EPA-600/2-76 143a
May 1976
Environmental Protection Technology Series
MEYERS PROCESS DEVELOPMENT FOR
CHEMICAL DESULFURIZATION OF COAL
Volume I
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-143a
May 1976
MEYERS PROCESS DEVELOPMENT
FOR CHEMICAL DESUL FURIZATION OF COAL
VOLUME I
by
E.P. Koutsoukos, M. L. Kraft, R.A. Orsini,
R.A. Meyers, M. J. Santy, andL.J. Van Nice
TRW Systems Group
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-1336
ROAPNo. 21AFJ-033
Program Element No. 1AB013
EPA Project Officer: L. Lorenzi, Jr.
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
The Meyers Process for chemical removal of sulfur from coal was tested
at bench-scale for desulfurization of both fine and coarse coal. In excess
of 90 percent of the pyrite was removed from run-of-mine fine coal and clean
coarse coal and over 80 percent of the pyrite from run-of-mine coarse coal.
Process unit improvements involving 1) increased process slurry solids
concentration for higher process throughput (33% w/w], 2) lowered filtration
requirements through use of larger top-size fine coal and 3) generated
elemental sulfur removal by vaporization from the coal matrix were demon-
\
strated. Pyrite leaching and reagent regeneration rate expressions were
validated. Engineering studies showed that 1) the process may be engineered
in a number of basic design configurations including simultaneous leach
and regeneration, separate leach and regeneration, use of oxygen or air for
regeneration, fine coal or coarse coal processing, and combination with
coal cleaning; 2) these process design schemes lead to stand-alone full
capital cost estimates of $30-80/KW of power plant name plate capacity;
3} estimated coal desulfurization costs, annualized on a utility financed
basis, range between $0.33 and $0.51/MM Btu, and 4) assuming ROM coal costs
of $20/ton, the costs of the desulfurized fuel are estimated to range
between $1.14 and $1.32/MM Btu.
m
-------
TABLE OF CONTENTS
Page
Abstract i i i
List of Figures ix
List of Tables xiv
Acknowledgments xvi
Metric Conversion Factors xvii
Conclusions 1
Recommendations 6
1. Introduction 7
2. Pyritic Sulfur Removal from Suspendable Coal H
2.1 Coal Selection and Sample Preparation H
2.2 Coal Processing Experimentation 17
2.3 Coal Processing Data 24
2.4 Data Discussion and Interpretation 35
2.4.1 Mixer Unit Operation 40
2.4.2 L-R Unit Operation 50
2.4.2.1 Pyrite Leaching Rates During L-R Processing 56
of L.K. Coal at 120°C
2.4.2.2 Reagent Regeneration During L-R Processing 68
2.4.2.3 Temperature Effects on L-R Processing of 76
of L.K. Coals
2.4.2.4 Coal Top Size Effects During L-R Processing 82
2.4.2.5 Slurry Concentration Effects 85
2.4.2.6 Reagent Composition Effects 86
2.4.2.7 Oxygen Partial Pressure Effects 95
2.4.3 Settler Unit Operation 95
-------
TABLE OF CONTENTS (continued)
2.4.4 Coal Washing Operation 99
2.4.5 Elemental Sulfur Recovery Operation 101
2.4.5.1 Elemental Sulfur Recovery with Toluene 101
2.4.5.2 Elemental Sulfur Recovery with Hexane 107
2.4.5.3 Elemental Sulfur Vaporization 108
2.5 Solid-Liquid Separation Unit Operations 112
2.6 Coal Drying Operation ' 114
2.7 Principal Conclusions from Bench-Scale Data 114
2.8 L-R Processing of Upper Freeport Seam Coal 122
3. Reagent Recycl ability-Trace Element Data 127
4. Processing of Coarse Coal 135
4.1 L-R Processing Data for Coarse Coal 136
4.2 Clean Coal Gravity Fraction Processing 139
4.3 Preliminary Data on Coarse Coal Processing by Size 141
Fraction
5. Process Engineering 145
5.1 Suspendable Coal Processing 145
5.1.1 Design Basis for Suspendable Coal 146
5.1.2 Process Baseline Design 153
5.1.2.1 Conceptual Design for Commercial Scale 157
5.1.2.2 Process Cost Estimate 172
5.1.3 Process Trade-off Studies 180
5.1.3.1 Reactor Model 181
5.1.3.2 Pressure Effect 183
5.1.3.3 Iron Concentration Effect 188
-------
TABLE OF CONTENTS (continued)
Page
5.1.3.4 Oxygen Purity 190
5.1.3.5 Compressed Air Regeneration 192
5.1.3.6 Three Reactor Configuration 196
5.1.3.7 Additional Studies 197
5.2 Coarse Coal (-1/4 inch) Processing 198
5.2.1 Concept Development 198
5.2.2 Conceptual Process Details 202
5.2.2.1 Reactor Section - Pit and Continuous 202
5.2.2.2 Regenerator Section - Pit and Continuous 206
5.2.2.3 Sulfate Removal Section 207
5.2.2.4 Coal Washing and Filtration Sections 209
5.2.2.5 Sulfur Removal Section 210
5.2.2.6 Energy Balance 210
5.2.3 Process Cost Estimate 210
5.3 Projection of Process Economics 215
5.3.1 Processing Option Description 215
5.3.2 Economics Model Description 219
5.3.3 Process Economics Evaluations 225
6. Chemical Analysis Studies 231
6.1 Adequacy of Standard Sulfur Analysis Techniques for 232
Determining Meyers Process Performance
6.2 Meyers Process Monitoring Techniques 242
7. Materials Compatibility 249
7.1 Literature Survey and Alloy Selection Criteria 250
7.2 Experimental Corrosion Static Test Program 253
vii
-------
TABLE OF CONTENTS (continued)
Page
7.2.1 Experimental Method 253
7.2.2 Initial Runs and Modifications 257
7.2.3 Results and Conclusions of the Bomb Test 261
7.2.4 Conclusions and Recommendations 281
7.3 Dynamic Testing 282
8. References 291
9. Glossary of Abbreviations and Symbols 293
vm
-------
LIST OF FIGURES
No. Page
1 Basic Bench-Scale Processing Scheme for the Removal of 18
Pyritic Sulfur from Coal
2 Bench-Scale Coal Leaching and Reagent Regeneration Apparatus 19
3 Pyritic Sulfur Removal from 14 Mesh Top Size L.K. Coal 38
Processed at 120°C with 5 Wt. % Fe Reagent
4 Typical Mixer Processing Conditions 44
5 100 Mesh Top Size Lower Kittanning Coal Leached with 5 Wt. % 59
Fe Reagent
6 Pyrite Leaching Rate Constant Data '66
7 Predicted and Experimental Y Values During L-R Processing 70
of L.K. Coal at 120°C
8 Predicted and Experimental Y Values During L-R Processing 71
of L.K. Coal at 110°C
9 Predicted and Experimental Y Values During L-R Processing 72
of L.K. Coal at 130°C
10 Temperature Effect on L-R Processing of 100 Mesh Top Size x 77
L.K. Coal (20 Wt. Percent Slurries)
11 Temperature Effect on L-R Processing of 14 Mesh Top Size 78
L.K. Coal (33 Wt. Percent Slurries)
12 Arrhenius Plots of Pyrite Leaching Rate Constants 83
13 Coal Top Size Effect on L-R Processing of Suspendable L.K. 84
Coal
14 Slurry Concentration Effect on L-R Processing of Suspendable 87
L.K. Coal
15 Effect of Total Iron Concentration on Pyrite Removal During 94
L-R Processing of Suspendable L.K. Coals
16 Effect of Oxygen Partial Pressure on L-R Processing of 96
Suspendable L.K. Coal
17 Vapor Pressure of Sulfur 110
18 Typical Process Design Curves 120
ix
-------
LIST OF FIGURES (continued)
No. Page
19 Upper Freeport Coal Leaching with 5 Wt. % Fe Reagent 123
20 Pyrite Leaching Rates from Upper Freeport and L.K. Mine Coals 124
21 Pyritic Sulfur Removal from ROM and Cleaned L.K. Coal at 102°C 140
22 Filtration Rate Correlation 154
23 Block Diagram for Fine Coal 155
24 Process Flow Diagram for Fine Coal 158
25 Simplified Reactor System 184
26 Reactor Residence Times as a Function of Oxygen Pressure in R-l 185
27 Effect of Pressure and Residence Time on Ferrous Iron Make 187
28 Reactor System Coat as a Function of Oxygen Pressure 189
29 Oxygen Circulation Diagram 191
30 Coarse Coal Process Schematic 201
31 Pit Reactor Schematic 203
32 Continuous Reactor Schematic 205
33 Sulfate Removal Section Schematic 208
34 Process Economics Case Block Diagram 217
35 Bomb Test Assembly 255
36 Composite Welded Tube Used in Test 255
37 Test Coupon Stack 256
38 Comparison of Corroded and Control Tubes 260
39 Microstructure of 304 Test Coupon After 1000 Hours Exposure 263
40 Microstructure of 304 Test Coupon at 400x Showing Sensiti- 263
zation at Grain Boundaries
41 Microsturcture of Unexposed Sample of 304 Test Coupon 264
-------
LIST OF FIGURES (continued)
Page
42 Micrograph at 100X Showing Very Extensive Pitting Type of 266
Corrosion Just Outside the Weld Zone in 316SS After Exposure
43 Micrograph at 100X Showing Very Extensive Pitting Type of 266
Corrosion Just Outside the Weld Zone in 316SS After Exposure
44 Micrograph at 100X Showing Very Extensive Pitting Type of 267
Corrosion Just Outside the Weld Zone in 316SS After Exposure
45 Edge of 316 Control Sample Showing Elongated Stringers Due 268
to the Tensile Test
46 Edge of Exposed Test Coupon of 316 Showing Transgranular 268
Cracks Extending Inward from Cut Edge
47 Edge of 31 6L Control 269
48 Edge of Exposed 31 6L Showing Crack Extension Along Crack 269
Sensitive Corroded Paths and Pitting
49 Edge of 347 Control 270
50 Edge of Exposed Coupon of 347 Showing Limited Extension of 270
Cracks in the Interior
51 Edge of Incoloy 825 Control 271
52 Edge of Exposed Incoloy 825 Test Coupon 271
53 Weld Between 304 and 304L in the Control Tube Sample 272
54 Weld Between 304 and 304L Tubes Showing Severe Corrosion 272
Effects After Exposure
55 Outer Diameter of 304 Tube After Exposure Showing Corrosion 273
Along the Edge and Pits on Surface
56 Inner Edge of Exposed 304L Tubing Showing Corroded Areas 274
57 Exposed Tube Sample 275
58 Exposed 31 6L Tube 276
59 I.D. of Exposed Armco Tubing Showing Some Cracks and Corroded 277
Areas
XI
-------
LIST OF FIGURES (continued)
Np_. Page
60 O.D. of Exposed Armco Tubing Showing Stringers Along the Tube 277
Drawing Axis and Light Pitting
61 O.D. of Exposed 316L Tubing Showing Corrosion Pits 278
62 I.D. of Exposed 316L Tubing Showing Less Severe Corrosion 278
63 Composite Tube Used in Dynamic Test 284
64 Composite Welded Tubing: Corrosion Effects in the Tube Interior 285
65 Section of Armco Tubing Interior 286
66 Section of 304 Tubing Interior 286
67 Weld Zone of Armco-304 Showing Chilled Structure and Sensiti- 287
zation by Chromium Carbide
68 Structure of 304 Away from Weld ^ 287
69 Weld Zone of 316L-321 Showing Sensitization 288
70 Weld Zone of 304L-Carpenter 20 Cb/3 Showing Chilled Structure 288
and Grain Boundary Sensitization
xii
-------
LIST OF TABLES
No. Page
1 Raw Lower Kittanning Coal Analyses 13
2 Lower Kittanning Coal Ash Composition 15
3 Lower Kittanning Coal Particle Size Distribution 16
4 Rate Data on L-R Processing of 100 Mesh Top Size L.K. Coal 27
with 3 Wt. % Iron Solution at 120°C and 100 Psig
5 Rate Data on L-R Processing of 100 Mesh Top Size L.K. Coal 31
with 5 Wt. % Iron Solution at 120°C and 100 Psig
6 Process Mass Balance Data 33
7 Pyrite Removal from Coal with Iron Sulfate Solution at 102°C 47
and Ambient Pressure
8 Measured and Predicted Pyrite Removal at the "Mixer Unit 51
Operation"
9 Analyzed and Predicted Pyritic Sulfur Content of 100 Mesh 61
Top Size L.K. Coal as a Function of L-R Processing Time
at 120°C
10 Analyzed and Predicted Pyritic Sulfur Content of 14 Mesh 62
Top Size L.K. Coal as a Function of L-R Processing Time
at 120°C
11 Analyzed and Predicted Sp of L.K. Coal as a Function of L-R 81
Processing Time at 110gC and 130°C
12 Pyrite Removal in the Settler Operation 98
13 Typical Wash Section Data from L-R Processed L.K. Coals 101
14 Elemental Sulfur Product Recovery from the L-R Processing 105
of L.K. Coal
15 Elemental Sulfur Recovery by Successive Stages of Hexane- 109
Toluene Leaching
16 Vaporization of Elemental Sulfur from Ferric Sulfate 111
Leached Coal
17 Trace Element Analysis Accuracy Verification 129
xi 11
-------
LIST OF TABLES (continued)
—' - Page
18 Coal and Reagent Trace and Minor Elements 131
19 Trace Cation Content of Calcium Precipitates Recovered 133
from Spent Reagent
20 L-R Processing of 3/8 Inch x 0 L.K. Coal in 5 Wt. % Fe 137
Reagent at 120°C and 100 Psig
21 Pyritic Sulfur Removal from 1/4 Inch x 0 L.K. Coal at 102°C 138
22 Chemical Removal of Pyritic Sulfur from Cleaned 8 x 14 Mesh 139
Lower Kittanning Coal at 102°C
23 Pyrite Leaching of 1/4 Inch x 0 Lower Kittanning Coal at 102°C 141
24 Preliminary Data on the Chemical Removal of Pyritic Sulfur 142
from Size-Fractions of Lower Kittanning Coal at 102°C
25 Process Mass Balance for Fine Coal 162
26 Coal Desulfurization Process Equipment List 173
27 Sources of Equipment Cost Information 178
28 Sources of Operating Cost Information 180
29 Effect of Inert Gas Buildup on Reactor Section Annual Cost 193
30 Effect of Compressed Air on Reactor Section Annual Cost 195
31 Equipment Costs for a Three-Reactor Configuration 197
32 Coarse Coal Process Equipment Lists 213
33 Annualized Costs for Battery Limits Desulfurization Plants 214
34 Economic Evaluation Criteria 224
35 Constants for Use in Economics Evaluations 225
36 Case 1, Cleaned Fine Coal 226
37 Case 2, ROM Coarse Coal 227
38 Case 3, Deep Cleaned Fine Coal with 50% Meyers Process Bypass 278
xiv
-------
LIST OF TABLES (continued)
Np_. Page
39 Case 4, Deep Cleaned Coarse Coal with 50% Meyers Process Bypass 229
40 Upgrading Processing Costs ^ 230
41 Comparison of Total Sulfur Analysis Techniques 236
42 Coal Sulfur Forms Analysis Investigations - Comparison 240
of Techniques and Effect of Coal Digestion Time
43 Comparison of Corrosion Rates of Armco and 316/316L Steels 252
44 Corrosion Data from Literature on 304 and 316SS 253
45 Corrosion Data from Literature on Various Alloys 254
46 List of Materials Tested 258
47 Ferric Sulfate Corrosive Medium 259
48 Compilation - Coal Corrosion Material Evaluation 262
49 Changes in Mechanical Properties of Alloys After Exposure to 280
Iron Sulfate Solution
50 Summary of Qualitatively Observed Events During Corrosion 281
Testing
xv
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the valuable assistance received in
this project from the following TRW personnel: L. Ledgerwood, S. Ziegler,
and D. Kilday for substantial assistance in the experimentation; C. Murray
and M. Wong for assistance rendered in process economics; D. Hopp for
assistance in computer programming; J. Blumenthal and B. Dubrow for mana-
gerial assistance and manuscript review; and L. Broberg, M. Ramirez, and
V. Melough for technical typing; special acknowledgment is due to V. Butler
who undertook the difficult and arduous task of report coordination and
finalization.
The authors owe appreciation to Lloyd Lorenzi, Jr., the monitoring
Project Officer for the Environmental Protection Agency under this contract
for his constant interest, cooperation and valuable comments on the project,
and to T. Kelly Janes, also of EPA, for guidance and encouragement.
Gratitude is due to A. W. Deurbrouck of the U.S. Bureau of Mines
(Bruceton, Pennsylvania) for providing the coals used in this bench-scale
development program and to R. Kaplan of the Commercial Testing & Engineer-
ing Company (Chicago, Illinois) for his cooperation in expediting coal
analyses for TRW.
xvi
-------
METRIC CONVERSION FACTORS
In compliance with EPA policy, metric units have been used extensively
in this report (followed by British units in parentheses). However, in
some cases, British units have been used for ease of comprehension. For
these cases, the following conversion table is provided:
British
Metric
1 Btu
1 Btu
1 kw
1 hp (electric)
1 psi
5/9 (°F-32)
1 inch
1 ft
1 ft2
1 ft3
1 gallon
1 pound
1 ton (short)
252 calories
2.93 x 10"4 kilowatt-hours
1,000 joules/sec
746 joules/sec
2
0.07 kilograms/cm
°C
2.54 centimeters
0.3048 meter
2
0.0929 meters
0.0283 meters3 or 28.3 liters
3.79 liters
0.4536 kilograms
0.9072 metric tons
xv 11
-------
CONCLUSIONS
General Conclusions
1. The Meyers Process applied to high-ash run-of-mine coal is capable
of removing in excess of 90 percent of the pyritic sulfur from 14
and 100 mesh top-size coal and at least 80 percent from coarser
coal (1/4 inch top-size)..*
2. Physical cleaning of coarse coal appears to enhance the pyrite
leaching rates of the Meyers Process. iPyrite removals in excess of
90 percent were attained with cleaned 1/4 inch top-size Lower
Kittanning coal.
3. Meyers Process operation under simultaneous coal Teaching-reagent
regeneration processing is feasible and efficient. It appears to
be the most economic mode of processing high pyritic sulfur coal
of at least up to 14 mesh top-size.
4. Simultaneous Teaching-regeneration processing of coal up to at
least 130°C and 150 psig (135 psi oxygen pressure) does not
measurably affect the organic matrix of the coaTs tested.
5. Pyrite Teaching rates are not affected by the quantity of coaT
present in the slurry. Processing of sTurries containing up to
at least 33 wt percent solids is feasible. Process coal through-
put doubled and filtration time (cost) was reduced when the slurry
solids concentration was upgraded from 20 to 33 wt. percent.
6. Solid-Tiquid separations of process sTurries are feasibTe by
commerciaTTy available equipment (vendor tests). The filtration
costs of 14 mesh top-size coal slurries are approximately one-half
those of 100 mesh top-size coal slurries while the difference in
pyrite leaching rates is no more than 20 percent.
7. The process may be engineered in-a number of basic design config-
urations including simultaneous leach and regeneration, separate
Teach and regeneration, use of oxygen or air for regeneration,
fine coaT or coarse coal processing, and in combination with coal
cleaning; these process design schemes lead to stand-alone full
capital cost estimates of $30-80/KW of power plant name capacity.
8. The estimated coal desulfurization costs, annualized on a utility
financed basis, range between $0.33 and $0.51/MM Btu. Assuming
ROM coaT costs of $20/ton, the costs of the desulfurized fuel are
estimated to range between $T.T4 and $T.32/MM Btu.
EPA policy is to express all data in Agency documents in metric units.
For those particular non-metric units utilized in this report conversion
factors have been provided. These factors are Tocated on page xvii.
1
-------
Specific Conclusions
1. The efficiency and selectivity of ferric sulfate solutions' in the
leaching of pyrite from coal observed during previous investiga-
tions! was reproduced in the current program with high ash
(30 percent) Run-of-Mine Lower Kittanning (ROM L.K.) coal containing
4 percent pyritic sulfur. Eighty-three percent of the pyrite was
removed from 100 mesh top-size L.K. coal during 8.5 hours of pro-
cessing a 20 wt percent coal slurry at 102°C under ambient pressure.
2. The L-R processing (simultaneous coal leaching-reagent regeneration)
scheme developed in this program led to substantial improvement
in the rate of pyrite removal from fine coal, up to at least 14
mesh top-size, over that obtained under ambient pressure processing.
Approximately eighty percent of the coal pyrite was removed in only
two hours when 20 wt percent slurries of 100 mesh top-size L.K.
coal were processed under L-R conditions at 120°C and 100 psig
pressure (85 psi oxygen). The large improvement in pyrite leach-
ing rates appears to be principally due to temperature increase.
The rate advantage of L-R processing ceased to apply when pyrite
removal exceeded approximately 80 percent.
3. The commercial scale processing scheme developed from bench-
scale data for the desulfurization of high pyrite coal consists
of the following unit operations connected in series: mixer,
L-R reactor, ambient pressure reactor, coal-reagent separation,
processed coal wash, coal-water separation, elemental sulfur
recovery, and coal drying. All unit operations were successfully
tested at bench-scale.
4. Pyrite removal occurs in the mixer, L-R reactor, and ambient
pressure reactor. The removal rate in all three reactors is gov-
erned by the empirical leaching rate expression
where
W = wt percent pyrite in coal,
Y = ferric ion- to- total iron ratio in the reactor
reagent, and»
K. = rate constant, a function of temperature and
coal top-size.
-------
5. Up to approximately 80 percent pyrite removal, the rate constant
shows a strong temperature dependence expressible by
= A, exp (-
EL/RT)
Beyond 80 percent removal, the KL value appears virtually unaffected
by temperature in the range investigated (90°C to 130°C). Appar-
ently, a change of reaction mechanism occurs when pyrite removal
exceeds 80 percent; it is speculated that a diffusion controlled
process takes over. This change in mechanism may be specific to
this particular high-ash coal.
6. The E|_ value is independent of temperature, coal top-size (up to
14 mesh), or reactor unit operation (processing conditions). The
value of AL is independent of temperature under mixer and L-R
operations, but different for each operation. The reasons for
this unexpected AL value dependence on mode of operation are not
apparent from the available data. The estimated EL and A Lvalues
for mixer and L-R reactor processing and those of the diffusion
rate constant are summarized in the report.
7. Reagent regeneration is governed by the rate expression
rR = - dFe
dt
• KR po
where
exp (-
ER/RT),
= oxygen partial pressure, and
+2
Fe = ferrous ion concentration in the reagent solution.
AR and
are constants.
The reagent regeneration rate operates simultaneously with the
leaching rate in the L-R reactor.
The KL value at 120°C under L-R processing conditions is five
times higher than the KL value for 102°C processing under ambient
pressure conditions (separate leaching and regeneration); both the
14 and 100 mesh top-size coals exhibit the same large increase in
KL value. The estimated EL justifies only a two- fold increase
in Ki for 20 degree temperature rise-observed between 80° and
102°C (separate Teaching-regeneration) and between 110° and 130°
(L-R processing). The 250 percent increase in the KL value,
which appears to be the result of changing the mode of processing,
could not be attributed to reaction of oxygen with pyrite.
-------
Additional rate data of more fundamental nature are required to
delineate the reasons for the unexplained increase in KL.
9. The K|_ value obtained with 100 mesh top-size L.K. coal was approx-
imately 20 percent larger than the K|_ obtained with 14 mesh top-
size coal. Since solid-liquid separation efficiency increases
substantially with coal top-size, it appears preferable to process
14 mesh coal in view of the small rate penalty involved. .
10. The coal top-size effect on K|_ becomes substantial with ROM coals
coarser than 14 mesh top-size. The ratio of K|_ values obtained
at 102°C with 100 mesh and 1/4 inch top-size L.K. coals is esti-
mated to be equal to 4 up to approximately 60 percent pyrite re-
moval and between 6 and 9 for higher pyrite removal. L-R
processing of coarse ROM coal at 120°C did not appear to improve
the above ratios. In fact, in the case of 3/8 inch top-size ROM
coal increasing the temperature from 102°C to 120°C has probably
no effect on the leaching rate when pyrite removal exceeds 20-25
percent.
11. Despite the adverse effect of coal particle size on pyrite leaching
rates, coarse ROM coal desulfurization by the Meyers Process is
feasible. Pyrite reduction in excess of 80 percent was attained
when 1/4 inch x 0 ROM L.K. coal was leached for 48 hours with
ferric sulfate solution at 102°C. The pyritic sulfur content of
the ROM coal was reduced from 3.48 wt. percent to 0.66 wt. percent
and its heat content increased from 11,822 to 13,023 btu per pound.
12. Preliminary data on cleaned coarse coal processing by the Meyers
Process indicate that physical cleaning of high pyrite ROM coal
prior to chemical desulfurization enhances substantially the
pyrite leaching rate. At 102°C, the pyrite leaching rate obtained
from processing the 8 x 14 mesh fraction of cleaned L.K. coal was
twice that obtained from processinq the same size fraction of ROM
L,K. coal. The pyritic sulfur content of the cleaned fraction was
reduced from 1.06 to 0.09 wt. percent.
13. In addition to temperature and coal particle size, the Ki value
may also be affected by the type or source of coal. Data obtained
from the L-R processing of Marion mine coal (Upper Freeport coal)
indicate that the pyrite leaching rate constant applicable to this
coal is at least an order of magnitude higher than corresponding
KL values applicable to L.K. coal. This comparison assumes that
the same leaching rate expression applies to both coals.
14. The concentration of coal in the slurry and the concentration of
iron in the reagent affect the leaching rate through their effect
on Y. These effects become negligible under continuous exchanqe
or L-R processing. Thus, in principle, the coal content of the
• slurry is limited only by equipment limitations in the transfer
of thick slurries and the iron content of the reagent is limited
only by the solubility of the iron sulfate in water
-------
15. The sulfate sulfur-to-elemental sulfur ratio of the product sulfur
of the Meyers Process is independent of the mode of processing or
the value of the processing parameters used within the ranges
investigated. The value of the ratio is approximately 1.5 indi-
cating that 60 percent of the reacted pyritic sulfur converts to
sulfate sulfur and 40 percent to elemental sulfur.
16. Elemental sulfur recovery is virtually complete by extraction of
processed coal with toluene. However, preliminary data revealed
that recovery by vaporization is also efficient. The latter
scheme combines sulfur recovery and coal drying into a simple
operation and, therefore, appears to be the more desirable method
of sulfur recovery. Additional experimentation is needed in
order to optimize sulfur product recovery techniques.
17. Special investigations revealed that the Meyers Process does not
affect standard coal analysis techniques.
18. Experimentation with potential reactor construction materials for
the Meyers Process revealed that 316L stainless steel could be
safely used, but 304 stainless steel was inadequate.
19. Process analyses for both fine and coarse coal processing show
that the reactor/regenerator section of the process accounts for
a large part of the equipment cost. At commercial scale it was
estimated that 40-60 percent of the total equipment cost occurs
in this section,
20. All liquid streams of the Meyers Process are being recycled.
Because of the closed loop design of the process, separate
optimization of the reactor section generally leads to inoperable
designs. The reactor trade-offs must be coupled to the wash and
by-product recovery sections in all process optimization efforts.
21. The process analyses showed that either low purity oxygen (95%)
or high purity oxygen (99.5%) could be used with no significant
process cost impact. It also showed that air can be competitive
if energy is efficiently recovered from the reactor vent gas. Air
regeneration may be applicable only to high pyrite coals where
the heat liberated by regeneration is sufficient to saturate the
large volume of inert vent gas.
-------
RECOMMENDATIONS
1. The principal unit operations of thevMeyers Process are suffi-
ciently developed at bench-scale to merit immediate scale-up to
the pilot plant stage.
2. Bench-scale testing should continue in support of pilot plant
testing for identification of process improvements and for the
evaluation of process application to new areas or coals. Bench
scale represents the most cost effective scale to investigate
process improvements and expanded utility.
3. In support of pilot plant testing it is recommended that bench-
scale data be generated to aid in coal selection and in test plan
formulation best suited to each of the selected coals. Bench-
scale investigations should also be used as an aid to understand
and explain unexpected observations and to resolve problems that
may occur.
4. In the area of process improvements, it is recommended that bench-
scale investigations be performed on improved techniques for sulfur
product recovery, on schemes for coal trace element recovery, on
means of product utilization or disposal, and on the identification
and preliminary evaluation of process modifications with the
potential for reducing processing costs; e.g., simultaneous sulfate
and elemental sulfur recovery and elimination of coal wash, combin-
ation of physical coal cleaning and chemical desulfurization.
5. In the area of expanded process utilization, bench-scale studies
are recommended to investigate the feasbility of using the Meyers
Process to render presently unusable coking coal acceptable to
metallurgical industry (high sulfur coal, middlings) and to exam-
ine the advantages of combining chemical desulfurization with
coal conversion. Simultaneous chemical desulfurization and non-
pyrite ash reduction should also be investigated.
6. Coal pyrite leaching rate studies of more fundamental nature than
previously performed are needed and recommended. These studies
should aim at prediction of the efficiencies of chemical coal
desulfurization from easily measured physical and/or chemical
properties of coal. At a minimum, studies should be undertaken
which identify the influence that the coal matrix exerts on coal
pyrite leaching rates.
7. Process designs should be prepared for low pyrite coals to identify
the ranges of parameters that should be included in the experi-
mental investigations. Emphasis should be placed on the reactor/
regenerator area which continues to be the major cost portion of
the process.
-------
1 . INTRODUCTION
The Meyers Process can be used to desulfurize a large number of coals
prior to combustion in order to meet govermental requirements for sulfur
oxide emissions control.
The process consists of several steps including crushing, chemical
treating, sulfur removal and solution regeneration. While the process is
new, the unit operations are similar to various existing technologies such
as processes for the heap leaching of copper, regeneration of steel mill
waste pickle liquor and recovery of elemental sulfur from volcanic ash.
The chemistry of the process is represented by the pyrite leaching and
reagent regeneration steps shown below:
FeS2 + 4.6 Fe2(S04)3 + 4.8
2.4
9.6 FeS0 + 4.8
10.2 FeS04 + 4.8 H2S
4.8 FeS0) + 4.8
0.8S
overall process reaction:
FeS2 + 2.4 02
0.6 FeS04 + 0.2 Fe2(S04)3 + 0.8S
Generated iron sul fates can be recovered by crystallization and stabilized
for disposal or a part of the iron sul fate product may be neutralized with
lime to yield gypsum. For the processing of fine or suspendable coal
(e.g., 14 mesh top size) leaching and regeneration are often economically
combined in a single process unit, while for the processing of coarse coal
(e.g., -3/8 inch) the two reactions are more economically performed
separately.
1 The Meyers Process is more applicable to coals rich in pyritic sulfur
rather than those in which organic sulfur is in high concentration. Such
coal is found in the Appalachian region of the United States which now
supplies 60 percent of the current U.S. production. An estimated one-third
of the Appalachian production can be lowered to sulfur contents of 0.6 to
0.9 percent at which level the emission requirements for new power plants
can be met. Additional Appalachian and Interior Basin coal can meet the
state emission standards by this process.
7
-------
An initial bench-scale testing program for the definition and obtaining
of critical design data for the Meyers Process for the chemical removal of
pyritic sulfur from coal was performed. In addition, a survey program
covering application of the process for desulfurization of pulverized, raw
run-of-mine coal from 35 U.S. coal mines has been completed. ' Run-of-mine
coal was used in all of these experimentation in order to accentuate any
possible processing difficulties which might be caused by the presence of
unusually large concentrations of pyrite and inorganic ash. The effort
demonstrated that the process, even when operating on ROM coal, has wide
applicability for the reduction of sulfur content of U.S. coal to levels
consistent with the Federal Standards for New Stationary Sources.
This bench-scale program was aimed at optimizing and improving the
critical leaching and regeneration process steps, evaluating analytical
procedures and studying other process improvements. As before, high ash
run-of-mine coal was used for the majority of the experimentation, while
studies aimed at determining the benefits of clean coal processing utilized
coal typical of washed Appalachian coal.
This effort resulted in the attainment of the necessary data and de-
finition of some significant improvements. In doing so, more than fifty
fully material balanced process simulation runs, requiring approximately
2000 solution analyses and 1000 individual coal analyses, were performed.
These data were assessed and evaluated in combination with the results re-
ported in the initial bench-scale program.
The resulting report is necessarily extensive, and therefore a guide
is provided to the reader who wishes to focus his attention on specific
program results. The results in this volume are presented in six major
sections as follows:
• Pyritic Sulfur Removal from Suspendable Coal
• Reagent Recyclability-Trace Flement Data
-------
t L-R Processing of Coarse Coal
• Process Engineering
• Chemical Analysis Studies
• Materials Compatibility
while the complete data base is presented in Volume II of this report.
Those readers desiring to review process unit operations for suspend-
able coal are directed to Sections 2.4 through 2.7 of this report. Those
readers desiring process design, estimated process capital cost and
estimated operating costs are directed to Sections 5.1 and 5.3. Coarse
coal depyritization data and the engineering analysis for a coarse coal
processing scheme are presented in Sections 4 and 5.2, respectively.
Section 6 of the report is devoted to sulfur analysis techniques for
processed coal and the identification of potential process monitoring
techniques. Section 7 presents data on materials compatibility to process
environment.
-------
2. PYRITIC SULFUR REMOVAL FROM SUSPENDABLE COAL
The bulk of the program effort was expended in the processing of
suspendable coal. Coal is defined here as suspendable if its top size
does not exceed 10 mesh. Two1 coal sizes were investigated: 100 mesh x 0
and 14 mesh x 0. The emphasis in experimentation was placed in optimization
of pressurized simultaneous coal Teaching-reagent regeneration (L-R) pro-
cessing and the parameters affecting this type of processing.
The ensuing report sections describe in detail the experimentation
performed, the procedures used, the data generated, and our interpretation
of the data.
2.1 COAL SELECTION AND SAMPLE PREPARATION
A Lower Kittanning coal seam feed was selected for the majority of
experimentation on this program while the Upper Freeport Seams Sample was
processed for comparison purposes. This selection was based on the avail-
ability of large deposits of these coals, the fact that a substantial data
bank existed on ambient pressure processing of Lower Kittanning coal by
1 ^
the Meyers Process, the availability of data on Upper Freeport coal from
the coal survey project, and the similarity of process behavior of these
seams to most Appalachian coals.
Raw run-of-mine Lower Kittanning coal was ground to 3/8 inch x 0 size
and shipped to TRW by the Bureau of Mines, Pittsburgh, Pennsylvania at
the request of the EPA Project Officer. The coal was received in three
55-gallon drums with each drum containing approximately 150 kilograms of
coal.
11
-------
The coal from each drum, when needed for experimentation, was suc-
cessively halved by the use of commercial rifflers to approximately 20
kilogram lots. This was the minimum lot size used for further grinding
(100 mesh and 14 mesh top sizes). Upon grinding, each 20 kilogram lot
was subjected to the same riffling procedure (halving) until the desired
coal sample siz.e was attained. The nominal coal sample sizes used in this
program were 2 kilograms (20 wt. percent slurry processing) and 4 kilograms
(33 wt. percent slurry processing). Samples submitted to analyses were in
the 200 to 300 grams range. The described riffling procedure is the most
commonly used method for coal sampling and it is ASTM approved (ASTM
sampling analysis procedures utilized in this program are listed in
Appendix G, Volume II). Both the "as received" and the unused portion
of the ground coal were stored under helium in order to inhibit weathering.
However, sulfur-forms analyses of the unprocessed coal performed periodi-
cally during the program indicated that some coal weathering took place
in spite of the precautions taken (see Table 1).
The unprocessed coal was completely characterized prior to initiation
of processing. The coal characterization procedure included complete coal
analyses and coal particle size distribution determinations. The coal
analyses involved short proximate (moisture, ash, and heat content), sulfur
forms, ash composition, and trace element determinations. In addition,
short proximate, sulfur forms, and ash-iron analyses were performed period-
ically on the unprocessed coal during the duration of the program in order
to assure proper sampling and to monitor coal weathering. The coal
analyses, with the exception of trace element analyses, were performed by
Commercial Testing & Engineering Company (CT&E), Chicago, Illinois. The
trace element analyses were performed by TRW's Applied Chemistry Department.
Table 1 summarizes the analysis data on the raw Lower Kittanning coal
used in this program. The data are presented chronologically with respect
to the dates of analysis in order to illustrate the increase in the sulfate
content of the coal with storage time and the associated decrease in pyrite.
12
-------
TABLE 1. RAW LOWER KITTANNING COAL ANALYSES
Date
Analyzed
Oct.
1973
Dec.
1973
Apri 1 -
June
1974
May
1974
June
1974
Sept.
1974
Oct.
1974
Coal
Top Size
100
Mesh
14
Mesh
100
Mesh
14
Mesh
14
Mesh
3/8 inch*
14
Mesh
Number
Of Samples
Analyzed
3
2
4
2
1
3
4
Coal Composition, Dry
Ash
(Wt.%)
30.08
+ .02
29.76
;K70
30.85
**
32.32
+ .35
30.12
28.91
+ .52
31.77
+ .71
Heat
Content
(Btu/Lb)
10,520
± 26
10,626
+ 173
10,275
**
10,109
+ 106
10,324
10,448
+ 106
10,020
t 77
Total
Sulfur
(Wt.%)
4.55
+.03
4.48
+ .01
4.53
+.06
4.77
+.01
4.50
4.28
+ 14
4.55
+.07
Pyritic
Sulfur
(Wt.%)
4.05
+ .01
4.01
+ .01
3.88
+.19
3.85
+.02
3.85
3.56
+ 18
3.77
+.04
Sulfate
Sulfur
(Wt.%)
0.08
+_.02
0.07
+ .01
0.24
+ .04
0.28
+.01
0.26
0.35
+_.03
0.41
+.02
Organic
Sulfur
(Wt.%)
0.41
+ .04
0.41
+ .01
0.41
+ .17
0.64
+ .01
0.39
0.37
+ .03
0.38
+ .05
Iron
(Wt.%)
3.91
+ .03
4.03
+ .12
3.98
+ .06
4.20
+ .04
3.89
3.68
+ .13
4.00
+ .09
Moisture
In Raw
Coal
(Wt.%)
1.20
+ 34
1.51
±.03
1.19
+ .22
1.81
1.23
1.16
+ .04
1.66
+_.19
These samples were taken from a different drum than the rest.
**Single analysis.
-------
The observed trend with time in the values of these two sulfur forms, the
internal consistency in individual sample analyses (each increase in
sulfate is associated with a corresponding decrease in pyrite), and the
good agreement in the total sulfur and iron values among coal samples
derived from different ground lots and analyzed at different times indicate
that the variations in pyritic and sulfate sulfur from sample to sample
are not due to sampling or analysis error but to coal weathering. With
one exception, the data in Table 1 verify the adequacy of raw coal sampling
procedures used by TRW and the consistency in analyses performed at CT&E.
The indicated standard deviations are within ASTM standards.
The exception to the overall consistency of data in Table 1 involves
the two samples of 14 mesh x 0 coal analyzed in May 1974. The higher ash,
lower heat content, higher total sulfur, and higher iron values of these
samples (compared to the rest of the data on coal derived from the same
drum) would be internally consistent and would indicate inappropriate
sampling if the analyzed higher total sulfur was due to higher inorganic
sulfur forms in the sample (higher inorganic sulfur leads to higher ash
and lower heat content in the sample). However, the sulfur forms data
from these two samples show that the unexpected high total sulfur content
was due to high organic sulfur. This analysis is not consistent with
the higher iron and ash content of the samples which suggest that the
pyrite or iron sulfate forms should have exhibited the high values. Thus,
in addition to sampling procedure the sample analyses became suscept.
The same lot of ground coal was re-sampled and reanalyzed in June and
yielded the expected values (Table 1).
Discrepancies in sulfur forms analyses were more frequent with pro-
cessed coal. The low sulfur-high ash content of the processed coal and/or
residual elemental sulfur left on the coal during processing could have
been the culprits of poor sample analysis. Whenever these discrepancies
occurred, the coal sample analysis data were subjected to the same internal
consistency scrutiny as that described for the raw coal.
s
Table 2 presents the data on the mineral content of the coal utilized
(ash composition). These data were supplemented with trace element
14
-------
analyses; the latter are presented in a later section of this report de-
voted exclusively to the fate of trace elements during coal processing.
The ash composition data appears to be typical of Pennsylvania bituminous
coals (the quantity of ash in the coal was not typical; its softening
temperature was 1385*C (2525°F)).
TABLE 2. LOWER KITTANNIN6 COAL ASH COMPOSITION
Ash Component Wt. %
Phos. pentoxide, p og 0.11
/
Silica, Si02 49.63
Ferric oxide, Fe203 18.66
Alumina, A120 25.64
Titania, Ti02 1.87
Lime, CaO 0.75
Magnesia, MgO 0.53
Sulfur trioxide, S03 0.79
Potassium oxide, K20 1.64
Sodium oxide, Na20 0.29
Undetermined 0-09
100.00
15
-------
Table 3 presents data on the particle size distribution of the re-
ceived 3/8 inch x 0 coal and of the two suspendable coal sizes investigated
in this program (14 mesh x 0, and 100 mesh x 0). At least three determi-
nations were performed on each of the coal top sizes; the sample to sample
reproducibility was very good. The particle size distribution presented
in Table 3 proved to be almost identical to that obtained from the Lower
Kittanning coal used in the previous program (EPA contract EHSD 71-7).
TABLE 3. LOWER KITTANNING COAL PARTICLE SIZE DISTRIBUTION
Size Range
(Mesh)
+6
-6 +14
-14 f28
-28 +42
-42 +60
-60 +80
-80 +100
-100 +115
-115 +150
-150 +200
-200 +325
-325 xO
Lost
Three sample ave
Weight Fractions*
100 Mesh x 0
.090
.051
.054
.136
.592
.071
.006
rages (maximum de
14 Mesh x 0
.020
.200
.160
.090
.080
.040
.040
.030
.100
.210
.020
.010
viation 10%)
3/8 Inch x 0
.248
.260
.150
.093
.054
.038
.018
.019
.013
.026
.054
.021
.006
16
-------
Coal grinding and riffling to the desired sample size for processing
(2 to 4 kilograms) were the only operations performed on the coal prior
to mixing it with the reagent fdr processing. Thus, the data generated
on this program was on ground run-of-the mine coal.
2.2 COAL PROCESSING EXPERIMENTATION
As a result of previous bench-scale investigations on the Meyers
Process, a conceptual process scheme was proposed, for processing suspend-
able coal, involving the following basic unit operations: slurry
preparation (mixer), simultaneous coal Teaching-reagent regeneration (L-R
reactor), ambient pressure residual pyrite leaching-coal settling (settler),
coal washing, elemental sulfur recovery, coal drying, and product sulfate
recovery. The major objective of the current program was to investigate
the validity of the above conceptual scheme. The principal difference
between the above scheme and the one defined in the previous bench-scale
program is L-R processing (simultaneous Teaching-regeneration under pres-
sure versus separate coal leaching and reagent regeneration unit operations
with coal leached at ambient pressure). Thus, the emphasis was placed on
L-R processing; however, its effects on the other unit operations as well
as potential improvements on these operations were also investigated.
The investigation of product sulfate recovery (process slip-stream treat-
ment) was postponed until the basic process parameters are finalized
(especially reagent composition). Preliminary examination of this unit
operation has been proposed to commence at the conclusion of this program.
Figure 1 is a block diagram of the basic processing scheme used in
the investigations of suspendabTe coaT. Every effort was made to simuTate
rthe envisioned full-scale processing to the degree allowable by the size
of bench-scale equipment. Figure 2 depicts in greater detail the front
end of the process and the apparatus utilized (mixer, L-R reactor, and
settler). The ensuing paragraphs describe the procedure and the equipment
used.
17
-------
00
COAL
11
COAL
EAGE
MIXER
PRODUCT
UUAL
DRIER
PRODUCT
SULFUR
LJLFUR
DISTILLATION
REAGENT
\
:NT
:R
10° C
SLURRY
y
COAL LEACH
REAGENT
REGEN.
110-130°C
A
SLURRY
| f
COAL
SETTLER
90° C
SLURRY ||^ER
•Mm
^ |
FILTER
\\\\\\
i
TOLUENE
WET COAL
ELEM.
SULFUR
RECOVERY
SPENT
TOLUENE
I
FILTER
HOT
WATER
SLURRY
FILTER
T/rrrn
HOT
WET WATER
COAL I
COAL
WASHING
80-90°C
WATER (DISPOSAL)
Figure 1. Basic Bench-Scale Processing Scheme for the Removal
of Pyritic Sulfur From Coal
-------
Pressure vJet
Control Test
Meter
Pump Seal
Purge System
1
1
1 '
1
1
1
1
f
i
90° - 95°c
•'.- '";- '- V •;>'•• ,.--••->.;-;.
ilSlliiflii
linSSffi^^Sli^
1 I
i
i
,1
21
1
^t
1
!',eactor-Sett ler •
riNAL PROCESSINGS
Sti rrer Reactor
_l
STEP #1 (-v.1.0 HOUR)
STEP #2 (1-8 HOURS)**
STEP #3 (18-24 HOURS)
•'Final processing includes elemental sulfur recovery, coal washing and drying.
**Reactor volume ^13 liters
Figure 2. Bench-Scale Coal Leaching and Reagent Regeneration Apparatus
-------
The ground coal (14 or 100 mesh top size) was mixed with hot spent or
fresh reagent in the "mixer" and the resulting slurry was brought to
boiling; slurry boiling was maintained until coal wetting was complete
(slurry foaming subsided). Actually, the "mixer" was comprised of 3 to 4
four-liter glass flasks equipped with heating tape and stirrer; each flask
contained approximately 2 liters of slurry. The "defoamed" slurry was
then transferred to the L-R reactor; the reactor was pressurized with
nitrogen gas to the desired operating pressure (50 to 150 psig) and the
slurry was heated to the desired L-R temperature (110 to 130°C). This
entire "slurry preparation and heating" operation ("mixer operation") re-
quired approximately one hour (mixing time, tm), although in early experi-
mentation it was substantially longer. Once the desired L-R temperature
was reached, a slurry sample was taken and then oxygen was introduced to
the reactor. The composition of the slurry sample taken at this point
(t,n = 0.0 hours) defined the starting coal and reagent composition of
L-R processing and simultaneously served as the means of determining the
extent of reaction during t .
The L-R reactor and its associated parts (feed and sampling lines,
pump, and slurry circulation loops) were constructed from 316 stainless
steel stock. The reactor was a cylinder 100 cm (39.4 inches) in length
with a 14 cm (5.5 inches) OD and 12.7 cm (5.0 inches) ID. The reactor
was flanged at the one end (top) and a hemispherical bottom was welded to
the other end. Both end pieces were equipped with 2.54 cm (1 inch)
openings.
The reactor was loaded with slurry through the top and drained through
a one inch ball valve attached to the reactor bottom. The reactor was
operated one-half to three-quarters full.
During L-R processing the slurry was circulated through a loop ex-
ternal to the reactor (Figure 2) at a rate of approximately 4 liters per
minute. Slurry circulation was necessary for oxygen introduction and for
sampling. Slurry was circulated with the aid of a 316 stainless steel
centrifugal pump having a 3.2 cm (1.25 inches) OD suction and an 1.9 cm
20
-------
(0.75 inches) OD discharge. The pump was equipped with a water irrigated
double mechanical seal. The slurry was withdrawn from a reactor opening
located 30 cm from its bottom and, after oxygenation, it was returned
through a tangential inlet to the reactor bottom. The main circulation
loop was equipped with two by-pass loops. One was used to trap slurry
samples at the desired reaction times. The second was used to return part
of the oxygen depleted slurry through the top of the reactor as a spray.
The principal function of the second by-pass was to control flow through
the main loop without the need of excessive throttling of the pump; an
intended second function during early experimentation was slurry foam
control. However, top spraying the reactor slurry did not prove to be
a very effective foam control device). This by-pass loop was assembled
from a variety of stainless steel tubing and as such it also served as a
means of testing the compatibility of these materials with L-R processing
(the materials involved and the data derived from these tests are de-
cribed in a separate section of this report).
Oxygen was introduced to the slurry immediately downstream of the
sample trap at a rate which varied from approximately 0.5 to 2.0 liters*
per minute; the oxygen feed rate depended on the quantity of coal being
processed and on L-R processing time (higher rates were fed at the start
of L-R to purge the nitrogen in the reactor and to accommodate the higher
reaction rates). In principle, due to the prevailing turbulent flow con-
ditions, intimate gas-slurry mixing took place in the recirculation loop
prior to slurry return to the reactor. The excess oxygen, oxygen not
consumed by reagent regeneration, percolated through the reactor, passed
through a condenser which removed water vapor and entrained reagent solu-
tion, and exited the reactor through an automatic pressure relief valve.
The quantity of reactor vent gas (principally oxygen) was measured Jay a
wet-test meter.
Several slurry samples were taken during L-R processing (normally
every 0.5 hours during the first two hours of processing and every hour
thereafter). The drawn slurry samples were weighed and immediately
i o "H^
filtered. The filtrate was analyzed for iron forms (Fe and Fe ). The
At Standard Temperature and Pressure (STP).
21
-------
reagent wet coal was spray-washed with hot water to remove the reagent
solution from it, extracted in toluene to leach out elemental sulfur,
dried at 100°C under vacuum, and weighed. The weight percents solids in
the slurry samples were determined and compared to the starting values as
a means of evaluating how representative the samples were of the reactor
mixture. The dry coal samples were analyzed for ash, heat content, total
sulfur, sulfur forms, and iron.
After the desired L-R processing time was completed, the reactor was
depressurized and drained. The L-R slurry was transferred immediately to
either a "settler" for further reaction or to the filtration apparatus or
to both. In all cases, because of equipment size restrictions, the slurry
was processed through the post L-R operations in two batches. Three pro-
cessing options were used: (a) both batches were "settler" processed prior
to filtration, washing, elemental sulfur recovery, and drying; (b) only
one of the batches (approximately one-half of the L-R slurry) was "settler"
processed prior to further processing; or (c) neither batch was "settler"
processed. In addition, there were two options in "settler" processing:
(a) reaction in an agitated "settler" or (b) reaction in a non-agitated
"settler". Whatever the option used, slurry processing through the re-
maining operations was the same. (There were a few exceptions where L-R
processed coal was used for special experimentation in washing and elemental
sulfur recovery operations; these will be noted as they occur in data pre-
sentation.)
"Settler" processing was conceived as a means of concentrating dilute
L-R slurries of fine coal while simultaneously removing the last few percent
of pyrite from L-R processed coal. Since it was realized that a relatively
quiescent reactor could be subject to reagent concentration gradients and
a consequent reduction in the pyrite leaching rate, an agitated "settler"
was also used in parallel for comparison. Indeed, the non-agitated "settler"
was not as effective a pyrite leaching reactor as the agitated one because
of reactant and product segregation. The non-agitated settler was abandoned,
during early experimentation, not because of reagent segregation problems
but because L-R processing of thick slurries proved efficient and a thickener
was judged unnecessary. The agitated "settler" was retained as an ambient
pressure reactor where the slow reacting, last few percent of pyrite could
22
-------
be leached out of coal. Thus, in the majority of experimentation one-half
of the L-R slurry was processed through an agitated settler and the other
half was transferred directly to the filtration and washing unit operations.
The coal derived from these two slurries (after washing, toluene extraction,
and drying) was used to determine L-R processing and "settler" processing
efficiencies.
The "settlers", covered 5-liter cylindrical glass vessels, were
equipped with heating tapes, condenser, and stirrer (used only for agitated
"settler" processing). The slurry was processed in these vessels for 18-24
hours at 90°C and at ambient pressure. Each settler was sampled at the
start and at the end of processing. The slurry samples were analyzed for
iron forms only.
The hot slurries from either the L-R reactor or "settler" were vacuum
filtered. The reagent-wet coal was subjected to a 2 stage washing scheme
with both stages consisting of the following: a slurry wash with a quan-
tity of water equal to from 2 to 4 times the weight of dry coal estimated
to be in the filter cake; a cake wash with two dry coal weights of water.
The slurry washes were performed at reflux temperatures for 30 minutes.
Additional water washing or coal washing with 0.1 N sulfuric was performed
on processed coals whose analysis indicated excessive iron sulfate de-
position. All filtrates were analyzed for iron forms.
The product elemental sulfur was extracted from the processed coal
with toluene (special experimentation was also performed with hexane).
The water-wet coal filter cake from the wash section was slurried in twice
the estimated dry coal weight of toluene. The slurry was heated to the
toluene-water azeotrope temperature (85°C) where it was maintained until
the water in the slurry (from the wet cake) was distilled-off (the toluene
was returned to the slurry continuously during distillation); the slurry
was then refluxed for 30 minutes at its reflux temperature (108°C). Subse-
quently, the slurry was filtered hot and the resulting coal cake was
rinsed with a small quantity of fresh toluene. The elemental sulfur in the
23
-------
toluene filtrate was recovered as a residue upon vaporization of the toluene.
The toluene residue contained small quantities of dissolved coal along with
the elemental sulfur. Each residue was analyzed (in triplicate) for sulfur.
In the majority of experiments a double toluene extraction was performed.
The toluene-wet coal was vacuum dried in a well trapped oven. The
dried processed coal was then sampled for analysis. Each batch of pro-
cessed coal was subjected to short proximate, sulfur forms, and iron
analyses.
The procedure described was the one followed with the majority of
coal processing experiments. There was, however, special experimentation
performed which involved single unit operations. These special require-
ments are described in conjunction with the data presentation.
2.3 COAL PROCESSING DATA
Approximately 50 batches of suspendable (100 and 14 mesh top size)
Lower Kittanning coal were subjected to L-R processing (simultaneous coal
Teaching-reagent regeneration processing) during the feasibility and
parametric investigations of the L-R mode of operation of the Meyers Pro-
cess. The following parametric effects were investigated over the ranges
indicated:
• Reaction time (Mixer, 1-2 hours; L-R, 1-8 hours; Settler, 18-24 hours)
• Slurry concentration (20 wt. percent and 33 wt. percent coal in
iron sulfate)
• Starting reagent iron concentration (near 0.0 to 5 wt. percent)
• L-R reaction temperature (110°C to 130°C)
• L-R reaction pressure (50 to 150 psig, with a corresponding oxygen
partial pressure variation of 35 to 135 psia)
24
-------
• pH (1-4 to 1.9); actually, this was an investigation of the effect
of sulfuric acid on L-R processing.
Special experimentation was performed in the following areas: ambient
pressure desulfurization of the Lower Kittanning coal by continuous
reagent exchange, slurry foam control, elemental sulfur recovery, coal
washing, solids-liquid separations, and reagent recyclability.
