&EPA
United States Industrial Environmental Research EPA-600//-79-013a
Environmental Protection Laboratory January 1979
Agency Research Triangle Park NC 27711
Reactor Test Project for
Chemical Removal of
Pyritic Sulfur from Coal;
Volume I. Final Report
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wideirange of energy-related environ-
mental issues. '
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endtirsemerit or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-013a
January 1979
Reactor Test Project for
Chemical Removal of Pyritic
Sulfur from Coal;
Volume I. Final Report
by
R.A. Meyers, M.J. Santy, W.D Hart,
L.C. McClanathan, and R.A. Orsini
TRW, Defense and Space Systems Group
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-1880
Program Element No. EHB527
EPA Project Officer: Lewis D. Tamny
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
Plant checkout and shakedown was completed at the end of September and
initial plant process performance was evaluated on an Appalachian coal. Oper-
ation of the plant through January of 1978 demonstrated that the Reactor Test
Unit (RTU) could be run continuously in three-shift operation to reduce the
coal from 2,4 Ibs S02/106 Btu to a level of 1.0 to 1.2 Ibs S02/106 Btu, after
rinsing and extraction of generated elemental sulfur. There was no measurable
coal oxidation during processing and leach rates in the plant were greatly
improved over bench-scale values. The leach solution/coal/oxygen environment
was found to be corrosive to the installed stainless steel reactor, necessi-
tating future upgrading to support additional testing. Bench-scale experimen-
tation showed that the leach solution can be used as a homogeneous dense-media
to efficiently gravity-separate coal prior to processing. Beneficial engineer-
ing cost improvements are obtained based on using this approach, resulting in
capital cost estimates of $68-$69/KW and with $0.44-$0.50/106 Btu processing
costs, including amortization of capital, for input coal costing $0.78-
$0.81/106 Btu. Overall energy efficiency was 93 to 96 percent.
ii
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the valuable assistance received in this
project from the following TRW personnel: Dwight Wever, Bill Bowes, John Hunt
and Ed Phipps for supervising operation of the test plant; William Bradshaw
and Allen Keenan, lead plant technicians; Jack Denson and Larry Ledgerwood for
control lab operation and experimental support; Edwin Moon, Debbie Hopp and
Betty Cruz for engineering analysis support; Lou Resales and Maurice Bianchi
for materials study support; Bernard Dubrow, Les Van Nice and Elias Koutsoukos
for managerial assistance and manuscript review; Roger McGough, subcontract
manager; Verna Melough, Sharon Cavin, Monique Tholke and Velnia Butler for
technical typing and Marilyn Jennings for report coordination and finalization.
The authors also want to acknowledge the support of Dave Tamny, Jim Kilgroe
and T, Kelly Janes, the monitoring project officers and sponsoring managers at
the Environmental Protection Agency.
Gratitude is due to the American Electric Power Service Corporation for
providing 300 tons of coal for this initial operation phase of the plant and
most particularly to Sam Ruggeri of that organization.
Ill
-------
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
1 Btu
1 Btu
1 kw
1 hp (electric)
1 psi
5/9 C°F-32)
1 inch
1 ft
1 ft2
1 gallon
1 pound
1 ton (short)
Metric
252 calories
2,93 x 10"4 kilowatt-hours
1,000 joules/sec
746 joules/sec
0.07 kilograms/cm2
oc
2,54 centimeters
0,3048 meter
0,0929 meters2
0.0283 meters3 or 28.3 liters
3.79 liters
0.4536 kilograms
0.9072 metric tons
iv
-------
CONTENTS
ABSTRACT li
ACKNOWLEDGEMENTS . . 111
METRIC CONVERSION FACTORS - 1v
FIGURES - viii
TABLES xi
1. INTRODUCTION 1
2. CONCLUSIONS 8
3. RECOMMENDATIONS 11
4. REACTOR TEST UNIT 12
4.1 Process Description 12
4.1.1 Coal Feed System 15
4-1.2 Leach Solution Feed System 15
4.1.3 Oxygen Feed System 20
4.1-4 Process Steam- 20
4.1.5 Coal-Reagent Mixing 22
4.1.6 Primary Reaction and Reagent Regeneration 22
4.1.7 Secondary Reaction 24
4.1.8 Filtration and Leach Solution Recovery 24
4.1.9 Coarse Coal Leaching and Reagent Depletion . 26
4.1.10 Instrumentation 28
4.1.11 Process Sampling and Analysis 28
4.2 Process Equipment 30
4.2.1 Coal Transport Bins 31
4.2.2 Bin Tilter 31
4.2.3 Coal Storage Tank 31
4.2.4 Weigh Belt 31
4.2.5 Tank Farm 32
4.2.6 Leach Solution/Slurry Pumps 32
4.2.7 Heat Exchangers 34
4.2.8 Slurry Mixing Tank 34
4.2.9 Reactors 35
4.2.10 Flash Drum 36
4.2.11 Scrubbers 36
4.2.12 Filter and Vacuum System 37
4.2.13 Og/Ng/Air Supply Systems 38
4.2.14 Steam Supply System 38
-------
CONTENTS (Continued) Dano
Kayc
•^o
4.2.15 Process Water System • ^°
4.2.16 Instrumentation -*°
4.2.17 Control Laboratory ™
4.2.18 Capistrano Test Site Support Facilities 41
5. REACTOR TEST UN'IT OPERATION 42
5.1 RTU Shakedown Activities 42
5.1.1 RTU Safety Review and Training • • • 43
5.1.2 Shakedown Operations 43
5.2 RTU Operation 48
5.2.1 Operating Procedures 48
5.2.2 Equipment Experience 50
5.3 RTU Operation Conclusions 55
6. REACTOR TEST UNIT DATA 57
6.1 Summary of Previous Bench Scale Experimentation 57
6.2 Data from the Reactor Test Unit 61
6.2.1 Mine-cleaned Martinka Coal Characterization 65
6.2.2 Coal Processing Data 72
6.2.3 Slurry Sampling 99
6.2.4 RTU Mass Balance Data 104
6.3 Data Analysis Conclusions 113
7. SUPPORTING BENCH-SCALE EXPERIMENTATION . . 117
7.1 Introduction 117
7.2 Gravichem Separation 117
7.2.1 Background 117
7.2.2 Equilibrium Gravichem Separation .... 119
7.2.3 Non-Equilibrium Separation 131
7.3 Processing of RTU Coal 143
7.3.1 Pressurized Bench-Scale Processing . 146
7.3.2 Ambient Pressure Bench Scale Processing 146
7.3.3 Weathering of RTU Coal 146
7.3.4 Density of Coal Slurry 148
7.3.5 Elemental Sulfur Recovery 150
7.4 Laboratory Study Conclusions 157
8. ENGINEERING DESIGN AND COST ESTIMATION 159
8.1 Introduction and Background 159
8.1.1 Historical Background 160
8.1.2 Current Generalized Process Concept 161
8.1.3 Base Case Design, Cost Estimate, and —
Economics Approach 169
vi
-------
CONTENTS (Continued)
Page
8.2 Design and Cost Estimate Base Cases 177
8.2.1 Base Case 1 -Mine Cleaned Martinka Coal 177
8.2.2 Base Case 2 - TVA Kentucky No. 9 Coal 190
8.3 Engineering Analysis Conclusions 203
8.3.1 Base Case Results 203
8.3.2 Cost Sensitivity Analysis 206
9. VENDOR TESTING 210
9.1 General Approach 210
9.1.1 Equipment Suppliers Survey and Selection 220
9.1.2 Vendor Test Implementation Plan 223
9.1.3 Vendor Testing Results 224
9.1.4 Coal Supply for Vendor Studies 225
9.2 Vendor Testing Conclusions 225
10. MATERIALS OF CONSTRUCTION EVALUATION 228
10.1 Inspection and Analysis of RTU Equipment 229
10.2 Analysis of Test Samples 241
10.3 Discussion 256
10.4 Corrosion Study Conclusions 258
11. REFERENCES 260
12. GLOSSARY OF ABBREVIATIONS AND SYMBOLS 263
vii
-------
FIGURES
Number
Page
1 Meyers Process 2
2 Reactor Test Unit 3
3 Reacted Coal on Belt Filter 5
4 Gravichem Process 6
5 RTU Schematic 14
6 RTU Bin Lift I6
7 RTU Bin Tilter 17
8 Coal Feed to Fine Coal Storage and the Coarse Coal Leacher .... 18
9 Coal Feed to the Mixing Unit . 19
10 RTU Tank Farm 21
11 RTU Mixer and Primary Reactor 23
12 Primary Reactor and Belt Filter 25
13 Vacuum Pump and Filtrate Receivers 27
14 RTU Control/Instrumentation Console 29
15 Comparison of Capistrano and Warner Laboratories Pyritic
Sulfur Analyses 68
16 Comparison of Capistrano and Warner Laboratories Sulfate
Sulfur Analyses 69
17 Mine-Cleaned Martinka Coal Size Distribution Data 71
18 Pyritic Sulfur Analyses of Processed Coal from Experi-
ment 01-03 76
19 Arrhenius Plot of Data from RTU Experiments 01-01 to 01-11 .... 81
20 Reagent Y Data from RTU Processing Experiments 01-01 and
01-02 90
21 Summary of Estimated Sampling Bias 101
22 Slurry Concentration Variation as a Function of Sample
Location 103
-------
FIGURES (Continued)
Number Page
23 Gravichem Process ng
24 Hot-Briquetted Coal 121
25 Bench-Scale Gravichem Test Apparatus Flow Diagram 123
26 Gravichem Processing of TVA Coal 125
27 Puritic Sulfur Leaching for 1.3 Specific Gravity Sink Coal
at 102°C 132
28 Coal Oxidation After Two Hours at 100°C as a Function of
Ash Content 133
29 Particle Size Distribution in Fractions of RTU 3 Coal 138
30 Bench-Scale Coal Leaching and Reagent Regeneration
Apparatus 133
31 Elemental Sulfur Recovery as a Function of Acetone
Extraction Stages 154
32 Effect of Water on Elemental Sulfur Recovered by Successive
Acetone Stages 155
33 History of the Gravichem Process 162
34 Gravichem Process Block Diagram Suspendable Coal Approach 163
35a TRW Coal Desulfurization Gravichem Process 178
35b TRW Coal Desulfurization Gravichem Process 179
36a TRW Coal Desulfurization Gravichem Process 195
36b TRW Coal Desulfurization Gravichem Process 196
37 Base Case 1. Upgrading Cost vs. Battery Limit Capital 208
38 Base Case 2. Upgrading Cost vs. Battery Limit Capital 209
39 Base Case 1. Upgrading Cost vs. ROM Coal Cost 210
40 Base Case 2. Upgrading Cost vs. ROM Coal Cost 211
41 Base Case 1. Upgrading Cost vs. % Pyritic Sulfur Removed 212
42 Base Case 2. Upgrading Cost vs. % Pyritic Sulfur Removed 213
43 Base Case 1. Upgrading Cost vs. ROM Coal Cost at Different
Levels of Pyritic Sulfur Removal 214
44 Base Case 2. Upgrading Cost vs. ROM Coal Cost at Different
Levels of Pyritic Sulfur Removal 215
45 Base Case 1. Upgrading Cost vs. % of Total Coal to Float
Processing . . .• 216
1x
-------
FIGURES (Continued)
Number
Page
46 Base Case 2. Upgrading Cost vs. % Total Cost to
Float Processing. . . ^'7
47 Base Case 1. Upgrading Cost vs. ROM Coal Cost 218
48 Base Case 2. Upgrading Cost vs. ROM Coal Cost 218
49 Flush Port Fitting 231
50 Porous Spongy Structure 232
51 Pipe Weld Corrosion 233
52 Crevice Corrosion in Reducer Fitting. .... 234
53 Flange Crevice Corrosion 235
54 Baffle Bolts and Nuts from R-l Cell 4 237
55 Weir Baffle 238
56 TE-56 Thermocouple Probe 239
57 R-l Reactor Internals After Experiment 03 Showing Pits
Where Grinders Had Operated 240
58 R-l Reactor Internals After Experiment 03, Showing Smooth
Bottom Pits at Welds and on Free Surfaces 242
59 Coupon Mounting Rack and Test Speciments 245
60 316L Cres Coupons Showing Signs of Crevice Attack During
Experiment 01 (L) and Smooth Bottom Pitting During
Experiment 03 (R) 246
61 304 Stainless Steel Coupons from Experiment 03 At and
Above Water Line (L) and Below Water Line (R) 247
62 Inconel and Incoloy Alloy Coupons, 7.5X: Inconel 601
(Upper L), Inconel 617 (Upper R), Inconel 625 (Lower L),
Incoloy 825 (Lower R) 249
63 Hastelloy C-276 Coupons From Experiment 03 At and Above
Water Line (L) and Below Water Line (R) 250
64 Titanium Coupons: T1-50A (L) and Ti-12 (R), from
Experiment 01 (Top) and Experiment 03 (Bottom) 251
65 Lead Coupons from Experiment 01. Surface Marked by
Cleaning Operation. No Pitting or Crevice
Corrosion. 1.5X 252
-------
TABLES
Number Page
1 RTU Pump Data 33
2 RTU Shakedown Sequence 45
3 Equipment Ranges Utilized 51
4 RTU Channel-Scan at Fixed Time 63
5 RTU Time-Scan by Channel 64
6 As Received Mine-Cleaned Martinka Coal Analyses 66
7 Size Distribution Data for Mine-Cleaned Martinka Coal
Processed in the RTU 70
8 Operating Parameters for Acidified Iron Sulfate Reagent
Leach Experimentation 75
9 Summary of Processed Coal Analyses Obtained During RTU
Experimentation with Acidified Iron Reagent 77
10 Operating Parameters for Low Iron Reagent Leach
Experimentation 83
11 Summary of Processed Coal Analyses Obtained During RTU
Experimentation with Low Iron Reagent 84
12 Excess Fe+2 Generation Data from Processing 14 Mesh x 0
Mine Cleaned Martinka Coal with 5% w/w Iron Reagent 89
13 RTU Reagent Regeneration Data 91
14 Elemental Sulfur Product Generation Data from RTU
Processing 95-96
15 Summary of Heat Content Changes in RTU Processed Coals 98
16 T-2 Mass Balance Data from RTU Processing Experimentation .... 109
17 Oxygen Mass Balance Data Obtained During RTU
Experimentation Ill
18 Equilibrium Gravichem Tests on 14 Mesh x 0 Martinka Mine
Coal with Variation of Medium and Separation Time 122
19 Equilibrium Gravichem Tests at 80°C on 3/8" Topsize
Kentucky No. 9 Mine Coal with Variation of Medium 126
20 Particle Size Distribution of Size-Reduced Kentucky No. 9
Sink Coal 127
x1
-------
TABLES (Continued)
Number
21 Gravichem Processed Kentucky Mine No. 9 Coal 128
22 Sulfur Removal from Coal with Ferric Sulfate Solutions 13°
23 Comparative Results of Cleaning 134
24 Analyses of Fractions of Martinka Coal Obtained from
Nonequilibrium Float/Sink Experiments 135
25 Analyses of Fractions of Martinka Coal Obtained from
Nonequilibrium Float/Sink Experiments • • • • 137
26 Relative Settling Rates of Coal and Pyrite Subjected to
Gravitational and Centrifugal Forces 140
27 Flotation and Centrifugal Separation of Coals 141
28 Yield of Martinka Coal Separated in Nonequilibrium Float/
Sink Experiments after 0.5 Hours @ 80°C (S.G. 1.3) 142
29 Leaching of Martinka Coal at 80°C 145
30 Bench-Scale Processing of Martinka Coal. . : 147
31 Distribution of Inorganic Sulfur in RTU Coals (%) . . 149
32 Elemental Sulfur Content of Nonprocessed Coal. . . , 150
33 Density of Coal/Water Slurries :151
34 Product Elemental Sulfur Recovery from Leach Martinka Coal .... 153
35 Sources of Equipment Cost Information 172
36 Sources of Operating Cost Information 173
37 Battery Limits Operating Costs Format 174
38 Economic Evaluation Criteria Utility Financing 176-177
39 Base Case 1 - Process Equipment List 184-188
40 Battery Limit Process Costs -Base Case 1 . 189
41 Base Case 1 - Process Economics 191
42 Base Case 2 -Process Equipment List 197-2C1
43 Battery Limits Process Costs - Base Case 2 202
44 Base Case 2 - Process Economics 204
45 Summary of Base Case Economics Results ..... 205
46 Vendor Selection Program 221-222
47 Coal Cake Collected for Vendor Testing 226
48 Coupon Corrosion Test Results - 29 November Inspection 244
xi i
-------
1. INTRODUCTION
The Meyers Process is a technology for chemically removing essentially all
of the pyritic sulfur from coal through a mild oxidative treatment. Important
pollutant trace elements of lead, cadmium, and arsenic are removed at the same
time. It is particularly cost effective for providing compliance coal for
industrial boilers and smaller electric utilities, and for recovering and
desulfurizing waste fine coal rejected from mining and washing operations.
A process schematic is shown in Figure 1. Coal is mixed with an aqueous
solution of ferric sulfate (Step 1), previously derived from the coal, to form
a slurry. The slurry is raised in temperature to 100°-130°C (Step 2) where the
ferric sulfate oxidizes the pyritic sulfur content of the coal to form elemen-
tal sulfur and additional iron sulfate. At the same time oxygen or air is
introduced to regenerate the reacted ferric sulfate. Iron sulfate dissolves
into the leach solution while the elemental sulfur is removed in a second
extraction (Step 3). The coal is dried and solvent recovered (Step 4). The
products of the process are iron sulfate, which may be limed to give a dry
gypsum and iron oxide material and elemental sulfur. Trace elements from the
coal are rejected from the leach solution with the stabilized gypsum-iron oxide
solid. Elemental sulfur is the most desirable product which can be obtained in
the process of controlling sulfur oxide pollution since it may be easily stored
without additional pollution or may be marketed. The gypsum-iron oxide product
is a safe and storable solid product.
M p 3}
After testing the process on some 40 U.S. coals* * * ' and performing
200 fully material balanced bench-scale extractions/ * ' the value of the
process for controlling the sulfur content of coal was firmly established and
the data necessary for design of a test plant was then available. Engineering
design background obtained through extensive studies at both TRW* ' ' and
various engineering organizations*''8' provided confidence that the process
-------
H2S04
Co O-
ro
COAL
(Fe S9)
Fe2(S04)3
FILTER
NEUTRAL-
IZATION
Co
Fe S2 + Fe
Fe S0 +
Fe SO4
(S04)
FILTER
COAL
SOLVENT
STILL
Figure 1. Meyers Process
-------
was economically attractive and indicated important engineering data which could
only be obtained at test plant scale.
It was determined that the initial test plant evaluation should concentrate
on the key process steps of coal-leach solution slurry formation, leaching,
regeneration and filtration. A test plant, termed the Reactor Test Unit, was
constructed at TRW's San Juan Capistrano site for the purpose of testing these
portions of the process (Figure 2). The plant, sized to process from 1/8 to
1/3 ton/hour of coal was dedicated in April, 1977.
Figure 2. Reactor Test Unit
-------
Plant checkout and shakedown was completed at the end of September and
initial process performance was evaluated on coal donated by the American
Electric Power Service Corporation (AEP) from its Martinka mine (Appalachian
coal). Operation of the plant through January of 1978 demonstrated that the
Reactor Test Unit (RTU) could be run continuously in three-shift operation to
reduce the AEP coal from 2.4 Ibs S02/106 Btu to a level of 1.0-1.2 Ibs S02/10
Btu, after rinsing and extraction of generated elemental sulfur. The coal
product is shown as a cake on the plant belt-filter in Figure 3. There was
no measurable coal oxidation. Leach rates in the RTU were greatly improved
over bench-scale values, allowing elimination of the use of a second process
reactor for completing pyrite removal. The leach solution/coal/oxygen environ-
ment was found to be corrosive in the installed stainless steel reactor,
necessitating future upgrading of the reactor material of construction to
support additional testing.
Supporting bench-scale experimentation showed that the iron sulfate-
sulfuric acid leach solution can be used as a homogeneous dense media to effi-
ciently gravity-separate fine coal at specific gravities of 1.2 to 1.35. Ben-
eficial engineering cost improvements are obtained, based on using this
gravity-separation effect, whereby a significant portion of the input coal
which floats in the leach solution and is almost pyrite free, may bypass the
reactor, elemental sulfur extraction and dryer portions of the Meyers Process
(Figure 4). This revised technology is termed the Gravichem Process. When
applied, at bench-scale, to a Tennessee Valley Authority (Interior Basin) coal
containing 12% ash and 7 Ibs of S02/106 Btu, two products are obtained, a 4%
ash float coal containing 3 Ibs S02/106 Btu and an 11-12% ash sink coal con-
taining 4 Ibs S02/106 Btu after treatment by the Meyers Process. Both of these
products meets State SO emission standards for this coal. Because of these
A
promising results, TVA has shipped 300 tons of coal to TRW to demonstrate
desulfurization in the RTU.
Laboratory experimentation also showed that the solvent system, acetone
and water, is the most economically attractive method thus far investigated for
removal of generated elemental sulfur from treated coal.
Process cost forecasts based on both RTU data and supporting laboratory
experimentation for the Gravichem Process are $68-69/KW capital cost with
-------
Figure 3. Reacted Coal on Belt Filter
-------
cn
H2S04-
Fe2(S04)3
H20
Fe
GRAVITY
SEPARATOR
(FeS)
Co O-
NEUTRAL-
IZATION
REACTOR
§>
SOLVENT
FILTER
SOLVENT
STILL
MEYERS PROCESS
A
SULFUR
Figure 4. Gravichem Process
-------
$0.44-0.50 10 /Btu, processing costs, including utility financed capital
amortization, for input coal costing $0,78"0,81/10b/Btu. Overall energy effi-
ciency for these process designs including both coal use for process heat and
electric energy for plant operation is 93% to 96%,
The report is organized into nine sections to follow. Sections 2 and 3
present the conclusions and recommendations, respectively, which resulted in
this study. The RTU is described in Section 4 and the operation of the plant
in Section 5. A complete set of RTU data together with analysis of the data is
presented in Section 6. The supporting bench-scale laboratory results are found
in Section 7 and engineering design and full-scale process cost estimation in
Section 8. A plan for vendor testing of process unit operations, not built into
the RTU, is described in Section 9 and a detailed plant materials evaluation is
presented in Section 10. The referenced appendices are continued in Volume II
of this report.
-------
2. CONCLUSIONS
PLANT OPERATION RESULTS
1. The Reactor Test Unit (RTU) can be operated continuously for
testing of Meyers Process units for coal-leach solution mixing,
simultaneous coal leaching and leach solution regeneration,
filtration of leach solution from treated coal and water wash-
ing of coal on the filter.
2. The input coal from American Electric Power Service Corpora-
tion's Martinka mine in Fairmont, West Virginia, containing
1% inorganic sulfur can be reliably and continuously reduced,
in the RTU, to a pyritic sulfur level of 0.16% without any
measurable coal loss and with coal heat content increases
averaging 350 Btu/lb.
3. RTU coal product, after bench-scale extraction of residual
sulfate and elemental sulfur, was continuously and reliably
reduced to a total sulfur content of 0.68-0.75% and 1.0-1.2
Ibs S02/106 Btu.
4. Leach rates in the RTU were improved over bench-scale values by
an average factor of five due mainly to favorable coal seg-
regation in the primary reactor.
5. Plant Teacher-regenerator operation at temperatures ranging
from 230°-270°F (1100-132°C), pressures of 30-80 psig and
residence time of 5-8 hours was demonstrated.
6. The use of a single reactor-regenerator is sufficient to meet
design basis pyrite removal and provide regenerated leach
solution. The use of a secondary reactor to complete the
reaction of pyrite with leach solution, found to be necessary
at bench-scale, was demonstrated to be unnecessary in the
Reactor Test Unit.
7. The leach solution/coal/oxygen environment caused corrosion
in the primary reactor-regenerator system indicating that
upgrading of the 316L material of construction for improved
corrosion resistance is needed to support further testing.
8. The following materials were found to be suitable for leach
solution-coal service at temperatures up to 90°C: fiber rein-
forced plastics, elastomers and 316L stainless steel. The
-------
following materials were determined to be suitable for service
at reactor-regenerator temperatures: up to 130°C, titanium,
Hastaloy, and rubber-lined brick over mild steel.
9. No significant corrosion was observed in the leach solution-
coal mix tank, flash-down tank, storage tanks or reactor-
regenerator pumps.
SUPPORTING EXPERIMENTATION RESULTS
10. The iron sulfate-sulfuric acid leach solution can be used as a
homogenous liquid to efficiently gravity-separate fine coal
from pyrite rich coal at specific gravities of 1,2 to 1,35.
11. Beneficial engineering cost improvements are obtained by
using this gravity-separation effect to bypass a significant
portion of the input coal around the reactor, elemental sulfur
extraction and dryer units of the Meyers Process. This
revised process is termed the Gravichem Process.
12. The Gravichem Process provides two products with no coal
reject, a float coal containing 2-4% ash with almost no pyritic
sulfur and a sink coal generally lower in ash than the input
coal and also nearly pyrite-free after treatment by the >
Meyers Process. The two products can be used separately or
combined,
13. Bench-scale testing of the Gravichem Process on the American
Electric Power Service Corporation (Appalachian) coal gave two
products: a float coal containing 1.0 Ibs S02/106 Btu and a
sink coal containing 1.1 Ibs SQ2/106 Btu after treatment by
the Meyers Process, Both products met New Source Performance
Standards.
14. Bench-scale testing of the Gravichem Process on a Tennessee
Valley Authority (Eastern Interior Basin) coal containing
12% ash and 7 Ibs of S02/106 Btu gave two products: a 4% ash
float coal containing 3 Ibs S02/106 Btu and an 11-12% ash
sink coal containing 4 Ibs S02/106 Btu after treatment by the
Meyers Process.
15. The solvent system, acetone and water, is the most economi-
cally attractive method thus far investigated for removal of
generated elemental sulfur from treated coal. This solvent
also dissolves and removes residual iron sulfate.
ENGINEERING DESIGN RESULTS
16. Process cost forecasts for the Gravichem Process are $68-69/KW
capital cost with $0.44-0.50/106 Btu, processing costs includ-
ing utility financed capital amortization, for input coal
costing $0.78-0,81/106 Btu.
-------
17. Coal energy efficiency is 94-97% for the Gravichem Process
including coal use for process heating,
18. Overall energy efficiency including both coal use and electric
energy for plant operation is 93-96%.
10
-------
3. RECOMMENDATIONS
1. Design and replace the corroded primary reactor with an
improved vessel constructed of either titanium or rubber-
lined brick over mild steel,
2. Construct and test in integrated operation, a leach-solution
gravity-separator unit to evaluate Gravichem separation and
Meyers Process treatment of sink coal.
3. Test the operation of the plant on a coal from the Eastern
Interior region of the United States, such as the Tennessee
Valley Authority coal, 300 tons of which is now on hand.
4. Construct and test in integrated operation, Meyers Process
units for continuous residual sulfate extraction, generated
elemental sulfur extraction, coal drying, recovery of sol-
vent, water and elemental sulfur by distillation and liming
out of generated iron sulfate.
11
-------
4. REACTOR TEST UNIT
During the past seven years, TRW has been involved in various stages of
development of the Meyers Coal Desulfurization Process. The early work
involved laboratory and bench scale testing^1'2'3'4'5' which was aimed at
obtaining process application information and process design data. Those
early efforts led to the design of the Reactor Test Unit (RTU). The design
of the RTU was carried out in two steps; first a preliminary design to
evaluate the overall approach and probable cost of the envisioned unit
and secondly, a detailed engineering and design effort which immediately
preceded equipment procurement and RTU construction. The detailed engineer-
ing, procurement, and construction of the RTU occurred over a 19-month period.
Immediately following plant construction, TRW initiated operator training,
safety check-out, procedure verification, and RTU shake-down in May 1977.
This Section presents a detailed description of the RTU as constructed.
Intended uses and capabilities of the integrated RTU are discussed as are
those of each major unit operation sub-system in Section 4.1. Section 4.2
details the specific equipment which is incorporated into the RTU. Infor-
mation such as equipment type, manufacturer, size, model, configuration,
capacity, operating limits, and materials of construction is presented.
4.1 PROCESS DESCRIPTION
The Reactor Test Unit is a pilot scale coal processing facility designed
to demonstrate those unit operations comprising the front end of the Meyers
Process, namely, coal-reagent mixing, primary pyrite reaction and reagent
regeneration, secondary (finishing) pyrite reaction and slurry filtration.
Designed for high flexibility, the RTU has the capability of processing a
range of suspendable coals up to approximately 8 mesh top-size and coarse
coals up to approximately 3/8-inch top-size. Spent reagent may be regenerated
either exclusive of, or simultaneously with, coal leaching. The primary
reactor may be used as either a five-stage or a three-stage reaction unit to
increase the available range of coal processing times.
12
-------
The most common modes of RTU operation are represented schematically in
Figure 5. Fine coal ground to the desired size is loaded into feed tank
T-l. Dry coal is fed continuously by a live bottom feeder A-2 to the weigh
belt A-3 which discharges through rotary valve A-4 to the three-stage mixer
T-2 (stream 1). Aqueous iron sulfate leach solution (stream 2) enters T-2
after preheating in a heat exchanger E-2 and passing through the foam scrubber
T-3. Steam is added (stream 3) to raise the slurry to its boiling point.
Foaming, which may occur during the early stages of mixing, ceases when coal
particle wetting is complete. The heated slurry (stream 4) is then pumped
to a five-stage pressure vessel R-l in which most of the pyrite is removed.
R-l Slurry heating is achieved by direct injection of steam into any or all
reaction stages. Reagent regeneration may be carried out simultaneously
with pyrite leaching by means of oxygen injection into any or all reaction
stages (stream 5). Unusued oxygen saturated with steam (stream 6) is con-
tacted in a foam knock out drum V-l with the feed reagent (stream 7) to pre-
heat the reagent and cool vent gases. Slurry in any stage of R-l may be
cooled by means of a cooling water heat exchanger E-l which may be applied
to slurry recirculation loops for removal of excess heat of reaction. Vent
gas from both T-3 and V-l are water scrubbed in T-4 to remove any traces of
acid mist.
Reacted coal slurry (stream 8), at elevated temperature and pressure, is
flashed into a flash drum T-5 for gas-liquid separation. Generated steam
(stream 9) is condensed in T-4, and the condensate plus any entrained acid
mist is removed with scrubber water. Reacted slurry (stream 10) is fed to the
belt filter S-l. The filtrate, which is regenerated leach solution, is
removed from the coal slurry through the evacuated filtrate receiver V-2 and
pumped (stream 12) to the leach solution storage tank T-7. Coal on the filter
belt is washed with water (stream 11) and discharged to coal storage. Wash
water is removed through the evacuated wash water receiver V-3 and pumped
(stream 13) to the liquid waste holding tank T-9 for subsequent disposal.
As a processing alternate, partially processed slurry from T-5 may be
loaded into the secondary reactor R-2 for final depyritization in a batch
mode. Slurry may be retained, heated and agitated, for extended periods of
time in R-2 prior to being pumpted to S-l.
13
-------
J I COAL
\7 MECHANICAL
v
ATM OS.
WATER
RETURN
P-12
P-M
Figure 5. RTU Schematic
-------
The coarse coal contacting vessel T-6 is a heated and insulated tank in
which hot reagent may flow through a bed of retained coarse coal. This unit
is used principally to convert regenerated leach solution in storage tanks
T-7 or T-8 to a more depleted solution simulating recycle reagent after
secondary reaction. This capability is required since R-2 is not utilized
during all experimentation. T-6 is basically a coarse coal reactor and, if
appropriate sampling ports and possibly some flow distribution internals were
added, it could be used to obtain design data for coarse coal processing.
Detailed descriptions of the principal RTU coal processing operations
are presented in the ensuing sections.
4.1.1 Coal Feed System
Process feed coal is transported from the Ultrasystem's coal grinding
facility located in Irvine, California, to the Capistrano Test Site in steel
bins which maintain a nitrogen atmosphere to prevent coal weathering. Coal
bins are hoisted to the top level of the RTU test stand and emptied by a
hydraulic bin tilter-vibrator A-l (Figures 5 anc| 7). DUSt control is main-
tained by coupling the bin vent to transport ducting with flexible sleeving.
Coal feed is routed according to the coal top-size. During coarse coal
utilization (1/4-inch top size or larger), feed coal is diverted from the bin
discharge to the T-6 leach tank where it is contacted with reagent to effect
leach solution depletion. Fine coal processing feed material is charged to
the coal storage tank T-l (Figure 8). A vibrating discharge unit A-2 at the
bottom of T-l prevents solids bridging or packing and maintains coal flow to
the weigh belt A-3. Fine coal feed is metered by A-3 to a moisture lock
rotary valve A-4 which feeds the coal reagent mixing vessel T-2 (Figure 9).
The weigh belt is instrumented to display coal feed rates and cumulative
coal flow on the plant control panel. All dry coal storage and transfer
units downstream of the bin discharge are operated under nitrogen pressure
to prevent oxidation of the coal feed.
4.1.2 Leach Solution Feed System
Leach reagents utilized during RTU operation were generally blended in
the secondary reactor R-2 and pumped to reagent storage tanks via P-7. Iron
15
-------
Figure 6. RTU Bin Lift
16
-------
Fiqure 7. RTU Bin Tilter
17
-------
Figure 8. Coal Feed to Fine Coal Storage
and the Coarse Coal Leacher
18
-------
FINE COAL DISCHARGE A-2
Figure 9. Coal Feed To The Mixing Unit
19
-------
sulfate reagent was stored in leach, solution surge tank. T~7 while distilled
water or iron free leach reagent was stored in surge tank T-8 (Figure 10).
Quantities of reagent in storage were monitored with buoyant-type liquid
level guages.
Reagent from either T-7 or T-8, as required, was pumped to the reactor
system by P-13. Prior to actual coal contacting and reaction, reagent from
P-13 was used to break down foam and demist primary reactor vent gases in
knock-out drum V-l. Reagent and any entrained mist or foam was brought up
to the desired temperature for the first stage of mixing (typically, 170 F)
by heat exchanger E-2 and fed to the foam scrubber T-3 to break down any foam
generated during the mixing operation. Reagent was gravity fed from T-3 to
the first stage of the T-2 mixer. Reagent feed rates are monitored by a
magnetic flow meter located between E-2 and T-3.
4.1.3 Oxygen Feed System
Oxygen is stored as a liquid at the RTU facility and vaporized as
required for process reagent regeneration. The total flow rate of gaseous
02 (200 psig) to the primary reactor R-l was monitored by target flow meter
FE-61 and feed 0,, was subsequently partitioned among the reaction stages with
the aid of rotameters FI-62 through FI-66. Oxygen-slurry contacting is
achieved by injecting the 02 stream into slurry recirculation loops associated
with each reactor stage. The point of 02 injection is located approximately
ten feet upstream from the point of slurry reentry into the reactor stage.
Slurry recirculation rates were maintained sufficiently high to ensure
turbulent flow and vigorous blending of reagent and oxygen.
Unreacted oxygen flows from R-l through the knock-out drum V-l and
pressure relief valve PV-43 to the vent gas scrubber T-4. Vent gas flow
rates from R-l are measured by target flow meter FE-44 which is located
between V-l and PV-43. The oxygen content of vent gas is measured by the
oxygen analyzer AE-171 which analyzes a dried slipstream of the V-l exhaust.
4.1.4 Process Steam
Steam for process heating was generated by a Clayton steam generator
which had the capability of generating 1000 pounds of 150 psig steam per
20
-------
Figure 10. RTU Tank Farm
-------
hour. A majority of the process steam required is used for direct contact
heating of slurry streams and wet coal cakes although several heat exchangers
are used in the RTU. Flow rates of steam into T-2 and R-l for direct contact
heating are measured with rotameters FI-16 through FI-18 and FI-67 through
FI-71, respectively. Steam feed rates are set and maintained manually. Flow
rates of steam into heat exchangers are regulated by temperature actuated
flow controllers.
4.1.5 Coal-Reagent Mixing
Initial coal-reagent contacting, coal wetting and slurry defoaming were
performed in the mixing unit T-2 under atmospheric pressure (Figure 11). The
mixer is partitioned into three stages by weirs which may be adjusted to vary
stage volumes and, consequently, slurry retention times. Each stage is
equipped with a variable speed mixer to promote coal wetting and prevent
settling. Slurry temperature was increased stagewise from approximately
170°F to the solution normal boiling point (nearly 214°F for iron sulfate
reagents) by direct contact heating. Shielded thermocouples are situated
in each stage of T-2 to permit continuous monitoring of slurry temperature.
Gases or foam generated in T-2 are treated by the foam scrubber T-3 and
subsequently by the vent gas scrubber T-4.
Thoroughly wetted slurry was pumped from T-2 to the primary reactor R-l
by slurry pump P-l. The rate at which slurry was pumped through P-l was
regulated by slurry level controller LIC-26 in the third stage of T-2.
4.1.6 Primary Reaction and Reagent Regeneration
Simultaneous coal leaching and reagent regeneration were performed in
the primary reactor at nominal temperatures up to 270°F and pressures up
to 100 psig. Design capabilities of R-l are somewhat higher with a maximum
temperature and pressure of 300°F and 150 psig, respectively. The R-l
reactor consists of five stages separated by weirs. All five stages may
be used for reaction or, alternately, the P-l slurry feed may be diverted
around the first two stages of R-l to effectively yield a three-stage reactor.
Each reaction stage is a continuous flow stirred tank reactor in which
agitation is achieved by means of a variable speed mixer and a slurry
recirculator. The slurry recirculation loops are also used for steam
22
-------
ro
to
PRIMARY REACTOR R-l
Figure 11. RTU Mixer and Primary Reactor
-------
injection to maintain reaction temperature and oxygen injection for reagent
regeneration.
Reaction stages of R-l have interconnected overhead ullage and, therefore,
all stages are operated at the same pressure. R-l vents to the knock-out
drum V-l which demists vent gases and breaks down any foam which may be
generated during processing. Vent gases leaving V-l are further processed
in the vent gas scrubber T-4.
Processed slurry Is flashed from the fifth stage of R-l into
the slurry flash drum T-5 which is maintained at atmospheric pressure. Vent
gas and flashed steam from T-5 are processed through T-4. Underflow slurry
from T-5 is gravity fed to either the secondary reactor for further processing
or the filtration unit S-l. Slurry is flashed from R-l in pulses having
durations of up to three seconds with approximately 25 seconds between pulses.
4.1.7 Secondary Reaction
Coals requiring further processing after primary reaction in R-l are
gravity fed from the flash drum T-5 to the secondary reactor R-2 for batch
processing. Reaction in R-2 may be carried out at temperatures up to 250°F
in the absence of oxygen. Alternately, at temperatures below 250°F, oxygen
partial pressures may be maintained such that the system,total pressure does
not exceed 14 psig. Reaction temperatures in R-2 are maintained by electrical
heating tapes at the reactor surface to enable prolonged processing without
the slurry dilution inherent to direct steam heating. Slurry agitation is
provided by a single variable speed mixer. R-2 vent gases are processed
through the vent gas scrubber T-4. Fully processed slurry may be pumped from
R-2 to either the filtration unit S-l or to waste disposal T-9. When not in
service as a reaction unit, R-2 may be utilized for reagent formulation and
mixing.
4-1.8 Filtration and Leach Solution Recovery
Processed slurries from the primary reactor R-l (via flash drum T-5)
or from the secondary reactor R-2 were fed to the filtration unit S-l for
reagent recovery and coal washing (Figure 12). S-l is a belt-type filter
with two filtration zones having discrete filtrate collection systems. The
first filtration zone is for recovery of concentrated reagent which is
24
-------
ro
en
Figure 12. Primary Reactor and Belt Filter
-------
collected in a filtrate receiver V-2 (Figure 13) and recycled back to the
appropriate leach solution surge tank T-7 or T-8. The second filtration
zone is for recovery of hot wash water which is sprayed on the coal filter
cake to effect further reagent recovery. Filtrate from coal cake washing
is collected in the wash water receiver V-3 and pumped to disposal tank T-9.
Vacuum for filtration was provided by vacuum pump K-l which couples to the
filtrate receivers V-2 and V-3. Steam is applied to the coal cake during
filtration and washing to prevent cracking of the filter cake and maintain
cake temperatures of approximately 180 F.
The spray-washed coal cake was scraped from the filter belt and loaded
into storage containers or disposal dumpsters. Residual solids adhering to
the filter belt were removed with water sprays directed at the inner and
outer belt surfaces. The low solids slurry resulting from belt washing may
be recycled for cake wash to minimize losses of process solids or disposed
of directly.
Those process products not required for process characterization or
vendor testing were disposed of at class "A" landfills. Low solids content
liquid wastes from T-9 were disposed of by Industrial Trucking Co., Wilmington,
California. Discard coal cake was disposed of by Removal Inc., Lawndale,
California.
4.1.9 Coarse Coal Leaching and Reagent Depletion
Depleted leach solution for recycle reagent simulation is generated in
the coarse coal leach tank T-6. Coarse coal is loaded into T-6 from the bin
tilter A-l and may be preheated with steam which is introduced at the reactor
bottom. Hot leach solution from heat exchanger E-2 contacts the coarse coal
bed at the bottom of T-6, flows upward through the bed and is pumped to
reagent storage in T-7 or T-8. Alternately, reagent may be retained in T-6
for batch leaching. Spent reagent may be drained from T-6 through a screened
drain port at the vessel bottom. Near complete recovery of spent reagent is
effected by washing with water which flows upward through the coal bed.
Processed coal is discharged from the bottom of T-6 into a dumpster for
sampling and/or disposal.
26
-------
r-o
FILTRATE
RECEIVER V-2
Figure 13. Vacuum Pump and Filtrate Receivers
-------
4.1.10 Instrumentation
RTU instrumentation may be classified according to relative importance
of instrument function during plant operation. Primary instrumentation
requires continuous monitoring for maintenance of plant operational safety
and/or acquisition of data required for process evaluation. Data from primary
instrumentation are displayed at the plant control panel either in digital,
gauge or printed tape form, and are also recorded on magnetic tapes for
subsequent evaluation (Figure 14). Additionally, primary data are generally
displayed on the test stand to facilitate plant operation and local monitoring.
Examples of primary data sources include the weigh belt coal feed rates,
magnetic flow meters measuring reagent and slurry feed rates, thermocouples
monitoring reaction temperature, transducers monitoring reactor pressures
and level gauges indicating slurry depths in the mixer and reactors.
Secondary data sources are those providing information which is not
critical to safe plant operation or process evaluation. These data sources
are not monitored continuously with some being checked on a daily or weekly
basis. Data acquisition from secondary sources usually consists of manual
recording and tabulation. Examples of secondary data sources include flow
meters indicating seal flush water flow rates, steam flow meters and reagent
storage tank level indicators.
4.1.11 Process Sampling and Analysis
All major RTU processed solid, liquid, slurry and gas streams were sampled
and analyzed during plant operation. Analyses of solid and liquid streams
were performed at the RTU control laboratory. Additionally, selected coal
samples were analyzed by Warner Laboratories, Cresson, Pennsylvania. Gas
sampling and analysis was performed on a continuous basis by RTU equipment.
Sampling ports available for acquisition of feed reagent samples are
located at the leach solution feed pump P-13 and at the foam scrubber T-3.
Feed solids are sampled with a full stream coal diverter located at a point
between the weigh belt A-3 and rotary valve A-4. Stream diversions of up to
one minute were used during normal sampling although longer periods of time
were used during A-3 calibration. Slurry samples were drawn from the mixer
T-2 through sampling ports located in the wall of each mixing stage. Primary
28
-------
UD
Figure 14. RTU Control/Instrumentation Console
-------
reactor R-l slurry samples may be drawn from each stage through wall sampling
ports or by full stream diversion of slurry passing through recirculation
loops. Processed slurry sampling ports are located in the slurry flash drum
T-5 and in the secondary reactor R-2. Processed coal samples may be taken
directly from the filter belt S-l after the spray wash operation.
Feed coal samples were submitted for short proximate analysis (moisture,
ash, heat content and total sulfur analyses), sulfur forms analysis (pyritic
and sulfate sulfur analysis), ash iron content analyses and size distribution
determination. Reagent samples, whether feed reagent or slurry sample fil-
trates, were analyzed for total iron, iron forms, sulfate content and pH.
Slurry samples were analyzed for solids content and slurry solids were water
washed, extracted with toluene and dried. Dry slurry solids received short
proximate, sulfur forms and ash iron analyses. Residues from toluene extracts
were analyzed for total sulfur content.
Vent gas from R-l was continuously sampled by means of a slipstream
taken immediately downstream from V-l. The sample slipstream was dried
and subsequently analyzed for oxygen content by a Taylor oxygen analyzer.
Gas analysis results were continuously displayed at the plant control panel.
4.2 PROCESS EQUIPMENT
Detailed mechanical design specifications, drawings, and vendor prints
for all RTU equipment have previously been submitted to the EPA (TRW-Procon
Design Drawings, Vendor Prints, and Design Specifications). These documents
contain plant plot plans, mechanical flow diagrams, civil and structural
drawings, mechanical, electrical and instrument drawings, piping diagrams
and vendor data. Detailed specifications for all equipment, piping, elec-
trical, and instrumentation installations are also included. Descriptions
of services provided on a regular basis in support of RTU operation as well
as stepwise procedures have been documented. This extensive documentation
is not replicated in this report but is mentioned for reference only. How-
ever, specifications of principal coal processing equipment utilized during
RTU operation are summarized in this section to provide definition of the
RTU processing capabilities. The ensuing sections will also describe the
various support facilities utilized during RTU operation.
30
-------
4.2.1 Coal Transport Bins
Seventy-five coal transport bins were fabricated of carbon steel. Bin
dimensions are 42" x 48" x 96" (H) which provide approximately 112 cubic feet
of storage space. Coal bins are designed to receive coal ranging from
3/8-inch top-size to 100 mesh top-size. Each bin is equipped with a 12-inch
diameter opening located at the top for coal loading and a 34-1/4-inch x
14-11/16-inch (H) trap door on the side of the bin near the bottom for coal
discharge. Coal bins are airtight and equipped with gas inlet and outlet
fittings to enable slight pressurization with nitrogen.
4.2.2 Bin Tilter
The bin tilter A-l is a steel frame capable of handling a fully loaded
bin and equipped with a pneumatically operated tilting mechanism. A-l is
equipped with leg clamps which automatically seal the bin to the feed hopper
gasket and prevent dust emission to the atmosphere during the coal discharge
operation. The A-l vibrator maintains coal solids flow once the bin discharge
door is opened.
4.2.3 Coal Storage Tank
Coal is stored in a vertical carbon steel tank T-l which is a 6-foot diam-
eter cylindrical vessel with a conical bottom to facilitate coal discharge.
T-l has a total volume of approximately 180 cubic feet. The T-l vessel receives
coal from the coal feed chute through a 1-foot square opening in its top. T-l
is equipped with a live bin bottom capable of maintaining variable solids dis-
charge flow rates without bridging or packing. The T-l discharge unit A-l
is equipped with a two-point cycle timer to regulate the discharge oscillator.
A-l is capable of maintaining coal discharge rates from 250 to 1000 pounds per
hour. T-l is maintained under a slight nitrogen pressure (up to 6 inches
water column) to prevent coal oxidation. Maintenance of an adequate quantity
of coal in T-l is ensured by monitoring the coal level with a sonic level
switch which activates a low level alarm.
4.2.4 Weigh Belt
An Autoweigh weigh belt A-3, capable of metering coal flow rates up to
1000 pounds per hour, transfers coal fed from the coal storage tank T-l to
the rotary feed valve A-4. A-3 is designed to transfer ground coal from
31
-------
8 mesh to 100 mesh in top-size. A belt speed monitor and strain-gauge load
cell are used to provide coal feed rate and total coal feed outputs, and to
signal a feedback controller which maintains preset coal feed rates by adjust-
ing the belt speed. Coal feed rate monitoring accuracy is within 0.5% of the
actual coal feed rate. The A-4 feed valve is airtight and acts as a moisture
barrier between A-3 and the first stage of the slurry mixing unit T-2. Both
A-3 and A-4 are operated under a nitrogen atmosphere. A full stream butterfly
valve coal diverter is located between A-3 and A-4 which enables acquisition
of feed coal samples.
4.2.5 Tank Farm
Leach solution and process liquid wastes are stored in a tank farm con-
sisting of three glass-fiber reinforced polyester tanks T-7 through T-9.
Each tank is 12 feet in diameter, 20 feet high and has a storage capacity of
16,000 gallons. These tanks are designed for atmospheric pressure service at
up to 200°F. Stored solution volumes are measured by Varec liquid level gauges
and are accurate to approximately 5 gallons. Tank farm plumbing is designed
to enable solution transfer among the storage tanks.
4.2.6 Leach Solution/Slurry Pumps
Sixteen process pumps are included in the RTU for either liquid (water
or reagent) or slurry service. These pumps are largely centrifugal and pro-
gressive cavity-type pumps although diaphragm and regenerative pumps are also
used. Pumps used in acid reagent service were constructed of 316L S.S. while
those in water service were fabricated of cast iron or other less noble mate-
rials. Specific data for each pump are listed in Table 1.
