&EPA
United States      Industrial Environmental Research EPA-600//-79-013a
Environmental Protection  Laboratory         January 1979
Agency        Research Triangle Park NC 27711
Reactor Test Project for
Chemical Removal of
Pyritic Sulfur from Coal;
Volume I. Final Report
         Interagency
         Energy/Environment
         R&D  Program Report

-------
                  RESEARCH REPORTING SERIES


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

    1. Environmental Health Effects Research

    2. Environmental Protection Technology

    3. Ecological Research

    4. Environmental Monitoring

    5. Socioeconomic Environmental Studies

    6. Scientific and Technical Assessment Reports  (STAR)

    7. Interagency Energy-Environment Research and Development

    8. "Special"  Reports

    9. Miscellaneous Reports

This report has been  assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series  result from the
effort funded  under the  17-agency  Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by  providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects;  assessments  of,  and development of, control technologies for  energy
systems; and integrated assessments of a wideirange of energy-related environ-
mental issues.  '
                        EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for  publication. Approval does not signify that the contents necessarily reflect
the  views and policies of the Government, nor does mention of trade names or
commercial products constitute endtirsemerit or recommendation for use.

This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                EPA-600/7-79-013a

                                      January 1979
    Reactor Test Project for
Chemical  Removal of Pyritic
         Sulfur from  Coal;
      Volume I.  Final Report
                      by

             R.A. Meyers, M.J. Santy, W.D Hart,
              L.C. McClanathan, and R.A. Orsini

            TRW, Defense and Space Systems Group
                  One Space Park
              Redondo Beach, California 90278
                Contract No. 68-02-1880
               Program Element No. EHB527
             EPA Project Officer: Lewis D. Tamny

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

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

-------
                                   ABSTRACT

     Plant checkout and shakedown was completed at the  end  of September and
initial plant process performance was evaluated on an Appalachian coal.  Oper-
ation of the plant through January of 1978 demonstrated that the Reactor Test
Unit (RTU) could be run continuously in three-shift operation to reduce the
coal from 2,4 Ibs S02/106 Btu to a level  of 1.0 to 1.2  Ibs  S02/106 Btu, after
rinsing and extraction of generated elemental  sulfur.  There was no measurable
coal oxidation during processing and leach rates in the plant were greatly
improved over bench-scale values.  The leach solution/coal/oxygen environment
was found to be corrosive to the installed stainless steel  reactor, necessi-
tating future upgrading to support additional  testing.   Bench-scale experimen-
tation showed that the leach solution can be used as a  homogeneous dense-media
to efficiently gravity-separate coal prior to  processing.   Beneficial  engineer-
ing cost improvements are obtained based on using this  approach, resulting in
capital cost estimates of $68-$69/KW and with  $0.44-$0.50/106 Btu processing
costs, including amortization of capital, for  input coal costing $0.78-
$0.81/106 Btu.  Overall energy efficiency was  93 to 96  percent.
                                       ii

-------
                               ACKNOWLEDGMENTS

     The authors wish to acknowledge the valuable assistance received in this
project from the following TRW personnel:  Dwight Wever, Bill Bowes, John Hunt
and Ed Phipps for supervising operation of the test plant; William Bradshaw
and Allen Keenan, lead plant technicians; Jack Denson and Larry Ledgerwood for
control lab operation and experimental support; Edwin Moon, Debbie Hopp and
Betty Cruz for engineering analysis support; Lou Resales and Maurice Bianchi
for materials study support; Bernard Dubrow, Les Van Nice and Elias Koutsoukos
for managerial assistance and manuscript review; Roger McGough, subcontract
manager; Verna Melough, Sharon Cavin, Monique Tholke and Velnia Butler for
technical typing and Marilyn Jennings for report coordination and finalization.
The authors also want to acknowledge the support of Dave Tamny, Jim Kilgroe
and T, Kelly Janes, the monitoring project officers and sponsoring managers at
the Environmental Protection Agency.
     Gratitude is due to the American Electric Power Service Corporation for
providing 300 tons of coal for this initial operation phase of the plant and
most particularly to Sam Ruggeri of that organization.
                                        Ill

-------
                          METRIC CONVERSION  FACTORS

     In compliance with EPA policy,  metric units  have  been  used  extensively in
this report (followed by British units  in parentheses).   However,  in  some cases,
British units have been used for ease of comprehension.   For  these cases, the
following conversion table is provided:
                British
            1  Btu
            1  Btu
            1  kw
            1  hp (electric)
            1  psi
            5/9  C°F-32)
            1  inch
            1 ft
           1 ft2
           1 gallon
           1 pound
           1 ton (short)
            Metric
252 calories
2,93 x 10"4 kilowatt-hours
1,000 joules/sec
746 joules/sec
0.07 kilograms/cm2
oc
2,54 centimeters
0,3048 meter
0,0929 meters2
0.0283 meters3 or 28.3 liters
3.79 liters
0.4536 kilograms
0.9072 metric tons
                                    iv

-------
                                  CONTENTS
ABSTRACT	   li
ACKNOWLEDGEMENTS	  .  .  111

METRIC CONVERSION FACTORS	-	   1v
FIGURES	  - viii

TABLES	   xi
 1.   INTRODUCTION  	    1
 2.   CONCLUSIONS	    8
 3.   RECOMMENDATIONS	   11
 4.   REACTOR TEST  UNIT	   12
      4.1  Process  Description  	   12

           4.1.1  Coal Feed System	   15
           4-1.2  Leach Solution Feed System	   15
           4.1.3  Oxygen Feed System	   20
           4.1-4  Process Steam-	   20
           4.1.5  Coal-Reagent Mixing	   22
           4.1.6  Primary Reaction  and Reagent Regeneration	   22
           4.1.7  Secondary Reaction 	   24
           4.1.8  Filtration and Leach Solution Recovery  	   24
           4.1.9  Coarse Coal Leaching and  Reagent Depletion  	  .   26
           4.1.10 Instrumentation	   28
           4.1.11 Process Sampling  and Analysis	   28
      4.2  Process  Equipment	   30

           4.2.1  Coal Transport Bins	   31
           4.2.2  Bin Tilter	   31
           4.2.3  Coal Storage Tank	   31
           4.2.4  Weigh Belt	   31
           4.2.5  Tank Farm	   32
           4.2.6  Leach Solution/Slurry Pumps	   32
           4.2.7  Heat Exchangers	   34
           4.2.8  Slurry Mixing Tank	   34
           4.2.9  Reactors	   35
           4.2.10 Flash Drum	   36
           4.2.11 Scrubbers	   36
           4.2.12 Filter and Vacuum System  	   37
           4.2.13 Og/Ng/Air Supply Systems  	   38
           4.2.14 Steam Supply System	   38

-------
                           CONTENTS (Continued)                            Dano
                                                                           Kayc

                                                                             •^o
          4.2.15 Process Water System 	  •    ^°
          4.2.16 Instrumentation	    -*°
          4.2.17 Control Laboratory 	    ™
          4.2.18 Capistrano Test Site Support Facilities	    41

5.   REACTOR TEST UN'IT OPERATION	    42
     5.1  RTU Shakedown Activities	    42
          5.1.1   RTU Safety Review and Training	•  •  •    43
          5.1.2   Shakedown Operations 	    43

     5.2  RTU Operation	    48
          5.2.1   Operating Procedures 	    48
          5.2.2   Equipment Experience 	    50

     5.3  RTU Operation Conclusions	    55

6.   REACTOR TEST UNIT DATA	    57
     6.1  Summary of Previous Bench Scale Experimentation  	    57

     6.2  Data from the Reactor Test Unit	    61
          6.2.1   Mine-cleaned Martinka Coal Characterization	    65
          6.2.2   Coal Processing Data	    72
          6.2.3  Slurry Sampling	    99
          6.2.4  RTU Mass Balance Data	104

     6.3  Data Analysis Conclusions	113

7.   SUPPORTING BENCH-SCALE EXPERIMENTATION 	  .  .   117
     7.1  Introduction	117
     7.2  Gravichem Separation	   117
          7.2.1   Background	   117
          7.2.2  Equilibrium Gravichem Separation  	  ....   119
          7.2.3  Non-Equilibrium Separation 	   131
     7.3  Processing of RTU Coal	   143

          7.3.1  Pressurized Bench-Scale Processing  .	   146
          7.3.2  Ambient  Pressure Bench Scale Processing	   146
          7.3.3  Weathering of RTU Coal	   146
          7.3.4  Density  of Coal Slurry	   148
          7.3.5  Elemental Sulfur Recovery	   150
     7.4  Laboratory Study Conclusions	   157

8.   ENGINEERING DESIGN AND COST ESTIMATION	   159
     8.1  Introduction and Background 	    159

          8.1.1  Historical Background	    160
          8.1.2  Current  Generalized Process Concept	    161
          8.1.3  Base Case Design, Cost Estimate,  and               —
                 Economics Approach 	  	    169