It should be noted that the major objective of^this program was to
test the feasibility of conceptual Meyers Process improvements and to de-
termine the processing scheme and operating parameters of the improved
process. Thus, the parametric investigations were limited to ranges con-
sidered as the most economically and technically practical for process
utilization. The number of parameter values investigated were limited to
the minimum necessary to (1) .identify trends, (2) establish the most
probable parameter values for use in the process design and (3) develop
empirical rate expressions which adequately describe coal pyrite extraction
under the L-R mode of operation of the Meyers Process.
The processing data is summarized in the 15 tables of Appendices A,
and C of this report (Volume 2). Appendix A includes the data derived from
processing 100 mesh top size coal; Appendix B includes the 14 mesh top size
coal data; and Appendix C presents the data derived from L-R processing of
weathered coal with low-iron reagent (the starting reagent does not contain
iron; it derives iron from coal during slurry preparation and processing).
-s
The experiments are numbered consecutively starting with Appendix A through
Appendix C. To a large degree, the numerical order of the experiments
corresponds with the order in which they were performed. Each table con-
tains experiments (usually more than one) performed under the set of nominal
conditions indicated in its title. Specific conditions pertaining to
individual experiments are included with the rate data derived from the
particular experiment. Two types of data are presented for each experiment:
rate data, indicating the change in slurry and coal composition as a
function of reaction time, and mass balance data, showing the over-
all process balance for solids, liquids, sulfur, and iron. Examples of the
25
-------
data presented in Appendices A through C and table explanations are given
below.
Table 4 presents the rate data derived from Experiment 1. The
coal was processed under the conditions indicated and in accordance with
the procedure described in the previous section (Section 2.2). The first
row of data indicate the actual coal (solids) content of the slurry
(as opposed to the nominal or intended value used in the table title),
the Y value of the starting reagent (the ferric ion to total iron ratio
in the reagent), and the composition of the starting coal on dry coal basis.
The next five rows present data derived from slurry samples (150-200
grams each) drawn from the L-R reactor during processing.
The first column of data in Table 4 indicates the L-R reaction times
(t, R) at which samples were drawn. It also shows the time, t^,
required to mix the slurry (to thoroughly wet the coal in order to
minimize slurry foaming) and to heat it to the desired L-R reaction
temperature, 120°C in this case. A sample was always drawn at the start
of L-R processing (end of "mixer operation"); a slurry sample was also
drawn at the conclusion of L-R operation (t, D= 2 hours, in this case).
L-K
The second column shows the solids content of the slurry samples; this data
served as an indication of the adequacy of the sampling procedure. As seen,
the solids content of each sample is nearly identical to that of the
starting slurry. The third column shows the reagent Y value of the liquid
phase of each sample as determined from iron forms analyses. Columns
four through ten present the analyzed dry coal composition data for each
of the slurry samples. The last two columns show the computed pyritic
sulfur, S , removal from coal as a function of reaction time. Pyritic
sulfur removal values were computed from coal sulfur forms analysis data;
the Sp removal during tm was also computed from ferrous ion production data.
The removal values "Based on Analyzed S " were computed from the
differences in the pyritic sulfur content of the starting coal and that
of the particular sample; the S values were determined from coal sample
26
-------
TABLE 4.
RATE DATA ON L-R PROCESSING OF TOO MESH TOP SIZE L.K. COAL WITH 3 WT. % IRON
SOLUTION AT 120°C AND TOO PSIG (85 PSI 02). NOMINAL SLURRY SOLIDS 20 WT. %
RATE DATA
EXPERIMENT NO. 1: tm = 1.5 Hours, tL_R = 2 Hours; tg = 0; Fe in Starting Reagent = 2.98 Wt. %; Starting Y = 1.0
Reaction Time, Hours
(t = time in mixer;
t. = time under L-R
processing;
tg = time in settler)
Starting Coal (5 Sample
Average) and Slurry
Sample
Slurry
Composition
Wt. %
Solids
18.9
Y
(Fe+3/Fe)
1.0
Coal Composition Wt. % (Except Heat Content), Dry
Ash
29.62
+.66
Heat
Content*
Btu/Lb
10,573
± 91
Total
Sulfur,
St
4.52
+.05
Pyritic
Sulfur,
SD
4.03
+_.03
Sulfate
Sulfur,
Ss
0.08
+.02
Organic
Sulfur,
So
0.41
+_.04
Iron,
Fe
3.97
+.12
% S Removal
Based on
Analyzed
SD
Based on
Corrected
SP
Reactor Samples
*L-R = °'° Hours
(tm =1.5 Hours)
t,D =0.5 Hours
•1.0 Hour
=1.5 Hours
=2.0 Hours
18.6
17.7
N.A.
19.0
18.5
.28
.58
.75
.84
.84
26.52
24.95
28.05
27.76
27.89
11,059
11,172
10,310
10,104
9,974
2.95
2.42
2.25
2.58
2.23
2.07
(3.09)*
1.52
1.17
0.45
0.49
0.29
0.42
0.82
1.09
1.39
0.59
0.48
0.26
1.04
0.35
2.74
2.60
3.10
3.55
4.04
49
(25)*
62
71
89
88
Processed Coal Composition
L-R Processed Coal { Jj** J
Reanalyzed Sample (Batch A + B)
30.60
26.22
25.93
10,179
10,993
11,090
2.00
1.46
1.65
0.15
0.08
0.67
0.32
0.31
0.18
1.53
1.07
0.79
3.64
3.17
1.33
96
98
83
45?
61
75
74
90
691 7fi
82 176
74
tv>
* Coal S content and % S removal based on FE2 (ferrous ion) production during t .
-------
analyses performed at CT&E (Commercial Testing & Engineering Laboratories).
Thus,
% S Removal = % Sp (starting coal) - % S (sample) x 10Q
% S (starting coal)
(1)
This method of computation assumes that the weight of the coal remains
constant during the reaction. This, of course, is not the case because of
ash (pyrite) dissolution, but the error introduced in the above computation
is insignificant (less than one percent error in the calculated removal).
The pyrite removal values "Based on Corrected S " were computed the
same way except that an adjusted sample S value was used instead of that
derived from coal pyrite analysis. The adjusted or corrected sample Sp
value was computed as follows:
S (corr.) = S (Anal.) + SQ (Anal.) - 0.43 (2)'
The above correction is based on the following assumptions. First, the organic
sulfur content (wt. percent) of the coal, S , was not affected by the Meyers
Process and that its value will change only because of changes in the
weight of the coal (ash dissolution). For the coal used in this program,
this assumption fixes the coal S range during processing between 0.41 and
0.44 weight percent (0.43 was selected as the "expected S " for all samples
in order to reduce the computations involved; the introduced error is in-
significant). Second, the analyzed values for total sulfur, St> and sulfate
sulfur, Ss, are correct; and third, the elemental sulfur produced during the
process was completely recovered.
Experimentation performed to date has shown that the Meyers Process
does not affect the SQ content in the coal other than through changes in
coal weight because of ash dissolution. However, the SQ value is determined
indirectly from $t, Sp, and Sg analyses (S0 = St~Sp~Ss~V under the assump-
tion that the unrecovered product elemental sulfur in processed coal, S is
virtually zero. Thus, the accuracy of the SQ value depends on the accuracy
of St, S , and S$ analyses and the completness of Sn recovery. Equation 2,
above, assumes that meaningful errors affecting S determinations occurred
28
-------
only in S analyses. Based on available data this assumption appears to be
statistically valid for the following reasons: S. and S analyses per-
V S
formed on the same sample consistently agreed within ±0.05 or better;
abnormalities in SQ determinations were observed at both high and low
(near zero) S. values; unrecovered SM can not account for S values below
s n o
0.43 wt percent. A more direct proof of the validity of Equation 2 can be
derived from the smoothing effect its application had on S removal rate
data. It should be emphasized, however, that application of Equation 2 to
every S value may not have been warranted; it was employed only for the
sake of consistency (occasional erroneous application of Equation 2 does
not affect conclusions drawn on the performance of the Meyers Process).
The last row of information in Table 4 presents data on the composition
of the processed coal. In this particular experiment the coal was not
"settler" processed; the L-R slurry was processed in two batches through
the rest of the unit operations (filtration, washing, elemental sulfur
recovery, and drying) because of equipment size limitations. Thus, in
principle, the two-hours reactor sample and the samples from Batches A and
B should have been of the same composition (identically processed coals).
Obviously they were not. Even though a part of the discrepancy could be
attributed to sulfur forms analysis errors, poor sampling appeared to also
be responsible. The two batches of L-R processed coal were mixed, refluxed
for 30 minutes in 0.1N H2S04, dried, sampled, and analyzed (washing with
acid was performed to remove deposited iron and sulfate sulfur thought to
interfere with the accuracy of analysis). The "corrected S " value of the
combined batch agreed well with the average of the "corrected S " values
of the individual batches, but not with the two-hour reaction sample.
A coal sampling problem appears to be responsible for the discrepancy
in the two S values of the starting L-R coal (tL_R = 0.0 sample). The top
value (2.07 wt. percent) was obtained from sulfur forms analysis of the
coal in the drawn slurry sample. The value in parenthesis was computed
from the measured quantity of ferrous ion produced during tm. According to
Meyers Process chemistry, repeatedly verified, 10.2 moles of Fe are pro-
duced per mole of pyrite oxidized. Fe+2 production from side reactions of
29
-------
Fe+3 with the organic matrix of the Lower Kittanning coal or with non-
pyritic ash components is negligible. Since at t Fe+ is not being
+2
regenerated dur to lack of oxygen, Fe production is an accurate measure
of pyrite removal from coal in the mixer (tm reaction time). Thus, the
S value in parenthesis and the corresponding S removal value were
assumed to be the correct values. The discrepancy between the directly
+9
analyzed and Fe computed S values have been attributed to poor reactor
sampling at t,R = 0.0. Because of inadequate slurry mixing during the
start of L-R operation, the sample drawn from the circulation loop con-
tained the smaller particle size coal representing the coal fraction with
higher pyrite depletion. Later reactor samples were substantially more
representative of the coal in the reactor.
Early in the program we experienced difficulties in obtaining repre-
sentative slurry samples (partially due to slurry foaming) and as a
consequence the decision was made not to rely completely on reactor samples
for the determination of S removal rates as a function of reaction time.
Thus, separate experiments were performed for a number of L-R times, t^_^,
ranging from one to eight hours. This decision greatly expanded the
planned experimental effort, but it also greatly improved data reliability.
Alleviation of foaming problems and improved sampling procedures resulted
in more representative samples as a scan of the data in Volume 2 and the
data in Table 5 (Experiment 7) below indicate. However, sampling,
even of processed dry coal, was not completely consistent. At least in
part, the sampling difficulties can be attributed to the unusually high
ash content of the coal utilized.
Table 5 is identical in format to Table 4 except that it includes the
composition of "settler" processed coal (last row of Table 5). Settler
processing, described in the previous section, was always conducted at 90°C.
The coal slurry residence time in the settler in hours is indicated by the
ts value in parentheses. The "AG" notation indicates that the "settler"
was stirred ("NAG" was used to denote quiescent "settler" processing).
The difference in the S removal values between "L-R Processed Coal" and
"L-R + Settler Processed Coal" represents the percent of starting coal
pyrite removed during "settler" processing.
30
-------
TABLE 5. RATE DATA ON L-R PROCESSING OF 100 MESH TOP SIZE L.K. COAL WITH 5 WT. % IRON
SOLUTION AT 120°C AND TOO PSIG
RATE DATA
cvDCDTMrwT wn •, tm = T'0 Hour' Si p = 1'° H°ur» te = 1 9 Hours (%50% of Coal); FET (Iron in Reagent) = 4.94 Wt. %;
EXPERIMENT NO. 7: smarting Y (Fe^Ffr) =0.95
Reaction Time, Hours
(t = time in mixer;
t = time under L-R
processing;
ts = time in settler)
Starting Coal (5 Sample
Average) and Slurry
Sample
Slurry
Composition
Wt. %
Solids
19.2
Y
(Fe+3/Fe)
.95
Coal Composition Wt. % (Except Heat Content), Dry
Ash
30.85
+.46
Heat
Content
Btu/Lb
10,275
+ 60
Total
Sulfur,
St -
4.53
+.06
Pyritic
Sulfur,
SD
3.88
±.09
Sulfate
Sulfur,
ss
0.24
+.04
Organic
Sulfur,
S0
0.41
+.07
Iron,
Fe
3.98
+.06
% S Removal
Based on
Analyzed
SP
Based on
Corrected
SD
Reactor Samples
t, _R = 0.0 Hours
(t =1.0 Hour)
m
tL_R = 0.5 Hours
» 1.0 Hour
20.2
20.4
19.1
.54
.72
.88
>8.97
>7.14
>5.15
10,558
10,823
11,147
3.68
2.77
1.75
3.12
(3.01)*
2.22
1.04
0.19
0.25
0.29
0.37
0.30
0.42
3.14
2.41
1.57
20
(24)*
43
73
21
46
73
Processed Coal Composition
L-R Processed Coal
L-R + "Settler" (AG) Proc. Coal (tg = 19)
28.55
25.64
10,831
11,072
1.98
1.21
1.43
0.49
0.23
0.27
0.32
0.45
1.88
1.25
63
87
66
87
CO
Coal S content and % Sn removal based on FE2 (ferrous ion) production during t .
p p m
-------
Table 6 presents "Process Mass Balance" data from Experiment 1.
The solids, liquids, sulfur, and iron overall process mass balances are
summarized in separate sections of the table. The table is a computer
print out summary of individual process unit operations. A mass balance
computer program was written which utilized detailed information on the
composition of feed and output streams (weights and analysis data) from
each process unit operation of each experiment, performed the mass balance
calculations on solids, liquids, and slurries around each unit operation
(mixer, L-R reactor, settler, wash, elemental sulfur recovery Teacher, and
coal drier), and compiled the data on overall process mass balance shown
in Table 6.
The "solids balance" in Table 6 refers to coal. The "input" and "dry
process coals" (completely processed coal plus samples) are direct
weight measurements (the moisture content of the feed coal was estimated
from short-proximate analyses performed on the "as received" raw coal; it
averaged very nearly one weight percent). The processed coal and samples
were corrected for deposited reagent. The quantity of deposited reagent
was estimated from the iron and sulfate sulfur analyses of the starting
and processed coals (including samples) and from the extent of pyrite re-
moval (S analyses). The excess iron on the processed coal (reagent
derived iron) was assumed to be FeSO. and Fe^OU. For the majority of
experiments this correction was small (little, if any, reagent deposition
that withstood washing). The dissolved coal mineral matter was determined
from starting and processed coal ash analyses converted to mineral matter
(assuming the ash to be Fe203 and the mineral matter FeSp); it is shown
in Table 6 as "pyrite dissolved" and as "non-pyritic ash dissolved". The
first quantity was determined from S coal analyses and the second by
difference. Table 6 indicates that 3 grams of "non-pyritic ash" was de-
posited on coal during processing. This is probably due to normal ash
analysis uncertainty (deviations). A scan of the data from all experiments
indicates that a very small quantity of non-pyritic mineral matter was
dissolved during processing. The "coal dissolved in solvent" (elemental
sulfur recovery unit) was determined from the weight and sulfur content of
the residue derived from the distillation of the spent organic solvent
(toluene or hexane, toluene in the large majority of experiments).
32
-------
TABLE 6. PROCESS MASS BALANCE DATA
EXR. 1, L.K.100XOMESH, 20PCT SLURRY(NOMINAL) . 120C, 100PS1G(85PS1 02)
OVERALL SOLIDS BALAKCI OVERALL COAL SULFUR BALANCE
DRY COAL INPUT 1882
HRY FRnr.ESS COALSCCORR. FOR OEF. REAG.l 1773
FYKI It LliSSOLVEO DURING REACTION 111
6SHUON-PYRITIC) DISSOLVED DURING REACTION -15
COAL DISSOLVED IN SOLVENT 3
TOTAL SOLIUb RECOVERED 1872
LOSS (INPUT - RECOVERED) 10
PERCENT RECOVERY OF SCLIOS 99
INPUT COAL SULFUR 85
PROCESS COAL SULFURCCORR. FOR OEP. REAG.I 24
SULFUR IN SAMPLES 6
SULFUR OISSOLVEtHAS SO'*) IN LEACH R-1AGENT 36
SULFUR RECOVERED WITH ORGANIC SOLVF-MT 18
TOTAL SULFUR RECOVERED 83
LOSS (INPUT - RECOVERED) 2
PERCENT RECOVERY OF SULFUR 98
co
CO
OVERALL LIQUIDS BALANCE
REAGENT
COAL MOISTURE
RINSE AND FILTER PQPER 4ATE*
DISSOLVED ASH AND PYRITE
HASH MATER
SOLVENT
TOTAL LIQUIDS IN
FILTPATE REAGENT
SAMPLE AND FOAM REAGENT
DEPOSITED REAGENT
hASH FILTRATES
CRGAMC FILTRATES
AZEOTROPE AND ORGANIC FILTRATE
LIQUIDS IN DRYER TRAPS
LIQUIDS ON EQUIPMENT
TOTAL LIQUID RECOVERED
LOSS (IN - RECOVERED)
l-ERCfcMT RECOVERY OF LIQUIDS
7747
19
H79
96
22844
8848
H0033
AT E R
992
64
22048
8337
274
660
53
38382
1651
96
OVERALL IRON BALANCE
STARTING COAL
STARTING REAGENT
TOTAL IRON INPUT
PROCESS COAL IRON
SAMPLE COAL IRON
REAGENT FILTRATE IRCN
SAMPLE FILTRATE IRON
FOAM MATERIALS IRON
REACTOR WASH IRON
WASH FILTRATE IRON
TOTAL IRON RECOVERED
LOSS (INPUT - RECOVERED)
PERCENT RECOVERY OF
75
231
306
55
7
152
2-9
0
0
36
279
27
91
-------
The "liquids balance" entries do not require explanation; they repre-
sent direct or by difference weight measurements. The entry "sample and
foam reagent" could be an exception. In the majority of experiments, the
weight of liquid shown for this entry represents the quantity of reagent
drawn during reactor slurry sampling. In early experimentation, however,
we experienced some slurry foaming during the early stages of L-R processing;
as a result a small quantity of foam left the reactor through the oxygen
exhaust valve. This foam was trapped, weighed, filtered, and analyzed.
Its components (principally reagent) were then entered in the appropriate
mass balances. It should also be noted that the "dissolved pyrite and ash"
entry in the "liquids balance" does not include the produced elemental
sulfur; the latter is part of the "organic filtrate" entry.
In the "overall coal sulfur balance" the "input", "process coal",
and "sample" sulfur values were determined from coal weights and Eschka
analyses. The processed coal and samples sulfur values were corrected for
deposited reagent sulfate. The "sulfur recovered with organic solvent" was
determined from sulfur analysis of the organic solvent residue. The "sul-
fur dissolved as sulfate" was computed from the measured pyritic sulfur
removal (coal S analyses) by assuming that 60 percent of the removed S
was dissolved as sulfate. This assumption was based on extensive experi-
mentation performed on Lower Kittanning coal in the previous bench-scale
program which indicated that pyrite removed from coal by the Meyers
Process was converted and recovered as sulfate sulfur and elemental sulfur
at a ratio estimated to be 1.5 (the ratio was estimated from coal S anal-
yses and ferrous ion production). Within experimental uncertainty, the
same ratio appears valid for this program's coal, even when processed under
L-R conditions at as high a temperature as 130°C. In the majority of
experiments, the recovered elemental sulfur equalled 32-35 percent of the
S removed from the coal instead of the expected 40 percent; this computes
to Ss/$n ratios in the range of 1.8 to 2.1 (Experiment 1 represents an
exception in that the recovered $n is less than 30 percent of the estimated
removed S ). However, complete elemental sulfur recovery is difficult and
the determination of recovered elemental sulfur is subject to error. During
the previous program, elemental sulfur recovery based on an S /S ratio of
*> n
1.5 range between 80-90 percent (32-36 percent of the removed pyritic sulfur),
34
-------
The same appears to be true with the present coal based on the quantity
of elemental sulfur recovered and the S removals as computed from coal
analyses before and after processing. (In the previous program the S /S
ratios were also verified from Fe+2 production; because of Fe+2 oxidation
during L-R operation this verification technique could not be used in the
majority of experiments performed in the current program.)
It should be noted that 80-90 percent elemental sulfur recovery does
not necessarily imply that this sulfur product was incompletely removed
from coal. Incomplete recovery from toluene (the extraction medium),
small errors in analysis or weighing of the sulfur residue, and small
errors in the computation of S removal can easily account for 10-20 percent
elemental sulfur deficiency.
;
The "iron balance" entries were computed from weight measurements
and iron analyses. The origin of each entry is evident from its label.
2.4 DATA DISCUSSION AND INTERPRETATION
The data in Appendices A through C (Volume 2) indicate that in the
majority of the L-R experiments performed the overall process mass balance
of solids, liquids, sulfur, and iron were better than +.5 percent; in-
dividual unit operation mass balances were even better. In view of the
fact that the overall process mass balances of each experiment involved
over one hundred weight measurements and analysis determinations, that
each experiment involved several manual slurry transfers, and that the
system was relatively small (e.g., the sulfur balance involves less than
100 grams of sulfur in most cases), the mass balances attained should be
considered excellent. In the rare cases where larger than 5 percent dis-
crepancies occurred in one or more of the "balanced" quantities, the
discrepancy was probably due to untraceable errors in weights and analyses
rather than losses during processing. The biggest losses were in liquids
and they represent water evaporation losses during coal washing. In
general, the mass balance results indicate that the data can be used with
confidence in drawing conclusions on the effectiveness of L-R processing
35
-------
on the leaching of pyrite from Lower Kittanning coal. Individual pieces
of data or data involving a small portion of the processed slurry could,
of course, be in substantial error and yet not influence the described
mass balances. In fact, this was the case with a number of reactor samples,
as indicated in the previous section.
The raw rate data in the same Appendices of this report (Volume 2) do
not exhibit the consistency of the mass balance data. However, these
apparent inconsistencies are explainable. It will be shown later that,
despite some scatter, the data are internally consistent and relatable to
that generated under ambient pressure processing in the previous bench-
scale program (separate Teaching-regeneration operations).
The scatter in rate data was traced to three causes: sampling (es-
pecially reactor slurry sampling), sulfur forms analyses, and oxygen
deficiency- The problems experienced in sampling and sulfur forms analyses
were alluded to earlier. In general, slurry sampling improved greatly
when "dead spots" were eliminated from the reactor sampling loop and when
slurry foaming was reduced substantially by "wetting" the coal in the
mixer section. The only consistently poor sample was the one taken at the
start of L-R operation, most likely because coal particle size distribution
homogeneity had not yet been accomplished in the reactor (this sample was
taken almost immediately after the start of slurry circulation). However,
the S content of coal at the start of L-R was determined from the quantity
of ferrous ion produced during tm and a poor coal sample at the start of
L-R did not present problems in data interpretation. The inconsistencies
in sulfur forms analyses were substantially reduced through the application
of the organic sulfur correction to S discussed in the previous section
(Equation (2), page 28). Oxygen deficiency during L-R processing was
probably the single most important reason for the observed rate data
scatter and it was certainly the one responsible for the larger deviations
in the L-R rate data. (Oxygen was used during L-R processing to regenerate
+3
the Fe being depleted by pyrite oxidation. Oxygen concentration in-
+2 +3
fluences the Fe to Fe conversion rate; the iron forms the pyrite
oxidation rate as explained below.)
36
-------
According to data derived from ambient pressure processing of Lower
Kittanning by the Meyers Process each mole of pyrite leached from coal
consumes (reduces) 9.2 moles of Fe"1"3 and produces 10.2 moles of Fe+2. The
quantity of elemental sulfur produced during L-R processing appears to
indicate that the same stoichiometry applies to this mode of processing.
Ambient pressure data also indicated that the Sn leaching rate depends on
+3 P
the square of the Fe -to-Fe ratio, Y. If the same rate dependence on Y
exists under L-R processing, a small change in the Y value should translate
into substantial effect on the S removal rate. This appears to be the
case upon analysis of the "rate data" Tables contained in Appendices A and
B, Volume 2. Since oxygen is used to regenerate Fe (one mole of oxygen
regenerates four moles of Fe ), even a small deficiency in oxygen during
L-R processing can have a pronounced effect on S leaching rate through
its influence on Y.
The above discussion is illustrated by the data depicted in Figure 3.
This figure summarizes the data from 14 experiments performed on 14 mesh
top size coal under the conditions indicated (data from Tables B-l, B-2,
B-3 and B-4, Appendix B, (Volume 2). The remaining data on this size coal
(11C°C and 130°C, Tables B-5 and B-6) were not included in order to avoid
additional clutter of the plotted data; however, it can be easily estab-
lished through inspection of Tables 5 and 6 that the majority of these data
fall within the band traced by the plotted data. In order to minimize data
scatter from errors in sulfur forms analyses the "corrected" S removals
during t were estimated from ferrous ion production; all other data was
derived from direct coal analyses of reactor samples and product coal
samples. The data designated by bars (samples from 4 experiments) and that
labeled as "outlier samples" were derived from experiments processed under
identical conditions except for small differences in the starting reagent
Y. The plotted data indicate that during L-R processing the only discernible
parametric effect is that of slurry concentration during the first hour of
processing which could be reasoned to be just an apparent effect due to the
higher $„ removal during t . In fact, the "outlier samples" deviate more
p 3 m
from the mean of the plotted data than any sample representing parameter!c
variations. These outliers represent samples from Experiments 30 and 31
(Table B-2). The Y values of these samples differ only by approximately
37
-------
90
80
o
u
1
Q
UJ
I
ex:
u.
U
40
5 30
LU
20
10
o
A
T
T 33 WT. % COAL SLURRY, 120°C. 100 PSIG
1 (SEVERAL EXPERIMENTS WITH THE D BEING
OUTLIER SAMPLES)
• 20 WT. % COAL SLURRY, 120°C, 100 PSIG
A 33 WT. % COAL SLURRY, 120°C, 50 PSIG (35 PSI 02)
*33 WT. % COAL SLURRY, 120°C, 150 PSIG (135 PSI 02)
m
1.0 2.0 3.0
L-R REACTION TIME, HOURS
4.0
5.
8.0
Figure 3.
Pyritic Sulfur Removal from 14 Mesh Top Size L.K. Coal
Processed at 120°C with 5 Wt. % Fe Reagent (Data Summary)
38
-------
0.10 from those of corresponding data from Experiment 28 (or from the
average Y of all the experiments represented by the bars in Figure 3);
however, the indicated S removals differ by nearly 100 percent at t,
0.5 hours (22 percent versus 42 percent for the average of the "bar" data)
and nearly.50 percent at tL_R =1.0 hour (39 or 40 percent versus 58 percent
for the averaged data). This example does not only illustrate the dif-
ficulty of sensing oxygen deficiency during L-R processing, but it also
underscores the importance of Y to pyrite leaching rates and it appears
to indicate that ferric ion is the oxidizing agent in pyrite removal
during L-R processing just as it was determined to be during ambient
pressure processing (separate Teaching-regeneration). It can be concluded,
therefore, that under efficient regeneration the S removal rates during
L-R should be at least as high as those indicated by the upper bounds of
the plotted data. This translates to 80-85 percent pyritic sulfur removal
from 14 mesh x 0 Lower Kittanning coal after approximately two hours of
L-R processing plus t . Since results from the previous bench-scale
program indicated that the same extent of S removal required 10-12 hours
of ambient pressure processing at 102°C the L-R S removal rate at the
110°-130°C range must be substantially higher than that obtained at 102°C
with separate regeneration. This conclusion is fully validated in Section
2.4.2.
Oxygen deficiency during part of the experimentation could have
been the result of insufficient feed rate or inadequate mixing due to
deficient apparatus design or inappropriate procedures; or it could have
been unavoidable because the demand of oxygen by the system exceeded theo-
retical limitations (e.g., maximum oxygen solubility in the slurry) or
practical means of oxygen incorporation into the slurry. The scatter in
the data of nearly identically processed coal tends to indicate that the
oxygen deficiency was due to inadequate experimentation rather than being
due to unavoidable causes. Thus, it is our opinion that conclusions con-
cerning the efficiency of L-R processing should be based on the assumption
of efficient regeneration (rate of Fe+3 regeneration as predicted by the
kinetics of the reaction).
39
-------
The scatter in the data, the apparent insensltivity to parametric
changes, and the nearly abrupt reduction in S removal rate after two
hours of L-R processing (Figure 3) hindered initial efforts to apply
previously derived rate expressions on leaching and on regeneration to
L-R processing or to develop new ones which adequately described the pro-
cess. It was not until virtually the complete set of data was generated
that the oxygen deficiency probability became apparent and that the flat-
tening of the S removal rate became statistically valid. Once these
unexpected occurrences were recognized, it became evident that a single
rate expression (or at least a single rate constant expression) can not
describe the entire leaching operation of this particular coal and that Y
values predicted from the simultaneous solution of previously derived Sn
+3 "
leaching rate and -Fe regeneration rate expressions would not predict the
generated data. The coal leaching process, therefore, was divided into
four parts (processing ranges) and the data derived from each one were
examined separately. The four processing ranges are: (a) mixer processing,
(b) L-R processing up to 80 percent S removal, (c) L-R processing between
80-90 percent S removal (2-8 hours L-R processing), and (d) "settler" pro-
cessing. Data discussion and process definition in the outlined four
processing ranges are presented in the ensuing report sections in which
process behavior in each major process unit operation is described and
discussed separately.
The above discussion and conclusions apply to the L-R processing of
100 mesh top size coal as well (data in Appendix A, Volume 2).
2.4.1 Mixer Unit Operation
Initially, the "Mixer Unit Operation" was defined as that part of
processing where hot, approximately 70°C, reagent was mixed with "as
received" coal and the resulting slurry, pressurized by nitrogen gas to
the desired L-R pressure, was heated to the desired L-R temperature. How-
ever, substantial slurry foaming during L-R processing, especially in the
early stages of L-R processing, required that the mixer section be modified
so as to furnish a substantially non-foaming slurry to the L-R section.
40
-------
An extensive small scale investigation (200-500 grams of coal) was
undertaken in order to identify the causes of coal slurry foaming, to
establish practical foam control methods, and to define the proper "mixer
section" for L-R processing of suspendable coal. The following parameters
were examined for their effect on coal slurry foaming and its control:
slurry pH, iron salt concentration and Y of reagent, slurry solids content,
coal particle size, coal moisture, slurry temperature, retention time
near slurry boiling temperatures, anti-foam agents, wetting agents, coal
washing (soaking in hot water), and slurry reflux. In addition, the
foaming characteristics of processed and raw coal slurries were compared
and the gas evolved (principally C02) from raw coal treated with acids
was determined. The quantity of foam (height in a reactor column) gener-
ated during boiling of the coal slurry after treatment, or after additives
had been incorporated, and its consistency (solids content and persistence)
served as criteria for measuring the various effects.
Slurry pH (in the range of 1 to 7), the Y of reagent, slurry temper-
ature up to boiling and "soaking" time up to 2 hours at any temperature in
this range, anti-foam agents, and coal washing did not affect slurry
foaming. Slurry foaming decreased with decreasing iron concentration, but
it remained at unacceptable levels even when iron was reduced to zero
(dilute sulfuric acid or water coal slurries). Foam volume, density, and
stability increased with increasing solids content in the slurry; however,
even very dilute slurries foamed. The maximum CO^ evolution measured
(approximately 150cc of gas per kilogram of coal) was not sufficient to
account for the volume of the generated foam. Slurry foaming increased
with decreasing coal top size in the range of 14 to 200 mesh, but it re-
mained at unacceptable levels even with the 14 mesh x 0 coal; removal of
the 200 x 0 fraction from the 14 mesh top size coal reduced slurry foaming
substantially and rendered the 14 x 200 mesh coal processable. Wetting
agents changed the nature (consistency) of the foam radically, but not its
volume; the normally thick, black slurry foam was converted to thin (soap
like), clean foam. Combinations of wetting and anti-foam agents arrested
slurry foaming completely. Slurry reflux at ambient pressure for approxi-
mately 30 minutes virtually eliminated slurry foaming. Finally, the
41
-------
moisture content of coal appeared to have an effect on the volume and
density of coal slurry foaming with slurry foaming decreasing with in-
creasing coal moisture content, i.e., the coal was in essence "pre-wetted".
The above investigations led to the conclusion that the principal
cause of slurry foaming is the presence of dry (unwetted) coal in the
slurry and that Lower Kittanning coal which equilibrates in air to only
about one percent moisture is difficult to wet. Three approaches
rendered Lower Kittanning coal slurries processable through L-R: (a)
coal wetting through the addition of wetting and anti-foam agents, (b)
slurry reflux for approximately 30 minutes and (c) removal of coal fines
(200 mesh x 0 fraction) prior to processing. All three methods were tested
in bench-scale L-R processing runs and all proved successful. However,
removal of the fines fraction (a substantial portion of suspendable coal)
was the least desirable method because a different technique had to be
identified for processing the fines by the Meyers Process and because
this approach was not as effective in foam control as coal wetting. The
wetting, anti-foam agent approach was rejected because of the cost of
these agents and because it appeared to adversely affect L-R processing
(refer to data in Tables A-7 and B-7, Volume 2); this method would have
been used if foam arrest by slurry reflux did not prove successful. The
slurry reflux method was judged to be by far the most desirable because
of its simplicity and easy adaption to L-R processing. This approach-
was therefore, selected for incorporation into the "mixer unit operation"
as standard practice.
Mixer processing, "mixer unit operation", in all suspendable coal
experiments involving simultaneous leaching-regeneration (except for the
two experiments in which wetting agents were used) consisted of the fol-
lowing steps: slurry preparation and heating to reflux temperature (102°C)
slurry defoaming by 30 minutes reflux, slurry transfer to the L-R reactor
and pressurization, and slurry heating to the desired L-R temperature.
Normal slurry mixer processing times, tm, ranged between 1.0 and 1.5
hours; occasionally, longer times were used either because foaming persisted
or because of equipment problems. The approximately 10 kilograms of slurry
42
-------
used in each experiment was prepared with 65° to 70°C reagent, split into
3-4 four-liter glass flasks equipped with condensers, and transferred
to the 13-liter stainless steel L-R reactor when defoaming was completed.
The temperature-pressure- time history of mixer processing was recorded
in detail during each experiment. Figure 4 depicts typical mixer
processing conditions.
Pyritic sulfur removal from coal during tm varied from near zero (less
than 5%) in experiments where no-iron containing starting reagent was used
to nearly 30 percent when coal was processed with 5 wt. percent iron re-
agent as 20 weight percent slurry for longer than normal t . Shorter t , 3
weight percent Fe reagent, and 33 percent coal slurries gave intermediate S
removals (see data in Appendices A, B, and C, Volume 2). This type of
parametric behavior was expected from previous experience with ambient
pressure processing of Lower Kittanning coals. The next step was to investi-
gate the predictability of the t data generated in this program by the
pyritic sulfur leaching rate expression developed in the previous bench-
scale program (EPA contract EHSD 71-7) from ambient pressure data generated
at 70°, 85°, and 102°C on a similar Lower Kittanning coal, namely,
„ _ dWp _ (/• ., 2 V2
rL ~ " dt L Wp Y '
where
r. is the pyrite leaching rate, expressed in weight of pyrite
removed per 100 weights of coal per hour (rate of coal
pyrite cone, reduction),
W is the pyrite concentration in coal at time t in wt. percent,
P
t is the reaction (leaching) time in hours,
Y is the ferric ion-to-total iron ratio in the Teacher at
time t, dimensionless, and
K, is the pyrite leaching rate constant (a function of tem-
L perature and coal particle size) expressed in (hours)-'
(wt. percent pyrite in coal)"'
43
-------
150*H
CD
i—i
CO
o.
a:
a.
o
DC
UJ
X
100*-
50*-
AIR
OXYGEN
0
e
«
UJ
OL
ra
i
UJ
Q_
UJ
1—
0
I-H
1 —
<
UJ
oc
:
LU
X
t— i
S
130*
120*
1 fm\J
110*
100
80-
60-
40-
20-
0-
i
SLURRY HEATING L'R PROCBSIN6 AT 12°°C
/"X^^ /
SLURRY REFLUX >V /
/
\ /
/ \ /
/ \ /
/ v
T
SLURRY TRANSFER TO L-R REACTOR
in ?'n Vn dh (in f»n ?'n
"MIXER" REACTION TIME, MINUTES
Figure 4. Typical Mixer Processing Conditions
Utilized L-R Temperatures and Pressures
44
-------
with
KL = AL exp (-EL/RT),
where
AL is the Arrhenius frequency factor in the units of K ,
EL is the apparent activation energy in calories/mole,
R is the gas constant in calories/mole°K, and
T is the absolute temperature, in °K
From data generated in the previous bench-scale program, the K, value for
pyrite leaching from 100 mesh x 0 UK. coal at 102°C was determined to be
between 0.12 and 0.15 (hours)"1 (wt. percent pyrite)"1. The corresponding K,
values for 85°C and 70°C leaching were 0.09 and 0.03, respectively. The
Arrhenius plot of these three rate constants did not furnish the expected
straight line; thus, the A. and E. values derived from these constants and
reported in Reference 1 were based on the 85°C and 102°C data. Recent
review of the 85°C and 70°C data revealed that the determined rate constant
for the 85°C run was 0.088 + 0.044 while that of the 70°C data was 0.03
+_ 0.013 (single experiment data for each temperature with 50 rate data
points taken in each experiment). It was, therefore, concluded that the
70°C rather than the 85°C K, value should have been used for the AL and EL
determinations. It was' also noted that a KL value of 0.07 for the 85°C
run gives a straight line Arrhenius plot for all three temperatures with
KL (102°C) =0.15. The new data assessment led to two sets of AL and EL
values depending on which end of the 102°C KL range was used with the 70°C
value:
For 102°C KL = 0.15
A = 4.7 x io6 (hours)"1 (wt. percent pyrite in coal)" and
EL = 12.8 Kcal/mole
For 102°C KL = 0.12
AL = 3.4 x TO5 (hours)"1 (wt. percent pyrite in coal)
EL = 11.1 Kcal/mole
45
-------
Similar but less extensive rate data generated with 14 mesh x 0 Lower
Kittanning coal revealed that the KL value for this size coal was approxi-
mately 20 percent lower than the corresponding KL for 100 mesh x 0 coal;
thus, at 102°C
0.10 < K14 £0.12 (hours)"1 (wt. percent pyrite in coal)" ;
the corresponding A^ is
2.7 x 105 <. Aj4 <. 3.8 x 106 (same units)
assuming that the activation energy is the same for both top size coals,
namely
11.1 1 EL <. 12.8 Kcal/mole
Prior to utilizing the above rate expressions to predict "mixer" pro-
cessing, it was considered necessary that at least one ambient pressure
leaching experiment be performed under continuous reagent exchange conditions
with the L.K. coal used in the current program in order to compare identi-
cally generated data from the two coals (present and previous program L.K.
coals). The experiment was performed at 102°C (slurry reflux temperature)
with 100 mesh x 0 coal slurried in ferric sulfate solution containing 4.9
weight percent iron. The total reaction time was 8.5 hours (8.0 at 102°C plus
34 minutes slurry heating time between 70°C and 102°C). The generated rate
data and overall process mass balance data are summarized in Table 7.
The first part of Table 7 (first page) shows the measured Fe+2/Fe
ratio (1-Y) and the computed change in coal weight, pyrite removed, average
rate of removal (from reaction start to the indicated reaction time), and
coal composition as a function of reaction time. The pyrite removal was
computed from measured ferrous ion production as a function of time as-
suming that pyritic sulfur was oxidized to sulfate sulfur and elemental
sulfur at the ratio of 1.5. All other computed quantities were derived
from the calculated pyrite removal plus a small correction for any non-
pyritic ash dissolved or added to the coal during reaction. (The
46
-------
TABLE 7. PYRITE REMOVAL FROM COAL WITH IRON SULFATE SOLUTION AT 102°C AND AMBIENT PRESSURE
SUHMARY OF REACTION DATA
SC«t/*:= l.SOO
LOH£R KITTANNTNG - 100 *ESH
MM
FPQM
START
34
49
64
79
94
109
124
139
154
169
134
199
214
229
244
259
274
289
304
319
334
349
364
379
394
409
4?4
439
454
469
484
499
514
FROM
FF+P/FE
WT. RATIO
.3333
.355$
.3626
.3294
.2791
.2682
.2752
.2P65
.2P.88
.2813
.2662
.247?
.2373
.2102
.2C85
.2C50
.le.88
.1839
.1868
.1825
.1705
.1544
.1429
.1329
.1250
.1265
.1245
.1?19
.1154
.1123
.1103
.0946
.0771
FINAL CCAL
COAL VT.
F9F." CP
SCELEf)* P
641 . ?7
634.05
62?. 44
622. 65
618.37
612.95
607.14
601.37
595.93
590. 61
585.37
580.3'=
575. 03
570.43
565.79
560.92
555.94
551.23
546.18
541.32
536.64
532.18
527.70
?23.26
518.60
513.81
509.14
504.60
500.08
495.53
490.95
486.83
482. 88
ANALYSTS
PYSITF HLMULflTIV
•?FI*OVrD
!CT OF TNIT
18.78
23. S5
?7.76
3?. 06
31. 90
35.03
39.43
43.63
'7, 01
tc. fir)
52 . 34
54.05
57.19
58.87
60.69
63.17
65.29
66.45
6C.73
70.41
71.54
72.01
7?. 66
73.76
75.11
76.90
78.3ft
79.1.5
P0.23
P1.15
62.13
PI. 93
81.24
P3.64
RATE
PCT/HR
33.13
29.33
26.02
24.35
20.36
19.28
19.08
18.83
18.3?
17.68
17.07
16.30
16.03
15.42
14.92
14.63
14.30
13.80
13.57
13.24
12.85
12.38
11.98
11.68
11.44
11.28
11.09
10.86
10.60
10.38
10.18
9.85
9.48
CALCULATED
TOTAL S
HT-PCT
3.791
3.591
3.442
7.274
3.28H
'.157
2.984
2.817
?.683
2.571
2.469
2.401
2.275
2.207
2.133
2.032
1.946
1.899
1.806
1.738
1.691
1.672
1.645
1.600
1.545
1.471
1.410
1.366
1.334
1.296
1.255
1.263
1.29?
1.670
PYR. S
HT-PCT
3.306
3.104
2.955
2.785
2.792
2.668
2.493
2.326
2.190
2.078
1.976
1.907
1.780
1.711
1.637
1.536
1.449
1.402
1.308
1.239
1.192
1.171
1.146
1.100
1.045
.971
.909
.865
.832
.794
.753
.761
.790
.690
CCAL ANALYSES STEP-BY-STEP
ORG. S
WT-PCT
.414
.415
.416
.417
.417
.418
.419
.420
.421
.421
.422
.422
.423
.423
.424
.424
.425
.425
.426
.426
.426
.426
.426
.427
.427
.427
.428
.428
.428
,428
.429
.429
.429
.750
S04 S
WT-PCT
.081
.081
.081
.081
.081
.082
..082
.082
.082
.082
.082
.082
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.084
.084
.084
.084
.084
.084
.230
ASH
WT-PCT
29.388
29.323
29.276
29.221
29.223
29.184
29.128
29.074
29.031
28.995
28.962
28.940
28.899
28.877
28.853
28.821
28.793
28.778
28.748
28.726
28.711
28.705
P8.696
28.682
28.664
28.640
28.620
28.606
?8.596
28.583
28.570
28.573
28.582
28.550
HT CONTENT
BTU/L8
10637
10655
10668
10683
10682
10693
10709
10723
10735
10745
10754
10760
10772
10778
10784
10793
10801
10805
10813
10819
10823
10825
10827
10831
10836
10843
10848
10852
10855
10859
10862
10861
10659
11013
Coal weight in grams
-------
TABLE 7. (CONTINUED)
Co
ICW^p KITTAKNTNG - 10Q
OVERALL PeOCrSS lALANCfS Ff>« "OLIO*;, LTIUTO1^, SULFUP flUO I»ON*
ANT
HRV CC«L TNPLT
IN
DISSOLVED HURUG
CC«L IN SOLVFNT
FRCV FILTER
CVCN DRIE[) COAL
TOTAL ^OLTPS ^ECOVFPir
EBLANC" (RFCCV^EO-INFUT)
"?FCOVCKY OF
T«CTTCN
7.?
3.n
(»7n. 0
? RALANCF
LEACH ";EAGPNT IMPUT
ASH AKP PYRITE OISSOLVFO DURING
TOTAL HATFR (WASHES AKC WFTTUG FILTFR5)
TOTAL CRGAMC SOLVENT TN°UT
TOTAL LIQUTOS IN
LFACH CFAGENT R^COVE^FC
HOUir RFMCVEC IN SAPPLF^ «ND FILTFPS
TOTAL WASH FILTRATES
SOLVENT DISTILLATE
LIQUirS IN DRIER TRAPS
FOUND CN ECUIFfENT ANO FINAL FILTFR PAPERS
TOTAL LIQUIDS RECOVERFT
PALANC1! (RECCVFRED-INFUT)
CCMPUTEO EVAPORATION LOSSES FffCM FILTERS
ESTIM*TEH OVERALL BAL*NHr
RF/COVFKY OF LICUIDS
-5.1
99.?
18893.0
IS. 3
1818.0
25057.3
17723.0
680.8
,0
,0
.0
120
3
2«»57J».fi
6
0.0
COfll SULFUR BALANCr
SULFUR IN COAL INPUT 2°.5
SULFUR TN PROCESSED COAL 7.°
SULFUR IN SAMPLES TAKEN DURING REACTION 5.1
SULFUR OISSCLVFC (AS SOU» IN LF«CH OEAGENT 11.5
SULFUR RECOVERED FRCM ORGANIC SOLVENT 7.0
BALANCE (RECOVERED-INPUT! 2.2
PERCFNT RECOVERY OF SULFUR 157.<«
IRON BALANCE
IRON IN COAL 2«».7
IRON IN LFACH REAGENT INPUT 927.6
TOTAL IPON INPUT 952.<*
IRON IN L^ACH REAGENT ORAMN FRO^ REACTOR 73*».S
IRON IN SAMPLES ORAfcN DURING REACTION 37.7
IRON IN INTERMEDIATE FILTRATES REMOVED 0.0
IRON IN FINAL FILTRATE AND WASHFS i7°.9
IRON IN PROCESSED COAL 7.3
TOT4L IRON RECOVERED 959.9
BALANCE (RECOVERED-INPUT* 7.5
PERCENT RECOVERY OF IRON 100.8
98.1
Weights are In grams
-------
correction is made by linearly adding to or subtracting from the coal being
processed any ash in excess or in deficiency from that calculated by sub-
tracting the pyritic ash removed from the ash in the starting coal; the
quantity to be added or subtracted is determined from ash analyses of the
starting and processed coals.) The last line in the table shows the
processed coal composition data obtained from direct coal analyses and
the percent pyrite removal based on pyrite analyses of starting and pro-
cessed coal.
The data in Table 7 show that the agreement in pyrite removal values
determined from ferrous ion production and from direct coal analyses was
very good. The discrepancy in the corresponding total sulfur content
values is believed to be due to reagent iron sulfate left on the coal be-
cause of insufficient washing. The higher than expected ash content of
the processed coal and the 2.2 grams of excess recovered sulfur (see "sul-
fur balance" on second page of Table 7) supports the above assumption and
indicates that the abnormally high organic sulfur value of the analyzed
processed coal must be due at least in part to unwashed sulfate. The
obtained process mass balance was, also, very good, especially if our
assumption concerning sulfate deposition is valid. Based on recovered
elemental sulfur, the Ss/$n = 1.7; or, conversely, if the true Ss/Sn is
1.5, 90 percent of the product elemental sulfur was recovered. Thus, in
every aspect the data was similar to that generated with the lower ash
Lower Kittanning coal used in the previous bench-scale program. In fact,
the data in Table 7 is nearly identical point-by-point with that generated
from Experiment 50 of the previous program.
The rate data from Table 7 were utilized in Equation (3) to determine
the value of K, at 102°C for this coal. The computed K. was equal to 0.12
li
(hours) (W_) which corresponds with the lower end of the KL range
determined in the previous program. It was, therefore, decided to use the
following A. and E, values in Expression (4) to predict coal processing in
the mixer:
49
-------
100 mesh x 0 coal: AL= 3.4 x 105 (hours)"1 (Wp)
14 mesh x 0 coal : AL= 2.7 x 105 (hours)"1 (Wp)-1
Both top sizes : EL= 11.1 kilocalories per mole
Table 8 summarizes experimentally derived and predicted data for
"mixer" processing of 100 mesh and 14 mesh top size Lower Kittanning coal.
The data from 15 experiments of each top size coal are presented. The
agreement between calculated Y, W , and extent of pyrite removal values
at t and those predicted by Expressions 3 and 4 is very good; the rare
exceptions are apparently due to errors in iron-forms analyses, it should
be noted that the experiments in Table 8 were performed with a number of
different starting slurry and reagent compositions (see data in Appendices
A and B, Volume 2) and for differing reaction times, tm, and temperature-
time history. The data in Table 8 justify the use of Expressions (3) and
(4) and the quoted.values of AL and EL for the design of the "mixer" for
processing Lower Kittanning coals up to 14 mesh top size.
A complete set of mixer processing data for all experiments in Ap-
pendices A and B is included in Appendix D, Volume 2 together with
predicted data on L-R processing.
2.4.2 L-R Unit Operation
The L-R unit operation combines the two major Meyers Process opera-
tions, coal pyrite leaching and reagent regeneration, into a single unit.
Its experimental investigation constituted the major thrust of this pro-
gram.
L-R processing (simultaneous coal Teaching-reagent regeneration)
starts when oxygen is introduced to the slurry at the conclusion of mixer
processing. During this program, L-R processing was performed at constant
temperature and pressure; however, three values of each of these parameters
50
-------
TABLE 8. MEASURED AND PREDICTED PYRITE REMOVAL AT THE "MIXER UNIT
OPERATION"
Exp.
No.
Starting
' Y
wp
Measured and Predicted Values at t = tm
1
Meas.
f
Pred.*
wp
Meas.**
Pred.*
% S Removal
Meas.**
Pred . *
A. 100 Mesh Top Size Coal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1.0
1.0
1.0
1.0
1.0
0.76
0.95
0.91
0.82
0.92
0.86
1.0
0.73
0.75
0.88
7.55
7.55
7.55
7.55
7.55
7.55
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
0.28
0.34
0.40
0.29
0.29
0.35
0.54
0.51
0.46
0.37
0.40
0.45
0.33
0.34
0.23
0.27
0.38
0.31
0.27
0.29
0.28
0.58
0.51
0.45
0.42
0.42
0.47
0.32
0.34
0.22
5.79
6.08
5.76
5.91
5.97
6.58
5.62
5.89
5.94
5.07
5.35
5.16
6.48
6.40
6.31
5.65
5.95
5.69
5.68
5.62
6.09
5.65
5.62
5.70
5.19
5.37
5.08
6.42
6.38
6.09
25
21
25
23
22
14
24
20
20
32
28
31
12
13
14
27
23
26
26
27
21
24
24
23
30
28
32
12
13
17
B. 14 Mesh Top Size Coal
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
1.0
0.87
0.94
0.95
0.94
0.71
0.80
0.67
0.62
0.84
0.84
0.45
0.77
0.52
0.70
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
0.59
0.49
0.56
0.55
0.57
0.33
0.36
0.30
0.30
0.41
0.39
0.26
0.37
0.29
0.43
0.56
0.47
0.53
0.53
0.55
0.28
0.31
0.26
0.28
0.35
0.31
0.25
0.33
0.26
0.32
5.55
5.69
5.65
5.59
5.75
6.53
6.35
6.49
6.63
6.41
6.35
6.87
6.46
6.81
6.71
5.44
5.58
5.63
5.50
5.57
6.37
6.23
6.40
6.56
6.24
6.16
6.85
6.34
6.70
6.45
25
23
24
25
22
11
14
11
9
13
14
6
12
7
8
27
25
24
26
25
13
15
13
10
15
16
6
14
8
12
Predicted values from r, = K, W.!;2M2) and K. = A, exp(-E/RT)
** L u r L L
Calculated values from measured ferrous ion production
51
-------
were used. The majority of experimentation was performed at 120°C and
100 psig (85 psi 02) with 100 mesh and 14 mesh top size Lower Kittanning
coal. The major objective was the investigation of coal pyrite leaching
efficiency under L-R processing; thus, reaction time was the parameter
most thoroughly studied. The generated data has been tabulated as a
function of this parameter (Appendices A through D, Volume 2). Slurry
composition (solids content, reagent composition, and reagent pH) was also
varied with each top size coal. The data from approximately 50 experiments
were utilized to determine the efficiency of this unit operation and its
sensitivity to changes in process parameters.
Pyrite leaching from coal during slurry preparation necessitated that
the L-R unit operation be investigated in conjunction with the mixer
operation. However, the data presented in the previous section (Section
2.4.1) demonstrated that process performance in the mixer could be ac-
curately defined; thus, the L-R operation could be evaluated separately
even though the experimentation involved both unit operations. In fact,
the predictability of mixer performance by the use of Equations (3) and (4)
suggested that the L-R operation performance should be also predictable
through the simultaneous use of the above equations and those for reagent
regeneration developed in the previous bench-scale program, namely
where
rD moles of ferric ion regenerated per unit time,
K
p oxygen partial pressure in atmospheres,
U2
Fe ferrous ion concentration in moles per liter,
KR rate constant, a function of temperature only,
in liters/mole-atm-unit time
52
-------
and
KR = AR exp (-ER/RT) (6)
with "
AD = 6.7 x 105 Liters/Mole-Atm-Min
- K
and
ER = 13.2 Kcal/Mole.
However, attempts to predict the S (pyritic sulfur content of coal)
versus t, D (L-R reaction time) data contained in Appendices A and B,
L-K
Volume 2, through the simultaneous solution of Equations (3) through (6)
failed. In general, the experimentally determined pyrite removal rates
were higher than the predicted rates up to approximately 80 percent re-
.J.O
moval and lower thereafter; also, the measured Y values (Fe -to-Fe ratio)
were lower than the predicted ...Y values.
Comparison of the predicted and experimentally derived data led to the
following observations:
• The pyrite leaching rate during L-R processing at constant
temperature depended only on W and Y. The magnitude of
the rate dependence on these variables appeared to be the
same as that determined for ambient pressure leaching of
Lower Kittanning coals; thus, Equation (3) should have pre-
dicted the experimental S versus t data in all Meyers
Process unit operations if correct K, and Y values were
used.
t The K, (leaching rate constant) values predicted through
Equation (4) for L-R leaching were substantially lower
than the experimental K, values suggested by the pyrite
removal data for up to nearly 80 percent removal (it was
indicated earlier that the experimental rates were higher
than the predicted rates even though the corresponding
measured Ys were lower than the predicted). This lack of
53
-------
predictability of the L-R KL from AL and EL values deter-
mined under ambient pressure processing of coal could be
suggesting that KL should not be treated as a single
constant, but as a sum of rate constants each of which
relates to a certain coal or pyrite particle size within a
given coal top size; however, the generated data did not
permit verification of this possibility.