Centrifugal pumps were utilized for low head pressure and moderate or
high flow rate applications such as R-l slurry recirculation and liquid waste
disposal. Nine centrifugal pumps were utilized in the RTU with eight being .
used for slurry/reagent pumping service (P-2 through P-6, P-9, P-12 and P-14)
and one being used in very dilute reagent service (P-10). Progressive cavity
pumps were utilized for pumping slurries at low flow rates against large head
pressures as encountered in feeding slurry to the pressurized primary reactor
through P-l. To provide possible replacement units for P-l, two progressive
cavity pumps were also used in less severe applications (P-7 and P-13). Two
32
-------
TABLE 1. RTU PUMP DATA
co
Pump
Identification
P-l
P-2 through P-6
P-7
P-8
P-9 and P-10
P-ll
P-12
P-13
P-14
P-15
P-16
Description
Slurry feed pump
Slurry recirculation
R-2 discharge pump
Filter wash pump
S-l filtrate pumps
Reagent recirculation pump
Reagent circulation pump
Reagent feed pump
Waste disposal pump
Cooling water pump
Seal flush water pump
Manufacturer
Moyno
Dean Brothers
Moyno
Warren Rupp
Duriron
Warren Rupp
Dean Brothers
Moyno
Dean Brothers
Paco
Aurora
Type
Progressive cavity
Centrifugal
Progressive cavity
Diaphragm
Centrifugal
Diaphragm
Centrifugal
Progressive cavity
Centrifugal
Turbine
Regenerative turbine
Design
Capacity,
gpm
1.1-4.6
25
1.1-4.6
10
20
5
20
0-8.4
200
100
12
AP,
psi
140
33
140
40
40
18
18
140
32
250
150
Driver,
H.P.
3
2
3
Air drive
5
Air drive
2
1.5
7.5
15
5
Material of
Construction
316 SS
316 SS
316 SS
316 SS
Durcomet 100
316 SS
316 SS
316 SS
316 SS
Cast iron
C.I. /Bronze
-------
diaphragm pumps were utilized in the RTU for low head pressure, low flow rate
applications (P-8 and P-ll). One regenerative turbine pump (P-16) was utilized
for a seal flush water pump due to its high head pressure, low flow rate char-
acteristics and one turbine pump (P-15) was used for cooling water circulation.
4.2.7 Heat Exchangers
Two process streams are heated by jacketed pipe heat exchangers: 1) feed
reagent to T-3 is heated in E-2 and 2) filter belt wash water is heated in
E-3. The heating fluid for both exchangers is 150 pound steam and each is
designed for a maximum temperature and pressure of 400 F and 200 psi, respec-
tively. Exchanger E-2 consists of two heat exchangers in series, each of which
is 30 inches long having a 3/4-inch Sch. 40 pipe of 316 S.S. and a 1.5-inch
Sch. 80 jacket of carbon steel. Exchanger E-3 consists of three heat exchangers
in series, each of which is 45 inches long having a 3/4-inch Sch. 80 pipe and
a 1.5-inch Sch. 80 jacket. E-3 is all carbon steel construction since it is
not subjected to a corrosive environment.
The primary reactor R-l is equipped with jacketed pipe heat exchanger
E-l to provide for removal of excess heat of reaction. E-l consists of four
heat exchangers in series which may be divided among the R-l slurry recircula-
tion loops. Each segment of E-l is 3.5 feet in length having 1.25-inch
Sch. 40 pipe of 316 S.S. and 2-inch standard weight carbon steel jacketing.
E-l is cooled with site cooling water and is designed for pipe temperatures
and pressures up to 300°F and 150 psi.
A shell and tube heat exchanger E-4 is used to cool flush water for pump
and mixer seals. This unit is cooled with site cooling water. Shell material
is bronze and tubing is admiralty brass. E-4 is designed for temperatures
and pressures of up to 300°F and 150 psi.
4.2.8 Slurry Mixing Tank
Coal-reagent mixing and preheating takes place in the mixing tank T-2.
The mixer is a three-stage, 316L S.S. horizontal vessel having a diameter of
2.5 feet and length of 7.5 feet. The stages are separated by adjustable weirs
with which the stage volumes and, therefore, slurry residence times may be
varied. Weir heights may be adjusted from approximately 10 to 23 inches.
Each stage of T-2 is equipped with a variable speed 3/4 horsepower Lightnin
34
-------
agitator capable of up to 350 rpm. Live steam for slurry heating is injected
into each stage of T-2 through nozzles in the vessel wall. All stages of T-2
vent through a common foam scrubber T-3 and through a vent gas scrubber T-4.
Each stage of the mixer is equipped with a slurry sampling port, drain
port and a thermocouple probe. The third stage of T-2 contains an Air Products
bubble tube-type level indicator and an auxiliary stilling well.
4.2.9 Reactors
Three reactors are utilized in the RTU: (1) the primary fine coal reactor
R-l; (2) the secondary fine coal reactor R-2; and (3) the coarse coal leacher
T-6. RTU reactors are constructed of 316L S.S. and are rated for processing
temperatures of up to 300°F. All RTU reactors are insulated with 1.5 inches
of urethane foam.
The primary fine coal reactor is a horizontal pressure vessel with a
diameter of 38 inches and a length of 14.75 feet. R-l is designed for pres-
sures up to 150 psig and for full vacuum (to accommodate hot lockup and sub-
sequent depressurization through cooling). Composed of five stages, R-l is
partitioned by stationary weirs which are 28.5 inches in height. Slurry feed
lines are plumbed to allow bypass of the first two R-l stages and provide for
three stage processing. Each stage of R-l is agitated by a variable speed
1.5 horsepower Lightnin mixer and a slurry recirculation system. Four
baffles are spaced around each stage to promote efficient slurry mixing.
Slurry level in the fifth stage is measured by an Ashcroft buoyant type level
gauge. R-l slurry is discharged from the fifth stage to flash drum T-5 either
by flashing through a 1-inch ball valve KV-241 or by pumping slurry through
the fifth-stage recirculation pump P-6.
Reactor heating and slurry oxygenation are performed by injecting steam
and oxygen into the slurry recirculation loop of each R-l reaction stage.
Reactor temperatures are monitored by thermocouple probes installed in each
R-l stage. R-l stages vent to a common foam knock-out drum V-l and vent gas
scrubber T-4. Reactor pressure is maintained by a Masoneilan control valve
and is monitored by a transducer PT-173 located at the reactor vent.
• R-l "is equipped with 1-inch sampling ports in the vessel wall and in the
slurry recirculation loops. Recirculation loop sampling ports enable full
stream bypass through a sampling unit which is ultimately isolated and drained.
35
-------
The secondary fine coal reactor is a vertically oriented cylindrical
vessel with a conical bottom. R-2 has a diameter of 3.5 feet and has a height
of 4.1 feet (T/T); the corresponding volume of R-2 is 365 gallons. R-2 is
designed to maintain 14 psig or full vacuum. Reactor pressures are monitored
by an Ashcroft pressure gauge PI-98. R-2 agitation is provided by a variable
speed 1-1/2 horsepower Lightnin mixer. Three baffles are distributed within
R-2 to prevent channeled flow and promote efficient mixing.
Slurry enters R-2 through a 2-inch nozzle at the reactor top and dis-
charges through a 1.5-inch nozzle at the reactor bottom. Slurry levels in
R-2 are monitored with an Air Products bubble tube level sensor LT-95. Slurry
temperatures are monitored with a thermocouple probe TE-99.
The coarse coal leach solution depleter T-6 is a vertically oriented
cylindrical vessel with a conical bottom. The vessel diameter is 5 feet and
the straight side height is 5 feet. T-6 is essentially an atmospheric pres-
sure vessel. Coal enters T-6 through a 12-inch square chute at the vessel top
and is discharged through a 12-inch knife valve at the bottom. Reagent enters
and discharges from T-6 through 3/4-inch ports at the bottom and top of the
vessel, respectively. T-6 is heated by live steam injected through a 1-inch
port at the vessel bottom. Reaction temperatures are monitored with a thermo-
couple probe in the vessel wall. Reagent may be drained from T-6 through a
3/4-inch drain port equipped with a 2-inch duplex strainer for retaining coal.
Reactor pressure is monitored with an Ashcroft mechanical pressure gauge PI-260.
4.2.10 Flash Drum
The slurry flash drum T-5 is a vertical vessel with a conical bottom.
T-5 has a 3-foot diameter and a height of 3.7 feet (T/T). T-5 is designed for
up to 14 psig pressure and full vacuum. Flash slurry enters T-5 through a
2-inch nozzle located 1.7 feet below the upper straight side end. A porous
demister pad is situated immediately above the flash slurry inlet to minimize
•;
slurry entrainment. Slurry level is monitored in T-5 and an alarm condition
results if slurry levels rise appreciably above the conical vessel bottom.
4.2.11 Scrubbers
Slurry foam control and acid mist elimination are obtained in the RTU
with three scrubbers: (1) the foam scrubber T-3; (2) the knock-out drum V-4
36
-------
and (3) the vent gas scrubber T-4. T-3 is a 316L S.S. vessel of 1-foot diam-
eter and 4 feet in height. A single bubble cap tray is used to break down
foam generated in the mixer T-2. Designed for gas flow rates of 30 SCFM and
liquid flow rates of 2.5 gpm, T-3 may be operated under full vacuum or at
pressures up to 14 psig with temperatures up to 300°F. V-l is a 316L S.S.
vessel having a 1-foot diameter and a height of 6.5 feet. Two bubble cap
trays are used to break down foam and remove acid mist from the R-l effluent
gas. V-l has the same design pressure and temperature requirements as R-l,
namely, full vacuum to 150 psig at temperatures up to 300°F. T-4 is a low
pressure FRP vessel having a diameter of 1.5 feet and a height of 4.5 feet.
Vent gases flow upward through a 3.5 cubic foot volume of 316 S.S. flex ring
packing which is sprayed with cooling water from an overhead nozzle. Operating
temperatures in T-4 may range up to 215°F.
4.2.12 Filter and Vacuum System
Solid-liquid separation is performed with a belt filter supplied by
Ametec. The filter belt consists of a polypropylene mesh cloth 1 foot in
width which is supported and conveyed by a channeled plastic belt. The filter
is driven by a two horsepower variable speed belt drive. Filter belt speeds
may be varied from 1 to 10 feet per minute producing coal cake thicknesses of
up to 3 inches. Suction is applied to the belt by means of two vacuum pans
over which the belt must pass sequentially. The first vacuum pan is for col-
lection of concentrated reagent while the second collects wash water from the
filter cake rinse. Feed slurry is distributed on the belt by a stationary
spreader. After initial dewatering, the filter cake is steamed and sprayed
with hot wash water by overhead nozzles. The filtration unit is completely
enclosed in a reinforced fiberglass hood to prevent steam losses. Sampling
ports are built into the hood to permit processed coal sampling before and
after the washing operation. Filter belt washing nozzles are located at the
belt return immediately beyond the processed coal discharge chute to remove
remaining coal particles from the belt and prevent blinding.
Vacuum is supplied to the filter vacuum pans through the respective filtrate
receiving vessel (V-2) and wash water receiving vessel (V-3). The filtrate
receiver is a 316L S.S. vessel with a 1.2-foot diameter and a height of 5 feet.
The wash water receiver is a 316L S.S. vessel with a diameter of 2 feet and a
height of 6 feet. Both receiving vessel^ are under vacuum provided by a
37
-------
50 horsepower 630 ACFM Nash pump K-l which is capable of maintaining vacuum
of 22 inches Hg. Each receiving vessel is equipped with a demister pad to
prevent liquid entrainment into K-l. The contents of V-2 are recycled to
reagent storage tanks T-7 or T-8 while the V-3 underflow is pumped to waste
disposal.
4.2.13 02/N2/Air Supply Systems
Gaseous oxygen is supplied to the primary reactor R-l from a liquid oxy-
gen (LOX) storage tank which is remote from the RTU coal processing facility.
The LOX is vaporized and supplied to the RTU at 200 psig. Oxygen feed rates
are monitored by a Ramapo target flow meter to ±0.5 percent. Process nitrogen
is stored in liquefied form and vaporized for plant use. Nitrogen flow moni-
toring equipment is identical to that used for oxygen monitoring. Nitrogen is
supplied to the plant at 200 psig. Air is supplied to the facility at 125 psig
by a 100 SCFM Rand air compressor and is used primarily for equipment drive and
valve actuation purposes.
4.2.14 Steam Supply System
Process steam is supplied by a Clayton steam generator with the capability
of generating 1000 pounds of 150 psig steam per hour. The steam generator is
equipped with self-contained water softener and boiler treatment subsystems.
The generator is powered by a propane burner also supplied by Clayton Manu-
facturing Co. Steam feed rates to the RTU are monitored by Brooks rotameters.
4.2.15 Process Water System
Two water sources were utilized during RTU operation. The first, an
existing well water supply, is used to supply the steam generator and to main-
tain the 100-gallon seal flush water supply tank. The second source of proc-
ess water is the RTU catch basin which is a large hypalon lined reservoir hav-
ing a width of 30 feet, a length of 80 feet and a depth of 5 feet. Water from
the catch basin is used for cooling water in heat exchangers, as scrubber
water in T-4 and for filter belt wash water.
4.2.16 Instrumentation
Detailed instrumentation listings and specifications are presented in the
"TRW-Procon Design Specification" manual. This document contains instrumenta-
tion tag numbers, manufacturers, materials of construction, operating ranges
38
-------
and, in many cases, unit schematics. Locations of the various instruments in
the RTU are fully detailed in the "TRW-Procon Design Drawings and Vendor
Prints" manual, drawing numbers X418455 through X418461. Both TRW-Procon
documents have been previously transmitted to EPA.
Principal types of instrumentation located on the RTU include pressure
gauges, temperature indicators, level gauges, and flow meters. Ashcroft
bellows type pressure gauges were used to monitor nitrogen pressures to coal
bin storage and instrument purge streams (i.e., PI-184 and 167). Ashcroft
bourdon gauges were utilized in monitoring pump slurry and liquid discharge
pressures (i.e., PI-73 through PI-77). Remote monitoring of feed gas pres-
sures and reactor pressures was performed with Viatran transducers. Trans-
ducers were also used in conjunction with digital display equipment to provide
local data readout.
Three principal types of temperature measurement equipment were utilized
during RTU operation: (1) thermocouples, (2) dial thermometers and (3) bulb
thermometers. Shielded type K (chromel-alumel) thermocouples provided by
Heat Technology were used to monitor mixer and reactor slurry temperatures
(TE-19, 20, 21 and TE-52 through 56). Bi-metal dial thermometers by Weston/
Disco were used for local display purposes such as monitoring wash water
temperatures and reagent storage temperatures. Taylor bulb thermometers in
thermowells were utilized in conjunction with temperature controllers such as
that controlling steam feed rates to heat exchanger E-2 (TIC-32).
A variety of level gauges were employed to monitor liquid, slurry and
coal depths in the RTU tanks and reactors. Foam levels in the mixer could be
determined by Jacoby-Tarbox sight glasses installed in the T-2 vent lines. A
Masoneilan buoyant level gauge was utilized to monitor slurry levels in the
fifth stage of the pressurized primary reactor while differential (bubble tube)
level indicators were used in the mixer and secondary reactor. Varec float-
type indicators were utilized for monitoring reagent and liquid waste levels
in tank farm units. Sonic level probes by National Sonics were utilized to
activate high and low level alarms in critical process vessels such as the
coal feed tank and the primary reactor.
Steam flow rates were monitored with Brooks rotameters. Water flow rates
were monitored with both Brooks and Wallace and Tiernan rotameters.
39
-------
Additionally, Wallace and Tiernan rotameters were used in some gas flow appli-
cations. Fischer and Porter magnetic flow meters were used for all slurry flow
rate measurements (FE-29, 83-87) as well as for reagent feed rate measurements
(FE-31). Oxygen feed and vent rates were monitored with Ramapo target flow
meters FE-44 and FE-61.
Outputs of all primary instrumentation and alarm instrumentation were
translated into engineering units and recorded by a Doric data processor.
The data processor is of modular construction with a solid-state channel
scanner, digitizer, microprocessor-based control circuits, channel display and
printer. A total of 99 channels for data storage are available and one addi-
tional channel serves a self-check function. The system contains a real-time
digital clock providing time data printout in days, hours, minutes and seconds.
Channel scan rates may be varied from 2 to 20 channels per second. The sys-
tem is capable of continuous scanning and logging or periodic logging at inter-
vals of 1, 5, 10, 15, 30 and 60 minutes. Alternately, the system may scan
continuously but output only those channels which are beyond preset tolerance
limits. When scanning at 2 channels per second, an integral digital strip
printer may be utilized to provide hard copy printout; all data are recorded
on seven-track magnetic tape, regardless of scan speed or frequency.
4.2.17 Control Laboratory
Sampling and analyses of all priocess solid and liquid streams were per-
formed by personnel from the control, laboratory at the Capistrano test site.
The analytical capabilities of the 760 square foot control laboratory include
coal moisture analyses, total sulfur analyses, sulfur forms analyses (pyritic
and sulfate sulfur), slurry solids determination, reagent ferrous iron and
total iron analyses, reagent sulfate, reagent pH and coal size distribution
analyses. Selected coal samples were submitted to Warner Laboratories,
Cresson, Pennsylvania, for coal short proximate analyses (moisture, ash, heat
content and total sulfur analyses). Toluene extract residues were also sub-
mitted to Warner for total sulfur analyses. All analyses are performed in
accordance with standard ASTM methods.
To ensure uniform sampling, sample preparation and sample analyses,
standard procedures were prepared and documented. The procedural format is
40
-------
essentially the same for sampling and analyses and includes the following
elements:
• Procedural scope describing the operation or operations to
be defined.
• Procedural conditions specifying personnel and safety require-
ments and listing all relevant reference documentation.
• Test requirements including equipment requirements and appa-
ratus configuration specifications.
• Detailed analytical and sampling procedures and operational
flow diagrams.
• Data recording format.
A tabulation of the standard analytical and sampling procedures is presented
in Volume II, Appendix B.
4.2.18 Capistrano Test Site Support Facilities
In addition to the control laboratory which was utilized on a continuous
basis, the Capistrano test site also provided a variety of services on an
as-needed basis in support of the RTU. The CIS weld, machine, and valve shops
were all utilized periodically for quick in situ and remote repairs, uncoded
fabrications and equipment maintenance. All RTU instrumentation was cali-
brated upon receipt by TRW metrology which also provided trouble shooting
assistance during plant shakedown and operation. CTS procurement, shipping
and receiving, and quality assurance assisted RTU operation by locating
replacement parts as needed, identifying specialty fabrication shops and pro-
curing all required chemicals. Site power, water and sanitation facilities
utilized during RTU operation were maintained by the CTS maintenance
department.
41
-------
5. REACTOR TEST UNIT OPERATION
On May 1, 1977, RTU shakedown activities were initiated. The overall
shakedown effort was comprised of a system readiness and safety review, oper-
ator training, operating procedure verification, laboratory procedure verifi-
cation, data acquisition system verification, RTU operability evaluation, and
Operational Test Plan revision. By late September, 1977, all objectives of
the RTU shakedown were met and the RTU was deemed ready to start the planned
operational phase of the project.
Operational testing of the RTU was started on October 1, 1977, and was
terminated on January 26, 1978. During that 4-month period, the unit processed
49,700 Ibs of coal over approximately 250 hours of operation in increments up
to 32 hours each. Sufficient data was acquired during operational periods to
allow verification of process chemistry (discussed in Section 6), to paramet-
rically evaluate process variables (Section 6), to determine equipment char-
acteristics, and to generate coal quantities sufficient for vendor testing
(Section 9).
i
The remainder of Section 5 presents a summary of achievements and obser-
vations relating to RTU shakedown and operation. Sections 5.1 and 5.2 discuss
shakedown and operation activities respectively. Additionally, Section 5.2
presents a detailed description of typical RTU start-up, operations, and shut-
down sequences. Equipment operating experience is also delineated. Conclusions
which were drawn from the first four months of RTU operational experience are
presented in Section 5.3.
5.1 RTU SHAKEDOWN ACTIVITIES
The primary objectives of the RTU shakedown activities were to verify unit
operability, train and familiarize personnel with process operations, facilities,
and procedures, and acquire preliminary system data. Necessary modifications
to the operational, smapling, analysis and calibration procedures and to the
reactor system itself were made during the shakedown period.
42
-------
5.1.1 RTU Safety Review and Training
A thorough pre-start-up safety review of the dual reactor test system was
initiated following completion of plant construction. The Safety Review Team
(SRT) was comprised of TRW personnel representing the TRW Health and Safety,
Structural Design and Fabrication, and Chemical Engineering Departments,
Capistrano Test Site management and other upper management representatives from
within TRW operations. Most SRT members were individuals not previously inti-
mately involved in the RTU design and operational decisions, thus limiting
possible bias. The review was conducted through study of prepared materials
(memos, reports, manuals, etc.) and visual site inspections.
The SRT reviewed all start-up, shutdown and operational procedures and
evaluated the unit operations from chemical (corrosion, flammability, acidity,
etc.) and toxicological standpoints. On-site safety measures such as fire
extinguishers, breathing equipment, protective clothing, etc., were reviewed
for adequacy. Additionally, all emergency procedures, process safety control,
emergency bypass controls ar.d monitoring equipment (both for personnel and
environmental safety) were evaluated. At the culmination of the safety review,
the team prepared a report detailing all findings and recommendations. The
safety review report was evaluated by RTU project management and all necessary
changes to plant operation, facilities, and procedures were made.
In parallel with the safety review, personnel training was conducted.
This pre-start-up training effort had as its primary goal the safe and orderly
operation of the RTU during start-up, run and shutdown sequences. To facili-
tate the achievement of that goal, all operating, laboratory and engineering
personnel were familiarized with the appropriate procedures, both normal and
emergency. The training consisted of both operating procedure study and
on-site experience with the processing equipment and area to familiarize per-
sonnel with types and locations of processing equipment, valves, service facil-
ities, instrumentation and control hardware, effluent monitoring equipment and
emergency control points.
5.1.2 Shakedown Operations
(9)
A detailed shakedown plan was submitted to EPA in June 1977, In
general, that document described in detail the planned objectives, approach,
43
-------
sequences of operations, and expected results of the plant shakedown effort.
The broad objectives of the RTU equipment shakedown period were as follows:
• Demonstration of equipment operability using coal and iron
sulfate leach solution.
t Demonstration of the operability of the RTU at steady-state
condition.
• Demonstration that controls, indicators and data acquisition
systems were functional and suitable for repeated start-up,
operation and shutdown sequences.
t Verification that sampling and analysis procedures were
reproducible, safe and effective for the locations proposed,
and that acceptable time delays between sampling and final
result determinations were obtained,
• Estimation of the reproducibility of a single set of process
conditions and the associated product characteristics.
t Standardization of settings of instruments, control devices
and equipment to obtain desired ranges of pressure, tempera-
ture, volume, flowrate, weight, agitation rate, transfer
rate, etc.
t Determination of the practical operational equipment limits
for preparation and feeding of coal slurry, slurry flow and
mixing and filtration operations.
• Demonstration that all RTU mechanical components and by-passes
operated without unscheduled interruption or mechanical fail-
ure over the proposed operational ranges of flowrate, pressure,
temperature and chemical environment.
The general approach to the shakedown of the RTU consisted of the sequen-
tial imposition of increasing numbers of stress-producing environments on the
unit operations of the process. In this manner, identification and correction
of any problems were expected. Also the gradual approach enabled personnel
development of familiarity with process controls, checkout procedures, etc.,
in an easily assimilated manner. Each set of stress-producing environments
was studied under continuous steady-state operation of the RTU over a period
of time Chours), followed by inspection of the equipment and an assessment of
its performance. The testing sequence which was followed during shakedown and
the resultant accomplishments are summarized in Table 2. The table indicates
the reactant media (i.e., coal type and size, liquid phase component, and gas-
eous phase component) and the primary activities carried out during each test run.
44
-------
TABLE 2. RTU SHAKEDOWN SEQUENCE
Run Reactant Media
No. Coal Type Liquid Phase Gas Phase
Major
Accomplishments
001
None
Process Water
Nitrogen System operated warm. System dynamically leak checked
under nitrogen gas pressure. Foreign material flushed
from system components.
002
Coal 1
14 mesh x 0
100 mesh x 0
Distilled water
(no added iron
sulfate)
Nitrogen System operated hot. Initial coal feed operations and
and oxygen sampling procedures evaluated at two feed coal sizes.
Effluent .gas analysis equipment and procedures checked.
Initial data acquisition system evaluated. Pumping
capabilities verified.
003/004 Coal 1 Water (no added
14 mesh x 0 iron sulfate)
Nitrogen Coal feed system operation and sampling verified. Reactor
operated at maximum temperatures and gas pressures.
Filter settings evaluated and filter cake samples obtained.
Two shift operation attained. Initial mass balances
obtained. Reactor wall and pump-around loop samples
evaluated. Process sampling procedures and control lab
analysis procedures evaluated.
005 Coal 1 Water (no added
14 mesh 1 0 iron sulfate)
Oxygen Obtained baseline (no added iron salt) reaction data.
Verified operability of instrumentation needed for mass
balance. Evaluated mixer and reactor sampling (wall vs.
loop) results and procedures. Verified coal analysis
reproducibility.
006
Coal 1
14 mesh x 0
Leach Solution
(water plus added
sulfuric acid and
iron sulfate)
Oxygen Verified proper mixer settings to effect efficient coal/
leach solution wetting without foaming. Verified operat-
ing, sampling, and analysis procedures using leach solution.
Obtained initial coal desulfurization data with leach
solution. Regeneration rate evaluated. Filter operation
evaluated with leach solution.
Test runs 003 and 004 were accomplished during one testing sequence.
-------
As was expected, during shakedown operations, some equipment related dif-
ficulties were encountered. Those which resulted in significant equipment
modifications and/or lengthy duration rework are summarized below,
• Turbine meters were initially installed in liquid flow moni-
toring service. During shakedown operations, most liquid
circulating lines tended to build up in particulate loading
(up to a few hundredths of a percent solids) due to the vari-
ous recycle loops and the closed loops water system. The
turbine meters were found to be inoperable when subjected to
any degree of particulate loading. They were, therefore,
replaced by magnetic flow meters which were capable of han-
dling virtually any stream, from particulate free liquids to
concentrated slurries.
• The Autoweigh weigh belt would not operate to the required
±0.5 percent accuracy as initially configured. This degree
of accuracy was required in order to achieve a reasonably
accurate mass balance around the system. Both hardware and
electronic difficulties were encountered during the early
shakedown period. Weigh belt internals and electronics were
exchanged several times (by the manufacturer) for redesigned/
modified replacements. Following supplier modifications, the
unit was found to operate within the specified tolerance
levels.
• Filter modifications were required after it was determined
that back flooding of the belt (slurry back-up and overflow
from the rear of the unit) occurred when trying to achieve
maximum cake thicknesses or when belt slippage and momentary
belt slowing occurred. This problem was solved by modifying
the slurry feed distributor and by adding a dam at the rear
of the belt.
• During the initial shakedown test, it was determined that the
reactor feed pump P-l (and its potential spare P-7) would not
pump cold solution at the required pressure. This was found
to be the result of internal pump rotor to stator tolerance
levels. The close tolerance required to enable the pump to
achieve high delivery pressures at the low flow rates and high
temperatures required for reactor operations grew to an inoper-
able tolerance level when pump parts ran cool. This occur-
rence was the direct result of thermal expansion/contraction
effects. The problem of cold feed pump inoperability was cir-
cumvented by revising procedures to include initial heat-up of
the unit. This was accomplished via closed loop recirculation
of heated (by direct steam injection) mix tank T-2 contents
through P-l and back to the mix tank. P-l was usable in recir-
culating service because of the low head required to pump from
P-l back to T-2 as opposed to the high head required to pump
against reactor R-l pressures.
46
-------
t Several hours into the initial leach solution run (Run 006)
a major leach solution leak (acid spray) occurred which '
forced the system into an emergency shut-down. The plant
was shut down and personnel evacuated the test stand within
minutes of the incident. There were no injuries or equip-
ment damage as the result of the rapid emergency procedure.
The cause of the acid spary was found to be an improperly
installed pipe flange on one of the leach solution supply
lines. The flange was determined to be of carbon steel con-
struction instead of the specified 316 Stainless Steel
required for acid service. As a result of the incident, a
thorough reinspection of the RTU was carried out to assure
that all piping, flanges, and valves had been installed as
specified. No other improperly installed equipment was
discovered.
• Following evaluation of initial leach solution regeneration
rate data, obtained during the first leach solution run, it
was determined that insufficient solution regeneration in
the reactor was occurring. This occurrence was thought to
be the result of low turbulence experienced in the reactor
pump-around loops where regenerant oxygen is blended with
circulating slurry. To enhance regeneration and thereby
increase coal desulfurization potential of the system, the
oxygen blending portion of the recirculation loops were mod-
ified. The original 1-1/4 inch Schedule 80 pipe was replaced
with 3/4-inch Schedule 40 pipe to increase flow velocities
and therefore increase turbulence. It was determined in
subsequent runs that reducing the pipe diameter did indeed
increase regeneration rates to acceptable levels.
In summary, the RTU was operating successfully during the final phase of
shakedown operations and most of the planned objectives of the overall effort
were met. The various subsystems of the RTU were found to operate in an inte-
grated fashion as anticipated. For instance, the three T-2 mixer stages were
found to be maintainable at uniform temperature or with increasing temperature
for each succeeding stage within the planned slurry flow rate ranges. Plant
startup was accomplished over an approximate 3-4 hour period, which was deemed
adequate to meet the objectives of the operational test plan. Reactor tempera-
tures were demonstrated controllable up to 275°F using the available steam
source, and a fully burdened system, including heat exchangers, belt filter,
mixer and reactor. The reactor pressurization and control system performed as
expected. The nitrogen, air, and water supply systems were of adequate capacity
and controllability for all purposes intended. The coal and slurry transfer
systems functioned satisfactorily, as desired, following equipment modifica-
tions. The belt filter performed as intended up to a maximum slurry feed rate
47
-------
of 1500 pounds per hour. Beyond that limit, belt flooding and incomplete fil-
tration was found to occur. An acceptable leach solution regeneration rate was
attained following modifications.
The objectives of the concurrent support activities were also met during
shakedown activities. All data generated during shakedown was evaluated with
acquisition hardware and software being modified and/or generated where appro-
priate. Suitable data recording formats were developed which allowed for fast
and accurate data retrieval. The data itself was evaluated primarily to deter-
mine baseline process chemistry. In the control laboratory, procedures were
verified and techniques developed which proved to yield accurate and reproduc-
ible laboratory analysis data. At the conclusion of shakedown, a revised
Operational Test Plan^10^ was generated and submitted to EPA. The plan was
revised in the context of the newly acquired data on a working knowledge of the
processing equipment operating limitations and capabilities. The submission
of that document signaled the end of RTU shakedown operations.
5.2 RTU OPERATION
The RTU was brought into operational status immediately following shake-
down activities. During the course of operations, it was demonstrated that the
RTU could be operated continuously to test Meyers Process units for coal-leach
solution mixing, simultaneous coal depyritization and leach solution rege'nera-
tion, filtration of leach solution from treated coal, and product coal water
washing. No difficulties were encountered in changing process variables such
as coal feed rates, leach solution feed rates, steam flow rates, reactor pres-
sure, oxygen flow rates, agitation and/or pump-around rates, or filter condi-
tions during RTU operations. Sixteen distinct sets of operational variables
were systematically studied while treating 49,700 pounds of coal over a total
operating period of 254 hours. Many of the variable changes were made during
continuous elongated run sequences, up to 32 hours in duration, demonstrating
the overall flexibility and controllability of the RTU.
5.2.1 Qperating Procedures
Prior to initiation of the planned experimental testing effort, the RTU
operational procedures were reviewed and revised in light of the previously
acquired shakedown experience. Those procedures are tabulated by name and >
48
-------
identification number in Volume U, Appendix B of this report, The following
paragraphs describe the nominal start-up, operating and shut-down sequences
followed during operation of the RTU.
Prior to RTU start-up, all auxiliary supply systems are activated. It is
the auxiliary supply systems which provide nitrogen, oxygen, utility/instrument
air, leach solution and process/cooling water to the RTU. Nitrogen is used
primarily as a purge gas in the coal feed and reaction systems. Oxygen is
employed for reagent regeneration in the primary reactor and, with minor piping
modifications, could be utilized in the secondary reactor also. Air is used to
drive equipment (i.e., pumps P-8 and P-ll) and to actuate various plant instru-
mentation. Start-up of auxiliary supply systems consists primarily of verifying
all associated valve and regulator settings, verifying the adequacy of material
supplies, and subsequent supply initiation. Once these supporting systems have
been activated, RTU start-up procedures are initiated.
RTU start-up procedures may be summarized by processing unit. The first
step in start-up of all units is verification of associated instrumentation.
Prior to start-up of the coal feed system, the coal level in storage tank T-l
is verified to be sufficient for test initiation or, if needed, additional coal
is transferred from tote bins to T-l. The storage tank T-l, weigh belt A-3 and
rotory valve A-4 are purged with nitrogen and then activated. Start-up of the
mixer 1-2 includes verification of weir heights and initiation of mixing with
M-l through M-3. Steam flow to T-2 is initiated and slurry pump P-l is acti-
vated in a recirculation mode to stage three of T-2; slurry recirculation is
maintained until the P-l effluent temperature exceeds 200°F. Start-up of the
primary reactor R-l requires verification of seal flush water flow, initiation
of slurry recirculation through pumps P-2 through P-6, and activation and
regulation of the R-l pressure relief valve PIC-43. Mixers M-4 through M-8 are
turned on to augment agitation provided by the slurry recirculation loops.
Steam and gas (either nitrogen, oxygen or a mixture thereof) flows to R-l are
initiated after verification of PIC-43. Activation of the belt filter S-l is
preceded by verification of all filter valving and activation of the vacuum
pump K-l and reagent filtrate and wash filtrate pumps P-9 and P-10. S-l belt
speeds are adjusted to a nominal setting and coal cake steaming and wash streams
are activated. Processed coal catch bins are positioned prior to initiation of
filtration,
49
-------
Upon pressurization of R-l and heat up of P-l, slurry flow through the RTU
may be initiated. The coal discharge vibrator A-2 is activated and coal is fed
to A-3 and, subsequently, to 1-2, Reagent feed pump P-13 is turned on providing
hot reagent to T-2 through V-l and T-3. Reactor effluent slurry flow rates
through P-l and KV-241 (R-l slurry vent valve) are adjusted to maintain desired
slurry levels in T-2 and R-l. Filter belt speeds are adjusted at this time to
yield a properly dewatered processed coal filter cake. Procedures for plant
operation after steady slurry and gas flows have been obtained consist largely
of control adjustments to maintain specified reaction temperatures, pressures,
flow rates and slurry levels. The complete start-up operation requires approx-
imately 4 to 6 hours, depending primarily on the steady state reactor temperature
desired.
Plant shut-down is initiated by terminating all RTU feed and transfer
streams including coal, reagent, slurry, gas and steam. Recirculation
pumps are shut off and the R-l unit is depressurized, Belt filter operation
proceeds until coal is no longer discharged into the catch bins and then S-l
operation is terminated. All vacuum and filtrate pumps are subsequently shut
off. The T-2 and R-l mixers are left operating to prohibit retained slurry
settling. After the RTU has been secured, all slurry transfer pumps (including
recirculation pumps) and piping are drained and flushed with clear water. The
final step in the RTU shut-down sequence is shut-down, venting and securing of
the auxiliary supply systems. The entire shut-down sequence requires approxi-
mately 2 hours to accomplish. The heated slurries contained within the
mixer T-2 and reactor R-l require an additional eight to ten hours to cool to
ambient temperatures.
5.2.2 Equipment Experience
During the 4-month operational test phase of the project, 16 discrete tests
were conducted. The specific test sequences and related processing conditions
over which the RTU was operated are delineated in Section 8. A summary of the
major equipment operational ranges utilized while carrying out the experimental
test sequences is presented in Table 3. In general, the equipment performed
well within the limits studied. There are, of course, several observations
relating to equipment operating experience worth noting.
50
-------
TABLE 3. EQUIPMENT RANGES UTILIZED
Equipment
Sen/fee
Weigh Belt A-3
Mixer T-2
Reactor R-l
Coal Feed
Coal - Leach Solution
Coal - Leach Solution - 0,
Pumps P-l, P-7 Slurry Feed Pumps
P-2/6
P-8
P-9
P-10
P-ll
P-12
P-l 3
P-14
P-l 5
P-16
Mixers M-l/3
M-4/8
Slurry Circulation Pumps
Belt Wash Pump
Filtrate Pump
Wash Pump
T-6 Discharge Pump
Leach Solution Pump
Leach Solution Feed Pump
Waste Disposal Pump
Cooling Water Pump
Seal Water Pump
Mix Tank Mixers
Reactor Mixers
Range Utilized During
Operational Testing
199 - 300 Ibs/hr
156°- 215 °F
1.1 - 1.8 hrs residence time
199°- 273 °F
23 - 80 PSIG
5.1 - 8.2 hrs residence time
58 - 141 SCFH 02
.8 - 4.2 gpm
23 - 80 PSI AP
18 - 25 gpm
0 -- 2 gpm
.6 - 4.6 gpm
1.1 - 2.0 gpm
Not Used
1 - 20 gpm
1.2 - 3.0 gpm
23 - 80 PSI AP
200 gpm
100 gpm
12 gpm
150 - 200 RPM
150 - 200 RPM
51
-------
Coal Feed Equipment. The coal transfer equipment performed well for coal
feed rates up to 300 pounds per hour. A minor modification to the weigh belt
unit was required to accommodate accidental coal spillage and dusting over belt
edges. The modification consisted of the addition of a dust collection chute
and flanged discharge port underneath the belt compartment. This modification
allowed quick periodic removal of coal particulate buildup. A second intermit-
tant situation requiring equipment modification occurred in this section of the
RTU. The dry coal delivery chute to the mixer T-2 occasionally plugged, requir-
ing short duration recirculation mode operation of T-2 while the plug was dis-
sipated either by external chute tamping or chute disassembly and unplugging.
The cause of this occurrence was found to be steam backflow from the first
stage of the mixer (which was normally operated at 185° to 190°F) which resulted
in coal wetting and agglomeration. The situation was circumvented by reducing
nominal first-stage mixer temperatures to 170°F, by increasing nitrogen down-
flow through the chute", and by installing a cooling coil to the outside of the
chute.
Leach Solution Feed Equipment. The leach solution feed equipment, includ-
ing the solution storage tanks T-7 and T-8, heat exchanger E-2, knock-out
drum V-l, associated piping, valves, control and instrumentation, performed
reliably during operation. There was, however, some difficulty experienced with
the leach solution feed pump P-13. Midway through operational testing, several
outer shaft seal failures were experienced in rapid succession. Following
thorough disassembly and inspection of pump internals, it was determined that
the pump contained a small inner connecting shaft of carbon steel construction.
It had been installed in the unit during pump manufacture. The inner shaft had
corroded and resulted in an inner shaft seal failure which allowed the acidic
leach solution to attack the main outer seals. A replacement of inner shaft 316L
was fabricated and installed in P-13. Also, near the end of operations a
P-13 rotor related problem was experienced. The problem was manifested as an
inability of the pump to meet desired throughputs at operating pressures. It
was determined upon inspection of pump internals, that the chrome plated
316L rotor had crazed and that the chrome plating had flaked off, thus result-
ing in too great a clearance between the pump rotor and stator. This oversize
clearance allowed backflow of the leach solution and therefore limited pumping
52
-------
ability. In order to minimize plant down time, the P-13 rotor was replated
with chrome and placed back into service. The rechromed rotor was still in
service at the conclusion of operations. A new 316L stainless steel rotor
(not chrome plated) was fabricated and is currently warehoused as a future
replacement spare.
Mixing Equipment. The coal mixing equipment, consisting of mix tank T-2,
mixers M-l through M-3, foam knock-out unit T-3, and all associated hardware
performed as expected with the exception of the steam inlet lines to T-2.
Early in operations, the lines were found to clog with coal particulate and
leach solution which had backflowed from T-2 during unit shut-down and cooling.
The apparent cause of the backflow was steam collapse in the lines upon cooling
combined with faulty check valves in the inlet lines. The check valves were
replaced and a steam venting sequence was added to the shut-down procedures.
Another observation worth noting is that there were no instances of foaming in
T-2 during the entire operational phase of the project.
Reactor Equipment. For the most part, the equipment associated with the
reactor R-l performed as expected. The slurry circulating pumps P-2 through
P-6, the reactor mixers M-4 through M-8f the knock-out drum T-5, the oxygen
and nitrogen systems, the temperature, pressure and level control systems» and
associated instrumentation performed satisfactorily. The slurry feed pump P-l
(and its spare P-7) experienced the same difficulties as experienced by the
leach solution feed pump P-13 (discussed in preceding paragraphs) at the con-
clusion of operations. In addition to the flaking and seal problems, the
316L pump rotor also experienced corrosion and/or errosion. The unit lost
pumping capability as was the case with P-13. The additional corrosion/erosion
factor was believed to be the result of the abrasive nature of the coal contain-
ing slurry as feed to R-l as opposed to the leach solution liquid handled by
P-13. As in the case of P-13, the reactor feed pump P-l (and P-7) was replated
with chrome to the specified tolerance limit and put back into service to mini-
mize RTU down time. The pump ran satisfactorily for several hours prior to
chrome plate flaking followed by pump failure. Following the second failure
it was determined that the thicker chrome plating (required to make up for
corrosi on/erosion loss) was unable to withstand the thermal expansion experi-
enced during plant start-up. A new Hastelloy C-276 rotor was fabricated and
53
-------
installed in P-l, The selection of Hastelloy C-276 was based on RTU corrosion
studies (discussed in Section 10) and pump manufacturer recommendations.
The R-l ball valve slurry sampling system was found to operate satisfac-
torily. Both wall and pump-around loop sampling equipment obtained representa-
tive and reproducible R-l slurry samples (discussed in Section 8). Therefore,
early in the testing sequence, circulation loop sampling was terminated due to
a tendency for those sample outlets to clog during use. The direct introduction
of steam into R-l to maintain desired reactor temperatures was found to dilute
the recycled leach solution over prolonged running periods. An appreciable
part of the dilution was due to repeated plant start-ups (the plant was not
operated over weekends, and was shut down during planned maintenance and repairs).
This dilution factor would be less, were the RTU in a continuous operation mode.
However, reactor leach solution regeneration capability was sufficient to main-
tain adequately concentrated ferric reagent supply throughout operation,
despite the dilution effect.
The reactor associated equipment which was subjected to elevated tempera-
ture (>215°F) and oxygen partial pressure operation did experience a systematic
corrosion related problem (discussed in detail in Section 10). During the
course of operations, the reactor R-l, reactor internals, and slurry circulation
loop piping experienced progressively severe corrosion, The reactor and its
internals experienced pitting and crevice corrosion. The pump-around loops
experienced crevice corrosion at flange interfaces and pitting. The damage to
the reactor and loop piping/flanges was sufficiently severe by mid-operations
to warrant reactor repair (filling of the deeper pits with weld rod material)
and replacement of reactor associated piping with new 316L material. At con-
clusion of operations, following unit inspection, the reactor and its associated
piping were again found to have experienced severe pitting corrosion. Based on
the RTU materials study effort (detailed in Section 10), it has been determined
that a new reactor and associated piping, flanges and valves should be con-
structed and put into service prior to further RTU operation. Based on the
corrosion study, the new reactor and piping, flanges and valves should be of
titanium construction. This construction material is predicated not only, by
material study results but also by the size of the equipment (the reactor and
piping are too small in size to be applicable to metallic and/or nonmetallic
54
-------
liners and/or acid brick lining) and the relatively low cost of titanium
replacement parts. For instance, a quote of $49,300 was obtained in January
1978 for a replacement reactor of titanium construction, as compared to $25,900
which was paid for the original 316L reactor in early 1976.
Filtration Equipment. The filtration equipment, including the belt fil-
ter S-l, the belt wash return pump P-8, the filtrate receiver V-2, the filtrate
pump P-9, the wash receiver V-3, the wash pump P-10, and the filter vacuum pump
performed as expected. The belt filter control variables (wash water tempera-
ture, flow rate, belt speed, wash location, vacuum level, etc.) were such that
adequate filtration and washing of 14 mesh x 0 coal was attained up to a total
slurry feed limit of 1500 pounds per hour. The belt filter therefore became
the capacity limiting unit in the RTU. Belt blinding did not occur to an appre-
ciable extent. However, the filter belt did fail (it was literally ripped
apart) at one point in the test program. Belt failure was attributed to a
softening of the belt rubber facing primarily as a result of prolonged elevated
temperature (180° to 190°F) exposure to acidic slurry feed. The belt, as a
result of the softening, experienced increased frictional forces at the belt
interface to the unit vacuum pans. Following belt failure, the belt was
repaired and placed back into service. The vacuum pan edges over which the
rubber belt travels were fitted with teflon overlays. The teflon-belt interface
resulted in less system frictional forces, thus allowing proper belt operation
for the duration of operations.
5.3 RTU OPERATION CONCLUSIONS
\, Pre-start-up RTU safety reviews and personnel training were
successfully completed.
2. Shakedown activities demonstrated equipment operability, and
equipment functionality over the desired test ranges.
3. Shakedown activities verified RTU sampling and analysis pro-
cedures and resulted in the acquisition of baseline process
chemistry data.
4, RTU operating procedures were modified prior to RTU opera-
tions in the context of shakedown derived experience.
5. .A revised RTU Operational Test Plan was generated based on
information obtained during shakedown.
55
-------
6. During RTU operations, 49,700 pounds of coal were treated over
a total of 254 hours.
7. Continuous RTU'operation was demonstrated for periods up to
32 hours in duration.
8. Sixteen distinct sets of operational variables were studied.
9. RTU controllability and flexibility was demonstrated.
10. Sufficient data was obtained during operations to allow proc-
ess chemistry evaluation and verification.
11. Most of the RTU equipment performed as expected without
incident.
12. Some RTU equipment experienced operating difficulties but in
general modification to the affected equipment and/or pro-
cedural changes resulted in acceptable unit operation.
13. Elevated temperature operation as experienced in the reactor
section of the RTU resulted in unexpectedly severe corrosion
to equipment and piping internals.
14. The reactor R-l and its associated piping, flanges and valves
need to be replaced prior to further RTU operation.
56
-------
6. REACTOR TEST UNIT DATA
6.1 SUMMARY OF PREVIOUS BENCH SCALE EXPERIMENTATION
Meyers Process chemistry has been characterized through bench scale
experimentation consisting of more than 200 mass balanced experiments. These
tests were performed primarily with Appalachian coals^4'5'^1^ although coals
from the Interior Basin and Western regions have also been tested^'4'. The
resultant data compilation shows that the pyrite leaching and reagent regen-
eration steps of the Meyers Process may be expressed as follows:
FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 -> 10.2 FeS04 + 4.8 H2S04 + 0.8 S
and
2.3 02 + 9.2 FeS04 + 4.6 H2S04 -* 4.6 Fe2(S04)3 + 4.6 H20.
Pyrite leach reaction products were initially determined by direct meas-
urement of sulfate sulfur and elemental sulfur generated during ferric chloride
leaching. Multi-stage toluene extractions were utilized to ensure complete
elemental sulfur recovery from processed coals. Sulfur balances generally
indicated better than 95% sulfur recovery and further verified the measured
product sulfur ratios. Later experimentation performed with ferric sulfate
reagent also verified preliminary findings through measurement of pyrite
removals and recovered elemental sulfur. Thus it was determined that each
mole of pyrite reacted yields 1.2 moles of reagent soluble sulfate sulfur
(FeS04 and H2S04) and 0.8 moles of elemental sulfur. This product sulfate
sulfur-to-elemental sulfur ratio of 1.5 was found to be independent of the
following processing variables: 1) reaction temperature up to 270 F, 2) reagent
anion (sulfate or chloride), 3) reagent acidity down to a pH of 1, 4) reagent
total from concentration to approximately 10% w/w and 5) oxygen partial pres-
sures to 85 psia. _,
57
-------
It should be noted that losses or incomplete extraction of elemental
sulfur will result in an apparent product ratio which is greater than 1.5.
For this reason, bench scale data obtained using single stage toluene extrac-
tions tend to indicate slightly high product ratios.
Pyrite leaching and reagent regeneration reactions have been found to
proceed at rates which are expressed by the following empirical equations:
-dW
= KL Y Wp
where
KL
Y
W_
is the pyrite leach reaction rate constant
is the ratio of reagent ferric ion-to-total iron concentration
is the weight percent of pyrite in coal
and
r -
rR
dt
where
KF
P0
Fe
2
+2
is the reagent regeneration rate constant
is the oxygen partial pressure
is the reagent ferrous ion concentration.
Rate constants for the pyrite leaching and reagent regeneration reactions,
KL and KR, have an Arrhenius type exponential temperature dependence which is
expressed as:
K = A exp [-E/RT]
where
A is the pre-exponential or frequency factor
E is the apparent overall activation energy
58
-------
R is the gas constant and
T is the reaction temperature.
The pyrite leach reaction frequency factor AL is a function of coal type
and coal particle size distribution. The value of AL increases with decreas-
ing coal particle size distribution for a specified coal. For Lower Kittanning
coals such as that from the Martinka mine, the value of A, obtained from 14
mesh top-size coal samples ranges from approximately 3 x 105 W ~* hour"1 to
7 x 10 Wp" hour" . The value of AL may be increased either by size reduc-
tion of the coal or by physical coal cleaning. Enhancement of A. by physical
coal cleaning may be attributed to the removal of coal particles containing
pyrite having a low specific surface area or low accessibility (isolation by
the matrix material). Thus, pyrite removal rates may be increased by coal
size reduction and/or physical cleaning.
An apparent activation energy for the pyrite leach reaction E, of approxi-
mately 11.1 kilocalories per mole of pyrite reacted was determined from bench
scale experimentation. Hence, toe pyrite leach reaction is strongly tempera-
ture dependent with reaction rates nearly doubling for each 36 F temperature
increment in the temperature range investigated. However, the reaction rate
temperature dependence has been observed to vanish after 80-90% pyrite removal
from some high pyrite ROM Lower Kittanning coals. This observation appeared
to indicate that pyrite leach rates can become diffusion limited and thus,
rather insensitive to temperature changes.