                                    vi

-------
                            CONTENTS (Continued)
                                                                            Page
      8.2  Design and Cost Estimate Base Cases	   177
           8.2.1   Base Case 1  -Mine Cleaned Martinka Coal	   177
           8.2.2  Base Case 2  - TVA Kentucky No. 9 Coal	   190
      8.3  Engineering Analysis Conclusions	   203
           8.3.1   Base Case Results	   203
           8.3.2  Cost Sensitivity Analysis	   206
 9.   VENDOR TESTING	   210
      9.1  General Approach	   210
           9.1.1   Equipment Suppliers Survey and Selection  	   220
           9.1.2  Vendor Test Implementation Plan	   223
           9.1.3  Vendor Testing Results	   224
           9.1.4  Coal Supply  for Vendor Studies	   225
      9.2  Vendor Testing Conclusions	   225
10.   MATERIALS OF CONSTRUCTION EVALUATION 	   228
      10.1  Inspection and Analysis of RTU Equipment	   229
      10.2  Analysis of Test Samples	   241
      10.3  Discussion	   256
      10.4  Corrosion Study Conclusions	   258
11.   REFERENCES	   260
12.   GLOSSARY OF ABBREVIATIONS AND SYMBOLS	   263
                                     vii

-------
                                   FIGURES
Number
                                                                            Page
   1    Meyers Process 	   2
   2    Reactor Test Unit	   3
   3    Reacted Coal on Belt Filter	   5
   4    Gravichem Process	   6
   5    RTU Schematic	  14
   6    RTU Bin Lift	  I6
   7    RTU Bin Tilter	  17
   8    Coal Feed to Fine Coal Storage and the Coarse  Coal  Leacher ....  18
   9    Coal Feed to the Mixing Unit	  .  19
  10    RTU Tank Farm	  21
  11    RTU Mixer and Primary Reactor	  23
  12    Primary Reactor and Belt Filter	  25
  13    Vacuum Pump and Filtrate Receivers	  27
  14    RTU Control/Instrumentation Console	  29
  15    Comparison  of Capistrano and  Warner  Laboratories  Pyritic
        Sulfur Analyses	  68
  16    Comparison  of Capistrano and  Warner  Laboratories  Sulfate
        Sulfur Analyses	  69
  17    Mine-Cleaned Martinka Coal Size  Distribution Data	  71
  18    Pyritic  Sulfur Analyses of Processed Coal  from Experi-
        ment  01-03	  76
   19    Arrhenius Plot of Data from RTU  Experiments 01-01 to 01-11 ....  81
   20    Reagent  Y Data from RTU Processing Experiments 01-01 and
        01-02	  90
   21    Summary  of  Estimated Sampling Bias	101
   22    Slurry Concentration Variation as a  Function of Sample
        Location	103

-------
                             FIGURES  (Continued)

Number                                                                      Page
  23    Gravichem Process	   ng
  24    Hot-Briquetted Coal	   121
  25    Bench-Scale Gravichem Test Apparatus Flow Diagram	   123
  26    Gravichem Processing of TVA Coal 	   125
  27    Puritic Sulfur Leaching for 1.3 Specific Gravity Sink Coal
        at 102°C	   132
  28    Coal Oxidation After Two Hours at 100°C as a Function of
        Ash Content	   133
  29    Particle Size Distribution in Fractions of RTU 3 Coal	   138
  30    Bench-Scale Coal Leaching and Reagent Regeneration
        Apparatus	   133
  31    Elemental Sulfur Recovery as a Function of Acetone
        Extraction Stages	   154
  32    Effect of Water on Elemental Sulfur Recovered by Successive
        Acetone Stages 	  155
  33    History of the Gravichem Process	  162
  34    Gravichem Process Block Diagram Suspendable Coal Approach	  163
  35a   TRW Coal Desulfurization Gravichem Process 	  178
  35b   TRW Coal Desulfurization Gravichem Process 	  179
  36a   TRW Coal Desulfurization Gravichem Process 	  195
  36b   TRW Coal Desulfurization Gravichem Process 	  196
  37    Base Case 1.  Upgrading Cost vs. Battery Limit Capital	  208
  38    Base Case 2.  Upgrading Cost vs. Battery Limit Capital	  209
  39    Base Case 1.  Upgrading Cost vs. ROM Coal Cost	  210
  40    Base Case 2.  Upgrading Cost vs. ROM Coal Cost	  211
  41    Base Case 1.  Upgrading Cost vs. % Pyritic Sulfur Removed	212
  42    Base Case 2.  Upgrading Cost vs. % Pyritic Sulfur Removed	213
  43    Base Case 1.  Upgrading Cost vs. ROM Coal Cost at Different
        Levels of Pyritic Sulfur Removal 	  214
  44    Base Case 2.  Upgrading Cost vs. ROM Coal Cost at Different
        Levels of Pyritic Sulfur Removal 	  215
  45    Base Case 1.  Upgrading Cost vs. % of Total Coal to Float
        Processing . . .•	216
                                      1x

-------
                             FIGURES (Continued)
Number
                                                                            Page
  46    Base Case 2.   Upgrading Cost vs. % Total Cost to
        Float Processing.  .  .	   ^'7
  47    Base Case 1.   Upgrading Cost vs. ROM Coal  Cost	   218
  48    Base Case 2.   Upgrading Cost vs. ROM Coal  Cost	   218
  49    Flush Port Fitting	   231
  50    Porous Spongy Structure 	   232
  51    Pipe Weld Corrosion	   233
  52    Crevice Corrosion in Reducer Fitting. ....	   234
  53    Flange Crevice Corrosion	   235
  54    Baffle Bolts  and Nuts from R-l Cell 4	   237
  55    Weir Baffle	   238
  56    TE-56 Thermocouple Probe	239
  57    R-l Reactor Internals After Experiment 03 Showing Pits
        Where Grinders Had Operated	   240
  58    R-l Reactor Internals After Experiment 03, Showing Smooth
        Bottom Pits at Welds and on Free Surfaces	   242
  59    Coupon Mounting Rack and Test Speciments	   245
  60    316L Cres Coupons Showing Signs of Crevice Attack During
        Experiment 01 (L) and Smooth Bottom Pitting During
        Experiment 03 (R)   	   246
  61    304 Stainless Steel  Coupons from Experiment 03 At and
        Above Water Line (L) and Below Water Line (R)	   247
  62    Inconel and Incoloy Alloy Coupons, 7.5X:  Inconel 601
        (Upper L), Inconel 617 (Upper R), Inconel  625 (Lower L),
        Incoloy 825 (Lower R)	   249
  63    Hastelloy C-276 Coupons From Experiment 03 At and Above
        Water Line (L) and Below Water Line (R)	   250
  64    Titanium Coupons:  T1-50A (L) and Ti-12 (R), from
        Experiment 01 (Top)  and Experiment 03 (Bottom)  	   251
  65    Lead Coupons from Experiment 01.  Surface Marked by
        Cleaning Operation.   No Pitting or Crevice
        Corrosion.  1.5X	                     252

-------
                                    TABLES

Number                                                                      Page
   1     RTU Pump Data	   33
   2     RTU Shakedown Sequence	   45
   3     Equipment Ranges Utilized 	   51
   4     RTU Channel-Scan at Fixed Time	   63
   5     RTU Time-Scan by Channel	   64
   6     As Received Mine-Cleaned Martinka Coal Analyses	   66
   7     Size Distribution Data for Mine-Cleaned Martinka Coal
         Processed in the RTU	   70
   8     Operating Parameters for Acidified Iron Sulfate Reagent
         Leach Experimentation	   75
   9     Summary of Processed Coal Analyses Obtained During RTU
         Experimentation with Acidified Iron Reagent 	   77
  10     Operating Parameters for Low Iron Reagent Leach
         Experimentation 	   83
  11     Summary of Processed Coal Analyses Obtained During RTU
         Experimentation with Low Iron Reagent	   84
  12     Excess Fe+2 Generation Data from Processing 14 Mesh x 0
         Mine Cleaned Martinka Coal with 5% w/w Iron Reagent	   89
  13     RTU Reagent Regeneration Data	   91
  14     Elemental Sulfur Product Generation Data from RTU
         Processing	95-96
  15     Summary of Heat Content Changes in RTU Processed Coals	   98
  16     T-2 Mass Balance Data from RTU Processing Experimentation ....  109
  17     Oxygen Mass Balance Data Obtained During RTU
         Experimentation 	  Ill
  18     Equilibrium Gravichem Tests on 14 Mesh x 0 Martinka Mine
         Coal with Variation of Medium and Separation Time	122
  19     Equilibrium Gravichem Tests at 80°C on 3/8" Topsize
         Kentucky No. 9 Mine Coal with Variation of Medium	126
  20     Particle Size Distribution of Size-Reduced Kentucky No. 9
         Sink Coal	  127

