EPA-600/2-76 143a
May 1976
Environmental Protection Technology Series
               MEYERS PROCESS DEVELOPMENT FOR
               CHEMICAL DESULFURIZATION  OF  COAL
                                                  Volume I
                                  Industrial Environmental Research Laboratory
                                        Office of Research and Development
                                       U.S. Environmental Protection Agency
                                 Research Triangle Park, North Carolina 27711

-------
                RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into five series. These five broad
 categories were established to facilitate further development and application of
 environmental technology. Elimination of traditional grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The five series are:

     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies

 This report  has been assigned to  the ENVIRONMENTAL PROTECTION
 TECHNOLOGY series. This series describes research performed to develop and
 demonstrate  instrumentation, equipment, and methodology to  repair or prevent
 environmental degradation from point  and non-point sources of pollution. This
 work provides  the new or improved technology required for the control and
 treatment of pollution sources to meet  environmental quality standards.
                     EPA REVIEW NOTICE

 This report has been reviewed by the U.S. Environmental
 Protection Agency, and approved for publication.  Approval
 does not signify that the contents necessarily reflect the
 views  and policy of the Agency,  nor does mention of trade
 names or  commercial products  constitute endorsement or
 recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                         EPA-600/2-76-143a

                                         May 1976



      MEYERS PROCESS  DEVELOPMENT

FOR  CHEMICAL DESUL FURIZATION OF COAL

                     VOLUME  I
                          by

       E.P. Koutsoukos, M. L. Kraft, R.A. Orsini,
      R.A. Meyers, M. J. Santy, andL.J. Van Nice

                  TRW Systems Group
                    One Space Park
            Redondo Beach, California  90278
                Contract No. 68-02-1336
                 ROAPNo. 21AFJ-033
              Program Element No. 1AB013


          EPA Project Officer: L.  Lorenzi, Jr.

       Industrial Environmental Research Laboratory
         Office of Energy, Minerals, and Industry
           Research Triangle Park, NC  27711


                     Prepared for

     U.S. ENVIRONMENTAL PROTECTION AGENCY
           Office of Research and Development
                 Washington, DC  20460

-------
                                ABSTRACT

     The Meyers Process for chemical removal  of sulfur from coal  was tested
at bench-scale for desulfurization of both fine and coarse coal.   In excess
of 90 percent of the pyrite was removed from run-of-mine fine coal and clean
coarse coal and over 80 percent of the pyrite from run-of-mine coarse coal.
Process unit improvements involving 1) increased process slurry solids
concentration for higher process throughput (33% w/w], 2) lowered filtration
requirements through use of larger top-size fine coal  and 3)  generated
elemental sulfur removal by vaporization from the coal matrix were demon-
                                            \
strated.  Pyrite leaching and reagent regeneration rate expressions were
validated.  Engineering studies showed that 1) the process may be engineered
in a number of basic design configurations including simultaneous leach
and regeneration, separate leach and regeneration, use of oxygen or air for
regeneration, fine coal or coarse coal processing, and combination with
coal cleaning;  2) these process design schemes lead to stand-alone full
capital cost estimates of $30-80/KW of power plant name plate capacity;
3} estimated coal desulfurization costs, annualized on a utility financed
basis, range between $0.33 and $0.51/MM Btu, and 4) assuming ROM coal costs
of $20/ton, the costs of the desulfurized fuel are estimated to range
between $1.14 and $1.32/MM Btu.
                                    m

-------
                             TABLE OF CONTENTS
                                                                     Page
Abstract                                                              i i i
List of Figures                                                        ix
List of Tables                                                        xiv
Acknowledgments                                                       xvi
Metric Conversion Factors                                            xvii
Conclusions                                                            1
Recommendations                                                        6

1.   Introduction                                                      7
2.   Pyritic Sulfur Removal from Suspendable Coal                     H
     2.1  Coal Selection and Sample Preparation                       H
     2.2  Coal Processing Experimentation                             17
     2.3  Coal Processing Data                                        24
     2.4  Data Discussion and Interpretation                          35
          2.4.1  Mixer Unit Operation                                 40
          2.4.2  L-R Unit Operation                                   50
                 2.4.2.1  Pyrite Leaching Rates During L-R Processing 56
                            of L.K. Coal at 120°C
                 2.4.2.2  Reagent Regeneration During L-R Processing  68
                 2.4.2.3  Temperature Effects on L-R Processing of    76
                            of L.K. Coals
                 2.4.2.4  Coal Top Size Effects During L-R Processing 82
                 2.4.2.5  Slurry Concentration Effects                85
                 2.4.2.6  Reagent Composition Effects                 86
                 2.4.2.7  Oxygen Partial Pressure Effects             95
          2.4.3  Settler Unit Operation                               95

-------
                         TABLE OF CONTENTS  (continued)
          2.4.4  Coal Washing  Operation                                99
          2.4.5  Elemental  Sulfur Recovery  Operation                  101
                 2.4.5.1   Elemental  Sulfur  Recovery with Toluene      101
                 2.4.5.2   Elemental  Sulfur  Recovery with Hexane       107
                 2.4.5.3   Elemental  Sulfur  Vaporization              108
     2.5  Solid-Liquid  Separation Unit Operations                    112
     2.6  Coal Drying Operation   '                                   114
     2.7  Principal  Conclusions from Bench-Scale Data                 114
     2.8  L-R  Processing  of Upper Freeport  Seam Coal                  122
3.   Reagent Recycl ability-Trace Element  Data                         127
4.   Processing of Coarse Coal                                        135
     4.1   L-R Processing Data for Coarse Coal                         136
     4.2  Clean Coal Gravity Fraction  Processing                      139
     4.3  Preliminary Data  on  Coarse Coal Processing by Size          141
            Fraction
5.   Process Engineering                                              145
     5.1  Suspendable Coal  Processing                                 145
          5.1.1  Design Basis  for  Suspendable Coal                    146
          5.1.2  Process  Baseline  Design                              153
                 5.1.2.1  Conceptual  Design  for Commercial  Scale      157
                 5.1.2.2  Process Cost Estimate                       172
          5.1.3  Process Trade-off Studies                            180
                 5.1.3.1  Reactor Model                               181
                 5.1.3.2  Pressure Effect                             183
                 5.1.3.3  Iron Concentration Effect                   188

-------
                        TABLE OF CONTENTS  (continued)
                                                                      Page
                 5.1.3.4  Oxygen Purity                               190
                 5.1.3.5  Compressed Air Regeneration                 192
                 5.1.3.6  Three Reactor Configuration                 196
                 5.1.3.7  Additional Studies                          197
     5.2  Coarse Coal (-1/4 inch) Processing                          198
          5.2.1  Concept Development                                  198
          5.2.2  Conceptual Process Details                           202
                 5.2.2.1  Reactor Section  - Pit and Continuous        202
                 5.2.2.2  Regenerator Section - Pit and Continuous    206
                 5.2.2.3  Sulfate Removal  Section                     207
                 5.2.2.4  Coal Washing and Filtration Sections        209
                 5.2.2.5  Sulfur Removal Section                      210
                 5.2.2.6  Energy Balance                              210
          5.2.3  Process Cost Estimate                                210
     5.3  Projection of Process Economics                             215
          5.3.1  Processing Option Description                        215
          5.3.2  Economics Model Description                          219
          5.3.3  Process Economics Evaluations                        225
6.  Chemical Analysis Studies                                         231
     6.1  Adequacy of Standard Sulfur Analysis Techniques for         232
            Determining Meyers Process Performance
     6.2  Meyers Process Monitoring Techniques                        242
7.   Materials Compatibility                                          249
     7.1  Literature Survey and Alloy Selection Criteria              250
     7.2  Experimental Corrosion Static Test Program                  253
                                    vii

-------
                       TABLE  OF  CONTENTS  (continued)
                                                                     Page
          7.2.1  Experimental Method                                 253
          7.2.2  Initial Runs and Modifications                      257
          7.2.3  Results and Conclusions of the Bomb Test            261
          7.2.4  Conclusions and Recommendations                     281
     7.3  Dynamic Testing                                            282
8.   References                                                      291
9.   Glossary of Abbreviations and Symbols                           293
                                    vm

-------
                              LIST OF FIGURES
 No.                                                                   Page
 1   Basic Bench-Scale Processing Scheme for the Removal of              18
       Pyritic Sulfur from Coal
 2   Bench-Scale Coal Leaching and Reagent Regeneration Apparatus        19
 3   Pyritic Sulfur Removal from 14 Mesh Top Size L.K. Coal              38
       Processed at 120°C with 5 Wt. % Fe Reagent
 4   Typical Mixer Processing Conditions                                 44
 5   100 Mesh Top Size Lower Kittanning Coal Leached with 5 Wt. %        59
       Fe Reagent
 6   Pyrite Leaching Rate Constant Data                                 '66
 7   Predicted and Experimental Y Values During L-R Processing           70
       of L.K. Coal at 120°C
 8   Predicted and Experimental Y Values During L-R Processing           71
       of L.K. Coal at 110°C
 9   Predicted and Experimental Y Values During L-R Processing           72
       of L.K. Coal at 130°C
10   Temperature Effect on L-R Processing of 100 Mesh Top Size    x       77
       L.K. Coal (20 Wt. Percent Slurries)
11   Temperature Effect on L-R Processing of 14 Mesh Top Size            78
       L.K. Coal (33 Wt. Percent Slurries)
12   Arrhenius Plots of Pyrite Leaching Rate Constants                   83
13   Coal Top Size Effect on L-R Processing of Suspendable L.K.          84
       Coal
14   Slurry Concentration Effect on L-R Processing of Suspendable        87
       L.K. Coal
15   Effect of Total Iron Concentration on Pyrite Removal During         94
       L-R Processing of Suspendable L.K. Coals
16   Effect of Oxygen Partial Pressure on L-R Processing of              96
       Suspendable L.K. Coal
17   Vapor Pressure of Sulfur                                           110
18   Typical Process Design Curves                                      120
                                     ix

-------
                         LIST OF FIGURES (continued)
  No.                                                                   Page
  19   Upper Freeport  Coal Leaching with 5 Wt. % Fe Reagent              123
  20   Pyrite Leaching Rates from Upper Freeport and L.K. Mine Coals     124
  21   Pyritic Sulfur  Removal from ROM and Cleaned L.K. Coal at 102°C    140
  22   Filtration  Rate Correlation                                       154
  23   Block Diagram for  Fine Coal                                       155
  24   Process Flow Diagram for Fine Coal                                158
  25   Simplified  Reactor System                                         184
  26   Reactor Residence  Times as a Function of Oxygen Pressure in R-l   185
  27   Effect of Pressure and Residence Time on Ferrous Iron Make        187
  28   Reactor System  Coat as a Function of Oxygen Pressure              189
  29   Oxygen Circulation Diagram                                        191
  30   Coarse Coal  Process Schematic                                     201
  31    Pit Reactor  Schematic                                             203
 32    Continuous Reactor Schematic                                      205
 33   Sulfate Removal  Section Schematic                                 208
 34   Process Economics  Case Block Diagram                              217
 35   Bomb Test Assembly                                                255
 36   Composite Welded Tube Used in Test                                255
 37   Test Coupon  Stack                                                 256
 38   Comparison of Corroded and Control Tubes                          260
 39   Microstructure of 304 Test Coupon After 1000 Hours Exposure       263
40    Microstructure of 304 Test Coupon at 400x Showing Sensiti-        263
       zation at  Grain Boundaries
41    Microsturcture of Unexposed Sample of 304 Test Coupon             264

-------
                        LIST OF FIGURES (continued)
                                                                      Page
42   Micrograph at 100X Showing Very Extensive Pitting Type of         266
       Corrosion Just Outside the Weld Zone in 316SS After Exposure
43   Micrograph at 100X Showing Very Extensive Pitting Type of         266
       Corrosion Just Outside the Weld Zone in 316SS After Exposure
44   Micrograph at 100X Showing Very Extensive Pitting Type of         267
       Corrosion Just Outside the Weld Zone in 316SS After Exposure
45   Edge of 316 Control Sample Showing Elongated Stringers Due        268
       to the Tensile Test
46   Edge of Exposed Test Coupon of 316 Showing Transgranular          268
       Cracks Extending Inward from Cut Edge
47   Edge of 31 6L Control                                              269
48   Edge of Exposed 31 6L Showing Crack Extension Along Crack          269
       Sensitive Corroded Paths and Pitting
49   Edge of 347 Control                                               270
50   Edge of Exposed Coupon of 347 Showing Limited Extension of        270
       Cracks in the Interior
51   Edge of Incoloy 825 Control                                       271
52   Edge of Exposed Incoloy 825 Test Coupon                           271
53   Weld Between 304 and 304L in the Control Tube Sample              272
54   Weld Between 304 and 304L Tubes Showing Severe Corrosion          272
       Effects After Exposure
55   Outer Diameter of 304 Tube After Exposure Showing Corrosion       273
       Along the Edge and Pits on Surface
56   Inner Edge of Exposed 304L Tubing Showing Corroded Areas          274
57   Exposed Tube Sample                                               275
58   Exposed 31 6L Tube                                                 276
59   I.D. of Exposed Armco Tubing Showing Some Cracks and Corroded     277
       Areas
                                    XI

-------
                       LIST OF FIGURES  (continued)
Np_.                                                                   Page
60   O.D. of Exposed Armco Tubing Showing Stringers Along the Tube     277
       Drawing Axis and Light Pitting
61   O.D. of Exposed 316L Tubing Showing Corrosion Pits                278
62   I.D. of Exposed 316L Tubing Showing Less Severe Corrosion         278
63   Composite Tube Used in Dynamic Test                               284
64   Composite Welded Tubing:  Corrosion Effects in the Tube Interior  285
65   Section of Armco Tubing Interior                                  286
66   Section of 304 Tubing Interior                                    286
67   Weld Zone of Armco-304 Showing Chilled Structure and Sensiti-     287
       zation by Chromium Carbide
68   Structure of 304 Away from Weld    ^                              287
69   Weld Zone of 316L-321 Showing Sensitization                       288
70   Weld Zone of 304L-Carpenter 20 Cb/3 Showing Chilled Structure     288
       and Grain Boundary Sensitization
                                    xii

-------
                              LIST OF TABLES
 No.                                                                   Page
 1   Raw Lower Kittanning Coal Analyses                                  13
 2   Lower Kittanning Coal Ash Composition                               15
 3   Lower Kittanning Coal Particle Size Distribution                    16
 4   Rate Data on L-R Processing of 100 Mesh Top Size L.K. Coal          27
       with 3 Wt. % Iron Solution at 120°C and 100 Psig
 5   Rate Data on L-R Processing of 100 Mesh Top Size L.K. Coal          31
       with 5 Wt. % Iron Solution at 120°C and 100 Psig
 6   Process Mass Balance Data                                           33
 7   Pyrite Removal from Coal with Iron Sulfate Solution at 102°C        47
       and Ambient Pressure
 8   Measured and Predicted Pyrite Removal at the "Mixer Unit            51
       Operation"
 9   Analyzed and Predicted Pyritic Sulfur Content of 100 Mesh           61
       Top Size L.K. Coal as a Function of L-R Processing Time
       at 120°C
10   Analyzed and Predicted Pyritic Sulfur Content of 14 Mesh            62
       Top Size L.K. Coal as a Function of L-R Processing Time
       at 120°C
11   Analyzed and Predicted Sp of L.K. Coal as a Function of L-R         81
       Processing Time at 110gC and 130°C
12   Pyrite Removal in the Settler Operation                             98
13   Typical Wash Section Data from L-R Processed L.K. Coals            101
14   Elemental Sulfur Product Recovery from the L-R Processing          105
       of L.K. Coal
15   Elemental Sulfur Recovery by Successive Stages of Hexane-          109
       Toluene Leaching
16   Vaporization of Elemental Sulfur from Ferric Sulfate               111
       Leached Coal
17   Trace Element Analysis Accuracy Verification                       129
                                     xi 11

-------
                         LIST OF TABLES (continued)
 —'                                                     -              Page
 18   Coal and Reagent Trace and Minor Elements                         131
 19   Trace Cation Content of Calcium Precipitates Recovered            133
        from Spent Reagent
 20   L-R Processing of 3/8 Inch x 0 L.K. Coal in 5 Wt.  % Fe            137
        Reagent at 120°C and 100 Psig
 21   Pyritic Sulfur Removal from 1/4 Inch x 0 L.K. Coal  at 102°C       138
 22   Chemical Removal of Pyritic Sulfur from Cleaned 8  x 14 Mesh       139
        Lower Kittanning Coal at 102°C
 23   Pyrite Leaching of 1/4 Inch x 0 Lower Kittanning Coal  at  102°C     141
 24   Preliminary Data on the Chemical Removal of Pyritic Sulfur         142
        from Size-Fractions of Lower Kittanning Coal  at  102°C
 25   Process Mass Balance for Fine Coal                                162
 26   Coal Desulfurization Process Equipment List                       173
 27   Sources of Equipment Cost Information                             178
 28   Sources of Operating Cost Information                             180
 29   Effect of Inert Gas Buildup on Reactor Section Annual Cost         193
 30   Effect of Compressed Air on Reactor Section Annual Cost           195
 31    Equipment Costs for a Three-Reactor Configuration                 197
 32    Coarse Coal  Process Equipment Lists                               213
 33    Annualized Costs for Battery Limits Desulfurization Plants        214
 34    Economic Evaluation Criteria                                      224
 35    Constants for Use in Economics Evaluations                        225
 36    Case 1,  Cleaned Fine Coal                                         226
 37    Case 2,  ROM  Coarse Coal                                           227
38    Case 3,  Deep Cleaned Fine Coal with 50% Meyers Process Bypass     278
                                     xiv

-------
                        LIST OF TABLES (continued)
Np_.                                                                   Page
39   Case 4, Deep Cleaned Coarse Coal with 50% Meyers Process Bypass   229
40   Upgrading Processing Costs                               ^        230
41   Comparison of Total Sulfur Analysis Techniques                    236
42   Coal Sulfur Forms Analysis Investigations - Comparison            240
       of Techniques and Effect of Coal Digestion Time
43   Comparison of Corrosion Rates of Armco and 316/316L Steels        252
44   Corrosion Data from Literature on 304 and 316SS                   253
45   Corrosion Data from Literature on Various Alloys                  254
46   List of Materials Tested                                          258
47   Ferric Sulfate Corrosive Medium                                   259
48   Compilation - Coal Corrosion Material Evaluation                  262
49   Changes in Mechanical Properties of Alloys After Exposure to      280
       Iron Sulfate Solution
50   Summary of Qualitatively Observed Events During Corrosion         281
       Testing
                                    xv

-------
                             ACKNOWLEDGMENTS

     The authors wish to acknowledge the valuable assistance received in
this project from the following TRW personnel:  L. Ledgerwood, S. Ziegler,
and D. Kilday for substantial assistance in the experimentation; C. Murray
and M. Wong for assistance rendered in process economics; D. Hopp for
assistance in computer programming; J. Blumenthal and B. Dubrow for mana-
gerial assistance and manuscript review; and L. Broberg, M. Ramirez, and
V. Melough for technical typing; special acknowledgment is due to V. Butler
who undertook the difficult and arduous task of report coordination and
finalization.

     The authors owe appreciation to Lloyd Lorenzi, Jr., the monitoring
Project Officer for the Environmental  Protection Agency under this contract
for his constant interest, cooperation and valuable comments on the project,
and to T. Kelly Janes, also of EPA, for guidance and encouragement.

     Gratitude is due to A. W. Deurbrouck of the U.S. Bureau of Mines
(Bruceton, Pennsylvania) for providing the coals used in this bench-scale
development program and to R. Kaplan of the Commercial Testing & Engineer-
ing Company  (Chicago, Illinois) for his cooperation in expediting coal
analyses for TRW.
                                    xvi

-------
                     METRIC CONVERSION FACTORS

     In compliance with EPA policy, metric units have been used extensively
in this report (followed by British units in parentheses).  However, in
some cases, British units have been used for ease of comprehension.  For
these cases, the following conversion table is provided:
     British
         Metric
     1 Btu
     1 Btu
     1 kw
     1 hp  (electric)
     1 psi
     5/9 (°F-32)
     1 inch
     1 ft
     1 ft2
     1 ft3
     1 gallon
     1 pound
     1 ton  (short)
252 calories
2.93 x 10"4 kilowatt-hours
1,000 joules/sec
746 joules/sec
                 2
0.07 kilograms/cm
°C
2.54 centimeters
0.3048 meter
             2
0.0929 meters
0.0283 meters3 or 28.3 liters
3.79 liters
0.4536 kilograms
0.9072 metric tons
                                    xv 11

-------
                                CONCLUSIONS

General Conclusions

     1.  The Meyers Process applied to high-ash run-of-mine coal is capable
         of removing in excess of 90 percent of the pyritic sulfur from 14
         and 100 mesh top-size coal and at least 80 percent from coarser
         coal (1/4 inch top-size)..*

     2.  Physical cleaning of coarse coal appears to enhance the pyrite
         leaching rates of the Meyers Process. iPyrite removals in excess of
         90 percent were attained with cleaned 1/4 inch top-size Lower
         Kittanning coal.

     3.  Meyers Process operation under simultaneous coal Teaching-reagent
         regeneration processing is feasible and efficient.  It appears to
         be the most economic mode of processing high pyritic sulfur coal
         of at least up to 14 mesh top-size.

     4.  Simultaneous Teaching-regeneration processing of coal up to at
         least 130°C and 150 psig (135 psi oxygen pressure) does not
         measurably  affect  the  organic matrix of the  coaTs  tested.

     5.  Pyrite Teaching rates are not affected by the quantity of coaT
         present in the slurry.  Processing of sTurries containing up to
         at least 33 wt percent solids is feasible.  Process coal through-
         put doubled and filtration time  (cost)  was reduced when the slurry
         solids concentration was upgraded from 20 to 33 wt. percent.

     6.  Solid-Tiquid separations of process sTurries are feasibTe by
         commerciaTTy available equipment (vendor tests).  The filtration
         costs of 14 mesh top-size coal slurries are approximately one-half
         those of 100 mesh top-size coal slurries while the difference in
         pyrite leaching rates is no more than 20 percent.


      7.  The process may be engineered  in-a number of basic design config-
         urations  including simultaneous leach and regeneration, separate
         Teach and regeneration,  use of oxygen or air for regeneration,
         fine coaT  or coarse coal  processing, and in  combination with coal
         cleaning;  these process  design  schemes lead  to stand-alone full
         capital cost estimates of $30-80/KW of power plant name capacity.

      8.  The estimated  coal  desulfurization costs, annualized on a utility
         financed  basis,  range between  $0.33 and $0.51/MM Btu.   Assuming
         ROM coaT  costs of  $20/ton,  the  costs of the  desulfurized fuel are
         estimated  to range between  $T.T4 and $T.32/MM Btu.
 EPA policy  is  to  express all  data in Agency documents in metric units.
 For those particular  non-metric units utilized in this report conversion
 factors  have been provided.   These factors are Tocated on page xvii.
                                     1

-------
Specific Conclusions

     1.  The efficiency and selectivity of ferric sulfate solutions' in the
         leaching of pyrite from coal observed during previous investiga-
         tions! was reproduced in the current program with high ash
         (30 percent)  Run-of-Mine Lower Kittanning (ROM L.K.)  coal  containing
         4 percent pyritic sulfur.   Eighty-three percent of the pyrite was
         removed from 100 mesh top-size L.K.  coal  during 8.5 hours of pro-
         cessing a 20 wt percent coal slurry at 102°C under ambient pressure.

     2.  The L-R processing (simultaneous coal leaching-reagent regeneration)
         scheme developed in this program led to substantial improvement
         in the rate of pyrite removal from fine coal, up to at least 14
         mesh top-size, over that obtained under  ambient pressure  processing.
         Approximately eighty percent of the coal pyrite was removed in only
         two hours when 20 wt percent slurries of 100 mesh top-size L.K.
         coal were processed under L-R conditions at 120°C and 100 psig
         pressure (85 psi oxygen).   The large improvement in pyrite leach-
         ing rates appears to be principally  due to temperature increase.
         The rate advantage of L-R processing ceased to apply when pyrite
         removal exceeded approximately 80 percent.

     3.  The commercial scale processing scheme developed from bench-
         scale data for the desulfurization of high pyrite coal consists
         of the following unit operations connected in series:  mixer,
         L-R reactor, ambient pressure reactor, coal-reagent separation,
         processed coal wash, coal-water separation, elemental sulfur
         recovery, and coal drying.  All unit operations were successfully
         tested at bench-scale.


     4.  Pyrite  removal  occurs  in the mixer, L-R  reactor, and ambient
         pressure  reactor.  The  removal rate in all three reactors is gov-
         erned by  the empirical  leaching rate expression
         where
              W   =  wt percent pyrite in coal,


              Y   =  ferric ion- to- total iron ratio in the reactor
                     reagent, and»

              K.   =  rate constant, a function of temperature and
                     coal top-size.

-------
5.  Up to approximately 80 percent pyrite removal, the rate constant
    shows a strong temperature dependence expressible by
             = A,  exp (-
                                  EL/RT)
    Beyond 80 percent removal, the KL value appears virtually unaffected
    by temperature in the range investigated (90°C to 130°C).  Appar-
    ently, a change of reaction mechanism occurs when pyrite removal
    exceeds 80 percent; it is speculated that a diffusion controlled
    process takes over.  This change in mechanism may be specific to
    this particular high-ash coal.

6.  The E|_ value is independent of temperature,  coal  top-size (up to
    14 mesh), or reactor unit operation (processing conditions).   The
    value of AL is independent of temperature under mixer and L-R
    operations, but different for each operation.  The reasons for
    this unexpected AL value dependence on mode of operation are  not
    apparent from the available data.  The estimated EL and A Lvalues
    for mixer  and L-R  reactor processing and those of the diffusion
    rate constant are  summarized  in  the report.

7.  Reagent regeneration is governed by the rate expression
                   rR = - dFe

                           dt
                        • KR po
    where
                    exp (-
                  ER/RT),
              =  oxygen partial pressure, and
           +2
         Fe   =  ferrous ion concentration in the reagent solution.
AR and
                   are constants.
    The reagent regeneration rate operates simultaneously with the
    leaching rate in the L-R reactor.

    The KL value at 120°C under L-R processing conditions is five
    times higher than the KL value for 102°C processing under ambient
    pressure conditions (separate leaching and regeneration); both the
    14 and 100 mesh top-size coals exhibit the same large increase in
    KL value.  The estimated EL justifies only a two- fold increase
    in Ki  for 20 degree temperature rise-observed between 80° and
    102°C (separate Teaching-regeneration) and between 110° and 130°
    (L-R processing).   The 250 percent increase in the KL value,
    which appears to be the result of changing the mode of processing,
    could not be attributed to reaction of oxygen with pyrite.

-------
     Additional  rate  data  of more  fundamental  nature  are  required to
     delineate  the  reasons  for  the unexplained increase  in  KL.

 9.   The K|_  value obtained  with  100 mesh  top-size  L.K.  coal  was  approx-
     imately 20 percent  larger  than the K|_  obtained with  14 mesh top-
     size coal.  Since solid-liquid separation efficiency increases
     substantially  with  coal top-size, it appears  preferable to  process
     14 mesh coal in  view  of the small rate penalty involved. .

10.   The coal top-size effect on K|_ becomes substantial with ROM coals
     coarser than 14  mesh  top-size.  The  ratio of  K|_  values  obtained
     at 102°C with  100 mesh and 1/4 inch  top-size  L.K.  coals is  esti-
     mated to be equal to  4 up  to  approximately 60 percent pyrite re-
     moval and between 6 and 9  for higher pyrite removal.  L-R
     processing of  coarse  ROM coal at 120°C did not appear to improve
     the above ratios.   In fact, in the case of 3/8 inch  top-size ROM
     coal increasing  the temperature from 102°C to 120°C  has probably
     no effect on the leaching  rate when  pyrite removal  exceeds  20-25
     percent.

11.   Despite the adverse effect of coal particle size on  pyrite  leaching
     rates, coarse  ROM coal desulfurization by the Meyers Process is
     feasible.  Pyrite reduction in excess  of 80 percent was attained
     when 1/4 inch  x  0 ROM L.K.  coal was  leached for  48 hours with
     ferric sulfate solution  at 102°C.  The pyritic sulfur content of
     the ROM coal was reduced  from 3.48 wt. percent to 0.66 wt.  percent
     and  its heat content increased from  11,822 to 13,023 btu per pound.

12.   Preliminary data on cleaned coarse coal processing by the  Meyers
     Process indicate that physical cleaning of high  pyrite ROM  coal
     prior to chemical  desulfurization enhances substantially the
     pyrite leaching  rate. At  102°C, the pyrite leaching rate  obtained
     from processing  the 8 x  14 mesh fraction of cleaned L.K.  coal was
     twice that obtained from processinq  the same  size fraction  of ROM
     L,K. coal.  The  pyritic  sulfur content of the cleaned fraction  was
     reduced from 1.06 to 0.09  wt. percent.

13.   In addition to temperature and coal  particle  size,  the Ki  value
     may also be affected by  the type or  source of coal.   Data  obtained
     from the L-R processing  of Marion mine coal (Upper Freeport coal)
     indicate that  the pyrite  leaching rate constant  applicable  to this
     coal is at least an order  of  magnitude higher than corresponding
     KL values applicable to  L.K.  coal.   This comparison assumes that
     the same leaching rate expression applies to  both coals.

14.   The concentration of coal  in  the slurry and the  concentration of
     iron in the reagent affect the leaching rate  through their effect
     on Y.  These effects  become negligible under  continuous exchanqe
     or L-R processing.   Thus,  in  principle, the coal content of the
   •  slurry is limited only by  equipment  limitations  in the transfer
     of thick slurries and the  iron content of the reagent is limited
     only by the solubility of  the iron sulfate in water

-------
15.   The sulfate sulfur-to-elemental sulfur ratio of the product sulfur
     of the Meyers Process is independent of the mode of processing or
     the value of the processing parameters used within the ranges
     investigated.  The value of the ratio is approximately 1.5 indi-
     cating that 60 percent of the reacted pyritic sulfur converts to
     sulfate sulfur and 40 percent to elemental sulfur.

16.   Elemental sulfur recovery is virtually complete by extraction of
     processed coal with toluene.  However, preliminary data revealed
     that recovery by vaporization is also efficient.  The latter
     scheme combines sulfur recovery and coal drying into a simple
     operation and, therefore, appears to be the more desirable method
     of sulfur recovery.  Additional experimentation is needed in
     order to optimize sulfur product recovery techniques.

17.   Special investigations revealed that the Meyers Process does not
     affect standard coal analysis techniques.

18.   Experimentation with potential reactor construction materials for
     the Meyers Process revealed that 316L stainless steel could be
     safely used, but 304 stainless steel was inadequate.

19.   Process analyses for both fine and coarse coal processing show
     that the reactor/regenerator section of the process accounts for
     a large part of the equipment cost.  At commercial scale it was
     estimated that 40-60 percent of the total equipment cost occurs
     in this section,

20.  All  liquid  streams  of the Meyers Process are being recycled.
     Because  of  the  closed loop  design of the process, separate
     optimization  of the  reactor section generally leads to inoperable
     designs.  The  reactor trade-offs must be coupled  to the wash and
     by-product  recovery sections  in all process optimization efforts.


21.   The process analyses showed that either low purity oxygen (95%)
     or high purity oxygen (99.5%) could be used with no significant
     process cost impact.  It also showed that air can be competitive
     if energy is efficiently recovered from the reactor vent gas.  Air
     regeneration may be applicable only to high pyrite coals where
     the heat liberated by regeneration is sufficient  to saturate the
     large volume of inert vent  gas.

-------
                        RECOMMENDATIONS

1.   The principal  unit  operations  of  thevMeyers  Process  are  suffi-
    ciently developed at bench-scale  to  merit immediate scale-up to
    the pilot plant  stage.

2.   Bench-scale  testing should  continue  in  support  of pilot  plant
    testing for identification  of process improvements and for the
    evaluation of process  application to new  areas  or coals.   Bench
    scale represents the most cost effective  scale  to investigate
    process improvements and expanded utility.

3.   In support of pilot plant testing it is recommended  that bench-
    scale data be generated to  aid in coal  selection  and in  test plan
    formulation  best suited to  each of the  selected coals.   Bench-
    scale investigations should also  be  used  as  an  aid to understand
    and explain  unexpected observations  and to  resolve problems that
    may occur.

4.   In the area of process improvements, it is  recommended that bench-
    scale investigations be performed on improved techniques for sulfur
    product recovery, on schemes for  coal  trace  element recovery, on
    means of product utilization or disposal, and on  the identification
    and preliminary evaluation  of process  modifications  with the
    potential for reducing processing costs;  e.g.,  simultaneous sulfate
    and elemental sulfur recovery and elimination of  coal wash, combin-
    ation of physical coal cleaning and  chemical desulfurization.

5.   In the area of expanded process utilization, bench-scale studies
    are recommended to investigate the feasbility of  using the Meyers
    Process to render presently unusable coking  coal  acceptable to
    metallurgical industry (high sulfur  coal, middlings) and to exam-
    ine the advantages of combining chemical  desulfurization with
    coal conversion.  Simultaneous chemical desulfurization  and non-
    pyrite ash reduction should also  be  investigated.

6.   Coal pyrite leaching rate studies of more fundamental nature than
    previously performed are needed and  recommended.   These studies
    should aim at prediction of the efficiencies of chemical coal
    desulfurization from easily measured physical and/or chemical
    properties of coal.  At a minimum, studies  should be undertaken
    which identify the influence that the coal  matrix exerts on coal
    pyrite leaching rates.

7.   Process designs should be prepared for low pyrite coals to  identify
    the ranges of parameters that should be included in the experi-
    mental investigations.  Emphasis  should be placed on the  reactor/
    regenerator area which continues  to  be the major cost portion of
    the process.

-------
                            1 .  INTRODUCTION

     The Meyers Process can be used to desulfurize a large number of coals
prior to combustion in order  to meet govermental requirements for sulfur
oxide emissions control.

     The process consists  of  several steps  including crushing, chemical
treating,  sulfur removal and  solution regeneration.  While the process is
new, the unit operations are  similar to  various  existing technologies such
as  processes for the  heap  leaching  of copper,  regeneration of steel mill
waste pickle liquor and recovery  of elemental  sulfur from volcanic ash.
The chemistry of the  process  is represented by the pyrite leaching and
reagent regeneration  steps shown  below:
FeS2 + 4.6 Fe2(S04)3 + 4.8
2.4
               9.6  FeS0  + 4.8
                                       10.2 FeS04 + 4.8 H2S
                                        4.8 FeS0)  + 4.8
                                                                0.8S
overall process reaction:
      FeS2 +  2.4  02
                                         0.6  FeS04  + 0.2  Fe2(S04)3 + 0.8S
Generated iron sul fates can  be recovered by crystallization and stabilized
for disposal or a part of the iron  sul fate product may be neutralized with
lime to yield gypsum.  For the processing of fine or suspendable coal
(e.g., 14 mesh top size) leaching and regeneration are often economically
combined in a single process unit,  while for the processing of coarse coal
(e.g., -3/8 inch) the two reactions are more economically performed
separately.

    1 The Meyers Process is more  applicable to  coals  rich in pyritic sulfur
rather than those in which organic  sulfur is in high concentration.  Such
coal is found in the Appalachian region of the United States which now
supplies 60 percent of the current  U.S. production.  An estimated one-third
of the Appalachian production can be lowered to sulfur contents of 0.6 to
0.9 percent at which level the emission requirements for new power plants
can be met.  Additional Appalachian and Interior Basin coal can meet the
state emission standards by  this process.

                                     7

-------
    An initial  bench-scale testing program for the definition and obtaining
of critical  design data for the Meyers Process for the chemical removal of
pyritic sulfur from coal  was performed.    In addition, a survey program
covering application of the process for desulfurization of pulverized, raw
run-of-mine coal  from 35 U.S. coal mines has been completed. '   Run-of-mine
coal was used in all of these experimentation in order to accentuate any
possible processing difficulties which might be caused by the presence of
unusually large concentrations of pyrite  and inorganic ash.  The effort
demonstrated that the process, even when operating on ROM coal, has wide
applicability for the reduction of sulfur content of U.S. coal to levels
consistent with the Federal Standards for New Stationary Sources.

      This bench-scale program was aimed at optimizing and improving the
critical leaching and regeneration process steps, evaluating analytical
procedures and studying other process improvements.  As before, high ash
run-of-mine coal was used for the majority of the experimentation,  while
studies aimed at determining the benefits of clean coal processing utilized
coal  typical of washed Appalachian coal.
      This effort resulted  in the  attainment  of  the  necessary data and  de-
 finition of some significant improvements.   In  doing  so,  more than fifty
 fully material  balanced  process  simulation runs,  requiring  approximately
 2000 solution analyses and 1000  individual coal analyses, were performed.
 These data were assessed and evaluated in combination with  the results re-
 ported in the initial  bench-scale program.

      The resulting  report  is necessarily extensive, and   therefore a guide
 is  provided  to  the  reader  who wishes to  focus his attention on specific
 program  results.  The results in  this  volume are  presented  in six major
 sections  as  follows:

      •   Pyritic  Sulfur Removal from Suspendable Coal
      •   Reagent  Recyclability-Trace Flement  Data

-------
     t  L-R Processing of  Coarse  Coal
     •  Process Engineering
     •  Chemical Analysis  Studies
     •  Materials  Compatibility

while the complete data base is presented in Volume II of this report.

     Those readers desiring to  review process unit operations for suspend-
able coal are directed to  Sections 2.4 through 2.7 of this report.  Those
readers desiring process design,  estimated process capital cost and
estimated operating costs  are directed to Sections 5.1 and 5.3.  Coarse
coal depyritization data and the  engineering analysis for a coarse coal
processing scheme  are presented in Sections 4 and 5.2, respectively.
Section 6 of the report is devoted to sulfur analysis techniques for
processed coal and the identification of potential process monitoring
techniques.  Section 7 presents data on materials compatibility to process
environment.

-------
            2.  PYRITIC SULFUR REMOVAL FROM SUSPENDABLE COAL

     The bulk of  the program effort was  expended  in  the processing of
suspendable coal.  Coal is defined here  as suspendable if its top size
does not exceed 10 mesh.  Two1 coal sizes were investigated:  100 mesh x 0
and 14 mesh x 0.  The emphasis in experimentation was placed in optimization
of pressurized simultaneous coal Teaching-reagent regeneration (L-R) pro-
cessing and the parameters affecting this type of processing.

     The ensuing  report sections describe in detail the experimentation
performed, the procedures used, the data generated, and our interpretation
of the data.

2.1  COAL SELECTION AND SAMPLE PREPARATION

     A Lower Kittanning coal seam feed was selected for the majority of
experimentation on this program while the Upper Freeport Seams Sample was
processed for comparison purposes.  This selection was based on the avail-
ability of large  deposits of these coals, the fact that a substantial data
bank existed on ambient pressure processing of Lower Kittanning coal  by
                   1                   ^
the Meyers Process,  the availability of data on Upper Freeport coal  from
the coal survey project,  and the similarity of process behavior of these
seams to most Appalachian coals.

     Raw run-of-mine Lower  Kittanning coal was ground to 3/8 inch x  0 size
and shipped to TRW by  the Bureau of Mines, Pittsburgh, Pennsylvania  at
the request of the EPA Project Officer.  The coal was received in three
55-gallon drums with each drum containing approximately 150  kilograms of
coal.
                                    11

-------
      The coal  from each  drum,  when  needed  for  experimentation,  was  suc-
 cessively halved by the  use  of commercial  rifflers  to  approximately 20
 kilogram lots.   This was the minimum  lot size  used  for further  grinding
 (100 mesh and  14 mesh top sizes).   Upon grinding, each 20  kilogram  lot
 was subjected  to the same riffling  procedure  (halving) until  the desired
 coal sample siz.e was attained.  The nominal coal sample sizes used  in this
 program were 2 kilograms (20 wt. percent slurry processing)  and 4 kilograms
 (33 wt. percent slurry processing).  Samples  submitted to  analyses  were  in
 the 200 to 300 grams range.  The described riffling procedure is the most
 commonly used method  for coal  sampling and it is ASTM approved  (ASTM
 sampling analysis  procedures  utilized in this program are listed in
 Appendix G, Volume II).  Both the  "as received" and the unused  portion
 of the ground coal  were  stored  under helium in order to inhibit weathering.
 However,  sulfur-forms analyses  of  the unprocessed coal performed periodi-
 cally  during the  program indicated that some coal  weathering took place
 in spite of the precautions taken  (see Table 1).

      The unprocessed coal was  completely characterized prior  to initiation
 of processing.   The coal characterization  procedure included  complete coal
 analyses and coal  particle size  distribution determinations.  The coal
 analyses involved  short  proximate (moisture, ash, and  heat content), sulfur
 forms,  ash composition,  and  trace element  determinations.  In addition,
 short proximate,  sulfur  forms, and  ash-iron analyses were  performed period-
 ically  on the unprocessed coal during the  duration  of  the  program in order
 to  assure proper sampling and  to monitor coal  weathering.  The  coal
 analyses,  with the  exception of  trace element  analyses, were  performed by
 Commercial  Testing  & Engineering Company (CT&E), Chicago,  Illinois.   The
 trace element analyses were  performed by TRW's  Applied Chemistry Department.

     Table  1 summarizes  the analysis data  on the raw Lower Kittanning coal
used in this program.  The data  are presented  chronologically with respect
to  the dates of analysis in order to  illustrate the increase in the sulfate
content of the coal with storage time and  the  associated decrease in pyrite.
                                     12

-------
TABLE 1.   RAW LOWER KITTANNING COAL ANALYSES
Date
Analyzed
Oct.
1973
Dec.
1973
Apri 1 -
June
1974
May
1974
June
1974
Sept.
1974
Oct.
1974
Coal
Top Size
100
Mesh
14
Mesh
100
Mesh
14
Mesh
14
Mesh
3/8 inch*
14
Mesh
Number
Of Samples
Analyzed
3
2
4
2
1
3
4
Coal Composition, Dry
Ash
(Wt.%)
30.08
+ .02
29.76
;K70
30.85
**
32.32
+ .35
30.12
28.91
+ .52
31.77
+ .71
Heat
Content
(Btu/Lb)
10,520
± 26
10,626
+ 173
10,275
**
10,109
+ 106
10,324
10,448
+ 106
10,020
t 77
Total
Sulfur
(Wt.%)
4.55
+.03
4.48
+ .01
4.53
+.06
4.77
+.01
4.50
4.28
+ 14
4.55
+.07
Pyritic
Sulfur
(Wt.%)
4.05
+ .01
4.01
+ .01
3.88
+.19
3.85
+.02
3.85
3.56
+ 18
3.77
+.04
Sulfate
Sulfur
(Wt.%)
0.08
+_.02
0.07
+ .01
0.24
+ .04
0.28
+.01
0.26
0.35
+_.03
0.41
+.02
Organic
Sulfur
(Wt.%)
0.41
+ .04
0.41
+ .01
0.41
+ .17
0.64
+ .01
0.39
0.37
+ .03
0.38
+ .05
Iron
(Wt.%)
3.91
+ .03
4.03
+ .12
3.98
+ .06
4.20
+ .04
3.89
3.68
+ .13
4.00
+ .09
Moisture
In Raw
Coal
(Wt.%)
1.20
+ 34
1.51
±.03
1.19
+ .22
1.81
1.23
1.16
+ .04
1.66
+_.19
These samples were taken from a different drum than the rest.
**Single analysis.

-------
The observed trend with time in the values of these two sulfur forms, the
internal consistency in individual sample analyses (each increase in
sulfate is associated with a corresponding decrease in pyrite), and the
good agreement in the total sulfur and iron values among coal samples
derived from different ground lots and analyzed at different times indicate
that the variations in pyritic and sulfate sulfur from sample to sample
are not due to sampling or analysis error but to coal  weathering.  With
one exception, the data in Table 1 verify the adequacy of raw coal sampling
procedures used by TRW and the consistency in analyses performed at CT&E.
The indicated standard deviations are within ASTM standards.

     The exception to the overall consistency of data in Table 1 involves
the two samples of 14 mesh x 0 coal analyzed in May 1974.  The higher ash,
lower  heat content, higher total sulfur, and higher iron values of these
samples (compared to the rest of the data on coal derived from the same
drum)  would be internally consistent and would indicate inappropriate
sampling  if the analyzed higher total sulfur was due to higher inorganic
sulfur forms  in the sample  (higher inorganic sulfur leads to higher ash
and lower heat content in the sample).  However, the sulfur forms data
from these two samples show that the unexpected high total sulfur content
was due to high organic sulfur.  This analysis is not consistent with
the higher iron and ash content of the samples which suggest that the
pyrite or iron sulfate forms should have exhibited the high values.  Thus,
in addition to sampling procedure the sample analyses became suscept.
The same lot of ground coal was re-sampled and reanalyzed in June and
yielded the expected values (Table 1).

     Discrepancies in sulfur forms analyses were more frequent with pro-
cessed coal.   The low sulfur-high ash content of the processed coal and/or
residual elemental sulfur left on the coal during processing could have
been the culprits of poor sample analysis.  Whenever these discrepancies
occurred,  the coal sample analysis data were subjected to the  same internal
consistency scrutiny as that described for the raw coal.
                                               s
    Table 2  presents the data on the mineral content of the coal utilized
(ash composition).  These data were supplemented with trace element
                                     14

-------
analyses; the latter are presented in a later section of this report de-
voted exclusively to the fate of trace elements during coal  processing.
The ash composition data appears to be typical of Pennsylvania bituminous
coals (the quantity of ash in the coal was not typical; its softening
temperature was 1385*C (2525°F)).
           TABLE 2.  LOWER KITTANNIN6 COAL ASH COMPOSITION

           Ash Component                          Wt. %

         Phos. pentoxide, p og                     0.11
                   /
         Silica, Si02                             49.63

         Ferric oxide, Fe203                      18.66

         Alumina, A120                            25.64

         Titania, Ti02                             1.87

         Lime, CaO                                 0.75

         Magnesia, MgO                             0.53

         Sulfur trioxide, S03                      0.79

         Potassium oxide, K20                      1.64

         Sodium oxide, Na20                        0.29

         Undetermined                              0-09
                                                 100.00
                                     15

-------
     Table 3 presents  data  on  the  particle  size  distribution of the re-
ceived 3/8 inch x 0 coal  and of the  two  suspendable  coal  sizes investigated
in this program (14 mesh  x  0,  and  100 mesh  x  0).   At least three determi-
nations were performed on each of  the coal  top sizes;  the sample to sample
reproducibility was very  good.   The  particle  size  distribution presented
in Table 3 proved to be almost identical  to that obtained from the Lower
Kittanning coal used in the previous program  (EPA  contract EHSD 71-7).
       TABLE 3.   LOWER KITTANNING COAL  PARTICLE  SIZE DISTRIBUTION
Size Range
(Mesh)
+6
-6 +14
-14 f28
-28 +42
-42 +60
-60 +80
-80 +100
-100 +115
-115 +150
-150 +200
-200 +325
-325 xO
Lost
Three sample ave
Weight Fractions*
100 Mesh x 0






.090
.051
.054
.136
.592
.071
.006
rages (maximum de
14 Mesh x 0

.020
.200
.160
.090
.080
.040
.040
.030
.100
.210
.020
.010
viation 10%)
3/8 Inch x 0
.248
.260
.150
.093
.054
.038
.018
.019
.013
.026
.054
.021
.006

                                   16

-------
      Coal  grinding and riffling to the desired sample size for processing
 (2 to 4 kilograms) were the only operations performed on the coal prior
 to mixing it with the reagent fdr processing.  Thus, the data generated
 on this program was on ground run-of-the mine coal.

 2.2  COAL PROCESSING EXPERIMENTATION

      As a result of previous bench-scale investigations on the Meyers
 Process, a conceptual process scheme was proposed, for processing suspend-
 able coal, involving the following basic unit operations:  slurry
 preparation (mixer), simultaneous coal Teaching-reagent regeneration (L-R
 reactor), ambient pressure residual pyrite leaching-coal settling (settler),
 coal washing, elemental sulfur recovery, coal drying, and product sulfate
 recovery.  The major objective of the current program was to investigate
 the validity of the above conceptual scheme.  The principal difference
 between the above scheme and the one defined in the previous bench-scale
 program is L-R processing (simultaneous Teaching-regeneration under pres-
 sure versus separate coal leaching and reagent regeneration unit operations
 with coal leached at ambient pressure).  Thus, the emphasis was placed on
 L-R processing; however, its effects on the other unit operations as well
 as potential improvements on these operations were also investigated.
 The investigation of product sulfate recovery (process slip-stream treat-
 ment) was postponed until the basic process parameters are finalized
 (especially reagent composition).  Preliminary examination of this unit
 operation has been proposed to commence at the conclusion of this program.

      Figure 1 is a block diagram of the basic processing scheme used in
 the investigations of suspendabTe coaT.  Every effort was made to simuTate
rthe envisioned full-scale processing to the degree allowable by the size
 of bench-scale equipment.  Figure  2 depicts in greater detail the front
 end of the process and the apparatus utilized (mixer, L-R reactor, and
 settler).  The ensuing paragraphs describe the procedure and the equipment
 used.
                                      17

-------
00
           COAL
            11
               COAL
               EAGE
               MIXER
           PRODUCT
             UUAL
               DRIER
    PRODUCT
    SULFUR
LJLFUR
             DISTILLATION
                                          REAGENT
\
:NT
:R
10° C



SLURRY

y
COAL LEACH
REAGENT
REGEN.
110-130°C
A


SLURRY

| f
COAL
SETTLER
90° C
SLURRY ||^ER



•Mm
^ |
FILTER
\\\\\\


i

                      TOLUENE
    WET COAL
                                 ELEM.
                                SULFUR
                               RECOVERY
                             SPENT
                             TOLUENE
I
                                          FILTER
                                                          HOT
                                                          WATER
SLURRY
                    FILTER
                   T/rrrn
                                                                                            HOT
                                                                                  WET       WATER
                                                                                  COAL  I
                COAL
               WASHING
               80-90°C
                                                                     WATER (DISPOSAL)
                        Figure 1.   Basic Bench-Scale Processing Scheme for the Removal
                                   of Pyritic Sulfur From Coal

-------
                           Pressure vJet
                           Control  Test
                                   Meter
                                                                       Pump Seal
                                                                      Purge System
1
1
1 '
1
1


1

1
f
i


90° - 95°c
•'.- '";- '- V •;>'•• ,.--••->.;-;.
ilSlliiflii
linSSffi^^Sli^


1 I
i
i
,1
21
1
^t
1
!',eactor-Sett ler •
                                                                                                            riNAL  PROCESSINGS

                                                                                                     Sti rrer  Reactor
                                                                                                   _l
STEP  #1  (-v.1.0 HOUR)
STEP #2  (1-8 HOURS)**
STEP #3 (18-24 HOURS)
          •'Final processing  includes elemental sulfur recovery, coal washing and drying.

         **Reactor volume  ^13  liters
              Figure  2.   Bench-Scale  Coal  Leaching and Reagent  Regeneration Apparatus

-------
      The ground coal  (14 or 100 mesh top size)  was mixed with hot spent or
 fresh reagent in the  "mixer" and the resulting  slurry was brought to
 boiling; slurry boiling was maintained  until  coal  wetting was complete
 (slurry foaming subsided).   Actually, the "mixer"  was comprised of 3 to 4
 four-liter glass flasks equipped with heating tape and stirrer; each flask
 contained approximately 2 liters of slurry.   The "defoamed"  slurry was
 then transferred to the L-R reactor; the reactor was pressurized with
 nitrogen gas to the desired operating pressure  (50 to 150 psig) and the
 slurry was heated to  the desired L-R temperature (110 to 130°C).  This
 entire "slurry preparation and heating" operation ("mixer operation") re-
 quired approximately  one hour (mixing time,  tm), although in early experi-
 mentation it was substantially longer.   Once the desired L-R temperature
 was reached, a slurry sample was taken  and then oxygen was introduced to
 the reactor.  The composition of the slurry sample taken at  this point
 (t,n = 0.0 hours) defined the starting coal  and reagent composition of
 L-R processing and simultaneously served as  the means of determining the
 extent of reaction during t .

      The L-R reactor  and its associated parts (feed and sampling lines,
 pump, and slurry circulation loops) were constructed from 316 stainless
 steel stock.  The reactor was a cylinder 100 cm (39.4 inches) in length
 with a 14 cm (5.5 inches) OD and 12.7 cm (5.0 inches) ID.  The reactor
 was flanged at the one end  (top) and a  hemispherical bottom was welded to
 the other end.   Both  end pieces were equipped with 2.54 cm  (1  inch)
 openings.

      The  reactor  was  loaded  with  slurry through the top and  drained through
 a one  inch  ball valve  attached  to the reactor bottom.   The reactor was
 operated one-half to three-quarters full.

     During L-R processing the  slurry was  circulated through a loop ex-
ternal to the reactor  (Figure 2) at a rate of approximately  4  liters per
minute.  Slurry circulation was necessary  for oxygen introduction and for
sampling.  Slurry was  circulated with the  aid of a 316 stainless steel
centrifugal  pump having a 3.2 cm  (1.25  inches)  OD  suction and an 1.9 cm

                                      20

-------
(0.75 inches) OD discharge.  The pump was equipped with a water irrigated
double mechanical seal.  The slurry was withdrawn from a reactor opening
located 30 cm from its bottom and, after oxygenation, it was returned
through a tangential inlet to the reactor bottom.  The main circulation
loop was equipped with two by-pass loops.  One was used to trap slurry
samples at the desired reaction times.  The second was used to return part
of the oxygen depleted slurry through the top of the reactor as a spray.
The principal function of the second by-pass was to control flow through
the main  loop without  the  need  of  excessive  throttling of  the  pump;  an
intended  second  function during early  experimentation was  slurry foam
control.   However,  top spraying the  reactor  slurry did not prove to  be
a very effective foam  control device).  This  by-pass loop  was  assembled
from  a variety of stainless  steel  tubing and  as  such it also served  as a
means of  testing the compatibility of  these materials with L-R processing
 (the  materials involved and  the data derived  from these tests  are de-
cribed in  a  separate section of this report).

      Oxygen was  introduced to the slurry immediately downstream of the
sample trap at a  rate which  varied from approximately 0.5  to 2.0 liters*
per minute; the  oxygen feed  rate depended on the quantity  of coal being
processed and on  L-R processing  time (higher rates were fed at the start
of L-R to purge  the nitrogen in  the  reactor and to accommodate the higher
reaction rates).  In principle,  due  to the prevailing turbulent flow con-
ditions, intimate  gas-slurry  mixing took place in the recirculation loop
prior to slurry  return to the reactor.  The excess oxygen, oxygen not
consumed by reagent regeneration,  percolated through the reactor, passed
through a condenser which removed water vapor and entrained reagent  solu-
tion, and exited  the reactor through an automatic pressure relief valve.
The quantity of  reactor vent gas  (principally oxygen) was  measured Jay a
wet-test meter.

     Several slurry samples  were taken during L-R processing (normally
every 0.5 hours during the first two hours of processing and every hour
thereafter).  The drawn slurry  samples were weighed and immediately
                                                        i o       "H^
filtered.  The filtrate was  analyzed for iron forms (Fe    and  Fe   ).  The
 At Standard Temperature and  Pressure (STP).
                                     21

-------
 reagent  wet  coal was spray-washed with hot water to remove the reagent
 solution from  it, extracted in toluene to leach out elemental sulfur,
 dried at 100°C under vacuum, and weighed.  The weight percents solids in
 the slurry samples were determined and compared to the starting values as
 a means  of evaluating  how representative the samples were of the reactor
 mixture.  The  dry coal samples were analyzed for ash, heat content, total
 sulfur,  sulfur forms,  and iron.

      After the desired L-R processing time was completed, the reactor was
 depressurized  and drained.  The L-R slurry was transferred immediately to
 either a "settler" for further reaction or to the filtration apparatus or
 to both.  In all cases, because of equipment size restrictions, the slurry
 was processed  through  the post L-R operations in two batches.  Three pro-
 cessing  options were used:  (a) both batches were "settler" processed prior
 to filtration, washing, elemental sulfur recovery, and drying; (b) only
 one of the batches (approximately one-half of the L-R slurry) was  "settler"
 processed prior to further processing;  or (c) neither batch was "settler"
 processed.   In addition, there were two options in "settler" processing:
 (a) reaction in an agitated "settler"  or  (b) reaction in a non-agitated
 "settler".   Whatever the option used, slurry processing through the re-
 maining  operations was the same.  (There were a few exceptions where L-R
 processed coal  was used for special experimentation in washing and elemental
 sulfur recovery operations; these will be noted as they occur in data pre-
 sentation.)

      "Settler"  processing was conceived as a means of concentrating dilute
 L-R slurries of fine coal while simultaneously removing the last few percent
 of  pyrite  from  L-R processed coal.  Since it was realized that a relatively
 quiescent  reactor could be subject to reagent concentration gradients and
 a consequent reduction in the pyrite leaching rate, an agitated "settler"
was also used in parallel for comparison.  Indeed, the non-agitated  "settler"
was not as effective a pyrite leaching reactor as the agitated one because
of reactant and product segregation.  The non-agitated settler was abandoned,
during early experimentation, not because of reagent segregation problems
but because L-R processing of thick slurries proved efficient and  a  thickener
 was judged unnecessary.  The agitated  "settler" was retained as  an ambient
 pressure  reactor where the slow reacting, last  few percent of pyrite could
                                     22

-------
be leached out  of  coal.   Thus,  in the majority of experimentation one-half
of the L-R slurry  was  processed through an agitated settler and  the  other
half was transferred directly to the  filtration and washing unit operations.
The coal derived from  these  two slurries (after washing,  toluene extraction,
and drying) was used to  determine L-R processing and "settler" processing
efficiencies.

     The  "settlers", covered 5-liter  cylindrical glass vessels,  were
equipped with heating  tapes, condenser, and stirrer (used only for agitated
"settler" processing).   The  slurry was processed in these vessels for  18-24
hours at 90°C and  at ambient pressure.  Each settler was  sampled at  the
start and at the end of  processing.   The slurry samples were analyzed  for
iron forms only.

     The hot slurries  from either the L-R reactor or "settler" were  vacuum
filtered.  The  reagent-wet coal  was subjected  to a 2 stage  washing scheme
with both stages consisting  of  the following:   a slurry wash with a  quan-
tity of water equal to from  2 to 4 times the weight of dry  coal  estimated
to be in the filter cake; a  cake wash with two dry coal weights  of water.
The slurry washes  were performed at reflux temperatures for 30 minutes.
Additional water washing or  coal washing with  0.1 N sulfuric was performed
on processed coals whose analysis indicated excessive iron  sulfate de-
position.  All  filtrates were analyzed for iron forms.

     The product elemental sulfur was extracted from the  processed coal
with toluene (special  experimentation was also performed  with hexane).
The water-wet coal filter cake  from the wash section was  slurried in twice
the estimated dry  coal weight of toluene.   The slurry was heated to  the
toluene-water azeotrope  temperature (85°C)  where it was maintained until
the water in the slurry  (from the wet cake)  was distilled-off (the toluene
was returned to the slurry continuously during distillation); the slurry
was then refluxed  for 30  minutes at its  reflux temperature  (108°C).  Subse-
quently,  the slurry was filtered  hot  and  the resulting coal  cake was
rinsed with a small quantity of  fresh toluene.   The elemental sulfur in the
                                     23

-------
toluene filtrate was recovered as a residue upon vaporization of the toluene.
The toluene residue contained small quantities of dissolved coal along with
the elemental sulfur.  Each residue was analyzed (in triplicate) for sulfur.
In the majority of experiments a double toluene  extraction was performed.

     The toluene-wet coal was vacuum dried in a well  trapped oven.   The
dried processed coal was then sampled for analysis.   Each batch of pro-
cessed coal was subjected to short proximate, sulfur  forms, and iron
analyses.

     The procedure described was the one followed with the majority of
coal processing experiments.  There was, however, special experimentation
performed which involved single unit operations.   These special require-
ments are described in conjunction with the data presentation.

