x T
t x W
T x W
t x T x W
Total
Error
Degrees
of Freedom
1
1
4
1
3
1
4
4
4
19
10
Sum of
Squares
0.0271
0.0058
0.1264
0.1187
0.0077
0.0018
0.0084
0.0139
0.0093
0.1927
0.0934
Mean
Square
0.0271
0.0058
0.0316
0.1187
0.0026
0.0018
0.0021
0.0035
0.0023
0.0093
Calculated
F-Ratio
2.91
0.6
3.4
12.8*
0.28
0.2
0.2
0.4
0.2
F-Ratio @
5% Level
4.96
4.96
3.48
4.96
3.71
4.96
3.48
3.48
3.48
244
-------
Table G-9a
EFFECT OF TYPE OF SINTERING ATMOSPHERE ON THE
DENSITY OF CaS/FLY ASH MIXES
SINTERED FOR FOUR HOURS
Weight %
Fly Ash
20
30
40
50
60
Mean Density
Type of Sintering Atmosphere
Flowing
Nitrogen
Temperature, °C
1000
1.555
1.485
1.449
1.386
1.343
1.444
1100
1.578
1.521
1.429
1.412
1.354
1.459
Static
Nitrogen3
Temperature, °C
1000
1.580
1.506
1.470
1.416
1.387
1.472
1100
1.590
1.540
1.403
1.284
1.223
1.408
Mean Density
1.574
1.513
1.438
1.374
1.327
1.445
aThe entries under static nitrogen are all averages of the two repeats
which were shown in Table G-4.
Table G-9b
ANOVA FOR THE DATA OF TABLE G-9a
Source of
Variation
Temperature, T
Atmosphere, A
Weight %, W
T x A
T x W
A x W
T x A x W
Total
Error
Degrees
of Freedom
1
1
4
1
4
4
4
19
10
Sum of
Squares
0.0028
0.0006
0.1616
0.0080
0.0093
0.0046
0.0065
0.1934
0.0934
Mean
Square
0.0028
0.0006
0.0404
0.0080
0.0023
0.0011
0.0016
0.0093
Calculated
F-Rat io
0.3
0.06
4.34*
0.86
0.2
0.1
0.2
F-Rat io
@ 5% Level
4.96
4.96
3.48
4.96
3.48
3.48
3.48
245
-------
Table G-lOa
EFFECT OF SINTERING TEMPERATURE AND FLY ASH CONTENT ON THE DENSITY
OF GYPSUM/FLY ASH MIXES SINTERED FOR TWO HOURS IN FLOWING NITROGEN
Weight %
Fly Ash
20
30
40
50
Mean Density
Temperature, °C
1000
1.770
1.741
1.509
1.248
1.567
1100
2.118
1.992
1.797
1.667
1.894
1200
2.091
1.783
2.318
1.572
1.941
Mean Density
1.993
1.839
1.875
1.496
1.801
Table G-lOb
ANOVA FOR THE DATA OF TABLE G-lOa
Source of
Variation
Temperature, T
Weight %, W
T x W
Total
Error
Degrees
of Freedom
2
3
6
11
10
Sum of
Squares
0.3316
0.4108
0.2123
0.9547
0.0934
Mean
Square
0.1658
0.1369
0.0354
0.0093
Calculated
F-Ratio
17.83*
14.7
3.81*
F-Ratio
@ 5% Level
4.10
3.71
3.22
1000 or 1200°C. At 40 percent fly ash the density trend from 1000 to
1100"C is the same but then shows a marked increase, rather than a
decrease, in going to 1200°C. A repeat measurement at these conditions
is needed to determine whether the density increase is real.
Tables G-lla and G-llb examined the effect of sintering time on the
density of gypsum/fly ash mixes sintered in flowing nitrogen at 1000°C.
246
-------
Table G-lla
EFFECT OF SINTERING TIME ON THE DENSITY OF GYPSUM/FLY ASH MIXES
SINTERED AT 1000°C IN FLOWING NITROGEN
Weight %
Fly Ash
20
30
40
50
60
Mean
Sintering Time, hr
1
1.761
1.735
1.552
1.386
1.202
1.527
2
1.770
1.741
1.509
1.248
1.184
1.490
4
1.400
1.226
1.044
0.977
0.999
1.129
Mean Density
1.644
1.567
1.368
1.204
1.128
1.382
Table G-llb
ANOVA FOR DATA IN TABLE G-lla
Source of
Variation
Time , t
Weight %, W
t x W
Total
Error
Degrees
of Freedom
2
4
8
14
10
Sum of
Squares
0.4837
0.5975
0.0507
1.1319
0.0934
Mean
Square
0.2419
0.1494
0.0063
0.0093
Calculated
F-Rat io
26*
16*
0.7
F-Rat io
05% Level
4.10
3.48
3.07
Time had a greater effect than fly ash content. The density decreases
or prolonged sintering may be attributed to the evolution of H?0, SO,,
and oxygen on decomposition of gypsum.
Tables G-12a and G-12b present an analysis of the effect of type of
sintering atmosphere on density. All three main effects were significant,
as was the interaction on temperature and atmosphere.
247
-------
Table G-12a
EFFECT OF TYPE OF SINTERING ATMOSPHERE OR THE DENSITY OF
GYPSUM/FLY ASH MIXES SINTERED FOR FOUR HOURS
Weight %
Fly Ash
20
30
40
50
60
Mean Density
Type of Sintering Atmosphere
Flowing Nitrogen
Temperature, °C
1000
1.400
1.226
1.044
0.977
0.999
1.129
1100
2.094
1.835
1.345
1.151
1.142
1.513
Static Nitrogen
Temperature, °C
1000
1.393
1.265
1.036
1.018
0.978
1.138
1100
1.644
1.254
0.998
0.962
1.045
1.181
Mean Density
1.633
1.395
1.106
1.027
1.041
1.240
Table G-12b
ANOVA FOR DATA IN TABLE G-12a
Source of
Variation
Temperature, T
Atmosphere, A
Weight %, W
T x A
T x W
A x W
T x A x W
Total
Error
Degrees
of Freedom
1
1
4
1
4
4
4
19
10
Sum of
Squares
0.2277
0.1312
1.125
0.1459
0.1168
0.0349
0.0427
1.8242
0.0934
Mean
Square
0.2277
0.1312
0.2813
0.1459
0.0292
0.0087
0.0107
0.0093
Calculated
F-Ratio
24.5*
14.1*
30.2*
15.7*
3.1
0.9
1.2
F-Ratio
@ 5% Level
4.96
4.96
3.48
4.96
3.48
3.48
3.48
248
-------
Another response to sintering on which data were obtained was
residual sulfide and sulfate content. Figures G-3 through G-6 show the
change in sulfur content as a function of fly ash content, which may be
summarized as follows:
• For gypsum in flowing nitrogen, at 1100°C/2 hr the residual
CaSO, decreased linearly with an increase in fly ash con-
tent. At 1200°C/2 hr, there was a sharp decrease in
CaSO, even at 30 percent fly ash and over 95 percent loss
of sulfate at 60 percent fly ash.