                                        x1

-------
                              TABLES (Continued)
Number
  21      Gravichem Processed Kentucky Mine  No.  9 Coal	128
  22      Sulfur Removal  from Coal  with Ferric Sulfate  Solutions 	  13°
  23      Comparative Results of Cleaning	134
  24      Analyses  of Fractions of  Martinka  Coal  Obtained from
         Nonequilibrium  Float/Sink Experiments	135
  25      Analyses  of Fractions of  Martinka  Coal  Obtained from
         Nonequilibrium  Float/Sink Experiments	•  •  •  •  137
  26      Relative  Settling  Rates of Coal  and  Pyrite  Subjected to
         Gravitational and  Centrifugal  Forces 	  140
  27      Flotation and Centrifugal  Separation of Coals	  141
  28      Yield of  Martinka  Coal  Separated in  Nonequilibrium Float/
         Sink Experiments after 0.5 Hours @ 80°C (S.G.  1.3)	  142
  29      Leaching  of Martinka Coal  at 80°C	  145
  30      Bench-Scale Processing of Martinka Coal.  .  :  	  147
  31      Distribution of Inorganic Sulfur in  RTU Coals  (%)	.  .  149
  32      Elemental Sulfur Content  of Nonprocessed Coal.  .  . ,	  150
  33      Density of Coal/Water Slurries  	  :151
  34      Product Elemental  Sulfur  Recovery  from Leach Martinka Coal ....  153
  35      Sources of Equipment Cost Information	  172
  36      Sources of Operating Cost Information	  173
  37      Battery Limits  Operating  Costs  Format	  174
  38      Economic  Evaluation Criteria Utility Financing  	 176-177
  39      Base Case 1  - Process Equipment List	184-188
  40      Battery Limit Process Costs -Base Case 1	  .  189
  41      Base Case 1  - Process Economics	  191
  42      Base Case 2 -Process Equipment List	197-2C1
  43      Battery Limits  Process  Costs - Base  Case 2  	  202
  44      Base Case 2 - Process Economics	  204
  45      Summary of Base Case Economics  Results 	  .....  205
  46      Vendor Selection Program  	 221-222
  47      Coal  Cake Collected for Vendor  Testing	   226
  48      Coupon Corrosion Test Results - 29 November Inspection 	   244
                                       xi i

-------
                              1.   INTRODUCTION

     The Meyers Process is a technology for chemically removing essentially all
of the pyritic sulfur from coal through a mild oxidative treatment.  Important
pollutant trace elements of lead, cadmium, and arsenic are removed at the same
time.  It is particularly cost effective for providing compliance coal  for
industrial boilers and smaller electric utilities, and for recovering and
desulfurizing waste fine coal rejected from mining and washing operations.
     A process schematic is shown in Figure 1.  Coal is mixed with an aqueous
solution of ferric sulfate (Step 1), previously derived from the coal,  to form
a slurry.  The slurry is raised in temperature to 100°-130°C (Step 2) where the
ferric sulfate oxidizes the pyritic sulfur content of the coal to form elemen-
tal sulfur and additional iron sulfate.  At the same time oxygen or air is
introduced to regenerate the reacted ferric sulfate.  Iron sulfate dissolves
into the leach solution while the elemental sulfur is removed in a second
extraction (Step 3).  The coal is dried and solvent recovered (Step 4).  The
products of the process are iron sulfate, which may be limed to give a  dry
gypsum and iron oxide material and elemental sulfur.  Trace elements from the
coal are rejected from the leach solution with the stabilized gypsum-iron oxide
solid.  Elemental sulfur is the most desirable product which can be obtained in
the process of controlling sulfur oxide pollution since it may be easily stored
without additional pollution or may be marketed.  The gypsum-iron oxide product
is a safe and storable solid product.
                                                     M p 3}
     After testing the process on some 40 U.S. coals*  *  *  ' and performing
200 fully material balanced bench-scale extractions/  *  '  the value  of the
process for controlling the sulfur content of coal was firmly established and
the data necessary for design of a test plant was then available.  Engineering
design background obtained through extensive studies at both TRW*  '  '   and
various engineering organizations*''8' provided confidence that the  process

-------
         H2S04
                                                   Co O-
ro
   COAL
   (Fe S9)
                                 Fe2(S04)3
                                                   FILTER
                              NEUTRAL-
                               IZATION
                                                              Co
Fe S2 + Fe

Fe S0 +
Fe SO4

(S04)
FILTER
COAL
                                                                   SOLVENT
                                                                               STILL
                                            Figure 1.  Meyers Process

-------
was economically attractive and indicated important engineering data which  could
only be obtained at test plant scale.

     It was determined that the initial test plant evaluation should concentrate
on the key process steps of coal-leach solution slurry formation, leaching,
regeneration and filtration.  A test plant, termed the Reactor Test Unit, was
constructed at TRW's San Juan Capistrano site for the purpose of testing these
portions of the process (Figure 2).  The plant, sized to process from 1/8 to
1/3 ton/hour of coal was dedicated in April, 1977.
                          Figure  2.  Reactor Test Unit

-------
     Plant checkout and shakedown  was  completed at the end of September and
initial  process  performance was  evaluated on coal  donated by the American
Electric Power Service Corporation (AEP)  from its  Martinka mine (Appalachian
coal).   Operation  of the plant through January of  1978 demonstrated that the
Reactor Test Unit  (RTU) could be run continuously  in  three-shift operation to
reduce  the AEP coal  from 2.4 Ibs S02/106  Btu to a  level  of 1.0-1.2 Ibs S02/10
Btu, after rinsing and extraction  of generated elemental  sulfur.  The coal
product is shown as a cake  on the  plant belt-filter in Figure 3.  There was
no measurable coal oxidation.    Leach rates in the RTU were greatly improved
over bench-scale values, allowing  elimination of the  use of a second process
reactor for completing pyrite removal. The  leach  solution/coal/oxygen environ-
ment was found to  be corrosive in  the  installed stainless steel reactor,
necessitating future upgrading of  the  reactor material  of construction to
support additional testing.
     Supporting bench-scale experimentation  showed that the iron sulfate-
sulfuric acid leach solution can be used  as  a homogeneous dense media to effi-
ciently gravity-separate fine coal  at  specific gravities  of 1.2 to 1.35.  Ben-
eficial  engineering cost improvements  are obtained, based on using this
gravity-separation effect,  whereby a significant portion  of the input coal
which floats in the leach solution and is almost pyrite  free, may bypass the
reactor, elemental sulfur extraction and  dryer portions  of the Meyers Process
(Figure 4).  This  revised technology is termed the Gravichem Process.  When
applied, at bench-scale, to a Tennessee Valley Authority (Interior Basin) coal
containing 12% ash and 7 Ibs of  S02/106 Btu, two products are obtained, a 4%
ash float coal  containing 3 Ibs  S02/106 Btu  and an 11-12% ash sink coal con-
taining 4 Ibs S02/106 Btu after  treatment by the Meyers  Process.  Both of these
products meets State SO  emission  standards  for this  coal.  Because of  these
                       A
promising results, TVA has  shipped 300 tons  of coal to TRW to demonstrate
desulfurization in the RTU.
     Laboratory  experimentation  also showed  that the  solvent system, acetone
and water,  is the  most  economically attractive method thus far investigated for
removal  of generated  elemental sulfur  from treated coal.
     Process  cost  forecasts  based  on both RTU data and  supporting  laboratory
experimentation  for  the Gravichem  Process  are $68-69/KW  capital  cost with

-------
Figure 3.  Reacted Coal  on Belt Filter

-------
cn
                     H2S04-
                                        Fe2(S04)3
                                                                                H20
             Fe
                              GRAVITY
                             SEPARATOR
                                (FeS)
                                                               Co O-

NEUTRAL-
IZATION
                                                  REACTOR

                                                     §>
             SOLVENT
            FILTER
                                                                                 SOLVENT
                                                                                            STILL
                                            MEYERS PROCESS
            A
                                                                                                SULFUR

                                                      Figure 4.   Gravichem Process

-------
$0.44-0.50 10 /Btu, processing costs, including utility financed capital
amortization, for  input coal costing $0,78"0,81/10b/Btu.  Overall energy effi-
ciency for these process designs  including  both coal use for process heat and
electric energy for plant operation is  93%  to 96%,
     The report is organized into nine  sections to follow.  Sections 2 and 3
present the  conclusions and recommendations, respectively, which resulted in
this study.  The RTU  is described in Section 4 and the operation of the plant
in  Section 5.  A complete set  of  RTU data together with analysis of the data is
presented in Section  6.  The supporting bench-scale  laboratory results are found
in  Section 7 and engineering design and full-scale process cost estimation in
Section 8.   A  plan for vendor  testing of process  unit operations, not built into
the RTU,  is  described in  Section  9 and  a detailed plant materials evaluation is
presented in Section  10.  The  referenced appendices  are continued in Volume II
of  this report.