2.3  COAL PROCESSING DATA

     Approximately 50 batches of suspendable (100 and 14 mesh top size)
Lower Kittanning coal were subjected to L-R processing (simultaneous coal
Teaching-reagent regeneration processing) during the  feasibility and
parametric investigations of the L-R mode of operation of the Meyers Pro-
cess.   The following parametric effects were investigated over the ranges
indicated:

     •  Reaction time  (Mixer, 1-2 hours; L-R, 1-8 hours; Settler, 18-24 hours)

     •  Slurry concentration (20 wt.  percent and 33 wt. percent coal in
        iron sulfate)

    •  Starting reagent  iron concentration  (near 0.0 to  5  wt.  percent)

    •  L-R reaction temperature  (110°C  to  130°C)

    •  L-R reaction pressure (50 to  150 psig,  with a corresponding oxygen
       partial pressure variation of 35 to  135 psia)
                                     24

-------
     •  pH (1-4 to 1.9); actually, this was an  investigation of the effect
        of sulfuric acid on  L-R processing.

Special experimentation was  performed in  the  following areas:   ambient
pressure  desulfurization  of the  Lower Kittanning coal  by continuous
reagent exchange,  slurry  foam control, elemental  sulfur recovery,  coal
washing,  solids-liquid separations,  and reagent recyclability.

      It should  be  noted that the  major objective of^this program was to
test  the  feasibility  of conceptual Meyers Process improvements and to de-
termine the  processing scheme and operating parameters of the improved
process.  Thus, the parametric investigations were limited to ranges con-
sidered as the  most economically and technically practical for  process
utilization.  The  number  of  parameter values  investigated were limited to
the minimum  necessary to  (1) .identify trends, (2) establish the most
probable  parameter values  for use in the  process  design and (3) develop
empirical rate  expressions which  adequately describe coal  pyrite extraction
under the L-R mode of operation of the Meyers Process.

     The processing data is  summarized in the 15  tables of Appendices A,
and C of this report  (Volume  2).   Appendix A  includes  the data derived from
processing 100 mesh top size coal; Appendix B includes the 14  mesh  top size
coal  data; and Appendix C presents the data derived from L-R processing of
weathered coal  with low-iron reagent  (the starting reagent does  not contain
iron;  it derives iron  from coal during slurry preparation and  processing).
  -s
The experiments are numbered consecutively starting with Appendix A through
Appendix C.   To a large degree, the numerical  order of the experiments
corresponds  with the order in which they were performed.  Each table con-
tains  experiments (usually more than one)  performed under the set of nominal
conditions indicated in its title.  Specific conditions pertaining  to
individual experiments are included with the rate data derived from the
particular experiment.  Two types of data are presented for each experiment:
rate data, indicating  the change in slurry and coal composition as  a
function of  reaction time, and mass balance data, showing the over-
all process  balance for solids, liquids,  sulfur,  and iron.  Examples of the
                                     25

-------
 data  presented in Appendices A through C and table explanations are given
 below.

      Table 4  presents the rate data derived from Experiment 1.  The
 coal  was  processed under the conditions indicated and in accordance with
 the procedure described in the previous section (Section 2.2).  The first
 row of  data indicate the actual coal (solids) content of the slurry
 (as opposed to the nominal or intended value used in the table title),
 the Y value of the starting reagent (the ferric ion to total iron ratio
 in the  reagent), and the composition of the starting coal on dry coal basis.
 The next  five rows present data derived from slurry samples (150-200
 grams each) drawn from the L-R reactor during processing.

      The  first column of data in Table 4 indicates the L-R reaction times
 (t, R)  at which samples were drawn.  It also shows the time, t^,
 required  to mix the slurry (to thoroughly wet the coal in order to
 minimize  slurry foaming) and to heat it to the desired L-R reaction
 temperature,  120°C in this case.  A sample was always drawn at the start
 of L-R  processing (end of "mixer operation"); a slurry sample was also
 drawn at  the  conclusion of L-R operation (t,  D= 2 hours, in this case).
                                           L-K
 The second column shows the solids content of the slurry samples; this data
 served  as an  indication of the adequacy of the sampling procedure.  As seen,
 the solids content of each sample is nearly identical to that of the
 starting slurry.   The third column shows the reagent Y value of the liquid
 phase of each sample as determined from iron forms analyses.  Columns
 four  through ten present the analyzed dry coal composition data for each
of the slurry samples.   The last two columns show the computed pyritic
sulfur,  S , removal  from coal  as a function of reaction time.  Pyritic
sulfur removal values were computed from coal sulfur forms analysis data;
the Sp removal during tm was also computed from ferrous ion production data.

      The removal values "Based on Analyzed S  " were  computed  from the
differences in the pyritic sulfur content of  the  starting  coal  and that
of the particular sample; the S  values were  determined  from  coal  sample

                                    26

-------
           TABLE 4.
RATE DATA  ON L-R PROCESSING  OF TOO MESH TOP SIZE  L.K.  COAL WITH 3 WT.  % IRON
SOLUTION AT 120°C AND TOO PSIG (85 PSI 02).   NOMINAL SLURRY SOLIDS  20  WT. %
                                                                RATE DATA
EXPERIMENT NO. 1: tm = 1.5 Hours, tL_R = 2 Hours; tg = 0; Fe in Starting Reagent = 2.98 Wt. %; Starting Y = 1.0
Reaction Time, Hours
(t = time in mixer;
t. = time under L-R
processing;
tg = time in settler)
Starting Coal (5 Sample
Average) and Slurry
Sample
Slurry
Composition
Wt. %
Solids
18.9
Y
(Fe+3/Fe)
1.0
Coal Composition Wt. % (Except Heat Content), Dry
Ash
29.62
+.66
Heat
Content*
Btu/Lb
10,573
± 91
Total
Sulfur,
St
4.52
+.05
Pyritic
Sulfur,
SD
4.03
+_.03
Sulfate
Sulfur,
Ss
0.08
+.02
Organic
Sulfur,
So
0.41
+_.04
Iron,
Fe
3.97
+.12
% S Removal
Based on
Analyzed
SD

Based on
Corrected
SP

Reactor Samples
*L-R = °'° Hours
(tm =1.5 Hours)
t,D =0.5 Hours
•1.0 Hour
=1.5 Hours
=2.0 Hours
18.6
17.7
N.A.
19.0
18.5
.28
.58
.75
.84
.84
26.52
24.95
28.05
27.76
27.89
11,059
11,172
10,310
10,104
9,974
2.95
2.42
2.25
2.58
2.23
2.07
(3.09)*
1.52
1.17
0.45
0.49
0.29
0.42
0.82
1.09
1.39
0.59
0.48
0.26
1.04
0.35
2.74
2.60
3.10
3.55
4.04
49
(25)*
62
71
89
88
Processed Coal Composition
L-R Processed Coal { Jj** J
Reanalyzed Sample (Batch A + B)
30.60
26.22
25.93
10,179
10,993
11,090
2.00
1.46
1.65
0.15
0.08
0.67
0.32
0.31
0.18
1.53
1.07
0.79
3.64
3.17
1.33
96
98
83
45?
61
75
74
90

691 7fi
82 176
74
tv>
       *  Coal  S  content and % S  removal based on FE2 (ferrous ion) production during t .

-------
analyses performed at CT&E (Commercial  Testing  &  Engineering  Laboratories).
Thus,
     % S  Removal  = % Sp (starting coal)  -  % S  (sample)  x 10Q
                           % S  (starting coal)
(1)
This method of computation assumes that the weight of the coal remains
constant during the reaction.  This, of course, is not the case because of
ash  (pyrite) dissolution, but the error introduced in the above computation
is insignificant  (less than one percent error in the calculated removal).

     The pyrite removal values "Based on Corrected S " were computed the
same way except that an adjusted sample S  value was used instead of that
derived from coal pyrite analysis.  The adjusted or corrected sample Sp
value  was  computed as follows:

     S (corr.) = S  (Anal.) + SQ (Anal.) - 0.43                      (2)'

The above  correction is based on the following assumptions.  First, the organic
sulfur content  (wt. percent) of the coal, S , was not affected by the Meyers
Process and that  its value will change only because of changes in the
weight of  the coal (ash dissolution).  For the coal used in this program,
this assumption fixes the coal S  range during processing between 0.41 and
0.44 weight percent (0.43 was selected as the "expected S " for all samples
in order to reduce the computations involved; the introduced error  is in-
significant).  Second, the analyzed values for total sulfur, St> and sulfate
sulfur, Ss, are correct; and third, the elemental sulfur produced during  the
process was completely recovered.

     Experimentation performed to date has shown that the Meyers Process
does not affect the SQ content in the coal other than through changes in
coal weight because of ash dissolution.  However, the SQ value is determined
indirectly from $t, Sp, and Sg analyses (S0 = St~Sp~Ss~V under the  assump-
tion that the unrecovered product elemental sulfur  in processed coal, S   is
virtually zero.   Thus, the accuracy of the SQ value depends on the  accuracy
of St,  S ,  and S$ analyses and the completness of Sn recovery.  Equation 2,
above,  assumes that meaningful errors affecting S  determinations occurred

                                     28

-------
only in S  analyses.  Based on available data this assumption appears to be
statistically valid for the following reasons:  S. and S  analyses per-
                                                  V      S
formed on the same sample consistently agreed within ±0.05 or better;
abnormalities in SQ determinations were observed  at both high and low
(near zero) S. values; unrecovered SM can not account for S  values below
             s                      n                      o
0.43 wt percent.  A more direct proof of the validity of Equation 2 can be
derived from the smoothing effect its application had on S  removal rate
data.  It should be emphasized, however, that application of Equation 2 to
every S  value may not have been warranted; it was employed only for the
sake of consistency (occasional erroneous application of Equation 2 does
not affect conclusions drawn on the performance of the Meyers Process).

     The last row of  information in Table 4 presents data on the composition
of the processed coal.   In this particular experiment the coal was not
"settler" processed;  the L-R  slurry was processed in two batches through
the rest of the unit  operations (filtration, washing, elemental sulfur
recovery, and drying) because of equipment size limitations.  Thus, in
principle, the two-hours reactor sample and the samples from Batches A and
B should have been of the same composition (identically processed coals).
Obviously they were not.  Even though a part of the discrepancy could be
attributed to sulfur  forms analysis errors, poor  sampling appeared to also
be responsible.  The  two batches of L-R processed coal were mixed, refluxed
for 30 minutes in 0.1N H2S04, dried, sampled, and analyzed (washing with
acid was performed to remove  deposited iron and sulfate sulfur thought to
interfere with the accuracy of analysis).  The "corrected S " value of the
combined batch agreed well with the average of the "corrected S " values
of the individual batches, but not with the two-hour reaction sample.

     A coal sampling  problem  appears to be responsible for the discrepancy
in the two S  values  of the starting L-R coal (tL_R = 0.0 sample).  The top
value (2.07 wt. percent) was  obtained from sulfur forms analysis of the
coal in the drawn slurry sample.  The value in parenthesis was computed
from the measured quantity of ferrous ion produced during tm.  According to
Meyers Process chemistry, repeatedly verified, 10.2 moles of Fe   are pro-
duced per mole of pyrite oxidized.  Fe+2 production from side reactions of
                                     29

-------
 Fe+3  with  the organic matrix of the Lower Kittanning coal or with non-
 pyritic  ash  components is negligible.  Since at t  Fe+  is not being
                                     +2
 regenerated  dur to lack of oxygen, Fe   production is an accurate measure
 of pyrite  removal from coal in the mixer (tm reaction time).  Thus, the
 S  value in  parenthesis and the corresponding S  removal value were
 assumed  to be the correct values.  The discrepancy between the directly
               +9
 analyzed and Fe   computed S  values have been attributed to poor reactor
 sampling at  t,R = 0.0.  Because of inadequate slurry mixing during the
 start of L-R operation, the sample drawn from the circulation loop con-
 tained the smaller particle size coal representing the coal fraction with
 higher pyrite depletion.  Later reactor samples were substantially more
 representative of the coal in the reactor.

      Early in the program we experienced difficulties in obtaining repre-
 sentative  slurry samples (partially due to slurry foaming) and as a
 consequence  the decision was made not to rely completely on reactor samples
 for the  determination of S  removal rates as a function of reaction time.
 Thus, separate experiments were performed for a number of L-R times, t^_^,
 ranging  from one to eight hours.  This decision greatly expanded the
 planned  experimental effort, but it also greatly improved data reliability.
 Alleviation  of foaming problems and improved sampling procedures resulted
 in more  representative samples as a scan of the data in Volume 2 and the
 data  in  Table 5  (Experiment 7)  below indicate.   However, sampling,
 even  of  processed dry coal, was not completely consistent.  At least in
 part,  the  sampling difficulties can be attributed to the unusually high
 ash content  of the coal utilized.

      Table 5 is identical in format to Table 4 except that  it  includes the
 composition  of "settler" processed coal (last row of Table  5).   Settler
 processing, described in the previous section, was always  conducted  at 90°C.
 The coal  slurry residence time in the settler in hours  is  indicated  by the
 ts value  in parentheses.   The "AG" notation indicates that  the  "settler"
was stirred  ("NAG" was used to denote quiescent  "settler"  processing).
The difference in the S  removal values between  "L-R Processed Coal" and
 "L-R  + Settler Processed Coal" represents  the  percent of starting coal
 pyrite removed during "settler" processing.

                                     30

-------
              TABLE 5.  RATE  DATA ON L-R PROCESSING OF 100 MESH TOP SIZE L.K.  COAL WITH 5 WT.  % IRON
                         SOLUTION AT 120°C AND TOO PSIG
                                                                RATE DATA
cvDCDTMrwT wn •, tm = T'0 Hour' Si p = 1'° H°ur» te = 1 9 Hours (%50% of Coal); FET (Iron in Reagent) = 4.94 Wt. %;
EXPERIMENT NO. 7: smarting Y (Fe^Ffr) =0.95
Reaction Time, Hours
(t = time in mixer;
t = time under L-R
processing;
ts = time in settler)
Starting Coal (5 Sample
Average) and Slurry
Sample
Slurry
Composition
Wt. %
Solids
19.2
Y
(Fe+3/Fe)
.95
Coal Composition Wt. % (Except Heat Content), Dry
Ash
30.85
+.46
Heat
Content
Btu/Lb
10,275
+ 60
Total
Sulfur,
St -
4.53
+.06
Pyritic
Sulfur,
SD
3.88
±.09
Sulfate
Sulfur,
ss
0.24
+.04
Organic
Sulfur,
S0
0.41
+.07
Iron,
Fe
3.98
+.06
% S Removal
Based on
Analyzed
SP

Based on
Corrected
SD

Reactor Samples
t, _R = 0.0 Hours
(t =1.0 Hour)
m
tL_R = 0.5 Hours
» 1.0 Hour
20.2
20.4
19.1
.54
.72
.88
>8.97
>7.14
>5.15
10,558
10,823
11,147
3.68
2.77
1.75
3.12
(3.01)*
2.22
1.04
0.19
0.25
0.29
0.37
0.30
0.42
3.14
2.41
1.57
20
(24)*
43
73
21
46
73
Processed Coal Composition
L-R Processed Coal
L-R + "Settler" (AG) Proc. Coal (tg = 19)
28.55
25.64
10,831
11,072
1.98
1.21
1.43
0.49
0.23
0.27
0.32
0.45
1.88
1.25
63
87
66
87
CO
        Coal  S  content and % Sn removal based on FE2  (ferrous ion) production during t .
             p               p                                                 m

-------
     Table 6 presents "Process Mass Balance" data from Experiment 1.
The  solids, liquids, sulfur, and iron overall process mass balances are
summarized in separate sections of the table.  The table is a computer
print out summary of individual process unit operations.  A mass balance
computer program was written which utilized detailed information on the
composition of feed and output streams (weights and analysis data) from
each process unit operation of each experiment, performed the mass balance
calculations on solids, liquids, and slurries around each unit operation
(mixer, L-R reactor, settler, wash, elemental sulfur recovery Teacher, and
coal drier), and compiled the data on overall process mass balance shown
in Table 6.

     The "solids balance" in Table 6 refers to coal.  The "input" and "dry
process coals" (completely processed coal plus samples) are direct
weight measurements (the moisture content of the feed coal was estimated
from short-proximate analyses performed on the "as received" raw coal; it
averaged very nearly one weight percent).  The processed coal and samples
were corrected for  deposited reagent.  The quantity of deposited reagent
was  estimated from the iron and sulfate sulfur analyses of the starting
and  processed coals (including samples) and from the extent of pyrite re-
moval (S  analyses).  The excess iron on the processed coal (reagent
derived iron) was assumed to be FeSO. and Fe^OU.  For the majority of
experiments this correction was small (little, if any, reagent deposition
that withstood washing).  The dissolved coal mineral matter was determined
from starting and processed coal ash analyses converted to mineral matter
(assuming the ash to be Fe203 and the mineral matter FeSp); it is shown
in Table 6 as "pyrite dissolved" and as "non-pyritic ash dissolved".  The
first quantity was determined from S  coal analyses and the second by
difference.  Table 6 indicates that 3 grams of "non-pyritic ash" was de-
posited on coal during processing.  This is probably due to normal ash
analysis uncertainty (deviations).  A scan of the data from all experiments
indicates that a very small quantity of non-pyritic mineral matter was
dissolved during processing.  The "coal dissolved in solvent"  (elemental
sulfur recovery unit) was determined from the weight and  sulfur  content of
the residue derived from the distillation of the spent organic  solvent
(toluene or hexane, toluene in the large majority of experiments).
                                     32

-------
                                      TABLE 6.  PROCESS MASS BALANCE DATA
     EXR.  1,   L.K.100XOMESH, 20PCT  SLURRY(NOMINAL) . 120C, 100PS1G(85PS1  02)

     OVERALL  SOLIDS BALAKCI                                      OVERALL  COAL  SULFUR BALANCE
       DRY COAL  INPUT                              1882

       HRY  FRnr.ESS COALSCCORR. FOR  OEF.  REAG.l     1773
       FYKI It  LliSSOLVEO DURING REACTION             111
       6SHUON-PYRITIC) DISSOLVED DURING REACTION  -15
       COAL  DISSOLVED IN SOLVENT                      3
       TOTAL SOLIUb RECOVERED                      1872
       LOSS  (INPUT - RECOVERED)                      10
       PERCENT RECOVERY OF SCLIOS                    99
                                                              INPUT COAL SULFUR                           85

                                                              PROCESS COAL SULFURCCORR.  FOR OEP. REAG.I  24
                                                              SULFUR IN SAMPLES                             6
                                                              SULFUR OISSOLVEtHAS SO'*)  IN LEACH R-1AGENT  36
                                                              SULFUR RECOVERED WITH  ORGANIC SOLVF-MT      18
                                                              TOTAL SULFUR RECOVERED                      83
                                                              LOSS (INPUT - RECOVERED)                      2
                                                              PERCENT RECOVERY OF SULFUR                 98
co
CO
OVERALL LIQUIDS  BALANCE

 REAGENT
 COAL MOISTURE
 RINSE AND  FILTER PQPER 4ATE*
 DISSOLVED  ASH  AND PYRITE
 HASH MATER
 SOLVENT
 TOTAL LIQUIDS  IN

 FILTPATE  REAGENT
 SAMPLE  AND FOAM REAGENT
 DEPOSITED  REAGENT
 hASH FILTRATES
 CRGAMC  FILTRATES
 AZEOTROPE  AND  ORGANIC FILTRATE
 LIQUIDS  IN DRYER TRAPS
 LIQUIDS  ON EQUIPMENT
 TOTAL  LIQUID RECOVERED
 LOSS  (IN - RECOVERED)
 l-ERCfcMT  RECOVERY OF LIQUIDS
                                                   7747
                                                     19
                                                    H79
                                                     96
                                                  22844
                                                   8848
                                                  H0033
                                        AT E R
  992
   64
22048
 8337
  274
  660
   53
38382
 1651
   96
                                                                  OVERALL  IRON  BALANCE
                 STARTING COAL
                 STARTING REAGENT
                 TOTAL IRON  INPUT
PROCESS COAL  IRON
SAMPLE COAL IRON
REAGENT FILTRATE  IRCN
SAMPLE FILTRATE IRON
FOAM MATERIALS  IRON
REACTOR WASH  IRON
WASH FILTRATE IRON
TOTAL IRON RECOVERED
LOSS (INPUT - RECOVERED)
PERCENT RECOVERY OF
 75
231
306

 55
  7
152
 2-9
  0
  0
 36
279
 27
 91

-------
     The "liquids balance" entries do not require explanation; they repre-
 sent direct or by difference weight measurements.  The entry "sample and
 foam reagent" could be an exception.  In the majority of experiments, the
 weight of liquid shown for this entry represents the quantity of reagent
 drawn during reactor slurry sampling.  In early experimentation, however,
 we experienced some slurry foaming during the early stages of L-R processing;
 as a result a small quantity of foam left the reactor through the oxygen
 exhaust valve.  This foam was trapped, weighed, filtered, and analyzed.
 Its components (principally reagent) were then entered in the appropriate
 mass balances.  It should also be noted that the "dissolved pyrite and ash"
 entry in the "liquids balance" does not include the produced elemental
 sulfur; the latter is part of the "organic filtrate" entry.

      In the "overall coal sulfur balance" the "input",  "process coal",
 and  "sample" sulfur values were determined from coal weights and Eschka
 analyses.  The processed coal and samples sulfur values were corrected for
 deposited reagent sulfate.  The "sulfur recovered with organic  solvent" was
 determined from sulfur analysis of the organic solvent residue.  The  "sul-
 fur dissolved as sulfate" was computed from the measured pyritic sulfur
 removal (coal S  analyses) by assuming that 60 percent of the removed  S
 was dissolved as sulfate.  This assumption was based on extensive experi-
 mentation performed on Lower Kittanning coal in the previous bench-scale
 program  which indicated that pyrite removed from coal by the Meyers
 Process was converted and recovered as sulfate sulfur and elemental sulfur
 at a ratio estimated to be 1.5 (the ratio was estimated from coal S   anal-
 yses and ferrous ion production).  Within experimental uncertainty, the
 same ratio appears valid for this program's coal, even when processed  under
 L-R conditions at as high a temperature as 130°C.   In the majority of
 experiments, the recovered elemental sulfur equalled 32-35  percent of the
 S  removed from the coal instead of the expected 40 percent;  this computes
 to Ss/$n ratios in the range of 1.8 to 2.1  (Experiment 1  represents  an
 exception in that the recovered $n is less than  30  percent  of the estimated
 removed S ).   However, complete elemental sulfur recovery is  difficult and
the determination of recovered elemental sulfur  is  subject  to error.   During
the previous program, elemental sulfur recovery  based on  an S /S  ratio of
                                                              *>   n
1.5 range  between 80-90 percent (32-36 percent of the removed pyritic sulfur),
                                     34

-------
The same appears to be true with the present coal based on the quantity
of elemental sulfur recovered and the S  removals as computed from coal
analyses before and after processing.   (In the previous program the S /S
ratios were also verified from Fe+2 production; because of Fe+2 oxidation
during L-R operation this verification technique could not be used in the
majority of experiments performed in the current program.)

     It should be noted that 80-90 percent elemental sulfur recovery does
not necessarily imply that this sulfur product was incompletely removed
from coal.  Incomplete recovery from toluene (the extraction medium),
small errors in analysis or weighing of the sulfur residue, and small
errors in the computation of S  removal can easily account for 10-20 percent
elemental sulfur deficiency.
                      ;

     The  "iron  balance"  entries were  computed  from weight measurements
 and  iron  analyses.   The  origin  of each  entry  is  evident from  its  label.

 2.4  DATA DISCUSSION AND INTERPRETATION

     The  data  in Appendices  A through C (Volume  2)  indicate that  in  the
 majority  of the L-R experiments performed  the  overall  process mass balance
 of solids,  liquids, sulfur,  and iron  were  better than  +.5 percent; in-
 dividual  unit  operation  mass balances were even  better.   In view  of  the
 fact that the  overall  process mass  balances of each  experiment  involved
 over one  hundred weight  measurements  and analysis determinations, that
 each experiment involved several  manual  slurry transfers, and that the
 system was  relatively small  (e.g.,  the  sulfur  balance  involves  less  than
 100 grams of sulfur in most  cases),  the mass  balances  attained  should be
 considered  excellent.   In the rare  cases where larger  than 5  percent dis-
 crepancies  occurred in one or more  of the  "balanced" quantities,  the
 discrepancy was probably due to untraceable errors  in  weights and analyses
 rather than losses  during processing.   The biggest  losses were  in liquids
 and they  represent  water evaporation  losses during  coal washing.  In
 general,  the mass balance results indicate that  the  data  can  be used with
 confidence  in drawing conclusions on  the effectiveness of L-R processing
                                      35

-------
on the leaching of pyrite from Lower Kittanning coal.  Individual pieces
of data or data involving a small portion of the processed slurry could,
of course, be in substantial error and yet not influence the described
mass balances.  In fact, this was the case with a number of reactor samples,
as indicated in the previous section.

     The raw rate data in the same Appendices of this report (Volume 2) do
not exhibit the consistency of the mass balance data.  However, these
apparent inconsistencies are explainable.  It will be shown later that,
despite some scatter, the data are internally consistent and relatable to
that generated under ambient pressure processing in the previous bench-
scale program (separate Teaching-regeneration operations).

     The  scatter  in  rate data was traced to three causes:   sampling (es-
 pecially  reactor  slurry sampling), sulfur forms analyses,  and  oxygen
deficiency-  The  problems experienced  in sampling and sulfur forms  analyses
were alluded to earlier.  In general,  slurry  sampling improved greatly
when  "dead  spots" were eliminated from the reactor sampling loop and  when
slurry  foaming was reduced  substantially by "wetting" the  coal  in the
mixer section.  The  only consistently  poor sample was the  one  taken at the
start of  L-R operation, most likely  because coal  particle  size distribution
homogeneity had not  yet been accomplished in  the  reactor  (this sample was
taken almost immediately after the start of slurry circulation).  However,
the S  content of coal at the start  of L-R was determined  from the  quantity
of ferrous  ion produced during tm and  a  poor  coal sample  at the start of
L-R did not present  problems in  data interpretation.  The  inconsistencies
in sulfur forms analyses were substantially reduced  through the application
of the organic sulfur correction to  S  discussed  in  the previous section
(Equation (2), page  28).    Oxygen deficiency  during  L-R processing  was
probably the single most important reason for the observed rate data
scatter and it was certainly the one responsible  for the  larger deviations
in the  L-R rate data.  (Oxygen was used during L-R processing  to  regenerate
      +3
the Fe    being depleted by pyrite oxidation.  Oxygen concentration  in-
               +2      +3
fluences the Fe   to Fe   conversion rate; the iron  forms  the  pyrite
oxidation  rate as explained below.)

                                     36

-------
     According to data derived from ambient  pressure  processing of Lower
Kittanning by the Meyers Process each mole of  pyrite  leached  from coal
consumes (reduces) 9.2 moles of Fe"1"3 and produces  10.2 moles  of Fe+2.  The
quantity of elemental sulfur produced during L-R processing appears to
indicate that the same stoichiometry applies to this  mode of  processing.
Ambient pressure data also  indicated that the  Sn leaching rate depends on
                    +3                         P
the square of the Fe  -to-Fe ratio, Y.  If the same rate dependence on Y
exists under L-R processing, a small change  in the Y  value should translate
into substantial effect on  the S   removal rate.  This appears to be the
case upon analysis of the  "rate data" Tables contained in Appendices A and
B,  Volume 2.  Since oxygen  is used to regenerate Fe   (one mole of oxygen
regenerates four moles of  Fe  ), even a small  deficiency in  oxygen during
L-R processing can have a pronounced effect  on S   leaching rate through
its influence on Y.
     The above discussion  is  illustrated  by  the  data  depicted  in  Figure 3.
This figure summarizes the data from 14 experiments performed on 14 mesh
top size coal under the conditions indicated (data from Tables B-l, B-2,
B-3 and B-4, Appendix B, (Volume 2).  The remaining data on this size coal
(11C°C  and 130°C, Tables B-5 and B-6) were not included in order to avoid
additional  clutter of the plotted data; however, it can be easily estab-
lished  through inspection of Tables 5 and 6 that the majority of these data
fall within the band traced by the plotted data.  In order to minimize data
scatter from errors in sulfur forms analyses the "corrected" S  removals
during t  were estimated from ferrous ion production; all other data was
derived from direct coal analyses of reactor samples  and product coal
samples.  The data designated by bars (samples from 4 experiments) and that
labeled as "outlier samples" were derived from experiments processed under
identical conditions except for small differences in  the starting reagent
Y.  The plotted data indicate that during L-R processing the only discernible
parametric effect is that of slurry concentration during the first hour of
processing which could be reasoned to be just an apparent effect due to the
higher $„ removal during t .  In fact, the "outlier samples" deviate more
        p              3  m
from the mean of the plotted data than any sample representing parameter!c
variations.  These outliers represent samples from Experiments 30 and 31
(Table  B-2).  The Y values of these samples differ only by approximately
                                     37

-------
   90
   80
o
u

1
Q
UJ

I
ex:

u.
U
   40
5  30
LU
   20
   10
                    o
                    A
           T
            T 33 WT. % COAL SLURRY, 120°C. 100 PSIG
            1 (SEVERAL EXPERIMENTS WITH THE D BEING
              OUTLIER SAMPLES)
            • 20 WT. % COAL SLURRY, 120°C, 100 PSIG

            A 33 WT. % COAL SLURRY, 120°C, 50 PSIG (35 PSI 02)

            *33 WT. % COAL SLURRY, 120°C, 150 PSIG (135 PSI 02)
          m
                         1.0        2.0        3.0
                               L-R REACTION TIME, HOURS
                                       4.0
5.
8.0
       Figure 3.
Pyritic Sulfur Removal from  14  Mesh Top Size L.K. Coal
Processed at 120°C with 5 Wt. % Fe Reagent (Data Summary)
                                        38

-------
0.10 from those of corresponding data from Experiment 28  (or from the
average Y of all the experiments represented by the bars  in Figure 3);
however, the indicated S  removals differ by nearly 100 percent at t,
0.5 hours (22 percent versus 42 percent for the average of the "bar" data)
and nearly.50 percent at tL_R =1.0 hour (39 or 40 percent versus 58 percent
for the averaged data).  This example does not only illustrate the dif-
ficulty of sensing oxygen deficiency during L-R processing, but it also
underscores the importance of Y to pyrite leaching rates  and it appears
to indicate that ferric ion is the oxidizing agent in pyrite removal
during L-R processing just as it was determined to be during ambient
pressure processing  (separate Teaching-regeneration).  It can be concluded,
therefore, that under efficient regeneration the S  removal rates during
L-R should be at least as high as those indicated by the  upper bounds of
the plotted data.  This translates to 80-85 percent pyritic sulfur removal
from 14 mesh x 0 Lower Kittanning coal after approximately two hours of
L-R processing plus  t .  Since results from the previous  bench-scale
program indicated that the same extent of S  removal required 10-12 hours
of ambient pressure  processing at 102°C the L-R S  removal rate at the
110°-130°C range must be substantially higher than that obtained at 102°C
with separate regeneration.  This conclusion is fully validated in Section
2.4.2.

      Oxygen  deficiency during part of the experimentation could  have
 been  the  result of insufficient feed rate or inadequate  mixing due to
 deficient apparatus  design  or inappropriate  procedures;  or  it  could have
 been  unavoidable because  the  demand  of oxygen  by  the  system exceeded  theo-
 retical limitations  (e.g.,  maximum oxygen  solubility  in  the slurry) or
 practical means of oxygen  incorporation into the  slurry.   The  scatter in
 the data of  nearly identically  processed  coal  tends  to indicate  that  the
 oxygen  deficiency  was due  to  inadequate experimentation  rather than being
 due to  unavoidable causes.  Thus,  it is our  opinion  that conclusions  con-
 cerning the  efficiency of  L-R processing  should be  based on the  assumption
 of efficient regeneration  (rate of Fe+3 regeneration  as  predicted by  the
 kinetics of  the reaction).
                                     39

-------
      The scatter  in  the data,  the apparent  insensltivity  to  parametric
 changes, and the  nearly abrupt reduction  in  S   removal  rate  after two
 hours of L-R processing (Figure 3)  hindered  initial efforts  to apply
 previously derived rate expressions  on  leaching and on  regeneration to
 L-R processing or to develop new ones which  adequately  described the pro-
 cess.  It was not until virtually the complete  set of data was generated
 that the oxygen deficiency  probability  became apparent  and that the flat-
 tening of the S  removal  rate  became statistically valid.  Once these
 unexpected occurrences were recognized, it became evident that a single
 rate expression (or  at least a single rate constant expression) can not
 describe the entire  leaching operation  of this  particular coal and that Y
 values predicted  from the simultaneous  solution of previously derived Sn
                     +3                                                 "
 leaching rate and -Fe  regeneration  rate expressions would not predict the
 generated data.   The coal leaching  process,  therefore,  was divided into
 four parts (processing ranges)  and  the  data  derived from each one were
 examined separately.  The four processing ranges are:   (a) mixer processing,
 (b) L-R processing up to  80 percent  S   removal, (c) L-R processing between
 80-90 percent S  removal  (2-8  hours  L-R processing), and  (d)  "settler" pro-
 cessing.  Data discussion and  process definition in the outlined four
 processing ranges are presented in  the  ensuing  report sections in which
 process behavior  in  each  major process  unit  operation is  described and
 discussed separately.

      The above discussion and  conclusions apply to the  L-R processing of
 100 mesh top size coal as well  (data in Appendix A, Volume 2).

 2.4.1   Mixer Unit Operation

      Initially, the  "Mixer  Unit Operation" was  defined  as that part of
 processing where  hot, approximately  70°C, reagent was mixed  with "as
 received" coal and the resulting slurry, pressurized by nitrogen gas to
the desired  L-R pressure, was  heated to the  desired L-R temperature.  How-
ever, substantial  slurry  foaming during L-R  processing, especially  in the
early stages of L-R processing, required that the mixer section  be modified
so as to furnish a substantially non-foaming slurry to  the L-R section.
                                    40

-------
     An extensive small scale  investigation  (200-500 grams of coal) was
undertaken in order to identify the causes of coal  slurry foaming, to
establish practical foam control methods, and to define the proper "mixer
section" for L-R processing of suspendable coal.  The following parameters
were examined for their effect on coal slurry foaming and its control:
slurry pH, iron salt concentration and Y of  reagent, slurry solids content,
coal particle size, coal moisture, slurry temperature, retention time
near slurry boiling temperatures, anti-foam  agents, wetting agents, coal
washing (soaking in hot water), and slurry reflux.  In addition, the
foaming characteristics of processed and raw coal slurries were compared
and the gas evolved (principally C02) from raw coal treated with acids
was determined.  The quantity  of foam (height in a reactor column) gener-
ated during boiling of the coal slurry after treatment, or after additives
had been incorporated, and its consistency (solids content and persistence)
served as criteria for measuring the various effects.

     Slurry pH (in the range of 1 to 7), the Y of reagent, slurry temper-
ature up to boiling and "soaking" time up to 2 hours at any temperature in
this range, anti-foam agents,  and coal washing did not affect slurry
foaming.  Slurry foaming decreased with decreasing  iron concentration, but
it remained at unacceptable levels even when iron was reduced to zero
(dilute sulfuric acid or water coal slurries).  Foam volume, density, and
stability increased with increasing solids content  in the slurry; however,
even very dilute slurries foamed.  The maximum CO^ evolution measured
(approximately 150cc of gas per kilogram of  coal) was not sufficient to
account for the volume of the  generated foam.  Slurry foaming increased
with decreasing coal top size  in the range of 14 to 200 mesh, but it re-
mained at unacceptable levels  even with the  14 mesh x 0 coal; removal of
the 200 x 0 fraction from the  14 mesh top size coal reduced slurry foaming
substantially and rendered the 14 x 200 mesh coal processable.  Wetting
agents changed the nature (consistency) of the foam radically, but not its
volume; the normally thick, black slurry foam was converted to thin (soap
like), clean foam.  Combinations of wetting  and anti-foam agents arrested
slurry foaming completely.  Slurry reflux at ambient pressure for approxi-
mately 30 minutes virtually eliminated slurry foaming.  Finally, the
                                    41

-------
 moisture content of coal appeared to have an effect on the volume and
 density of coal  slurry foaming with slurry foaming decreasing with in-
 creasing coal moisture content, i.e., the coal was in essence "pre-wetted".

      The above investigations led to the conclusion that the principal
 cause of slurry foaming is the presence of dry (unwetted) coal in the
 slurry and that Lower Kittanning coal  which equilibrates in air to only
 about one percent moisture is difficult to wet.  Three  approaches
 rendered Lower Kittanning coal slurries processable through L-R:  (a)
 coal  wetting  through the addition of wetting and anti-foam agents, (b)
 slurry reflux for approximately 30 minutes and (c) removal of coal fines
 (200  mesh x 0 fraction) prior to processing.  All three methods were tested
 in bench-scale L-R processing runs and all proved successful.  However,
 removal of the fines fraction (a substantial portion of suspendable coal)
 was the least desirable method because  a  different technique had to be
 identified for processing the fines by the Meyers Process and because
 this  approach was not as effective in foam control as coal wetting.  The
 wetting, anti-foam agent approach was rejected because of the cost of
 these agents  and because it appeared to adversely affect L-R processing
 (refer to data in Tables A-7 and B-7, Volume 2); this method would have
 been  used if  foam arrest by slurry reflux did not prove successful.  The
 slurry reflux method was judged to be by far the most desirable because
 of its simplicity and easy adaption to L-R processing.  This approach-
 was therefore, selected for incorporation into the "mixer unit operation"
 as standard practice.

      Mixer processing, "mixer unit operation", in all suspendable coal
 experiments involving simultaneous leaching-regeneration (except for  the
 two experiments in which wetting agents were used) consisted of the fol-
 lowing steps:   slurry preparation and heating to reflux temperature  (102°C)
 slurry defoaming by 30 minutes reflux, slurry transfer to the L-R reactor
and pressurization, and slurry heating to the desired  L-R temperature.
Normal slurry mixer processing times, tm, ranged between 1.0  and  1.5
hours; occasionally,  longer times were used either because foaming persisted
or because  of equipment problems.  The approximately 10  kilograms of slurry
                                    42

-------
used  in each experiment was  prepared with 65° to 70°C reagent,  split into
3-4 four-liter glass  flasks  equipped with condensers, and transferred
to the 13-liter  stainless  steel  L-R reactor when defoaming was  completed.
The temperature-pressure- time  history of mixer processing was recorded
in detail during each experiment.   Figure 4 depicts  typical mixer
processing conditions.

     Pyritic  sulfur removal from coal during tm varied from near zero  (less
than 5%)  in experiments where no-iron containing starting reagent was  used
to nearly 30 percent when coal was  processed with 5 wt. percent iron re-
agent as  20 weight  percent slurry  for longer than normal  t  .  Shorter t , 3
weight percent Fe reagent, and 33  percent coal  slurries gave intermediate S
removals  (see data in Appendices A, B, and  C, Volume 2).  This type of
parametric behavior was expected from previous experience with ambient
pressure processing of Lower Kittanning coals.  The next step was to investi-
gate the predictability of the t   data generated in this program by the
pyritic sulfur leaching rate expression developed in the previous bench-
scale program (EPA contract EHSD 71-7) from ambient pressure data generated
at 70°, 85°,  and 102°C on a similar Lower  Kittanning coal, namely,

                       „  _     dWp   _   (/• ., 2 V2
                       rL ~  "  dt         L Wp  Y '
where
     r.      is the pyrite leaching rate, expressed in weight of pyrite
            removed per 100 weights of coal per hour (rate of coal
            pyrite cone, reduction),

     W      is the pyrite concentration in coal at time t in wt. percent,
      P
     t      is the reaction (leaching) time in hours,

     Y      is the ferric ion-to-total iron ratio in the Teacher at
            time t, dimensionless, and

     K,      is the pyrite leaching rate constant (a function of tem-
      L     perature and coal particle size) expressed in (hours)-'
            (wt.  percent pyrite in coal)"'
                                   43

-------
   150*H
CD
i—i
CO
o.
a:
a.


o
DC

UJ

X
   100*-
     50*-
                       AIR
OXYGEN

0
e
«
UJ
OL
ra
i
UJ
Q_
UJ
1—
0
I-H
1 —
<
UJ
oc
:
LU
X
t— i
S


130*
120*
1 fm\J
110*
100


80-
60-
40-
20-
0-
i
SLURRY HEATING L'R PROCBSIN6 AT 12°°C
/"X^^ /
SLURRY REFLUX >V /
/
\ /
/ \ /
/ \ /
/ v

T
SLURRY TRANSFER TO L-R REACTOR


in ?'n Vn dh (in f»n ?'n
             "MIXER"  REACTION TIME, MINUTES



       Figure 4.   Typical  Mixer Processing Conditions
  Utilized L-R Temperatures and Pressures


                                     44

-------
with
                       KL =  AL  exp  (-EL/RT),

where

     AL     is the Arrhenius frequency factor in the units of K ,

     EL     is the apparent activation energy in calories/mole,
     R      is the gas constant  in calories/mole°K, and
     T      is the absolute temperature, in °K
 From data  generated  in the  previous  bench-scale program, the K,  value  for
 pyrite  leaching  from 100 mesh  x  0 UK.  coal  at 102°C was determined  to be
 between 0.12  and 0.15 (hours)"1  (wt.  percent pyrite)"1.   The corresponding K,
 values  for 85°C  and  70°C leaching were 0.09 and 0.03, respectively.  The
 Arrhenius  plot of these three  rate constants did not furnish the expected
 straight line; thus, the A.  and  E.  values  derived  from these constants and
 reported in Reference 1  were based on the  85°C and 102°C data.   Recent
 review  of  the 85°C and 70°C data revealed  that the determined rate constant
 for the 85°C  run was 0.088  + 0.044 while that  of the 70°C data  was 0.03
 +_ 0.013 (single  experiment  data  for  each temperature with 50 rate data
 points  taken  in  each experiment).  It was, therefore, concluded that the
 70°C rather than the 85°C K, value should  have been used for the AL  and EL
 determinations.   It  was' also noted that a  KL value of 0.07 for  the 85°C
 run gives  a straight line Arrhenius  plot for all three temperatures  with
 KL (102°C) =0.15.   The new data assessment led to two sets of  AL and  EL
 values  depending on  which end  of the 102°C KL  range was used with the  70°C
   value:
 For  102°C KL =  0.15
     A   =  4.7  x  io6  (hours)"1  (wt.  percent  pyrite  in  coal)"  and
     EL  =  12.8 Kcal/mole
 For  102°C  KL =  0.12
     AL = 3.4 x  TO5  (hours)"1  (wt. percent  pyrite  in  coal)
     EL = 11.1 Kcal/mole
                                     45

-------
Similar but less extensive rate data  generated  with 14 mesh x 0 Lower
Kittanning coal  revealed that the  KL  value  for  this size coal  was  approxi-
mately 20 percent lower than the corresponding  KL for 100 mesh x 0 coal;
thus, at 102°C

     0.10 < K14 £0.12 (hours)"1 (wt.  percent pyrite in coal)" ;

the corresponding A^ is

     2.7 x 105 <. Aj4 <. 3.8 x 106       (same units)

assuming that the activation energy is the  same for both top size  coals,
namely

     11.1 1 EL <. 12.8 Kcal/mole

     Prior to utilizing the above  rate expressions to predict "mixer" pro-
cessing, it was considered necessary that at  least one ambient pressure
leaching experiment be performed under continuous reagent exchange conditions
with the L.K. coal used in the current program  in order to compare identi-
cally generated data from the two  coals (present and previous program L.K.
coals).  The experiment was performed at 102°C  (slurry reflux temperature)
with 100 mesh x 0 coal slurried in ferric sulfate solution containing 4.9
weight  percent  iron.  The total reaction time was  8.5  hours  (8.0  at  102°C plus
34 minutes slurry heating time between 70°C and 102°C).  The generated rate
data and overall process mass balance data are summarized  in Table 7.

     The first part of Table 7  (first page) shows  the  measured  Fe+2/Fe
ratio (1-Y) and the computed change in coal weight,  pyrite removed,  average
rate of removal (from reaction  start to the indicated  reaction  time),  and
coal composition as a function  of reaction time.   The  pyrite removal was
computed from measured ferrous  ion production  as a function of time  as-
suming that pyritic sulfur was oxidized to sulfate sulfur and elemental
sulfur at the ratio of 1.5.  All other computed  quantities were derived
from the calculated pyrite removal plus a small  correction for any non-
pyritic ash dissolved or  added to  the coal during reaction.  (The

                                     46

-------
   TABLE 7.    PYRITE REMOVAL FROM COAL WITH IRON SULFATE SOLUTION AT 102°C AND AMBIENT PRESSURE
SUHMARY OF REACTION DATA
                           SC«t/*:= l.SOO
LOH£R KITTANNTNG - 100  *ESH
MM
FPQM
START
34
49
64
79
94
109
124
139
154
169
134
199
214
229
244
259
274
289
304
319
334
349
364
379
394
409
4?4
439
454
469
484
499
514
FROM

FF+P/FE
WT. RATIO
.3333
.355$
.3626
.3294
.2791
.2682
.2752
.2P65
.2P.88
.2813
.2662
.247?
.2373
.2102
.2C85
.2C50
.le.88
.1839
.1868
.1825
.1705
.1544
.1429
.1329
.1250
.1265
.1245
.1?19
.1154
.1123
.1103
.0946
.0771
FINAL CCAL
COAL VT.
F9F." CP
SCELEf)* P
641 . ?7
634.05
62?. 44
622. 65
618.37
612.95
607.14
601.37
595.93
590. 61
585.37
580.3'=
575. 03
570.43
565.79
560.92
555.94
551.23
546.18
541.32
536.64
532.18
527.70
?23.26
518.60
513.81
509.14
504.60
500.08
495.53
490.95
486.83
482. 88
ANALYSTS
PYSITF HLMULflTIV
•?FI*OVrD
!CT OF TNIT
18.78
23. S5
?7.76
3?. 06
31. 90
35.03
39.43
43.63
'7, 01
tc. fir)
52 . 34
54.05
57.19
58.87
60.69
63.17
65.29
66.45
6C.73
70.41
71.54
72.01
7?. 66
73.76
75.11
76.90
78.3ft
79.1.5
P0.23
P1.15
62.13
PI. 93
81.24
P3.64
RATE
PCT/HR
33.13
29.33
26.02
24.35
20.36
19.28
19.08
18.83
18.3?
17.68
17.07
16.30
16.03
15.42
14.92
14.63
14.30
13.80
13.57
13.24
12.85
12.38
11.98
11.68
11.44
11.28
11.09
10.86
10.60
10.38
10.18
9.85
9.48

CALCULATED
TOTAL S
HT-PCT
3.791
3.591
3.442
7.274
3.28H
'.157
2.984
2.817
?.683
2.571
2.469
2.401
2.275
2.207
2.133
2.032
1.946
1.899
1.806
1.738
1.691
1.672
1.645
1.600
1.545
1.471
1.410
1.366
1.334
1.296
1.255
1.263
1.29?
1.670
PYR. S
HT-PCT
3.306
3.104
2.955
2.785
2.792
2.668
2.493
2.326
2.190
2.078
1.976
1.907
1.780
1.711
1.637
1.536
1.449
1.402
1.308
1.239
1.192
1.171
1.146
1.100
1.045
.971
.909
.865
.832
.794
.753
.761
.790
.690
CCAL ANALYSES STEP-BY-STEP
ORG. S
WT-PCT
.414
.415
.416
.417
.417
.418
.419
.420
.421
.421
.422
.422
.423
.423
.424
.424
.425
.425
.426
.426
.426
.426
.426
.427
.427
.427
.428
.428
.428
,428
.429
.429
.429
.750
S04 S
WT-PCT
.081
.081
.081
.081
.081
.082
..082
.082
.082
.082
.082
.082
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.083
.084
.084
.084
.084
.084
.084
.230
ASH
WT-PCT
29.388
29.323
29.276
29.221
29.223
29.184
29.128
29.074
29.031
28.995
28.962
28.940
28.899
28.877
28.853
28.821
28.793
28.778
28.748
28.726
28.711
28.705
P8.696
28.682
28.664
28.640
28.620
28.606
?8.596
28.583
28.570
28.573
28.582
28.550
HT CONTENT
BTU/L8
10637
10655
10668
10683
10682
10693
10709
10723
10735
10745
10754
10760
10772
10778
10784
10793
10801
10805
10813
10819
10823
10825
10827
10831
10836
10843
10848
10852
10855
10859
10862
10861
10659
11013
Coal weight in grams

-------
                                                 TABLE 7.    (CONTINUED)
Co
                                     ICW^p KITTAKNTNG - 10Q
                             OVERALL PeOCrSS lALANCfS Ff>« "OLIO*;, LTIUTO1^, SULFUP flUO I»ON*
          ANT
HRV CC«L TNPLT
                IN
                DISSOLVED HURUG
           CC«L  IN  SOLVFNT
           FRCV  FILTER
      CVCN DRIE[) COAL
TOTAL ^OLTPS  ^ECOVFPir
EBLANC"  (RFCCV^EO-INFUT)
         "?FCOVCKY OF
                                      T«CTTCN
                                                     7.?
                                                     3.n
                                                   (»7n. 0
               ?  RALANCF
LEACH ";EAGPNT  IMPUT
ASH AKP PYRITE OISSOLVFO DURING
TOTAL HATFR  (WASHES  AKC WFTTUG FILTFR5)
TOTAL CRGAMC  SOLVENT  TN°UT
TOTAL LIQUTOS  IN
LFACH CFAGENT  R^COVE^FC
HOUir RFMCVEC IN  SAPPLF^ «ND FILTFPS
TOTAL WASH FILTRATES
SOLVENT DISTILLATE
LIQUirS IN DRIER TRAPS
FOUND CN ECUIFfENT ANO  FINAL  FILTFR PAPERS
TOTAL LIQUIDS  RECOVERFT
PALANC1! (RECCVFRED-INFUT)
CCMPUTEO EVAPORATION LOSSES FffCM FILTERS
ESTIM*TEH OVERALL  BAL*NHr
        RF/COVFKY OF  LICUIDS
                                                    -5.1
                                                    99.?
                                                 18893.0
                                                    IS. 3
                                                  1818.0
                                                 25057.3
                                                 17723.0
                                                   680.8
,0
,0
.0
                                                   120
                                                     3
                                                 2«»57J».fi
                                                       6
                                                     0.0
        COfll SULFUR BALANCr
SULFUR IN COAL INPUT                          2°.5
SULFUR TN PROCESSED COAL                       7.°
SULFUR IN SAMPLES TAKEN DURING REACTION        5.1
SULFUR OISSCLVFC (AS SOU»  IN LF«CH OEAGENT    11.5
SULFUR RECOVERED FRCM ORGANIC SOLVENT          7.0
BALANCE (RECOVERED-INPUT!                       2.2
PERCFNT RECOVERY OF SULFUR                   157.<«

        IRON BALANCE
IRON IN COAL                                  2«».7
IRON IN LFACH REAGENT INPUT                  927.6
TOTAL IPON INPUT                             952.<*
IRON IN L^ACH REAGENT ORAMN FRO^ REACTOR     73*».S
IRON IN SAMPLES ORAfcN DURING REACTION         37.7
IRON IN INTERMEDIATE FILTRATES REMOVED         0.0
IRON IN FINAL FILTRATE AND WASHFS            i7°.9
IRON IN PROCESSED COAL                         7.3
TOT4L IRON RECOVERED                         959.9
BALANCE (RECOVERED-INPUT*                       7.5
PERCENT RECOVERY OF IRON                     100.8
                                                    98.1
      Weights are  In grams

-------
correction is made by  linearly  adding  to  or  subtracting from the coal being
processed any ash in excess or  in deficiency from that calculated by sub-
tracting the pyritic ash removed from  the ash in the starting coal; the
quantity to be added or subtracted  is  determined from ash analyses of the
starting and processed coals.)  The last  line in the table shows the
processed coal composition data obtained  from direct coal analyses and
the percent pyrite removal based on pyrite analyses of starting and pro-
cessed coal.

     The data  in  Table 7  show that  the agreement in pyrite removal values
determined from ferrous  ion  production and from direct coal analyses was
very good.  The discrepancy  in the  corresponding total sulfur content
values is believed to  be  due to reagent iron sulfate left on the coal be-
cause of insufficient  washing.   The higher than expected ash content of
the processed  coal and the  2.2 grams of excess  recovered sulfur  (see "sul-
fur balance" on second page  of Table 7) supports the above assumption and
indicates that the abnormally high  organic sulfur value of the analyzed
processed coal must  be due  at least in part  to  unwashed sulfate.  The
obtained process  mass  balance was,  also,  very good, especially if our
assumption  concerning sulfate deposition is  valid.  Based on recovered
elemental sulfur, the  Ss/$n  = 1.7;  or, conversely,  if  the true Ss/Sn is
1.5, 90 percent of the product elemental  sulfur was recovered.   Thus, in
every aspect the  data  was similar to that generated with the lower ash
Lower Kittanning  coal  used  in the previous  bench-scale program.   In  fact,
the data in Table 7  is nearly identical point-by-point with that generated
from  Experiment 50 of the previous  program.

     The rate  data from  Table 7 were utilized in  Equation  (3)  to determine
the value of K, at 102°C  for this coal.  The computed  K. was equal to 0.12
        li
(hours)   (W_)    which corresponds  with the  lower end  of the  KL  range
determined in  the previous  program.  It was, therefore, decided  to use the
following A.  and  E,  values  in Expression  (4) to predict coal processing  in
the mixer:
                                     49

-------
      100 mesh x 0 coal:  AL= 3.4 x 105 (hours)"1 (Wp)
      14 mesh x 0 coal  :  AL= 2.7 x 105 (hours)"1 (Wp)-1
      Both  top sizes    :  EL= 11.1 kilocalories per mole

      Table 8  summarizes  experimentally derived and predicted data for
 "mixer" processing of  100 mesh and 14 mesh top size Lower Kittanning coal.
 The data from 15 experiments of each top size coal are presented.  The
 agreement  between calculated Y, W  , and extent of pyrite removal values
 at t  and  those predicted by Expressions 3 and 4 is very good; the rare
 exceptions are apparently due to errors in iron-forms analyses,  it should
 be noted that the experiments in Table 8 were performed with a number of
 different  starting slurry and reagent compositions (see data in Appendices
 A and B, Volume 2) and for differing reaction times, tm, and temperature-
 time history.  The data  in Table 8 justify the use of Expressions (3) and
 (4) and the quoted.values of AL and EL for the design of the "mixer" for
 processing Lower  Kittanning coals  up to 14 mesh top size.

      A complete set  of mixer processing data for all experiments in Ap-
 pendices A and B is  included in Appendix D, Volume 2 together with
 predicted  data on L-R  processing.

 2.4.2 L-R Unit Operation

      The L-R unit operation combines the two major Meyers Process opera-
 tions, coal pyrite leaching and reagent regeneration, into a single unit.
 Its  experimental investigation constituted the major thrust of this pro-
 gram.

      L-R processing  (simultaneous  coal  Teaching-reagent regeneration)
starts when oxygen is  introduced to the slurry at the conclusion of  mixer
processing.  During  this program, L-R processing was performed  at  constant
temperature and pressure; however, three values of each of these parameters

                                    50

-------
  TABLE 8.   MEASURED AND PREDICTED  PYRITE  REMOVAL  AT  THE  "MIXER  UNIT
             OPERATION"
Exp.
No.
Starting
' Y
wp
Measured and Predicted Values at t = tm
1
Meas.
f
Pred.*
	 wp
Meas.**
Pred.*
% S Removal
Meas.**
Pred . *
A. 100 Mesh Top Size Coal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1.0
1.0
1.0
1.0
1.0
0.76
0.95
0.91
0.82
0.92
0.86
1.0
0.73
0.75
0.88
7.55
7.55
7.55
7.55
7.55
7.55
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
0.28
0.34
0.40
0.29
0.29
0.35
0.54
0.51
0.46
0.37
0.40
0.45
0.33
0.34
0.23
0.27
0.38
0.31
0.27
0.29
0.28
0.58
0.51
0.45
0.42
0.42
0.47
0.32
0.34
0.22
5.79
6.08
5.76
5.91
5.97
6.58
5.62
5.89
5.94
5.07
5.35
5.16
6.48
6.40
6.31
5.65
5.95
5.69
5.68
5.62
6.09
5.65
5.62
5.70
5.19
5.37
5.08
6.42
6.38
6.09
25
21
25
23
22
14
24
20
20
32
28
31
12
13
14
27
23
26
26
27
21
24
24
23
30
28
32
12
13
17
B. 14 Mesh Top Size Coal
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
1.0
0.87
0.94
0.95
0.94
0.71
0.80
0.67
0.62
0.84
0.84
0.45
0.77
0.52
0.70
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
7.27
0.59
0.49
0.56
0.55
0.57
0.33
0.36
0.30
0.30
0.41
0.39
0.26
0.37
0.29
0.43
0.56
0.47
0.53
0.53
0.55
0.28
0.31
0.26
0.28
0.35
0.31
0.25
0.33
0.26
0.32
5.55
5.69
5.65
5.59
5.75
6.53
6.35
6.49
6.63
6.41
6.35
6.87
6.46
6.81
6.71
5.44
5.58
5.63
5.50
5.57
6.37
6.23
6.40
6.56
6.24
6.16
6.85
6.34
6.70
6.45
25
23
24
25
22
11
14
11
9
13
14
6
12
7
8
27
25
24
26
25
13
15
13
10
15
16
6
14
8
12
   Predicted values  from  r,  =  K,  W.!;2M2) and K.  = A,  exp(-E/RT)
**                         L     u  r            L    L
   Calculated values from measured ferrous ion production
                                     51

-------
were used.  The majority of experimentation was performed at 120°C and
100 psig  (85 psi 02) with 100 mesh and 14 mesh top size Lower Kittanning
coal.  The major objective was the investigation of coal pyrite leaching
efficiency under L-R processing; thus, reaction time was the parameter
most thoroughly studied.  The generated data has been tabulated as a
function  of this parameter (Appendices A through D, Volume 2).  Slurry
composition (solids content, reagent composition, and reagent pH) was also
varied with each top size coal.  The data from approximately 50 experiments
were utilized to determine the efficiency of this unit operation and its
sensitivity to changes in process parameters.

     Pyrite leaching from coal during slurry preparation necessitated that
the L-R  unit operation be investigated in conjunction with the mixer
operation.  However, the data presented in the previous section (Section
2.4.1) demonstrated that process performance in the mixer could be ac-
curately  defined; thus, the L-R operation could be evaluated separately
even though the experimentation involved both unit operations.  In fact,
the predictability of mixer performance by the use of Equations (3) and (4)
suggested that the L-R operation performance should be also predictable
through the simultaneous use of the above equations and those for reagent
regeneration developed in the previous bench-scale program, namely
where
      rD     moles of ferric ion regenerated per unit time,
       K
      p      oxygen partial pressure in atmospheres,
       U2
      Fe     ferrous ion concentration in moles per liter,
      KR     rate constant, a function of temperature only,
             in liters/mole-atm-unit time
                                    52

-------
and
        KR =  AR exp (-ER/RT)                                     (6)
with   "
        AD =  6.7 x 105 Liters/Mole-Atm-Min
        - K
and
        ER =  13.2 Kcal/Mole.

     However, attempts to predict the S  (pyritic sulfur content of coal)
versus t,  D  (L-R reaction time) data contained in Appendices A and B,
        L-K
Volume 2, through the simultaneous solution of Equations (3) through (6)
failed.  In  general, the experimentally determined pyrite removal rates
were higher  than the predicted rates up to approximately 80 percent re-
                                                           .J.O
moval and lower thereafter; also, the measured Y values (Fe  -to-Fe ratio)
were lower than the predicted ...Y values.