• For gypsum in static nitrogen, sulfur losses were greater
at all levels of fly ash content than were those for
flowing nitrogen.
• Sulfide content reached levels to 1.5 percent, but there was
no clear effect of fly ash content.
• For CaS in flowing nitrogen, there was a decrease of about
12 percent in sulfide sulfur, independent of fly ash con-
tent in the range of 20 to 60 percent. There appeared to
be a small increase in sulfate content with increase in
fly ash content to the level of 1.7 percent CaS.
• For CaS in static nitrogen there was a sharp decrease in
sulfide sulfur with an increase in fly ash content with
over 99 percent rejection at 60 percent fly ash. The
1100°C/4 hr data showed about 9 percent
One can only speculate at this point on the gypsum results. If
there were to be any CaSO, decomposition, use of flowing nitrogen to
sweep away the S0?/0? formed should have favored it. The loss of sul-
fate sulfur was not via reduction to sulfide and must, therefore, have
been as SO™ . One possibility is that in static nitrogen the temperature
of the sinter may have been higher than the thermocouple indication.
This reasoning is supported by cross plots in Figures G-7 and G-8 showing
the effect of temperature on the sulfur retention of gypsum fly ash mixes
versus CaS/ fly ash mixes. Increasing temperature from 1000 to 1100°C
led to a small loss of sulfate sulfur from the gypsum mixes, but on
249
-------
Curve 690963-A
Curve 690959-A
to
O
80
70
60
140
i »
20
10
0
Initial Mix
< 1000'C.l hr
v 1000°C. 2 hrs
10
%
E
_3
u
O
o llOO'C. 2 hrs
0 1100°C, 4 hrs
o 1200°C. 2 hrs
20 40
Wt * Fly Ash
60
20 40
Wt * Fly Ash
A Initial Mix
> 10M°C. 4 hr
0 1100°C. 4 hr
o 1ZOO°C, 2 hr
10
a
3 4
20 40
Wt % Fly Ash
20 40
Wt % Fly Ash
60
Figure G-3 - Sulfur Retention in Gypsum/Fly
Ash Mixes Sintered in Flowing
Nitrogen
Figure G-4 - Sulfur Retention in Gypsum/Fly
Ash Mixes Sintered in Static
Nitrogen
-------
Curve 690960-A
Curve 690961-A
N>
Ul
70
60 -
-| 50
E
.2 40 (-
o
5 30
20
10 r
0
A Initial Mix
v 1000°C. 2 hr
a 1100°C. 2 hr
0 1100°C. 4 hr
o 12fJO°C. 2 hr
0 20 40
Wt. % Fly Ash
60
20 40 60
Wt. % Fly Ash
80
70
60
O
s
12
10
8
6
4
2
0
0
0
-
0
-
0
-
&
0
i 1 1
20 40 60
Wt % Fly Ash
20 40 60
Wt % Fly Ash
Figure G-5 - Sulfur Retention in Calcium
Sulfide/Fly Ash Mixes Sin-
tered in Flowing Nitrogen
Figure G-6 - Sulfur Retention in Calcium
Sulfide/Fly Ash Mixes Sin-
tered in Static Nitrogen
-------
Curve 69OT62-A
Curve 690%4-A
S3
01
ISJ
70
60
| 50
3
I 40
s30
20
10
0
A 20* Fly Ash. 2hr
v 30* Fly Ash. 2hr
c> 60* Fly Ash. 2hr
6
900 1000 1100 1200
Temperature, °C
1
3
ISI
900 1000 1100 1200
Temperature, °C
70
60
3
2»
20
10
A 20* Fly Ash
v 30* Fly Ash
> 60* Fly Ash
I 4
900 1000 1100 1200
Temperature. °C
900 1000 1100 1200
Temperature, °C
Figure G-7 - Effect of Temperature on Sulfur
Content of Gypsum/Fly Ash Mixes
Sintered in Flowing Nitrogen
Figure G-8 - Effect of Temperature on Sulfur
Content of Calcium Sulfide/Fly Ash
Mixes Sintered in Flowing Nitrogen
-------
further increase to 1200°C, major to over 95 percent loss of sulfate
sulfur occurred. In contrast, temperature was without effect on the
CaS mixes except at the lowest level (20%) of fly ash. Here a linear
increase with temperature was demonstrated. This might better be
regarded as a loss at 1000°C, decreasing with an increase in temperature.
Since no effect of temperature on CaS was demonstrated for 20 percent
fly ash, the losses in the case of gypsum are attributed to decomposition.
Another technique for carrying out the spent sorbent/fly ash reac-
tion is hot pressing, a modified sintering operation in which the speci-
men is subjected simultaneously to heat and pressure. A higher density
was expected with this technique than with conventional atmospheric
pressure sintering. Figure G-9 is a sketch of the apparatus. A reduc-
ing atmosphere is probably created at the surface of the pellet being
made, permitting a test of the hypothesis that air leakage in the first
set of sintering tests was responsible for the loss of sulfide sulfur.
The sketch shows a top plunger of graphite; in subsequent work, this was
changed to an aluminum rod.
Two compositions were evaluated:
• 80 percent CaS plus 20 percent fly ash (75-CF--17 series)
• 80 percent gypsum plus 20 percent fly ash (75-CF-18 series).