-------
                               2.   CONCLUSIONS
PLANT OPERATION RESULTS
        1.   The  Reactor  Test  Unit  (RTU) can  be  operated  continuously for
            testing  of Meyers Process  units  for coal-leach  solution  mixing,
            simultaneous coal  leaching and leach solution regeneration,
            filtration of leach  solution  from treated  coal  and water wash-
            ing  of coal  on the filter.

        2.   The  input coal from  American  Electric Power  Service  Corpora-
            tion's Martinka mine in  Fairmont, West Virginia,  containing
            1% inorganic sulfur  can  be reliably and continuously reduced,
            in the RTU,  to a  pyritic sulfur  level  of 0.16%  without any
            measurable coal loss and with coal  heat content increases
            averaging 350 Btu/lb.

        3.   RTU  coal product, after  bench-scale extraction  of residual
            sulfate and  elemental  sulfur, was continuously  and reliably
            reduced to a total sulfur  content of 0.68-0.75% and  1.0-1.2
            Ibs  S02/106  Btu.

        4.   Leach rates  in the RTU were improved over  bench-scale values by
            an average factor of five  due mainly to favorable coal seg-
            regation in  the primary  reactor.

        5.   Plant Teacher-regenerator  operation at temperatures  ranging
            from 230°-270°F (1100-132°C), pressures of 30-80 psig and
            residence time of 5-8  hours was  demonstrated.

        6.   The  use of a single  reactor-regenerator is sufficient to meet
            design basis pyrite  removal and  provide regenerated  leach
            solution. The use of  a  secondary reactor  to complete the
            reaction of  pyrite with  leach solution, found to be  necessary
            at bench-scale, was  demonstrated to be unnecessary in the
            Reactor Test Unit.

        7.   The  leach solution/coal/oxygen environment caused corrosion
            in the primary reactor-regenerator  system  indicating that
            upgrading of the  316L  material of construction  for improved
            corrosion resistance is  needed  to  support further testing.

        8.   The  following materials  were  found  to be suitable for leach
            solution-coal service  at temperatures up to  90°C: fiber rein-
            forced plastics,  elastomers and  316L stainless  steel. The

-------
            following materials were determined to be suitable for service
            at reactor-regenerator temperatures:  up to 130°C, titanium,
            Hastaloy, and rubber-lined brick over mild steel.

        9.  No significant corrosion was observed in the leach solution-
            coal mix tank, flash-down tank, storage tanks or reactor-
            regenerator pumps.

SUPPORTING EXPERIMENTATION RESULTS

       10.  The iron sulfate-sulfuric acid leach solution can be used as a
            homogenous liquid to efficiently gravity-separate fine coal
            from pyrite rich coal at specific gravities of 1,2 to 1,35.

       11.  Beneficial engineering cost improvements are obtained by
            using this gravity-separation effect to bypass a significant
            portion of the input coal around the reactor, elemental sulfur
            extraction and dryer units of the Meyers Process.  This
            revised process  is  termed the Gravichem Process.

       12.  The Gravichem Process provides two  products with no coal
            reject, a float  coal containing 2-4% ash with almost no pyritic
            sulfur and a sink coal generally lower in ash than the input
            coal and also nearly pyrite-free  after treatment by the >
            Meyers  Process.   The two products can  be  used separately or
            combined,

       13.  Bench-scale testing of the Gravichem Process on the American
            Electric Power Service Corporation  (Appalachian) coal gave two
            products:  a float  coal  containing  1.0 Ibs S02/106 Btu and a
            sink coal containing 1.1 Ibs SQ2/106 Btu after treatment by
            the Meyers Process,  Both products  met New Source Performance
            Standards.

       14.  Bench-scale testing of the Gravichem Process on a Tennessee
            Valley Authority (Eastern Interior  Basin) coal containing
            12% ash and 7 Ibs of S02/106 Btu gave two products:  a 4% ash
            float coal containing 3  Ibs S02/106 Btu and an 11-12% ash
            sink coal containing 4 Ibs S02/106  Btu after treatment by the
            Meyers Process.

       15.  The solvent system, acetone and water, is the most economi-
            cally attractive method  thus far investigated for removal of
            generated elemental  sulfur from treated coal.  This solvent
            also dissolves and  removes residual iron sulfate.

ENGINEERING DESIGN RESULTS

       16.  Process cost forecasts for the Gravichem Process are $68-69/KW
            capital cost with $0.44-0.50/106 Btu, processing costs includ-
            ing utility financed capital amortization, for input coal
            costing $0.78-0,81/106 Btu.

-------
17.   Coal  energy efficiency is 94-97% for the Gravichem Process
     including coal  use for process heating,

18.   Overall  energy efficiency including both coal  use and electric
     energy for plant operation is 93-96%.
                              10

-------
                     3.  RECOMMENDATIONS
1.   Design and replace the corroded primary reactor with an
    improved vessel constructed of either titanium or rubber-
    lined brick over mild steel,

2.   Construct and test in integrated operation, a leach-solution
    gravity-separator unit to evaluate Gravichem separation and
    Meyers Process treatment of sink coal.

3.   Test the operation of the plant on a coal from the Eastern
    Interior region of the United States, such as the Tennessee
    Valley Authority coal, 300 tons of which is now on hand.

4.   Construct and test in integrated operation, Meyers Process
    units for continuous residual sulfate extraction, generated
    elemental sulfur extraction, coal drying, recovery of sol-
    vent, water and elemental sulfur by distillation and liming
    out of generated iron sulfate.
                              11

-------
                           4.   REACTOR TEST UNIT

     During the past seven years,  TRW has  been  involved in various stages of
development of the Meyers Coal  Desulfurization  Process.   The early work
involved laboratory and bench  scale testing^1'2'3'4'5' which was aimed at
obtaining process application  information  and process  design data.  Those
early efforts led to the design of the Reactor  Test Unit (RTU).   The design
of the RTU was carried out in  two  steps; first  a preliminary design to
evaluate the overall approach  and  probable cost of the envisioned unit
and secondly, a detailed engineering and design effort which immediately
preceded equipment procurement and RTU construction.   The detailed engineer-
ing, procurement, and construction of the  RTU occurred over a 19-month period.
Immediately following plant construction,  TRW initiated operator training,
safety check-out, procedure verification,  and RTU shake-down in  May 1977.
     This Section presents a detailed description of the RTU as  constructed.
Intended uses and capabilities of  the integrated RTU are discussed as are
those of each major unit operation sub-system in Section 4.1.  Section 4.2
details the specific equipment which is incorporated into the RTU.  Infor-
mation such as equipment type, manufacturer, size, model, configuration,
capacity, operating limits, and materials  of construction is presented.
4.1  PROCESS DESCRIPTION
     The Reactor Test Unit is  a pilot scale coal processing facility designed
to demonstrate those unit operations comprising the front end of the Meyers
Process, namely, coal-reagent  mixing, primary pyrite reaction and reagent
regeneration, secondary (finishing) pyrite reaction and slurry filtration.
Designed for high flexibility,  the RTU has the  capability of processing a
range of suspendable coals up  to approximately  8 mesh top-size and coarse
coals up to approximately 3/8-inch top-size. Spent reagent may be regenerated
either exclusive of, or simultaneously with, coal leaching.  The primary
reactor may be used as either  a five-stage or a three-stage reaction unit to
increase the available range of coal processing times.
                                     12

-------
     The most common modes of RTU operation are represented schematically in
Figure 5.  Fine coal ground to the desired size is loaded into feed tank
T-l.  Dry coal is fed continuously by a live bottom feeder A-2 to the weigh
belt A-3 which discharges through rotary valve A-4 to the three-stage mixer
T-2 (stream 1).  Aqueous iron sulfate leach solution (stream 2) enters T-2
after preheating in a heat exchanger E-2 and passing through the foam scrubber
T-3.  Steam is added (stream 3) to raise the slurry to its boiling point.
Foaming, which may occur during the early stages of mixing, ceases when coal
particle wetting is complete.  The heated slurry (stream 4) is then pumped
to a five-stage pressure vessel R-l in which most of the pyrite is removed.
R-l Slurry heating is achieved by direct injection of steam into any or all
reaction stages.  Reagent regeneration may be carried out simultaneously
with pyrite leaching by means of oxygen injection into any or all reaction
stages  (stream 5).  Unusued oxygen saturated with steam (stream 6) is con-
tacted  in a foam knock out drum V-l with the feed reagent (stream 7) to pre-
heat the reagent and cool vent gases.  Slurry in any stage of R-l may be
cooled  by means of a cooling water heat exchanger E-l which may be applied
to slurry recirculation loops for removal of excess heat of reaction.  Vent
gas from both T-3 and V-l are water scrubbed in T-4 to remove any traces of
acid mist.
     Reacted coal slurry  (stream 8), at elevated temperature and pressure, is
flashed  into a flash drum T-5 for gas-liquid separation.  Generated steam
(stream 9) is condensed in T-4, and the condensate plus any entrained acid
mist is  removed with scrubber water.  Reacted slurry (stream 10) is fed to the
belt filter S-l.  The filtrate, which is regenerated leach solution, is
removed  from the coal slurry through the evacuated filtrate receiver V-2 and
pumped  (stream 12) to the leach solution storage tank T-7.  Coal on the filter
belt is washed with water  (stream 11) and discharged to coal storage.  Wash
water is removed through the evacuated wash water receiver V-3 and pumped
(stream  13) to the liquid waste holding tank T-9 for subsequent disposal.
     As a processing alternate, partially processed slurry from T-5 may be
loaded  into the secondary reactor R-2 for final depyritization in a batch
mode.  Slurry may be retained, heated and agitated, for extended periods of
time in R-2 prior to being pumpted to S-l.