     Comparison of the predicted and experimentally derived data led to the
following observations:

     •  The  pyrite leaching rate during L-R processing at constant
        temperature depended only on W  and Y.  The magnitude of
        the  rate dependence on these variables appeared to be the
        same as that determined for ambient pressure leaching of
        Lower Kittanning coals; thus, Equation (3) should have pre-
        dicted the experimental S  versus t data in all Meyers
        Process unit operations if correct K,  and Y values were
        used.

     t  The  K,  (leaching rate constant) values predicted through
        Equation (4) for L-R leaching were substantially lower
        than the experimental K,  values suggested by the pyrite
        removal data for up to nearly 80 percent removal (it was
        indicated earlier that the experimental rates were higher
        than the predicted rates even though the corresponding
        measured Ys were lower than the predicted).  This lack of

                                     53

-------
   predictability of the L-R KL from AL and EL values deter-
   mined under ambient pressure processing of coal  could  be
   suggesting that KL should not be treated as a  single
   constant, but as a sum of rate constants each  of which
   relates to a certain coal or pyrite particle size within a
   given coal top size; however, the generated data did not
   permit verification of this possibility.

•  The experimental data indicated that the L-R pyrite
   leaching rate was dropping faster than predicted by
   Equation (3) with a single valued KL (isothermal pro-
   cessing) when pyrite removal exceeded approximately 80
   percent.  This observation implied  a possible change  in
   reaction mechanism; however, the data indicated contin-
   ued leaching rate dependence on W   and  Y.  Again, a
   change in KL value was indicated which  because of the
   isothermal processing was not predictable  by Equation
   (4) (in principle, this  change in K, value could have
   been  predicted if K, was indeed a sum of rate constants
   of known values).

•  The low measured Ys in comparison to Ys predicted through
   the simultaneous solution of Equations  (3) through (6)
   and the large variances  in the deltas between measured
   and predicted Ys, especially during early  L-R processing
   when pyrite leaching rate was the  highest, led  to the
   conclusion that reagent  regeneration during many of the
   experiments was inefficient due  to  incomplete mixing
   (e.g., inappropriate Reynolds Number in the slurry cir-
   culation loop), although lack of adequate  oxygen supply
   could  not be ruled out for certain  experiments involving
   33 wt. percent coal slurries.  Equations (5) and (6)
   proved repeatedly valid  during separate reagent  regen-
   eration  experimentation performed in the parametric
   ranges that  L-R  processing was  performed.   There  is
                               54
no

-------
        reason to believe that the presence of coal could have had
        an inhibiting effect on regeneration rate unless it af-
        fected oxygen mixing (there was no evidence of direct
        reaction of oxygen with coal).

The above observations led to the following tentative conclusions con-
cerning pyrite leaching data predictability by Equations (3) through (6):

     t  Pyrite removal versus leaching time data generated from
        suspendable L.K. coal in the Meyers Process unit oper-
        ations investigated in this program (Mixer, L-R, and
        Settler Operations) should be predictable by Equations
        (3) through (6).  The only input data required are:
        starting coal pyrite concentration, starting reagent iron
        and ferrous ion concentrations, the oxygen partial pres-
        sure during L-R processing, the temperature-time curve
        of the process, and the values of the Arrhenius constants
        A and E for Equations (4) and (6).  If the L-R processing
        behavior of the coal used in this program is typical of
        L.K. coals, then more than one pair of A and E values
        will be needed for Equation (4) for each coal top size.
        This is tantamount to saying that Equation (3) should
        be most properly expressed as
                             dW
                                                                 (7)

                                                                 (8)
        with each K^ depending on a different pair of A and E
        values.  The K^ and W ^ values are specific either to
        narrow coal or pyrite particle sizes within a  given top
        size coal or to elements of coal (shells) whose thick-
        ness is the same for all particle sizes and depends only
        on reagent accessibility rates (distance from coal sur-
        face and coal porosity).

                                    55

-------
     •  The L-R data generated  in this program could not be
        predicted by Equations  (3) through  (6) because of the
        change in the  KL value  during L-R processing and be-
        cause of inconsistent regeneration  during the early
        stages of L-R  processing.  However,  if the conclusions
        reached in the previous paragraph are valid, the
        generated L-R  data  should be predictable by Equation
                                                                •
        (3) through the use of  experimentally determined K^
        values from L-R data and through the use of measured
        Y  values.

      In order to test  the validity of the above conclusions, the L-R data
 in  Appendices A and B, Volume 2 of this report were treated as  if derived
 from two separate L-R  operations (L-R Sections) describable by  Equation (3)
 but with different reaction rate constants  (KL values).  The first "L-R
 Section" involved pyrite leaching rate data  up to approximately 80 percent
 overall pyrite removal (pyrite  removal based on starting coal pyrite con-
 tent); the second "L-R Section" involved the remaining L-R rate data
 (pyrite removal beyond 80 percent) which apparently required a  lower K,
 value in order to be predictable by Equation (3).  A K, value was determined
 for each of the two L-R Sections which  was   then used in conjunction with
 measured Y values in Equation (3) to test the predictability of the data in
 Appendices A and B.  The KL dependence on W during its transition from the
 high to the low value  was also  determined from the experimental data.  The
 detailed procedures used and the results of the L-R rate development in-
 vestigations are presented  in the next section of this report.

 2.4.2.1  Pyrite Leaching Rates  During L-R Processing of L.K. Coal at 120eC

     On the assumption that  Equation (3) would prove to be the  correct
empirical   rate expression by which pyrite removal from suspendable L.K.
coal during L-R processing could be predicted, its integrated form was
used to determine the  KL values  applicable  to 120°C L-R processing of the
two top size coals (14 and 100 mesh).  Equation (9) is the integrated form
of Equation (3):

                                    56

-------
                                        WP
where
     K,     is the L-R rate constant for the particular coal top
           size applicable up to nearly 80 percent pyrite re-
                                          i      -i          '
           moval and expressed in (hours)"  (W )" ,

     t.  D  is the L-R reaction time in hours,
      L-K
     7     is the average Y value of the reagent during t. _D»
           dimensionless quantity,

     W     is the concentration of pyrite in coal at t,  D, in
      p                                               L-K
           wt.  percent ,

     w"J    is the concentration of pyrite in coal at tL_R • 0
           (end of Mixer processing and start of L-R), in  wt.
           percent.

     The Y values were determined from measured Y versus t, _n curves plot-
ted from data in Appendices A and B, Volume 2.  The W  values were computed
from the S  values (pyritic sulfur concentration in coal) in Appendices A
and B (W  =  (120/64.1) S ) and the w"J values were obtained from Table 8,
Section 2.4.1.
     Over 20 data points derived from eight different experiments were
used in Equation  (9)  in order to determine the values of  K, for  each
of the two top size coals and to check their constancy with changes in
processing parameters other than temperature and coal particle size.  These
computations revealed that the most probable K,  value for L-R processing of
                                              '-1-1
the 100 mesh top size coal at 120°C was 0.6 (hours)   (W  )   and the cor-
responding KL value for the 14 mesh top size coal was 0.5  (hours)"  (W  )"  .
With the exception of occasional/outliers, the deviation  in the  K, values
was less than + 0.1.  These computations also revealed that the  use of  S
values corrected for anomalies in the organic sulfur content of  the coal
(Equation 2, page 28) tended to minimize the deviation in  the computed  K,
values.                              57
                                                                         1

-------
      The estimated KL values for 120°C L-R processing of 100 mesh and  14
 mesh top size L.K. coals are approximately 5 times larger than K. values
 determined for pyrite leaching from the corresponding  coal  top sizes at 102°C
 (Section 2.4.1).  This unexpectedly high factor of five difference between
 the 120°C and the 102°C KL values  (Equation 4 predicts only a factor  of 2
 difference) was further verified from data plots similar to that illus-
 trated in Figure 5 for 100 mesh top size coal.

      In Figure 5, the top curve presents the S  removal versus L-R reaction
 time data generated in Experiments 7, 8, and 9; these experiments differ
 only in overall L-R reaction time  used (see Appendix A, Volume 2  for de-
 tails on experimental conditions).  The second curve presents the same
 type of data from Experiment 10 which because of insufficient supply of
 oxygen or inefficient mixing of oxygen into the slurry  was  performed with
 reagent whose Y remained low for a large part of L-R processing.   It is
 noted that the lower Y values are reflected in  lower S   removals, as Equa-
 tion (3) predicts should happen.   "Corrected" S  values were used and
 the present S  removal was computed from

                                   i-  (s_/s";)
                   % Sn Removal  =  ,—//os /fi/i  i \    *  TOO.       (10)
                      P            I "*  \ 1 ȣv
      The lower two curves in Figure 5 represent the expected S  versus t
 (coal  leaching time) data from Experiments 7 through 10 if they were per-
 formed at 102°C (ambient pressure) with Y versus Sp traces being  identical
 at  102°C and 120°C for the corresponding experiments.   These curves were
 generated by using a rearranged form of Equation (9),  namely

                   t =(_!	\  t   64.1	U)               (11)
                   t  % Y2 J  (  120 Sn    jn ;               v   ;
                        KLY             p    wp

     The required  leaching time, t, was computed for attaining each of the
S   values indicated in  the top two curves; curve rather than actual Sp
values were  used which  is equivalent to using averaged data.

*Fnuat1on (10)  Is  a more exact form of Equation  1  (page 28) because It ac-
 Sunts  for  the reduction in coal  weight  during  reaction (pyrite dissolution);
 iS derivltion, based  on pyrite mass  balance, is  given in Appendix E, Volume 2.

-------
en
                                                                              EXP. 7,8,9 (1Q2°C, AMBIENT
                                                                                              PRESSURE)
                                                                                 EXP. 10  (102°C AMBIENT
                                                                                               PRESSURE)
                                                                O  EXP. No. 7
                                                                   EXP. No. 8
                                                                   EXP. No. 9
                                                                   EXP. No. 10
120°C,  100 PSIG
                                              345
                                                LEACHING TIME, HOURS
                         Figure 5.  100 Mesh Top  Size Lower Kittanning Coal Leached
                                   with 5 Wt,%  Fe  Reagent (20% Slurries)
                  8

-------
The Y  values
                      r	„„ K, CYiuuaijr expiainea from the measured re-
agent  Y values during the performance of  Experiments 7 through 10 at 120«C;
a  Y was computed for each level of Sp removal.  The K,  value used was that
determined in Section 2.4.1  for 102°C pyrite leaching from 100 mesh top
size  L.K. coal  (K,_ = 0.12 hours"1 W^1).  v/J, which is the starting L-R
pyrite concentration in  the  coal  used  in  each of  the above experiments
 (Table 8, Section 2.4.1), was assumed  to  be the starting  pyrite  concentra-
tion  of the corresponding coals  processed at 102°C.

      The data  in  Figure  5 show that the leaching  time  required to obtain a
 given   Sp  removal at  102°C  (ambient pressure processing) is five times
 greater than that required for the  same S  removal at  120°C (L-R processing)
 provided the leaching  took place  under  identical  Y values.  Thus, the pre-
 viously estimated five fold  increase in K.  value  between 102°C and 120°C
 is  further validated.   In addition, the  data indicate that  KL  remains
constant up to  at least  75-80 percent S   removal and  that  the rate of S
removal depends on the Y value of the reagent during  L-R processing  just
 as was proven  to be the  case with ambient pressure processing (Section
2.4.1).  Identical observations were made from similar  treatment  of  14 mesh
coal  top size data.   It was,  therefore, concluded  that  the KL values deter-
mined  in this Section for the L-R processing of the two top  size  coals  at
120°C  are valid for  up to approximately 80 percent S   removal.  Thus, their
use in Equation (3),  in conjunction  with experimentally derived Y versus
t    data,  should  lead to the prediction of the  120°C Sp versus tL_R data
in~Appendices  A and  B  of  Volume 2 regardless of  the experimental  conditions
used to generate these data  (except  for temperature), provided  that  Equa-
tion  (3)  is  a  valid  rate  expression  for L-R processing  of  suspendable L.K.
coal as it  has  been determined to  be for separate  leaching-regeneration.

     Equation  (3) was  subjected to the  validity  test  and the results are
summarized  in Tables 9  (100 mesh  top size  coal)  and 10  (14 mesh top  size
   •D   The  test proved conclusively that  the rate of pyrite removal from
suspendable  Lower Kittanning  coal during L-R processing is predictable by
Equation  (3) up to at  least 80 percent Sp  removal.
                                     60

-------
TABLE 9.  ANALYZED AND PREDICTED PYRITIC SULFUR CONTENT OF 100 MESH  TOP  SIZE
          L.K. COAL AS A FUNCTION OF L-R PROCESSING TIME AT 120°C
Exp.
No
A. 3 Wt.%
1
2
3
4
5
6
B. 5 Wt.%
7
8
9
10
11
12
C. 5 Wt.%
13
14
15
16
D. 5 Wt.%
17
VR-
sulfur

-------
                    TABLE 10.    ANALYZED AND PREDICTED PYRITIC SULFUR CONTENT  OF 14 MESH  TOP SIZE  L.K.

                                  COAL AS A FUNCTION OF L-R PROCESSING  TIME AT 120°C (5 WT% Fe REAGENT)
Exp.
No.
A. 20 Wt.'
B.
C.
D.

23
24
25
26
27
33 Wt.
28
29
30
31
32
33
33 Wt.
34
35
33 Wt.'
36
t, n = 0.5 Hours
L~K
Y < (VP
tL_R = 1.0 Hour
T SP* 
-------
     Tables 9 and 10 summarize all of the $„ and Y versus t, D data gener-
                                           P               L~K
ated during the first 3 hours of L-R processing of L.K. coal with iron
containing starting reagent at 120°C.  The data are divided into five L-R
reaction time intervals UL_R) with the end of each time interval repre-
senting a slurry sampling point.  The average reagent Y and the experimental
and predicted S   values are listed for each tL_R.  Also, the data are
grouped with respect to reagent iron concentration, slurry coal content,
and oxygen partial pressure.  Each 7 value represents the average reagent
Y  during the indicated interval and it was computed from plots of the Y
versus t, R data listed in the tables of Appendices A and B.  The S* values
represent the pyritic sulfur concentration in the coal at the end of each
reaction interval; they were derived from sulfur forms analyses of the coal
in slurry samples or from analyses on the product coal (the pyritic sulfur
value was adjusted for anomalies in the organic sulfur content of the coal
sample).  The (Sp)  values represent the predicted pyritic sulfur content
of the coal at the end of each t. _R as computed from
                                               P
 Equation  (12)  is an  integrated form of Equation (3); all of its terms have
 been defined previously.  The values of KL> Y, and tL_R are listed in
 Tables 9  and 10; the values of W? for each experiment are listed in Table
 8  (Section 2.4.1).

     Comparison of analyzed and predicted S   values in Tables 9 and 10 re-
 veals good agreement in these majority of cases; in fact, when normal  sam-
 pling and analyses uncertainties are taken into consideration the agreement
 can be considered excellent for S  removals up to nearly 80 percent (resid-
 ual S  0.8 wt. percent).  It should be noted that these experimental data
were generated under a variety of processing conditions involving starting
reagent iron and sulfate concentrations, starting reagent Y, slurry coal
content, slurry pH,  total pressure, and oxygen partial  pressure.
     The large majority of the Sj$ values  1n Tables 9 and 10 were  derived

                                     63

-------
from analyses of coal present in reactor samples drawn during processing.
The S* values for which a deviation is indicated represent analyses per-
formed on samples taken at the end of the run where the entire batch of
 L-R processed  coal  was  sampled;  the  deviation was  derived  either  from
duplicate analyses  on the same sample or from analyses of duplicate samples.
S*  values which appeared to be grossly inconsistent with respect to the
rest of the data derived from a particular experiment or with respect to
data from similar experiments are placed in parentheses in the above tables.
Discrepancies between analyzed and predicted S  are not limited to the
values in parentheses;  in certain cases, entire experiments appear not to
obey Equation 3  (e.g.,  Experiments 2 and 11), but  the reasons  for the dis-
crepancy  are  not readily  apparent.

     The majority of the  discrepancies between analyzed and predicted S
appear in the first group of experiments in Table 9 (Experiments  1 through
6).  In these experiments coal was processed with iron sulfate solutions
which did not contain added sulfuric acid.  Apparently, without the added
acid the pH during  L-R  processing was not sufficiently low to inhibit iron
sulfate deposition  on the coal (see data in Appendix A, Table A-l).  It
is  conceivable that the deposited iron sulfate complicated the sulfur forms
analyses performed  on these coal samples and led to errors in pyritic sul-
fur determinations  (10  to 20 percent error in pyritic sulfur analyses can
account for most of the observed discrepancies).  Another reason that the
discrepancies between analyzed and predicted S  are concentrated  in this
group of experiments may be inadequate sampling.  Experiments 1  through 6
represent early experimentation with L-R processing and it could be that
sampling techniques were not as good as later in the program.    It
should be noted, however, that there is nothing in this group of data, or
in  any of the other groups of data in Tables 9 and 10, which relates dis-
crepancies between  analyzed and predicted S  values to the value of a
certain processing  parameter (at least within the  indicated ranges of the
processing parameters).

     Consistent discrepancies between predicted and analyzed S  values
appear when the pyritic sulfur concentration in coal has been reduced to

                                    64

-------
approximately 0.8 weight percent (nearly 80 percent pyrite removal).  Appar-
ently, further S  reduction during L-R processing takes place at rates
which are appreciably lower than those predicted by Equation (3) and the
KL values in Tables 9 and 10.  This discrepancy between S* and (S )  be-
comes more apparent when predicted S  values for three hours L-R processing
(Tables 9 and 10) are compared to S  values derived from analyses of coal
processed in excess of three hours (Experiments 2, 4, 5, 6, 11, 12, 16, 24
through 27, 32, and 33 in Appendices A and B, Volume 2).  It is also
illustrated by the 14 mesh top size data in Figure 6 (120°C rate data from
Appendix B).

    In Figure 6 the reciprocal pyritic sulfur concentration in coal (coal
analysis values corrected for organic sulfur anomalies) is plotted against
L-R reaction time which was normalized for inconsistencies in reagent
regeneration.  According to Equation (12), the slope of the data in Figure
6 should be constant and proportional to the KL value of Equation (3) for
120°C processing of 14 mesh top size coal.  Indeed, the data illustrates
that this is the case up to the 1/S  value of approximately 1.2 (S  of 0.83
wt. percent and W  of 1.56 wt. percent), but a definite change in K,  value
occurs thereafter.  Actually, a change in reaction mechanism is indicated
by the data which Equation (3) did not predict (speculation on possible
causes of data unpredictability in this S  region and potential modifi-
cations of Equation (3) which could possibly rectify this discrepancy were
presented in the previous Section).  However, in the absence of a more
precise rate expression, Equation (3) can be utilized to predict L-R pro-
cessing even beyond 80 percent pyrite removal by allowing for a change in
the isothermal value of K, .

    The data in Figure 6 indicate that L?R processing can best be represented
by Equation (3) through the use of three K,  values.  For the 14 mesh top
size L.K. coal processed at 120°C the indicated K.  values are as follows:

    KU  = 0.5 (hours)"1 (% pyrite in coal)"1 for W  >.1.6 wt.%
    KL2 = f(V " Wp " 1<1 (nours)"  (% Pyrite)   for I-6 lWp 1 1-2 wt.%
    KL3 = 0.1  (hours)"1 (% pyrite in coal)"1 for VI  < 1.2 wt.35
                                     65

-------
o
V
z
I
      1.0
      1.2
       1.5
               2.2
               2.0
O   1-8
U
Z
"   1-6
u

R   1.4
            u
            OS
=J   1.2
CO
y
fc   1.0
Q:
            5   °-8
            O
          - Q£

            G   °'6
            LU
    0.4


    0.2


     0
                                   KL  =f(wp)
                                                                         PREDICTED CURVES FROM
                                                                         EQUATION (3) AND INDICATED KL
                                                                         EXPERIMENTAL DATA
                                           ty , NORMALIZED REACTION TIME IN HOURS

                                          Figure 6.  Pyrite  Leaching Rate Constant Data

-------
The deviation in the KL-J  is estimated  to, be +  10  percent,  but  the  deviation
in the KL3 value is closer to +. 50  percent.  However,  the  large  deviation
in KL3 is understandable  since  it can  result from only a ±6.1 deviation
in the S  values which comprise the concentration region to which  K. 3
applies; such deviation is normal for  pyritic  sulfur analyses.

      It is interesting to note  that the  estimated K-3  value is identical to
the KL value found applicable for ambient pressure processing  of the same
size  coal at 102°C.  Within experimental  uncertainty,  the  same observation
was made with the 100 mesh top  size coal  where K, - was estimated to  be 0.12
        i       i *
(hours)"   (W )~ .   The implication of this observation is that  if KL3 is
temperature insensitive during  small temperature  changes  (e.g.,  a  diffusion
constant)  it is possible  to have an observable change  in reaction  mechanism
at  120°C but not at  102°C, provided that at 102°C KLI  and  KLS  are  nearly
identical.  This may be the case with  suspendable L.K. coal since  the L.K.
coal  used  in the previous bench-scale  program  indicated no change  in the
pyrite  leaching rate mechanism  up to 95  percent pyrite removal at  102°C
 (see  also  discussion in Section 2.4.1).

     It is  believed that the  preceding discussion proves conclusively that
Equation (3)  can be used to predict the Meyers Process performance during
the L-R processing  of suspendable L.K.  coal at 120°C provided  the experi-
mentally derived KL values are used and the Y versus tL_R history of the
process is  known.  The  latter requirement limits the utility of Equation
(3)  severely unless the change in Y during L-R processing can  be predicted
through the use of  a valid reagent regeneration rate expression.   In the
next Section  an attempt is made to prove that Equation  (5) is the appro-
priate reagent  regeneration rate expression for L-R processing when oxygen
supply to  the  slurry is adequate and proper mixing has occurred.
 The KIT  for 100 mesh top size coal was determined earlier to be 0.6 (hours)"
         the KL£ was estimated to be the same as that of the 14 mesh top size
 coal.
                                    67

-------
2.4.2.2  Reagent Regeneration During L-R Processing

     Reagent regeneration involves the oxidation of the Fe+2 generated
                                         +3
during the pyrite leaching reaction to Fe   by oxygen.  During L-R pro-
cessing the two operations (pyrite leaching-reagent regeneration) take
place simultaneously and their rates are related through the iron forms in
                                                     +2
the reagent.  The regeneration rate depends on the Fe   concentration in
                                                    +3
the reagent solution and the leaching rate on the Fe  -to-Fe ratio, which
                           +3                  +2
is defined as Y.*   Since Fe   is equal to Fe-Fe   , Y  is also a measure of
regeneration rate.
                                                               •
     Review of the  Y versus t, R data in Appendices A and B, Volume 2 re-
veals that in most  experiments the rise in Y was rapid during the early
stages of L-R processing.  This observation during the performance of the
experiments led to  the  erroneous conclusion that reagent regeneration was
taking place efficiently during L-R processing.  When valid pyrite leaching
rate constants were determined, it became possible to generate predicted Y
versus t, R data which  when compared to the experimental data led to the
following observations:
     •   In the majority of experiments reagent regeneration was
         less efficient  than predicted during the first hour of
         L-R processing.
     •   Reagent regeneration was inconsistent in a number of
         experiments.
     •   Equations  (5) and  (6) appeared to be valid rate ex-
         pressions  for spent reagent regeneration, but the
         activation  energy value computed from the regeneration
         of synthetic spent reagents for use in Equation  (6)  is
         probably higher than the activation energy applicable
         to the regeneration of actual spent reagent in the
         presence of coal.
     •   The reagent regeneration rate was probably catalyzed  by
         cations leached from the coal  or by the coal  itself.
*                       +2
 Note that because of Fe   production  in  the  mixer the  starting L-R Y was
 always low; therefore, high  regeneration rates  were expected during the
 early stages of L-R processing.
                                     68

-------
 The  validity of the  above observations becomes evident from the data pre-.
 sented  in  Figures  7  through 9 where predicted Y versus t,  D data (solid
                                                         L~K
 curves)  are compared to measured Y values as a function of L-R processing
 time (dashed curves).   The predicted data in these figures was abstracted
 from Table D-l, Appendix D, Volume 2 of this report.

      Table D-l  lists predicted data for both the Mixer and L-R Operations
 which correspond to  the experimental  data from every  experiment in Appen-
 dices A and B of Volume 2.  The data for each experiment are presented in
 two  sections.  The top section lists Mixer Operation  data which were dis-
 cussed  in  detail in  Section 2.4.1.  The bottom section lists the predicted
 Wn (pyrite concentration in coal), Y, and x,  D (fraction of pyrite removed
  p                                         L-K
 during  L-R) as a function of t,  „ (L-R reaction time); it also lists the
 overall  pyrite removal, x, . , at each of the tL_R values shown  (over-
 all  pyrite removal is based on starting coal  W  and it includes pyrite
 removal  in both the  Mixer and L-R Operations).  The predicted L-R data were
 generated  from known starting coal and reagent compositions and Equations
 (3), (5),  and (6).  The K, values used in Equation (3) were those deter-
 mined as applicable  to L-R data up to 80 percent pyrite removal.  The KR
 values  (regeneration rate constants) were computed from the Arrhenius
 constants, AR and  ER, through the use of Equation (6); these constants were
 determined from synthetic spent reagent solutions.

      Figure 7 illustrates measured and predicted Y values derived from L-R
 processing at 120°C  of 20 wt. percent slurries of 100 mesh top size L.K.
 coal.  The solid curve depicts the expected Y variation with L-R reaction
 time from  Experiments 8 through 12 (predicted data) and the dashed curves
.show the corresponding experimentally derived Y data.  If small differences
 in starting Y values and Mixer reaction times are disregarded, the only
 meaningful difference among these experiments was L-R reaction time.  Thus,
 at any  given reaction time the slurry composition, and therefore Y, should
 be the  same for all  the experiments in this group (a  small variation could
 be explained on the  basis of the observed differences in starting L-R Y
 values).   The predicted data1 curve confirms that these experiments should
 have yielded identical Y versus tL_R data if performed reproducibly.  The

                                      69

-------
    100 MESH TOP SIZE COAL, 20 WT. % SLURRIES, 5 WT. % Fe REAGENT
                                       PREDICTED DATA


                                          EXPERIMENTAL
                    1.0         1.5         2.0

                     L-R REACTION TIME, HOURS

Figure 7.   Predicted and Experimental Y  Values During L-R Processing
           of L.K. Coal at 120°C

                              70

-------
      14 MESH TOP SIZE COAL, 33 WT % SLURRY. .5 WT % Fe REAGENT
                                      PREDICTED DATA EXP. 37 & 38
                                      EXPERIMENTAL DATA
                                     EXP. No. 37

                                     EXP. No. 38
1.0        1.5        2.0


   L-R REACTION TIME, HOURS
                                                    2.5
3.0
Figure 8.  Predicted and Experimental Y Values During L-R Processing
          of L.K.  Coal at 110°C
                             71

-------
    14 MESH TOP SIZE COAL, 33  WT % SLURRY, 5 WT % Fe REAGENT
                                     PREDICTED DATA EXP. 39 & 40
                                     EXPERIMENTAL DATA

                                      EXP. No. 39

                                     EXP. No. 40
                  1.0        1.5         2.0

                     L-R REACTION TIME,  HOURS
Figure 9-  Predicted  and Experimental  Y Values During  L-R Processing
          of L.K.  Coal at 130°C
                             72

-------
 experimental data show that Experiments 8 and 9 yielded reproducible data
 which, however, differed substantially from the data generated in Experi-
 ment 10 or in  Experiments  11 and  12.  The lack of consistency  in Y data
 reproducibility could have only resulted from inconsistent oxygen supply
 to the slurry (insufficient feed or inefficient mixing) since there was no
 evidence of errors in iron mass balance.   (Note that the Y value depends
           +2
 only on Fe   and oxygen concentration when pyrite leaching is performed at
 constant temperature; starting Y has a negligible effect as indicated  by
 the predicted curve in Figure 7).

     As would be expected from  inefficient regeneration, the data in Figure
 7  reveal that oxygen  starvation was more pronounced during  the early stages
                         +2
 of L-R processing when  Fe   generation was high;  however, the data also
 reveal that  oxygen  starvation need not have taken place  if  experimentation
 was  performed in accordance with the  procedures used  in  Experiment 10.*
 In Experiment 10 the  oxygen supply to the slurry  appears to have been ade-
 quate  to permit reagent regeneration to  take  place under the  kinetically
 controlled rate.   This conclusion  is based on the observation that the
 experimental Y  versus tL_R data from this  experiment traces a curve which
 is identical in shape to the  predicted curve  (the two curves will coincide
 if translated on the  time axis  by  a  constant  multiplier).   Conversely,  this
 observation leads  to  the conclusion  that  Equation (5) is a  valid  reagent
 regeneration rate expression  for use in  the  prediction of  L-R processing
 data.   The fact that  the predicted and measured reaction times  for the
 same Y value differs  by a constant factor  indicates  that the  actual  KR
 value  for this  temperature is higher than  that used  in Equation  (5)  to
 generate the predicted data curve.  Since  Equation  (6) is  also  probably
 valid  on the basis of the above data comparisons, a  wrong  KR  value  implies
 that at  least one of  the Arrhenius constants  used to predict  the  L-R data
 was wrong.

     Equations  (5) and  (6) were derived from  a substantial  bank of data  on
the regeneration of synthetic spent reagent solutions  (previous bench-scale
*
 Oxygen feed procedures were not intentionally varied, but precise control
 of slurry circulation fates was difficult in the experimental set-up used
 in this program, principally because of slurry pump size.
                                     73

-------
program).   It  is  highly unlikely that  the Arrhenius constants deterged
from these  experiments were .wrong.   It was, therefore, assumed that the
apparent higher KR values observed  here under  L-R processing were due to
catalytic effects emanating from some  unknown  substance leached  from the
coal (probably metallic cations and very likely copper).*  The catalytic
effect would be expected to manifest itself through the lowering of the
activation  energy of the reaction,  higher KR values, and different, most
likely lower,  KR  dependence on temperature.   In fact, if the AR value was
not affected by the catalyst, it would be expected  that the catalytic
effect on KR to be proportionally greater at the lower of two reaction
temperatures.  The data in Figures  7,  8, and 9 tend to show that this was
actually the case.  Until further verified, this catalytic effect should
be considered  as  an assumption.

     In Figure 8, predicted data are compared to experimental L-R data
generated at 110°C.  In Figure 9, the  same type of data are  compared from
130°C experimentation.  In both cases  experimentation involved the L-R pro-
cessing of  14  mesh top size L.K. coal  as 33 wt. percent slurries in 5 wt.
percent iron reagent.   Very good reproducibility of data  is  exhibited by
each pair of experiments depicted in these two Figures.   However,  it appears
that the regenerationYate did not  reach its kinetic limits  during at least
the first hour of L-R processing at 130°C probably due to  insufficiency in
oxygen.  Thus, the 130°C experimental  curves in Figure 9  are not expected
to obey Equation  (5) and, therefore, they are not strictly comparable to
the data from  Experiment 10 (Figure 7) and Experiments 37  and 38  (Figure 8).
Yet it can  still  be argued that the delta between "catalyzed" and  "un-
catalyzed"  KR  values is smaller at  the lower temperatures  (this  is espe-
cially true if it is assumed the regeneration  rate at t^_R = 1.5 hours is
no longer limited by oxygen).   Of course, the  effect of temperature on the
difference  in  predicted and experimental  (suggested from measured  Y)  KR
values is readily apparent from comparison of  the data in  Figures  7 and 8.

     Predicted and experimental data were compared  for all  the  experiments
in Appendices  A and B.   (The experimental  Ys are listed in  Appendices  A and
B and the corresponding predicted Ys in Appendix D).  These comparisons
 It is possible,  but not very probable, that the catalyst was leached  from
 the reactor,  since different type  stainless steel  reactors  were used  during
 L-R and synthetic reagent experimentation.
                                     74

-------
strengthened further the conclusions drawn from the data in Figures 7
through 9.   However, the number of experiments during which regeneration
was not limited by oxygen insufficiency was small and, therefore, no
attempt was made to estimate either the activation energy of the assumed
catalyzed reaction or the KR value at any given L-R temperature.  Attempts
to discern any variation in the KR value as a function of reagent recycle
time (due to accumulation of catalyst) were inconclusive for the same
reason.

     The above analysis on the reagent regeneration data derived from L-R
processing furnishes the reasons for (a) the scatter in the pyrite removal
versus leaching time data, (b) the difficulty to discern any sensitivity
in pyrite removal rates to changes in processing parameters, and (c) the
requirement that experimental and not predicted Y values had to be used
in Equation (3) in order to predict the generated S_ versus t.  D data.
                                                   p         L—K
Thus, it became clear that parametric effects on pyrite leaching must be
examined on the basis of predicted data rather than experimental data in
order that the real effect of each parameter becomes apparent.   Since
Equation (3) was proven to be a valid leaching rate expression under
virtually all  processing conditions examined when experimentally determined
KL values were used with it  (Section 2.4.2.1) and since Equations (5) and
(6) appeared to be valid on  the basis of the data analysis  performed  in  this
section, the predicted data  listed in Table D-l, Appendix D, Volume 2 were
used wherever appropriate in the examination of parametric effects.   It
should be noted, however, that the data in Table D-l are only valid up to
80 percent pyrite removal because they were generated with KL values which
were proven valid only up to this level of pyrite removal.  Also, these
data were generated using AR and ER values derived from uncatalyzed reagent
regeneration reactions; thus, the predicted pyrite removal values listed  in
Table D-l are lower than the expected pyrite removal values from L-R  pro-
cessing under efficient reagent regeneration.  However, at a given L-R
processing temperature these values are uniformally lower and should  not
affect data comparison.

     The higher (uncatalyzed) ER value was also used in process engineering
analysis because the lower ("catalyzed") ER value could not be  determined
                                     75

-------
with adequate precision from the available data and because it is not
known if similar ER depression would occur during the L-R processing of
other coal, even L.K, coals from different mine than the one used in the
current program.  The use of the higher ER value should not substanially
affect the size of the reactors or the computed process costs, except
perhaps when L-R operation is performed at low temperatures and pressures
(e.g., 100°C and 50 psig).

2.4.2.3  Temperature Effects on L-R Processing of L.K. Coals

      Figures 10 and  11 illustrate the temperature effect on pyrite removal
during the L-R processing of 100 and 14 mesh top size L.K. coals in the
temperature range of 110*C to 130°C. The temperature effect was  investigated
on 20 wt. percent slurries of 100 mesh coal and on 33 wt. percent slurries
of 14 mesh coal; thus, the data plotted in these figures differ  in two pro-
cessing variables.   However, the temperature effect should be identical
for the two top size coals and, therefore, the observed difference in
the magnitude of the temperature effect in Figures 10 and 11 is entirely due
to the difference in slurry concentrations.*

      The data in Figures 10 and 11 show that the rate of pyrite  removal
from suspendable L.K. coals during L-R processing increases with increasing
temperature.  In addition, the temperature effect on pyrite removal  is
larger between 110° and 120°C than between 120° and 130°C regardless of
slurry composition and it is more pronounced on 33 wt. percent slurrios
than on 20iWt. percent slurries.  The data also reveal that the  130°C
curves of the two slurry concentrations are nearly identical.  These ob-
servations lead to the conclusion that in terms of temperature effect alone,
the most efficient processing conditions, among those represented in Figures
10 and 11, are the conditions under which Experiment 40 was performed.   In
addition,  the data in these figures indicate that 33 wt. percent slurry
processing is preferable to 20 wt. percent slurry processing at  any  of the
three temperatures since twice as much coal per unit time  is processed at
*
 Pyrite removal rates from the two size fraction coals differ only  in
 value (Arrhenius frequency factor) which is temperature  independent.
                                     76

-------
   90
O
Z   70
CO
CO
LU
U
O
    60
O
Z
oi
<  50
£  40
D_
H-
z
    30
    20
   10
                                             130°C (EXP. 20)
                                                        120°C (EXP. 10)


                                                        110°C(EXP. 18)
                0.5
1.0         1.5         2.0

   L-R REACTION TIME,  HOURS
2.5
3.0
        Figure 10.  Temperature Effect on L-R Processing of 100 Mesh Top
                    Size L.K. Coal (20 Wt. Percent Slurries)
                                      77

-------
90
                         1.0        1.5         2.0

                          L-R REACTION TIME, HOURS
2.5
3.0
     Figure 11.  Temperature  Effect on L-R Processing of 14 Mesh Top
                 Size L.K.  Coal  (33 Wt. Percent Slurries)
                                   78

-------
the higher slurry concentration than at the lower one while the penalty in
reaction time for processing the more concentrated slurry never exceeds a
factor of 1.5 when comparison is made at the same temperature.

     It should be noted that the discussed temperature effects apply only
up to 80 percent pyrite removal; beyond this pyrite removal level the
temperature effect on L-R processing becomes negligible.  Also, the effects
depicted in Figures 10 and 11 represent the overall temperature influence
on suspendable coal desulfurization during L-R processing and not the de-
pendence on temperature of the individual rates controlling the process
pyrite leaching and reagent regeneration.

     The temperature effect on the reagent regeneration rate during L-R
processing is adequately defined through Equation (6) using the AR and ER
values listed in  Section  2.4.2.   It  will  be  shown  that  the  temperature
effect on the pyrite leaching rate, up to 80 percent pyrite removal, can
also be defined by an Arrhenius type expression, but with at least one of
the Arrhenius constants being different for ambient pressure processing
from that for L-R processing.  Beyond 85 percent pyrite removal the
leaching rate appears to be insensitive to temperature in the 102° to 130°C
range.  A rate constant transition region occurs between 80 and 85 percent
pyrite removal, but the leaching rate temperature dependence in this region
was not defined.

     The inability to predict the values of L-R leaching rate constants
from the A,  and EL values estimated from ambient pressure processing of
L.K. coals (see Mixer Section) necessitated that empirical L-R KL values
be determined for each top size coal from L-R data generated at each of
the three temperatures.  The methodology used to determine these KL values
was described in detail in Section 2.4.2.1 in conjunction with the deter-
mination of KL values for the 120°C L-R processing of 100 mesh and 14 mesh
top size L.K. coals.  Using this approach the following KL values were
estimated for llOtand 130°C processing of suspendable L.K. coal:
                                     79

-------
     (a)  110°C L-R Processing
          K,  =0.42 (hours)"  (pyrite cone, in coal)"  for TOO mesh top
                     size coal
          K,  = 0.35 (hours)"  (pyrite cone, in coal)~  for 14 mesh top
                     size coal
     (b)  130°C L-R Processing
          K,  = 0.84 (hours)"  (pyrite cone, in coal)"  for TOO mesh top
                     size coal
          K,  = 0.70 (hours)"1 (pyrite cone, in coal)"  for 14 mesh top
                     size coal

These K, values were determined from randomly selected experimental data
points  and subsequently tested through Equation (12) (an integrated form
of the  leaching rate expression Equation 3) against all the rate data
generated at these temperatures (Tables A-5, A-6, B-5, and B-6  in Volume
2).  The results of this test are shown in Table 11.  The agreement between
coal pyritic sulfur content values derived from coal sample analyses, S*,
and those predicted by Equation (12), (S_)D, is, in general, very good.
Thus, the K.  values listed above appear to be valid (detailed discussion
on the  validity of these comparisons was presented in Section 2.4.2.1 in
conjunction with the review of the data in Tables 9 and 10).

     It is noted that the KL values for 130°C leaching are exactly double
those for 110°C leaching.  A factor of 2 was also estimated to be the dif-
ference in KL values for ambient pressure pyrite leaching at temperatures
differing by 20°C.  This observation implies that the apparent activation
energies governing the leaching reaction rates at ambient pressures and
under L-R conditions are the same.  However, the K, values between 102°C
(ambient pressure) and 110°C (L-R processing) differ by a factor of 3.5.
This unexpectedly large factor is not derivable from the A. and E, values
listed on page 45.
                                    80

-------
                         TABLE 11.    ANALYZED AND PREDICTED Sn  OF L.K.  COAL AS  A FUNCTION OF  L-R
                                       PROCESSING TIME AT  110°CP(94 PSI 09)  AND 130°C (76  PSI 09)
Exp.
No.
t, n = 0.5 Hours
T SP* P
t,n = 1.5 Hours
T SP* (SP'P
t, n = 2.0 Hours
L-K
T SP* 
-------
     Figure 12 is an Arrhenius plot of KL values derived from both L-R and
ambient pressure processing of suspendable L.K. coals.  The ambient pres-
sure curve and the Arrhenius constants derived from it were discussed in
Section 2.4.1.  In contrast to the ambient pressure processing data, the
L-R data exhibit very good compliance to the Arrhenius Equation in the
investigated range.  The Arrhenius constants derived from these data were
as follows:

     (a)  For 100 mesh top size coal
          E.  = 11.1 Kcal/mole pyrite removed
          AL = 8.9 x 105 (hours)"1 (pyrite cone, in coal)"

     (b)  For 14 mesh top size coal
          E,  = 11.1 Kcal/mole pyrite removed
          AL = 7.4 x 10  (hours)'  (pyrite cone, in coal)"

2.4.2.4  Coal Top Size Effects During L-R Processing

     In previous sections it was demonstrated that the effect of coal top
size on suspendable L.K. coal desulfurization by the Meyers Process could
be confined to the leaching rate constant.  It was also shown that  the
KL value of the 100 mesh top size coal was 20 percent larger than that of
the 14 mesh coal, regardless of process operating conditions.  This delta
in KL values represents the predicted maximum coal particle size effect on
pyrite leaching rates when the above two top size coals are processed iden-
tically.

     Under L-R processing, assuming  identical  starting  systems, the coal
top size effect is tempered by the fact that the  Y value  of the coarser
coal slurry at a given reaction time  is higher than the Y of  the  less
coarse coal slurry (this is because the finer  coal reacts at  a higher  rate
generating more Fe+  and the difference in  Fe+2 is not  entirely compensated
by the higher regeneration rate).  Figure 13 illustrates  the  coal  top  size
effect on L-R processing efficiency of suspendable L.K. coal  with typical
L-R data abstracted from Table D-l, Volume  2.   It can be  concluded from

                                    82

-------
oo
to
   0.80
   0.70
   0.60
   0.50

   0.40

   0.30
1
z
8
                   0.20
0.10
                 O
                    0.05
                 8  0.03

                    0.02
                       2.4
                                130°    120°    110°   102°
                          2.6
                                                                        O 100 MESH TOP SIZE
                                                                        A  14 MESH TOP SIZE
                                                                        • 100 MESH TOP SIZE
                                                                              AT AMBIENT
                                                                        COAL PROCESSED
                                                                      PRESSURE
                                               85°
                                                               70° C
                                                                       COAL
                                                                       COAL
                                                                                  L-R PROCESSED
                                                  2.8
3.0 X 10'
                                                                                                _o
                                                                          ,-1
                             Figure  12.
                          RECIPROCAL TEMPERATURE IN (°K)
                     Arrhenius Plots  of  Pyrite  Leaching Rate Constants

-------
     90
Q_



UJ
z
UJ
      80
      70
 I
_l


o


i     60
      50
40
      30
      20
      10
       0.0
            100 MESH X 0 (EXP. 16),



           14 MESH X 0 (EXP. 29)
            0.5         1.0       '  1.5         2.0


                  L-R REACTION TIME,  HOURS
2.5
  Figure  13.  Coal Top Size Effect on L-R Processing of Suspendable

             L.K. Coal
                                 84

-------
these data that the coal top size effect on pyrite removal during L-R pro-
cessing is negligible up to at least the 14 mesh top size.  Since in
general liquid-solids separation efficiency increases with increase in
solids particle size, the data in Figure 13 suggest that desulfurization
of L.K. coal should be preferably performed on the 14 mesh top size rather
than on finer coal sizes.

2.4.2.5  Slurry Concentration Effects

     Slurry concentration, defined as the weight of coal in 100 weights
of slurry, does not appear as a variable in either the leaching or the
regeneration rates.  However, this parameter affects the Y value of the
slurry and, therefore, it is an implicit variable of the leaching rate.
The higher the coal concentration per unit weight of reagent solution the
            +2
higher is Fe   production from pyrite oxidation and therefore the larger
the influence on reagent solution Y which is the Fe  -to-Fe ratio with
  +3        +2
Fe   = Fe-Fe  .  Thus, the slurry concentration effect is maximum during
ambient pressure, batch coal leaching without reagent exchange (e.g.,
Mixer and Settler Operations in this program) and zero if coal is pro-
cessed under constant Y (e.g., continuous, variable rate reagent exchange
leaching).  During L-R processing this parameter affects also the regen-
eration rate since an increase in the slurry concentration will result
            +2
in higher Fe   concentration in the reagent solution of the slurry which
in turn will lead to higher regeneration rates.  The magnitude of the
effect of slurry concentration on pyrite removal depends on the values
                                                              +2
of L-R processing parameters which affect the delta between Fe   production
                                 +2
from the leaching reaction and Fe   consumption in the reagent regeneration
reaction.

     In Section 2.4.2.3 the slurry concentration effect was discussed in
conjunction with the temperature effect on pyrite removal  from suspendable
L.K. coal during L-R processing.  Comparison of the data  in  Figures 10 and
11 led to the conclusion that in the temperature range  of 110° to 130°C  it
was more efficient to desulfurize L.K. coal as 33 wt. percent slurries than
as 20 wt. percent slurries, even though at 110°C the pyrite  leaching  rate
                                     85

-------
was appreciably lower at the higher slurry concentration (note that twice
as much coal is processed per unit time when 33 wt. percent slurries are
substituted for 20 wt. percent slurries).

     Figure 14 compares data generated from the processing of 33 and 20 wt.
percent slurries with all other processing conditions being virtually
identical (120°C, 100 psig, 85 psi 02> 5 wt. percent Fe reagent, 14 mesh
top size coals).  These data show that at the indicated processing con-
ditions, which represent nearly the center of the range of operational
parameters considered feasible for the Meyers Process, the slurry concen-
tration effect on pyrite removal rates is insignificant in the range of
33 to 20 wt. percent slurries.  Thus, the data in this Figure suggest that
the 33 wt. percent slurry represents the minimum desirable slurry concen-
tration for use in conjunction with the indicated L-R processing conditions.

     Review of the entire bank of data in Table D-l, Volume 2 reveals that
the slurry concentration effect on pyrite removal rate during the L-R
processing of suspendable L.K. coal is principally affected by temperature
and oxygen partial pressure.  The effect of this parameter increases as the
values of either temperature or oxygen partial pressure are reduced within
the investigated ranges.

2.4.2.6  Reagent Composition Effects

     The Meyers Process reagent is an iron sulfate solution which under
certain processing conditions require the addition of sulfuric acid.  Thus,
its effects on coal pyrite leaching rates should be related to total  iron
concentration, iron forms (Fe+2 and Fe*3) concentrations, sulfate ion con-
centration, and pH.

     The effects of  iron forms  on leaching  and  regeneration  rates have been
discussed in detail  in  previous  sections.   It has  been  repeatedly demon-
strated that one of  the most  important  parameters  to coal desulfurization
                                                                 +3
by the Meyers Process is the  quantity Y, which  represents the Fe  -to-Fe
ratio in the reagent solution.   The  second  order dependence of the pyrite
                                     86

-------
     90
     80
O
to
LU
U
O
O
Z
Q2

a
Q.

I-


UJ

U

UJ
Q.
     70
     60
     50
     40
     30
     20
                                       20 WT. % SLURRY (EXP. 24)
                                     33 WT. % SLURRY (EXP. 32)
     10
      0.0
   Figure  14.
                 0.5
1.0
1.5
2.0
                     L-R REACTION TIME, HOURS
2.5
               Slurry  Concentration Effect on L-R Processing of
               Suspendable  L.K. Coal
                                  87

-------
 leaching rate on this parameter held valid regardless of processing scheme
 or processing conditions used, provided that the starting reagent contained
 a soluble  iron  salt.  It was also shown that the regeneration rate ex-
 hibits a second order dependence on ferrous ion concentration.  Thus,
 reagent composition in terms of iron forms, especially the ratio of iron
 forms, has a  significant effect on pyrite removal from coal by the Meyers
 Process.

      The sulfate concentration and pH effects of the starting reagent
 solution on Meyers Process performance were investigated concurrently.
 Both  these parameters were varied through the iron sulfate concentration
 in the starting reagent and through added sulfuric acid.  The sulfate
 concentration was varied from zero (when the starting reagent was distilled
 water) to  15 wt. percent (when 5 wt. percent Fe starting reagent of Y = 1
 acidified  by 2 wt. percent sulfuric acid was used).  The corresponding
 starting reagent pH range was 7 to approximately 0.7.  It should be noted,
 however, that even when the starting pH was 7 its value dropped to below
 2 within a few minutes of reaction time because pyrite oxidation generates
 sulfuric acid.  The drop in pH was experienced even under L-R processing
 where the generated sulfuric acid is, to a large extent, quickly consumed
                                                      +2      +^
 by the regeneration of the ferric ion (oxidation of Fe  -to-Fe  ).
                                                         /
      According to rate Equations (3) through (6) sulfate concentration and
 pH should not affect directly either the leaching or regeneration rates.
 The data verified this expectation, except that at L-R reaction times in
 excess of approximately one hour these two parameters exhibited a small
 indirect effect on pyrite removal through their influence on the Y of the
 reagent solution.  The data in Appendices A, B, and C show that when  L.K.
 coal was processed under L-R conditions with reagent solutions which  did
 not contain added sulfuric acid, the iron content of the coal during  pro-
 cessing did not decrease in proportion to pyrite removed (Experiments 1
through 11, 15,  18 through 22, 41, 42, 43, and 45).  Comparison of pre-
dicted (Table D-l) and experimental  Y values indicated that iron was
depositing  on the coal  at a higher Fe+3-to-Fe+2 ratio than that in the
reagent solution;  thus,  the measured Y values in these experiments at t,  R

-------
values in excess of one hour were increasing at lower rates than expected
and at longer reaction times the Y actually dropped with increasing t,
in certain experiments (e.g., Experiment 1).  It should be noted, however,
that at the tL_R range where iron begins to have an effect on Y, the
reagent Y values are high and pyrite leaching is principally limited by
the low pyrite concentration in coal; therefore, iron deposition has an
insignificant effect on pyrite removal during L-R processing.

     The major sulfate concentration-pH effect on the Meyers Process was
not on pyrite removal  but on sulfate deposition on  processed coal and on
processed  coal washing.  The data  in Appendices A and B  show that,  in the
experiments where  sulfuric  acid was not  added to the starting reagent,
iron  and sulfate deposited  on  the  coal during L-R processing when pyrite
removal exceeded 40-50 percent and the reagent Y value exceeded  approxi-
mately 0.5.   The deposition of iron  and  sulfate  occurred to  approximately
the same extent when 3 wt.  percent starting Fe reagent solution was used
(10.7 wt.  percent  ferric sulfate) as when 5 wt. percent  Fe reagent  (17.9 wt.
percent ferric sulfate) was used.  The same observation  is made, with re-
spect  to iron deposition, from the data  in Appendix C which were generated
with  reagent  solutions whose iron  sulfate concentration  was derived from
coal  (iron sulfate was not  added to the  starting reagent  solution); during
these  experiments  sulfate did  not  deposit on the coal.   Iron and sulfate
deposition during  coal processing  under  separate leaching regeneration
conditions at ambient  pressure was minimal  (< 0.2 wt. percent sulfate
sulfur) even with  no added  acid to the starting reagent  solution and
regardless of the  iron sulfate concentration in the slurry (up to 10 wt.
percent Fe provided the Y was  maintained above 0.8).

     The above experimental observations are consistent with the Meyers
Process chemistry  formulated on the basis of elemental sulfur recovery
and iron mass balance  (ferrous  ion production from  known  quantities of
oxidized coal pyrite).  According  to this chemistry each mole of pyrite is
converted to one mole Fe+2, 1.2 moles of sulfate, and 0.8 moles of ele-
mental sulfur through oxidation by Fe+3 according to the following
reactions:
                                    89

-------
     0.4 FeS2 + 0.4 Fe2(S04)3           * 1.2 FeS04 + 0.8S
     0.6 FeS2 + 4.2 Fe2(S04)3 + 4.8 H20 -   9 FeS04 + 4.8 H2S04

The sum of these two reactions yield the following net reaction for pyrite
leaching in the absence of oxygen (separate Teaching-regeneration):

     FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 + 10.2 FeS04 + 4.8 H2$04 + 0.8S

Thus,  each mole of pyrite reacted yields 4.8 moles of sulfuric acid and
hence  the observed immediate drop in pH.  The low pH inhibits the hydrolysis
of the iron salts and, therefore, no iron deposition occurs on the coal
unless iron sulfate solubilities are exceeded.

     Under L-R processing oxygen is added to the process during coal
                                                               +3
pyrite leaching for the purpose of regenerating the consumed Fe   (ferric
sulfate).  Thus, the following reaction takes place simultaneously with
the leaching reaction:

     2 FeS04 + 0.5 02 + H2$04 ->  Fe2(S04)3 + H20

At steady-state, the leaching and regeneration reactions yield the fol-
lowing net reaction:

     FeS2 + 2.4 02 •*•  0.6 FeS04 + 0.2 Fe2(S04)3 + 0.8S

In order to maintain steady-state operation, the product sulfate must  be
removed continuously from the system at the sul fate-to-iron ratio produced
which  in molar terms is 1.2.  At bench-scale, experimentation was per-
formed in batch mode and the product sulfate was not removed from the
                '                +2                                  +3
system until  virtually all the Fe   in the system was oxidized to  Fe
(Y >0.90).   Thus,  the 0.6 moles of FeS04 product per mole of reacted pyrite
was virtually all  oxidized to the ferric state.  In experiments wbere
acid was added to the starting reagent the expected reaction  is

     0.6 FeS04 + 0.3 H2S04 + 0.15 02 ->  0.3 Fe(S0)  +  0.3 H0
                                     90

-------
with the product sulfate remaining  soluble.   The data  in Appendix B, Volume
2 verifies this expectation  (acidified  runs).   In  the  absence of added
acid, the ferrous sulfate oxidation would  be  expected  to yield ferric
oxide according to the following  reaction

     0.6 FeS04 + 0.15 02 -*•   0.2 Fe2(S04)3  + 0.1  Fe203

with the product iron oxide  remaining on the  processed coal unless acid
washed.  The data in Appendix A verifies this  expected iron deposition
on the processed coal.

     The above analysis is totally  based on the iron-to-sulfate ratio in
the system with no regard to pH.  While it adequately  explains the iron
oxide deposition on the processed coal  it  does not explain the observed
simultaneous sulfate deposition when non-acidified starting reagents were
used.  One explanation that  may be  advanced is occlusion of sulfate in the
deposited iron.  A more probable  explanation  is the formation of insoluble
basic iron sulfate.  Chemical analyses  on  the  deposits revealed that the
iron-to-sulfate ratio in them was 1.5.  According  to the literature, the
most insoluble form of basic iron sulfate  is  3 Fe20,-4 S03'9 H20 which has
the same iron-to-sulfur ratio as  the deposit  on the processed coal.  Further-
more, phase diagram data on  the iron oxide-sulfate-water system appears
to indicate that formation of the above basic  iron sulfate could be expected,
in small quantities, at the  L-R processing conditions  utilized in the per-
formed experimentation when  the starting reagent did not contain added acid.
Addition of sulfuric acid eliminates formation of  the  basic iron sulfate.

     The total iron concentration  should  also have no direct effect on
the Meyers Process according to the formulated rate expressions, Equations
(3) through (6), except as it affects Y in operations  where neither re-
generation nor reagent exchange are operative, e.g., Mixer and Settler
                                                                 +3
Operations.  In these operations, the portion of Fe present as Fe   is
being irreversibly reduced to Fe+2.  Thus, it  is expected that for a given
Fe+3-to-Fe+2 ratio in the starting  reagent solution, the higher the
starting Fe concentration in the  reagent is per mole of pyrite present,
the higher the pyrite removal would be  at  any reaction time interval.  The

                                    91

-------
data in Appendices A through C verify  this expectation.   For example,
pyrite removal  in the Mixer Operation from 33 wt. percent coal slurries,
after approximately one hour of processing, ranged from 9-12 percent when
5 wt. percent Fe reagent solution was used; 20 wt. percent coal slurries
with the same reagent processed for equivalent periods yielded 18-24 per-
cent pyrite removal.  Also as expected, the experiments processed with
zero iron concentration in the starting reagent led to less than one percent
pyrite removal  during approximately one hour of Mixer Operation (Appendix C
data).  The data from the Settler Operation were analogous, but less pro-
nounced and more difficult to distinguish  because pyrite  removal based  on
starting coal pyrite was very small in  this unit operation and even small
noise in the data can easily mask effects  on rates or on  extent of pyrite
removal.

     During L-R processing the total  iron  concentration of the reagent
should have no effect on Y, except at the  start of the operation if the
starting L-R Y is different than 1.0.  Since reagent regeneration depends
     +2                                 +>2
on Fe   concentration, the higher the Fe   concentration  in the starting
L-R reagent solution is, the higher would  be the regeneration rate at the
                                              +2
start of the operation.  Once, however, the Fe   ions in  the starting
solution are regenerated, the regeneration rate depends solely on the
  +2
Fe   generation rate from the pyrite leaching reaction which does not de-
pend on total iron concentration.  For example, a starting 33 wt. percent
coal slurry of 5.6 wt. percent Fe reagent  with a Y of 0.4 contains 0.6
        +2
moles Fe   per liter; after 15 minutes  of  L-R reaction, an additional 0.4
           +2
moles of Fe    have been added to each  liter of reagent solution from the
pyrite leaching reaction and yet, because of regeneration,  less  than 0.5
  +2
Fe   moles are present in the reagent solution at the end of this reaction
interval and only 0.35 moles at the end of 30 minutes reaction time even
                                 +2
though approximately 0.7 moles Fe   were generated by the leaching  reaction
(see Experiment 32, Table D-l, Volume 2).   Thus, it is apparent  that the
          +2
bulk of Fe   present  in the starting reagent was  regenerated during the
first few minutes of  L-R operation.  Similar observations are made from
the L-R processing of 20 wt. percent slurries  in  5  and 3 wt.  percent iron
reagent; in  these experiments the Y of the  reagent increases even faster
because of the lower pyrite concentration  on a per  liter of reagent basis.
                                    92

-------
     Equation (5), the regeneration rate expression, and the Y versus t,  R
data generated with starting reagent to which no iron was added (Table C-l
Volume 2) indicate that there is an Fe concentration below which its
effect on the Y values during L-R processing cease to be insignificant.
Yet the data in Figure 15 show that even in the "no added iron" case,
where the required,Fe+  for the initiation of the pyrite leaching reaction
is assumed to have been derived from non-pyritic iron in the coal, the
effect of Fe concentration on pyrite removal rates during L-R  processing
was negligible.   It is believed that the explanation is that the concen-
tration  of iron in the reagent solution adjacent to the reaction sites is
substantially higher  than that in the  bulk and  so is the Y  if  it is assumed
that oxygen diffusion to the vicinity  of the reaction sites is faster than
iron ion diffusion to the bulk reagent solution.  Assuming  that this ex-
planation is valid, it is reasonable to assume  that the data in Table C-l
could  have been predicted by Equation  (3)  if the Y values near the reaction
sites were known.

     Pyrite leaching  by direct oxidation with oxygen was also considered as
an alternate explanation for the unpredictability of pyrite removal rates
during the "no added  iron" experimentation.  However, examination of the
data indicated that the reaction product mix was independent of the iron
concentration in  the  feed reagent (see Section  2.4.5.1 page 101).   It was,
therefore, concluded  that a change  in  reaction  mechanism with no apparent
change in the process chemistry was very unlikely.  It should be noted
that the role of  iron in Meyers Process chemistry was unequivocally proven
from separate Teaching-regeneration experimentation.