The reagent regeneration rate constant KR is a function of the regenera-
tion pre-exponential factor AR, apparent activation energy ER and reaction
temperature. The value of AR is dependent on the efficiency with which oxygen
and reagent are contacted. That is AR contains an efficiency factor (i.e.,
AR = £A') which relates to the effective reagent-oxygen contact area and
contact time. The efficiency factor e of a specified regeneration apparatus
may have a slight temperature dependence since solution viscosity and specific
volume are temperature dependent. However, within the temperature range
investigated temperature effects upon e are considered to be small. During
bench scale experimentation, values of AR up to 6.7 x 10 liters/mole-
atmosphere-minute were obtained. Since larger values of AR were not attainable
59
-------
at bench scale, this value is assumed to correspond to an efficiency factor
of unity. The value of ED was found to be a constant 13.2 ki local on" es per
+2
mole of Fe oxidized.
Pyrite leach rates are proportional to the square of reagent Y. Thus a
high ratio of ferric ion-to-total iron must be maintained during processing
to maximize pyrite leach rates. In order to maintain high reagent Y values
during processing with low iron reagent or during processing with high solids
content slurries, reagent regeneration may be performed simultaneously with
the leach reaction. Bench scale tests have demonstrated that reagent Y values
and pyrite leach rates are increased by simultaneous leach-regeneration (L-R)
processing with no apparent adverse affect on regeneration efficiency.
Further, L-R processing has not been found to affect the product sulfate
sulfur-to-elemental sulfur ratio.
Selectivity of ferric sulfate toward the oxidation of coal pyrite has
been verified by measurements of excess ion consumption and by measurements
of mineral matter free (mmf) coal heat content changes. The extent of both
excess ferric ion consumption and mmf heat content degradation have been
found to decrease with increasing coal rank. Bituminous coals of Appalachia
may show approximately 1% decrease in mmf coal heat content (very nearly the
limit of analytical precision) as a result of Meyers processing at 212 F while
the mmf heat content of Western sub-bituminous coals has shown an average
decrease of nearly 7% under identical processing conditions. Bench scale
experimentation performed with 14 mesh x 0 mine cleaned Martinka coal and
acidified iron sulfate reagent indicated an mmf heat content decrease of less
than 1% for up to 48 hours of 212°F processing, although L-R processing at
250 F for 6 hours or more resulted in an mmf heat content decrease of approxi-
mately 2%(''), Thus, it appears that selectivity of the iron sulfate reaction
with pyrite may decrease somewhat with increasing reaction temperature and/or
increasing oxygen partial pressure.
Further verification of leach reaction selectivity was provided by mate-
rial balance data which, for Lower Kittanning coals processed at temperatures
of 212 F to 250 F with 0 to 5% w/w iron reagents, indicated virtually complete
solids recovery with a balance precision of approximately 2%.
60
-------
Product sulfate sulfur and elemental sulfur are generally recovered from
Meyers Process leached coal by water washing and subsequent solvent extraction
or vaporization. The sulfate sulfur product is readily dissolved in the
reagent solution and may be recovered from recycled reagent slip-streams either
by liming or by a combination of liming and evaporative concentration of
reagent to precipitate ferrous sulfate. Elemental sulfur recovery can be
effected by solvent extraction with numerous solvents which are non-reactive
with sulfur at moderate temperatures including toluene, acetone and hexane.
Toluene has been most commonly used for routine laboratory extractions. Bench
scale extraction experimentation performed with 14 mesh x 0 mine cleaned
Martinka coal shows approximately 97% elemental sulfur recovery with a single
stage toluene extraction. Acetone has also been demonstrated to be an effec-
tive sulfur extraction solvent which, unlike toluene, is completely miscible
with water and can therefore be utilized in conjunction with coal washing.
As such, acetone may be preferable to toluene for large scale extraction appli-
cations. Alternate approaches to elemental sulfur recovery include techniques
involving vaporization into an inert gas stream and sulfur recovery by vacuum
distillation. However, vaporization techniques were not employed during the
RTU program.
Specific details of previous bench scale experimentation as they apply
to the RTU program are presented in the ensuing data reduction sections.
6.2 DATA FROM THE REACTOR TEST UNIT
Operation of the RTU was continuously monitored and recorded either by
the Doric data logger or manually. Primary data sources, those which required
monitoring for plant operation or evaluation of the Meyers Process kinetics,
were recorded by the data logger on magnetic tapes. These tapes were shipped
to Space Park where they were converted by computer program from the Doric
format to engineering formats. Secondary data from the RTU (i.e., flow rates
not critical to plant operation or kinetic evaluation) were monitored and
recorded manually. Analytical data relating to slurry and coal analyses were
also recorded manually although the raw data was processed through computer
programs to yield the various analytical results in selected formats. Computer
programs developed for data treatment during this program are presented in
Volume II, Appendix A.
61
-------
Magnetic tapes generated by the Doric data logger are capable of storing
data from 99 channels plus one self-check channel. At the conclusion of the
program, 49 channels were in use to record numerical data and 28 channels were
in service as indicators (alarm channels). Remaining channels were available
for allocation as required. Data stored on these tapes were converted from
the Doric format to two engineering formats which are presented in Tables 4
and 5. The format presented in Table 4 represents a scan of all channels in
use at the time indicated by the heading (Julian day, hour:minute:second) and
is essentially self-explanatory. Only active channels are displayed. The
Ramapo excitation monitor (Channel 40) is used as a panel-mounted voltmeter
which facilitates calibration checks of the R-l feed and vent oxygen streams,
Fe-61 and Fe-44. The transducer pressure sensor monitor (Channel 99) serves
a similar function with regard to pressure measurement equipment. The format
presented in Table 5 provides all the data taken for each channel during a
specified experimental period. The latter format is somewhat more abbreviated
than that presented in Table 4 and improves the ease of data evaluation and
interpretation in terms of identifying trends. Thus, the format presented in
Table 5 was used most frequently.
After conversion of the magnetic tape contents to engineering units, the
data were available in storage for use in any of various computer programs
prepared during this project to facilitate data reduction (i.e., plotting,
averaging, mass balances, rate determinations, etc.). Similarly, all control
lab analytical data in storage was available for recall into the data reduc-
tion programs. Manually recorded pilot plant data required as input for data
reduction programs were input to central data storage through the computer
terminals or, for large inputs, were submitted as a keypunch card deck. Thus,
nearly all data obtained from the RTU was centrally stored and readily avail-
able for recall. Only those manually recorded data relating solely to plant
operation were not included in the centralized data compilation.
A complete file of the raw data obtained from the RTU facility is stored
at TRW. Data which are pertinent to process evaluation are presented in the
ensuing subsections of Section 6.
62
-------
TABLE 4. RTU CHANNEL-SCAN AT FIXED TIME
RUN 03-03-00 DAY 10 513*100
CHANNEL
0
1
i
3
4
5
6
7
e
9
10
11
12
13
15
16
17
18
19
20
.21
22
23
24
25
26
27
28
29
30
31
32
11
34
35
34
37
38
39
40
41
42
43
44
.45.
46
47
48
90
91
99
PAPAMlTfR
SELF CHECK
T-Z CELL 1 TEW«>
T-2 CEIL 2 TE*P
T-2 CELL 3 TEMP
R-l CELL 1 TFMP
R-J Cf LL ? TEMP
R-l CELL 3 TEMP
R-l CELL 4 TEM»
R-l CELL 5 TEMP
OXYGEN TFMP TO R-l
V-l DISCH GAS TPMP
LEACH TO V-l TP.MP
LEACH FROM V-l TFMP
R-2 TEMP
QMGEN FLOW TO R-l
V-l DISCH GAS FLOW
02 ANAL REFERENCE
02 ANAL SAMPLE
AUTOWEIGH TOTAL
AUTOWEIGH RATE
R-i PRESSURE
OXYGEN TO R-I
V-l DISCH GAS PPFSS
P-l DISCHARGE FLOW
P-2 DISCHARGE FLOW
P-3 DISCHARGE FLOW
P-4 DISCHARGE FLOW
P-5 DISCHARGE FLOW
P-6 DISCHARGE FLnw
T-3 LEACH FLOW
P-9 DISCHARGE FLOW
P-10 DISCHARGE FLOW
?-7 DI5CHAR6E FLOW
T-2 SLURRY LEVFL
R-l SLURRY LEVFL
pyMP p-i JJISCH ol»FSs
N2 SUPPLY TO R-l
S-l VACUUM PAN A
S-l VACUUM PAN 9
RAMAPO EX. MONITOR
LEACH SL TEMP TO T-3
T-4 TANK TE*P
v-i TANK TEMP
T-4 H20 IN TEMP
T-JO FLUSH H20 TPMP
P-l DISCH TEMP
SI INLET TEMP SLURRY
T-4 H20 OUT TE*»
PUMP P-7 DISCH PRESS
if-2 LIQUID LEVFL
XOUCFR P»S. MONITOR
TAG
UJR-170
T*-l<>
TF-20
TP-?1
TF-5?
TP-53
TF-54
TF-55
TF-S6
TP-173
TF-149
TF-10Q
TF-?00
TF-90
FF-fcl
PP-44
A P-l 71
AF-171
WT-9
WT-9A
PT-173
PT-17'
PT-150
FP-?«»
FP-33
FF-»»4
F*-«5
FF-«6
FP-ft?
FP-31
FF-157
FP-15P
FF-101
PT-26
PT-5B
PT-?<»
<»T-107
PT-249
PT-?50
TP-3?
TF.-?Ol
TF-202
TF-203
TF-195
TP-?«
TP-
TP-204
pT-ibd
PT-95
DATA
4.999
135.7
134,3
1?7.Q
'1?.3
?02,5
20"5.4
?06.2
205.7
•55.0
55.0
50.5
71.6
56.1
*.l
o.o
OR. 61
*°*15
5661
-.2
-57.50
0,0
?1.8
.920
.869
.674
1*842
.465
-.077
-.014
.006
.005
-1.250
13.3
11.0
-50,00
60.10
14.560
-3,750
-6.606
5^.5
65.1
60.0
6ft. 6
69.2
77.3
5". 7
IS". 2
-7.96
2?, 3
14.781
UNITS
VOLTS
DEC- F
•^fc *• f
PEG F
9* & w r
DEC F
OEG F
DEG F
DEG F
DEG F
DEG F
OEG F
OEG F
DEG F
OEG F
DEG F
SCFH
SCFH
PERCENT
PERCENT
POUNDS
POUNDS/MR
PSIG
PSIG
PSIG
GPU
GPP
6PM
GPf
GPM
GPU
GPP
GPM
GPP
GPM
INCHES
INCHES
PSIG
PSIG
PSIA
PSIA
VOLTS
DFG F
DEG F
DEG F
DEG F
PEG F
DEIS F
DtG F
DEG F
PSIG
INCHES
ny
ALARM
LOWER
LOWER
63
-------
TABLE 5. RTU TIME-SCAN BY CHANNEL
en
RUN
DAY
10
10
10
_m_
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
03-03-00
TIME
1)23)00
1(26(00
1(29)00
1(35)00
1(38)00
K4KOO
1(44(00
1(47(00
1)50(00
1(53(00
1)56(00
1)59)00
2102(00
2(05(00
2(08(00
2(11(00
2d7(00
2)20100
2(23(00
2(26(00
2(29(00
2(32(00
2(35(00
2(38100
2(41(00
2144(00
2(47(00
2(50(00
2(53)00
2(56:00
2(59100
3(02(00
3(05(00
3)08(00
3(11(00
3114(00
CHANNEL
0
UJR-170
4.999
C_HAMEJL_
1
TE-19
156.5
CHANNEL CHANNEL
2 3
TE-20 TE-21
207.9 213.9
4.999 159.8 205.1
4.999 159.5 200.7
4.999 158.1 JO.4.,4
4.999 157.8 209.4
4.999 159.3 211.2
4.999 161.3 210.1
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4,999
4.999
4,999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
163.9
166.2
166.8
166,2
166.0
165.0
164.8
164.4
163.9
163.0
162.6
163.0
164.6
165.4
164.9
163.4
162.2
163.0
164.4
166.2
167.8
170.0
171.8
174.4
177,4
173,6
176.6
171,6
166. B
161.3
159.6
211.2
209. A
207,2
205.5
206.5
204.7
203.6
203. S
202.9
203.0
208.8
210.1
208.9
204.3
202.4
205.7
20B.1
210.8
211.6
212.9
213.2
213.8
233.9
212.9
213,3
211.7
207.4
202.7
199.4
193.9
190,1
213.8
213.8
2JL3.J>_
813.8
214.0
213.9
213.9
213.7
213.8
213.8
213.8
ai3.7
213.7
213.7
213.4
213.4
213.6
213.9
213.3
213.8
213.6
213.5
213.8
213.8
Z14.0
213.9
£14.0
214.0
214.1
214.0
214.0
214.0
213.8
213.9
213.6)
212.9
211.6
CHANNEL C
4
TE-52 T
235.9
234.5
233.5
zYi. 7 "~
231.5
230.3
228.7
226.6
£26.1
226.3
223.4
223.2
223.9
224.6
223.5
222.0
221.5
219.7
E19.7
219.5
219.9
219.9
219.2
215.5
217.3
219.2
219.7
216.5
219.0
219.3
217.5
217,3
217.2
217.9
216.7
216.4
;HANNE_L a
5
'£-51 Tl
230.0
229.1
229.3
"228"". 3
227.7
227.2
226.0
225.9
225.5
224.9
224.2
223.8
223.0
222.1
221.6
221.0
220.8
221.1
221.0
220.5
221.2
221.4
221.1
221.0
221.5
221.2
220.3
220.4
219.9
219.1
219.5
219.5
219.4
219.3
219.4
219.4
218.9
HANNEL
6
E-54
236.7
235.4
235.0
234,3
233.7
232.9
232.1
231.1
230.4
229.7
229.6
226.5
226.4
227.2
226.6
226.7
226.8
230.2
233.5
236.3
236.7
240.6
241.9
241.6
238.7
236.9
234.5
233.3
232.2
233.3
235.1
235.7
236.6
237.5
238.5
239.3
239.4
239.9
CHANNEL
7
TE-55
237.5
237.5
236.8
2.36. 4_
236.0
235.5
234.9
234.5
233.1
232.5
231.3
231.7
230.8
230.0
229.6
229.1
229.1
228.4
228.1
228.5
228.9
229.3
230.3
230.9
231.5
231.6
232.4
232.0
232.1
231.3
231.4
231.5
231.3
231.5
231.6
231.8
232.0
232.1
CHA N tiii. -CHAUHEi _ £iLAJ4N£L Jl H AJJ N E L
6 9 10 11
TE-56 TE-173 TE-148 TE-199
240.4 62.6 57.6 59.3
240.1
236.9
. 23,8.. 2
237.4
236.6
235.9
235.6
235.0
234.5
233.2
233.0
232.0
231.2
230.9
229.8
234.3
229.0
226.6
228.2
227.3
228.7
236.1
238.5
237.2
236.1
235.7
234.4
233.7
233.6
232.8
232.3
231.8
231.2
230.6
230.2
229.8
229.8
62.2
62.1
61.9
61.9
61,9
61.8
62.1
61,7
61.8
61.6
61.8
61.6
61.4
61.4
61.4
61.3
62.1
61.2
60.6
60.6
60.4
60.3
60.2
60.3
60.2
60.2
60.1
60.4
60.4
60.2
60.4
60.2
60.5
60.7
60.3
59.5
57.8
57.3
5ii«L_
57.4
57.3
51*4
57.9
57.5
97.9
57.7
57.6
57.6
57.7
58.2
57.8
58.2
58.2
58.1
57.6
57.7
58.0
58.0
57.4
57.4
57.6
57.6
57.8
57.3
57.5
58.0
58.0
57.6
57.5
57.4
57.4
57.4
57.8
59.7
59.3
59.2
59.2
59.4
5A.3
59.5
59.4
~59.4~
59.4
59.2
59.6
59.1
59.7
59.2
59.0
59.4
58. 8
59.4
59.5
59.2
59.4
58.7
59.4
59.4
59.3
59.4
59.6
59.6
59.7
59.7
59.1
59.2
58.4
59.4
-------
6.2.1 Mine-cleaned Martinka Coal Characterization
Samples of mine-cleaned Martinka coal processed in the RTU were provided
by American Electric Power (AEP) at three different times: (l) coal No. 1
was received in November 1976 and was utilized for RTU shakedown operations,
(2) coal Nos. 2 and 3 were received in June and July of 1977, respectively,
and were utilized during the RTU operational phase of the current program.
Samples of each coal were obtained for analysis from the A-3 weigh belt at
frequent intervals during plant operation. All samples were analyzed at the
TRW control lab for sulfate and pyritic sulfur (sulfur forms) and a portion
of these were additionally analyzed by Warner Laboratories for moisture, ash,
heat content, and total sulfur (short proximate analysis), as well as for
sulfur forms. Warner Laboratories also performed coal iron analyses on all
samples submitted.
Starting coal analyses of the mine-cleaned Martinka coals are presented
in Table 6. Results from each laboratory are presented separately and then
averaged for each coal. The first column of the Table lists the identifying
number of the coal lot from which analytical samples were drawn. Columns two
and three list the analytical laboratory and number of samples analyzed by
each. Short proximate, sulfur forms, and coal iron analyses are listed in the
remainder of the columns.
Analyses of coal No. 2 are presented in two parts since approximately
half of coal No. 2 (Coal 2A) was stored under a partially enclosed shelter
prior to processing while the other half (Coal 2B) was fully exposed to the
atmospheric environment. Coal Nos. 2A and 2B were treated as distinct RTU feed
coals prior to complete analysis since the extent of weathering in these two
coals, and consequently their starting pyritic sulfur content, was expected
to differ.
To compare coal Nos. 1, 2, and 3, consider first their respective dry
short proximate analyses (ash, heat content, and total sulfur). Coal No. 1
differs significantly from coal Nos. 2A, 2B, and 3 with respect to ash content;
this difference exceeds three times the largest indicated standard deviation
in ash measurement. On the other hand, coals 2A, 2B, and 3 are statistically
identical with respect to ash analyses. As would be expected for coals from
the same mine, the same set of observations holds true with respect to coal
65
-------
TABLE 6. AS RECEIVED MINE-CLEANED MART IN KA COAL ANALYSES
Coal ID Lab "^
1 Warner 3
CTS 14
Average
2A Warner 8
CTS 10
Average
2B Warner 7
CTS 16
Average
2 (2A Warner plus
2B Warner & CTS)
3 Warner 10
CTS 12
Average
Moisture
1.32
±0.173
1.32
±0.173
0.93
±0.285
0.93
±0.286
0.79
±0.183
0.79
±0.183
0.86
±0.170
1.01
±0.204
1.01
±0.204
Ash
16.11
±0.261
16.11
±0.261
14.05
±0.339
14.05
±0.339
14.24
±0.354
14.24
±0.354
14.15
±0.245
14.32
±0.469
14.32
±0.469
Heat
Content
12508
± 77
12508
± 77
12928
± 103
12928 '
± 103
12919
± 33
12919
± 33
12923
± 54
12907
± 58
12907
± 58
ST
1.73
±0.045
1.73
±0.045
1.61
±0.089
1.61
±0.089
1.57
±0.059
1.57
±0.059
1.59
±0.053
1.49
±0.081
1.49
±0.081
sp
0.63
±0.063
0.70
±0.069
0.68
±0.071
0.62
±0.062
0.79
±0.061
0.71
±0.107
0.56
±0.065
0.57
±0.072
0.57
±0.069
0.58
±0.038
0.60
±0.023
0.67
±0.043
0.64
±0.051
Ss
0.47
±0.035
0.43
±0.076
0.43
±0.071
0.32
±0.058
0.30
±0.041
0.31
±0.049
0.35
±0.054
0.35
±0.073
0.35
±0.067
0.34
±0.036
0.31
±0.029
0.31
±0.026
0.31
±0.027
So
0.63
±0.011
0.62
0.67
±0.080
0.59
0.66
±0.054
0.65
0.67
0.59
±0.067
0.54
Fe
1 .33
±0.010
1 .33
±0.010
1.14
±0.063
1.14
±0.063
1 .11
±0.059
1.11
±0.059
1 .13
±0.043
1.02
±0.071
1 .02
±0.071
-------
heat content. Considering coal total sulfur analyses, coals 2A and 2B are
statistically identical, yet differ significantly from coals 1 and 3 which
differ significantly from each other. Hence, based on the short proximate
analysis, coal Nos. 1, 2, and 3 differ in total sulfur while 2A and 2B are
identical. These results are not surprising since the three coal samples
represent output of the Martinka mine cleaning facility at three different
times. Note that while the short proximate analyses of coal Nos, 1, 2, and
3 are significantly different in a statistical sense, they may not represent
significant variance in output for a commercial cleaning facility.
Best estimates of the starting coal sulfur forms were generally obtained
by averaging mean results from both analytical laboratories. This approach
was applied to obtain sulfate sulfur estimates for all starting coals and
pyritic sulfur estimates for coal Nos. 1 and 3. However, considerable bias
was found between CTS and Warner analyses of S in coal 2A; CTS and Warner
analyses of 0.79 and 0.62, respectively, differ by approximately three standard
deviations. Since in all other respects coal Nos. 2A and 2B appear to be
identical, they should have identical S contents. Considering that Warner
analyzed coal No. 2A to have 0.62% w/w S and that both labs agree on a
0.57% w/w S content for coal No. 2B, the CTS analyses showing coal No. 2A
to contain 0.79% w/w S are assumed to be Incorrect. Hence, the best estimate
of coal No. 2 S content was taken to be the average of available mean values,
disregarding the CTS analysis of coal No. 2A. Average coal analyses thus
obtained for coal Nos. 1,2, and 3 were utilized in all ensuing leach rate
evaluations.
Measured bias in sulfur forms analyses performed by CTS and Warner
Laboratories is summarized graphically in Figures 15 and 16 for starting and
RTU processed mine-cleaned Martinka coals. Warner analyses are plotted as a
function of CTS analyses in these Figures and the zero bias line is provided
for reference. Pyritic sulfur analyses are seen to be only slightly biased
with most deviations from zero bias falling within the limits of combined
analytical and sampling uncertainty (approximately 0.06% w/w during this
experimentation as indicated by starting coal analyses). The one data group-
ing with bias which significantly exceeds analytical and sampling uncertainty
is the previously mentioned starting analyses of coal No. 2A. Indeed, analyses
67
-------
COAL 1
COAL 2A
COAL 2B
o COAL 3
.1 .2 .3 .4 .5
CIS ANALYSES
.6
.8 .9
Figure 15. Comparison of Capistrano and Warner Laboratories
Pyritic Sulfur Analyses
of both starting and processed coal No. 2A indicate that CIS S analyses are
biased somewhat high with respect to Warner analyses. Whether this is due to
the analytical apparatus or personnel variance is not known. The generally
slight bias observed is substantially smaller than the ASTM acceptable-differ-
ence between two laboratories analyzing the same coal with less than 2% S ,
namely 0.30% w/w. Hence, obtaining best estimates of the true coal S content
by averaging results from both laboratories appears justifiable on the basis
of both starting and processed coal analyses.
Bias associated with S$ analyses is negligible as indicated by Figure 16.
Deviations from zero bias are generally less than 0.04% w/w which corresponds
68
-------
A COAL 1
• COAL 2
o COAL 3
0 o.l 0.2 0.3 0.4 0.5 0.6 0.7 0.8
CTS ANALYSES
Figure 16. Comparison of Capistrano and Warner Laboratories
Sulfate Sulfur Analyses
to the ASTM acceptable difference between analyses for S by two different
o
laboratories.
Coal samples were provided by AEP in 70 to 90 ton lots at a 1.25-inch
top size. These coal samples were received by Ultrasysterns, Irvine, California,
where they were size reduced to approximate the nominal 14 mesh x 0 size dis-
tribution previously tested at bench scale. Size distribution data from coal
Nos. 1, 2, and 3 are presented in Table 7. These data indicate good coal
preparation reproducibility. A graphical comparison of these data and pre-
viously obtained bench scale size distribution data is presented in Figure 17.
Coals prepared for the RTU are seen to be only slightly finer than the nominal
14 mesh x 0 mine-cleaned Martinka coal utilized during bench scale testing.
69
-------
TABLE 7. SIZE DISTRIBUTION DATA FOR MINE-CLEANED MARTINKA COAL PROCESSED IN THE RTU
Screen
(Tyler Equiv.)
35
48
100
150
200
Opening
(Microns)
420
297
149
105
74
Cumulative Percentages by Weight
Coal 1 (6 samples)
7.9
18.1
43.4
53.0
63.5
± 0.49
± 0.68
± 1.13
± 1.10
± 1.13
Coal 2 (
6.5
14.7
38.5
52.0
62.9
8 samples)
± 1.14
± 1.66
± 2.41
± 1.68
± 1.42
Coal 3 (1
8.6
16.6
40.9
54.3
68.6
2 samples)
±0.63
±0.89
± 2.06
± 3.20
± 4.66
-------
CUMULATIVE PERCENTAGE BY WEIGHT
o
o
OJ
o
CO
o
in
o
vo
o
00
LU
o.
o
.ObO
742
.525
071
. 31 I
.263
IOC
.131
• uz/o
nee
.0328
.0232
.0164
.0116
nno9
.UUo£
0058
.0041
no?q
KB
*
>DO
*
w
p
»
*
JO
i— C
•J M
4 K
^ r
n -
'
cm
o
X
DC
LU
LU
sj CO i
J — 1 O
£ < CO
3 0
Jr > -T-
3 =3 Z
- |— LU
c D: CQ
a n ^
^
X
n:
CO
s:
o
0
LU
3;
C_3
•sz
LU '
CQ
1
A
— i.
s
o
p m
i— «
in o
1 U j"
1 A ^<
_Li_ ;=
rn
20 3
m
on
48
cc
_a2._
100
150
200
Figure 17. Mine-Cleaned Martinka Coal Size Distribution Data
71
-------
6.2.2 Coal Processing Data
RTU experimentation was performed under a variety of operating conditions
which, at bench scale, were indicated to be suitable for demonstration of both
plant operability and viability of the Meyers Process at large scale. Param-
eters of primary interest were varied as follows:
• Reaction temperature - 230°F to 270°F
• Coal residence time - 5 to 10 hours
• Reagent total iron concentration - 0.0 to 4.5% w/w
t Reagent acid concentration - 0 to 4% w/w
• Oxygen partial pressure — 20 to 50 psia.
Reaction times were varied by varying the number of stages utilized in
the primary reactor R-l and/or varying the slurry feed rate. Coal feed rates
of 200 to 300 pounds per hour were utilized with reagent feed rates of 1 to
2 gpm. This corresponds to a range of slurry solids concentrations from 25-
33% which was tested during the RTU operational phase.
Process evaluation was based primarily on analyses of product coal samples
taken from the RTU filter belt S-l. Product coal samples were generally taken
from the filter belt every hour during transient plant operation and more
frequently during steady state operation. Slurry samples drawn from various
stages of the mixer T-2 and the reactor R-l provided data on reagent total
+3
iron concentration, reagent Y (Fe /total Fe ratio) and slurry solids.
Coal samples obtained from S-l received a water rinse on the filter belt
and were essentially free of bulk reagent. Final coal washing was performed
at the CTS control lab and consisted of a slurry wash and subsequent cake wash.
Slurry samples were filtered and the solids were washed; slurry filtrates were
analyzed for total iron, iron forms, and sulfate content.
The water wet coal was vacuum dried at 212°F for approximately 4 hours
prior to toluene extraction of product elemental sulfur. The vacuum drying
process was repeated after toluene extraction of the coal. Product elemental
sulfur was recovered from the toluene extract by solvent distillation. Fully
processed coal samples were then subjected to short proximate, sulfur forms,
and iron analyses. The specific detailed procedures followed in accomplishing
72
-------
these operations were identical to those developed during previous bench-
scale programs f4'5'11).
Bench scale testing of vacuum distillation^ ' indicates that under the
conditions of vacuum drying employed during the RTU program, little or no loss
of elemental sulfur is incurred. Hence, vacuum drying prior to toluene extrac-
tion of processed coals was not expected to significantly affect measured
product sulfate sulfur-to-elemental sulfur ratios.
Raw data from the workup and analyses of coal and reagent samples were
reduced and tabulated in the desired format by computer. Computer programs
written during this project to perform these computations are included in
Volume II, Appendix A.
6.2.2.1 Pyrite Leach Rate Data--
Experimentation performed with the RTU consisted primarily of two types:
1) tests performed with acidified iron sulfate reagent and 2) tests performed
with initially iron free aqueous dilute ^SO^. In the latter case, reagent
iron was derived entirely from the coal being processed.
During testing with acidified iron reagent, the leach solution was
recycled back to the reaction system via the leach solution storage tank T-7.
This mode of operation resulted in reagent dilution from the live steam used
to maintain reactor temperature since salt was not continuously added to the
reagent recycle stream. Limited salt addition was employed to reduce this
dilution effect. However, since reagent total iron concentration does not
directly affect pyrite leach rates during simultaneous Teaching-regeneration
operation, reagent iron concentrations were allowed to range from 4.5% w/w
to 2.7% w/w.
Results from Processing with Acidified Iron Sulfate Reagent
Eleven sets of experimental conditions were tested (Experiment 01) with
acidified iron sulfate reagent during the operational phase of the program.
The initial reagent feed was composed of 4.5% w/w iron (as ferric sulfate)
with 4.2% w/w H2S04. Owing to dilution, the reagent total iron concentration
ranged down to 2.7% w/w. However, a sulfate-to-iron mole ratio of approxi-
mately 2 was maintained for all tests.
73
-------
A summary of the principal parameters associated with that set of experi-
mental conditions is presented in Table 8. The experiment numbers are listed
in the first column. Columns two and three list the experimental start and
stop times, specified in terms of Julian day and hour. Reactor temperature,
pressure and oxygen partial pressure are presented in columns four, five and
six. Feed rates of coal, leach reagent and oxygen are listed in columns seven
through nine. Listed in column ten are the flow rates of slurry through the
R-l circulation/oxygenation loop. Columns eleven and twelve contain the esti-
mated mean slurry residence times in T-2 and R-l based on the ratio of reactor
volume-to-volumetric slurry flow rate. All experiments were performed utilizing
only three of the available five stages of R-l, thereby maintaining relatively
short residence times.
Evaluation of this experimentation was performed on the basis of S analy-
ses of steady state coal samples taken from the RTU filter belt. The minimum
processing time required to achieve steady state operation under constant
operating conditions was estimated as the ratio of utilized vessel volume (T-2
and R-l) to slurry volumetric flow rate. Verification of this steady state
estimate through product coal analyses is typified by Experiment 01-03, the
data from which are presented in Figure 18. Steady state for Experiment 01-03
was estimated to be attained after 7.7 hours of plant operation at the speci-
fied test conditions. Note that the product coal S content reached a value
which is constant, to within analytical precision, after only 5 hours of
operation. These data are typical of all experiments performed to date with
the RTU. Hence, a plug flow type steady state estimation was found to be a
suitable guideline for processed coal data acquisition from the RTU.
Steady state processed coal analyses from RTU processing with iron reagent
are summarized in Table 9. All steady state data are included in the presented
averages. The number of processed coal samples submitted for short proximate
and sulfur forms analyses are designated by xP and yF, respectively, and listed
in column two of the Table. Mean R-l reaction temperatures are listed in col-
umn three. And columns four through ten contain the average short proximate,
sulfur forms and coal iron analyses. Experiments 01-04 and 01-08 are considered
to be merely extensions of Experiments 01-03 and 01-07, respectively, since no
effect was observed by increasing slurry recirculation rates from 20 gpm to
25 gpm. Hence, these experiments are not listed separately in Table 9.
74
-------
TABLE 8. OPERATING PARAMETERS FOR ACIDIFIED IRON SULFATE REAGENT LEACH EXPERIMENTATION
-J
en
Experiment Run time
No. ' (day/hour)
Start Stop
01-01
01-02
01-03
01-04
01-06
01-05
01-07
01-08
01-09
01-10
01-11
299/16:15
300/03:15
305/10:00
305/21:30
305/00:00
314/15:45
321/17:00
322/02:50
322/05:03
322/13.54
325/15:30
300/02:00
300/10:10
305/21:30
305/23:30
306/10:45
315/02:00
322/02:27
322/04:50
322/13:33
323/02:30
326/03:00
R-l*
Temp
(°F)
230
231
251
251
251
250
245
245
270
270
251
R-l
Pressure
(psig)
28
55
54
53
67
40-54
25-68
65
77
51
67
02 Coal
Pressure Feed
Rate
(psla) (Ib/hr)
22
49
39
38
52
25-39
12-55
52
50
24
52
299
299
299
299
299
298
298
298
298
298
298
Leach
Feed
Rate
(gpm)
1.0
1.0
0.8
0.8
0.8
1.0
1.0
1.0
1.0
1.0
1.0
02
Feed
Rate
(SCFH)
56
57
54
59
123
165
59
57
56
56
128
Reel re.
Rates
(gpm)
20
20
20
25
20
20
20
25
20
20
20
Residence time**
T-2 R-l
(hrs) (hrs)
1.6
1.6
1.83
1.83
1.83
1.6
1.6
1.6
1.6
1.6
1.6
5.1
5.1
5.9
5.9
5.9
5.1
5.1
5.1
5.1
5.1
5.1
* The temperature of the primary reactor was varied as required for parametric investigations. T-2
mixer temperatures were the same for all experiments and ranged from approximately 170°F in the first
stage to 214°F 1n the third stage.
** Slurry residence time 1s computed as the ratio of reactor volume-to-volumetric slurry flow rate.
-------
0.20
01 ft
. 1 o
01 r
, 1 b
01 /i
. 1 *\
•3.
^ 0 12
Q.
00 o in
_, U-IU
0
-_. n np
tJ U , Uo
LU
oo
UO
LU
S^ n nfi
o U .UD
OL.
On/i
. U'l
n n?
0
.
•
m
-H
i — i
—1
- n
0
CO
—I
m
a
-H
m
—I
t— <
m
I
•
AVE. =
0.165
t o.c
129 %
W/W
0123
456 7 8 9 10 11 12
EXPERIMENTAL RUN TIME, HOURS
Figure 18. Pyritic Sulfur Analyses of Processed Coal from Experiment 01-03
-------
TABLE 9. SUMMARY OF PROCESSED COAL ANALYSES OBTAINED DURING
RTU EXPERIMENTATION WITH ACIDIFIED IRON REAGENT
Exp. No. No. of* Processing
Samples Temp.,
(°F)
Starting 15P; 31 F
Coal 2
01-01 2P; 4F 230
01-02 IP; 7F 231
01-03 3P; 6F 251
01-05 IP; 4F 250
01-06 2P; 4F 251
01-07 2P; 7F 245
01-09 IP; 3F 270
01-10 2P; 12F 270
01-11 2P; 11F 251
Coal Composition, % w/w (Except Heat Content), Dry Basis
Ash
14.15
±0.245
12.28
±0.212
13.14
12.73
+0.085
13.96
12.69
±2.14
13.32
±0.863
10.65
13.56
±0.636
13.80
±0.636
Heat
Content,
Btu/lb
12923
± 54
13397
± 52
13300
13286
+ 65
12888
13163
± 299
13123
± 52
13487
13158
± 119
13071
± 96
Total
Sulfur,
St
1.59
±0.053
0.94
±0.000
0.85
1.00
+0.006
0.85
0.90
±0.014
0.87
±0.035
0.93
0.91
±0.092
0.83
±0.021
Pyri ti c
Sulfur,
SP
0.58
±0.038
0.23
±0.083
0.25
±0.055
0.16
+0.033
0.16
±0.068
0.15
±0.027
0.19
±0.075
0.13
±0.006
0.17
±0.037
0.16
±0.044
Sulfate Organic
Sulfur, Sulfur,
Ss So
0.34 0.67
±0.036
0.09 0.62
±0.033
0.13 0.47
±0.120
0.09 0.75
+ .040
0.11 0.58
±0.025
0.11 0.64
±0.015
0.12 0.56
±0.023
0.18 0.62
±0.015
0.12 0.62
±0.029
0.13 0.54
±0.025
Iron,
Fe
1.13
±0.043
0.52
±0.071
0.46
0.45
±0.007
0.53
0.42
±0.029
0.47
±0.021
0.50
0.49
±0.035
0.55
±0.007
The number of short proximate analyses (ash, heat content, total sulfur, and coal iron) and sulfur
forms analyses (sulfate and pyritic sulfur) will be designated by xP and yF, respectively, where x
and y are the number of analyses included in the indicated average. The total number of coal samples
analyzed is y minus x.
-------
Data presented in Table 9 indicate that all experiments, except for those
run at the nominal 230°F temperature, resulted in product coal Sp contents
between 0.13% w/w and 0.19% w/w. While the 0.06% w/w spread is very nearly
the limit of pyritic sulfur analytical precision, the indicated Sp values are
means of distinct populations obtained under different operating conditions
(i.e., temperature, pressure, reagent Y) and must be treated as discrete data.
Experiments 01-01 and 01-02, which were performed at nominally 230 F, resulted
in product coal S values of 0.23% w/w and 0.25% w/w, respectively.
P
Product coal S contents show an average reduction of 64% corresponding
to a reduction of S from 0.34% w/w to approximately 0.12% w/w. The slightly
higher product coal S content obtained from Experiment 01-09 may be the result
of combined high temperature and high oxygen partial pressure (49 psia com-
pared to 23 psia for Experiment 01-10) since the solubilities of basic ferric
sulfates, ferrous sulfate and gypsum all decrease with increasing temperature
above about 212°F. However, the S data presented in Table 9 do not generally
indicate a temperature effect upon product coal sulfate contents.
The total sulfur content of the mine-cleaned Martinka coal was reduced
from 1.59% w/w to 0.83-1.01% w/w under the conditions tested. Additional water
washing of these product coals to reduce their S contents to the design level
o
of 0.02% w/w would enable products of the majority of these experiments to meet
federal standards of 1.2 pounds of S02 per 106 Btu.
Ash removals above those anticipated on the basis of pyrite and iron
sulfate removed from ROM Martinka coals have previously been found to exceed
(1)
four percentv . However, the data presented in Table 9 shows a mean ash
decrease of approximately 1.25% w/w which may be compared to the mean theo-
retical ash removal (based on AS and AS ) of 1.04% w/w. Thus, only 0.2% w/w
of excess ash was removed. This is not surprising since the concentrations of
reagent soluble non-pyritic mineral matter (i.e., iron oxides, iron sulfate and
gypsum) are expected to be reduced by physical cleaning. That this is the case
for the Martinka coals provided for this program is indicated by the fact that
these coals contain little or no iron other than that associated with pyrite
and iron sulfate. Although little excess ash was removed from mine cleaned
Martinka coal, the removal of 1.25% w/w ash orginating from pyrite and sulfate
represents an upgrading of the coal with respect to both the coal sulfur content
and the coal heat value.
78
-------
Correlation of these data was obtained with the previously determined
pyrite leaching rate expression, namely:
-dW
-E =
dt
2 U2
(1)
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,
t is the reaction (leaching) time in hours,
Y is the ferric ion-to-total iron ratio in the leacher at time
t, dimensionless, and
K. is the pyrite leaching rate constant (a function of temperature
and coal particle size) expressed in (hours)-l (wt. percent
pyrite in coal)"1.
with
KL = AL exp (-EL/RT).
(2)
where
R
T
is the Arrhenius frequency factor in the units of K. ,
is the apparent activation energy in calories/mole,
is the gas constant in calories/mole°K, and
is the absolute temperature, in K.
This rate expression was used in conjunction with a model of the reaction
system as a series of continuous-flow stirred-tank reactors. Measured slurry
temperatures and reagent Y values for each reaction stage were input for each
experimental condition and the value of K, was determined at the various
reaction temperatures tested.
79
-------
Equation (2) indicates that a plot of In KL vs. 1/T should yield a
straight line having a slope of -EL/R and an intercept of In A,_. An Arrhenius
plot of the data from Experiments 01-01 through 01-11 is presented in Fig-
ure 19. A very good linear correlation was obtained. The apparent activation
energy and frequency factor indicated by these data are:
E. = 26.9 Kcalories/mole, and
14 -1 -1
AL - 9.2 x 101 W * hr .
These values may be compared with bench scale coal processing results
which yielded an E (Bench) of 11.1 Kcalorie/mole and AL (Bench) of 2.95 x
105 W"1 hr"1. Hence, K, (RTU) was measured to be 3 to 8 times greater than
K (Bench) for 14 mesh top-size coal at temperatures between 230 F and 270 F.
The deviations of A. and E, from the values obtained during batch reactor
(bench scale) operation were only partially anticipated. The apparent value
of A. was expected to increase during stagewise operation as the mixing in
reaction stages deviated from ideal and the average particle residence time
per stage increased. Also, under non-ideal mixing, pyrite rich particles
might be expected to have longer residence times than pyrite lean particles
which have a lower density. Thus, coal processing in a continuous flow
stirred tank reactor is expected to result in a higher apparent A. value than
would identical coal processing in a batch reactor. However, the apparent
doubling of the activation energy E. above E. (bench) was not anticipated.. Jo
effect such an increase in EL requires that deviations from ideal mixing or
segregation of pyrite rich coal particles increase with increasing temperature.
A decrease in reagent density could increase coal segregation despite mechan-
ical mixing and, perhaps, selectively cause the pyrite rich coal particles to
remain longer in each reaction stage. While density changes due to tempera-
ture alone are small, this effect coupled with reagent dilution can be sig-
nificant. Note that the conditions of Experiment 01 were performed in the
order of increasing temperature (Table 8) and that, as previously stated, the
reagent total iron was reduced from 4.5% w/w to 2.7% w/w by dilution with
process steam. The combined temperature and dilution effects are estimated
to have resulted in mean reagent densities of 1.15 gm/cc during 230°F tests,
1.08 gm/cc during 250°F tests and 1.06 gm/cc during 270°F tests. Thus, sig-
nificant reagent density differences occurred between the 230°F and 270°F
80
-------
3.
2.
1.0
o xo
Q
,«• .8
\
{_>
UJ
a:
270°F
250°F
230°F
2.4
2.5
2.6 x 10~a
1/T, '
Figure 19. Arrhenius Plot of Data from RTU Experiments
01-01 to 01-11
81
-------
experimentation which may have resulted in the apparent increase in the pyrite
leaching activation energy. Whether or not this density difference is in
itself sufficient to cause the observed factor of 2 increase in activation
energy cannot be determined from available data. However, if the observed
increase is the result of a density effect then reagent density would become
an important reaction design parameter for systems utilizing stirred tank
reactors.
Results from RTU Processing with Low Iron Reagent
Two groups of experiments were performed in the RTU with low iron reagent:
(1) shakedown experimentation with coal 1 using pure water feed reagent and
(2) processing of coal 3 with acidified reagent during the RTU operational
phase (Experiment 03). Both groups of experiments were performed utilizing
all five stages of R-l rather than the three stages used in Experiment 01.
Thus, longer reaction times were obtained for similar slurry feed rates.
Operating parameters for low iron reagent experimentation are listed in
Table 10. Shakedown experiments 8 and 9 were performed with pure water feed
and coal derived iron and sulfate. Experiment 03 was performed with a nominal
1% w/w H2S04 feed solution. However, due to inadequate reagent mixing .in
tank T-8, acid segregation in the reagent solution resulted in a 2% w/w acid
feed reagent for experimental condition 03-03. The difference between 1% and
2% starting reagent acid concentration is not significant since acid addition
serves only to prevent iron oxide deposition from the reagent and possible sub-
sequent sulfate occlusion on the coal or blinding of the pyrite surfaces. The
pH of these solutions ranges from 1.0 to 1.2 and either pH is sufficiently low
to suppress precipitation of iron oxide from the reagent.
A summary of processed coal analyses from low iron reagent processing is
presented in Table 11. Replicate shakedown experiments 8 and 9 show an S
reduction of 57* from 0.68% w/w to approximately 0.27% w/w during 265°F proc-
essing. The sulfur forms analyses problem observed during Experiment 8
(i.e., the high SQ and low Sp) is most likely attributable to inefficient
pyrite extraction; the expected SQ based on ash removal is 0.63% w/w which
would indicate a true Sp value of 0.26% w/w. Experimentation performed with
acidified reagent at 230°F indicates S removals from 0.64% w/w to approxi-
mately 0.16% w/w or 75%.
82
-------
TABLE 10. OPERATING PARAMETERS FOR LOW IRON REAGENT LEACH EXPERIMENTATION
00
CO
Experiment
No.
8
(Shakedown)
9
(Shakedown)
03-01
03-02
03-03
Run time Reagent R-1 R-l 02 Coal Reagent 02 Recirc. Residence time*
(day/hour) H2S04 TemP Pressure Pressure Feed Feed Feed Rates T-2 R-l
Concentration Rate Rate Rate
Start Stop (* w/w) (°F) (psig) (psia) (Ib/hr) (gpm) (SCFH) (gpm) (hrs) (hrs)
238/12:00
249/11:15
6/17:04
9/12:00
10/17:07
238/22:41 0
249/22:31 0
7/03:00 1
10/03:20 1
11/07:28 2
265 50 26 198 1.0 94 23** 1.9 9.5
265 50 26 198 1.0 103 16** 1.9 9.5
222 31 28 298 1.7 94 20 1.1 5.7
232 41 34 199 1.2 88 20 1.6 8.2
234 41 33 199 1.2 150 20 1.6 8.2
* Slurry residence time is computed as the ratio of reactor volume-to-volumetric slurry flow rate.
** Redrculation piping used during shakedown processing was 1.25 inch sch. 80 pipe while that used during Phase III
processing was 0.75 inch sch. 40 pipe.
-------
TABLE 11. SUMMARY OF PROCESSED COAL ANALYSES OBTAINED DURING
RTU EXPERIMENTATION WITH LOW IRON REAGENT
00
Exp. No. No. of* Processing
Samples Temp.,
(°F)
Starting 3P; 14F
Coal 1
8 IP; 5F 265
(Shakedown}
9 IP; 5F 265
(Shakedown)
Starting 10P; 22F
Coal 3
03-01** 4P; 8F 222
03-02** 4P; 11F 232
03-03** 4P; 9F 234
Coal Composition, % w/w
Ash
16.11
±0.261
13.83
12.33
14.32
±0.469
12.97
±0.894
12.02
±0.636
12.51
±0.953
Heat
Content,
Btu/lb
12508
± 77
12985
13213
12907
± 58
13258
± 162
13388
± 91
13265
± 155
Total
Sulfur,
St
1.73
±0.045
1.03
1.07
1.49
±0.081
0.68
±0.044
0.78
±0.065
0.75
±0.033
(Except Heat Content), Dry Basis
Pyri ti c
Sulfur,
SP
0.68
±0.071
0.17?
±0.061
0.29
±0.021
0.64
±0.051
0.18
±0.064
0.17
±0.042
0.14
±0.031
Sulfate
Sulfur,
Ss
0.43
±0.071
0.14
±0.025
0.15
±0.007
0.31
±0.027
0.01
±0.010
0.01
±0.006
0.02
±0.007
Organic Iron,
Sulfur,
So Fe
0.62 1.33
±0.010
0.72? 0.88
0.63 0.71
0.54 1.02
±0.071
0.49 0.23
±0.029
0.60 0.26
±0.049
0.59 0.25
±0.060
**
The number of short proximate analyses (ash, heat content, total sulfur and coal iron) and sulfur
forms analyses (sulfate and pyritic sulfur) will be designated by xP and yF, respectively, where
x and y are the number of analyses included in the indicated average. The number of coal samples
analyzed is y minus x.
These experiments were performed with 1 to 2% w/w H2SO» reagent.
-------
These results are somewhat surprising since higher S removals are
expected at higher processing temperatures for coals from the same mine having
nearly identical starting S contents and processed under similar oxygen par-
tial pressures and throughputs. The observed difference could be due to sev-
eral things, including: (1) differences between the pyrite particle sizes
and distributions in coal Nos, 1 and 3; (2) differences in oxygen availability
and (3) the effect of reagent acidification. That coal Nos. 1 and 3 differ
with respect to pyrite particle size and distribution appears feasible since
they are known to differ in other respects (i.e. , ash and organic sulfur con-
tents). The higher ash coal No. 1 may contain pyrite which is difficult to
remove owing to its low specific surface area. Oxygen availability is impor-
tant to pyrite leach rates obtained during low iron reagent processing.
Because the bulk reagent contains very little iron, continuous and efficient
regeneration of reagent ferric ion is necessary to maintain high pyrite leach
rates. The recirculation/oxygenation loop used during shakedown was 1.25-inch
schedule 80 pipe while that used during Phase III operation was 0.75-inch
schedule 40 pipe. Thus, flow velocities in the regenerator during shakedown
were approximately 40% of those obtained during Phase III operation (the asso-
ciated shakedown Reynolds number was approximately 66% of that in Phase III).
Since gas-reagent mixing efficiency, and therefore oxygenation, may relate to
recirculation flow velocities, slurry oxygenation obtained during shakedown
experimentation may have been less efficient than that obtained during Phase III
operation.
Thus, the specific cause of differences in pyrite leach rates obtained
during low iron processing of coal Nos. 1 and 3 cannot be accurately defined
on the basis of available data. Further experimentation, including leaching
of coal Nos. 1 and 3 with high iron reagent, would be required before confi-
dent conclusions can be drawn.