                                      13

-------
J I COAL
\7 MECHANICAL
 v
                                                      ATM OS.
                                                                      WATER
                                                                      RETURN
                 P-12
                                                   P-M
                            Figure 5.   RTU Schematic

-------
     The coarse coal contacting  vessel  T-6  is  a  heated and  insulated  tank  in
which hot reagent may flow through a  bed  of retained  coarse coal.   This  unit
is used principally to  convert regenerated  leach solution  in storage  tanks
T-7 or T-8 to a more depleted solution  simulating recycle  reagent  after
secondary reaction.  This  capability  is required since R-2  is not  utilized
during all experimentation.  T-6 is basically  a  coarse coal  reactor and, if
appropriate sampling ports and possibly some flow distribution internals were
added, it could be used to obtain design  data  for coarse coal processing.
     Detailed descriptions of the principal  RTU  coal  processing operations
are presented in the ensuing sections.
4.1.1  Coal Feed System
     Process feed coal  is  transported from  the Ultrasystem's coal  grinding
facility located in  Irvine, California, to  the Capistrano  Test Site in steel
bins which maintain  a nitrogen atmosphere to prevent  coal weathering.  Coal
bins are hoisted to  the top level of  the  RTU test stand and emptied by a
hydraulic bin tilter-vibrator A-l (Figures  5 anc| 7).   DUSt  control  is main-
tained by coupling the  bin vent  to transport ducting  with  flexible sleeving.
     Coal feed is routed according to the coal top-size.  During coarse  coal
utilization (1/4-inch top  size or larger),  feed  coal  is diverted from the bin
discharge to the T-6 leach tank  where it  is contacted with  reagent to effect
leach solution depletion.   Fine  coal  processing  feed  material is charged to
the coal storage tank T-l  (Figure 8).  A  vibrating discharge unit  A-2 at the
bottom of T-l prevents  solids bridging  or packing and maintains coal  flow to
the weigh belt A-3.  Fine  coal feed is  metered by A-3 to a  moisture lock
rotary valve A-4 which  feeds the coal reagent  mixing  vessel  T-2 (Figure  9).
The weigh belt is instrumented to display coal feed rates  and cumulative
coal flow on the plant  control panel.  All  dry coal storage and transfer
units downstream of  the bin discharge are operated under nitrogen  pressure
to prevent oxidation of the coal feed.
4.1.2  Leach Solution Feed System
     Leach reagents  utilized during RTU operation were generally blended in
the secondary reactor R-2  and pumped  to reagent  storage tanks via  P-7.   Iron
                                       15

-------
Figure 6.   RTU  Bin  Lift




          16

-------
Fiqure 7.   RTU Bin Tilter
             17

-------
Figure 8.   Coal  Feed to Fine Coal Storage
           and the Coarse Coal Leacher

                   18

-------
                                FINE COAL  DISCHARGE A-2
Figure 9.   Coal  Feed To  The Mixing Unit
                     19

-------
sulfate reagent was stored in leach, solution  surge  tank. T~7 while distilled
water or iron free leach reagent was stored in  surge  tank T-8 (Figure 10).
Quantities of reagent in storage were monitored with  buoyant-type liquid
level guages.
     Reagent from either T-7 or T-8, as required, was pumped to the reactor
system by P-13.  Prior to actual coal contacting and  reaction,  reagent from
P-13 was used to break down foam and demist primary reactor vent gases in
knock-out drum V-l.  Reagent and any entrained  mist or foam was brought up
to  the desired temperature for the first  stage  of mixing (typically, 170 F)
by  heat exchanger E-2 and fed to the foam scrubber  T-3 to break down any foam
generated during the mixing operation.  Reagent was gravity fed from T-3 to
the first stage of the T-2 mixer.  Reagent feed rates are monitored by a
magnetic flow meter located between E-2 and T-3.
4.1.3  Oxygen  Feed System
     Oxygen is  stored as a liquid at the  RTU  facility and vaporized as
 required for process reagent regeneration.  The total flow rate of gaseous
 02 (200  psig)  to the primary reactor R-l  was  monitored by target flow meter
 FE-61  and  feed  0,, was subsequently partitioned  among  the reaction stages with
 the aid  of  rotameters FI-62 through FI-66.  Oxygen-slurry contacting is
 achieved by injecting the 02 stream into  slurry recirculation loops associated
 with each  reactor  stage.  The point of 02 injection is located  approximately
 ten feet upstream  from the point of slurry reentry  into the reactor stage.
 Slurry recirculation rates were maintained sufficiently high to ensure
 turbulent  flow  and vigorous blending of reagent and oxygen.
      Unreacted  oxygen flows from R-l through  the knock-out drum V-l and
 pressure relief  valve PV-43 to the vent gas scrubber  T-4.   Vent gas flow
 rates  from  R-l  are measured by target flow meter FE-44 which is located
 between V-l and  PV-43.  The oxygen content of vent  gas is measured by the
 oxygen analyzer AE-171 which analyzes a dried slipstream of the V-l exhaust.
 4.1.4   Process  Steam
      Steam  for process heating was generated  by a Clayton steam generator
 which  had the capability of generating 1000 pounds  of 150 psig  steam per
                                     20

-------
Figure 10.   RTU Tank Farm

-------
hour.  A majority of the process  steam required is used for direct contact
heating of slurry streams and wet coal  cakes  although several heat exchangers
are used in the RTU.  Flow rates  of steam into T-2 and R-l for direct contact
heating are measured with rotameters FI-16 through FI-18 and FI-67 through
FI-71, respectively.  Steam feed  rates are set and maintained manually.  Flow
rates of steam into heat exchangers are regulated by temperature actuated
flow controllers.
4.1.5  Coal-Reagent Mixing
     Initial coal-reagent contacting,  coal wetting and slurry defoaming were
performed in the mixing unit T-2  under atmospheric pressure (Figure 11).  The
mixer is partitioned into three stages by weirs which may be adjusted to vary
stage volumes and, consequently,  slurry retention times.  Each stage is
equipped with a variable speed mixer to promote coal wetting and prevent
settling.  Slurry temperature was increased stagewise from approximately
170°F to the solution normal boiling point (nearly 214°F for iron sulfate
reagents) by direct contact heating.  Shielded thermocouples are situated
in each stage of T-2 to permit continuous monitoring of slurry temperature.
Gases or foam generated in T-2 are treated by the foam scrubber T-3 and
subsequently by the vent gas scrubber T-4.
     Thoroughly wetted slurry was pumped from T-2 to the primary reactor R-l
by slurry pump P-l.  The rate at which slurry was pumped through P-l was
regulated by slurry level controller LIC-26 in the third stage of T-2.
4.1.6  Primary Reaction and Reagent Regeneration
     Simultaneous coal  leaching and reagent regeneration were performed in
the primary reactor at nominal temperatures up to 270°F and pressures up
to 100 psig.  Design capabilities of R-l are somewhat higher with a maximum
temperature and pressure of 300°F and 150 psig, respectively.  The R-l
reactor consists of five stages separated by weirs.  All five stages may
be used for reaction or, alternately,  the P-l slurry feed may be diverted
around the first two stages of R-l to effectively yield a three-stage reactor.
Each reaction stage is  a continuous flow stirred tank reactor in which
agitation is achieved by means of a variable speed mixer and a slurry
recirculator.  The slurry recirculation loops are also used for steam

                                    22

-------
ro
to
                                                                               PRIMARY  REACTOR R-l
                                    Figure  11.   RTU Mixer and  Primary Reactor