     In conclusion, starting reagent composition did not exhibit any ap-
preciable effect on pyrite removal rates during L-R processing, but it did
affect iron deposition on the coal during processing.  The latter effect
was traced to insufficiency of sulfate ion correctable by the addition of
sulfuric acid.   For the particular coal processed in this program, the
sulfuric acid requirement was estimated at 1.5  percent of the weight of
reagent used in 33 wt. percent coal slurries (maximum slurry concentration
investigated).   Two weight percent sulfuric acid was actually used.
                                    93

-------
    90
                20 WT % SLURRIES OF 100 MESH TOP SIZE L. K. COAL
    80
    70
O

(S)
3  60
O
Z  50
C£
O
    40
Of
    30
    20
                              5 WT % Fe REAGENT (EXP. 7)
                              3 WT % Fe REAGENT (EXP. 6)
                              NO ADDED Fe REAGENT
                                               (EXP. 44)
    10
     0.0
0.5
1.0
1.5
2.0
2.5
3.0
                           L-R REACTION TIME, HOURS
         Figure 15.   Effect of Total  Iron  Concentration on Pyrite Removal
                    During L-R Processing of Suspendable L.K.  Coals
                                     94

-------
2.4.2.7  Oxygen Partial Pressure  Effects

     Figure 16 illustrates the  oxygen  partial  pressure  effect  on pyrite
removal from suspendable  L.K. coal  during  L-R  processing with  data from
Table D-l, Volume 2.  Equation  (5), the regeneration  rate  expression,
predicts a nearly four-fold  increase in regeneration  rate  when the oxygen
partial pressure is increased from  35  psi  to 135  psi; the  data in Figure 16
show that such an increase in regeneration rate translates to  a nearly 70
percent increase in the L-R  pyrite  leaching rate  at 120°C.  Thus, the oxygen
partial pressure used during L-R  processing has an appreciable effect on
pyrite leaching from suspendable  L.K.  coals.   It  is noted  that the experi-
mental data in Appendix B, Volume indicate a substantially reduced effect
from that depicted in Figure 16.  Analysis of  the data  revealed that the
apparent negligible oxygen partial  pressure effect on L-R  pyrite removal
was due to inconsistent regeneration.  Comparison of  experimental and pre-
dicted L-R Y values indicated that  regeneration was efficient  throughout
the experimentation at 50 psig, but inefficient during  the early stages of
L-R processing at the higher pressures (refer  to  Section 2.4.2.2 for details
on comparison of predicted and  experimental Y  values  during L-R processing).

2.4.3  Settler Unit Operation

     In early program experimentation, approximately  one-half  of the slurry
at the conclusion of L-R  processing was transferred to  a surge tank where it
was permitted to react further  at 90°C and ambient pressure while at the
same time the settling behavior of  the coal was being examined; this
practice originated the title of  this  unit operation.   However, laboratory-
scale tests on liquid-solids separation of suspendable  L.K. coal slurries
in 5 wt. percent Fe reagent  solution revealed  that a  settler operation was
not necessary even for 20 wt. percent  slurries of 100 mesh top size coal.
The high pyrite removal rates from  33  wt.  percent slurries of  14 mesh top
size coal assured that the process  did not require a  settler.  However, the
practice of further processing  part of the L-R processed slurry in an
ambient pressure reactor  continued  when it became apparent that L-R pyrite
leaching rates dropped substantially at pyrite removal  levels  exceeding 80
percent.   The settler now became  a  stirred, tail-end  reactor after the L-R
reactor.
                                     95

-------
O
U
g
D-
O
5
Q
UJ
U
    90
    80
    70
    60
    50
    40
     30
     20
             MESH TOP SIZE COAL  PROCESSED AT  120°C  AS 33 WT %  SLURRIES
PSIG (135 PSIO2), EXP.
PSIG ( 85 PSI 02), EXP.
PSIG ( 35 PSI O,), EXP.
                                                            36
                                                            30
                                                            35
0.0        0.5        1.0        1.5        2.0       2.5
                      L-R REACTION TIME,  HOURS
                                                                   3.0
                               3.5
    Figure 16.  Effect  of Oxygen Partial Pressure on L-R  Processing
                 of Suspendable L.K.  Coal
                                      96

-------
     Settler performance was monitored by sulfur forms analyses of coal
samples taken at the start and at the conclusion of the approximately 20
hour operation and by the ferrous ion production during the operation.
The percent of the raw coal pyritic sulfur removed during the "settler"
operation is easily determined from the rate data listed in Appendices A
through C, Volume 2; it is the difference in the S  removal values listed
for "L-R processed coal" and "L-R + settler processed coal".  In the
majority of experiments pyrite removal in the Settler Operation ranged
between 5 and 10 percent of the pyrite in the raw coal or 20 to 40 percent
of the pyrite present in the coal at the start of this unit operation.  As
expected, pyrite removal was nearly zero when the experiment's starting
coal slurry was prepared with iron-free reagent (data in Table C-l, Volume
2)-

     Pyrite removal in the Settler Operation varied because the unit's
starting slurry composition varied.  Since the reagent was neither ex-
changed nor regenerated in this unit operation, Meyers Process chemistry
dictates that the quantity of pyrite expected to be leached from the coal
will be proportional to the quantity of ferric ion present in the reagent
of the feed slurry.  The rate of removal is governed by Equation  (3).
Since the operating conditions of this unit (temperature, pressure, re-
action time) and the Y value of the feed slurry were very nearly the same
in the majority of experiments which included  the Settler Operation, the
differences in pyrite removal in this operation among these experiments
should be principally due to the different values of the reagent  Fe to coal
pyrite ratio of the feed slurry.  The larger this ratio, the higher should
be the pyrite removal in the "settler" obtained in the pertinent  experiments
listed in Appendices A through C, Volume 2.  Thus, the less concentrated
slurries with the highest Fe content should have yielded the highest pyrite
removals after 20 hours of settler processing, and vice versa.  The latter
expectation is certainly validated by comparing the low iron settler data
contained in Appendix C to corresponding data  from any of the experiments
listed in Appendices A or B.  Less extreme comparisons between Experiment
12, Table C-l, which was performed on 20 wt. percent coal slurry, and
Experiments 13 through 15, Table C-2, which were performed on 33  wt.  percent
slurries, validated the above expectations, also.  Analyses of the  processed
                                     97

-------
coals show that 47  percent  of  the  feed  pyrite  to  the  settler was  removed
in Experiment 12, while the corresponding  pyrite  removals  from Experiments
13, 14, and 15 were 17, 14, and  21  percent,  respectively.

     In general, the experimental  data  derived from "settler" processed coal
were in qualitative agreement  with the  predicted  behavior  based on the
chemistry of the Meyers Process  and on  Equation (3).   However, quantitative
comparisons were substantially less successful as the data in Table 12 in-
dicate.  Table 12 compares  percentages  of  pyrite  removal during the Settler
Operation computed  from (a) sulfur forms analyses of  the feed and product
coals and (b) the  quantity  of  ferrous ion  produced in the  "settler",
assuming that the chemistry of the Meyers  Process applies  to this unit
                        4.9
operation (10.2 moles Fe   produced from each  mole of pyrite reacted).

              TABLE 12.   PYRITE REMOVAL  IN THE SETTLER OPERATION
Experiment Nos.
% Pyrite Removed Based on
(a) Coal Analyses
(b) Fe+2 Production
23

10
32
28

14
29
29

17
28
30

47
32
31

13
20
33

52
23
Obviously the agreement between the two sets of pyrite removal values is
poor, but the difference is equivalent to only 0.1 to 0.2 percent pyritic
sulfur in coal which is not too far in excess of expected deviations in
the chemical analyses involved.  Note that each pyrite removal computation
from coal analyses (set (a) in Table 12) required two coal sample analyses
for pyrite content, settler feed and settler processed coal samples, each
of which contained less than one percent pyritic sulfur with allowable
deviation 0.1 percent.  On the other hand, Fe+2 production was determined
from large changes in Fe+  concentration (100 to 200 percent change  in  Fe+2
concentration in the reagent solution, depending on slurry concentration)
and, therefore, should yield a more reliable pyrite removal valve.   Using
equation (3) and a KL value of 0.07 (hours)"1 (VJJ"1 for  90°C processing
of 14 mesh top size coal, estimated from the Arrhenius plots  in  Figure  12,
predicted "settler" pyrite removal values were computed for the  experiments
                                    98

-------
in Table 12.  The predicted  values  ranged from 30 to 36 percent showing
relatively good agreement with the  pyrite removal values derived from
ferrous ion production.

     From the above analysis  of the "settler" data it can be  concluded that
the Settler Operation exhibited compliance to the formulated  Meyers Process
chemistry and obeyed the developed  leaching rate expression with rate con-
stant values which are derivable from the ambient pressure A,  and E,
values computed in Section 2.4.1.

2.4.4  Coal Washing Operation

     Several processed coal washing schemes were investigated during the
early stages of the program  in an effort to identify an  efficient scheme
for washing deposited stil fates from coal  processed with  reagent which did
not contain added sulfuric acid.  Multiple stage water washing (3 to 4
stages with water equal to twice the weight of coal  per  stage),  expanded
wash times (0.5 to 1.0 hours), and  washing with spent reagent did not
materially reduce the deposited sulfate.   The deposited  sulfate could
only be removed by agents which simultaneously removed the deposited iron
oxide; for example, a single  stage  coal  washing with 1 N sulfuric acid
solution followed by two stages of  water wash dissolved  completely both
deposits.  This observation combined with Meyers Process Chemistry pre-
dictions on iron deposition led to  the conclusion that the deposited
sulfate was either occluded in or chemically bound with  the deposited
iron (basic iron sulfate) and, therefore, it could not be removed without
the simultaneous dissolution  of the iron  oxide (see Section 2.4.2.6 for
details).

     As predicted, the addition of  2 wt.  percent sulfuric acid to the
starting reagent solution helped to maintain large majority of the process
iron and sulfate products in  solution and inhibited  to a large extent the
formation of water insoluble  sulfates.   Thus,  coal  processed  with acidified
reagent required washing principally for  the removal  of  reagent  retained
by the coal  after slurry filtrations;  hot water wash accomplished this
task effectively as the data  (acidified  runs)  in Appendices A through C,

                                     99

-------
 Volume 2 and Table 18  below  indicate.  However, the problem  of  sulfate
 deposition on processed  coal was  not completely solved with  the addition
 of the 2 wt.  percent sulfuric acid to the starting reagent.  At long  L-R
 processing reaction times  the residual sulfate on processed  and washed
 coal  remained unacceptably high  (see data in Tables B-l and  B-2,  Volume 2).
 Extrapolation of phase diagram data on the  Fe^-SOg-^O system available
 in literature appear to  indicate  that higher sulfuric acid concentrations
 may be required  in the starting reagent than the 2 wt. percent  utilized
 in this program  when L-R reaction temperatures exceed 110°C.  Efforts
 to reduce residual sulfate sulfur on processed coal to below 0.2  wt.  per-
 cent continue.   It should  be noted that coal leaching with (a)  low  iron
 concentration reagent, (b) for less than 3  hours, and (c) at temperatures
 below 120°C does not present the  sulfate deposition problem.

      A two-stage wash  scheme was  adopted as adequate for washing coal
 processed with acidified reagent  (note exceptions in previous paragraph).
 Each stage consisted of  a  slurry  wash and a cake wash and for each  wash
 a water-to-dry coal weight ratio  of 2 was employed.  During  reagent re-
 cycle testing the two-stage wash  scheme was preceded by a brief cake  wash
 with  a quantity  of hot water equal to 50 percent the estimated  weight of
 the dry coal.  The collected cake wash filtrate was then used as part of
 the make-up reagent solution and, thus, reagent recycling was more
 complete.

      Table  13 summarizes Wash Section data  derived from the  washing of L.K.
coals  processed  under  a  variety of experimental conditions;  the data  are
typical of  the results obtained in this unit operation.  The data in  the
third  column of Table  13 represent the estimated quantity of reagent  iron
on the wet  coal after  filtration  of the L-R or Settler reactor  slurries.
The quantity of spent  reagent retained by the processed coal after  filtra-
tion was estimated from the difference in weights of the wet filter cake
and of the dry product coal plus  recovered  sulfur residue; the  iron was
then computed from the measured iron concentration in the filtrate.  The
fourth column lists the Y value of the spent reagent solution retained  by
the coal determined from iron analyses on the filtrates.  Columns 5 through
                                    100

-------
                      TABLE 13.   TYPICAL WASH SECTION DATA FROM  L-R PROCESSED  L.K.  COALS
Experiment
No.
12 (AG)*
17**
25***
30 (AG)****
Coal
Top
Size
(Mesh)
100
100
14
14
Estimated
Reagent Fe
on Coal
(gin)
15.2
21.8
43.2
31.4
Reagent
Y
0.85
0.84
0.96
0.56
First staoe Wash Filtrates
Slurry Wash
Fe (gm)
12.7
17.7
35.4
23.0
Y
0.81
0.69
0.82
0.41
Cake Wash
Fe (gm)
2.8
3.5
7.0
4.8
Y
0.81
0.43
0.80
0.42
Second stage wash nitrates
Slurry Wash
Fe (gm)
0.1
0.7
0.8
1.1
Y
0
0.03
0.16
0.11
Cake Wash
Fe (gm)
0.0
0.2
0.3
0.4
Y
-
0
0.15
0.12
Recovered
Fe
(Percent)
102.6
106.0
100.7
93.3
*     Coal  processed  as  a  20 wt. percent slurry in 5 wt.  percent  Fe  reagent solution for 5 hours at 120°C
      and 100 psig;  the  slurry was also Settler processed.

**    Coal  processed as  a  33 wt. percent slurry in 5 wt.  percent  Fe  reagent solution for 2 hours at 120°C
      and 50 psig.

***   Coal processed as  a  20 wt. percent slurry in 5 wt.  percent  Fe  reagent solution for 6 hours at 120°C
      and 100 psig.

****  Coal processed as a  33 wt. percent slurry in 5 wt.  percent  Fe  reagent solution for 2 hours at 120°C
      and 100 psig; the slurry was  also Settler processed.

-------
 12 show the total  iron  and  the  forms  of iron  (Y  value)  recovered from
 each wash.   The last column shows  the iron  recovered  in the  Wash Section
 as a percentage of the  iron in  Column 3.

      The data in Table  13 show  that better  than  95  percent of the reagent
 iron on the processed coal  was  recovered  in the  first wash stage and that
 the remaining was completely recovered in the  second  stage;  the sulfate
 recovery values should  be analogous.   The low  Y  values  in  the filtrates
 of the second stage washes  indicate that  the  iron salt  dissolved in  this
 stage was principally in  the ferrous  state  which could  mean  deposited
 ferrous sulfate; however, the quantities  of iron salt present in these
 filtrates are insignificant in  terms  of sulfate  deposition of the coal.
 In fact, the quantity of  iron present in  the  second stage  wash filtrates
 renders the need for a  second wash stage  questionable.

 2.4.5  Elemental Sulfur Recovery Operation

      The product elemental  sulfur  from the  experiments  listed in Appendices
 A through C was recovered by leaching the processed and washed coal  with
 toluene.  However, special  proof-of-principal  experimentation was performed
 on elemental  sulfur recovery by hexane and  by  vaporization in inert  gas.
 The three techniques and  the data  obtained  from  them  are discussed below in
 separate sections.

 2.4.5.1  Elemental Sulfur  Recovery with Toluene

      The procedure  used to  recover the product elemental sulfur during L-R
 experimentation  was  described in Section  2.2.  In outline, the wet coal
 from  the  Wash Section was slurried in twice the  estimated  dry coal weight
 toluene,  the water on the coal  was azeotroped  off at  85°C, the toluene
 slurry was refluxed  at 108°C  for 30 minutes and  filtered hot, and the
 filtrate evaporated  at low  temperatures to  recover  the  dry sulfur residue.
 This procedure constituted  a single stage sulfur recovery  operation; in the
majority of experiments the above  procedure was  repeated (two stage  sulfur
 recovery).

                                    102

-------
     The sulfur residues from either stage analyzed to between 80 and 90
percent by weight elemental sulfur depending on the attained degree of
dryness; a small portion of the residue was carbonaceous material.  The
weight of residue obtained from the first stage recovery was approximately
nine times that recovered from the second stage for virtually every experi-
ment; thus, 90 percent of the recovered elemental  sulfur was leached from
the coal in the first stage.  The quantity of elemental sulfur recovered
in each experiment is listed in the sulfur balance section of the mass
balance data tables contained in Appendices A through C, Volume 2.

     According to Meyers Process chemistry, formulated on the basis of data
generated during the ambient pressure  leaching of  coals with iron salt
solutions, coal pyritic sulfur during  leaching with iron salts is oxidized
to sulfate sulfur (S ) and elemental sulfur (S ) at the molar ratio of 1.5.
This ratio was computed from the quantity of ferrous ion produced per mole
of pyrite removed from coal and from the sulfate sulfur produced in a small
number of carefully mass balanced coal extractions with ferric chloride.
On the basis of this ratio, 40 percent of the pyritic sulfur removed from
the coal should be recovered as elemental sulfur.  Actual Sn recovery from
these ambient pressure experiments, after 60 minutes reflux of the pro-
cessed coal in toluene, was between 80 and  90  percent of that expected from
the $s/Sn  ratio of 1.5.   Conversely,  the S /S  ratios derived from the  quantv
ties of pyrite removed and elemental sulfur recovered ranged in value from
1.8 to 2.1.  This  discrepancy in the S /Sn values  was attributed to incom-
plete S  recovery principally due to operational losses rather than leaching
inefficiency.  In fact, it was concluded that an 80-90 percent recovery of
product elemental sulfur represented a reasonable  expectation from a bench-
scale unit operation which involves a  numbers of steps and yields 20 to 40
grams of product from 8 kilograms of slurry.
     Simultaneous coal  Teaching-reagent regeneration does not permit  the
determination of S /S   ratios from ferrous ion production.   Thus,  elemental
sulfur recovery was  used  to determine, at least through comparison with
ambient pressure data discussed above, the validity of Meyers Process
                                     103

-------
chemistry during  L-R  processing of L.K. coals.  Percent sulfur recovery
and the S /S  ratio were  computed for each experiment.  Percent $n recovery
was computed from Equation  (13)

                               S  recovered
              % S recovered  =                    x 100         (13)
                 n
                                    x S  removed
on the assumption that the Ss/Sn ratio of 1.5 was valid.   The Ss/$n values
were computed from Equation (14)

                        S  removed   S  recovered
              S /S  =  — £ - ~-JL-A -                 04)
              °s' n          S  recovered                       x
                              n

on the assumption that the quantity of recovered $n represented the entire
yield of S  in the experiment (100 percent Sn recovery).   Table 14 sum-
marizes the results.

     The data in Table 14 show appreciable scatter, but the range of values
is nearly identical to the range of the corresponding values obtained from
ambient pressure processed coal (one 30 minute toluene extraction was
performed in the first six experiments; therefore, Sn recovery was
approximately 10 percent lower than it was in the rest of experiments).
The correspondence of ambient pressure and L-R processing S  recovery data
permits at least the Inference that the actual S /S  ratio during L-R pro-
cessing is also 1.5.  The data scatter appears to be the same within each
group of experiments as it is among groups.  Since each group of experi-
ments was performed under different experimental conditions, the 1.5 value
of the Ss/$n ratio evidently applies to the entire spectrum of experimental
conditions used in this program.  It may be argued that the sulfur  recovery
from coal processed at 130°C, Experiments 20, 21, 34, and 40, and that from
coal processed with iron-free starting reagent, Experiments 42,  44, and 45,
points to a higher $s/Sn ratio than that applicable to the rest  of  the
experiments.  At least for Experiments 42 through 45 such conjecture does
not appear valid on the basis of sulfate recovery.  Because  in these
experiments the starting reagent solution did not contain iron sulfate, it
was possible to estimate the sulfate product of the pyrite leaching reaction;
in all  four experiments the Sg value pointed to a S /S  ratio  of 1.5  (the
actual  values were 1.7 + 0.1).
                                     104

-------
TABLE 14.  ELEMENTAL SULFUR PRODUCT RECOVERY FROM THE L-R
           PROCESSING OF U.K. COAL
Exp.
Nos.
1
2
3
4
5
6

7
8
9
10
11
12

13
14
15
16

17
18
19
20
21
42***
44***
45***
*
**
***
100 Mesh Top Size Coal
% Sn if
Recovered* v~
75.9
77.3
77.2
69.8
76.2
74.5

92.7
80.7
89.3
73.5
74.4
95.5

90.0
85.1
78.9
88.0

84.5
85.6
(65.5)
70.7
80.1
70.0
72.4
72.0
s/snj**
2.29
2.23
2.23
2.58
2.28
2.36

1.70
2.10
1.80
2.40
2.36
1.62

1.78
1.94
2.17
1.84

1.96
1.92

2.54
2.12
2.57
2.45
2.47
14
Nos!
23
24
25
26
27

28
29
30
31
32
33

34
35

36

37
38
39
40
41


Mesh Top Size Coal
Recoverec
83.6
89.2
85.2
(71.4)
86.8

85.8
84.8
87.8
86.8
73.1
(60.8)

79.9
83.0

84.9

71.2
81.4
70.2
68.1
43.7



Percent of expected product Sn, the latter computed
of pyrite removed and S /S = 1-5.
Ss/Sn value computed from the quantities of pyrite
covered .
Coal processed with
lifc" * C* **/
1.99
1.80
1.93

1.88

1.91
1.94
1.85
1.88
2.42


2.13
2.01

1.95

2.51
2.07
2.56
2.67
(Coal leaching was
performed with
def earning agents)


from the quantity
removed and Sn re-
iron-free starting reagent.
                           105

-------
     The data in Table 14 proves conclusively that L.K.  coal pyrite leaching
during L-R processing, at least within the parametric ranges investigated,
always yields the two sulfur products specified by the Meyers Process
chemistry, sulfate sulfur and elemental sulfur.  In addition, these data
strongly suggest that the reaction mechanism under all leaching conditions
examined to date, including ambient pressure processing, is the same, even
though the process may become diffusion limited at some point.  This obser-
vation lends additional validity to the single rate expression applicability
concept already used for describing pyrite leaching rates from suspendable
L.K. coal regardless of processing conditions.

     Table 14 does not include data from Experiment 43 because toluene
leaching was performed subsequent to vacuum drying of the processed coal
rendering the recovered S  value unreliable.  The data from Experiment 22
are nearly identical to that derived from Experiment 41 shown in the table.
In these two experiments wetting and defoaming additives were used which
apparently distorted the sulfur analyses performed on the processed coal
rendering meaningful data interpretation impossible; it is believed that
the unusually low $n recovery indicated that the pyrite removal during these
experiments was over estimated due to errors in sulfur forms analyses (see
also Tables A-7 and B-7, Volume 2).  The values in parentheses indicate in-
complete recovery of Sn due to traced residue loss, the weight of which
could not be accurately determined.

     Toluene retention on the product coal was also determined through
specially designed experimentation.  Approximately 500 grams of dry coal
were wetted with 50 percent its weight water and refluxed for 60 minutes in
500 grams of toluene in a well trapped apparatus; the water was azeotroped
off prior to reflux.  The toluene slurry was then filtered  in a closed
system and the toluene-wet coal cake was vacuum dried at 100°C overnight
in a well trapped oven.  A careful mass balance was performed on the coal,
water, and toluene from the precise weights of feed and product coal and
liquids.   The experiment was repeated with twice the weight of feed  mate-
rials.  The mass balances from both experiments were virtually identical
and indicated a 0.9 percent increase in the weight of coal,  a  0.3  percent
                                    106

-------
 increase in the weight of water, and a 2 percent decrease in the weight of
 toluene.  Thus, the total system mass balance shows that 0.9 percent of
 toluene was not accountable.  In terms of coal mass balance, the coal re-
 tained 0.9 percent of the toluene; in terms of toluene mass balance it
 retained 1.8 percent of its weight toluene (note that the water gain
 represents approximately 0.2 percent toluene).  Thus, our estimate of
 toluene retention on processed coal is one percent of the coal weight.


 2.4.5.2   Elemental  Sulfur Recovery with Hexane
                                  
-------
temperature did not  furnish meaningful  improvements.   The  data  in  Table  15
illustrate the relative $n leaching  efficiencies  of hexane and  toluene.
These data show that only a small  percentage  of the elemental sulfur  in  the
coal was recovered  in the two  hexane leaching stages  preceding  the toluene
stage.  It is also  noted that  in three  of  the four  batches of coal  less
sulfur was recovered in the first  stage than  in the second;  this observation
appears to be consistent with  the  contention  that hexane's inefficiency  was
due to its inability to displace water  from coal  at reasonable  leaching
times.

2.4.5.3  Elemental  Sulfur Vaporization

      Although  the atmospheric boiling point of sulfur is 444.6°C,  which is
 somewhat  above the  decomposition temperature  of coal, an appreciable vapor
 pressure  (sufficient for engineering design of equipment for removal  of
 sulfur by vaporization)  begins about 250°C (Figure 17).  At 350°C  (still
 below the thermal decomposition point of coal), the vapor pressure of
 sulfur is substantial  M50 mm Hg).  Therefore,  it seems  reasonable that
 elemental  sulfur could  be removed from coal by vaporization either in
 inert gas stream such  as steam or nitrogen or under vacuum, provided that
 the sulfur does not extensively react with coal during the  thermal process.

     An initial demonstration of the removal  of pyritic sulfur from coal,
via ferric ion oxidation, was performed utilizing vaporization for
removal of elemental sulfur.   In this case, vaporization was effected at
100-120°C and  150 mm of Hg pressure with no detectable coal decomposition
or elemental sulfur reaction  with coal.  In fact, greater than 99  percent
pure  sulfur was obtained as condensate on a heat-exchanged  surface.  How-
ever,  vacuum vaporization in commercial process equipment  is thought to be
relatively expensive, so that the possibility of vaporizing sulfur from
coal at atmospheric pressure  and in an inert gas stream was investigated
(Table 16).

     A coal sample was treated with ferric sulfate under  the usual atmos-
pheric leaching conditions  but neither water washed  to remove  iron sulfate
from the coal pore structure  nor extracted with  solvent to remove elemental

                                   108

-------
    TABLE 15.  ELEMENTAL SULFUR  RECOVERY  BY  SUCCESSIVE STAGES OF
               HEXANE-TOLUENE  LEACHING
                      L-R Processed Coal  Filter Cakes
                  Sample A

             riOOO Gr.  Dry  Coal)
                Wash
            1/2
     Washed Coal (Wet)

       -1/2	,
     3:1 Sol vent-To-Coal   4:1
          Extractions
   1st
   Hexane
   2nd
   Hexane
            Sulfur
            Out,Gr
0.74
1.10
    Toluene    3.24


TOTAL SULFUR
   RECOVERED:* 5.08
                  Sulfur
                  Out,Gr
                         1st
                         Hexane   ~0.0
                         2nd
                         Hexane
0.50
                        Toluene   3.21
                                  3.71
                                             f"
                                                      Sample  B

                                                 (~600 Gr.  Dry  Coal)

                                                       I

                                                      Wash

                                                        Mashed  Coal Wet

                                                 1/2	
                              3:1  Sol vent-to-Coal 4:1
                                     Extractions
                                                    Sulfur
                                                    Out,Gr
        1st
        Hexane    0.29
2nd
Hexane    0.31
                            Toluene   1.38
                                      1.98
                                    Sulfur
                                    Out.Gr
                     1st
                     Hexane    0.13
2nd
Hexane    0.31
                             Toluene    1.57
                                                                        2.01
Estimated  overall  S  recovery efficiency >90 percent in all cases.
                                    109

-------
bO
W
to
LLJ
                       300°             400°

                         TEMPERATURE,  °C
             Figure 17.  Vapor Pressure of Sulfur
                              110

-------
              TABLE 16.  VAPORIZATION OF ELEMENTAL SULFUR FROM FERRIC SULFATE LEACHED COAL*
Sample
No.
1
2
3
4
5
6
Water
Washed
0
X
X
X
X
x.
Toluene Thermal % w/w Sulfur
Extracted Treatment Sulfate Organic
0-
X
X
0
0
0
0
0
220°C/15mm/17 hrs
Hg
200°C/3 hrs
250*0/2 hrs
350°C/2 hrs
0.34
0.13
0.05
0.03
0.01
0.01
1.18
0.52
0.51
0.92
0.63
0.56
% Weight Loss
During Vaporization
-
-
2.7
^/
1.7
1.2
2.8
200cj TOO mesh x 0 Lower Kittanning Seam coal was extracted twice with 2.5a 1,N Fe2($04)3 solution, total
residence time 22 hrs.  Cursory water wash to remove surface iron sulfate; dried at 110°C under vacuum
to remove moisture and as little volatile elemental sulfur as possible.  The coal was subsequently
split into lOg samples and subjected to the treatment indicated in columns 2 through 4 (X indicates
treatment, 0 indicates no treatment).  The organic sulfur content of the ROM coal was 0.41  +_ 0.04 wt.
percent.

-------
 sulfur (Sample No.  1).   Drying  was  effected  at  relatively  low  temperature
 in an attempt to remove water but  not  vaporize  large  amounts of the  elemen-
 tal sulfur which had been  formed.   Another sample  (No.  2)  was  washed with
 water and toluene  to remove  sulfate and elemental  sulfur, respectively.
 The decrease in "organic"  sulfur between samples  1  and  2 (0.66  percent w/w)
 is a measure of the elemental sulfur which had  remained adsorbed  on  the coal
 after a cursory drying. Another sample, (No. 3),  was further  treated
 under vacuum at 200°C to remove any additional' residual sulfur.  There
 was no further decrease in organic sulfur indicating  that  a maximum  of
 the  elemental sulfur had been removed by the toluene extraction; however
 there was  a  decrease in the sulfate content.

     Temperature Programmed Thermogravimetric Analysis studies  show that
 iron sulfates quantitatively decompose at 450-600°C in an  inert atmosphere
 giving  a residue of iron oxide.   Apparently, this degradation occurs at
 somewhat lower  temperatures under vacuum and in the presence of a reducing
 coal environment.  Note that the coal  weight loss during vacuum treatment was
 only 2.7 percent of which  0.67  percent could be ascribed to elemental sulfur.

     Three samples were contacted with an  inert gas stream Cargon) in
 small ceramic boats inserted into a Burrell tube furnace at the specified
 temperatures and residence times (Nos.  4-6).   It can  be seen that increasing
 the  temperature of vaporization from 200PC to 350°C increases  the quantity of
 sulfur vaporize.  A vaporization temperature of 250-350°C  is indicated for
 use  in process  design.  Note the the weight losses during  vaporization did
 not  exceed 2.8  percent  (including the loss due to volatilization of  sulfur)
 indicating that sulfur  can be relatively selectively  vaporized  from  coal.
 The  method utilized for contacting the coal with the  inert gas  stream
was  rather primitive compared to that which can be obtained in  commercial
 size equipment, so that residence times  should be substantially reduced.

2.5  SOLID-LIQUID SEPARATION UNIT OPERATIONS

    The Meyers Process scheme utilized in  the current program  involved
three solid-liquid separation unit operations:  (!) coal separation  from

                                   112

-------
spent iron sulfate solutions,  (2)  coal separation from spent wash water,
and (3) coal separation from  spent toluene or hexane.   All  three operations
were performed by vacuum  filtration with conventional, 4 liter  size labora-
tory apparatuses.  The size of the system did not lend itself to quantita-
tive evaluation of parametric  effects on these unit operations, nor was there
a guarantee that such evaluations  would be meaningful  to a  scaled-up
operation using commercial equipment; thus, a systematic study  of these
unit operations was  not undertaken.  Qualitatively, filtration  efficiency
in terms of rates and dryness  of the filter cake increased  with decreasing
liquid density, as expected for constant filter cake depth.  Thus, toluene-
coal slurries filtered faster  than water-coal slurries and  the  latter
appeared to filter faster than spent reagent-coal slurries.  Also, while
the toluene content of the filter  cake was  approximately 30 percent of the
dry coal  weight, the water content  of spent reagent  and  water filter
cakes was estimated to be in the 35-40 percent  range.  The effects of coal
top size and slurry concentration  in  the  ranges  of  100 to 14 mesh and 33
to 20 wt. percent, respectively, were not easily discernable.  It appeared
that the 33 wt.  percent slurries of 14 mesh top  size coal filtered at a
higher rate than 20 wt. percent slurries  of 100  mesh top size coal, but an
effect on filtration rate was  not  apparent  when  only one of the above
parameters differed.   Filtration rates and  the moisture  content of the
filter cake appeared unaffected by variations in L-R processing parameter
values.

    The feasibility of conducting  the described  solid-liquid separations on
a scaled-up process by commercially available equipment  and the type of
equipment recommended were determined by  outside vendors.  Bird Machine
Company representatives performed  bench-scale tests  at the TRW site on
actual  process slurries and concluded that  they  can  be separated by avail-
able commercial  equipment at reasonable rates even  at  solids concentrations
as low as 17 percent.  They recommended centrifugal  separation for the
organic solvent  coal  slurries  and  either  centrifugal separation or rotary
drum filtration  for the aqueous coal  slurries depending  on concentration.
Details on these evaluations are included in  a Bird  Machine Company memo-
randum  to TRW reproduced in Appendix  F, Volume 2 of  this report.  Envirotech

                                     113

-------
Corporation, Salt Lake City,  Utah,  arrived at similar conclusions  using
simulated Meyers Process slurries of 100 mesh top size L.K. coal.

2.6  COAL DRYING OPERATION

    The toluene-wet coal from the elemental sulfur recovery operation was
vacuum dried at 100CC for approximately 16 hours (overnight).  This was the
only coal drying procedure used in this program.  The efficiency of this
type of drying operation was determined from the special experimentation
performed on processed coal toluene retention described in Section 2.4.5.1.

    Selection of drying equipment for the process scheme presented in the
process design section of this report was based on information furnished by
outside vendors.

2.7  PRINCIPAL CONCLUSIONS FROM BENCH-SCALE DATA

     In this report section an attempt is made to summarize the principal
conclusions drawn throughout the discussion of the data derived from the
L-R processing of suspendable Lower Kittanning coals by the Meyers Process.
The emphasis in this summary is placed on information and conclusions vital
to the full scale process design presented in Section 5.  Of course, the
most important information derived from the data generated in this program
was the conclusive proof that L-R processing (simultaneous coal Teaching-
reagent regeneration) of L.K. coals up to at least 14 mesh top size was
both feasible and desirable.  The expected advantages of L-R processing  -
combination of two unit operations into one, elimination of the need for
solid-liquid separations during leaching, and higher pyrite leaching rates -
were attained without any detectable adverse effects on the composition  or
physical properties of the processed coal.   Furthermore, these advantages
were augmented by the attained pyrite leaching  rates which, under relatively
mild temperature and pressure conditions  (1100-130°C5 50-150 psig),  were
higher than expected.
                                    114

-------
     The performed experimentation and  resulting data suggested that the
L-R processing scheme may be defined  by a  process train comprised of the
following main unit operations connected in  series:  mixer, L-R reactor,
ambient pressure reactor, coal-reagent  separation, processed coal wash,
coal-water separation, elemental  sulfur recovery, and coal drying.  The
last two unit operations become a single operation when elemental sulfur
is recovered by vaporization.  Elemental sulfur recovery by organic solvent
extraction requires an additional solids-liquid separation operation.
Multiple stage processing may be  desirable or  necessary for certain unit
operations (e.g., L-R reactor or  coal wash).

     In addition to the described process  train, the L-R processing scheme
requires one or more unit operations  in the  reagent recycle loop for sulfate
and trace element recovery.   It is probable  that these unit operations
would best be employed on process slip  streams rather than the entire re-
agent recycle stream.  The  generated  data  were insufficient for the defini-
tion of the sulfate recovery  unit operations.

     In terms of processing function, the  identified unit operations may be
grouped into four process sections:   the reactor section, the wash section,
the elemental sulfur recovery section,  and the sulfate recovery section.
The reactor section - comprised of the  mixer,  L-R reactor, and ambient
pressure reactor - is the core section  of the  Meyers Process scheme for
L-R processing; therefore,  it was the most thoroughly investigated of the
four process sections.

     The three reactors were  operated in a batch mode with respect to coal
slurry, but oxygen was continuously fed to the L-R reactor.  Reactor slurry
composition homogeneity was assumed at  all times  (well-mixed reactors).
Coal pyrite leaching took place in all  three reactors with the bulk of the
pyrite being removed in the L-R reactor, as  intended.  Throughout the
leaching process and regardless of processing  conditions coal pyrite was
oxidized to sulfate sulfur, S , and elemental  sulfur, Sn> in accordance with
                              s
Meyers Process chemistry.  The  SJSn ratio in the process  sulfur  product
                                 s   n
was estimated to be equal to  1.5  regardless of processing conditions.
     f $*? . •'

                                     115

-------
     Pyrite leaching rates in all three unit operations of the reactor
 section were adequately described by the empirical  rate expression
               'L '

where W  is the concentration of pyrite in coal at reaction time t in
weight percent and Y is the ferric ion-to-total iron ratio in the reagent
solution at the same reaction time.  The validity of Equation (3) was
proven from bulk measurements of Y and Wp during the leaching of both 100
and  14 mesh top size L.K. coals under a variety of processing conditions.
Bulk Y measurements were inadequate for use in Equation (3) only when the
total  iron concentration in the reagent was well below one percent (e.g.,
in experiments performed with iron-free starting reagent with the process
iron being derived from the coal).

     In the temperature range investigated, 70° to 130°C, the reaction con-
stant  of  Equation  (3) depended only on temperature and coal particle top size.
Between zero  and 80 percent pyrite removal the temperature dependence of
K. could  be expressed by the Arrhenius-type equation

                 KL =  AL exp (-EL/RT).                      (4)

Beyond approximately 80 percent pyrite removal the K,  value appeared to
be virtually  insensitive to temperature change in the range of 102°-130°C;
obviously, Equation (4) did not apply to this residual pyrite region.
Apparently, a change in the rate occurred when a certain minimum pyrite
concentration was reached in the coal.  It is believed that this apparent
change in mechanism is due to the high ash content of the particular ROM
L.K.  coal used in this program.  Indirect evidence for the validity of
this statement is furnished by the coarse coal data in Figure 21, Section
4  of this report, where ROM and cleaned L.K. coal data are compared.  These
data indicate that a substantial reduction in coarse coal pyrite  leaching
rate occurred when pyrite removal from the ROM coal samples reached 70  per-
cent, but an equivalent drop in rate was not observed with the cleaned
coal.  For process design purposes, rate constant values  specific  to  the
80-95 percent pyrite removal region were estimated from the experimental
rate  data.
                                     116

-------
    For reasons  which could not be delineated by the available data,  two
different AL  values were required in Equation (4) to express the K depen-
dence on temperature during mixer processing and L-R processing.  Possible
explanations  are offered in Sections 2.4.2 and 2.4.2.3.

    For reagent  regeneration during L-R processing, the rate expression
developed in  the previous bench-scale program with synthetic spent reagent
proved valid.  The data, however, suggested that the activation energy of
the regeneration reaction may be lower for actual spent reagent solution
regenerated  in the presence of coal than that computed from synthetic
spent reagent studies.  This apparent catalytic effect could not be quan-
tified from  the  available data.  Qualitative analysis of the data revealed
that neglecting  this catalytic effect on the regeneration rate would  not
significantly influence the engineering analysis of the process.

     In  summary,  the empirical rate expressions derived during the current
and previous  bench-scale programs can be used with confidence for the  de-
sign of  the  reactor section for depyritizing suspendable ROM Lower Kittan-
ning coals  (higher K, values may be applicable to cleaned L.K. coals).
The following rate expressions and rate constants are applicable to each
of the three  unit  operations:
Mixer:

        rL = KL Wp Y2  and KL = AL exp  (-EL/RT)

  with  AL = 3.4 x 105 (hours)"1  (W  J'1  for  100 mesJi x 0 L.K. coal

        AL = 2.7 x 105 (hours)"1  (Wp)"]  for  14 mesh x 0 L.K. coal

        E,  = 11.1 kilocalories per mole pyrite reacted for both
             top size coals

  The mixer is the unit operation where  the  coal slurry is prepared and
  brought  to the desired L-R processing  conditions (temperature and pressure),
  Slurry residence time in this unit depends largely on the time required to

                                    117

-------
 wet the feed coal.   Wetting of  Lower  Kittanning  coals was  accomplished by
 refluxing their freshly mixed slurries  for approximately 30  minutes.   Thus,
 slurry residence time in the mixer  should be  less  than  one hour.

    For a given top  size coal, pyrite  removal  in  this unit  operation depends
 on the square of Y  and, therefore,  on any parameter that affects  Y.   Since
 reagent regeneration or exchange  does not take place in this unit operation,
 pyrite removal is highest when  the  Fe+3-to-pyrite  ratio in the  slurry is
 highest (high Fe+3  starting reagent concentration, low  coal  concentration
 in the slurry).  Typical pyrite removal  values for L.K. coal  were:  20
 percent with 20 wt.  percent slurries  and 5 wt. percent  Fe  reagent; 10 per-
 cent with 33 wt.  percent slurries and 5  wt. percent reagent;  approximately
 one percent with slurries prepared  from iron-free  starting reagent.

L-R Reactor (no°-130°C, 50-150  psig,  14-100 mesh coal top  size):

        r,  = K.  W_Y          where  Y  *  1-Fe   /Fe
         L    L  P

        rR = KRP0   (pe+2)2  wnere  KR = AR exP  (~ER/RT)
  with  AR = 6.7 x 105 Uters/mole-atm-mi mites

  and    ER = 13.2  kilocalories per mole  of Fe+2 oxidized

  The  above  expressions  operate  simultaneously on the process  in the L-R
  unit operation.  The value  of  KL to be  used in the leaching  rate expres-
  sion depends on  temperature and coal  top size up  to approximately 80
  percent  pyrite removal,  on  Wp  for  a short transition period, and it  is
  probably a pure constant with  respect to changes  in processing parameters
  in the 80-95 percent pyrite removal region.  Thus,

       KL = AL exp (-EL/RT) for  wp >_ 1.6 wt.%

 with   AL (14 mesh)  = 7.4 x 105  (Hours)'1 (% Pyrite in Coal)'1

      AL (100 mesh)  = 8.9 x 105 (Hours)'1  (% Pyrite in Coal)'1

      EL  (both  top sizes) = ll.l  kilocalories/mole pyrite removed;

                                    118

-------
 also  KL =   (Wp)  = Wp - 1.1  (Hours)"1 (% Pyrite)"1  for 1.6 >W  >  1.2 wt.%

 and   KL =  0.1  (Hours)"1  (% Pyrite in Coal)"1 for W  _ KL i 0.07 (Hours)"1 (% Pyrite)"1 for 90-1Q2°C

  The available data did not permit the determination of precise KL values
  or the KL temperature dependence, if any, for temperatures lower than
  9Q*C  1n this unit operation,  The value of 0,07 was estimated from
  "settler" data generated at 90"C.  The 0,10 value 1s equal to that de-
  termined for pyrlte leaching from 14 mesh top size L.K. coal at 102°C
  1n the mixer where the reaction appeared to be klnetlcally controlled,
  but it is also equal to the estimated value of the diffusion rate KL
  determined applicable in the 110°-130°C temperature range after 80-85
  percent of pyrite has been removed from coal.

     This unit operation serves  a dual function:   (a) the last few percent
  of coal pyrite is removed  in it and  (b) ferrous  ion is being generated
  which can subsequently be  used to recover the iron and sulfate products
  of the process as ferrous  sulfate, since ferrous sulfate is less soluble
  than ferric sulfate.

     Figure 18 presents typical  process design curves and illustrates the
predominance of the L-R unit operation on the overall performance of the
Meyers Process.  It is noted that under L-R processing, the parametric
effects on the individual rates  and differences in pyrite removal in the
mixer are substantially tempered because of the simultaneous  influence of
                                     119

-------
   80
  70
O
U

Q
Z  50

O

Q
LLJ
to
CO
i
   40
   30
  20
  10
SUSPENDABLE  LOWER KITTANNING COALS



1. 33 WT. % SLURRY, 130°C, 100 PSIG

2. 20 WT. % SLURRY, 120°^ 100 PSIG

3. 33 WT. % SLURRY, 120°C, 100 PSIG
                          1.0        2.0         3.0

                            L - R REACTION TIME, HOURS
                Figure  18.   Typical Process Design Curves
                             4.0
5.0
                                   120

-------
the leaching and regeneration  rates  on process performance in  the  L-R
reactor.  Each curve in  Figure 18 represents both 14 and 100 mesh  top size
coals; thus, the coal particle size  effect for coal  top sizes  up to 14
mesh is undetectable under  L-R processing.   Also, the temperature  effect
on the leaching rate is  less pronounced.   Finally, the slurry  coal con-
centration in the 20-33  wt. percent  range had virtually no effect  on
overall process performance even  though twice as much coal  per unit time
was processed when 33 wt. percent slurries were employed as when 20 wt.
percent slurries were used.  The  effect of pressure (oxygen partial pres-
sure) on overall process performance was also small  under L-R  processing.
The apparent process insensitivity to narrow parametric changes, even
though individual rates  are substantially affected by similar  changes in
the same parameters, is  the result of compensating effects  brought about
when the individual leaching and  regeneration rates operate simultaneously
(independent variables become  interdependent).

     The data in Figure  18 were derived from  the  rate expressions and  rate
constant values presented above.   For  up  to 80  percent  pyrite removal,  data
from Appendix D, Volume  2 were used.   Appendix  D  contains similar data  for
every experiment performed with suspendable coal  in the current program.
Data in this Appendix are tabulated  for up to 90  percent  pyrite removal
but, at least for the particular  L.K.  coal used  in this program, are only
valid up to 80 percent pyrite  removal.  Note, also, that  the data in Appen-
dix D and Figure 18 apply to well  mixed reactors  only  (batch operation).

     The Meyers Process  design information  summarized  in  this  section should
only be used with confidence to predict  pyrite  removal  from ROM, run-of-the-
mine, Lower Kittanning coals.   Limited data  on  cleaned  Lower Kittanning coal
(Section 4), Illinois No. 5 coal,1 and Upper Freeport  seam  coal appear to
indicate that the empirical pyrite leaching  rate expression, Equation (3),
derived from the study of ROM  L.K. coal  may  also be  applicable  to these
coals provided empirical KL values are determined for  each  coal.  That the
KL values for different  coals  could  be substantially  different  is dramati-
cally illustrated in the next  section where  pyrite removal  from L.K. and
Upper Freeport coals is  compared.  It is  estimated that the Upper  Freeport
coal KL is at least an order of magnitude larger than  that  of  L.K. coal.
                                     121

-------
 2.8  L-R PROCESSING OF UPPER  FREEPORT SEAM COAL

      A sample of low pyrite ROM Upper Freeport seam coal  was available
 from previously performed  screening  tests  (pyrite leaching under ambient
 pressure) as part of the Survey Program on the Meyers Process.^  '  The
 coal contained 1.7 wt.  percent  pyrite (0.9 wt.  percent pyritic sulfur)
 and thus it was considered a  good  candidate for testing the effectiveness
 of L-R processing on coal  with  low pyrite  content.   Testing of this coal
 was principally motivated  by  the substantial  reduction in the value of the
 leaching rate constant observed during the L-R processing of ROM L.K.
 coal when its pyrite content  was reduced by approximately 80 percent to
 1.2 wt.  percent (the pyrite concentration  in  the ROM L.K. coal was 7.3 wt.
 percent).

      The available 60 mesh top  size  ROM Upper Freeport coal  was  slurried in
four times its weight 65°C ferric sulfate solution containing  5 wt. percent
iron and  immediately processed in the  identical manner that  L.K.  coal was
processed.  Figure 19 summarizes the data obtained from the  L-R processing
of Upper  Freeport coal.  Note that approximately 82 percent  of the  pyrite
in coal was removed during t  , 1.5 hours slurry residence time in the mixer;
at this point the pyritic sulfur content of the coal was  only  0.16  wt. per-
cent.  An additional 1.5 hours of L-R  processing at 120°C increased the
overall pyrite removal to 93 percent and reduced the pyritic  sulfur content
of the coal to 0.07 wt. percent.  Obviously, pyrite removal  from  Upper
Freeport coal by the Meyers Process appears to be efficient  down  to virtually
zero pyrite.

     In Figure 20 pyrite removal data  from nearly identical  processing  of
Upper Freeport and L.K. coals are being compared.   It  is  obvious  that  pyrite
removal from the Upper Freeport coal  takes  place  at  higher  rates  than
those for L.K. coal.  Since only a single rate experiment was  performed with
the Upper Freeport coal, it was not possible  to  determine if pyrite leaching
from this coal could be expressed by  Equation  (3),  the empirical  leaching
rate expression developed  from  the L.K. investigations,  or if a different
                                    122

-------
    90
    80
    70
o

5
tu
Q£

Qi
D
LL
u
Qi



Q.
    60
    50
   40
   30
   20
                    60 MESH TOP SIZE COAL
                    65 - 120°C

                    —   t
                        m
                          120°C, 100 PSIG
                          	
                                                            'L-R
_L
J.
0.0        0.5
                            1.0          1.5         2.0

                                  REACT ION TIME, HOURS
                                   2.5
                                  3.0
          figure  19.  Upper Freeporc Coal Leaching with 5 Wt.  %  Fe Reagent
                                         123

-------
l\5
•£>
                           SUSPENDABLE COAL LEACHING WITH 5 WT % Fe REAGENT SOLUTION

                                              20 WT % ROM COAL SLURRIES
                                                   1    UPPER FREEPORT COAL (60 MESH TOP SIZE)

                                                   2    L. K. COAL (14 MESH TOP SIZE)
                                                                120°C, 100 PSIG
                                                                	  f,
                                                                      L-R
                                                                    J_
                                             3.0         4.0        5.0

                                                REACTION TIME, HOURS
6.0
7.0
8.0
                       Figure  20.  Pyrite Leaching Rates from  Upper Freeport and L.K. Mine Coals

-------
expression was required.   If,  however,  it  is  assumed that Equation (3)
applies to Upper Freeport  coal,  then  the data in  Figure 20 appear to indi-
cate that the K, value  for the Upper  Freeport coal  is  between one and two
orders of magnitude  higher than  the equivalent KL value for the L.K. coal.
                                       125

-------
                       3.   REAGENT  RECYCLABILITY-TRACE  ELEMENT DATA

     A total of 31 L-R experiments  were  performed  utilizing a single batch
of recycled reagent.  Twenty-nine of  these experiments  are summarized in
Appendices A and B of Volume  2  (Experiments 7  through 11, 13 through 15,
17 through 19,and 23 through  40) and  provided  for  91 hours of L-R pro-
cessing exposure under a variety of operating  conditions as specified in
the Appendices.  An additional  10 hours  of L-R processing time was accumu-
lated on the recycled solution  during special  experimentation.  Recycled
reagent processing exposure over the  101 hours of  L-R processing time
includes 380 hours of settler processing at 90°C and ambient pressure (for
one half of the reagent in  nearly all  experiments), and 46 hours of mixing
and heat-up time prior to L-R processing.   Approximately 70 percent of
recycled reagent exposure time  was  accrued subsequent to addition of 2 weight
percent H^SO. to the reagent.   During the reagent  recycle experiment!'on
nearly 77 kilograms of dry  coal were  processed.  An average of 18 percent
reagent makeup was required to  compensate for  evaporative losses incurred
during processing as well as  reagent  losses associated with water washing
of the leached coal.  Aqueous filtrate from the first stage coal wash was
utilized in the preparation of  make up reagent for the majority of the
experiments.

      Comparison of the data  generated from experiments performed with re-
 cycled and fresh reagent indicate  no  apparent degradation in the activity
 of the reagent with age or quantity  of  coal processed, (e.g., compare data
 from Experiments 23, fresh reagent,  and  31, reagent recycled for 40 hours
 under L-R conditions).  A more conclusive proof of the stability of reagent
 activity during recycling  is the fact that the data generated at the
 beginning and at the end of  the reagent recycle experimentation are equally
 predictable by Equation (3).   It is  believed  that the reagent recyclability
 tests performed in this program were of sufficient time duration and
 spanned the range of potential processing conditions of the Meyers Process
 adequately enough to permit  the conclusion that reagent solution recycl-
 ability is viable.
                                    127

-------
      The  reagent recycle investigation also furnished a means of verifying
 earlier conclusions  (Contract No. 68-02-0647)  that the Meyers Process
 removes,  in addition to pyrite, non-pyritic mineral matter usually present
 in  coal in minor or trace quantities.  This extensive recycling of reagent
 provided  for  the build up of trace elements in the reagent and, therefore,
 furnished the potential of analyzing the reagent for these elements with
 increased accuracy and of attempting a cursory mass balance of each element
 in  conjunction with coal analyses.

      The  majority of the elements selected for analysis either have been
 proven to be  hazardous pollutants or are recognized as being potentially
 hazardous.  Included in these categories are As, Be, Cd, Cr, Cu, Li, Mn,
 Ni,  Pb, V and Zn.  Other elements were selected because of their corrosion
 promoting tendencies during processing and/or combustion or because they
 were previously shown to be amenable to leaching by the Meyers Process.

      Analyses were performed on starting and processed samples of both
 coal  and  iron sulfate reagent.  The detailed description and evaluation of
 the  analytical procedures used can be found in Reference 2.  Briefly, all
 coal  and  reagent samples were analyzed by atomic absorption (AA) spectro-
 scopy for Be,  Ca, Cd, Cr, K, Li, Mg, Mn, Na, Ni, Pb, V and Zn.  Arsenic
 was  also  determined utilizing a special spectrophotometric procedure.  The
 chloride  content of the leach solution was determined using a specially
 modified  silver chloride turbidimetric procedure.

      An NBS certified coal (Standard Reference Coal No. 1632) was analyzed
 concurrently with the selected Lower Kittanning samples as a check on the
 accuracy  of the analytical system.  The results are presented in Table 17.
A comparison  of NBS certified values with analyzed values  shows  good
to excellent agreement for As, Be, Cr, Cu, Mn, Pb, V and Zn as was the case
with coals tested earlier and reported in Reference 2.
                                     128

-------
                             TABLE 17.  TRACE ELEMENT ANALYSIS ACCURACY VERIFICATION
                                        (Concentrations are Shown as ppm in Dry Coal)
ro
10
        NBS
                    As
Be
Cd
Cr
Ca    Li     Mn     Ni
Pb
Zn
     Certified   5.9 +_ 0.6   (1.5)   0.19 + 0.03  20.2 + 0.5  18+2   -   40+3  15 + 1  30 + 9  35 + 3  37 +_
     Analysis                             ~"            ~         ~            _       _       _
     Analysis of  NBS  Sample  in  Duplicate:
                    6.1         2          2
                    5.4         2          4
                        16
                        16
                       17    12      33     25
                       17    13      33     28
        Zinc  determination  performed  on  a  plasma ashed sample.
      Note:   Elements in parenthesis  are  not  certified and are for reference only.
                                       28      24      42*
                                       24      30      35*

-------
     One sample of starting coal and one sample of processed Lower Kittanning
coal were subjected to duplicate trace element analyses.  The processed
coal sample was derived from Experiment 31.   In this experiment 14 mesh
top size coal underwent 3 hours L-R processing at 120°C and 20 hours
"settler" processing at 90°C with reagent solution which had already been
recycled for approximately 40 hours at 120°C and for nearly 200 hours at
90°C.  One starting reagent sample and two recycled reagent samples were
also analyzed for trace element content.  One of the recycled reagent
samples was drawn after 98 hours of L-R processing (87 hours at 120°C and
11 hours at 110°C).  The second sample was drawn at the conclusion of the
reagent recycle investigations (101 hours L-R processing); it differed
from the first by 3 hours L-R processing at 130°C.  The second sample was
analyzed principally to determine if trace element leaching from coal was
enhanced at 130°C.  The data from these analyses are summarized in Table 18.

     In general, the data in Table 18 show that the decrease in coal trace
element concentrations during processing corresponds directly to trace
element build up in the recycled iron sulfate leach solution.  Elements
which apparently are not subject to leaching by the Meyers Process include
Be, Cd, K, Li, Na, Pb, and V.  Elements appearing to be most susceptible
to leaching are As, Ca, Cu, Mn and Zn, with at least 50 percent removal
being observed for each.  It is also noted that the Cr and Ni concentrations
of both coal and reagent solution increased during processing; apparently,
this was due to a slight corrosion of the stainless steel apparatus used
in these experiments.  A similar occurrence is observed with Mg which can
not be explained on the basis of available data.  It is assumed to be the
result of errors in analyses and probably in the analysis of the starting
coal which was not performed in duplicate.

     An attempt was made at mass balancing the trace elements which exhib-
ited the highest percent removal from coal (As, Cu, Mn, and Zn).  The mass
balance was performed through the detailed bookkeeping of  liquids and solids
entering and leaving the process during each experiment performed with  re-
cycled  reagent.   The quantities of trace elements removed  from the coal
processed in each experiment were computed from the values  listed  in  column 7

                                    130

-------
             TABLE  18.   COAL  AND  REAGENT  TRACE AND MINOR ELEMENTS
Element
As
Be
Ca
Cd
Cl
Cr
Cu
K
Li
Mg
Mn
Na
Ni
Pb
V
Zn
Trace Element Concentrations in PPM
Starting
Coal*
13 16
- 2 2
1590
3 1
NA NA
39 36
27 27
4030 -
37 37
946 -
26 25
637 -
48 43
17 18
55 50
30 26
Processed
Coal*
5 5
2 2
7.6 7.6
1 1
NA NA
55 59
7 7
3170 3160
42 37
4171 4414
8 8
734 753
50 50
22 25
50 55
13 14
Starting
Reagent
1
0.06
5
0.07
37
6
0.09
0.6
NA
6
28
4
6
0.44
NA
0.78
Recycled Reagent
(98 hrs. at
110-120°C)
19 (31)**
0.9
318
0.14
175
30
64 (66)
1.0
NA
238
30 (88)
36
154
0.65
NA
52 (46)
Recycled Reagent
(101 hrs. at
110-130"C)
19 (27)
0.6
293
0.14
no
30
58 (57)
0.7
NA
224
32 (76)
20
136
0.58
NA
52 (40)
Estimated
Element
Removal
(Mo/Kg Coal)
10 + 2
0
1582 + 159
1 +0.3
ND
ND
20 + 3
865 + 360
0
ND
18 + 3
0
ND
0
0
14 + 4
**
   Duplicate analysis performed for most elements.

   Values in parentheses  represent the expected trace element concentrations In the
   reagent based on coal  analyses and on the quantity of coal processed  with the
   same batch of raagent  (adjusted for make-up).

NA -  Not analyzed.
ND -  Not determinable from available data.
                                          131

-------
 of Table  18.   In turn, these values were computed from the analyses per-
 formed  on the  starting coal and the processed coal from Experiment 31
 (columns  2 and 3, Table 18).  The assumption was made that the quantity of
 trace element  removal per unit weight of coal processed was the same for
 all  experiments and equal to that determined for Experiment 31; that is,
 L-R reaction time beyond one hour of processing and changes in slurry
 concentration, pressure, and temperature had no effect on trace element
 removal  (note, however, that only a limited number of experiments were
 performed at temperatures,  pressures, and slurry concentrations different
 from those of  Experiment 31).  The trace element bookkeeping process began
 with the  batch of pure reagent solution used as the feed to the reagent
 recycle tests  and ended 30  experiments later when the reagent recycle
 investigations were completed.  Trace element build up in the reagent
 during  each experiment was  computed by adding to the known reagent com-
 position  at the start of the experiment the quantity of trace elements
 leached from the coal during the experiment; the new reagent solution
 composition was appropriately adjusted for trace element and liquid re-
 movals  (losses) during transfers and coal wash and for additions through
 reagent make up.

     As a result of the above bookkeeping operation, it was possible to
 compute the expected concentration of the selected trace elements in the
 reagent solution as a function of reagent recycle time.  In Table 18, com-
 puted trace element concentrations (values in parentheses, columns 5 and 6)
 are  compared with the corresponding trace element concentrations obtained
 from direct trace element analyses of the recycled reagent solutions.
 Reasonably good agreement is seen for As, Cu, and Zn, but not for Mn.  The
 reason for the discrepancy  between computed and analyzed Mn concentrations
 could not be discerned from the available data; it is believed to be due
 to the analyzed high Mn concentration in the starting reagent solution
which,  if erroneously high as believed to be, would make the computed con-
centrations higher than should actually be.  The agreement between computed
and analyzed  trace element concentrations for three out of four mass
balanced trace elements adds substantial credence to the values  in column  7
of Table 18.   These latter values are not only important to the design  of

                                    132

-------
a product recovery system  for  the  Meyers  Process,  but  also to the evalu-
ation of the environmental  benefits  derived  from the utilization of the
Meyers Process for the desulfurization  of a  particular coal.