• The experimental data indicated that the L-R pyrite
leaching rate was dropping faster than predicted by
Equation (3) with a single valued KL (isothermal pro-
cessing) when pyrite removal exceeded approximately 80
percent. This observation implied a possible change in
reaction mechanism; however, the data indicated contin-
ued leaching rate dependence on W and Y. Again, a
change in KL value was indicated which because of the
isothermal processing was not predictable by Equation
(4) (in principle, this change in K, value could have
been predicted if K, was indeed a sum of rate constants
of known values).
• The low measured Ys in comparison to Ys predicted through
the simultaneous solution of Equations (3) through (6)
and the large variances in the deltas between measured
and predicted Ys, especially during early L-R processing
when pyrite leaching rate was the highest, led to the
conclusion that reagent regeneration during many of the
experiments was inefficient due to incomplete mixing
(e.g., inappropriate Reynolds Number in the slurry cir-
culation loop), although lack of adequate oxygen supply
could not be ruled out for certain experiments involving
33 wt. percent coal slurries. Equations (5) and (6)
proved repeatedly valid during separate reagent regen-
eration experimentation performed in the parametric
ranges that L-R processing was performed. There is
54
no
-------
reason to believe that the presence of coal could have had
an inhibiting effect on regeneration rate unless it af-
fected oxygen mixing (there was no evidence of direct
reaction of oxygen with coal).
The above observations led to the following tentative conclusions con-
cerning pyrite leaching data predictability by Equations (3) through (6):
t Pyrite removal versus leaching time data generated from
suspendable L.K. coal in the Meyers Process unit oper-
ations investigated in this program (Mixer, L-R, and
Settler Operations) should be predictable by Equations
(3) through (6). The only input data required are:
starting coal pyrite concentration, starting reagent iron
and ferrous ion concentrations, the oxygen partial pres-
sure during L-R processing, the temperature-time curve
of the process, and the values of the Arrhenius constants
A and E for Equations (4) and (6). If the L-R processing
behavior of the coal used in this program is typical of
L.K. coals, then more than one pair of A and E values
will be needed for Equation (4) for each coal top size.
This is tantamount to saying that Equation (3) should
be most properly expressed as
dW
(7)
(8)
with each K^ depending on a different pair of A and E
values. The K^ and W ^ values are specific either to
narrow coal or pyrite particle sizes within a given top
size coal or to elements of coal (shells) whose thick-
ness is the same for all particle sizes and depends only
on reagent accessibility rates (distance from coal sur-
face and coal porosity).
55
-------
• The L-R data generated in this program could not be
predicted by Equations (3) through (6) because of the
change in the KL value during L-R processing and be-
cause of inconsistent regeneration during the early
stages of L-R processing. However, if the conclusions
reached in the previous paragraph are valid, the
generated L-R data should be predictable by Equation
•
(3) through the use of experimentally determined K^
values from L-R data and through the use of measured
Y values.
In order to test the validity of the above conclusions, the L-R data
in Appendices A and B, Volume 2 of this report were treated as if derived
from two separate L-R operations (L-R Sections) describable by Equation (3)
but with different reaction rate constants (KL values). The first "L-R
Section" involved pyrite leaching rate data up to approximately 80 percent
overall pyrite removal (pyrite removal based on starting coal pyrite con-
tent); the second "L-R Section" involved the remaining L-R rate data
(pyrite removal beyond 80 percent) which apparently required a lower K,
value in order to be predictable by Equation (3). A K, value was determined
for each of the two L-R Sections which was then used in conjunction with
measured Y values in Equation (3) to test the predictability of the data in
Appendices A and B. The KL dependence on W during its transition from the
high to the low value was also determined from the experimental data. The
detailed procedures used and the results of the L-R rate development in-
vestigations are presented in the next section of this report.
2.4.2.1 Pyrite Leaching Rates During L-R Processing of L.K. Coal at 120eC
On the assumption that Equation (3) would prove to be the correct
empirical rate expression by which pyrite removal from suspendable L.K.
coal during L-R processing could be predicted, its integrated form was
used to determine the KL values applicable to 120°C L-R processing of the
two top size coals (14 and 100 mesh). Equation (9) is the integrated form
of Equation (3):
56
-------
WP
where
K, is the L-R rate constant for the particular coal top
size applicable up to nearly 80 percent pyrite re-
i -i '
moval and expressed in (hours)" (W )" ,
t. D is the L-R reaction time in hours,
L-K
7 is the average Y value of the reagent during t. _D»
dimensionless quantity,
W is the concentration of pyrite in coal at t, D, in
p L-K
wt. percent ,
w"J is the concentration of pyrite in coal at tL_R • 0
(end of Mixer processing and start of L-R), in wt.
percent.
The Y values were determined from measured Y versus t, _n curves plot-
ted from data in Appendices A and B, Volume 2. The W values were computed
from the S values (pyritic sulfur concentration in coal) in Appendices A
and B (W = (120/64.1) S ) and the w"J values were obtained from Table 8,
Section 2.4.1.
Over 20 data points derived from eight different experiments were
used in Equation (9) in order to determine the values of K, for each
of the two top size coals and to check their constancy with changes in
processing parameters other than temperature and coal particle size. These
computations revealed that the most probable K, value for L-R processing of
'-1-1
the 100 mesh top size coal at 120°C was 0.6 (hours) (W ) and the cor-
responding KL value for the 14 mesh top size coal was 0.5 (hours)" (W )" .
With the exception of occasional/outliers, the deviation in the K, values
was less than + 0.1. These computations also revealed that the use of S
values corrected for anomalies in the organic sulfur content of the coal
(Equation 2, page 28) tended to minimize the deviation in the computed K,
values. 57
1
-------
The estimated KL values for 120°C L-R processing of 100 mesh and 14
mesh top size L.K. coals are approximately 5 times larger than K. values
determined for pyrite leaching from the corresponding coal top sizes at 102°C
(Section 2.4.1). This unexpectedly high factor of five difference between
the 120°C and the 102°C KL values (Equation 4 predicts only a factor of 2
difference) was further verified from data plots similar to that illus-
trated in Figure 5 for 100 mesh top size coal.
In Figure 5, the top curve presents the S removal versus L-R reaction
time data generated in Experiments 7, 8, and 9; these experiments differ
only in overall L-R reaction time used (see Appendix A, Volume 2 for de-
tails on experimental conditions). The second curve presents the same
type of data from Experiment 10 which because of insufficient supply of
oxygen or inefficient mixing of oxygen into the slurry was performed with
reagent whose Y remained low for a large part of L-R processing. It is
noted that the lower Y values are reflected in lower S removals, as Equa-
tion (3) predicts should happen. "Corrected" S values were used and
the present S removal was computed from
i- (s_/s";)
% Sn Removal = ,—//os /fi/i i \ * TOO. (10)
P I "* \ 1 ȣv
The lower two curves in Figure 5 represent the expected S versus t
(coal leaching time) data from Experiments 7 through 10 if they were per-
formed at 102°C (ambient pressure) with Y versus Sp traces being identical
at 102°C and 120°C for the corresponding experiments. These curves were
generated by using a rearranged form of Equation (9), namely
t =(_! \ t 64.1 U) (11)
t % Y2 J ( 120 Sn jn ; v ;
KLY p wp
The required leaching time, t, was computed for attaining each of the
S values indicated in the top two curves; curve rather than actual Sp
values were used which is equivalent to using averaged data.
*Fnuat1on (10) Is a more exact form of Equation 1 (page 28) because It ac-
Sunts for the reduction in coal weight during reaction (pyrite dissolution);
iS derivltion, based on pyrite mass balance, is given in Appendix E, Volume 2.
-------
en
EXP. 7,8,9 (1Q2°C, AMBIENT
PRESSURE)
EXP. 10 (102°C AMBIENT
PRESSURE)
O EXP. No. 7
EXP. No. 8
EXP. No. 9
EXP. No. 10
120°C, 100 PSIG
345
LEACHING TIME, HOURS
Figure 5. 100 Mesh Top Size Lower Kittanning Coal Leached
with 5 Wt,% Fe Reagent (20% Slurries)
8
-------
The Y values
r „„ K, CYiuuaijr expiainea from the measured re-
agent Y values during the performance of Experiments 7 through 10 at 120«C;
a Y was computed for each level of Sp removal. The K, value used was that
determined in Section 2.4.1 for 102°C pyrite leaching from 100 mesh top
size L.K. coal (K,_ = 0.12 hours"1 W^1). v/J, which is the starting L-R
pyrite concentration in the coal used in each of the above experiments
(Table 8, Section 2.4.1), was assumed to be the starting pyrite concentra-
tion of the corresponding coals processed at 102°C.
The data in Figure 5 show that the leaching time required to obtain a
given Sp removal at 102°C (ambient pressure processing) is five times
greater than that required for the same S removal at 120°C (L-R processing)
provided the leaching took place under identical Y values. Thus, the pre-
viously estimated five fold increase in K. value between 102°C and 120°C
is further validated. In addition, the data indicate that KL remains
constant up to at least 75-80 percent S removal and that the rate of S
removal depends on the Y value of the reagent during L-R processing just
as was proven to be the case with ambient pressure processing (Section
2.4.1). Identical observations were made from similar treatment of 14 mesh
coal top size data. It was, therefore, concluded that the KL values deter-
mined in this Section for the L-R processing of the two top size coals at
120°C are valid for up to approximately 80 percent S removal. Thus, their
use in Equation (3), in conjunction with experimentally derived Y versus
t data, should lead to the prediction of the 120°C Sp versus tL_R data
in~Appendices A and B of Volume 2 regardless of the experimental conditions
used to generate these data (except for temperature), provided that Equa-
tion (3) is a valid rate expression for L-R processing of suspendable L.K.
coal as it has been determined to be for separate leaching-regeneration.
Equation (3) was subjected to the validity test and the results are
summarized in Tables 9 (100 mesh top size coal) and 10 (14 mesh top size
•D The test proved conclusively that the rate of pyrite removal from
suspendable Lower Kittanning coal during L-R processing is predictable by
Equation (3) up to at least 80 percent Sp removal.
60
-------
TABLE 9. ANALYZED AND PREDICTED PYRITIC SULFUR CONTENT OF 100 MESH TOP SIZE
L.K. COAL AS A FUNCTION OF L-R PROCESSING TIME AT 120°C
Exp.
No
A. 3 Wt.%
1
2
3
4
5
6
B. 5 Wt.%
7
8
9
10
11
12
C. 5 Wt.%
13
14
15
16
D. 5 Wt.%
17
VR-
sulfur
-------
TABLE 10. ANALYZED AND PREDICTED PYRITIC SULFUR CONTENT OF 14 MESH TOP SIZE L.K.
COAL AS A FUNCTION OF L-R PROCESSING TIME AT 120°C (5 WT% Fe REAGENT)
Exp.
No.
A. 20 Wt.'
B.
C.
D.
23
24
25
26
27
33 Wt.
28
29
30
31
32
33
33 Wt.
34
35
33 Wt.'
36
t, n = 0.5 Hours
L~K
Y < (VP
tL_R = 1.0 Hour
T SP*
-------
Tables 9 and 10 summarize all of the $„ and Y versus t, D data gener-
P L~K
ated during the first 3 hours of L-R processing of L.K. coal with iron
containing starting reagent at 120°C. The data are divided into five L-R
reaction time intervals UL_R) with the end of each time interval repre-
senting a slurry sampling point. The average reagent Y and the experimental
and predicted S values are listed for each tL_R. Also, the data are
grouped with respect to reagent iron concentration, slurry coal content,
and oxygen partial pressure. Each 7 value represents the average reagent
Y during the indicated interval and it was computed from plots of the Y
versus t, R data listed in the tables of Appendices A and B. The S* values
represent the pyritic sulfur concentration in the coal at the end of each
reaction interval; they were derived from sulfur forms analyses of the coal
in slurry samples or from analyses on the product coal (the pyritic sulfur
value was adjusted for anomalies in the organic sulfur content of the coal
sample). The (Sp) values represent the predicted pyritic sulfur content
of the coal at the end of each t. _R as computed from
P
Equation (12) is an integrated form of Equation (3); all of its terms have
been defined previously. The values of KL> Y, and tL_R are listed in
Tables 9 and 10; the values of W? for each experiment are listed in Table
8 (Section 2.4.1).
Comparison of analyzed and predicted S values in Tables 9 and 10 re-
veals good agreement in these majority of cases; in fact, when normal sam-
pling and analyses uncertainties are taken into consideration the agreement
can be considered excellent for S removals up to nearly 80 percent (resid-
ual S 0.8 wt. percent). It should be noted that these experimental data
were generated under a variety of processing conditions involving starting
reagent iron and sulfate concentrations, starting reagent Y, slurry coal
content, slurry pH, total pressure, and oxygen partial pressure.
The large majority of the Sj$ values 1n Tables 9 and 10 were derived
63
-------
from analyses of coal present in reactor samples drawn during processing.
The S* values for which a deviation is indicated represent analyses per-
formed on samples taken at the end of the run where the entire batch of
L-R processed coal was sampled; the deviation was derived either from
duplicate analyses on the same sample or from analyses of duplicate samples.
S* values which appeared to be grossly inconsistent with respect to the
rest of the data derived from a particular experiment or with respect to
data from similar experiments are placed in parentheses in the above tables.
Discrepancies between analyzed and predicted S are not limited to the
values in parentheses; in certain cases, entire experiments appear not to
obey Equation 3 (e.g., Experiments 2 and 11), but the reasons for the dis-
crepancy are not readily apparent.
The majority of the discrepancies between analyzed and predicted S
appear in the first group of experiments in Table 9 (Experiments 1 through
6). In these experiments coal was processed with iron sulfate solutions
which did not contain added sulfuric acid. Apparently, without the added
acid the pH during L-R processing was not sufficiently low to inhibit iron
sulfate deposition on the coal (see data in Appendix A, Table A-l). It
is conceivable that the deposited iron sulfate complicated the sulfur forms
analyses performed on these coal samples and led to errors in pyritic sul-
fur determinations (10 to 20 percent error in pyritic sulfur analyses can
account for most of the observed discrepancies). Another reason that the
discrepancies between analyzed and predicted S are concentrated in this
group of experiments may be inadequate sampling. Experiments 1 through 6
represent early experimentation with L-R processing and it could be that
sampling techniques were not as good as later in the program. It
should be noted, however, that there is nothing in this group of data, or
in any of the other groups of data in Tables 9 and 10, which relates dis-
crepancies between analyzed and predicted S values to the value of a
certain processing parameter (at least within the indicated ranges of the
processing parameters).
Consistent discrepancies between predicted and analyzed S values
appear when the pyritic sulfur concentration in coal has been reduced to
64
-------
approximately 0.8 weight percent (nearly 80 percent pyrite removal). Appar-
ently, further S reduction during L-R processing takes place at rates
which are appreciably lower than those predicted by Equation (3) and the
KL values in Tables 9 and 10. This discrepancy between S* and (S ) be-
comes more apparent when predicted S values for three hours L-R processing
(Tables 9 and 10) are compared to S values derived from analyses of coal
processed in excess of three hours (Experiments 2, 4, 5, 6, 11, 12, 16, 24
through 27, 32, and 33 in Appendices A and B, Volume 2). It is also
illustrated by the 14 mesh top size data in Figure 6 (120°C rate data from
Appendix B).
In Figure 6 the reciprocal pyritic sulfur concentration in coal (coal
analysis values corrected for organic sulfur anomalies) is plotted against
L-R reaction time which was normalized for inconsistencies in reagent
regeneration. According to Equation (12), the slope of the data in Figure
6 should be constant and proportional to the KL value of Equation (3) for
120°C processing of 14 mesh top size coal. Indeed, the data illustrates
that this is the case up to the 1/S value of approximately 1.2 (S of 0.83
wt. percent and W of 1.56 wt. percent), but a definite change in K, value
occurs thereafter. Actually, a change in reaction mechanism is indicated
by the data which Equation (3) did not predict (speculation on possible
causes of data unpredictability in this S region and potential modifi-
cations of Equation (3) which could possibly rectify this discrepancy were
presented in the previous Section). However, in the absence of a more
precise rate expression, Equation (3) can be utilized to predict L-R pro-
cessing even beyond 80 percent pyrite removal by allowing for a change in
the isothermal value of K, .
The data in Figure 6 indicate that L?R processing can best be represented
by Equation (3) through the use of three K, values. For the 14 mesh top
size L.K. coal processed at 120°C the indicated K. values are as follows:
KU = 0.5 (hours)"1 (% pyrite in coal)"1 for W >.1.6 wt.%
KL2 = f(V " Wp " 1<1 (nours)" (% Pyrite) for I-6 lWp 1 1-2 wt.%
KL3 = 0.1 (hours)"1 (% pyrite in coal)"1 for VI < 1.2 wt.35
65
-------
o
V
z
I
1.0
1.2
1.5
2.2
2.0
O 1-8
U
Z
" 1-6
u
R 1.4
u
OS
=J 1.2
CO
y
fc 1.0
Q:
5 °-8
O
- Q£
G °'6
LU
0.4
0.2
0
KL =f(wp)
PREDICTED CURVES FROM
EQUATION (3) AND INDICATED KL
EXPERIMENTAL DATA
ty , NORMALIZED REACTION TIME IN HOURS
Figure 6. Pyrite Leaching Rate Constant Data
-------
The deviation in the KL-J is estimated to, be + 10 percent, but the deviation
in the KL3 value is closer to +. 50 percent. However, the large deviation
in KL3 is understandable since it can result from only a ±6.1 deviation
in the S values which comprise the concentration region to which K. 3
applies; such deviation is normal for pyritic sulfur analyses.
It is interesting to note that the estimated K-3 value is identical to
the KL value found applicable for ambient pressure processing of the same
size coal at 102°C. Within experimental uncertainty, the same observation
was made with the 100 mesh top size coal where K, - was estimated to be 0.12
i i *
(hours)" (W )~ . The implication of this observation is that if KL3 is
temperature insensitive during small temperature changes (e.g., a diffusion
constant) it is possible to have an observable change in reaction mechanism
at 120°C but not at 102°C, provided that at 102°C KLI and KLS are nearly
identical. This may be the case with suspendable L.K. coal since the L.K.
coal used in the previous bench-scale program indicated no change in the
pyrite leaching rate mechanism up to 95 percent pyrite removal at 102°C
(see also discussion in Section 2.4.1).
It is believed that the preceding discussion proves conclusively that
Equation (3) can be used to predict the Meyers Process performance during
the L-R processing of suspendable L.K. coal at 120°C provided the experi-
mentally derived KL values are used and the Y versus tL_R history of the
process is known. The latter requirement limits the utility of Equation
(3) severely unless the change in Y during L-R processing can be predicted
through the use of a valid reagent regeneration rate expression. In the
next Section an attempt is made to prove that Equation (5) is the appro-
priate reagent regeneration rate expression for L-R processing when oxygen
supply to the slurry is adequate and proper mixing has occurred.
The KIT for 100 mesh top size coal was determined earlier to be 0.6 (hours)"
the KL£ was estimated to be the same as that of the 14 mesh top size
coal.
67
-------
2.4.2.2 Reagent Regeneration During L-R Processing
Reagent regeneration involves the oxidation of the Fe+2 generated
+3
during the pyrite leaching reaction to Fe by oxygen. During L-R pro-
cessing the two operations (pyrite leaching-reagent regeneration) take
place simultaneously and their rates are related through the iron forms in
+2
the reagent. The regeneration rate depends on the Fe concentration in
+3
the reagent solution and the leaching rate on the Fe -to-Fe ratio, which
+3 +2
is defined as Y.* Since Fe is equal to Fe-Fe , Y is also a measure of
regeneration rate.
•
Review of the Y versus t, R data in Appendices A and B, Volume 2 re-
veals that in most experiments the rise in Y was rapid during the early
stages of L-R processing. This observation during the performance of the
experiments led to the erroneous conclusion that reagent regeneration was
taking place efficiently during L-R processing. When valid pyrite leaching
rate constants were determined, it became possible to generate predicted Y
versus t, R data which when compared to the experimental data led to the
following observations:
• In the majority of experiments reagent regeneration was
less efficient than predicted during the first hour of
L-R processing.
• Reagent regeneration was inconsistent in a number of
experiments.
• Equations (5) and (6) appeared to be valid rate ex-
pressions for spent reagent regeneration, but the
activation energy value computed from the regeneration
of synthetic spent reagents for use in Equation (6) is
probably higher than the activation energy applicable
to the regeneration of actual spent reagent in the
presence of coal.
• The reagent regeneration rate was probably catalyzed by
cations leached from the coal or by the coal itself.
* +2
Note that because of Fe production in the mixer the starting L-R Y was
always low; therefore, high regeneration rates were expected during the
early stages of L-R processing.
68
-------
The validity of the above observations becomes evident from the data pre-.
sented in Figures 7 through 9 where predicted Y versus t, D data (solid
L~K
curves) are compared to measured Y values as a function of L-R processing
time (dashed curves). The predicted data in these figures was abstracted
from Table D-l, Appendix D, Volume 2 of this report.
Table D-l lists predicted data for both the Mixer and L-R Operations
which correspond to the experimental data from every experiment in Appen-
dices A and B of Volume 2. The data for each experiment are presented in
two sections. The top section lists Mixer Operation data which were dis-
cussed in detail in Section 2.4.1. The bottom section lists the predicted
Wn (pyrite concentration in coal), Y, and x, D (fraction of pyrite removed
p L-K
during L-R) as a function of t, „ (L-R reaction time); it also lists the
overall pyrite removal, x, . , at each of the tL_R values shown (over-
all pyrite removal is based on starting coal W and it includes pyrite
removal in both the Mixer and L-R Operations). The predicted L-R data were
generated from known starting coal and reagent compositions and Equations
(3), (5), and (6). The K, values used in Equation (3) were those deter-
mined as applicable to L-R data up to 80 percent pyrite removal. The KR
values (regeneration rate constants) were computed from the Arrhenius
constants, AR and ER, through the use of Equation (6); these constants were
determined from synthetic spent reagent solutions.
Figure 7 illustrates measured and predicted Y values derived from L-R
processing at 120°C of 20 wt. percent slurries of 100 mesh top size L.K.
coal. The solid curve depicts the expected Y variation with L-R reaction
time from Experiments 8 through 12 (predicted data) and the dashed curves
.show the corresponding experimentally derived Y data. If small differences
in starting Y values and Mixer reaction times are disregarded, the only
meaningful difference among these experiments was L-R reaction time. Thus,
at any given reaction time the slurry composition, and therefore Y, should
be the same for all the experiments in this group (a small variation could
be explained on the basis of the observed differences in starting L-R Y
values). The predicted data1 curve confirms that these experiments should
have yielded identical Y versus tL_R data if performed reproducibly. The
69
-------
100 MESH TOP SIZE COAL, 20 WT. % SLURRIES, 5 WT. % Fe REAGENT
PREDICTED DATA
EXPERIMENTAL
1.0 1.5 2.0
L-R REACTION TIME, HOURS
Figure 7. Predicted and Experimental Y Values During L-R Processing
of L.K. Coal at 120°C
70
-------
14 MESH TOP SIZE COAL, 33 WT % SLURRY. .5 WT % Fe REAGENT
PREDICTED DATA EXP. 37 & 38
EXPERIMENTAL DATA
EXP. No. 37
EXP. No. 38
1.0 1.5 2.0
L-R REACTION TIME, HOURS
2.5
3.0
Figure 8. Predicted and Experimental Y Values During L-R Processing
of L.K. Coal at 110°C
71
-------
14 MESH TOP SIZE COAL, 33 WT % SLURRY, 5 WT % Fe REAGENT
PREDICTED DATA EXP. 39 & 40
EXPERIMENTAL DATA
EXP. No. 39
EXP. No. 40
1.0 1.5 2.0
L-R REACTION TIME, HOURS
Figure 9- Predicted and Experimental Y Values During L-R Processing
of L.K. Coal at 130°C
72
-------
experimental data show that Experiments 8 and 9 yielded reproducible data
which, however, differed substantially from the data generated in Experi-
ment 10 or in Experiments 11 and 12. The lack of consistency in Y data
reproducibility could have only resulted from inconsistent oxygen supply
to the slurry (insufficient feed or inefficient mixing) since there was no
evidence of errors in iron mass balance. (Note that the Y value depends
+2
only on Fe and oxygen concentration when pyrite leaching is performed at
constant temperature; starting Y has a negligible effect as indicated by
the predicted curve in Figure 7).
As would be expected from inefficient regeneration, the data in Figure
7 reveal that oxygen starvation was more pronounced during the early stages
+2
of L-R processing when Fe generation was high; however, the data also
reveal that oxygen starvation need not have taken place if experimentation
was performed in accordance with the procedures used in Experiment 10.*
In Experiment 10 the oxygen supply to the slurry appears to have been ade-
quate to permit reagent regeneration to take place under the kinetically
controlled rate. This conclusion is based on the observation that the
experimental Y versus tL_R data from this experiment traces a curve which
is identical in shape to the predicted curve (the two curves will coincide
if translated on the time axis by a constant multiplier). Conversely, this
observation leads to the conclusion that Equation (5) is a valid reagent
regeneration rate expression for use in the prediction of L-R processing
data. The fact that the predicted and measured reaction times for the
same Y value differs by a constant factor indicates that the actual KR
value for this temperature is higher than that used in Equation (5) to
generate the predicted data curve. Since Equation (6) is also probably
valid on the basis of the above data comparisons, a wrong KR value implies
that at least one of the Arrhenius constants used to predict the L-R data
was wrong.
Equations (5) and (6) were derived from a substantial bank of data on
the regeneration of synthetic spent reagent solutions (previous bench-scale
*
Oxygen feed procedures were not intentionally varied, but precise control
of slurry circulation fates was difficult in the experimental set-up used
in this program, principally because of slurry pump size.
73
-------
program). It is highly unlikely that the Arrhenius constants deterged
from these experiments were .wrong. It was, therefore, assumed that the
apparent higher KR values observed here under L-R processing were due to
catalytic effects emanating from some unknown substance leached from the
coal (probably metallic cations and very likely copper).* The catalytic
effect would be expected to manifest itself through the lowering of the
activation energy of the reaction, higher KR values, and different, most
likely lower, KR dependence on temperature. In fact, if the AR value was
not affected by the catalyst, it would be expected that the catalytic
effect on KR to be proportionally greater at the lower of two reaction
temperatures. The data in Figures 7, 8, and 9 tend to show that this was
actually the case. Until further verified, this catalytic effect should
be considered as an assumption.
In Figure 8, predicted data are compared to experimental L-R data
generated at 110°C. In Figure 9, the same type of data are compared from
130°C experimentation. In both cases experimentation involved the L-R pro-
cessing of 14 mesh top size L.K. coal as 33 wt. percent slurries in 5 wt.
percent iron reagent. Very good reproducibility of data is exhibited by
each pair of experiments depicted in these two Figures. However, it appears
that the regenerationYate did not reach its kinetic limits during at least
the first hour of L-R processing at 130°C probably due to insufficiency in
oxygen. Thus, the 130°C experimental curves in Figure 9 are not expected
to obey Equation (5) and, therefore, they are not strictly comparable to
the data from Experiment 10 (Figure 7) and Experiments 37 and 38 (Figure 8).
Yet it can still be argued that the delta between "catalyzed" and "un-
catalyzed" KR values is smaller at the lower temperatures (this is espe-
cially true if it is assumed the regeneration rate at t^_R = 1.5 hours is
no longer limited by oxygen). Of course, the effect of temperature on the
difference in predicted and experimental (suggested from measured Y) KR
values is readily apparent from comparison of the data in Figures 7 and 8.
Predicted and experimental data were compared for all the experiments
in Appendices A and B. (The experimental Ys are listed in Appendices A and
B and the corresponding predicted Ys in Appendix D). These comparisons
It is possible, but not very probable, that the catalyst was leached from
the reactor, since different type stainless steel reactors were used during
L-R and synthetic reagent experimentation.
74
-------
strengthened further the conclusions drawn from the data in Figures 7
through 9. However, the number of experiments during which regeneration
was not limited by oxygen insufficiency was small and, therefore, no
attempt was made to estimate either the activation energy of the assumed
catalyzed reaction or the KR value at any given L-R temperature. Attempts
to discern any variation in the KR value as a function of reagent recycle
time (due to accumulation of catalyst) were inconclusive for the same
reason.
The above analysis on the reagent regeneration data derived from L-R
processing furnishes the reasons for (a) the scatter in the pyrite removal
versus leaching time data, (b) the difficulty to discern any sensitivity
in pyrite removal rates to changes in processing parameters, and (c) the
requirement that experimental and not predicted Y values had to be used
in Equation (3) in order to predict the generated S_ versus t. D data.
p L—K
Thus, it became clear that parametric effects on pyrite leaching must be
examined on the basis of predicted data rather than experimental data in
order that the real effect of each parameter becomes apparent. Since
Equation (3) was proven to be a valid leaching rate expression under
virtually all processing conditions examined when experimentally determined
KL values were used with it (Section 2.4.2.1) and since Equations (5) and
(6) appeared to be valid on the basis of the data analysis performed in this
section, the predicted data listed in Table D-l, Appendix D, Volume 2 were
used wherever appropriate in the examination of parametric effects. It
should be noted, however, that the data in Table D-l are only valid up to
80 percent pyrite removal because they were generated with KL values which
were proven valid only up to this level of pyrite removal. Also, these
data were generated using AR and ER values derived from uncatalyzed reagent
regeneration reactions; thus, the predicted pyrite removal values listed in
Table D-l are lower than the expected pyrite removal values from L-R pro-
cessing under efficient reagent regeneration. However, at a given L-R
processing temperature these values are uniformally lower and should not
affect data comparison.
The higher (uncatalyzed) ER value was also used in process engineering
analysis because the lower ("catalyzed") ER value could not be determined
75
-------
with adequate precision from the available data and because it is not
known if similar ER depression would occur during the L-R processing of
other coal, even L.K, coals from different mine than the one used in the
current program. The use of the higher ER value should not substanially
affect the size of the reactors or the computed process costs, except
perhaps when L-R operation is performed at low temperatures and pressures
(e.g., 100°C and 50 psig).
2.4.2.3 Temperature Effects on L-R Processing of L.K. Coals
Figures 10 and 11 illustrate the temperature effect on pyrite removal
during the L-R processing of 100 and 14 mesh top size L.K. coals in the
temperature range of 110*C to 130°C. The temperature effect was investigated
on 20 wt. percent slurries of 100 mesh coal and on 33 wt. percent slurries
of 14 mesh coal; thus, the data plotted in these figures differ in two pro-
cessing variables. However, the temperature effect should be identical
for the two top size coals and, therefore, the observed difference in
the magnitude of the temperature effect in Figures 10 and 11 is entirely due
to the difference in slurry concentrations.*
The data in Figures 10 and 11 show that the rate of pyrite removal
from suspendable L.K. coals during L-R processing increases with increasing
temperature. In addition, the temperature effect on pyrite removal is
larger between 110° and 120°C than between 120° and 130°C regardless of
slurry composition and it is more pronounced on 33 wt. percent slurrios
than on 20iWt. percent slurries. The data also reveal that the 130°C
curves of the two slurry concentrations are nearly identical. These ob-
servations lead to the conclusion that in terms of temperature effect alone,
the most efficient processing conditions, among those represented in Figures
10 and 11, are the conditions under which Experiment 40 was performed. In
addition, the data in these figures indicate that 33 wt. percent slurry
processing is preferable to 20 wt. percent slurry processing at any of the
three temperatures since twice as much coal per unit time is processed at
*
Pyrite removal rates from the two size fraction coals differ only in
value (Arrhenius frequency factor) which is temperature independent.
76
-------
90
O
Z 70
CO
CO
LU
U
O
60
O
Z
oi
< 50
£ 40
D_
H-
z
30
20
10
130°C (EXP. 20)
120°C (EXP. 10)
110°C(EXP. 18)
0.5
1.0 1.5 2.0
L-R REACTION TIME, HOURS
2.5
3.0
Figure 10. Temperature Effect on L-R Processing of 100 Mesh Top
Size L.K. Coal (20 Wt. Percent Slurries)
77
-------
90
1.0 1.5 2.0
L-R REACTION TIME, HOURS
2.5
3.0
Figure 11. Temperature Effect on L-R Processing of 14 Mesh Top
Size L.K. Coal (33 Wt. Percent Slurries)
78
-------
the higher slurry concentration than at the lower one while the penalty in
reaction time for processing the more concentrated slurry never exceeds a
factor of 1.5 when comparison is made at the same temperature.
It should be noted that the discussed temperature effects apply only
up to 80 percent pyrite removal; beyond this pyrite removal level the
temperature effect on L-R processing becomes negligible. Also, the effects
depicted in Figures 10 and 11 represent the overall temperature influence
on suspendable coal desulfurization during L-R processing and not the de-
pendence on temperature of the individual rates controlling the process
pyrite leaching and reagent regeneration.
The temperature effect on the reagent regeneration rate during L-R
processing is adequately defined through Equation (6) using the AR and ER
values listed in Section 2.4.2. It will be shown that the temperature
effect on the pyrite leaching rate, up to 80 percent pyrite removal, can
also be defined by an Arrhenius type expression, but with at least one of
the Arrhenius constants being different for ambient pressure processing
from that for L-R processing. Beyond 85 percent pyrite removal the
leaching rate appears to be insensitive to temperature in the 102° to 130°C
range. A rate constant transition region occurs between 80 and 85 percent
pyrite removal, but the leaching rate temperature dependence in this region
was not defined.
The inability to predict the values of L-R leaching rate constants
from the A, and EL values estimated from ambient pressure processing of
L.K. coals (see Mixer Section) necessitated that empirical L-R KL values
be determined for each top size coal from L-R data generated at each of
the three temperatures. The methodology used to determine these KL values
was described in detail in Section 2.4.2.1 in conjunction with the deter-
mination of KL values for the 120°C L-R processing of 100 mesh and 14 mesh
top size L.K. coals. Using this approach the following KL values were
estimated for llOtand 130°C processing of suspendable L.K. coal:
79
-------
(a) 110°C L-R Processing
K, =0.42 (hours)" (pyrite cone, in coal)" for TOO mesh top
size coal
K, = 0.35 (hours)" (pyrite cone, in coal)~ for 14 mesh top
size coal
(b) 130°C L-R Processing
K, = 0.84 (hours)" (pyrite cone, in coal)" for TOO mesh top
size coal
K, = 0.70 (hours)"1 (pyrite cone, in coal)" for 14 mesh top
size coal
These K, values were determined from randomly selected experimental data
points and subsequently tested through Equation (12) (an integrated form
of the leaching rate expression Equation 3) against all the rate data
generated at these temperatures (Tables A-5, A-6, B-5, and B-6 in Volume
2). The results of this test are shown in Table 11. The agreement between
coal pyritic sulfur content values derived from coal sample analyses, S*,
and those predicted by Equation (12), (S_)D, is, in general, very good.
Thus, the K. values listed above appear to be valid (detailed discussion
on the validity of these comparisons was presented in Section 2.4.2.1 in
conjunction with the review of the data in Tables 9 and 10).
It is noted that the KL values for 130°C leaching are exactly double
those for 110°C leaching. A factor of 2 was also estimated to be the dif-
ference in KL values for ambient pressure pyrite leaching at temperatures
differing by 20°C. This observation implies that the apparent activation
energies governing the leaching reaction rates at ambient pressures and
under L-R conditions are the same. However, the K, values between 102°C
(ambient pressure) and 110°C (L-R processing) differ by a factor of 3.5.
This unexpectedly large factor is not derivable from the A. and E, values
listed on page 45.
80
-------
TABLE 11. ANALYZED AND PREDICTED Sn OF L.K. COAL AS A FUNCTION OF L-R
PROCESSING TIME AT 110°CP(94 PSI 09) AND 130°C (76 PSI 09)
Exp.
No.
t, n = 0.5 Hours
T SP* P
t,n = 1.5 Hours
T SP* (SP'P
t, n = 2.0 Hours
L-K
T SP*
-------
Figure 12 is an Arrhenius plot of KL values derived from both L-R and
ambient pressure processing of suspendable L.K. coals. The ambient pres-
sure curve and the Arrhenius constants derived from it were discussed in
Section 2.4.1. In contrast to the ambient pressure processing data, the
L-R data exhibit very good compliance to the Arrhenius Equation in the
investigated range. The Arrhenius constants derived from these data were
as follows:
(a) For 100 mesh top size coal
E. = 11.1 Kcal/mole pyrite removed
AL = 8.9 x 105 (hours)"1 (pyrite cone, in coal)"
(b) For 14 mesh top size coal
E, = 11.1 Kcal/mole pyrite removed
AL = 7.4 x 10 (hours)' (pyrite cone, in coal)"
2.4.2.4 Coal Top Size Effects During L-R Processing
In previous sections it was demonstrated that the effect of coal top
size on suspendable L.K. coal desulfurization by the Meyers Process could
be confined to the leaching rate constant. It was also shown that the
KL value of the 100 mesh top size coal was 20 percent larger than that of
the 14 mesh coal, regardless of process operating conditions. This delta
in KL values represents the predicted maximum coal particle size effect on
pyrite leaching rates when the above two top size coals are processed iden-
tically.
Under L-R processing, assuming identical starting systems, the coal
top size effect is tempered by the fact that the Y value of the coarser
coal slurry at a given reaction time is higher than the Y of the less
coarse coal slurry (this is because the finer coal reacts at a higher rate
generating more Fe+ and the difference in Fe+2 is not entirely compensated
by the higher regeneration rate). Figure 13 illustrates the coal top size
effect on L-R processing efficiency of suspendable L.K. coal with typical
L-R data abstracted from Table D-l, Volume 2. It can be concluded from
82
-------
oo
to
0.80
0.70
0.60
0.50
0.40
0.30
1
z
8
0.20
0.10
O
0.05
8 0.03
0.02
2.4
130° 120° 110° 102°
2.6
O 100 MESH TOP SIZE
A 14 MESH TOP SIZE
• 100 MESH TOP SIZE
AT AMBIENT
COAL PROCESSED
PRESSURE
85°
70° C
COAL
COAL
L-R PROCESSED
2.8
3.0 X 10'
_o
,-1
Figure 12.
RECIPROCAL TEMPERATURE IN (°K)
Arrhenius Plots of Pyrite Leaching Rate Constants
-------
90
Q_
UJ
z
UJ
80
70
I
_l
o
i 60
50
40
30
20
10
0.0
100 MESH X 0 (EXP. 16),
14 MESH X 0 (EXP. 29)
0.5 1.0 ' 1.5 2.0
L-R REACTION TIME, HOURS
2.5
Figure 13. Coal Top Size Effect on L-R Processing of Suspendable
L.K. Coal
84
-------
these data that the coal top size effect on pyrite removal during L-R pro-
cessing is negligible up to at least the 14 mesh top size. Since in
general liquid-solids separation efficiency increases with increase in
solids particle size, the data in Figure 13 suggest that desulfurization
of L.K. coal should be preferably performed on the 14 mesh top size rather
than on finer coal sizes.
2.4.2.5 Slurry Concentration Effects
Slurry concentration, defined as the weight of coal in 100 weights
of slurry, does not appear as a variable in either the leaching or the
regeneration rates. However, this parameter affects the Y value of the
slurry and, therefore, it is an implicit variable of the leaching rate.
The higher the coal concentration per unit weight of reagent solution the
+2
higher is Fe production from pyrite oxidation and therefore the larger
the influence on reagent solution Y which is the Fe -to-Fe ratio with
+3 +2
Fe = Fe-Fe . Thus, the slurry concentration effect is maximum during
ambient pressure, batch coal leaching without reagent exchange (e.g.,
Mixer and Settler Operations in this program) and zero if coal is pro-
cessed under constant Y (e.g., continuous, variable rate reagent exchange
leaching). During L-R processing this parameter affects also the regen-
eration rate since an increase in the slurry concentration will result
+2
in higher Fe concentration in the reagent solution of the slurry which
in turn will lead to higher regeneration rates. The magnitude of the
effect of slurry concentration on pyrite removal depends on the values
+2
of L-R processing parameters which affect the delta between Fe production
+2
from the leaching reaction and Fe consumption in the reagent regeneration
reaction.
In Section 2.4.2.3 the slurry concentration effect was discussed in
conjunction with the temperature effect on pyrite removal from suspendable
L.K. coal during L-R processing. Comparison of the data in Figures 10 and
11 led to the conclusion that in the temperature range of 110° to 130°C it
was more efficient to desulfurize L.K. coal as 33 wt. percent slurries than
as 20 wt. percent slurries, even though at 110°C the pyrite leaching rate
85
-------
was appreciably lower at the higher slurry concentration (note that twice
as much coal is processed per unit time when 33 wt. percent slurries are
substituted for 20 wt. percent slurries).
Figure 14 compares data generated from the processing of 33 and 20 wt.
percent slurries with all other processing conditions being virtually
identical (120°C, 100 psig, 85 psi 02> 5 wt. percent Fe reagent, 14 mesh
top size coals). These data show that at the indicated processing con-
ditions, which represent nearly the center of the range of operational
parameters considered feasible for the Meyers Process, the slurry concen-
tration effect on pyrite removal rates is insignificant in the range of
33 to 20 wt. percent slurries. Thus, the data in this Figure suggest that
the 33 wt. percent slurry represents the minimum desirable slurry concen-
tration for use in conjunction with the indicated L-R processing conditions.
Review of the entire bank of data in Table D-l, Volume 2 reveals that
the slurry concentration effect on pyrite removal rate during the L-R
processing of suspendable L.K. coal is principally affected by temperature
and oxygen partial pressure. The effect of this parameter increases as the
values of either temperature or oxygen partial pressure are reduced within
the investigated ranges.
2.4.2.6 Reagent Composition Effects
The Meyers Process reagent is an iron sulfate solution which under
certain processing conditions require the addition of sulfuric acid. Thus,
its effects on coal pyrite leaching rates should be related to total iron
concentration, iron forms (Fe+2 and Fe*3) concentrations, sulfate ion con-
centration, and pH.
The effects of iron forms on leaching and regeneration rates have been
discussed in detail in previous sections. It has been repeatedly demon-
strated that one of the most important parameters to coal desulfurization
+3
by the Meyers Process is the quantity Y, which represents the Fe -to-Fe
ratio in the reagent solution. The second order dependence of the pyrite
86
-------
90
80
O
to
LU
U
O
O
Z
Q2
a
Q.
I-
UJ
U
UJ
Q.
70
60
50
40
30
20
20 WT. % SLURRY (EXP. 24)
33 WT. % SLURRY (EXP. 32)
10
0.0
Figure 14.
0.5
1.0
1.5
2.0
L-R REACTION TIME, HOURS
2.5
Slurry Concentration Effect on L-R Processing of
Suspendable L.K. Coal
87
-------
leaching rate on this parameter held valid regardless of processing scheme
or processing conditions used, provided that the starting reagent contained
a soluble iron salt. It was also shown that the regeneration rate ex-
hibits a second order dependence on ferrous ion concentration. Thus,
reagent composition in terms of iron forms, especially the ratio of iron
forms, has a significant effect on pyrite removal from coal by the Meyers
Process.
The sulfate concentration and pH effects of the starting reagent
solution on Meyers Process performance were investigated concurrently.
Both these parameters were varied through the iron sulfate concentration
in the starting reagent and through added sulfuric acid. The sulfate
concentration was varied from zero (when the starting reagent was distilled
water) to 15 wt. percent (when 5 wt. percent Fe starting reagent of Y = 1
acidified by 2 wt. percent sulfuric acid was used). The corresponding
starting reagent pH range was 7 to approximately 0.7. It should be noted,
however, that even when the starting pH was 7 its value dropped to below
2 within a few minutes of reaction time because pyrite oxidation generates
sulfuric acid. The drop in pH was experienced even under L-R processing
where the generated sulfuric acid is, to a large extent, quickly consumed
+2 +^
by the regeneration of the ferric ion (oxidation of Fe -to-Fe ).
/
According to rate Equations (3) through (6) sulfate concentration and
pH should not affect directly either the leaching or regeneration rates.
The data verified this expectation, except that at L-R reaction times in
excess of approximately one hour these two parameters exhibited a small
indirect effect on pyrite removal through their influence on the Y of the
reagent solution. The data in Appendices A, B, and C show that when L.K.
coal was processed under L-R conditions with reagent solutions which did
not contain added sulfuric acid, the iron content of the coal during pro-
cessing did not decrease in proportion to pyrite removed (Experiments 1
through 11, 15, 18 through 22, 41, 42, 43, and 45). Comparison of pre-
dicted (Table D-l) and experimental Y values indicated that iron was
depositing on the coal at a higher Fe+3-to-Fe+2 ratio than that in the
reagent solution; thus, the measured Y values in these experiments at t, R
-------
values in excess of one hour were increasing at lower rates than expected
and at longer reaction times the Y actually dropped with increasing t,
in certain experiments (e.g., Experiment 1). It should be noted, however,
that at the tL_R range where iron begins to have an effect on Y, the
reagent Y values are high and pyrite leaching is principally limited by
the low pyrite concentration in coal; therefore, iron deposition has an
insignificant effect on pyrite removal during L-R processing.
The major sulfate concentration-pH effect on the Meyers Process was
not on pyrite removal but on sulfate deposition on processed coal and on
processed coal washing. The data in Appendices A and B show that, in the
experiments where sulfuric acid was not added to the starting reagent,
iron and sulfate deposited on the coal during L-R processing when pyrite
removal exceeded 40-50 percent and the reagent Y value exceeded approxi-
mately 0.5. The deposition of iron and sulfate occurred to approximately
the same extent when 3 wt. percent starting Fe reagent solution was used
(10.7 wt. percent ferric sulfate) as when 5 wt. percent Fe reagent (17.9 wt.
percent ferric sulfate) was used. The same observation is made, with re-
spect to iron deposition, from the data in Appendix C which were generated
with reagent solutions whose iron sulfate concentration was derived from
coal (iron sulfate was not added to the starting reagent solution); during
these experiments sulfate did not deposit on the coal. Iron and sulfate
deposition during coal processing under separate leaching regeneration
conditions at ambient pressure was minimal (< 0.2 wt. percent sulfate
sulfur) even with no added acid to the starting reagent solution and
regardless of the iron sulfate concentration in the slurry (up to 10 wt.
percent Fe provided the Y was maintained above 0.8).
The above experimental observations are consistent with the Meyers
Process chemistry formulated on the basis of elemental sulfur recovery
and iron mass balance (ferrous ion production from known quantities of
oxidized coal pyrite). According to this chemistry each mole of pyrite is
converted to one mole Fe+2, 1.2 moles of sulfate, and 0.8 moles of ele-
mental sulfur through oxidation by Fe+3 according to the following
reactions:
89
-------
0.4 FeS2 + 0.4 Fe2(S04)3 * 1.2 FeS04 + 0.8S
0.6 FeS2 + 4.2 Fe2(S04)3 + 4.8 H20 - 9 FeS04 + 4.8 H2S04
The sum of these two reactions yield the following net reaction for pyrite
leaching in the absence of oxygen (separate Teaching-regeneration):
FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 + 10.2 FeS04 + 4.8 H2$04 + 0.8S
Thus, each mole of pyrite reacted yields 4.8 moles of sulfuric acid and
hence the observed immediate drop in pH. The low pH inhibits the hydrolysis
of the iron salts and, therefore, no iron deposition occurs on the coal
unless iron sulfate solubilities are exceeded.
Under L-R processing oxygen is added to the process during coal
+3
pyrite leaching for the purpose of regenerating the consumed Fe (ferric
sulfate). Thus, the following reaction takes place simultaneously with
the leaching reaction:
2 FeS04 + 0.5 02 + H2$04 -> Fe2(S04)3 + H20
At steady-state, the leaching and regeneration reactions yield the fol-
lowing net reaction:
FeS2 + 2.4 02 •*• 0.6 FeS04 + 0.2 Fe2(S04)3 + 0.8S
In order to maintain steady-state operation, the product sulfate must be
removed continuously from the system at the sul fate-to-iron ratio produced
which in molar terms is 1.2. At bench-scale, experimentation was per-
formed in batch mode and the product sulfate was not removed from the
' +2 +3
system until virtually all the Fe in the system was oxidized to Fe
(Y >0.90). Thus, the 0.6 moles of FeS04 product per mole of reacted pyrite
was virtually all oxidized to the ferric state. In experiments wbere
acid was added to the starting reagent the expected reaction is
0.6 FeS04 + 0.3 H2S04 + 0.15 02 -> 0.3 Fe(S0) + 0.3 H0
90
-------
with the product sulfate remaining soluble. The data in Appendix B, Volume
2 verifies this expectation (acidified runs). In the absence of added
acid, the ferrous sulfate oxidation would be expected to yield ferric
oxide according to the following reaction
0.6 FeS04 + 0.15 02 -*• 0.2 Fe2(S04)3 + 0.1 Fe203
with the product iron oxide remaining on the processed coal unless acid
washed. The data in Appendix A verifies this expected iron deposition
on the processed coal.
The above analysis is totally based on the iron-to-sulfate ratio in
the system with no regard to pH. While it adequately explains the iron
oxide deposition on the processed coal it does not explain the observed
simultaneous sulfate deposition when non-acidified starting reagents were
used. One explanation that may be advanced is occlusion of sulfate in the
deposited iron. A more probable explanation is the formation of insoluble
basic iron sulfate. Chemical analyses on the deposits revealed that the
iron-to-sulfate ratio in them was 1.5. According to the literature, the
most insoluble form of basic iron sulfate is 3 Fe20,-4 S03'9 H20 which has
the same iron-to-sulfur ratio as the deposit on the processed coal. Further-
more, phase diagram data on the iron oxide-sulfate-water system appears
to indicate that formation of the above basic iron sulfate could be expected,
in small quantities, at the L-R processing conditions utilized in the per-
formed experimentation when the starting reagent did not contain added acid.
Addition of sulfuric acid eliminates formation of the basic iron sulfate.
The total iron concentration should also have no direct effect on
the Meyers Process according to the formulated rate expressions, Equations
(3) through (6), except as it affects Y in operations where neither re-
generation nor reagent exchange are operative, e.g., Mixer and Settler
+3
Operations. In these operations, the portion of Fe present as Fe is
being irreversibly reduced to Fe+2. Thus, it is expected that for a given
Fe+3-to-Fe+2 ratio in the starting reagent solution, the higher the
starting Fe concentration in the reagent is per mole of pyrite present,
the higher the pyrite removal would be at any reaction time interval. The
91
-------
data in Appendices A through C verify this expectation. For example,
pyrite removal in the Mixer Operation from 33 wt. percent coal slurries,
after approximately one hour of processing, ranged from 9-12 percent when
5 wt. percent Fe reagent solution was used; 20 wt. percent coal slurries
with the same reagent processed for equivalent periods yielded 18-24 per-
cent pyrite removal. Also as expected, the experiments processed with
zero iron concentration in the starting reagent led to less than one percent
pyrite removal during approximately one hour of Mixer Operation (Appendix C
data). The data from the Settler Operation were analogous, but less pro-
nounced and more difficult to distinguish because pyrite removal based on
starting coal pyrite was very small in this unit operation and even small
noise in the data can easily mask effects on rates or on extent of pyrite
removal.
During L-R processing the total iron concentration of the reagent
should have no effect on Y, except at the start of the operation if the
starting L-R Y is different than 1.0. Since reagent regeneration depends
+2 +>2
on Fe concentration, the higher the Fe concentration in the starting
L-R reagent solution is, the higher would be the regeneration rate at the
+2
start of the operation. Once, however, the Fe ions in the starting
solution are regenerated, the regeneration rate depends solely on the
+2
Fe generation rate from the pyrite leaching reaction which does not de-
pend on total iron concentration. For example, a starting 33 wt. percent
coal slurry of 5.6 wt. percent Fe reagent with a Y of 0.4 contains 0.6
+2
moles Fe per liter; after 15 minutes of L-R reaction, an additional 0.4
+2
moles of Fe have been added to each liter of reagent solution from the
pyrite leaching reaction and yet, because of regeneration, less than 0.5
+2
Fe moles are present in the reagent solution at the end of this reaction
interval and only 0.35 moles at the end of 30 minutes reaction time even
+2
though approximately 0.7 moles Fe were generated by the leaching reaction
(see Experiment 32, Table D-l, Volume 2). Thus, it is apparent that the
+2
bulk of Fe present in the starting reagent was regenerated during the
first few minutes of L-R operation. Similar observations are made from
the L-R processing of 20 wt. percent slurries in 5 and 3 wt. percent iron
reagent; in these experiments the Y of the reagent increases even faster
because of the lower pyrite concentration on a per liter of reagent basis.
92
-------
Equation (5), the regeneration rate expression, and the Y versus t, R
data generated with starting reagent to which no iron was added (Table C-l
Volume 2) indicate that there is an Fe concentration below which its
effect on the Y values during L-R processing cease to be insignificant.
Yet the data in Figure 15 show that even in the "no added iron" case,
where the required,Fe+ for the initiation of the pyrite leaching reaction
is assumed to have been derived from non-pyritic iron in the coal, the
effect of Fe concentration on pyrite removal rates during L-R processing
was negligible. It is believed that the explanation is that the concen-
tration of iron in the reagent solution adjacent to the reaction sites is
substantially higher than that in the bulk and so is the Y if it is assumed
that oxygen diffusion to the vicinity of the reaction sites is faster than
iron ion diffusion to the bulk reagent solution. Assuming that this ex-
planation is valid, it is reasonable to assume that the data in Table C-l
could have been predicted by Equation (3) if the Y values near the reaction
sites were known.
Pyrite leaching by direct oxidation with oxygen was also considered as
an alternate explanation for the unpredictability of pyrite removal rates
during the "no added iron" experimentation. However, examination of the
data indicated that the reaction product mix was independent of the iron
concentration in the feed reagent (see Section 2.4.5.1 page 101). It was,
therefore, concluded that a change in reaction mechanism with no apparent
change in the process chemistry was very unlikely. It should be noted
that the role of iron in Meyers Process chemistry was unequivocally proven
from separate Teaching-regeneration experimentation.
In conclusion, starting reagent composition did not exhibit any ap-
preciable effect on pyrite removal rates during L-R processing, but it did
affect iron deposition on the coal during processing. The latter effect
was traced to insufficiency of sulfate ion correctable by the addition of
sulfuric acid. For the particular coal processed in this program, the
sulfuric acid requirement was estimated at 1.5 percent of the weight of
reagent used in 33 wt. percent coal slurries (maximum slurry concentration
investigated). Two weight percent sulfuric acid was actually used.