Each powder sample was thoroughly mixed, placed in the die, nnd heated
while pressure was applied to the upper plunger. Specimens were held
at selected temperatures and pressures for 30 minutes. The whole
assembly was then cooled, and measurements were made on the specimens
from which their density was calculated. These results are in
TabLes G-13 and C-14 and plotted in Figures G-10 and (".-11.
The specimens were obtained in the form of cylinders, 1.27 cm
(1/2 in) diameter and about 2.54 cm (1 in) long, with a graphite coating
outside. They appeared well compacted and sintered and did not dis-
integrate with aging as had specimens made by conventional sintering.
Both compositions showed an increase in density with an increase in hot-
pressing temperature. The effect of temperature, however, is more
253
-------
Graphite
Spacer
O
Dwg. 6358A32
Graphite Plunger (i" diameter)
Graphite Die
ZrC^Grog
Powder Mix
RF Heating
Graphite
O Crucible
Silica
Crucible
Figure G-9 - Schematic of Hot Pressing Unit
254
-------
KJ
Ol
Table G-13 0*9.1706806
CHARACTERISTICS OF HOT-PRESSED 80% CALCIUM SULFIDE/20% FLY ASH MIXES. SERIES 75-CF-l
Hot Pressing Conditions
Force.
N(lb)
3559 (800)
4448 (1000)
5338 (1200)
Nominal
Pressure.
MPa(psi)
28.1 (4074)
35.1 (5093)
42.1 (6112)
Temperature.
°C
850
950
1050
1200&
850
950
1050
850
950
1050
Product
Density,
g/cc
1.76
2.43
2.47
2.74
2.67
1.89
2.30
2.54
2.78
2.01
2.47
2.74
Compressive
Strength.
MPa(psi)
—
-
65.2(9450)
—
122.4(17750)
37.2(5400)
23.4(3400)
34.1(4950)
—
—
-
54.8(7950)
Composition. wt%
Total
Sulfur
17.98
17.86
-
15.64
—
—
—
—
—
25.54
18.65
16.04
Sulfide
0.11
0.16
0.48-.0.71
0.24
2.47;2.77
6.06
11.41
0.56
0.52
—
18.70;17.85
19. 68; 18. 82
0.40;0.56
1.55;1.12
Sulfate Calc
54.03;54.28
55.05
50.42
58.90
Calculated
Total Sulfur
ium Wt%3
18.18
18.53
17.31
19.90
41.21 30.06 16.37
27.45 35.
20.85
50.40
57.34
—
20.11;20.80
22.02
50.90
49.10
27 15.22
18.37
17.38
19.66
—
25.11
26.60
17.46
17.72
Notes:
a Initialvalue35.59wt%
b Specimen hot pressed for 15 minutes; all others. 30 minutes
-------
Dwg. 1706807
Table G-14
CHARACTERISTICS OF HOT-PRESSED 80% GYPSUM/20% FLY ASH MIXES. SERIES 75-CF-183
Hot Pressing Conditions
Force.
N(lb)
3559 (800)
5338(1200)
Nominal
Pressure,
MPa (psi)
28.1 (4074
42.1 (6112)
Temperature.
°C
Product
Density.
g/cc
850 2.03
1050 2.70
850 2.21
1050 2.85
Compressed
Strength.
MPa (psi)
Composition. wt%
Total
Sulfur
15.73
16.39
17.34
16.72
Sulfide
0.24
1.75
1.56
Sulfate Cak
56.90;54.26
46.23;43.95 29
57.76
56. 08; 56. 89
54.77C
42.88:45.08
Total Sulfur.
:ium wt%b
18.78
.66 16.80
18.99
16.24
Ui
All specimens hot pressed for 30 minutes
b Initial value 14.94 wt%
c Gravimetric determination
-------
Curve 716806-A
2.8
2.7
2.6
2.5
2.4
"s 2-3
"5-2.2
•t 2.1
C
0> 0 f,
o 2. u
1.9
1.8
1.7
1.6
1.5
-Highest Density
in Conventional
Sintering
-1200 Ib
-lOOOIb
-SOOIb
950° C
1*50° C
i i i
750 850 950 1050
Temperature, °C
600 800 1000 1200
Compacting Load, Ib
2.8
2.7
2.6
2.5
2.4
2.3'
2.2
2.1
2.0
1.9
1.8
1.7
1.6
1.5
s.
o
0>
o
1400
Figure G-10 - Density Data of the Hot-Pressed Specimens Having
the Composition of 80% CaS and 20% Fly Ash
i.urve 716807
i. u
2.9
2.8
2.7
2.6
%
e 2.5
o
^2.4
•| 2.3
S 2.2
2.1
2.0
1.9
1 8
1 | 1 |
-(3)
1 1 '
/
//
/ Aoo ib
1200lb /
/ /'
/ / -f-
/
/
/ V
/ / ^Highest Density
/in Conventional
Sintering
i i
i ! i
1 ' 1 ' 1 ' 1 '
(b)
^^^^ 1050° C
,
:
xx'xl50°C
1 1 L. I 1 ! ! i
3. U
2.9
2.8
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2.0
1.9
1 8
E
en
QJ
O
750 850 950 1050
Temperature, °C
1400
Compacting Load, Ib
Figure G-ll - Density Data of the Hot-Pressed Specimens Having
the Composition of 80% CaSO, and 20% Fly Ash
257
-------
pronounced than the effect of pressure. Figure G-10 indicates that for
the CaS/fly ash composition the effect of pressure is negligible when
the temperature is high, 950 to 1050°C, but significant at 850°C.
Figure G-ll for the CaSO,/fly ash compositions again shows a strong
effect of temperature. The effect of pressure is somewhat greater at
1050°C and somewhat smaller at 850°C than for CaS/fly ash.
For comparison the highest densities obtained by conventional
sintering are included in both the figures. Note that hot pressing
yields higher density when conducted at high temperature, for example,
850°C.
The specimens showed no degradation upon aging for several weeks,
in contrast with sintered specimens, which crumbled on aging. Table G-13
shows compressive strength data for one composition: 80 wt % CaS plus
20 percent fly ash. The values obtained are comparable to those of typical
cement composites. The data show considerable scatter but the combined
effect of higher temperature and pressure appears to be increased compres-
sive strength. The value of 122.4 MPa (17,750 psi) obtained at 800 lb/
1200°C may not be representative. Compressive strengths on the gypsum/fly
ash specimens were not measured; a comparison is made below in a subsequent
experiment.