     Comparison of the data in columns  5  and 6,  Table  18, appears to in-
dicate that there was no significant increase in the leaching of trace
elements when L-R processing temperature  was increased from  120°C to 130°C.
Both, the analyzed and computed trace element concentrations lead to the
same conclusion.

     Calcium, one of the minor elements in coal, was also substantially
leached from coal during processing.  Because of the high quantity of
calcium leached from the coal, relative to trace elements, and because of
its low solubility in the  reagent, this element  was periodically removed
from the recycling reagent solution  as  a  precipitate.   Approximately one
gram of white crystalline  precipitate was recovered from the reagent
solution per kilogram of coal  processed.   Emission spectrographic analyses
of six precipitates revealed that  Ca was  the only  cation present in the
precipitates in larger than trace  quantities. The anion of  the precipitates
was not determined, but in all probability it was  sulfate.   One precipitate
was analyzed in duplicate  for  trace  cation content; the data are presented
in Table 19.
    TABLE 19.   TRACE CATION CONTENT OF CALCIUM PRECIPITATES RECOVERED
               FROM SPENT REAGENT
 Sample
     Trace Cations  Present  in  the  Calcium  Precipitate, ppm
As   Be   Cd   Cr    Cu    Li    Mg    Mn    Ni    Pd   Si   Sr   V   Zn
   A

   B*
<1 <0,5    5   10    8    <1     5    12    67    7   10   50  <1    9

<1 <0.5    8   14    6    <1     5    12    67    <1   10   30  <1 <0.2
 r*
 Sample  B  is  a  duplicate of A
                                     133

-------
     Attempts to mass balance Ca using the procedure followed for trace
element mass balancing described earlier were not very successful.  Less
than 25 percent of the Ca removed from coal, estimated from before and
after processing coal analyses, could be accounted for by the recovered
precipitates (assumed to be CaSO.'ZHpO) and by the build up of dissolved
Ca in the reagent solution.  This was not completely unexpected since a
small amount of precipitate could have deposited on the coal during pro-
cessing and have been leached out and not recovered during coal  wash (note
that the quantities involved are 3-4 grams of precipitate per kilogram of
coal  subjected to washing).
                                   134

-------
                       4.  PROCESSING OF COARSE COAL

      Coarse coal  can easily be separated from the leaching solution thus
 L-R processing (simultaneous coal Teaching-reagent regeneration) is not
 necessarily advantageous to coarse coal unless it improves substantially
 depyritization rates over those obtained during ambient pressure processing.
 The substantial increases of pyritic sulfur leaching rates with increasing
 temperature observed with suspendable coal slurries suggested that coarse
 coal  pyrite removal rates may indeed improve substantially under L-R pro-
 cessing.   However, a brief experimentation with 3/8 inch x 0 L.K. coal
 revealed  that L-R processing of coarse coal results in only marginal  gains
 in rate over those attainable at 102°C under ambient pressure processing
 (separate Teaching-regeneration) as the data in Section 4.1 indicate.

     In general, L-R processing of coarse coal was similar  to that described
for suspendable coal.  The major difference being that coarse coal was pro-
cessed as a semi-fluidized bed while suspendable coal was processed as a
well mixed, partially circulating slurry.   In order to accommodate this
difference, the existing  suspendable coal processing reactor was modified
by the addition of a reinforced wire mesh basket and lid to support and
confine the coal during  reagent circulation.  Two additional thermocouples
were installed to monitor internal coal  bed temperatures.   The coal bed
density was approximately 0.7 kg/liter.  Also, the reagent  circulation
rate was decreased to 3  liters per minute  (as determined by indirect
measurement).  Procedural changes were  as follows:   (1) slurry was charged
into the reactor through  the reactor top flange opening, (2) only reagent
composition was monitored during L-R processing since the circulated re-
agent contained less than 1.5 percent solids  (these solids  being non-
representative of the whole coal), and  (3)  reactor drainage consisted of
reagent removal through  the reactor bottom while the retaining basket and
coal were removed through the reactor top.

     All additional processing (water washing, elemental sulfur extraction
and coal drying) was carried out in a manner identical to that empToyed
during suspendabTe coaT processing.

                                    T35

-------
     A major difficulty encountered during coarse coal processing was main-
tenance of uniform reagent flow throughout the coal bed.  To ensure uniform
flow, the retaining basket sides were removed so that coal-reactor wall
contact prevented channeling around the basket.  Additionally, process coal
weights were reduced from the planned 4 to 5 Kg range to a 1.5 to 3 Kg
range.  Verification of uniform flow was made by positioning thermocouples
within the coal bed and evaluating their response to temperature changes
in  the circulating reagent  (slow response was taken to indicate regions of
low flow).

4.1  L-R  PROCESSING DATA  FOR  COARSE  COAL

      Three coarse coal leaching experiments were performed under pro-
cessing conditions using acidified reagent.  These consisted of two
experiments of 8 hours each and a single experiment of 2 hours.  Settler
processing was not employed during any of these experiments.  One 8 hour
experiment (Experiment 46) consisted of processing approximately 3200  grams
of  coal using 2 weights of reagent per weight of coal while the other
 (Experiment 47) consisted of  processing nearly 1900 grams of coal using 4
weights of reagent per weight of coal.  In the 2 hour experiment (Experi-
ment 48), approximately 1700  grams of coal were processed using 4 weights
of  reagent per weight of coal.  Processed coal analyses for these experi-
ments are presented in Table 20.  The analyses are seen to be internally
consistent, although Experiment 46 results seem to indicate slight iron
oxide deposition (probably due to slightly inadequate reagent acidification),

     All three experiments resulted in rather low pyrite removals.  Pyrite
removals during mixing (tm)', calculated on the basis of ferrous iron pro-
duction, were 6, 15 and 14 percent for Experiments 46, 47 and 48, respec-
tively.  The low percentage pyritic sulfur removal observed during t   in
                                                                    m
Experiment 46 was due to a low starting reagent Y value (0.65) and the
high coal-to-reagent ratio used in this experiment (0.5); the other two
experiments were performed at a coal-to-reagent ratio of  0.25 and at
starting Ys >0.9.
                                   136

-------
                 TABLE 20.  L-R PROCESSING OF 3/8 INCH x 0 L.K. COAL IN 5 WT. % FE REAGENT*
                            AT 120eC AND 100 PSIG (85 PSI 02)

Experiment
No.
Starting
Coal
46
47
48
Coal Com
Ash
28.91
+ .52
31.16
29.40
30.27
Heat
Content
Btu/Lb
10,448
± 106
10,246
10,638
10,465
osition Wt. % (Except Heat Content), Dry
Total
Sulfur
St
4.28
+.14
2.86
2.82
3.16
Pyri ti c
Sulfur
SP
3.56
+_.18
2.11
2.27
2.50
sulfate
Sulfur
Ss
0.35
+ .03
0.25
0.22
0.14
Organic
Sulfur
So
0.37
+.03
0.50
0.33
0.52
Iron,
Fe
3.68
±-13
3.21
2.67
2.84
% Sp Removal
Based on
Analyzed
SP

41
36
30
Based on
Corrected
SP
-
37
38
26
GO
       The reagent was acidified with 2 Wt.  %

-------
      Comparable 102°C data for 3/8 inch x 0 L.K. coal are not available.1
 However,  102°C data are available for the 1/4 inch x 0 L.K. coal used in
 the  previous bench-scale program.  These data are presented in Table 21.
 This 1/4  inch x 0 L.K. coal had a starting pyritic sulfur content of 3.48
 percent and attained 13 percent pyrite removal during mixing and heat up.
 High reagent Y values were maintained by frequent reagent exchange.

    TABLE  21.  PYRITIC SULFUR REMOVAL FROM 1/4 INCH x 0 L.K. COAL AT 102°C
Process Time (hrs)
2
4
6
8
10
Average Y Value
'vO.S
^0.9
M).9
^0.9
*J0.9
Pyrite Removal, Percent
32
46
55
62
65
     A comparison of the 3/8 inch x 0 L.K. coal data with the 1/4 inch x 0
 L.K. coal data indicates that the average pyritic sulfur removal rate
 during the first 2 hours of processing was slightly lower for the 3/8 inch
 x 0 coal than for the 1/4 inch x 0 coal.  Average reagent Y values were
 approximately equal for both experiments during this time.  Between 2 and
 8 hours of processing, however, the average pyritic sulfur removal rates
 are seen to be substantially higher for the 1/4 inch x 0 L.K. coal (102eC
 processing) than for the 3/8 inch x 0 L.K. coal (120°C processing) despite
 the lower pyrite content of the 1/4 inch x 0 coal (reagent Y values being
 nearly equal for both experiments during this time).  Thus, the pyrite
 leaching rate constant of 3/8 inch x 0 L.K. coal is lower at 120°C than
 the rate constant of 1/4 inch x 0 L.K. coal at 102°C.  This seems to in-
 dicate that pyrite removal rates quickly become diffusion limited during
 processing of coarse coals.

     While the benefits of elevated temperature processing of run-of-mine
coarse coal  appear to be minimal, it will be shown in the following sections
that 102°C processing data indicate potential  for application of  the  L-R
 processing mode to specific coarse coal  size  fractions.
                                   138

-------
4.2-  CLEAN COAL GRAVITY FRACTION PROCESSING
     The relative rates of desulfurization  of  run-of-mine fine L.K. coal
(100 x 0 mesh) and coarse L.K. coal  (1/4  inch  x 0) are compared in Figure
21   along with  the  pyrite  removal rate curves of narrow size cuts  of  clean and
run-of-mine coarse L.K. coal,  illustrating  the enhancement of rate effected
by cleaning coal.  These data  on coarse coal were generated with  100-500
grams 1/4  inch x 0 Lower Kittanning  coal  (or size fractions therefrom)
slurried in iron sulfate solution (5 wt.  %  Fe, 15-20 wt. % solids).  The ROM
8 x 14 mesh coal fraction is seen to have lower pyrite removal than 1/4
inch x 0 coal  for all  times prior to 40 hours  leaching, at which time both
their pyrite contents  and pyrite removal  rates are seen to become nearly
equal.  Upon cleaning  the 8 x  14 mesh size  fraction  (float-sink removal of
20  percent of  the total coal), however, its pyrite removal rate closely
approximates that of  1/4 inch  x  0 coal during  the first 9 hours of process-
ing.  After the  initial 9 hours  of processing, pyrite  removal rates of the
cleaned 8  x 14 mesh coal are seen to approach  that of  a much finer mesh
coal.  Coal analyses  from these  experiments are presented in Table 22.
The results show that:   (1)  clean coal reacts  substantially  faster than
ROM (raw)  coal,  (2) coarse coal  (in  this  case, a cleaned narrow size
fraction)  can  be desulfurized  to near zero  pyrite  (0.09 percent pyritic
sulfur), and (3) the  desulfurized clean coal  has 0.92  percent total sulfur
at  a 14736 Btu/lb  heat content,  which is  very  close  to the  Federal Stan-
dards for  New  Stationary Sources (0.89 percent sulfur  at this heat content).

TABLE 22.  CHEMICAL REMOVAL  OF PYRITIC SULFUR FROM CLEANED  8 x 14 MESH
           LOWER KITTANNING  COAL AT 102°C
Coal
Whole Coal
Sink (1.75) (20.4%)
Float (1.75) (79.6%)
Float - 48 hr leach
Whole - 48 hr leach
•^••••^^^^^^^^^^^^^^^^^^^^^•H
% W/W SULFUR
Total
4.43
13.7
2.02
0.92
1.39
—-n—H^^^^^^^M
Pyritic
3.37
12.4
1.06
0.09
0.57
•HBl^^MM^H
r Sulfate
0.43
1.22
0.23
0.03
0.04
•^•^^•••"""•i
Organic
0.68
0.04
0.73
0.80
0.78
••••••••••
Ash
20.11
71.6
6.92
5.38
14.88
•••^^•^••i
BTU
12066
	
13899
14736
13257
mmm^mmm
                                     139

-------
              8   12   16  20   24   28   32  36   40   44   48
                    REACTION TIME, HOURS
Figure 21.   Pyritic Sulfur Removal from ROM and Cleaned L.K.  Coal
            at  102°C
                               140

-------
4.3  PRELIMINARY DATA ON COARSE COAL  PROCESSING BY SIZE FRACTION

     Typical Sp removal values  from 1/4  inch x 0  coal are presented in
Table 23 along with the ratio of  the  effective leaching rate constants
measured for -100 mesh and  1/4  inch x 0  Lower Kittanning coal with iden-
tical reagent Y and coal pyrite values.   It is interesting to note that
this ratio does not remain  constant with extent of pyrite removal; thus,
the leaching rate constants of  various top sizes  of the same coal cannot
be related by a single proportionality constant.  This observation can be
attributed to the fact that a given top  size coal is comprised of a wide
range of coal particle sizes with each particle size being associated with
a different pyrite leaching rate  constant.

TABLE 23.  PYRITE LEACHING  OF 1/4 INCH x 0 LOWER  KITTANNING COAL AT 102°C
Reaction
'ime, Hours
4
6
24
48
Y
Avg. Value
^0.8
^0.9
^0.8
^0.9
% Sp
Removal
50
57
75
81
i
KL 100 Mesh x 0 KL 1/2 Inch x 0
2.7
3.0
6.0
9.6
     A sample of  1/4  inch  x  0 Lower Kittanning coal was riffled into sev-
eral narrow size  fractions.   Pyrite removal data  for  100  x  0, 14 x 28.
and 4 x 8 mesh fractions are shown in Table 24.   The  pyritic sulfur removal
rates differed substantially, but  in  all  cases high  pyritic sulfur re-
movals were obtained.  While these size  fractions are all observed to
react at acceptable pyrite leaching rates  (from a commercial processing
standpoint) during 102°C processing, the 14 x 28  mesh and 100 mesh x 0
fractions are seen to react  at significantly higher rates than the 4 x 8
mesh fraction.  These pyrite removal rates may be taken as  a relative
measure of the fraction of readily accessible pyrite  which  is associated
                                    141

-------
             TABLE 24.   PRELIMINARY DATA ON THE  CHEMICAL  REMOVAL OF PYRITIC SULFUR FROM SIZE-FRACTIONS
                        OF LOWER KITTANNING COAL AT  102°C
Coal Fraction
Mesh
1/4"xO


4x8*


14 x 28*




i
TOO x 0*




wu
100


25


20




11




Run No.
_
1&2
354
_
2
3
_
2
1
1
2
••
2
2
1
1
Time
Hours
0*
24
48
0**
48
120
0**
12
24
48
72
0**
6.5
12
24
48
% w/w Sulfur Forms
Total
4.45
1.68
1.50
5.06
2.03
1.53
3.51
1.24
1.06
1.08
1.08
4.62
1.50
1.33
l.lc
1.'.3
Pyntic
3.48
0.93
0.66
3.95
1.29
0.57
2.46
0.53
0.30
0.30
0.18
2.62
0.25
0.15
0.05
0.05
Sulfate
0.50
0.12
0.14
0.43
0.09
0.14
0.40
0.07
0.06
0.08
0.10
1.02
0.24
0.23
0.25
0.25
Organic
0.47
0.63
0.70
0.68
0.66
0.82
0.65
0.64
0.70
0.70
0.86
0.98
1.01
1.00
0.89
0.93
Ash
20.84
17.02
15.54
25.44
21.86
20.30
16.56
11.82
9.13
13.86
12.30
20.79
13.66
13.00
12.40
12.55
titu
11822
12728
13023
10927
11970
12032
12643
13693
14129**
13258**
13533
'11543
13285
13321
13280
13178
t> Pyrite
Removal
0
73
81
0
67
86
0
78
* 88
' 88
93
0
90
9&
98
98
ro
       *  Size cuts of 1/4 inch x 0 coal
       ** Starting coal
       ***Sampling problem is suspected here
                               V

-------
with each size fraction.  Thus,  it  is  seen  that  the  finer cuts of coarse
coal (those containing pyrite. which is largely of the accessible type)
would probably be receptive  to elevated temperature  processing and,
perhaps, benefit could be realized  by  processing these  coarse coal frac-
tions in an L-R mode with the  coarser  cuts  being processed at 102°C under
ambient pressure.
                                     143

-------
                            5.   PROCESS ENGINEERING

     Process design  studies for chemical  removal  of pyritic  sulfur from
coal have  indicated  that the process may be laid  out using a number of
alternative processing methods.  Some of the variations which have been
tested and considered in preliminary engineering  designs include the
following:

     t  Air vs oxygen  for  regeneration
     •  Coal top sizes  from 1/4 inch to 100 mesh
     •  Leaching and regeneration temperatures up to  265°F (130°C)
     •  Leaching and regeneration in the  same vessel  and  in  separate
        vessels
     •  Removal of generated elemental sulfur by vaporization or
        solvent extraction

All of the above conditions are effective and their utilization involves
economic trade-offs.   Because  processing  steps and equipment needed for
removing sulfur from fine  or suspendable  coal sizes  (up to about 8 mesh top
size) are  significantly different from those needed for coarser coal,  they
are separately described, respectively, in  Sections 5.1 and 5.2.  A summary
of projected process economics for commercialization  of both fine and
coarse coal processes  is given in Section 5.3.

5.1  SUSPENDABLE COAL  PROCESSING

     Suspendable coal  is coal  of a small  enough particle  size that it  may
be processed as a substantially uniform slurry with moderate mixing energy.
Although no sharp top  size specification  can be given,  it appears that coals
with top sizes up to about 8 mesh may be  classed as suspendable.  Bench-
scale experiments were conducted using 14 mesh and 100 mesh  top size coals
as representative of the suspendable type.   Either of these  sizes are often
referred to as fine coal to differentiate them from the coarse coal described
in Section 5.2.
                                    145

-------
      A conceptual  full  scale process  design for the chemical removal of
 pyriti-c sulfur  from fine coal  is described in three sub sections.  Section
 5.1.1  gives the design  basis which relies heavily on the bench-scale ex-
 perimental data, but also  incorporates  information provided by equipment
 vendors and data obtained  from the literature.  Section 5.1.2  presents the
 baseline design and equipment  list with capital and operating cost estimates.
 Section 5.1.3 summarizes the major trade-offs examined in arriving at the
 baseline design and shows  some of the cost and operating sensitivities of
 the  process.
 5.1.1   Design Basis for Suspendable Coal

      Processing coal to remove pyritic sulfur using aqueous iron sulfate
 involves four major process sections, each containing several unit operations.

      The reactor section which includes mixing and solution regeneration has
 three  main process requirements which are:
     •   Providing mixing and wetting of ground coal with the aqueous
         ferric sulfate  leach solution and raising the slurry to the
        operating temperature  and pressure.

     t  Providing the residence time and reaction conditions which
        remove a nominal 95 percent of  the pyrite  orginally  con-
        tained in feed  coal.

     •  Providing the residence time and reaction conditions which
        regenerate the ferric  sulfate solution from the spent iron
        sulfate leach solution.

     The washing section which includes several stages of coal washing and
coal  dewatering has two main process requirements which are:

     •  Providing for contact of the leach solution-wet coal with a
        minimum quantity of wash water to remove water soluble iron
        sulfates.
                                    146

-------
     •  Providing  for  separation of coal  from the  leach  solution
        and the wash water.

     The sulfur removal  section which removes both elemental  sulfur and
excess water  from  the  product coal  has four main process requirements which
are:

     •  Providing  conditions such as heat or solvent  contact  to
        remove elemental  sulfur from the  processed coal.

     t  Providing  the  thermal  environment necessary to reduce the
        moisture level of the coal  to the desired  value.

     •  Providing  the  means  for recovery  of the  elemental sulfur
        for subsequent marketing, storage or disposal.

     •  Providing  for  separation of coal  product from solvent,
        if used.

     The sulfate removal  section which removes excess iron sulfate from the
recycle leach solution has four main process requirements which are:

     t  Providing  for  the removal of iron sulfate  from the aqueous
        spent leach solution by crystallization  and/or neutralization.

     •  Providing  for  the recovery of wash water from the wash section
        effluents.

     •  Providing  for  maintaining the correct acid level  by neutral-
        izing excess acid if required.

     •  Providing  for  separation of the by-product iron  sulfate and
        neutralization product from the recycle  streams.

     Specific information and  data  for the steps or operations which are
important to  the process  design are presented in the  following paragraphs.
These data rely heavily on the information given in Section 2.7 but also
include qualitative observations made during the bench-scale  experimental
efforts.
                                    147

-------
 Mixing -  The  present  bench-scale effort has demonstrated that there  is
 a more critical aspect to the mixing operation than simply surface wetting
 the  particles and suspending them in the leach solution.  Preparing the
 slurry can be readily accomplished with mixing times of 15 minutes or less,
 but  it was found that severe foaming of the slurry will occur when it is
 pressurized and raised in temperature.  Based on the experience described in
 Section 2.4.1, the mixing time for a high rank, high ash, dry coal should be
 between 30 and 60 minutes at the normal boiling point of the solution if
 subsequent foaming is to be avoided.  Lesser times may be possible with
 moist  or  low rank coal.  The quantity of foam produced seems to decrease
 with increasing coal particle size and to decrease with lower solids content
 in the slurry.  These are secondary parameters which should not be considered
 of major  importance in the process design.

 Leach  Reaction - The net overall reaction between pyrite and the ferric
 sulfate leach solution is represented by:

     FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 - - 10.2 FeS04 + 4.8 H2$04 + 0.8S  (15)

     AH = -55 Kcal/g-mole Fe$2 = -0.10 MM btu/lb-mole Fe$2 reacted

 The  reaction rate was found to have a second order dependence on both the
 fraction of pyrite (or pyritic sulfur) in the coal and the fraction of the
 total  iron in the leach solution which is in the ferric ion form.  The leach
 rate at temperatures of interest is represented as follows :
where         [Wp] = wt% pyrite in dry coal at time t.
              [Y]  = fraction of iron as ferric  ion at  time  t.  and
               K,   = leach rate constant (a function  of temperature
                L    and Wp).

KL is independent of total iron concentration at least in the immediate
vicinity of 3 percent to 5 percent total iron.   Physical  considerations such
as increased solution density and viscosity and  the  limited  solubility of
                                    148

-------
ferrous  sulfate  in  the  ferric sulfate solution  become  increasingly important
to the design of the pyrite leacher when total  iron  concentration approaches
10 percent.

     The leach rate  constant  as a function of temperature can be adequately
represented by
            KL =  AL  x exp (-EL/RT)

where       E, = 11,100 cal/mole,
            R  = 1.987 cal/ mole - °K,
            T  = temperature in °K, and
            AL = a function of size, temperature and Wp-

     For 14 mesh top size coal at atmospheric pressure and temperatures be-
tween 70°C and the solution boiling point (about 102°C), the value of A. is
2.7 x 10  for all values of Wp.  At temperatures between 110°C  and 130°C
under oxygen and steam pressure up to about 150 psig, the value of A. is
7.4 x 10  when Wp is  large  (above 1.6) and 2.7 x 10  when Wp is small  (below
1.2).  Additional refinement of the data in the transition region would be
desirable, but an adequate  representation of the data can be obtained  by  a
linear decrease  in AL from  7.4 x  105  for Wp =  1.6 to  the AL value of 2.9  x 10
for Wp = 1.2.

     For 100 mesh top size coal,  the ranges of applicability for  AL  appear
to be the same,  but its value  is about 25 percent higher.  Thus at low
temperature or low Wp, the  value of AL is 3.4  x 105 while at higher temper-
ature and high Wp the value is 8.9 x 10 .

     These leach rate constants have  been defined only  for the high pyrite,
high ash Lower Kittanning coal used in the bench-scale  programs.  They
should not be applied to  other coals  or coal seams.   The Lower Kittanning
coals investigated in the bench-scale programs had starting values for this
coal of Wp between 6 and  8.  The  transition to a lower  rate constant,  thus,
occurs at about  75 percent  to  80  percent pyrite removal  (Wp M.6).  Based
on a single test of a coal  with a lower ash and low  starting Wp  (Upper
Freeport seam, Wp = 1.7), a high  leach rate was found at least to 80 per-
cent or 90 percent removal  (i.e., Wp  about 0.2).
                                    149

-------
 Regeneration - The leach reaction produces both ferrous-suIfate and
 sulfuric acid which must be processed for continuous recycle operation.
 For each mole of pyrite reacted 9.6 moles of ferrous sulfate  must  be  re-
 generated to maintain the acid at a constant level.  This gives by-products
 for disposal of 0.2 moles of  Fe2(S04)3,  0.6 moles  of FeS04 and 0.8 moles
 of elemental sulfur.  Alternately, regeneration of 9.2 moles of ferrous
 sulfate can be considered if some acid is neutralized to give by-products
 of 1.0 mole of FeSO/,, 0.2 moles of I-LSO,, and 0.8 moles of elemental sulfur.
                   4                c.  4
 The choice of the extent of regeneration should be made on the basis of the
 by-product  preference and economics within process design constraints.

     The regeneration reaction is:

     1.0 FeS04 + 0.5 H2S04 + 0.25 02 —* 0.5 Fe£(S04)3 + 0.5 H20   (18)

     AH = -18.6 Kcal/g-mole FeS04 = -.0335 MM btu/lb-mole FeS04

     If hydrolysis of a portion of the ferric sulfate to iron oxide should
 occur as

     Fe2(S04)3 + 3 H20  	* Fe203 + 3 H2S04,         (19)

 then additional acid neutralization or regeneration of ferrous ion would be
 required to remove the acidity produced from the hydrolysis reaction.  The
 extent of hydrolysis at temperatures below 250°F appears to be small,  but
 at higher temperatures there is some evidence of precipitation of ferric
oxide and possibly a low hydrate or anhydrous ferrous sulfate.  The hy-
drolysis products and/or precipitates^formed at 265°F were found to redis-
solve slowly in ambient temperature spent leach solution and  do not remain
as permanent products.   No data was obtained above 265°F.

     The regeneration rate was found  to be second order in the molar con-
centration of ferrous ion over the range of ferrous concentration  from 100
percent to less than 1  percent of the total iron.  The  rate  is
                       +2
            rR =  "d[Fdt 3  = KR [pe+2]2 C°2]'                       (20)

                                    150

-------
              +2
 where      [Fe  ] = concentration of ferrous ion, mole/liter,
           [02]    = oxygen partial pressure, atm, and
           KR     = 1.832 liters/mole-atm-hour at 248°F.

 Over the  range  of temperatures studied (212°F to 265°F) the rate  constant
 was found to vary exponentially with temperature as

           KR =  40.2 x 106 exp (-13,200/RT),                         (2])
 which  gives

                     Temp   °F (°C)       ^R  I1ters/m°le~at;ni-noijr
                       212    (100)               0.74
                       230    (110)               1.18
                       248    (120)               1.83
                       266    (130)               2.79

 The  ferric  sulfate  regeneration rates were obtained under conditions where
 oxygen  in the  form  of minute air or oxygen bubbles was  dispersed through-
 out  the ferrous  sulfate  solution.   Thus,  all  of the solution was continually
 saturated with oxygen at the partial  pressure of oxygen present in the
 regeneration gas.   At bench-scale, the minute bubbles were formed by pumping
 a portion of the liquid  in  turbulent flow (NR  >3000) through a pipe whose
 length was  50  or more times its diameter.  Gas containing oxygen was added
 to the  liquid  in an amount ranging from less than 1 percent to  greater than
 10 percent  by  volume at  flow conditions.   The method is very similar to
 aeration equipment  used  to  reduce the biological or chemical  oxygen demand
 of chemical plant effluent  streams, except that ferric  sulfate  regeneration
 is conducted at  higher temperatures.

 Separation  - The major separation step requires treated, fine coal  to
 be separated from the spent leach solution.  The four principal methods
 which could be employed  are hydrocyclones, centrifuges, filters and
thickeners.   Suspendable coal has  a  large fraction of particles smaller
than  100 microns in diameter and  in  general  hydrocyclones are not useful
for particle sizes below several  hundred  microns.   Centrifuges would require
                                    151

-------
 very high power input and  recycle rates  to  separate  the  coal  from the  leach
 solution because of the fine particle size  and  the small  liquid-solid  den-
 sity difference.  Filtration is  applicable,  but for  slurries  less than 30
 or 35 percent solids, the  filter area requirements increase rapidly.
 Typically, a 10 percent solids  slurry needs  more than  ten times  the filter
 area needed for a 35 to 55 percent slurry.   Thickeners have been used  on
 commercial scale to remove coal  fines from water and other aqueous media.
 Data for similar density solutions and coal  sizes were reported  in Reference
                                                          2
 4.  It was estimated that  a thickener area of about  20 ft per ton/day of
 coal with an edge depth of about 8 feet  would provide  an  underflow with
 greater than 35 percent solids  and an overflow  containing only  a few  tenths
 percent (or less) solids when the feed contains 10 to  20 percent of 100
 mesh coal.  Since the thickener  slurry can be maintained  near the leach
 solution boiling point, the time spent in the thickener  could be used  to
 carry the leaching reaction to  greater degree of completion and  to re-
 dissolve any solids formed during leaching/regeneration.

 Filtration - The two important  design values relating  to filtration
 are the filtration rate and the  coal  "moisture" content.   These  values are
 not independent and are both highly dependent on the specific coal and its
 properties.   Generalized correlations reported  in the  literature were  re-
 viewed and a data point was obtained from a  filter manufacturer  for a  bench-
 scale slurry of the high ash, Lower Kittanning  coal.  The vendor report
 given in Appendix F showed that  rates equivalent to  about 25  Ib  of dry
           2
 coal/hr/ft  were obtained  with  a 20 percent  slurry of  -100 mesh  coal  in  a
 5  percent iron  leach solution.   They projected  a 60  percent  increase  in
 rate for a slurry with a solids  concentration of about 33 percent.

      One reported correlation  plots rate against a  parameter which  is the
product of the percent  ash  in the  -200 mesh  fraction times the square  root
of the weight percent of the  -200  mesh fraction.  For  the Lower  Kittanning
sample tested in  the above  filtration test,  the parameter has a  value  of
about 200:
          (25% ash) x (67% of -200)1/2 = 205
                                     152

-------
      Figure  22  shows  this measured point and the literature  data with
 extrapolations  to  a typical  14 mesh top size coal.   The  expected filtra-   -
 tion  rate  for a cleaned  14 mesh top size coal  is expected  to be near 200
 lb/hr/ff  in a  33  percent slurry and about 150 lb/hr/ft2 in  a 20 percent
 slurry.  For a  cleaned  100 mesh top size coal  the filtration rate is ex-
 pected to  be about 100  lb/hr/ft2 for a 33 percent slurry and about 70
 lb/hr/ft   for a 20 percent slurry.

      Data  on the moisture content of the filter cake was taken by the
 filter manufacturer at  the same time as rate was measured.   Appendix F
 shows the  projected cake moisture to be 40 to 45 percent.  This is higher
 than  either  reported  in  the  literature (26 to 34%)  or  found  in a typical
 bench-scale  filtration  (about 32%) for the high ash, 100 mesh top size
 coal.  The literature values are for water wet rather  than leach solution
 wet cakes.   The data  for both water and leach solution (5  to 5.5% iron)
 are summarized  as  follows:

                                 parts  of liquid.per  100  parts of dry coal
                                     100 mesh            14  mesh
                                     high ash            high ash

      Bench Scale
         leach  solution               45-50               45-50
         water                        35-40               35-40
      Vendor  Test
         leach  solution               65-80              not  tested
      Reference  5
         water                        35-50               15-25

      In order to provide  for an  adequate amount of coal  washing to remove
 the sulfate  leach  solution it was  decided that 50 parts  of liquid per 100
 parts of dry coal  would  be used  for both leach solution  and  water on both
 coal sizes.

 5.1.2  Process Baseline Design

     A block diagram of the  Meyers  Process as  applied  to coal of about 8 mesh
top size  or finer is shown in  Figure 23.   The  block diagram  shows the main
                                     153

-------
   240


   220


   200


   180


«*  160
I


"I. 140
LU


S  120
u
   100


    80


    60


    40


    20


     0
                                       DATA PLOTTED FROM
                                       FIGURE 12-30 OF
                                       REFERENCE 5 AND
                                       EXTRAPOLATED TO
                                       LOWER RANGES
          -14 MESH
          8-10% ASH
                                   \
                                   LOWER KITT.
                                   -100 MESH TEST
                                    25% ASH
30    40
60
80   100
                                            150    200
                                                             300
               (% ASH IN -200 MESH) X (% OF -200 MESH)
         Figure  22.  Filtration Rate Correlation
                                154

-------
operations and the  interconnections  between each of the four process
sections.  Before discussing  the  process  flow diagrams and mass balance
in detail, the block diagram  will  be described to give a brief process
overview.

Reactor Section - ground coal,.with  a  nominal top size of 14 mesh, is
mixed with hot recycled iron  sulfate leach solution.  After wetting is
complete at the solution boiling  temperature, the slurry is introduced
into a vessel where the majority  of  the pyrite reaction is accomplished at
elevated temperature and pressure.   Oxygen is simultaneously added to re-
generate the  leach  solution.  Heat of  reaction is removed and is used to
reheat the recycle  leach solution.   The slurry is passed to a secondary
reactor operated at atmospheric pressure  and near the solution boiling
temperature,  where  the remaining  pyrite reaction occurs.

Wash Section  - The  iron sulfate leach  solution is removed from the powdered
coal in a series of counter current  flow  contactors and separators.  The
slurry from the secondary reactor is first filtered and the cake is washed
on the filter.  Both the filtrate and  wash liquids are sent to the Sulfate
Removal Section.  The first filter cake is reslurried, filtered a second
time,and then reslurried with recovered clean water and finally dewatered
in a centrifuge.

Sulfur Removal Section - Moist coal  from  the centrifuge is flash dried by
high temperature steam which  simultaneously vaporizes the elemental sulfur
produced in the leach reaction.   The dry  coal is separated from the hot
steam and sulfur vapor stream in  a cyclone and cooled to give the clean
product coal.  The  hot sulfur vapor-steam effluent from the cyclone is
scrubbed with large quantities recycled hot water and the liquid sulfur  is
drawn off to by-product storage.  A  small part of the hot water is used  in
the Wash Section with the remainder  circulated to the evaporator.

Sulfate Removal Section - The major  function of this section is the evapora-
tion of wash water to concentrate leach solution for recycle.  The filtrate
from the wash section and a portion  of the spent wash water  from the  first
filter is fed to a  three effect evaporator which recovers  most  of the wash
water.  The by-product iron sulfate  crystals which  form in the  final  stage
                                     156

-------
 of  evaporation  are  separated  from  the concentrated  leach  solution  and
 stored.   The  remaining wash water  from  the  first  filter is  partially
 neutralized with  lime to yield  a gypsum by-product.  The  separated  and
 partially neutralized wash water is combined with the concentrated  solution
 from the  crystal  separator and  recycled to  the Reactor Section.  Overall,
 the pyrite is reacted with oxygen  and water to give ferrous sulfate, sul-
 furic acid and  sulfur.  These by-products are removed as  shown.  The fuel
 requirement is  equal to a few percent of the product coal and makeup water
 is  needed to  replace water of crystallization and water vapor loss through
 the vacuum filters  and the vacuum  evaporator.

 5.1.2.1   Conceptual Design for Commercial Scale

     Process  engineering studies and trade-offs produced  a baseline flow
 diagram for a commercial scale  plant.   The  flow sheet, which is divided
 into its  four major sections  is given in Figure 24.  The  corresponding mass
 balance and stream  properties are  given in  Table  25.  The baseline  plant
 size was  chosen equal to 100  tons  of dry coal feed  per hour equivalent to
 about 250 MW  power  plant feed.  This size is about  the maximum size for
 a single  train  based on available  commercial equipment.

 Feed and Mixer - Crushed coal, nominally 14 mesh top size, is feed from
 feed hopper A-l.  The coal  is assumed to have 3.2 percent pyritic sulfur
and 10 percent moisture on a dry basis;  thus, the total  solids feed rate
 is  110 tons per hour (TPH)  at room temperature,  assumed to be 77°F.  The
coal feed, stream 1, is brought to the mix tank, T-l, by conveyor, C-l,
 and introduced through the rotary feed valve, RV-1.   Recycled leach
 solution,  stream 4, at its boiling point (215°F)  is introduced to the first
mixer stage after first passing through the gas scrubber SP-1.  Steam,
 streams 2 and 3, is needed to raise the feed coal from 77°F to the 215°F
mixer temperature.  Approximately 5.6 TPH of atmospheric  pressure steam is
required to heat the coal while 6.5 TPH  is  available from the flash drum,
T-2.  It  is possible that the steam would actually be added to the enclosed
conveyor to provide heated coal with an effective 15.6 percent moisture
content.   The excess 0.9 TPH would be vented through SP-1 along with any
flash steam formed in stream 4.
                                    157

-------
                                   A-l

                                   FEED
                                   HOPPER
C-l
FEED
CONVEYOR
ROTARY
VALVE
T-l

MIX
TANK
M-IA/C
MIX
TANK
MIXERS
                                                            M-2A/E   K-l
SCRUBBER-    FLASH
MIST        DRUM
ELIMINATOR
PRIMARY
REACTOR
PRIMARY
REACTOR
MIXERS
PRIMARY
RECYCLE
COMPRESSOR
V-l

KNOCK-OUT
DRUM
R-2

SECONDARY
REACTOR
en
oo
                                                           SLURRY FEED
                                                           PUMP
                                              P-22A/J
                                              CIRCULATING
                                              PUMPS
                                                                 P-2
                                                                 REACTOR DISCHARGE
                                                                 PUMP
                                                                                                                                                       COAL DESULFURIZATION PROCESS
                                                                                                                                                           REACTOR SECTION
                                                                                                                                                    6-10-75              NO. 1335- Ol
                                                            Figure  24.    Process  Flow Diagram for  Fine  Coal

-------
                    VP-1
                              B-l
                                        F-l    V-2
                                                          V-3
                                                                 T-3
                                                                             M-4 .
                                                                                         B-2
                                                                                                    F-2    V-4
                                                                                                                     V-5
                                                                                                                               VP-2     T-4
                                                                                                                                                    M-5
                                                                                                                                                                CG-1
                                                                                                                                                                             T-5
tn
vo
                    VACUUM   BAROMETRIC FILTER FILTRATE      WASH   CONTACTOR  CONTACTOR   BAROMETRIC  FILTER  FILTRATE    WASH     VACUUM CONTACTOR   CONTACTOR  CENTRIFUGE   CENTRATE
                    PUMP     CONDENSER       RECEIVER     RECEIVER             MIXER       CONDENSER         RECEIVER    RECEIVER   PUMP                 MIXER                    RECEIVER
                     P-5

                     LEACH FILTRATE
                     PUMP


                     P-10

                     COOLING WATER
                     RETURN PUMP
P-8          P-ll            P-4           P-9
WASH WATER  COOLING WATER  CONTACTOR   CENTRATE
PUMP        RETURN PUMP     SLURRY PUMP   PUMP
                                                  COAL DESULFURIZATION PROCESS
                                                       WASH SECTION
                                              6-10-75                 NO. 1335-O2
                                                                           Figure  24.     (Continued)

-------
                             SC-1

                             SCREW
                             CONVEYOR
D-l

FLASH
DRYER
RECYCLE
GAS
HEATER
CYCLONE
SEPARATOR
             K-3

             COMPRESSOR
PRESSURE
LET DOWN
SCREW CONVEYOR
SC-3

COAL
COOLER
GAS
COOLER
S-2

CYCLONE
SEPARATOR
5-3
PHASE
SEPARATOR
O
                                                                                                                        P-13

                                                                                                                        SULFUR
                                                                                                                        PUMP
                                                                                    P-12

                                                                                    PROCESS
                                                                                    WATER
                                                                                    PUMP
                                                                                                                                                1   COAL DESULFURIZATION PROCESS
                                                                                                                                                     SULFUR REMOVAL SECTION
                                                                                                                                                6-10-75               NO. 1335-O3
                                                                                 Figure  24.    (Continued)

-------
                                    T-7

                                    ACID
                                    NEUTRAL-
                                    IZATION
                                    TANK
M-6

NEUTRALIZER
MIXER
EV-1

FIRST STAGE
EVAPORATOR
EV-2

SECOND STAGE
EVAPORATOR
                                    EV-3
                                                  E-1
THIRD STAGE    CONCENTRATE
EVAPORATOR    RECYCLE
              REBOILER
CG-2

SULFATE
CRYSTAL
CENTRIFUGE
1-6

CENTRATE
RECEIVER
B-3

BAROMETRIC
CONDENSER
VP-3

VACUUM
PUMP
CTl
                                              P-19
©






M
er;
, (a
T-7
— »
                                                                                                                                                                    P-17
                                                                                          Co SO,
                                                                           P-20
                                    P-19
                                                     P-14
                                    LEACH SOLUTION   EVAPORATOR
                                    RETURN PUMP      CONCENTRATE
                                                     PUMP
                      P-20

                      CaSO4

                      SLURRY PUMP
                         P-15                P-16             P-17             P-18

                         EVAPORATOR         EVAPORATOR      EVAPORATOR      COOLING
                         CONCENTRATE        CONCENTRATE     LEACH SOLUTION   WATER
                         PUMP               PUMP            RETURN PUMP      RETURN PUMP
                                                                              COAL DESULFURIZATION PROCESS
                                                                                SULFATE REMOVAL SECTION
                                                                           6-10-75                NO. 1335-O4
                                                                                   Figure  24.    (Continued)

-------
CT>
ro
                                       TABLE  25.   PROCESS MASS BALANCE FOR FINE COAL

                                                    (Stream Flows in Tons Per Hour)

Water
FeS04
Fe2(S04)3
H2S04
Pyrite
Sulfur
Coal
Oxygen
Inert
Total, TPH
T, °F
P, Psig
gpm
P, lb/ft3
Fe, %
Y
S04/Fe
COAL MAKEUP
FEED STEAM
1 2
10.0 -.9*
6.0

94.0


TTO" ^oTg"
77 215
0 -9
50.0
-
-
_
FLASH FEED
STEAM SOLN.
3 4
6.5 144.1
3.9
30.6
5.7




O" 184.4
215 215
.9 0
614
74.8
5.4
.86
1.75
R-l 02 RECYCLE
FEED MAKEUP ~ GAS
56 7
159.2 1.5
14.4
18.2
8.8
5.2
.2
94.0
3.9 13.1
Tr .8
3oO 379 TO"
215 77 264
28.8 53.8 53.8
968
77.3
5.2
.49 -
1.73 -
R-l COMPRESSOR R-l
GAS FEED EXIT
8 9 10
14.0 1.5 147.4
5.8
37.0
5.0
.7
1.1
94.0
13.4 > 13.1
.8 .8
oo o i c ^ 9Q1 fl
£O*£ 1 Q • O £~7 1 • U
250 177 250
28.8 28.8 28.8
907
80.0
6.3
.83
1.64
R-2
FEED
16
140.9
5.8
37.0
5.0
.7
1.1
94.0


284TB"
215
0
873
81.3
6.6
.83
1.64
R-2
EXIT
17
140.7
11.1
30.6
6.6
.3
1.2
94.0


28175"
215
0
874
81.2
6.7
.68
1.64
     Excess steam to vent

-------
TABLE 25.  (CONTINUED)

Water
FeS04
Fe2(S04)3
Pyri te
Sulfur
Coal
—j Oxygen
CT>
co Inert
Total, TPM
T, °F
P, Psig
gpm
p, lb/ft3
Return Vent
Soln. 00
18 19
131.7 Tr
3.9
30.6
5.7



.4
Tr
171.9 0.4
127 177
28.8 0
548
78.2
Crystallizer
Centrate
20
50.3
2.2
26.0
5.2




83.7
200
30. G
221
95.4
Neutral izer
Return
21
81.4
1.7
4.6
.5




88.2
160
15.0
335
65.6
Neutral izer
to
Crystallizer
22
54.3
1.1
3.1
0.3




58.8
160
15.0
224
65.6
Filtrate
to Cake
Crystal! izer F-l
23
105.2
8.3
22.9
4.9




141.3
160
10.0
443
79.6
24
43.1
1.0
3.0
.6
.3
1.2
94.0

143.2
160
0
;
Wash Contactor
F-l Vents
25 26
143.1 Tr
1.0
3.0
.6




147.7 0.0
160 77
10.0 0
589
62.5
F-2
To Contractor Feed
Neutral izer Feed T-3 Slurr
27
135.5
2.8
7.7
1.7




147.7
160
5.0
561
65.7
28 29
146.3 189.4
.3 1.3
.9 3.9
.2 .8
.3
1.2
94.0

147.7 290.9
1 60 1 60
5.0 15.0
601 1065
61.3 68.1

-------
TABLE 25.  (CONTINUED)


Water
FeS04
Fe2(S04)3
H2S04
Pyrite
Sulfur
Coal
Oxygen
Inert
Total, TPH
T, °F
P, Psig
gpm
P, lb/ft3
Dryer
Cake Wash Water Makeup Evaporator
F-2 F-2 Return Water Return
30 31 32 33 34
47.2 147.2 14.3 27.1 72.9
.1 .1
.3 .3
.1 .1
.3
1.2
94.0


143.2 147.7 14.3 27.1 72.9
160 160 215 77 180
0 10.0 30.0 30.0 30.0
607 60 108 473
60.7 59.8 62.3 60.2
Centrifuge
Feed
Slurry
35
161.5
.1
.3
.1
.3
1.2
94.0


257.5
180
15.0
942
68.2
Cake Dryer Dryer
Centrifuge Gas Output
36
14.3
Tr
Tr
Tr
0.3
1.2
94.0


109.8
180
0
-
-
37 38
258.9 273.2
Tr
Tr
Tr
0.1
Tr 1.2
Tr 41.4


258.9 315.9
650 450
20.0 18.0
-
-
Dryer
Coarse
Cut
39
Tr
Tr
Tr
Tr
0.2
Tr
52.6


52.8
450
18.0
-
-
Cyclone Purge
Sol ids Steam
40 41
Tr 0.1
Tr
Tr
Tr
0.1
Tr
41.3


41.4 0.1
450 300
18.0 18.0
-
50.0
Coal
Product
42
0.1
Tr
Tr
Tr
0.3
Tr
93.9


94.3
150
0
-
50.0
Cooler
Cyclone Water
Gas Feed
43 44
273.2 1137.4




1.2
.1


274.5 1137.4
450 215
17.8 30.0
4745
59.8

-------
TABLE 25.  (CONTINUED)

Mater
FeS04
Fe2(S04)3
Pyrite
Sulfur
Coal
Oxygen
Inert
Lime
Gypsum
Total , TPH
T, °F
P, Psig
gpm
P, lb/ft3
Cooler Gas
Effluent Effluent
45 46
1410.6 258.9


1.2 Tr
.1 Tr




1411.9 258.9
250 250
15.0 15.0
Separator
Liquid
47
1151.7


1.2
.1




1153.0
250
15.0
4851
59.2
Reboiler
Feed
48
1151.7








1151.7
250
40.0
4851
59.2
Water
Return
49
1151.7








1151.7
215
30.0
4802
59.8
Sulfur
Product
50



1.2
0.1




1.3
215
25.0
2.9
112.0
From
EV-1
51
124.1
9.4
26.0
5.2







164.7
120
5.0
516
79.6
to
Vacuum
52
35.4








35.4
115
(1.5
Psia)
From From
EV-2 EV-2
53 54
35.2 88.9
9.4
26.0
5.2







35.2 129.5
1 50 155
(3.7 5.0
Psia) 3?6
85.9
From
EV-3
55
37.7








37.7
207
(13
Psia)
Centrifuge Sulfate
Feed Product
56 57
51.2 0.9
9.4 7.2
26.0
5.2







91.8 8.1
210 200
5.0 0
230
99.7
Calcium
Lime Sulfate
58 59
0.4






0.5
1.6
0.9 1.6
77 77
0 5.0
2.8
144.8

-------
     The mixer vessel T-l was sized for three stages of mixing at 0.25 hours
per stage.  Under the design constraint that the vessel is 75 percent full,
the cost model used for vessel sizing found a field fabricated vessel 18.7
feet in diameter by 32.9 feet long has minimum cost.  The selected vessel
size (18 x 36) gives three stages each about 12 feet long and 12.6 feet
deep with slightly less than 15000 gallons in each stage.  Any foam gen-
erated during coal wetting will be broken down and the entrapped air will
be scrubbed in SP-1 by the returning leach solution.  The actual air flow
through SP-1 is very low and will probably not exceed the air in the bulk
coal (50 cubic feet per minute).

Primary Reactor - The fully wetted and deaerated coal slurry from the mixer
is pumped by slurry pump P-l  (stream 5) into the first stage of the primary
reactor R-l.  Both removal of pyrite and oxidation of ferrous to ferric
iron sulfate occur in this reactor.  A five stage reactor was selected
since the cost model showed the minimum cost field fabricated vessel had
length to diameter ratios near five.  Under the design constraint that the
reactor must have five stages and operates about 85 percent full, the cost
model found a reactor 25.9 feet in diameter by 127.7 feet long operated  at
15 psi of oxygen was minimum cost.  The selected vessel size (26 x 125)
gives five  stages each about 25 feet long by 23 feet deep and holding about
80,000 gallons of slurry.  At the residence time of 1.5 hours per stage, a
temperature of 250°F and an oxygen partial pressure of 15 psi, the pyrite
is reduced to 88 percent of the original level and the leach solution is
regenerated to a Y (ferric iron to total iron ratio) of 0.83 in the primary
reactor.

Oxygen Loop - Excess oxygen saturated with steam and containing an equilib-
rium level of inert gas (mainly argon) leaves the primary reactor in stream
8.  The gas is contacted with returning leach solution, stream 18, in a
knock-out drum, vessel V-l.  The leach solution is warmed to 215°F (stream
4) by condensing steam from the oxygen stream.  The gaseous effluent, which
was assumed to leave  V-l  50°F warmer than the feed leach solution is split to
give a small vent stream 19 and a recycle oxygen stream 9.  The vent rate
is selected to maintain the inert gas at the design level; namely 5  percent
on a dry basis.   The recycle oxygen is compressed by K-l to the reactor feed
pressure.   Makeup oxygen, stream 6, is added to balance the oxygen used for
                                    166

-------
 regeneration  in  R-l  and  that  vented  to  remove  inerts.

     Assuming 15 psia  oxygen  pressure the  gas  pressures in  reactor R-l  at
 250®F  are as  follows:

     Oxygen          15.0  psia
     Inert Gas          .8  psia
     Steam          27.7  psia
                     43.5  psia  (28.8 psig)

 Since  the recycle gas  must also  overcome the liquid  head  in the reactor
 (about 13 psi),  the  control valve/injector drop  (about  10 psi) and  other
 line losses,  the recycle compressor  was sized  to  provide  a 25 psi  pressure
 increase.  For the baseline case this results  in a 300  horsepower compressor
 operating at  a 1.58  compression  ratio and a compressor  outlet pressure of
 53.8 psig.

 Flash  Steam  - The heat of  reaction and  regeneration  is  accommodated in
 three  ways:   temperatures  of  the recycled  oxygen  and the  feed slurry are
 raised in R-l, heat  is lost from the insulated walls of the mixer and
 reactors, and water  is evaporated from  the solutions.   Part of the steam
 (13.4  TPH) is removed  from the  recycle  oxygen  to  provide  an isothermal
 primary reactor  R-l  at 250°F  and part of the steam (6.5 TPH) is removed
 by flash drum T-2 in dropping the slurry temperature and  pressure from
 reactor R-l  (250°F)  to reactor  R-2 (215°F).  The  heat is  almost entirely
 utilized in  heating  the feed  coal and the  recycle leach solution.

 Secondary  Reactor  -  The  secondary reactor, R-2, is operated near  the  atmo-
 spheric  boiling  point with  a  residence  time of 36  hours.  During  this time,
 additional pyrite  is removed  from the coal to  provide an  overall  pyrite
 removal  of 95 percent while the  Y of the solution  is decreased to a  value
 of 0.68  in the reactor effluent.   The low value of Y  is desired to  provide
 sufficient ferrous sulfate  for removal  as the  by-product  iron form.   The
cost model found the minimum  cost reactor was  27.9 feet in diameter  by 465.9
feet long.  The  final equipment  list and costing  used three field fabricated
vessels each 28  feet in diameter  and  160 feet  long.  The  reactors contain no

                                    167

-------
 internal  stages,  but  have  circulating  pumps to avoid large vertical con-
 centration  gradients  from  occurring  in the solution.  The slurry from the
 secondary reactor,  stream  17,  is  pumped  by P-2 to the first filter, F-l.

 Coal  Washing  -  Bench-scale experience  with removal of the sulfate  leach
 solution  from coal  shows that  the  solution may be treated as consisting of
 two  types.  Surface solution is readily  removed by flushing with water or
 may  be readily  displaced by a  more dilute wash solution.  Solution in
 the  pores of  the  coal particles requires a definite residence  time to reach
 equilibrium with  the  bulk  or surface liquid.  The coal washing section,
 therefore,  consists of filtration, washing on the filter, equilibration
 with dilute solution, a second filtration and wash, equilibration  with
 wash water  and  finally dewatering in a centrifuge.

 First Filter  -  Coal slurry from the  secondary reactor, stream  17,  containing
 approximately 33  percent solids is fed to a 12 foot diameter by 24 foot
 long rotary vacuum filter, F-l.   The filtrate from vacuum receiver V-2,
 stream 23,  is pumped, P-5,  to the  sulfate removal section.  Dilute  wash
 solution  from the second filter,  stream  25, is used to wash the filter  cake
 and  displace  the  surface solution on the coal particles.  This sulfate  rich
 wash solution,  stream 27,  is pumped, P-6, from the vacuum receiver V-3  to
 the  sulfate removal section.   Vacuum is  provided by a 3000 standard  cubic
 feet per  minute (SCFM) vacuum  pump,  VP-1, which is vented, stream  26, back
 to the enclosed filter F-l.  The  vapors  and gases removed from the vacuum
 receivers,  V-2  and  V-3, pass through a barometric condenser, B-l,  before
 entering  the  vacuum pump.  In  B-l  most of the flash steam is condensed  and
 enters the  cooling  water loop  where  it is pumped to the  cooling water tower
 by P-10.

First Stage Repulping  - The washed filter cake from the  first  filter, stream
 24, and dilute  wash water  from the second  filter  are  gravity fed through a
 closed chute  to a stirred  tank, T-3.  This  40,000 gallon tank  is operated
 about  three-fourths full to1 give  an  average residence time of  30 minutes
 to equilibrate  pore solution with the  bulk liquid.   The slurry, stream 29,
 is pumped,  P-3, to  the second  stage  filter.   Any gases introduced with the
cake are  vented to  the scrubbing  system, stream 26.
                                     168

-------
Second Filter - The partially washed slurry, stream 29, containing approxi-
mately 33 percent solids, is filtered and washed on a second filter of the
same size and type as the first filter.  Filtrate, stream 25, is pumped, P-7,
from the vacuum receiver, V-4, to the first filter wash.  Wash water for
the second filter, stream 31, is obtained from the centrate receiver.   The
partially spent wash water is pumped, P-8, from the vacuum receiver V-5 to
the first stage contactor.  Vacuum is provided by vacuum pump VP-2 operating
through the barometric condenser B-2.

Second Stage Repulping - The washed filter cake from the second filter,
stream 30, is contacted with water in a 40,000 gallon stirred tank, T-4.
The wash water, streams 32, 34, and 33 is obtained from the dryer, the
evaporators, and makeup, respectively.

Dewatering - The slurry from the second contactor, stream 35, is pumped,
P-4, to the dewateririg centrifuge, CG-1.  The slurry with approximately 33
percent solids is separated in the 36 inch diameter by 90 inch long solid
bowl centrifuge to provide a dewatered coal, stream 36.   According to  vendor
literature and discussions, the dewatered coal  is expected to have about 15
percent moisture.   The centrate from receiver T-5 is pumped, P-9, to provide
the wash, stream 31, for the second filter.

Drying - Coal from the centrifuge, stream 36, is fed to a flash dryer, D-l,
by a screw feeder, SC-1.  In this dryer concept the coal is heated to about
450°F by superheated steam, stream 37, and carried upward to the enlarged
top area of the dryer.  The larger particles are removed from the dryer,
stream 39, while the fine particles and gas, stream 38, are fed to a cyclone,
S-l.  During the drying in D-l sulfur is also vaporized from the coal  and
is present along with water vapor in the cyclone effluent gas stream 43.
Thefine coal from the cyclone, stream 40, and coarse coal,  stream 39, are
let down to atmospheric pressure by screw conveyor SC-2 which is back purged
with a small quantity of steam to prevent the sulfur containing  gas in the
cyclone from leaving the system with the coal.  The coal, stream 42,  is
then transported and cooled to product storage temperature  by the  screw  con-
veyor, SC-3 which rejects heat either to cooling water  or to  the  atmosphere.

                                    169

-------
Sulfur Removal - The cyclone effluent gas,stream 43, at about 450°F is
cooled by a large spray of water, stream 44, in gas cooler  C-l.  The water
is obtained from return stream 49 from the sulfate removal section.
The gas and liquid, stream 45, cooled to 250°F is separated in cyclone S-2
to  give vapor stream 46 and liquid  stream 47.   The  liquid stream 47
contains the water fed to the gas cooler, stream 44, the water vaporized
from  the coal  in the dryer, and the sulfur vaporized from the coal.  The
liquid is phase separated in vessel S-3.  The liquid sulfur by-product,
stream 50,  is  pumped, P-13, to storage while the hot water, stream 48, is
pumped, P-12,  to the sulfate removal section.

Steam Circulation - Saturated steam at 250°F from the cyclone, stream 46,
is compressed  by K-3, reheated by H-l, and fed to the dryer as stream 37.
Compression is accomplished by two 3500 HP series compressors which make up
the 10 psi  pressure drop around the gas circulation loop.  The heater pro-
vides nearly 100 million Btu per hour (MM Btu/hr) to the steam to supply
the heat required to heat the dryer feed, stream 36, to 450°F and vaporize
the water and  sulfur.  Slightly more than 80 MM Btu/hr are rejected to the
hot water loop,  stream  48,  for  use  in the  sulfate removal  section
while about 15 MM Btu/hr are lost from the equipment and lines or rejected
as sensible heat in the hot coal and liquid sulfur.   The circulating water
is kept in balance by returning a portion of the water, stream 32, to the
wash  section equal to the water vaporized from the feed coal, stream 36.

Neutralization - Sulfate rich wash solution from the wash section, stream
27, is fed to  a stirred tank, T-7, and a lime slurry, stream 58,  is added
to neutralize  part of the sulfuric acid.  The tank is sized for  about 10
minutes of residence time and has a baffled settling zone.  Gypsum slurry
stream 59 is withdrawn for disposal and the partially neutralized liquid
is removed by  pump P-19.  A portion of the liquid, stream 21, is returned
to the reactor section while the remainder, stream 22, is combined with
the filtrate,  stream 23, as feed to the triple effect evaporators.

Evaporation - Evaporator EV-1 is operated at partial vacuum  (about 0.1 atmo-
spheres)  and uses condensing steam from the second evaporator,  stream  53,
to evaporate water,  stream 52, in the first evaporator.   The evaporated
                                    170

-------
water is condensed in the barometric condenser,  B-3,  and any residual  gas  is
removed by vacuum pump VP-3.  The partially concentrated leach solution,
stream 51, is pumped, P-14, to the second evaporator, EV-2.  The second
evaporator operates at about 155°F and 0.2 atmospheres using steam from
the third evaporator, stream 55, to evaporate the water, stream 53.  The
two condensate streams from the reboilers of the first and second evapo-
rators (streams 53 and 55) are combined, stream 34, to provide clean wash
water for the wash section.  The leach solution from the second evaporator,
stream 54, which has been concentrated to 8.3 percent iron, is at a
temperature where the solubility of ferrous sulfate is a maximum and is a
solids free solution.  This stream is feed to the third evaporator, EV-3,
which is operated at atmospheric pressure and at the normal boiling point
of the solution.  Heat to vaporize water is provided to the reboiler, E-l,
by the hot water loop from the wash section (streams 48 and 49).  The over-
head steam, stream 55, is used in the second evaporator as previously de-
scribed.  The leach, solution in EV-3 is concentrated to a total iron
concentration of nearly 12 percent which exceeds the solubility of ferrous
sulfate.  Thus, crystalline ferrous sulfate forms in EV-3 and a portion of
the slurry, stream 56, is fed to a centrifuge CG-2 to separate the crystals,
stream 57, from the concentrated leach solution, stream 20.  The concentrated
leach solution is pumped, P-17, to the reactor section.