Differences in product coal S contents indicated in Table 11 are due to
the fact that samples from Experiments 8 and 9 were not washed in the labora-
tory prior to analysis. However, bench scale experimentation performed to
date with acidified and non-acidified water reagents has not indicated an
appreciable difference in processed coal S contents. It, therefore, appears
that further washing would effect essentially complete sulfate removal from
85
-------
Experiments 8 and 9 as was the case for Experiments 03-01 to 03-03. Note that
all processed coals from Experiment 03-03 contain less than 1.2 pounds S02
per 10 Btu.
Data from low iron reagent processing may be compared with 4% iron reagent
processing data obtained at similar temperatures and 02 partial pressures.
Experiments 01-10 and 01-01 were performed at 270°F and 230 F, respectively,
with 02 pressures of approximately 23 psia. A residence time of 5.1 hours at
elevated temperature and pressure was used for both Experiment 01 conditions.
A comparison of data from Experiment 01-10 (Table 9) and Experiments 8 and 9
indicates that similar AS values were obtained from both groups of experiments
although only half as much residence time was required with iron reagent.
Thus, higher pyrite leach rates were obtained during processing with iron
reagent. On the other hand, extrapolation of the results from Experiment 01-01
(residence time of 5.1 hours) to the 5.7 and 8.2 hours of residence time uti-
lized in Experiments 03-01 through 03-03 yields essentially the same results,
which were obtained in Experiment 03. That is, assuming that the effective
pore reagent Y value for low iron reagent processing experiments was similar
to that of Experiment 01-01, processed coal analyses from Experiment 03 appear
to indicate pyrite leach rates similar to those obtained during the higher
iron reagent Experiment 01-01.
As mentioned previously, Coal No. 1 reacted at a lower rate than did
Coal No. 3 under low iron reagent processing conditions. The comparison of
Experiments 8 and 9 with Experiment 01-10 indicates that Coal No. 1 also
reacted slower than did Coal No. 2 which was processed with a higher iron con-
tent reagent. The comparison of Experiment 03 with Experiment 01-01 indicates
that Coal No. 3 processed with acidified low iron reagent reacted at a rate
comparable to that obtained from higher iron reagent processing of Coal No. 2.
The conclusion drawn from these observations depends on the assumed reaction
rate of Coal No. 2 relative to Coal Nos. 1 and 3. Since all three coals are
products of the same mine and physical cleaning facility and were processed
at the same top-size, it seems reasonable to expect that these coals would
react at similar rates under similar processing conditions. If this expecta-
tion is correct then the lower rates obtained during Experiments 8 and 9 are
most probably due to inefficient oxygenation. Further, Experiments 01-01 and
03 indicate that reaction rates measured in the RTU were not affected by
reagent total iron content up to a concentration of 4.5% w/w.
86
-------
Coal processed in the RTU with acidified low iron reagent was found to
contain approximately 0.1% w/w less $s after the control lab washing stages
than did coals processed with 4% iron reagent. This observation is not con-
sidered to be reflective of differences in process chemistry but, rather,
indicates the necessity of effective pore solution displacement during the
washing of coals processed with high iron reagent.
6.2.2.2 Reagent Regeneration Data--
Each stage of the R-l reactor is provided with a slurry recirculation/
oxygenation loop which promotes reagent regeneration concurrently with pyrite
leaching. During RTU operation the oxygen feed was proportionated among the
reaction stages according to estimated regeneration requirements occurring at
each stage. An excess of oxygen was fed to each stage of the RTU which cor-
responded to between 4 and 7 times the actual stoichiometric requirement.
Bench scale experimentation has shown that under such operation, the
reagent regeneration rate may be expressed as follows:
where
rR is the moles of ferric ion regenerated per unit time,
P02 is the partial pressure of oxygen in atmospheres,
Fe+2 is the ferrous ion concentration in moles per liter,
KR is the rate constant, a function of temperature only, in liters/
mole-atm-unit time,
and
' KR = AR exp (-ER/RT) (4)
with
AR = 6.7 x 10 liters/mole-atm-min., and
ER = 13.2 Kcal/mole.
87
-------
The bench scale determined ER value should apply to RTU regeneration since
reagent, unlike coal particles, is not prone to segregation and ideal flow
through the staged reactor system should be closely approximated. On the other
hand, values of AD computed from RTU data may differ from that determined at
K
bench scale since the extent or efficiency of regeneration may be different.
Hence, the ER value of 13.2 Kcal/mole was assumed to be correct for RTU proc-
essing and the value of AD (RTU) was computed from process data.
K
Computation of reagent regeneration rates in R-l requires data on inlet
and outlet reagent total iron concentrations and Y values, steady state Y data
+2 '
for each reactor stage, (L pressure data, and Fe generation rates. Of these
+2
required data, only the Fe generation rates are not directly measured during
+2
processing. The major source of Fe generation is the pyrite leach reaction
+2
which was discussed in Section 6.2.2.1. A secondary source of Fe is reac-
tion between ferric ion and the coal matrix. Although the extent of this oxi-
dation reaction is too small to be detected by heat content analysis of the
+2
coal, sufficient Fe is produced to warrant accounting during regeneration
+2
rate computation. The quantity of Fe generated above that expected from
+2
pyrite leaching is termed excess Fe .
Bench scale data from processing 14 mesh x 0 mine cleaned Martinka coal
with 5% w/w iron reagent at 212°F are presented in Table 12 (Reference 8).
+2
These data indicate an excess Fe generation rate of approximately 0.005
+2
weights of Fe per weight of coal per hour during the first 5 hours of proc-
essing. Subsequent processing resulted in little or no additional excess Fe+2
generation. These data provided the basis for estimating the rate of excess
+2
Fe generation during RTU coal processing.
Equations (1) through (4) were used with steady state data from Experi-
ment 01 to determine regeneration efficiency in the RTU. Since attainment of
steady state with respect to reagent Y required long processing times for con-
ditions of Experiment 01, not all conditions of this experiment provided true
steady state reagent Y data. Conditions 02, 03, 06, and 10 all provided steady
state or near steady state Y data while conditions 01, 05, 07, 09, and 11
yielded only transient data. Experiments 01-04 and 01-08 were short term
tests involving pulse response,measurements only in the second stage of reac-
tion and cannot be used for overall regeneration rate evaluation.
-------
TABLE 12. EXCESS Fe+2 GENERATION DATA FROM PROCESSING 14 MESH x 0
MINE CLEANED MARTINKA COAL WITH 5% w/w IRON REAGENT
=., _____ =
+2
Excess Fe Generation
Exp.
B.S.
B.S.
B.S.
No.
35 '
27
28
Temp.
(°F)
212
212
212
Process
Time
(hrs)
5
24
48
+2
weight Fe
weight coal
0.0258
0.0197
0.0294
+2
weight Fe
weight coal-hr
0.0052
0.0008
0.0006
It should be noted that rapid attainment of steady state processed coal
S contents relative to times required to obtain steady state reagent Y values
is indicative of the fact that product coal analyses are rather insensitive to
small changes in reagent Y. This is particularly true when coal S contents
are reduced to 0.20% w/w or less.
Examples of transient and near steady state reagent Y data are illus-
trated in Figure 20. Reagent Y data from Experiments 01-01 and 01-02 are
presented in the Figure as a function of process time. Data from Experiment
01-01 indicate that the reagent Y is decreasing with time at a substantial rate
in all stages. Increasing the partial pressure of oxygen from 22 psia to
49 psia (Experiment 01-02) is seen to have essentially stabilized the reagent
Y value in each stage. Although data from Experiment 01-02 indicate that the
first and second stage Y values may have still been changing at the test con-
clusion, the rate of change appears to be rather small. Note that the stage
two Y value should not exceed that of stage three and that the first stage Y
+2
value should not exceed that of stage two since Fe generation due to S
removal decreases with each successive stage. Thus, the true steady state
Y values of the first two stages probably do not differ from the measured
values by more than about 0.02 which will not substantially affect regenera-
tion rate estimation.
Regeneration rate estimates computed from Experiments 01-02, 01-03, 01-06,
and 01-10 are presented in Table 13. Reaction temperature, oxygen partial
pressure and oxygen feed rate data are listed in columns two, three, and four.
89
-------
o
o
VO
en
o
CM
•
o
00
00
CO
o
o
•
o
I-
5 vo
ii I O
C£
CM
•
O
CO
o
O 1ST STAGE R-l
0 2ND STAGE R-l
A 3RD STAGE R-l
EXP.
01-01
EXP. 01-02
o
to
22 psia
po.
49 p
la
17.00 18.00 19.00 20.00 21.00 22.00 23.00 24.00 01.00 02.00 03.00 04.00 05.00 06.00 07.00 08.00 09.00 10.00 11.00
CLOCK TIME (HOURS)
Figure 20. Reagent Y Data From RTU Processing Experiments 01-01 And 01-02
-------
TABLE 13. RTU REAGENT REGENERATION DATA
Exp. Temp. Pn
No. (°F) °2
(psia)
01-02 231 49
01-03 251 39
01-06 251 52
01-10 270 23
02 XS Fe2
Feed Generation
Rate / wt. Fe2 \
(SCFH) Vrft. coal-hr;
57 0.0
0.005 (bench scale)
0.010
54 0.0
0.005 (bench scale)
0.010
123 0.0
0.005 (bench scale)
0.010
56 0.0
0.005 (bench scale)
0.010
0.020
ARxlO"5
, liter _>
Stage
1
1.33
2.70
4.08
0.72
1.50
2.29
0.58
1.16
1.72
0.01
0.60
1.18
2.34
vmole-atm-min
Stage Stage
2
1.16
2.51
3.85
0.66
1.44
2.21
0.79
1.47
2.14
0.41
0.95
1.47
2.52
3
0.78
2.09
3.70
1.10
2.18
3.25
1.04
2.06
3.07
0.29
0.81
1.33
2.35
I
Average
1.09
2.43
3.88
0.83
1.71
2.58
0.80
1.56
2.31
0.24
0.79
1.33
2.40
% of
Bench
Scale
AR
16
36
57
12
25
39
12
23
35
4
12
20
36
Average KpxlO
, liter V ,
vmole-atm-min'
3.28
7.32
11.68
4.07
8.39
12.65
3.92
7.65
11.33
1.82
6.01
10.11
18.24
-------
+2
The value of AR was estimated for each reaction stage based on an excess Fe
generation rate of from zero to four times the value measured at bench scale,
and an ER value of 13.2 Kcal/mole; the assumed excess Fe generation and cor-
responding AR values are listed in columns five through nine. The average AR
values obtained in the RTU are listed in column ten as a percentage of the
bench scale AD value. The average measured KR values from this testing are
R "
listed for reference in the last column.
+2
Rate constant data presented in Table 13 for a fixed level of excess Fe
generation indicate little temperature dependence (Experiments 01-02, 01-03,
and 01-06) or show a rate decrease with increasing temperature (Experiments
01-02, 01-03, and 01-06 compared to Experiment 01-10). Since the temperature
effect on reagent regeneration is known with confidence, these data must indi-
+?
cate that excess Fe generation is not a constant, but rather a function of
temperature which increases with increasing temperature. As such, higher
+2
excess Fe generation rates and, therefore, higher KD values apply for higher
+2
processing temperatures. If the bench scale excess Fe generation rate is
assumed to be correct up to 230°F, then the excess rate must double with each
additional 20°F increment for the data to support an ED value of 13.2 Kcal/
o
mole. Twice the bench scale excess must apply at 250 F and four times the
bench scale excess must apply at 270°F. This observation is in agreement with
bench scale data indicating that excess ferric ion consumption increases by at
least a factor of two between 212°F and 250°F when processing 14 mesh x 0 mine
cleaned Martinka coal ^ '. Thus, these data appear to indicate that regenera-
tion in the RTU proceeded with an AR (RTU) of approximately 2.4 x 105 liters/
mole-atm-min or at 37% of the rate obtained at bench scale. This represents
an estimate of the minimum average regeneration rate obtained in the RTU since
excess Fe+ generation rates at 230°F may actually be higher than those meas-
ured at 212°F.
These data also indicate that increasing the oxygen throughput from a
four-fold to a fourteen-fold stoichiometric excess did not improve regenera-
tion rates (compare Experiments 01-03 and 01-06).
If indeed the regeneration efficiency in the RTU is 37% with respect to
bench scale regeneration, then a comparison of the two regeneration systems is
appropriate. The primary parameters of concern are the oxygen-reagent blending
92
-------
path length, Reynolds number and gas-reagent contact frequency (stage turn-
around time through the regenerator loop). These parameters compare as
follows:
• Both the RTU and bench scale systems have gas-slurry mixing
path lengths on the order of 10 feet. Both systems also
have bends up to about 90° at some point. The bench unit,
however, has a valved constriction in the mixing path which
may double the fluid velocity at the valve. The RTU unit
contains no flow constrictions.
• The RTU system circulates slurry at a rate of 20 gpm through
0.82-inch I.D. pipe while the bench unit circulates slurry at
a rate of 1 gpm through a 0.75-inch I.D. tube. A ratio of the
Reynolds numbers obtained in the RTU and bench unit mixing
regions may be computed to avoid estimating viscosities of
slurries at the reaction temperature. The ratio Re(RTU)/
Re(Bench) based on an estimated 50% constriction in the bench
scale tubing is approximately nine.
• Each stage of the RTU contained approximately 150 gallons of
slurry which was circulated through the regeneration loop at
a rate of 20 gpm giving a slurry turnaround time of approxi-
mately 7.5 minutes. The bench scale reactor generally con-
tained about 1.6 gallons which was circulated at 1 gpm for a
turnaround time of approximately 1.5 minutes.
From these facts, it appears that oxygen-slurry mixing was at least as
vigorous in the RTU system as it was in the bench scale apparatus. Any addi-
tional gas-slurry blending due to the restriction in the bench circulation
line should be compensated for, at least in part, by the higher overall
Reynolds number obtained in the RTU circulation lines. Hence, the principal
difference between the two systems resides in their respective stage turnaround
times through slurry circulation/oxygenation loops; bench scale reactor turn-
around times through the oxygenation loop were only one-fifth as long as RTU
turnaround times.
That the regeneration efficiency should relate to slurry turnaround time
through the regeneration loop appears reasonable if the oxygen bubbles coalesce
and disentrain from the slurry at a reasonably rapid rate (on the order of
minutes) after slurry reentry into the reactor. The regeneration rate expres-
sion (Equation 3) is expressed in terms of oxygen partial pressure but, more
fundamentally, the regeneration rate is proportional to the concentration of
93
-------
dissolved oxygen. The concentration of dissolved oxygen depends on the rate
of pyrite leaching and the rate of oxygen dissolution into the reagent. The
latter depends on the partial pressure of oxygen and the gas-reagent contact
area. Since, under conditions of rapid oxygen disentrainment the gas-reagent
contact area outside the circulation loop decreases rapidly with time, regen-
eration efficiency would be expected to increase with decreasing stage turn-
around time through the recirculation loop. Assuming that this hypothesis is
correct, it would appear that to increase regeneration rates in the RTU would
require an increase in the slurry recirculation rate (though not necessarily
the Reynolds number) and/or the partial pressure of oxygen.
6.2.2.3 Elemental Sulfur Recovery-
All coal samples drawn from the RTU were extracted with toluene to recover
product elemental sulfur (S ) prior to sulfur forms analyses. The quantity of
S generated during processing is determined by analyses of toluene extracts
from filter belt coal samples. Additionally, S recovery has been determined
for a limited number of samples from the primary reactor.
Results of S extractions performed during RTU shakedown and Experiments
01 and 03 are presented in Table 14. Included in this table are the quantity
of residue composed of sulfur and tar, sulfur content of the residue (based on
one analysis per sample) and S content of the coal based on extraction data.
Results from double toluene extractions are indicated by two tar and sulfur
content values. Sample coal analysis data used in computation of the product
ratio S /S are also presented. The last column of the table lists the product
sulfur ratio Ss/Sn which is estimated as (AS - S )/S . These data were obtained
identically for all experimentation with the exception that twice as much
toluene extract was distilled to generate the sulfur-tar residue for Experi-
ment 03 than was used for the preceding experimentation. Also, balance pre-
cision was increased from 0.005 gram to 0.00005 gram for Experiment 03. The
resultant improved precision is evidenced by the relative S /S data scatter
obtained before and after initiation of Experiment 03, namely ±1.66 compared
to ±0.71. Thus, the substantial scatter in S /S data obtained prior to
Experiment 03 is principally due to extraction residue weighing inaccuracies.
That these inaccuracies were essentially random is verified by the fact that
the mean Ss/Sn ratio from combined shakedown and Experiment 01 testing and from
94
-------
TABLE 14. ELEMENTAL SULFUR PRODUCT GENERATION DATA FROM RTU PROCESSING
VO
en
Experiment # Sample ID
8 (Shakedown) SI 1900
SI 2000
SI 2200
9 (Shakedown) SI 1800
SI 2000
SI 2101
SI 2331
01-01 M7 0130
M7 2105
SI 2345
SI 0157
01-02 SI 0745
SI 0942
01-03 M7 2103
SI 2125
01-05 M7 1940
SI 1910
SI 1950
01-06 SI 1050
01-07 SI 0144
SI 0235
Sulfur + Tar
Residue
(% w/w)
0.67
0.92
0.76
0.76
0.88
0.97
0.73
0.41
0.32
0.52
0.31
0.55
0.55
0.78
0.65
0.75
0.45
0.52
0.32
0.17
0.15
Sulfur
Content
11.31
13.28
25.62
12.61
12.95
9.11
22.33
26.2
44.8
39.2
41.1
54.8
56.8
42.6
11.9
57.7
44.7
48.8
50.9
31.1
48.6
Elemental
Sulfur, S
(% w/w) "
0.08
0.12
0.19
0.10
0.11
0.09
0.16
0.11
0.14
0.20
0.13
0.30
0.31
0.33
0.08
0.43
0.20
0.25
0.16
0.05
0.07
Coal Analysis,
S°
0.68
0.68
0.68
0.68
0.68
0.68
0.68
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
Sl
0.21
0.17
0.07
0.32
0.32
0.23
0.33
0.24
0.21
0.23
0.29
0.33
0.22
0.30
0.17
0.13
0.17
0.16
0.13
0.20
0.33
% w/w
ASP
0.47
0.51
0.61
0.36
0.36
0.44
0.35
0.34
0.37
0.35
0.29
0.25
0.36
0.28
0.41
0.45
0.36
0.37
0.45
0.38
0.25
*
VSn
5.18
3.19
2.12
2.60
2.27
4.00
1.15
2.09
1.64
0.75
1.23
0
0.13
0
4.13
0.05
0.80
0.48
1.81
6.60
2.57
- Continued -
-------
TABLE 14. (Continued)
CTv
Experiment 1 Sample ID
01-09 M8
SI
01-10 SI
01-11 SI
SI
03-01 SI
SI
si
03-02 SI
SI
SI
SI
03-03 ' SI
SI
SI
1445
1255
0108
0224
0301
2258
2356
0200
2213
0015
0119
0335
0431
0527
0727
Sulfur + Tar
Residue
(% w/w)
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.33
0.36
0.38
0.23
0.35
47, 0
48, 0
60, 0
42, 0
0.45
45, 0
52, 0
50, 0
47, 0
38, 0
.23
.13
.22
.21
.21
.18
.14
.22
.31
Sulfur
Content
33
28
36
51
46
26.6
26.2
20.1
41.4
41
36.8
22.5
36.0
33.0
27.5
.0
.1
.2
.5
.9
, 0.8
, 3.7
, 1.1
, 6.3
.3
, 4.1
, 3.8
, 7.7
, 4.8
, 4.6
Elemental
Sulfur, S
(% w/w) n
0.11
0.10
0.14
0.13
0.16
0.13
0.13
0.13
0.19
0.19
0.18
0.13
0.19
0.16
0.12
Coal Analysis , 3
S°P
0.58
0.58
0.58
0.58
0.58
0.64
0.64
0.64
0.64
0.64
0.64
0.64
0.64
0.64
0.64
SJ
0.11
0.13
0.08
0.07
0.14
VSn
0.25
0.16
0.17
0.19
0.22
0.19
0.15
0.17
0.15
0.13
} W/W
ASP
0.47
0.45
0.50
0.51
0.44
Ave =
0.39
0.48
0.47
0.45
0.42
0.45
0.49
0.47
0.49
0.51
*
VSn
3.27
3.5
2.57
3.25
1.75
2.2 ± 1.66
2.00
2.69
2.61
1.37
1.21
1.50
2.77
1.47
2.06
3.25
VSn Ave * 2J * °'71
Ss/Sn Cum Ave = 2.2 t 1.45
The ratio Sg/Sn
is computed as (AS - Sn)/Sn,
-------
Experiment 03 are 2.2 ±1.66 and 2.1 ±0.71, respectively; these values are sta-
tistically indistinguishable.
Collectively, the data presented in Table 14 indicate a product S /S
^ s n
ratio of 2.2 which may be compared to the previously determined S /S ratio
/c 11 \ <; n
of 1.5^ '. The difference between these two values may be the result of
incomplete $n recovery from the processed coal (leaving 0.04% S on the coal
or losing it during recovery would suffice for this experimentation), loss of
S during the recovery process (i.e., during vacuum drying), or a systematic
error in any of the numerous analyses and measurements used in computation of
the Ss/Sn ratio such as original and final S , residue weight, purity, etc.
Therefore, the higher product Ss/$n ratio obtained from S recovery data is
not considered to be reflective of process chemistry but of the increasing
difficulty in obtaining an S balance from low S coal processing experiments.
A comparison of experiments performed with iron reagent (Experiment 01)
and experiments performed with initially iron free reagent (Experiments 8, 9,
and 03) indicates that reagent iron concentration does not effect the product
s^/Sm ratio. This finding is in agreement with previously obtained bench
(f\\
scale data* '.
Most coal samples obtained during Experiment 03 were treated with two
stages of toluene extraction rather than the nominal single stage extraction.
Data from this experiment shows that 94 percent of the extracted $n was
recovered in the first stage of extraction. Thus, approximately 0.01% w/w Sn
was recovered during the second toluene extraction of processed mine cleaned
Martinka coal.
6.2.2.4 Process Effects on Coal Heat Content-
Mine cleaned Martinka feed coals and processed coals obtained from the
RTU may be compared on the basis of mineral matter free (mmf) heat content
to determine effects of the Meyers Process on the coal matrix. An mmf heat
content was computed based on starting coal analyses and average processed
coal analyses for each experiment using the Parr formula (ASTM designation
D388-66). Results of these computations are presented in Table 15. - These
data indicate a mean heat content increase of 352 Btu per pound based on dry
coal analyses; this corresponds to a 2.7% increase in coal heating value. On
97
-------
TABLE 15. SUMMARY OF HEAT CONTENT CHANGES IN RTU PROCESSED COALS
•£>
CO
Experiment
No.
8 (Shakedown)
9 (Shakedown)
01-01
01-02
01-03
01-05
01-06
01-07
01-09
01-10
oi-n
03-01
03-02
03-03
Average Heat
Content Change
Heat Content,
Dry Basis, Btu/lb
Initial Final
12508
12508
12923
12923
12923
12923
12923
12923
12923
12923
12923
12907
12907
12907
12985
13213
13397
13300
13268
12888
13163
13123
13487
13158
13071
13258
13388
13265
Changes
Btu/lb Percent
477
705
474
377
345
-35
240
200
566
235
148
351
481
358
352
±188
3.8
5.6
3.7
2.9
2.7
-0.3
1.9
1.5
4.4
1.8
1 .1
2.7
3.7
2.8
2.7
±1.5
Heat Content, Dry Mineral
Matter Free Basis, Btu/lb
Initial Final
15172
15172
15283
15283
15283
15283
15283
15283
15283
15283
15283
15293
15293
15293
15279
15260
15459
15508
15396
15185
15264
15337
15255
15426
15367
15420
15393
15342
Changes
Btu/lb Percent
107
88
176
225
113
-98
-19
54
-28
143
84
127
100
49
80
±85
0.7
0.6
1.1
1 .5
0.7
-0.6
-0.1
0.3
0.2
0.9
0.5
0.8
0.7
0.3
0.5
±0.5
-------
a dry mmf basis a mean heat content increase of 80 Btu per pound is indicated
(0.5% of the mmf heat content). The observed rnrnf heat content increase is not
believed to be meaningful in the sense of coal upgrading but, rather, is con-
sidered to reflect inaccuracies associated with analyses and with the Parr
formula itself. However, on the basis of these data, it appears reasonable
to conclude that no measurable degradation of the coal matrix occurred
during RTU L-R processing at temperatures of 230°F and 270°F with reagent
iron concentrations of 0-4.5% w/w.
These data are in agreement with bench scale results indicating that the
Meyers Process has little or no effect on the organic matrix of Appalachian
bituminous coals. However, bench scale L-R experimentation performed with
14 mesh x 0 mine cleaned Martinka coal resulted in an mmf heat content loss
of approximately 2% during 250 F processing for six hours or more. Thus,
while no matrix oxidation was observed during L-R processing in the RTU, slight
matrix oxidation was observed during bench scale L-R operation.
This discrepancy may be associated with coal weathering occurring during
the 8-month period between obtaining a coal sample for bench scale processing
and initiation of RTU coal processing. The unweathered coal (as indicated by
the lack of sulfate sulfur) utilized during bench scale processing had an mmf
heat content of 15,448 Btu per pound while that of the RTU feed coal was
15,249 Btu per pound (average of all three RTU feed coals). The 239 Btu per
pound difference between bench scale and RTU feed coals represents 1.5% of the
unweathered coals' mmf heat content. Thus, it is feasible that 1-2% of the
coal organic matrix was oxidized during weathering and that this portion of the
matrix is readily oxidizable by any oxygen bearing process. Data from the RTU
would then indicate that no measurable coal oxidation can be further incurred
during short periods of time under moderate conditions. It should be noted
that if the preceding assumptions are correct, oxidation of 1-2% of the mmf
coal is inevitable and will occur either during coal storage at the respective
point of utilization or during coal processing.
6.2.3 Slurry Sampling
Slurry samples were drawn from each stage of T-2 and R-l during RTU
operation for the purpose of obtaining reagent Y data and coal analyses.
Sampling ports in the RTU are located at two positions: (1) in the vessel
wall below the slurry level and (2) in the slurry recirculation loop (R-l
only). Wall samples are taken by allowing slurry to pass through a ball
99
-------
valve into an evacuated sample bomb. One bomb is used to clear out stagnant
coal which may accumulate in the sampling lines. The contents of this bomb
are discarded and a second sample is immediately drawn for analysis. Loop
samples are obtained from R-l stages by full stream diversion of circulating
slurry through a sampling tube which can be isolated and subsequently drained
of slurry. Each sample drawn contained approximately one liter of slurry.
Coal analysis data obtained from slurry samples have been evaluated in
conjunction with reagent Y data and the reaction model presented in Sec-
tion 6.2.2. This treatment is limited to results from Experiment 01 since a
pyrite leach rate expression is required. A plot of the difference between
computed and analyzed S values (AS ) versus sample location is presented in
Figure 21. The sampling locations are spaced along the abscissa according to
the theoretical time required to reach steady state compositions at each point.
Data are presented at two times for most stages since two slightly different
slurry residence times were utilized during Experiment 01. Data from wall
samples and recirculation loop samples are indicated in the Figure by open
and closed symbols, respectively.
Data from the three stages of mixing (T2-1, 2, and 3) indicate that coal
samples from the first two stages contain significantly more pyrite than com-
puted, while stage three samples are lower in pyrite than computed. Samples
from the first two stages of T-2 were found to consistently contain more
pyrite than the feed coal (1.4 ±0.66% w/w compared to a starting coal S con-
tent of 0.58% w/w). Stage three mixer samples can be shown to be non-
representative and low in S by using measured ferrous iron generation data.
Measured S values from R-l wall samples are also seen to be low with respect
to computed S values although the AS approximated analytical precision after
the first stage of pressurized processing. The limited data from R-l loop
samples are generally consistent with wall sampling data although high meas-
ured pyrite values were obtained from Rl-3 samples at 270°F,
These data suggest major segregation of the feed coal prior to reagent
reflux in the third stage of mixing. Bench scale experimentation performed
to determine causes of coal slurry foaming has indicated that reagent reflux
is key to effective coal wetting at atmospheric pressure ^ ^. Hence, high
pyrite coal appears to segregate in the lower portion of T-2 (the vicinity
100
-------
o
LU
M
1 ?
. 1
On
Or
Oc
0 4
0.3
•0.2
•0.1
0
0.1
0.2
0.3
0.4
o
a
a
A
o
o
A
0 °
A
o
A
O
a
o
8
£
A *
A @
o , * - 230°F TEST SA
MPLES (WALL, LOOP)
A, A - 250°F TEST SAMPLES (WALL, LOOP)
o,« - 270°F TEST SAMPLES (WALL, LOOP)
1
1
i
i
A
A
1
(100) (200) (300) (400 MINUTES)
T2-1 T2-2 T2-3 Rl-3 R1-* Rl-5
SAMPLE LOCATION (TIME)
Figure 21. Summary of Estimated Sampling Bias
-------
of the sampling ports) prior to effective coal wetting. The reverse segrega-
tion effect (drawing low pyrite samples rather than high pyrite samples)
observed in the third stage of mixing is not understood, but may be associated
with a froth flotation effect caused by reagent boiling and the injection of
steam. This phenomena is also observed in R-l where both oxygen and steam are
injected into the slurry.
The apparent reduction of sampling bias to the level of analytical pre-
cision during the last two stages of R-l processing probably relates to the
fact that the S contents of all size-gravity coal fractions are very low.
That is, non-representative sampling of a depyritized coal is not likely to
yield samples having substantially different S contents from the mean S
content of the whole coal.
Coal segregation indicated by AS data is verified by slurry solids con-
tent data which are summarized in Figure 22. Slurry solids data from each
sampling location are presented in the Figure as a function of AS . The nom-
inal solids content of the feed slurry was 33% for all conditions of Experi-
ment 01 and dilution down to approximately 29% was effected by the addition of
live steam for heating. Slurry solids data indicate that samples which were
high in S (i.e., samples from T2-1, T2-2, and the Rl-3 loop) also tended to
be solids rich with respect to the feed slurry. On the other hand, samples
which were low in S tended to have reasonable or slightly low slurry solids
contents. Thus, as would be expected, whether or not a slurry sample contains
coal which is representative of the whole coal is reflected in the solids con-
tent of the sample. If the sample slurry solids content is not representative
of the system as a whole, the coal contained in the sample probably is not
representative of the whole coal either.
Despite the fact that slurry samples drawn from the RTU were biased with
respect to S and solids contents, reasonably good sampling precision was
obtained. Replicate sampling generally yielded S agreement to within
±0.07% w/w, Ss agreement to within ±0.03% w/w and slurry solids agreement to
within ±3%. Such precision coupled with sampling bias data would tend to
indicate that only specific slurry streams or eddys are accessible through
specific sampling ports and that the composition of the accessible slurry did
not change significantly with time. Thus, improvement of slurry sampling most
102
-------
O
CO
0.2
o
-0 2
n A
3
3
** -n fi
•
e?
M _n 8
>- u-°
_i
<
— -10
CL
' -1 2
0
LJ
^ 11
=s -1 .*
<_)
_J
^ -1 6
Q.
-1 ft
-2 Q
U
AA
V
.
I
X A 9b
^^j
A^
L
i . '
D-
i—
LU
1-
O
0
OL
in
UJ
A
&
D
+ *•
+
+
+
ZERO SAMP
LING BIAS
X^ X
y
X
.
X
• - T2-1
+ - T2-2
A - T2-3
O - Rl-3
n - Rl-4
O - Rl-5
X - Rl-3
A - Rl-4
* - Rl-5
(WALL)
(WALL)
(WALL)
(LOOP)
(LOOP)
(LOOP)
10
30 40 50
SLURRY SOLIDS, PERCENT
60
70
80
Figure 22. Slurry Concentration Variation as a
Function of Sample Location
-------
likely requires more efficient slurry agitation within the various stages and/
or relocation of the sampling ports away from localized flow patterns (i.e.,
wall effects).
6.2.4 RTU Mass Balance Data
Flow rates of RTU process streams were monitored wherever possible t,o
enable mass balance determinations to be made around each of the major unit
operations, namely coal-reagent mixing, pyrite reaction and slurry filtration.
Primary flow measurement equipment utilized in the RTU to facilitate plant
operation and provide mass balance data were: (1) a weigh belt for metering
ground coal to the system, (2) magnetic flow meters for measuring liquid and
slurry flow rates, (3) rotameters for measuring steam, oxygen and wash water
flow rates and (4) target meters for measuring oxygen feed and exhaust flow
rates. Output from the primary flow measurement equipment was generally
dependent upon the input of one or more auxiliary devices which provide such
data as stream temperature, pressure, gas purity, etc. The majority of this
data was recorded at 3-minute intervals by the Doric data logger although
some data were recorded manually (i.e., steam and wash water feed rates) at
intervals commensurate with the data reduction needs.
With the exception of the magnetic flow meters, all primary flow measure-
ment devices used in the RTU measured mass flow directly. The weight belt
meters the coal feed utilizing a load cell which measures the weight of coal
per unit area of belt. Output from the weigh belt is transmitted as total coal
weight and also as a feed rate. Calibration can be obtained through direct
weighing of the belt effluent. Rotameters and target meters are essentially
momentum measurement devices which may be calibrated at either standard con-
ditions or operating conditions; the former requires the application of tem-
perature and pressure corrections. However, magnetic flow meters are true
velocity meters and, as such, yield accurate volumetric flow rates. In order
to convert from the magnetic flow meter volumetric flow to mass flow, a gen-
eral slurry density expression was formulated in terms of temperature, total
iron sulfate concentration, acid concentration and slurry solids content.
Considering the liquid reagent components as ideal solutions and neglecting
the specific volume of dissolved salt, the slurry density was expressed as:
104
-------
p " riQO - CA + 3.578 TFe) ^ A 1 |\ . » ~nM',Z ZTT
L 62.248 113366 J [l + 0-0003(T - 77) j T 84(l . Q>Qls)
where
3
p is the slurry density, Ib/ft ,
A is the sulfuric acid concentration of the reagent, wt. %,
TFe is the total iron concentration of the reagent, wt. %,
S is the slurry solids content, wt. %, and
T is the slurry temperature, °F.
The densities of water and sulfuric acid at 77°F (25°C) were taken to
be 62.25 and 113.97 Ib/ft3, respectively. A value of 84 Ib/ft3 was used for
the density of the coal.
This expression yields density estimates which agree with measured values
to within about 5% or less for the solutions used in the RTU. Hence, the pre-
cision of mass flow rates obtained from magnetic flow meters is approximately
5% rather than the 0.5-1% precision associated with volumetric flow readings.
Mass flow rates which are of importance to mass balance efforts, but are
not measurable by primary flow measurement equipment, include the steam flash
rate from R-l pressure let-down, slurry accumulation or depletion in T-2 and
R-l, and the flow rate of wet product coal cake from the S-l filter. These
quantities were either inferred from the other data normally obtained from
the RTU or obtained by special measurement during specified experimentation.
The quantity of steam flashed during R-l pressure let-down can be estimated
from the increase in the temperature of vent gas scrubber water in T-4. Thus,
assuming perfect scrubber efficiency, the mass flow rate of flash steam was
calculated from the flow rate of scrubber water and its temperature change
resulting from steam condensation. Slurry accumulation or depletion in T-2
and R-l is computed from slurry level data generated by sensors located in
the last stage of each vessel. Special measurements are required to obtain
the mass flow rate of wet coal cake from S-l. This involves collecting the
S-l solid effluent in a tared dumpster and determining weight increases as a
105
-------
function of time. This procedure is both time consuming and cumbersome and
was performed only for special mass balance experimentation. Thus, mass bal-
ances around S-l were not generally performed.
Computer programs which are presented in Volume II, Appendix A were
written to perform mass balances around T-2 and R-l. These programs account
for all process streams and utilize both Doric data and manually recorded
data. Data from Experiment 10 (a special shakedown mass balance test),
Experiment 01 and Experiment 03 were evaluated with the aid of these programs
to identify inconsistencies among the RTU flow measurement outputs.
A low temperature and pressure test (Experiment No. 10) was performed
during RTU shakedown to obtain a complete system mass balance and verify the
major flow measurement devices. This experiment was performed with a water-
coal slurry using flow rates .of 3 gpm and 500 Ib. of coal per hour. These
flow rates correspond to residence times of approximately 0.6 hour in the
mixer and 3.2 hours in the reactor (all five stages of R-l were utilized).
Mixer heating was utilized to maintain the slurry at 190°F (sufficiently hot
for proper P-l operation but below the solution boiling point). Slurry tem-
peratures in R-l were allowed to drop in order to minimize evaporative losses
during pressure let-down. The plant was operated for 10 hours under the nom-
inal conditions with the last 3 hours being considered to be at steady state.
The solids balance between the A-3 plant coal feed and S-l filter belt
output indicates that 160 Ib. of dry coal was "lost". This represents 11% of
the feed coal. However, filter belt flooding conditions prevailed during the
entire steady state operational period and solids removed during the belt wash
were discarded to T-9 rather than being returned to the filter cake as a fil-
ter belt wash water slurry per normal operation. Post-run observation indi-
cated that a "large build-up of coal" was found in the filter belt drip pan.
Thus, the 160 Ib. coal "loss" appears to be a reasonable and explainable
occurrence.
The overall liquid balance included units A-3, T-2, R-l, T-5 and S-l.
The vent steam could not be computed since the AT of T-4 scrubber water was
not measured. The vent steam from T-2 (operating at 190°F) was assumed to
be negligible while the steam from R-l (operating at ^160°^) was computed to
to be negligible on the basis of inlet and outlet reagent temperature values
106
-------
from the knockout drum V-l. This balance indicated an 830 Ib. liquid gain
(17% of input liquids). This result may be partially or totally due to the
following: (1) error in reagent filtrate flow rates, which, for this experi-
ment, were both FE-157 and FE-158, or (2) leakage from water flushed pump and
mixer seals. The possibility of incorrect reagent feed flow rates has been
discarded since a good mass balance was obtained around the 1-2 mixer. These
data combine to yield the total system mass balance which includes units A-3,
T-2, R-l, T-5 and S-l. The net result is a 662 Ib. gain which represents 11%
of the input material.
A mass balance around T-2 based on coal, reagent and steam feed rates and
slurry flow calibrations of FE-29 (slurry out) resulted in a total mass balance
of better than 93%. This result appeared to be reasonable since the accuracy
of each of the two magnetic flow meters is estimated to be approximately 5%.
A heat balance around T-2 essentially verified the validity of FI-16/18 steam
flow indicators. A mass balance around the combined units T-2 and R-l indi-
cated a loss of 1755 Ib. (28% of the input material). This result was found
to be caused by improper operation of the R-l outlet slurry flow meter (FE-60).
Subsequent experimentation revealed that signal response times of some mag-
netic flow meter components were sufficiently slow compared to slurry pulse
flow times (pulse lengths are generally less than two seconds) that entire
pulses were sometimes not detected. Attempts were made to resolve this
response problem throughout Experiment 01. However, satisfactory performance
of FE-60 was never obtained during pulse flow operation and the flow meter
was ultimately removed from service.
Data from the Experiment 10 mass balance test gave rise to the following
observations:
t Stream flow measurement equipment associated with T-2 was
operational and accurate. Subsequent experimentation indi-
cated that steam flow meters FI-16 through 18 were somewhat
undersized which necessitated frequent use of flow meter
bypass lines to maintain desired T-2 temperatures. Thus,
under normal operating conditions, steam feed rates to T-2
were not accurately known at all times.
• Magnetic flow meters could not be utilized under conditions
of pulse flow. Thus, the quantity of slurry leaving R-l and
entering S-l could not be directly measured with available
107
-------
flow measurement equipment. Although the mass flow rate of
slurry from R-l can in principle be estimated from the
increase in temperature of T-4 scrubber water, adequate
monitoring of scrubber water temperatures was never achieved
during RTU operation. As such, mass balances could not be
performed on either R-l or S-l with available data.
• An accurate accounting of system solids cannot be made unless
the S-l filter belt is operating properly. Filter belt opera-
tion under flooding or near flooding conditions necessitates
discard of the belt wash solution and its entrained coal
component.
Experimentation performed in the RTU did provide meaningful mass balance
data relating to the T-2 mixing unit. Results of the T-2 mass balances from
RTU experimentation are presented in Table 16. The first two columns of the
table list the experiment number and day. The second two columns list the
run times included in the mass balance. These times do not necessarily reflect
actual experimental run times since periods of time which included data noise
on one or more of the required data channels have been eliminated. The remain-
ing columns in Table 16 list mass balance data for each experiment. The accu-
mulation column lists material "gains" or "losses" associated with slurry level
changes within T-2. Slurry level data from Experiment 01 was recorded man-
ually and, therefore, in limited quantity; available data is insufficient to
permit computation of material accumulation. During Experiment 03 the slurry
level of T-2 was recorded by the Doric Data Logger and included in the mass
balance. The accumulation term is generally seen to be small as would be
expected since every effort was made to maintain a constant slurry level
throughout each experiment.
Data presented in Table 16 indicate that a good material balance was
obtained for the T-2 unit. An average recovery of 100% was obtained and most
experiments balanced to within 7%. Thus, the precision obtained in mass bal-
ancing the T-2 unit is comparable to that of the flow meters monitoring the
liquid flow streams. The trend from slightly low material recovery to slightly
high material recovery during the period of RTU operation may relate to instru-
ment calibration drift or an interaction between magnetic flow meter electrodes
and the leach reagent during the course of experimentation (i.e., scaling or
corrosion). However, iron concentration, in this case, should have no effect
on the magnetic flow meter calibration; that this is correct is evidenced by
108
-------
TABLE 16. T-2 MASS BALANCE DATA FROM RTU PROCESSING EXPERIMENTATION
o
<£>
Experiment
No.
10 (Shakedown)
01-01
01-02
01-03
01-04
01-06
01-05
01-07
01-08
01-09
01-10
01-11
03-01
03-02
03-02
Day
250
299
300
305
305
306
314
319
321
322
322
322
325
6
9
10
Mass Balance Time
Start
20:59
16:36
03:03
10:02
16:26
21:32
00:02
12:31
15:52
17:22
22:18
17:00
02:51
04:51
13:51
17:31
17:40
12:02
20:05
21:38
17:10
Stop
23:56
02:57
09:00
15:29
21:29
23:29
10:38
13:48
16:46
01:13
00:48
02:18
04:48
13:48
01:48
02:58
01:58
19:14
21:14
03:23
07:28
Coal
1468
3080
1770
1621
1506
580
3139
388
272
2327
745
2718
580
2663
3557
2820
2469
1430
228
1145
2852
Material
Reagent
4317
6195
3565
3089
2871
1108
5976
732
519
4519
1394
4974
1056
4797
6366
7739
7651
4174
666
3304
8313
In, Ibs.
Steam
150
288
165
151
140
54
293
40
27
238
70
129
27
124
106
131
569
500
80
399
1012
Total
5935
9563
5500
4861
4517
1742
9408
1160
818
7084
2209
7821
1663
7584
10029
10690
10689
6104
974
4848
12177
Material Accum. Total
out, Ibs out, Ibs
5609 -74 5535
9369
5185
4656
4252
1781
ti_
9369
5185
4656
4252
1781
j
9307 a 9307
1019 d 1019
807 3 807
6892 fe 6892
2215 < 2215
7840 § 7840
1663
7741
10492
8535
1663
7741
10492
8535
11267 279 11546
6533 -64 6469
1123 -20 1103
5160 55 5215
13350 62 13412
Sain,
Ibs
-400
-194
-315
-205
-265
39
-101
-141
- 11
-192
6
19
0
157
463
-2155
857
365
129
367
1235
Percent
Recovery
93
98
94
96
94
102
99
88
99
97
100
100
100
102
105
80
108
106
113
107
110
Material accumulated in the mixer.
-------
the fact that Experiment 10 was performed with pure water and still indicates
a slightly low material recovery as do nearly all early experiments.
An oxygen mass balance around R-l for an actual coal processing experi-
ment requires the following data: (1) oxygen flow rates into R-l, (2) V-l
vent gas flow rates, (3) the oxygen content of V-l vent gas, (4) Fe genera-
tion rates from both pyrite reaction and matrix oxidation, (5) starting and
• O
final Fe concentrations for each reaction stage, and (6) inlet and outlet
reagent Fe concentrations. Since the oxygen mass balance is reflective of
both the instrument and analytical precisions as well as either inferred or
extrapolated data (i.e., excess Fe+2 generation rates), an accurate oxygen
balance of a coal processing experiment is not readily obtained. However, an
effort was made to perform oxygen balances for all experimentation performed
in the RTU.
Estimates of the average oxygen consumption rates during processing were
+2
made based on pyrite removal data and excess Fe generation rates presented
in Section 6.2.2 and on reagent Y data. These values are compared to oxygen
consumption values indicated from oxygen flow data and R-l pressure data in
Table 17. Experimental conditions and run times are reiterated in the first
five columns of the table. Values of total input and vent oxygen obtained by
integrating the output of FE-61 and FE-44 are listed in columns six and seven.
Quantities of oxygen accumulated or depleted in R-l as indicated by starting
and final reactor pressures are listed in column eight. The difference between
oxygen input (column six) and the sum of vent oxygen and oxygen accumulated in
the reactor (columns seven and eight) is the oxygen consumption indicated by
RTU flow and pressure equipment and is listed in column nine. The average
measured oxygen consumption rate for each experiment (listed in column ten)
may be compared to the average oxygen consumption rates computed from AS and
+2 P
excess Fe data (listed in column eleven) to indicate the validity of acquired
mass balance data.
Data presented in Table 17 show little correlation between measured and
computed oxygen consumption values. The observed discrepancies do not appear
to be associated with the estimated excess Fe+2 generation rates since esti-
mation errors of this type should result in comparable differences between
measured and computed oxygen consumptions for experiments performed at similar
110
-------
TABLE 17. OXYGEN MASS BALANCE DATA OBTAINED DURING RTU EXPERIMENTATION
Experiment
No.
01-01
01-02
01-03
01-04
01-05
01-06
01-07
01-08
01-09
01-10
01-11
03-01
03-02
03-03
Temp.
°F
230
231
251
251
250
251
245
245
270
270
251
222
232
234
V
psia
22
49
39
39
25-39
52
10-53
50
49
23
52
28
34
33
0? Feed
Rate
SCFH
56
57
54
59
NA
123
59
57
56
56
128
94
88
150
Run
Time,
Mrs
9.7
5.9
11.5
1.9
10.2
10.7
9.5
1.9
8.5
12.3
11.5
10.1
15.3
14.3
Oxygen, SCF Average Q£
Rate,
In
541
395
21
113
1318
557
109
479
693
1478
909
1200
2254
Out
543
128
178
28
911
172
9
0
594
1133
953
1085
3189
Accum
27
2
11
-2
-3
36
-3
7
3
-4
-42
5
9
Consumed
-29
265
432
87
DATA NOT AVAILABLE
410
349
103
472
96
349
-2
no
-944
Measured
-3
38
38
43
38
37
51
55
8
30
0
7
-66
Consumption
SCFH
Computed
9
9
18
18
23
14
14
27
20
21
15
15
22
-------
temperatures. Results from Experiments 01-01 and 01-02, 01-09 and 01-10, and
03-02 and 03-03 do not indicate that this is the case.
Oxygen accumulation quantities are not sufficiently large to have a sig-
nificant impact on the measured consumption rates and therefore do not repre-
sent a major source of error. Oxygen feed rates measured with FE-61 should
be accurate since meter calibration was performed under conditions identical
to those used during plant operation and rough flow rate verification is pro-
vided by downstream rotameters used to proportion oxygen among the reactor
stages (Fl-62 through Fl-66). Thus, the most probable source of the observed
discrepancy appears to be the vent oxygen flow rate measurement.
The rate at which oxygen is vented from V-l is computed from the vent gas
oxygen content indicated by AE-171 and flow rate indicated by FE-44. The
oxygen analyzer provides continuous analysis of a dried slipstream of the V-l
discharge gas. Analyses from AE-171 cannot be wholly responsible for observed
discrepancies since to bring Experiments 01-02 through 01-09 and 01-11 into
agreement with computed consumption rates would require stream oxygen purities
greater than 100%. However, there are reasons why FE-44 may be singly respon-
sible for observed data discrepancies:
• Water saturated vent gas was cooled or in the process of cool-
ing when its flow rate was being measured. Thus, entrained
water droplets or condensate may have contacted the flow tar-
get and thereby altered the meter performance. This possi-
bility can be eliminated during future experimentation by
locating the flow meter in a heated region with a large cross
section to prevent condensation and to decrease the likeli-
hood of entrainment.
• Owing to the pressure regulator used in the RTU, reactor
venting was often pulsed rather than continuous. Pulses
occurring during venting were up to approximately 9 minutes
in duration and had amplitudes ranging to nearly 400 SCFH.
Since the data logger was recording at 3-minute intervals,
true pulse shapes are not known and the possibility exists
that entire pulses were not detected. It should be noted
however that venting during Experiments 01-03, 01-09, 01-10,
01-11 and 03-03 was reasonably smooth and pulse free. Yet,
data from these experiments do not correspond to computed
oxygen consumption estimates.
It therefore appears that an additional design effort is required in
conjunction with further elevated temperature processing experimentation
112
-------
performed in the absence of coal to resolve the problems associated with
obtaining a balance of oxygen in the RTU. Additional bench scale tests are
+2
also required to define excess Fe generation rates at temperatures above
212°F.
6.3 DATA ANALYSIS CONCLUSIONS
1. Three distinct coal samples (coal Nos. 1, 2 and 3) were sup-
plied by American Electric and Power for coal processing
experimentation in the RTU. These coals differed with respect
to coal ash, heat content, total sulfur, and sulfur forms
analyses.