Chemical analyses for some of the specimens were obtained as shown
also in Tables G-13 and G-14. The four gypsum/fly ash specimens showed
total sulfur contents in the range of 15.7 to 17.3 wt %, corresponding
to 66.8 to 73.6 wt % CaSO, versus 80 percent initially. The CaS/fly ash
specimens showed total sulfur contents in the range of 15.6 to 25.5 per-
cent, or 35.2 to 57.5 percent CaS versus 80 percent initially. Sulfide
was determined by an iodometric method, while sulfate was determined by
an ion-exchange method.. As a check on the validity of these methods,
the total sulfur was calculated from these values and compares well with
the total sulfur determined by a bomb method. As a further check, the
258
-------
sulfate was determined gravimetrically for one specimen (Table G-14, (c)).
While the gravimetric value was somewhat lower than the ion-exchange
values, the relative agreement indicates no serious interferences from
the anions.
The sulfur data are plotted in Figures G-12 through G-14. The
gypsum results are as expected: low CaS content and decreasing CaSO,
content with an increase in temperature. The CaS results were unexpected.
The 78 wt % CaSO, found at 850°C/5.52 MPa (800 psi) corresponds to
41.3 wt % CaS or about 52 percent of the original CaS content, so despite
the expectation that hot pressing would exclude atmospheric oxygen,
about half the CaS was oxidized to the sulfate and the rest of the sul-
fide sulfur was lost.
For explanations of the above changes, atmospheric oxygen seems
most likely compared to hydrolysis loss by moisture or reduction of
ferric oxide (Fe 0 ) as per the following equations:
4 Fe-0 -> 8 FeO + 2 02
CaS + 2 02 + CaSO^ .
These appear thermodynamically possible, since the net free energy
change at 1000°C is -7782 J/g mole of CaS. Stoichiometry, however,
limits the conversion to less than 1 percent.
In another test, 75-CF-32, CaS was hot pressed with Fe 0 . The
cylindrical specimen produced was magnetic, confirming the ability of
CaS to reduce Fe 0 . Additional observations should be made to clarify
what chemical changes are occurring.
Further experiments were conducted to explore the effect of particle
size. Pure CaS/fly ash and CaSO,-2H 0/fly ash mixes were compared
with the CAFB regenerator stone/fly ash mixes. Three compositions were
made with each system, as shown in Table G-15. Each composition was
ball milled for two hours, screened to -125 urn, and subjected to hot
pressing for one hour in a 1.9-cm (3/4 in) graphite die at approximately
3310 MPa (4800 psi) and 1050°C. A flowing nitrogen atmosphere was
259
-------
Curve 690933-A
80
o>
CO
o 60
0>
5
3
00
I 40
o
» 20
o>
O>
CO
o
o>
CO
60
I 40
f, 20
'35
800
CaSO,
8.27MPa
(1200 psi)
5.52MPa
(800 psi)
CaS ^
1
900 1000 1100
Temperature, °C
1200
1300
Figure G-12 -
Residual Sulfur Content of Hot-Pressed Gypsum/Fly
Ash Mixes. Series 75-CF-18
260
-------
Curve 690935-A
c/n
o
E
.5
'o
fa
O
80
60
40
t 2o
T
CaSO,
(1200 psi)
- 1
80
° 60
TO
.S 40
_o
s
--
f 20
(800 psi)
o vcas°4
1000 1100
Temperature, °C
1200 1300
0-13 - Residual Sulfur Content of Hot-Pressed Calcium
Sulfide/Fly Ash Mixes. Series 75-CF-17
261
-------
Curve 690934-A
80
3
CO
o 60
o>
•S3
3
CO
<3
40
.f> 20
o>
6.89 M Pa
(1000 psi)
-I-
CaS
I
800 900 1100 1200
Temperature, °C
1300
Figure G-14 - Residual Sulfur Content of Hot-Pressed Calcium
Sulfide/Fly Ash Mixes. Series 75-CF-17
maintained during the experiment. After the samples were hot pressed,
they were calipered and weighed. The calculated densities of the CAFB/
fly ash specimen were slightly higher than those of the other two systems.
All three systems show a decrease in density as the fly ash content
increases from 20 to 80 wt %, consistent with the fact that the density
of fly ash is lower than that of the other three components.
While all of the mixes were successfully hot pressed, when aged
in air two of the specimens crumbled within 30 days. Table G-15 also
presents compressive strength data on the hot-pressed mixes. The values
were much higher than expected so a further test was devised.
Six of the specimens were pulverized by a combination of methods,
including ball milling. The samples were first hand crushed to
-841 um (-20 mesh) and then ball milled with Al 0 balls in an aluminium-
lined mill. Dry milling did not produce any significant size reduction,
so a liquid vehicle was added, as noted in Table G-16. Except for one
262
-------
Table G-15
COMPRESS I VE STRENGTH OF HOT-PRESSED MIXES OF FLY ASH
VARIOUS CALCIUM COMPOUNDS
AND
Specimen
Id ent i f ic at ion
Composition,
CaS |
wt %
Fly Ash
Condition
after Aging
Compressive
Strength
MPa | psi
Density,
g/cc
75-CF-22
75-CF-23
75-CF-24
75-CF-25
PO
£ 75-CF-26
75-CF-27
75-CF-28
75-CF-29
75-CF-30
80
50
20
CaSOA.2H20
80
50
20
CAFB Stone
80
50
20
20
50
80
Fly Ash
20
50
80
Fly Ash
20
50
80
Good
Good
Good
Crumbled
Good
Good
Crumbled
Good
Good
195.8
243.4
199.0
45.2
120.0
—
261.2
28,400
35,300
28,860
6,550
17,400
—
37,880
2.55
2.46
2.40
2.43
2.35
2.22
2.57
2.55
2.46
aFly Ash was from the Duquesne Light Co., Phillips Plant, Pittsburgh, PA.
Minimum aging time about one month.
-------
gypsum/fly ash mix, methanol was used to avoid setting in the presence
o
of water. A 240 cm (1/2 pt) capacity ball mill was used to grind
approximately 15 g of the sample. The grinding was done overnight,
following which the slurry was passed through a 44 ym (325 mesh) screen.