Solubilities - Since the solubility of ferrous sulfate in the presence of
ferric sulfate, sulfuric acid and trace ions is not yet fully defined, the
baseline process flows may require some adjustment when pilot scale data
have been evaluated.   Nevertheless, the planned mode of operation which
takes advantage of the reported solubility characteristics of ferrous sul-
fate in aqueous solution should be applicable.  Below about 150°F, the
equilibrium crystalline phase is FeS04-7H20 which has an increasing solu-
bility with temperature.  It reaches a maximum solubility of nearly 60 grams
of FeSO, (anhydrous basis) per 100 grams of water.  Above about 150°F the
equilibrium solid phase is FeSO^-^O which has a decreasing solubility in
water with increasing temperature.  Both the first and second stages of
evaporation are below the saturation limits and are expected to remain
solids free.   Only the final stage operates as a crystallizer and produces
                                     171

-------
crystalline ferrous sulfate both from a decreased solubility at the higher
temperature and from an increased concentration because of evaporation.

5.1.2.2  Process Cost Estimate

      Throughout bench-scale development, process costs have frequently been
 reviewed with  an objective of focusing experimental effort in  the  areas of
 greatest cost  sensitivity.  The  capital cost of  equipment required to  per-
 form the pyrite leaching must be carefully controlled  to maintain  a low
 processing cost per ton of coal  product.  As will be seen in the capital
 estimate presented in  the following  discussion,  the major capital  cost
 continues  to be in the reactor  section of the  process.  This section of the
 unit accounts  for  approximately  48 percent of  the total installed  equipment
 capital  cost.   The sulfur removal section of the process accounts  for  22
 percent  while  the  wash section  represents 17 percent and the sulfate removal
 section  13 percent of  the total  equipment capital requirements.  It there-
 fore becomes apparent  that the  reactor section of the  process  represents
 the most likely  area of future  process economic  gains  as the design data
 base broadens  and  other innovative process schemes  (relative to  the reaction
 section  of the process) are evaluated.

      As  the process development progressed and additional experimental data
 were obtained, some complications were identified and  some  process simplifi-
 cations  were demonstrated.  The  net  result is  that  at  the  conclusion of this
 bench-scale effort, the process  for  removing pyritic sulfur from coal  remains
 highly attractive  and  sufficient data  has been obtained to  provide confidence
 in the economic viability of  the process.

 Baseline Capital Cost  Estimate  - The previous  section  of this  report pre-
 sented a conceptual  process design and process flow sheet  for  removing 95
 percent of the pyritic sulfur from a high ash  coal  which  initially contained
 3.2  percent pyritic sulfur.  The major equipment for the  process is given in
 Table 26 and identified with the equipment of  the flow sheet (Figure 24).
 The  equipment  was  selected and  sized to approach the optimum cost  for pro-
 cessing this high  pyrite coal to the 95 percent  removal  level.

                                     172

-------
                       TABLE 26.   COAL DESULFURIZATION PROCESS EQUIPMENT LIST
REACTOR SECTION    $3.26 MM FOB. $6.36 MM INSTALLED*
                                                              o
   1           A-l            Ground Coal  Feed Hopper - 5000 ft
   2           C-l            Feed Conveyor - 20 in. Wide x 20 ft, 5 hp,  200  ft/min.
   3           K-l            Oxygen Recycle Compressor - 300 hp,  1.6 Compression  Ratio
   4           M-1A/C         Mix Tank Mixers (3) - 15 hp, Stainless Steel  (SS)
   5           M-2A/E         Primary Reactor Mixers (5) - 200 hp, SS
   6           P-l            Slurry Feed  Pump - 1000 gpm, 60 psi, 50 hp,  SS
   7           P-2            Reactor Discharge Pump - 875 gpm, 5  psi, 3.5 hp, SS
   8           P-22A/J        Circulation  Pumps (12) - 1000 gpm, 5 psi, 4.0 hp,  SS
   9           R-l            Primary Reactor - 26 ft  0 x 125 ft, Carbon Steel  (CS)  with  SS  clad,  30 psig
  10           R-2            Secondary Reactor (3) - 28 ft 0 x 165 ft, SS, 0 psig
  11           RV-1           Rotary Valve -  .5 hp, 18 in. x 18 in., 20 RPM
  12           SP-1           Scrubber-Mist Eliminator - 3 ft 0 x  10 ft,  SS,  0 psig,  Baffles, Demister  Pad
  13           T-l            Mix Tank - 18 ft 0 x 36 ft, SS, 0 psig
  14           T-2            Flash Drum - 5 ft 0 x 10 ft, SS, 5 psig
  15           V-l            Knock-Out Drum - 5 ft 0 x 25 ft, SS, 30 psig, 15 ft  Packing, Demister Pad
  Installed costs for each process section were derived through the application of the appropriate
  Guthrie factor^ to the FOB cost of individual pieces of equipment.

-------
TABLE 26.  (CONTINUED)
WASH SECTION
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
$1.16 MM FOB,
B-l
B-2
CG-1
F-l
F-2
M-4
M-5
P-3
P-4
P-5
P-6
P-7
P-8
P-9
P-10
P-ll
T-3
T-4
T-5
V-2
V-3
V-4
V-5
VP-1
VP-2
2.28 MM INSTALLED
Barometric Condenser - SS, Condensation Rate = 13 Ton/Hr
Barometric Condenser - SS, Condensation Rate = 2. 5 Ton/Hr
Centrifuge (4) - 36 in. 0 x 90 in. Solid Bowl, SS, 150 hp
Rotary Vacuum Filter - 12 ft 0 x 24 ft Drum, 912 ft2, SS, 8 hp
Rotary Vacuum Filter - 12 ft 0 x 24 ft Drum, 912 ft2, SS, 8 hp
Contactor Mixer, 35 hp, SS
Contactor Mixer, 35 hp, SS
Contactor Slurry Pump - 1065 gpm, 15 psi, 15 hp, SS
Contactor Slurry Pump - 950 gpm, 15 psi, 10 hp, SS
Leach Filtrate Pump - 450 gpm, 10 psi, 3.5 hp, SS
Leach Wash Water Pump - 560 gpm, 5 psi, 2.5 hp, SS
Filtrate Pump - 590 gpm, 10 psi, 5 hp, SS
Wash Water Pump - 560 gpm, 5 psi, 2.5 hp, SS
Centrate Pump (4) - 150 gpm, 10 psi, 1 hp, SS
Cooling Water Return Pump - 1200 gpm, 5 psi, 5 hp
Cooling Water Return Pump - 200 gpm, 5 psi, 1 hp, CS
Contactor- 40,000 gal, 0 psig, SS
Contactor- 40,000 gal, 0 psig, SS
Centrate Receiver (4) - 650 gal, 0 psig, SS
Filtrate Receiver- 2,000 gal, Vac, SS
Wash Receiver- 2,500 gal, Vac, SS
Filtrate Receiver- 2,500 gal, Vac, SS
Wash Receiver- 2,500 gal, Vac, SS
Vacuum Pump- 3000 SCFM, 200 hp, CS
Vacuum Pump - 3000 SCFM, 200 hp, CS

-------
—1
in
                                         TABLE 26.  (CONTINUED)


 SULFUR REMOVAL SECTION   $1.42 MM FOB,  $2.94 MM INSTALLED

 1            C-l             Gas Cooler - 7 ft 0 x 100 ft,  Water Sprays, SS,  15 psig
 2            D-l             Flash Dryer - 11  ft (2 x 65 ft  Drying Section,  22 ft  0 x 20 ft De-entrainment Section,
                             SS, 20 psig
                                                                                         o
 3            H-l             Recycle Gas Heater - 97  MM Btu/Hr,  Radiant Section = 6000 ft, Convective Section =
                             12,000 ft2, SS Tubes
 4            K-3             Compressor (2) -  1.15 Compression Ratio,  3500  hp
 5            P-12            Process Water Pump - 4850 gpm, 40 psi,  150 hp, CS
 6            P-13            Sulfur Pump - 3 gpm, 25 psi, 0.5 hp, SS
 7            S-l             Cyclone Separator - SS, 15 psig, 120,000  ACFM  Capacity
 8            S-2             Cyclone Separator - SS, 15 psig, 107,000  ACFM  Capacity
 9            S-3             Phase Separator - 50,000 gal,  15 psig, SS
10            SC-1            Screw Conveyor -  20 ft x 14 in. 0, 2 hp,  SS
11            SC-2            Pressure Let Down Screw Conveyor - 20 ft  x 14  in. 0, 2 hp, CS
12            SC-3            Coal Cooler - Screw Type, 20 ft x 14 in.  0, Cooled Shell, CS, 2 hp

-------
                                      TABLE 26.  (CONTINUED)

 SULFATE  REMOVAL  SECTION   $0.97 MM  FOB.  $1.68 MM INSTALLED

 1            B-3             Barometric Condenser - SS, Condensation Rate =35.4 Ton/Hr
 2            CG-2             Sulfate Crystal  Centrifuge - 36 in. 0 x 72 in. Solid  Bowl,  SS,  125 hp
 3            E-l              Concentrate  Recycle Reboiler - 10,000 ft2, SS/SS
 4            EV-1             First Stage  Evaporator - Evaporation Rate = 35 Ton/Hr,  1.5  psia, SS
 5            EV-2             Second Stage Evaporator - Evaporation Rate = 35 Ton/Hr,  3.7 psia,  SS
 6            EV-3             Third Stage  Evaporator - Evaporation Rate = 38 Ton/Hr,  13 psia, SS
 7            M-6             Neutralizer  Mixer - 5 hp, SS
 8            P-14             Evaporator Concentrate Pump - 520 gpm, 5 psi, 2.0 hp, SS
 9            P-15             Evaporator Concentrate Pump - 380 gpm, 5 psi, 1.5 hp, SS
10            P-16             Evaporator Concentrate Pump - 1380 gpm, 5 psi, 5.0 hp,  SS
11             P-17             Leach Solution Return Pump - 220 gpm, 30 psi, 5.0 hp, SS
12            P-18             Cooling Water Return Pump - 8000 gpm, 5 psi, 30 hp, CS
13            P-19             Leach.Solution Return Pump - 560 gpm, 30 psi, 10 hp,  SS
14            P-20             Calcium Sulfate Slurry Pump - 3 gpm, 5 psi, 0.5 hp, SS
15            T-6              Centrate Receiver - 900 gal, SS, 0 psig
16            T-7              Neutralizer  Tank - 7,500 gal, SS, 0 psig
17            VP-3             Vacuum Pump  - 700 CFM, 50 hp

                               TOTAL ESTIMATED  CAPITAL - $6.SIMM  FOB. $13.26MM INSTALLED

-------
     Capital  equipment costs were obtained  from various sources:  technical
literature,  equipment suppliers and internal  (TRW)  costing data.   The
 specific  sources of  data  for the  various classes of equipment are presented
 in  Table  27  .  When  cost  data were obtained  from literature or other non-
 current sources, appropriate cost escalation  factors,  based on the Marshall
 and Swift Equipment  Cost  Index (to escalate  costs  from date of publication
 to  June 1975), were  applied.   The capital  equipment cost  for each processing
 section is  presented in  Table 26.   The costs  are  presented in terms of FOB
 equipment cost and installed equipment cost.   The  FOB  equipment  cost is the
 base,  uninstailed  cost at  point  of manufacture or point  of shipment.  The
 installed equipment  cost  includes the following elements:
                       •   FOB Equipment Cost
                       •   Field Materials
                          -   Equipment
                          -   Piping
                          -   Concrete
                          -   Steel
                          -   Instruments
                          -   Electrical
                          -   Insulation
                          -   Paint
                       •   Material Erection
                       •   Direct Field Labor
                       •   Indirect Costs
                          -   Freight
                          -   Taxes
                          -   Construction Overhead
                          -   Fringe Benefits
                          -   Labor Burden
                          -   Field Supervision
                          -   Temporary Facilities
                          -   Construction Equipment
                          -   Small Tools
                          -   Miscellaneous  Field Costs
                          -   Contractor Engineering Costs
The installed equipment cost does not include  a contingency factor.
                                     177

-------
       TABLE 27.   SOURCES OF EQUIPMENT COST INFORMATION
        Equipment Type
Hoppers
Conveyors
Compressors
Mixers
Pumps
Reactors
Vessels >40,000 gal
Vessels <40,000 gal
Tanks >40,000 gal
Tanks <40,000 gal
Drums
Centrifuges and Support Equipment
Filters and Support Equipment
Heat Exchangers
Evaporators
Gas Cooler
Dryer
Heater
Cyclone Separators
Rotary Valves
 Information Source
TRW Data
Reference 7
Elliott Company
Reference 8
Reference 6
TRW Data
TRW Data
Reference 6
TRW Data
Reference 6
Reference 6
Bird Machine Company
Ametec Company
Reference 6
Reference 9
TRW Data
TRW Data
Reference 6
TRW Data
Reference 6
                              178

-------
Operating Cost Estimate - The process operating costs have also  been  esti-
mated.   The basis for these estimates was technical  literature and  informal
supplier quotes.   Specific sources of information  are presented  in  Table 28,

     The  total estimated  processing  cost has been determined as follows:
     Capital  Related  Costs:                             Annual  Cost, $1000
        Depreciation -  10% straight line                       1,326
        Maintenance, insurance, taxes, interest (15% of        1,989
        Labor:                                      caP1tal)
              Labor, 8  operating positions                     1,200
        Utilities:
              Electrical power - 7500  KW  (25 mil/Kw-hr)        1,500
              Cooling water - 20°F rise; 9500 gpm                228
                             (5«/1000 gal)
              Heating  - 97 MM Btu/hr;  coal, 4T/hr
              Process water - 110 gpm (25<£/1000 gal)              13
        Materials:
              Oxygen 99.5%,  3.9T/hr ($25/T)                      780
              Lime -  .5T/hr ($28/T)                              112
        TOTAL  COST                                             7,148
        Processing  Cost  (100 T/hr;  0.8 MM T/yr) $8.94/T of feed coal
        Coal yield  (weight basis)     90%
        Coal yield  (Btu basis)        94%
     The added cost of energy may  also be considered for the baseline coal.
 If  the  baseline coal is similar to the Lower Kittanning coal utilized in
 our laboratory studies, it will contain about 20 percent ash and have a
 heating value of 12,300 Btu/lb as  fed.  After processing the coal will be
 89  percent recovered, have 16 percent ash, and have a heating value of
 12,900  Btu/lb.  With feed coal prices at $15.00/T the feed costs 6U/MMBtu.
 After processing the available energy costs $1.04/MMBtu.

     Based on the  current conceptual process design, it is concluded that
 a broad spectrum of Eastern coals  can be processed at costs of about $9.00
 per ton.  It was assumed in developing these costs that the pyrite removal
 plant is coordinated with a power  plant which will have the principal
                                   179

-------
             TABLE 28.   SOURCES OF OPERATING COST INFORMATION

               Cost  Element                           Information Source
 Maintenance,  Insurance, Taxes and Interest           Reference 8
 Labor Requirement (Number of Positions)              Reference 8
 Labor Cost                                          TRW Data
 Utilities
    Electrical  Power                                 Reference 8
    Cooling Water                                    TRW Data
    Process Water                                    Reference 8
    Oxygen                                            Linde, Division of
                                                       Union Carbide
    Lime                                             Reference 10
 off-site facilities  such  as  coal grinding facilities, change house, offices,
 rail  facilities,  etc.   To the extent that these off-sites are not available
 or for bookkeeping purposes  are prorated to the coal processing cost, the
 direct costs  given above  will be increased.

 5.1.3   Process Trade-Off  Studies

     At  the start of this bench-scale program a conceptual process design
 was available based on data  generated in the previous program.   Early
 bench-scale experiments showed that in the elemental sulfur recovery
 section, the desired displacement of the aqueous leach solution by an
organic  sulfur solvent could not be reliably performed.  Similarily, dis-
placement of the organic  solvent by wash water did not occur.  The
demonstrated fall-back process using parafinic or aromatic hydrocarbon
solvents as used in the current bench-scale effort is thought to  be  less
attractive at a commercial scale than sulfur vaporization for three main
reasons:
                                    180

-------
     •  Higher  equipment  capital cost for the multiple  extraction,
        solvent distillation and sulfur crystallization.

     •  Higher energy use, estimated at 4 to 5 percent  of the coal,
        because the heat  is rejected at too low a temperature level
        to be used in the sulfate removal section.

     §  Loss of solvent on the coal  (less than 1 percent and probably
        only a  few tenths of a percent).
 Therefore, the more favorable alternate shown  in the baseline design  (Section
 5.1.2) was developed and demonstrated to be operable by laboratory experi-
 ments.
                                *
     With the coal washing and  sulfur removal  process  identified, emphasis
 centered on the reactor section cost.  It was  pointed  out in Section  5.1.2.2
 that nearly one-half of the process equipment  cost  is  associated with the
 reactor section in the baseline design.  Consequently, the majority of the
 trade-off studies conducted during the course  of the experimental phase of
 the program and the more concentrated effort performed during the subsequent
 conceptual process design were  concentrated on the  reaction section.  A
 number of the more pertinent results are summarized in the following  para-
 graphs.

 5.1.3.1  Reactor Model
     Bench-scale results showed the initial rate of pyrite reaction increased
with increased temperature  (and pressure).  This increased rate only applied
to removal of the first 70 or 80 percent of the pyrite after which the rate
at high temperature was about the same as the rate obtained at the normal
boiling point of the leach solution.  A computer model was prepared de~
cribing the reaction in a continuous well stirred reactor.  Since the coal
particles have a statistical distribution of residence times ranging from
zero to infinity, some particles have very little removal while others with
long residence time have very high removals.  The coal was grouped into  a
                                     181

-------
 variable number of equal weight  fractions  and  reacted  for  average  residence
 time calculated as the  arithematic  mean  time for  the group,   For ten  time
 groups the numerical  error  in  pyrite  removal compared  with the  integrated
 results was about one percent, the  error was negligible  with  100 or 1000
 groups.  When this computer routine was  incorporated into  the overall
 reactor computer model, forty  time  groups  were selected  to provide results
 with virtually no numeric error.

      In a  multistage  reactor, coal  from the first well  stirred stage con-
 tinuously  enters  the  next stage.   Thus, if it  is assumed that the  first
 stage  is fed  with a coal of uniform pyrite concentration, the calculated
 effluent from the first stage will contain coal with 40 different  pyrite
 concentrations.   Since  each of these coal fractions reacts at a different
 rate,  the  effluent from the second  stage would have 40 times 40 different
 pyrite concentrations.  After 3 stages,, the 64,000 different pyrite con-
 centrations would  exceed the storage capacity of the time sharing  computer.
 It was  decided  that the feed to a stage could be adequately represented by
 five groups of  coal of different pyrite concentrations.  Each coal  group
 was reacted for the 40 time increments to give 200 output pyrite concen-
 centrations would  strain the storage capacity of typical time sharing
 computer systems  and  a  few more stages would exceed the capacity of large
 computers.  It  was decided that the feed to a stage, other than the first
 stage,  could  be adequately represented by five groups of coal of different
 pyrite  concentrations.  Each coal group was reacted for the 40 time in-
 crements to give  200  output pyrite concentrations for a reactor stage.
 These  200  values  were sorted and rearranged by increasing  pyrite concentra-
 tion.   The average pyrite concentrations for the lowest 40, the next  lowest
 40, etc. was  used  to  provide the five starting concentrations for  the next
 stage calculation.  The numeric error introduced by this procedure was not
 determined, but is believed to be negligible.  Greater error  will  certainly
 be caused  in  a  real coal by the inhomogeneity  of pyrite  concentration in
 the starting  coal particles and variation  in particle  residence time  caused
by particle size and density variation.  The departure of  a real coal  from
 ideal particle behavior assumed in  the model would tend  to increase the
real  removal efficiency of a stage  by retarding the progress  of the larger
particles and the more dense high pyrite particles.

                                    182

-------
5.1.3.2  Pressure Effect

     The cost-pressure trade-off was expected to  be a critical  study  in the
process design.  The overall trade-off  is  somewhat complex  since there are
a number of interactions  among the  streams and  equipment that make up the
reactor system.  A simplified diagram is shown  in Figure 25.

     Coal and recycled leach solution are  fed to  a multistage mixer where
some pyrite reaction occurs.  The leach solution  in stream  1 has increased
in total iron and decreased in Y (ratio of ferric iron to total iron) as a
result of the reaction occurring in the mixer.  In the first (primary)
reactor the temperature is set at the design value and oxygen is present
at the design pressure to oxidize ferrous  iron  to ferric iron and thereby
increase the solution Y.  Both the  pyrite  reaction and the  iron oxidation
liberate heat.  Part of the heat is lost from the reactor walls and part is
needed to heat stream 1 from the mixer  temperature to the reactor tempera-
ture.  The balance of the heat is removed  to maintain the reactor at the
design temperature.  This heat is removed  by vaporizing water which is
condensed outside the reactor by contact with the recycle leach solution.
The reactor total pressure is a sum of  the steam  pressure at the temperature
of reaction, the specified oxygen pressure and  the partial  pressure of inert
gas which builds up in the recycle  gas  loop.

     When the reactor oxygen pressure is high,  the fraction of  steam in the
gas is low and high circulation rates are  required to remove the steam
formed to maintain constant reactor temperature.  Thus, high oxygen pressure
increases the pressure rating of the reactor and  the amount of  recycle gas
compression, but it increases the rate  of  reaction by providing a higher Y
and thereby decreases the required residence time and reactor size.

     The effluent from the primary reactor, stream 2, is fed to the secondary
reactor which removes additional pyrite and decreases the Y of  the leach
solution.   If the residence times in the mixer  and the primary  reactor are
preselected in reasonable ranges, then  a residence time for the secondary
reactor can be found which gives the desired pyrite removal.  Figure  26 shows
the time trade-offs between the primary reactor,  R-l, and the secondary
                                    183

-------
00
COAL

OXYGEN

MIXER
i




RECYCLE
SOLUTION
RECYCLE
COMPRESSION
i
J_




PRIMARY
REACTOR
I

HEAT
RECOVERY
©. SECONDARY ©, FeSO^ FeSO4r
REACTOR REMOVAL


                                           Figure 25.   Simplified Reactor System

-------
                                    MIXER: 3 STAGE, 215F, 0.75 HR
                                    R-l:  5 STAGE, 250F
                                    R-2:  10 STAGE, 215F
                                    95% PYRITE REMOVAL
                                    5%FE, L/C =2.0
0
               5             10            15

                  RESIDENCE TIME IN R-l, HOURS
             Figure 26.  Reactor Residence Times as a Function
                        of Oxygen Pressure in R-l
                              185

-------
reactor, R-2, at three oxygen pressures.   A primary reactor is always re-
quired to regenerate the leach solution  since there is not sufficient ferric
iron to react with the desired pyrite.   In principal  it is possible to avoid
a secondary reactor if sufficient time  is provided in the primary reactor.
For example, the figure shows that at 15 psi  oxygen pressure, 18.8 hours of
reaction time in the primary reactor will remove 95 percent of the pyrite
without a secondary reactor.

     Pyrite removal is not the only criteria for selecting pressure and
residence times.  It is also necessary to remove the iron sulfate formed
from the pyrite reaction.  The baseline  design removes as ferrous sulfate
all of the iron formed in the pyrite leaching reaction.  Therefore, the
reactor effluent must contain enough iron in the ferrous form (low enough
Y) to provide this quantity of ferrous  iron.   Referring again to the
simplified reactor system diagram, if the secondary reactor effluent,
stream 3, has a Y of 0.67, then removal  of the ferrous sulfate from re-
action will increase the Y to 0.85 for recycle.  Returning a recycle
solution with a Y higher than about 0.85 presents problems in crystallizing
the ferrous sulfate from a solution which is too rich in ferric sulfate.
Figure 27 shows the influence of residence time in the secondary reactor
on the effluent ferrous iron.  Instead of presenting the data in terms of
Y, the figure gives the available ferrous iron per 100 moles of pyrite
fed based on a recycle Y of 0.85.  Thus, if the coal has  95  percent  pyrite
removal, 95 moles of ferrous iron must be available.   It  is  evident  that
secondary reactor residence times in the vicinity of 35 to 40 hours are
needed to provide this quantity of ferrous iron.

     To find the effect of pressure, computer runs were made at the correct
residence time for 95 percent removal with 95 moles of available  ferrous
iron at pressures from 10 to 50 psi.  The results are as  follows:
                                   186

-------
    REQUIRED AT
    95% REMOVAL
                             R-l TEMPERATURE = 250 F
                             R-l SIZED FOR 95% REMOVAL
                             MIXER FEED: 5% FE, Y = .85
                             FEED SO./FE RATIO =  1.75
                             FEED LIQUID/COAL RATIO = 2
                      30            35

               RESIDENCE TIME IN R-2, HOURS

Figure  27.  Effect of Pressure and Residence Time on Ferrous
           Iron Make

                        187

-------
     Oo  Pressure        Residence Times        Reactor System Cost*
         psi          R-1,  hr     R-2, hr      	$000	
         10           9.21        34.8               2835
         15           7.39       36.7               2814
         20           6.40       37.9               2824
         30           5.30       39.3               2858
         50           4.27       40.7               2938
*
 Includes:   Mixer,  R-1,  R-2,  Knockout  drum  and  recycle  compressor.
These results are presented in Figure  28 which shows that most of the cost
is in the primary and secondary reactors.  It may be seen that the cost
minimum is very broad.  Although minimum cost is near 15 psi of oxygen, it
was concluded that any pressure in this range can be used if it was found
to be more desirable for other reasons.

5.1.3.3  Iron Concentration Effect

     The baseline process has a leach  solution containing 5 weight percent
iron in the mixer feed.  After pyrite  leaching and water evaporation the
iron concentration increases to 6.7 percent.   Likewise the feed ratio of
sulfate to iron is 1.75 which decreases to 1.64 at the secondary reactor
effluent.  Calculations were also performed with a mixer feed of 4 percent
which gave an effluent iron level of 5.6 percent.  The comparative cost
results are as follows (for 15 psi oxygen pressure):
                                            v

                                      5% iron feed         4% iron feed
       Mixer, 3 stage                  0.75 hour            0.75 hour
       R-1, 5 stage                    7.39 hour            8.91 hour
       R-2, 10 stage                  36.7  hour           38.2  hour
       Equipment cost (FOB)
       Mixer                            $118,000             $120,000
       R-1                                896,000            1,074,000
       R-2                              1,708,000            1,804,000
       R-1  Knockout                       28,000                28,000
       Compressor                         64.000                65,000
       Total                           $2,814,000           $3,091,000
                                   188

-------
             TOTAL COST
Z
o
l-

LU

Q_

O
LU
U_
O
LU
o

QL
LU
                  R-2
                  R-l
                MIX VESSEL
              10
                                                            COMPRESSOR
                                                            R-l KNOCKOUT
                         20         30         40
                            OXYGEN PRESSURE, PSI
50
                      Figure 28.  Reactor System Cost as a Function
                                 of Oxygen Pressure
                                   189

-------
The purchased equipment cost increase in the reactor section resulting from
a decrease of 1  percent in iron concentration is seen to be $277,000 or
nearly 10 percent.   Cost impact on downstream portions of the process were
qualitatively examined.  A negligible cost impact on equipment in the sul-
fate and sulfur removal sections was found, although about 10 percent less
wash water would be required to remove the aqueous sulfate leach solution.
The evaporators with a cost of about $0.5 million could be reduced in size
to produce less wash water, but the overall cost reduction would appear to
be not more than $50,000.  It was concluded that total downstream process
equipment cost reduction will be much less than the $277,000 increase in
reactor cost.  Therefore reducing iron concentration will result in a small
increase in process equipment cost.

 5.1.3.4  Oxygen  Purity

      The  evaluation of the effects  of oxygen purity  includes the  effects
 of inert  gas  concentration in  the  purchased oxygen and  the  effects  of inert
 gas  buildup  in  the recycle stream,  but  ignores  any inert gas that may be
 generated  from  the coal  during  leaching.   A simple diagram  of  the oxygen
 system  is  shown  in Figure  29.   At  steady state  the recycle  flow rate is
 set  by  the requirement to  remove  steam  from the reactor for temperature
 control as described in Section 5.1.3.2.   Also  at steady state, the flow
 rate of oxygen  in  the feed must equal the rate  of oxygen consumption plus
 the  flow  rate of oxygen in the vent stream and  the quantity of inert gas
 vented  must  equal  the quantity of inert gas  in  the oxygen feed.  The oxygen
 vented  (wasted)  depends on its concentration  in the  vent stream and there-
 fore its  partial pressure  in  the  reactor.   For  a given coal and reactor
 temperature  and  oxygen pressure,  the oxygen  concentration in the vent
 stream  depends  on  the inert gas buildup in the  reactor.  The larger the
buildup of inerts  in the reactor  the lower will be  the oxygen waste through
the  vent.  However, since  the  reactor temperature and oxygen pressure are
fixed,  the higher  the inert gas buildup the higher  the total pressure and
thus the  reactor pressure  rating  requirement.   Obviously there is a trade-
off  to  be made  between the cost of oxygen wasted an- the cost of the reactor.
For  the purpose  of comparing  the  operating cost of purchased oxygen with
the  capital  cost of the reactor,  the following criteria, based on cost data
 in Section 5.1.2.2, were applied:
                                     190

-------
                                                             -»~VENT
OXYGEN FEED
                      REACTOR
                      3.49 T/HR
                      O2USE
  RECYCLE
COMPRESSOR
                   Figure 29.  Oxygen Circulation Diagram

-------
     •  Oxygen at 99.5 percent purity by gas volume up to 30C psi
        is priced at $25/T.

     •  Annual  cost of capital  is  a  total  of 25  percent which in-
        cluded  depreciation,  maintenance,  insurance,  taxes  and
        interest.

     •  The capital  for installed  reactor equipment is twice  the
        purchase price;  the actual ratio for the baseline case is
        6.36/3.26  -  1.95.
                                                                 J
Table 29 shows the influence of oxygen purity both in the reactor gas and
in the purchased gas on the cost variable components of the reactor system.
For comparison, purchased purities of 99.5 percent and 95 percent were
selected.  The 95 percent pure oxygen is usually priced lower (based on
contained oxygen) because the argon/oxygen separation step is not required.
At a  price of  $20/ton  of contained oxygen for 95 percent purity, the compar-
ison  shows that slightly lower cost can be obtained with the  lesser purity
oxygen gas if the  lower purity oxygen is recycled to  about  50 percent con-
sumption.   The  difference,  which is  only $50,000/yr  (about  $.06/ton of
coal), disappears  if the differential  in oxygen  is actually $3/T of
contained oxygen instead of the $5/T used in Table 29.  When high purity
oxygen is used  recycle ratios can  be set to  give inert gas  buildup in the
range of 5 to 50 percent inert gas in the reactor at  little change in process
cost.   With  95  percent oxygen the  range  is about 15  to 70 percent inert
gas  in the reactor.   The baseline  case was chosen at  5 percent inert gas
in the reactor  using 99.5 percent  oxygen feed to allow for any coal derived
inert gas  generation during the leaching operation.

5.1.3.5  Compressed Air Regeneration

     The cost of oxygen has  increased two or three fold  in the  last five
years and it is necessary to reexamine  the  substitution  of compressed air
for oxygen.  Compressed air could be  of interest if  its  cost in the Meyers
Process were less than oxygen  or  if  it  is similar in cost  but significantly
reduces the energy consumption of the process.

                                    192

-------
          TABLE 29.  EFFECT OF INERT GAS BUILDUP ON REACTOR
                     SECTION ANNUAL COST
    Oxygen Gas Containing 0.5% Inert ($25/T of Op)
Reactor inert gas, %
Reactor pressure, psig
Oxygen vented, T/hr
Equipment cost, $000 (FOB)
R-l
R-l knockout
Compressor
Total
Annual cost, $000*
Capital related
Total Op cost
Total cost, $000
Oxygen Gas Containing
Reactor inert gas, %
Reactor pressure, psig
Oxygen vented, T/hr
Equipment cost, $000 (FOB)
R-l
R-l knockout
Compressor
Total
Annual cost, $000*
Capital related
Total 02 cost
Total cost, $000
2
53
1.14

901
28
60
989

494
913
1407
5% Inert
, 7
54
8.13

908
27
29
964

482
1838
2320
5
54
.37

905
28
64
997

498
760
1258
)
($20/T
10
55
3.15

913
29
54
996

498
1049
1547
10
55
.17

913
29
67
1009

504
712
1225
of 0,)
15
56
1.49

922
30
63
1015

508
786
1294
20
57
.07

931
31
72
1034

517
701
1218
20
57
.93

931
31
68
1030

515
697
1212
50
68
.02

1031
40
90
1161

580
691
1271
50
68
.19

1031
40
89
1160

580
580
1160
80
113
.00

1429
56
123
1608

804
687
1491
80
113
.04

1429
56
123
1608

804
553
1357
90
163
.00

2087
111
142
2340

1170
687
1857
90
163
.00

2087
m
142
1857

1170
550
1720
Power cost for the recycle compression was neglected because it was too

small to affect the cost comparison.
                                   193

-------
     The reactor section was  sized  and  costed  in a  manner similar to the
method used for examining oxygen purity (Section 5.1.3.4).   A summary of
the results is shown in Table 30.   Three oxygen partial  pressures were
included, namely:  15, 10 and 5 psi and for each oxygen  partial  pressure
three levels of oxygen utilization  were examined.  It had been the plan to
examine 75, 50 and 25 percent use of the oxygen, but convergence logic in
the computer model sometimes  prevented  the 25  percent point from being
obtained.  Basically the problem relates to water balancing the reactor
system.

      It is evident from the results that the major trade-off occurs between
reactor cost and compression cost.   For example, at 15 psi of oxygen
partial pressure when the oxygen is 75  percent consumed, the reactor pres-
sure  needed is 338 psig largely from the high  concentration of nitrogen in
the reactor vent gas.   If the oxygen in the air is only 25 percent used the
reactor pressure decreases to 138 psig.  However, while the reactor cost
decreases with decreasing pressure, the compression cost increases and the
optimum at 15 psi of oxygen pressure is between these extremes.

      A second factor is water vapor  loss from the reactor.  The reactor
operating at 250°F and  sized for an overall removal of 95 percent of the
pyrite from a 3.2 percent pyritic sulfur coal  produces excess heat equiva-
lent  to about 12 tons per hour of water vaporized in the reactor.  As  either
the oxygen partial pressure decreases or the percent oxygen utilization
decreases the water vapor in the vent gas increases.  Thus at 15 psi oxygen
pressure oxygen utilization can be decreased to less than 25 percent before
all excess steam is removed, but at 5 psi of oxygen pressure something less
than  45 percent  (probably 35 to  40%) is the lowest utilization  possible.
If the reactor temperature is increased above 250°F or a coal with  less
starting pyrite is used, even greater oxygen utilization than  shown in
Table 30 will  be required.

     The cost comparison shown in Table 30 was made on the assumption  that
the large quantity of pressurized vent gas can provide one-half of  the
energy needed to compress the feed air.  Even with this  optimistic  assumption,
it may be seen that the process costs are higher with air than  the  baseline

                                    194

-------
                           TABLE 30.   EFFECT OF COMPRESSED  AIR  ON  REACTOR SECTION ANNUAL COST
01
Oxygen pressure, psi
Reactor pressure, psig
Oxygen used, % of feed
Water vented, T/hr
Recycle compressor, HP
Air compressor, HP    ^
Equipment cost, $000 (FOB)
   R-l
   R-l Knockout
   Recycle compressor
   Air compressor
   Total
Annual cost, $000
   Capital related
   Total  02  cost
   Compressor  power*
Baseline
99.5% Oo
15
54
90
.0
315
-
905
28
64
-
997
498
760
63
1321
15
338
75
.2
880
3235
3405
242
146
414
4207
2104
-
500
2604
15
188
50
.7
685
4095
2087
109
120
501
2817
1409
-
546
1955
Regeneration
15
138
25
3.1
310
6600
1648
70
64
733
2515
1253
-
722
1975
10
238
75
.4
810
3235
3064
149
137
414
3764
1882
-
486
2368
with Compressed Air
10
138
50
1.1
550
3485
1988
73
100
440
2601
1301
-
459
1760
10
109
30
4.1
200
4880
1680
48
44
576
2348
1174
-
528
1702
5
138
75
.7
625
2455
2909
75
111
333
3428
1714
-
371
2085
5
88
50
2.4
230
2730
2096
55
50
362
2563
1282
-
319
1601
5
83
45
3.5
125
2910
2022
50
30
381
2483
1242
-
316
1558
      Power  cost  assumes $0.025/kw-hr and that 50 percent of air compression  power is recovered by
      expanding the  vent gas through a turbine drive on the compressor.

-------
process with oxygen.   The factor of overall  power use also shows that air
generally is a significantly higher power consumer.   Oxygen is produced at
a total energy consumption of about 345 Kw-hr per ton of oxygen.  This is
equivalent to nearly 1800 horsepower for the oxygen  needed in the baseline
case.  Thus, the baseline which uses a total of about 2100 horsepower for
oxygen production and gas recycle is well below any  of the air fed cases
and only the low pressure, low oxygen utilization case is competitive if
the optimistic recovery of energy from the vent gas  is assumed.  Since no
process testing has been conducted at these low oxygen pressure, high gas
throughput rates, it is entirely possible that solution foaming may occur
or that much greater reactor volumes or residence times may be needed.
The current analysis tends to show that oxygen continues to be much pre-
ferred over compressed air in the process design.

5.1.3.6  Three Reactor Configuration

     In the early stages of the process design, trade-offs were made between
the baseline design and a three reactor configuration.  The concept of the
three reactor configuration was that the first reactor (R-1A) would use
oxygen at the highest pressure in a once-through mode to give partial
reaction and partial regeneration.  The second reactor (R-1B) would con-
tinue the reaction and regeneration using the vent oxygen at a lower pres-
sure and lower reactor temperature also in the once-through mode.  The
third reactor (R-2) would be like the second reactor of the baseline process.
Although all of the process constraints were not known in the early cost
trade-off studies, the trend was for the first reactor of the three reactor
configuration to move toward low pressure and the second reactor to move
toward atmospheric pressure and thus become the two reactor  (baseline)
configuration.  Upon completion of the baseline definition a brief exami-
nation was made of the cost of a three reactor configuration meeting  the
processing constraints (pyrite removal and adequate ferrous  sulfate  in  the
effluent).   Table 31 shows three cases which meet these constraints.   It
can be seen that in each case the equipment cost is higher than the  base-
line case.   A full study which is both costly and time consuming was not
undertaken  since it appeared unlikely that there is a configuration  with a
significantly lower cost than the baseline case.

                                    196

-------
        TABLE 31.   EQUIPMENT  COSTS  FOR A THREE-REACTOR CONFIGURATION

     Operating Conditions
                  -' 	R-IA	             R-1B
     Baseline       7.49 hrs, 250°F, 15 psi 02           None
     Case 1         2.50 hrs, 250°F, 50 psi 02   10 hrs, 220°F, 10 psi02
     Case 2         2.25 hrs, 250°F, 50 psi 02   14 hrs, 220°F, 5  psi 0£
     Case 3         2.00 hrs, 250°F, 50 psi 02   14 hrs, 225°F, 5  psi02
     Equipment Cost, $000  (FOB)
                    Baseline
     Mixer            118
     R-1A             905
     R-1B
     R-2             1677
     Knockout          28
     Compressor        64
Case 1
118
490
825
1677
Case 2
118
450
964
1677
Case 3
118
409
1008
1677
                     2792        3110       3209         3212

 5.1.3.7   Additional  Studies

     While the process design effort attempted to examine all of the para-
meters which were of major  importance to providing a low cost process
design, it considered only  the bench-scale Lower Kittanning coal which had a
fairly high level of pyrite and rather poor filtering qualities.  As data
becomes available on other  coals, new designs should be prepared to ac-
commodate other separation  rates, different pyrite levels, other levels of
pyrite removal, other methods for by-product removal and additional para-
meters identified during the testing and process design.  Since coals may
differ significantly in reaction rate and processing parameters, it is
likely that several different process flow sheets will be needed to accom-
modate the heat and mass balances for a wide spectrum of coals.  It is
likely that the improvements in reaction and filtration rates that appears
to come with mechanical precleaning of coal will lead to a still lower cost
process.

                                    197

-------
 5.2   COARSE  COAL  (-1/4  INCH) PROCESSING

      The bench-scale efforts relative to the processing of coarse coal
 (see  Section 4) have been principally directed toward the acquisition of
 data  and information on pressurized reaction systems.  Comparisons of data
 for pressurized and unpressurized systems indicate that the imposition of
 pressure does not significantly influence the pyrite leach rate from coarse
 coal.  Because the use of non-pressurized reaction vessels and support
 equipment significantly reduces capital investment requirements, a decision
 was made to  formulate a conceptual coarse coal process design on that basis.
 The rationale, design details and estimated economics of this design are
 presented in the following subsections.

 5.2.1 Concept Development

      The physical characteristics and laboratory-derived leach reaction
 rates of coarse coal required the development of a completely different
 processing concept than that conceived for fine (suspendable) coal.   It
 is, of course, understood that one of the primary virtues of coarse coal
 processing is that the product is shippable, handleable and storable by
 conventional means.  In terms of processing, coarse coal usage implies less
 pre-leaching preparation, the elimination of requirements for product pellet-
 izing or briquetting operations and the potential application of conven-
 tional physical separation systems to augment the productivity/investment
 ratio of the Meyers Process.

     The major physical characteristic of coarse coal which necessitates
modification of the fine coal processing concept is the tendency of the
 1/4" x 0 coal grind to separate.  The fines (approximately 48 mesh x  0)
are nominally slurryable, while the larger mesh fraction (about 1/4"  x
48 mesh)  settles out quite rapidly.  It is therefore impractical to consider
slurry transportation of the coal; because of the tendency for the coal  and
leach or wash liquors to separate it is in fact much more efficient to
retain the coal in a fixed position and allow the liquors to  pass  through
the reacting mass (as in a packed bed) and to transport the coal mechanically
by means  of conveyor systems.
                                    198

-------
     While the basic leaching reaction  established  during  fine  coal  in-
vestigations is fully applicable, the range  of  particle  sizes included in
the nominal 1/4" x 0 grind gives rise to  a corresponding range  of reaction
rates; the fines react in basic accordance with the rate constants derived
in the fine coal evaluations, but the coarse particles are considerably
slower.  Since, as indicated earlier, there  is  little rate constant  ad-
vantage offered by the use of pressurized reaction  vessels, the coarse
coal conceptual process must allow  for  relatively long coal retention
periods.  The leaching residence time may be shortened by  preprocessing
the coarse ROM coal through conventional  physical cleaning (scalping)
which offers a means of segregating a small, extremely pyrite rich and
difficult to leach, fraction of coal for  direct disposal.  Further economic
advantages are indicated when the scalped coal  is deep cleaned  into  two
fractions, one lean in pyrite that  would  not require depyritization  and
one rich in pyrite to serve as feed stream to the Meyers Process.  The
recombined product would meet the sulfur  standards  even  though  only  approxi-
mately one half of the ROM coal was subjected to the Meyers Process.

     A total of four reactor configurations  were identified as  potentially
viable.  These include an above-ground  batch reactor; a  lined pit batch
reactor; a continuous countercurrent reactor; and a continuous  cocurrent
reactor.  All of these possibilities were conceived as atmospheric pressure
units in which only the leach reaction  occurs;  because of  the relatively
easy coarse coal-from-liquid separation,  the regeneration  reaction can be
accomplished in a separate vessel optimized  for that purpose.

     The four reactor configurations were simulated  with computer models
and evaluated in detail.  It quickly became  apparent that  the above-ground
batch reactor scheme would be the most  expensive and complex of the group
and it was  therefore eliminated from consideration  on that basis.  It was
also found  that the continuous countercurrent reactor requires  a smaller
vessel  and  a lower leach solution throughput than the cocurrent reactor,
therefore the continuous cocurrent  reactor scheme was also eliminated.
                                    199

-------
      Both the  continuous countercurrent and batch pit reactor  schemes were
 expanded to  complete  conceptual flow diagrams for further  study.  The pro-
 cess, as described  in the following paragraphs, could incorporate either
 reactor type and  could be preceded by a physical separation  train.  The
 conceptual design is  presented schematically in Figure  30.   A  brief
 description  of the  conceptual design follows  with  details of  the key ele-
 ments being  presented in subsequent subsections.

      Coal is fed  to the reactor,  stream 1, at a design  rate  of 100  TPH
 where it contacts a continuously  regenerated stream of  leach solution,
 stream 2, of YENTER = 0.95.   Spent leach  solution,  containing  a portion
 of the fines fraction from  the coal feed  is withdrawn from the reactor,
 stream 3, and  blended with  the liquid  stream 17 recovered  from the  coal
 drain.  Stream 17 also contains a portion of the fines  fed to  the reactor
 via stream 1 which  are acquired during the drain and rinse cycle.   The
 weak leach solution-coal slurry is bled,  stream 5,  at a rate sufficient to
 maintain the total  iron content of 4 percent and to maintain the fines
 content of the leach solution loop at  a maximum of  10 percent.  The fines
 in bleed streams  are removed  by filtration with the filter cake being con-
 veyed, stream  7,  for reinjection  into  the coal processing  sequence  at the
 water wash step.  The filtrate is pumped,  stream 6, to  a unit  which converts
 the major portion of the Fe  to  Fe    by  iron reduction.  The  Fe+2  rich
 stream is then sent to an evaporator/crystallizer unit, stream 9, which
 produces the clean  water used in  the coal wash, stream  12.  The ferrous
 salt is filtered  from the crystal!izer output,  stream  13,  and  disposed  of
 through liming, stream 14.  Residual solution, taken as filtrate,  is routed
 to the regenerator, stream  11, together with the spent  leach solution which
 bypasses the bleed  stream diverter, stream 10.  Regeneration of the leach
 solution is  accomplished as outlined in Section 5.1.1  relative to fine
 coal  processing;  oxygen is  supplied, stream 15, to  convert the Fe+2 ion
 and  sulfuric acid to  Fe  and water.   The freshly regenerated  leach solu-
 tion,  stream 2, completes the loop through reinjection  into  the reactor.

     The leach  solution-wet coal  leaving the reactor, stream 4, is  drained
                                         w
on a conveyor and  spray rinsed with leach solution-contaminated water,
                                    200

-------
                                                            COAL FINES
r\>
o
                       SCRAP IRON
                         STEAM
                                      COAL-
                   AMBIENT
                   REACTION
                   SYSTEM
                   (PIT OR
                   CONTINUOUS)
LEACH
SOLUTION
DRAIN AND
RINSE
WATER
WASH AND
DRAIN
                                                                                                                    CENTRIFUGE
EVAPORATOR
CRYSTALLIZER
                                        WASTE SALT
                                        TO LIME PIT
                                                                                                                       DRYER
                                                                              SULFUR
                                                                              DISTILLATION
                                                                              AND COOL
                                                                                                                      PRODUCT
                                                                                                                      COAL TO
                                                                                                                      STORAGE
                                                                                                                      AND
                                                                                                                      SHIPPING
                                                         SULFUR
                                                         CONDENSE
                                                         AND
                                                         DISPOSAL
                                              Figure 30.   Coarse  Coal  Process  Schematic

-------
 stream 18.  This rinse water stream is the filtrate obtained by removing
 fines from stream 19.  The wet fines are conveyed directly to the dryer
 via  stream 20.

     The wash vessel is sized for a nominal 1-hour residence time to permit
 transfer of the leach solution trapped in coal pores to the passing water
 stream, which originates as evaporator and dryer condensates, streams 12 and
 24.  Following a conveyor draining step similar to that employed earlier to
 remove excess leach solution, the coal, stream 22, is transported to a
 continuous centrifuge which reduces the total water content to about 20-22
 percent with the centrate, stream 25, going to the fines filter.  The coal
 is conveyed, stream 23, to a two stage inert atmosphere dryer/distillation
 unit where water  is  first  removed,  stream  24,  followed by  elemental  sulfur
 removal,  stream 26.  The elemental  sulfur  is cast  into blocks for disposal
 and  the  coal,  stream 27, is  cooled  for preshipment storage.

 5.2.2  Conceptual Process Details

 5.2.2.1  Reactor Section - Pit and Continuous

     Both the batch pit and continuous countercurrent reactor concepts were
 sized to accommodate 100 TPH of 1/4" x 0 ground coal.  In the case of the
 pit  reactor, three identical units are required such that filling, leach
 reaction and raw product coal withdrawal  can proceed concurrently; the
 withdrawal rate, which sizes the balance of the process units, is 100 TPH.
 The  pit reactor configuration, upon which the conceptual  process design is
 based, is shown schematically in Figure 31  .  As can be seen, the sectional
 view indicates a basically triangular pit with the sides angled at 30°
 such that the natural angle of repose of wet coal is exceeded and the coal
 can  be withdrawn via the conveyor at the bottom.  The pit and drain  pump
 lining material is acid-resistant concrete and the metal parts and various
 exposed equipment items are of stainless-clad mild steel.  Each pit  is  sized
 to hold sufficient coal for 5 days of operation, i.e., 12000 tons at the  100
TPH withdrawal  rate.  Using a bulk density of 47 lb/ft3, the reactor volume
 is 511,000 cubic feet; for a 300 foot length,  the  tank width  is  about 43 feet
 and the depth about 80 feet.

                                    202

-------
                               REMOVABLE
                             LEACH SOLUTION
                               SPRAY RACK
ro
o
CO
                                                               ACID RESISTANT CONCRETE
                                                       LEACH SOLUTION COLLECTION SUMP
                                               Figure 31.    Pit Reactor Schematic

-------
     The leach solution is distributed over the upper surface of the coal
mass by means of a hinged spray rack which can be moved for pit load
operations.  The leach solution drains through the coal bed, is collected
in a full-length sump, and is circulated back to the regenerator.  Fol-
lowing completion of the leach reaction, coal is withdrawn from the pit
reactor by opening the hydraulically-operated doors comprising the lower
one-third of the reactor wall; this allows the coal to fall onto a nominal
48 mesh conveyor and to drain residual leach solution while being trans-
ported to the rinse section of the plant.

     The five day pit capacity imposes a similar limit on the leach reaction
period.  Based on bench-scale and laboratory data, rate constants were es-
timated which show that about three days of leaching will be required to
reduce pyrite to target level.  The design over-capacity allows for the use
of coals having a higher pyrite content or the processing of a pyrite-rich
fraction from a front-end physical separation/pyrite concentration unit as
described earlier.

     All three pit reactors are serviced by a track-mounted reclaim-
conveyor system having a 120 TPH capacity; this unit reclaims coal from a
pile and distributes it along the length of each pit.  The 120 TPH capacity
assures maintenance of any operating cycle time within the 5-day maximum
imposed by the reactor design.

     The ambient-pressure continuous countercurrent reactor system concept
was also sized for a 100 TPH feed rate.  Each reactor  (see Figure 32) con-
sists of a cylindrical outer shell having a conical bottom and a discharge
lock hopper at the lower end of the cone.  A raised lip  surrounds the upper
periphery of the cylindrical shell, stainless steel clad is used over all
interior surfaces, a spreader cone at the top distributes the feed coal
over the cylindrical area and a series of internal baffle plates limits
the formation of any flow channels which may develop.  The conical bottom
section houses the leach solution distribution orifices  through which
leach solution is forced under pressure to flow upward through the coal bed.

     Operationally, coal is loaded at the top of each  reactor at the design
feed rate and is distributed by the spreader plate; continuous withdrawal
                                    204

-------
LEACH SOLUTION
 OVERFLOW TROUGH
COAL FILL DISTRIBUTOR
PLATE
 SUPPORT
STRUCTURE
                                                TO
                                           LOCK-HOPPER FOR
                                           COAL REMOVAL
                                                             COAL MIXING
                                                             BAFFLES
                                                               LEACH SOLUTION
                                                               INJECTOR MANIFOLD
                   Figure 32.  Continuous  Reactor Schematic


                                      205

-------
through the base lock hopper provides a nominally uniform downward travel
of the coal.  The design residence time for processing feed coal having a
3.2 percent pyritic sulfur content is 50 hours.  The reactor system con-
sists of five vessels each 20 feet in diameter and holding 1000 tons of
coal.  With feed and withdrawal rates of 20 TPH, coal leaching occurs in
about 80 feet of vessel length.  Leach solution at a rate of 286 TPH flows
upward through the coal at a rate of less than 0.01 ft/sec.  The feed
leach solution has a Y of 0.95 and 4 percent iron while the effluent
solution has a Y of 0.5 and 4.2 percent iron.

5.2.2.2  Regenerator Section - Pit and Continuous

     As the regeneration reaction is conducted in a separate vessel for
both the pit and continuous leaching concepts, it is possible to optimize
the design for that purpose.  This is in contrast to the fine coal process
wherein both the leaching and regeneration reactions take place concurrently
in the same vessel, necessitating off-optimum conditions for each.  As
mentioned earlier, the leach solution is actually a 10 percent fine coal
slurry therefore both regenerator designs must incorporate sufficient
mechanical agitation to maintain the slurry and minimize settling tendencies.

     The pit reactor, being a batch-mode operation, will  deplete the leach
solution to different degrees throughout the period of the leach reaction.
This in turn imposes a variable load on the regenerator with the peak load
occurring at the start of the leaching reaction.  The regenerator has been
sized to accommodate the peak, resulting in excess capacity for the major
portion of the coal reaction period.

     As a result of a trade-off/optimization study, the regenerator for the
pit reactor process has been sized at a total capacity of 62,000 cubic feet
at a working pressure of 150 psia and a working temperature of 250°F.  The
vessel  is  rubber and acid brick-lined and contains three wier-separated
stages  of equal  volume.  The vessel  accepts the 1620 TPH of leach solution
flow rate  required by the pit reactor and regenerates it to a minimum Y  =
0.95.
                                    206

-------
     The regenerator required by the  continuous reactor Is designed to accept
a constant input Y of 0.5 and to regenerate  the leach  solution to an out-
put level of Y = 0.95.  As was the case with the pit system regenerator, the
vessel is rubber and acid brick-lined and  is designed  for a 150 psia working
pressure at 250°F, and contains three equal  volume stages.  However, the
leach solution flow rate is lower  (1430 TPH) requiring a vessel of about
55000 cubic feet.

     Both regenerators are operated at well over the atmospheric  boiling
point of the solution and are therefore good sources of low-pressure steam.
In addition to process heat, a portion of the low-pressure steam  is  used to
drive an aspirator which pulls a mild vacuum on the two fines filters,  as
will be discussed in Section 5.2.2.4.

5.2.2.3  Sulfate Removal Section

     The reaction with pyrite produces 1.0 mole of iron and 1.2 moles of
sulfate from each mole of pyrite reacted.  In the fine coal process,
Section 5.1.2, the iron and 1.0 mole  of the  sulfate are removed as FeS04
while the excess 0.2 mole of sulfate  is neutralized with lime to  form
gypsum.   An alternative approach was  incorporated in the coarse coal  pro-
cess.  This approach involves dissolving scrap-iron into the leach solution
by the reaction:
                     o
                               +3   .  n c c +2
               0.2 Fe  + 0.4 Fe *  •*•  0.6 Fe*                 (22)

Then all the reaction products  are  removed  as  1.2 moles of  Fes04-  This
section of the process  (see  Figure  33)  is the  same for both the pit and
continuous leaching concepts.   In processing coarse coal, the  best access
to the spent leach solution  is  just downstream of the fines filter (Stream
6, Figure 30) at which  time  essentially no  fines are present.  The entire
fines-free bleed stream is pumped to  a  heater  conversion tank  having a 3-
hour residence capacity.  This  tank is  filled  with scrap iron.  The Fe
+ Fe+2 conversion is based on an  input  Y of 0.5 and an effluent Y near
zero.  The converter effluent stream  is pumped to an evaporator/crystallizer
                                    207

-------
ro
o
CO
        FROM REACTOR,     _
        DRAIN/RINSE    _   (TO)
                                  REGENERATOR
                                                           TO WASH
                      -Y = 0.5
                                                                   EVAPORATOR
                                   CONVERSION TANK
                                            v\
          TO WATER WASH -4
STEAM,
SCRAP IRON
y
                                        TO REGENERATOR
                                                      
-------
train which reduces the water content and  increases  the  iron content to
above the saturation level.  The overhead  water  product,  Stream 12, is
sent to the coal wash line and the salt-laden  stream to  a leaf filter
                      +2
which separates the Fe   salt.  All residual streams are  combined and
pumped to the regenerator.  The ferrous  sulfate  salt is  sent to disposal.

5.2.2.4  Coal Washing and Filtration Sections

     In both the pit and continuous reactor systems, similar schemes for
the separation of coal fines and for the washing of  the  raw leached coal
product are envisioned.  Fine coal must  be separated at  two points in each
of the conceptual designs (see Figure 30).  The  first provides a means of
controlling the fines content of the leach s'olution  at a  maximum of 10
percent; this involves filtration of the spent leach solution bleed stream,
using a drum-type vacuum filter.  The wet  fines  cake is  conveyed to the
water wash line and the filtrate is transferred  to the excess salt removal
section.

     The second fines recovery operation,  again  employing a rotary drum
vacuum filter, operates on the wash water  circuit  (Stream 19, Figure 30).
This filter separates a water-wet fines  cake which is transported to the
dryer.  The filtrate is recycled for use as coal rinse.

     Analyses of typical 1/4" x 0 coals  have indicated that the fines con-
tent can be as high as 30 percent of the bulk  mass although 15 to 25 percent
is typical.  The filters were sized in order to  accommodate the extreme,
but with the understanding that this represents  an overcapacity in terms
of most applications.  It should be noted  that considerable in-process
separation of fines is to be expected in the reaction, washing, rinsing and
centrifugation steps.  The degree of fine  particle entrapment to be ex-
pected is not known at this time.  In some instances the  fines (in the
continuous reactor) may separate upon feeding  from the distribution plate,
thereby immediately loading the leach circuit  return line, in other cases
the fines may be carried through to the  rinse/wash line  and separated at
that point.  It therefore follows that a degree  of overcapacity is war-
ranted throughout the fines removal and  transportation operations in order
to allow for normal feed coal variations.
                                    209

-------
 5.2.2.5   Sulfur Removal Section

     Both the residual  water and free sulfur are removed from the coal  in
a two-stage inert atmosphere dryer system.   In the first stage, water
amounting to a nominal  16.5 TPH is quantitatively removed as the feed is
heated to 480°F.   In the second stage the coal mass is held at 480°F for
                                                                  x
1 hour to permit distillation of the elemental sulfur.  The clean water is
recovered, cooled and pumped to the wash  train water feed; sulfur is
recovered from the inert gas atmosphere and cast into blocks for sale or
disposal.  The product  coal is cooled and transported to a temporary
storage area to await on-site use or shipment.

5.2.2.6  Energy Balance

     An energy balance around the process indicates it is possible to
eliminate all battery limit heat supply requirements with the exception of
the dryer.  The reactor (either pit or continuous) temperature is maintained
by conserving the heat of reaction and by the sensible heat supplied by the
freshly regenerated leach solution.  The evaporator, conversion tank and
crystal!izer heat requirements are supplied by the heat captured from the
leach solution flash and blowdown vessel.  The regeneration reaction is
quite exothermic and by operating the regenerator at  150  psi and 250°F,
well above the solution boiling point, the subsequent pressure let down to
ambient pressure results in the generation of large quantities of low-
pressure steam.  The quantity of  steam is sufficient  to operate  an ejector
which pulls a mild vacuum on the evaporator and the two fine-coal filters
                                                                           s
and also to fulfill the latent and sensible heat requirements  of the
hardware items mentioned earlier.  The dryer  heat requirements,  expressed
in terms of the quantity of product coal required, were  found  to be
9.5 TPH coal.