-------
injection to maintain reaction temperature and oxygen injection  for  reagent
regeneration.
     Reaction stages of R-l have interconnected overhead ullage  and,  therefore,
all stages are operated at the same pressure.  R-l vents to the  knock-out
drum V-l which demists vent gases and breaks down any foam which may  be
generated during processing.  Vent gases leaving V-l are further processed
in the vent gas scrubber T-4.
     Processed slurry Is flashed from the fifth stage of R-l into
the slurry flash drum T-5 which is maintained at atmospheric pressure.  Vent
gas and flashed steam from T-5 are processed through T-4.  Underflow  slurry
from T-5 is gravity fed to either the secondary reactor for further processing
or the filtration unit S-l.  Slurry is flashed from R-l in pulses having
durations of up to three seconds with approximately 25 seconds between pulses.
4.1.7  Secondary Reaction
     Coals requiring further processing after primary reaction in R-l are
gravity fed from the flash drum T-5 to the secondary reactor R-2 for  batch
processing.  Reaction in R-2 may be carried out at temperatures up to 250°F
in the absence of oxygen.  Alternately, at temperatures below 250°F,  oxygen
partial pressures may be maintained such that the system,total pressure does
not exceed 14 psig.  Reaction temperatures in R-2 are maintained by electrical
heating tapes at the reactor surface to enable prolonged processing without
the slurry dilution inherent to direct steam heating.  Slurry agitation is
provided by a single variable speed mixer.  R-2 vent gases are processed
through the vent gas scrubber T-4.  Fully processed slurry may be pumped from
R-2 to either the filtration unit S-l or to waste disposal T-9.  When not in
service as a reaction unit, R-2 may be utilized for reagent formulation and
mixing.
4-1.8  Filtration and Leach Solution Recovery
     Processed slurries from the primary reactor R-l (via flash drum  T-5)
or from the secondary reactor R-2 were fed to the filtration unit S-l for
reagent recovery and coal washing (Figure 12).  S-l is a belt-type filter
with two filtration zones having discrete filtrate collection systems.  The
first filtration zone is for recovery of concentrated reagent which is

                                     24

-------
ro
en
                                Figure  12.   Primary Reactor and Belt Filter

-------
collected in a filtrate receiver V-2 (Figure 13)  and recycled back to the
appropriate leach solution surge tank T-7 or T-8.   The second filtration
zone is for recovery of hot wash water which is  sprayed on the coal filter
cake to effect further reagent recovery.   Filtrate from coal cake washing
is collected in the wash water receiver V-3 and  pumped to disposal tank T-9.
Vacuum for filtration was provided by vacuum pump  K-l which couples to the
filtrate receivers V-2 and V-3.   Steam is applied  to the coal cake during
filtration and washing to prevent cracking of the  filter cake and maintain
cake temperatures of approximately 180 F.
     The spray-washed coal cake  was scraped from  the filter belt and loaded
into storage containers or disposal dumpsters.   Residual  solids adhering to
the filter belt were removed with water sprays directed at the inner and
outer belt surfaces.  The low solids slurry resulting from belt washing may
be recycled for cake wash to minimize losses of process solids or disposed
of directly.
     Those process products not  required  for process characterization or
vendor testing were disposed of  at class  "A" landfills.  Low solids content
liquid wastes from T-9 were disposed of by Industrial  Trucking Co., Wilmington,
California.  Discard coal cake was disposed of by  Removal  Inc., Lawndale,
California.
4.1.9  Coarse Coal  Leaching and  Reagent Depletion
     Depleted leach solution for recycle  reagent  simulation is generated in
the coarse coal leach tank T-6.   Coarse coal is loaded into T-6 from the bin
tilter A-l and may be preheated  with steam which  is  introduced at the reactor
bottom.  Hot leach solution from heat exchanger E-2  contacts the coarse coal
bed at the bottom of T-6, flows  upward through the bed and is pumped to
reagent storage in T-7 or T-8.   Alternately, reagent may be retained in T-6
for batch leaching.  Spent reagent may be drained  from T-6 through a screened
drain port at the vessel  bottom.   Near complete  recovery of spent reagent is
effected by washing with water which flows upward  through the coal bed.
Processed coal  is discharged from the bottom of  T-6  into a dumpster for
sampling and/or disposal.
                                    26

-------
r-o
                                                                                         FILTRATE
                                                                                       RECEIVER V-2
                                     Figure  13.   Vacuum Pump and  Filtrate Receivers

-------
4.1.10  Instrumentation
     RTU instrumentation may be classified according to relative importance
of instrument function during plant operation.   Primary instrumentation
requires continuous monitoring for maintenance  of plant operational safety
and/or acquisition of data required for process evaluation.  Data from primary
instrumentation are displayed at the plant control  panel either in digital,
gauge or printed tape form, and are also recorded on magnetic tapes for
subsequent evaluation (Figure 14).  Additionally, primary data are generally
displayed on the test stand to facilitate plant operation and local monitoring.
Examples of primary data sources include the weigh  belt coal  feed rates,
magnetic flow meters measuring reagent and slurry feed rates, thermocouples
monitoring reaction temperature, transducers monitoring reactor pressures
and level gauges indicating slurry depths in the mixer and reactors.
     Secondary data sources are those providing information which is  not
critical to safe plant operation or process evaluation.  These data sources
are not monitored continuously with some being  checked on a daily or weekly
basis.  Data acquisition from secondary sources usually consists of manual
recording and tabulation.   Examples of secondary data sources include flow
meters indicating seal flush water flow rates,  steam flow meters and  reagent
storage tank level indicators.
4.1.11  Process Sampling and Analysis
     All major RTU processed solid, liquid, slurry  and gas streams were sampled
and analyzed during plant operation.  Analyses  of solid and liquid streams
were performed at the RTU control  laboratory.   Additionally,  selected coal
samples were analyzed by Warner Laboratories, Cresson, Pennsylvania.   Gas
sampling and analysis was performed on a continuous basis by RTU equipment.
     Sampling ports available for acquisition of feed reagent samples are
located at the leach solution feed pump P-13 and at the foam scrubber T-3.
Feed solids are sampled with a full stream coal diverter located at a point
between the weigh belt A-3 and rotary valve A-4.  Stream diversions of up to
one minute were used during normal sampling although longer periods of time
were used during A-3 calibration.   Slurry samples were drawn from the mixer
T-2 through sampling ports located in the wall  of each mixing stage.   Primary

                                    28

-------
UD
                                 Figure  14.   RTU  Control/Instrumentation  Console

-------
reactor R-l slurry samples may be drawn from each stage through wall sampling
ports or by full stream diversion of slurry passing through recirculation
loops.  Processed slurry sampling ports are located in the slurry flash drum
T-5 and in the secondary reactor R-2.   Processed coal  samples may be taken
directly from the filter belt S-l after the spray wash operation.
     Feed coal samples were submitted for short proximate analysis (moisture,
ash, heat content and total sulfur analyses), sulfur forms analysis (pyritic
and sulfate sulfur analysis), ash iron content analyses and size distribution
determination.  Reagent samples, whether feed reagent or slurry sample fil-
trates, were analyzed for total  iron,  iron forms, sulfate content and pH.
Slurry samples were analyzed for solids content and slurry solids were water
washed, extracted with toluene and dried.  Dry slurry solids received short
proximate, sulfur forms and ash  iron analyses.   Residues from toluene extracts
were analyzed for total sulfur content.
     Vent gas from R-l was continuously sampled by means of a slipstream
taken immediately downstream from V-l.  The sample slipstream was dried
and subsequently analyzed for oxygen content by a Taylor oxygen analyzer.
Gas analysis results were continuously displayed at the plant control  panel.
4.2  PROCESS EQUIPMENT
     Detailed mechanical  design  specifications, drawings, and vendor prints
for all RTU equipment have previously been submitted to the EPA (TRW-Procon
Design Drawings, Vendor Prints,  and Design Specifications).  These documents
contain plant plot plans, mechanical flow diagrams, civil and structural
drawings, mechanical, electrical and instrument drawings, piping diagrams
and vendor data.  Detailed specifications for all equipment, piping, elec-
trical, and instrumentation installations are also included.  Descriptions
of services provided on a regular basis in support of RTU operation as well
as stepwise procedures have been documented.  This extensive documentation
is not replicated in this report but is mentioned for reference only.  How-
ever, specifications of principal coal processing equipment utilized during
RTU operation are summarized in  this section to provide definition of the
RTU processing capabilities.  The ensuing sections will also describe the
various support facilities utilized during RTU operation.