There was a mild smell of sulfur after grinding in some cases.
The powders were filtered and dried and then mixed with Ottawa
sand and water. The sand-to-powder ratio was 2.67 and the water-to-
powder ratio 0.5. A control specimen was made with Portland cement and
Ottawa sand. The pastes were allowed to set in 1.9 cm diameter by
1.27 cm high (3/4 in diameter by 1/2 in high) Teflon cylindrical molds.
After 24 hr the specimens were taken out of the mold so we could see
whether they remained as solid blocks or crumbled to powder. This result
is recorded in Table G-16, along with the composition of the hot-pressed
powder. The powder compositions containing CaS and fly ash showed a
cementlike setting property. The strongest block qualitatively was
obtained with 50 percent CaS + 50 percent fly ash.
Compressive strengths for the CaS/fly ash mixes and the Portland
cement controls after curing in air for 7 days are given in Table G-16.
While the values obtained were lower than those for Portland cement,
the CaS/fly ash hot-pressed specimens do have a cementlike setting
property.
264
-------
Table G-16
SETTING CHARACTERISTICS OF MIXES OF SAND AND
POWDERED HOT-PRESSED MATERIALS3*
Specimen
Identification
c
Powder Composition,
Wt %
CaS
75-CF-22 80
75-CF-23 50
Fly Ash
Observation
After 24 Hrs
7-Day
Compressive
Strength
kPa DSI
20 Solid block, strong 965 140
50 Solid block, very 1721 7-^n
75-CF-24
20
80
75-CF-26
75-CF-27
CaSO. -2H-0
4 2
50
20
Fly ,
50
80
CAFB-9 Stone Fly Ash
75-CF-30 20 80
Control- Portland cement
strong
Solid block, moder- 620 90
ately strong
Crumbled to powder
Crumbled to powder
Crumbled to powder —
Solid block, strong 5860 850
The mixes contained 22 g of Ottawa sand and powdered specimens of hot-
pressed materials in weight ratio of 2.67/1 plus 3 cc of water. The
pastes were allowed to set in Teflon molds for 24 hours.
The hot-pressed specimens were all ground in methanol except for
75-CF-27, which was ground in water prior to being mixed with sand
and water.
:The "powder composition" is before hot pressing.
265
-------
APPENDIX H
CALCIUM SULFIDE STUDIES
The first run made in the 10-cm laboratory fluidized-bed test unit,
Batch L-l, served to check operability and to produce a test quantity
of sulfated limestone. Supplies of spent sorbent from the CAFB pilot
plant regenerator were extremely limited, and the initial tests were
therefore planned around use of a simulated stone produced by sulfiding
calcined limestone and then oxidizing this to the sulfate.
In L-l, 1600 g of -595 4420 ym Limestone 1359 were calcined, sul-
fided, and air oxidized in separate operations. Tables H-l and H-2
contain the operating conditions for the run while Table H-3 contains
product analyses. Table H-4 presents an analysis of the performance,
based on the chemical compositions of the samples in various stages.
We assumed that no material was lost by elutriation, since we had found
less than 1 g of solids in the cyclone and essentially nothing in the
'final filter. This was a sintered metal filter with a mean pore size
of 65 ytn and rated at 100 percent removal of 20 ym particles and 98 per-
cent of 8 ym particles in gas service. The active filtering area was
2 2
334 cm (0.36 ft ). The filter pressure drop normally was essentially
zero.
Calcination was high but incomplete in the first step and essen-
tially completed during sulfidation. Oxidation (70.8% of the CaS) was
also high but significantly less than the complete conversion desired.
Figures H-l and H-2 show photomicrographs of sulfided and oxidized
particles. The particles were embedded in plastic and then ground and
polished to expose the interior of the grains. The outer layer is plastic
while the darker layer inside is CaS. There was no evidence of pores
266
-------
containing CaS that had penetrated deep into the particles. Further,
there was no evidence of CaS deep within the particles of oxidized
stone.
Table H-l
MATERIAL BALANCE DATA FOR LIMESTONE SULFIDATION STUDIES
BATCH L-l
Fluid ization
Calcination
Sulfidation
Oxidation
BATCH L-4
Fluidization
Calcination
Sulfidation
Oxidation
BATCH L-6
Calcination
Sulfidation
Oxidation
Gas Flow Rates (2,/min) 15
N2 H2S H2
89
134
43
43 0.90 4.08
37
194
202
45
55 1.2 3.0
52-18
36
18 2.51 5.15
22
°C, 1 atm
Batch Charge,
Air g
800
1600
6.1
1000
2000
1100
9.4-5.5
1476
8.8 300a
aSulfided limestone. Bulk density - 1.276 g/cc.
Thus it appeared that in the fluidized bed most of the sulfide sul-
fur was laid down initially close to the surface of each particle, at
least to the level of 21 percent Sulfidation. This suggested that if
the CaS were present in a sufficiently thin layer, it might be possible
to oxidize it completely to CaSO,.
267
-------
Table H-2
TIME/TEMPERATURE DATA FOR LIMESTONE SULFIDATION STUDIES
Heating
Reacting
Cooling
BATCH L-l. Particle Size -595 + 420 yro
Calcination
Sulfidation
Oxidation
125 min @ 600-730°C 195 min @ 730-760°C
170 min @ 659-766°C 120 min @ 766°C
130 min @ 651-810°C 150 min @ 810-795°C
BATCH L-4. Particle Size -1190 + 500 ym
Calcination
Sulfidation
Oxidation
1260 min @ 600-730°C 315 min (? 730-906°C
300 min (? 25-762°C 20 min @ 762°C
40 min @ 766-589°C
40 min @ 795-520°C
40 min @ 906-600°C
50 min @ 762-457°C
345 min @ 480-870°C 155 min @ 870-935°C 15 min @ 935-835°C
BATCH L-6. Particle Size -595 + 149 ym
60 min @ 548-743°C 255 min @ 743-904°C
Calcination
Sulfidation
Oxidation
125 min @ 864-880°C 15 min @ 866-760°C
Standby overnight
@ 378°C
200 min @ 370-859°C 180 min @ 859-872°C 120 min @ 872-349°C
Environmentally, there would still be a problem because the core
of the particles would be CaO. On the other hand, if the calcium could
be completely sulfided, it might be possible to encapsulate a CaS core
with a relatively impermeable CaSO, shell. A process material balance
on a no-loss basis is shown in Table H-5.