2. RTU processing times required to attain steady state operation
with respect to reagent Y can be substantially larger than
would be estimated by this ratio of reactor volume-to-volu-
metric slurry flow rate. However, since product coal Sp con-
tents are relatively insensitive to small changes in reagent Y
(particularly when Sp is reduced to 0.2% w/w or less), a plug
flow type estimate of time required to achieve steady state
appears to be effective with respect to product coal Sp
contents.
3. All RTU processing experiments performed with 14 mesh top-size
mine cleaned Martinka coal and acidified iron reagent at
temperatures above 230°F yielded product coals having Sp con-
tents of 0.13% w/w to 0.19% w/w after nominal residence times
of 5 to 6 hours. Similar processing at 230°F yielded product
coals with Sp contents of 0.24% w/w.
4. RTU coal processing experiments performed with acidified iron
sulfate reagent reduced coal sulfate sulfur from 0.34% w/w to
approximately 0.12% w/w after application of a preliminary
filter belt wash followed by the standard laboratory washing
and work-up. Coal processing with acidified iron free reagent
reduced coal sulfate sulfur from 0.31% w/w to approximately
0.02% w/w after a filter belt wash and standard laboratory
work-up.
5. Little or no excess ash removal was observed during RTU proc-
essing of mine cleaned Martinka coal. Ash depletion from the
coal generally corresponded directly to the quantity of pyrite
and sulfate removed during processing.
6. Coal processing data obtained from the RTU was correctable
with the previously determined leach rate expression, namely:
2 2
PL - KL Y2 Wp2
113
-------
where
K is the leach reaction rate constant
Y is the ratio of reagent ferric ion-to-total iron
concentration, and
W is the weight percent of pyrite in the coal.
7. Coal desulfurization rates in the RTU exhibited an Arrhenius
type exponential temperature dependence as did previous bench
scale experimentation. That is,
KL = AL exp [-EL/RT]
where
A. is the pre-exponential factor or frequency factor
E. is the activation energy
R is the gas constant, and
T is the reaction temperature.
However, rate constants K|_ observed during RTU processing were
3 to 8 times greater than those obtained during bench-scale
experimentation. This is attributed in part to non-ideal
mixing of the reaction system which is a series of continu-
ous flow stirred tank reactors.
8. Desulfurization rates obtained during coal processing with
acidified water reagent appear to be essentially the same
as those obtained by processing with acidified iron sulfate
reagent. Thus, reagent iron concentration does not affect
pyrite leach rates with acidified reagents.
9. Mine cleaned Marti nka coal can be processed through the RTU
to yield a product coal having less than 0.6 pound of sulfur
per 10b Btu.
10. Reagent regeneration in the RTU appeared to be adequately
described by the previously determined regeneration rate
expression:
\
114
-------
where
KR is an Arrhenius type rate constant
PQ2 is the oxygen partial pressure
+2
Fe is the concentration of ferrous ion in reagent.
The reagent regeneration rate constant obtained in the RTU
appeared to be at least 37% of the rate constant obtained at
bench-scale. This possible disparity may be the result of
relatively rapid disentrainment of oxygen bubbles from the
slurry and the fact that the regenerator volume (slurry recir-
culation loop volume) per unit volume slurry is lower in the
RTU than in the bench-scale unit.
11. Reagent regeneration rates could not be improved by increas-
ing the oxygen throughput from a fourfold to a fourteenfold
stoichiometric excess.
12. Excess ferric ion consumption appears to be a temperature
dependent phenomenon. The degree to which this dependence is
quantified will determine the precision with which reagent
regeneration rates can be measured during L-R processing of
low pyrite coals (i.e., the mine cleaned Martinka coal).
13. An average dry nmf heat content increase of 80 Btu per pound
(±85 Btu per pound) was obtained during L-R coal processing
of 14 mesh top-size mine cleaned Martinka coal in the RTU at
temperatures of 230° to 270°F with 0 to 4.5% w/w iron reagent.
Thus, no coal matrix degradation was observed during RTU coal
processing.
14. RTU coal processing experimentation performed with acidified
reagents having 0 to 4.5% w/w total iron concentrations indi-
cated a product sulfate sulfur-to-elemental sulfur ratio of
2.2. Because of inefficiencies associated with the recovery
of small quantities of elemental sulfur (less than 0.3% w/w)
and physical losses which may be incurred, this data is con-
sidered to be essentially in agreement with previous bench-
scale data indicating a 1.5 product sulfur ratio. Thus,
reagent total iron concentration did not appear to affect
the ratio of sulfur products during RTU operation.
15. Slurry sampling during RTU operation was generally of high
precision but was not representative of the bulk slurry.
It, therefore, appears that meaningful slurry sampling will
require more efficient slurry agitation and/or relocation of
the sampling ports away from areas subject to localized flow
patterns, such as wall effects and eddys.
16. Magnetic flow meters are not suitable for measurement of
slurry flow rates under pulse flow conditions encountered
115
-------
in the R-l effluent. To obtain a material balance of the
R-l unit will require an alternate flow measurement device
with faster response times or an estimation of slurry effluent
from the quantity of steam flashed during R-l pressure let-down.
17. An excellent material balance was obtained for the mixer 1-2
with an average material recovery of 100% being obtained.
Material balance precision was approximately 7% which is
comparable to the precision of flow meters used to monitor
the liquid flow streams.
18. The acquisition of an oxygen mass balance for the R-l unit
appears to require modification of the vent gas flow measure-
ment system.
116
-------
7. SUPPORTING BENCH-SCALE EXPERIMENTATION
7.1 INTRODUCTION
Supporting bench-scale experimentation has been conducted concurrent with
the RTU program to expand coal cleaning technology utilizing the Meyers Process
and to assist on-line RTU operations by supplying technical processing data to
assist in the overall program.
This section principally contains experimental work involving the use of
the Meyers Process reagent for gravity separation of coal prior to chemical
cleaning as well as a review of related studies performed earlier, in the U.S.S.R.
Included in the section are experimental studies examining gravity separa-
tion of coal under equilibrium and non-equilibrium conditions and their atten-
dant effect upon recovery of clean coal fractions.
Additional experimental studies are presented herein which include: (1)
chemical processing (Meyers Process) of sink fractions of gravity separated
coal; (2) comparisons with studies made in the U.S.S.R.; and (3) general exper-
imentation to investigate processing changes, weathering of coal, density
determination of coal slurries, and elemental sulfur recovery by acetone.
7.2 GRAVICHEM SEPARATION
7.2.1 Background
A practical method for actual float-sink cleaning of coal in a dense liquid
has long been sought as an alternative to mechanical cleaning. Heavy liquids,
such as zinc chloride-water, or chlorinated, brominated, or fluorinated hydro-
carbons are useful for prediction of yields which can theoretically be obtained
in mechanical washing plants, but are impractical for actual production because
they are expensive and add pollutants to the coal, atmosphere or water table.
We find that the density of aqueous ferric sulfate leach solution, as utilized
in the Meyers Process, is ideal for accomplishing a practical gravity separation
117
-------
of coal for specific gravities between 1.2 and 1.4. For example, for about 40%
of Appalachian coal, the float coal (which is often 40-60% by weight of the
total) averages 0.8-1.2 Ibs S02/106 Btu and needs no further processing to meet
standards, while the sink coal slurry contains most of the coal pyrite. This
sink slurry can then be processed through the Meyers Process to produce coal
containing 1.2-1.5 S02/106 Btu, This approach allows production of coal which
will meet SO emissions standards for both new and existing stationary
J\
sources, and reduces processing costs relative to the straight through Meyers
Process by allowing 40-60% of the coal to bypass reactor, elemental sulfur
extraction and dryer units. This float-sink approach was tested many years ago
by A.Z. Yurovskii in the U.S.S.R. but only published in the open literature very
recently^2). Yurovskii subsequently treated the sink coal with a mixture of
nitric acid and ferric sulfate to remove pyritic sulfur, not realizing that the
separation medium itself was sufficiently active to accomplish near total pyritic
sulfur removal.
7.2.1.1 Literature Background--
The cited work of Yurovskii reports laboratory (bench-scale) work for the
removal of iron pyrite (FeS2) from coal by reacting it with ferric sulfate solu-
tions under a variety of conditions, and scale-up studies at pilot plant level
which demonstrates ferric sulfate solution gravity separation using centrifugal
separators with simultaneous sulfur and ash reduction. Principally, he inves-
tigated the effects of temperature, multiple ferric sulfate treatments and ash
content on the degree of pyrite removal. His data support the observations
from Gravichem processing (Sections 7.2.2 and 7.2.3) that significant reduction
in ash content and pyrite is observed upon preliminary or sustained treatment
of coal with ferric sulfate solution. Yurovskii, however, only reported 43%
removal as his highest value for pyrite depletion in his bench-scale studies
after twofold one-hour desulfurizations, which is low when compared to the
Meyers Process. Yurovskii's work is discussed in more detail in Section 7.2.2
and 7.2.3.
7.2.1.2 Process Application--
The Gravichem variation of the Meyers Process involves utilization of
ferric sulfate leach solution to gravity separate input coal into a very low
118
-------
pyrite float material and a high pyrite sink material. The sink material is
subsequently heated to reaction temperature and processed through the Meyers
Process. The Gravichem process modification is shown schematically in Fig-
ure 23. Run-of-mine coal, often containing 5-10 Ibs S02/MM Btu, is crushed and
screened to remove rock and easily removed pyrite. This rough-cleaned coal is
mixed with leach solution and pumped to a tank, where coal lighter than the
selected specific gravity of the leach solution floats and the heavier coal
sinks. The float fraction is filtered and washed but not dried, while the sink
fraction is treated by the Meyers Process to give a dry coal product. The
processed-sink fraction may be hot-briquetted without binder (see Figure 24)
to give a near 0% moisture product suitable for shipping in an open hopper-car.
For stoker-boiler (industrial-type boiler) applications, it may be recombined
with the wet float coal filter cake (about 20% moisture) to give a combined
product of a shippable 1" x 0 grind with a combined moisture content of 10%.
7.2.2 Equilibrium Gravichem Separation
A series of float-sink separations of Martinka mine coal (14 mesh x 0) in
ferric sulfate leach solution of varying specific gravity were performed
(Table 18) which demonstrate the feasibility and the physical and chemical basis
of the Gravichem concept.
These tests were of an equilibrium type in which coal was mixed with hot
leach solution (80°C) in an agitator for 15 minutes to ensure wetting, then
transferred to a float sink apparatus maintained at 80°C (Figure 25) and allowed
to separate without stirring. After the specified times, the float product was
skimmed from the top of the vessel and the sink product was separated from the
remaining leach solution. The results (Table 18) show that float coal is
obtained in 23-48% w/w yield, in each case meeting the NSPS standard of 1.2 Ibs
StyiO6 Btu - i.e., 0.94-1.16 Ibs S02/106 Btu, while the sink fraction treated
by the Meyers Process gave a product of about 1.2-1.3 Ibs S02/10 Btu.
The first separation (Experiment 1 of Table 18) was performed utilizing a
standard mixed organic solvent of specific gravity 1.30 for purposes of com-
parison with leach solution gravity separations (Experiments 2-7). It can be
seen that the yield of leach-solution float at 1.28 specific gravity was equiv-
alent to that obtained with the organic solvent, indicating that the flotation
utilizing ferric sulfate leach solution is essentially a physical separation
119
-------
ro
o
CRUSH/SCREEN
GRAVI-
SEPARATOR
1.3 SP. GR.
LEACH
.SOLUTION,
FILTER/ WASH
MEYERS PROCESS
CaSO.
ROCK
Figure 23. Gravichem Process
-------
ro
MEYERS PROCESSED COAL VIHICH HAS
BEEN BRIQUETTED WITHOUT BINDER
Figure 24. Hot-Briquetted Coal (1" -Size)
-------
TABLE 18. EQUILIBRIUM GRAVICHEM TESTS ON 14 MESH x 0 MARTINKA MINE COAL WITH VARIATION
OF MEDIUM AND SEPARATION TIME
Fraction, % w/w
of Coal***
Sp Gr
1. 1.30*
2. 1.28**
3. 1.28**
4. 1.28**
!\3 5. 1.33**
6. 1.33**
7. 1.43**
*Mixture of
Separation
Time, hrs
0.25
3
16
46
3
16
3
Float
27.6
23.0
31.0
24.0
33.0
40.0
48.0
Sink
72.4
77.0
69.0
76.0
67.0
60.0
52.0
toluene and perchloroethylene, wt
**Aqueous ferric sulfate at
80°C with
specific
Total Sulfur;
Float
0.75
0.80
0.72
0.69
0.77
0.76
0.83
0.80
ratio 1:1
gravity
Sink
1.89
1.45
1.39
--
1.56
1.56
2.11
2.26
.35.
obtained as
, % w/w Heat
Processed
Sink-* Float
14711
0.79 15057
0.86 14628
14829
14604
14748
14336
14413
Fe+++
Fe2(S(
(wt 5
follows: 1.28 = 7.5
1.33 = 5.0
1.43 = 5.0
Content, Btu/lb
Sink
12041
12917
12810
--
12665
12435
12108
11956
as
I) (wt %
4
21
30
Processed
Sink+
—
13105
12907
--
--
--
—
--
H20
) (wt %)
88.5
74
65
Sulfur Content, Ibs S/106 Btu
Float
0.51
0.53
0.49
0.47
0.53
0.52
0.58
0.56
Processed
Sink Sink*
1.57
1.12 0.60
1.09 0.67
-.
1.23
1.25
1.74
1.89
fHeated at 100°C for 24 hrs in aqueous ferric sulfate separation media, washed with water, then extracted with toluene and dried.
*Fiitered from separation medium, washed of leached solution treated and dried to constant weight in a vacuum oven at 100°C.
-------
LEACH
COAL SOLUTION
AGITATOR
SETTLER -(4 INCH DIAMETER X
24 INCH LENGTH)
FLOAT
ZONE 1
ZONE 2
-»- SINK
Figure 25. Bench-Scale Gravichem Test Apparatus
Flow Diagram
123
-------
equivalent to that obtained with organic solvents of the same density. However,
the sulfur content of the Gravichem float products, at 1.28 specific gravity,
decreases with increasing time of gravity separation (Experiments 2-4) indicat-
ing that leaching of pyritic sulfur is taking place, as expected, during the
gravity separation. Similarly, the sink fraction is undergoing significant
pyrite leaching - e.g., a sulfur content of 1.89% w/w .is seen for the sink coal
obtained from organic solvent but a sulfur content of 1.39-1.45% w/w is seen
for the Gravichem sink. The Gravichem sink material has greatly enhanced heat
content relative to the organic solvent sink coal because of dissolution of
coal mineral matter in addition to pyrite (Experiments 2-6).
Increasing the specific gravity of the iron sulfate leach solution gives
increased yields of float coal with only slightly higher sulfur content and
slightly less heat content (Experiments 5-7). No definite trend is seen for
the relative amounts of float material as separation is varied from 3 to 46 hours
(Experiments 2-4), while a possible trend is seen at 1.33 specific gravity
between 3 hours and 16 hours.
Coal from the Kentucky No. 9 Mine, supplied by the TVA Cumberland Power
Plant, was likewise processed by the Gravichem method. Ground coal at 3/8-inch
topsize was mixed with iron sulfate leach solutions of 1.28-1.34 specific grav-
ity (Figure 26), heated to 80°C and allowed to gravity separate in a holding
tank to give a 38-52% w/w yield of float product, after removal of residual
iron sulfate leach solution. The float product (Table 19) from all of the sep-
arations is a power plant fuel containing 3.09-3.50 Ibs S02/106 Btu, having a
heat content range between 13774 and 14354 Btu/lb. The sink fractions
(48-62% w/w) contain most of the coal pyrite.
The sink coal was size-reduced while still in leach solution to a 14 mesh x
0 coal/Ieach solution slurry (Table 20), then treated at 102°C according to
Meyers Process procedures. The product contained less than 4 Ibs S0?/106 Btu.
Thus, both float and processed sink coals meet the Tennessee State Implementa-
tion Standard requirements of 4 Ibs S02/106 Btu. A substantial decrease in
coal pyrite actually has occurred.
Gravity separation and chemical processing of Kentucky Mine No. 9 coal
produced a significant reduction in ash and total sulfur and an increase in
heat content as noted in Table 19. A more extensive analysis summary is
124
-------
3/8 INCH X 0 LEACH
COAL SOLUTION
MIX
TANK
ro
en
FLOAT
FILTERED, WASHED
GRAVITY
SEPARATION
TANK
AGITATOR
TANK
3/8" X 0 FLOAT PRODUCT
MEYERS PROCESS
14 MESH X 0
PROCESSED SINK PRODUCT
Figure 26. Gravichem Processing of TVA Coal
-------
TABLE 19. EQUILIBRIUM GRAVICHEM TESTS AT 80°C ON 3/8" TOP-SIZE KENTUCKY NO. 9 MINE COAL
WITH VARIATION OF MEDIUM
ro
o>
Sample
As Received
Fraction, %
w/w of Coal***
Exp
No.
1
2
3
4
5
6
7
*
**
Sp. Gr. Separation Float
Time, hrs.
1.30
1.28
1 .28
1 .28
1.33-1
1.30-1
1.30-1
Mixture
Aqueous
0.5
0.5
3.0
**
3.0
.34 3.0
.34 3.0
.34 3.0
of toluene and
ferric sulfate
40
35
38
41
45
48
52
Sink
60
65
62
59
54
52
48
Total Pyritic Ash
Sulfur Sulfur X w/w
ST sp
#1 4.40
12 4.49
13 3.99
Total Sulfur,
X w/w
2.13 12.79
1.81 12.79
*1.93 10.78
Heat
Content
Btu/lb.
12408
12414
12687
Heat Content,
Btu/lb
Sulfur Content
Ibs SO^/
106 Btu
7.09
7.24
6.29
Ash,
% w/w
Float Sink Processed Float Sink ; Processed Float Sink Processed
Sink* Sink+ Sink+
2.40 5.61
2.33 2
2.22 4.92
2.41 ,, 4.78 (2
(2.29)f
2.27 2
2.35
2.44 2
perchloroethylene, wt ratio 1:1.35.
at 80°C with specific gravity obtained
13889 11369
.42 14128
14354 12295
.54) 14046 . 12004
(14322)"
.57 13774
14205
.51 13960
- +++
Fe as
as follows: Fe2(SO^)3
(wt %)
3.
12827 3
3.
3.
(2.
12973 3.
3.
12335 4.
H2S04
(wt X)
93 17.50
48 9.51
19 14.83
22 , 14.78
43)*
38 10.40
52
13 13.04
H20
(wt X)
Sulfur Content,
Ibs S02/105 Btu
Float Sink
3
3
3
3
(3
3
3
.46 9.87
.30
.09 8.00
.30 j, 7.96
.20)*
.30
.31
3.50
Processed
Sink+
3.77
3.96
4.0
•>. 7.5 4 88.5
*** Filtered from separation medium, washed if leached solution treated and dried to constant weight in a vacuum over at 100°C.
+ Processed in 1.3 S.G. leach solution (above) for 48 hours at 102°C subsequent to size reduction in Waring blender for 5 minutes @ 15,000 rpm.
I Processed float.
-------
TABLE 20. PARTICLE SIZE DISTRIBUTION OF
SIZE-REDUCED KENTUCKY NO. 9*
SINK COAL
Screen Size
14
35
48
100
150
200
Pan
Retained,
% w/w
1.75
3.73
9.25
23.72
14.10
12.92
34.32
*
3/8 inch x 0 sink coal in Gravichem
leach solution, size-reduced in a Waring
blender for 5 minutes at 15,000 rpm.
presented in Table 21 which, in addition to the above noted parameters, also
contains sulfur forms analyses, including elemental sulfur (Sn) on three float
samples and four processed sink fractions.
Referring to Table 21 the reduction in total sulfur is related to both
pyritic and sulfate sulfur removal during gravity separation, as was noted for
the Martinka coal (above), e.g., compare Experiment 3 with the organic sink in
Experiment 1. Additionally, it is seen that elemental sulfur is formed during
the separation process. In Experiment 3, correcting for approximately 0.10% w/w
elemental sulfur* on the starting coal (Section 7.3.3) due to weathering, the
amount of S formed is greater than would be expected (0.40% w/w ASp) based on
the amount of pyrite (S ) removed. Sp formed in Experiments 4 and 5 is as
expected. For AS of 2.38 and 2.32% w/w during Gravichem processing, 0.95 and
0.93% w/w S should be produced. Based on the two analyses (1.0% w/w) and cor-
recting for 0.1% w/w Sn on the starting coal, it is seen that these data are
'Assumption is that sink fraction contents the majority of the elemental
sulfur formed during weathering.
127
-------
TABLE 21. GRAVICHEM PROCESSED KENTUCKY MINE NO. 9 COAL
ro
oo
Exp
No.
Coal Gravi ty
Fraction
%
Control Total coal
100
Sample
(As Received)
1*
2
3
4
5
6
7
Organic Float**
Organic Sink**
Organic Float**
Organic Sink***
Gravi chem Float**
Gravi chem Sink**
Gravi chem Float**
Gravi chem Float***
Gravi chem Sink**
Gravi chem Sink***
Gravi chem Float**
Gravi chem Sink***
Gravi chem Float**
Gravi chem Sink***
Gravi chem Float**
Gravi chem Sink**
Gravi chem Sink***
40
60
35
65
38
62
41
41
59
59
46
54
48
52
52
48
48
Ash
% w/w
12.12
±1.16
3.93
17.50
3.48
9.51
3.19
14.83
3.22
2.43
14.78
—
3.38
10.40
3.52
11.11
4.13
16.20
13.04
Heat
Content,
Btu/lb
12503
± 159
13889
11369
14128
12827
14354
12295
14046
14322
12004
—
13774
12973
14205
12627
13960
11767
12335
Total
Sulfur,
% w/w
4.27
±0.22
2.40
5.61
2.33
2.42
2.22
4.92
2.41
2.29
4.92
2.54
2.27
2.57
2.35
2.73
2.44
4.99
2.51
Pyri ti c
Sulfur,
% w/w
1.96
±0.13
0.54
2.59
0.33
0.11
0.39
2.30
0.38
0.12
2.45
0.21
0.52
0.27
0.35
0.48
0.50
2.20
0.33
Sulfate
Sulfur,
% w/w
0.38
±0.15
0.18
0.92
0.02
0.33
0.01
0.10
0.05
0.20
0.26
0.37
0.06
0.41
0.01
0.39
0.04
0.34
0.31
Organic
Sulfur,
% w/w
1.92
±0.20
1.68
2.10
1.98
1.98
1.82
2.52
1.98
1.96
2.21
1.97
1.69
1.89
1.99
1.86
1.91
2.45
1.89
Elemental
Sulfur,
% w/w
.03
.02
.35
1.0
1.0
0.05
1.40
* Refer to Table 5-3.
** Gravity Separation only.
*** Gravichem processed.
+ Elemental sulfur removed with toluene before coal analysis
-------
in agreement with postulated Meyers Process chemistry. Experiment 7, however,
indicates a much higher formation of Sn (1.30% w/w)* than would be expected from
the ASn (2.26% w/w), i.e., 0.90% w/w S would be expected based on 0.4% w/w AS
p " p"
With respect to the formation of Sn during oxidation of pyrite by ferric
ion, Yurovskii reports that it is only formed at low Fe2(S04)3 concentrations.
Many of his experiments were conducted using 16% Fe2(S04)3 or 4.48% Fe+++, which
he considered high, and the presence of Sn was not confirmed. The experiments
on Kentucky Mine No. 9 coal, shown in Table 21 were conducted in leach solution
containing 7.5% Fe . Thus, Yurovskii's error in not observing S leads to an
erroneous conclusion with respect to the efficiency of Fe2(S04)3 as an oxidizing
agent for removal of pyrite from coal. He based his conclusions for determining
the degree of pyritic sulfur reduction on total sulfur (Eschka method) differ-
ence before and after processing.
Therefore the elemental sulfur, which was undoubtedly formed, increased
the total sulfur of the coal and was assumed to be unreacted pyrite, thus lower-
ing the observed efficiency of the reaction. Table 22 is a summary of
Yurovskii's work on three coals of varying sulfur ash content. Also shown in
the Table are data obtained on Kentucky Mine No. 9 coal for various reaction
times at 102°C. Referring to Table 22, the degree of oxidation of coal pyrite
(FeS2), as reported by Yurovskii, based on total sulfur reduction, has been cor-
rected for Sn, i.e., the theoretical amount of Sn formed (0.4% w/w AS ) has
also been converted to pyrite loss and reported in the table as corrected data.
Corrections were made as follows:
Yurovskii assumed S^
where
= original coal total sulfur, %,
= treated coal total sulfur, %,
= pyritic sulfur reduction, %,
Corrected for initial sulfur on coal.
129
-------
TABLE 22. SULFUR REMOVAL FROM COAL WITH FERRIC SULFATE SOLUTIONS
Ash
Coal Content Treatment
of Coal Time
% w/w hrs.
Lutuginugol 14.0 1
1
1
2
7.1 2
Gorlovskugol 8.3 2
4.2 2
o Krasnodonugol 21.0 2
9.0 2
Kentucky No. 9 15.5 2
(1.3 sink) 4
8
14
32
48
Temp
°C
80
90
100
100
100
100
100
100
100
102
.102
102
102
102
102
Solution
Fe
Content
%
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
7.5
7.5
7.5
7.5
7.. 5
7.5
Sulfur Content of Coal , % w/w
Original Treated
Total
S
4.47
4.47
4.47
4.47
5.11
3.60
4.23
4.20
5.10
4.92
4.92
4.92
4.92
4.92
4.92
Pyri ti c
S
2.97
2.97
2.97
2.97
3.38
2.34
2.74
2.72
3.37
2.45
2.45
2.45
2.45
2.45
2.45
Total
S
4.03
3.70
3.61
3.45
3.80
2.63
2.73
3.90
3.80
3.16
2.91
2.65
2.43
2.45
2.54
Pyri ti c
S
2.53
2.20
2.11
1.96
2.07
1.33
1.24
2.42
2.07
1.28
1.01
0.62
0.37
0.25
0.21
Degree of Oxidation
of Coal FeS2, % w/w
Yurovskii
*
15
26
29
34
40
43
54
11
40
Corrected
**
25
43
48
57
67
72
90
18
67
Mpyers
+
48
59
75
85
90
91
* Yurovskii's data as presented.
** Yurovskii's data corrected for elemental sulfur.
f Experimental Meyers Process.
-------
However, from theory (above) sti - (Stf - Sn) = ASp (Assume Sn =0.4 ASp)
sti - Stf » ASp - 0.4 ASp
Stl - Stf
ASP — o —
AS
Degree of oxidation, % = original Pyritic S. %
It is also seen in Table 22 that Kentucky Mine No. 9 sink coal has a com-
parable, although lower, degree of oxidation of coal pyrite (compare with 2 hours
@ 100°C - Lutuginugol, 14% ash). Comparing the reaction rate constant (KL) in
Figure 27 of Kentucky Mine No. 9 coal (KL = 0.14 hours"1 W ~l) with Yurovskii's
Lutuginugol coal (0.21 hours ~1 W "-1), it is seen that the pyrite oxidation rate
is 67% of the Lutuginugol coal KL> Y = 0.96 was estimated for Yurovskii's work
since he used 12 grams of coal in 100 ml of 16% Fe2(S04)3 leach solution.
The effect of ash content on coal pyrite oxidation rate was noted by
Yurovskii. Plotting the corrected data from Table 22 (2 hours 0 100°C) for
each coal (degree of oxidation versus ash content), it is seen in Figure 28
that a correlation does exist. It may be noted that Kentucky Mine No. 9 coal
supports the correlation. Based on these data, the KL for Kentucky Mine No. 9
coal (15.5% ash) should be lower than 14% Lutuginugol, although not as low as
was observed.
A final comparison of the work of Yurovskii and data obtained on Kentucky
Mine No. 9 coal is seen in Table 23 . As was noted earlier for Kentucky Mine
No. 9 coal, Yurovskii also observed a large reduction in ash content of the coal
fraction separated by his chemicogravitational process. It may be noted that
a comparable degree of cleaning is noted for the two methods although Kentucky
No. 9 was cleaned at a larger particle size (0-9 mm).
7.2.3 Non-Equilibrium Separation
Following the equilibrium Gravichem separation studies (Section 7.2.2) it
was desirable to investigate non-equilibrium Gravichem float-sink separation
which would more nearly be the case in a process plant.
131
-------
to
ro
KENTUCKY MINE NO. 9
48%*LUTUGINUGOL
/K =0.21
50%* (2.4 HOURS)
•REDUCTION IN
PYRITE
15 20 25
PROCESS TIME, HOURS
Figure 27. Puritic Sulfur Leaching Data for 1.3 Specific Gravity
Sink Coal at 102°C (Kentucky Mine No. 9)
-------
IUU
# 90
tb
^-*
X
O 80
u
UJ
E 70
t
£ 60
u
u_
O 50
O
i *
X
O 30
u.
O
g 20
0
UJ
Q 10
n
\.
X
• X
% CQX = 103-3.88 ASH (%)
CORRELATION COEFFICIENT
• YUROVSKII DATA
0 KENTUCKY MINE NO. 9
V
>*. •
\
N
\
• S
ASH, %
Figure 28. Coal Oxidation After Two Hours at 100°C
as a Function of Ash Content
133
-------
TABLE 23. COMPARATIVE RESULTS OF CLEANING
Coal
Size,
Method mm
Chemicograv itational 0-3
(Yurovskii)
Gravichem 0-9
(Kentucky No. 9)
Original Coal,
% w/w
Ash
25.79
12.12
12.12
12.12
Total
Sulfur
5.82
4.27
4.27
4.27
Float, % w/w
Yield
60
52
48
46
Ash
4.50
4.13
3.52
3.38
Sulfur
2.17
2.44
2.35
2.27
Sampling ports were added to the bench scale Gravichem separator (Fig-
ure 25) and the float fraction was skimmed off the top of the leach solution
as before, while coal particles and leach solution in zones 1 and 2 were removed
through valves and finally the sink fraction was obtained by separating from
the filtrate. This allows isolation of coal "in transit" for a first prelimi-
nary look at the behavior of float and sink particles in a dynamic situation.
Two series of non-equilibrium Gravichem float/sink separations were con-
ducted with Martinka coal at 80°C using 1.2 and 1.3 specific gravity (S.G.)
leach solutions. Other experimental parameters include:
• Weight of coal: 900 g (nominal).
t Weight of leach solution: 3600 g (nominal).
§ Coal size: 14 mesh.
The results for 1.3 S.G. leach solution (Table 24) show that the samples
taken were indeed non-equilibrium, in that times greater than 2 hours would be
required for the slurry to have attained equilibrium.
It may be noted that the combined float and zone 1 fractions (1.3 S.G.)
are representative of the float portion taken in equilibrium Gravichem separa-
tion (Btu, ash, total sulfur, pyritic sulfur, etc.) as shown in Table 18.
Likewise, the combined zone 2 and sink portions are comparable to equilibrium
sink percent recovery (coal % w/w) in Table 18; however, the zone 2 fraction
in Table 24 is more nearly like the combined float and zone 1 fractions in the
same study. Thus, non-equilibrium Gravichem shows a greater yield of clean
134
-------
TABLE 24. ANALYSES OF FRACTIONS OF MARTINKA COAL OBTAINED FROM
NONEQUILIBRIUM FLOAT/SINK EXPERIMENTS
Fraction
Float
Zone 1
Zone 2
Sink
Total Coal
(Weighted
Average)
Starting
Coal
*
Trace;
*+
Time S.G. Coal
Hrs. of Leach % w/w
Solution
0.5
1.0
2.0
0.5
1.0
2.0
0.5
1.0
2.0
0.5
1.0
2.0
0.5
1.0
2.0
...
values for S
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
—
were
10.0
14.0
14.0
21.3
21.0
23.9
30.3
23.0
17.0
38.4
42.0
46.0
— -
< 0.01; 0.01
Sulfur Analysis - !
St
0.79
0.87
1.01
0.91
0.79
0.81
0.96
0.92
0.84
1.76
1.72
1.66
1.24
1.24
1.24
1.61
was used
SP
0.13
0.13
0.18
0.16
0.12
0.14
0.26
0.27
0.16
1.15
0.98
0.92
0.57
0.54
0.51
0.72
forS0
Ss
Tr*
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
0.18
I w/w
So
0.65
0.73
0.82
0.74
0.66
0.66
0.69
0.65
0.67
0.60
0.73
0.73
0.66
0.69
0.72
0.71
calculation
Heat
Content
Btu/lb.
14851
14650
14454
14604
14642
14635
14069
14105
14136
11357
11132
11732
13220
12951
13321
12950
(starting coal
Ash
% w/w
4.22
5.35
6.25
6.39
5.41
5.31
8.46
8.27
7.84
23.95
23.85
21.87
13.54
14.30
13.54
14.70
excepted).
Sulfur Content - Ibs
Individual 04.7!**
Fraction hL i]
0.53
0.59
0.70
0.62 0.59
0.54 0.56
0.55 0.61
0.68
0.65
0.59
1.55
1.55
1.41
0.94
0.96
0.93
1.24
S/106 Btu
FL+Z1+Z2
0.63
0.60
0.60
Weighted average of Float (FL) and Zone 1 (Zl) or Float and Zone 1 and Zone 2 (Z2).
-------
coal may be attained at 1.3 S.G. than was noted in the preceding equilibrium
studies (Section 7.2.2).
Table 25 includes the non-equilibrium data obtained at 1,2 S.G. Forty-
three percent of the coal is retained in zone 1 and zone 2 at 0.2 hours, which
has Btu, ash and total sulfur content comparable to zone 1 and zone 2 fractions
obtained at 1.3 S.G. after 0.5 hours. The recovered coal, however, is less for
the 1.2 S.G. separation (43% versus 52%), plus an additional 10% of float was
also recovered from the 1.3 S.G. 0.5 hours separation. With respect to the
1.2 S.G. separation, although more coal was recovered in zone 1 and zone 2 (51%)
for 0.1 hours at 1.2 S.G. than for 0.2 hours (43%), the total sulfur and heat
content, however, indicate a less-clean coal fraction compared to the 0.2 hours
1.2 S.G. (0.70 and 0.63 Ibs S/106 Btu, respectively).
Particle size distributions for each coal density fraction listed in
Table 24 are shown in Figure 29. It is seen that although the percentage dis-
tribution of particles is comparable between 1 and 2 hours for each mean parti-
cle size (mesh) within the float and zone 1 and 2 fractions, the decrease in
60 mesh and increase in pan (-200 mesh), within the sink fraction, suggests that
equilibrium has not been achieved.
Based on the work of Yurovskii (op cit) and the application of Stokes Law,
it is not to be expected that small particles will attain equilibrium within
the system due to gravity alone over short periods of time. It is well known
that the velocity at which fine particles settle in heavy liquids is given by
Stokes Law:
2 r2
v = -(p - P
where
r = mean particle radius, cm,
PI = density of the liquid, g/cc
P2 = density of the settling particle, g/cc,
n = viscosity of liquid, poise, and
g = acceleration due to gravity.
136
-------
TABLE 25. ANALYSES OF FRACTIONS OF MARTINKA COAL OBTAINED FROM
NONEQUILIBRIUM FLOAT/SINK EXPERIMENTS
CO
Fraction
Float
Zone 1
Zone 2
Sink
Total Coal
(Weighted
Average)
Starting
Coal
Time
Hrs.
0.2
0.1
0.2
0.1
0.2
0.1
0.2
0.1
0.2
-..
S.G.
of Leach
Solution
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
Coal
% w/w
0.2
22.5 0
14.0 0
28.5 0
28.6 0
49.0 1
57.2 1
1
1
1
Sulfur Analysis - °,
St
**
.98
.90
.99
.88
.55
.49
.26
.23
.61
0
0
0
0
0
0
0
0
0
SP Ss
Tr*
.22 Tr
.18 Tr
.23
.22
.89
.86
.55
.58
.72 0.18
£ w/w
So
0.75
0.71
0.75
0.65
0.65
0.62
0.70
0.64
0.71
Heat
Content
Btu/lb.
14050
14234
13992
14189
12250
12446
13151
13157
12950
Ash
% w/w
7.97
7.61
8.45
8.03
18.83
18.06
13.43
13.67
14.70
Sulfur Content - Ibs S/106 Btu
individual FL+Z1*** FL+Z1+Z2—
Fraction
0.
0.
0.
0.
1.
1.
0.
0.
1.
70 0.70
63 0.63
71 0.71
62 0.62
26
20
95
93
24
Trace; values for S were « 0.01; 0.01 was used for S calculation (starting coal excepted).
Insufficient coal to process for analysis.
Weighted average of Float (FL) and Zone 1 (Zl) or Float and Zone 1 and Zone 2 (Z2).
-------
00
70
60
50
S 40
Z
30
20
10
FLOAT ZONE 1 J I ZONE 2 H SINK 1
1 HR 2 HR
60 100 200 PAN
60 100 200 PAN
MESH
60 100 200 PAN
60 100 200 PAN
Figure 29. Particle Size Distribution in Fractions of RTU 3 Coal
-------
It was pointed out by Yurovskii that for particles of a given diameter and
specific gravity, an increase in the speed of particle movement is possible
only when the speed will depend not on acceleration due to gravity, g, but on
the acceleration due to centrifugal force, i.e.,
2
Vc = 97T(p2 ' pl} Cg
where
= _ _
g 90
rc = rotational radius of centrifuge, cm,
n = number of revolutions of the centrifuge/min (rpm).
Table 26 illustrates the relative sensitivity of small particles of Martinka
coal and pyrite to gravitational and centrifugal forces with respect to veloc-
ity of settling. It may be noted that the interaction effects (particle-
particle) are not included in this comparison; thus, the velocities are in
reality lower than those given.
It is seen that the centrifugal method is particularly effective and would
enhance coal separation in the Gravichem process. The effectiveness of the
centrifugal method is demonstrated by Yurovskii 's data in Table 27, wherein he
compares a flotation and centrifugal separation of a coal containing originally
24.5% w/w ash and 3.21% w/w sulfur. It is seen that for essentially equal con-
centrate yield from 1.5 S.G. solution, the centrifugal method gives a much
cleaner concentrate (float).
It was desirable to investigate the non-equilibrium float/sink process
further by assessing the effect upon coal separation (yield) when discrete par-
ticle size fractions of Martinka coal are used, i.e., the standard input coal
(14 mesh x 0) used in the preceding experiments (Tables 24, 25) was separated
into four particle size fractions (+60, -60 x 100, -100 x 200, -200 mesh).
Non-equilibrium float sink experiments were conducted at 80°C on each of the
four fractions for 30 minutes in 1.3 S.G. leach solution, as described above.
It is seen in Table 28 that the combined yield for float, zone 1 and zone 2
fractions is similar to the previous experiment (Table 24) using the whole coal.
139
-------
TABLE 26. RELATIVE SETTLING RATES OF COAL AND PYRITE SUBJECTED TO
GRAVITATIONAL AND CENTRIFUGAL FORCES
Particle Particle
(mesh)
Coal 100
200
Pyri te 1 00
200
V - fl_(p2-Pl) g
** 9 K-2 r
V - 2 r fr, n }
V ' 9 n Ip2 pl'
Size
(cm)
0.015
0.0074
0.015
0.0074
c»z
90
Density of
Particle
(g/cc)
1.45
1.45
5.0
5.0
*
Gravitational
Vel oci ty
cm/sec
9 x 10"3
2 x 10"5
0.22
5.6 x 10"4
**
Centrifugal Velocity
(1000 rpm)
cm/sec
r = 10 r = 300
c c
1.02 30
2 x 10"3 6 x 10"2
24.9 747
6.3 x 10"2 1.89
-------
TABLE 27. FLOTATION AND CENTRIFUGAL SEPARATION ON COALS*
Original Coal, % w/w
Sample Source Ash Sulfur
Irminskaya 31.0 5.30
TsOF
Kalmiusskaya 24.5 3.21
TsOF
Cleaning Method
Flotation proper
Cleaning flotation (with a
depressor)
Centrifugal separation in
sulfate solution
Centrifugal separation after
preliminary chemical
treatment
Flotation proper
Cleaning flotation (with a
depressor)
Centrifugal separation in
sulfate solution
Concentrate, %
Yield
71.9
62.0
62.8
62.4
79.8
69.5
68.4
Ash
14.7
9.9
6.2
3.7
11.0
8.2
4.4
w/w
Sulfur
3.70
3.55
2.60
1.84
2.60
2.61
1.86
* Table 80, Yurovskii"2'
-------
-p.
ro
TABLE 28. YIELD OF MARTINKA COAL SEPARATED IN NONEQUILIBRIUM FLOAT/SINK
EXPERIMENTS AFTER 0.5 HOURS @ 80°C (S.G. 1.3)
Feed Coal
Size Fraction
(Mesh)
-200
-100 x +200
- 60 x +100
- 14 x +60
Cumulative Fraction
Recovered
Fraction Recovered
from 14 x 0 mesh Coal
Feed Coal
Weight
(gm)
537
725
776
792
*
Float
5
3
160
194
13%
10%
Recovered
Zone 1
177
302
194
126
29%
21%
Coal (gm)
Zone 2
200
161
76
32
17%
30%
Sink
130
235
344
425
41%
38%
Recovery
%
95
97
99
98
* 0.5 hour, 1.3 S.G, from Table 23.
-------
Likewise, the sink fraction is comparable. It is noted, however, that the yield
of zone 1 and zone 2 is approximately reversed which indicates that some lower
specific gravity coal may be contained in zone 2 in the previous experiment.
Referring again to Figures 29 and 30, one notes that the total sulfur,
based on the weighted average for total coal, decreases ^0.35% w/w. This is
seen at either 0.1 hours (S.G. 1.2) or 2 hours (S.G. 1.3). Such a decrease
cannot be accounted for (including a correction for removal of S ) on the basis
of the rate constant (KL) for removal of pyritic sulfur (S ) from Martinka coal.
However, the data show the decrease to be consistent for all of the experiments
in both tables. A series of experiments were conducted at 80°C with 14 mesh
Martinka coal wherein samples were wetted by 30 minutes prefoaming, leached
separately in water and 4% w/w aqueous H2S04 solution for 1 hour, and for 0.5,
1 and 3 hours in 4% w/w aqueous H2S04 containing 7.5% iron by weight.
Subsequent coal analyses from these experiments are presented in Table 29.
The following observations are made from these data.
t The leached coal total sulfur decreases primarily as a result
of reductions in sulfate and organic sulfur forms.
t Since organic sulfur is not water soluble and since the organic
sulfur value of the feed coal analyzed abnormally high, it can
can be concluded that the reduction in total sulfur by water
and by dilute H2S04 represents sulfate leaching (the reduction
in organic sulfur is attributed to errors by sulfur forms
analyses).
t Elemental sulfur formed indicates a maximum of 0.12% w/w reduc-
tion in pyrite during treatment with acid-Fe reagent (assuming
0.03% present as a result of weathering).
7.3 PROCESSING OF RTU COAL
Experimental studies were conducted at bench-scale level to evaluate proc-
essing variables and their attendant effect on sulfur removal from Rartinka
mine cleaned coals in support of the RTU. A series of experiments were com-
pleted which include pressurized bench-scale processing at 120°C and 135°C,
and unpressurized processing at 102°C. Additional tasks included determination
of the density of leach solution/coal (slurry) and evaluation of the weathering
of RTU candidate coals.
143
-------
Pressure
Wet
Control Test
Meter
Gas
Analysis
[ PT3 L
MIXER
AND
DEFOAMER
N,
Pump Seal
Purge System
lC_OAL._j
SLURRY
I FINAL PROCESSING*
Reactor-Settler Stirrer Reactor
STEP #1 (* 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 i> 13 liters.
Figure 30. Bench-Scale Coal Leaching and Reagent Regeneration Apparatus
-------
TABLE 29. LEACHING OF MARTINKA COAL AT 80°C
Reagent
Feed Coal
Water
4% H2S04 in
Water
4% H2S04 +7.5%
Fe in H20
Time
Hrs.
1.0
1.0
1 0.5
1 1.0
( 3.0
Total
Sulfur
s.
t
1.62
1.35
1.27
1.30
1.15
1.23
Coal
Pyri ti c
Sulfur
S
P
0.72
0.84
0.76
0.75
0.68
0.64
Analyses
Sulfate
Sulfur
S
s
0.18
**
Tr
Tr
Tr
Tr
Tr
X, w/w
Organic
Sulfur
S
0
0.72
0.50
0.50
0.55
0.47
0.59
Elemental
Sulfur
S
n
0.03***
0.03
0.04
0.06
0.06
0.08
**
***
Coal analyses performed after removal of elemental sulfur (except feed coal)
<0.01%, reported as trace
Included in the organic sulfur column
-------
7.3.1 Pressurized Bench-Scale Processing
Experimental studies were conducted in the pressurized bench-scale reactor
(Figure 30) as described previously'5', using 14 mesh Martinka coal processed
at 135°C and 100 psig 02 pressure. The object of the study was to investigate
the effect of reduced Fe+++ and acid (H2S04) concentrations on the removal of
sulfur from this coal under simulated RTU conditions. It is seen in Table 30
(Experiments 1 and 2) that the amount of sulfur removal is not affected by a
_i __ ii
reduction in Fe and acid concentrations when compared to coal processed in
-
5% Fe and 4% H2S04 standard RTU leach solution [Reference Experiment 01-10
(2P; 12F)].
7.3.2 Ambient Pressure Bench Scale Processing
Gravichem separation of coal using leach solution with various densities
achieved by increasing the acid or Fe content, required the investigation of
these solutions with respect to their suitability as reagents for the Meyers
Process.* It was demonstrated in Section 7.2 that 7.5% Fe and 4% acid is a
suitable leach solution as shown by the reduction of inorganic sulfur in Ken-
tucky Mine No. 9 coal. It has also been shown (Table 30, Experiment 3) that
high (20%) I^SO^ concentration does not change the response of inorganic sulfur
to the oxidizing reaction of the Meyers Process. A significant reduction in
ash and increase in heat content is noted, however, with the higher acid
concentrations.
7.3.3 Weathering of RTU Coal
A study was conducted to determine the effect of coal particle size on
weathering of Martinka coal. RTU coal No. 1 was used for these tests since
high sulfate values (MD.40% w/w sulfate sulfur) were found on this coal which
was stored for sometime after grinding. The objective was to determine if the
sulfate was due to weathering of the as-received 1-1/2" x 0 coal or the result
of post-grinding storage. A representative sample of the coal pile was taken
at the storage area and particle size fractions were obtained from the sample
to determine the extent of weathering for three fractions after grinding to
3/8" topsize (3/8" x 14 mesh; 14 x 100 mesh; 100 mesh x 0).
Standard starting leach solution contains 18% Ferric Sulfate (5% Fe),
4% acid, 78% H20.
146
-------
TABLE 30. BENCH-SCALE PROCESSING OF MARTINKA COAL
Experiment
No.
Control
RTU Processed
Exp. 01-10
*
*
**
3
*
Processed
**
Processed
***
Temp Tine 02
Pressure
(°C) (Mrs) (psi)
132
135 12 72
135 12 72
***
102 48 0
at 100 psig.
at ambient pressure.
Fe H0SO,, Coal Analysis,
" Ash Total Pyritic
(%) (%) Sulfur Sulfur
St SP
5 4 — - 0.91 0.17
0 2 15.76 0.91 0.23
0 0 16.15 0.90 0.20
5 20 11.80 0.96 0.29
% w/w
Sulfate Organic
Sulfur Sulfur
Ss So
0.12 0.65
0.17 0.51
0.01 0.69
0.06 0.61
Meat
Content
(Btu/lb)
___
12360
12509
13456
Equivalent to ^ 14 hours 0 135°C.
-------
Analytical data are summarized in Table 31, Included in the Table are
similar data for the two additional RTU coals (designated RTU #2 and RTU #3)
which were received from the mine in July 1977. Comparing RTU 1* with RTU 2*
and RTU 3* in Table 31, it is seen that a significant decrease in pyritic sulfur
and an increase in sulfate sulfur occurs for RTU 1.
The data indicate that the majority of the observed weathering in the RTU
ground coal occurred prior to the grinding operation. Additionally, the weath-
ering is not a pile surface phenomenon since it is seen that the RTU 1 (Interior)
sample likewise shows decreased pyritic sulfur and increased sulfate sulfur.
Elemental sulfur analyses of coals processed in the program studies have
been included in Table 32. Samples of nonprocessed input coal were ground to
14 mesh and extracted with toluene to remove elemental sulfur. It is seen that
uniformally sampled RTU coals have 0.03% w/w elemental sulfur whereas Kentucky
Mine No. 9 coal (a Gravichem coal) has 0.10% w/w elemental sulfur. These data
show that elemental sulfur is produced during weathering, however, not to the
extent that is is produced (M3.4% w/w AS ) during Meyers Processing.
During the RTU coal sampling studies it was noted (Table 32) that random
(yellow) deposits were present throughout the coal pile. Sampling of these
deposits with subsequent processing for removal of elemental sulfur revealed
the presence of higher than normal elemental sulfur. Additionally, qualitatively,
high presence of sulfate was determined.
7.3.4 Density of Coal Slurry
A series of density determinations were made on 100 mesh x 0 and 14 mesh x 0
Martinka mine cleaned coal in slurry concentrations ranging from 19% to 42% coal
by weight in distilled water. Densities of the slurry combinations are tabu-
lated in Table 33. It is seen, in addition, that the average density of the
14 mesh x 0 coal is 1.4241 gm/cc ±0.0052 gm/cc. Similar values for the
100 mesh x 0 coal (1.4221 gm/cc ±0.0053 gm/cc) indicate that comparable wetting
of the coal is achieved for both size distributions.