93
-------
90
20 WT % SLURRIES OF 100 MESH TOP SIZE L. K. COAL
80
70
O
(S)
3 60
O
Z 50
C£
O
40
Of
30
20
5 WT % Fe REAGENT (EXP. 7)
3 WT % Fe REAGENT (EXP. 6)
NO ADDED Fe REAGENT
(EXP. 44)
10
0.0
0.5
1.0
1.5
2.0
2.5
3.0
L-R REACTION TIME, HOURS
Figure 15. Effect of Total Iron Concentration on Pyrite Removal
During L-R Processing of Suspendable L.K. Coals
94
-------
2.4.2.7 Oxygen Partial Pressure Effects
Figure 16 illustrates the oxygen partial pressure effect on pyrite
removal from suspendable L.K. coal during L-R processing with data from
Table D-l, Volume 2. Equation (5), the regeneration rate expression,
predicts a nearly four-fold increase in regeneration rate when the oxygen
partial pressure is increased from 35 psi to 135 psi; the data in Figure 16
show that such an increase in regeneration rate translates to a nearly 70
percent increase in the L-R pyrite leaching rate at 120°C. Thus, the oxygen
partial pressure used during L-R processing has an appreciable effect on
pyrite leaching from suspendable L.K. coals. It is noted that the experi-
mental data in Appendix B, Volume indicate a substantially reduced effect
from that depicted in Figure 16. Analysis of the data revealed that the
apparent negligible oxygen partial pressure effect on L-R pyrite removal
was due to inconsistent regeneration. Comparison of experimental and pre-
dicted L-R Y values indicated that regeneration was efficient throughout
the experimentation at 50 psig, but inefficient during the early stages of
L-R processing at the higher pressures (refer to Section 2.4.2.2 for details
on comparison of predicted and experimental Y values during L-R processing).
2.4.3 Settler Unit Operation
In early program experimentation, approximately one-half of the slurry
at the conclusion of L-R processing was transferred to a surge tank where it
was permitted to react further at 90°C and ambient pressure while at the
same time the settling behavior of the coal was being examined; this
practice originated the title of this unit operation. However, laboratory-
scale tests on liquid-solids separation of suspendable L.K. coal slurries
in 5 wt. percent Fe reagent solution revealed that a settler operation was
not necessary even for 20 wt. percent slurries of 100 mesh top size coal.
The high pyrite removal rates from 33 wt. percent slurries of 14 mesh top
size coal assured that the process did not require a settler. However, the
practice of further processing part of the L-R processed slurry in an
ambient pressure reactor continued when it became apparent that L-R pyrite
leaching rates dropped substantially at pyrite removal levels exceeding 80
percent. The settler now became a stirred, tail-end reactor after the L-R
reactor.
95
-------
O
U
g
D-
O
5
Q
UJ
U
90
80
70
60
50
40
30
20
MESH TOP SIZE COAL PROCESSED AT 120°C AS 33 WT % SLURRIES
PSIG (135 PSIO2), EXP.
PSIG ( 85 PSI 02), EXP.
PSIG ( 35 PSI O,), EXP.
36
30
35
0.0 0.5 1.0 1.5 2.0 2.5
L-R REACTION TIME, HOURS
3.0
3.5
Figure 16. Effect of Oxygen Partial Pressure on L-R Processing
of Suspendable L.K. Coal
96
-------
Settler performance was monitored by sulfur forms analyses of coal
samples taken at the start and at the conclusion of the approximately 20
hour operation and by the ferrous ion production during the operation.
The percent of the raw coal pyritic sulfur removed during the "settler"
operation is easily determined from the rate data listed in Appendices A
through C, Volume 2; it is the difference in the S removal values listed
for "L-R processed coal" and "L-R + settler processed coal". In the
majority of experiments pyrite removal in the Settler Operation ranged
between 5 and 10 percent of the pyrite in the raw coal or 20 to 40 percent
of the pyrite present in the coal at the start of this unit operation. As
expected, pyrite removal was nearly zero when the experiment's starting
coal slurry was prepared with iron-free reagent (data in Table C-l, Volume
2)-
Pyrite removal in the Settler Operation varied because the unit's
starting slurry composition varied. Since the reagent was neither ex-
changed nor regenerated in this unit operation, Meyers Process chemistry
dictates that the quantity of pyrite expected to be leached from the coal
will be proportional to the quantity of ferric ion present in the reagent
of the feed slurry. The rate of removal is governed by Equation (3).
Since the operating conditions of this unit (temperature, pressure, re-
action time) and the Y value of the feed slurry were very nearly the same
in the majority of experiments which included the Settler Operation, the
differences in pyrite removal in this operation among these experiments
should be principally due to the different values of the reagent Fe to coal
pyrite ratio of the feed slurry. The larger this ratio, the higher should
be the pyrite removal in the "settler" obtained in the pertinent experiments
listed in Appendices A through C, Volume 2. Thus, the less concentrated
slurries with the highest Fe content should have yielded the highest pyrite
removals after 20 hours of settler processing, and vice versa. The latter
expectation is certainly validated by comparing the low iron settler data
contained in Appendix C to corresponding data from any of the experiments
listed in Appendices A or B. Less extreme comparisons between Experiment
12, Table C-l, which was performed on 20 wt. percent coal slurry, and
Experiments 13 through 15, Table C-2, which were performed on 33 wt. percent
slurries, validated the above expectations, also. Analyses of the processed
97
-------
coals show that 47 percent of the feed pyrite to the settler was removed
in Experiment 12, while the corresponding pyrite removals from Experiments
13, 14, and 15 were 17, 14, and 21 percent, respectively.
In general, the experimental data derived from "settler" processed coal
were in qualitative agreement with the predicted behavior based on the
chemistry of the Meyers Process and on Equation (3). However, quantitative
comparisons were substantially less successful as the data in Table 12 in-
dicate. Table 12 compares percentages of pyrite removal during the Settler
Operation computed from (a) sulfur forms analyses of the feed and product
coals and (b) the quantity of ferrous ion produced in the "settler",
assuming that the chemistry of the Meyers Process applies to this unit
4.9
operation (10.2 moles Fe produced from each mole of pyrite reacted).
TABLE 12. PYRITE REMOVAL IN THE SETTLER OPERATION
Experiment Nos.
% Pyrite Removed Based on
(a) Coal Analyses
(b) Fe+2 Production
23
10
32
28
14
29
29
17
28
30
47
32
31
13
20
33
52
23
Obviously the agreement between the two sets of pyrite removal values is
poor, but the difference is equivalent to only 0.1 to 0.2 percent pyritic
sulfur in coal which is not too far in excess of expected deviations in
the chemical analyses involved. Note that each pyrite removal computation
from coal analyses (set (a) in Table 12) required two coal sample analyses
for pyrite content, settler feed and settler processed coal samples, each
of which contained less than one percent pyritic sulfur with allowable
deviation 0.1 percent. On the other hand, Fe+2 production was determined
from large changes in Fe+ concentration (100 to 200 percent change in Fe+2
concentration in the reagent solution, depending on slurry concentration)
and, therefore, should yield a more reliable pyrite removal valve. Using
equation (3) and a KL value of 0.07 (hours)"1 (VJJ"1 for 90°C processing
of 14 mesh top size coal, estimated from the Arrhenius plots in Figure 12,
predicted "settler" pyrite removal values were computed for the experiments
98
-------
in Table 12. The predicted values ranged from 30 to 36 percent showing
relatively good agreement with the pyrite removal values derived from
ferrous ion production.
From the above analysis of the "settler" data it can be concluded that
the Settler Operation exhibited compliance to the formulated Meyers Process
chemistry and obeyed the developed leaching rate expression with rate con-
stant values which are derivable from the ambient pressure A, and E,
values computed in Section 2.4.1.
2.4.4 Coal Washing Operation
Several processed coal washing schemes were investigated during the
early stages of the program in an effort to identify an efficient scheme
for washing deposited stil fates from coal processed with reagent which did
not contain added sulfuric acid. Multiple stage water washing (3 to 4
stages with water equal to twice the weight of coal per stage), expanded
wash times (0.5 to 1.0 hours), and washing with spent reagent did not
materially reduce the deposited sulfate. The deposited sulfate could
only be removed by agents which simultaneously removed the deposited iron
oxide; for example, a single stage coal washing with 1 N sulfuric acid
solution followed by two stages of water wash dissolved completely both
deposits. This observation combined with Meyers Process Chemistry pre-
dictions on iron deposition led to the conclusion that the deposited
sulfate was either occluded in or chemically bound with the deposited
iron (basic iron sulfate) and, therefore, it could not be removed without
the simultaneous dissolution of the iron oxide (see Section 2.4.2.6 for
details).
As predicted, the addition of 2 wt. percent sulfuric acid to the
starting reagent solution helped to maintain large majority of the process
iron and sulfate products in solution and inhibited to a large extent the
formation of water insoluble sulfates. Thus, coal processed with acidified
reagent required washing principally for the removal of reagent retained
by the coal after slurry filtrations; hot water wash accomplished this
task effectively as the data (acidified runs) in Appendices A through C,
99
-------
Volume 2 and Table 18 below indicate. However, the problem of sulfate
deposition on processed coal was not completely solved with the addition
of the 2 wt. percent sulfuric acid to the starting reagent. At long L-R
processing reaction times the residual sulfate on processed and washed
coal remained unacceptably high (see data in Tables B-l and B-2, Volume 2).
Extrapolation of phase diagram data on the Fe^-SOg-^O system available
in literature appear to indicate that higher sulfuric acid concentrations
may be required in the starting reagent than the 2 wt. percent utilized
in this program when L-R reaction temperatures exceed 110°C. Efforts
to reduce residual sulfate sulfur on processed coal to below 0.2 wt. per-
cent continue. It should be noted that coal leaching with (a) low iron
concentration reagent, (b) for less than 3 hours, and (c) at temperatures
below 120°C does not present the sulfate deposition problem.
A two-stage wash scheme was adopted as adequate for washing coal
processed with acidified reagent (note exceptions in previous paragraph).
Each stage consisted of a slurry wash and a cake wash and for each wash
a water-to-dry coal weight ratio of 2 was employed. During reagent re-
cycle testing the two-stage wash scheme was preceded by a brief cake wash
with a quantity of hot water equal to 50 percent the estimated weight of
the dry coal. The collected cake wash filtrate was then used as part of
the make-up reagent solution and, thus, reagent recycling was more
complete.
Table 13 summarizes Wash Section data derived from the washing of L.K.
coals processed under a variety of experimental conditions; the data are
typical of the results obtained in this unit operation. The data in the
third column of Table 13 represent the estimated quantity of reagent iron
on the wet coal after filtration of the L-R or Settler reactor slurries.
The quantity of spent reagent retained by the processed coal after filtra-
tion was estimated from the difference in weights of the wet filter cake
and of the dry product coal plus recovered sulfur residue; the iron was
then computed from the measured iron concentration in the filtrate. The
fourth column lists the Y value of the spent reagent solution retained by
the coal determined from iron analyses on the filtrates. Columns 5 through
100
-------
TABLE 13. TYPICAL WASH SECTION DATA FROM L-R PROCESSED L.K. COALS
Experiment
No.
12 (AG)*
17**
25***
30 (AG)****
Coal
Top
Size
(Mesh)
100
100
14
14
Estimated
Reagent Fe
on Coal
(gin)
15.2
21.8
43.2
31.4
Reagent
Y
0.85
0.84
0.96
0.56
First staoe Wash Filtrates
Slurry Wash
Fe (gm)
12.7
17.7
35.4
23.0
Y
0.81
0.69
0.82
0.41
Cake Wash
Fe (gm)
2.8
3.5
7.0
4.8
Y
0.81
0.43
0.80
0.42
Second stage wash nitrates
Slurry Wash
Fe (gm)
0.1
0.7
0.8
1.1
Y
0
0.03
0.16
0.11
Cake Wash
Fe (gm)
0.0
0.2
0.3
0.4
Y
-
0
0.15
0.12
Recovered
Fe
(Percent)
102.6
106.0
100.7
93.3
* Coal processed as a 20 wt. percent slurry in 5 wt. percent Fe reagent solution for 5 hours at 120°C
and 100 psig; the slurry was also Settler processed.
** Coal processed as a 33 wt. percent slurry in 5 wt. percent Fe reagent solution for 2 hours at 120°C
and 50 psig.
*** Coal processed as a 20 wt. percent slurry in 5 wt. percent Fe reagent solution for 6 hours at 120°C
and 100 psig.
**** Coal processed as a 33 wt. percent slurry in 5 wt. percent Fe reagent solution for 2 hours at 120°C
and 100 psig; the slurry was also Settler processed.
-------
12 show the total iron and the forms of iron (Y value) recovered from
each wash. The last column shows the iron recovered in the Wash Section
as a percentage of the iron in Column 3.
The data in Table 13 show that better than 95 percent of the reagent
iron on the processed coal was recovered in the first wash stage and that
the remaining was completely recovered in the second stage; the sulfate
recovery values should be analogous. The low Y values in the filtrates
of the second stage washes indicate that the iron salt dissolved in this
stage was principally in the ferrous state which could mean deposited
ferrous sulfate; however, the quantities of iron salt present in these
filtrates are insignificant in terms of sulfate deposition of the coal.
In fact, the quantity of iron present in the second stage wash filtrates
renders the need for a second wash stage questionable.
2.4.5 Elemental Sulfur Recovery Operation
The product elemental sulfur from the experiments listed in Appendices
A through C was recovered by leaching the processed and washed coal with
toluene. However, special proof-of-principal experimentation was performed
on elemental sulfur recovery by hexane and by vaporization in inert gas.
The three techniques and the data obtained from them are discussed below in
separate sections.
2.4.5.1 Elemental Sulfur Recovery with Toluene
The procedure used to recover the product elemental sulfur during L-R
experimentation was described in Section 2.2. In outline, the wet coal
from the Wash Section was slurried in twice the estimated dry coal weight
toluene, the water on the coal was azeotroped off at 85°C, the toluene
slurry was refluxed at 108°C for 30 minutes and filtered hot, and the
filtrate evaporated at low temperatures to recover the dry sulfur residue.
This procedure constituted a single stage sulfur recovery operation; in the
majority of experiments the above procedure was repeated (two stage sulfur
recovery).
102
-------
The sulfur residues from either stage analyzed to between 80 and 90
percent by weight elemental sulfur depending on the attained degree of
dryness; a small portion of the residue was carbonaceous material. The
weight of residue obtained from the first stage recovery was approximately
nine times that recovered from the second stage for virtually every experi-
ment; thus, 90 percent of the recovered elemental sulfur was leached from
the coal in the first stage. The quantity of elemental sulfur recovered
in each experiment is listed in the sulfur balance section of the mass
balance data tables contained in Appendices A through C, Volume 2.
According to Meyers Process chemistry, formulated on the basis of data
generated during the ambient pressure leaching of coals with iron salt
solutions, coal pyritic sulfur during leaching with iron salts is oxidized
to sulfate sulfur (S ) and elemental sulfur (S ) at the molar ratio of 1.5.
This ratio was computed from the quantity of ferrous ion produced per mole
of pyrite removed from coal and from the sulfate sulfur produced in a small
number of carefully mass balanced coal extractions with ferric chloride.
On the basis of this ratio, 40 percent of the pyritic sulfur removed from
the coal should be recovered as elemental sulfur. Actual Sn recovery from
these ambient pressure experiments, after 60 minutes reflux of the pro-
cessed coal in toluene, was between 80 and 90 percent of that expected from
the $s/Sn ratio of 1.5. Conversely, the S /S ratios derived from the quantv
ties of pyrite removed and elemental sulfur recovered ranged in value from
1.8 to 2.1. This discrepancy in the S /Sn values was attributed to incom-
plete S recovery principally due to operational losses rather than leaching
inefficiency. In fact, it was concluded that an 80-90 percent recovery of
product elemental sulfur represented a reasonable expectation from a bench-
scale unit operation which involves a numbers of steps and yields 20 to 40
grams of product from 8 kilograms of slurry.
Simultaneous coal Teaching-reagent regeneration does not permit the
determination of S /S ratios from ferrous ion production. Thus, elemental
sulfur recovery was used to determine, at least through comparison with
ambient pressure data discussed above, the validity of Meyers Process
103
-------
chemistry during L-R processing of L.K. coals. Percent sulfur recovery
and the S /S ratio were computed for each experiment. Percent $n recovery
was computed from Equation (13)
S recovered
% S recovered = x 100 (13)
n
x S removed
on the assumption that the Ss/Sn ratio of 1.5 was valid. The Ss/$n values
were computed from Equation (14)
S removed S recovered
S /S = — £ - ~-JL-A - 04)
°s' n S recovered x
n
on the assumption that the quantity of recovered $n represented the entire
yield of S in the experiment (100 percent Sn recovery). Table 14 sum-
marizes the results.
The data in Table 14 show appreciable scatter, but the range of values
is nearly identical to the range of the corresponding values obtained from
ambient pressure processed coal (one 30 minute toluene extraction was
performed in the first six experiments; therefore, Sn recovery was
approximately 10 percent lower than it was in the rest of experiments).
The correspondence of ambient pressure and L-R processing S recovery data
permits at least the Inference that the actual S /S ratio during L-R pro-
cessing is also 1.5. The data scatter appears to be the same within each
group of experiments as it is among groups. Since each group of experi-
ments was performed under different experimental conditions, the 1.5 value
of the Ss/$n ratio evidently applies to the entire spectrum of experimental
conditions used in this program. It may be argued that the sulfur recovery
from coal processed at 130°C, Experiments 20, 21, 34, and 40, and that from
coal processed with iron-free starting reagent, Experiments 42, 44, and 45,
points to a higher $s/Sn ratio than that applicable to the rest of the
experiments. At least for Experiments 42 through 45 such conjecture does
not appear valid on the basis of sulfate recovery. Because in these
experiments the starting reagent solution did not contain iron sulfate, it
was possible to estimate the sulfate product of the pyrite leaching reaction;
in all four experiments the Sg value pointed to a S /S ratio of 1.5 (the
actual values were 1.7 + 0.1).
104
-------
TABLE 14. ELEMENTAL SULFUR PRODUCT RECOVERY FROM THE L-R
PROCESSING OF U.K. COAL
Exp.
Nos.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
42***
44***
45***
*
**
***
100 Mesh Top Size Coal
% Sn if
Recovered* v~
75.9
77.3
77.2
69.8
76.2
74.5
92.7
80.7
89.3
73.5
74.4
95.5
90.0
85.1
78.9
88.0
84.5
85.6
(65.5)
70.7
80.1
70.0
72.4
72.0
s/snj**
2.29
2.23
2.23
2.58
2.28
2.36
1.70
2.10
1.80
2.40
2.36
1.62
1.78
1.94
2.17
1.84
1.96
1.92
2.54
2.12
2.57
2.45
2.47
14
Nos!
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Mesh Top Size Coal
Recoverec
83.6
89.2
85.2
(71.4)
86.8
85.8
84.8
87.8
86.8
73.1
(60.8)
79.9
83.0
84.9
71.2
81.4
70.2
68.1
43.7
Percent of expected product Sn, the latter computed
of pyrite removed and S /S = 1-5.
Ss/Sn value computed from the quantities of pyrite
covered .
Coal processed with
lifc" * C* **/
1.99
1.80
1.93
1.88
1.91
1.94
1.85
1.88
2.42
2.13
2.01
1.95
2.51
2.07
2.56
2.67
(Coal leaching was
performed with
def earning agents)
from the quantity
removed and Sn re-
iron-free starting reagent.
105
-------
The data in Table 14 proves conclusively that L.K. coal pyrite leaching
during L-R processing, at least within the parametric ranges investigated,
always yields the two sulfur products specified by the Meyers Process
chemistry, sulfate sulfur and elemental sulfur. In addition, these data
strongly suggest that the reaction mechanism under all leaching conditions
examined to date, including ambient pressure processing, is the same, even
though the process may become diffusion limited at some point. This obser-
vation lends additional validity to the single rate expression applicability
concept already used for describing pyrite leaching rates from suspendable
L.K. coal regardless of processing conditions.
Table 14 does not include data from Experiment 43 because toluene
leaching was performed subsequent to vacuum drying of the processed coal
rendering the recovered S value unreliable. The data from Experiment 22
are nearly identical to that derived from Experiment 41 shown in the table.
In these two experiments wetting and defoaming additives were used which
apparently distorted the sulfur analyses performed on the processed coal
rendering meaningful data interpretation impossible; it is believed that
the unusually low $n recovery indicated that the pyrite removal during these
experiments was over estimated due to errors in sulfur forms analyses (see
also Tables A-7 and B-7, Volume 2). The values in parentheses indicate in-
complete recovery of Sn due to traced residue loss, the weight of which
could not be accurately determined.
Toluene retention on the product coal was also determined through
specially designed experimentation. Approximately 500 grams of dry coal
were wetted with 50 percent its weight water and refluxed for 60 minutes in
500 grams of toluene in a well trapped apparatus; the water was azeotroped
off prior to reflux. The toluene slurry was then filtered in a closed
system and the toluene-wet coal cake was vacuum dried at 100°C overnight
in a well trapped oven. A careful mass balance was performed on the coal,
water, and toluene from the precise weights of feed and product coal and
liquids. The experiment was repeated with twice the weight of feed mate-
rials. The mass balances from both experiments were virtually identical
and indicated a 0.9 percent increase in the weight of coal, a 0.3 percent
106
-------
increase in the weight of water, and a 2 percent decrease in the weight of
toluene. Thus, the total system mass balance shows that 0.9 percent of
toluene was not accountable. In terms of coal mass balance, the coal re-
tained 0.9 percent of the toluene; in terms of toluene mass balance it
retained 1.8 percent of its weight toluene (note that the water gain
represents approximately 0.2 percent toluene). Thus, our estimate of
toluene retention on processed coal is one percent of the coal weight.
2.4.5.2 Elemental Sulfur Recovery with Hexane
-------
temperature did not furnish meaningful improvements. The data in Table 15
illustrate the relative $n leaching efficiencies of hexane and toluene.
These data show that only a small percentage of the elemental sulfur in the
coal was recovered in the two hexane leaching stages preceding the toluene
stage. It is also noted that in three of the four batches of coal less
sulfur was recovered in the first stage than in the second; this observation
appears to be consistent with the contention that hexane's inefficiency was
due to its inability to displace water from coal at reasonable leaching
times.
2.4.5.3 Elemental Sulfur Vaporization
Although the atmospheric boiling point of sulfur is 444.6°C, which is
somewhat above the decomposition temperature of coal, an appreciable vapor
pressure (sufficient for engineering design of equipment for removal of
sulfur by vaporization) begins about 250°C (Figure 17). At 350°C (still
below the thermal decomposition point of coal), the vapor pressure of
sulfur is substantial M50 mm Hg). Therefore, it seems reasonable that
elemental sulfur could be removed from coal by vaporization either in
inert gas stream such as steam or nitrogen or under vacuum, provided that
the sulfur does not extensively react with coal during the thermal process.
An initial demonstration of the removal of pyritic sulfur from coal,
via ferric ion oxidation, was performed utilizing vaporization for
removal of elemental sulfur. In this case, vaporization was effected at
100-120°C and 150 mm of Hg pressure with no detectable coal decomposition
or elemental sulfur reaction with coal. In fact, greater than 99 percent
pure sulfur was obtained as condensate on a heat-exchanged surface. How-
ever, vacuum vaporization in commercial process equipment is thought to be
relatively expensive, so that the possibility of vaporizing sulfur from
coal at atmospheric pressure and in an inert gas stream was investigated
(Table 16).
A coal sample was treated with ferric sulfate under the usual atmos-
pheric leaching conditions but neither water washed to remove iron sulfate
from the coal pore structure nor extracted with solvent to remove elemental
108
-------
TABLE 15. ELEMENTAL SULFUR RECOVERY BY SUCCESSIVE STAGES OF
HEXANE-TOLUENE LEACHING
L-R Processed Coal Filter Cakes
Sample A
riOOO Gr. Dry Coal)
Wash
1/2
Washed Coal (Wet)
-1/2 ,
3:1 Sol vent-To-Coal 4:1
Extractions
1st
Hexane
2nd
Hexane
Sulfur
Out,Gr
0.74
1.10
Toluene 3.24
TOTAL SULFUR
RECOVERED:* 5.08
Sulfur
Out,Gr
1st
Hexane ~0.0
2nd
Hexane
0.50
Toluene 3.21
3.71
f"
Sample B
(~600 Gr. Dry Coal)
I
Wash
Mashed Coal Wet
1/2
3:1 Sol vent-to-Coal 4:1
Extractions
Sulfur
Out,Gr
1st
Hexane 0.29
2nd
Hexane 0.31
Toluene 1.38
1.98
Sulfur
Out.Gr
1st
Hexane 0.13
2nd
Hexane 0.31
Toluene 1.57
2.01
Estimated overall S recovery efficiency >90 percent in all cases.
109
-------
bO
W
to
LLJ
300° 400°
TEMPERATURE, °C
Figure 17. Vapor Pressure of Sulfur
110
-------
TABLE 16. VAPORIZATION OF ELEMENTAL SULFUR FROM FERRIC SULFATE LEACHED COAL*
Sample
No.
1
2
3
4
5
6
Water
Washed
0
X
X
X
X
x.
Toluene Thermal % w/w Sulfur
Extracted Treatment Sulfate Organic
0-
X
X
0
0
0
0
0
220°C/15mm/17 hrs
Hg
200°C/3 hrs
250*0/2 hrs
350°C/2 hrs
0.34
0.13
0.05
0.03
0.01
0.01
1.18
0.52
0.51
0.92
0.63
0.56
% Weight Loss
During Vaporization
-
-
2.7
^/
1.7
1.2
2.8
200cj TOO mesh x 0 Lower Kittanning Seam coal was extracted twice with 2.5a 1,N Fe2($04)3 solution, total
residence time 22 hrs. Cursory water wash to remove surface iron sulfate; dried at 110°C under vacuum
to remove moisture and as little volatile elemental sulfur as possible. The coal was subsequently
split into lOg samples and subjected to the treatment indicated in columns 2 through 4 (X indicates
treatment, 0 indicates no treatment). The organic sulfur content of the ROM coal was 0.41 +_ 0.04 wt.
percent.
-------
sulfur (Sample No. 1). Drying was effected at relatively low temperature
in an attempt to remove water but not vaporize large amounts of the elemen-
tal sulfur which had been formed. Another sample (No. 2) was washed with
water and toluene to remove sulfate and elemental sulfur, respectively.
The decrease in "organic" sulfur between samples 1 and 2 (0.66 percent w/w)
is a measure of the elemental sulfur which had remained adsorbed on the coal
after a cursory drying. Another sample, (No. 3), was further treated
under vacuum at 200°C to remove any additional' residual sulfur. There
was no further decrease in organic sulfur indicating that a maximum of
the elemental sulfur had been removed by the toluene extraction; however
there was a decrease in the sulfate content.
Temperature Programmed Thermogravimetric Analysis studies show that
iron sulfates quantitatively decompose at 450-600°C in an inert atmosphere
giving a residue of iron oxide. Apparently, this degradation occurs at
somewhat lower temperatures under vacuum and in the presence of a reducing
coal environment. Note that the coal weight loss during vacuum treatment was
only 2.7 percent of which 0.67 percent could be ascribed to elemental sulfur.
Three samples were contacted with an inert gas stream Cargon) in
small ceramic boats inserted into a Burrell tube furnace at the specified
temperatures and residence times (Nos. 4-6). It can be seen that increasing
the temperature of vaporization from 200PC to 350°C increases the quantity of
sulfur vaporize. A vaporization temperature of 250-350°C is indicated for
use in process design. Note the the weight losses during vaporization did
not exceed 2.8 percent (including the loss due to volatilization of sulfur)
indicating that sulfur can be relatively selectively vaporized from coal.
The method utilized for contacting the coal with the inert gas stream
was rather primitive compared to that which can be obtained in commercial
size equipment, so that residence times should be substantially reduced.
2.5 SOLID-LIQUID SEPARATION UNIT OPERATIONS
The Meyers Process scheme utilized in the current program involved
three solid-liquid separation unit operations: (!) coal separation from
112
-------
spent iron sulfate solutions, (2) coal separation from spent wash water,
and (3) coal separation from spent toluene or hexane. All three operations
were performed by vacuum filtration with conventional, 4 liter size labora-
tory apparatuses. The size of the system did not lend itself to quantita-
tive evaluation of parametric effects on these unit operations, nor was there
a guarantee that such evaluations would be meaningful to a scaled-up
operation using commercial equipment; thus, a systematic study of these
unit operations was not undertaken. Qualitatively, filtration efficiency
in terms of rates and dryness of the filter cake increased with decreasing
liquid density, as expected for constant filter cake depth. Thus, toluene-
coal slurries filtered faster than water-coal slurries and the latter
appeared to filter faster than spent reagent-coal slurries. Also, while
the toluene content of the filter cake was approximately 30 percent of the
dry coal weight, the water content of spent reagent and water filter
cakes was estimated to be in the 35-40 percent range. The effects of coal
top size and slurry concentration in the ranges of 100 to 14 mesh and 33
to 20 wt. percent, respectively, were not easily discernable. It appeared
that the 33 wt. percent slurries of 14 mesh top size coal filtered at a
higher rate than 20 wt. percent slurries of 100 mesh top size coal, but an
effect on filtration rate was not apparent when only one of the above
parameters differed. Filtration rates and the moisture content of the
filter cake appeared unaffected by variations in L-R processing parameter
values.
The feasibility of conducting the described solid-liquid separations on
a scaled-up process by commercially available equipment and the type of
equipment recommended were determined by outside vendors. Bird Machine
Company representatives performed bench-scale tests at the TRW site on
actual process slurries and concluded that they can be separated by avail-
able commercial equipment at reasonable rates even at solids concentrations
as low as 17 percent. They recommended centrifugal separation for the
organic solvent coal slurries and either centrifugal separation or rotary
drum filtration for the aqueous coal slurries depending on concentration.
Details on these evaluations are included in a Bird Machine Company memo-
randum to TRW reproduced in Appendix F, Volume 2 of this report. Envirotech
113
-------
Corporation, Salt Lake City, Utah, arrived at similar conclusions using
simulated Meyers Process slurries of 100 mesh top size L.K. coal.
2.6 COAL DRYING OPERATION
The toluene-wet coal from the elemental sulfur recovery operation was
vacuum dried at 100CC for approximately 16 hours (overnight). This was the
only coal drying procedure used in this program. The efficiency of this
type of drying operation was determined from the special experimentation
performed on processed coal toluene retention described in Section 2.4.5.1.
Selection of drying equipment for the process scheme presented in the
process design section of this report was based on information furnished by
outside vendors.
2.7 PRINCIPAL CONCLUSIONS FROM BENCH-SCALE DATA
In this report section an attempt is made to summarize the principal
conclusions drawn throughout the discussion of the data derived from the
L-R processing of suspendable Lower Kittanning coals by the Meyers Process.
The emphasis in this summary is placed on information and conclusions vital
to the full scale process design presented in Section 5. Of course, the
most important information derived from the data generated in this program
was the conclusive proof that L-R processing (simultaneous coal Teaching-
reagent regeneration) of L.K. coals up to at least 14 mesh top size was
both feasible and desirable. The expected advantages of L-R processing -
combination of two unit operations into one, elimination of the need for
solid-liquid separations during leaching, and higher pyrite leaching rates -
were attained without any detectable adverse effects on the composition or
physical properties of the processed coal. Furthermore, these advantages
were augmented by the attained pyrite leaching rates which, under relatively
mild temperature and pressure conditions (1100-130°C5 50-150 psig), were
higher than expected.
114
-------
The performed experimentation and resulting data suggested that the
L-R processing scheme may be defined by a process train comprised of the
following main unit operations connected in series: mixer, L-R reactor,
ambient pressure reactor, coal-reagent separation, processed coal wash,
coal-water separation, elemental sulfur recovery, and coal drying. The
last two unit operations become a single operation when elemental sulfur
is recovered by vaporization. Elemental sulfur recovery by organic solvent
extraction requires an additional solids-liquid separation operation.
Multiple stage processing may be desirable or necessary for certain unit
operations (e.g., L-R reactor or coal wash).
In addition to the described process train, the L-R processing scheme
requires one or more unit operations in the reagent recycle loop for sulfate
and trace element recovery. It is probable that these unit operations
would best be employed on process slip streams rather than the entire re-
agent recycle stream. The generated data were insufficient for the defini-
tion of the sulfate recovery unit operations.
In terms of processing function, the identified unit operations may be
grouped into four process sections: the reactor section, the wash section,
the elemental sulfur recovery section, and the sulfate recovery section.
The reactor section - comprised of the mixer, L-R reactor, and ambient
pressure reactor - is the core section of the Meyers Process scheme for
L-R processing; therefore, it was the most thoroughly investigated of the
four process sections.
The three reactors were operated in a batch mode with respect to coal
slurry, but oxygen was continuously fed to the L-R reactor. Reactor slurry
composition homogeneity was assumed at all times (well-mixed reactors).
Coal pyrite leaching took place in all three reactors with the bulk of the
pyrite being removed in the L-R reactor, as intended. Throughout the
leaching process and regardless of processing conditions coal pyrite was
oxidized to sulfate sulfur, S , and elemental sulfur, Sn> in accordance with
s
Meyers Process chemistry. The SJSn ratio in the process sulfur product
s n
was estimated to be equal to 1.5 regardless of processing conditions.
f $*? . •'
115
-------
Pyrite leaching rates in all three unit operations of the reactor
section were adequately described by the empirical rate expression
'L '
where W is the concentration of pyrite in coal at reaction time t in
weight percent and Y is the ferric ion-to-total iron ratio in the reagent
solution at the same reaction time. The validity of Equation (3) was
proven from bulk measurements of Y and Wp during the leaching of both 100
and 14 mesh top size L.K. coals under a variety of processing conditions.
Bulk Y measurements were inadequate for use in Equation (3) only when the
total iron concentration in the reagent was well below one percent (e.g.,
in experiments performed with iron-free starting reagent with the process
iron being derived from the coal).
In the temperature range investigated, 70° to 130°C, the reaction con-
stant of Equation (3) depended only on temperature and coal particle top size.
Between zero and 80 percent pyrite removal the temperature dependence of
K. could be expressed by the Arrhenius-type equation
KL = AL exp (-EL/RT). (4)
Beyond approximately 80 percent pyrite removal the K, value appeared to
be virtually insensitive to temperature change in the range of 102°-130°C;
obviously, Equation (4) did not apply to this residual pyrite region.
Apparently, a change in the rate occurred when a certain minimum pyrite
concentration was reached in the coal. It is believed that this apparent
change in mechanism is due to the high ash content of the particular ROM
L.K. coal used in this program. Indirect evidence for the validity of
this statement is furnished by the coarse coal data in Figure 21, Section
4 of this report, where ROM and cleaned L.K. coal data are compared. These
data indicate that a substantial reduction in coarse coal pyrite leaching
rate occurred when pyrite removal from the ROM coal samples reached 70 per-
cent, but an equivalent drop in rate was not observed with the cleaned
coal. For process design purposes, rate constant values specific to the
80-95 percent pyrite removal region were estimated from the experimental
rate data.
116
-------
For reasons which could not be delineated by the available data, two
different AL values were required in Equation (4) to express the K depen-
dence on temperature during mixer processing and L-R processing. Possible
explanations are offered in Sections 2.4.2 and 2.4.2.3.
For reagent regeneration during L-R processing, the rate expression
developed in the previous bench-scale program with synthetic spent reagent
proved valid. The data, however, suggested that the activation energy of
the regeneration reaction may be lower for actual spent reagent solution
regenerated in the presence of coal than that computed from synthetic
spent reagent studies. This apparent catalytic effect could not be quan-
tified from the available data. Qualitative analysis of the data revealed
that neglecting this catalytic effect on the regeneration rate would not
significantly influence the engineering analysis of the process.
In summary, the empirical rate expressions derived during the current
and previous bench-scale programs can be used with confidence for the de-
sign of the reactor section for depyritizing suspendable ROM Lower Kittan-
ning coals (higher K, values may be applicable to cleaned L.K. coals).
The following rate expressions and rate constants are applicable to each
of the three unit operations:
Mixer:
rL = KL Wp Y2 and KL = AL exp (-EL/RT)
with AL = 3.4 x 105 (hours)"1 (W J'1 for 100 mesJi x 0 L.K. coal
AL = 2.7 x 105 (hours)"1 (Wp)"] for 14 mesh x 0 L.K. coal
E, = 11.1 kilocalories per mole pyrite reacted for both
top size coals
The mixer is the unit operation where the coal slurry is prepared and
brought to the desired L-R processing conditions (temperature and pressure),
Slurry residence time in this unit depends largely on the time required to
117
-------
wet the feed coal. Wetting of Lower Kittanning coals was accomplished by
refluxing their freshly mixed slurries for approximately 30 minutes. Thus,
slurry residence time in the mixer should be less than one hour.
For a given top size coal, pyrite removal in this unit operation depends
on the square of Y and, therefore, on any parameter that affects Y. Since
reagent regeneration or exchange does not take place in this unit operation,
pyrite removal is highest when the Fe+3-to-pyrite ratio in the slurry is
highest (high Fe+3 starting reagent concentration, low coal concentration
in the slurry). Typical pyrite removal values for L.K. coal were: 20
percent with 20 wt. percent slurries and 5 wt. percent Fe reagent; 10 per-
cent with 33 wt. percent slurries and 5 wt. percent reagent; approximately
one percent with slurries prepared from iron-free starting reagent.
L-R Reactor (no°-130°C, 50-150 psig, 14-100 mesh coal top size):
r, = K. W_Y where Y * 1-Fe /Fe
L L P
rR = KRP0 (pe+2)2 wnere KR = AR exP (~ER/RT)
with AR = 6.7 x 105 Uters/mole-atm-mi mites
and ER = 13.2 kilocalories per mole of Fe+2 oxidized
The above expressions operate simultaneously on the process in the L-R
unit operation. The value of KL to be used in the leaching rate expres-
sion depends on temperature and coal top size up to approximately 80
percent pyrite removal, on Wp for a short transition period, and it is
probably a pure constant with respect to changes in processing parameters
in the 80-95 percent pyrite removal region. Thus,
KL = AL exp (-EL/RT) for wp >_ 1.6 wt.%
with AL (14 mesh) = 7.4 x 105 (Hours)'1 (% Pyrite in Coal)'1
AL (100 mesh) = 8.9 x 105 (Hours)'1 (% Pyrite in Coal)'1
EL (both top sizes) = ll.l kilocalories/mole pyrite removed;
118
-------
also KL = (Wp) = Wp - 1.1 (Hours)"1 (% Pyrite)"1 for 1.6 >W > 1.2 wt.%
and KL = 0.1 (Hours)"1 (% Pyrite in Coal)"1 for W _ KL i 0.07 (Hours)"1 (% Pyrite)"1 for 90-1Q2°C
The available data did not permit the determination of precise KL values
or the KL temperature dependence, if any, for temperatures lower than
9Q*C 1n this unit operation, The value of 0,07 was estimated from
"settler" data generated at 90"C. The 0,10 value 1s equal to that de-
termined for pyrlte leaching from 14 mesh top size L.K. coal at 102°C
1n the mixer where the reaction appeared to be klnetlcally controlled,
but it is also equal to the estimated value of the diffusion rate KL
determined applicable in the 110°-130°C temperature range after 80-85
percent of pyrite has been removed from coal.
This unit operation serves a dual function: (a) the last few percent
of coal pyrite is removed in it and (b) ferrous ion is being generated
which can subsequently be used to recover the iron and sulfate products
of the process as ferrous sulfate, since ferrous sulfate is less soluble
than ferric sulfate.
Figure 18 presents typical process design curves and illustrates the
predominance of the L-R unit operation on the overall performance of the
Meyers Process. It is noted that under L-R processing, the parametric
effects on the individual rates and differences in pyrite removal in the
mixer are substantially tempered because of the simultaneous influence of
119
-------
80
70
O
U
Q
Z 50
O
Q
LLJ
to
CO
i
40
30
20
10
SUSPENDABLE LOWER KITTANNING COALS
1. 33 WT. % SLURRY, 130°C, 100 PSIG
2. 20 WT. % SLURRY, 120°^ 100 PSIG
3. 33 WT. % SLURRY, 120°C, 100 PSIG
1.0 2.0 3.0
L - R REACTION TIME, HOURS
Figure 18. Typical Process Design Curves
4.0
5.0
120
-------
the leaching and regeneration rates on process performance in the L-R
reactor. Each curve in Figure 18 represents both 14 and 100 mesh top size
coals; thus, the coal particle size effect for coal top sizes up to 14
mesh is undetectable under L-R processing. Also, the temperature effect
on the leaching rate is less pronounced. Finally, the slurry coal con-
centration in the 20-33 wt. percent range had virtually no effect on
overall process performance even though twice as much coal per unit time
was processed when 33 wt. percent slurries were employed as when 20 wt.
percent slurries were used. The effect of pressure (oxygen partial pres-
sure) on overall process performance was also small under L-R processing.
The apparent process insensitivity to narrow parametric changes, even
though individual rates are substantially affected by similar changes in
the same parameters, is the result of compensating effects brought about
when the individual leaching and regeneration rates operate simultaneously
(independent variables become interdependent).
The data in Figure 18 were derived from the rate expressions and rate
constant values presented above. For up to 80 percent pyrite removal, data
from Appendix D, Volume 2 were used. Appendix D contains similar data for
every experiment performed with suspendable coal in the current program.
Data in this Appendix are tabulated for up to 90 percent pyrite removal
but, at least for the particular L.K. coal used in this program, are only
valid up to 80 percent pyrite removal. Note, also, that the data in Appen-
dix D and Figure 18 apply to well mixed reactors only (batch operation).
The Meyers Process design information summarized in this section should
only be used with confidence to predict pyrite removal from ROM, run-of-the-
mine, Lower Kittanning coals. Limited data on cleaned Lower Kittanning coal
(Section 4), Illinois No. 5 coal,1 and Upper Freeport seam coal appear to
indicate that the empirical pyrite leaching rate expression, Equation (3),
derived from the study of ROM L.K. coal may also be applicable to these
coals provided empirical KL values are determined for each coal. That the
KL values for different coals could be substantially different is dramati-
cally illustrated in the next section where pyrite removal from L.K. and
Upper Freeport coals is compared. It is estimated that the Upper Freeport
coal KL is at least an order of magnitude larger than that of L.K. coal.
121
-------
2.8 L-R PROCESSING OF UPPER FREEPORT SEAM COAL
A sample of low pyrite ROM Upper Freeport seam coal was available
from previously performed screening tests (pyrite leaching under ambient
pressure) as part of the Survey Program on the Meyers Process.^ ' The
coal contained 1.7 wt. percent pyrite (0.9 wt. percent pyritic sulfur)
and thus it was considered a good candidate for testing the effectiveness
of L-R processing on coal with low pyrite content. Testing of this coal
was principally motivated by the substantial reduction in the value of the
leaching rate constant observed during the L-R processing of ROM L.K.
coal when its pyrite content was reduced by approximately 80 percent to
1.2 wt. percent (the pyrite concentration in the ROM L.K. coal was 7.3 wt.
percent).
The available 60 mesh top size ROM Upper Freeport coal was slurried in
four times its weight 65°C ferric sulfate solution containing 5 wt. percent
iron and immediately processed in the identical manner that L.K. coal was
processed. Figure 19 summarizes the data obtained from the L-R processing
of Upper Freeport coal. Note that approximately 82 percent of the pyrite
in coal was removed during t , 1.5 hours slurry residence time in the mixer;
at this point the pyritic sulfur content of the coal was only 0.16 wt. per-
cent. An additional 1.5 hours of L-R processing at 120°C increased the
overall pyrite removal to 93 percent and reduced the pyritic sulfur content
of the coal to 0.07 wt. percent. Obviously, pyrite removal from Upper
Freeport coal by the Meyers Process appears to be efficient down to virtually
zero pyrite.
In Figure 20 pyrite removal data from nearly identical processing of
Upper Freeport and L.K. coals are being compared. It is obvious that pyrite
removal from the Upper Freeport coal takes place at higher rates than
those for L.K. coal. Since only a single rate experiment was performed with
the Upper Freeport coal, it was not possible to determine if pyrite leaching
from this coal could be expressed by Equation (3), the empirical leaching
rate expression developed from the L.K. investigations, or if a different
122
-------
90
80
70
o
5
tu
Q£
Qi
D
LL
u
Qi
Q.
60
50
40
30
20
60 MESH TOP SIZE COAL
65 - 120°C
— t
m
120°C, 100 PSIG
'L-R
_L
J.
0.0 0.5
1.0 1.5 2.0
REACT ION TIME, HOURS
2.5
3.0
figure 19. Upper Freeporc Coal Leaching with 5 Wt. % Fe Reagent
123
-------
l\5
•£>
SUSPENDABLE COAL LEACHING WITH 5 WT % Fe REAGENT SOLUTION
20 WT % ROM COAL SLURRIES
1 UPPER FREEPORT COAL (60 MESH TOP SIZE)
2 L. K. COAL (14 MESH TOP SIZE)
120°C, 100 PSIG
f,
L-R
J_
3.0 4.0 5.0
REACTION TIME, HOURS
6.0
7.0
8.0
Figure 20. Pyrite Leaching Rates from Upper Freeport and L.K. Mine Coals
-------
expression was required. If, however, it is assumed that Equation (3)
applies to Upper Freeport coal, then the data in Figure 20 appear to indi-
cate that the K, value for the Upper Freeport coal is between one and two
orders of magnitude higher than the equivalent KL value for the L.K. coal.
125
-------
3. REAGENT RECYCLABILITY-TRACE ELEMENT DATA
A total of 31 L-R experiments were performed utilizing a single batch
of recycled reagent. Twenty-nine of these experiments are summarized in
Appendices A and B of Volume 2 (Experiments 7 through 11, 13 through 15,
17 through 19,and 23 through 40) and provided for 91 hours of L-R pro-
cessing exposure under a variety of operating conditions as specified in
the Appendices. An additional 10 hours of L-R processing time was accumu-
lated on the recycled solution during special experimentation. Recycled
reagent processing exposure over the 101 hours of L-R processing time
includes 380 hours of settler processing at 90°C and ambient pressure (for
one half of the reagent in nearly all experiments), and 46 hours of mixing
and heat-up time prior to L-R processing. Approximately 70 percent of
recycled reagent exposure time was accrued subsequent to addition of 2 weight
percent H^SO. to the reagent. During the reagent recycle experiment!'on
nearly 77 kilograms of dry coal were processed. An average of 18 percent
reagent makeup was required to compensate for evaporative losses incurred
during processing as well as reagent losses associated with water washing
of the leached coal. Aqueous filtrate from the first stage coal wash was
utilized in the preparation of make up reagent for the majority of the
experiments.
Comparison of the data generated from experiments performed with re-
cycled and fresh reagent indicate no apparent degradation in the activity
of the reagent with age or quantity of coal processed, (e.g., compare data
from Experiments 23, fresh reagent, and 31, reagent recycled for 40 hours
under L-R conditions). A more conclusive proof of the stability of reagent
activity during recycling is the fact that the data generated at the
beginning and at the end of the reagent recycle experimentation are equally
predictable by Equation (3). It is believed that the reagent recyclability
tests performed in this program were of sufficient time duration and
spanned the range of potential processing conditions of the Meyers Process
adequately enough to permit the conclusion that reagent solution recycl-
ability is viable.
127
-------
The reagent recycle investigation also furnished a means of verifying
earlier conclusions (Contract No. 68-02-0647) that the Meyers Process
removes, in addition to pyrite, non-pyritic mineral matter usually present
in coal in minor or trace quantities. This extensive recycling of reagent
provided for the build up of trace elements in the reagent and, therefore,
furnished the potential of analyzing the reagent for these elements with
increased accuracy and of attempting a cursory mass balance of each element
in conjunction with coal analyses.
The majority of the elements selected for analysis either have been
proven to be hazardous pollutants or are recognized as being potentially
hazardous. Included in these categories are As, Be, Cd, Cr, Cu, Li, Mn,
Ni, Pb, V and Zn. Other elements were selected because of their corrosion
promoting tendencies during processing and/or combustion or because they
were previously shown to be amenable to leaching by the Meyers Process.
Analyses were performed on starting and processed samples of both
coal and iron sulfate reagent. The detailed description and evaluation of
the analytical procedures used can be found in Reference 2. Briefly, all
coal and reagent samples were analyzed by atomic absorption (AA) spectro-
scopy for Be, Ca, Cd, Cr, K, Li, Mg, Mn, Na, Ni, Pb, V and Zn. Arsenic
was also determined utilizing a special spectrophotometric procedure. The
chloride content of the leach solution was determined using a specially
modified silver chloride turbidimetric procedure.
An NBS certified coal (Standard Reference Coal No. 1632) was analyzed
concurrently with the selected Lower Kittanning samples as a check on the
accuracy of the analytical system. The results are presented in Table 17.
A comparison of NBS certified values with analyzed values shows good
to excellent agreement for As, Be, Cr, Cu, Mn, Pb, V and Zn as was the case
with coals tested earlier and reported in Reference 2.
128
-------
TABLE 17. TRACE ELEMENT ANALYSIS ACCURACY VERIFICATION
(Concentrations are Shown as ppm in Dry Coal)
ro
10
NBS
As
Be
Cd
Cr
Ca Li Mn Ni
Pb
Zn
Certified 5.9 +_ 0.6 (1.5) 0.19 + 0.03 20.2 + 0.5 18+2 - 40+3 15 + 1 30 + 9 35 + 3 37 +_
Analysis ~" ~ ~ _ _ _
Analysis of NBS Sample in Duplicate:
6.1 2 2
5.4 2 4
16
16
17 12 33 25
17 13 33 28
Zinc determination performed on a plasma ashed sample.
Note: Elements in parenthesis are not certified and are for reference only.
28 24 42*
24 30 35*
-------
One sample of starting coal and one sample of processed Lower Kittanning
coal were subjected to duplicate trace element analyses. The processed
coal sample was derived from Experiment 31. In this experiment 14 mesh
top size coal underwent 3 hours L-R processing at 120°C and 20 hours
"settler" processing at 90°C with reagent solution which had already been
recycled for approximately 40 hours at 120°C and for nearly 200 hours at
90°C. One starting reagent sample and two recycled reagent samples were
also analyzed for trace element content. One of the recycled reagent
samples was drawn after 98 hours of L-R processing (87 hours at 120°C and
11 hours at 110°C). The second sample was drawn at the conclusion of the
reagent recycle investigations (101 hours L-R processing); it differed
from the first by 3 hours L-R processing at 130°C. The second sample was
analyzed principally to determine if trace element leaching from coal was
enhanced at 130°C. The data from these analyses are summarized in Table 18.
In general, the data in Table 18 show that the decrease in coal trace
element concentrations during processing corresponds directly to trace
element build up in the recycled iron sulfate leach solution. Elements
which apparently are not subject to leaching by the Meyers Process include
Be, Cd, K, Li, Na, Pb, and V. Elements appearing to be most susceptible
to leaching are As, Ca, Cu, Mn and Zn, with at least 50 percent removal
being observed for each. It is also noted that the Cr and Ni concentrations
of both coal and reagent solution increased during processing; apparently,
this was due to a slight corrosion of the stainless steel apparatus used
in these experiments. A similar occurrence is observed with Mg which can
not be explained on the basis of available data. It is assumed to be the
result of errors in analyses and probably in the analysis of the starting
coal which was not performed in duplicate.
An attempt was made at mass balancing the trace elements which exhib-
ited the highest percent removal from coal (As, Cu, Mn, and Zn). The mass
balance was performed through the detailed bookkeeping of liquids and solids
entering and leaving the process during each experiment performed with re-
cycled reagent. The quantities of trace elements removed from the coal
processed in each experiment were computed from the values listed in column 7
130
-------
TABLE 18. COAL AND REAGENT TRACE AND MINOR ELEMENTS
Element
As
Be
Ca
Cd
Cl
Cr
Cu
K
Li
Mg
Mn
Na
Ni
Pb
V
Zn
Trace Element Concentrations in PPM
Starting
Coal*
13 16
- 2 2
1590
3 1
NA NA
39 36
27 27
4030 -
37 37
946 -
26 25
637 -
48 43
17 18
55 50
30 26
Processed
Coal*
5 5
2 2
7.6 7.6
1 1
NA NA
55 59
7 7
3170 3160
42 37
4171 4414
8 8
734 753
50 50
22 25
50 55
13 14
Starting
Reagent
1
0.06
5
0.07
37
6
0.09
0.6
NA
6
28
4
6
0.44
NA
0.78
Recycled Reagent
(98 hrs. at
110-120°C)
19 (31)**
0.9
318
0.14
175
30
64 (66)
1.0
NA
238
30 (88)
36
154
0.65
NA
52 (46)
Recycled Reagent
(101 hrs. at
110-130"C)
19 (27)
0.6
293
0.14
no
30
58 (57)
0.7
NA
224
32 (76)
20
136
0.58
NA
52 (40)
Estimated
Element
Removal
(Mo/Kg Coal)
10 + 2
0
1582 + 159
1 +0.3
ND
ND
20 + 3
865 + 360
0
ND
18 + 3
0
ND
0
0
14 + 4
**
Duplicate analysis performed for most elements.
Values in parentheses represent the expected trace element concentrations In the
reagent based on coal analyses and on the quantity of coal processed with the
same batch of raagent (adjusted for make-up).
NA - Not analyzed.
ND - Not determinable from available data.
131
-------
of Table 18. In turn, these values were computed from the analyses per-
formed on the starting coal and the processed coal from Experiment 31
(columns 2 and 3, Table 18). The assumption was made that the quantity of
trace element removal per unit weight of coal processed was the same for
all experiments and equal to that determined for Experiment 31; that is,
L-R reaction time beyond one hour of processing and changes in slurry
concentration, pressure, and temperature had no effect on trace element
removal (note, however, that only a limited number of experiments were
performed at temperatures, pressures, and slurry concentrations different
from those of Experiment 31). The trace element bookkeeping process began
with the batch of pure reagent solution used as the feed to the reagent
recycle tests and ended 30 experiments later when the reagent recycle
investigations were completed. Trace element build up in the reagent
during each experiment was computed by adding to the known reagent com-
position at the start of the experiment the quantity of trace elements
leached from the coal during the experiment; the new reagent solution
composition was appropriately adjusted for trace element and liquid re-
movals (losses) during transfers and coal wash and for additions through
reagent make up.
As a result of the above bookkeeping operation, it was possible to
compute the expected concentration of the selected trace elements in the
reagent solution as a function of reagent recycle time. In Table 18, com-
puted trace element concentrations (values in parentheses, columns 5 and 6)
are compared with the corresponding trace element concentrations obtained
from direct trace element analyses of the recycled reagent solutions.
Reasonably good agreement is seen for As, Cu, and Zn, but not for Mn. The
reason for the discrepancy between computed and analyzed Mn concentrations
could not be discerned from the available data; it is believed to be due
to the analyzed high Mn concentration in the starting reagent solution
which, if erroneously high as believed to be, would make the computed con-
centrations higher than should actually be. The agreement between computed
and analyzed trace element concentrations for three out of four mass
balanced trace elements adds substantial credence to the values in column 7
of Table 18. These latter values are not only important to the design of
132
-------
a product recovery system for the Meyers Process, but also to the evalu-
ation of the environmental benefits derived from the utilization of the
Meyers Process for the desulfurization of a particular coal.