The next run in this series was L-4, in which -1190 +500 ym mesh
particles of Limestone 1359 were calcined, sulfided to a low degree,
and then oxidized with air. Sulfidation was set by stoichiometry. The
data in Table H-3 indicate that the sulfidation achieved was uniform
across the bed at about 5.1 mol % of the calcium, as planned. The
reactor product after oxidation wtih 3.2 percent oxygen showed 1.6 to
268
-------
Table H-3
CHEMICAL ANALYSIS OF STONE SAMPLES FROM LIMESTONE
SULFIDATION STUDIES
Batch Ca S
Batch L-l
Calcined 61.7
Component wt %
SOA~ zs C02
11.9
Sulfided 64.6 10.9 0.15 - 2,1
Oxidized 55.0 3.3 24.0
Batch L-4
Feed 40.19
Calcined 68.48
43.38
4.19
Sulfided, center of bed 65.93 2.70 - 2.45 4.93
Sulfided, edge of bed - 2.65
Oxidized
4-1190 ym 65.32
+841 ym 67.39
0.86 4.84
1.04
+595 ym 66.69 0.05 - 1.90
+500 ym 67.27 0.
-500 ym 67.00 0.
Batch L-6
Feed 38.76
Sulfided, powder 48.61 32.
Sulfided, sinter 42.23 30.
Oxidized 49.74 27.
05 - 1.12
05 - 1.32
43.41
69 - 28.09
65
95 4.01 27.80
3.6 mol % of the calcium was sulfated. The highest sulfation was in
the -841 +595 ym fraction. These results showed that the thin layer
of CaS was not readily oxidizable under the conditions used without
losing a substantial amount of sulfur (30-70%). The S0« monitor on the
269
-------
Figure H-l - Sulfided Limestone, -595 +420 ym
Figure H-2 - Oxidized Sulfided Limestone, -595 +420 um
270
RM-61343
-------
Table H-4
CALCULATED RESULTS FOR BATCH L-l
Calcined
Stone
Composition, wt %
CaC03 27.05
CaO 71.23
CaS
CaS04
Inerts 1.72
100.00
Ratios
Inert content, g/g mol Ca 1.116
% of CaC03 calcined 82.5
% of Ca sulfided
% of CaS oxidized
Sulfided Oxidized
Stone Stone
4.77
68.67 57.25
24.48 7.41
0.21 34.00
1.87 1.34
100.00 100.00
1.158 0.974
97.04
21.0 7.48
70.8
Stoichiometry
Mols of H2S fed
Mols of CaS made, theor.
Mols of 02 fed
Mols of CaSO, made, theor.
4.57
3.32
8.13
2.87
off-gas during oxidation showed about 0.5 percent SO within 5 minutes
after cutting in the air stream, rising to about 1 percent, and then
dropping to 0.1 percent within 20 minutes.
Following the completion of the oxidation based on quantity of air
fed, an attempt was made to simulate exposure to the CAFB regenerator
conditions of 1070°C in a mildly oxidizing atmosphere. Flow rates were
cut back to minimize possible carry-over. The total excess oxygen used
271
-------
Table H-5
PROCESS MATERIAL BALANCE FOR BATCH L-l
Raw Calc
Stone Sto
Material Balance, g mols
CaC03 15.822 2
CaO 13
CaS
CaSO,.
Totals 15.822 15
Material Balance, g
CaC03 1582.19 277
CaO 730
CaS
CaS04
Inerts 17.81 17
Totals 1600.00 1026
Composition, wt %
Calcium 39.55 61
Carbon dioxide 43.51 11
Sulfide
Sulfate
Inerts 1.11 1
ined Sulfided Oxidized
ne Stone Stone
.774 0.468
.048 12.012 11.765
3.327 1.183
0.015 2.874
.822 15.822 15.822
.41 46.76
.68 672.68 658.86
239.88 85.29
2.08 391.12
.81 17.81 17.81
.00 979.21 1153.08
.68 64.63 54.89
.90 2.10
10.91 3.29
0.15 23.95
.74 1.82 1.54
was 10.1 moles at a concentration ranging from 3.2 to 4.9 percent. The
maximum temperature attainable with the equipment appeared to be 935°C
after two hours, so the run was terminated.
Some change in particle size distribution was also noted. The final
oxidized product contained 1 percent +1190 urn and 4 percent -500 ym versus
272
-------
none of these fractions in the original limestone charge. The total
cyclone catch was 15.8 g or 0.08 percent of the original charge. The
filter catch was 0.2 g.
Limestone batch L-6 was aimed at 100 percent sulfidation of -595
+149 um stone followed by 100 percent oxidation to CaSO,. Significant
corrosion attack was observed for the first time. The reactor bed
thermocouple Inconel sheath had been penetrated and had to be replaced.
Some hard deposits on the reactor wall had formed and, by scraping, a
small amount of scale was removed. The concentration of H~S used was
10 percent versus 2 percent in previous runs. The temperature level
was 870 to 900°C versus 760°C. The hydrogen concentration was also
increased to 20 percent versus 5 to 8.5 percent to continue suppression
of the dissociation of H~S.
The average sulfide content found was 30.48 wt % versus an average
calcium of 45.42 percent. On a mole basis, there were 0.951 moles of
sulfur, 1.133 moles of calcium, corresponding to 84 mole % sulfidation.
The oxidation level achieved with 6 percent oxygen was lower than
had been expected or desired: 3.4 mole % of the calcium sulfated. The
retention of sulfur was about 83 percent. There was no problem of con-
trolling bed temperature with the conditions used. Further exploration
of CaS production was deferred in favor of dry sulfation studies.
273
-------
APPENDIX I
TEST PROCEDURES FOR IN SITU LEACHING OF SPENT SORBENTS WITH SEAWATER
TEST I
This test is to determine the leaching of spent stone by mixing with
natural seawater under field winter conditions at an offshore mid-Atlantic
location. Two hundred grams of material are placed in 2 liters of seawater
in a 4 liter container and placed on a magnetic stirrer for 24 hours.