*Ground to 3/8" x 0 before separation into fractions.
148
-------
TABLE 31. DISTRIBUTION OF INORGANIC SULFUR IN RTU COALS (%)
vo
Coal
RTU 1***
RTU 1
(Interior)
RTU 2***
RTU 2
(Interior)
RTU 3***
RTU 1
RTU 1A
RTU 2
RTU 3
Combined 3/8" x 14 Mesh TOO Mesh x 14 100 Mesh x 0
Coal Fractions Sn Sc S_ S S S
s*s** p s p ps
P s
0.92 0.59 0.78 0.44 0.66 0.50 1.88 1.3
0.70 0.58
1.16 0.01 1.05 0.01 1.00 0.11 2.18 0.36
1.16 0.01
1.02 0.05 0.82 0 0.92 0 2.00 0.11
(COAL SIZE FRACTIONS)
58 26 16
51 18 6
63 25 13
67 22 11
S = Pyritic Sulfur
** P
S = Sulfate Sulfur
Ground to 3/8" x 0 before separating into fractions
-------
TABLE 32. ELEMENTAL SULFUR CONTENT OF
NONPROCESSED COAL
Elemental Sulfur,
Coal* % w/w
RTU 1** 0.01
RTU 2** 0.02
RTU 3** 0.03
Kentucky Mine No. 9** 0.10
RTU 1 (Yellow)*** 0.07
RTU 2 (Yellow)*** 0.10
*Toluene extracted to remove elemental sulfur
**Uniform sampling
***Speciany sampled yellow deposits
7.3.5 Elemental Sulfur Recovery
The product elemental sulfur from the oxidation of pyrite in coal by the
Meyers Process has been routinely recovered by extraction with toluene and sub-
sequent distillation of the solvent. Sulfur recovery is virtually complete by
a single stage extraction of leached coal with toluene provided the coal is
predried or azeotropically dewatered. Experimentation has demonstrated that
acetone is preferable to toluene for product sulfur recovery because (a) it
does not require that the coal be dewatered prior to recovery, (b) it extracts
iron sulfates as well as elemental sulfur from wet coal, and (c) it is easily
(56°C) and completely (>99.5%) recoverable from coal. Acetone has been selected
as the most cost effective technique for elemental sulfur recovery from Meyers
Process leached coal.
Investigations on acetone efficiency in recovery of product elemental
sulfur, Sn, from the oxidation of pyrite in coal (Meyers Process) was performed
at bench-scale with bench-scale leached and RTU leached Martinka coal samples.
Parameters examined for appreciable effect on recovery efficiency were: (a) the
moisture content of leached coal and feed solvent (from 1% to 72% water based
150
-------
TABLE 33. DENSITY OF COAL/WATER SLURRIES
Coal Slurry Temp
Size Concentration °C
(Mesh) %
14 21 40
60
80
28 40
60
80
42 40
60
80
100 23 40
60
80
Density
of Slurry
(gm/cc)
1.0594
1.0510
1.0418
1.0849
1.0768
1.0685
1.1363
1.1286
1.1215
1.0920
1.0835
1.0760
Density
of Coal
(gm/cc)
1.4265
1.4187
1.4235
1.4325
1.4256
1.4301
1.4243
1.4171
1.4190
x 1.4241
a 0.0052
1.4260
1.4161
1.4241
x 1.4221
a 0.0053
Slurry
EQUATIONS FOR CALCULATING DENSITY (D), gm/cc
Function of Temp, °C
21 D = 1.07713 - 0.00044 Temp
28 D = 1.10133 - 0.00041 Temp
42 D = 1.15100 - 0.00037 Temp
Temp
°C
40
60
80
Function of Slurry
Concentration, %
D = 0.9824 + 0.00366 Slurry
D = 0.9734 + 0.00370 Slurry
D = 0.9622 + 0.00379 Slurry
151
-------
on dry coal weight*), (b) coal particle size (14 and 100 mesh top-size), (c)
extraction residence time per stage (0,5, 1.0, and 2.0 hours), (d) number of
extraction stages (1 to 3), and (e) sample prehistory. The latter included
leached coal samples extracted immediately after taken from the RTU belt filter
(water rinse only) and samples that were stored for as long as 30 days; the
study also included thoroughly washed coal samples and oven dried samples. The
majority of experiments were performed with 14 mesh top-size, RTU leached coal
slurried in twice its dry weight acetone. This coal contained 0.30 grams ele-
mental sulfur per 100 grams of dry coal; one of the bench-scale leached Martinka
coals contained 0.40% elemental sulfur. Acetone extractions were performed at
the reflux temperature of the slurry (56°C).
The data generated in elemental sulfur recovery by acetone investigations
are summarized in Table 34 and Figures 31 and 32. Table 34 presents data from
19 experiments involving two-and three-stage extractions of leached coal with
acetone. Data are also presented from toluene extractions of dry leached coal
for comparison purposes. The table shows the estimated weight of elemental
sulfur and of moisture on 100 weights (dry basis) of leached coal fed to the
first extraction stage (columns 2 and 3), the weight of the input sulfur
recovered in each of the three stages, the total S recovered, and the percent
recovery (recovery efficiency). The elemental sulfur on the feed coal (leached
coal) was estimated by three independent techniques: (1) the sulfur recovered
by three toluene extractions, (2) forty percent of the chemically leached pyrite,
and (3) from the increase in organic sulfur of leached but not extracted coal;
the three techniques agreed within 10%. The asterisked experiments were performed
with leached coal which was not completely washed from the iron sulfate reagent;
thus, it is conceivable that a small quantity of sulfate sulfur may have been
analyzed as elemental sulfur which could have exaggerated slightly the quoted
efficiencies.
Figure 31 depicts the expected elemental sulfur recovery efficiencies, as
a function of acetone extraction stages, from up to 14 mesh top-size chemically
leached Martinka coal. The data were generated at reflux temperatures with
Water was present on the feed coal and in a limited number of experiments
in the acetone feed.
152
-------
TABLE 34. PRODUCT ELEMENTAL SULFUR RECOVERY FROM LEACHED MARTINKA COAL
01
to
Sulfur (Sn), Grams, Per TOO Grams of Coal
Experiment
No.
A. Acetone
1**
2**
3
4**
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Estimated Sn
on Coal
(% w/w)
Extractions
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.40
0.40
0.40
0.30
0.40
0.30
0.30
0.30
0.30
Moisture on
Feed Coal
(% w/w)
1
1
1
1
1
20
20
20
20
20
24
28
42
42
52
72
72
72
72
1st Stage
Extraction
Sn Recovered
0.23
0.29
0.21
0.27
0.27
0.27
0.22
0.20
0.28
0.22
0.26
0.27
0.20
0.18
0.20
0.13
0.17
0.14
0.14
2nd Stage
Extraction
S Recovered
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.05
0.03
-
0.08
0.07
0.12
-
0.13
_
_
_
-
3rd Stage
Extraction
S Recovered
_
0.02
_
.
0.01
.
0.001
_
0.01
-
0.05
0.05
0.05
-
0.08
.
_
_
-
Total Sn
Recovery
(% w/w)
0.27
0.35
.25
0.31
0.32
0.31
0.26
.25
0.32
-
0.39
0.39
0.37
-
0.41
_
_
_
-
%
Recovery
90
117
83
103
107
103
87
83
107
-
98
98
93
-
103
_
_
_
-
B. Toluene Extractions
20
21
0.30
0.30
11
1
0.28
0.26
0.01
0.01
_
~
0.29
0.27
97
90
One hour extractions of 33% coal slurries at reflux temperatures (56°C for acetone
slurries, 110°C for toluene).
Filter belt water rinse only on process coal.
-------
WATER CONTENT OF COAL
t FED TO FIRST STAGE
• FfH-f H H t Hi 11 -H- H-H-H -ftt-f ttt-H+H-H
20 (23 SAMPLES)
50 (10 SAMPLES)
72 ( 7 SAMPLES)
EXTRACTION STAGES EMPLOYED
Figure 31. Elemental Sulfur Recovery as a Function
of Acetone Extraction Stages
154
-------
-tr rr!T tft;
1st STAGE
ON COAL FED TO FIRST STAGE
Figure 32. Effect of
Recovered
Water on Elemental Sulfur
by Successive Acetone Stages
155
-------
coal which was leached chemically under a variety of conditions and up to
30 days prior to acetone extraction. Acetone extraction time was varied between
30 minutes and 2 hours per stage. The ±10 percent band about the first stage
S recovery value indicated for the 0-20 percent moisture coal samples applies
also to the higher moisture coal samples also. This band represents the maxi-
mum uncertainty expected in the quoted average value and it is due partly to
experimental and analysis uncertainties and partly to minor parametric effects
(including effects due to small differences in water content). Similar or even
larger relative uncertainties in the second and third extraction stages are
masked by the small percentage of original elemental sulfur recovered in these
stages.
Figure 32 summarizes the effect of water in the coal-acetone slurry on
elemental sulfur recovery from chemically leached coal. The moisture effect
is depicted for each successive extraction stage.
The data presented here complemented with chemical analyses on the recovered
elemental sulfur residues and determinations on acetone retention by extracted
and dried coal led to the following observations:
• Elemental sulfur recovery from Meyers Process leached coal by
a single stage acetone extraction ranged from 49% to 85%
depending on the moisture content of the leached coal.
• The average elemental sulfur recovery by a single stage acetone
extraction of processed coal containing 1-20% moisture was 83%.
An additional 13% (approximately 80% of the residual Sn) was
recovered in the second acetone extraction stage and another
2-3% in the third stage. The corresponding elemental sulfur
recoveries from wet coal containing 50-70% moisture were 50,
30, and 20%.
• The water content of the chemically leached coal or of the feed
(recycled) acetone had a pronounced effect on sulfur recovery
efficiency during the first extraction stage. The moisture
effect diminished in subsequent stages.
t Coal particle size up to 14 mesh top-size and extraction res-
idence time beyond 30 minutes did not exhibit measurable effect
on elemental sulfur recovery efficiencies by acetone.
• The purity of the sulfur ranged recovered from the first stage
ranged between 50-80% and that of the second stage was between
40-60%; the sulfur content of the residue recovered in the
third stage was approximately 10%.
156
-------
• Approximately 0.5% of the coal matrix was dissolved in the
three acetone extraction stages; at least 60% was dissolved
in the first stage. The quantity of coal matrix dissolved
in acetone appeared to be independent of the amount of sulfur
recovered. Thus, product sulfur purity should increase with
increasing pyrite concentration in the feed coal (higher
concentration of product sulfur in the coal).
The above observations lead to the following conclusions:
t Elemental sulfur can be readily recovered from wet leached
coal by acetone.
t Two 1-hour acetone extraction stages should be adequate for
better than 85% sulfur recovery provided the water content
of the acetone-coal slurry does not exceed 40% of the weight
of the coal in the slurry. (A third stage may be required
if the water concentration of the slurry is appreciably
higher).
t Water concentration is the only parameter of those examined
and in the ranges investigated that had a pronounced effect
on elemental sulfur recovery by acetone.
• Acetone retention by the coal is expected to be less than
0.5% upon drying in commercial driers (extrapolation of
data generated at TRW and at Wyssmont Co., Inc. laboratories).
• Acetone extraction is selective to elemental sulfur. Less
than 0.5% of the coal matrix was dissolved in three stages
of extraction (most of it was dissolved in the first stage).
7.4 LABORATORY STUDY CONCLUSIONS
1. Equilibrium float-sink separations of Martinka mine coal
(14 mesh x 0) in ferric sulfate leach solution of varying
specific gravity demonstrate the feasibility and the physical
and chemical basis of the Gravichem method.
2. The sulfur content of Gravichem float and sink coal fractions
is reduced during gravity separation because of leaching of
pyritic sulfur.
3. Gravichem sink material has greatly enhanced heat content
because of dissolution of coal mineral matter in addition
to pyrite.
4. Equilibrium Gravichem processing of 3/8-inch top-size coal
from Kentucky No. 9 Mine (TVA coal) generated a float product
(38-52% w/w yield) containing 3.09-3.50 Ibs SOp/HP Btu, hav-
ing a heat content range between 13774 and 14354 Btu/lb.
This coal meets Tennessee State sulfur standards.
157
-------
5. The sink fraction of the TVA coal when treated at 102°C^accord-
ing to Meyers process procedures, yields a coal containing less
than 4 Ibs S02/1Q6 Btu.
6. Non-equilibrium Gravichem separation of coal into four frac-
tions revealed that a greater yield of clean coal may be
attained than was noted for simple float/sink fractions (two
fractions).
7. Data from coal cleaning tests in the U.S.S.R. using ferric sul-
fate solution as the separation medium corroborates the TRW
data and verifies the applicability of Gravichem as a Meyers
process modification.
8. The amount of pyritic sulfur removal by chemical leaching is
not affected by Fe+++ and acid concentration under simultaneous
coal leaching-reagent regeneration conditions.
9. High (20%) H2S04 concentration does not change the response of
inorganic sulfur to the oxidizing reaction of the Meyers process.
10. Weathering of RTU coals during storage prior to processing was
observed. Weathering was manifested by reduction in the pyritic
sulfur content of the coal, by increase in sulfate sulfur con-
tent, and by formation of elemental sulfur. <
11. Elemental sulfur can be readily recovered from wet leached coal
by acetone.
158
-------
8. ENGINEERING DESIGN AND COST ESTIMATION
This section of the report documents the results of a process engineering
study performed in support of the current RTU Project. The primary objectives
of the process engineering effort were twofold:
1) To prepare conceptual full-scale process designs specific to
several coal feed stocks incorporating updated process data
from the RTU.
2) To evaluate the new Gravichem process option on a conceptual
full-scale basis combining data obtained at lab or bench-
scale with RTU plant data.
The detailed results of the current process engineering task are presented on
the following pages.
Section 8.1 contains a brief introduction and the historical background
which lead to the current chemical coal desulfurization process scheme. Also
described in this section is the generalized Gravichem Process concept. The
balance of Section 8.1 deals with the approach which was utilized in generat-
ing conceptual full-scale process designs, determining battery limits process
capital and operating costs, and evaluating integrated grass roots process
economics.
Section 8.2 details the specific base case designs, flow sheets, equip-
ment lists, battery limits costs, and integrated grass roots facility costs
and economics evaluations. Section 8.3 presents the data analysis conclusions
drawn from the engineering effort.
8.1 INTRODUCTION AND BACKGROUND
Throughout bench-scale development, processing techniques and their asso-
ciated costs have frequently been reviewed with an objective of focusing experi
mental effort in the areas of greatest cost sensitivity. The capital cost of
equipment required to perform the pyrite leaching must be carefully controlled
159
-------
to maintain a low processing cost per ton of coal product. A considerable
effort has been made to reduce costs in the core processing steps. While some
improvements in cost have been made in the basic Meyers process compared to
previous designs, the major emphasis has centered on obtaining the large cost
advantages possible by gravity separation of coal into clean (float) and high
pyrite (sink) fractions and chemically processing only the high pyrite frac-
tion. The goal has therefore been to obtain maximum benefit from the new
Gravichem float/sink technology presently under development. During the
course of the current engineering task, two coal specific process designs were
generated and evaluated. The designs incorporate Meyers process technology
into recently demonstrated gravity separation techniques (which utilize Meyers
process leach solution as a separation media) yielding improved desulfuriza-
tion schemes defined as the "Gravichem Process". The two designs presented
herein are specific to bench-scale and RTU tested mine cleaned Martinka coal
and bench-scale tested TVA Kentucky Number 9 coal.
8.1.1 Historical Background
Over the last six years, there have been several process concepts devel-
oped by TRW which involve Meyers' Coal Desulfurization Process Technology.
The earliest of the conceptual process designs was generated in 1972 and has
been reported in detail previously'4). That design was based on processing
a generalized 3.2% w/w pyritic sulfur ROM Lower Kittanning coal (pulverized
to -100 mesh) in a batch mode to remove 95% w/w of the contained pyritic sul-
fur. The specific design was based on early laboratory-scale test data and
the processing scheme represented a direct Meyers process treatment of coal.
The next series of designs were generated in 1975 and 1976(5'. Those
designs also related to the desulfurization of a generalized ROM Lower Kit-
tanning coal. However, the designs generated in that time period pertained
to continuous mode coarse coal processing (0.25-inch top-size) and continuous
mode fine coal processing (16 mesh x 0). Of the four design cases which were
developed, one represented a straight-through Meyers Process, one involved a
physical cleaning pretreatment and two of the schemes coupled physical clean-
ing, conventional gravity separation and Meyers process treatment.
The most recent of the reported bench-scale based design studies^11' was
completed in late 1977. That design was specific to mine cleaned Martinka
coal containing 1.2% w/w pyritic sulfur. The design was based on continuous mode
160
-------
operation utilizing 14 mesh x 0 coal. The overall process design incorporated
Meyers process treatment with the newly developed float/sink gravity separation
technology which utilized a 1.3 specific gravity Meyers process leach solution
as the liquid media and takes advantage of concurrent pyrite leaching. The
overall process is termed the Gravichem Process. This technology differs from
the previous float/sink Meyers Process option^ in that the conventional sep-
aration technique utilizes stabilized magnetite water solutions as the liquid
media thus requiring a separate processing circuit. A series of block dia-
grams is presented in Figure 33 which indicate the genesis of the current
process technology scheme.
8.1.2 Current Generalized Process Concept
As mentioned in the preceding paragraphs, the most recent generalized
coal desulfurization processing technique, Gravichem, utilizes the previously
demonstrated Meyers Process in conjunction with a newly developed leach solu-
tion gravity separation approach. A block diagram of the Gravichem Process
is presented as Figure 34. It should be noted that the block diagram presents
a processing concept which is generally applicable to any high pyrite contain-
ing suspendable coal (nominally 14 mesh top size). The scheme presented in
Figure 34 represents the current state of Gravichem Process technology. It is
based on past developmental experience^ ' ' ' as well as information obtained
during the current RTU program and its associated laboratory/bench scale sup-
port activities (discussed in Sections 6 and 7 of this report). The following
paragraphs describe the Gravichem Process (Figure 34) and discuss the intercon-
nections between each of the main process operations.
Ground coal, with a nominal top-size of 14 mesh, is mixed with hot recy-
cled leach solution. Process required make-up sulfuric acid is also added to
the mix tank such that the final leach solution specific gravity is maintained
at the desired level. After coal wetting and mixing is complete at the solu-
tion boiling temperature, the now partially reacted coal slurry is cooled and
introduced into a float/sink separator. The lighter low pyrite, low ash, coal
particles float to the top of the separator in a relatively dilute slurry
while the more dense high pyrite, high ash, coal particles in a much thicker
slurry are pumped from the bottom of the unit.
161
-------
1972 - MEYERS PROCESS (100 MESH TOP SIZE)
(REF. 4)
ROM UK. COAL'
100 T/HR COAL
3.2% PYRITIC SULFUR
20% ASH.
19.7 X 10° MM BTUAR"
MIXING
WASH
SULFUR
EXTRACTION
94.2 T/HR
16% PYRITIC SULFUR
15% ASH
19.4 X 10° MM BTUAR
1976 - CLEANED FINE COAL (IX MESH TOP SIZE)
(REF. 7)
ROML.K. COAL
120 T/HR COAL
20% ASH
»- 4% npmc SULFUR ^
PHYSICAL
CLEANING
DISCARD
MM BTUAR 20 T/HR COAL
75% ASH
10 - 14% .PYRITIC SUL
i.4x IO'MMBTUA
100 T/HR COAL
9% ASH
MEYERS PROCESS
FINE COAL
CONFIGURATION
1.5 -2% PYRITIC SULFUR
22.2X10° MM BTUAR
FUR
!
COAL PRODUCT
93 T/HR COAL
6% ASH
. 1% PYRITIC SULFUR.
21. 3 X 10° MM BTUAR
1976 - RUN-OF-MINE COARSE COAL (1/4 IN. TOP SIZE)
(REF. 7)
ROM L. K. COAL
100 T/HR COAL
20% ASH
3-4% PYRITIC SULFUR
19.7 X 10*MM BTUAR
MEYERS PROCESS
COARSE COAL
CONFIGURATION
COAL PRODUCT
85 T/HR COAL
15% ASH
.2% PYRITIC SULFUR
17.5 X 10* MM BTU/YR
no
1976 - DEEP CLEANED FINE AND COARSE COAL WITH 50% MEYERS PROCESS BYPASS (14 MESH TOP SIZE)
(REF. 7)
ROMl.K. COAL
240 T/HR COAL
20% ASH
3 - 4% PYRITIC SULFUR
A AST
PHYSICAL
CLEANING
DISCARD
47.2X10" MM BTU/YR jJT/MRCQ),
75% ASH
10 - 14% PYR
2.8 X 10° M(i
200 T/HR COAL
1.5-2% PYRITIC SULFUR
L 44.4 X 10* MM BTUAR
1TIC SULFUR
* BTUAR
CONVENTIOh
GRAVITY
SEPARATION
1
ML
100 T/HR COAL
MEYERS PROCESS
CONFIGURATION
3 - 4% PYRITIC
SULFUR
COAL PRODUCT _
85-90 T/HR COAL
10- 15% ASH
.2% PYRITIC SULFUR
BYPASS 100 T/HR COAL
LOW ASH
LOW SULFUR
185 - 190 T/HR COAL
6% ASH
.2% PYRITIC SULFUR
FINE ,
43.4 X 10° MM BTUAR
COARSE .
42.2X10° MM BTUAR
1977 GRAVICHEM PROCESSING WITH 55% FLOAT PRODUCT (14 MESH TOP SIZE)
(REF. 8)
MINE CLEANED
MARTI NKA COAL
225 T/HR COAL
14% ASH
1% PYRITIC SULFUR
46.4 X 10* MM BTUAR
*ROML.K. COAL - RUN-OF-MINE LOWER KITTANNING COAL
"8000 OFHMING HOURS PER YEAR, DRY COAL BASIS.
FLOAT _ 1 WATER 1
123 T/HR CO;
6% ASH
.26% PYRfTIC
28.1 X10*M
SINK
01 T/HR COAL
3.7% ASH
.53% PYRITIC
8.2 X 10° MH
^L | WASH [
SULFUR
M BTUAR
EXCEPT MIXING
SULFUR
K BTUAR
123 T/HR COAL
6% ASH
.26% PYRITIC SULFUR
28.1 X 10* MM BTUAR
98 T/HR COAL 223 T/HR COAL
22.2% ASH 13.2% ASH
. 15% PYRITIC SULFUR .21% PYRITIC SULFUR
IB. 1 X 10° MM BTUAR 46.2 X 10° MM BTUAR
Figure 33. History of the Gravichem Process
-------
01
to
COAL/IEACH
SOLUTION
MIXING
Figure 34. Gravtchem Process Block Diagram Suspendable Coal
(14 mesh x 0) Approach
-------
The clean (low pyrite, low ash) coal float fraction is next sent through
a series of filtrations and water contacting operations to wash the iron
sulfate-sulfuric acid leach solution from the coal. The washed coal cake is
then ready for final blending into the chemically desulfurized sink fraction.
The recombined fractions comprise the product coal stream. The float cake
wash water is supplied from internally process generated steam condensate and
make-up water. The concentrated leach solution filtrate is sent through a
regenerator where the ratio of Fe+3 to total Fe is adjusted via oxygen induced
oxidation. The regenerated reagent stream is then recycled back to the mixing
unit.
The high pyrite, high ash sink fraction from the separator is introduced
into a reaction vessel. It is here that the majority of the exothermic pyrite
leach reaction takes place at elevated temperatures and pressures in the pres-
ence of oxygen. The oxygen is simultaneously added to exothermally regenerate
(adjust the Fe to total iron ratio) the leach solution.
The regenerated, fully reacted slurry is then subjected to a filtration
and cake wash operation. The water used for sink cake washing is obtained
from internally generated process steam condensate and make-up water. The
concentrated leach solution filtrate is recycled directly to the mix tank.
The wash filtrate (very dilute leach solution) is split with the majority of
the steam going to the neutralization operation. It is during the neutraliza-
tion operation that lime is added to the dilute leach solution stream to effect
recovery of product iron and sulfate as a rusty gypsum. The remainder of the
wash filtrate is concentrated via evaporation, mixed with the concentrated
leach solution, and recycled to the mix tank. The steam generated during
evaporation/concentration is utilized within the battery limits process in
various process heating applications with the condensate being recycled as
cake wash water. Utility purchased high pressure steam is used as the heat
source for evaporation.
The water washed sink cake is next contacted with an appropriate organic
solvent, namely acetone, to effect elemental sulfur extraction from the cake.
The solvent rich coal slurry is then subjected to either filtration or centrif-
ugation, depending on specific process requirements. The solvent -miscible
water -elemental sulfur-sulfate sulfur solution is then separated from the
164
-------
acetone wet cake. The multi-component solution is then distilled under ele-
vated pressure and temperature such that molten elemental sulfur is recovered
for disposal. The recovered acetone is recycled to contacting while the sul-
fate laden water stream is sent to neutralization for sulfate removal and water
recycle to cake wash operations.
The solvent wet sink cake is next sent through a drying operation where
all but very small quantities of the acetone are flashed from the cake, con-
densed and recycled to solvent contacting. High pressure utility steam is
used to heat the circulation inert gas (N2) stripping media utilized in the
dryer. The hot dry sink cake is next blended with the cooler water wet float
cake to produce a low pyrite, reduced ash, Gravichem Process clean coal
product.
8.1.3 Base Case Design^ Cost Estimate, and Economics Approach
During the course of the current RTU project, two base case conceptual
full scale process designs were generated. Base Case 1 is a process designed
to treat mine cleaned Martinka coal, containing 1.51% w/w total sulfur (1.0%
w/w pyritic sulfur) containing 2.3 Ibs S02/MM Btu to a sulfur content level
equivalent to 1.2 Ibs S02/MM Btu. Base Case 2 is specific to the treatment
of TVA Kentucky No. 9 coal. The TVA coal process objective was to decrease
the initial 4.3% w/w total sulfur (2.4% w/w pyritic sulfur) containing
6.94 Ibs S02/MM Btu coal down to an equivalent sulfur content of 4 Ibs S02/MM
Btu.
8.1.3.1 Design Bases-
Presented in this section are the design bases and related assumptions
which were utilized in generating the two base case conceptual full scale
Gravichem Process designs. Unless otherwise noted, the information deline-
ated below is common to both base cases.
Feed Coal .
Base Case 1
• 14 mesh x 0 coal from the Martinka Mine, Lower Kittanning seam.
• 225 TPH total coal feed rate, chosen such that 100 TPH is
processed in leach reactor
165
-------
• Coal contains 1.51% w/w total sulfur with 1% w/w pyritic
sulfur
t 12,900 Btu/lb HHV
• 3% w/w moisture
Base Case 2
• 14 mesh x 0 TVA coal from Hopkins County, Kentucky Number 9
seam
t 200 TPH total coal feed rate such that 100 TPH is processed
in each leach reactor
• Coal contains 4.3% w/w total sulfur with 2.4% w/w pyritic
sulfur
• 12,400 Btu/lb HHV
• 10% w/w moisture
Product Coal.
t Base Case 1 combined float and sink coal to meet 1.2 Ib
SOp/MM Btu specification
• Base Case 2 combined float and sink coal to meet 4.0 Ib
S02/MM Btu specification
Mixing.
• Coal slurry mixed to 25% w/w coal based on requirements for
float/sink separation
• Coal mixed at 215°F for 45-minute residence time in agitated
vessel. Based on RTU experience for adequate wetting and
deaeration
• 3 stage mixer to ensure residence time
• Leach solution contains 7.5% w/w Fe and 4% H2$04 resulting
in a 1.3 specific gravity required for float/sink separation
Float/sink separation.
• One hour residence time in separator at 176°F based on bench-
scale studies discussed in Section 5
• Coal fractons based on bench-scale experimentation (Section 5)
Base Case 1 - 125 TPH float/100 TPH sink
Base Case 2 - 100 TPH float/100 TPH sink
166
-------
• Sink slurry is 33% w/w coal, float slurry is about 20% w/w
coal based on information from suppliers of coal cleaning
equipment and bench-scale studies
Reaction. The net overall reaction between pyrite and ferric sulfate
leach solution is represented by:
FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 + 10.2 FeS04 + 4.8 H2S04 + 0.8 S
AH = -55 Kcal/g mole Fe$2 = -0.10 MM Btu/lb mole Fe$2 reacted
The reaction rate, having second order dependence on the pyrite concentration
and fraction c
rate equation
and fraction of total iron as Fe+3, is represented by the following empirical
-d[W ]
rL - -1^ ' KL
where
W = wt % pyrite in coal at time t
Y = fraction of total iron as ferric ion at time t
K|_ = leach rate constant (function of temperature and coal top size)
KR = reactor constant (function of reactor size, configuration, etc.)
Base Case 1
• 6-hour residence time in reactor at 250°F and 35 psig based on
RTU experience and bench-scale determination of rate constants
(Section 5 and Reference 11). Reaction rates were found to be
higher in the RTU than those found at bench-scale thus requir-
ing the use of KR in the rate expression.
• Extent of reaction carried out in the mixer and separator is
based on rate constants determined at bench-scale adjusted
for the respective temperatures.
t Extent of ash removal on the leached coal is based on the
amount of pyrite reacted.
167
-------
Base Case 2
t 6-hour residence time at 250°F and 35 psig based on bench-
scale studies of TVA coal resulting in equal or faster reac-
tion rates than Martinka coal. Reaction rate constants were
adjusted by KR assuming the same increase in rate would be
observed at larger scale on the TVA coal as was observed on
Martinka coal at RTU scale.
• Extent of reaction in mixer and separator are based on
bench-scale determined rate constants adjusted for the
operating temperatures of the mixer and separator.
• Extent of ash removed is based on the amount of pyrite
reacted although bench scale data (Section 7) indicates
a removal of other ash on the processed sink coal. The
implications of this difference are discussed in Section 8.3.1,
Regeneration. The regeneration of the leach solution is represented by
the following reaction:
FeS04 + 0.5 H2S04 + 0.25 02 -> 0.5 Fe2(S02)3 + 0.5 H20
AH = -18.6 Kcal/g-mole FeS04 = 0.0335 MM Btu/lb-mole FeS04
• Regeneration in the reactor at 250°F and 35 psig maintains a
Y = 0.95 based on previously determined bench-scale data(4>5),
• 0.5-hour residence time in regenerator to maintain Y at 0.90
for recycled leach solution at the mixer based on bench-scale
studies(4,5).
Filtration and cake washing. Based on RTU and bench-scale experimentation
(Sections 7 and 8 and Reference 11) the filter/washing operation is modeled as
follows:
• At the interface on the filter drum between filtration and
cake washing the coal cake contains 50 pounds of water plus
its associated salt per 100 pounds of coal.
• 90% of the salt remaining on the cake at the interface is sur-
face salt which is removed by washing. The .other 10% of the
salt is in the coal pores and is not removed by cake washing,
• The minimum wash water required is equal to 1.4 times the
amount of water on the coal cake at the interface.
168
-------
• The "dry" coal cake contains 50 pounds water plus the residual
salt per 100 pounds of coal.
• Subsequent reslurrying of the coal cake with water dilutes the
salt contained in the coal pores to a homogeneous solution of
salt and water which can be removed by additional filtration
and washing.
Evaporation.
• Water removal carried out to the extent that recycled leach
solution contains 7.5% w/w iron.
• Evaporator operated at 290°F and 35 psig such the steam pro-
duced can be utilized for process heating requirements.
Solvent contacting.
• The water washed coal cake is contacted with acetone to make
up a slurry which is 40% by weight coal.
• Pore moisture and salt comes to equilibrium with the acetone,
based on bench-scale experimentation (Section 7).
Solvent separation.
Base Case 1
• Elemental sulfur on coal is reduced to 0.08% w/w based on
bench-scale studies (Section 1).
• Final separation of coal by centrifuge reduces acetone on
coal cake to 12% w/w based on vendor supplied information
(Bird Machine Co.).
Base Case 2
• Elemental sulfur on coal reduced to 0.2% w/w based on bench-
scale studies (Section 1).
• Final separation of acetone from coal by rotary pan filter
resulting in a coal cake containing 27% w/w acetone (Bird
Machine Co.).
Drying.
• Requires 3 driers each with 54 trays and 60-minute residence
time based on vendor testing and experience (Section 9).
• Inert gas temperature of 400°F heating coal to 225 F.
169
-------
• Residual dry coal acetone levels of 0.1% w/w were based on
vendor testing (discussed in Section 9) and additional manu-
facturer supplied information. While the vendor tests indi-
cated residual acetone levels of 0.5% w/w, Wyssmont Inc.
personnel indicated that a residual level of 0.1% w/w acetone
would be attainable through proper adjustment of dryer gas
recirculating rate and temperature, and solids residence time.
Solvent distillation.
• Column operated at 15 psig and 250°F to keep elemental sul-
fur molten in bottoms.
• Column diameter sized to maintain 5 ft/sec vapor flow.
• Column height is 65 feet using 20 trays based on suppliers
information (Artisan Industries).
Neutralization.
• Stoichiometric conversion of H-SO. and iron sulfates to
gypsum and iron oxides.
Materials of Construction.
Based on RTU experience, materials studies and supplier recommendations
the following materials were determined suitable for the various leach solution-
coal services:
t Fiber reinforced plastics and elastomers up to 180°F.
• 316 L stainless up to atmospheric boiling temperatures (215°F).
• Hastelloy C and rubber/acid brick lined carbon steel for ser-
vice up to 265°F.
• Titanium for service up to 300°F.
8.1.3.2 Battery Limits Process Costs--
For each of the base case designs generated, a battery limits capital and
operating cost was determined. Equipment lists, tabulating the required major
equipment were generated. The associated equipment numbers presented on the
lists are keyed to the process flow diagrams. The selected equipment was sized
to approach the optimum cost for processing the specific coal to the desired
product sulfur content.
Battery limits capital cost. Capital equipment costs adjusted to mid-
1978, were obtained from various sources: technical literature, equipment
suppliers and internal (TRW) costing data. The specific sources of data for
170
-------
the various classes of equipment are presented in Table 35. When cost data
were obtained from literature or other noncurrent sources, appropriate cost
escalation factors, based on the Marshall and Swift Equipment Cost Index (to
escalate costs from date of publication to mid-1978), were applied. The
capital equipment cost is also tabulated on the equipment list. The costs
are presented in terms of FOB equipment cost and installed equipment cost.
The FOB equipment cost is the base, uninstalled cost at point of manufacture
or point of shipment. The installed equipment cost includes the following
elements:
t 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
The installed equipment cost does not include a contingency factor.
Battery limits process operating cost. The process operating costs have
also been estimated for each of the base cost designs. The basis for these
estimates was technical literature and informal supplier quotes. Specific
sources of information are presented in Table 36. The total estimated battery
limits processing cost was determined as presented in Table 3Z.
171
-------
TABLE 35. SOURCES OF EQUIPMENT COST INFORMATION
Equipment Type
Information Source
Hoppers
Conveyors
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
Dryers and Support Equipment
Rotary Valves
Separators
TRW Data
Reference 13
Reference 14
Reference 15
TRW Data
TRW Data
Reference 15
TRW Data
Reference 15
Reference 15
Bird Machine Company
Ametec Company
Reference 15
Reference 16
Wyssmont Incorporated
Reference 15
TRW Data
8.1.3.3 Overall Process Economics--
The overall Gravichem Process economics were evaluated for each of the
base cases. The battery limit plants were incorporated with their associated
off-sites so that integrated grass roots processing facilities could be defined,
The following paragraphs describe the overall integrated Gravichem process
model as well as the bases for the economic analyses performed.
172
-------
TABLE 36. SOURCES OF OPERATING COST INFORMATION
. : Cost Element Information Source
Maintenance, Insurance, Taxes and Interest Reference 14
Labor Requirement (Number of Positions) Reference 14
Labor Cost TRW Data
Utilities
Electrical Power TRW Data
Cooling Water TRW Data
Process Water Reference 14
Materials
Oxygen Linde, Division of
Union Carbide
Nitrogen Linde, Division of
Union Carbide
Sulfuric Acid Reference 17
Lime Reference 17
Acetone Reference 17
Waste Disposal TRW Data
Integrated facility description. The integrated desulfurization facility
includes all battery limit Gravichem equipment plus the required off-sites.
The off-sites include such items as:
• Feed and product coal storage, handling and transport
equipment.
t Shallow physical coal cleaning facilities and size separation
equipment.
• By-product handling and storage facilities.
t Waste treatment (physical cleaning and process generated)
and storage facilities.
• Process water treatment, storage and pumping facilities.
• Cooling water treatment and pumping equipment.
173
-------
TABLE 37. BATTERY LIMITS OPERATING COSTS FORMAT
Capital Related Costs: Annual Cost, $1000
Depreciation - IQ% straight line XXXX
Maintenance, insurance, taxes XXXX
interest - 15% of capital
Labor:
Labor, XX operating positions XXXX
at $160,000/position
Utilities:
Electric power, XXX Kw XXXX
at 5^/Kw-hr
Cooling water, XXXX gpm, XXX
30°F rise at 5<£/1000 gal
Process water, XX gpm XX
at 25^/1000 gal
Heating, XX MM Btu/hr, fired duty,
XX TPH coal product equivalent
Materials:
Oxygen, XX TPH at $50/ton XXX
Nitrogen, XX TPH (included in oxygen cost)
Sulfuric acid, XX TPH at $55/ton X
Lime, XX TPH at $35/ton X
Acetone, X TPH at $320/ton X
Waste disposal, X TPH at $6/ton X
TOTAL COST XXXXXX
• Power and steam generation facilities
• Site office buildings and shop structures.
t Other site improvements such as roads, fences, railroad spurs,
etc.
174
-------
It should be noted that the economic evaluations presented herein do not include
land costs and assume that oxygen and nitrogen is purchased as an over-the-fence
utility item (i.e., neither battery limit nor off-site equipment include an oxy-
gen plant). For an integrated process facility of this type, it has been esti-
mated that the installed off-site capital cost will be equivalent to approximately
50% of the installed battery limits
Process economics model . The technique used in capitalizing the integrated
Gravichem process is the primary determinant in calculating the required unit
price of the product. Due to the nature of the Gravichem process (clean energy
production), only utility financing methods were considered in this evaluation.
The specific method utilized in this analysis is based on the technique used by
the Federal Power Commission Synthetic Gas-Coal Task Force in their report on
MO)
synthetic gasv . The recommended economic evaluation criteria are given in
Table 38.
The annual i zed revenue expressions for utility financing which were devel-
oped by the FPC Task Force can be combined to give a simplified expression of
the following form:
^
P = (0.140X + 0.121Y + 1.006Z)/E
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
E = annual energy output, MM Btu/yr
When this expression is utilized in conjunction with the criteria presented in
Table 38, the required sales price for the processed coal is obtained. From
the required sales price and the cost of the feed coal, an equivalent upgrading
cost can be determined.
175
-------
TABLE 38. ECONOMIC EVALUATION CRITERIA UTILITY FINANCING*
Operating Cost Criteria
Raw material - coal at $15/ton, $20/ton, $30/ton
Utilities
Electricity at 5^/kw-hr
Oxygen at $50/ton (nitrogen included in cost)
Cooling water at 5£/1000 gal.
Process water at 25
-------
TABLE 38. (Continued)
Capital Cost Criteria
Battery limits capital as discussed in text
Off-site capital at 50% of battery limits capital
Overhead and profit at 22% of battery limits + off-sites
Engineering and design at 10% of battery limits + off-sites
Contingency at 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 at 9% of total plant investment x 1.875
Start-up cost at 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
8.2 DESIGN AND COST ESTIMATE BASE CASES
The following pages present the results of two process engineering studies
relating to Gravichem application to two specific coal feedstocks. The coal
feedstocks were mine cleaned Martinka Coal and TVA Kentucky No. 9 coal. Each
of the base case designs are presented in terms of a brief process description,
process flow diagrams, a major equipment list, battery limits process costs,
and overall process economics. The detailed mass balance and their associated
flow diagrams as well as steam balances for each of the design cases are presented
in Appendix C.
8.2.1 Base Case 1 -Mine Cleaned Marti nka Coal
Base Case 1 deals with a conceptual full-scale Gravichem Process design
based on a processing facility capable of treating coal fed at a rate of 225
TPH from the Martinka Mine, Lower Kittanning seam. The coal has an "as
received" heating value of 12,900 Btu/lb. The coal contains 1.51% w/w total
sulfur with a ayritic sulfur level of 1.0% w/w. Process flow diagrams (Dwg.
Nos. 2473-3 and 4) for the design are presented as Figures 35a and 35b. Pre-
sented on the flovf diagrams are equipment, key process temperatures and pres-
sures, flow rates and composition of selected coal streams, and flow rates for
steam and raw materials.
177
-------
00
WASH FltJWrE CONTACIO» CONT*CIO« FILTIATE fILTM WMH FILl
MCEIVH M)XCK (ECEIVIR «CEIVt«
3NS/HK
3TAL SULFUR, %
SULFATt SULFUR. %
ELEMENt L SIAfUR, 1i
MOISTURE, %
HEM VALUE. BT
LB SO./W Wl
M COMKSITlONS (DitY aAilSl_
_P-Lr-z_p-a_p-i_ p^iLfiiS. F-IO
SLUBRV RfACTOFL REACTOR LEACH LEACH WASH FILTER FILTRATE WASH W,
FEED RED DISCHARGE SOLUTION FILTRATE FILTRATE FEED PUMP FEED PUT
PUMP PUMP PUMP FEED PUMP PUMP PUMP PUMP
WCQAL DESLRFUKiZATIOl
GRAV10HEM PBOCESS
NUMBM, W3-0)
Figure 35a. TRW Coal Desulfurization Gravichem Process
-------
10
— - -blS. £3! «-> £d T-IL T-ia coi T-U
ntJMK WASHFILIRATE CONTACTOR SLUWIV flLTIR TfLHATf CONTACTOR CENTRIFUGE CENTW1
MCIIVIR naiwt COOLER utxiv» RECBVEI
JErf JS-i t-\^ T-W^ T-17 Ma_ T;W_ MU .._ ..._, ^^^ ___ ^^
IIRimR ACETONE SIWP«» STRIPKR STRIfft* NEUTULIICR AGO NEUTkAUZK CONTACTOR CONTACTOR NEUTRALIZE!! CARE DM ADSORPTION SOUBMK STRIPPER
PtEIKAIER JTRIfPER REBDILER CONDENSATE BOTTOMS RED COOLER TANK MIXIR MIXER MIXED OKUM OVHHIAO
5K5T OK& CVCIONI S^RHWO ^NDEHSAJE
COMPENSER RECEIVE ft
SWAM NUMBED
TONS/™
TDTAlSUlfUlLIt
PVDTIC SUIPJR. %
OICANIC SULFUK, %
SIX FATE iltiFUS %
^o^feii™AI
i
:u
LIB
b
ITS
"ifS
7
'w
.08
"ijS
9
1
SKS
'^
SS3"""
IKW COAl WSULFUHIIATIO.
GMVICHEM PftOCISi
Figure 35b. TRW Coal Desulfurizatlon Gravichem Process
-------
8.2.1.1 Process Description--
Feed and mixer. Crushed coal, nominally 14 mesh top-size, is fed from
feed hopper A-l to the mix tank T-l by conveyor C-l. The coal, containing
1.0% w/w pyritic sulfur and 3% w/w moisture is fed at a rate of 225 TPH on a
dry basis. Recycled leach solution at its boiling point (215°F) is introduced
to the first mixer stage after passing through the knock-out drum T-2. Flash
steam from T-3 and low pressure steam are used to heat the coal on the conveyor
to 215°F. The mixer T-l was sized for three stages at 0.25-hour residence time
per stage to provide adequate wetting of the coal and partial reaction of the
pyri te.
Separator. The slurry from the mixer which is 25% w/w coal is pumped
through the slurry cooler to the separator S-l. The slurry is cooled to 176°F
in the slurry cooler E-l. The separator provides 1-hour residence time where
the heavy pyrite rich coal will settle and the lighter low pyrite coal will
float in the 1.3 specific gravity leach solution (see Section 7). Based on
current bench scale desulfurization studies, about 55% of the coal feed will
leave the separator as the float portion in a slurry which is about 21% w/w
coal. After washing to remove leach solution, this coal will meet the total
sulfur specification without further pyrite removal. The sink portion, con-
taining 45% of the coal feed, is pumped to the primary reactor, R-l, as a 33%
w/w coal slurry for further pyrite reduction.
Float coal washing and dewaterinq. The float slurry from the separator
S-l containing about 125 TPH coal is filtered, washed and recontacted with
water in a 3-stage countercurrent arrangement of .three filters (F-l, F-2 and
F-5) and two contactors (J-6 and T-22). The coal cake from the last stage con-
tains 33% w/w moisture and residual sulfate sulfur has been reduced to less
than 0.01% w/w coal. The coal cake from this filter is the product coal from
the float portion which will be recombined with the sink coal after it is pro-
cessed for pyrite removal. The float coal contains 0.83% w/w total sulfur with
a heating value of 14,290 Btu/lb, which is below the specification of 1.2 Ib
S02/106 Btu.
Primary reactor. The sink slurry from the separator containing about
ICO TPH coal is pumped to the primary reactor R-l where the leach reaction for
the removal of pyrite takes place. The reactor was sized to provide 6-hour
180
-------
residence-time at 250 F and 35 psig oxygen partial pressure which, based on
RTUexperimentations, will adequately remove pyrite to a level of 0.15% w/w
pyritic sulfur in the coal (see Section 6). The reactor is maintained at
250°F by the generation of heat from heats of reaction and regeneration and the
injection of 300 psig saturated steam into the reactor. The reactor is divided
into 5 stages each with a mixer. Simultaneous regeneration of the leach solu-
+3
tion to maintain a ratio of Fe to total iron equal to 0.95 is accomplished
by feeding about 2 TPH oxygen to the reactor. The inert gas leaving the
reactor is vented through a knock-out drum T-4 where it is contacted with
recycled leach solution before venting to the atmosphere.
Sink coal filtration and washing. The slurry effluent from the reactor
is flashed to about 1 psig and then pumped to the rotary drum vacuum filter
F-3 where the leach solution is separated and the coal cake washed to remove
residual sulfate salts. The wash water for the filter is recycled water from
the wash water hold tank T-8. Make-up water from the pond is required. The
coal cake from F-3 contains 33% w/w moisture.
Leach solution recycle loop. The wash filtrates from filters F-l and F-3
containing iron sulfates are collected and fed to the evaporator EV-1 for water
removal. In order to remove the excess iron produced from the leach reaction
a portion of the wash filtrate F-3 is sent to the lime neutralizer for disposal.
The evaporator feed stream is heated to 233°F in the feed/bottoms exchanger
E-3. The evaporator is operated at 35 psig and about 290°F such that the water
removed as steam may be used for process heating service. The leach solution
.in the evaporator bottoms is concentrated in iron to the extent that the
recycled leach solution will be 7.5% w/w total iron at the mixer T-l. The
filtrate from filter F-l is heated to 250°F in exchanger E-2 and fed to the
regenerator R-2. The regenerator R-2, operating at 250°F and 35 psig, is
designed for 0.5-hour residence time to regenerate the F-l filtrate such that
the combined recycle streams will have a ratio of Fe+3 to total iron equal to
0.90 at the mixer T-l. The oxygen requirement for the regenerator is about
0.25 TPH. The evaporator bottoms, regenerator effluent, F-3 filtrate and the
makeup sulfuric acid are combined and fed to the knock-out drum T-4, then
pumped to the mixer T-l through knock-out drum T-2, thus completing the
recycle loop.
181
-------
Acetone washing and elemental sulfur removal. The coal cake from filter
F-3 containing elemental sulfur produced in the leach reaction and residual
sulfate salts from the leach solution is contacted with about 100 TPH of ace-
tone in the contactor T-ll. The coal slurry is then filtered on a rotary pan
filter F-4 where the cake is washed with acetone. The coal cake from F-4 is
again recontacted with acetone (110 TPH) in T-13 and the slurry fed to a centri-
fuge CG-1 where the resulting coal cake is reduced to about 12% w/w acetone.
Bench-scale experience with two-stage acetone washing shows that the elemental
sulfur on the coal is reduced to 0.08% w/w (see Section 7). The sulfate sul-
fur on the cake is reduced to about 0.02$ w/w after the filter F-4. All of
the acetone contacting equipment are vented to the atmosphere through a water
scrubber GS-1 and carbon bed T-20 for residual acetone removal.
Acetone recovery and sulfur removal. The acetone streams from filter F-4
and the centrifuge CG-1 which contain elemental sulfur, water, and salts are
preheated in E-9 and fed to the acetone stripper SS-1. The overhead vapor
from the stripper, operating at 250°F and 15 psig, containing essentially all
of the acetone feed is condensed and recycled to the contactor T-ll. The
stripper bottoms containing water, elemental sulfur and salts is fed to the
stripper bottoms separator T-17 where the elemental sulfur is removed. The
sulfate rich solution is cooled by E-12 and fed to the neutralizer.
Lime Neutra1ization. The sulfate rich streams from the acetone stripper
and the wash filtrate from F-3 are contacted with a lime slurry in the acid
neutralizer T-18 where the sulfuric acid is neutralized and the salts reacted
to produce gypsum. The gypsum slurry is dumped to a settling pond for disposal,
Following settling, the pond water is returned to filter F-3 as part of the
wash water following water treatment to remove any trace acetone.