                                    30

-------
4.2.1  Coal Transport Bins
     Seventy-five coal transport bins were  fabricated  of carbon  steel.   Bin
dimensions are 42" x 48" x 96"  (H) which  provide  approximately  112  cubic feet
of storage space.  Coal bins are designed to  receive coal  ranging from
3/8-inch top-size to 100 mesh top-size.   Each bin is equipped with  a  12-inch
diameter opening located at the top  for coal  loading and a 34-1/4-inch  x
14-11/16-inch (H) trap door on  the side of  the bin  near  the bottom  for  coal
discharge.  Coal bins are airtight and equipped with gas inlet and  outlet
fittings to enable slight pressurization  with nitrogen.
4.2.2  Bin Tilter
     The bin tilter A-l is a steel frame  capable  of handling a fully  loaded
bin and equipped with a pneumatically operated tilting mechanism.   A-l  is
equipped with leg clamps which  automatically  seal the  bin  to the feed hopper
gasket and prevent dust emission to  the atmosphere  during  the coal  discharge
operation.  The A-l vibrator maintains coal solids  flow  once the bin  discharge
door is opened.
4.2.3  Coal Storage Tank
     Coal is stored in a vertical carbon  steel  tank T-l  which is a  6-foot diam-
eter cylindrical vessel with a  conical bottom to  facilitate  coal discharge.
T-l has a total volume of approximately 180 cubic feet.  The T-l vessel receives
coal from the coal feed chute through a 1-foot square  opening in its top.  T-l
is equipped with a live bin bottom capable of maintaining  variable  solids dis-
charge flow rates without bridging or packing.  The T-l  discharge unit A-l
is equipped with a two-point cycle timer  to regulate the discharge  oscillator.
A-l is capable of maintaining coal discharge  rates  from  250  to 1000 pounds per
hour.  T-l is maintained under a slight nitrogen  pressure  (up to 6  inches
water column) to prevent coal oxidation.  Maintenance  of an  adequate quantity
of coal in T-l is ensured by monitoring the coal  level with  a sonic level
switch which activates a low level alarm.
4.2.4  Weigh Belt
     An Autoweigh weigh belt A-3, capable of  metering  coal flow rates up to
1000 pounds per hour, transfers coal fed  from the coal storage tank T-l to
the rotary feed valve A-4.   A-3 is designed to  transfer  ground coal from

                                     31

-------
8 mesh to 100 mesh in top-size.   A belt speed monitor and strain-gauge load
cell are used to provide coal  feed rate and total  coal  feed outputs, and to
signal a feedback controller which maintains preset coal  feed rates by adjust-
ing the belt speed.  Coal feed rate monitoring accuracy is within 0.5% of the
actual coal feed rate.   The A-4  feed valve is airtight  and acts as a moisture
barrier between A-3 and the first stage of the slurry mixing unit T-2.  Both
A-3 and A-4 are operated under a nitrogen atmosphere.  A full stream butterfly
valve coal diverter is  located between A-3 and A-4 which enables acquisition
of feed coal samples.
4.2.5  Tank Farm
     Leach solution and process  liquid wastes are  stored in a tank farm con-
sisting of three glass-fiber reinforced polyester  tanks T-7 through T-9.
Each tank is 12 feet in diameter, 20 feet high and has  a storage capacity of
16,000 gallons.  These  tanks are designed for atmospheric pressure service at
up to 200°F.  Stored solution volumes are measured by Varec liquid level  gauges
and are accurate to approximately 5 gallons.  Tank farm plumbing is designed
to enable solution transfer among the storage tanks.
4.2.6  Leach Solution/Slurry Pumps
     Sixteen process pumps are included in the RTU for  either liquid (water
or reagent) or slurry service.  These pumps are largely centrifugal and pro-
gressive cavity-type pumps although diaphragm and  regenerative pumps are  also
used.   Pumps used in acid reagent service were constructed of 316L S.S. while
those in water service  were fabricated of cast iron or  other less noble mate-
rials.  Specific data for each pump are listed in  Table 1.
     Centrifugal pumps  were utilized for low head  pressure and moderate or
high flow rate applications such as R-l slurry recirculation and liquid waste
disposal.  Nine centrifugal pumps were utilized in the  RTU with eight being  .
used for slurry/reagent pumping  service (P-2 through P-6, P-9, P-12 and P-14)
and one being used in very dilute reagent service  (P-10).  Progressive cavity
pumps were utilized for pumping  slurries at low flow rates against large head
pressures as encountered in feeding slurry to the  pressurized primary reactor
through P-l.  To provide possible replacement units for P-l, two progressive
cavity pumps were also  used in less severe applications (P-7 and P-13).  Two
                                     32

-------
                                               TABLE  1.   RTU  PUMP  DATA
co
Pump
Identification
P-l
P-2 through P-6
P-7
P-8
P-9 and P-10
P-ll
P-12
P-13
P-14
P-15
P-16
Description
Slurry feed pump
Slurry recirculation
R-2 discharge pump
Filter wash pump
S-l filtrate pumps
Reagent recirculation pump
Reagent circulation pump
Reagent feed pump
Waste disposal pump
Cooling water pump
Seal flush water pump
Manufacturer
Moyno
Dean Brothers
Moyno
Warren Rupp
Duriron
Warren Rupp
Dean Brothers
Moyno
Dean Brothers
Paco
Aurora
Type
Progressive cavity
Centrifugal
Progressive cavity
Diaphragm
Centrifugal
Diaphragm
Centrifugal
Progressive cavity
Centrifugal
Turbine
Regenerative turbine
Design
Capacity,
gpm
1.1-4.6
25
1.1-4.6
10
20
5
20
0-8.4
200
100
12
AP,
psi
140
33
140
40
40
18
18
140
32
250
150
Driver,
H.P.
3
2
3
Air drive
5
Air drive
2
1.5
7.5
15
5
Material of
Construction
316 SS
316 SS
316 SS
316 SS
Durcomet 100
316 SS
316 SS
316 SS
316 SS
Cast iron
C.I. /Bronze

-------
diaphragm pumps were utilized in the RTU for low head pressure, low flow rate
applications (P-8 and P-ll).   One regenerative turbine pump (P-16) was utilized
for a seal flush water pump due to its high head pressure, low flow rate char-
acteristics and one turbine pump (P-15) was used for cooling water circulation.
4.2.7  Heat Exchangers
     Two process streams are heated by jacketed pipe heat exchangers:  1) feed
reagent to T-3 is heated in E-2 and 2) filter belt wash water is heated in
E-3.  The heating fluid for both exchangers is 150 pound steam and each is
designed for a maximum temperature and pressure of 400 F and 200 psi, respec-
tively.  Exchanger E-2 consists of two heat exchangers in series, each of which
is 30 inches long having a 3/4-inch Sch. 40 pipe of 316 S.S. and a 1.5-inch
Sch. 80 jacket of carbon steel.  Exchanger E-3 consists of three heat exchangers
in series, each of which is 45 inches long having a 3/4-inch Sch. 80 pipe and
a 1.5-inch Sch. 80 jacket.  E-3 is all carbon steel construction since it is
not subjected to a corrosive environment.
     The primary reactor R-l is equipped with jacketed pipe heat exchanger
E-l to provide for removal of excess heat of reaction.  E-l consists of four
heat exchangers in series which may be divided among the R-l slurry recircula-
tion loops.  Each segment of E-l is 3.5 feet in length having 1.25-inch
Sch. 40 pipe of 316 S.S. and 2-inch standard weight carbon steel jacketing.
E-l is cooled with site cooling water and is designed for pipe temperatures
and pressures up to 300°F and 150 psi.
     A shell and tube heat exchanger E-4 is used to cool flush water for pump
and mixer seals.  This unit is cooled with site cooling water.  Shell material
is bronze and tubing is admiralty brass.  E-4 is designed for temperatures
and pressures of up to 300°F and 150 psi.
4.2.8  Slurry Mixing Tank
     Coal-reagent mixing and preheating takes place in the mixing tank T-2.
The mixer is a three-stage, 316L S.S. horizontal vessel having a diameter of
2.5 feet and length of 7.5 feet.  The stages are separated by adjustable weirs
with which the stage volumes and, therefore, slurry residence times may be
varied.  Weir heights may be adjusted from approximately 10 to 23 inches.
Each stage of T-2 is equipped with a variable speed 3/4 horsepower Lightnin

                                     34

-------
agitator capable of up to 350 rpm.   Live  steam for slurry heating  is  injected
into each stage of T-2 through nozzles  in the vessel  wall.   All  stages of T-2
vent through a common foam scrubber  T-3 and  through a vent gas  scrubber T-4.
     Each stage of the mixer is  equipped  with a slurry sampling port, drain
port and a thermocouple probe.   The  third stage of T-2 contains an Air Products
bubble tube-type level indicator and an auxiliary stilling well.
4.2.9  Reactors
     Three reactors are utilized in  the RTU:   (1) the primary fine coal reactor
R-l; (2) the secondary fine coal  reactor  R-2; and (3) the coarse coal leacher
T-6.  RTU reactors are constructed of 316L S.S. and are rated for  processing
temperatures of up to 300°F.  All RTU reactors are insulated with  1.5 inches
of urethane foam.
     The primary fine coal reactor is a horizontal pressure vessel  with a
diameter of 38 inches and a length of 14.75  feet.  R-l is designed for pres-
sures up to 150 psig and for full vacuum  (to accommodate hot lockup and sub-
sequent depressurization through cooling).  Composed of five stages,  R-l is
partitioned by stationary weirs  which are 28.5 inches in height.   Slurry feed
lines are plumbed to allow bypass of the  first two R-l stages and  provide for
three stage processing.  Each stage  of R-l is agitated by a variable  speed
1.5  horsepower Lightnin mixer and a  slurry recirculation system.   Four
baffles are spaced around each  stage to promote efficient slurry mixing.
Slurry level in the fifth stage  is measured  by an Ashcroft buoyant type level
gauge.  R-l slurry is discharged from the fifth stage to flash  drum T-5 either
by flashing through a 1-inch ball valve KV-241 or by pumping slurry through
the  fifth-stage recirculation pump P-6.
     Reactor heating and slurry  oxygenation  are performed by injecting steam
and  oxygen into the slurry recirculation  loop of each R-l reaction stage.
Reactor temperatures are monitored by thermocouple probes installed in each
R-l  stage.  R-l stages vent to  a common foam knock-out drum V-l and vent gas
scrubber T-4.  Reactor pressure  is maintained by a Masoneilan control valve
and  is monitored by a transducer PT-173 located at the reactor  vent.
  •   R-l "is equipped with 1-inch sampling ports in the vessel wall  and in the
slurry recirculation loops.  Recirculation loop sampling ports  enable full
stream bypass through a sampling unit which  is ultimately isolated and drained.
                                      35