Temperature and pH are recorded for every 15 minutes in the first hour
and every 6 hours after the start of the test for 24 hours. After 6 hours
of the test, 1 liter is removed and sealed in a glass jar for analysis.
After 24 hours, the remaining liter is sealed for analysis. (Note: When
removing the samples of water, stop the stirrer and allow the particles
to settle out before removing the water samples.) Perform this test on
all three chemicals. A control sample must be kept for each different
seawater type used.
TEST II
This test is to determine the leaching of spent stone as a function
of concentration in natural seawater. Four concentrations (20, 100, 200,
and 400 g) are placed in 1 liter of seawater in 2-liter breakers. Pour
each packet of material in such a way as to form a mound of spent stone
in each beaker. Do not stir. Any rolling of the ship will provide all
the mixing necessary. Temperature and pH are recorded every 15 minutes
in the first hour and every 6 hours after the start of the test for
24 hours. After 24 hours, the water is decanted off to be sealed for
analysis, a scraping of the surface of each mount is placed in a vial
for analysis, and a core of each mound is removed for analysis. Ten grams
are needed of each for analysis. A control sample of seawater must be
taken for analysis.
274
-------
Table 1-1
CHEMICAL ANALYSES OF SPENT SORBENT SAMPLES
Source ANL
Type of stone sorbent Dolomite
Composition, wt %
CaSO, 57
CaC03 9
CaO 2
CaS <0.05
MgO 20
Inerts 12
100
Westinghouse
ERCA R&D
Limestone Dolomite D-2
CAFB 9
4.35 53.83
—
84.26 13.83
5.04 0.91
25.99
6.35 5.44
100.00 100.00
Table 1-2
TEST I. EFFECT OF LEACHING TIME ON LIQUOR TEMPERATURE AND pH
Stone
CAFB-9
1 Argonne
| WestinRhouse D-2
Initial Seawater
Conditions
Temperature,
°C
PH
Time,
hr
0
0.25
0.50
0.75
1.0
6.0
12.0
18.0
24.0
14.5 ± 2.1
8.5
Temperature
Rise, °C i
0.0 9.
0.3 9
0.7 9
1.1 10
1.4 10
4.6 10
5.3 10
5.2 9
5.3 9
18.45 ± 1.1
7.95
Temperature
>H Rise, °C p
0 0.70 9.
3 1.95 10.
5 3.50 10.
2 4.00 10.
.3 4.90 10.
.3 6.20 9.
.0 6.70 9.
.9 6.70 10.
.9 7.30 10.
18.45 ± 1.
7.95
Temperature
H Rise, °C
0 1.0
1 1.85
1 2.50
1 2.80
1 3.20
9 4.50
9 5.00
0 5.50
0 5.80
I
pH
10.4
10.5
10.5
10.9
11.8
12.0
12.0
12.0
11.9
275
-------
Table 1-3
TEST IT. EFFECT OF TREAT RATIO ON LEACHING OF SPENT SORBENTS
Treat
n _ «. J -.
Ratio ,
g/
20
100
200
400
Stone
Time , hrs
0.00
0.25
0.50
0.75
1.00
6.0
12.0
18.0
24.0
0.00
0.25
0.50
0.75
1.00
6.0
12.0
18.0
24.0
0.00
0.25
0.50
0.75
1.00
6.0
12.0
18.0
24.0
0.00
0.25
0.50
0.75
1.00
6.00
12.00
18.0
24.0
CAFB-9
AT, °C
0.0
0.25
0.80
1.19
1.47
2.42
2.69
2.37
2.30
*
*
*
*
*
*
*
*
*
2.51
23.9
26.43
25.36
23.47
22.93
22.84
22.26
22.2
70.25
51.92
41.64
36.5
33.38
33.39
32.98
32.19
32.22
PH
8.5
8.75
8.9
8.95
9.05
9.65
9.70
10.0
10.0
*
*
*
*
*
*
*
*
*
9.9
9.6
9.5
9.5
9.55
10.10
10.10
10.20
10.20
_ —
11.3
11,25
11.25
11.35
11.8
11.8
11.9
11.9
Argonne
AT, °C
0.25
1.45
3.10
3.50
4.40
6.20
7.90
7.80
7.60
0 05
1.35
3.0
3.40
4.4
6.3
7.5
7.7
7.4
0.0
1.45
3.0
3.4
4.4
6.3
7.5
7.7
7.5
0.0
1.55
3.1
3.5
4.3
5.6
9.3
8.6
7.6 j
pH
7.7
8.05
8.13
8.15
8.18
8.3
8.6
8.62
8.6
7.7
8.58
8.65
8.68
8.70
8.70
9.20
9.20
9.15
7.7
8.92
9.0
9.0
9.3
9.08
9.5
9.55
9.48
7.7
9.5
9.52
9.55
9.58
9.3
10.05
10.1
9.72
Westinghouse
D-2
AT, °C
0.8
0.95
1.6
1.8
2.3
4.1
5.3
5.9
5.3
0.6
1.25
1.9
2.2
2.7
4.0
5.1
5.7
5.2
t
t
t
t
t
t
t
t
t
0.7
2.65
3.6
4.0
4.5
4.2
5.2
5.7
5.4
PH
7.7
8.6
8.8
8.9
9.0
9.85
9.97
10.12
9.95
7.7
9.65
10.05
10.13
10.15
9.91
10.4
10.55
10.2
t
t
t
t
t
t
t
t
t
7.7
10.28
10.29
10.25
10.2
9.92
10.7
11.2
11.82
*Sample lost
tlnsufficient sample for test
27*
-------
APPENDIX J
ANALYTICAL PROCEDURES USED ON SAMPLES FROM OCEAN DUMPING TESTS
SAMPLE PREPARATION
Solids
A representative fraction of the sample was transferred to a
platinum vessel and dried at 110°C in excess of 12 hours (overnight).
It was reduced in a tungsten carbide mortar and pestle to a <100 mesh
size and finally diluted with high-purity graphite prior to spectro-
chemical determination.
Liquids
Samples of the liquids were obtained by agitating the bottle or
jar until all sediment was uniformly distributed whence a known volume
was removed and immediately filtered through a millipore. The filtrate
was transferred to a 100 ml tared plastic beaker and evaporated to
dryness over low heat on an electric hot plate.