Coal drying. The coal cake from the centrifuge CG-1 is fed through a
rotary valve RF-2 to the coal drier D-l where the coal is heated to 225°F by
a 400°F nitrogen gas stream. The drier gas system is a closed loop with a
condenser to remove the acetone, a recycle gas blower B-l, and two steam
heaters E-7 and E-8 to heat the recycled nitrogen stream back to 400°F. The
recovered acetone is returned to the system for further use. The dried coal
from the drier is discharged through a rotary valve RV-3 and combined with the
coal product from the float portion. The coal from the drier contains 0.69%
182
-------
w/w total sulfur and has a heating value of 11,570 Btu/lb. The combined coal
product contains 0.77% w/w total sulfur with a heating value of 13,080 Btu/lb
resulting in 1.18 Ib S02/106 Btu material.
8.2.1.2 Battery Limit Costs—
Equipment list. The previous paragraphs of this document presented a con-
ceptual process design and flow diagram (Figures 35a and 35b) for a battery
limits processing facility treating 225 TPH of feed coal (221.8 TPH of desul-
furized coal product). An equipment list, tabulating the required major equip-
ment, is presented in Table 39. The selected equipment was sized to process
the 1.5% total sulfur feed coal such that the combined Gravichem product would
contain no more than 1.2 Ibs S02/MM Btu. As may be seen in Table 39, the total
battery limits FOB equipment cost is 9.6 million dollars while the total
installed equipment cost is 17.4 million dollars.
Battery limits process operating cost. Process operating costs have been
estimated as presented in Table 40. This estimated total annualized operating
cost results in an equivalent battery limit production cost of $6.61/ton of
coal product. This production cost is based on an apparent product coal rate
of 209.5 TPH (221.8 TPH of desulfurized product less 12.3 TPH for process
steam generation). The battery limit processing facility therefore operates
with an overall coal yield of 93% on a dry weight basis and an overall energy
yield of 94% on a Btu basis. Also of interest is the net overall plant energy
efficiency when the coal required for producing the electric power loads for
the process and oxygen plant is accounted for along with the plant steam
requirements. This efficiency, assuming a 33% efficiency for electric power
from coal, is equal to 93.5% where the coal requirement for power production
is 2.1 TPH. Another number of interest is the noncapitalized battery limits
operating cost. This is essentially the total process operating cost less the
annual depreciation charge (10% of installed equipment). The noncapitalized
battery limit operating cost is $5.18/ton of coal feed (20*/MM Btu equivalent)
or $5.57/ton of coal product (21tf/MM Btu equivalent). The noncapitalized oper-
ating cost represents a pure operating cost of the battery limit plant.
183
-------
TABLE 39. BASE CASE 1 - PROCESS EQUIPMENT LIST
00
.£»
MAJOR
A-l
B-l
C-l
CG-1
CY-1
D-l
E-l
E-2
E-3
E-5
E-6
E-8
E-9
E-10
E-ll
E-12
EV-1
PROCESS EQUIPMENT $9.64 MM FOB, $17.39 MM INSTALLED
Feed Hopper - 11,000 ft3
Recycle Gas Blower (3) - CS, 1.7 Compression Ratio, 40 hp
Feed Conveyor - 40 in. wide x 20 ft x 10 hp, 200 ft/min.
Centrifuge Package (2) - 44" * x 132" Screen Bowl, CS, 200 ph
Cyclone (3) - 5 psig, 850 ft3, CS
Coal Dryer (3) - 30' $ 55', CS, 5 psig, 15 hp, 54 trays, 60 min.
Slurry Cooler - 3550 ft2, CS/SS
Regenerator Feed Heater - 1840 ft2, CS/Hastelloy C
Evaporator Feed/Bottoms Exchanger - 4120 ft2, Hastelloy C/Hastelloy C
Slurry Cooler - 3400 ft2, CS/SS
p
Dryer Overhead Condenser (3) - 700 ft , CS/CS
Recycle Gas Heater (3) - 602 ft2, CS/CS
Stripper Preheater - 125 ft2, CS/SS
Stripper Reboiler - 7,560 ft2, Hastelloy C
Stripper Overhead Condenser (2) - 9,420 ft , CS/CS
Neutral izer Feed Cooler - 800 ft2, CS/SS
Evaporator - 2000 ft , 35 psig, CS with titanium clad, 20 hp
$K
FOB
22
16
599
2100
80
50
112
78
13
80
260
31
900
$K
Installed
26
as D-l
27
1197
as D-l
2625
217
136
303
212
as D-l
as D-l
48
223
780
90
1440
(Continued)
-------
oo
in
F-l
F-2
F-3
F-4
F-5
GS-1
M-l A/C
M-2 A/E
M-3
M-4
M-5
M-6
M-7
P-l
P-2
Rotary Drum Vacuum Filter Package (2) -
Rotary Drum Vacuum Filter Package (2) -
Rotary Drum Vacuum Filter Package -
Rotary Pan Filter Package (2) - 24' ,
Rotary Drum Vacuum Filter Package (2) -
Scrubber - 5' x 30' , 20 trays, 0 psig
Mix Tank Mixers (6) - 25 hp, SS
Reactor Mixers (5) - 190 hp, Hastelloy
Contactor Mixer (2) - 25 hp, SS
Contactor Mixer - 50 hp, SS
Contactor Mixer (2) - 20 hp, CS
Neutral izer Mixer - 15 hp, SS
Contactor Mixer (2) - 25 hp, SS
Slurry Feed Pump (2) -1500 gpm, 15 psi,
12' $ x 16', 608 ft2, SS,
130 hp
12' <(> x 16', 608 ft2, SS,
130 hp
12 <() x 24, 912 ft2, SS,
250 hp (inc. vac. sys.)
445 ft2, SS, 105 hp (inc. vac.
sys.)
12' x 16', 608 ft2, SS,
130 hp
, CS
C
15 hp, SS
Reactor Feed Pump - 1000 gpm, 15 psi, 10 hp, SS
$K
FOB
247
247
184
941
247
13
57
194
20
13
8
7
20
15
5
$K
Installed
392
392
292
1491
392
21
92
315
35
21
14
11
35
45
15
(Continued)
-------
TABLE 39. (Continued)
$K $K
FOB Installed
P-3 Reactor Discharge Pump - 1000 gpm, 15 psi, 10 hp, SS 5 15
P-4 Leach Solution Feed Pump (2) - 1100 gpm, 15 psi, 15 hp, SS 8 21
P-5 Leach Filtrate Pump (2) - 600 gpm, 65 psi, 30 hp, SS as F-l
P-6 Wash Filtrate Pump (2) - 300 gpm, 70 psi, 15 hp, SS as F-l
P-7 Filter Feed Pump (2) - 550 gpm, 10 psi, 5 hp, SS 4 13
P-8 Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS as F-2
£ P-9 CaS04 Slurry Pump - 320 gpm, 4 hp, SS, 5 psi 3 8
P-10 Wash Water Feed Pump - 825 gpm, 15 psi, 10 hp, CS 4 13
P-ll Leach Filtrate Pump - 500 gpm, 5 psi, 1.5 hp, SS as F-3
P-12 Wash Filtrate Pump - 250 gpm, 10 psi, 2 hp, SS as F-3
P-13 Contactor Slurry Pump - 1000 gpm, 15 psi, 15 hp, SS 6 17
P-14 Filtrate Pump (2) - 350 gpm, 7.5 psi, 3 hp, SS as F-4
P-15 Contactor Slurry Pump (2) - 550 gpm, 7.5 psi, 6 hp, CS 3 n
P-16 Acetone Centrate Pump (2) - 450 gpm, 7.5 psi, 5 hp, CS as CG-1
P-17 Acetone Return Pump (3) - 20 gpm, 0.5 hp, CS, 15 psi as D-l
P-18 Scrubber Water Return Pump - 150 gpm, 10 psi, 2 hp, CS 2 16
(Continued)
-------
TABLE 39. (Continued)
CXI
P-19
P-20
P-21
P-22
P-23
R-l
R-2
RV-1
RV-2
RV-3
S-l
SS-1
T-l
T-2
T-3
T-4
T-5
Acetone Condensate Pump - 1400 gpm, 15 hp, CS5 15 psi
Wash Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS
Filter Feed Pump (2) - 550 gpm, 10 psi, 5 hp, SS
Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS
Wash Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS
Reactor - 23' $ x 115', CS with acid-brick lining, 45 psig
Regenerator - 10.5' 4. x 63', 50 psig, CS with acid-brick lining
Rotary Valve (2) - 1 hp, 24" x 24", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Separator - 0.5 hr Residence Time
Acetone Stripper - 13' <|> x 65', SS, 20 Trays, 15 psig
Mix Tank (2) - 21' 4 x 41', SS, 0 psig, 102,000 gal
Knock-Out Drum (2) - 3200 gal, SS, 0 psig
Flash Drum - 6100 gal, 7' 4 x 21', SS, 2 psig
Knock-Out Drum - 13500 gal, 81 4 x 24' , SS, 0 psig
Wash Filtrate Receiver (2) - 1600 gal., Vac, SS
$K
FOB
5
4
808
219
39
25
25
855
192
288
38
40
42
$K
Installed
15
as F-2
13
as F-5
as F-5
1975
534
43
27
27
1710
327
504
87
91
81
as F-l
(Continued)
-------
TABLE 39. (Continued)
$K $K
FOB Installed
T-6 Contactor (2) - 25,000 gal, 0 psig, SS 130 260
T-7 Filtrate Receiver (2) - 1600 gal, Vac, SS as F-2
T-8 Wash Water Hold Tank - 72,000 gal, 0 psig, FRP 50 100
T-9 Filtrate Receiver - 2000 gal, Vac, SS as F-3
T-10 Wash Filtrate Receiver - 2500 gal, Vac, SS as F-3
T-ll Contactor - 40,000 gal, 0 psi, SS 81 185
T-12 Filtrate Receiver (2) - 1300 gal, Vac, SS as F-4
T-13 Contactor (2) - 20,000 gal, 0 psig, CS 33 66
£ T-14 Centrate Receiver (2) - 1300 gal, Vac, CS as CG-1
00
T-15 Dryer Condensate Receiver (3) - 140 gal, CS, 0 psig as D-l
T-16 Stripper Condensate Receiver - 9500 gal, CS, 15 psig 9 17
T-17 Stripper Bottoms Separator - 2700 gal, Hastelloy C, 15 psig 19 43
T-18 Acid Neutralizer Tank - 11,000 gal, FRP, 0 psig 7 15
T-19 Filtrate Receiver (2) - 3000 gal, Vac, SS as F-l
T-20 Carbon Adsorption Drum - 5' x 10', FRP, 0 psig 3 10
T-21 Wash Filtrate Receiver (2) - 1600 gal, Vac, SS as F-2
T-22 Contactor (2) - 25,000 gal, 0 psig, SS 130 260
T-23 Filtrate Receiver (2) - 1600 gal, Vac, SS ~ as F-5
T-24 Wash Filtrate Receiver - 1600 gal, Vac, SS as F-5
*CS/SS - Carbon Steel Sheet/Stainless Steel Tubes
-------
TABLE 40. BATTERY LIMIT PROCESS COSTS - BASE CASE 1
Capital Related Costs: Annual Cost, $1000
Depreciation - 10% straight line Ij739
Maintenance, insurance, taxes, ? cnq
interest - 15% of capital '
Labor:
Labor, 13 operating positions 2 nsn
at $160,000/position
Utilities:
Electric power, 2850 Kw at 5^/Kw-hr 1,128
Cooling water, 16,700 gpm, 30°F rise 428
at 5^/1000 gal
Process water, 216 gpm, at 25^/1000 gal 26
Heating, 321 MM Btu/hr, fired duty, 12.3 TPH
coal product equivalent
Materials:
Oxygen, 2.4 TPH at $50/ton 940
Nitrogen, 0.25 TPH (included in oxygen cost)
Sulfuric acid, 1.4 TPH at $55/ton 615
Lime, 2.6 TPH at $35/ton 728
Acetone, 0.1 TPH at $320/ton 253
Waste disposal, 9 TPH at $6/ton 428
TOTAL COST 10,974
8.2.1.3 Process Economics--
The equipment list (Table 39) presented earlier shows that a 225 TPH bat-
tery limit Gravichem process train has an installed cost estimated to be 17.4
million dollars. That battery limit plant when incorporated with its associ-
ated off-sites represents an integrated grass roots coal desulfurization facil-
ity. The off-sites for a plant of this type are estimated to be approximately
50% of the installed battery limits cost^ or in this case 8.7 million dollars.
189
-------
The results of the integrated grass roots Gravichem process economic eval-
uation are presented in Table 41. Calculations were performed using three
assumed run-of-mine (ROM) coal costs; $15, $20 and $30 per ton. These values
were selected since they represent the broad range of currently reported ROM
coal costs ($15/ton at mine mouth to $30/ton reported delivery price at some
plant sites).
As may be determined from the data (Table 41), the required market value
of the treated coal product ranges from $1.01/MM Btu with $15/ton ROM feed
coal to $1.66/MM Btu with $30/ton ROM feed material. An equivalent upgrading
cost is found by deducting the cost of the ROM "dirty" coal energy ($Q.58/MM
Btu, $0.78/MM Btu, and $1.16/MM Btu at $15, $20 and $30/ton ROM coal respec-
tively). The upgrading costs ranged from $0.43/MM Btu to $0.50/MM Btu when
ROM coal feed varied from $15/ton to $30/ton.
8.2.2 Base Case 2 - TVA Kentucky No. 9 Coal
Base Case 2 involves the desulfurization of TVA Kentucky No. 9 coal via
the Gravichem Process. The conceptual design described is based on a coal feed
rate of 200 TPH for TVA coal from Hopkins County, Kentucky Number 9 seam, with
a heating value of 12,400 Btu/lb. The coal contains 4.3% w/w total sulfur with
a pyritic sulfur level of 2.4% w/w. Process flow diagrams (Dwg. Nos. 2473-5
and 6) for the design are presented as Figures 36a and 36b. The flow diagrams
indicate equipment, key process temperatures and pressures, flow rates and
compositions of key coal streams, and steam and raw material flow rates.
8.2.2.1 Process Description--
Feed and mixer. 200 TPH (dry basis) coal are fed to the mixer T-l from
the feed hopper A-l by conveyor C-l after heating to 215°F on the conveyor by
flash steam and low pressure process steam. At the mixer the coal is contacted
with recycled leach solution at 215°F where the 0.75-hour residence time, in
three stages, provides thorough wetting of the coal and partial reaction of the
pyrite. The feed coal contains 10% moisture and 2.4% w/w pyritic sulfur on a
dry basis.
Separator. The mixer slurry containing 24% coal is pumped to the sepa-
rator S-l through the slurry cooler E-l where it is cooled to 176°F. The
190
-------
TABLE 41.
BASE CASE 1 - Pi?nrFSS Frflf)n
Annual Product, 1.66 MM Tons/yr
Energy Value, 43.4 x 106 MM Btu/yr
Capital Related Requi rements. $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)
Total Capital Related Costs (Utility)
Operating Costs. $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Sulfuric Acid, Acetone)
Supplies
Disposal
Utilities
Labor (19 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
Utility Financing
Upgrading Cost, $/MM Btu
Utility Financing
$15/Ton
17.39
8.70
5.74
2.61
5.17
39.61
6.68
7.40
7.04
60.73
26.70
1.60
0.88
0.43
3.09
2.83
1.47
37.00
ROM Coal Cost
$20/Ton
17.39
8.70
5.74
2.61
5.17
39.61
6.68
9.18
9.05
64.52
35.60
1.60
0.88
0.43
3.09
2.83
1.47
45.90
$30/Ton
17.39
8.70
5.74
2.61
5.17
39.61
6.68
12.74
13.07
72.10
53.40
1.60
0.88
0.43
3.09
2.83
1.47
63.70
1.01
0.43
1.22
0.44
1.66
0.50
* Assumes, debt/equity = 75/25, interest on debt = 9%, return on
equity = 15%.
**Equivalent to a plant capital investment of $68/kw.
191
-------
separator is sized for one-hour residence time where the heavier high pyrite
coal is separated as the sink fraction in the 1.3 specific gravity leach solu-
tion and the light low pyrite coal goes to the float fraction. 50% of the
coal separates as the sink portion in a slurry containing 33% w/w coal based
on current float/sink studies using TVA coal (Section 7). The sink slurry is
pumped to the reactor R-l for pyrite removal via the leach reaction. The float
coal slurry, containing about 20% w/w coal, meets the total sulfur specifica-
tion without further pyrite removal after washing and dewatering.
Float coal washing and dewatering. The 100 TPH coal in the float slurry
is fed to a 3-stage countercurrent arrangement of three filters (F-l, F-2 and
F-5) and two contactors (T-6 and T-22) where the leach solution is separated
from the coal and the coal water washed such that the residual sulfate salts
on the coal cake are less than 0.01% w/w sulfate sulfur. The coal cake from
the last filtration stage, containing 33% moisture and 2.56% w/w total sulfur
with a heating value of 13,830 Btu/lb, is well below the product specification
of 4.0 Ib S02/MM Btu.
Primary reactor. The 100 TPH coal in the sink slurry, containing 3.13%
w/w pyritic sulfur, is pumped to the primary reactor R-l where pyrite removal
by the leach reaction takes place at 250°F and 35 psig. Based on bench-scale
and RTU experimentation (Sections 6 and 7) the reactor is sized for six-hour
residence time where the pyritic sulfur is reduced to a level of 0.53% w/w
(dry basis). Simultaneous regeneration of the leach solution to maintain the
+3
ratio of Fe to total iron at 0.95 is accomplished by the injection of 3.9
TPH oxygen. The inert gas leaving the reactor contacts the recycled leach
solution in the knockout drum T-4 prior to venting to the atmosphere. The
heats of reaction and regeneration are removed in a pump-around loop where
the reactor slurry is exchanged with cold recycle leach solution in exchanger
E-4, maintaining the reactor temperature at 250°F.
Sink coal filtration and washing. The reactor slurry is cooled to 215°F
by flashing to about 1 psig then pumped to a rotary drum vacuum filter F-3
where the leach solution is separated for recycle and the coal cake washed
for the removal of sulfate salts. Wash water is provided from condensate of
low pressure steam and recycled pond water. The coal cake from F-3 contains
33% w/w moisture.
192
-------
Leach solution recycle loop. The leach solution filtrate stream from both
the float and sink coal filters are recycled to the mixer T-l. The wash fil-
trates from F-l and F-3 are dilute in iron sulfates and must be dewatered in
the evaporator EV-1. A portion of the F-3 wash stream is pumped to the neutral-
izer in order to remove from the system the excess iron produced in the leach
reaction. The feed to the evaporator is preheated to 264°F by the feed/bottoms
exchanger E-3 and the evaporator feed preheater E-9. The evaporator is operated
at 35 psig and 290 F such that the steam produced can be utilized for process
heating service and the condensate recycled as filter wash water. The evapo-
rator bottoms are concentrated in salt to the extent that the total leach solu-
tion recycle stream at the mixer T-l is 7.5% w/w iron. The F-l filtrate stream
is heated to 237°F in exchangers E-4 and E-2 and fed to the regenerator R-2,
operating at 35 psig and 250°F, where 0.5-hour residence time is provided to
regenerate the leach solution to the extent that the combined recycle streams
have a ratio of Fe to total iron equal to 0.90 at the mixer T-l. The oxygen
requirement for regeneration is 0.92 TPH. The evaporator bottoms, regenerator
effluent, F-3 filtrate and 2.8 TPH makeup sulfuric acid are combined and fed
to the knockout drum T-4, then pumped to the mixer T-l through knockout drum
T-2 thus completing the recycle loop.
Acetone washing and elemental sulfur removal. The sink coal cake from
filter F-3 is fed to the contactor T-ll where it is slurried with acetone to
dissolve the elemental sulfur produced in the leach reaction. The resulting
402 coal slurry is then filtered on a rotary pan filter F-4 where the cake is
washed with acetone. The elemental sulfur on the coal cake is reduced to
about 0.2% w/w based on bench-scale studies (Section 7) and the sulfate sulfur
further reduced to about 0.02% w/w. The acetone contacting equipment are
vented to the atmosphere through a water scrubber GS-1 and a carbon bed T-20
for removal of acetone vapors.
Acetone recovery and sulfur removal. The acetone-water filtrate from fil-
ter F-4 containing elemental sulfur and sulfate salts are fed to the acetone
stripper SS-1 operating at 250°F and 15 psig. The stripper bottoms consisting
of water, molten elemental sulfur and sulfate salts are fed to the bottoms sep-
arator where the molten sulfur is removed and the water phase decanted. The
sulfate laden water phase is cooled by E-12 and fed to the neutral izer. The
193
-------
stripper overhead vapor containing essentially all of the acetone feed is con-
densed and recycled to the contactor T-ll.
Lime neutralization. The bottoms stream from the acetone stripper and a
portion of the F-3 wash filtrate are removed from the system at the neutralizer
in order to maintain the iron concentration in the leach solution at balance.
The neutralizer feed streams are contacted with lime forming gypsum and iron
oxides and pumped to the settling pond for solids removal. The solids-free
pond water is treated and recycled to the system as wash water for the filters.
Coal drying. The acetone wet coal cake from filter F-4 containing about
27% acetone is fed to the coal drier F-l where the coal is heated to 225°F and
the acetone flashed by a 400°F stream of nitrogen gas. The acetone is condensed
from the gas stream, which is a closed loop, and returned to the system for cake
washing. The gas stream is compressed and reheated to 400°F by blower B-l and
recycle gas heaters E-7 and E-8. The dried coal cake is discharged from the
drier and recombined with the float coal cake. The dried sink coal portion
contains 2.5% w/w total sulfur and has a heating value of 11,990 Btu/lb. The
combined coal product contains 2.53% total sulfur with a heating value of
12,930 Btu/lb resulting in a coal with 3.91 Ib S02/MM Btu.
8.2.2.2 Battery Limit Process Costs—
Equipment list. The previous section of this document presented a concep-
tual process design and flow diagram (Figures 36a and 36b) for a battery limits
processing facility treating 200 TPH of feed coal (193.3 TPH of desulfurized
coal product). The process was designed to reduce the 4.3% w/w total sulfur
feed coal to a combined Gravichem product containing no more than 4.0 Ibs
S02/MM Btu. An equipment list, tabulating the required major equipment, is
presented as Table 42. The total battery limits FOB equipment cost was deter-
mined to be 8.7 million dollars while the total installed equipment cost is
15.3 million dollars.
Battery limits process operating costs. The total estimated battery limit
processing cost is presented in Table 43. As may be determined from Table 43,
the estimated total annualized operating cost results in an equivalent battery
limit production cost of $8.63/ton of coal product. This production cost is
based on cleaned coal production 185.5 TPH (193.3 TPH of desulfurized product
194
-------
10
in
Figure 36a. TRW Coal Desulfurization Gravichem Process
-------
a*
Figure 36b. TRW Coal Desulfurizatlon Gravichem Process
-------
TABLE 42. BASE CASE 2 - PROCESS EQUIPMENT LIST
MAJOR PROCESS EQUIPMENT $8.67 MM FOB, $15.31 MM INSTALLED
A-l
B-l
C-l
CY-1
D-l
E-l
E-2
E-3
E-4
E-5
E-6
E-7
E-8
E-9
E-10
E-ll
Feed Hopper - 10,000 ft3
Recycle Gas Blower (3) - CS, 1.7 Compression Ratio, 85 hp
Feed Conveyor - 40 in. wide x 20 ft x 10 hp, 200 ft/min.
Cyclone (3) - 5 psig, 850 ft3, CS
Coal Dryer (3) - 30' <(> 55', CS, 5 psig, 15 hp, 54 trays, 60 min.
Slurry Cooler - 3185 ft2, CS/SS
Regenerator Feed Heater - 1040 ft , CS/Hastelloy C
Evaporator Feed/Bottoms Exchanger - 2890 ft , Hastelloy C/
Hastell oy C
2
Reactor Pumparound/Regenerator Feed Exchanger - 1360 ft , Hastelloy
C/ Hastelloy C
Slurry Cooler - 3400 ft2, CS/SS
Dryer Overhead Condenser (3) - 1770 ft2, CS/CS
Recycle Gas Preheater (3) - 1500 ft2, CS/SS
Recycle Gas Heater (3) - 1050 ft2, CS/CS
2
Evaporator Feed Preheater - 1140 ft , CS/Hastelloy C
Stripper Reboiler - 3120 ft2, Hastelloy C
Stripper Overhead Condenser - 7421 ft , CS/CS
$K
FOB
22
as
16
as
2450
76
37
95
65
78
as
as
as
37
51
114
$K
Installed
26
D-l
27
D-l
3062
205
102
258
177
212
D-l
D-l
D-l
102
142
362
(Continued)
-------
TABLE 42. (Continued)
E-12
E-13
EV-1
F-l
F-2
-, F"3
00
F-4
F-5
GS-1
M-l A/C
M-2 A/E
M-3
M-4
M-6
M-7
Neutral izer Feed Cooler - 800 ft2, CS/SS
Pondwater Heater - 50 ft2, CS/SS
2
Evaporator - 1400 ft , 35 psig, CS with titanium clad
Rotary Drum Vacuum Filter Package - 12' x 24'
250 hp (inc
Rotary Drum Vacuum Filter Package - 12' x 24'
250 hp (inc
Rotary Drum Vacuum Filter Package - 12' $ x 24'
250 hp (inc
Rotary Pan Filter Package (2) - 24' 4, 445 ft2,
(inc. vac. sys.
Rotary Drum Vacuum Filter Package - 12' $ x 24'
250 hp (inc
Scrubber - 5' 4 x 30', 20 trays, 0 psig, CS
Mix Tank Mixers (6) - 25 hp, SS
Reactor Mixers (5) - 190 hp, Hastelloy C
Contactor Mixer - 50 hp, SS
Contactor Mixer - 50 hp, SS
Neutral izer Mixer - 15 hp, SS
Contactor Mixer - 50 hp, SS
, 912 ft2, SS,
. vac . sys . )
, 912 ft2, SS,
. vac. sys.)
, 912 ft2, SS,
. vac. sys.)
SS, 105 hp
)
, 912 ft2, SS,
. vac. sys.)
$K
FOB
31
8
722
184
184
184
941
184
13
57
194
13
13
7
13
$K
Installed
90
25
1155
292
292
292
1491
292
21
92
315
21
21
11
21
(Continued)
-------
TABLE 42. (Continued)
$K $K
FOB Installed
P-l
P-2
P-3
P-4
P-5
P-6
P-7
P-8
P-9
P-10
P-ll
P-12
P-13
P-14
P-17
P-18
P-19
Slurry Feed Pump (2) - 1350 gpm, 15 psi, 15 hp, SS
Reactor Feed Pump - 1000 gpm, 15 psi, 10 hp, SS
Reactor Discharge Pump - 960 gpm, 15 psi, 10 hp, SS
Leach Solution Feed Pump (2) - 960 gpm, 15 psi, 15 hp, SS
Leach Filtrate Pump - 1120 gpm, 75 psi, 75 hp, SS
Wash Filtrate Pump - 400 gpm, 70 psi, 20 hp, SS
Filter Feed Pump - 850 gpm, 10 psi, 7.5 hp, SS
Filtrate Pump - 340 gpm, 7.5 psi, 3 hp, SS
CaSO^ Slurry Pump - 320 gpm, 2 hp, SS, 5 psi
Wash Water Feed Pump - 730 gpm, 15 psi, 10 hp, CS
Leach Filtrate Pump - 500 gpm, 5 psi, 1.5 hp, SS
Wash Filtrate Pump - 250 gpm, 10 psi, 2 hp, SS
Contactor Slurry Pump - 1000 gpm, 15 psi, 15 hp, SS
Filtrate Pump (2) - 350 gpm, 7.5 psi, 3 hp, SS
Acetone Return Pump (3) - 65 gpm, 1.5 hp, CS, 15 psi
Scrubber Water Return Pump - 150 gpm, 10 psi, 2 hp, CS
Acetone Condensate Pump - 540 gpm, 7.5 hp, CS, 15 psi
15 45
5 15
5 15
8 21
as F-l
as F-l
5 18
as F-2
3 8
4 13
as F-3
as F-3
6 17
as F-4
as D-l
2 16
3 9
(Continued)
-------
TABLE 42. (Continued)
o
o
P-20
P-21
P-22
P-23
P-24
R-l
R-2
RV-1
RV-2
RV-3
S-l
SS-1
T-l
T-2
T-3
T-4
T-5
T-6
Wash Filtrate Pump - 370 gpm, 7.5 psi, 5 hp, SS
Filter Feed Pump - 850 gpm, 10 psi, 7.5 hp, SS
Filtrate Pump - 340 gpm, 7.5 psi, 3 hp, SS
Wash Filtrate Pump - 370 gpm, 7.5 psi, 5 hp, SS
Reactor Pumparound Pump - 325 gpm, 15 psi, 5 hp, Hastelloy C
Reactor - 23' $ x 115', CS with acid-brick lining, 45 psig
Regenerator - 10' $ x 61', 50 psig, CS with acid-brick lining
Rotary Valve (2) - 1 hp, 24" x 24", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Separator - 0.5 hr Residence Time (Acid Brick)
Acetone Stripper - 8' x 54' , SS, 20 Trays, 15 psig
Mix Tank (2) - 20' $ x 40', SS, 0 psig, 91,330 gal
Knock-Out Drum (2) - 3200 gal, SS, 0 psig
Flash Drum - 6100 gal, 7' x 21', SS, 2 psig
Knock-Out Drum - 13,500 gal, 8' 4, x 24', SS, 0 psig
Wash Filtrate Receiver - 2000 gal, Vac, SS
Contactor - 40,000 gal, 0 psig, SS
$K
FOB
as
5
as
as
4
808
219
39
25
25
855
139
274
38
40
42
as
81
$K
Installed
F-2
18
F-5
F-5
12
1975
534
43
27
27
1710
236
480
87
91
81
F-l
185
(Continued)
-------
TABLE 42. (Continued)
$K $K
FOB Installed
T-7 Filtrate Receiver - 2000 gal, Vac, SS as F-2
T-8 Wash Water Hold Tank - 40,000 gal, 0 psig, FRP 34 67
T-9 Filtrate Receiver - 2000 gal, Vac, SS as F-3
T-10 Wash Filtrate Receiver - 2500 gal, Vac, SS as F-3
T-ll Contactor - 40,000 gal, 0 psig, SS 81 185
T-12 Filtrate Receiver (2) - 1300 gal, Vac, SS as F-4
T-15 Dryer Condensate Receiver (3) - 450 gal, CS, 0 psig as D-l
2 T-16 Stripper Condensate Receiver - 3400 gal, CS, 15 psig 5 9
T-17 Stripper Bottoms Separator - 2700 gal, Hastelloy C, 15 psig 19 43
T-18 Acid Neutralizer Tank - 15,000 gal, FRP, 0 psig 9 19
T-19 Filtrate Receiver - 5500 gal, Vac, SS as F-l
T-20 Carbon Adsorption Drum - 5' <|> x 10', FRP, 0 psig 3 10
T-21 Wash Filtrate Receiver - 2500 gal, Vac, SS as F-2
T-22 Contactor - 40,000 gal, 0 psig, SS 81 185
T-23 Filtrate Receiver- 2000 gal, Vac, SS as F-5
T-24 Wash Filtrate Receiver - 2500 gal, Vac, SS as F-5
CS/SS - Carbon Steel Sheet/Stainless Steel Tubes
-------
TABLE 43. BATTERY LIMITS PROCESS COSTS - BASE CASE 2
Capital Related Costs: Annual Cost, $1000
Depreciation - 10% straight line 1,531
Maintenance, insurance, taxes, 2,296
interest - 15% of capital
Labor:
Labor, 12 operating positions at 1,920
$160,000/position
Utilities:
Electric power, 2280 Kw at 5<£/Kw-hr 902
Cooling water, 11,500 gpm, 30°F rise at 273
5£/1000 gal
Process water, 100 gpm at 25<£/1000 gal 12
Heating, 201 MM Btu/hr, fired duty, 7.8 TPH
coal product equivalent
Materials:
Oxygen, 4.8 TPH at $507ton 1,909
Nitrogen, 0.53 TPH tincluded in oxygen cost)
Sulfuric acid, 2.7 TPH at $55/ton 1,194
Lime, 5.5 TPH at $35/ton 1,530
Acetone, 0.1 TPH at $3207ton 253
Waste disposal, 18 TPH at $6/ton 846
TOTAL COST ^ 12,666
less 7.8 TPH for process steam generation). The battery limit processing facil-
ity therefore operates with an overall coal yield of 93% on a dry weight basis
and an overall energy yield of 97% on a Btu basis. The coal required for pro-
cess and oxygen plant electric power consumption is 2.1 TPH resulting in a
95.6% net overall energy efficiency when included with the coal usage for plant
steam generation. The noncapitalized battery limit operating cost, the total
202
-------
process operating cost less the annual depreciation charge (10% of installed
equipment), is $7.03/ton of coal feed (28<£/MM Btu equivalent) or $7.58/ton of
coal product (29<£/MM Btu equivalent). This noncapitalized operating cost essen-
tially represents a pure battery limit operating cost.
8.2.2.3 Process Economics—
The Base Case 2 design, as discussed earlier, resulted in an installed
battery limit plant costing 15.3 million dollars, Using rational presented
in Section 8.2.1.3, it is therefore expected that the associated off-sites
would cost an additional 7.7 million dollars. The resultant integrated grass
roots facility economics are then determined to be as shown in Table 44.
As may be seen, the required market value of the coal product ranges from
$1.09/MM Btu with $15/ton ROM feed coal to $1.75/MM Btu with $30/ton ROM feed
material. The equivalent upgrading cost, found by deducting the cost of the
ROM "dirty" coal energy ($0.60/MM Btu, $0.81/MM Btu, and $1.21/MM Btu at $15,
$20 and $307ton ROM coal respectively) from the required market value, ranges
from $0.49/MM Btu to $0.54/MM Btu.
8.3 ENGINEERING ANALYSIS CONCLUSIONS
As discussed earlier, two conceptual full-scale process designs and cost
estimates were generated based on the most current technical data available.
Each of the two design packages, termed "Base Cases", were aimed at evaluating
the recently conceived Gravichem Process as applied to two differing coals.
Furthermore, each Base Case deals with desulfurizing coal to different residual
sulfur levels. Base Case 1 is representative of Gravichem application to a
relatively low sulfur coal (1.51% w/w total sulfur with 1.035 w/w pyritic sulfur)
when processed to a residual total sulfur content equivalent to 1.2 Ibs S02/MM
Btu. Base Case 2 is typical of Gravichem application to a relatively high sul-
fur coal (4.3% w/w total sulfur with 2.4% w/w pyritic sulfur) when processed
to the less stringent equivalent sulfur limit of 4 Ibs S02/MM Btu.
8.3.1 Base Case Results
Table 45 presents a summary of the Gravichem process economics as they
relate to processing low (Base Case 1) and high (Base Case 2) sulfur content
coals. It should be noted that the economics picture, as presented, is specific
203
-------
TABLE 44. BASE CASE 2 - PROCESS ECONOMICS'
Annual Product, 1.46 MM Tons/yr
Energy Value, 38.0 x 106 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)
Total Capital Related Costs (Utility)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Sulfuric Acid, Acetone)
Supplies
Disposal
Utilities
Labor (18 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
Utility Financing
Upgrading Cost, $/MM Btu
Utility Financing
$15/Ton
15.31
7.66
5.05
2.30
4.55
34.87
5.88
7.08
6.35
54.18
23.76
2.98
0.79
0.85
3.10
2.61
1.29
35.38
1.09
0.49
ROM Coal Cost
$20/Ton
15.31
7.66
5.05
2.30
4.55
34.87
5.88
8.66
8.13
57.54
31.68
2.98
0.79
0.85
3.10
2.61
1.29
43.30
1.31
0.50
$30/Ton
15.31
7.66
5.05
2.30
4.55
34.87
5.88
11.83
11.71
64.29
47.52
2.98
0.79
0.85
3.10
2.61
1.29
59.14
1.75
0.54
*Assumes, debt/equity = 75/25, interest on debt = 9%, return on
equity = 15%.
**
Equivalent to a plant capital investment of $69/kw.
204
-------
TABLE 45. SUMMARY OF BASE CASE ECONOMICS RESULTS
ro
o
Base Case 1
Cost -Marti nka Coal-
225 TPH Coal Feed Basis
ROM Coal Cost 15 20 30
$/Ton
Battery Limit 17.4 17.4 17.4
Capital - $ MM
Total Plant 39.6 39.6 39.6
Investment - $ MM
Total Operating 37.0 45.9 63.7
Costs - $ MM/Yr
Required Product 1.01 1.22 1.66
Market Price - $/MM Btu
Upgrading Cost .43 .44 .50
$/MM Btu
Base Case 2** Adjusted Base Case 2***
-Kentucky No. 9 Coal- -Kentucky No. 9 Coal-
200 TPH Coal Feed Basis 225 TPH Coal Feed Basis
15 20 30 15 20
15.3 15.3 15.3 17.2 17.2
34.9 34.9 34.9 39.2 39.2
35.4 43.3 59.1 39.8 48.7
1.09 1.31 1.75 1.09 1.31
.49 .50 .54 .49 .50
30
17.2
39.2
66.5
1.75
.54
**
Mine cleaned Martinka coal, 1.51% w/w total sulfur as fed, 1.2 Ibs S02/MM Btu as processed, 1.66 MM TPY product,
43.4 X 106 MM Btu/yr energy value.
TVA supplied Kentucky No. 9 coal, 4.3% w/w total sulfur as fed, 4 Ibs S02/MM Btu as processed, 1.46 MM TPY product,
3ft n Y ir>6 MM Rtu/vr *>nerav value.
38.0 X 105 MM Btu/yr energy value.
*** Base Case 2 adjusted to an equivalent 1.64 MM TPY product, 42.8 X 106 MM Btu/yr energy value.
-------
to utility financing under the ground rules as detailed in Section 8.1.3.3. As
may be seen from the table, the resultant upgrading costs range from $0.43/MMBtu
to $0.50/MM Btu for the low sulfur coal case while the range is $0.49/MM Btu
to $0.54/MM Btu for the higher sulfur Kentucky coal case. The low sulfur Base
Case 1, which requires 0.79% w/w pyritic sulfur removal, has essentially the
same capital requirement as Base Case 2 which effects a 1.88% w/w pyritic sul-
fur removal. This finding is substantiated when the Base Case 2 economics are
adjusted to reflect comparable plant throughputs. The adjusted Base Case 2
economics are also presented on Table 45.
The apparent higher processing costs (upgrading cost) associated with
Base Case 2 are generally the direct result of reacting and removing greater
quantities of pyrite. The additional pyrite removal required by the Kentucky
No. 9 coal (1.88% w/w as compared to 0.79% w/w for the Martinka coal) requires
a directly proportional increase in certain operating costs. Specific costs
which increased in Base Case 2 are associated with oxygen consumption as a
regenerant, lime utilization for neutralization of sulfate, sulfuric acid
makeup, and sulfur and gypsum disposal.
Bench-scale processing of Kentucky No. 9 sink (Section 7), performed dur-
ing the latter stages of the design effort, indicate that ash other than pyrite
is also leached from the coal during processing. The bench-scale data shows
that the sink processed material will have an ash content ranging from 9.5%
to 10.5% w/w while the process design material balance, which considers only
pyrite removal, shows an ash content of 16.3% w/w. Incorporating this inform-
ation into the overall process design and assuming a 10% ash content in the
processed sink would result in a heating value of 12,890 Btu/lb for the sink
coal. The combined float/sink product would therefore have an ash content of
6.5% w/w with a heating value of 13,390 Btu/lb. The net energy output of the
plant would remain unchanged although processing costs would increase some-
what owing to increased neutralization and disposal costs for the removal of
the additional ash. This new information would, however, indicate that a sig-
nificantly cleaner (less ash) and therefore more desirable, desulfurized coal
product would be generated than is indicated by the Base Case 2 process design.
8.3.2 Cost Sensitivity Analysis
In addition to each of the specific Base Case studies (Sections 8.2.1
and 8.2.2), a parametric analysis relating to process economics was performed
206
-------
to evaluate cost sensitivity. As part of the sensitivity study, three process
cost related elements were varied. For each Base Case, installed battery limit
capital cost was varied to evaluate the dependence of the bottom line upgrading
cost on equipment cost estimates. The assumed battery limits cost was allowed
to vary by approximately ±50% while all operating costs were held constant.
Figures 37 and 38 present the effects of battery limit capital on upgrading
costs as a function of ROM coal cost for each of the Base Cases. Figures 39
and 40 are cross plots showing the effects of ROM coal cost on upgrading cost
as a function of battery limit capital. As may be seen from the Figures,
upgrading costs can vary from approximately -9i/m Btu to +10tf/MM Btu over the
range of ±50% of the Base Case battery limit capital for any given ROM coal
cost. This holds true for both high and low sulfur coal processing.
Another variable evaluated was that of percent pyrite or pyritic sulfur
removed. In this study, the feed coal for each Base Case was varied with
respect to starting pyrite content. The resultant pyritic sulfur removal
requirements were then determined in order to meet the appropriate residual
sulfur levels for each of the cases. In this evaluation, the capital costs
were assumed to remain constant and the appropriate operating costs (oxygen,
lime, sulfuric acid, and disposal) were modified. The results of that eval-
uation are presented in Figures 41, 42, 43 and 44. Figures 41 and 42 present
the upgrading costs for each case as a function of percent pyritic sulfur
removed for each of the three ROM coal costs. Figures 43 and 44 indicate the
upgrading costs as a function of ROM coal cost at several pyritic sulfur
removal levels. The sulfur removal levels ranged from 0.5% w/w to 4% w/w
for each case. As may be seen from the Figures, the resultant upgrading costs
vary significantly with the amount of pyritic sulfur removed. In fact, over
the range evaluated for both the high and low sulfur cases, the upgrading
costs increase by 65% to 77% of the 0.5% w/w removal values as the amount
of pyritic sulfur removed increases to 4% w/w.
The third variable evaluated was the percent float which is allowed to,
in essence, bypass the conventional Meyers Process. For each Base Case, the
float/sink ratio was allowed to vary from 33/66 to 66/33 (Base Case 1 is a
55/45 split while Base Case 2 is a 50/50 split). For this analysis, essen-
tially all operating costs were held constant while the capital related costs
207
-------
.60
55
.50
0.45
O
z
Q
o
Q.
D
.40
,35
,30
ROM COAL
$AON
BASE CASE 1 $17.4 MM
10 12 14 16 18 20 ,22
BATTERY LIMIT CAPITAL $/MM
24 26
Figure 37. Base Case 1. Upgrading Cost
vs. Battery Limit Capital
208
-------
ROM COAL
$/TON
BASE CASE 2 $15.31 MM
10 12 14 16 18 20
BATTERY LIMIT CAPITAL $MM
22 24
Figure 38. Base Case 2. Upgrading Cost
vs. Battery Limit Capital
209
-------
,60
,55
fe
.50
<-> .45
O
Z
.40
.35
.30
BATTERY LIMIT
CAPITAL COST
$25 MM
BASE CASE 1
$17.4 MM
$17 MM
$10 MM
0 10 20 30
ROM COAL COST $/TON
Figure 39. Base Case 1. Upgrading Cost
vs. ROM Coal Cost
210
-------
.65,
BATTERY LIMIT
CAPITAL COST
.60
.55
VJ-
to
o
U
O
z
5 .50
o
Q_
.45
.401
$23 MM
$15.3 MM
(BASE
CASE 2)
$8 MM
0 10 20
ROM COAL COST - $/TON
30
Figure 40. Base Case 2. Upgrading Cost
vs. ROM Coal Cost
211
-------
ROM COAL
COST $/TON
BASE CASE 1
.78% SP REMOVAL
.40
1 2
% SP REMOVED
F-fgure 41. Base Casel. Upgrading Cost vs. % Pyritic Sulfur Removed
212
-------
BASE CASE 2
1.88% SP REMOVAL
ROM COAL
COST - $AON
30
1 2
%SP REMOVED
Figure 42. Base Case 2. Upgrading Cost vs. % Pyritic Sulfur Removed
213
-------
.80
.70
8 -60
o
z
o
o
^ =50
,40
%SP
REMOVED
4.0-
2.0-
1.0,
0.79 (BASE CASE 1).
i I 0.5i
0 5 10 15 20 25
ROM COAL COST $/TON
30
Figure 43. Base Case 1. Upgrading
Cost vs. ROM Coal Cost
at Different Levels of
Pyritic Sulfur Removal
(i.e., battery limit costs, off-sites, etc.) were adjusted. The capital costs
were adjusted to reflect the appropriate percent treatment in the more costly
Meyers Process portion of the plant, and the less costly bypass float washing.
In all cases, the costs associated with mixing, float/sink separation, and
wash water generator (evaporator) were held constant.
Figures 45 and 46 show the effect of float processing on the upgrading
costs as a function of ROM coal cost. Figures 47 and 48 present the cross
plots indicating upgrading costs as a function of ROM coal costs at several
float/sink ratios. As may be seen from the Figures, the overall processing
costs do indeed drop with increasing float percentage. The costs drop by
approximately 5<£/MM Btu over the float/sink ratio range of 0.5 to 2 Ci.e.,
214
-------
33/66 to 66/33) for each of the ROM coal costs evaluated. This finding holds
true for both the high and low sulfur coal base cases.
.30
5 10 15 20 25
ROM COAL COST $/TON
30
Figure 44, Base Case 2. Upgrading Cost vs,
ROM Coal Cost at Different
Levels of Pyritic Sulfur
Removal
215
-------
ROM COAL
COST $/TON
BASE CASE 1
55% FLOAT
10 20 30 40 50 60 70 80
% OF TOTAL COAL TO FLOAT PROCESSING
Figure 45. Base Case 1. Upgrading Cost vs,
Coal to Float Processing
% of Total
216
-------
BASE CASE 2 - 50%
ROM
I COAL COST
I— $/TON -
0 10 20 30 40 50 60 70
% OF TOTAL COAL TO FLOAT PROCESSING
Figure 46.,
Base Case 2. Upgrading Cost vs.
Total Coal to Float Processing
217
-------
.55
-5°
to
o
u
o
RATIO OF FLOAT COAL
TO SINK COAL
1/2
o
0.
(BASE CASE 1)
55/45 '
2/1
.40
Figure 47.
Base Case 1.
Upgrading Cost
vs. ROM Coal
Cost
5 10 15 20 25
ROM COAL COST $/FON
30
.60
.55
RATIO OF FLOAT COAL
TO SINK COAL
O
u
o
z
-50
2
.45
(BASE CASE 2)
1/1
2/1
Figure 48.
Base Case 2.
Upgrading Cost
vs. ROM Coal
Cost
5 10 15 20 25
ROM COAL COST $AON
30
218
-------
9. VENDOR TESTING
9.1 GENERAL APPROACH
The current RTU equipment is intended to emphasize study of the desulfur-
ization reaction portion of the Meyers Process. Operations which may be eval-
uated are coal transfer and leach solution contacting, elevated temperature-
pressure depyritization, and solution regeneration. A filtration operation is
also included in the RTU to facilitate coal and solution separation and handling
as are solution storage facilities. The specific equipment which would be
required to evaluate operations relating to efficient coal-leach solution sep-
aration, elemental sulfur removal via solvent extraction, leach solution concen-
tration and neutralization, and coal drying is not included in the RTU. These
operations would be required to demonstrate an integrated plant facility uti-
lizing the Meyers Process.
The vendor testing task was intended to provide the necessary information
on those unit operations not in the RTU through utilization of equipment supplier
testing services. It was reasoned that basic data and scale-up information
could be obtained in a timely cost-effective manner by allowing equipment vendors
to apply their unit operation experience in testing coal prepared in the RTU.
Thus, depyritized coal from the RTU would be packaged and transferred to each
vendor. The vendor, following an approved test plan, would treat and return the
• coal for detailed analysis by TRW. The success of the vendor study was thought
to be closely associated with the initial vendor selection and the detailed
coordination, planning, and review of activities both in preparation and testing
of the coal.
A vendor survey was completed and a vendor coordination and test plan was
generated. An initial vendor coal supply was also obtained. Actual tests have
been provided by one vendor to date. A processed coal supply is currently
stored at the RTU which can be used for further vendor testing. The details of
the above mentioned activities are provided in the following sections.
219
-------
9.1.1 Equipment Suppliers Survey and Selection
A critical part of the vendor testing task was the selection of suitable
vendors to be utilized. The general approach taken was to initially select
vendor candidates for filtration, centrifugation, solvent stripping, coal dry-
ing and solution neutralization unit operations based on previous contacts.
Then, the most promising vendor's, (based on scale-up capability, cost, sched-
ule, ability to duplicate actual process conditions, sample sizes required and
versatility for different tests at the same facility) would be determined.
Finally a single vendor would be selected to test each unit operation and a
specific cost, test plan, schedule etc., would be generated. Table 46 provides
a list of the potential vendors considered for this program and the findings
obtained from telephone inquiries and written correspondence. In general, each
vendor was supplied with essentially the same type of information consisting of:
• Description of feed (composition, physical and chemical
properties.
• Description of product (desired composition, physical and
chemical properties).
• Description of tests desired (a matrix was provided showing
the desired variables and parameters to be measured).
• Description of possible acceptable approaches (types of
unit operations, sequences of operations, possible
coreactant/contacting solution compositions, etc.).
By the conclusion of Phase 3 of the current program, initial vendor selec-
tions were completed. The selected vendors and the associated services which
were anticipated are as follows:
Envirotech Corporation
• Filtration of coal slurry in leach or water solution, 4 tests,
40 gallons of slurry required (bench scale).
• Acetone slurry make-up, contacting and filtration, 4 tests
(bench scale).
• Solvent stripping of acetone filter cake, 8 tests (bench
scale).
• Neutralization of coal slurry leach solution filtrate,
8 tests (bench scale).
220
-------
TABLE 46. VENDOR SELECTION PROGRAM
Vendor
Address
Unit Operation
Test Requirement
(Per Test)
Time Quantity
(Days) (Gallon)
Comments
Envirotech Corp.