Comparison of the data in columns 5 and 6, Table 18, appears to in-
dicate that there was no significant increase in the leaching of trace
elements when L-R processing temperature was increased from 120°C to 130°C.
Both, the analyzed and computed trace element concentrations lead to the
same conclusion.
Calcium, one of the minor elements in coal, was also substantially
leached from coal during processing. Because of the high quantity of
calcium leached from the coal, relative to trace elements, and because of
its low solubility in the reagent, this element was periodically removed
from the recycling reagent solution as a precipitate. Approximately one
gram of white crystalline precipitate was recovered from the reagent
solution per kilogram of coal processed. Emission spectrographic analyses
of six precipitates revealed that Ca was the only cation present in the
precipitates in larger than trace quantities. The anion of the precipitates
was not determined, but in all probability it was sulfate. One precipitate
was analyzed in duplicate for trace cation content; the data are presented
in Table 19.
TABLE 19. TRACE CATION CONTENT OF CALCIUM PRECIPITATES RECOVERED
FROM SPENT REAGENT
Sample
Trace Cations Present in the Calcium Precipitate, ppm
As Be Cd Cr Cu Li Mg Mn Ni Pd Si Sr V Zn
A
B*
<1 <0,5 5 10 8 <1 5 12 67 7 10 50 <1 9
<1 <0.5 8 14 6 <1 5 12 67 <1 10 30 <1 <0.2
r*
Sample B is a duplicate of A
133
-------
Attempts to mass balance Ca using the procedure followed for trace
element mass balancing described earlier were not very successful. Less
than 25 percent of the Ca removed from coal, estimated from before and
after processing coal analyses, could be accounted for by the recovered
precipitates (assumed to be CaSO.'ZHpO) and by the build up of dissolved
Ca in the reagent solution. This was not completely unexpected since a
small amount of precipitate could have deposited on the coal during pro-
cessing and have been leached out and not recovered during coal wash (note
that the quantities involved are 3-4 grams of precipitate per kilogram of
coal subjected to washing).
134
-------
4. PROCESSING OF COARSE COAL
Coarse coal can easily be separated from the leaching solution thus
L-R processing (simultaneous coal Teaching-reagent regeneration) is not
necessarily advantageous to coarse coal unless it improves substantially
depyritization rates over those obtained during ambient pressure processing.
The substantial increases of pyritic sulfur leaching rates with increasing
temperature observed with suspendable coal slurries suggested that coarse
coal pyrite removal rates may indeed improve substantially under L-R pro-
cessing. However, a brief experimentation with 3/8 inch x 0 L.K. coal
revealed that L-R processing of coarse coal results in only marginal gains
in rate over those attainable at 102°C under ambient pressure processing
(separate Teaching-regeneration) as the data in Section 4.1 indicate.
In general, L-R processing of coarse coal was similar to that described
for suspendable coal. The major difference being that coarse coal was pro-
cessed as a semi-fluidized bed while suspendable coal was processed as a
well mixed, partially circulating slurry. In order to accommodate this
difference, the existing suspendable coal processing reactor was modified
by the addition of a reinforced wire mesh basket and lid to support and
confine the coal during reagent circulation. Two additional thermocouples
were installed to monitor internal coal bed temperatures. The coal bed
density was approximately 0.7 kg/liter. Also, the reagent circulation
rate was decreased to 3 liters per minute (as determined by indirect
measurement). Procedural changes were as follows: (1) slurry was charged
into the reactor through the reactor top flange opening, (2) only reagent
composition was monitored during L-R processing since the circulated re-
agent contained less than 1.5 percent solids (these solids being non-
representative of the whole coal), and (3) reactor drainage consisted of
reagent removal through the reactor bottom while the retaining basket and
coal were removed through the reactor top.
All additional processing (water washing, elemental sulfur extraction
and coal drying) was carried out in a manner identical to that empToyed
during suspendabTe coaT processing.
T35
-------
A major difficulty encountered during coarse coal processing was main-
tenance of uniform reagent flow throughout the coal bed. To ensure uniform
flow, the retaining basket sides were removed so that coal-reactor wall
contact prevented channeling around the basket. Additionally, process coal
weights were reduced from the planned 4 to 5 Kg range to a 1.5 to 3 Kg
range. Verification of uniform flow was made by positioning thermocouples
within the coal bed and evaluating their response to temperature changes
in the circulating reagent (slow response was taken to indicate regions of
low flow).
4.1 L-R PROCESSING DATA FOR COARSE COAL
Three coarse coal leaching experiments were performed under pro-
cessing conditions using acidified reagent. These consisted of two
experiments of 8 hours each and a single experiment of 2 hours. Settler
processing was not employed during any of these experiments. One 8 hour
experiment (Experiment 46) consisted of processing approximately 3200 grams
of coal using 2 weights of reagent per weight of coal while the other
(Experiment 47) consisted of processing nearly 1900 grams of coal using 4
weights of reagent per weight of coal. In the 2 hour experiment (Experi-
ment 48), approximately 1700 grams of coal were processed using 4 weights
of reagent per weight of coal. Processed coal analyses for these experi-
ments are presented in Table 20. The analyses are seen to be internally
consistent, although Experiment 46 results seem to indicate slight iron
oxide deposition (probably due to slightly inadequate reagent acidification),
All three experiments resulted in rather low pyrite removals. Pyrite
removals during mixing (tm)', calculated on the basis of ferrous iron pro-
duction, were 6, 15 and 14 percent for Experiments 46, 47 and 48, respec-
tively. The low percentage pyritic sulfur removal observed during t in
m
Experiment 46 was due to a low starting reagent Y value (0.65) and the
high coal-to-reagent ratio used in this experiment (0.5); the other two
experiments were performed at a coal-to-reagent ratio of 0.25 and at
starting Ys >0.9.
136
-------
TABLE 20. L-R PROCESSING OF 3/8 INCH x 0 L.K. COAL IN 5 WT. % FE REAGENT*
AT 120eC AND 100 PSIG (85 PSI 02)
Experiment
No.
Starting
Coal
46
47
48
Coal Com
Ash
28.91
+ .52
31.16
29.40
30.27
Heat
Content
Btu/Lb
10,448
± 106
10,246
10,638
10,465
osition Wt. % (Except Heat Content), Dry
Total
Sulfur
St
4.28
+.14
2.86
2.82
3.16
Pyri ti c
Sulfur
SP
3.56
+_.18
2.11
2.27
2.50
sulfate
Sulfur
Ss
0.35
+ .03
0.25
0.22
0.14
Organic
Sulfur
So
0.37
+.03
0.50
0.33
0.52
Iron,
Fe
3.68
±-13
3.21
2.67
2.84
% Sp Removal
Based on
Analyzed
SP
41
36
30
Based on
Corrected
SP
-
37
38
26
GO
The reagent was acidified with 2 Wt. %
-------
Comparable 102°C data for 3/8 inch x 0 L.K. coal are not available.1
However, 102°C data are available for the 1/4 inch x 0 L.K. coal used in
the previous bench-scale program. These data are presented in Table 21.
This 1/4 inch x 0 L.K. coal had a starting pyritic sulfur content of 3.48
percent and attained 13 percent pyrite removal during mixing and heat up.
High reagent Y values were maintained by frequent reagent exchange.
TABLE 21. PYRITIC SULFUR REMOVAL FROM 1/4 INCH x 0 L.K. COAL AT 102°C
Process Time (hrs)
2
4
6
8
10
Average Y Value
'vO.S
^0.9
M).9
^0.9
*J0.9
Pyrite Removal, Percent
32
46
55
62
65
A comparison of the 3/8 inch x 0 L.K. coal data with the 1/4 inch x 0
L.K. coal data indicates that the average pyritic sulfur removal rate
during the first 2 hours of processing was slightly lower for the 3/8 inch
x 0 coal than for the 1/4 inch x 0 coal. Average reagent Y values were
approximately equal for both experiments during this time. Between 2 and
8 hours of processing, however, the average pyritic sulfur removal rates
are seen to be substantially higher for the 1/4 inch x 0 L.K. coal (102eC
processing) than for the 3/8 inch x 0 L.K. coal (120°C processing) despite
the lower pyrite content of the 1/4 inch x 0 coal (reagent Y values being
nearly equal for both experiments during this time). Thus, the pyrite
leaching rate constant of 3/8 inch x 0 L.K. coal is lower at 120°C than
the rate constant of 1/4 inch x 0 L.K. coal at 102°C. This seems to in-
dicate that pyrite removal rates quickly become diffusion limited during
processing of coarse coals.
While the benefits of elevated temperature processing of run-of-mine
coarse coal appear to be minimal, it will be shown in the following sections
that 102°C processing data indicate potential for application of the L-R
processing mode to specific coarse coal size fractions.
138
-------
4.2- CLEAN COAL GRAVITY FRACTION PROCESSING
The relative rates of desulfurization of run-of-mine fine L.K. coal
(100 x 0 mesh) and coarse L.K. coal (1/4 inch x 0) are compared in Figure
21 along with the pyrite removal rate curves of narrow size cuts of clean and
run-of-mine coarse L.K. coal, illustrating the enhancement of rate effected
by cleaning coal. These data on coarse coal were generated with 100-500
grams 1/4 inch x 0 Lower Kittanning coal (or size fractions therefrom)
slurried in iron sulfate solution (5 wt. % Fe, 15-20 wt. % solids). The ROM
8 x 14 mesh coal fraction is seen to have lower pyrite removal than 1/4
inch x 0 coal for all times prior to 40 hours leaching, at which time both
their pyrite contents and pyrite removal rates are seen to become nearly
equal. Upon cleaning the 8 x 14 mesh size fraction (float-sink removal of
20 percent of the total coal), however, its pyrite removal rate closely
approximates that of 1/4 inch x 0 coal during the first 9 hours of process-
ing. After the initial 9 hours of processing, pyrite removal rates of the
cleaned 8 x 14 mesh coal are seen to approach that of a much finer mesh
coal. Coal analyses from these experiments are presented in Table 22.
The results show that: (1) clean coal reacts substantially faster than
ROM (raw) coal, (2) coarse coal (in this case, a cleaned narrow size
fraction) can be desulfurized to near zero pyrite (0.09 percent pyritic
sulfur), and (3) the desulfurized clean coal has 0.92 percent total sulfur
at a 14736 Btu/lb heat content, which is very close to the Federal Stan-
dards for New Stationary Sources (0.89 percent sulfur at this heat content).
TABLE 22. CHEMICAL REMOVAL OF PYRITIC SULFUR FROM CLEANED 8 x 14 MESH
LOWER KITTANNING COAL AT 102°C
Coal
Whole Coal
Sink (1.75) (20.4%)
Float (1.75) (79.6%)
Float - 48 hr leach
Whole - 48 hr leach
•^••••^^^^^^^^^^^^^^^^^^^^^•H
% W/W SULFUR
Total
4.43
13.7
2.02
0.92
1.39
—-n—H^^^^^^^M
Pyritic
3.37
12.4
1.06
0.09
0.57
•HBl^^MM^H
r Sulfate
0.43
1.22
0.23
0.03
0.04
•^•^^•••"""•i
Organic
0.68
0.04
0.73
0.80
0.78
••••••••••
Ash
20.11
71.6
6.92
5.38
14.88
•••^^•^••i
BTU
12066
13899
14736
13257
mmm^mmm
139
-------
8 12 16 20 24 28 32 36 40 44 48
REACTION TIME, HOURS
Figure 21. Pyritic Sulfur Removal from ROM and Cleaned L.K. Coal
at 102°C
140
-------
4.3 PRELIMINARY DATA ON COARSE COAL PROCESSING BY SIZE FRACTION
Typical Sp removal values from 1/4 inch x 0 coal are presented in
Table 23 along with the ratio of the effective leaching rate constants
measured for -100 mesh and 1/4 inch x 0 Lower Kittanning coal with iden-
tical reagent Y and coal pyrite values. It is interesting to note that
this ratio does not remain constant with extent of pyrite removal; thus,
the leaching rate constants of various top sizes of the same coal cannot
be related by a single proportionality constant. This observation can be
attributed to the fact that a given top size coal is comprised of a wide
range of coal particle sizes with each particle size being associated with
a different pyrite leaching rate constant.
TABLE 23. PYRITE LEACHING OF 1/4 INCH x 0 LOWER KITTANNING COAL AT 102°C
Reaction
'ime, Hours
4
6
24
48
Y
Avg. Value
^0.8
^0.9
^0.8
^0.9
% Sp
Removal
50
57
75
81
i
KL 100 Mesh x 0 KL 1/2 Inch x 0
2.7
3.0
6.0
9.6
A sample of 1/4 inch x 0 Lower Kittanning coal was riffled into sev-
eral narrow size fractions. Pyrite removal data for 100 x 0, 14 x 28.
and 4 x 8 mesh fractions are shown in Table 24. The pyritic sulfur removal
rates differed substantially, but in all cases high pyritic sulfur re-
movals were obtained. While these size fractions are all observed to
react at acceptable pyrite leaching rates (from a commercial processing
standpoint) during 102°C processing, the 14 x 28 mesh and 100 mesh x 0
fractions are seen to react at significantly higher rates than the 4 x 8
mesh fraction. These pyrite removal rates may be taken as a relative
measure of the fraction of readily accessible pyrite which is associated
141
-------
TABLE 24. PRELIMINARY DATA ON THE CHEMICAL REMOVAL OF PYRITIC SULFUR FROM SIZE-FRACTIONS
OF LOWER KITTANNING COAL AT 102°C
Coal Fraction
Mesh
1/4"xO
4x8*
14 x 28*
i
TOO x 0*
wu
100
25
20
11
Run No.
_
1&2
354
_
2
3
_
2
1
1
2
••
2
2
1
1
Time
Hours
0*
24
48
0**
48
120
0**
12
24
48
72
0**
6.5
12
24
48
% w/w Sulfur Forms
Total
4.45
1.68
1.50
5.06
2.03
1.53
3.51
1.24
1.06
1.08
1.08
4.62
1.50
1.33
l.lc
1.'.3
Pyntic
3.48
0.93
0.66
3.95
1.29
0.57
2.46
0.53
0.30
0.30
0.18
2.62
0.25
0.15
0.05
0.05
Sulfate
0.50
0.12
0.14
0.43
0.09
0.14
0.40
0.07
0.06
0.08
0.10
1.02
0.24
0.23
0.25
0.25
Organic
0.47
0.63
0.70
0.68
0.66
0.82
0.65
0.64
0.70
0.70
0.86
0.98
1.01
1.00
0.89
0.93
Ash
20.84
17.02
15.54
25.44
21.86
20.30
16.56
11.82
9.13
13.86
12.30
20.79
13.66
13.00
12.40
12.55
titu
11822
12728
13023
10927
11970
12032
12643
13693
14129**
13258**
13533
'11543
13285
13321
13280
13178
t> Pyrite
Removal
0
73
81
0
67
86
0
78
* 88
' 88
93
0
90
9&
98
98
ro
* Size cuts of 1/4 inch x 0 coal
** Starting coal
***Sampling problem is suspected here
V
-------
with each size fraction. Thus, it is seen that the finer cuts of coarse
coal (those containing pyrite. which is largely of the accessible type)
would probably be receptive to elevated temperature processing and,
perhaps, benefit could be realized by processing these coarse coal frac-
tions in an L-R mode with the coarser cuts being processed at 102°C under
ambient pressure.
143
-------
5. PROCESS ENGINEERING
Process design studies for chemical removal of pyritic sulfur from
coal have indicated that the process may be laid out using a number of
alternative processing methods. Some of the variations which have been
tested and considered in preliminary engineering designs include the
following:
t Air vs oxygen for regeneration
• Coal top sizes from 1/4 inch to 100 mesh
• Leaching and regeneration temperatures up to 265°F (130°C)
• Leaching and regeneration in the same vessel and in separate
vessels
• Removal of generated elemental sulfur by vaporization or
solvent extraction
All of the above conditions are effective and their utilization involves
economic trade-offs. Because processing steps and equipment needed for
removing sulfur from fine or suspendable coal sizes (up to about 8 mesh top
size) are significantly different from those needed for coarser coal, they
are separately described, respectively, in Sections 5.1 and 5.2. A summary
of projected process economics for commercialization of both fine and
coarse coal processes is given in Section 5.3.
5.1 SUSPENDABLE COAL PROCESSING
Suspendable coal is coal of a small enough particle size that it may
be processed as a substantially uniform slurry with moderate mixing energy.
Although no sharp top size specification can be given, it appears that coals
with top sizes up to about 8 mesh may be classed as suspendable. Bench-
scale experiments were conducted using 14 mesh and 100 mesh top size coals
as representative of the suspendable type. Either of these sizes are often
referred to as fine coal to differentiate them from the coarse coal described
in Section 5.2.
145
-------
A conceptual full scale process design for the chemical removal of
pyriti-c sulfur from fine coal is described in three sub sections. Section
5.1.1 gives the design basis which relies heavily on the bench-scale ex-
perimental data, but also incorporates information provided by equipment
vendors and data obtained from the literature. Section 5.1.2 presents the
baseline design and equipment list with capital and operating cost estimates.
Section 5.1.3 summarizes the major trade-offs examined in arriving at the
baseline design and shows some of the cost and operating sensitivities of
the process.
5.1.1 Design Basis for Suspendable Coal
Processing coal to remove pyritic sulfur using aqueous iron sulfate
involves four major process sections, each containing several unit operations.
The reactor section which includes mixing and solution regeneration has
three main process requirements which are:
• Providing mixing and wetting of ground coal with the aqueous
ferric sulfate leach solution and raising the slurry to the
operating temperature and pressure.
t Providing the residence time and reaction conditions which
remove a nominal 95 percent of the pyrite orginally con-
tained in feed coal.
• Providing the residence time and reaction conditions which
regenerate the ferric sulfate solution from the spent iron
sulfate leach solution.
The washing section which includes several stages of coal washing and
coal dewatering has two main process requirements which are:
• Providing for contact of the leach solution-wet coal with a
minimum quantity of wash water to remove water soluble iron
sulfates.
146
-------
• Providing for separation of coal from the leach solution
and the wash water.
The sulfur removal section which removes both elemental sulfur and
excess water from the product coal has four main process requirements which
are:
• Providing conditions such as heat or solvent contact to
remove elemental sulfur from the processed coal.
t Providing the thermal environment necessary to reduce the
moisture level of the coal to the desired value.
• Providing the means for recovery of the elemental sulfur
for subsequent marketing, storage or disposal.
• Providing for separation of coal product from solvent,
if used.
The sulfate removal section which removes excess iron sulfate from the
recycle leach solution has four main process requirements which are:
t Providing for the removal of iron sulfate from the aqueous
spent leach solution by crystallization and/or neutralization.
• Providing for the recovery of wash water from the wash section
effluents.
• Providing for maintaining the correct acid level by neutral-
izing excess acid if required.
• Providing for separation of the by-product iron sulfate and
neutralization product from the recycle streams.
Specific information and data for the steps or operations which are
important to the process design are presented in the following paragraphs.
These data rely heavily on the information given in Section 2.7 but also
include qualitative observations made during the bench-scale experimental
efforts.
147
-------
Mixing - The present bench-scale effort has demonstrated that there is
a more critical aspect to the mixing operation than simply surface wetting
the particles and suspending them in the leach solution. Preparing the
slurry can be readily accomplished with mixing times of 15 minutes or less,
but it was found that severe foaming of the slurry will occur when it is
pressurized and raised in temperature. Based on the experience described in
Section 2.4.1, the mixing time for a high rank, high ash, dry coal should be
between 30 and 60 minutes at the normal boiling point of the solution if
subsequent foaming is to be avoided. Lesser times may be possible with
moist or low rank coal. The quantity of foam produced seems to decrease
with increasing coal particle size and to decrease with lower solids content
in the slurry. These are secondary parameters which should not be considered
of major importance in the process design.
Leach Reaction - The net overall reaction between pyrite and the ferric
sulfate leach solution is represented by:
FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 - - 10.2 FeS04 + 4.8 H2$04 + 0.8S (15)
AH = -55 Kcal/g-mole Fe$2 = -0.10 MM btu/lb-mole Fe$2 reacted
The reaction rate was found to have a second order dependence on both the
fraction of pyrite (or pyritic sulfur) in the coal and the fraction of the
total iron in the leach solution which is in the ferric ion form. The leach
rate at temperatures of interest is represented as follows :
where [Wp] = wt% pyrite in dry coal at time t.
[Y] = fraction of iron as ferric ion at time t. and
K, = leach rate constant (a function of temperature
L and Wp).
KL is independent of total iron concentration at least in the immediate
vicinity of 3 percent to 5 percent total iron. Physical considerations such
as increased solution density and viscosity and the limited solubility of
148
-------
ferrous sulfate in the ferric sulfate solution become increasingly important
to the design of the pyrite leacher when total iron concentration approaches
10 percent.
The leach rate constant as a function of temperature can be adequately
represented by
KL = AL x exp (-EL/RT)
where E, = 11,100 cal/mole,
R = 1.987 cal/ mole - °K,
T = temperature in °K, and
AL = a function of size, temperature and Wp-
For 14 mesh top size coal at atmospheric pressure and temperatures be-
tween 70°C and the solution boiling point (about 102°C), the value of A. is
2.7 x 10 for all values of Wp. At temperatures between 110°C and 130°C
under oxygen and steam pressure up to about 150 psig, the value of A. is
7.4 x 10 when Wp is large (above 1.6) and 2.7 x 10 when Wp is small (below
1.2). Additional refinement of the data in the transition region would be
desirable, but an adequate representation of the data can be obtained by a
linear decrease in AL from 7.4 x 105 for Wp = 1.6 to the AL value of 2.9 x 10
for Wp = 1.2.
For 100 mesh top size coal, the ranges of applicability for AL appear
to be the same, but its value is about 25 percent higher. Thus at low
temperature or low Wp, the value of AL is 3.4 x 105 while at higher temper-
ature and high Wp the value is 8.9 x 10 .
These leach rate constants have been defined only for the high pyrite,
high ash Lower Kittanning coal used in the bench-scale programs. They
should not be applied to other coals or coal seams. The Lower Kittanning
coals investigated in the bench-scale programs had starting values for this
coal of Wp between 6 and 8. The transition to a lower rate constant, thus,
occurs at about 75 percent to 80 percent pyrite removal (Wp M.6). Based
on a single test of a coal with a lower ash and low starting Wp (Upper
Freeport seam, Wp = 1.7), a high leach rate was found at least to 80 per-
cent or 90 percent removal (i.e., Wp about 0.2).
149
-------
Regeneration - The leach reaction produces both ferrous-suIfate and
sulfuric acid which must be processed for continuous recycle operation.
For each mole of pyrite reacted 9.6 moles of ferrous sulfate must be re-
generated to maintain the acid at a constant level. This gives by-products
for disposal of 0.2 moles of Fe2(S04)3, 0.6 moles of FeS04 and 0.8 moles
of elemental sulfur. Alternately, regeneration of 9.2 moles of ferrous
sulfate can be considered if some acid is neutralized to give by-products
of 1.0 mole of FeSO/,, 0.2 moles of I-LSO,, and 0.8 moles of elemental sulfur.
4 c. 4
The choice of the extent of regeneration should be made on the basis of the
by-product preference and economics within process design constraints.
The regeneration reaction is:
1.0 FeS04 + 0.5 H2S04 + 0.25 02 —* 0.5 Fe£(S04)3 + 0.5 H20 (18)
AH = -18.6 Kcal/g-mole FeS04 = -.0335 MM btu/lb-mole FeS04
If hydrolysis of a portion of the ferric sulfate to iron oxide should
occur as
Fe2(S04)3 + 3 H20 * Fe203 + 3 H2S04, (19)
then additional acid neutralization or regeneration of ferrous ion would be
required to remove the acidity produced from the hydrolysis reaction. The
extent of hydrolysis at temperatures below 250°F appears to be small, but
at higher temperatures there is some evidence of precipitation of ferric
oxide and possibly a low hydrate or anhydrous ferrous sulfate. The hy-
drolysis products and/or precipitates^formed at 265°F were found to redis-
solve slowly in ambient temperature spent leach solution and do not remain
as permanent products. No data was obtained above 265°F.
The regeneration rate was found to be second order in the molar con-
centration of ferrous ion over the range of ferrous concentration from 100
percent to less than 1 percent of the total iron. The rate is
+2
rR = "d[Fdt 3 = KR [pe+2]2 C°2]' (20)
150
-------
+2
where [Fe ] = concentration of ferrous ion, mole/liter,
[02] = oxygen partial pressure, atm, and
KR = 1.832 liters/mole-atm-hour at 248°F.
Over the range of temperatures studied (212°F to 265°F) the rate constant
was found to vary exponentially with temperature as
KR = 40.2 x 106 exp (-13,200/RT), (2])
which gives
Temp °F (°C) ^R I1ters/m°le~at;ni-noijr
212 (100) 0.74
230 (110) 1.18
248 (120) 1.83
266 (130) 2.79
The ferric sulfate regeneration rates were obtained under conditions where
oxygen in the form of minute air or oxygen bubbles was dispersed through-
out the ferrous sulfate solution. Thus, all of the solution was continually
saturated with oxygen at the partial pressure of oxygen present in the
regeneration gas. At bench-scale, the minute bubbles were formed by pumping
a portion of the liquid in turbulent flow (NR >3000) through a pipe whose
length was 50 or more times its diameter. Gas containing oxygen was added
to the liquid in an amount ranging from less than 1 percent to greater than
10 percent by volume at flow conditions. The method is very similar to
aeration equipment used to reduce the biological or chemical oxygen demand
of chemical plant effluent streams, except that ferric sulfate regeneration
is conducted at higher temperatures.
Separation - The major separation step requires treated, fine coal to
be separated from the spent leach solution. The four principal methods
which could be employed are hydrocyclones, centrifuges, filters and
thickeners. Suspendable coal has a large fraction of particles smaller
than 100 microns in diameter and in general hydrocyclones are not useful
for particle sizes below several hundred microns. Centrifuges would require
151
-------
very high power input and recycle rates to separate the coal from the leach
solution because of the fine particle size and the small liquid-solid den-
sity difference. Filtration is applicable, but for slurries less than 30
or 35 percent solids, the filter area requirements increase rapidly.
Typically, a 10 percent solids slurry needs more than ten times the filter
area needed for a 35 to 55 percent slurry. Thickeners have been used on
commercial scale to remove coal fines from water and other aqueous media.
Data for similar density solutions and coal sizes were reported in Reference
2
4. It was estimated that a thickener area of about 20 ft per ton/day of
coal with an edge depth of about 8 feet would provide an underflow with
greater than 35 percent solids and an overflow containing only a few tenths
percent (or less) solids when the feed contains 10 to 20 percent of 100
mesh coal. Since the thickener slurry can be maintained near the leach
solution boiling point, the time spent in the thickener could be used to
carry the leaching reaction to greater degree of completion and to re-
dissolve any solids formed during leaching/regeneration.
Filtration - The two important design values relating to filtration
are the filtration rate and the coal "moisture" content. These values are
not independent and are both highly dependent on the specific coal and its
properties. Generalized correlations reported in the literature were re-
viewed and a data point was obtained from a filter manufacturer for a bench-
scale slurry of the high ash, Lower Kittanning coal. The vendor report
given in Appendix F showed that rates equivalent to about 25 Ib of dry
2
coal/hr/ft were obtained with a 20 percent slurry of -100 mesh coal in a
5 percent iron leach solution. They projected a 60 percent increase in
rate for a slurry with a solids concentration of about 33 percent.
One reported correlation plots rate against a parameter which is the
product of the percent ash in the -200 mesh fraction times the square root
of the weight percent of the -200 mesh fraction. For the Lower Kittanning
sample tested in the above filtration test, the parameter has a value of
about 200:
(25% ash) x (67% of -200)1/2 = 205
152
-------
Figure 22 shows this measured point and the literature data with
extrapolations to a typical 14 mesh top size coal. The expected filtra- -
tion rate for a cleaned 14 mesh top size coal is expected to be near 200
lb/hr/ff in a 33 percent slurry and about 150 lb/hr/ft2 in a 20 percent
slurry. For a cleaned 100 mesh top size coal the filtration rate is ex-
pected to be about 100 lb/hr/ft2 for a 33 percent slurry and about 70
lb/hr/ft for a 20 percent slurry.
Data on the moisture content of the filter cake was taken by the
filter manufacturer at the same time as rate was measured. Appendix F
shows the projected cake moisture to be 40 to 45 percent. This is higher
than either reported in the literature (26 to 34%) or found in a typical
bench-scale filtration (about 32%) for the high ash, 100 mesh top size
coal. The literature values are for water wet rather than leach solution
wet cakes. The data for both water and leach solution (5 to 5.5% iron)
are summarized as follows:
parts of liquid.per 100 parts of dry coal
100 mesh 14 mesh
high ash high ash
Bench Scale
leach solution 45-50 45-50
water 35-40 35-40
Vendor Test
leach solution 65-80 not tested
Reference 5
water 35-50 15-25
In order to provide for an adequate amount of coal washing to remove
the sulfate leach solution it was decided that 50 parts of liquid per 100
parts of dry coal would be used for both leach solution and water on both
coal sizes.
5.1.2 Process Baseline Design
A block diagram of the Meyers Process as applied to coal of about 8 mesh
top size or finer is shown in Figure 23. The block diagram shows the main
153
-------
240
220
200
180
«* 160
I
"I. 140
LU
S 120
u
100
80
60
40
20
0
DATA PLOTTED FROM
FIGURE 12-30 OF
REFERENCE 5 AND
EXTRAPOLATED TO
LOWER RANGES
-14 MESH
8-10% ASH
\
LOWER KITT.
-100 MESH TEST
25% ASH
30 40
60
80 100
150 200
300
(% ASH IN -200 MESH) X (% OF -200 MESH)
Figure 22. Filtration Rate Correlation
154
-------
operations and the interconnections between each of the four process
sections. Before discussing the process flow diagrams and mass balance
in detail, the block diagram will be described to give a brief process
overview.
Reactor Section - ground coal,.with a nominal top size of 14 mesh, is
mixed with hot recycled iron sulfate leach solution. After wetting is
complete at the solution boiling temperature, the slurry is introduced
into a vessel where the majority of the pyrite reaction is accomplished at
elevated temperature and pressure. Oxygen is simultaneously added to re-
generate the leach solution. Heat of reaction is removed and is used to
reheat the recycle leach solution. The slurry is passed to a secondary
reactor operated at atmospheric pressure and near the solution boiling
temperature, where the remaining pyrite reaction occurs.
Wash Section - The iron sulfate leach solution is removed from the powdered
coal in a series of counter current flow contactors and separators. The
slurry from the secondary reactor is first filtered and the cake is washed
on the filter. Both the filtrate and wash liquids are sent to the Sulfate
Removal Section. The first filter cake is reslurried, filtered a second
time,and then reslurried with recovered clean water and finally dewatered
in a centrifuge.
Sulfur Removal Section - Moist coal from the centrifuge is flash dried by
high temperature steam which simultaneously vaporizes the elemental sulfur
produced in the leach reaction. The dry coal is separated from the hot
steam and sulfur vapor stream in a cyclone and cooled to give the clean
product coal. The hot sulfur vapor-steam effluent from the cyclone is
scrubbed with large quantities recycled hot water and the liquid sulfur is
drawn off to by-product storage. A small part of the hot water is used in
the Wash Section with the remainder circulated to the evaporator.
Sulfate Removal Section - The major function of this section is the evapora-
tion of wash water to concentrate leach solution for recycle. The filtrate
from the wash section and a portion of the spent wash water from the first
filter is fed to a three effect evaporator which recovers most of the wash
water. The by-product iron sulfate crystals which form in the final stage
156
-------
of evaporation are separated from the concentrated leach solution and
stored. The remaining wash water from the first filter is partially
neutralized with lime to yield a gypsum by-product. The separated and
partially neutralized wash water is combined with the concentrated solution
from the crystal separator and recycled to the Reactor Section. Overall,
the pyrite is reacted with oxygen and water to give ferrous sulfate, sul-
furic acid and sulfur. These by-products are removed as shown. The fuel
requirement is equal to a few percent of the product coal and makeup water
is needed to replace water of crystallization and water vapor loss through
the vacuum filters and the vacuum evaporator.
5.1.2.1 Conceptual Design for Commercial Scale
Process engineering studies and trade-offs produced a baseline flow
diagram for a commercial scale plant. The flow sheet, which is divided
into its four major sections is given in Figure 24. The corresponding mass
balance and stream properties are given in Table 25. The baseline plant
size was chosen equal to 100 tons of dry coal feed per hour equivalent to
about 250 MW power plant feed. This size is about the maximum size for
a single train based on available commercial equipment.
Feed and Mixer - Crushed coal, nominally 14 mesh top size, is feed from
feed hopper A-l. The coal is assumed to have 3.2 percent pyritic sulfur
and 10 percent moisture on a dry basis; thus, the total solids feed rate
is 110 tons per hour (TPH) at room temperature, assumed to be 77°F. The
coal feed, stream 1, is brought to the mix tank, T-l, by conveyor, C-l,
and introduced through the rotary feed valve, RV-1. Recycled leach
solution, stream 4, at its boiling point (215°F) is introduced to the first
mixer stage after first passing through the gas scrubber SP-1. Steam,
streams 2 and 3, is needed to raise the feed coal from 77°F to the 215°F
mixer temperature. Approximately 5.6 TPH of atmospheric pressure steam is
required to heat the coal while 6.5 TPH is available from the flash drum,
T-2. It is possible that the steam would actually be added to the enclosed
conveyor to provide heated coal with an effective 15.6 percent moisture
content. The excess 0.9 TPH would be vented through SP-1 along with any
flash steam formed in stream 4.
157
-------
A-l
FEED
HOPPER
C-l
FEED
CONVEYOR
ROTARY
VALVE
T-l
MIX
TANK
M-IA/C
MIX
TANK
MIXERS
M-2A/E K-l
SCRUBBER- FLASH
MIST DRUM
ELIMINATOR
PRIMARY
REACTOR
PRIMARY
REACTOR
MIXERS
PRIMARY
RECYCLE
COMPRESSOR
V-l
KNOCK-OUT
DRUM
R-2
SECONDARY
REACTOR
en
oo
SLURRY FEED
PUMP
P-22A/J
CIRCULATING
PUMPS
P-2
REACTOR DISCHARGE
PUMP
COAL DESULFURIZATION PROCESS
REACTOR SECTION
6-10-75 NO. 1335- Ol
Figure 24. Process Flow Diagram for Fine Coal
-------
VP-1
B-l
F-l V-2
V-3
T-3
M-4 .
B-2
F-2 V-4
V-5
VP-2 T-4
M-5
CG-1
T-5
tn
vo
VACUUM BAROMETRIC FILTER FILTRATE WASH CONTACTOR CONTACTOR BAROMETRIC FILTER FILTRATE WASH VACUUM CONTACTOR CONTACTOR CENTRIFUGE CENTRATE
PUMP CONDENSER RECEIVER RECEIVER MIXER CONDENSER RECEIVER RECEIVER PUMP MIXER RECEIVER
P-5
LEACH FILTRATE
PUMP
P-10
COOLING WATER
RETURN PUMP
P-8 P-ll P-4 P-9
WASH WATER COOLING WATER CONTACTOR CENTRATE
PUMP RETURN PUMP SLURRY PUMP PUMP
COAL DESULFURIZATION PROCESS
WASH SECTION
6-10-75 NO. 1335-O2
Figure 24. (Continued)
-------
SC-1
SCREW
CONVEYOR
D-l
FLASH
DRYER
RECYCLE
GAS
HEATER
CYCLONE
SEPARATOR
K-3
COMPRESSOR
PRESSURE
LET DOWN
SCREW CONVEYOR
SC-3
COAL
COOLER
GAS
COOLER
S-2
CYCLONE
SEPARATOR
5-3
PHASE
SEPARATOR
O
P-13
SULFUR
PUMP
P-12
PROCESS
WATER
PUMP
1 COAL DESULFURIZATION PROCESS
SULFUR REMOVAL SECTION
6-10-75 NO. 1335-O3
Figure 24. (Continued)
-------
T-7
ACID
NEUTRAL-
IZATION
TANK
M-6
NEUTRALIZER
MIXER
EV-1
FIRST STAGE
EVAPORATOR
EV-2
SECOND STAGE
EVAPORATOR
EV-3
E-1
THIRD STAGE CONCENTRATE
EVAPORATOR RECYCLE
REBOILER
CG-2
SULFATE
CRYSTAL
CENTRIFUGE
1-6
CENTRATE
RECEIVER
B-3
BAROMETRIC
CONDENSER
VP-3
VACUUM
PUMP
CTl
P-19
©
M
er;
, (a
T-7
— »
P-17
Co SO,
P-20
P-19
P-14
LEACH SOLUTION EVAPORATOR
RETURN PUMP CONCENTRATE
PUMP
P-20
CaSO4
SLURRY PUMP
P-15 P-16 P-17 P-18
EVAPORATOR EVAPORATOR EVAPORATOR COOLING
CONCENTRATE CONCENTRATE LEACH SOLUTION WATER
PUMP PUMP RETURN PUMP RETURN PUMP
COAL DESULFURIZATION PROCESS
SULFATE REMOVAL SECTION
6-10-75 NO. 1335-O4
Figure 24. (Continued)
-------
CT>
ro
TABLE 25. PROCESS MASS BALANCE FOR FINE COAL
(Stream Flows in Tons Per Hour)
Water
FeS04
Fe2(S04)3
H2S04
Pyrite
Sulfur
Coal
Oxygen
Inert
Total, TPH
T, °F
P, Psig
gpm
P, lb/ft3
Fe, %
Y
S04/Fe
COAL MAKEUP
FEED STEAM
1 2
10.0 -.9*
6.0
94.0
TTO" ^oTg"
77 215
0 -9
50.0
-
-
_
FLASH FEED
STEAM SOLN.
3 4
6.5 144.1
3.9
30.6
5.7
O" 184.4
215 215
.9 0
614
74.8
5.4
.86
1.75
R-l 02 RECYCLE
FEED MAKEUP ~ GAS
56 7
159.2 1.5
14.4
18.2
8.8
5.2
.2
94.0
3.9 13.1
Tr .8
3oO 379 TO"
215 77 264
28.8 53.8 53.8
968
77.3
5.2
.49 -
1.73 -
R-l COMPRESSOR R-l
GAS FEED EXIT
8 9 10
14.0 1.5 147.4
5.8
37.0
5.0
.7
1.1
94.0
13.4 > 13.1
.8 .8
oo o i c ^ 9Q1 fl
£O*£ 1 Q • O £~7 1 • U
250 177 250
28.8 28.8 28.8
907
80.0
6.3
.83
1.64
R-2
FEED
16
140.9
5.8
37.0
5.0
.7
1.1
94.0
284TB"
215
0
873
81.3
6.6
.83
1.64
R-2
EXIT
17
140.7
11.1
30.6
6.6
.3
1.2
94.0
28175"
215
0
874
81.2
6.7
.68
1.64
Excess steam to vent
-------
TABLE 25. (CONTINUED)
Water
FeS04
Fe2(S04)3
Pyri te
Sulfur
Coal
—j Oxygen
CT>
co Inert
Total, TPM
T, °F
P, Psig
gpm
p, lb/ft3
Return Vent
Soln. 00
18 19
131.7 Tr
3.9
30.6
5.7
.4
Tr
171.9 0.4
127 177
28.8 0
548
78.2
Crystallizer
Centrate
20
50.3
2.2
26.0
5.2
83.7
200
30. G
221
95.4
Neutral izer
Return
21
81.4
1.7
4.6
.5
88.2
160
15.0
335
65.6
Neutral izer
to
Crystallizer
22
54.3
1.1
3.1
0.3
58.8
160
15.0
224
65.6
Filtrate
to Cake
Crystal! izer F-l
23
105.2
8.3
22.9
4.9
141.3
160
10.0
443
79.6
24
43.1
1.0
3.0
.6
.3
1.2
94.0
143.2
160
0
;
Wash Contactor
F-l Vents
25 26
143.1 Tr
1.0
3.0
.6
147.7 0.0
160 77
10.0 0
589
62.5
F-2
To Contractor Feed
Neutral izer Feed T-3 Slurr
27
135.5
2.8
7.7
1.7
147.7
160
5.0
561
65.7
28 29
146.3 189.4
.3 1.3
.9 3.9
.2 .8
.3
1.2
94.0
147.7 290.9
1 60 1 60
5.0 15.0
601 1065
61.3 68.1
-------
TABLE 25. (CONTINUED)
Water
FeS04
Fe2(S04)3
H2S04
Pyrite
Sulfur
Coal
Oxygen
Inert
Total, TPH
T, °F
P, Psig
gpm
P, lb/ft3
Dryer
Cake Wash Water Makeup Evaporator
F-2 F-2 Return Water Return
30 31 32 33 34
47.2 147.2 14.3 27.1 72.9
.1 .1
.3 .3
.1 .1
.3
1.2
94.0
143.2 147.7 14.3 27.1 72.9
160 160 215 77 180
0 10.0 30.0 30.0 30.0
607 60 108 473
60.7 59.8 62.3 60.2
Centrifuge
Feed
Slurry
35
161.5
.1
.3
.1
.3
1.2
94.0
257.5
180
15.0
942
68.2
Cake Dryer Dryer
Centrifuge Gas Output
36
14.3
Tr
Tr
Tr
0.3
1.2
94.0
109.8
180
0
-
-
37 38
258.9 273.2
Tr
Tr
Tr
0.1
Tr 1.2
Tr 41.4
258.9 315.9
650 450
20.0 18.0
-
-
Dryer
Coarse
Cut
39
Tr
Tr
Tr
Tr
0.2
Tr
52.6
52.8
450
18.0
-
-
Cyclone Purge
Sol ids Steam
40 41
Tr 0.1
Tr
Tr
Tr
0.1
Tr
41.3
41.4 0.1
450 300
18.0 18.0
-
50.0
Coal
Product
42
0.1
Tr
Tr
Tr
0.3
Tr
93.9
94.3
150
0
-
50.0
Cooler
Cyclone Water
Gas Feed
43 44
273.2 1137.4
1.2
.1
274.5 1137.4
450 215
17.8 30.0
4745
59.8
-------
TABLE 25. (CONTINUED)
Mater
FeS04
Fe2(S04)3
Pyrite
Sulfur
Coal
Oxygen
Inert
Lime
Gypsum
Total , TPH
T, °F
P, Psig
gpm
P, lb/ft3
Cooler Gas
Effluent Effluent
45 46
1410.6 258.9
1.2 Tr
.1 Tr
1411.9 258.9
250 250
15.0 15.0
Separator
Liquid
47
1151.7
1.2
.1
1153.0
250
15.0
4851
59.2
Reboiler
Feed
48
1151.7
1151.7
250
40.0
4851
59.2
Water
Return
49
1151.7
1151.7
215
30.0
4802
59.8
Sulfur
Product
50
1.2
0.1
1.3
215
25.0
2.9
112.0
From
EV-1
51
124.1
9.4
26.0
5.2
164.7
120
5.0
516
79.6
to
Vacuum
52
35.4
35.4
115
(1.5
Psia)
From From
EV-2 EV-2
53 54
35.2 88.9
9.4
26.0
5.2
35.2 129.5
1 50 155
(3.7 5.0
Psia) 3?6
85.9
From
EV-3
55
37.7
37.7
207
(13
Psia)
Centrifuge Sulfate
Feed Product
56 57
51.2 0.9
9.4 7.2
26.0
5.2
91.8 8.1
210 200
5.0 0
230
99.7
Calcium
Lime Sulfate
58 59
0.4
0.5
1.6
0.9 1.6
77 77
0 5.0
2.8
144.8
-------
The mixer vessel T-l was sized for three stages of mixing at 0.25 hours
per stage. Under the design constraint that the vessel is 75 percent full,
the cost model used for vessel sizing found a field fabricated vessel 18.7
feet in diameter by 32.9 feet long has minimum cost. The selected vessel
size (18 x 36) gives three stages each about 12 feet long and 12.6 feet
deep with slightly less than 15000 gallons in each stage. Any foam gen-
erated during coal wetting will be broken down and the entrapped air will
be scrubbed in SP-1 by the returning leach solution. The actual air flow
through SP-1 is very low and will probably not exceed the air in the bulk
coal (50 cubic feet per minute).
Primary Reactor - The fully wetted and deaerated coal slurry from the mixer
is pumped by slurry pump P-l (stream 5) into the first stage of the primary
reactor R-l. Both removal of pyrite and oxidation of ferrous to ferric
iron sulfate occur in this reactor. A five stage reactor was selected
since the cost model showed the minimum cost field fabricated vessel had
length to diameter ratios near five. Under the design constraint that the
reactor must have five stages and operates about 85 percent full, the cost
model found a reactor 25.9 feet in diameter by 127.7 feet long operated at
15 psi of oxygen was minimum cost. The selected vessel size (26 x 125)
gives five stages each about 25 feet long by 23 feet deep and holding about
80,000 gallons of slurry. At the residence time of 1.5 hours per stage, a
temperature of 250°F and an oxygen partial pressure of 15 psi, the pyrite
is reduced to 88 percent of the original level and the leach solution is
regenerated to a Y (ferric iron to total iron ratio) of 0.83 in the primary
reactor.
Oxygen Loop - Excess oxygen saturated with steam and containing an equilib-
rium level of inert gas (mainly argon) leaves the primary reactor in stream
8. The gas is contacted with returning leach solution, stream 18, in a
knock-out drum, vessel V-l. The leach solution is warmed to 215°F (stream
4) by condensing steam from the oxygen stream. The gaseous effluent, which
was assumed to leave V-l 50°F warmer than the feed leach solution is split to
give a small vent stream 19 and a recycle oxygen stream 9. The vent rate
is selected to maintain the inert gas at the design level; namely 5 percent
on a dry basis. The recycle oxygen is compressed by K-l to the reactor feed
pressure. Makeup oxygen, stream 6, is added to balance the oxygen used for
166
-------
regeneration in R-l and that vented to remove inerts.
Assuming 15 psia oxygen pressure the gas pressures in reactor R-l at
250®F are as follows:
Oxygen 15.0 psia
Inert Gas .8 psia
Steam 27.7 psia
43.5 psia (28.8 psig)
Since the recycle gas must also overcome the liquid head in the reactor
(about 13 psi), the control valve/injector drop (about 10 psi) and other
line losses, the recycle compressor was sized to provide a 25 psi pressure
increase. For the baseline case this results in a 300 horsepower compressor
operating at a 1.58 compression ratio and a compressor outlet pressure of
53.8 psig.
Flash Steam - The heat of reaction and regeneration is accommodated in
three ways: temperatures of the recycled oxygen and the feed slurry are
raised in R-l, heat is lost from the insulated walls of the mixer and
reactors, and water is evaporated from the solutions. Part of the steam
(13.4 TPH) is removed from the recycle oxygen to provide an isothermal
primary reactor R-l at 250°F and part of the steam (6.5 TPH) is removed
by flash drum T-2 in dropping the slurry temperature and pressure from
reactor R-l (250°F) to reactor R-2 (215°F). The heat is almost entirely
utilized in heating the feed coal and the recycle leach solution.
Secondary Reactor - The secondary reactor, R-2, is operated near the atmo-
spheric boiling point with a residence time of 36 hours. During this time,
additional pyrite is removed from the coal to provide an overall pyrite
removal of 95 percent while the Y of the solution is decreased to a value
of 0.68 in the reactor effluent. The low value of Y is desired to provide
sufficient ferrous sulfate for removal as the by-product iron form. The
cost model found the minimum cost reactor was 27.9 feet in diameter by 465.9
feet long. The final equipment list and costing used three field fabricated
vessels each 28 feet in diameter and 160 feet long. The reactors contain no
167
-------
internal stages, but have circulating pumps to avoid large vertical con-
centration gradients from occurring in the solution. The slurry from the
secondary reactor, stream 17, is pumped by P-2 to the first filter, F-l.
Coal Washing - Bench-scale experience with removal of the sulfate leach
solution from coal shows that the solution may be treated as consisting of
two types. Surface solution is readily removed by flushing with water or
may be readily displaced by a more dilute wash solution. Solution in
the pores of the coal particles requires a definite residence time to reach
equilibrium with the bulk or surface liquid. The coal washing section,
therefore, consists of filtration, washing on the filter, equilibration
with dilute solution, a second filtration and wash, equilibration with
wash water and finally dewatering in a centrifuge.
First Filter - Coal slurry from the secondary reactor, stream 17, containing
approximately 33 percent solids is fed to a 12 foot diameter by 24 foot
long rotary vacuum filter, F-l. The filtrate from vacuum receiver V-2,
stream 23, is pumped, P-5, to the sulfate removal section. Dilute wash
solution from the second filter, stream 25, is used to wash the filter cake
and displace the surface solution on the coal particles. This sulfate rich
wash solution, stream 27, is pumped, P-6, from the vacuum receiver V-3 to
the sulfate removal section. Vacuum is provided by a 3000 standard cubic
feet per minute (SCFM) vacuum pump, VP-1, which is vented, stream 26, back
to the enclosed filter F-l. The vapors and gases removed from the vacuum
receivers, V-2 and V-3, pass through a barometric condenser, B-l, before
entering the vacuum pump. In B-l most of the flash steam is condensed and
enters the cooling water loop where it is pumped to the cooling water tower
by P-10.
First Stage Repulping - The washed filter cake from the first filter, stream
24, and dilute wash water from the second filter are gravity fed through a
closed chute to a stirred tank, T-3. This 40,000 gallon tank is operated
about three-fourths full to1 give an average residence time of 30 minutes
to equilibrate pore solution with the bulk liquid. The slurry, stream 29,
is pumped, P-3, to the second stage filter. Any gases introduced with the
cake are vented to the scrubbing system, stream 26.
168
-------
Second Filter - The partially washed slurry, stream 29, containing approxi-
mately 33 percent solids, is filtered and washed on a second filter of the
same size and type as the first filter. Filtrate, stream 25, is pumped, P-7,
from the vacuum receiver, V-4, to the first filter wash. Wash water for
the second filter, stream 31, is obtained from the centrate receiver. The
partially spent wash water is pumped, P-8, from the vacuum receiver V-5 to
the first stage contactor. Vacuum is provided by vacuum pump VP-2 operating
through the barometric condenser B-2.
Second Stage Repulping - The washed filter cake from the second filter,
stream 30, is contacted with water in a 40,000 gallon stirred tank, T-4.
The wash water, streams 32, 34, and 33 is obtained from the dryer, the
evaporators, and makeup, respectively.
Dewatering - The slurry from the second contactor, stream 35, is pumped,
P-4, to the dewateririg centrifuge, CG-1. The slurry with approximately 33
percent solids is separated in the 36 inch diameter by 90 inch long solid
bowl centrifuge to provide a dewatered coal, stream 36. According to vendor
literature and discussions, the dewatered coal is expected to have about 15
percent moisture. The centrate from receiver T-5 is pumped, P-9, to provide
the wash, stream 31, for the second filter.
Drying - Coal from the centrifuge, stream 36, is fed to a flash dryer, D-l,
by a screw feeder, SC-1. In this dryer concept the coal is heated to about
450°F by superheated steam, stream 37, and carried upward to the enlarged
top area of the dryer. The larger particles are removed from the dryer,
stream 39, while the fine particles and gas, stream 38, are fed to a cyclone,
S-l. During the drying in D-l sulfur is also vaporized from the coal and
is present along with water vapor in the cyclone effluent gas stream 43.
Thefine coal from the cyclone, stream 40, and coarse coal, stream 39, are
let down to atmospheric pressure by screw conveyor SC-2 which is back purged
with a small quantity of steam to prevent the sulfur containing gas in the
cyclone from leaving the system with the coal. The coal, stream 42, is
then transported and cooled to product storage temperature by the screw con-
veyor, SC-3 which rejects heat either to cooling water or to the atmosphere.
169
-------
Sulfur Removal - The cyclone effluent gas,stream 43, at about 450°F is
cooled by a large spray of water, stream 44, in gas cooler C-l. The water
is obtained from return stream 49 from the sulfate removal section.
The gas and liquid, stream 45, cooled to 250°F is separated in cyclone S-2
to give vapor stream 46 and liquid stream 47. The liquid stream 47
contains the water fed to the gas cooler, stream 44, the water vaporized
from the coal in the dryer, and the sulfur vaporized from the coal. The
liquid is phase separated in vessel S-3. The liquid sulfur by-product,
stream 50, is pumped, P-13, to storage while the hot water, stream 48, is
pumped, P-12, to the sulfate removal section.
Steam Circulation - Saturated steam at 250°F from the cyclone, stream 46,
is compressed by K-3, reheated by H-l, and fed to the dryer as stream 37.
Compression is accomplished by two 3500 HP series compressors which make up
the 10 psi pressure drop around the gas circulation loop. The heater pro-
vides nearly 100 million Btu per hour (MM Btu/hr) to the steam to supply
the heat required to heat the dryer feed, stream 36, to 450°F and vaporize
the water and sulfur. Slightly more than 80 MM Btu/hr are rejected to the
hot water loop, stream 48, for use in the sulfate removal section
while about 15 MM Btu/hr are lost from the equipment and lines or rejected
as sensible heat in the hot coal and liquid sulfur. The circulating water
is kept in balance by returning a portion of the water, stream 32, to the
wash section equal to the water vaporized from the feed coal, stream 36.
Neutralization - Sulfate rich wash solution from the wash section, stream
27, is fed to a stirred tank, T-7, and a lime slurry, stream 58, is added
to neutralize part of the sulfuric acid. The tank is sized for about 10
minutes of residence time and has a baffled settling zone. Gypsum slurry
stream 59 is withdrawn for disposal and the partially neutralized liquid
is removed by pump P-19. A portion of the liquid, stream 21, is returned
to the reactor section while the remainder, stream 22, is combined with
the filtrate, stream 23, as feed to the triple effect evaporators.
Evaporation - Evaporator EV-1 is operated at partial vacuum (about 0.1 atmo-
spheres) and uses condensing steam from the second evaporator, stream 53,
to evaporate water, stream 52, in the first evaporator. The evaporated
170
-------
water is condensed in the barometric condenser, B-3, and any residual gas is
removed by vacuum pump VP-3. The partially concentrated leach solution,
stream 51, is pumped, P-14, to the second evaporator, EV-2. The second
evaporator operates at about 155°F and 0.2 atmospheres using steam from
the third evaporator, stream 55, to evaporate the water, stream 53. The
two condensate streams from the reboilers of the first and second evapo-
rators (streams 53 and 55) are combined, stream 34, to provide clean wash
water for the wash section. The leach solution from the second evaporator,
stream 54, which has been concentrated to 8.3 percent iron, is at a
temperature where the solubility of ferrous sulfate is a maximum and is a
solids free solution. This stream is feed to the third evaporator, EV-3,
which is operated at atmospheric pressure and at the normal boiling point
of the solution. Heat to vaporize water is provided to the reboiler, E-l,
by the hot water loop from the wash section (streams 48 and 49). The over-
head steam, stream 55, is used in the second evaporator as previously de-
scribed. The leach, solution in EV-3 is concentrated to a total iron
concentration of nearly 12 percent which exceeds the solubility of ferrous
sulfate. Thus, crystalline ferrous sulfate forms in EV-3 and a portion of
the slurry, stream 56, is fed to a centrifuge CG-2 to separate the crystals,
stream 57, from the concentrated leach solution, stream 20. The concentrated
leach solution is pumped, P-17, to the reactor section.