5.2.3  Process Cost Estimate

     Processing costs,  including both estimates of capital  requirements  and
annualized costs, have been calculated for both the pit and continuous pro-
cess  concepts.   The basis for calculation was June 1975  in  both  instances

                                    210

-------
For purposes of comparison to process  costs  for  fine  coal  processing (as
presented earlier in Section 5.1.2.2), the  plant capital  requirement was
assumed to be that for a "battery  limits" facility.   That  is, the economics
to be presented here do not account  for plant  off-sites and do not repre-
sent a  "grass roots" operation.   The  logic  behind  this economic treatment,
the information sources, and the method of treatment  are the same as pre-
sented in the fine coal section.

Coarse Coal Capital Cost Estimate
     Process cost estimates are presented for  the four process schemes
employing the two reactor system concepts described earlier.  The four
approaches, pit, continuous, pit/float-sink  combination, and continuous/
float-sink combination were evaluated  on the basis  of a Meyers Process
feed rate of 100 TPH coal containing 3.2 percent pyritic sulfur.

     The objective of front-end physical separation equipment in the float
sink combination options is to isolate a low pyrite segment of the total
feed which does not require further  treatment  and to  concentrate the pyrite
in a "sink" fraction which can be  fed  to the coarse coal Meyers Process.
Float-sink separation data for approximately 400 Appalachian coals were
examined.  It was found that on the  average  each of these  coals could be
separated into two approximately equal weight  portions by  float/sink
separation in a liquid medium with a specific  gravity of 1.3.  Further,
it was noted that in 227 of the coals  the float  fraction contained less
than 1 percent total sulfur.  For  these 227  coals the data shows that
the 1.3 specific gravity "float" fraction amounts to  nearly 55 percent of
the 3/8 inch by zero sample with a standard  deviation of about 15 percent.
The average total sulfur content of  this float fraction is 0.8 percent with
a standard deviation of 0.1 percent

     Float-sink equipment trains are presently in widespread use, but use
higher specific gravity media to achieve float fractions in the 90 percent
range, thereby minimizing losses and maximizing  the saleable product.  It
is apparent that if a separation were  made at  low specific gravity as
suggested above but without the benefit of a Meyers Process for treating
the "sink" fraction, the gross effect  would  be a virtual doubling of coal
prices as well  as the generation of  about five times  as much spoil re-
quiring acceptable disposal.

-------
     In combining the float-sink treatment with the Meyers coarse coal
process it was assumed that a 50-50 float-sink division of the feed would
represent a reasonable design value, based on the analysis described above.
Further analysis of these coals indicated that the sink fraction to be
used as process feed would contain about 3.8 percent pyritic sulfur.  In
a commercial application an economic trade-off would be made at this point
to establish the most advantageous position in terms of the adviseability of
blending the "float" fraction with the process product (thereby reducing
the degree of pyrite removal in the process) versus adjustment of the media
specific gravity to control the quantity of "sink" fraction fed to the pro-
cess.  In the conceptual designs considered here, no reblending is considered
in order to facilitate direct comparison with the fine coal process.

     An equipment list detailing installed equipment costs has been prepared
for 100 TPH pit and continuous coarse coal battery limits process options
and is presented in Table 32.  As may be seen, the installed equipment capital
required for the pit process option is $4.17 million while that required for
the continuous process option is about 50 percent higher.  It should be
noted that for the pit and continuous battery limits processes, the reactor
sections account for approximately $1.54 million and $3.95 million,  respec-
tively, or 37 percent and 60 percent of the total capital requirement
(compared to 48 percent for the fine coal processing base case).  The
additional installed capital investment required for the addition of float-
sink options is estimated to be $.42 million.

Operating Cost Estimate

     The process operating costs have been estimated for the four options
and are summarized in Table  33.   It may be seen that the operating  costs
for the coarse coal options range from $2.94 per ton of coal processed  to
$6.28 per ton.  These  are comparable to the $8.94 per ton determined for the
fine coal  processing scheme  (Section 5.1.2.2).   However,  it  must be stressed
that the operating costs presented in Table  33   are based on process
designs with slightly varying bases.   These differences  are the following:
                                    212

-------
           TABLE 32.   COARSE  COAL  PROCESS  EQUIPMENT  LISTS
                                            Pit Reactor    Continuous  Reactor
      Item                                     $1000       	$1000
Reactor                                         728                3175
Regenerator and Exchanger                       658                 625
Compressor                                        6                   6
Wash Vessel                                      91                 91
Centrifuge                                       79                 79
Knock-out Pot                                    10                 10
Regenerator Agitators                           106                 106
Flash Vessel  and Exchanger                       21                 21
Regenerator Pumps                                 8                   8
Leach Solution Storage                           18                 24
Leach Surge Tank                                  8                   8
Iron Conversion Tank                             24                 24
Iron Conversion Tank Exchanger                   70                 70
Iron Conversion Tank Pump                         2                   2
Fines Filter (Leach Circuit)                     65                 65
Salt Filter (Plate)                              32                 32
Evaporator                                      556                 556
Crystal!izer                                     71                 71
Fines Cake Conveyors (2)                         11                 H
Crystallizer Feed/Exit Pumps                      4                   4
Evaporator Feed/Exit Pumps                       11                 H
Evaporator/Filter Ejector                        12                 12
Evaporator Condensate Vessel                      4                   ^
Coal Rinse Conveyor and Sump                     62                 62
Wash Water Feed/Exit Pumps                        6                   6
Wash Rack Pump                                    1                   1
Centrifuge Fines Filter                          64                 64
Wash Vessel Drain Conveyor                       30
Centrifuge Liquid Exit Pump                       1                   ^
Sulfur Trap                                      18                 18
Evaporator and Dryer Condensate Hold Tank        12
Dryer (2 Units)                               _1381. -              1381
       Process Equipment TOTALS               4,170                '

                                     213

-------
                             TABLE 33.   ANNUALIZED  COSTS FOR BATTERY  LIMITS DESULFURIZATION  PLANTS

Cost Element - $1000/Yr

Capital Related Costs
Depreciation - 10%
Maintenance, Insurance, Taxes
Labor

Utilities
Electricity (25 Mil/KW-HR)

Heating-Coal Equivalent

Water-Process and Cooling
Materials
Oxygen 99.5% ($25/Ton)
Lime ($28/Ton)
Scrap Iron ($50/Ton)
TOTAL
Cost/Ton Feed - $/Ton
% Coal Yield (Weight Basis)d
% Coal Yield (Btu Basis)d
PROCESS OPTION
Pit Reactor
Process9

417
626
900
(6 Positions)

600
(3,000 KW)
(9 TPH)
(247 MM Btu/Hr)
458C

800
112
280
4,193
$5.25/Ton
85%
90%
Continuous Reactor
Process3

659
989
900
(6 Positions)

600
(3,000 KW)
(9 TPH)
(247 MM Btu/Hr)
458C

800
112
280
4,798
$6. 01 /Ton
85%
90%
Pi t/Fl oat-Sink
Combination^

459
689
1,200
(8 Positions)

700
(3,500 KW)
(9 TPH)
(247 MM But/Hr)
458C

800
112
280
4,698
$2.94/Ton
92%
95%
Conti nuous/Fl oat-Si nk
Combination'3

701
1052
1,200
(8 Positions)

700
(3,500 KW)
(9 TPH)
(247 MM Btu/Hr)
458C

800
112
280
5,303
$3. 31 /Ton
92%
95% •
Fine Coal
Process8

1 ,326
1,989
1,200
(8 Positions)

1,500
(7,500 KW)
(4 TPH)
(97 MM Btu/Hr)
241 c

780
112
-
7,148
$8.94/Ton
90%
94%
 PO
,4s.
                    Requirements for 100 TPH coal  feed
                    Requirements for 200 TPH combined coal  feed
                    Based on evaporation to reject waste heat
                    Coal  yield values reflect the  heating requirements of the process

-------
     1)  The pit reactor, continuous  reactor  and  fine  coal  processes
         are 100 TPH processing facilities  for  3.2  percent  pyritic
         sulfur coal while the two  float-sink options  are 200 TPH
         units using cleaned coal.

     2)  The pit/float-sink and continuous/float-sink  processes as-
         sume the feed coal was precleaned  to about 2  percent
         pyritic sulfur prior to  gravity separation.   The resultant
         gravity separated fractions  are assumed  to contain 3.8
         percent pyritic  sulfur  (Meyers  Process treated) and less
         than 0.5 percent pyritic sulfur (by-passed around  the
         Meyers Process).

     The coal yield data  (weight  basis)  presented in Table  33  indicates a
 higher yield for the combined float-sink options  (92 percent) than for the
 pit and continuous reactor processes  options  (85  percent).   This is based
 on the design criteria that all of the material bypassed around the Meyers
 Process, 50 percent of the 200 TPH feed, is 100 percent recovered and
                                                                            I
 blended with coarse coal  Meyers   Process product.  The coal  yield in terms
 of energy recovery is also based  on the  above stated design criteria.

 5.3 PROJECTION OF PROCESS ECONOMICS

     Several Meyers Process options were evaluated  to  determine their over-
 all  process economics.  The processing options  evaluated included both fine
 and  coarse coal approaches (Sections  5.1  arid  5.2) integrated into grass roots
 coal desulfurization facilities.   This section  presents a description of the
 various integrated process options, a discussion  of the economics evaluation
 approach,  and results of  the analyses.

 5.3.1  Processing Option  Description

     The  integrated desulfurization facilities  include  all  battery limits
Meyers  Process  equipment  in either  fine  or  coarse coal  treatment configura-
tion, discussed in Sections 5.1 and 5.2  plus  the  required off-sites.  The
off-sites  include such items as:
                                    215

-------
     •  Feed and product  coal  storage,  handling  and  transport equipment.

     •  Physical coal  cleaning facilities  and  size separation equipment.

     •  By-product handling and storage facilities.

     •  Waste treatment (physical  cleaning and process generated)
        and storage facilities.

     t  Process water treatment, storage and pumping facilities.

     •  Cooling water treatment and pumping equipment.

     •  Power and steam generation facilities.

     •  Site office buildings and shop  structures.

     •  Other site improvements such as roads, fences, railroad
        spurs, etc.

It should be noted that the economic evaluations do  not include land costs
and assume that oxygen is purchased as  an  over-the-fence utility item (i.e.,
neither battery limit nor off-site equipment include an oxygen plant).

     For purposes of economic evaluation,  four differing process configura-
tion cases were developed.   The central basis for each case was a 100 ton
per hour Meyers Process unit.   A brief  description of each case is presented
below and block diagrams  containing simplified mass  balances for each case
are presented in Figure 34.

Case 1 - Cleaned Fine Coal  Case
                                                                       /
      Run-of-mine  (ROM) coal  is physically cleaned and then reduced to
14 mesh top  size.  The unit feed  rate  is  120  tons per hour of coal con-
taining about 20  percent ash  and  3  to  4 percent pyritic sulfur.   The
cleaning plant  refuse  (20  tons per  hour)  is assumed to contain approxi-
mately 75 percent ash and  10  to 14  percent  pyritic  sulfur.   The Meyers
                                    216

-------
          CASE 1 - CLEANED FINE COAL (14 MESH TOP SIZE)
FEED COAl
120 T/HR COAL
20% ASH
3 - 4% PYRITIC
SULFUR
23.6 x 106
MM BTU/YR

ASH D
PHYSICAL
CLEANING
ISCARD
20 T/HR COAL
75% ASH
10 - 14% PYRITIC SULFl
1.4x 10 MM BTU/YR

100 T/HR COAL
9% ASH
1.5 -2% PYRITIC SULFUR
22. 2 x 106MM BTU/YR
JR
MEYERS PROCESS
FINE COAL
CONFIGURATION

COAL PRODUCT
93 T/HR COAL
6% ASH
.1% PYRITIC SULFUR
21. 3 x 106 MM BTU/YR
          CASE 2 - RUN-OF-MINE COARSE COAL (1/4 IN. TOP SIZE)
FEED COAL ^
100 T/HR COAL
20% ASH
3 - 4% PYRITIC SULFUR
19.7x 106 MM BTUAR
MEYERS PROCESS
COARSE COAL
CONFIGURATION

COAL PRODUCT
85 T/HR COAL
15% ASH
.2% PYRITIC SULFUR
17.5x 106 MM BTU/YR
ro
          CASES 3 AND 4,  - DEEP CLEANED FINE AND COARSE COAL WITH 50% MEYERS PROCESS BYPASS
FEED COAL
240 T/HR COAL
20% ASH
3 - 4% PYRITIC SULFUR
47. 2 x 106 MM BTU/YR
ASH D1S
PHYS
CLEAt
CARD
ICAL
>
-------
Process feed consists of 100 tons per hour of coal containing approximately
9 percent ash and 1.5 to 2 percent pyritic sulfur.  The fine coal Meyers
Process configuration utilized in this case is essentially that previously
described in Section 5.1.2 with the exception that the reaction and filtra-
associated equipment requirements are significantly reduced due to expected
reaction rate improvements (2 to 3 times ROM fine coal processing base
case described in Section 5.1.1) and lowered ash contents.  The processing
uses about 100 MM Btu/hr equal to about 4 TPH and the pyrite removed equals
3 TPH.  Therefore the product rate is 93 tons per hour"of coal containing
about 6 percent ash and 0.1 percent pyritic sulfur (93-95 percent removal
of pyritic sulfur).

Case 2 - ROM Coarse Coal Case
     ROM 1/4 inch top size coal is fed directly to a coarse coal continuous
Meyers Process (described in Section 5.2).  The coal feed rate is 100 tons
per hour and the coal consists of 20 percent ash and 3 to 4 percent pyritic
sulfur.  Processing removes about 6 TPH of pyrite and requires about 9 TPH
of coal to supply the 250 MM Btu/hr required for process heat.  The pyritic
sulfur is approximately 90-95 percent removed yielding a product rate of
85 tons per hour of coal containing 15 percent ash and about 0.2 percent
pyritic sulfur.

Case 3 - Deep Cleaned Fine Coal with 50 Percent Meyers Process Bypass
     ROM coarse coal containing 20 percent ash and 3 to 4 percent pyritic
sulfur is fed to a physical cleaning plant at a rate of 240 tons per hour.
The ash discard (40 tons per hour) contains 75 percent ash and 10 to 14
percent pyritic sulfur.  The cleaned  coal containing  9 percent  ash and  1.5
to 2 percent pyritic sulfur is fed to a gravity separation unit  at a rate
of 200 tons per hour.  The heavy fraction consists of  100 tons per hour  of
coal containing 15-18 percent ash and 3 to 4 percent pyritic  sulfur while
the light portion consists of 100 tons per hour of coal with  low ash and
pyritic sulfur content.  The heavy fraction then  reduced  to 14 mesh top
size and fed to a fine coal Meyers Process unit  (described  in Section  5.1)
which produces 90 tons per hour of product coal containing  10-13 percent
ash and 0.2 percent pyritic sulfur.  The Meyers Process  product  when

                                    218

-------
blended with the bypassed light coal fraction yields  a  combined product
stream of 190 tons per hour of coal containing  about  6  percent ash and
0.2 percent pyritic sulfur (overall 90-95  percent  removal of  pyritic sul-
fur from feed coal).

Case 4 - Deep Cleaned Coarse Coal with 50  Percent  Process Bypass
     Two hundered and forty tons per hour  of 1/4 inch top size ROM coal
is treated as previously discussed  in Case 3.   The only difference in the
treatment scheme is that a continuous coarse coal  Meyers Process configura-
tion (discussed in Section 5.2) is  utilized.  The  coarse coal process
requires more coal for internal process  heat and therefore yields only
185 TPH of product.

5.3.2  Economi cs Model Descri pti on

     The technique used in capitalizing  the process options (Cases 1, 2, 3
and 4) is a primary determinant in  calculating  the unit price of the product.
Both utility financing and investor financing methods were considered in
determining the price of the product coal.

     The method used  to determine the product cost for  the different
financing techniques  was based on the technique used  by the FPC Synthetic
Gas-Coal Task Force in their report on synthetic gas.    The applicable
equations used in this analysis are summarized  below.

     The investor capitalization method  used in this  analysis was the
"Discounted Cash Flow" (DCF) financing method with an assumed discounted
cash flow rate of return of 20 percent after taxes.   In essence, this
technique determines  the annual revenue  during  the  plant life which will
generate a discounted cash flow equal to the total capital invested for
the plant.  For this  analysis, it was assumed that the  total capital re-
quirement was used prior to the plant start-up.  Other  assumptions used in
this analysis include the following:
                                    219

-------
     •  Plant life was assumed to be 20 years with no cash value at
        the end of life.

     •  A straight line method was used to calculate annual depre-
        ciation.

     •  Operating costs and working capital requirements were
        assumed to be constant during the life of the process.

     •  Annual revenue and production were assumed to be constant
        during plant life (i.e., no start-up period).

     t  The present value of the investment includes the cost of
        capital during the construction period.  The construction
        period cost of capital, Ic, was approximated by the
        following expression:
                           Ic = 1.875  (O.Olr)  (C-W),
        where r is the discounted rate of return (percent), C is
        the total capital requirement  ($MM) and W is the annual
        working capital  ($MM).

     •  100 percent equity capital .

     Using the above assumptions to modify the basic equations, the required
annual  revenue requirements for the DCF method can be calculated from the
expression given for total capital requirements:
where ,
      C = Total capital requirement, $MM,
     Rj = Annual revenue requirement for DCF investor  financing,  $MM,
      D = Annual depreciation, $MM,
                                    220

-------
    FAn  = Annuity  factor for the life of the process at the DCF rate
         of  return,
      W  = Annual working capital, $MM,
    f   = Single payment present value factor at the DCF rate of
         return for  a  single quantity at the end of the plant life,
      N  = Net annual  operating costs of the plant operation (exclud-
         ing depreciation), and
      T  = Annual income tax rate, percent.

     As  a regulated  industry, the utility industry has different capitali-
zation,  tax and return-on-investment requirements which differ significantly
from a nonregulated  industry.  The basic equations derived in Appendix I
of the referenced  document were used along with all of the applicable in-
vestor financing assumptions listed above.  In addition, the following
assumptions were made for the utility financing calculations and were
obtained directly  from Appendix I of the FPC report:

     •   The debt/equity  ratio was  assumed  to be  75/25.

     •   The interest on  the  debt was  assumed to  be  9 percent.

     t   The required return  on the equity  was  assumed  to be  15 percent.

     •   The corresponding return on rate base  is  10.5  percent.


     Because  the rate base for utility financing varies during the life of
the plant,  this would result in a variable revenue flow (and subsequent coal
price) for  each year  of operation.  For this analysis, an average coal  price
during the  plant life was calculated to provide a basis of comparison.
Based on this assumption, the average annual revenue requirement could be
calculated  from the  following expression;
                                    221

-------
    Ru = N + 0.05 (C-W)  +  0.01  Fp +  TQQ  I  T  (l-d)ru IjC  -  0.5  (C-W)J>

where N, C, W and T are  the  same  as for the investor financing,  and
     R  = Average annual  revenue  requirement  for  utility financing.
      p = Return on rate  base,  percent per  year,
      d = Debt fraction  (0.75), and
     r  = Return on equity,  percent per year.

The capitalization requirements were  obtained from integrated Meyers Process
cost estimates which were based on June 1975  dollars.

     All process capital  cost data are based  on the detailed cost estimates
for battery limits Meyers Process units sized for 100  tons per hour  of coal
feed (presented in Sections  5.1 and  5.2).  The battery limits cost for each
specific economics case was  adjusted  to account for the varying coal input
compositions and characteristics.  For instance,  the battery limits  fine
coal Meyers Process unit cost estimated for Case  1, is significantly less
($5 MM  less) than that presented in  the base case cost estimate (Section
5.1.2.2).  This cost adjustment is warranted due  to increased reaction rates
(2 to 3 times base case), faster filtration rates observed for cleaned coal
and lower circulating gas rates in the  sulfur vaporization section.   These
increased rates lead to less capital  investment in reactor, filter  and sul-
fur removal associated operations.

       Off-site investment for  each of the  four cases  was  estimated  as a
percentage of the battery limit  cost.  The basic off-site investment was
then  adjusted for special characteristics  of the case. For fine coal pro-
cessing,  the basic off-site investment was estimated  at 50 percent  of the
battery limits cost given in Table 26 of Section 5.1  or $6.63 million.  It
was  expected that Case  1  would have  a reduction  in off-site capital relative
to  the  base case resulting  from  less by-products, waste disposal, power and
steam.   The reduction was estimated  to equal the $0.8-$1.0 million cost for
the  physical  cleaning.   Case 2 off-sites were estimated to be 100  percent
of  the  battery limits cost  given in  Section  5.2  or $4.2 million.   Case  3  has
                                     222

-------
a 240 TPH physical cleaning  section  and  a gravity separation  section  and
grinder with a total estimated  capital cost of $2.5 million.   This  added
to the basic process off-site gives  a total off-site cost of  $9.13  million.
Case 4 has the physical  cleaning train and gravity separation without
grinding.  This gives a  total off-site capital  estimate of $6.5 million.

     Operating costs were estimated  from  base case  operating  costs.  Ad-
justments were made to account  for such things  as varying oxygen and chemicals
(lime and scrap iron) requirements due to varying Meyers Process feed pyritic
sulfur contents.   The operating  labor requirement was adjusted to properly
reflect the varying complexities  represented by the  four cases.  Disposal
costs were estimated for each case to reflect the differing requirements
(primarily due to the use of physical cleaning  in some cases  and not in
others and Meyers Process operation with  coals  of varying pyritic sulfur con-
tent).  Disposal  costs were estimated to  be $5.90 per ton of  refuse.

     All  other capital  related and operating related costs were determined
as percentages of estimated battery limit  costs, off-site equipment costs,
raw materials costs, and labor costs.  Table 34 presents a summary of the
economic  evaluation criteria.

     The  economic evaluation criteria given in  Table 34 and the annualized
revenue expressions for investor  and utility financed process plants
described in this section can be  combined  to give an expression of the
following form:

          P = (aX + bY + cZ)/d

where,
     P =   is the  required sales  price for processed  coal, $/MM Btu
     X =   working capital for raw materials and supplies, $
     Y =   sum of  the total plant  investment and start-up cost, $
     Z =   annual  total  operating  cost, $/year
     a, b, c = constants given  in Table  35
     d =   annual  energy output,  MM Btu/yr
                                    223

-------
                    TABLE 34.  ECONOMIC EVALUATION CRITERIA


Operating Cost  Criteria

   Raw material  -  coal @  $10/ton, $20/ton, $30/ton

   Utilities
      Electricity  @  2.5£/kw-hr
      Oxygen  @  $25/ton
      Cooling water  @ 5<£/1000 gal.
      Process water  @ 25<£/1000 gal.

   Chemicals
      Lime @  $28/ton
      Scrap iron @ $50/ton

   Disposal @ $5.90/tpn of  refuse

   Labor
      Process operating labor @ positions/shift x 8304 man-hours/yr  x  $5/man-hour
      Maintenance  labor @ 1.5%/yr of total plant investment
      Supervision  @  15% of  operating labor and maintenance labor
      General overhead and  administration at 60% of labor and supervision


   Supplies
      Operating @  30% of  process  operating  labor
      Maintenance  @  1.5%  of total  plant investment

   Taxes and insurance  @  2.7%/yr of total plant investment

   Total operating cost = sum of raw material  + utilities +  chemicals  + disposal
     + labor +  supplies + taxes  and insurance

Capital Cost Criten'a

   Battery limits  capital as discussed in text

   Off-site capital  as  discussed in text

   Overhead and profit  @  22% of battery limits +  off-sites

   Engineering and design @ 10% of battery  limits  + off-sites

   Contingency  @ 15% of battery limits + off-sites  + overhead and profit +
     engineering and design

   Total plant investment = sum of battery  limits  + off-sites + overhead and
     profit + engineering and design + contingency

   Interest for construction @ 9% of total  plant  investment x 1.875

   Start-up cost @ 20%  of total  operating costs

   Working capital = sum  of raw materials  inventory of 60 days at full  rate +
     materials  and supplies at 0.9% total  plant investment + 1/24 annual
     product  revenue
                                      224

-------
           TABLE  35.   CONSTANTS FOR USE IN ECONOMICS EVALUATIONS

a -
b
c
Investor Financing
.391
.505
1.016
Utility Financing
.140
.121
1.006
5.3.3  Process Economics Evaluations

     The process economics evaluations  were  carried  out  as  described in the
previous section.  The results of  those evaluations  are  presented  in Tables
36, 37, 38, and 39.  Calculations  were  performed  at  three assumed  ROM coal
costs and using both utility and investor  financing  criteria.  Coal costs of
$10 per ton, $20 per ton and $30 per ton were  selected since they  represent
the broad range of currently reported ROM  coal  costs  ($10 per ton  at mine
mouth to $25 .per ton reported delivered price  at  some  plant sites).  As
may be determined from the data, the required  market value  of the  treated
coal ranges from a low of 66<£/MM Btu for Case  4 (assuming $10/ton  ROM coal
cost and utility financing) to a high of $2.39/MM Btu  for Case 1 (assuming
$30/ton ROM coal cost and investor financing).   In all cases, utility finan-
cing yielded market costs on the order  of  60 percent to  80  percent of market
costs required when utilizing investor  financing.

     An equivalent upgrading cost  has also been determined.  The upgrading
cost is found by deducting the cost of  the dirty  energy  ($.41/MM Btu for 20%
ash, 3-4 percent sulfur coal at $10/ton; $.81  at  $20/ton and $1.22 at $30/
ton) from the cost shown in Tables 36 through  39  for the clean energy ($.66
to $2.39/MM Btu).  These upgrading or processing  costs,  which range from
$.25/MM Btu to $1.17/MM Btu, are presented in  Table 40.  For all cases the
processing cost includes ash reduction  as  well  as sulfur reduction.  Except
for Case 2, physical cleaning was  assumed  to be coupled  with pyrite removal
which results in a major reduction in ash  from about 20  percent to about
6 percent.
                                    225

-------
                   TABLE  36.  CASE 1, CLEANED FINE COAL
Product Annual
Energy Value, 21.3 x 10 MM Btu/yr
Capital Related Requirements, $MM
Battery Limit Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total Plant Investment^ '
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Working Capital (Investor Financing)
Total Capital Related Costs (Utility)
Total Capital Related Costs (Investor)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Scrap)
Supplies
Disposal
Utilities
Labor (13 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
\
Utility Financing
Investor Financing

$10/Ton
8.26
6.63
3.28
1.49
2.95
22.61
3.82
2.89
2.54
2.98
31.86
32.30

9.60
.06
.50
.42
1.63
1.62
.61
14.44


.84
1.33
ROM Coal Cost
•$20/Ton
8.26
6.63
3.28
1.49
2.95
22.61
3.82
4.81
4.57
5.05
35.81
36.29

19.20
.06
.50
.42
1.63 ,
1.62
.61
24.04


1.31
1.86

$30/Ton
8.26
6.63
3.28
1.49'
2.95
22.61
3.82
6.73
6.59
7.12
39.75
40.28

28.80
.06
.50
.42
1.63
1.62
.61
33.64


1.79
2.39
(1)
   Equivalent to a plant capital  investment of $79.60/kw.
                                      226

-------
                    TABLE 37.  CASE 2, ROM COARSE  COAL

Product Annual         fi
Energy  Value,  17.5  x 10b MM Btu/yr                      RQM
Capital  Related Requirements, $MM             $ IP/Ton      $20/Ton      $30/Ton
Battery Limit  Capital                            4.20          4.20         4.20
Off-site Capital                                 4.20          4.20         4.20
Overhead and Profit                             1.85          1.85         1.85
Engineering and Design                           .84           .84           .84
Contingency                                     1.66          1.66         1.66
Total  Plant Investment^1)                     12.75         12.75        12.75
Interest for Construction                       2.15          2.15         2.15
Start-up Costs                                  2.51          4.11         5.71
Working Capital (Utility Financing)             2.06          3.74         5.42
Working Capital (Investor Financing)            2.33          4.05         5.77
Total  Capital  Related Costs  (Utility)         19.47         22.75        26.03
Total  Capital  Related Costs  (Investor)        19.74         23.06        26.38

Operating Costs, $MM/Yr
Raw Material (Coal)                             8.00         16.00        24.00
Chemicals (Lime, Scrap)                          .39           .39           .39
Supplies                                         .30           .30           -30
Disposal                                         -64           .64           -64
Utilities                                       1-86          1.86         1.86
Labor (9 Positions)                             1-02          1.02         1.02
Taxes  and Insurance                              -34           -34           .34
Total  Operating Costs                          12.55         20.55        28.55
Required Coal Market Price, $/MM  Btu
Utility Financing
Investor Financing
Utility  Financing                                -84          1>32         Il8°
                                                1.20          T-74         2.28
(1)
   Equivalent to a plant capital investment  of $54.60/kw.
                                      227

-------
                   TABLE 38.  CASE  3,  DEEP CLEANED FINE COAL
                              WITH  50% MEYERS  PROCESS BYPASS
 Product Annual
Energy Value, 43.4 x 10° MM Btu/yr
Capital Related Requirements, $MM
Battery Limit Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total Plant Investment^ '
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Working Capital (Investor Financing)
Total Capital Related Costs (Utility)
Total Capital Related Costs (Investor)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Scrap)
Supplies
Disposal
Utilities
Labor (14 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
Utility Financing
Investor Financing

IIP/Ton
13.26
9.13
4.93
2.24
4.43
33.99
5.74
5.28
4.83
5.51
49.84
50.52

19.20
.11
.69
.84
2.62
2.01
.92
26.39

.73
1.11
ROM Coal Cost
$20/Ton
13.26
9.13
4.93
2.24
4.43
33.99
5.74
9.12
8.88
9.66
57.73
58.51

38.40
.11
.69.
.84
2.62
2.01
.92
45.59

1.20
1.63

$30/ Ton
13.26
9.13
4.93
2.24
4.43
33.99
5.74
12.96
12.92
13.80
65.61
66.49

57.60
.11
.69
.84
2.62
2.01
.92
64.79

1.66
2.15
(1)
   Equivalent to a plant capital  investment of $58.70/kw.
                                   228

-------
                  TABLE 39.  CASE 4, DEEP CLEANED COARSE COAL
                             WITH 50% MEYERS PROCESS BYPASS
                     .(1)
Product Annual         <-
Energy Value,  42.2  x 10° MM Btu/yr
Capital  Related Requirements, $MM
Battery Limit  Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total  Plant Investment^
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Working Capital (Investor Financing)
Total  Capital  Related Costs (Utility)
Total  Capital  Related Costs (Investor)

Operating Costs,  $MM/Yr
Raw Material (Coal)
Chemicals (Lime,  Scrap)
Supplies
Disposal
Utilities
Labor (11 Positions)
Taxes  and Insurance
Total  Operating Costs

Required Coal  Market Price. $/MM  Btu
Utility Financing
Investor Financing

$10/Ton
4.20
6.50
2.35
1.07
2.11
16.23
2.74
4.92
4.51
4.89
28.40
28.78
19.20
.39
.38
.96
1.96
1.29
.44
24.62
.66
.88
ROM Coal Cost
$20/Ton
4.20
6.50
2.35
1.07
2.11
16.23
2.74
8.76
8.55
9.04
36.28
36.77
38.40
.39
.38
.96
1.96
1.29
.44
43.82
1.14
1.41

$30/Ton
4.20
6.50
2.35
1.07
2.11
16.23
2.74
12.60
12.59
13.18
44.16
44.75
57.60
.39
.38
.96
1.96
1.29
.44
63.02
1.62
1.95
0)
   Equivalent to a plant capital investment  of  $28.80/kw.
                                    229

-------
                    TABLE 40.  UPGRADING (PROCESSING) COSTS
Case 1
   Utility Financed
   Investor Financed
  $10/Ton

$.43/MM Btu
$.92/MM Btu
ROM Coal Cost
   $20/Ton

 $.50/MM Btu
$1 .05/MM Btu
   $30/Ton

 $.57/MM Btu
$1.17/MM Btu
Case 2
   Utility Financed
   Investor Financed
$.43/MM Btu
$.79/MM Btu
 $.51/MM Btu
 $.93/MM Btu
 $.58/MM Btu
S1.06/MM Btu
Case 3
   Utility Financed
   Investor Financed
$.32/MM Btu
$.70/MM Btu
 $.39/MM Btu
 $.82/MM Btu
 $.44/MM Btu
 $.93/MM Btu
Case 4
   Utility Financed
   Investor Financed
$.25/MM Btu
$.47/MM Btu
 $.33/MM Btu
 $.60/MM Btu
 S.40/MM Btu
 $.73/MM Btu
                                     230

-------
                      6.  CHEMICAL ANALYSIS STUDIES

     The chemical  analysis studies had two objectives:   (1) to experimen-
tally investigate the adequacy of established coal sulfur analysis techniques
for the accurate determination of Meyers Process  performance and  (2) to iden-
tify and screen potential Meyers Process efficiency monitoring techniques
for use in full-scale operations.  The need for these studies was dictated
by frequent inconsistencies in sulfur analysis data derived from processed
coal and by the desirability to identify process  monitoring techniques which
are simpler and faster than the standard ASTM coal sulfur analysis methods.

     In the Meyers Process the pyritic sulfur in  coal is oxidized to elemen-
tal sulfur and sulfate by the reduction of ferric ions to the ferrous state;
the latter are regenerated to the ferric state by oxygen.  The sulfate sulfur
product dissolves in the reagent solution.  The elemental sulfur can be
recovered by extraction with hydrocarbon solvents (e.g., toluene or naphtha)
or by vapo>ization; in bench-scale experimentation hydrocarbon leaching was
used almost exclusively.  Thus, Meyers Process chemistry permits, in principle,
the use of three independent approaches for the estimation of process per-
formance all of which involve standardized chemical analysis techniques.
The most direct approach is the determination of  pyrite in coal before and
after processing.   The second approach involves the determination of total
sulfur in coal before and after processing.  The  third approach is based on
the estimation of ferrous ion production during leaching.  Since 10.2 moles
of ferrous ion are produced per mole of pyrite leached, this approach of
measuring the performance of the process is very  accurate provided the
leaching reaction is totally selective, as it is  with most coals, or ferrous
ion production from side reactions can be independently estimated.  This
approach is also an excellent means of monitoring process performance during
processing because it is simple and quick.  Unfortunately, it is only useful
when the leaching and regenerations operations are performed separately
whereas simultaneous Teaching-regeneration is contemplated for most coals.

     In principle, the first two approaches are equivalent since the Meyers
Process should not affect the organic sulfur in coal and it should have
                                    231

-------
 insignificant effect on the sulfate sulfur concentration of unweathered
 coal which normally is 0.1 wt. percent or less.  In practice, direct
 equivalency of these two approaches has been very inconsistent and infre-
 quent.  Probable causes for the observed inconsistencies may be grouped
 into three categories:  (1) performer inconsistencies in the analyses of
 coal, a very real possibility with processed coal, the sulfur content of
 which is  low,  (2) incomplete  recovery of process sulfur products or high
 sulfate starting coal (weathered coal), and (3) inadequacy of standard coal
 sulfur analyses techniques to accurately determine the sulfur and sulfur
 forms content of processed coal (Meyers Process effects on sulfur analysis
 techniques).  Though much of the evidence from hundreds of coal analyses
 pointed to performer inconsistency as the most probable cause for incon-
 sistencies in analysis data, there was sufficient uncertainty in such
 conclusion to warrant investigation of probable Meyers Process effects on
 standard  sulfur analysis techniques.

     The  investigation of the coal sulfur analysis problems was combined
 with the  effort to identify efficient process monitoring techniques for use
 in  full scale Meyers Process plants.   The results and conclusions are pre-
 sented in separate sections below.

 6.1  ADEQUACY OF STANDARD SULFUR ANALYSIS TECHNIQUES FOR DETERMINING
     MEYERS PROCESS PERFORMANCE

     The  total sulfur in ROM, run-of-the-mine, coal is comprised of three
 sulfur forms:  pyritic,  sulfate, and organic sulfur.  Product coal from
 the Meyers Process may contain only one or as many as four sulfur forms.
 Complete  leaching of pyrite, complete recovery of process sulfur products
 (elemental sulfur and sulfate), and dissolution of the sulfate in the ROM
would result in processed coal the total sulfur of which is equal to the
organic sulfur present in the starting coal.  Less than 100 percent leaching
of pyrite and sulfate and incomplete recovery of the elemental sulfur product
would result in processed coal with four sulfur forms.  Pyritic  sulfur  and
sulfate sulfur can be determined directly by standard ASTM techniques.   Or-
ganic sulfur and unrecovered elemental sulfur are computed as  a  single  value
                                    232

-------
from the difference between the total  sulfur  content of the coal, deter-
mined by Eschka analysis, and the sum  of  the  pyritic and sulfate sulfur
values.

     The Meyers Process affects only the  pyritic  sulfur in coal directly
unless the coal is severely weathered  in  which case sulfate removal also
occurs during processing.  Indirectly,  however, the Meyers Process may
affect all three sulfur forms in coal  because of  incomplete sulfur product
recovery and because elemental sulfur  is  computed as organic sulfur.
Pyritic sulfur is always reduced; sulfate sulfur  remains unchanged or in-
creases slightly during processing, depending on  processing conditions,
when low sulfate ROM coal is used; organic sulfur remains unchanged unless
elemental sulfur recovery is incomplete in which  case the "organic" sulfur
content of the processed coal should always be higher with the increase
being equal to unrecovered elemental sulfur.  Thus, in the cases where
sulfur removals computed from total sulfur and pyritic sulfur analyses do
not agree, the delta in the two sulfur removal values should equal the
quantity of unrecovered sulfur product.   Complementing the total and pyritic
sulfur analyses by the determination of the sulfate sulfur in the sample
should lead to the quantitative identification of each of the sulfur forms
left on the coal.

     The data in Appendices A and B, Volume 2 show that the coal sulfur
analyses were infrequently consistent  with the above expectations.  The
discrepancies are manifested in the value of  organic sulfur, and they were
obviously caused by inaccuracies in analyses  either of total sulfur or of one
or both of the sulfur forms (pyrite and sulfate).  As indicated earlier,
the extensive experimentation on the Meyers Process with Lower Kittanning
coals has not shown any evidence that  the process affects the organic sulfur
content of the coal.  Thus, indicated  reductions in the organic sulfur
of coals during processing must be due  to errors  in analyses.  The
same conclusion is drawn from coal samples exhibiting relatively large
increases in organic sulfur which could not be justified in terms of
unrecovered elemental  sulfur.
                                    233

-------
     The discrepancies are in the 0.1  to 0.5 wt.  percent sulfur range and
therefore, of relatively minor importance to the  measurement of Meyers
Process efficiency when applied to high sulfur,  high pyrite coals as the
one used in this program (though they are very important in determining
whether the processed coal met a specific sulfur  standard or not).  These
discrepancies become very important in the determination of Meyers Process
efficiency and rates as the pyrite content of the feed coal to the process
decreases (e.g., pyritic sulfur in the 1 to 2 wt. percent range).

     Problems with coal sampling and analysis have been encountered by
virtually everyone requiring accurate coal composition data on a regular
basis.  The consensus of opinion (from mine quality control personnel to
Bureau of Mines chemists) appears to be that the  standard ASTM procedures
for coal analysis are adequate when properly utilized, but that the con-
sistency of accurate data suffers when these techniques are applied rou-
tinely in an assembly line operation (necessary for holding analysis costs
to reasonable levels).  The purpose of this task  was to investigate whether
or not the Meyers Process aggravated these problems by direct effects on
the accuracy of standard sulfur analysis procedures.  In the case of total
sulfur analyses the concern centered on potential effects from organic
solvents used to recover elemental sulfur and from unrecovered elemental
sulfur itself.  In the case of sulfur forms the concern was whether the
reflux times specified by the ASTM procedures for sulfate and pyrite re-
covery were adequate for use on  chemically  leached  coal  whose  sulfate and
pyritic sulfur values are approximately equal and very low.

     Three different techniques were used in the total sulfur coal analysis
investigations:  Eschka, Bomb Wash, and Leco (Eschka being the standard ASTM
technique).  The rationale was that since these three sulfur analysis  techni-
ques differ substantially in procedure it would be very  unlikely that  the
Meyers Process could have the same effect on the accuracy  of each of them,
if indeed processed coal by the Meyers Process affected  established  sulfur
analysis techniques.  Eschka and Bomb Wash are commonly  used sulfur  analysis
techniques for coal; Leco was selected as the third method because it is an
adequately developed sulfur analysis technqiue and  because it  offers the
potential for use as a Meyers Process monitoring technique.
                                    234

-------
     The matrix of experiments  began  with comparison of the  three total
sulfur analysis methods on NBS  certified  coal  samples ranging  in sulfur
values from 0.546 to 3.020 wt.  percent.   The  three  analyses  methods were
then used on:

     t  Run-of-the-mine, ROM, Lower Kittanning coal  samples  from
        the coal utilized in this  program.

     •  L-R processed coals at  two different  reaction times.

     •  Organic solvent treated L.K.  coals  (the coal  was refluxed
        for one hour in either  hexane, or heptane,  or toluene
        without prior leaching  with ferric  sulfate).

     •  ROM and organic solvent treated coals  doped  with 2 to 3
        wt. percent elemental sulfur.

     The data generated from these experiments are  summarized in Table 41.

     The first group of data in Table 41  compares NBS certified total sulfur
values to total sulfur analyses performed on the same coal  using Eschka,
Bomb Wash, and Leco techniques.  The  ASTM acceptable reproducibility
spreads for total sulfur analyses  performed on the  same coal at two dif-
ferent laboratories are 0.1 wt. percent sulfur for  coal containing 2 wt.
percent sulfur or less and 0.2  wt. percent  sulfur for coal with higher than
two percent sulfur content.  All  three techniques yielded values within
the ASTM allowable spreads.  Thus, any one  of these  techniques may be
utilized for the determination  of  sulfur  concentration in ROM coals up to
at least 3 wt. percent.  It should be noted,  however,  that the Eschka
technique yielded values consistently closer  to the  certified sulfur values.
At coal sulfur concentrations in the  2 to 3 wt.  percent range the Bomb Wash
and Leco technique values indicate a  low  bias  with  Leco furnishing the
lowest sulfur concentrations.   On  the basis of these observations, the
Eschka technique must be assumed to be the  most accurate of  the three for
analyzing  high sulfur coal.  However, if the  NBS samples were certified
by the Eschka technique the above  conclusion may be  unwarranted.
                                    235

-------
       TABLE  41.    COMPARISON  OF TOTAL SULFUR ANALYSIS TECHNIQUES
                     (WEIGHT PERCENT SULFUR  IN COAL)
 I.    NBS CERTIFIED COALS:
Certified Values Eschka Bomb Wash Leco
0.546 + 0.003 0.54 +_ 0.02 0.55 +_ 0.
2.016 + 0.014 2.03+^0.04 1.92+_0.
3.020 + 0.008 2.94+^0.03 2.93 +_ 0.
II. RON (RUN-OF-KINE) LOWER KITTANNING COAL:
4.43 + .06 4.24
III. L-R PROCESSED LOWER KITTANNING COALS:
Experiment No.
S2 (Toluene*) 1.44 + 0.005 1.52+_0.
S3 (Toluene*) 1.49 1.53 + 0.
S4 (Toluene*) 2.37 + 0.01 2.39 + 0.
S5 (Hexane-Toluene**) 2.22 +_ 0.04 2.34 + 0.
IV. ORGANIC SOLVENT TREATED ROM LOWER KITTANNING COAL:
Solvent Used
00 0.59 +_0.04
01 1.90+0.03
03 2.86 +_ 0.01
4.08 +0.06
05 1.54^0.02
01 1.54 + 0.01
03 2.36 +_ 0.005
07 2. 33 +_ 0.06

Hexane 4.37 4.28 3.84+0.11
Heptane 4.48 +_ 0.05 4.30 4.19 + 0.08
Toluene 4.32 + 0.01 4.23 3.95 +_ 0.03
V. ROM LOWER KITTANNIHG COAL DOPED WITH ELEMENTAL SULFUR:
Sample Doped Bomb Wash Leco
Kith Elemental Calculated Analyzed Calculates Analyzed
Sulfur Sulfur Sulfur Sulfur Sulfur
ROM Coal 6.96 6.83 6.80 6.53+0.10
Hexane Refluxed 6.97 6.82 6.73 6.44+0.001
Heptane Refluxed 7.20 7.20 7.09 6.70+0.01
Toluene Refluxed 7.16 7.11 6.83 6.74+0.00
Eschka
Cal cul ate3 Analyzed
Sulfur Sulfur
7.15 6.96+0.03
7.16 6.77
7.39 7.10+0.06
7.35 7.00
% Recovery
Bomb Leco Eschka
Wash
98 96 97
98 96 95
100 94 96
99 99 95
 Elemental  sulfur on  processed coal  was extracted by toluene.
Elemental sulfur on processed coal was extracted first by hexane and then by toluene.
                                       236

-------
     The discrepancy  among sulfur values generated by the three techniques
 investigated  in this  program becomes meaningful when the total sulfur con-
 centration in the  coal  exceeds 4 wt. percent, as demonstrated by the second
 group of data in Table  41.  These data were generated on identical samples
 of ROM  Lower  Kittanning coal.   The spread between Eschka and Leco generated
 sulfur  values is unacceptable (too large to be considered a normal deviation
 between values of  valid sulfur analysis techniques).  A careful examination
 of the  procedures  prescribed for the Bomb Wash and Leco techniques did not
 result  in the identification of inherent technique shortcomings which could
 predict the apparent  bias toward lower sulfur values.  However, small
 systematic errors  in  the application of these procedures and defective equip-
 ment or operator biases  could not be completely discounted as probable
 causes  for the observed low sulfur value bias.  On the basis of the available
 data it must  be concluded that the Leco technique can not be  used with
 confidence to determine the sulfur content of coals when it exceeds 3 wt.
 percent.  At  lower coal sulfur concentrations the Leco technique exhibited
 adequate accuracy  and excellent reproducibility to render it useful in
 monitoring the efficiency of the Meyers Process.

     The data in Sections III  through V in Table 41  represent  the  results
 of studies performed  to determine if the Meyers Process  affects the accuracy
 of total sulfur determinations on coal.   Several  samples of L-R processed
 Lower Kittanning coal from four experiments were analyzed for  total sulfur
 content by each of the  three sulfur analysis techniques.   The  data are
 summarized in Section III of Table 41.   The data show that if the Meyers
 Process affects the accuracy of total sulfur determinations, its effect is
 virtually identical on  all  three techniques.   This  was considered  unlikely
 and it was, therefore,  concluded that the sulfur content of processed coal
 can be accurately  determined by any of the three techniques, if properly
 used.   The validity of  this conclusion was verified further by the data in
 Sections IV and V.

     Organic  solvents and elemental  sulfur are the  only  substances  poten-
tially  present in coal  subjected to the  Meyers Process and  not  present  in
the raw coal  (the former  would be present only if the  product  elemental
sulfur  is  recovered through organic solvent leaching of  the processed coal).
                                     237

-------
Thus, it was decided to investigate whether organic solvents of potential
use to the Meyers Process and whether elemental  sulfur not completely re-
covered from processed coal  could measurably affect the accuracy of coal
sulfur analysis techniques.   Samples of ROM Lower Kittanning coal (Section
II, Table 41) with known sulfur content were refluxed for one hour in each
of the three organic solvents listed in Table 41  and subsequently analyzed
by each of the three total sulfur analysis techniques.  Comparison of the
data in Sections II and IV reveals that these solvents did not affect the
accuracy of sulfur analyses.  In separate experiments, ROM coal samples
and organic solvent refluxed samples were doped  with known quantities of
elemental sulfur and then reanalyzed.  The data  in Section V demonstrate
that the elemental sulfur added to the coal was  accurately reflected by the
total sulfur analyses performed on the coal samples regardless of the
analysis technique used.

     In summary, the total  sulfur analysis investigations revealed that
the Meyers Process does not affect the accuracy of coal total sulfur de-
terminations by Eschka, Bomb Wash, or Leco techniques.  Within acceptable
deviations, these three analysis techniques proved equally accurate in the
determination of coal sulfur concentrations up to approximately  3 wt. per-
cent.  At higher coal sulfur concentrations the Leco technique demonstrated
an unacceptable low bias, yielding sulfur values which were lower by as
much "as 10 percent than those obtained by Eschka; however, it was not
established conclusively whether the low values were inherent to the tech-
nique or due to systematic errors in its utilization.

     The sulfur forms investigations led also to the conclusion  that the
Meyers Process does not affect the accuracy of standard coal analysis
techniques.   In these investigations potential interferences of  elemental
sulfur on sulfate analyses and sulfate on pyrite analyses were examined.

     In the  case of elemental sulfur the concern was that,  if  incompletely
recovered,  it could be physically leached during the hydrochloric  acid
extraction  and subsequently be oxidized to sulfate in  the process  of sulfate
analysis.   Experiments performed on elemental sulfur derived  from  the

                                     238

-------
Meyers Process as well  as  on coal  doped with elemental sulfur showed no
evidence that this  sulfur  form oxidizes to sulfate during the standard ASTM
procedure for coal  sulfate analysis.   In addition, sulfur balance data de-
rived from bench-scale  experimentation on the Meyers Process indicated that
elemental sulfur recovery  was virtually complete in the majority of experi-
ments and, therefore, could not account for the apparent inconsistencies in
sulfur forms analyses  (Section 2,  Volume 1 and Appendices A, B,  and C,
Volume 2 of this report).

     A more serious concern was that  associated with potential Meyers Process
effects on pyrite analysis data, either because of sulfate interference or
due to incomplete leaching of pyrite.   According to ASTM procedure,  pyrite
in coal is determined from the difference in the iron content of the nitric
acid and hydrochloric acid solutions  derived by refluxing for 30 minutes
identical coal samples,  except for size, in the respective acids.   The acid
soluble iron compounds  of  coal, principally iron sulfates and oxides, are
leached by the hydrochloric acid;  the  nitric acid being an oxidizing agent
leaches all of the  iron  in coal which  includes the pyritic iron.   Normally,
the soluble iron in ROM coals represents a small  percentage of the pyritic
iron; thus, incomplete  leaching of the sulfate by hydrochloric acid  intro-
duces an insignificant  error to the determination of the pyrite  content of
coal.  In processed coal the pyrite content is low and substantial  errors
in its determination could result  from improper leaching of the  iron forms
in coal, especially in cases where sulfate recovery is incomplete  as it was
the case in a number of  experiments performed during this program  (Appendices
A and B, Volume 2).  The concern was whether the  acid reflux times  prescribed
by ASTM procedures  were  adequate to remove the iron from coal to  the in-
tended extent.  The principal  question was whether 30 minutes reflux time
in hydrochloric acid was sufficient for the leaching of process  derived
sulfate product (higher  concentration  and conceivably different  form of iron
sulfate than typically found in ROM coals).

     A summary of the data generated on the effect of acid digestion time on
pyritic and sulfate sulfur analyses is presented  in Table 42.  ASTM  and
modified ASTM analysis procedures  were used in these investigations.  The

                                     239

-------
                      TABLE 42.
COAL SULFUR FORMS ANALYSIS  INVESTIGATIONS - COMPARISON OF
TECHNIQUES AND EFFECT OF  COAL DIGESTION TIME^a)
      Sample
 ROM (Weathered)
 L.  K.  Coal
 Processed  Coal
 (Exp.  S2)
Processed  Coal
(Exp. S4)
Processed Coal
(Exp. S2)
Doped with Pyrite
Digestion
(Reflux)
Time
0.5 Hours
1 Hour
2 Hours
3 Hours
0.5 Hours
1 Hour
2 Hours
3 Hours
0.5 Hours
1 Hour
2 Hours
3 Hours
1 Hour
2 Hours
3 Hours
Pyritic Sulfur, Wt. %

ASTM
2.66 + .12
2.66 + .04
2.63 + .07
2.56 + .04
0.39 + .02
0.44 + .01
0.41 + .02
0.44 + .02
1.00 + .06
0.94 + .04
1.01 + .02
1.07 + .02
(0.67}k 0.59
(0.66), 0.57
(0.66), 0.57

AA
2.76 +. .13
_
_
2.61 +. .02
0.38 + .01
0.40 + .04
0.39 + .06
0.34 + .05
1.00 + .00
0.91 + .04
0.91 + .04
1.07 + .02
(0.64)1? 0.63
(0.63), 0.50
(0.64), 0.59
Sulfate Sulfur, Wt. %

ASTM
0.61 + .03
0.62 + .01
0.64 + .01
0.67 +. .01
0.29 + .02
0.31 + .01
0.35 + .01
0.39 +_ .02
0.60
0.63 + .01
0.64 +\00
0.65 + .01
(0.60)^ 0.61
(0.60), 0.62
(0.60), 0.62
Turbidimetric
Peroxide - Bromine

0.14 + .03 0.15 + .02
0.18 + .05 0.14 + .05
0.24 + .08 0.24 +_ .01
0.16 + .01
<0.01 <0.01
0.16 + .01 0.19 + .02
0.04 +_ .01 0.04
0.27 + .00
0.24 + .02 0.27 + .04
0.30 + .03 0.28 + .01
0.19 + .02 0.31 + .04
0.41, 0.61
-------
modified techniques  are  simpler than the corresponding ASTM techniques  and
in principle more  accurate since they eliminate the subjectivity inherent
in the determination of  titration end points.  However, the data in  Table
42 do not appear to  justify the latter expectation.  Both the ASTM and  the
modified techniques  require the acid digestion step.  The modified tech-
niques used atomic absorption, AA instead of iron titration in pyrite
analyses and turbidimetric instead of gravimetric analyses in sulfate
determinations; two  different oxidants were used in the turbidimetric
sulfate analyses.  Comparison of ASTM and AA pyritic sulfur data reveals
that the two techniques  yielded equivalent values with slightly better
reproducibility evident  in data generated by the ASTM technique.   In
multiple sample analyses,  the AA technique may merit consideration.  The
turbidimetric technique  yielded consistently low values until  a cation  ex-
change column was  used to  remove interferences (last column of last  three rows
of data in Table 42).  The turbidimetric technique merits serious  consider-
ation since it simplifies  appreciably the determination of sulfate;  however,
cation interferences must  be eliminated which implies that the technique
may require frequent calibrations against gravimetrically derived  data.

     The first group of  data in Table 42 were obtained from analyses per-
formed in triplicate on  weathered ROM Lower Kittanning coal.   Substantially
weathered coal was selected in order that its sulfate content  be sufficiently
high to detect changes in  its determination as a function of digestion time.
The data indicate  a  trent  toward lower pyritic sulfur values and correspond-
ingly higher sulfate values with increasing digestion times.   The  trend
appears to be real,  but  the difference in the determined  sulfur form values
in coal samples digested for 0.5 and 3 hours, respectively,  is within
expected analysis  reproducibility limits.

     The second and  third  groups of data in Table 42 represent analyses
performed also in  triplicate on coals processed by the Meyers  Process.  Two
coals processed under  different conditions were selected;  thus,  the extent
of pyrite leaching and product sulfate recovery differed  in  the two experi-
ments.   Again, digestion time did not significantly affect sulfur  form
analysis values.    In addition, these data suggest that the Meyers  Process

                                     241

-------
product sulfate is as soluble  in  acids  as  natural  coal  sulfate,  at  least
under the conditions of standard  ASTM procedure  for sulfur forms analyses
(the feed coal  to Experiments  S2  and S4 contained  <0.2  wt.  percent  sulfate;
thus, at least  part of the sulfate  in the  processed coals  was  process-
derived).

     The last group of data in Table 42 were  obtained from analyses on
processed coal  doped with additional pyrite and  iron sulfate.  They were
principally performed to compare  the relative accuracy  of  ASTM and  modified
ASTM techniques.   However, they also show  that digestion time  does  not
affect the accuracy of sulfur  form  analyses  (the analyzed  pyritic sulfur
values were consistently lower than the computed values indicating  that
the pyrite doping technique used  led to a  systematic error—probably due
to incomplete incorporation of the  pyrite  into the coal  samples).

     In summary,  the sulfur analysis investigations performed  in this pro-
gram indicated  that the Meyers Process  does not  affect  the accuracy of
standard coal sulfur analysis  techniques even when recovery of the  process
sulfur products is incomplete. The conclusion is  that  carefully performed
sulfur analyses should yield the  true sulfur  forms composition in coal
whether processed through the  Meyers Process  or  not.  Indirectly, however,
by reducing the pyrite concentration of coal  to  low levels the Meyers
Process aggravates the frequently encountered inconsistencies  in coal
sulfur analyses.

6.2  MEYERS PROCESS MONITORING TECHNIQUES

     Functionally, the Meyers  Process  consists of three major operations:
coal pyrite oxidation, reagent regeneration,  and product recovery.    For
every mole of pyrite oxidized  in  the  leaching operation, one mole  of coal
iron is dissolved into the reagent  as  Fe+2 and 9.2 moles of reagent  Fe+3
                 +2
are reduced to  Fe  .  Thus, the reagent total iron  versus reaction time
trace or the ferrous ion versus reaction time trace can be directly trans-
lated into coal pyrite concentration  versus reaction time information.
         +2
Either Fe   or  total iron can  be  used  to monitor process performance in
the leacher.  Preferably the ratio  of these iron forms should be used
                                   242

-------
because it is not subject  to  errors resulting from sample water loss  due
to vaporization and because the  Fe+2/Fe ratio when accompanied by tempera-
ture data furnish complete information on all parameters affecting the
leaching rate of a given particle size coal.  The leaching rate depends on
the type of coal and  particle size, on temperature, on the pyrite concen-
tration in the coal,  and on Y;  the Fe+2/Fe ratio is equal to 1-Y.   Thus,
the monitoring of temperature,  Fe  , and Fe, and knowledge of the pyrite
concentration in the  feed  coal  and its particle size furnish the required
information to properly monitor  the pyrite leaching operation.   Temperature
and Y are also the appropriate  parameters for monitoring the reagent
regeneration operation when used in conjunction with oxygen partial pres-
sure monitoring.  The sulfur  product recovery operations can be easily
monitored from the weight  of  product generated as a function of time  with
occasional sulfate and elemental sulfur determinations performed to assure
the stability of product composition.

     Monitoring the Meyers Process by the described techniques  presents no
problems.  The techniques  are simple and fast; they require little labor
skill and they use inexpensive,  commercially available equipment.  Weights,
temperature, and oxygen partial  pressure can be monitored continuously.
The total iron and ferrous ion concentrations in the reagent would have to
be performed by grab  sampling, but the analysis time is short and  the
reliability high.  The ferrous  ion analysis performed by the standard ASTM
technique requires approximately 10 minutes; the corresponding  Fe  analysis
can be performed in approximately 20 minutes.  Several  analyses  can be
performed simultaneously by a single individual.   From experience  gained
from sampling and analysis of several  hundred coal  slurry samples, it was
estimated that the maximum time  required from slurry sampling to Y com-
putation is one hour  when  standard ASTM techniques  are used for  iron  forms
analyses.  This response time, though  adequate, can be reduced to  less
than 30 minutes if slurry  filtration and weighing of the reagent sample
are replaced by pipetting  the required sample size directly from the
slurry and if AA is substituted  for Fe titration (see Section 6.1)
Guarding against increases in the top  size of the coal  requires  only  a
single screen which is a part of the coal  feed train of the plant.

                                    243

-------
Monitoring the pyrite concentration in the feed coal  and the composition
of the product coal  could present problems in special situations where
quick analysis answers are required.   Under normal  processing conditions
sampling of feed and product coals need not be frequent (once or twice per
shift unless there is a more frequent change in coal  lots) and immediate
analysis data would not be required.