The insoluble fraction was transferred to a tared platinum
crucible and dried at 110°C. The dried weight corrected for the filter
was used in calculating the solid to liquid relationship. This dried
residue was diluted with graphite in the same manner as for solids.
SPECTROSCOPY
All of the spectrochemistry was done on a Jarrell Ash 3.4 meter,
Ebert mount spectrograph. Direct current arc excitation was employed
in an argon-oxygen atmosphere supplied by a modified Stallwood type jet.
The samples were compared against standards made by adulterating at
various levels high purity calcium carbonate or sodium chloride with the
elements of interest. Visual comparisons were made against these
standards with an estimated accuracy of 1/2 to 2X the determined value.
Determination of selenium was done by flameless atomic absorption.
277
-------
A preliminary analysis of some of the samples had been made to
generally characterize them. Impurities incorporated in graphite were
used in this instance.
278
-------
Table J-l
EXTRACTION OF TRACE ELEMENTS FROM CABF REGENERATOR STONE BY SEAWATER
Sample Identification (1)
II S L2
I C 6 LM
I C 24 LM
I C 6 LI
I C 24 LI
to
~J II C 20 LI
vO
II C 100 LI
II C 200 LI
II C 400 LI
II C 20 CS
II C 100 CS
II C 200 CS
II C 400 CS
Background
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Element Concentration, ppm
Cr I Ni
<0.5 0.05
<0.2 2000
Solids Sample
Concentration Volume,
| Cu | Pb g/e liters
0.1 <0.15
0.2 <0.2 3.3 1.0
<30 '30 7.9
0.2 <0.2 3.4 1.0
30 <30 11.9
<1
-------
Table J-l (Cont)
EXTRACTION OF TRACE ELEMENTS FROM ARGONNE SPENT SORBENT BY SEAWATER
Sample Identification (1)
I A 6 L
1 A 24 L
II A 20 L
II A 100 L
II A 400 L
00 II A 20 CS
O
II A 400 CS
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
EXTRACTION OF TRACE
I D 6 L
I D 24 L
II D 20 L
11 D 100 L
II D 400 L
II D 20 CS
II D 400 CS
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Filtrate
Floe
Element Concentration,
Cr I Ni
<1 <1
100 20
<1 <1
100 20
1 0.1
300 200
<0.5 0.05
<30 20
<0.2,<1 <0.2,<0.
100 ' 50
70 20
70 20
ELEMENTS FROM
-------
1.
NOTES TO TABLE J-l.
Sample code numbers have the following meanings:
First character
Second character
Third group
Fourth group
I or II
C
A
D
6,24
20,100,200,
400
LI, L2, L3
LM
CS,SS
SM
Refer to test conditions I or II
CAFB regenerator stone (limestone)
Argonne dolomite
Dolomite from Westinghouse test unit
Applies to Test I and designates the
time at which the sample was taken
Applies to Test II and designates the
grams of sorbent used per liter of
sea water
Indicates liquid sample 1, 2, and 3
Indicates liquid sample using Maryland
offshore water
Indicates core solids or surface solids
Indicates solids from test with
Maryland offshore water
281
-------
APPENDIX K
EXTRACTION OF SELECTED TRACE ELEMENTS BY
SEAWATER FROM SPENT SORBENTS
Mercur
II SL 1 <1
II A 400 L <1
II C 400 L
II A 400 CS <10
II C 400 CS
y, ppb Selenium, ppm
ND (1)
ND (1)
ND (2)
Fluorine, ppm
0.68
0.23
0.41
34
46
(1) Not detected by flameless atomic absorption: level is < 0.1 ppm
(2) Not detected by flameless atomic absorption: level is < 1 ppm
282
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
I REPORT NO.
EPA-600/7-79-158b
3 RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Chemically Active Fluid Bed for SOx Control:
Volume H. Spent Sorbent Processing for Disposal/
Utilization
15. REPORT DATE
December 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
C.H. Peterson
I. PERFORMING ORGANIZATION REPORT NO.
9.
FORMING ORGANIZATION NAME AND ADDRESS
Westinghouse Research and Development Center
1310 Beulah Road
Pittsburgh, Pennsylvania 15235
10. PROGRAM ELEMENT NO.
EHB536
11. CONTRACT/GRANT NO.
68-02-2142
12. SF
ND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
ERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/13
16, SUPPLEMENTARY
919/541-2825.
IERL-RTP project officer is Samuel L. Rakes, Mail Drop 61,
BTABSTR cTThe report describes the processing of spent calcium-based sulfur sorbents
(limestones or dolomites) from an atmospheric-pressure, chemically active fluid
bed (CAFB) gasification process, using a regenerative sulfur sorbent process that
produces low- to intermediate-Btu gas. Data are developed to provide a basis for
evaluating process concepts to minimize the environmental impact (heat release
H2S release, and potential leachates) or possibly for spent sorbent utilization '
Flow diagrams and cost estimates were prepared for five processing options ' A
dry sulfation process operating at 850 C to produce spent solids containing CaSO4
acceptable for disposal and low-temperature ash blending to produce a material for
disposal or utilization is recommended for further development A concept for
briquetting to produce aggregate is presented as a low-temperature blending option
based on laboratory tests that produced compacts with compressive strengths UD to
80 MPa. Direct disposal dead-burning for disposal by heating at 1250 C and reducing
the sulfide content to < 0.03%, and sintering at 1550 C to release the sulfur for
recovery and produce a possible source of lime containing < 0.15% sulfur are also
investigated. Processing sorbent from a once-through sorbent process containine
CaS is also considered. *
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
pollution Aggregates
Fiuidized Bed Processing
Coal Gasification Briquetting
Calcium Carbonates Waste Disposal
Degeneration Combustion
Sulfation
DISTRIBUTION STATEMENT
He lease to Public
Form 2220-1 (t-73)
b.lDENTIFIERS/OPEN ENDED TERMS
Pollution Control
Stationary Sources
Chemically Active Fluid
Bed Process
Spent Sorbent Processing
Dead Burning
19. SECURITY CLASS (This Report}
Unclassified
2O. SECURITY CLASS (This paft)
Unclassified
c. COSATI Field/Group
ISB UG
13H,07A
07C,07B
2 IB
22. PRICE
-------