Eimco BSP
Division
Salt Lake City,
Utah
Bird Machine Co.
ro
ro
So. Wai pole,
Mass.
Filtration and
Solvent Stripping
Neutralization
Centrifuge and
Solvent Stripping
15 10 Small pilot scale equipment
Slurry for leach solution and ace-
tone slurry filtration and
stripping.
2 Will use material provided
for filtration study for
neutralization tests.
3 300 Able to do testing including
(est.) Slurry acetone slurry repulping at
(est.) pilot scale on 18" solid
bowl centrifuge. Would run
bench scale tests prior to
pilot tests to determine
operating parameters.
Artisan Industries, Waltham, Mass. Solvent Distillation 4
Inc.
WyssmontCo., Inc. Ft. Lee, N.J. Drying
Drying
Able to run tests on 4"
glass column. Artisan will
provide acetone and makeup
feed at their facilities.
5 6000 Test can be performed at the
facilities in Ft. Lee on a
4 5 small pilot scale dryer or
on a large pilot dryer at
another location. Either
drier will provide design
information for scale-up to
full size.
(Continued)
-------
TABLE 46. (Continued)
Vendor
Address
Ametek Process
Temecula, Ca.
ro
l\3
ro
Dorr-Oliver
Oakland, Ca.
Unit Operation
Test Requirement
(Per Test)
Time Quantity
(Days) (Gallon)
Comments
Filtration
Centrifuge
Drying
Filtration
Centrifuge
Joy Manufacturing
Co.
Denver Equipment
Division
Denver, Col,
Filtration
Solvent Stripping
2 5 Can only perform bench scale
tests with acetone since
2 5 pilot equipment is not
explosion proof. Able to do
leach solution tests at
pilot scale.
2 5 Test could be done using a
small pilot dryer with
solvent recovery capability.
5 300 Cannot handle acetone slurry
due to inadequate ventila-
tion. Leach solution tests
could be done at pilot scale.
Testing could be done at
bench scale for acetone
slurry.
Testing can only be done at
bench scale since the pilot
equipment is not corrosion
resistant to leach solution.
-------
Bird Machine Company
• Centrifugation of leach solution or water coal slurry,
4 tests, 1200 gallons of slurry required (pilot scale).
t Centrifugation of acetone slurry, 4 tests, 4 gallons
acetone slurry required (bench scale).
Myssmont Corporation
• Drying of acetone wet cake, 6 tests, 30 gallons acetone wet
cake required (bench scale).
Artisan Industries
• Acetone stripping, 6 tests (bench scale).
Test plans and implementation strategy needed to develop firm agreements for
the studies were also developed and are discussed in the following paragraphs.
9.1.2 Vendor Test Implementation Plan
A detailed test plan was developed to define the test and data require-
ments for the RTU vendor testing effort^10). Briefly, the approach presented in
the plan was to initially contact vendors with testing capabilities, select the
most promising vendor, prepare specific test objectives, prepare and package
RTU processed coal for testing, process the coal at the vendor test facility
under several parametric sets of conditions, and to observe and evaluate the
tests through on-site test monitoring and by analysis of feed and product
materials.
At the conclusion of each test series, the vendor was to provide documen-
tation of the following:
• Run date, duration.
• Type and size of coal used.
• Equipment model, operating conditions, and controls.
• Feed rate, mass closure, energy consumption where obtainable.
• Observations during the run, such as uniformity of product,
thickness, apparent wetness, and problems.
t Conclusions and recommendations in terms of feed rates,
effectiveness of the unit operation for purpose intended,
and limiting factors for equipment usage.
223
-------
It was also planned to obtain full-scale process equipment costs and utilities
requirements based on the vendor test data for use in conceptual full-scale
design work.
9.1.3 Vendor Testing Results
During RTU operation (Phase 3 of the current program), initial small-scale
laboratory drying tests were conducted by Wyssmont. These studies were per-
formed to obtain preliminary drying characteristics of acetone wet Martinka
coal in tray dryer type equipment. It was anticipated that this preliminary
data would minimize the chances of nonproductive testing at the more costly
larger pilot scale. For purposes of this preliminary study, one gallon of
14 mesh x 0 coal product (Martinka, Coal #1), was obtained from the RTU on
30 June 1977. The RTU processing consisted of coal-water contacting and mixing
at 212°F for approximately 6 hours followed by filtration. Since there was no
added iron sulfate in solution and since oxygen was not utilized in the reactor,
it was anticipated that the coal would experience negligable reaction. The RTU
processed coal was further processed in TRW laboratory facilities to simulate
downstream Meyers Process operations prior to the final drying (i.e., acetone
extraction) step. The acetone wet cake was then sent to the Wyssmont labora-
tories for testing. The specific procedure followed in preparing the cake for
shipment to Wyssmont was as outlined below.
1. Water wet cake from the RTU was dried in vacuum overnight.
2. To 5.43 Ibs. of the dry coal, 0.81 Ibs. of acetone was added,
resulting in a 14.9% w/w acetone wet cake.
3. The cake was placed in a 1-gallon can in an inert N2 atmo-
sphere and was then shipped to Wyssmont.
The Wyssmont processed coal samples were returned to TRW for chemical anal-
ysis. Those samples were evacuated under hard vacuum for 24 hours at 212°F.
During the 24 hours all condensables were captured in cold traps. The condensed
volatiles represented 0.59 percent of the starting material by weight. The
condensed volatiles were subjected to mass spectrometric analysis and were found
to be 83% w/w acetone and 17% w/w water. These results indicate that the orig-
inal 14,9% w/w acetone wet cake was reduced to 0.49% w/w acetone in the small-
scale test apparatus.
224
-------
In addition to the input/output analysis and results described above,
samples of the cake were removed from the drying apparatus at Wyssmont as a
function of drying time and examined for weight loss (presumed to be acetone-
water volatilization). Samples were removed from the 250°F drying apparatus
after 15, 30, 40, 50 and 60 minutes of total drying time. They were found to
contain approximately 2.0% w/w, 1.8% w/w, 0.9% w/w, 1.1% w/w and 0.6% w/w
moisture respectively. It should be noted that these results are somewhat
crude in nature due to the lack of sophistication in equipment at the Wyssmont
facility and the inherent nature of the test apparatus (dust evolution could
not be detected if minor and therefore may interfere with results). However,
the final 60-minute weight loss result agrees very well with the more precise
mass spectrometric analytical results.
With regard to full scale applicability, Wyssmont personnel indicated that
although the small scale vendor tests indicated residual acetone levels of
0.5% w/w, a residual level of 0.1% w/w acetone would be attainable through
proper adjustment of dryer gas recirculating rate, and temperature, and solids
residence time.
9.1.4 Coal Supply for Vendor Studies
In preparation for the vendor studies plan discussed in 9.1.2, coal was
collected from the RTU filter belt during experiments 03-01, 03-02, and 03-03,
after steady state operation had been attained. Table 47 describes the coal
compositions and the coal quantities obtained for future vendor testing.
9.2 VENDOR TESTING CONCLUSIONS
An obvious need currently exists to demonstrate total Meyers Process oper-
ability beyond that of the limited RTU capability. The vendor testing approach
still appears to be a very advantageous method of obtaining the desired proc-
essing information and scale-up data, Based on the enthusiastic response of
the vendors contacted and the degree of flexibility found in several of the
selected test facilities, it is anticipated that a great deal of meaningful
information could be obtained in an efficient, cost effective manner. A system
encompassing RTU sample.generation, sample collection, sample documentation,
sample analysis before and after testing, and TRW-vendor communication and tech-
nical support has been demonstrated. Samples of RTU processed coal are stored
225
-------
ro
ro
TABLE 47. COAL CAKE COLLECTED FOR VENDOR TESTING
(14 MESH X 0 MARTINKA, RTU COAL #3)
Average*Coal Cake Composition, % w/w (Except Heat Content), Dry Basis
Container No. of Cake Wt.
*
Designation Containers Ibs.
0301-1/6 6 2700
0302-1/6 6 2800
0303-1/8 8 3900
Heat
Ash Content,
Btu/lb
13.0 13260
12.0 13390
12.5 13265
Total
Sulfur
ST
.91
1.03
1.04
Pyrite
Sulfur
SP
.18
.17
.14
Sulfate
Sulfur
SS
.06
.07
.11
Organic
Sulfur
S0
.49
.60
.59
Elemental
Sulfur
SE
.18
.19
.20
Cake
Iron Moisture
Fe
.23 15.5
.26 19.8
.25 19.1
55 gallon steel open head drums with insertable polyethylene liners and header.
Exp. 03-01, average of 12 samples and their analyses.
Exp. 03-02, average of 15 samples and their analyses.
Exp. 03~03, average of 13 samples and their analyses.
-------
and awaiting possible future shipment to the selected vendors. In summary,
the availability of vendors, vendor facilities, plans, and material is such
that the vendor test program could be started at any time. It is therefore
recommended that vendor testing of RTU generated coal product be reinstituted
at the earliest possible time to obtain the desired process scale-up data.
227
-------
10. MATERIALS OF CONSTRUCTION EVALUATION
Tests were performed on the TRW Coal Desulfurization Process RTU to
evaluate the effects of the process reactant environment on selected materials
of construction. Following initial shakedown tests in which acidified leach
solution and/or coal-water solutions were used, the plant was put into
operation using the full operational cycle. Experimental test sequences
were conducted over the periods 12 October to 23 November 1977 (Experiment
01) and 5 January to 11 January 1978 (Experiment 03).
The corrosive medium was an abrasive slurry/leach solution of variable
composition containing coal, water, 1 to 5 percent sulfuric acid, 0.1 to
5 percent ferric and ferrous sulfate, and blended oxygen. Oxygen partial
pressure ranged from zero up to about 45 psig. Chemical analysis of the
leach solution also showed the presence of traces of chloride ions (28 to
107 ppm) and fluoride ions (15 to 32 ppm). The reactor was operated at
temperatures up to 275°F and total pressures up to about 80 psig.
The RTU is generally constructed of 316L Cres (stainless steel). There-
fore the plant equipment itself serves as an extensive metallurgical specimen.
Other materials of construction were evaluated through the use of metallur-
gical test speciments which were installed inside the R-l pressure vessel
(the primary reactor) just prior to Experiments 01 and 03. The test specimens
included coupon test samples and stress-corrosion test samples made of a
variety of metals, non-metals and non-metallic coatings and liner materials.
Metallic and elastomeric erosion test orifice plates were place in the
reactor pump-around loops.
Experiment 01 ran in product test environment for approximately 170 hours
until leaks developed in the piping/valve system, forcing shutdown. The RTU was
inspected and the metallurgical test specimens were removed for analysis in late
228
-------
November 1977. During shutdown, the R-l reactor was repaired. Deeply corroded
areas were filled with 316L weld rod material and smoothed by grinding. Shallow
pits were removed via grinding only. New test specimens were installed and the
plant was restarted on 5 January 1978. Experiment 03 ran for approximately
80 hours before an unexpected pump failure forced shutdown. After draining
and cleaning, the internals of the R-l reactor and the metallurgical test
specimens were again inspected. The results of the inspections and analyses
of the R-l reactor and associated piping, and of the metallurgical test
specimens installed in the RTU are reported in Section 10.1 and 10.2
respectively.
10.1 INSPECTION AND ANALYSIS OF RTU EQUIPMENT
After each shutdown, an inspection and analysis of corroded or failed
RTU components was carried out. Of particular interest were parts of the
piping system, the R-l reactor, and the R-2 mix tank.
The first inspection, in November 1977, showed the effects of approxi-
mately 250 hours of operation at elevated temperatures. The first 80 hours
were shakedown runs. Some of these runs were without coal or oxygen. In all
but pne, a very low concentration (less than 0.2%) of iron existed in the
system. Sulfuric acid concentrations during shakedown ranged from zero to
4 percent. During Experiment 01, the plant ran for about 170 hours at normal
operating temperature (230°F to 275°F) under conditions of high iron and acid
concentrations. The specific operating conditions are delineated in Section 6
of this report.
The second inspection, in January 1978, showed the effects of about 80
hours of operation during Experiment 03 at 230°F. A slurry containing only
a small amount of iron (less than 0.2%) was used. The types of corrosion
noted after Experiment 03 were different than those found after Experiment 01,
suggesting that different methods of attack were present.
10.1.1 Piping System
After Experiment 01, various components from the discharge lines and
return loops for Cells 3, 4, and 5 of R-l were inspected where leaks were
observed. These lines were exposed to leach solution at 230 F to 270 F. The
flush port fittings in the discharge lines all contained leaks, some obviously
229
-------
associated with welds. The Cell 4 return line had a leak at the reducer
weld while the Cell 5 line had a leak in a formed bend area.
Figure 49 shows the Cell 5 flush port fitting. Salts were visible at
leaks in the fitting. Large pits were present on the inside surface. In
addition, many smaller pits formed. Cross sectioning the sample at one of
the smaller pits revealed a porous "spongy" structure, Figure 50. The porous
structure penetrated about 25% of the way through the wall. Little evidence
of weld sensitization was found while some degree of incipient attack was
noted on machined inside surfaces.
The Cell 5 return line was sectioned and a large pitted area was found
(Figure 51) which was associated with the pipe weld. No sensitization was
noted and the metallurgy of the parent metal and weld appeared normal. The
attack in the weld area appeared to follow grain boundaries. While the inside
surface appeared to be shal lowly pitted when viewed normal to the surface,
cross sectioning revealed that the porous material extended the full thickness
of the pipe and caused perforation.
Pitting was observed in reducers where a crevice condition existed,
Figure 52. These pits were not associated with welds. Portions of pipe
which fit into the reducers were badly attacked due to crevice corrosion also.
Flanges throughout the system showed evidence of crevice corrosion beneath
their gaskets as shown in Figure 53. In contrast, inspection after Experi-
ment 03 showed no crevice corrosion on the flanges.
Inspection of the failed P-l feed pump after Experiment 03 indicated that
the chrome plated 316L rotor had been severely attacked by the slurry. The
plating had flaked away, exposing the 316L, which then eroded and/or corroded
away until pump tolerances grew and pumping capability was lost.
Electron microprobe analysis (EMP) was performed on several failed 316L
components. The flange in Figure 53 was probed to determine the chemistry
of an area pitted by crevice corrosion. Both the pitted area and a deposited
crystal were analyzed. The EMP indicated high Si and 0 and significant P,
Fe, and Cr in the corroded areas. Analysis of a pit from a pipe end,
High = >10%
Significant = 1-10%
Trace = <1% 23Q
-------
OJ
Figure 49. Flush Port Fitting
-------
Figure 50. Porous Spongy Structure
indicative of pitting corrosion, showed a different chemistry. High Ca, K,
Fe, S, and 0 and significant amounts of Cl, Mn, Cr, Si, and Na were found, A
common feature of all the pits analyzed was that Mo and Ni, if present, were
only found in trace amounts, even though they comprised 2 and 12 percent
respectively of the starting metal (316L).
R-l Reactor
The R-l reactor is constructed of 316L stainless steel and is approxi-
mately 3 feet in diameter and 15 feet long. It is divided into five cells
Connections between the vessel's many nozzles and ports and the plant's
piping/valve system are welded and/or bolted at gasketed flanges.
Visual inspection of the reactor after Experiment 01 showed crevice
corrosion at the manways and at weir/ring interfaces, and pitting corrosion
near the "salt line" (water line) in the reactor. The area below the water
line was green indicating a ferrous sulfate deposit while that above the
water line was orange-red indicating ferric sulfate. A cell by cell summary
of the observations is given below.
232
-------
CO
CO
Figure 51. Pipe Weld Corrosion
-------
Figure 52. Crevice Corrosion in Reducer Fitting
Cell 1. There was no evidence of pitting or crevice attack. Wall
deposits especially on the manway side, made inspection difficult. Some small
pits were detected at the mixer port flange.
Cell 2. No evidence of pitting.
Cell 3. Several large pits were noted at or near the water line. A
.38-inch x .12-inch x .015-inch deep* pit was found at the weld near the #2
bulkhead on the opposite side from the port. Other small pits were visible
in this area. The metal around these pits crumbled when probed with a knife
blade revealing a porous structure similar to that noted on the piping/valve
components. There were about 30 of these small pits ranging from .015-inch
to .030-inch diameter. After probing, the dimensions of the larger exposed
areas were .188-inch x .125-inch x .060-inch deep, .125-inch diameter x
.125-inch deep and .100-inch diameter x .040-inch deep. Other pits were
noted near the center of Cell 3 at a small port. One of these pits was
.250-inch x .188-inch x .030-inch deep. There were also six interconnected
Note: Depths were estimated.
234
-------
ro
CO
en
Figure 53. Flange Crevice Corrosion
-------
pits in this area all approximately .030-inch deep. The largest of these was
.125-inch diameter. These various pits were determined to be typical of those
found in numerous other regions within Cell 3.
Cell 4. Heavy deposits were found in Cell 4 making inspection difficult.
Two pits were found near the #3 bulkhead on the far side. Their sizes were
.188-inch x .125-inch x .090-inch deep and .250 inch x .090 inch x .030-inch
deep. On the port side, a .125-inch diameter x .030-inch deep pit was found
near the #5 bulkhead. The bulkheads (weir rings) between Cells 3 and 4 and
between Cells 4 and 5 were attacked on the edges. Large areas could be
scraped away with the knife blade revealing a porous, crumbling, structure.
The weir nuts were attacked, especially where they were welded to the bulkheads,
The baffle nuts and bolts were severely attacked and could be separated by
hand, see Figure 54. The weir baffle was attacked at the edge, Figure 55.
Cell 5. Porosity and pitted structure which could be scraped away was
found near the #4 bulkhead on the far side from the manway. Exposed areas
were .188-inch x .125-inch x .060-inch deep, .125-inch diameter x .030-inch
deep, .250-inch x .125-inch x .030-inch deep and .250-inch diameter x .030-inch
deep. The bolts holding the baffles crumbled away as did the nuts. The lower
part of the bracket was pitted and the bracket-to-wall weld had a .125-inch
diameter x .030-inch deep pit. Two pits were found on the port side near the
dome weld. These were .250-inch diameter x .030-inch deep and .125-inch diam-
eter x .020-inch deep.
The TE 56 temperature probe, although appearing solid and shiny, crumbled
when scraped with the knife blade exposing a large area of porous, heavily
attacked material (Figure 56), The corrosion apparently originated at the
tip-to-probe weld. EMP analysis showed the probe base metal to be 316L. High
amounts of Si and 0 were found in the corroded area. Iron, Cr, and Ni were
either undetected or present in only trace amounts.
Inspection of the R-l reactor internals after Experiment 03 showed only
minor signs of the spongy pitting that was observed after Experiment 01. This
type of attack was visually evident only at the water line and on a welded
attachment. However, many smooth bottom pits were observed. General pitting
occurred where grinders had been operated tn repairing the reactor (Figure 57).
236
-------
INJ
CO
Figure 54. Baffle Bolts and Nuts from R-l Cell 4
-------
Figure 55. Weir Baffle
-------
ro
CJ
10
Figure 56. TE-56 Thermocouple Probe
-------
Figure 57. R-l Reactor Internals AFter Experiment 03 Showing Pits Where
Grinders Had Operated
-------
Pitting was evident at welds and on free surfaces (Figure 58). No crevice
corrosion was visible.
10.1.3 T-2 Mix Tank
The T-2 mix tank was inspected on 9 December 1977. Some crevice corrosion
was noted on the tnanway flange and door in the seal area. The mixer port
flanges were unattacked. In general, the internal surfaces of the mix tank
were not attacked by pitting or crevice corrosion. No crevice corrosion was
noted on the baffles, bolts or nuts. The temperature probes were also not
attacked. The only attack noted was under a piece of masking tape which had
been left adhered to the wall in Cell Number 3. A shallow pit had formed in
the crevice caused by the tape. The pit was .250-inch x .188-inch x .015-inch
deep. Two or three pits (about .100-inch diameter) were found on a blade from
one of the mixers. Incipient pitting was noted on the bore of the propeller.
The weir actuator shaft was attacked in the area of attachment to the weir
plate (crevice condition) and the corresponding weir plate guide showed a
possible pit. It should be noted that while the T-2 mix tank experienced the
identical coal/slurry stream that the R-l reactor saw, the temperatures in T-2
never exceeded 212°F since it is an atmospheric pressure vessel. Also, there
is no oxygen injection into T-2. These differences are believed to relate to
the much lesser degree of corrosion experienced in T-2 as compared with the
reactor R-l.
10.2 ANALYSIS OF TEST SAMPLES
The primary types of attack which RTU materials of construction must
resist are erosion-corrosion in high flowrate areas, stress corrosion crack-
ing (SCC) in high stress areas, pitting corrosion, crevice corrosion, and
general corrosion. The tests described in this section were designed to
predict the performance of a variety of materials used in the corrosive
environment of the R-l reactor.
Materials samples in the form of metal and alloy corrosion coupons,
stress-corrosion coupons, coated and lined coupons, and .non-metallic coupons
were mounted in the R-l reactor in Cell 3. A 316L and Teflon rack was con-
structed and the specimens were mounted using Ti-6Al-4V bolts, 316 Cres bolts,
and 304 Cres nuts. The bolts and nuts were sealed (overrated) using a
Hypalon elastomer coating. In addition, metallic and elastomeric orifice
241
-------
rvi
-f^
ro
Figure 58. R-l Reactor Internals After Experiment 03, Showing Smooth
Bottom Pits at Welds and on Free Surfaces
-------
plates were installed in the cell discharge and return lines to assess erosion
characteristics of these materials. All coupon and orifice plates had areas
where crevice conditions existed. A summary of the results of the inspection
and analysis after removal on 29 November 1977 are given in Table 48. The
observations noted for samples removed during both the November 1977 and
January 1978 inspections are presented below.
10.2.1 Corrosion Coupons and SCC Specimens
Figure 59 shows corrosion coupons and SCC samples attached to the mount-
ing rack prior to installation in the R-l reactor. After removal, the samples
were cleaned, weighed, and inspected for signs of corrosive attack.
Metal Corrosion Coupons. Inspections of samples tested in Experiment 01
suggest there are at least two distinct corrosion mechanisms operating, namely
corrosion in crevices and pit formation on free surfaces. Once corrosion has
initiated, a spongy porous structure forms which can tunnel under the surface
to honeycomb the substructure. The crevice attack appear to be caused by a
concentration celP ', with oxygen the most probable species experiencing the
concentration gradient. The pitting corrosion appears to be triggered by CSL~
ion and sustained by the presence of Fe + ion and possibly complicated by
localized deposits causing crevice conditions.
A third type of pitting was observed during Experiment 03. Pits which
formed did not have a spongy, porous structure but were smooth with rounded
bottoms. Test conditions for Experiment 03 were less severe than Experiment 01
(lower temperature, lower iron and halide ion concentration, shorter duration).
However, it is unclear what mechanism of attack caused the round bottomed pits.
The effects of the corrosive environment on specific coupon materials is
summarized below.
• 316L Cres Stainless Steel. A 316 Cres sample was put in as
a control. Crevice pitting occured under the bolt head and
the nut during Experiment 01 resulting in a weight loss of
0 14 percent. Inspection of the coupons after Experiment 03
revealed smooth bottomed pits but no crevice corrosion
(see Figure 60).
t 304 Stainless Steel. The samples exhibited smooth bottom
pits, but no crevice corrosion when inspected after
Experiment 03 (see Figure 61).
243
-------
TABLE 48. COUPON CORROSION TEST RESULTS - 29 NOVEMBER INSPECTION
Material
Result
Metals:
316 Cres
Pure Nickel
Inconel 601
Inconel 617
Inconel 625
Incoloy 825
Titanium
Lead
Non-Metal lies:
FRP
Ashland 197/3 + AT-8-4
Ashland 800/801L-68-26
Furfuryl Alcohol Resin
Ashland 800-10-19
Ashland 7240-4-37
Unfilled Polyester Resin
Ashland 7241-6-27
Unfilled Polyester Resin
Ashland 72L + 5b-l-23
FRP Pultrusion
Plasites 4020, 4030, 4092
Vinyl Ester Coatings on Steel
Ceil cote 252
Coatings:
Kynar 202/205
EA919 Epoxy
Viton A
AF-E-332(EPDM)
Crevice Corrosion
Completely Dissolved
Crevice Corrosion
Noncrevice Pitting
Crevice Corrosion
Noncrevice Pitting
Crevice Corrosion
Crevice Corrosion
Slight Non Crevice Pitting
No Attack
No Attack
No Pitting
Yellow - Red Coating
Edge Cracking
Severe Weight Loss
Resin Bond Attack
Blistered
Embrittled
Cracked
Surface Etched
Crazed
Pitted
Severe Weight Loss
Surface Etching
Blistered
Discolored
Blistered
Edge Cracking
Discolored
Severe Weight Loss
Etched
Embrittled
Cracked
Some Etching
Discoloration
Etched
Cracked
Delaminated
Peeled
Coating Undamaged
Badly Attacked
Unaffected
Unaffected
244
-------
ro
Figure 59. Coupon Mounting Rack and Test Specimens
-------
fNJ
-P>
CTl
Figure 60. 31
r „
316L Cres Coupons Showing Signs of Crevice Attack
Experiment 01 (L) and Smooth Bottom Pitting During
Experiment 03 (R)
-------
ro
Figure 61. 304 Stainless Steel Coupons from Experiment 03 at and
Above Water Line (L) and Below Water Line (R)
-------
• Nickel Base Metals and Alloys. Coupons of pure nickel and
of various Inconel and Incoloy alloys were tested. The pure
nickel coupon dissolved completely during Experiment 01.
Inconel 601 showed a weight loss in excess of 4%. It pitted
badly around the bolts which attached it to the mounting
rack, and also showed significant attack in noncrevice areas.
The Inconel 617 coupon was subject to similar crevice cor-
rosion and noncrevice pitting, but experienced only a
0.18% weight loss. Inconel 625 and Incoloy 825 samples
were not as prone to noncrevice attack, but did show
significant crevice corrosion (Figure 62). Weight losses
of 0.055% and 0.31%, respectively, were recovered. Coupons
of Hastelloy G and Hastelloy C-276 mounted in the R-l
reactor during Experiment 03 showed no visible signs of
corrosion when removed and examined (Figure 63).
t Titanium. Samples of commercially pure titanium (Ti-6&A)
and of Ti-12 alloy were tested in Experiment 01. Neither
coupon was visibly attacked (Figure 64). Weight losses were
less than 0.01% for both coupons. Ti-50A coupons tested
in Experiment 03 are also shown in Figure 64). They showed
no visible signs of corrosion. -Normally, titanium can
be attacked by low concentrations of sulfuric acid.
However, ferric ion suppresses this attack and acts as
an effective corrosion inhibitor(21). The corrosive
effects of fluorides are not known, however, the pres-
ence of calcium is known to tie up fluorides and thus
act as a corrosion inhibitor in titanium applications^2) _
It is expected that enough calcium is present in the
desulfurization environment to prevent fluoride attack.
• Lead. The lead samples had a yellow and red coating
which was adherent. This was probably lead sulfate. The
surfaces of the lead samples were not pitted (Figure 65).
Non-Metal Coupons. The nonmetallic coupons were removed from the R-l
reactor on 29 November 1977. The specimens were examined and evaluated for
evidence of degradation and the results are as follows:
• Ashland 197/3-400-7, FRP High Temperature Polyester
Resin. The fiberglass specimen underwent a 1.17%
weight loss. The coated edges of the sample were
cracked as was the surface gel coat. The fiberglass-
to-resin bond was obviously attacked as evidenced by
the pronounced fiber prominence at the specimen
surface. The portion of the sample covered by the
Teflon mounting plate was not discolored whereas the
exposed portion was red-brown in color.
248
-------
Figure 62. Inconel and Incoloy Alloy Coupons, 7.5X-'
Inconel 601 (Upper L), Inconel 617 (Upper
R), Inconel 625 (Lower L), Incoloy 825
(Lower R)
249
-------
tNJ
cn
O
Figure 63. Hastelloy C-276 Coupons from Experiment 03 at and Above
Water Line (L) and Below Water Line (R)
-------
Figure 64.
Titanium Coupons: Ti-50A (L) and
Ti-12 (R), from Experiment 01 (Top)
and Experiment 03 (Bottom)
251
-------
Figure 65. Lead Coupons from Experiment 01.
Surface Marked by Cleaning Operation.
No Pitting or Crevice Corrosion. 1.5X
252
-------
• Ashland 197/3 + AT-8-4 (Flame Retarded Version of 197/3).
The surface of the specimen was blistered, but otherwise'
showed no visible signs of surface attack. The weight
loss was almost negligible (0.50%). The sample, however,
was drastically embrittled and the surface gel coat
cracked during handling.
• Ashland 800/801L-68-26, Furfuryl Alcohol (Furan) Resin.
The weight loss on this sample was low, 0.95%. The sur-
face however, was badly etched except where covered by the
Teflon mounting plate. Close examination of the surface
showed it to be badly crazed and pitted. Surface veil
fibers were exposed in the etched areas.
• Ashland 800FR-10-19 (Flame Retarded Version of 800/801).
The weight loss on this sample was much more severe (2.50%).
The surface etching and veil fiber exposure was similar
to that of the 800/801 specimen. Again, the area under
the mounting plate was largely untouched.
• Ashland 7240-4-37, Unfilled Polyester Resin. The weight
gain on this sample was almost negligible (0.01%). How-
ever the sample was badly degraded. The surface was
blistered and the glass reinforcement was very prominent
due to degradation of the glass fiber-to-resin bond. The
surface, except for the portion under the mounting plate,
was a dark brown in color.
• Ashland 7241-6-27, Unfilled Polyester Resin. This sample
showed degradation similar to that of the 7240 resin
sample. The edge coating was cracked and in many places
nonexistent. Additionally, the sample was warped.
• Ashland 72L + Sb-1-23, (Flame Retarded Version of 7240
and 7241 Resin). This specimen showed a 2.50% weight
loss. The surface was badly cracked and etched except
under the mounting plate. Additionally, small blisters
were evident over all the sample surface and the edge
resin coating was cracked.
• Fiberglass Reinforced Polyester Pultrusion. The specimen
was badly embrittled and cracked during handling. The
area around the mounting hole ripped during removal from
the mounting plate.
Reinforced Coatings
• Plasites 4020, 4030, and 4092 Filled Vinyl Ester Coatings
on Steel. These specimens, although showing some surface
etching and discoloration, seem to survive. However, their
long term resistance is impossible to predict on the basis
of this short exposure.
253
-------
Coatings
Ceil cote 252, Flakeglass Filled Polyester on Carbon Steel.
The surface of the coating appeared to be etched and the
flakeglass filler exposed. The coating was cracked in one
corner and the steel substrate exposed. This was probably
the result of exposure to high temperature.
t Kynar 202/205 Coating on Carbon Steel (Approximately .006-
inch thick). The coating delaminated and cracked, probably
because of differential thermal expansion and loss of
adhesion due to moisture permeation. The steel almost
completely corroded away, but the coating film looked like
new.
• Kynar 202/205 Coating on 304 Stainless Steel (Approximately
.006-inch thick). The coating was also cracked and peeled
from this specimen although some was still adhering well to
the surface. Some crevice corrosion was exhibited under
the attachment hole after the coating was peeled back,
however the 304 was largely protected by the Kynar.
t Kynar Sheet Stock Heat Sealed Coating on Carbon Steel
(Pennwalt Sample). This specimen was thicker than the
other Kynar samples. It showed no signs of degradation
or substrate corrosion. The greater thickness of the
sample enabled it to withstand thermally induced stresses.
t EA 919 Epoxy Coating On 304 Stainless Steel. This specimen
was badly attacked. The coating could not be found and
the stainless steel substrate crumbled into powder. The
epoxy coating probably failed by hydrolytic breakdown of
the resin in the hot oxidizing acid solution as well as
thermal cracking and moisture permeation, which caused
loss of adhesion to the steel substrate.
• Viton A Sheet Stock Bonded with RTV 732 Silicon Adhesive
(Dow-Corning) to Carbon Steel. The Viton A was unaffected
with no change in hardness and the steel substrate suffered
minor corrosion because of a small leak in the RTV peri-
pheral bond seal. The silicone showed signs of edge
attack - small cracks and chalking. Viton A may be consid-
ered as a liner material; however, a suitable bonding and
sealing agent must be found.
• AF-E-332 Sheet Stock (EPDM), Bonded to Carbon Steel with
RTV 731 (Dow-Corning). The silicone rubber bond failed
in adhesion to the EPDM (Ethylene Propylene Diene Monomer)
surface and the carbon steel substrate completely dis-
solved. The EPDM showed no signs of degradation and change
in hardness. The silicone seal suffered edge attack
similar to the RTV 732; that is, extensive edge cracking
254
-------
and chalking. Although the EPDM itself was unaffected by
process conditions, bonding and sealing of the material
to the metal substrate presents a very difficult problem.
Stress Corrosion Cracking Specimens. None of the five SCC specimens
(titanium, welded titanium, Incoloy 825, 316L Cres, Inconel 625) showed attack.
However, the 304 Cres nuts failed, thus relieving the stress at some point
during the run so that the test results are inconclusive. Some crevice attack
occurred on the 316L Cres and the Incoloy 825 samples in the attachment area.
Coupon Mounting Rack. The 316L Cres support rack exhibited crevice
corrosion at the mounting holes and localized pitting attack on its edges.
This attack was not associated with welds. The Teflon portion of the rack
was unattacked although cold flow had occurred under bolt heads which had
been torqued to a high value. The PTFE Teflon mounting plates showed no
signs of degradation, either from chemical attack or abrasion.
The Ti-6Al-4V rack mounting bolts appeared not to be attacked during the
test but the 304 Cres nuts and 316 Cres bolts were severely attacked and, in
some cases, dissolved.
The Hystl 6793-175 (glass reinforced, alumina trihydrate filler polybuta-
diene resin) mounting plate was softened and discolored. The plate cracked
during handling and a cross section showed almost complete penetration of the
leach solution into the material. Again, the hot oxidizing acid conditions
appear to be the cause of the failure.
10.2.2 Orifice Plate Specimens
Five metallic and two elastomeric orifice plate specimens were tested
during Experiment 01. The metallic plates were mounted using gaskets while
the elastomeric plates acted as gaskets. The metallic materials tested were
Inconels 601, 617, and 625, Incoloy 825, and Titanium 50A. The Inconels
exhibited crevice corrosion under the gasket, although only a few pits were
noted in the Inconel 625 plate. The corrosion was found on both sides of the
Inconel 601 plate and on the downstream side of the Inconel 617 and 625
plates. Pitting on the exposed surface was noted for Inconel 601 and incipient
pitting (mottled surfaces) for Inconel 617. Incoloy 825 was severely attacked
in the crevice area on both sides of the orifice plate. The Titanium 50A plate
was unaffected by the exposure and no corrosive attack was detected.
255
-------
The Hypalon (chlorosulfonated polyethylene) orifice plate showed no
signs of abrasion or tearing. The elastomer showed no decrease in cut resist-
ance and in Shore A hardness. The Neoprene (chloroisoprene) orifice plate
showed no signs of abrasion. However the Shore A hardness jumped ten points
and material subjected to the hot sulfuric acid solution in the orifice area
was torn due to a dramatic decrease in its localized tear resistance.
10.3 DISCUSSION
The observed corrosion in the coal desulfurization reactor and pump
circuit components was extremely severe. The short operating time under full
process conditions indicates that very rapid 316L corrosion rates are being
experienced. The spongy, porous nature of some corroded areas suggests that
serious damage can result even though visual inspection of the surface
indicates little or no attack. In many cases, scraping what appeared to be
a slightly attacked region resulted in material crumbling away to reveal
large, deep pits. In one instance, a pit was detected in the R-l reactor
which penetrated about 25% through the wall.
The porous nature of the observed pitting is of critical concern. It
appears that once a pit has been initiated, either by crevice conditions or
by chloride attack on a free surface, that the corrosion follows metallurgical
paths offering the least resistance, such as grain boundaries, stringers, cold
worked areas, etc. This could result in multiple very small but relatively
deep subpits within the main pit. At the root of these subpits, the local
conditions would be much like a crevice so that an oxygen concentration cell
can exist. This is, in essence, an electrolytic cell where the driving force
to cause corrosion is an oxygen concentration gradient between the root of
the subpit (low concentration) and the bulk of the solution (high concentra-
tion). These conditions would cause the subpit to rapidly extend producing
the sponge-like structure. This possible mechanism could be worsened by
high oxygen contents in the main slurry. It is interesting to note that
neither crevice corrosion nor spongy structure was prevalent in the T-2 mixer
which is oxygen free. While some crevice initiated pits were found in T-2,
such as under the masking tape, the pits appeared to be shallow and smooth
bottomed.
256
-------
The coal desulfurization environment is extremely agressive to stainless
steel, especially for the type of attack observed. The relatively high tem-
peratures, low PH (PH of 1 to 2), crevice conditions, and high oxygen content
of the slurry greatly accelerate crevice (concentration cell) attack. Pitting
attack is greatly accelerated by the presence of halide ions and the reaction
sustained by the presence of a metal ion, such as ferric or cupric ion, which
can be reduced to provide the cathode reaction. Pitting rates are also
greatly accelerated by high temperature, low pH, high velocity, and high
oxygen content. Rough surfaces or surface deposits, internal or external
stresses on metal parts, and metallurgical defects, such as inclusions, pores,
and stringers, are especially susceptable to pitting attack. All of these
factors play a part in the corrosion behavior of the desulfurization unit to
varying degrees in different parts of the system.
During the course of the RTU corrosion study, it was discovered that
similar material problems had been encountered in a Sherrit-Cominco copper
ore processing pilot plant in Fort Saskatchewan, Alberta^18', In this process,
aggressive slurries must be handled, which are similar to those encountered
in the Meyers Coal Desulfurization Process. Significant corrosion problems
were experienced before final material selections were made. Results from the
Sherrit-Cominco pilot plant indicate that near ambient temperature storage
tanks and thickeners constructed of fiberglass reinforced plastic have per-
formed well. Where temperatures do not exceed 190°F, rubber lined mild steel
has been used satisfactorily. At temperatures up to 230°F and 1% to 5%
sulfuric acid concentrations, 316 Cres can be used if a high (50 to 100 g/i)
concentration of CuSCL is maintained. (Areas which do not contain significant
copper sulfate concentration experience rapid failure of 316 Cres components.)
Where temperatures reach 350°F to 400°F, titanium lines, valves, and heat
exchangers are used and perform well. Pumps and lines subjected to severe
abrasion and aggressive slurry compositions are constructed of Hastelloy
C-276, selected as a result of a development program during which several
other materials were eliminated. The above mentioned information appears to
be very consistent with RTU related observations.
257
-------
10.4 CORROSION STUDY CONCLUSIONS
The obvious conclusion drawn from the inspection and analysis of the
system components and test samples is that unprotected 316L Cres is an
unacceptable material for the R-l reactor and pump around loops when they are
operated at nominal process conditions (02 present, .2% to 5% iron, .2% to
4% sulfuric acid, and temperatures greater than 215°F).
Commercially pure titanium (Ti-50A) and titanium alloy Ti-12 appear to
be suitable materials of construction for the R-l reactor. The Ti-12 alloy
may be required at flanges, mounting rings, etc., since it is more resistant
to crevice attack than Ti-50A. The cost of pure titanium and Ti-12 is about
2 to 2-1/2 times the cost of 316 stainless but less expensive than Hastelloy
(23)
Cv '. This cost differential has been verified by a comparison of the
existing RTU reactor cost and a 1978 vendor quote for a replacement reactor
constructed of Ti-12. It should be noted that while titanium has performed
well in short term test runs on the RTU, long term tests are needed to verify
its acceptability.
Based on tests conducted during Experiment 03 and similar experience at
Sherrit-Cominco, Hastelloy C-276 may also be able to withstand the hostile
chemical environment of the Meyers Process reactor system. Hastelloy alloys
are prime candidates for pump-around loop components.
Various metallic and non-metallic materials may be considered for compon-
ent liners. Lead is a possibility as a lining material but long term corrosion
rates and erosion resistance must be further evaluated. Most of the non-
metallic samples were attacked to some degree by the high temperature sulfuric
acid/iron sulfate solution. Notable exceptions were Viton A, EPDM, and Teflon.
An effective means of bonding and sealing these materials to a metal substrate
must be found before they can be considered practical as liners.
For applications in Meyers process equipment operating at or near leach
solution normal boiling points (up to 213°F) in the absence of oxygen, results
indicate that 316L and Fiberglass reinforced plastics (FRP) may be suitable
materials of construction. There are also a number of coatings which also
appear to be satisfactory in this application as discussed in Section 10.2.1.
It should be stressed that all results are based on relatively short test
258
-------
durations. Therefore, an organized long term, materials testing activity
would be desirable to conclusively determine the merits of the studied
materials in Meyers Process applications.
259
-------
11. REFERENCES
1. Hamersma, J. W. and M. L. Kraft. Applicability of the Meyers Process for
Chemical Desulfurization of Coal: Survey of Thirty-Five Coals. Report No.
EPA-650/2-74-025a prepared by TRW Systems and Energy for the U.S. Environ-
mental Protection Agency under Contract No. 68-02-0647, Washington, D.C.,
1975.
2. Hamersma, 0. W., M. L. Kraft, R. A. Meyers, C. A. Flegal and A. A. Lee.
Applicability of the Meyers Process for Chemical Desulfurization of Coal:
Initial Survey of Fifteen Coals. Report No. EPA-650/2-74-025 prepared by
TRW Systems and Energy for the U.S. Environmental Protection Agency under
Contract No. 68-02-0647, Washington, D.C., 1974.
3. Hamersma, J. W., M. L. Kraft, E. P. Koutsoukos and R. A. Meyers. Chemical
Removal of Pyritic Sulfur from Coal. In: Advances in Chemistry, Series
No. 127, American Chemical Society, Washington, D.C., 1973.
4. Hamersma. J. W,, E. P. Koutsoukos, M. L. Kraft, R. A. Meyers, G. J. Ogle and
L. J. Van Nice. Program for Processes for the Selective Chemical Extraction
of Organic and Pyritic Sulfur from Fossil Fuels. Report No. 17270-6011-
RO-00, Vol. I and II, prepared by TRW Systems and Energy for the U.S.
Environmental Protection Agency, Research Triangle Park, under Contract
No. EHSD 71-7, No. Carolina, 1973.
5. Meyers, R. A., E. P. Koutsoukos, M. L. Kraft, R. A. Orsini, M. J. Santy and
L. J. Van Nice. Meyers Process Development for Chemical Desulfurization of
Coal. Report No. EPA-600/2-76-143a, Vol. I and II, prepared by TRW Systems
and Energy for the U.S. Environmental Protection Agency, Research Triangle
Park, No. Carolina, 1976.
6. Van Nice, L. J., and M. J. Santy. Pilot Plant Design for Chemical Desulfuriza-
tion of Coal. Environmental Protection Technology Series, EPA-600/2-77-080,
1977.
260
-------
11. (Continued)
7. Nekervis, W. F., and E. F. Hensley. Conceptual Design of a Commercial Scale
Plant for Chemical Desulfurization of Coal. Environmental Protection Technology
Series, EPA-600/2-75-051, 1975.
8. McGee, E. M. Evaluation of Pollution Control in Fossil Fuel Conversion
Processes, Coal Treatment: Section 1, Meyers Process. Environmental
Protection Technology Series, EPA-650/2-74-009K, 1975.
9. Meyers, R. A., M. J. Santy, E. P. Koutsoukos, L. J. Van Nice and R. A. Crsini.
Reactor Test Project for Chemical Removal of Pyritic Sulfur from Coal.
Report No. 25305-6023-TU-OO prepared by TRW Systems and Energy for U.S.
Environmental Protection Agency, under Contract No. 68-02-1880, Research
Triangle Park, No. Carolina, 1977.
10. Meyers, R. A. Reactor Test Project for Chemical Removal of Pyritic Sulfur
from Coal. Report No. 25305-6038-TU-OO prepared by TRW Systems and Energy
for the U.S. Environmental Protection Agency under Contract No. 68-02-1880,
Research Triangle Park, No. Carolina, 1977.
11. Meyers. R. A., L. J. Van Nice, E. P. Koutsoukos, M. J. Santy and R. A. Orsini.
Bench-Scale Development of Meyers Process for Coal Desulfurization. Report
No. 28456-6025-TU-OO prepared by TRW Systems and Energy for the U.S.
Environmental Protection Agency under Contract No. 68-02-2121, Research
Triangle Park, No. Carolina, 1977.
12. Yurovskii, A. J. Sulfur in Coals. U.S. Department of Commerce, National
Technical Information Service, Springfield, Virginia, 1974.
13. Mills, H. E. Costs of Process Equipment. In: Modern Cost Engineering
Techniques, H. Popper, eds. McGraw-Hill, 1970. pp 111-134.
261
-------
11. (Continued)
14. Happel, J. and D- Jordan. Chemical Process Economics, 2nd ed., Marcel
Dekker, Inc., 1975.
15. Guthrie, K. M. Capital Cost Estimating. In: Modern Cost Engineering
Techniques, H. Popper, eds. McGraw-Hill, 1970. pp 80-108.
16. Perry, R. and C. Chilton, Chemical Engineer's Handbook, 5th ed., McGraw-
Hill, 1973.
17. Chemical Marketing Reporter, Schnell Publishing Co., 1977.
18. TRW Correspondence with Sherrit-Cominco, Fort Saskatchewan, Alberta, Canada.
19. The Synthetic Gas-Coal Task Force for the Supply-Technical Advisory Com-
mittee, National Gas Survey, Federal Power Commission. Final Report -
The Supply - Technical Advisory Task Force - Synthetic Gas - Coal. 1973.
20. Graver, D. L. Pitting Corrosion of Stainless Steel. In: Technical Horizons,
Allegheney Ludlum Steel Corp., Pittsburgh, Pennsylvania, 1961.
21. Covington, L. C. The Role of Multi-Valent Metal Ions in Suppressing Crevice
Corrosion of Titanium. In: Titanium Science and Technology, R. I. Jaffee
and H. M. Burte, eds. Plenum Press, New York, 1973. pp 2395-2403.
22. TRW Correspondence with L. C. Covington, Timet, Henderson, Nevada.
23. Outlook for Titanium Brightens with CPI Gains. In: Chemical Engineering,
December 19, 1977. pp 40-42.
262
-------
12. GLOSSARY OF ABBREVIATIONS AND SYMBOLS
Abbreviations
Abs
ASTM
Btu
cal
eq
Exp.
Kcal
No.
wt
Symbol s
\
AR
C
A
E
u
p
absolute
American Society of Testing Materials
British Thermal Unit
calories
equation
experiment
kilocalories
number
weight
Arrhenius constant in leach reaction (hours)"1
(wt % pyrite in coal)"1
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 AR)
micron
total pressure, atmospheres
oxygen partial pressure
gas constant, cal/mole - K
263
-------
Symbols (cont'd)
r. pyritic sulfur leaching rate, weight of pyrite
removed per 100 wts of coal per hour
rR ferric ion regeneration rate, moles per liter
per minute
S elemental sulfur
S organic sulfur
S pyritic sulfur
S sulfate sulfur
S total sulfur
U -_
T absolute temperature, K
t time, hours (leaching)-minutes (regeneration)
W pyrite concentration in coal, wt %
Y ferric ion to total iron ratio
264
-------
1. REPORT NO.
EPA-600/7-79-013a
TECHNICAL REPORT DATA
iriease read Instructions on the reverse before completing}
3. RECIPIENT'S ACCESSION NO.
Reactor Test Project for Chemical Removal of
Pyritic Sulfur from Coal; Volume I. Final Report
5. REPORT DATE
January 1979
6. PERFORMING ORGANIZATION CODE
R.A.Meyers, M.J.Santy, W.D.Hart,
L.C.McClanathan, and R. A.Orsini
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
TRW, Defense and Space Systems Group
One Space Park
Redondo Beach, California 90278
10. PROGRAM ELEMENT NO.
EHB527
11. CONTRACT/GRANT NO.
68-02-1880
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final: 6/75 - 6/78
14. SPONSORING AGENCY CODE
EPA/600/13
^SUPPLEMENTARY NOTES ffiRL-RTP project officer is Lewis D. Tamny, Mail Drop 61, 919/
& 109.
is. ABSTRACT The report giveg results of an evaluation of the initial performance of the
Reactor Test Unit (RTU) for chemical removal of pyritic sulfur from an Appalachian
coal. Operation of the plant--from its checkout and shakedown in September 1977
through January 1978--demonstrated that the RTU could be run continuously in three-
shift operation to reduce th$ coal from 2.4 Ib SO2/million Btu to a level of 1.0 to
1. 2 Ib SO2/million Btu, after rinsing and extraction of generated elemental sulfur.
There was no measurable coal oxidation during processing and leach rates in the
plant were greatly improved over bench-scale values. The leach solution/coal/oxy-
gen environment was found to be corrosive to the installed stainless steel reactor,
necessitating future upgrading to support additional testing. Bench-scale experiments
showed that the leach solution can be used as a homogeneous dense-media to effi-
ciently gravity-separate coal prior to processing. Beneficial engineering cost impro-
vements are obtained based on using this approach, resulting in capital cost estima-
tes of $68-69/kW and with #0.44-0. 50/million Btu processing costs, including amor-
tization of capital, for input coal costing $0. 78-0. 81/million Btu. Overall energy ef-
ficiency was 93-96%.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Pollution
Coal
Desulfurization
Pyrite
Chemical Cleaning
Rinsing
Extract
Pollution Control
Stationary Sources
Pyritic Sulfur
13B
08G,21D
07A,07D
13H
18. DISTRIBUTION STATEMENT
Unclassified
277
Unlimited
EPA Form 2220-1 (9-73)
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
265
------- |