     All the elements of the Meyers Process monitoring scheme described
above, except product weight sulfate  monitoring, have been extensively
tested and used at bench-scale.   The  scheme proved  both efficient and re-
liable.  However, its use is predicated on the assumption that the process
is totally selective with respect to  ferric ion reaction with pyrite and
that the coal leaching and reagent regeneration operations are performed
separately.  Side reactions between reagent iron and  coal mineral matter
other than pyrite or the organic matrix of the coal will introduce errors
in the computation of pyrite concentration in coal  from ferrous iron con-
centration or Y measurements.  Experience gained from processing at least
50 different coals by the Meyers Process revealed that for the majority
of coals pyrite leaching by ferric sulfate was totally selective within
the accuracy of the measurements involved.  In cases  where excess ferric
ion utilization was observed, over that attributable  to pyrite oxidation,
the excess ferric ions consumed or ferrous ions produced per unit time
was the same in different samples of  the same coal.  Thus, a single deter-
mination of excess ferric ion utilization per coal  type is sufficient to
render the described process monitoring scheme usable with that coal when
pyrite leaching is not totally selective.

     The described process monitoring scheme becomes inadequate when the
chemical desulfurization of coal is performed by L-R processing  (simultaneous
coal  Teaching-reagent regeneration).   The L-R processing scheme  (Section  5)
is expected to be used with the high pyrite coals which can not  be  effi-
ciently desulfurized in coarse sizes.  In the L-R  scheme the major  fraction
of the pyrite in coal is removed in the L-R reactor where pyrite  leaching
and ferric ion regeneration take place simultaneously.  Thus,  the con-
centration of the individual iron forms in the slurry are affected  simulta-
neously by both reactions, leaching and regeneration, and the  value of Y can

                                   244

-------
no longer serve as a measure  of coal  pyrite depletion during processing in
the L-R reactor (the value  of Y can still  be used to monitor pyrite de-
pletion in the mixer and  ambient pressure  reactor of the L-R processing
scheme).  In order to monitor process performance in the L-R reactor,
direct coal sulfur analyses are required either for pyrite or for total
sulfur.  At least the pyrite  content  of the coal  in the L-R reactor
effluent stream must be monitored (the pyrite content of the feed coal to
the L-R reactor is attainable from the pyrite concentration in the process
feed coal and the delta in  the Y values of the mixer).

     Examination of time  requirements for  the generation of coal sulfur
data from coal slurries  (analysis response time)  using conventional sample
preparation techniques and  standard ASTM analysis methods revealed that
such procedure would be totally inadequate for monitoring the L-R reactor.
The response time would be  several hours long with sample preparation
being the major time consumer.   Sample preparation entails slurry filtra-
tion, coal washing, coal  drying, and  coal  sampling and weighing; coal
grinding may also be necessary.   Coal drying requires a minimum of four
hours, if performed according to acceptable ASTM  procedures; the remaining
operations require a minimum  of one hour.   The minimum time required for
the performance of the standard ASTM pyrite analysis is 2 hours; a total
sulfur analysis by Eschka could be performable in 1  hour.   Thus, the
minimum response time for total  sulfur data is 6  hours  and for pyritic
sulfur data 7 hours.  Such  response time was considered too long to serve
as the means of obtaining process performance information which would
permit timely changes in  process parameters to correct  processing defi-
ciencies and prevent the  generation of a final  coal  product with a sulfur
content higher than desired.   This conclusion is  valid  even though slurry
residence time in the ambient pressure reactor, in series with the L-R
reactor, could be as long as  40 hours, because the ambient pressure reactor
will  not normally have the  flexibility of  affecting large changes in pyrite
leaching rates.

     After careful examination  of the sample preparation procedure, it was
decided that it could be  shortened considerably,  with less than 10 percent
                                    245

-------
potential  penalty in  the  accuracy  of the  analysis,  if coal  washing and
coal drying were eliminated.   In the case of pyrite analysis, the elimi-
nation of these two operations affects only the accuracy of the weight of
the sample subjected  to analysis.   Data from over one hundred coal slurry
filtrations revealed  that the  quantity of reagent remaining on the coal
did not vary by more  than + 5  percent of  the weight of dry coal, even if
no effort was made to exactly  reproduce filtration  times;  in the majority
of experiments the deviation was less than +_3 percent.   Thus, sample
preparation time was  reduced to approximately 0.5 hours  and the response
time for pyrite analysis  using the standard ASTM technique (ASTM D-2492)
to approximately 2.5  hours.  This  response time was reduced to approxi-
mately 2 hours when the iron determinations in the  hydrochloric and nitric
acid extracts were performed by AA instead of titrimetric procedures
(see Section 6.1).  A two hour response time was  considered adequate for
monitoring the L-R reactor.

     The analysis response time could be  reduced to approximately 1 hour,
if total sulfur rather than pyritic sulfur is used  as the means of moni-
toring pyrite depletion in the L-R reactor.  However, the accuracy of the
data may be appreciably lower than that obtainable  with the shortened
pyrite analysis procedure described above.  The total sulfur analysis
procedure utilizes the Leco sulfur analysis technique, but on wet coal.
The data presented in Section  6.1  showed  that the Leco technique furnishes
high accuracy sulfur  values when the sulfur concentration in coal  (dry) is
less than 3 wt. percent and that  its accuracy is not affected  if a portion
of the sulfur in coal is  in the  elemental form.  However, similar  sulfur
accuracy determinations were not  performed with wet coal, though no
problems are anticipated.  The potential  source of error is associated
with the fact that the determined  sulfur  value must be corrected  for  the
product elemental sulfur, which  was not recovered prior to analysis,  and
for the reagent sulfate on the wet coal  (the coal sample may  be washed
prior to analysis, but then the  response time increases to approximately
1.5 hours).  The sulfate  correction can be made from the estimated weight
of reagent solution on the coal  and the Fe concentration and  Y values
determined routinely as elements of the process monitoring scheme described
earlier in this report section.   Assuming that 40 percent  of the  reacted
                                    246

-------
 pyritic  sulfur converts  to elemental sulfur, a valid assumption based on
 data generated to date from several coals, the determined depletion in
 total  sulfur by the  described procedure represents 60 percent of the
 leached  pyrite during the  reaction time interval examined.  The feasibility
 of this  procedure was not  tested experimentally in an integrated form.

      In  parallel with the  above investigations, a literature search was
 undertaken aiming at the identification of novel, rapid procedures for
 pyrite determination in  coal.  The chemical abstracts were searched for
 the last decade with limited success.  Three insufficiently developed new
 pyrite analysis techniques were discovered, but with no demonstrated
 potential to serve as monitoring techniques for the Meyers Process.

      One of these techniques was the result of a feasibility study per-
 formed by Barringer  Research in Ontario, Canada under government contract
 in 1968;12 it  involves  the sensing of sulfides in coal by a microwave
 detection system.  Though  an interesting approach because it permits the
 determination  of pyrite  in the coal without leaching, the data on the
 technique is insufficient  to evaluate its accuracy.

     A second  technique  developed and used by the Illinois State Geological
 Survey Institute is  a lithium aluminum hydride reduction procedure  for
 Pyrite in coal.13  This  technique is affected by the presence  of elemental
 sulfur in the coal (positive interference) and, therefore, it  is not better
 than a total sulfur  analysis technique (Leco or Eschka)  for monitoring the
 Meyers Process performance.   In fact, it is more complicated and less
 tested.

     The third technique is  a soft X-ray procedure  developed at  Pennsylvania
 State University.14  It  is reportedly capable of determining the concen-
 tration of pyritic,  sulfate,  and  organic sulfur in  coal  using measurements
of X-ray fluorescence from four discrete wavelengths.  This  promising
Procedure is  a very  recent development-first  reported  in December 1974-
and therefore inadequately tested.   Attempts  in our laboratory to duplicate
                                    247

-------
the reported data using a pentaerythritol analyzing crystal and an XRD-5
spectrophotometer were not successful because this system was incapable
of resolving the sulfur Kg doublet on which the procedure is based.  Per-
sonal contact with Dr. Eugene White at Penn State University (co-author
of the paper) revealed that the instrument they used had to be specially
modified to obtain the desired resolution and the surface temperature of
the crystal (Germanimum) had to be regulated within +_ 0.1°C to maintain
dispersion stability.   At the present time, the procedure though potentially
viable, would require  extensive instrument and procedure development be-
fore being amenable to use as a process  monitoring technique.   However,
X-ray fluoresence is an attractive analytical  technique because it is
potentially capable of being used to analyze more than one element and
because it is amenable to automation.  Its drawback, as a monitoring
technique involving coal slurries, is that it requires that the coal  be
dried prior to analysis and coal  drying  is a time-consuming operation.
                                   248

-------
                       7.   MATERIALS  COMPATIBILITY
                                      i

     Static testing of common  stainless  steel  alloys  in an environment
simulating that of the Meyers  Process  but with low solution oxygen concen-
tration  revealed that many stainless  materials  had severe corrosion
problems.  However, dynamic testing in the L-R apparatus under actual
processing conditions indicated  a  much wider range of stainless material
stability to  corrosion.

     The following composition and operating conditions are typical for
the Meyers desulfurization  process:
                                   V
     Fe2(S04)3 approximately 18%

     H2S04 approximately  10-20%

     FeSO. approximately  1%

     Finely dispersed 02

     Possibly chloride contamination
     Temperature 275°F, Pressure 50-100  psi
     Agitation and vigorous 02 bubbling
     Coal abrasive charge

     Obviously, materials to be  used  for the construction of the desulfur-
ization process plant must  possess the necessary mechanical strength,
abrasion resistance and corrosion  resistance properties.  Of course,
design life,  material cost, safety and maintenance are factors for trade-
off consideration with material  performance.

     Specific data on corrosion  rates  under the  conditions of the leaching
process considered here are unavailable  from literature.  It is recognized
that in view  of the complexities of the  electrochemical processes associated
with corrosion, published data on  corrosion  tests  under similar environments

                                   249

-------
may not be directly applicable.   No  assumptions  or extrapolations of avail-
able data may be possible.   Corrosion  rates  have to be determined experi-
mentally under the specific conditions  of the  process  in  consideration.
However, there is adequate  understanding and published data on corrosion
to limit the choice to certain  classes  of materials.

     An initial literature  survey and  alloy  selection  discussion, the
results of static testing for screening of applicable  materials and exper-
imentation on materials behavior in  the bench-scale reactor under dynamic
conditions are presented in the three  sections to follow.

7.1  LITERATURE SURVEY AND ALLOY SELECTION CRITERIA

Discussion of Alloy Composition and  Corrosion  Mechanisms

     In terms of the more common types  of corrosion, the  desirable metal
composition and metallurgical structure are  discussed  below.

Corrosion

     In the baseline Meyers Process  design the physical size of the welded
assemblies may be such that it  may not  be convenient or practicable to
solution anneal or carry out other post-welding  metallurgical  heat treat-
ment of the finished structure.   Hence, the  material chosen must resist
sensitization (preferred precipitation  of chromium carbide in the grain
boundaries).   Alloys containing low  C and/or some Nb,  Ta  are required to
avoid a sensitized condition which is highly susceptible  to intergranular
corrosion.

     Chloride ions present in small  concentrations in water,  and probably
also in the coal charge, may penetrate passivating films  and  initiate
pitting corrosion.  The critical temperature at which pitting attack is
appreciable increases with  increased Mo content.  At  an operational
temperature of 275°F, a high Mo content is desirable.  Mo is  the commonly
used alloying element to impart resistance to pitting  corrosion.

                                  250

-------
     Near boiling temperatures,  H2$04 between  20-80%  acid  concentration
causes serious attack on most  stainless  steels.   Addition  of ferric sulfate,
CuS04 or 02 reduces this attack  greatly.  Alloys  containing Mo, Co/Si + high
level of Ni show good corrosion  resistance  in  these environments.

     Chloride ions, although present in  small  concentrations, would cause
severe stress-corrosion cracking in  the  presence of welding and mechanical
stresses and the tensile stresses  in a pressure  vessel.  Abrasive grains
may also cause notches that initiate Griffith  crack propagation.  A high
level of Ni provides good  resistance to  chloride ion  stress corrosion
cracking.

     High Ni + Cr and a high Mo  content  are useful in the  presence of hot
oxidizing ferric salt solutions  of fluctuating concentrations.

     Differential aeration or  oxygen concentration initiates corrosion in
crevices.  Hence, good agitation and dispersion  of 02 bubbles in practice
would not only improve the sulfur  leaching  process but also reduce corrosion.

Stainless Steels, Coatings and Inhibitors

     The advantages of Mn-Ni substitution in austenitic stainless steels
are increased resistance to stress corrosion,  carbide precipitation (hence
sensitization) and pitting corrosion.  These alloys are further improved
by addition of Mo to provide increased resistance against  pitting corrosion
and Nb-Ta for increased stabilization against  sensitization.  One such
alloy is Armco 22-13-5 containing  5% Mn  with additional Mo and Nb.  The
Carpenter 20Cb-3 is also a similar alloy but with a low Mn content.

     Armco 22-13-5 (Nitronic-50) austenitic  stainless steel is superior to
316 or 316L in corrosion resistance  and  costs  more than 316L.  It also
combines high strength with corrosion resistance and  can be useful in
stressed parts such as pumps,  valves,  screens  and wire supports.

    Table  43 compares the corrosion  rates of Armco 22-13-5 with 316 and 316L
in different corrosive media.
                                   251

-------
  TABLE 43.   COMPARISON  OF  CORROSION RATES OF ARMCO  AND  316/316L STEELS
     Test Medium
Armco 22-13-5              316/316L
                                                                 2
  10% Fed-, Amb.  Temp.   <0.001  gm/in2,  no pits      0.0112  gm/in  pitted
          O
  10% H2S04, boiling      0.356  Ipy  (in  per year)    0.73  Ipy
  20% H2S04, boiling      1.64  Ipy                   2.2 Ipy
  Nitric acid test        0.0072 Ipy
  Ferric sulfate +        0.0108 Ipy
  Sulfuric acid                                           	
     This alloy is recommended for stressed parts  and if chloride ion stress
corrosion cracking or intergranular corrosion  is  a problem with conventional
300 series stainless steel  alloys.

     Types 347 (Nb + Ta), 321  (Ti), 309 Cb, and 310 Nb are stabilized
steels which can be used in the as-welded condition.   Data on their corrosion
properties in H2S04 - Fe2(S04)3 are not readily available.

     Organic inhibitors  used in pickling tanks may be useful  in minimizing
corrosion problems.  Some examples are thiourea,  dibutyl sulfoxide, amines,
0-polythiorea, quinoline derivatives,  ketones, etc.  Proteins such as milk
albumin are useful in providing a protective film.  Also 0.2% AS in FLSO.
greatly reduces corrosion.   These may  be considered if they do not adversely
affect the leaching process.

     Teflon or Kynar (polyvynilidene fluoride) coatings, may provide
additional protection in view  of the fluctuating corrosion conditions.
Kynar can be coated inside small tubings.  Tough, uniform teflon coatings
are possible to apply on large tubings and inside reaction vessels.
Sleeving with thin-walled teflon tubing is also a possibility.  Teflon is
completely stable in boiling H,,S04 and  Fe2(S04)3  solutions and Kynar is
also stable to a high degree.
                                   252

-------
 Survey of Literature  Data  on  Corrosion

     Tables 44 and 45  summarize available data on materials for conditions
 somewhat similar to those  under consideration.  A comparison of these data
 with the results of the  tests conducted in this study (Table 48, page     )
 indicated the importance of generating corrosion data under specific con-
 ditions.

        TABLE 44.  CORROSION  DATA FROM LITERATURE ON 304 AND 316SS15
Ferrous
sulfate
-
-
Mostly

H2S04
150 ppm
270 ppm
With free
acid (pH=l)
6% free
acid
Ferric
sulfate
603 ppm
9424 ppm
-
1.5%
Temp
Corrosion Rate, Ipy
Time
61 days
61 days
100°F 67 days
160°F
7 days
*Inches
304
0.017 <0
<0.0001 <0
0.0005 0
<0.0001
per year
316
.0001
.0001
.0004
-

 7.2   EXPERIMENTAL  CORROSION STATIC TEST PROGRAM

 7.2.1  Experimental  Method

      Figure  35  shows the stainless steel bomb assembly used  to  carry out
 the static corrosion test  under simulated plant conditions.   Four 1-1/2"
 sections of  1/2" O.D. tubings  of different candidate  materials  were
 welded together using the  TIG  process  without filler  rods as  seen in
 Figure 36.   This served  as a support for the  sheet  candidate  materials to
 be tested, which were machined into flat tensile test coupons after welding
 in pairs.  Thirty-two specimens were assembled to the configuration shown
 in Figure 37.  The coupons were held in  radial  slots  on the periphery of
teflon support rings fixed to  the  tube with 304 S.S.  screws.  All the
test coupons were machined flat and  polished  with 3/0 emery paper MO
P-inch finish) and were  used in the  polished  condition after  welding with
                                    253

-------
                               TABLE 45.   CORROSION DATA FROM LITERATURE ON VARIOUS  ALLOYS15'16'17'18
Corrosive
Medium
(1) FeS04, H2S04
PH=1
(2) 50% H2S04
42 g/1 Fo2 (S04)3
(3) 1.5% Ferric Sulfate
+4$ Free Acid
(4) 15% H2S04
(5) 15% H2S04
(6) 80% H2S04
(7) Stress Corrosion Test
42% MgCl2 Soln
(8) 4% Ferrous Sulfate
Ferric Sulfate
H2S04
PH 2.5
(9) 60% Solids g/1
28-55 gm/R H2S04
5-10 g/1 Fe+++
Some ferrous
0.1% NaClOj
Temp
F°
100

Boiling
160
176
104
176
Boiling
162


113


Time
Days
67

1
31
—
1
—
1000
hrs.
27


41


Aera-
ation
X

—
—
—

—
— —
X





Agita-
tion
X

—
X
—

—
— —
X






Type 304
0.0005


.0001
—
0.03
—
——
0.017





Average Corrosion Rate (ipy)*
Type 316
0.0004


.0001
—
nil
—
——
0.0006





Incoloy
825
—


—

—
—
"• •"
__


.0001


Inconel
625
—

0.070
--
0.007
—
0.090
"" "™
--





Hastelloy C-4
Un-
Weld-'d


0.111
—

—
—
•• —
--





As weld-


0.114
--

—
—
No
Cracks
--


—


ro
en
     Inches  per year

-------
 Figure 35. Bomb Test Assembly
                                   :j  ..... :   i
                                              . Ujiiii
Figure 36.  Composite Welded Tube Used In Test'
            304L, Armco and 315L
From Left to Right 304,
                                   255

-------
Figure 37.   Test Coupon Stack:  32 Tensile Coupons Assembled Around The
            Central Welded Tube With Teflon Support Rings
                                256

-------
no special annealing heat  treatment.   The tubes were also used in  the as-
welded condition.  All  test  coupons were carefully weighed and sized prior
to loading in the bomb.

     The  bomb was filled with ferric sulfate leaching solution used in  the
process.  The flanges  were bolted together with a high temperature gasket
seal.  The valve  fitting on  the outside of the top flange enabled  the
bomb to be pressurized with  02-  Heating tape was used to attain the
temperature.

     Table 46 lists the materials and conditions used in the  test.

7.2.2  Initial  Runs and Modifications

     A chromel-alumel  thermocouple (TC)  encased in a  304 S.S.  protective tube
was used  initially and the TC being immersed in the corrosive  solution with
the aid of a T-fitting on  the flange.  After only 16 hours  of  testing, the
bomb had  to be  disassembled  because the thermocouple sheathing (304 S.S.)
was completely  corroded away.  The 304 S.S.  screws used to  hold the teflon
rings showed signs of  corrosion.   The bomb was cleaned and  reassembled
without the TC  which was now placed on the outer bomb wall.  The gradient
across the S.S. wall of the  bomb  was  assumed to be negligible.

     After 300  hours  of testing at temperature and pressure,  the bomb was
disassembled  for replacement of the gasket.   It was found that the bomb
interior  and  test coupon stacks were covered by a heavy brown  deposit, and the
304 S.S.  screws that were  used to keep the teflon support rings in position
were very severely  corroded, with some so completely reacted  on that they
crumbled  to a powder  on touching.  The coupons themselves showed no signs
of corrosion.   The  stressed contact points between S.S. screws and the
metal tubing were also pitted.

     For further continuation of  the  test, the  304 screws were replaced
and new solution was used.   The test  coupons and bomb were cleaned of all
deposits.   The  bomb was  heated for a  second  period of 700 hours (solution
exchange at 300 hours  should not  have had a  material  impact on the test).

                                   257

-------
                TABLE 46.   LIST OF  MATERIALS  TESTED
                                                WELD.
 Composite Support Tube  (welded)

             Tensile  Test Coupons:
Temperature
Pressure
Time
             *260°F
              TOO psig
                                      316L
                                   Helded Metal Pairs
                                   304 - 304
                                   316 - 316
                                   316L - 316L
                                   347 - 347
                                   321 - 321
                                   Armco - Armco
                                   Carp - Carp
                                   Inconel -  Inconel
                                   Incoloy -  Incoloy
                                   316 - 304
                                   316L - 304
                                   Armco - 316
                                   Armco - 316L
                                   Incoloy -  Armco
                                   Carp - 316
                                   Carp -  316L
                                   321  -  316L
                                    321  -  316
                                    347  -  316
                                    347  -  316L
                                    347  -  304
              1000 hours (300 hours - first run + 700 hours  -  second run)
* A temperature excursion to 320 F for an unknown time between 24 and 60
  hours occurred during the course of the test initially
                                  258

-------
     At the end of the second period of exposure for 700 hours (total of
1000 hours on the test coupons) it was observed that the bomb and the
samples were covered by a very heavy brown to blackish-brown deposit.  Part
of the 304L portion of the composite tube between the weld areas 304/304L
and 304L/Armco was completely corroded away as shown in Figure 38.   The
heavy deposit is partly from the corrosion products and partly also from
the solution, which was now dark green in color, with the total iron content
down by 44% and nearly all ferrous.  The change in solution characteristics
is shown in Table 47.

                 TABLE 47.  FERRIC SULFATE CORROSIVE MEDIUM
Total Fe
r ++
Fe
Initial
50
13
(mg/g)


Brown heavy deposit None
Color
Cr/Ni
S04 and
Cl - ions
Brown
-NOT AN
300 hrs.
46
21
Considerable
Green
A L Y Z E D -
700 hrs.
28.4
28.1
Very heavy
Darker Green

     In order to obtain weight losses, the tough adhering deposits had to
be removed.  The procedure used was as follows.  The heavier buildups were
removed by mechanical brushing with a nylon brush and acetone in an ultra-
sonic bath.  Some of the more adherent deposits were loosened by thermal
cycling between 100°C and 0°C.  The samples weighed after the mechanical
removal still showed a weight gain as compared to the original,  indicating
incomplete removal of the oxide deposits.  Hence the following pickling
procedure was used:

     a)  60 minute dip in 20% NaOH + 5% KMn04 at 200°F
     b)  10 to 30 seconds in 15% H2S04 at 180°F
                                    259

-------
                                        Weld
                                                         S 2/3X
      Figure 38.  Comparison of Corroded And Control  Tubes.
BOTTOM:


TOP:
Tube After 1000 Hours Exposure to Corrosive Environment.
304L Tube Disintegration
                                           NOTE:
Control Tube:
The tubes have been aligned such that weld zones are
one below the other.  The missing portion (% 1 inch
of 304L) of the exposed  tube nas been corroded away.
From left to right, the sections are 316L, Armco, 304L,
and 304.  The threaded portion at the left end of the
exposed tube was cut off before the test.  The teflon
support ring is seen on the left.
                                    260

-------
This procedure removed the oxide deposit effectively without attacking
the bare metal.  Type 316 S.S. was attacked somewhat by the H2$04 solution.
A set of control samples was also subjected to identical treatments in
order to correct for weight loss from the bare metal due to the pickling
procedure.  Since the oxide deposits on the coupons could only have
inhibited the reaction of the pickling bath on the metal substrate, the
weight losses obtained after correction must represent a minimum figure.
However, the corrections are small and this effect is not likely to be
appreciable.

7.2.3  Results and Conclusions of the Bomb Test

     The corrosion rates are calculated from the weight loss and surface area
and density for each coupon in units of milligram (loss) per square
declmeter (mdd) per day.  This can be converted to the more conventional
inches penetration per year units (ipy) using the factor mdd x *    .  —' =
ipy where d - density (gms/cc).  The corrosion rates were found to vary by
a factor of three in duplicate samples of the same material.  This large
variation is due to the position of the sample in the bomb stack and the
degree of protection from additional corrosion afforded by the accumulation
of the surface deposits.  But the differences between two types of materials
is marked.  It should be emphasized that due to the static conditions in the
bomb, the corrosion rates obtained are probably lower than for the actual
dynamic conditions and are more meaningful as relative figures than as
absolute corrosion rates.

     Table 48 gives a summary of the corrosion rates, visual appearance
and tensile test results obtained.

     Several metallographic mounts were prepared from the test and control
samples to study the extent of corrosion and the corrosion mechanism.
Mounts were made from near the weld zone as well as near each end of the
grip section.  Figures 39 and 40 show the microstructure of the exposed
(corrosion tested) 304 S.S. coupons and  Figure 41 shows the microstructure
of a 304 S.S. control sample.  It was observed that while sensitization
                                   261

-------
TABLE 48.  COMPILATION - COAL CORROSION MATERIAL EVALUATION
ISpecimer
1 Number
LiA
MB
U
I 2B
3A
1 38
1 3C
4A
1 4B
1 SA
1 SB
1 6A
6B
L»
1 7B
8
i 9
I 10A
I 10B
1 11A
|m
hr-
1 13
,4
1
u
P7_
F8
ri9
p°
r 2i
^^^^•^M
Hatl.
304/304
316/316
~316/3T6~~
316/316
316L/316L
316L/316L
316L/316L
347/347
347/347
321/321
321/321
ARM/ARM
ARM/AWT
CARP/CARP
CARP/CARP
NCONEL/
INCONEL
NCOLOY/
TNTOI 0V
316/304
316/304
316L/304
316L/304
ARH73TT
ARM73T6L
INCOLOY/
ARM
CARP/316
CARP/31 6L
321/316L
321/316
341/316
"347/316"
" 347/304
^MMBMM^^H
m^mmmtm
Y.S.
44 ."4
35.7
1O
36.5
44.4
44.6
47.6
43.0
41.0
65.0
43.1
58~.0
"62.2
56.5
56.6
74.5
56.7
39.2
39.0
"46.4
46.9
To^r
' 49'rr
56.1
"3O
51.0
44.6
35.8
39.0
45.3
45.1
^^^^^
MMMUM
^^iH^BM
UTS
92.5
78.4
~74."6"
64.7
79.8
78.5
84.0
88.2
70.0
88.4
~W.T
100". 2
99.1
94.2
"96.3
1T7."7
91.8
77.4
79.4
87.4
90.2
^7¥.¥
8SL6
95.6"
"73.r
83.6
82.8
65.3
80.0
76.7
88.6
^^>BHMHMB
•^•MMM
s
M^MHI
Elong.
40
~ 43
34"
15
26
25
27
35
24
23
-~26"
23"
TT
20
23
18
15
35
41
34
36
~3T
"31
23
--34-
29
25
22
30
20
40
^^^•MMM
^^•MMM
JBJECT
^WM
Break
M
316
~3~16
316
W
W
H
H
347
M
321
W
U
W
H
W
W
316
316
W
w"
"3T6"
W"
W
316
H
W
321
W
U
347
l^^MM
^— ^— —
ED SPECIMENS
^^~^~~i~^^^
Defect
Few sml voids
O.K.
Few sml voids
outside welds
o'.k.
Few sml voids
outbids wel ds
O.K.
O.K.
O.K.
Couple voids
... _at..we1d .....
Few sml voids
.outside welds.
Small spot
at uejd
Small speck
at wel d .
Nothing
of sia.
O.K.
Small voids
large @ weld
O.K.
Various" voids
large @ weld
O.K.
O.K.
O.K.
O.K.
Numerous voids
at weld
Large void at
weld
Couple voids
outside weld
Couple sml.
voids. out. wld.
Small
abt.weld
O.K
O.K.
O.K.
O.K.
O.K.
O.K.
MM^^MM^^^^^^H
^MHBMMI
^••^^•MMB
Poor
Poor
Poor
Poor
Poor
Good
Good
V.Good
Poor
"Poor
'Good
V.Good
V.Good
T.G6ocr
V.Good
V~:Sood
V.Good
V.Good
>oor
V.Poor
Fair
Fair
Fair
Good
V.Good
Poor"
V.Good
Fair
Poor
V.Poor
Fair
Poor
•^^^MV
^^•WH
^HM^HM
0.0680
0.0510
0.1189
0.3783
0.3934
0.2477
0.3432
0.0832
0.1889
0.4019
0.1665
0.0632
0.0056
Not'Used
0.0127
6."0022
0.0170
0.0072
0.2974
6.2399
0.0630
0.0576
0.7649
0.0190
~ff."D014
0.1965
0.0105
0.0975
0.4524
0.3346
0.0833
0.2921
^^^^^^^^H
^^•••^
Gra/Cn3
7.89
7.91
7.98
7.95
7;98
7.99
8.03
7.91
7.93
7.93
7.91 "
7.86
7.92
7.91
8.07
8.04 ""
8.46
8.16
7.92
7.95
7.94
7.94
7.95
7.95
8.04
8.00
8.00
7.99
7.93
7.96
7.96
7.92
^•^^^"
••Ml
"spT!
1A
IB
2A
2B
2C
3A
3B
3C
4
5
6
7
8
9
10
11
12
T3
14
15
16
17
18
19
20
21
I^^H
•^•MMH^
^••^••ii^
304
304
316
316
316
31 6L
31 6L
31 6L
347
321
Armed
Carp
Inconel
Incoloy
316/304
316L/304
ARM/316
ARM/31 6L
INCOLOY/
ARM
CARP/316
CARP/3161
321/316L
321/316
347/316
347/316L
347/304
^^••^^MH
••••••
^^^^m
48.4
44.1
42.3
39.9
40.7
44.1
49.0
44.3
44.7
45.5
56.6
50.4
74. 2
48.1
42.4
47.7
43.7
49.6
52.8
42.3
49.1
41.7
42.5
40.8
40.7
45.0
H^MM
i^MHMMM
^ .
92.7
92.2
80.0
84.7
78.5
81.4
83.9
80.4
73.8
90.4
103.9
92.5
123.9
86.1
84.5
87.0
85.2
89.2
93.2
76.9
89.5
80.6
83.2
82.5
78.0
92.1
•MMM
CONTRC
»/
V
V
>l
V
V
V
V
V
>/
\
V
-J
V
V
V
Co
>/
V
V
V
V
V
V
V
V
V
V
V
>l
>/
V
V
V
V
V
V
>l
V
V
V
rrosion
Rate
.0010
.0024
.0075
.0079
.0049
.0068
.0017
.0037
.0080
.0034
.0014
.0001
0.0002
;0.0003
0.0001
0.0059
0.0048
0.0013
0.0011
0.0032
0.0004
0.003
d.ooo
0.002
0.009
0.006
0.001
0.006
^^^™
1 	






^••••Ml
RATINGS
^orrosiorT
5
8
7
9
6
4
2
3
1
Ratio;
1 - t
MM^^^^^^^^M
Mech 1 Edge Cracking!
Prop A |Hetallographic|ylsual
A
B
A
?
A
A
A
A
A
1 Codes:
jest, large
A =
B =
idM^BMO^^^B^
1
10
4
3
1
1
1
1
1
st number « v
No apparent c
Slight change
.L— ^—
9
8
6
7
5
3
4
2
1
*orst
hange
\

-------
                                                        200X
Figure 39.  Microstructure of 304 Test Coupon  After 1000  Hours Exposure
                      '

                                                 '-
                     -.                    ,.
                                              .:' .

                       -    -   ••      ..        -:"••     -
                                                    ' /^
         * ,
        •
                                 •
                                                    .  -,„  •
                                                        ».


                                             •
        •

          .        i              !    -   :   "•        'If

         -;:                     ".                  ^

         :4,^  •  '      -    •    '  •                 •     :.
              •                   *

           ""- >••%.  /               .*"       ,  «         "-v
              :      /            •                 r
           i  •- '.                                 *.  -

           \ '- •

           v                        '      • ,
                                           * •" ." ,  .  • ,   •»».

                                       •• W  ...;...  . V •




                                                      400X


Figure 40.   Microstructure of 304 Test Coupon at 400X  Showing

            Sensitization at Grain Boundaries


                              263

-------
                                 •;->A  i-    •'     -   -~Y
                                       <         .. -.   .
                                                     - .    •••,-*-

                    ' -i-  >- -  "\  ;       '  -.'-   •       -r^'-;-'  i


                                                           200X
Figure 41.  Microstructure of Unexposed  Sample of 304 Test Coupon
                                264

-------
had indeed occurred at some distance from the weld zone, intergranular
corrosive attack was barely evident.  Pitting corrosion was not seen here.

     Figures 42, 43 and 44 show that very severe pitting corrosion occurred
just outside the weld area in 316.   Both large and small pits  were observed,
and some of the large corroded areas were considerable in size.

      Figures 45 and 46 show the edges of the mounted specimens of a con-
trol  and exposed sample of 316.  Cracks near the outer  edge due to an
embrittling effect of the corrosion are evident in Figure 46.   Figures 47
and 48  similarly show the edges of a corroded and exposed sample of 316L.
Cracking and pitting due to corrosion are evident.

      Figures 49 and 50  show cracked  and exposed edges  of 347.  Other areas
of corroded 347 showed evidence of pitting similar to 316 though not quite
as extensive.

      Figures  51 and  52 show  areas of Incoloy 825  in the  unexposed and ex-
posed conditions.  There is evidence of either cracking or corrosion.

      Figures 53 and 54 show the unexposed and exposed areas near the weld
of the  304L/304 joint.  It is  recalled  that the 304L tube near the weld
was so  severely attacked that  part of the tube  fell apart as seen in
Figure  38.   Figure 55 shows the corroded area  on 304 tube.   Figure 56
shows the corroded inner-edge  of the 304L tubing.  Figure 57 shows the
weld  and parts  of 316L and Armco tubing.  Severe pitting exists on 316L
with  much less  on the Armco tubing.

     Figure 58  shows the uncovered (top) portion of the 316L tubing and
bottom  portion  which was covered tightly with teflon tape.  The much more
severe  pitting  in the unprotected area  is evident.

     Figure 59  shows cracks in the I.D. of Armco tube while the O.D.
Figure 60 shows only pits.   Figures  61  and  62 shows,  respectively, the
O.D.  and I.D. of the exposed 316L tubing.
                                   265

-------
          • V 4 " ^*'w * " ' '•"'
                                      «••'     ^ -


                                                        m
                                                         100X
Figure 42.  Micrograph  at 100X Showing Very Extensive Pitting Type of
            Corrosion Just Outside the Weld Zone  in 316SS After  Exposure
                            • --
                                                         100X

Figure 43.  Micrograph   at  100X  Showing  Very  Extensive  Pitting  Type  of
            Corrosion Just  Outside  the Weld Zone  in 316SS After Exposure
                                  266

-------
                                         •'  '  '-'•*
        _  •    ....     •         -         ,r-> ..,,*
       •-.--..•                    ••       -»* • -j

                   3     ..        -^

        ',.*':""•:-. •-,
                       - ••                  -•>'*•**$!&
        •  ^         •"        • .        > -   .!•. - ":»^S
         *-~'     • •                        "%^j -x-'?ri
      |.^-'--
        ' •
      ,

      '
          *• '**•
          ---«,.
      *•-••  .•  ;"-^          ;-;^li
                                              100X
Figure 44. Micrograph at 100X Showing Very Extensive Pitting Type of
         Corrosion Just Outside the Weld Zone in 316SS After Exposure
                           267

-------
                  ~*  ~
                ;  .1*5


                                                        .
                                                               .
                  ,
                                                            200X

Figure 45.  Edge of 316 Control Sample Showing Elongated Stringers Due to
            the Tensile Test
                                                     .
                                                           200X

Figure 46,   Edge of Exposed Test Coupon of 316 Showing Transgranular Cracks
            Extending Inward From Cut Edge
                                  268

-------
                                                        200X
 Figure 47.  Edge  of  316L Control
          i  ,•      •  •,.-,-•:,.   -  «      .  •;;         -

                                          .--.•-•
                       --•   •  •'•              :*.
         -„ , ^  •• •               . ..            —
                    -                        •  4  '    ;' '  i-
                        • • v  -   •   -         •
                         •                 "          *  -v- '

                              • • •' :             -       -•'
                   -- .        •>      •       -.  -       -  •.

                  •  •  - •
                  •  •
                                 -

                        1   -•
>.
           1



                 •

           '•:•• ,

         S3* i  ,",
               , » -

                       -^  •
        *cr  -.
                                   . •              .       .





                                                         200X
Figure  48.   Edge of  Exposed 31 6L Showing  Crack Extension Along Crack

            Sensitive Corroded Paths and  Pitting
                               269

-------
       J^ T-'-iV '• -   •••—•-*..  -       -  ?*~'   ~
       |fe>         •• -•- --.:,                ^*r-
       &4i
-------
          '

                                       •
                                           -
          (-*
          '










                                                      • J^
                                                      •
                                                           400X
Figure 51.   Edge  of Incoloy 825  Control
          >






                 '  -
                             -
        c

   -





£
                          .
                                                    -
                      •
                -
                                                           400X
 Figure  52.   Edge Of Exposed Incoloy 825 Test Coupon
                               271

-------
                                      •.


                                  '
            ,
           •' *'
             '
         '—,y •* • >   -"• -            .•          ,
                                   -  #.*,
        . >,~^-<~  *
                                  *    *
                         •  '• - u ••  -:#*•*&.  .
                 ...— • f
                                                                  .304L

                                                                  Parent

                                                                  Metal
                                                           200X
Figure 53.  Weld Between 304 And 304L  In  The Control  Tube Sample
                                                             1
                            L-          ?  •'                -/*
                                    '•  .'>-
                                                          200X


Figure 54.  Weld Between 304 and 304L  Tubes Showing Severe Corrosion

            Effects After Exposure
                                     272

-------
                                    -
                                                -  ,
                                                           200X
Figure 55.  Outer Diameter of 304 Tube After Exposure Showing Corrosion Along
           The Edge And Pits On Surface
                                      273

-------
            . .
                                                             200X
Figure 56.   Inner Edge Of Exposed 304L Tubing Showing Corroded Areas
                                   274

-------
                                     ARMCO
                                    •316L
Figure 57.   Exposed Tube Sample
              275

-------

                                  Exposed
                                  Section
                                  Teflon Tape
                                  Protected
                                  Section
                              7X
Figure 58.   Exposed 316L Tube
            276

-------
                     ...

                                                           100X
Figure 59.   I.D. of Exposed Armco Tubing Showing Some Cracks And Corroded
            Areas
              . - :

                      V
                                                          100X

Figure 60.  O.D.  of Exposed Armco Tubing Showing Stringers Along The Tube
            Drawing Axis and Light Pitting
                                 277

-------
                                                         100X
Figure 61.  O.D.  of Exposed 316L  Tubing  Showing  Corrosion Pits
                                                        100X
Figure 62.   I.D.  of Exposed  316L  Tubing  Showing  Less  Severe Corrosion
                               278

-------
     Table 49 gives change in yield strength  and  elongation  after exposure
based on an arbitrary ±  5% allowed variation  in yield  strength and a ± 10%
allowed variation in elongation.  The  only material  that  showed signifi-
cant degradation in both strength and  ductility was  316SS.
     Table  50 gives a summary of qualitatively observed events during the
course of the test.  It is evident that most of the 300 series stainless
steels showed alarmingly poor corrosion resistance in the desulfurization
process environments.

     The results of the present study show that the rates of corrosion of
the 300 series stainless steels are extremely sensitive to the metallurgical
condition of the sample, particularly residual stress.  This is brought
out dramatically in the case of the 304L tubing (Figure 38) which was com-
pletely corroded away due to the built-in stresses in the tube drawing
process.  It is well known that in an acidic environment, the stresses
could cause increase in corrosion rate several-fold especially at tempera-
tures in the region of 100°C.  A possible mechanism is the precipitation of
carbides and nitrides or agglomeration of C or N atoms at dislocation
sites which causes local action cells in plastically deformed material.
This is also borne out by the observation that 316 corrodes just outside
the weld zone where cooling stresses are greater and the fact that the
welded seam in Armco tubing, where some of the drawing stresses must have
been annealed out, is free of corrosion.

     The cracking observed along the cut edges of 316, 316L, 347 and
Armco tubing also is indicative of embrittling due to stress corrosion.
The cracking appears to be transgranular, following crack-sensitive paths
determined by loci of local action cells formed due to residual stress or
cold working.   Whether Cl" ions play a role in aiding stress corrosion by
adsorption is  not known since the solution has not been analyzed for Cl
ions.   The sorption may occur selectively along paths of pinned dislocations
or vacancy-alloying element clusters formed due to plastic deformation or
residual  cooling stresses after welding.  The fact that 304L (extra low
                                   279

-------
ro
oo
o
                        TABLE 49.  CHANGES  IN  MECHANICAL PROPERTIES OF ALLOYS AFTER EXPOSURE
                                   TO IRON  SULFATE  SOLUTION
Material
Yield Strength
Control + 5%
304

316
31 6L

321
Arraco

Carp
Inconel

Incoloy
347
45.98

38.95
43.5

43.2
53.8

47.9
70.5

45.7
42.5
- 50.82

- 43.05
- 48.09

- 47.80
- 59.40

- 52.9
- 77.9

- 50.5
- 46.9
Exposed
44.5
(~)
36.0
45.5
(0)
54.1
60.1
(")
56.5
( +)
74.5
(0)
56.7
42.0
Elongation
Control t 10£
35.6

33.9
23.4

27.9
23.4

18.9
19.8

15.3
13.5
- 43.40

- 41.50
- 28.6

- 34.1
- 28.6

- 23.1
- 24.2

- 18.7
- 16.5
Exposed*
40.0
/ \
(o)
30.0
26.0
m \
(o)
25.0
(-)
22.0
(\
-0)
21.5
(0)
18.0
(_ \
-0)
15.0
(-0)
30.0
               (+) indicates increase
               (-) indicates decrease
               (o) indicates that the values lie within the allowed range

-------
       TABLE 50.   SUMMARY  OF  QUALITATIVELY OBSERVED EVENTS DURING
                   CORROSION TESTING
    •  304 thermocouple  shield  tube  disintegrated  after  a  short  timers.)
       (16 hrs) exposure
    t  304 screws  used  initially  were  severely  corroded

    •  304L tubing was  severely corroded  in  the composite welded tube

    t  Type of Corrosion  (Plate Materials)
              Pitting type:       316,  316L,  347
              General type:       304,  321, Armco, Carpenter, Inconel,
                                  Incoloy
carbon, ELC, grade by Sandvik  Corporation was completely corroded away also
suggests the above mechanism as no sensitization can occur in the very low
carbon grade stainless  steel.

     The pits seen in 316 and  347 may be due to accelerated corrosion at
isolated stress points and/or  local action cells due to concentration
differences.

     The extreme sensitivity of corrosion rates to residual stresses is
also apparent from the fact that 304 was considerably more stable than 316
and 347 in sheet form while it was severely corroded in drawn tubes and
machined screws.

7.2.4  Conclusions and Recommendations

          •  304, 304L and 316 appeared to be marginal  as construction
             materials to be used in the coal desulfurization reactor.
          a  Dynamic  testing in an environment where oxygen is in
             plentiful  supply may improve the stability of some of
             these alloys.

                                   281

-------
 7.3  DYNAMIC TESTING

      A composite welded tube, made up of 1-1/2" sections of tubing of
 different candidate hardware materials for the coal desulfurization process
 was tested in  the  L-R  rig for approximately 50 hours of simultaneous coal
 leach and reagent  regeneration.  The experimental tubing replaced a section
 of the flow pipes  in the L-R rig; hence the tube interior was exposed to
 the erosive environment of flowing mixture of coal fines and ferrous/ferric
 sulfate solution.  The objective of this test was to evaluate and compare
 the corrosion  resistance of the various candidate materials under dynamic
 processing conditions  and  in the presence of  finely dispersed oxygen.

      Post-exposure examination was conducted  on the test tubing.  The tubing
 was first cut  into two halves longitudinally  to expose the interior.  After
 cleaning superficial dirt by blowing with compressed air, the different
 sections were  examined with a low power microscope for gross evidence of
 corrosion. Figure 63 shows a picture of the composite tubing before it was
 cut. Figure 64 shows the results of low magnification  microscopic examin-
 ations at various  locations.  There were no visible changes in  areas where
 no remarks are noted.  Faint corrosion spots  and adhering stains were ob-
 served but no  gross corrosion was evident on  even the worst area.   Figures
 65 and 66 show the 304 and Armco sections, respectively, under  low magnifi-
 cation.   It is evident that no apparent corrosion has occurred  in any of
 the tubing materials.
                                    s.
      Three sections were selected for microscopic examination:
           (1)   304-Armco weld area
           (2)   316-321 weld area
           (3)   Carpenter 20Cb/3-304L weld area.

     Figures 67 through 70 show the microstructure of the parent metal  and
the weld areas.  Intergranular corrosion found in some alloys in previous
investigations  was  not  observed in  any of the materials evaluated in this
investigation.   Sensitization (shown as black spots) was seen in the weld
                                   282

-------
areas as shown by the presence of precipitated chromium carbide but
corrosion at the grain boundaries  is  not discernible.

     Negligible corrosion was observed for all alloys tested for 50 hours
in the coal desulfurization test unit.  Apparently, the presence of
quantities of available oxygen may inhibit the corrosion of the stainless
materials investigated.  However, the test duration was too short to be
definitive.  Much longer times in excess of several hundred hours would
probably be required before a measurable amount of corrosion can be
observed.
                                    283

-------
                                                                                               Carp
                                                                                              ZO/16/3
304
IN >
00
I •
                               Figure 63.    Composite Tube Used  in  Dynamic Test
         Note:  two end  fittings are 304 SS.  Adapters Al  through A4
                are 304  SS

-------
                                Small
                              Corrosion
Stains
c * Sta1
Spots
I i
j ^^
304 1 1 Armco
304ELC
ns
1 ^
Armco 1 316

i l&^f^3 I
L 1 /321 ! Ai
' / ; 4
/* V \, I
rmco 1 347 I
i 1 1

      Thin Crack Loose Black    Small Holes
               Deposit on Weld  in Weld
     Small      Spotted-,
     Stain      Faint Stains
                                                                                     Al
                                              Inconel
A2
ro
03
01
                                         Faint Spots
Incoloy
I - 	
A3

1; 	 , 	 .-
CARP

. 1
3

r
04L
_i

A4
304

                                                                                  A1-A4:   304 Steel  adapters.
                                                                                           Sections  used for
                                                                                           mounts.
                     Figure 64.   Composite  Welded Tubing:  Corrosion Effects in the Tube  Interior.

-------

    Figure  65.    Section  of Armco  Tubing Interior.
                  Corrosion  Effects
 2X


No Gross
                                                     "U» • ,
                                                     2X

Figure 66.   Section  of 304 Tubing Interior.   No Evidence
             of Corrosion
                              286

-------
•r*  \      \ s   ,  • ,  .

,
            -  -   ..   •
 .
                   :

                   .'
      •••.  •     '.•,-•   .
    -..  .
 • .
               .     '  f-
     .
            '
               •  •  -••  .-
                                 .

                                              • » • .-'•'*

                                       '":•> •
                                                        .*•  • •
Figure 67.   Weld Zone of Armco-304 Showing  Chilled  Structure

             and Sensitization by Chromium Carbide
                                                        400X
  Figure 68.   Structure of 304 Away  from  Weld
                               287

-------
             -
               •'  •

               .

.

                                            ••->•..;<•.«  •  v
                                             '  .'." >'.' .'  •  , -


                                             •  H'; •"''•'•:,"•/;•" "•'.' '

                                                     •:;-'  '}•;'
      .
.
                                                       .
                                     • ,  .  •   ••. • . ,'i .  • -•. .->-:- .
                       •  •-                 .--.  •.  ;.,- ,;,.-/•. i
                               ••. v  .  •   •'..•• ,.<\t*'&V.-y
                       , .         .-:•"• ' -  . ..  '  -.  ...;   ':-.-'.'-..  .
                                         .•^.t'fi-K r.:"- ;
                                S •• •••• •••: • rf-fyQ &M
                                •-•••/.'     *   -   •   "/ v* ,»  . •
                             ••/-::•  r^t .:'•-•; •/••••••.-•.-••• \X. &/•*•
                                                        400X

Figure  69.   Weld  Zone of 316L-321  Showing  Sensitization
                    •-*      . --'             - .  '    /      .  ;

                                          *      -  •*' ^   ,'*
       .'   1         • •     •  •      .- ./        •-,..'.
                      • ••
                                       -• .' ^        ^ '    .
                               - '   ^ -    -   -•         ^  '
                                                        400X
    Figure  70.    weld Zone  of 304L-Carpenter 20 Cb/3  Showing Chilled
                  Structure  and Grain Boundary Sensitization
                              288

-------
                       APPENDIX TO SECTION 7

Corrosion Characteristics of  Steel Alloying  Elements
      Element(s)                             Property
Ni + Cr                        Resists  oxidizing  chemicals
High Ni + Mo                   Resists  non-oxidizing solns
High Mo + Cr/Si +              Resists  boiling H0SO.
High Ni + Cr                                    2 4
High Ni + Mo                   Resists  fluctuating oxidizing salt
                               solutions
High Co, Ta, Ti, Low C         Resists  sensitization during welding
High Ni                        Freedom  from  chloride ion stress
                               corrosion cracking
Mn + High Ni                   Improved resistance to pitting, stress
                               corrosion cracking and increased strength
                               and stabilization
                                   289

-------
                                       APPENDIX  TO  SECTION  7
                            CHEMICAL COMPOSITION OF CANDIDATE  MATERIALS TESTED  IN WT. PERCENT

Inconel 625
316
31 6L
304L
§ Armco 22-13-5
Carp 20-C5-3
304
321
347
Incoloy 825
Ref
(17)
(15)
(15)
(15)
(19)
(20)
(15)
(15)
(15)
(16)
Cr
20-30
16-18
17-21
17-21
20-23
19-21
18-20
17-19
17-19
19-23.5
Mi
57-60
10-14
9-13
8-12
11-13
32-38
8-12
9-12
9-13
38-46
Fe _
Mo (Nb+Ta) c
5 8-10 3-4 o.l
Rem* 2-3 0.08
Rem 2-3 0.03
Rem
Rem
Rera
Rem
Rera
Rem
Rem
0.03
1.5-3.0 0.1-0.3 0.06 Max
2-3 SxC-1.0 Q.07 Max
0.08
0.08
lOxC Min o.08
2.5-3.5 - 0.05
Ii li
0.4 0.5
1.0
1.5
2.0
1.0 Max
1.0
1.0
5xC Min 1.0
1.0
0.0-12 0.5
Mn
2.0 Max

4.
2.
2.
2.
2
1

0-6
0
0
,0
.0

.0




.0 Max
Rem = remainder

-------
                              8.   REFERENCES
 1.    Hamersma, J. VI., E. P. Koutsoukos, M. L. Kraft, R. A. Meyers, G.  J.
      Ogle, and L. J. Van Nice, "Chemical Desulfurization of Coal:   Report
      of Bench-Scale Developments",-EPA R2-73-173 a and b, prepared for
      the Office of Research and Monitoring of the Environmental  Protection
      Agency, Research Triangle Park, N. C., February, 1973.

 2.    Hamersma, J. W., M. L. Kraft, C. A. Flegal, A. A. Lee,  and  R. A.
      Meyers, "Applicability of the Meyers Process for Chemical Desulfuri-
      zation of Coal:  Initial Survey of Fifteen Coals", EPA-650/2-74-025,
      prepared for the Office of Research and Monitoring of the Environ-
      mental Protection Agency, Research Triangle Park, N. C., April, 1974.

 3.    Hamersma, J. W., and M. L. Kraft, "Applicability of the Meyers
      Process for Chemical Desulfurization of Coal:  Survey of Fifteen
      Coals", EPA-650/2-74-025-a, prepared for the Office of Research and
      Monitoring of the Environmental Protection Agency, Research Triangle
      Park, N. C., September 1975.

 4.    "Program for Processes for the Selective Chemical Extraction  of
      Organic and Pyritic Sulfur from Fossil Fuels", Document No. 17270-
      6011-RO-OO, Contract EHSD 71-7, prepared for the Office of  Research
      and Monitoring of the Environmental Protection Agency,  Research
      Triangle Park, N. C., January 15, 1975.

 5.    Leonard, J. W., and D. R. Mitchell (Editors); Coal Preparation; Am.
      Inst. of Mining, Metallurgical and Petroleum Engineers, Inc.,
      New York, N. Y., 1968.

 6.    Guthrie, K. M., "Capital Cost Estimating", in Modern Cost Engineering
      Techniques, ed. H. Popper, McGraw-Hill, 1970, pp 80 to  108.

 7.    Mills, H. E.,  "Costs of Process Equipment", in Modern Cost  Engineering
      Techniques, ed. H. Popper, McGraw-Hill, 1970, pp 111 to 134.

 8.    Happel, J., and D. Jordan, Chemical Process Economics, 2nd  ed.,
      Marcel Dekker, Inc.,  1975.

 9.    Perry, R., and C. Chilton, Chemical Engineers' Handbook, 5th  ed.,
      McGraw-Hill, 1973.

10.    "Chemical Marketing Reporter", Schnell Publishing Company,  June 2,
      1975, p. 44.

11.    "Final Report - The Supply-Technical Advisory Task  Force - Synthetic
      Gas-Coal", prepared by the Synthetic Gas-Coal Task  Force for the
      Supply-Technical Advisory Committee, National Gas Survey,  Federal
      Power Commission, dated April, 1973.
                                     291

-------
                          REFERENCES (CONTINUED)


12.   Barringer Research Ltd., Rexdale, Ontario, Canada, "Feasibility Study
      fdr Sensing Sulfides in Coal", Final Report, Natl. Center for Air
      Pollution Control Contract, PH86-67-270, Proj.  385, Report TR-68-55.

13.   Comparison of Oxidation and Reduction Methods in the Determination
      of Forms of Sulfur in Coal:  Environmental Geology Notes #66,
      Illinois State Geological  Survey, December 1973.

14.   Hurley, R. G., and E. W. White, New Soft X-Ray Method for Determining
      the Chemical Forms of Sulfur in Coal, Anal. Chem., Vol. 46, #14,
      December, 1974.

15.   "Corrosion Resistance of the Austenitic Chromium - Nickel Stainless
      Steels in Chemical Environments".  The International  Nickel Co.,
      N.Y., 1963.

16.   Incoloy 825 Spec Sheet, Huntington Alloy Products Division,
      International Nickel Co.,  Huntington, W. Virginia.

17.   Inconel 625 Spec Sheet, Huntington Alloy Products Divisions,
      International Nickel Co.,  Huntington, W. Virginia.

18.   Hastelloy C-4 Spec Sheet,  Cabot Corporation, Kokomo,  Indiana.

19.   ARMCO-Product Data Bulletin S-45a, Armco Steel  Corporation, Baltimore,
      Md.

20.   CarTech Selecting Carpenter Stainless Steels, Carpenter Tech Corpo-
      ration, Reading, Pa.
                                    292

-------
                 9.  GLOSSARY  OF ABBREVIATIONS AND SYMBOLS
Abbreviations
    Abs
    ASTM
    btu
    cal
    eq
    Exp.
    Kcal
    No.
    wt
absolute
American Society of Testing Materials
British Thermal Unit
calories
equation
experiment
kilocalories
number
weight
Symbols
    A,
    C
    A
    EL

    ER

    KL
    K
     R
    v
    P
    P(
    R
    r,
Arrhenius constant in leach reaction (hours)"
(wt% pyrite in coal)~'
Arrhenius constant in regeneration reaction
(minutes)'1 (atm)"1 (liters/mole)
concentration
difference in quantity following delta
activation energy for pyritic sulfur leaching
reaction, Kcal /mole
activation energy for ferric ion regeneration
reaction, Kcal /mole
pyritic sulfur leaching rate constant (units
same as A|_)
ferric ion regeneration rate constant (units
same as
     R
micron.
total pressure, atmospheres
oxygen partial pressure

gas constant, cal /mole, °K
pyritic sulfur leaching rate, weight of pyrite
removed per 100 wts of coal per hour
ferric ion regeneration rate, moles per liter
per minute
                                    293

-------
Symbols (cont'd)
    S                     elemental  sulfur
     n
    S                     organic  sulfur
    S                     pyritlc  sulfur
    S$                     sulfate  sulfur
    S.                     total  sulfur
    T                     absolute temperature,  °K
    t                     time,  hours  (leaching)-minutes  (regeneration)
    W                     pyrite concentration in coal, wt%
    Y                     ferric ion to  total iron  ratio
                                   294

-------
                                 TECHNICAL REPORT DATA
                           (Please read Inunctions on the reverse before completing!
 1. REPORT NO.
  EPA-600/2-76-143a
                            2.
                                                        3. RECIPIENT'S ACCESSION-NO.
 4. TITLE AND SUBTITLE
 Meyers Process Development for Chemical
 Desulfurization of Coal, Volume I
                5. REPORT DATE
                 May 1976
                6. PERFORMING ORGANIZATION CODE
 7.AUTHOR.S, E.P.Koutsoukos, M.L.Kraft, R.A.Orsini,
 R. A. Meyers, M. J.Santy, and L. J. Van Nice
                8. PERFORMING ORGANIZATION REPORT NO
 9. PERFORMING ORQANIZATION NAME AND ADDRESS
 TRW Systems Group
 One Space Park
 Redondo Beach, California  90278
                10. PROGRAM ELEMENT NO.
                1AB013; ROAP 21AFJ-033
                11. CONTRACT/GRANT NO.

                68-02-1336
 12. SPONSORING AGENCY NAME AND ADDRESS
                                                        13. TYPE OF REPORT AND PERIOD COVERED
  EPA, Office of Research and Development
  Industrial Environmental Research Laboratory
  Research Triangle Park, NC  27711
                13. TYPE OF REPORT AND I
                Final; 6/73-12/75
                14. SPONSORING AGENCY CODE
                 EPA-ORD
 15. SUPPLEMENTARY NOTES Project officer L.Lorenzi, Jr. is no longer with EPA; for details
 contact Lewis D. Tamny, Mail Drop 61, Ext 2851.
 16. ABSTRACT
               repOrf. gives results of bench- scale development of the Meyers Pro-
 cess (for chemical removal of sulfur from coal) for desulfurization of both fine and
 coarse coal.  More than 90% of the pyrite was  removed from run-of-mine (ROM) fine
 coal and clean coarse coal,  and more than 80% of the pyrite from ROM coarse coal.
 Process improvements were demonstrated involving:  increased process slurry solids
 concentration for higher process  throughput (33% w/w); lowered  filtration require-
 ments through use  of larger top- size fine coal; and generated elemental sulfur remo-
 val by vaporization from the coal matrix.  Pyrite leaching and reagent regeneration
 rate expressions were validated.   Engineering studies  showed that:  the process may
 be engineered in a  number of basic designs including simultaneous leach and regener-
 ation, separate  leach and regeneration, use of oxygen or air for regeneration, fine
 or coarse coal  processing, and combined with coal cleaning; these process designs
 lead to stand-alone full capital cost estimates of $30-80/KW of power plant name-
 plate capacity for the various process plant cases.  Assuming ROM coal costs of
 $0. 81 /MM Btu,  the required market price of the desulfurized coal ranges from
 $1 . 14 to $1 . 32/MM Btu  on a utility financed basis and $1 .41 to $1 . 86 /MM Btu on
 an investor financed basis.
 7.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b.lDENTIFIERS/OPEN ENDED TERMS
                            c. COSAT1 Field/Group
 Air Pollution
 Coal
 Desulfurization
 Pyrite
   Air Pollution Control
   Stationary Sources
   Meyers Process
   Chemical Coal Cleaning
13B
08G,21D
07A,07D
 8. DISTRIBUTION STATEMENT

 Unlimited
    19. SECURITY CLASS (ThisReport)
    Unclassified
!1. NO. OF P
    312
    20. SECURITY CLASS (Thispage)
    Unclassified
                            22. PRICE
EPA Form 2220-1 (9-73)
295

-------