Solubilities - Since the solubility of ferrous sulfate in the presence of
ferric sulfate, sulfuric acid and trace ions is not yet fully defined, the
baseline process flows may require some adjustment when pilot scale data
have been evaluated. Nevertheless, the planned mode of operation which
takes advantage of the reported solubility characteristics of ferrous sul-
fate in aqueous solution should be applicable. Below about 150°F, the
equilibrium crystalline phase is FeS04-7H20 which has an increasing solu-
bility with temperature. It reaches a maximum solubility of nearly 60 grams
of FeSO, (anhydrous basis) per 100 grams of water. Above about 150°F the
equilibrium solid phase is FeSO^-^O which has a decreasing solubility in
water with increasing temperature. Both the first and second stages of
evaporation are below the saturation limits and are expected to remain
solids free. Only the final stage operates as a crystallizer and produces
171
-------
crystalline ferrous sulfate both from a decreased solubility at the higher
temperature and from an increased concentration because of evaporation.
5.1.2.2 Process Cost Estimate
Throughout bench-scale development, process costs have frequently been
reviewed with an objective of focusing experimental effort in the areas of
greatest cost sensitivity. The capital cost of equipment required to per-
form the pyrite leaching must be carefully controlled to maintain a low
processing cost per ton of coal product. As will be seen in the capital
estimate presented in the following discussion, the major capital cost
continues to be in the reactor section of the process. This section of the
unit accounts for approximately 48 percent of the total installed equipment
capital cost. The sulfur removal section of the process accounts for 22
percent while the wash section represents 17 percent and the sulfate removal
section 13 percent of the total equipment capital requirements. It there-
fore becomes apparent that the reactor section of the process represents
the most likely area of future process economic gains as the design data
base broadens and other innovative process schemes (relative to the reaction
section of the process) are evaluated.
As the process development progressed and additional experimental data
were obtained, some complications were identified and some process simplifi-
cations were demonstrated. The net result is that at the conclusion of this
bench-scale effort, the process for removing pyritic sulfur from coal remains
highly attractive and sufficient data has been obtained to provide confidence
in the economic viability of the process.
Baseline Capital Cost Estimate - The previous section of this report pre-
sented a conceptual process design and process flow sheet for removing 95
percent of the pyritic sulfur from a high ash coal which initially contained
3.2 percent pyritic sulfur. The major equipment for the process is given in
Table 26 and identified with the equipment of the flow sheet (Figure 24).
The equipment was selected and sized to approach the optimum cost for pro-
cessing this high pyrite coal to the 95 percent removal level.
172
-------
TABLE 26. COAL DESULFURIZATION PROCESS EQUIPMENT LIST
REACTOR SECTION $3.26 MM FOB. $6.36 MM INSTALLED*
o
1 A-l Ground Coal Feed Hopper - 5000 ft
2 C-l Feed Conveyor - 20 in. Wide x 20 ft, 5 hp, 200 ft/min.
3 K-l Oxygen Recycle Compressor - 300 hp, 1.6 Compression Ratio
4 M-1A/C Mix Tank Mixers (3) - 15 hp, Stainless Steel (SS)
5 M-2A/E Primary Reactor Mixers (5) - 200 hp, SS
6 P-l Slurry Feed Pump - 1000 gpm, 60 psi, 50 hp, SS
7 P-2 Reactor Discharge Pump - 875 gpm, 5 psi, 3.5 hp, SS
8 P-22A/J Circulation Pumps (12) - 1000 gpm, 5 psi, 4.0 hp, SS
9 R-l Primary Reactor - 26 ft 0 x 125 ft, Carbon Steel (CS) with SS clad, 30 psig
10 R-2 Secondary Reactor (3) - 28 ft 0 x 165 ft, SS, 0 psig
11 RV-1 Rotary Valve - .5 hp, 18 in. x 18 in., 20 RPM
12 SP-1 Scrubber-Mist Eliminator - 3 ft 0 x 10 ft, SS, 0 psig, Baffles, Demister Pad
13 T-l Mix Tank - 18 ft 0 x 36 ft, SS, 0 psig
14 T-2 Flash Drum - 5 ft 0 x 10 ft, SS, 5 psig
15 V-l Knock-Out Drum - 5 ft 0 x 25 ft, SS, 30 psig, 15 ft Packing, Demister Pad
Installed costs for each process section were derived through the application of the appropriate
Guthrie factor^ to the FOB cost of individual pieces of equipment.
-------
TABLE 26. (CONTINUED)
WASH SECTION
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
$1.16 MM FOB,
B-l
B-2
CG-1
F-l
F-2
M-4
M-5
P-3
P-4
P-5
P-6
P-7
P-8
P-9
P-10
P-ll
T-3
T-4
T-5
V-2
V-3
V-4
V-5
VP-1
VP-2
2.28 MM INSTALLED
Barometric Condenser - SS, Condensation Rate = 13 Ton/Hr
Barometric Condenser - SS, Condensation Rate = 2. 5 Ton/Hr
Centrifuge (4) - 36 in. 0 x 90 in. Solid Bowl, SS, 150 hp
Rotary Vacuum Filter - 12 ft 0 x 24 ft Drum, 912 ft2, SS, 8 hp
Rotary Vacuum Filter - 12 ft 0 x 24 ft Drum, 912 ft2, SS, 8 hp
Contactor Mixer, 35 hp, SS
Contactor Mixer, 35 hp, SS
Contactor Slurry Pump - 1065 gpm, 15 psi, 15 hp, SS
Contactor Slurry Pump - 950 gpm, 15 psi, 10 hp, SS
Leach Filtrate Pump - 450 gpm, 10 psi, 3.5 hp, SS
Leach Wash Water Pump - 560 gpm, 5 psi, 2.5 hp, SS
Filtrate Pump - 590 gpm, 10 psi, 5 hp, SS
Wash Water Pump - 560 gpm, 5 psi, 2.5 hp, SS
Centrate Pump (4) - 150 gpm, 10 psi, 1 hp, SS
Cooling Water Return Pump - 1200 gpm, 5 psi, 5 hp
Cooling Water Return Pump - 200 gpm, 5 psi, 1 hp, CS
Contactor- 40,000 gal, 0 psig, SS
Contactor- 40,000 gal, 0 psig, SS
Centrate Receiver (4) - 650 gal, 0 psig, SS
Filtrate Receiver- 2,000 gal, Vac, SS
Wash Receiver- 2,500 gal, Vac, SS
Filtrate Receiver- 2,500 gal, Vac, SS
Wash Receiver- 2,500 gal, Vac, SS
Vacuum Pump- 3000 SCFM, 200 hp, CS
Vacuum Pump - 3000 SCFM, 200 hp, CS
-------
—1
in
TABLE 26. (CONTINUED)
SULFUR REMOVAL SECTION $1.42 MM FOB, $2.94 MM INSTALLED
1 C-l Gas Cooler - 7 ft 0 x 100 ft, Water Sprays, SS, 15 psig
2 D-l Flash Dryer - 11 ft (2 x 65 ft Drying Section, 22 ft 0 x 20 ft De-entrainment Section,
SS, 20 psig
o
3 H-l Recycle Gas Heater - 97 MM Btu/Hr, Radiant Section = 6000 ft, Convective Section =
12,000 ft2, SS Tubes
4 K-3 Compressor (2) - 1.15 Compression Ratio, 3500 hp
5 P-12 Process Water Pump - 4850 gpm, 40 psi, 150 hp, CS
6 P-13 Sulfur Pump - 3 gpm, 25 psi, 0.5 hp, SS
7 S-l Cyclone Separator - SS, 15 psig, 120,000 ACFM Capacity
8 S-2 Cyclone Separator - SS, 15 psig, 107,000 ACFM Capacity
9 S-3 Phase Separator - 50,000 gal, 15 psig, SS
10 SC-1 Screw Conveyor - 20 ft x 14 in. 0, 2 hp, SS
11 SC-2 Pressure Let Down Screw Conveyor - 20 ft x 14 in. 0, 2 hp, CS
12 SC-3 Coal Cooler - Screw Type, 20 ft x 14 in. 0, Cooled Shell, CS, 2 hp
-------
TABLE 26. (CONTINUED)
SULFATE REMOVAL SECTION $0.97 MM FOB. $1.68 MM INSTALLED
1 B-3 Barometric Condenser - SS, Condensation Rate =35.4 Ton/Hr
2 CG-2 Sulfate Crystal Centrifuge - 36 in. 0 x 72 in. Solid Bowl, SS, 125 hp
3 E-l Concentrate Recycle Reboiler - 10,000 ft2, SS/SS
4 EV-1 First Stage Evaporator - Evaporation Rate = 35 Ton/Hr, 1.5 psia, SS
5 EV-2 Second Stage Evaporator - Evaporation Rate = 35 Ton/Hr, 3.7 psia, SS
6 EV-3 Third Stage Evaporator - Evaporation Rate = 38 Ton/Hr, 13 psia, SS
7 M-6 Neutralizer Mixer - 5 hp, SS
8 P-14 Evaporator Concentrate Pump - 520 gpm, 5 psi, 2.0 hp, SS
9 P-15 Evaporator Concentrate Pump - 380 gpm, 5 psi, 1.5 hp, SS
10 P-16 Evaporator Concentrate Pump - 1380 gpm, 5 psi, 5.0 hp, SS
11 P-17 Leach Solution Return Pump - 220 gpm, 30 psi, 5.0 hp, SS
12 P-18 Cooling Water Return Pump - 8000 gpm, 5 psi, 30 hp, CS
13 P-19 Leach.Solution Return Pump - 560 gpm, 30 psi, 10 hp, SS
14 P-20 Calcium Sulfate Slurry Pump - 3 gpm, 5 psi, 0.5 hp, SS
15 T-6 Centrate Receiver - 900 gal, SS, 0 psig
16 T-7 Neutralizer Tank - 7,500 gal, SS, 0 psig
17 VP-3 Vacuum Pump - 700 CFM, 50 hp
TOTAL ESTIMATED CAPITAL - $6.SIMM FOB. $13.26MM INSTALLED
-------
Capital equipment costs were obtained from various sources: technical
literature, equipment suppliers and internal (TRW) costing data. The
specific sources of data for the various classes of equipment are presented
in Table 27 . When cost data were obtained from literature or other non-
current sources, appropriate cost escalation factors, based on the Marshall
and Swift Equipment Cost Index (to escalate costs from date of publication
to June 1975), were applied. The capital equipment cost for each processing
section is presented in Table 26. The costs are presented in terms of FOB
equipment cost and installed equipment cost. The FOB equipment cost is the
base, uninstailed cost at point of manufacture or point of shipment. The
installed equipment cost includes the following elements:
• FOB Equipment Cost
• Field Materials
- Equipment
- Piping
- Concrete
- Steel
- Instruments
- Electrical
- Insulation
- Paint
• Material Erection
• Direct Field Labor
• Indirect Costs
- Freight
- Taxes
- Construction Overhead
- Fringe Benefits
- Labor Burden
- Field Supervision
- Temporary Facilities
- Construction Equipment
- Small Tools
- Miscellaneous Field Costs
- Contractor Engineering Costs
The installed equipment cost does not include a contingency factor.
177
-------
TABLE 27. SOURCES OF EQUIPMENT COST INFORMATION
Equipment Type
Hoppers
Conveyors
Compressors
Mixers
Pumps
Reactors
Vessels >40,000 gal
Vessels <40,000 gal
Tanks >40,000 gal
Tanks <40,000 gal
Drums
Centrifuges and Support Equipment
Filters and Support Equipment
Heat Exchangers
Evaporators
Gas Cooler
Dryer
Heater
Cyclone Separators
Rotary Valves
Information Source
TRW Data
Reference 7
Elliott Company
Reference 8
Reference 6
TRW Data
TRW Data
Reference 6
TRW Data
Reference 6
Reference 6
Bird Machine Company
Ametec Company
Reference 6
Reference 9
TRW Data
TRW Data
Reference 6
TRW Data
Reference 6
178
-------
Operating Cost Estimate - The process operating costs have also been esti-
mated. The basis for these estimates was technical literature and informal
supplier quotes. Specific sources of information are presented in Table 28,
The total estimated processing cost has been determined as follows:
Capital Related Costs: Annual Cost, $1000
Depreciation - 10% straight line 1,326
Maintenance, insurance, taxes, interest (15% of 1,989
Labor: caP1tal)
Labor, 8 operating positions 1,200
Utilities:
Electrical power - 7500 KW (25 mil/Kw-hr) 1,500
Cooling water - 20°F rise; 9500 gpm 228
(5«/1000 gal)
Heating - 97 MM Btu/hr; coal, 4T/hr
Process water - 110 gpm (25<£/1000 gal) 13
Materials:
Oxygen 99.5%, 3.9T/hr ($25/T) 780
Lime - .5T/hr ($28/T) 112
TOTAL COST 7,148
Processing Cost (100 T/hr; 0.8 MM T/yr) $8.94/T of feed coal
Coal yield (weight basis) 90%
Coal yield (Btu basis) 94%
The added cost of energy may also be considered for the baseline coal.
If the baseline coal is similar to the Lower Kittanning coal utilized in
our laboratory studies, it will contain about 20 percent ash and have a
heating value of 12,300 Btu/lb as fed. After processing the coal will be
89 percent recovered, have 16 percent ash, and have a heating value of
12,900 Btu/lb. With feed coal prices at $15.00/T the feed costs 6U/MMBtu.
After processing the available energy costs $1.04/MMBtu.
Based on the current conceptual process design, it is concluded that
a broad spectrum of Eastern coals can be processed at costs of about $9.00
per ton. It was assumed in developing these costs that the pyrite removal
plant is coordinated with a power plant which will have the principal
179
-------
TABLE 28. SOURCES OF OPERATING COST INFORMATION
Cost Element Information Source
Maintenance, Insurance, Taxes and Interest Reference 8
Labor Requirement (Number of Positions) Reference 8
Labor Cost TRW Data
Utilities
Electrical Power Reference 8
Cooling Water TRW Data
Process Water Reference 8
Oxygen Linde, Division of
Union Carbide
Lime Reference 10
off-site facilities such as coal grinding facilities, change house, offices,
rail facilities, etc. To the extent that these off-sites are not available
or for bookkeeping purposes are prorated to the coal processing cost, the
direct costs given above will be increased.
5.1.3 Process Trade-Off Studies
At the start of this bench-scale program a conceptual process design
was available based on data generated in the previous program. Early
bench-scale experiments showed that in the elemental sulfur recovery
section, the desired displacement of the aqueous leach solution by an
organic sulfur solvent could not be reliably performed. Similarily, dis-
placement of the organic solvent by wash water did not occur. The
demonstrated fall-back process using parafinic or aromatic hydrocarbon
solvents as used in the current bench-scale effort is thought to be less
attractive at a commercial scale than sulfur vaporization for three main
reasons:
180
-------
• Higher equipment capital cost for the multiple extraction,
solvent distillation and sulfur crystallization.
• Higher energy use, estimated at 4 to 5 percent of the coal,
because the heat is rejected at too low a temperature level
to be used in the sulfate removal section.
§ Loss of solvent on the coal (less than 1 percent and probably
only a few tenths of a percent).
Therefore, the more favorable alternate shown in the baseline design (Section
5.1.2) was developed and demonstrated to be operable by laboratory experi-
ments.
*
With the coal washing and sulfur removal process identified, emphasis
centered on the reactor section cost. It was pointed out in Section 5.1.2.2
that nearly one-half of the process equipment cost is associated with the
reactor section in the baseline design. Consequently, the majority of the
trade-off studies conducted during the course of the experimental phase of
the program and the more concentrated effort performed during the subsequent
conceptual process design were concentrated on the reaction section. A
number of the more pertinent results are summarized in the following para-
graphs.
5.1.3.1 Reactor Model
Bench-scale results showed the initial rate of pyrite reaction increased
with increased temperature (and pressure). This increased rate only applied
to removal of the first 70 or 80 percent of the pyrite after which the rate
at high temperature was about the same as the rate obtained at the normal
boiling point of the leach solution. A computer model was prepared de~
cribing the reaction in a continuous well stirred reactor. Since the coal
particles have a statistical distribution of residence times ranging from
zero to infinity, some particles have very little removal while others with
long residence time have very high removals. The coal was grouped into a
181
-------
variable number of equal weight fractions and reacted for average residence
time calculated as the arithematic mean time for the group, For ten time
groups the numerical error in pyrite removal compared with the integrated
results was about one percent, the error was negligible with 100 or 1000
groups. When this computer routine was incorporated into the overall
reactor computer model, forty time groups were selected to provide results
with virtually no numeric error.
In a multistage reactor, coal from the first well stirred stage con-
tinuously enters the next stage. Thus, if it is assumed that the first
stage is fed with a coal of uniform pyrite concentration, the calculated
effluent from the first stage will contain coal with 40 different pyrite
concentrations. Since each of these coal fractions reacts at a different
rate, the effluent from the second stage would have 40 times 40 different
pyrite concentrations. After 3 stages,, the 64,000 different pyrite con-
centrations would exceed the storage capacity of the time sharing computer.
It was decided that the feed to a stage could be adequately represented by
five groups of coal of different pyrite concentrations. Each coal group
was reacted for the 40 time increments to give 200 output pyrite concen-
centrations would strain the storage capacity of typical time sharing
computer systems and a few more stages would exceed the capacity of large
computers. It was decided that the feed to a stage, other than the first
stage, could be adequately represented by five groups of coal of different
pyrite concentrations. Each coal group was reacted for the 40 time in-
crements to give 200 output pyrite concentrations for a reactor stage.
These 200 values were sorted and rearranged by increasing pyrite concentra-
tion. The average pyrite concentrations for the lowest 40, the next lowest
40, etc. was used to provide the five starting concentrations for the next
stage calculation. The numeric error introduced by this procedure was not
determined, but is believed to be negligible. Greater error will certainly
be caused in a real coal by the inhomogeneity of pyrite concentration in
the starting coal particles and variation in particle residence time caused
by particle size and density variation. The departure of a real coal from
ideal particle behavior assumed in the model would tend to increase the
real removal efficiency of a stage by retarding the progress of the larger
particles and the more dense high pyrite particles.
182
-------
5.1.3.2 Pressure Effect
The cost-pressure trade-off was expected to be a critical study in the
process design. The overall trade-off is somewhat complex since there are
a number of interactions among the streams and equipment that make up the
reactor system. A simplified diagram is shown in Figure 25.
Coal and recycled leach solution are fed to a multistage mixer where
some pyrite reaction occurs. The leach solution in stream 1 has increased
in total iron and decreased in Y (ratio of ferric iron to total iron) as a
result of the reaction occurring in the mixer. In the first (primary)
reactor the temperature is set at the design value and oxygen is present
at the design pressure to oxidize ferrous iron to ferric iron and thereby
increase the solution Y. Both the pyrite reaction and the iron oxidation
liberate heat. Part of the heat is lost from the reactor walls and part is
needed to heat stream 1 from the mixer temperature to the reactor tempera-
ture. The balance of the heat is removed to maintain the reactor at the
design temperature. This heat is removed by vaporizing water which is
condensed outside the reactor by contact with the recycle leach solution.
The reactor total pressure is a sum of the steam pressure at the temperature
of reaction, the specified oxygen pressure and the partial pressure of inert
gas which builds up in the recycle gas loop.
When the reactor oxygen pressure is high, the fraction of steam in the
gas is low and high circulation rates are required to remove the steam
formed to maintain constant reactor temperature. Thus, high oxygen pressure
increases the pressure rating of the reactor and the amount of recycle gas
compression, but it increases the rate of reaction by providing a higher Y
and thereby decreases the required residence time and reactor size.
The effluent from the primary reactor, stream 2, is fed to the secondary
reactor which removes additional pyrite and decreases the Y of the leach
solution. If the residence times in the mixer and the primary reactor are
preselected in reasonable ranges, then a residence time for the secondary
reactor can be found which gives the desired pyrite removal. Figure 26 shows
the time trade-offs between the primary reactor, R-l, and the secondary
183
-------
00
COAL
OXYGEN
MIXER
i
RECYCLE
SOLUTION
RECYCLE
COMPRESSION
i
J_
PRIMARY
REACTOR
I
HEAT
RECOVERY
©. SECONDARY ©, FeSO^ FeSO4r
REACTOR REMOVAL
Figure 25. Simplified Reactor System
-------
MIXER: 3 STAGE, 215F, 0.75 HR
R-l: 5 STAGE, 250F
R-2: 10 STAGE, 215F
95% PYRITE REMOVAL
5%FE, L/C =2.0
0
5 10 15
RESIDENCE TIME IN R-l, HOURS
Figure 26. Reactor Residence Times as a Function
of Oxygen Pressure in R-l
185
-------
reactor, R-2, at three oxygen pressures. A primary reactor is always re-
quired to regenerate the leach solution since there is not sufficient ferric
iron to react with the desired pyrite. In principal it is possible to avoid
a secondary reactor if sufficient time is provided in the primary reactor.
For example, the figure shows that at 15 psi oxygen pressure, 18.8 hours of
reaction time in the primary reactor will remove 95 percent of the pyrite
without a secondary reactor.
Pyrite removal is not the only criteria for selecting pressure and
residence times. It is also necessary to remove the iron sulfate formed
from the pyrite reaction. The baseline design removes as ferrous sulfate
all of the iron formed in the pyrite leaching reaction. Therefore, the
reactor effluent must contain enough iron in the ferrous form (low enough
Y) to provide this quantity of ferrous iron. Referring again to the
simplified reactor system diagram, if the secondary reactor effluent,
stream 3, has a Y of 0.67, then removal of the ferrous sulfate from re-
action will increase the Y to 0.85 for recycle. Returning a recycle
solution with a Y higher than about 0.85 presents problems in crystallizing
the ferrous sulfate from a solution which is too rich in ferric sulfate.
Figure 27 shows the influence of residence time in the secondary reactor
on the effluent ferrous iron. Instead of presenting the data in terms of
Y, the figure gives the available ferrous iron per 100 moles of pyrite
fed based on a recycle Y of 0.85. Thus, if the coal has 95 percent pyrite
removal, 95 moles of ferrous iron must be available. It is evident that
secondary reactor residence times in the vicinity of 35 to 40 hours are
needed to provide this quantity of ferrous iron.
To find the effect of pressure, computer runs were made at the correct
residence time for 95 percent removal with 95 moles of available ferrous
iron at pressures from 10 to 50 psi. The results are as follows:
186
-------
REQUIRED AT
95% REMOVAL
R-l TEMPERATURE = 250 F
R-l SIZED FOR 95% REMOVAL
MIXER FEED: 5% FE, Y = .85
FEED SO./FE RATIO = 1.75
FEED LIQUID/COAL RATIO = 2
30 35
RESIDENCE TIME IN R-2, HOURS
Figure 27. Effect of Pressure and Residence Time on Ferrous
Iron Make
187
-------
Oo Pressure Residence Times Reactor System Cost*
psi R-1, hr R-2, hr $000
10 9.21 34.8 2835
15 7.39 36.7 2814
20 6.40 37.9 2824
30 5.30 39.3 2858
50 4.27 40.7 2938
*
Includes: Mixer, R-1, R-2, Knockout drum and recycle compressor.
These results are presented in Figure 28 which shows that most of the cost
is in the primary and secondary reactors. It may be seen that the cost
minimum is very broad. Although minimum cost is near 15 psi of oxygen, it
was concluded that any pressure in this range can be used if it was found
to be more desirable for other reasons.
5.1.3.3 Iron Concentration Effect
The baseline process has a leach solution containing 5 weight percent
iron in the mixer feed. After pyrite leaching and water evaporation the
iron concentration increases to 6.7 percent. Likewise the feed ratio of
sulfate to iron is 1.75 which decreases to 1.64 at the secondary reactor
effluent. Calculations were also performed with a mixer feed of 4 percent
which gave an effluent iron level of 5.6 percent. The comparative cost
results are as follows (for 15 psi oxygen pressure):
v
5% iron feed 4% iron feed
Mixer, 3 stage 0.75 hour 0.75 hour
R-1, 5 stage 7.39 hour 8.91 hour
R-2, 10 stage 36.7 hour 38.2 hour
Equipment cost (FOB)
Mixer $118,000 $120,000
R-1 896,000 1,074,000
R-2 1,708,000 1,804,000
R-1 Knockout 28,000 28,000
Compressor 64.000 65,000
Total $2,814,000 $3,091,000
188
-------
TOTAL COST
Z
o
l-
LU
Q_
O
LU
U_
O
LU
o
QL
LU
R-2
R-l
MIX VESSEL
10
COMPRESSOR
R-l KNOCKOUT
20 30 40
OXYGEN PRESSURE, PSI
50
Figure 28. Reactor System Cost as a Function
of Oxygen Pressure
189
-------
The purchased equipment cost increase in the reactor section resulting from
a decrease of 1 percent in iron concentration is seen to be $277,000 or
nearly 10 percent. Cost impact on downstream portions of the process were
qualitatively examined. A negligible cost impact on equipment in the sul-
fate and sulfur removal sections was found, although about 10 percent less
wash water would be required to remove the aqueous sulfate leach solution.
The evaporators with a cost of about $0.5 million could be reduced in size
to produce less wash water, but the overall cost reduction would appear to
be not more than $50,000. It was concluded that total downstream process
equipment cost reduction will be much less than the $277,000 increase in
reactor cost. Therefore reducing iron concentration will result in a small
increase in process equipment cost.
5.1.3.4 Oxygen Purity
The evaluation of the effects of oxygen purity includes the effects
of inert gas concentration in the purchased oxygen and the effects of inert
gas buildup in the recycle stream, but ignores any inert gas that may be
generated from the coal during leaching. A simple diagram of the oxygen
system is shown in Figure 29. At steady state the recycle flow rate is
set by the requirement to remove steam from the reactor for temperature
control as described in Section 5.1.3.2. Also at steady state, the flow
rate of oxygen in the feed must equal the rate of oxygen consumption plus
the flow rate of oxygen in the vent stream and the quantity of inert gas
vented must equal the quantity of inert gas in the oxygen feed. The oxygen
vented (wasted) depends on its concentration in the vent stream and there-
fore its partial pressure in the reactor. For a given coal and reactor
temperature and oxygen pressure, the oxygen concentration in the vent
stream depends on the inert gas buildup in the reactor. The larger the
buildup of inerts in the reactor the lower will be the oxygen waste through
the vent. However, since the reactor temperature and oxygen pressure are
fixed, the higher the inert gas buildup the higher the total pressure and
thus the reactor pressure rating requirement. Obviously there is a trade-
off to be made between the cost of oxygen wasted an- the cost of the reactor.
For the purpose of comparing the operating cost of purchased oxygen with
the capital cost of the reactor, the following criteria, based on cost data
in Section 5.1.2.2, were applied:
190
-------
-»~VENT
OXYGEN FEED
REACTOR
3.49 T/HR
O2USE
RECYCLE
COMPRESSOR
Figure 29. Oxygen Circulation Diagram
-------
• Oxygen at 99.5 percent purity by gas volume up to 30C psi
is priced at $25/T.
• Annual cost of capital is a total of 25 percent which in-
cluded depreciation, maintenance, insurance, taxes and
interest.
• The capital for installed reactor equipment is twice the
purchase price; the actual ratio for the baseline case is
6.36/3.26 - 1.95.
J
Table 29 shows the influence of oxygen purity both in the reactor gas and
in the purchased gas on the cost variable components of the reactor system.
For comparison, purchased purities of 99.5 percent and 95 percent were
selected. The 95 percent pure oxygen is usually priced lower (based on
contained oxygen) because the argon/oxygen separation step is not required.
At a price of $20/ton of contained oxygen for 95 percent purity, the compar-
ison shows that slightly lower cost can be obtained with the lesser purity
oxygen gas if the lower purity oxygen is recycled to about 50 percent con-
sumption. The difference, which is only $50,000/yr (about $.06/ton of
coal), disappears if the differential in oxygen is actually $3/T of
contained oxygen instead of the $5/T used in Table 29. When high purity
oxygen is used recycle ratios can be set to give inert gas buildup in the
range of 5 to 50 percent inert gas in the reactor at little change in process
cost. With 95 percent oxygen the range is about 15 to 70 percent inert
gas in the reactor. The baseline case was chosen at 5 percent inert gas
in the reactor using 99.5 percent oxygen feed to allow for any coal derived
inert gas generation during the leaching operation.
5.1.3.5 Compressed Air Regeneration
The cost of oxygen has increased two or three fold in the last five
years and it is necessary to reexamine the substitution of compressed air
for oxygen. Compressed air could be of interest if its cost in the Meyers
Process were less than oxygen or if it is similar in cost but significantly
reduces the energy consumption of the process.
192
-------
TABLE 29. EFFECT OF INERT GAS BUILDUP ON REACTOR
SECTION ANNUAL COST
Oxygen Gas Containing 0.5% Inert ($25/T of Op)
Reactor inert gas, %
Reactor pressure, psig
Oxygen vented, T/hr
Equipment cost, $000 (FOB)
R-l
R-l knockout
Compressor
Total
Annual cost, $000*
Capital related
Total Op cost
Total cost, $000
Oxygen Gas Containing
Reactor inert gas, %
Reactor pressure, psig
Oxygen vented, T/hr
Equipment cost, $000 (FOB)
R-l
R-l knockout
Compressor
Total
Annual cost, $000*
Capital related
Total 02 cost
Total cost, $000
2
53
1.14
901
28
60
989
494
913
1407
5% Inert
, 7
54
8.13
908
27
29
964
482
1838
2320
5
54
.37
905
28
64
997
498
760
1258
)
($20/T
10
55
3.15
913
29
54
996
498
1049
1547
10
55
.17
913
29
67
1009
504
712
1225
of 0,)
15
56
1.49
922
30
63
1015
508
786
1294
20
57
.07
931
31
72
1034
517
701
1218
20
57
.93
931
31
68
1030
515
697
1212
50
68
.02
1031
40
90
1161
580
691
1271
50
68
.19
1031
40
89
1160
580
580
1160
80
113
.00
1429
56
123
1608
804
687
1491
80
113
.04
1429
56
123
1608
804
553
1357
90
163
.00
2087
111
142
2340
1170
687
1857
90
163
.00
2087
m
142
1857
1170
550
1720
Power cost for the recycle compression was neglected because it was too
small to affect the cost comparison.
193
-------
The reactor section was sized and costed in a manner similar to the
method used for examining oxygen purity (Section 5.1.3.4). A summary of
the results is shown in Table 30. Three oxygen partial pressures were
included, namely: 15, 10 and 5 psi and for each oxygen partial pressure
three levels of oxygen utilization were examined. It had been the plan to
examine 75, 50 and 25 percent use of the oxygen, but convergence logic in
the computer model sometimes prevented the 25 percent point from being
obtained. Basically the problem relates to water balancing the reactor
system.
It is evident from the results that the major trade-off occurs between
reactor cost and compression cost. For example, at 15 psi of oxygen
partial pressure when the oxygen is 75 percent consumed, the reactor pres-
sure needed is 338 psig largely from the high concentration of nitrogen in
the reactor vent gas. If the oxygen in the air is only 25 percent used the
reactor pressure decreases to 138 psig. However, while the reactor cost
decreases with decreasing pressure, the compression cost increases and the
optimum at 15 psi of oxygen pressure is between these extremes.
A second factor is water vapor loss from the reactor. The reactor
operating at 250°F and sized for an overall removal of 95 percent of the
pyrite from a 3.2 percent pyritic sulfur coal produces excess heat equiva-
lent to about 12 tons per hour of water vaporized in the reactor. As either
the oxygen partial pressure decreases or the percent oxygen utilization
decreases the water vapor in the vent gas increases. Thus at 15 psi oxygen
pressure oxygen utilization can be decreased to less than 25 percent before
all excess steam is removed, but at 5 psi of oxygen pressure something less
than 45 percent (probably 35 to 40%) is the lowest utilization possible.
If the reactor temperature is increased above 250°F or a coal with less
starting pyrite is used, even greater oxygen utilization than shown in
Table 30 will be required.
The cost comparison shown in Table 30 was made on the assumption that
the large quantity of pressurized vent gas can provide one-half of the
energy needed to compress the feed air. Even with this optimistic assumption,
it may be seen that the process costs are higher with air than the baseline
194
-------
TABLE 30. EFFECT OF COMPRESSED AIR ON REACTOR SECTION ANNUAL COST
01
Oxygen pressure, psi
Reactor pressure, psig
Oxygen used, % of feed
Water vented, T/hr
Recycle compressor, HP
Air compressor, HP ^
Equipment cost, $000 (FOB)
R-l
R-l Knockout
Recycle compressor
Air compressor
Total
Annual cost, $000
Capital related
Total 02 cost
Compressor power*
Baseline
99.5% Oo
15
54
90
.0
315
-
905
28
64
-
997
498
760
63
1321
15
338
75
.2
880
3235
3405
242
146
414
4207
2104
-
500
2604
15
188
50
.7
685
4095
2087
109
120
501
2817
1409
-
546
1955
Regeneration
15
138
25
3.1
310
6600
1648
70
64
733
2515
1253
-
722
1975
10
238
75
.4
810
3235
3064
149
137
414
3764
1882
-
486
2368
with Compressed Air
10
138
50
1.1
550
3485
1988
73
100
440
2601
1301
-
459
1760
10
109
30
4.1
200
4880
1680
48
44
576
2348
1174
-
528
1702
5
138
75
.7
625
2455
2909
75
111
333
3428
1714
-
371
2085
5
88
50
2.4
230
2730
2096
55
50
362
2563
1282
-
319
1601
5
83
45
3.5
125
2910
2022
50
30
381
2483
1242
-
316
1558
Power cost assumes $0.025/kw-hr and that 50 percent of air compression power is recovered by
expanding the vent gas through a turbine drive on the compressor.
-------
process with oxygen. The factor of overall power use also shows that air
generally is a significantly higher power consumer. Oxygen is produced at
a total energy consumption of about 345 Kw-hr per ton of oxygen. This is
equivalent to nearly 1800 horsepower for the oxygen needed in the baseline
case. Thus, the baseline which uses a total of about 2100 horsepower for
oxygen production and gas recycle is well below any of the air fed cases
and only the low pressure, low oxygen utilization case is competitive if
the optimistic recovery of energy from the vent gas is assumed. Since no
process testing has been conducted at these low oxygen pressure, high gas
throughput rates, it is entirely possible that solution foaming may occur
or that much greater reactor volumes or residence times may be needed.
The current analysis tends to show that oxygen continues to be much pre-
ferred over compressed air in the process design.
5.1.3.6 Three Reactor Configuration
In the early stages of the process design, trade-offs were made between
the baseline design and a three reactor configuration. The concept of the
three reactor configuration was that the first reactor (R-1A) would use
oxygen at the highest pressure in a once-through mode to give partial
reaction and partial regeneration. The second reactor (R-1B) would con-
tinue the reaction and regeneration using the vent oxygen at a lower pres-
sure and lower reactor temperature also in the once-through mode. The
third reactor (R-2) would be like the second reactor of the baseline process.
Although all of the process constraints were not known in the early cost
trade-off studies, the trend was for the first reactor of the three reactor
configuration to move toward low pressure and the second reactor to move
toward atmospheric pressure and thus become the two reactor (baseline)
configuration. Upon completion of the baseline definition a brief exami-
nation was made of the cost of a three reactor configuration meeting the
processing constraints (pyrite removal and adequate ferrous sulfate in the
effluent). Table 31 shows three cases which meet these constraints. It
can be seen that in each case the equipment cost is higher than the base-
line case. A full study which is both costly and time consuming was not
undertaken since it appeared unlikely that there is a configuration with a
significantly lower cost than the baseline case.
196
-------
TABLE 31. EQUIPMENT COSTS FOR A THREE-REACTOR CONFIGURATION
Operating Conditions
-' R-IA R-1B
Baseline 7.49 hrs, 250°F, 15 psi 02 None
Case 1 2.50 hrs, 250°F, 50 psi 02 10 hrs, 220°F, 10 psi02
Case 2 2.25 hrs, 250°F, 50 psi 02 14 hrs, 220°F, 5 psi 0£
Case 3 2.00 hrs, 250°F, 50 psi 02 14 hrs, 225°F, 5 psi02
Equipment Cost, $000 (FOB)
Baseline
Mixer 118
R-1A 905
R-1B
R-2 1677
Knockout 28
Compressor 64
Case 1
118
490
825
1677
Case 2
118
450
964
1677
Case 3
118
409
1008
1677
2792 3110 3209 3212
5.1.3.7 Additional Studies
While the process design effort attempted to examine all of the para-
meters which were of major importance to providing a low cost process
design, it considered only the bench-scale Lower Kittanning coal which had a
fairly high level of pyrite and rather poor filtering qualities. As data
becomes available on other coals, new designs should be prepared to ac-
commodate other separation rates, different pyrite levels, other levels of
pyrite removal, other methods for by-product removal and additional para-
meters identified during the testing and process design. Since coals may
differ significantly in reaction rate and processing parameters, it is
likely that several different process flow sheets will be needed to accom-
modate the heat and mass balances for a wide spectrum of coals. It is
likely that the improvements in reaction and filtration rates that appears
to come with mechanical precleaning of coal will lead to a still lower cost
process.
197
-------
5.2 COARSE COAL (-1/4 INCH) PROCESSING
The bench-scale efforts relative to the processing of coarse coal
(see Section 4) have been principally directed toward the acquisition of
data and information on pressurized reaction systems. Comparisons of data
for pressurized and unpressurized systems indicate that the imposition of
pressure does not significantly influence the pyrite leach rate from coarse
coal. Because the use of non-pressurized reaction vessels and support
equipment significantly reduces capital investment requirements, a decision
was made to formulate a conceptual coarse coal process design on that basis.
The rationale, design details and estimated economics of this design are
presented in the following subsections.
5.2.1 Concept Development
The physical characteristics and laboratory-derived leach reaction
rates of coarse coal required the development of a completely different
processing concept than that conceived for fine (suspendable) coal. It
is, of course, understood that one of the primary virtues of coarse coal
processing is that the product is shippable, handleable and storable by
conventional means. In terms of processing, coarse coal usage implies less
pre-leaching preparation, the elimination of requirements for product pellet-
izing or briquetting operations and the potential application of conven-
tional physical separation systems to augment the productivity/investment
ratio of the Meyers Process.
The major physical characteristic of coarse coal which necessitates
modification of the fine coal processing concept is the tendency of the
1/4" x 0 coal grind to separate. The fines (approximately 48 mesh x 0)
are nominally slurryable, while the larger mesh fraction (about 1/4" x
48 mesh) settles out quite rapidly. It is therefore impractical to consider
slurry transportation of the coal; because of the tendency for the coal and
leach or wash liquors to separate it is in fact much more efficient to
retain the coal in a fixed position and allow the liquors to pass through
the reacting mass (as in a packed bed) and to transport the coal mechanically
by means of conveyor systems.
198
-------
While the basic leaching reaction established during fine coal in-
vestigations is fully applicable, the range of particle sizes included in
the nominal 1/4" x 0 grind gives rise to a corresponding range of reaction
rates; the fines react in basic accordance with the rate constants derived
in the fine coal evaluations, but the coarse particles are considerably
slower. Since, as indicated earlier, there is little rate constant ad-
vantage offered by the use of pressurized reaction vessels, the coarse
coal conceptual process must allow for relatively long coal retention
periods. The leaching residence time may be shortened by preprocessing
the coarse ROM coal through conventional physical cleaning (scalping)
which offers a means of segregating a small, extremely pyrite rich and
difficult to leach, fraction of coal for direct disposal. Further economic
advantages are indicated when the scalped coal is deep cleaned into two
fractions, one lean in pyrite that would not require depyritization and
one rich in pyrite to serve as feed stream to the Meyers Process. The
recombined product would meet the sulfur standards even though only approxi-
mately one half of the ROM coal was subjected to the Meyers Process.
A total of four reactor configurations were identified as potentially
viable. These include an above-ground batch reactor; a lined pit batch
reactor; a continuous countercurrent reactor; and a continuous cocurrent
reactor. All of these possibilities were conceived as atmospheric pressure
units in which only the leach reaction occurs; because of the relatively
easy coarse coal-from-liquid separation, the regeneration reaction can be
accomplished in a separate vessel optimized for that purpose.
The four reactor configurations were simulated with computer models
and evaluated in detail. It quickly became apparent that the above-ground
batch reactor scheme would be the most expensive and complex of the group
and it was therefore eliminated from consideration on that basis. It was
also found that the continuous countercurrent reactor requires a smaller
vessel and a lower leach solution throughput than the cocurrent reactor,
therefore the continuous cocurrent reactor scheme was also eliminated.
199
-------
Both the continuous countercurrent and batch pit reactor schemes were
expanded to complete conceptual flow diagrams for further study. The pro-
cess, as described in the following paragraphs, could incorporate either
reactor type and could be preceded by a physical separation train. The
conceptual design is presented schematically in Figure 30. A brief
description of the conceptual design follows with details of the key ele-
ments being presented in subsequent subsections.
Coal is fed to the reactor, stream 1, at a design rate of 100 TPH
where it contacts a continuously regenerated stream of leach solution,
stream 2, of YENTER = 0.95. Spent leach solution, containing a portion
of the fines fraction from the coal feed is withdrawn from the reactor,
stream 3, and blended with the liquid stream 17 recovered from the coal
drain. Stream 17 also contains a portion of the fines fed to the reactor
via stream 1 which are acquired during the drain and rinse cycle. The
weak leach solution-coal slurry is bled, stream 5, at a rate sufficient to
maintain the total iron content of 4 percent and to maintain the fines
content of the leach solution loop at a maximum of 10 percent. The fines
in bleed streams are removed by filtration with the filter cake being con-
veyed, stream 7, for reinjection into the coal processing sequence at the
water wash step. The filtrate is pumped, stream 6, to a unit which converts
the major portion of the Fe to Fe by iron reduction. The Fe+2 rich
stream is then sent to an evaporator/crystallizer unit, stream 9, which
produces the clean water used in the coal wash, stream 12. The ferrous
salt is filtered from the crystal!izer output, stream 13, and disposed of
through liming, stream 14. Residual solution, taken as filtrate, is routed
to the regenerator, stream 11, together with the spent leach solution which
bypasses the bleed stream diverter, stream 10. Regeneration of the leach
solution is accomplished as outlined in Section 5.1.1 relative to fine
coal processing; oxygen is supplied, stream 15, to convert the Fe+2 ion
and sulfuric acid to Fe and water. The freshly regenerated leach solu-
tion, stream 2, completes the loop through reinjection into the reactor.
The leach solution-wet coal leaving the reactor, stream 4, is drained
w
on a conveyor and spray rinsed with leach solution-contaminated water,
200
-------
COAL FINES
r\>
o
SCRAP IRON
STEAM
COAL-
AMBIENT
REACTION
SYSTEM
(PIT OR
CONTINUOUS)
LEACH
SOLUTION
DRAIN AND
RINSE
WATER
WASH AND
DRAIN
CENTRIFUGE
EVAPORATOR
CRYSTALLIZER
WASTE SALT
TO LIME PIT
DRYER
SULFUR
DISTILLATION
AND COOL
PRODUCT
COAL TO
STORAGE
AND
SHIPPING
SULFUR
CONDENSE
AND
DISPOSAL
Figure 30. Coarse Coal Process Schematic
-------
stream 18. This rinse water stream is the filtrate obtained by removing
fines from stream 19. The wet fines are conveyed directly to the dryer
via stream 20.
The wash vessel is sized for a nominal 1-hour residence time to permit
transfer of the leach solution trapped in coal pores to the passing water
stream, which originates as evaporator and dryer condensates, streams 12 and
24. Following a conveyor draining step similar to that employed earlier to
remove excess leach solution, the coal, stream 22, is transported to a
continuous centrifuge which reduces the total water content to about 20-22
percent with the centrate, stream 25, going to the fines filter. The coal
is conveyed, stream 23, to a two stage inert atmosphere dryer/distillation
unit where water is first removed, stream 24, followed by elemental sulfur
removal, stream 26. The elemental sulfur is cast into blocks for disposal
and the coal, stream 27, is cooled for preshipment storage.
5.2.2 Conceptual Process Details
5.2.2.1 Reactor Section - Pit and Continuous
Both the batch pit and continuous countercurrent reactor concepts were
sized to accommodate 100 TPH of 1/4" x 0 ground coal. In the case of the
pit reactor, three identical units are required such that filling, leach
reaction and raw product coal withdrawal can proceed concurrently; the
withdrawal rate, which sizes the balance of the process units, is 100 TPH.
The pit reactor configuration, upon which the conceptual process design is
based, is shown schematically in Figure 31 . As can be seen, the sectional
view indicates a basically triangular pit with the sides angled at 30°
such that the natural angle of repose of wet coal is exceeded and the coal
can be withdrawn via the conveyor at the bottom. The pit and drain pump
lining material is acid-resistant concrete and the metal parts and various
exposed equipment items are of stainless-clad mild steel. Each pit is sized
to hold sufficient coal for 5 days of operation, i.e., 12000 tons at the 100
TPH withdrawal rate. Using a bulk density of 47 lb/ft3, the reactor volume
is 511,000 cubic feet; for a 300 foot length, the tank width is about 43 feet
and the depth about 80 feet.
202
-------
REMOVABLE
LEACH SOLUTION
SPRAY RACK
ro
o
CO
ACID RESISTANT CONCRETE
LEACH SOLUTION COLLECTION SUMP
Figure 31. Pit Reactor Schematic
-------
The leach solution is distributed over the upper surface of the coal
mass by means of a hinged spray rack which can be moved for pit load
operations. The leach solution drains through the coal bed, is collected
in a full-length sump, and is circulated back to the regenerator. Fol-
lowing completion of the leach reaction, coal is withdrawn from the pit
reactor by opening the hydraulically-operated doors comprising the lower
one-third of the reactor wall; this allows the coal to fall onto a nominal
48 mesh conveyor and to drain residual leach solution while being trans-
ported to the rinse section of the plant.
The five day pit capacity imposes a similar limit on the leach reaction
period. Based on bench-scale and laboratory data, rate constants were es-
timated which show that about three days of leaching will be required to
reduce pyrite to target level. The design over-capacity allows for the use
of coals having a higher pyrite content or the processing of a pyrite-rich
fraction from a front-end physical separation/pyrite concentration unit as
described earlier.
All three pit reactors are serviced by a track-mounted reclaim-
conveyor system having a 120 TPH capacity; this unit reclaims coal from a
pile and distributes it along the length of each pit. The 120 TPH capacity
assures maintenance of any operating cycle time within the 5-day maximum
imposed by the reactor design.
The ambient-pressure continuous countercurrent reactor system concept
was also sized for a 100 TPH feed rate. Each reactor (see Figure 32) con-
sists of a cylindrical outer shell having a conical bottom and a discharge
lock hopper at the lower end of the cone. A raised lip surrounds the upper
periphery of the cylindrical shell, stainless steel clad is used over all
interior surfaces, a spreader cone at the top distributes the feed coal
over the cylindrical area and a series of internal baffle plates limits
the formation of any flow channels which may develop. The conical bottom
section houses the leach solution distribution orifices through which
leach solution is forced under pressure to flow upward through the coal bed.
Operationally, coal is loaded at the top of each reactor at the design
feed rate and is distributed by the spreader plate; continuous withdrawal
204
-------
LEACH SOLUTION
OVERFLOW TROUGH
COAL FILL DISTRIBUTOR
PLATE
SUPPORT
STRUCTURE
TO
LOCK-HOPPER FOR
COAL REMOVAL
COAL MIXING
BAFFLES
LEACH SOLUTION
INJECTOR MANIFOLD
Figure 32. Continuous Reactor Schematic
205
-------
through the base lock hopper provides a nominally uniform downward travel
of the coal. The design residence time for processing feed coal having a
3.2 percent pyritic sulfur content is 50 hours. The reactor system con-
sists of five vessels each 20 feet in diameter and holding 1000 tons of
coal. With feed and withdrawal rates of 20 TPH, coal leaching occurs in
about 80 feet of vessel length. Leach solution at a rate of 286 TPH flows
upward through the coal at a rate of less than 0.01 ft/sec. The feed
leach solution has a Y of 0.95 and 4 percent iron while the effluent
solution has a Y of 0.5 and 4.2 percent iron.
5.2.2.2 Regenerator Section - Pit and Continuous
As the regeneration reaction is conducted in a separate vessel for
both the pit and continuous leaching concepts, it is possible to optimize
the design for that purpose. This is in contrast to the fine coal process
wherein both the leaching and regeneration reactions take place concurrently
in the same vessel, necessitating off-optimum conditions for each. As
mentioned earlier, the leach solution is actually a 10 percent fine coal
slurry therefore both regenerator designs must incorporate sufficient
mechanical agitation to maintain the slurry and minimize settling tendencies.
The pit reactor, being a batch-mode operation, will deplete the leach
solution to different degrees throughout the period of the leach reaction.
This in turn imposes a variable load on the regenerator with the peak load
occurring at the start of the leaching reaction. The regenerator has been
sized to accommodate the peak, resulting in excess capacity for the major
portion of the coal reaction period.
As a result of a trade-off/optimization study, the regenerator for the
pit reactor process has been sized at a total capacity of 62,000 cubic feet
at a working pressure of 150 psia and a working temperature of 250°F. The
vessel is rubber and acid brick-lined and contains three wier-separated
stages of equal volume. The vessel accepts the 1620 TPH of leach solution
flow rate required by the pit reactor and regenerates it to a minimum Y =
0.95.
206
-------
The regenerator required by the continuous reactor Is designed to accept
a constant input Y of 0.5 and to regenerate the leach solution to an out-
put level of Y = 0.95. As was the case with the pit system regenerator, the
vessel is rubber and acid brick-lined and is designed for a 150 psia working
pressure at 250°F, and contains three equal volume stages. However, the
leach solution flow rate is lower (1430 TPH) requiring a vessel of about
55000 cubic feet.
Both regenerators are operated at well over the atmospheric boiling
point of the solution and are therefore good sources of low-pressure steam.
In addition to process heat, a portion of the low-pressure steam is used to
drive an aspirator which pulls a mild vacuum on the two fines filters, as
will be discussed in Section 5.2.2.4.
5.2.2.3 Sulfate Removal Section
The reaction with pyrite produces 1.0 mole of iron and 1.2 moles of
sulfate from each mole of pyrite reacted. In the fine coal process,
Section 5.1.2, the iron and 1.0 mole of the sulfate are removed as FeS04
while the excess 0.2 mole of sulfate is neutralized with lime to form
gypsum. An alternative approach was incorporated in the coarse coal pro-
cess. This approach involves dissolving scrap-iron into the leach solution
by the reaction:
o
+3 . n c c +2
0.2 Fe + 0.4 Fe * •*• 0.6 Fe* (22)
Then all the reaction products are removed as 1.2 moles of Fes04- This
section of the process (see Figure 33) is the same for both the pit and
continuous leaching concepts. In processing coarse coal, the best access
to the spent leach solution is just downstream of the fines filter (Stream
6, Figure 30) at which time essentially no fines are present. The entire
fines-free bleed stream is pumped to a heater conversion tank having a 3-
hour residence capacity. This tank is filled with scrap iron. The Fe
+ Fe+2 conversion is based on an input Y of 0.5 and an effluent Y near
zero. The converter effluent stream is pumped to an evaporator/crystallizer
207
-------
ro
o
CO
FROM REACTOR, _
DRAIN/RINSE _ (TO)
REGENERATOR
TO WASH
-Y = 0.5
EVAPORATOR
CONVERSION TANK
v\
TO WATER WASH -4
STEAM,
SCRAP IRON
y
TO REGENERATOR
-------
train which reduces the water content and increases the iron content to
above the saturation level. The overhead water product, Stream 12, is
sent to the coal wash line and the salt-laden stream to a leaf filter
+2
which separates the Fe salt. All residual streams are combined and
pumped to the regenerator. The ferrous sulfate salt is sent to disposal.
5.2.2.4 Coal Washing and Filtration Sections
In both the pit and continuous reactor systems, similar schemes for
the separation of coal fines and for the washing of the raw leached coal
product are envisioned. Fine coal must be separated at two points in each
of the conceptual designs (see Figure 30). The first provides a means of
controlling the fines content of the leach s'olution at a maximum of 10
percent; this involves filtration of the spent leach solution bleed stream,
using a drum-type vacuum filter. The wet fines cake is conveyed to the
water wash line and the filtrate is transferred to the excess salt removal
section.
The second fines recovery operation, again employing a rotary drum
vacuum filter, operates on the wash water circuit (Stream 19, Figure 30).
This filter separates a water-wet fines cake which is transported to the
dryer. The filtrate is recycled for use as coal rinse.
Analyses of typical 1/4" x 0 coals have indicated that the fines con-
tent can be as high as 30 percent of the bulk mass although 15 to 25 percent
is typical. The filters were sized in order to accommodate the extreme,
but with the understanding that this represents an overcapacity in terms
of most applications. It should be noted that considerable in-process
separation of fines is to be expected in the reaction, washing, rinsing and
centrifugation steps. The degree of fine particle entrapment to be ex-
pected is not known at this time. In some instances the fines (in the
continuous reactor) may separate upon feeding from the distribution plate,
thereby immediately loading the leach circuit return line, in other cases
the fines may be carried through to the rinse/wash line and separated at
that point. It therefore follows that a degree of overcapacity is war-
ranted throughout the fines removal and transportation operations in order
to allow for normal feed coal variations.
209
-------
5.2.2.5 Sulfur Removal Section
Both the residual water and free sulfur are removed from the coal in
a two-stage inert atmosphere dryer system. In the first stage, water
amounting to a nominal 16.5 TPH is quantitatively removed as the feed is
heated to 480°F. In the second stage the coal mass is held at 480°F for
x
1 hour to permit distillation of the elemental sulfur. The clean water is
recovered, cooled and pumped to the wash train water feed; sulfur is
recovered from the inert gas atmosphere and cast into blocks for sale or
disposal. The product coal is cooled and transported to a temporary
storage area to await on-site use or shipment.
5.2.2.6 Energy Balance
An energy balance around the process indicates it is possible to
eliminate all battery limit heat supply requirements with the exception of
the dryer. The reactor (either pit or continuous) temperature is maintained
by conserving the heat of reaction and by the sensible heat supplied by the
freshly regenerated leach solution. The evaporator, conversion tank and
crystal!izer heat requirements are supplied by the heat captured from the
leach solution flash and blowdown vessel. The regeneration reaction is
quite exothermic and by operating the regenerator at 150 psi and 250°F,
well above the solution boiling point, the subsequent pressure let down to
ambient pressure results in the generation of large quantities of low-
pressure steam. The quantity of steam is sufficient to operate an ejector
which pulls a mild vacuum on the evaporator and the two fine-coal filters
s
and also to fulfill the latent and sensible heat requirements of the
hardware items mentioned earlier. The dryer heat requirements, expressed
in terms of the quantity of product coal required, were found to be
9.5 TPH coal.
5.2.3 Process Cost Estimate
Processing costs, including both estimates of capital requirements and
annualized costs, have been calculated for both the pit and continuous pro-
cess concepts. The basis for calculation was June 1975 in both instances
210
-------
For purposes of comparison to process costs for fine coal processing (as
presented earlier in Section 5.1.2.2), the plant capital requirement was
assumed to be that for a "battery limits" facility. That is, the economics
to be presented here do not account for plant off-sites and do not repre-
sent a "grass roots" operation. The logic behind this economic treatment,
the information sources, and the method of treatment are the same as pre-
sented in the fine coal section.
Coarse Coal Capital Cost Estimate
Process cost estimates are presented for the four process schemes
employing the two reactor system concepts described earlier. The four
approaches, pit, continuous, pit/float-sink combination, and continuous/
float-sink combination were evaluated on the basis of a Meyers Process
feed rate of 100 TPH coal containing 3.2 percent pyritic sulfur.
The objective of front-end physical separation equipment in the float
sink combination options is to isolate a low pyrite segment of the total
feed which does not require further treatment and to concentrate the pyrite
in a "sink" fraction which can be fed to the coarse coal Meyers Process.
Float-sink separation data for approximately 400 Appalachian coals were
examined. It was found that on the average each of these coals could be
separated into two approximately equal weight portions by float/sink
separation in a liquid medium with a specific gravity of 1.3. Further,
it was noted that in 227 of the coals the float fraction contained less
than 1 percent total sulfur. For these 227 coals the data shows that
the 1.3 specific gravity "float" fraction amounts to nearly 55 percent of
the 3/8 inch by zero sample with a standard deviation of about 15 percent.
The average total sulfur content of this float fraction is 0.8 percent with
a standard deviation of 0.1 percent
Float-sink equipment trains are presently in widespread use, but use
higher specific gravity media to achieve float fractions in the 90 percent
range, thereby minimizing losses and maximizing the saleable product. It
is apparent that if a separation were made at low specific gravity as
suggested above but without the benefit of a Meyers Process for treating
the "sink" fraction, the gross effect would be a virtual doubling of coal
prices as well as the generation of about five times as much spoil re-
quiring acceptable disposal.
-------
In combining the float-sink treatment with the Meyers coarse coal
process it was assumed that a 50-50 float-sink division of the feed would
represent a reasonable design value, based on the analysis described above.
Further analysis of these coals indicated that the sink fraction to be
used as process feed would contain about 3.8 percent pyritic sulfur. In
a commercial application an economic trade-off would be made at this point
to establish the most advantageous position in terms of the adviseability of
blending the "float" fraction with the process product (thereby reducing
the degree of pyrite removal in the process) versus adjustment of the media
specific gravity to control the quantity of "sink" fraction fed to the pro-
cess. In the conceptual designs considered here, no reblending is considered
in order to facilitate direct comparison with the fine coal process.
An equipment list detailing installed equipment costs has been prepared
for 100 TPH pit and continuous coarse coal battery limits process options
and is presented in Table 32. As may be seen, the installed equipment capital
required for the pit process option is $4.17 million while that required for
the continuous process option is about 50 percent higher. It should be
noted that for the pit and continuous battery limits processes, the reactor
sections account for approximately $1.54 million and $3.95 million, respec-
tively, or 37 percent and 60 percent of the total capital requirement
(compared to 48 percent for the fine coal processing base case). The
additional installed capital investment required for the addition of float-
sink options is estimated to be $.42 million.
Operating Cost Estimate
The process operating costs have been estimated for the four options
and are summarized in Table 33. It may be seen that the operating costs
for the coarse coal options range from $2.94 per ton of coal processed to
$6.28 per ton. These are comparable to the $8.94 per ton determined for the
fine coal processing scheme (Section 5.1.2.2). However, it must be stressed
that the operating costs presented in Table 33 are based on process
designs with slightly varying bases. These differences are the following:
212
-------
TABLE 32. COARSE COAL PROCESS EQUIPMENT LISTS
Pit Reactor Continuous Reactor
Item $1000 $1000
Reactor 728 3175
Regenerator and Exchanger 658 625
Compressor 6 6
Wash Vessel 91 91
Centrifuge 79 79
Knock-out Pot 10 10
Regenerator Agitators 106 106
Flash Vessel and Exchanger 21 21
Regenerator Pumps 8 8
Leach Solution Storage 18 24
Leach Surge Tank 8 8
Iron Conversion Tank 24 24
Iron Conversion Tank Exchanger 70 70
Iron Conversion Tank Pump 2 2
Fines Filter (Leach Circuit) 65 65
Salt Filter (Plate) 32 32
Evaporator 556 556
Crystal!izer 71 71
Fines Cake Conveyors (2) 11 H
Crystallizer Feed/Exit Pumps 4 4
Evaporator Feed/Exit Pumps 11 H
Evaporator/Filter Ejector 12 12
Evaporator Condensate Vessel 4 ^
Coal Rinse Conveyor and Sump 62 62
Wash Water Feed/Exit Pumps 6 6
Wash Rack Pump 1 1
Centrifuge Fines Filter 64 64
Wash Vessel Drain Conveyor 30
Centrifuge Liquid Exit Pump 1 ^
Sulfur Trap 18 18
Evaporator and Dryer Condensate Hold Tank 12
Dryer (2 Units) _1381. - 1381
Process Equipment TOTALS 4,170 '
213
-------
TABLE 33. ANNUALIZED COSTS FOR BATTERY LIMITS DESULFURIZATION PLANTS
Cost Element - $1000/Yr
Capital Related Costs
Depreciation - 10%
Maintenance, Insurance, Taxes
Labor
Utilities
Electricity (25 Mil/KW-HR)
Heating-Coal Equivalent
Water-Process and Cooling
Materials
Oxygen 99.5% ($25/Ton)
Lime ($28/Ton)
Scrap Iron ($50/Ton)
TOTAL
Cost/Ton Feed - $/Ton
% Coal Yield (Weight Basis)d
% Coal Yield (Btu Basis)d
PROCESS OPTION
Pit Reactor
Process9
417
626
900
(6 Positions)
600
(3,000 KW)
(9 TPH)
(247 MM Btu/Hr)
458C
800
112
280
4,193
$5.25/Ton
85%
90%
Continuous Reactor
Process3
659
989
900
(6 Positions)
600
(3,000 KW)
(9 TPH)
(247 MM Btu/Hr)
458C
800
112
280
4,798
$6. 01 /Ton
85%
90%
Pi t/Fl oat-Sink
Combination^
459
689
1,200
(8 Positions)
700
(3,500 KW)
(9 TPH)
(247 MM But/Hr)
458C
800
112
280
4,698
$2.94/Ton
92%
95%
Conti nuous/Fl oat-Si nk
Combination'3
701
1052
1,200
(8 Positions)
700
(3,500 KW)
(9 TPH)
(247 MM Btu/Hr)
458C
800
112
280
5,303
$3. 31 /Ton
92%
95% •
Fine Coal
Process8
1 ,326
1,989
1,200
(8 Positions)
1,500
(7,500 KW)
(4 TPH)
(97 MM Btu/Hr)
241 c
780
112
-
7,148
$8.94/Ton
90%
94%
PO
,4s.
Requirements for 100 TPH coal feed
Requirements for 200 TPH combined coal feed
Based on evaporation to reject waste heat
Coal yield values reflect the heating requirements of the process
-------
1) The pit reactor, continuous reactor and fine coal processes
are 100 TPH processing facilities for 3.2 percent pyritic
sulfur coal while the two float-sink options are 200 TPH
units using cleaned coal.
2) The pit/float-sink and continuous/float-sink processes as-
sume the feed coal was precleaned to about 2 percent
pyritic sulfur prior to gravity separation. The resultant
gravity separated fractions are assumed to contain 3.8
percent pyritic sulfur (Meyers Process treated) and less
than 0.5 percent pyritic sulfur (by-passed around the
Meyers Process).
The coal yield data (weight basis) presented in Table 33 indicates a
higher yield for the combined float-sink options (92 percent) than for the
pit and continuous reactor processes options (85 percent). This is based
on the design criteria that all of the material bypassed around the Meyers
Process, 50 percent of the 200 TPH feed, is 100 percent recovered and
I
blended with coarse coal Meyers Process product. The coal yield in terms
of energy recovery is also based on the above stated design criteria.
5.3 PROJECTION OF PROCESS ECONOMICS
Several Meyers Process options were evaluated to determine their over-
all process economics. The processing options evaluated included both fine
and coarse coal approaches (Sections 5.1 arid 5.2) integrated into grass roots
coal desulfurization facilities. This section presents a description of the
various integrated process options, a discussion of the economics evaluation
approach, and results of the analyses.
5.3.1 Processing Option Description
The integrated desulfurization facilities include all battery limits
Meyers Process equipment in either fine or coarse coal treatment configura-
tion, discussed in Sections 5.1 and 5.2 plus the required off-sites. The
off-sites include such items as:
215
-------
• Feed and product coal storage, handling and transport equipment.
• Physical coal cleaning facilities and size separation equipment.
• By-product handling and storage facilities.
• Waste treatment (physical cleaning and process generated)
and storage facilities.
t Process water treatment, storage and pumping facilities.
• Cooling water treatment and pumping equipment.
• Power and steam generation facilities.
• Site office buildings and shop structures.
• Other site improvements such as roads, fences, railroad
spurs, etc.
It should be noted that the economic evaluations do not include land costs
and assume that oxygen is purchased as an over-the-fence utility item (i.e.,
neither battery limit nor off-site equipment include an oxygen plant).
For purposes of economic evaluation, four differing process configura-
tion cases were developed. The central basis for each case was a 100 ton
per hour Meyers Process unit. A brief description of each case is presented
below and block diagrams containing simplified mass balances for each case
are presented in Figure 34.
Case 1 - Cleaned Fine Coal Case
/
Run-of-mine (ROM) coal is physically cleaned and then reduced to
14 mesh top size. The unit feed rate is 120 tons per hour of coal con-
taining about 20 percent ash and 3 to 4 percent pyritic sulfur. The
cleaning plant refuse (20 tons per hour) is assumed to contain approxi-
mately 75 percent ash and 10 to 14 percent pyritic sulfur. The Meyers
216
-------
CASE 1 - CLEANED FINE COAL (14 MESH TOP SIZE)
FEED COAl
120 T/HR COAL
20% ASH
3 - 4% PYRITIC
SULFUR
23.6 x 106
MM BTU/YR
ASH D
PHYSICAL
CLEANING
ISCARD
20 T/HR COAL
75% ASH
10 - 14% PYRITIC SULFl
1.4x 10 MM BTU/YR
100 T/HR COAL
9% ASH
1.5 -2% PYRITIC SULFUR
22. 2 x 106MM BTU/YR
JR
MEYERS PROCESS
FINE COAL
CONFIGURATION
COAL PRODUCT
93 T/HR COAL
6% ASH
.1% PYRITIC SULFUR
21. 3 x 106 MM BTU/YR
CASE 2 - RUN-OF-MINE COARSE COAL (1/4 IN. TOP SIZE)
FEED COAL ^
100 T/HR COAL
20% ASH
3 - 4% PYRITIC SULFUR
19.7x 106 MM BTUAR
MEYERS PROCESS
COARSE COAL
CONFIGURATION
COAL PRODUCT
85 T/HR COAL
15% ASH
.2% PYRITIC SULFUR
17.5x 106 MM BTU/YR
ro
CASES 3 AND 4, - DEEP CLEANED FINE AND COARSE COAL WITH 50% MEYERS PROCESS BYPASS
FEED COAL
240 T/HR COAL
20% ASH
3 - 4% PYRITIC SULFUR
47. 2 x 106 MM BTU/YR
ASH D1S
PHYS
CLEAt
CARD
ICAL
>
-------
Process feed consists of 100 tons per hour of coal containing approximately
9 percent ash and 1.5 to 2 percent pyritic sulfur. The fine coal Meyers
Process configuration utilized in this case is essentially that previously
described in Section 5.1.2 with the exception that the reaction and filtra-
associated equipment requirements are significantly reduced due to expected
reaction rate improvements (2 to 3 times ROM fine coal processing base
case described in Section 5.1.1) and lowered ash contents. The processing
uses about 100 MM Btu/hr equal to about 4 TPH and the pyrite removed equals
3 TPH. Therefore the product rate is 93 tons per hour"of coal containing
about 6 percent ash and 0.1 percent pyritic sulfur (93-95 percent removal
of pyritic sulfur).
Case 2 - ROM Coarse Coal Case
ROM 1/4 inch top size coal is fed directly to a coarse coal continuous
Meyers Process (described in Section 5.2). The coal feed rate is 100 tons
per hour and the coal consists of 20 percent ash and 3 to 4 percent pyritic
sulfur. Processing removes about 6 TPH of pyrite and requires about 9 TPH
of coal to supply the 250 MM Btu/hr required for process heat. The pyritic
sulfur is approximately 90-95 percent removed yielding a product rate of
85 tons per hour of coal containing 15 percent ash and about 0.2 percent
pyritic sulfur.
Case 3 - Deep Cleaned Fine Coal with 50 Percent Meyers Process Bypass
ROM coarse coal containing 20 percent ash and 3 to 4 percent pyritic
sulfur is fed to a physical cleaning plant at a rate of 240 tons per hour.
The ash discard (40 tons per hour) contains 75 percent ash and 10 to 14
percent pyritic sulfur. The cleaned coal containing 9 percent ash and 1.5
to 2 percent pyritic sulfur is fed to a gravity separation unit at a rate
of 200 tons per hour. The heavy fraction consists of 100 tons per hour of
coal containing 15-18 percent ash and 3 to 4 percent pyritic sulfur while
the light portion consists of 100 tons per hour of coal with low ash and
pyritic sulfur content. The heavy fraction then reduced to 14 mesh top
size and fed to a fine coal Meyers Process unit (described in Section 5.1)
which produces 90 tons per hour of product coal containing 10-13 percent
ash and 0.2 percent pyritic sulfur. The Meyers Process product when
218
-------
blended with the bypassed light coal fraction yields a combined product
stream of 190 tons per hour of coal containing about 6 percent ash and
0.2 percent pyritic sulfur (overall 90-95 percent removal of pyritic sul-
fur from feed coal).
Case 4 - Deep Cleaned Coarse Coal with 50 Percent Process Bypass
Two hundered and forty tons per hour of 1/4 inch top size ROM coal
is treated as previously discussed in Case 3. The only difference in the
treatment scheme is that a continuous coarse coal Meyers Process configura-
tion (discussed in Section 5.2) is utilized. The coarse coal process
requires more coal for internal process heat and therefore yields only
185 TPH of product.
5.3.2 Economi cs Model Descri pti on
The technique used in capitalizing the process options (Cases 1, 2, 3
and 4) is a primary determinant in calculating the unit price of the product.
Both utility financing and investor financing methods were considered in
determining the price of the product coal.
The method used to determine the product cost for the different
financing techniques was based on the technique used by the FPC Synthetic
Gas-Coal Task Force in their report on synthetic gas. The applicable
equations used in this analysis are summarized below.
The investor capitalization method used in this analysis was the
"Discounted Cash Flow" (DCF) financing method with an assumed discounted
cash flow rate of return of 20 percent after taxes. In essence, this
technique determines the annual revenue during the plant life which will
generate a discounted cash flow equal to the total capital invested for
the plant. For this analysis, it was assumed that the total capital re-
quirement was used prior to the plant start-up. Other assumptions used in
this analysis include the following:
219
-------
• Plant life was assumed to be 20 years with no cash value at
the end of life.
• A straight line method was used to calculate annual depre-
ciation.
• Operating costs and working capital requirements were
assumed to be constant during the life of the process.
• Annual revenue and production were assumed to be constant
during plant life (i.e., no start-up period).
t The present value of the investment includes the cost of
capital during the construction period. The construction
period cost of capital, Ic, was approximated by the
following expression:
Ic = 1.875 (O.Olr) (C-W),
where r is the discounted rate of return (percent), C is
the total capital requirement ($MM) and W is the annual
working capital ($MM).
• 100 percent equity capital .
Using the above assumptions to modify the basic equations, the required
annual revenue requirements for the DCF method can be calculated from the
expression given for total capital requirements:
where ,
C = Total capital requirement, $MM,
Rj = Annual revenue requirement for DCF investor financing, $MM,
D = Annual depreciation, $MM,
220
-------
FAn = Annuity factor for the life of the process at the DCF rate
of return,
W = Annual working capital, $MM,
f = Single payment present value factor at the DCF rate of
return for a single quantity at the end of the plant life,
N = Net annual operating costs of the plant operation (exclud-
ing depreciation), and
T = Annual income tax rate, percent.
As a regulated industry, the utility industry has different capitali-
zation, tax and return-on-investment requirements which differ significantly
from a nonregulated industry. The basic equations derived in Appendix I
of the referenced document were used along with all of the applicable in-
vestor financing assumptions listed above. In addition, the following
assumptions were made for the utility financing calculations and were
obtained directly from Appendix I of the FPC report:
• The debt/equity ratio was assumed to be 75/25.
• The interest on the debt was assumed to be 9 percent.
t The required return on the equity was assumed to be 15 percent.
• The corresponding return on rate base is 10.5 percent.
Because the rate base for utility financing varies during the life of
the plant, this would result in a variable revenue flow (and subsequent coal
price) for each year of operation. For this analysis, an average coal price
during the plant life was calculated to provide a basis of comparison.
Based on this assumption, the average annual revenue requirement could be
calculated from the following expression;
221
-------
Ru = N + 0.05 (C-W) + 0.01 Fp + TQQ I T (l-d)ru IjC - 0.5 (C-W)J>
where N, C, W and T are the same as for the investor financing, and
R = Average annual revenue requirement for utility financing.
p = Return on rate base, percent per year,
d = Debt fraction (0.75), and
r = Return on equity, percent per year.
The capitalization requirements were obtained from integrated Meyers Process
cost estimates which were based on June 1975 dollars.
All process capital cost data are based on the detailed cost estimates
for battery limits Meyers Process units sized for 100 tons per hour of coal
feed (presented in Sections 5.1 and 5.2). The battery limits cost for each
specific economics case was adjusted to account for the varying coal input
compositions and characteristics. For instance, the battery limits fine
coal Meyers Process unit cost estimated for Case 1, is significantly less
($5 MM less) than that presented in the base case cost estimate (Section
5.1.2.2). This cost adjustment is warranted due to increased reaction rates
(2 to 3 times base case), faster filtration rates observed for cleaned coal
and lower circulating gas rates in the sulfur vaporization section. These
increased rates lead to less capital investment in reactor, filter and sul-
fur removal associated operations.
Off-site investment for each of the four cases was estimated as a
percentage of the battery limit cost. The basic off-site investment was
then adjusted for special characteristics of the case. For fine coal pro-
cessing, the basic off-site investment was estimated at 50 percent of the
battery limits cost given in Table 26 of Section 5.1 or $6.63 million. It
was expected that Case 1 would have a reduction in off-site capital relative
to the base case resulting from less by-products, waste disposal, power and
steam. The reduction was estimated to equal the $0.8-$1.0 million cost for
the physical cleaning. Case 2 off-sites were estimated to be 100 percent
of the battery limits cost given in Section 5.2 or $4.2 million. Case 3 has
222
-------
a 240 TPH physical cleaning section and a gravity separation section and
grinder with a total estimated capital cost of $2.5 million. This added
to the basic process off-site gives a total off-site cost of $9.13 million.
Case 4 has the physical cleaning train and gravity separation without
grinding. This gives a total off-site capital estimate of $6.5 million.
Operating costs were estimated from base case operating costs. Ad-
justments were made to account for such things as varying oxygen and chemicals
(lime and scrap iron) requirements due to varying Meyers Process feed pyritic
sulfur contents. The operating labor requirement was adjusted to properly
reflect the varying complexities represented by the four cases. Disposal
costs were estimated for each case to reflect the differing requirements
(primarily due to the use of physical cleaning in some cases and not in
others and Meyers Process operation with coals of varying pyritic sulfur con-
tent). Disposal costs were estimated to be $5.90 per ton of refuse.
All other capital related and operating related costs were determined
as percentages of estimated battery limit costs, off-site equipment costs,
raw materials costs, and labor costs. Table 34 presents a summary of the
economic evaluation criteria.
The economic evaluation criteria given in Table 34 and the annualized
revenue expressions for investor and utility financed process plants
described in this section can be combined to give an expression of the
following form:
P = (aX + bY + cZ)/d
where,
P = is the required sales price for processed coal, $/MM Btu
X = working capital for raw materials and supplies, $
Y = sum of the total plant investment and start-up cost, $
Z = annual total operating cost, $/year
a, b, c = constants given in Table 35
d = annual energy output, MM Btu/yr
223
-------
TABLE 34. ECONOMIC EVALUATION CRITERIA
Operating Cost Criteria
Raw material - coal @ $10/ton, $20/ton, $30/ton
Utilities
Electricity @ 2.5£/kw-hr
Oxygen @ $25/ton
Cooling water @ 5<£/1000 gal.
Process water @ 25<£/1000 gal.
Chemicals
Lime @ $28/ton
Scrap iron @ $50/ton
Disposal @ $5.90/tpn of refuse
Labor
Process operating labor @ positions/shift x 8304 man-hours/yr x $5/man-hour
Maintenance labor @ 1.5%/yr of total plant investment
Supervision @ 15% of operating labor and maintenance labor
General overhead and administration at 60% of labor and supervision
Supplies
Operating @ 30% of process operating labor
Maintenance @ 1.5% of total plant investment
Taxes and insurance @ 2.7%/yr of total plant investment
Total operating cost = sum of raw material + utilities + chemicals + disposal
+ labor + supplies + taxes and insurance
Capital Cost Criten'a
Battery limits capital as discussed in text
Off-site capital as discussed in text
Overhead and profit @ 22% of battery limits + off-sites
Engineering and design @ 10% of battery limits + off-sites
Contingency @ 15% of battery limits + off-sites + overhead and profit +
engineering and design
Total plant investment = sum of battery limits + off-sites + overhead and
profit + engineering and design + contingency
Interest for construction @ 9% of total plant investment x 1.875
Start-up cost @ 20% of total operating costs
Working capital = sum of raw materials inventory of 60 days at full rate +
materials and supplies at 0.9% total plant investment + 1/24 annual
product revenue
224
-------
TABLE 35. CONSTANTS FOR USE IN ECONOMICS EVALUATIONS
a -
b
c
Investor Financing
.391
.505
1.016
Utility Financing
.140
.121
1.006
5.3.3 Process Economics Evaluations
The process economics evaluations were carried out as described in the
previous section. The results of those evaluations are presented in Tables
36, 37, 38, and 39. Calculations were performed at three assumed ROM coal
costs and using both utility and investor financing criteria. Coal costs of
$10 per ton, $20 per ton and $30 per ton were selected since they represent
the broad range of currently reported ROM coal costs ($10 per ton at mine
mouth to $25 .per ton reported delivered price at some plant sites). As
may be determined from the data, the required market value of the treated
coal ranges from a low of 66<£/MM Btu for Case 4 (assuming $10/ton ROM coal
cost and utility financing) to a high of $2.39/MM Btu for Case 1 (assuming
$30/ton ROM coal cost and investor financing). In all cases, utility finan-
cing yielded market costs on the order of 60 percent to 80 percent of market
costs required when utilizing investor financing.
An equivalent upgrading cost has also been determined. The upgrading
cost is found by deducting the cost of the dirty energy ($.41/MM Btu for 20%
ash, 3-4 percent sulfur coal at $10/ton; $.81 at $20/ton and $1.22 at $30/
ton) from the cost shown in Tables 36 through 39 for the clean energy ($.66
to $2.39/MM Btu). These upgrading or processing costs, which range from
$.25/MM Btu to $1.17/MM Btu, are presented in Table 40. For all cases the
processing cost includes ash reduction as well as sulfur reduction. Except
for Case 2, physical cleaning was assumed to be coupled with pyrite removal
which results in a major reduction in ash from about 20 percent to about
6 percent.
225
-------
TABLE 36. CASE 1, CLEANED FINE COAL
Product Annual
Energy Value, 21.3 x 10 MM Btu/yr
Capital Related Requirements, $MM
Battery Limit Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total Plant Investment^ '
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Working Capital (Investor Financing)
Total Capital Related Costs (Utility)
Total Capital Related Costs (Investor)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Scrap)
Supplies
Disposal
Utilities
Labor (13 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
\
Utility Financing
Investor Financing
$10/Ton
8.26
6.63
3.28
1.49
2.95
22.61
3.82
2.89
2.54
2.98
31.86
32.30
9.60
.06
.50
.42
1.63
1.62
.61
14.44
.84
1.33
ROM Coal Cost
•$20/Ton
8.26
6.63
3.28
1.49
2.95
22.61
3.82
4.81
4.57
5.05
35.81
36.29
19.20
.06
.50
.42
1.63 ,
1.62
.61
24.04
1.31
1.86
$30/Ton
8.26
6.63
3.28
1.49'
2.95
22.61
3.82
6.73
6.59
7.12
39.75
40.28
28.80
.06
.50
.42
1.63
1.62
.61
33.64
1.79
2.39
(1)
Equivalent to a plant capital investment of $79.60/kw.
226
-------
TABLE 37. CASE 2, ROM COARSE COAL
Product Annual fi
Energy Value, 17.5 x 10b MM Btu/yr RQM
Capital Related Requirements, $MM $ IP/Ton $20/Ton $30/Ton
Battery Limit Capital 4.20 4.20 4.20
Off-site Capital 4.20 4.20 4.20
Overhead and Profit 1.85 1.85 1.85
Engineering and Design .84 .84 .84
Contingency 1.66 1.66 1.66
Total Plant Investment^1) 12.75 12.75 12.75
Interest for Construction 2.15 2.15 2.15
Start-up Costs 2.51 4.11 5.71
Working Capital (Utility Financing) 2.06 3.74 5.42
Working Capital (Investor Financing) 2.33 4.05 5.77
Total Capital Related Costs (Utility) 19.47 22.75 26.03
Total Capital Related Costs (Investor) 19.74 23.06 26.38
Operating Costs, $MM/Yr
Raw Material (Coal) 8.00 16.00 24.00
Chemicals (Lime, Scrap) .39 .39 .39
Supplies .30 .30 -30
Disposal -64 .64 -64
Utilities 1-86 1.86 1.86
Labor (9 Positions) 1-02 1.02 1.02
Taxes and Insurance -34 -34 .34
Total Operating Costs 12.55 20.55 28.55
Required Coal Market Price, $/MM Btu
Utility Financing
Investor Financing
Utility Financing -84 1>32 Il8°
1.20 T-74 2.28
(1)
Equivalent to a plant capital investment of $54.60/kw.
227
-------
TABLE 38. CASE 3, DEEP CLEANED FINE COAL
WITH 50% MEYERS PROCESS BYPASS
Product Annual
Energy Value, 43.4 x 10° MM Btu/yr
Capital Related Requirements, $MM
Battery Limit Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total Plant Investment^ '
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Working Capital (Investor Financing)
Total Capital Related Costs (Utility)
Total Capital Related Costs (Investor)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Scrap)
Supplies
Disposal
Utilities
Labor (14 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
Utility Financing
Investor Financing
IIP/Ton
13.26
9.13
4.93
2.24
4.43
33.99
5.74
5.28
4.83
5.51
49.84
50.52
19.20
.11
.69
.84
2.62
2.01
.92
26.39
.73
1.11
ROM Coal Cost
$20/Ton
13.26
9.13
4.93
2.24
4.43
33.99
5.74
9.12
8.88
9.66
57.73
58.51
38.40
.11
.69.
.84
2.62
2.01
.92
45.59
1.20
1.63
$30/ Ton
13.26
9.13
4.93
2.24
4.43
33.99
5.74
12.96
12.92
13.80
65.61
66.49
57.60
.11
.69
.84
2.62
2.01
.92
64.79
1.66
2.15
(1)
Equivalent to a plant capital investment of $58.70/kw.
228
-------
TABLE 39. CASE 4, DEEP CLEANED COARSE COAL
WITH 50% MEYERS PROCESS BYPASS
.(1)
Product Annual <-
Energy Value, 42.2 x 10° MM Btu/yr
Capital Related Requirements, $MM
Battery Limit Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total Plant Investment^
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Working Capital (Investor Financing)
Total Capital Related Costs (Utility)
Total Capital Related Costs (Investor)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Scrap)
Supplies
Disposal
Utilities
Labor (11 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price. $/MM Btu
Utility Financing
Investor Financing
$10/Ton
4.20
6.50
2.35
1.07
2.11
16.23
2.74
4.92
4.51
4.89
28.40
28.78
19.20
.39
.38
.96
1.96
1.29
.44
24.62
.66
.88
ROM Coal Cost
$20/Ton
4.20
6.50
2.35
1.07
2.11
16.23
2.74
8.76
8.55
9.04
36.28
36.77
38.40
.39
.38
.96
1.96
1.29
.44
43.82
1.14
1.41
$30/Ton
4.20
6.50
2.35
1.07
2.11
16.23
2.74
12.60
12.59
13.18
44.16
44.75
57.60
.39
.38
.96
1.96
1.29
.44
63.02
1.62
1.95
0)
Equivalent to a plant capital investment of $28.80/kw.
229
-------
TABLE 40. UPGRADING (PROCESSING) COSTS
Case 1
Utility Financed
Investor Financed
$10/Ton
$.43/MM Btu
$.92/MM Btu
ROM Coal Cost
$20/Ton
$.50/MM Btu
$1 .05/MM Btu
$30/Ton
$.57/MM Btu
$1.17/MM Btu
Case 2
Utility Financed
Investor Financed
$.43/MM Btu
$.79/MM Btu
$.51/MM Btu
$.93/MM Btu
$.58/MM Btu
S1.06/MM Btu
Case 3
Utility Financed
Investor Financed
$.32/MM Btu
$.70/MM Btu
$.39/MM Btu
$.82/MM Btu
$.44/MM Btu
$.93/MM Btu
Case 4
Utility Financed
Investor Financed
$.25/MM Btu
$.47/MM Btu
$.33/MM Btu
$.60/MM Btu
S.40/MM Btu
$.73/MM Btu
230
-------
6. CHEMICAL ANALYSIS STUDIES
The chemical analysis studies had two objectives: (1) to experimen-
tally investigate the adequacy of established coal sulfur analysis techniques
for the accurate determination of Meyers Process performance and (2) to iden-
tify and screen potential Meyers Process efficiency monitoring techniques
for use in full-scale operations. The need for these studies was dictated
by frequent inconsistencies in sulfur analysis data derived from processed
coal and by the desirability to identify process monitoring techniques which
are simpler and faster than the standard ASTM coal sulfur analysis methods.
In the Meyers Process the pyritic sulfur in coal is oxidized to elemen-
tal sulfur and sulfate by the reduction of ferric ions to the ferrous state;
the latter are regenerated to the ferric state by oxygen. The sulfate sulfur
product dissolves in the reagent solution. The elemental sulfur can be
recovered by extraction with hydrocarbon solvents (e.g., toluene or naphtha)
or by vapo>ization; in bench-scale experimentation hydrocarbon leaching was
used almost exclusively. Thus, Meyers Process chemistry permits, in principle,
the use of three independent approaches for the estimation of process per-
formance all of which involve standardized chemical analysis techniques.
The most direct approach is the determination of pyrite in coal before and
after processing. The second approach involves the determination of total
sulfur in coal before and after processing. The third approach is based on
the estimation of ferrous ion production during leaching. Since 10.2 moles
of ferrous ion are produced per mole of pyrite leached, this approach of
measuring the performance of the process is very accurate provided the
leaching reaction is totally selective, as it is with most coals, or ferrous
ion production from side reactions can be independently estimated. This
approach is also an excellent means of monitoring process performance during
processing because it is simple and quick. Unfortunately, it is only useful
when the leaching and regenerations operations are performed separately
whereas simultaneous Teaching-regeneration is contemplated for most coals.
In principle, the first two approaches are equivalent since the Meyers
Process should not affect the organic sulfur in coal and it should have
231
-------
insignificant effect on the sulfate sulfur concentration of unweathered
coal which normally is 0.1 wt. percent or less. In practice, direct
equivalency of these two approaches has been very inconsistent and infre-
quent. Probable causes for the observed inconsistencies may be grouped
into three categories: (1) performer inconsistencies in the analyses of
coal, a very real possibility with processed coal, the sulfur content of
which is low, (2) incomplete recovery of process sulfur products or high
sulfate starting coal (weathered coal), and (3) inadequacy of standard coal
sulfur analyses techniques to accurately determine the sulfur and sulfur
forms content of processed coal (Meyers Process effects on sulfur analysis
techniques). Though much of the evidence from hundreds of coal analyses
pointed to performer inconsistency as the most probable cause for incon-
sistencies in analysis data, there was sufficient uncertainty in such
conclusion to warrant investigation of probable Meyers Process effects on
standard sulfur analysis techniques.
The investigation of the coal sulfur analysis problems was combined
with the effort to identify efficient process monitoring techniques for use
in full scale Meyers Process plants. The results and conclusions are pre-
sented in separate sections below.
6.1 ADEQUACY OF STANDARD SULFUR ANALYSIS TECHNIQUES FOR DETERMINING
MEYERS PROCESS PERFORMANCE
The total sulfur in ROM, run-of-the-mine, coal is comprised of three
sulfur forms: pyritic, sulfate, and organic sulfur. Product coal from
the Meyers Process may contain only one or as many as four sulfur forms.
Complete leaching of pyrite, complete recovery of process sulfur products
(elemental sulfur and sulfate), and dissolution of the sulfate in the ROM
would result in processed coal the total sulfur of which is equal to the
organic sulfur present in the starting coal. Less than 100 percent leaching
of pyrite and sulfate and incomplete recovery of the elemental sulfur product
would result in processed coal with four sulfur forms. Pyritic sulfur and
sulfate sulfur can be determined directly by standard ASTM techniques. Or-
ganic sulfur and unrecovered elemental sulfur are computed as a single value
232
-------
from the difference between the total sulfur content of the coal, deter-
mined by Eschka analysis, and the sum of the pyritic and sulfate sulfur
values.
The Meyers Process affects only the pyritic sulfur in coal directly
unless the coal is severely weathered in which case sulfate removal also
occurs during processing. Indirectly, however, the Meyers Process may
affect all three sulfur forms in coal because of incomplete sulfur product
recovery and because elemental sulfur is computed as organic sulfur.
Pyritic sulfur is always reduced; sulfate sulfur remains unchanged or in-
creases slightly during processing, depending on processing conditions,
when low sulfate ROM coal is used; organic sulfur remains unchanged unless
elemental sulfur recovery is incomplete in which case the "organic" sulfur
content of the processed coal should always be higher with the increase
being equal to unrecovered elemental sulfur. Thus, in the cases where
sulfur removals computed from total sulfur and pyritic sulfur analyses do
not agree, the delta in the two sulfur removal values should equal the
quantity of unrecovered sulfur product. Complementing the total and pyritic
sulfur analyses by the determination of the sulfate sulfur in the sample
should lead to the quantitative identification of each of the sulfur forms
left on the coal.
The data in Appendices A and B, Volume 2 show that the coal sulfur
analyses were infrequently consistent with the above expectations. The
discrepancies are manifested in the value of organic sulfur, and they were
obviously caused by inaccuracies in analyses either of total sulfur or of one
or both of the sulfur forms (pyrite and sulfate). As indicated earlier,
the extensive experimentation on the Meyers Process with Lower Kittanning
coals has not shown any evidence that the process affects the organic sulfur
content of the coal. Thus, indicated reductions in the organic sulfur
of coals during processing must be due to errors in analyses. The
same conclusion is drawn from coal samples exhibiting relatively large
increases in organic sulfur which could not be justified in terms of
unrecovered elemental sulfur.
233
-------
The discrepancies are in the 0.1 to 0.5 wt. percent sulfur range and
therefore, of relatively minor importance to the measurement of Meyers
Process efficiency when applied to high sulfur, high pyrite coals as the
one used in this program (though they are very important in determining
whether the processed coal met a specific sulfur standard or not). These
discrepancies become very important in the determination of Meyers Process
efficiency and rates as the pyrite content of the feed coal to the process
decreases (e.g., pyritic sulfur in the 1 to 2 wt. percent range).
Problems with coal sampling and analysis have been encountered by
virtually everyone requiring accurate coal composition data on a regular
basis. The consensus of opinion (from mine quality control personnel to
Bureau of Mines chemists) appears to be that the standard ASTM procedures
for coal analysis are adequate when properly utilized, but that the con-
sistency of accurate data suffers when these techniques are applied rou-
tinely in an assembly line operation (necessary for holding analysis costs
to reasonable levels). The purpose of this task was to investigate whether
or not the Meyers Process aggravated these problems by direct effects on
the accuracy of standard sulfur analysis procedures. In the case of total
sulfur analyses the concern centered on potential effects from organic
solvents used to recover elemental sulfur and from unrecovered elemental
sulfur itself. In the case of sulfur forms the concern was whether the
reflux times specified by the ASTM procedures for sulfate and pyrite re-
covery were adequate for use on chemically leached coal whose sulfate and
pyritic sulfur values are approximately equal and very low.
Three different techniques were used in the total sulfur coal analysis
investigations: Eschka, Bomb Wash, and Leco (Eschka being the standard ASTM
technique). The rationale was that since these three sulfur analysis techni-
ques differ substantially in procedure it would be very unlikely that the
Meyers Process could have the same effect on the accuracy of each of them,
if indeed processed coal by the Meyers Process affected established sulfur
analysis techniques. Eschka and Bomb Wash are commonly used sulfur analysis
techniques for coal; Leco was selected as the third method because it is an
adequately developed sulfur analysis technqiue and because it offers the
potential for use as a Meyers Process monitoring technique.
234
-------
The matrix of experiments began with comparison of the three total
sulfur analysis methods on NBS certified coal samples ranging in sulfur
values from 0.546 to 3.020 wt. percent. The three analyses methods were
then used on:
t Run-of-the-mine, ROM, Lower Kittanning coal samples from
the coal utilized in this program.
• L-R processed coals at two different reaction times.
• Organic solvent treated L.K. coals (the coal was refluxed
for one hour in either hexane, or heptane, or toluene
without prior leaching with ferric sulfate).
• ROM and organic solvent treated coals doped with 2 to 3
wt. percent elemental sulfur.
The data generated from these experiments are summarized in Table 41.
The first group of data in Table 41 compares NBS certified total sulfur
values to total sulfur analyses performed on the same coal using Eschka,
Bomb Wash, and Leco techniques. The ASTM acceptable reproducibility
spreads for total sulfur analyses performed on the same coal at two dif-
ferent laboratories are 0.1 wt. percent sulfur for coal containing 2 wt.
percent sulfur or less and 0.2 wt. percent sulfur for coal with higher than
two percent sulfur content. All three techniques yielded values within
the ASTM allowable spreads. Thus, any one of these techniques may be
utilized for the determination of sulfur concentration in ROM coals up to
at least 3 wt. percent. It should be noted, however, that the Eschka
technique yielded values consistently closer to the certified sulfur values.
At coal sulfur concentrations in the 2 to 3 wt. percent range the Bomb Wash
and Leco technique values indicate a low bias with Leco furnishing the
lowest sulfur concentrations. On the basis of these observations, the
Eschka technique must be assumed to be the most accurate of the three for
analyzing high sulfur coal. However, if the NBS samples were certified
by the Eschka technique the above conclusion may be unwarranted.
235
-------
TABLE 41. COMPARISON OF TOTAL SULFUR ANALYSIS TECHNIQUES
(WEIGHT PERCENT SULFUR IN COAL)
I. NBS CERTIFIED COALS:
Certified Values Eschka Bomb Wash Leco
0.546 + 0.003 0.54 +_ 0.02 0.55 +_ 0.
2.016 + 0.014 2.03+^0.04 1.92+_0.
3.020 + 0.008 2.94+^0.03 2.93 +_ 0.
II. RON (RUN-OF-KINE) LOWER KITTANNING COAL:
4.43 + .06 4.24
III. L-R PROCESSED LOWER KITTANNING COALS:
Experiment No.
S2 (Toluene*) 1.44 + 0.005 1.52+_0.
S3 (Toluene*) 1.49 1.53 + 0.
S4 (Toluene*) 2.37 + 0.01 2.39 + 0.
S5 (Hexane-Toluene**) 2.22 +_ 0.04 2.34 + 0.
IV. ORGANIC SOLVENT TREATED ROM LOWER KITTANNING COAL:
Solvent Used
00 0.59 +_0.04
01 1.90+0.03
03 2.86 +_ 0.01
4.08 +0.06
05 1.54^0.02
01 1.54 + 0.01
03 2.36 +_ 0.005
07 2. 33 +_ 0.06
Hexane 4.37 4.28 3.84+0.11
Heptane 4.48 +_ 0.05 4.30 4.19 + 0.08
Toluene 4.32 + 0.01 4.23 3.95 +_ 0.03
V. ROM LOWER KITTANNIHG COAL DOPED WITH ELEMENTAL SULFUR:
Sample Doped Bomb Wash Leco
Kith Elemental Calculated Analyzed Calculates Analyzed
Sulfur Sulfur Sulfur Sulfur Sulfur
ROM Coal 6.96 6.83 6.80 6.53+0.10
Hexane Refluxed 6.97 6.82 6.73 6.44+0.001
Heptane Refluxed 7.20 7.20 7.09 6.70+0.01
Toluene Refluxed 7.16 7.11 6.83 6.74+0.00
Eschka
Cal cul ate3 Analyzed
Sulfur Sulfur
7.15 6.96+0.03
7.16 6.77
7.39 7.10+0.06
7.35 7.00
% Recovery
Bomb Leco Eschka
Wash
98 96 97
98 96 95
100 94 96
99 99 95
Elemental sulfur on processed coal was extracted by toluene.
Elemental sulfur on processed coal was extracted first by hexane and then by toluene.
236
-------
The discrepancy among sulfur values generated by the three techniques
investigated in this program becomes meaningful when the total sulfur con-
centration in the coal exceeds 4 wt. percent, as demonstrated by the second
group of data in Table 41. These data were generated on identical samples
of ROM Lower Kittanning coal. The spread between Eschka and Leco generated
sulfur values is unacceptable (too large to be considered a normal deviation
between values of valid sulfur analysis techniques). A careful examination
of the procedures prescribed for the Bomb Wash and Leco techniques did not
result in the identification of inherent technique shortcomings which could
predict the apparent bias toward lower sulfur values. However, small
systematic errors in the application of these procedures and defective equip-
ment or operator biases could not be completely discounted as probable
causes for the observed low sulfur value bias. On the basis of the available
data it must be concluded that the Leco technique can not be used with
confidence to determine the sulfur content of coals when it exceeds 3 wt.
percent. At lower coal sulfur concentrations the Leco technique exhibited
adequate accuracy and excellent reproducibility to render it useful in
monitoring the efficiency of the Meyers Process.
The data in Sections III through V in Table 41 represent the results
of studies performed to determine if the Meyers Process affects the accuracy
of total sulfur determinations on coal. Several samples of L-R processed
Lower Kittanning coal from four experiments were analyzed for total sulfur
content by each of the three sulfur analysis techniques. The data are
summarized in Section III of Table 41. The data show that if the Meyers
Process affects the accuracy of total sulfur determinations, its effect is
virtually identical on all three techniques. This was considered unlikely
and it was, therefore, concluded that the sulfur content of processed coal
can be accurately determined by any of the three techniques, if properly
used. The validity of this conclusion was verified further by the data in
Sections IV and V.
Organic solvents and elemental sulfur are the only substances poten-
tially present in coal subjected to the Meyers Process and not present in
the raw coal (the former would be present only if the product elemental
sulfur is recovered through organic solvent leaching of the processed coal).
237
-------
Thus, it was decided to investigate whether organic solvents of potential
use to the Meyers Process and whether elemental sulfur not completely re-
covered from processed coal could measurably affect the accuracy of coal
sulfur analysis techniques. Samples of ROM Lower Kittanning coal (Section
II, Table 41) with known sulfur content were refluxed for one hour in each
of the three organic solvents listed in Table 41 and subsequently analyzed
by each of the three total sulfur analysis techniques. Comparison of the
data in Sections II and IV reveals that these solvents did not affect the
accuracy of sulfur analyses. In separate experiments, ROM coal samples
and organic solvent refluxed samples were doped with known quantities of
elemental sulfur and then reanalyzed. The data in Section V demonstrate
that the elemental sulfur added to the coal was accurately reflected by the
total sulfur analyses performed on the coal samples regardless of the
analysis technique used.
In summary, the total sulfur analysis investigations revealed that
the Meyers Process does not affect the accuracy of coal total sulfur de-
terminations by Eschka, Bomb Wash, or Leco techniques. Within acceptable
deviations, these three analysis techniques proved equally accurate in the
determination of coal sulfur concentrations up to approximately 3 wt. per-
cent. At higher coal sulfur concentrations the Leco technique demonstrated
an unacceptable low bias, yielding sulfur values which were lower by as
much "as 10 percent than those obtained by Eschka; however, it was not
established conclusively whether the low values were inherent to the tech-
nique or due to systematic errors in its utilization.
The sulfur forms investigations led also to the conclusion that the
Meyers Process does not affect the accuracy of standard coal analysis
techniques. In these investigations potential interferences of elemental
sulfur on sulfate analyses and sulfate on pyrite analyses were examined.
In the case of elemental sulfur the concern was that, if incompletely
recovered, it could be physically leached during the hydrochloric acid
extraction and subsequently be oxidized to sulfate in the process of sulfate
analysis. Experiments performed on elemental sulfur derived from the
238
-------
Meyers Process as well as on coal doped with elemental sulfur showed no
evidence that this sulfur form oxidizes to sulfate during the standard ASTM
procedure for coal sulfate analysis. In addition, sulfur balance data de-
rived from bench-scale experimentation on the Meyers Process indicated that
elemental sulfur recovery was virtually complete in the majority of experi-
ments and, therefore, could not account for the apparent inconsistencies in
sulfur forms analyses (Section 2, Volume 1 and Appendices A, B, and C,
Volume 2 of this report).
A more serious concern was that associated with potential Meyers Process
effects on pyrite analysis data, either because of sulfate interference or
due to incomplete leaching of pyrite. According to ASTM procedure, pyrite
in coal is determined from the difference in the iron content of the nitric
acid and hydrochloric acid solutions derived by refluxing for 30 minutes
identical coal samples, except for size, in the respective acids. The acid
soluble iron compounds of coal, principally iron sulfates and oxides, are
leached by the hydrochloric acid; the nitric acid being an oxidizing agent
leaches all of the iron in coal which includes the pyritic iron. Normally,
the soluble iron in ROM coals represents a small percentage of the pyritic
iron; thus, incomplete leaching of the sulfate by hydrochloric acid intro-
duces an insignificant error to the determination of the pyrite content of
coal. In processed coal the pyrite content is low and substantial errors
in its determination could result from improper leaching of the iron forms
in coal, especially in cases where sulfate recovery is incomplete as it was
the case in a number of experiments performed during this program (Appendices
A and B, Volume 2). The concern was whether the acid reflux times prescribed
by ASTM procedures were adequate to remove the iron from coal to the in-
tended extent. The principal question was whether 30 minutes reflux time
in hydrochloric acid was sufficient for the leaching of process derived
sulfate product (higher concentration and conceivably different form of iron
sulfate than typically found in ROM coals).
A summary of the data generated on the effect of acid digestion time on
pyritic and sulfate sulfur analyses is presented in Table 42. ASTM and
modified ASTM analysis procedures were used in these investigations. The
239
-------
TABLE 42.
COAL SULFUR FORMS ANALYSIS INVESTIGATIONS - COMPARISON OF
TECHNIQUES AND EFFECT OF COAL DIGESTION TIME^a)
Sample
ROM (Weathered)
L. K. Coal
Processed Coal
(Exp. S2)
Processed Coal
(Exp. S4)
Processed Coal
(Exp. S2)
Doped with Pyrite
Digestion
(Reflux)
Time
0.5 Hours
1 Hour
2 Hours
3 Hours
0.5 Hours
1 Hour
2 Hours
3 Hours
0.5 Hours
1 Hour
2 Hours
3 Hours
1 Hour
2 Hours
3 Hours
Pyritic Sulfur, Wt. %
ASTM
2.66 + .12
2.66 + .04
2.63 + .07
2.56 + .04
0.39 + .02
0.44 + .01
0.41 + .02
0.44 + .02
1.00 + .06
0.94 + .04
1.01 + .02
1.07 + .02
(0.67}k 0.59
(0.66), 0.57
(0.66), 0.57
AA
2.76 +. .13
_
_
2.61 +. .02
0.38 + .01
0.40 + .04
0.39 + .06
0.34 + .05
1.00 + .00
0.91 + .04
0.91 + .04
1.07 + .02
(0.64)1? 0.63
(0.63), 0.50
(0.64), 0.59
Sulfate Sulfur, Wt. %
ASTM
0.61 + .03
0.62 + .01
0.64 + .01
0.67 +. .01
0.29 + .02
0.31 + .01
0.35 + .01
0.39 +_ .02
0.60
0.63 + .01
0.64 +\00
0.65 + .01
(0.60)^ 0.61
(0.60), 0.62
(0.60), 0.62
Turbidimetric
Peroxide - Bromine
0.14 + .03 0.15 + .02
0.18 + .05 0.14 + .05
0.24 + .08 0.24 +_ .01
0.16 + .01
<0.01 <0.01
0.16 + .01 0.19 + .02
0.04 +_ .01 0.04
0.27 + .00
0.24 + .02 0.27 + .04
0.30 + .03 0.28 + .01
0.19 + .02 0.31 + .04
0.41, 0.61
-------
modified techniques are simpler than the corresponding ASTM techniques and
in principle more accurate since they eliminate the subjectivity inherent
in the determination of titration end points. However, the data in Table
42 do not appear to justify the latter expectation. Both the ASTM and the
modified techniques require the acid digestion step. The modified tech-
niques used atomic absorption, AA instead of iron titration in pyrite
analyses and turbidimetric instead of gravimetric analyses in sulfate
determinations; two different oxidants were used in the turbidimetric
sulfate analyses. Comparison of ASTM and AA pyritic sulfur data reveals
that the two techniques yielded equivalent values with slightly better
reproducibility evident in data generated by the ASTM technique. In
multiple sample analyses, the AA technique may merit consideration. The
turbidimetric technique yielded consistently low values until a cation ex-
change column was used to remove interferences (last column of last three rows
of data in Table 42). The turbidimetric technique merits serious consider-
ation since it simplifies appreciably the determination of sulfate; however,
cation interferences must be eliminated which implies that the technique
may require frequent calibrations against gravimetrically derived data.
The first group of data in Table 42 were obtained from analyses per-
formed in triplicate on weathered ROM Lower Kittanning coal. Substantially
weathered coal was selected in order that its sulfate content be sufficiently
high to detect changes in its determination as a function of digestion time.
The data indicate a trent toward lower pyritic sulfur values and correspond-
ingly higher sulfate values with increasing digestion times. The trend
appears to be real, but the difference in the determined sulfur form values
in coal samples digested for 0.5 and 3 hours, respectively, is within
expected analysis reproducibility limits.
The second and third groups of data in Table 42 represent analyses
performed also in triplicate on coals processed by the Meyers Process. Two
coals processed under different conditions were selected; thus, the extent
of pyrite leaching and product sulfate recovery differed in the two experi-
ments. Again, digestion time did not significantly affect sulfur form
analysis values. In addition, these data suggest that the Meyers Process
241
-------
product sulfate is as soluble in acids as natural coal sulfate, at least
under the conditions of standard ASTM procedure for sulfur forms analyses
(the feed coal to Experiments S2 and S4 contained <0.2 wt. percent sulfate;
thus, at least part of the sulfate in the processed coals was process-
derived).
The last group of data in Table 42 were obtained from analyses on
processed coal doped with additional pyrite and iron sulfate. They were
principally performed to compare the relative accuracy of ASTM and modified
ASTM techniques. However, they also show that digestion time does not
affect the accuracy of sulfur form analyses (the analyzed pyritic sulfur
values were consistently lower than the computed values indicating that
the pyrite doping technique used led to a systematic error—probably due
to incomplete incorporation of the pyrite into the coal samples).
In summary, the sulfur analysis investigations performed in this pro-
gram indicated that the Meyers Process does not affect the accuracy of
standard coal sulfur analysis techniques even when recovery of the process
sulfur products is incomplete. The conclusion is that carefully performed
sulfur analyses should yield the true sulfur forms composition in coal
whether processed through the Meyers Process or not. Indirectly, however,
by reducing the pyrite concentration of coal to low levels the Meyers
Process aggravates the frequently encountered inconsistencies in coal
sulfur analyses.
6.2 MEYERS PROCESS MONITORING TECHNIQUES
Functionally, the Meyers Process consists of three major operations:
coal pyrite oxidation, reagent regeneration, and product recovery. For
every mole of pyrite oxidized in the leaching operation, one mole of coal
iron is dissolved into the reagent as Fe+2 and 9.2 moles of reagent Fe+3
+2
are reduced to Fe . Thus, the reagent total iron versus reaction time
trace or the ferrous ion versus reaction time trace can be directly trans-
lated into coal pyrite concentration versus reaction time information.
+2
Either Fe or total iron can be used to monitor process performance in
the leacher. Preferably the ratio of these iron forms should be used
242
-------
because it is not subject to errors resulting from sample water loss due
to vaporization and because the Fe+2/Fe ratio when accompanied by tempera-
ture data furnish complete information on all parameters affecting the
leaching rate of a given particle size coal. The leaching rate depends on
the type of coal and particle size, on temperature, on the pyrite concen-
tration in the coal, and on Y; the Fe+2/Fe ratio is equal to 1-Y. Thus,
the monitoring of temperature, Fe , and Fe, and knowledge of the pyrite
concentration in the feed coal and its particle size furnish the required
information to properly monitor the pyrite leaching operation. Temperature
and Y are also the appropriate parameters for monitoring the reagent
regeneration operation when used in conjunction with oxygen partial pres-
sure monitoring. The sulfur product recovery operations can be easily
monitored from the weight of product generated as a function of time with
occasional sulfate and elemental sulfur determinations performed to assure
the stability of product composition.
Monitoring the Meyers Process by the described techniques presents no
problems. The techniques are simple and fast; they require little labor
skill and they use inexpensive, commercially available equipment. Weights,
temperature, and oxygen partial pressure can be monitored continuously.
The total iron and ferrous ion concentrations in the reagent would have to
be performed by grab sampling, but the analysis time is short and the
reliability high. The ferrous ion analysis performed by the standard ASTM
technique requires approximately 10 minutes; the corresponding Fe analysis
can be performed in approximately 20 minutes. Several analyses can be
performed simultaneously by a single individual. From experience gained
from sampling and analysis of several hundred coal slurry samples, it was
estimated that the maximum time required from slurry sampling to Y com-
putation is one hour when standard ASTM techniques are used for iron forms
analyses. This response time, though adequate, can be reduced to less
than 30 minutes if slurry filtration and weighing of the reagent sample
are replaced by pipetting the required sample size directly from the
slurry and if AA is substituted for Fe titration (see Section 6.1)
Guarding against increases in the top size of the coal requires only a
single screen which is a part of the coal feed train of the plant.
243
-------
Monitoring the pyrite concentration in the feed coal and the composition
of the product coal could present problems in special situations where
quick analysis answers are required. Under normal processing conditions
sampling of feed and product coals need not be frequent (once or twice per
shift unless there is a more frequent change in coal lots) and immediate
analysis data would not be required.
All the elements of the Meyers Process monitoring scheme described
above, except product weight sulfate monitoring, have been extensively
tested and used at bench-scale. The scheme proved both efficient and re-
liable. However, its use is predicated on the assumption that the process
is totally selective with respect to ferric ion reaction with pyrite and
that the coal leaching and reagent regeneration operations are performed
separately. Side reactions between reagent iron and coal mineral matter
other than pyrite or the organic matrix of the coal will introduce errors
in the computation of pyrite concentration in coal from ferrous iron con-
centration or Y measurements. Experience gained from processing at least
50 different coals by the Meyers Process revealed that for the majority
of coals pyrite leaching by ferric sulfate was totally selective within
the accuracy of the measurements involved. In cases where excess ferric
ion utilization was observed, over that attributable to pyrite oxidation,
the excess ferric ions consumed or ferrous ions produced per unit time
was the same in different samples of the same coal. Thus, a single deter-
mination of excess ferric ion utilization per coal type is sufficient to
render the described process monitoring scheme usable with that coal when
pyrite leaching is not totally selective.
The described process monitoring scheme becomes inadequate when the
chemical desulfurization of coal is performed by L-R processing (simultaneous
coal Teaching-reagent regeneration). The L-R processing scheme (Section 5)
is expected to be used with the high pyrite coals which can not be effi-
ciently desulfurized in coarse sizes. In the L-R scheme the major fraction
of the pyrite in coal is removed in the L-R reactor where pyrite leaching
and ferric ion regeneration take place simultaneously. Thus, the con-
centration of the individual iron forms in the slurry are affected simulta-
neously by both reactions, leaching and regeneration, and the value of Y can
244
-------
no longer serve as a measure of coal pyrite depletion during processing in
the L-R reactor (the value of Y can still be used to monitor pyrite de-
pletion in the mixer and ambient pressure reactor of the L-R processing
scheme). In order to monitor process performance in the L-R reactor,
direct coal sulfur analyses are required either for pyrite or for total
sulfur. At least the pyrite content of the coal in the L-R reactor
effluent stream must be monitored (the pyrite content of the feed coal to
the L-R reactor is attainable from the pyrite concentration in the process
feed coal and the delta in the Y values of the mixer).
Examination of time requirements for the generation of coal sulfur
data from coal slurries (analysis response time) using conventional sample
preparation techniques and standard ASTM analysis methods revealed that
such procedure would be totally inadequate for monitoring the L-R reactor.
The response time would be several hours long with sample preparation
being the major time consumer. Sample preparation entails slurry filtra-
tion, coal washing, coal drying, and coal sampling and weighing; coal
grinding may also be necessary. Coal drying requires a minimum of four
hours, if performed according to acceptable ASTM procedures; the remaining
operations require a minimum of one hour. The minimum time required for
the performance of the standard ASTM pyrite analysis is 2 hours; a total
sulfur analysis by Eschka could be performable in 1 hour. Thus, the
minimum response time for total sulfur data is 6 hours and for pyritic
sulfur data 7 hours. Such response time was considered too long to serve
as the means of obtaining process performance information which would
permit timely changes in process parameters to correct processing defi-
ciencies and prevent the generation of a final coal product with a sulfur
content higher than desired. This conclusion is valid even though slurry
residence time in the ambient pressure reactor, in series with the L-R
reactor, could be as long as 40 hours, because the ambient pressure reactor
will not normally have the flexibility of affecting large changes in pyrite
leaching rates.
After careful examination of the sample preparation procedure, it was
decided that it could be shortened considerably, with less than 10 percent
245
-------
potential penalty in the accuracy of the analysis, if coal washing and
coal drying were eliminated. In the case of pyrite analysis, the elimi-
nation of these two operations affects only the accuracy of the weight of
the sample subjected to analysis. Data from over one hundred coal slurry
filtrations revealed that the quantity of reagent remaining on the coal
did not vary by more than + 5 percent of the weight of dry coal, even if
no effort was made to exactly reproduce filtration times; in the majority
of experiments the deviation was less than +_3 percent. Thus, sample
preparation time was reduced to approximately 0.5 hours and the response
time for pyrite analysis using the standard ASTM technique (ASTM D-2492)
to approximately 2.5 hours. This response time was reduced to approxi-
mately 2 hours when the iron determinations in the hydrochloric and nitric
acid extracts were performed by AA instead of titrimetric procedures
(see Section 6.1). A two hour response time was considered adequate for
monitoring the L-R reactor.
The analysis response time could be reduced to approximately 1 hour,
if total sulfur rather than pyritic sulfur is used as the means of moni-
toring pyrite depletion in the L-R reactor. However, the accuracy of the
data may be appreciably lower than that obtainable with the shortened
pyrite analysis procedure described above. The total sulfur analysis
procedure utilizes the Leco sulfur analysis technique, but on wet coal.
The data presented in Section 6.1 showed that the Leco technique furnishes
high accuracy sulfur values when the sulfur concentration in coal (dry) is
less than 3 wt. percent and that its accuracy is not affected if a portion
of the sulfur in coal is in the elemental form. However, similar sulfur
accuracy determinations were not performed with wet coal, though no
problems are anticipated. The potential source of error is associated
with the fact that the determined sulfur value must be corrected for the
product elemental sulfur, which was not recovered prior to analysis, and
for the reagent sulfate on the wet coal (the coal sample may be washed
prior to analysis, but then the response time increases to approximately
1.5 hours). The sulfate correction can be made from the estimated weight
of reagent solution on the coal and the Fe concentration and Y values
determined routinely as elements of the process monitoring scheme described
earlier in this report section. Assuming that 40 percent of the reacted
246
-------
pyritic sulfur converts to elemental sulfur, a valid assumption based on
data generated to date from several coals, the determined depletion in
total sulfur by the described procedure represents 60 percent of the
leached pyrite during the reaction time interval examined. The feasibility
of this procedure was not tested experimentally in an integrated form.
In parallel with the above investigations, a literature search was
undertaken aiming at the identification of novel, rapid procedures for
pyrite determination in coal. The chemical abstracts were searched for
the last decade with limited success. Three insufficiently developed new
pyrite analysis techniques were discovered, but with no demonstrated
potential to serve as monitoring techniques for the Meyers Process.
One of these techniques was the result of a feasibility study per-
formed by Barringer Research in Ontario, Canada under government contract
in 1968;12 it involves the sensing of sulfides in coal by a microwave
detection system. Though an interesting approach because it permits the
determination of pyrite in the coal without leaching, the data on the
technique is insufficient to evaluate its accuracy.
A second technique developed and used by the Illinois State Geological
Survey Institute is a lithium aluminum hydride reduction procedure for
Pyrite in coal.13 This technique is affected by the presence of elemental
sulfur in the coal (positive interference) and, therefore, it is not better
than a total sulfur analysis technique (Leco or Eschka) for monitoring the
Meyers Process performance. In fact, it is more complicated and less
tested.
The third technique is a soft X-ray procedure developed at Pennsylvania
State University.14 It is reportedly capable of determining the concen-
tration of pyritic, sulfate, and organic sulfur in coal using measurements
of X-ray fluorescence from four discrete wavelengths. This promising
Procedure is a very recent development-first reported in December 1974-
and therefore inadequately tested. Attempts in our laboratory to duplicate
247
-------
the reported data using a pentaerythritol analyzing crystal and an XRD-5
spectrophotometer were not successful because this system was incapable
of resolving the sulfur Kg doublet on which the procedure is based. Per-
sonal contact with Dr. Eugene White at Penn State University (co-author
of the paper) revealed that the instrument they used had to be specially
modified to obtain the desired resolution and the surface temperature of
the crystal (Germanimum) had to be regulated within +_ 0.1°C to maintain
dispersion stability. At the present time, the procedure though potentially
viable, would require extensive instrument and procedure development be-
fore being amenable to use as a process monitoring technique. However,
X-ray fluoresence is an attractive analytical technique because it is
potentially capable of being used to analyze more than one element and
because it is amenable to automation. Its drawback, as a monitoring
technique involving coal slurries, is that it requires that the coal be
dried prior to analysis and coal drying is a time-consuming operation.
248
-------
7. MATERIALS COMPATIBILITY
i
Static testing of common stainless steel alloys in an environment
simulating that of the Meyers Process but with low solution oxygen concen-
tration revealed that many stainless materials had severe corrosion
problems. However, dynamic testing in the L-R apparatus under actual
processing conditions indicated a much wider range of stainless material
stability to corrosion.
The following composition and operating conditions are typical for
the Meyers desulfurization process:
V
Fe2(S04)3 approximately 18%
H2S04 approximately 10-20%
FeSO. approximately 1%
Finely dispersed 02
Possibly chloride contamination
Temperature 275°F, Pressure 50-100 psi
Agitation and vigorous 02 bubbling
Coal abrasive charge
Obviously, materials to be used for the construction of the desulfur-
ization process plant must possess the necessary mechanical strength,
abrasion resistance and corrosion resistance properties. Of course,
design life, material cost, safety and maintenance are factors for trade-
off consideration with material performance.
Specific data on corrosion rates under the conditions of the leaching
process considered here are unavailable from literature. It is recognized
that in view of the complexities of the electrochemical processes associated
with corrosion, published data on corrosion tests under similar environments
249
-------
may not be directly applicable. No assumptions or extrapolations of avail-
able data may be possible. Corrosion rates have to be determined experi-
mentally under the specific conditions of the process in consideration.
However, there is adequate understanding and published data on corrosion
to limit the choice to certain classes of materials.
An initial literature survey and alloy selection discussion, the
results of static testing for screening of applicable materials and exper-
imentation on materials behavior in the bench-scale reactor under dynamic
conditions are presented in the three sections to follow.
7.1 LITERATURE SURVEY AND ALLOY SELECTION CRITERIA
Discussion of Alloy Composition and Corrosion Mechanisms
In terms of the more common types of corrosion, the desirable metal
composition and metallurgical structure are discussed below.
Corrosion
In the baseline Meyers Process design the physical size of the welded
assemblies may be such that it may not be convenient or practicable to
solution anneal or carry out other post-welding metallurgical heat treat-
ment of the finished structure. Hence, the material chosen must resist
sensitization (preferred precipitation of chromium carbide in the grain
boundaries). Alloys containing low C and/or some Nb, Ta are required to
avoid a sensitized condition which is highly susceptible to intergranular
corrosion.
Chloride ions present in small concentrations in water, and probably
also in the coal charge, may penetrate passivating films and initiate
pitting corrosion. The critical temperature at which pitting attack is
appreciable increases with increased Mo content. At an operational
temperature of 275°F, a high Mo content is desirable. Mo is the commonly
used alloying element to impart resistance to pitting corrosion.
250
-------
Near boiling temperatures, H2$04 between 20-80% acid concentration
causes serious attack on most stainless steels. Addition of ferric sulfate,
CuS04 or 02 reduces this attack greatly. Alloys containing Mo, Co/Si + high
level of Ni show good corrosion resistance in these environments.
Chloride ions, although present in small concentrations, would cause
severe stress-corrosion cracking in the presence of welding and mechanical
stresses and the tensile stresses in a pressure vessel. Abrasive grains
may also cause notches that initiate Griffith crack propagation. A high
level of Ni provides good resistance to chloride ion stress corrosion
cracking.
High Ni + Cr and a high Mo content are useful in the presence of hot
oxidizing ferric salt solutions of fluctuating concentrations.
Differential aeration or oxygen concentration initiates corrosion in
crevices. Hence, good agitation and dispersion of 02 bubbles in practice
would not only improve the sulfur leaching process but also reduce corrosion.
Stainless Steels, Coatings and Inhibitors
The advantages of Mn-Ni substitution in austenitic stainless steels
are increased resistance to stress corrosion, carbide precipitation (hence
sensitization) and pitting corrosion. These alloys are further improved
by addition of Mo to provide increased resistance against pitting corrosion
and Nb-Ta for increased stabilization against sensitization. One such
alloy is Armco 22-13-5 containing 5% Mn with additional Mo and Nb. The
Carpenter 20Cb-3 is also a similar alloy but with a low Mn content.
Armco 22-13-5 (Nitronic-50) austenitic stainless steel is superior to
316 or 316L in corrosion resistance and costs more than 316L. It also
combines high strength with corrosion resistance and can be useful in
stressed parts such as pumps, valves, screens and wire supports.
Table 43 compares the corrosion rates of Armco 22-13-5 with 316 and 316L
in different corrosive media.
251
-------
TABLE 43. COMPARISON OF CORROSION RATES OF ARMCO AND 316/316L STEELS
Test Medium
Armco 22-13-5 316/316L
2
10% Fed-, Amb. Temp. <0.001 gm/in2, no pits 0.0112 gm/in pitted
O
10% H2S04, boiling 0.356 Ipy (in per year) 0.73 Ipy
20% H2S04, boiling 1.64 Ipy 2.2 Ipy
Nitric acid test 0.0072 Ipy
Ferric sulfate + 0.0108 Ipy
Sulfuric acid
This alloy is recommended for stressed parts and if chloride ion stress
corrosion cracking or intergranular corrosion is a problem with conventional
300 series stainless steel alloys.
Types 347 (Nb + Ta), 321 (Ti), 309 Cb, and 310 Nb are stabilized
steels which can be used in the as-welded condition. Data on their corrosion
properties in H2S04 - Fe2(S04)3 are not readily available.
Organic inhibitors used in pickling tanks may be useful in minimizing
corrosion problems. Some examples are thiourea, dibutyl sulfoxide, amines,
0-polythiorea, quinoline derivatives, ketones, etc. Proteins such as milk
albumin are useful in providing a protective film. Also 0.2% AS in FLSO.
greatly reduces corrosion. These may be considered if they do not adversely
affect the leaching process.
Teflon or Kynar (polyvynilidene fluoride) coatings, may provide
additional protection in view of the fluctuating corrosion conditions.
Kynar can be coated inside small tubings. Tough, uniform teflon coatings
are possible to apply on large tubings and inside reaction vessels.
Sleeving with thin-walled teflon tubing is also a possibility. Teflon is
completely stable in boiling H,,S04 and Fe2(S04)3 solutions and Kynar is
also stable to a high degree.
252
-------
Survey of Literature Data on Corrosion
Tables 44 and 45 summarize available data on materials for conditions
somewhat similar to those under consideration. A comparison of these data
with the results of the tests conducted in this study (Table 48, page )
indicated the importance of generating corrosion data under specific con-
ditions.
TABLE 44. CORROSION DATA FROM LITERATURE ON 304 AND 316SS15
Ferrous
sulfate
-
-
Mostly
H2S04
150 ppm
270 ppm
With free
acid (pH=l)
6% free
acid
Ferric
sulfate
603 ppm
9424 ppm
-
1.5%
Temp
Corrosion Rate, Ipy
Time
61 days
61 days
100°F 67 days
160°F
7 days
*Inches
304
0.017 <0
<0.0001 <0
0.0005 0
<0.0001
per year
316
.0001
.0001
.0004
-
7.2 EXPERIMENTAL CORROSION STATIC TEST PROGRAM
7.2.1 Experimental Method
Figure 35 shows the stainless steel bomb assembly used to carry out
the static corrosion test under simulated plant conditions. Four 1-1/2"
sections of 1/2" O.D. tubings of different candidate materials were
welded together using the TIG process without filler rods as seen in
Figure 36. This served as a support for the sheet candidate materials to
be tested, which were machined into flat tensile test coupons after welding
in pairs. Thirty-two specimens were assembled to the configuration shown
in Figure 37. The coupons were held in radial slots on the periphery of
teflon support rings fixed to the tube with 304 S.S. screws. All the
test coupons were machined flat and polished with 3/0 emery paper MO
P-inch finish) and were used in the polished condition after welding with
253
-------
TABLE 45. CORROSION DATA FROM LITERATURE ON VARIOUS ALLOYS15'16'17'18
Corrosive
Medium
(1) FeS04, H2S04
PH=1
(2) 50% H2S04
42 g/1 Fo2 (S04)3
(3) 1.5% Ferric Sulfate
+4$ Free Acid
(4) 15% H2S04
(5) 15% H2S04
(6) 80% H2S04
(7) Stress Corrosion Test
42% MgCl2 Soln
(8) 4% Ferrous Sulfate
Ferric Sulfate
H2S04
PH 2.5
(9) 60% Solids g/1
28-55 gm/R H2S04
5-10 g/1 Fe+++
Some ferrous
0.1% NaClOj
Temp
F°
100
Boiling
160
176
104
176
Boiling
162
113
Time
Days
67
1
31
—
1
—
1000
hrs.
27
41
Aera-
ation
X
—
—
—
—
— —
X
Agita-
tion
X
—
X
—
—
— —
X
Type 304
0.0005
.0001
—
0.03
—
——
0.017
Average Corrosion Rate (ipy)*
Type 316
0.0004
.0001
—
nil
—
——
0.0006
Incoloy
825
—
—
—
—
"• •"
__
.0001
Inconel
625
—
0.070
--
0.007
—
0.090
"" "™
--
Hastelloy C-4
Un-
Weld-'d
0.111
—
—
—
•• —
--
As weld-
0.114
--
—
—
No
Cracks
--
—
ro
en
Inches per year
-------
Figure 35. Bomb Test Assembly
:j ..... : i
. Ujiiii
Figure 36. Composite Welded Tube Used In Test'
304L, Armco and 315L
From Left to Right 304,
255
-------
Figure 37. Test Coupon Stack: 32 Tensile Coupons Assembled Around The
Central Welded Tube With Teflon Support Rings
256
-------
no special annealing heat treatment. The tubes were also used in the as-
welded condition. All test coupons were carefully weighed and sized prior
to loading in the bomb.
The bomb was filled with ferric sulfate leaching solution used in the
process. The flanges were bolted together with a high temperature gasket
seal. The valve fitting on the outside of the top flange enabled the
bomb to be pressurized with 02- Heating tape was used to attain the
temperature.
Table 46 lists the materials and conditions used in the test.
7.2.2 Initial Runs and Modifications
A chromel-alumel thermocouple (TC) encased in a 304 S.S. protective tube
was used initially and the TC being immersed in the corrosive solution with
the aid of a T-fitting on the flange. After only 16 hours of testing, the
bomb had to be disassembled because the thermocouple sheathing (304 S.S.)
was completely corroded away. The 304 S.S. screws used to hold the teflon
rings showed signs of corrosion. The bomb was cleaned and reassembled
without the TC which was now placed on the outer bomb wall. The gradient
across the S.S. wall of the bomb was assumed to be negligible.
After 300 hours of testing at temperature and pressure, the bomb was
disassembled for replacement of the gasket. It was found that the bomb
interior and test coupon stacks were covered by a heavy brown deposit, and the
304 S.S. screws that were used to keep the teflon support rings in position
were very severely corroded, with some so completely reacted on that they
crumbled to a powder on touching. The coupons themselves showed no signs
of corrosion. The stressed contact points between S.S. screws and the
metal tubing were also pitted.
For further continuation of the test, the 304 screws were replaced
and new solution was used. The test coupons and bomb were cleaned of all
deposits. The bomb was heated for a second period of 700 hours (solution
exchange at 300 hours should not have had a material impact on the test).
257
-------
TABLE 46. LIST OF MATERIALS TESTED
WELD.
Composite Support Tube (welded)
Tensile Test Coupons:
Temperature
Pressure
Time
*260°F
TOO psig
316L
Helded Metal Pairs
304 - 304
316 - 316
316L - 316L
347 - 347
321 - 321
Armco - Armco
Carp - Carp
Inconel - Inconel
Incoloy - Incoloy
316 - 304
316L - 304
Armco - 316
Armco - 316L
Incoloy - Armco
Carp - 316
Carp - 316L
321 - 316L
321 - 316
347 - 316
347 - 316L
347 - 304
1000 hours (300 hours - first run + 700 hours - second run)
* A temperature excursion to 320 F for an unknown time between 24 and 60
hours occurred during the course of the test initially
258
-------
At the end of the second period of exposure for 700 hours (total of
1000 hours on the test coupons) it was observed that the bomb and the
samples were covered by a very heavy brown to blackish-brown deposit. Part
of the 304L portion of the composite tube between the weld areas 304/304L
and 304L/Armco was completely corroded away as shown in Figure 38. The
heavy deposit is partly from the corrosion products and partly also from
the solution, which was now dark green in color, with the total iron content
down by 44% and nearly all ferrous. The change in solution characteristics
is shown in Table 47.
TABLE 47. FERRIC SULFATE CORROSIVE MEDIUM
Total Fe
r ++
Fe
Initial
50
13
(mg/g)
Brown heavy deposit None
Color
Cr/Ni
S04 and
Cl - ions
Brown
-NOT AN
300 hrs.
46
21
Considerable
Green
A L Y Z E D -
700 hrs.
28.4
28.1
Very heavy
Darker Green
In order to obtain weight losses, the tough adhering deposits had to
be removed. The procedure used was as follows. The heavier buildups were
removed by mechanical brushing with a nylon brush and acetone in an ultra-
sonic bath. Some of the more adherent deposits were loosened by thermal
cycling between 100°C and 0°C. The samples weighed after the mechanical
removal still showed a weight gain as compared to the original, indicating
incomplete removal of the oxide deposits. Hence the following pickling
procedure was used:
a) 60 minute dip in 20% NaOH + 5% KMn04 at 200°F
b) 10 to 30 seconds in 15% H2S04 at 180°F
259
-------
Weld
S 2/3X
Figure 38. Comparison of Corroded And Control Tubes.
BOTTOM:
TOP:
Tube After 1000 Hours Exposure to Corrosive Environment.
304L Tube Disintegration
NOTE:
Control Tube:
The tubes have been aligned such that weld zones are
one below the other. The missing portion (% 1 inch
of 304L) of the exposed tube nas been corroded away.
From left to right, the sections are 316L, Armco, 304L,
and 304. The threaded portion at the left end of the
exposed tube was cut off before the test. The teflon
support ring is seen on the left.
260
-------
This procedure removed the oxide deposit effectively without attacking
the bare metal. Type 316 S.S. was attacked somewhat by the H2$04 solution.
A set of control samples was also subjected to identical treatments in
order to correct for weight loss from the bare metal due to the pickling
procedure. Since the oxide deposits on the coupons could only have
inhibited the reaction of the pickling bath on the metal substrate, the
weight losses obtained after correction must represent a minimum figure.
However, the corrections are small and this effect is not likely to be
appreciable.
7.2.3 Results and Conclusions of the Bomb Test
The corrosion rates are calculated from the weight loss and surface area
and density for each coupon in units of milligram (loss) per square
declmeter (mdd) per day. This can be converted to the more conventional
inches penetration per year units (ipy) using the factor mdd x * . —' =
ipy where d - density (gms/cc). The corrosion rates were found to vary by
a factor of three in duplicate samples of the same material. This large
variation is due to the position of the sample in the bomb stack and the
degree of protection from additional corrosion afforded by the accumulation
of the surface deposits. But the differences between two types of materials
is marked. It should be emphasized that due to the static conditions in the
bomb, the corrosion rates obtained are probably lower than for the actual
dynamic conditions and are more meaningful as relative figures than as
absolute corrosion rates.
Table 48 gives a summary of the corrosion rates, visual appearance
and tensile test results obtained.
Several metallographic mounts were prepared from the test and control
samples to study the extent of corrosion and the corrosion mechanism.
Mounts were made from near the weld zone as well as near each end of the
grip section. Figures 39 and 40 show the microstructure of the exposed
(corrosion tested) 304 S.S. coupons and Figure 41 shows the microstructure
of a 304 S.S. control sample. It was observed that while sensitization
261
-------
TABLE 48. COMPILATION - COAL CORROSION MATERIAL EVALUATION
ISpecimer
1 Number
LiA
MB
U
I 2B
3A
1 38
1 3C
4A
1 4B
1 SA
1 SB
1 6A
6B
L»
1 7B
8
i 9
I 10A
I 10B
1 11A
|m
hr-
1 13
,4
1
u
P7_
F8
ri9
p°
r 2i
^^^^•^M
Hatl.
304/304
316/316
~316/3T6~~
316/316
316L/316L
316L/316L
316L/316L
347/347
347/347
321/321
321/321
ARM/ARM
ARM/AWT
CARP/CARP
CARP/CARP
NCONEL/
INCONEL
NCOLOY/
TNTOI 0V
316/304
316/304
316L/304
316L/304
ARH73TT
ARM73T6L
INCOLOY/
ARM
CARP/316
CARP/31 6L
321/316L
321/316
341/316
"347/316"
" 347/304
^MMBMM^^H
m^mmmtm
Y.S.
44 ."4
35.7
1O
36.5
44.4
44.6
47.6
43.0
41.0
65.0
43.1
58~.0
"62.2
56.5
56.6
74.5
56.7
39.2
39.0
"46.4
46.9
To^r
' 49'rr
56.1
"3O
51.0
44.6
35.8
39.0
45.3
45.1
^^^^^
MMMUM
^^iH^BM
UTS
92.5
78.4
~74."6"
64.7
79.8
78.5
84.0
88.2
70.0
88.4
~W.T
100". 2
99.1
94.2
"96.3
1T7."7
91.8
77.4
79.4
87.4
90.2
^7¥.¥
8SL6
95.6"
"73.r
83.6
82.8
65.3
80.0
76.7
88.6
^^>BHMHMB
•^•MMM
s
M^MHI
Elong.
40
~ 43
34"
15
26
25
27
35
24
23
-~26"
23"
TT
20
23
18
15
35
41
34
36
~3T
"31
23
--34-
29
25
22
30
20
40
^^^•MMM
^^•MMM
JBJECT
^WM
Break
M
316
~3~16
316
W
W
H
H
347
M
321
W
U
W
H
W
W
316
316
W
w"
"3T6"
W"
W
316
H
W
321
W
U
347
l^^MM
^— ^— —
ED SPECIMENS
^^~^~~i~^^^
Defect
Few sml voids
O.K.
Few sml voids
outside welds
o'.k.
Few sml voids
outbids wel ds
O.K.
O.K.
O.K.
Couple voids
... _at..we1d .....
Few sml voids
.outside welds.
Small spot
at uejd
Small speck
at wel d .
Nothing
of sia.
O.K.
Small voids
large @ weld
O.K.
Various" voids
large @ weld
O.K.
O.K.
O.K.
O.K.
Numerous voids
at weld
Large void at
weld
Couple voids
outside weld
Couple sml.
voids. out. wld.
Small
abt.weld
O.K
O.K.
O.K.
O.K.
O.K.
O.K.
MM^^MM^^^^^^H
^MHBMMI
^••^^•MMB
Poor
Poor
Poor
Poor
Poor
Good
Good
V.Good
Poor
"Poor
'Good
V.Good
V.Good
T.G6ocr
V.Good
V~:Sood
V.Good
V.Good
>oor
V.Poor
Fair
Fair
Fair
Good
V.Good
Poor"
V.Good
Fair
Poor
V.Poor
Fair
Poor
•^^^MV
^^•WH
^HM^HM
0.0680
0.0510
0.1189
0.3783
0.3934
0.2477
0.3432
0.0832
0.1889
0.4019
0.1665
0.0632
0.0056
Not'Used
0.0127
6."0022
0.0170
0.0072
0.2974
6.2399
0.0630
0.0576
0.7649
0.0190
~ff."D014
0.1965
0.0105
0.0975
0.4524
0.3346
0.0833
0.2921
^^^^^^^^H
^^•••^
Gra/Cn3
7.89
7.91
7.98
7.95
7;98
7.99
8.03
7.91
7.93
7.93
7.91 "
7.86
7.92
7.91
8.07
8.04 ""
8.46
8.16
7.92
7.95
7.94
7.94
7.95
7.95
8.04
8.00
8.00
7.99
7.93
7.96
7.96
7.92
^•^^^"
••Ml
"spT!
1A
IB
2A
2B
2C
3A
3B
3C
4
5
6
7
8
9
10
11
12
T3
14
15
16
17
18
19
20
21
I^^H
•^•MMH^
^••^••ii^
304
304
316
316
316
31 6L
31 6L
31 6L
347
321
Armed
Carp
Inconel
Incoloy
316/304
316L/304
ARM/316
ARM/31 6L
INCOLOY/
ARM
CARP/316
CARP/3161
321/316L
321/316
347/316
347/316L
347/304
^^••^^MH
••••••
^^^^m
48.4
44.1
42.3
39.9
40.7
44.1
49.0
44.3
44.7
45.5
56.6
50.4
74. 2
48.1
42.4
47.7
43.7
49.6
52.8
42.3
49.1
41.7
42.5
40.8
40.7
45.0
H^MM
i^MHMMM
^ .
92.7
92.2
80.0
84.7
78.5
81.4
83.9
80.4
73.8
90.4
103.9
92.5
123.9
86.1
84.5
87.0
85.2
89.2
93.2
76.9
89.5
80.6
83.2
82.5
78.0
92.1
•MMM
CONTRC
»/
V
V
>l
V
V
V
V
V
>/
\
V
-J
V
V
V
Co
>/
V
V
V
V
V
V
V
V
V
V
V
>l
>/
V
V
V
V
V
V
>l
V
V
V
rrosion
Rate
.0010
.0024
.0075
.0079
.0049
.0068
.0017
.0037
.0080
.0034
.0014
.0001
0.0002
;0.0003
0.0001
0.0059
0.0048
0.0013
0.0011
0.0032
0.0004
0.003
d.ooo
0.002
0.009
0.006
0.001
0.006
^^^™
1
^••••Ml
RATINGS
^orrosiorT
5
8
7
9
6
4
2
3
1
Ratio;
1 - t
MM^^^^^^^^M
Mech 1 Edge Cracking!
Prop A |Hetallographic|ylsual
A
B
A
?
A
A
A
A
A
1 Codes:
jest, large
A =
B =
idM^BMO^^^B^
1
10
4
3
1
1
1
1
1
st number « v
No apparent c
Slight change
.L— ^—
9
8
6
7
5
3
4
2
1
*orst
hange
\
-------
200X
Figure 39. Microstructure of 304 Test Coupon After 1000 Hours Exposure
'
'-
-. ,.
.:' .
- - •• .. -:"•• -
' /^
* ,
•
•
. -,„ •
».
•
•
. i ! - : "• 'If
-;: ". ^
:4,^ • ' - • ' • • :.
• *
""- >••%. / .*" , « "-v
: / • r
i •- '. *. -
\ '- •
v ' • ,
* •" ." , . • , •»».
•• W ...;... . V •
400X
Figure 40. Microstructure of 304 Test Coupon at 400X Showing
Sensitization at Grain Boundaries
263
-------
•;->A i- •' - -~Y
< .. -. .
- . •••,-*-
' -i- >- - "\ ; ' -.'- • -r^'-;-' i
200X
Figure 41. Microstructure of Unexposed Sample of 304 Test Coupon
264
-------
had indeed occurred at some distance from the weld zone, intergranular
corrosive attack was barely evident. Pitting corrosion was not seen here.
Figures 42, 43 and 44 show that very severe pitting corrosion occurred
just outside the weld area in 316. Both large and small pits were observed,
and some of the large corroded areas were considerable in size.
Figures 45 and 46 show the edges of the mounted specimens of a con-
trol and exposed sample of 316. Cracks near the outer edge due to an
embrittling effect of the corrosion are evident in Figure 46. Figures 47
and 48 similarly show the edges of a corroded and exposed sample of 316L.
Cracking and pitting due to corrosion are evident.
Figures 49 and 50 show cracked and exposed edges of 347. Other areas
of corroded 347 showed evidence of pitting similar to 316 though not quite
as extensive.
Figures 51 and 52 show areas of Incoloy 825 in the unexposed and ex-
posed conditions. There is evidence of either cracking or corrosion.
Figures 53 and 54 show the unexposed and exposed areas near the weld
of the 304L/304 joint. It is recalled that the 304L tube near the weld
was so severely attacked that part of the tube fell apart as seen in
Figure 38. Figure 55 shows the corroded area on 304 tube. Figure 56
shows the corroded inner-edge of the 304L tubing. Figure 57 shows the
weld and parts of 316L and Armco tubing. Severe pitting exists on 316L
with much less on the Armco tubing.
Figure 58 shows the uncovered (top) portion of the 316L tubing and
bottom portion which was covered tightly with teflon tape. The much more
severe pitting in the unprotected area is evident.
Figure 59 shows cracks in the I.D. of Armco tube while the O.D.
Figure 60 shows only pits. Figures 61 and 62 shows, respectively, the
O.D. and I.D. of the exposed 316L tubing.
265
-------
• V 4 " ^*'w * " ' '•"'
«••' ^ -
m
100X
Figure 42. Micrograph at 100X Showing Very Extensive Pitting Type of
Corrosion Just Outside the Weld Zone in 316SS After Exposure
• --
100X
Figure 43. Micrograph at 100X Showing Very Extensive Pitting Type of
Corrosion Just Outside the Weld Zone in 316SS After Exposure
266
-------
•' ' '-'•*
_ • .... • - ,r-> ..,,*
•-.--..• •• -»* • -j
3 .. -^
',.*':""•:-. •-,
- •• -•>'*•**$!&
• ^ •" • . > - .!•. - ":»^S
*-~' • • "%^j -x-'?ri
|.^-'--
' •
,
'
*• '**•
---«,.
*•-•• .• ;"-^ ;-;^li
100X
Figure 44. Micrograph at 100X Showing Very Extensive Pitting Type of
Corrosion Just Outside the Weld Zone in 316SS After Exposure
267
-------
~* ~
; .1*5
.
.
,
200X
Figure 45. Edge of 316 Control Sample Showing Elongated Stringers Due to
the Tensile Test
.
200X
Figure 46, Edge of Exposed Test Coupon of 316 Showing Transgranular Cracks
Extending Inward From Cut Edge
268
-------
200X
Figure 47. Edge of 316L Control
i ,• • •,.-,-•:,. - « . •;; -
.--.•-•
--• • •'• :*.
-„ , ^ •• • . .. —
- • 4 ' ;' ' i-
• • v - • - •
• " * -v- '
• • •' : - -•'
-- . •> • -. - - •.
• • - •
• •
-
1 -•
>.
1
•
'•:•• ,
S3* i ,",
, » -
-^ •
*cr -.
. • . .
200X
Figure 48. Edge of Exposed 31 6L Showing Crack Extension Along Crack
Sensitive Corroded Paths and Pitting
269
-------
J^ T-'-iV '• - •••—•-*.. - - ?*~' ~
|fe> •• -•- --.:, ^*r-
&4i
-------
'
•
-
(-*
'
• J^
•
400X
Figure 51. Edge of Incoloy 825 Control
>
' -
-
c
-
£
.
-
•
-
400X
Figure 52. Edge Of Exposed Incoloy 825 Test Coupon
271
-------
•.
'
,
•' *'
'
'—,y •* • > -"• - .• ,
- #.*,
. >,~^-<~ *
* *
• '• - u •• -:#*•*&. .
...— • f
.304L
Parent
Metal
200X
Figure 53. Weld Between 304 And 304L In The Control Tube Sample
1
L- ? •' -/*
'• .'>-
200X
Figure 54. Weld Between 304 and 304L Tubes Showing Severe Corrosion
Effects After Exposure
272
-------
-
- ,
200X
Figure 55. Outer Diameter of 304 Tube After Exposure Showing Corrosion Along
The Edge And Pits On Surface
273
-------
. .
200X
Figure 56. Inner Edge Of Exposed 304L Tubing Showing Corroded Areas
274
-------
ARMCO
•316L
Figure 57. Exposed Tube Sample
275
-------
Exposed
Section
Teflon Tape
Protected
Section
7X
Figure 58. Exposed 316L Tube
276
-------
...
100X
Figure 59. I.D. of Exposed Armco Tubing Showing Some Cracks And Corroded
Areas
. - :
V
100X
Figure 60. O.D. of Exposed Armco Tubing Showing Stringers Along The Tube
Drawing Axis and Light Pitting
277
-------
100X
Figure 61. O.D. of Exposed 316L Tubing Showing Corrosion Pits
100X
Figure 62. I.D. of Exposed 316L Tubing Showing Less Severe Corrosion
278
-------
Table 49 gives change in yield strength and elongation after exposure
based on an arbitrary ± 5% allowed variation in yield strength and a ± 10%
allowed variation in elongation. The only material that showed signifi-
cant degradation in both strength and ductility was 316SS.
Table 50 gives a summary of qualitatively observed events during the
course of the test. It is evident that most of the 300 series stainless
steels showed alarmingly poor corrosion resistance in the desulfurization
process environments.
The results of the present study show that the rates of corrosion of
the 300 series stainless steels are extremely sensitive to the metallurgical
condition of the sample, particularly residual stress. This is brought
out dramatically in the case of the 304L tubing (Figure 38) which was com-
pletely corroded away due to the built-in stresses in the tube drawing
process. It is well known that in an acidic environment, the stresses
could cause increase in corrosion rate several-fold especially at tempera-
tures in the region of 100°C. A possible mechanism is the precipitation of
carbides and nitrides or agglomeration of C or N atoms at dislocation
sites which causes local action cells in plastically deformed material.
This is also borne out by the observation that 316 corrodes just outside
the weld zone where cooling stresses are greater and the fact that the
welded seam in Armco tubing, where some of the drawing stresses must have
been annealed out, is free of corrosion.
The cracking observed along the cut edges of 316, 316L, 347 and
Armco tubing also is indicative of embrittling due to stress corrosion.
The cracking appears to be transgranular, following crack-sensitive paths
determined by loci of local action cells formed due to residual stress or
cold working. Whether Cl" ions play a role in aiding stress corrosion by
adsorption is not known since the solution has not been analyzed for Cl
ions. The sorption may occur selectively along paths of pinned dislocations
or vacancy-alloying element clusters formed due to plastic deformation or
residual cooling stresses after welding. The fact that 304L (extra low
279
-------
ro
oo
o
TABLE 49. CHANGES IN MECHANICAL PROPERTIES OF ALLOYS AFTER EXPOSURE
TO IRON SULFATE SOLUTION
Material
Yield Strength
Control + 5%
304
316
31 6L
321
Arraco
Carp
Inconel
Incoloy
347
45.98
38.95
43.5
43.2
53.8
47.9
70.5
45.7
42.5
- 50.82
- 43.05
- 48.09
- 47.80
- 59.40
- 52.9
- 77.9
- 50.5
- 46.9
Exposed
44.5
(~)
36.0
45.5
(0)
54.1
60.1
(")
56.5
( +)
74.5
(0)
56.7
42.0
Elongation
Control t 10£
35.6
33.9
23.4
27.9
23.4
18.9
19.8
15.3
13.5
- 43.40
- 41.50
- 28.6
- 34.1
- 28.6
- 23.1
- 24.2
- 18.7
- 16.5
Exposed*
40.0
/ \
(o)
30.0
26.0
m \
(o)
25.0
(-)
22.0
(\
-0)
21.5
(0)
18.0
(_ \
-0)
15.0
(-0)
30.0
(+) indicates increase
(-) indicates decrease
(o) indicates that the values lie within the allowed range
-------
TABLE 50. SUMMARY OF QUALITATIVELY OBSERVED EVENTS DURING
CORROSION TESTING
• 304 thermocouple shield tube disintegrated after a short timers.)
(16 hrs) exposure
t 304 screws used initially were severely corroded
• 304L tubing was severely corroded in the composite welded tube
t Type of Corrosion (Plate Materials)
Pitting type: 316, 316L, 347
General type: 304, 321, Armco, Carpenter, Inconel,
Incoloy
carbon, ELC, grade by Sandvik Corporation was completely corroded away also
suggests the above mechanism as no sensitization can occur in the very low
carbon grade stainless steel.
The pits seen in 316 and 347 may be due to accelerated corrosion at
isolated stress points and/or local action cells due to concentration
differences.
The extreme sensitivity of corrosion rates to residual stresses is
also apparent from the fact that 304 was considerably more stable than 316
and 347 in sheet form while it was severely corroded in drawn tubes and
machined screws.
7.2.4 Conclusions and Recommendations
• 304, 304L and 316 appeared to be marginal as construction
materials to be used in the coal desulfurization reactor.
a Dynamic testing in an environment where oxygen is in
plentiful supply may improve the stability of some of
these alloys.
281
-------
7.3 DYNAMIC TESTING
A composite welded tube, made up of 1-1/2" sections of tubing of
different candidate hardware materials for the coal desulfurization process
was tested in the L-R rig for approximately 50 hours of simultaneous coal
leach and reagent regeneration. The experimental tubing replaced a section
of the flow pipes in the L-R rig; hence the tube interior was exposed to
the erosive environment of flowing mixture of coal fines and ferrous/ferric
sulfate solution. The objective of this test was to evaluate and compare
the corrosion resistance of the various candidate materials under dynamic
processing conditions and in the presence of finely dispersed oxygen.
Post-exposure examination was conducted on the test tubing. The tubing
was first cut into two halves longitudinally to expose the interior. After
cleaning superficial dirt by blowing with compressed air, the different
sections were examined with a low power microscope for gross evidence of
corrosion. Figure 63 shows a picture of the composite tubing before it was
cut. Figure 64 shows the results of low magnification microscopic examin-
ations at various locations. There were no visible changes in areas where
no remarks are noted. Faint corrosion spots and adhering stains were ob-
served but no gross corrosion was evident on even the worst area. Figures
65 and 66 show the 304 and Armco sections, respectively, under low magnifi-
cation. It is evident that no apparent corrosion has occurred in any of
the tubing materials.
s.
Three sections were selected for microscopic examination:
(1) 304-Armco weld area
(2) 316-321 weld area
(3) Carpenter 20Cb/3-304L weld area.
Figures 67 through 70 show the microstructure of the parent metal and
the weld areas. Intergranular corrosion found in some alloys in previous
investigations was not observed in any of the materials evaluated in this
investigation. Sensitization (shown as black spots) was seen in the weld
282
-------
areas as shown by the presence of precipitated chromium carbide but
corrosion at the grain boundaries is not discernible.
Negligible corrosion was observed for all alloys tested for 50 hours
in the coal desulfurization test unit. Apparently, the presence of
quantities of available oxygen may inhibit the corrosion of the stainless
materials investigated. However, the test duration was too short to be
definitive. Much longer times in excess of several hundred hours would
probably be required before a measurable amount of corrosion can be
observed.
283
-------
Carp
ZO/16/3
304
IN >
00
I •
Figure 63. Composite Tube Used in Dynamic Test
Note: two end fittings are 304 SS. Adapters Al through A4
are 304 SS
-------
Small
Corrosion
Stains
c * Sta1
Spots
I i
j ^^
304 1 1 Armco
304ELC
ns
1 ^
Armco 1 316
i l&^f^3 I
L 1 /321 ! Ai
' / ; 4
/* V \, I
rmco 1 347 I
i 1 1
Thin Crack Loose Black Small Holes
Deposit on Weld in Weld
Small Spotted-,
Stain Faint Stains
Al
Inconel
A2
ro
03
01
Faint Spots
Incoloy
I -
A3
1; , .-
CARP
. 1
3
r
04L
_i
A4
304
A1-A4: 304 Steel adapters.
Sections used for
mounts.
Figure 64. Composite Welded Tubing: Corrosion Effects in the Tube Interior.
-------
Figure 65. Section of Armco Tubing Interior.
Corrosion Effects
2X
No Gross
"U» • ,
2X
Figure 66. Section of 304 Tubing Interior. No Evidence
of Corrosion
286
-------
•r* \ \ s , • , .
,
- - .. •
.
:
.'
•••. • '.•,-• .
-.. .
• .
. ' f-
.
'
• • -•• .-
.
• » • .-'•'*
'":•> •
.*• • •
Figure 67. Weld Zone of Armco-304 Showing Chilled Structure
and Sensitization by Chromium Carbide
400X
Figure 68. Structure of 304 Away from Weld
287
-------
-
•' •
.
.
••->•..;<•.« • v
' .'." >'.' .' • , -
• H'; •"''•'•:,"•/;•" "•'.' '
•:;-' '}•;'
.
.
.
• , . • ••. • . ,'i . • -•. .->-:- .
• •- .--. •. ;.,- ,;,.-/•. i
••. v . • •'..•• ,.<\t*'&V.-y
, . .-:•"• ' - . .. ' -. ...; ':-.-'.'-.. .
.•^.t'fi-K r.:"- ;
S •• •••• •••: • rf-fyQ &M
•-•••/.' * - • "/ v* ,» . •
••/-::• r^t .:'•-•; •/••••••.-•.-••• \X. &/•*•
400X
Figure 69. Weld Zone of 316L-321 Showing Sensitization
•-* . --' - . ' / . ;
* - •*' ^ ,'*
.' 1 • • • • .- ./ •-,..'.
• ••
-• .' ^ ^ ' .
- ' ^ - - -• ^ '
400X
Figure 70. weld Zone of 304L-Carpenter 20 Cb/3 Showing Chilled
Structure and Grain Boundary Sensitization
288
-------
APPENDIX TO SECTION 7
Corrosion Characteristics of Steel Alloying Elements
Element(s) Property
Ni + Cr Resists oxidizing chemicals
High Ni + Mo Resists non-oxidizing solns
High Mo + Cr/Si + Resists boiling H0SO.
High Ni + Cr 2 4
High Ni + Mo Resists fluctuating oxidizing salt
solutions
High Co, Ta, Ti, Low C Resists sensitization during welding
High Ni Freedom from chloride ion stress
corrosion cracking
Mn + High Ni Improved resistance to pitting, stress
corrosion cracking and increased strength
and stabilization
289
-------
APPENDIX TO SECTION 7
CHEMICAL COMPOSITION OF CANDIDATE MATERIALS TESTED IN WT. PERCENT
Inconel 625
316
31 6L
304L
§ Armco 22-13-5
Carp 20-C5-3
304
321
347
Incoloy 825
Ref
(17)
(15)
(15)
(15)
(19)
(20)
(15)
(15)
(15)
(16)
Cr
20-30
16-18
17-21
17-21
20-23
19-21
18-20
17-19
17-19
19-23.5
Mi
57-60
10-14
9-13
8-12
11-13
32-38
8-12
9-12
9-13
38-46
Fe _
Mo (Nb+Ta) c
5 8-10 3-4 o.l
Rem* 2-3 0.08
Rem 2-3 0.03
Rem
Rem
Rera
Rem
Rera
Rem
Rem
0.03
1.5-3.0 0.1-0.3 0.06 Max
2-3 SxC-1.0 Q.07 Max
0.08
0.08
lOxC Min o.08
2.5-3.5 - 0.05
Ii li
0.4 0.5
1.0
1.5
2.0
1.0 Max
1.0
1.0
5xC Min 1.0
1.0
0.0-12 0.5
Mn
2.0 Max
4.
2.
2.
2.
2
1
0-6
0
0
,0
.0
.0
.0 Max
Rem = remainder
-------
8. REFERENCES
1. Hamersma, J. VI., E. P. Koutsoukos, M. L. Kraft, R. A. Meyers, G. J.
Ogle, and L. J. Van Nice, "Chemical Desulfurization of Coal: Report
of Bench-Scale Developments",-EPA R2-73-173 a and b, prepared for
the Office of Research and Monitoring of the Environmental Protection
Agency, Research Triangle Park, N. C., February, 1973.
2. Hamersma, J. W., M. L. Kraft, C. A. Flegal, A. A. Lee, and R. A.
Meyers, "Applicability of the Meyers Process for Chemical Desulfuri-
zation of Coal: Initial Survey of Fifteen Coals", EPA-650/2-74-025,
prepared for the Office of Research and Monitoring of the Environ-
mental Protection Agency, Research Triangle Park, N. C., April, 1974.
3. Hamersma, J. W., and M. L. Kraft, "Applicability of the Meyers
Process for Chemical Desulfurization of Coal: Survey of Fifteen
Coals", EPA-650/2-74-025-a, prepared for the Office of Research and
Monitoring of the Environmental Protection Agency, Research Triangle
Park, N. C., September 1975.
4. "Program for Processes for the Selective Chemical Extraction of
Organic and Pyritic Sulfur from Fossil Fuels", Document No. 17270-
6011-RO-OO, Contract EHSD 71-7, prepared for the Office of Research
and Monitoring of the Environmental Protection Agency, Research
Triangle Park, N. C., January 15, 1975.
5. Leonard, J. W., and D. R. Mitchell (Editors); Coal Preparation; Am.
Inst. of Mining, Metallurgical and Petroleum Engineers, Inc.,
New York, N. Y., 1968.
6. Guthrie, K. M., "Capital Cost Estimating", in Modern Cost Engineering
Techniques, ed. H. Popper, McGraw-Hill, 1970, pp 80 to 108.
7. Mills, H. E., "Costs of Process Equipment", in Modern Cost Engineering
Techniques, ed. H. Popper, McGraw-Hill, 1970, pp 111 to 134.
8. Happel, J., and D. Jordan, Chemical Process Economics, 2nd ed.,
Marcel Dekker, Inc., 1975.
9. Perry, R., and C. Chilton, Chemical Engineers' Handbook, 5th ed.,
McGraw-Hill, 1973.
10. "Chemical Marketing Reporter", Schnell Publishing Company, June 2,
1975, p. 44.
11. "Final Report - The Supply-Technical Advisory Task Force - Synthetic
Gas-Coal", prepared by the Synthetic Gas-Coal Task Force for the
Supply-Technical Advisory Committee, National Gas Survey, Federal
Power Commission, dated April, 1973.
291
-------
REFERENCES (CONTINUED)
12. Barringer Research Ltd., Rexdale, Ontario, Canada, "Feasibility Study
fdr Sensing Sulfides in Coal", Final Report, Natl. Center for Air
Pollution Control Contract, PH86-67-270, Proj. 385, Report TR-68-55.
13. Comparison of Oxidation and Reduction Methods in the Determination
of Forms of Sulfur in Coal: Environmental Geology Notes #66,
Illinois State Geological Survey, December 1973.
14. Hurley, R. G., and E. W. White, New Soft X-Ray Method for Determining
the Chemical Forms of Sulfur in Coal, Anal. Chem., Vol. 46, #14,
December, 1974.
15. "Corrosion Resistance of the Austenitic Chromium - Nickel Stainless
Steels in Chemical Environments". The International Nickel Co.,
N.Y., 1963.
16. Incoloy 825 Spec Sheet, Huntington Alloy Products Division,
International Nickel Co., Huntington, W. Virginia.
17. Inconel 625 Spec Sheet, Huntington Alloy Products Divisions,
International Nickel Co., Huntington, W. Virginia.
18. Hastelloy C-4 Spec Sheet, Cabot Corporation, Kokomo, Indiana.
19. ARMCO-Product Data Bulletin S-45a, Armco Steel Corporation, Baltimore,
Md.
20. CarTech Selecting Carpenter Stainless Steels, Carpenter Tech Corpo-
ration, Reading, Pa.
292
-------
9. GLOSSARY OF ABBREVIATIONS AND SYMBOLS
Abbreviations
Abs
ASTM
btu
cal
eq
Exp.
Kcal
No.
wt
absolute
American Society of Testing Materials
British Thermal Unit
calories
equation
experiment
kilocalories
number
weight
Symbols
A,
C
A
EL
ER
KL
K
R
v
P
P(
R
r,
Arrhenius constant in leach reaction (hours)"
(wt% pyrite in coal)~'
Arrhenius constant in regeneration reaction
(minutes)'1 (atm)"1 (liters/mole)
concentration
difference in quantity following delta
activation energy for pyritic sulfur leaching
reaction, Kcal /mole
activation energy for ferric ion regeneration
reaction, Kcal /mole
pyritic sulfur leaching rate constant (units
same as A|_)
ferric ion regeneration rate constant (units
same as
R
micron.
total pressure, atmospheres
oxygen partial pressure
gas constant, cal /mole, °K
pyritic sulfur leaching rate, weight of pyrite
removed per 100 wts of coal per hour
ferric ion regeneration rate, moles per liter
per minute
293
-------
Symbols (cont'd)
S elemental sulfur
n
S organic sulfur
S pyritlc sulfur
S$ sulfate sulfur
S. total sulfur
T absolute temperature, °K
t time, hours (leaching)-minutes (regeneration)
W pyrite concentration in coal, wt%
Y ferric ion to total iron ratio
294
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing!
1. REPORT NO.
EPA-600/2-76-143a
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Meyers Process Development for Chemical
Desulfurization of Coal, Volume I
5. REPORT DATE
May 1976
6. PERFORMING ORGANIZATION CODE
7.AUTHOR.S, E.P.Koutsoukos, M.L.Kraft, R.A.Orsini,
R. A. Meyers, M. J.Santy, and L. J. Van Nice
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORQANIZATION NAME AND ADDRESS
TRW Systems Group
One Space Park
Redondo Beach, California 90278
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21AFJ-033
11. CONTRACT/GRANT NO.
68-02-1336
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND I
Final; 6/73-12/75
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES Project officer L.Lorenzi, Jr. is no longer with EPA; for details
contact Lewis D. Tamny, Mail Drop 61, Ext 2851.
16. ABSTRACT
repOrf. gives results of bench- scale development of the Meyers Pro-
cess (for chemical removal of sulfur from coal) for desulfurization of both fine and
coarse coal. More than 90% of the pyrite was removed from run-of-mine (ROM) fine
coal and clean coarse coal, and more than 80% of the pyrite from ROM coarse coal.
Process improvements were demonstrated involving: increased process slurry solids
concentration for higher process throughput (33% w/w); lowered filtration require-
ments through use of larger top- size fine coal; and generated elemental sulfur remo-
val by vaporization from the coal matrix. Pyrite leaching and reagent regeneration
rate expressions were validated. Engineering studies showed that: the process may
be engineered in a number of basic designs including simultaneous leach and regener-
ation, separate leach and regeneration, use of oxygen or air for regeneration, fine
or coarse coal processing, and combined with coal cleaning; these process designs
lead to stand-alone full capital cost estimates of $30-80/KW of power plant name-
plate capacity for the various process plant cases. Assuming ROM coal costs of
$0. 81 /MM Btu, the required market price of the desulfurized coal ranges from
$1 . 14 to $1 . 32/MM Btu on a utility financed basis and $1 .41 to $1 . 86 /MM Btu on
an investor financed basis.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSAT1 Field/Group
Air Pollution
Coal
Desulfurization
Pyrite
Air Pollution Control
Stationary Sources
Meyers Process
Chemical Coal Cleaning
13B
08G,21D
07A,07D
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
!1. NO. OF P
312
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
295
------- |