-------
     The secondary fine coal  reactor is a vertically oriented cylindrical
vessel with a conical bottom.  R-2 has a diameter of 3.5 feet and has a height
of 4.1 feet (T/T); the corresponding volume of R-2 is 365 gallons.  R-2 is
designed to maintain 14 psig or full vacuum.  Reactor pressures are monitored
by an Ashcroft pressure gauge PI-98.  R-2 agitation is provided by a variable
speed  1-1/2  horsepower  Lightnin mixer.   Three baffles are distributed within
R-2 to prevent channeled flow and promote efficient mixing.
     Slurry enters R-2 through a 2-inch nozzle at the reactor top and dis-
charges through a 1.5-inch nozzle at the reactor bottom.  Slurry levels in
R-2 are monitored with an Air Products bubble tube level sensor LT-95.  Slurry
temperatures are monitored with a thermocouple probe TE-99.
     The coarse coal leach solution depleter T-6 is a vertically oriented
cylindrical vessel with a conical  bottom.  The vessel diameter is 5 feet and
the straight side height is 5 feet.  T-6 is essentially an atmospheric pres-
sure vessel.  Coal enters T-6 through a 12-inch square chute at the vessel top
and is discharged through a 12-inch knife valve at the bottom.  Reagent enters
and discharges from T-6 through 3/4-inch ports at the bottom and top of the
vessel, respectively.  T-6 is heated by live steam injected through a 1-inch
port at the vessel bottom.  Reaction temperatures are monitored with a thermo-
couple probe in the vessel wall.   Reagent may be drained from T-6 through a
3/4-inch drain port equipped with a 2-inch duplex strainer for retaining coal.
Reactor pressure is monitored with an Ashcroft mechanical pressure gauge PI-260.
4.2.10  Flash Drum
     The slurry flash drum T-5 is a vertical vessel with a conical bottom.
T-5 has a 3-foot diameter and a height of 3.7 feet (T/T).  T-5 is designed for
up to 14 psig pressure and full vacuum.  Flash slurry enters T-5 through a
2-inch nozzle located 1.7 feet below the upper straight side end.  A porous
demister pad is situated immediately above the flash slurry inlet to minimize
                                                                            •;
slurry entrainment.  Slurry level  is monitored in T-5 and an alarm condition
results if slurry levels rise appreciably above the conical vessel bottom.
4.2.11  Scrubbers
     Slurry foam control and acid mist elimination are obtained in the RTU
with three scrubbers:  (1) the foam scrubber T-3; (2) the knock-out drum V-4

                                     36

-------
and (3)  the vent gas scrubber T-4.  T-3  is  a  316L S.S.  vessel  of 1-foot diam-
eter and 4 feet in height.  A single  bubble cap tray is used  to  break  down
foam generated in the mixer T-2.   Designed  for gas flow rates of 30  SCFM and
liquid flow rates of 2.5 gpm, T-3  may be operated under full  vacuum  or at
pressures up to 14 psig with temperatures up  to 300°F.   V-l  is a 316L  S.S.
vessel having a 1-foot diameter  and a height  of 6.5 feet.  Two bubble  cap
trays are used to break down foam  and remove  acid mist  from  the  R-l  effluent
gas.  V-l has the same design pressure and  temperature  requirements  as R-l,
namely, full vacuum to 150 psig  at temperatures up to 300°F.   T-4 is a low
pressure FRP vessel having a diameter of 1.5  feet and a height of 4.5  feet.
Vent gases flow upward through a 3.5  cubic  foot volume  of  316 S.S. flex ring
packing which is sprayed with cooling water from an overhead  nozzle.   Operating
temperatures in T-4 may range up to 215°F.
4.2.12  Filter and Vacuum System
     Solid-liquid separation is  performed with a belt filter supplied  by
Ametec.  The filter belt consists  of  a polypropylene mesh  cloth  1 foot in
width which is supported and conveyed by a  channeled plastic  belt.   The filter
is  driven by a two horsepower variable speed  belt drive.   Filter belt  speeds
may be  varied from 1 to 10 feet  per minute  producing coal  cake thicknesses of
up  to 3 inches.  Suction is applied to the  belt by means of  two  vacuum pans
over which the belt must pass sequentially.  The first  vacuum pan is for col-
lection of concentrated reagent  while the second collects  wash water from the
filter  cake rinse.  Feed slurry  is distributed on the belt by a  stationary
spreader.  After initial dewatering,  the filter cake is steamed  and  sprayed
with hot wash water by overhead  nozzles.  The filtration unit is completely
enclosed in a reinforced fiberglass hood to prevent steam  losses. Sampling
ports are built into the hood to permit  processed coal  sampling  before and
after the washing operation.  Filter  belt washing nozzles  are located  at the
belt return immediately beyond the processed  coal discharge  chute to remove
remaining coal particles from the  belt and prevent blinding.
      Vacuum is supplied to the filter vacuum  pans through  the respective filtrate
receiving vessel (V-2) and wash  water receiving vessel  (V-3).  The filtrate
receiver is a 316L S.S. vessel with a 1.2-foot diameter and  a height of 5 feet.
The wash water receiver is a 316L  S.S. vessel with a diameter of 2 feet and a
height  of 6 feet.  Both receiving  vessel^ are under vacuum provided  by a
                                       37

-------
50 horsepower 630 ACFM Nash pump K-l which is capable of maintaining vacuum
of 22 inches Hg.  Each receiving vessel is equipped with a demister pad to
prevent liquid entrainment into K-l.  The contents of V-2 are recycled to
reagent storage tanks T-7 or T-8 while the V-3 underflow is pumped to waste
disposal.
4.2.13  02/N2/Air Supply Systems
     Gaseous oxygen is supplied to the primary reactor R-l from a liquid oxy-
gen  (LOX) storage tank which is remote from the RTU coal processing facility.
The  LOX is vaporized and supplied to the RTU at 200 psig.  Oxygen feed rates
are  monitored by a Ramapo target flow meter to ±0.5 percent.  Process nitrogen
is stored in liquefied form and vaporized for plant use.  Nitrogen flow moni-
toring equipment is identical to that used for oxygen monitoring.  Nitrogen is
supplied to the plant at 200 psig.  Air is supplied to the facility at 125 psig
by a 100 SCFM Rand air compressor and is used primarily for equipment drive and
valve actuation purposes.
4.2.14   Steam Supply System
     Process steam is supplied by a Clayton steam generator with the capability
of generating 1000 pounds of 150 psig steam per hour.  The steam generator is
equipped with self-contained water softener and boiler treatment subsystems.
The  generator is powered by a propane burner also supplied by Clayton Manu-
facturing Co.  Steam feed rates to the RTU are monitored by Brooks rotameters.
4.2.15   Process Water System
     Two water sources were utilized during RTU operation.  The first, an
existing well water supply, is used to supply the steam generator and to main-
tain the 100-gallon seal flush water supply tank.  The second source of proc-
ess  water is the RTU catch basin which is a large hypalon lined reservoir hav-
ing  a width of 30 feet, a length of 80 feet and a depth of 5 feet.  Water from
the  catch basin is used for cooling water in heat exchangers, as scrubber
water in T-4 and for filter belt wash water.
4.2.16   Instrumentation
     Detailed instrumentation listings and specifications are presented in the
"TRW-Procon Design Specification" manual.  This document contains instrumenta-
tion tag numbers, manufacturers, materials of construction, operating ranges
                                       38

-------
and, in many cases, unit schematics.   Locations  of the  various  instruments in
the RTU are fully detailed in the  "TRW-Procon  Design  Drawings and  Vendor
Prints" manual, drawing numbers X418455  through  X418461.   Both  TRW-Procon
documents have been previously transmitted  to  EPA.
     Principal types of instrumentation  located  on  the  RTU include pressure
gauges, temperature indicators, level  gauges,  and flow  meters.  Ashcroft
bellows type pressure gauges were  used to monitor nitrogen pressures to coal
bin storage and instrument purge streams (i.e.,  PI-184  and 167).   Ashcroft
bourdon gauges were utilized in monitoring  pump  slurry  and liquid  discharge
pressures  (i.e., PI-73 through PI-77).   Remote monitoring  of feed  gas pres-
sures and  reactor pressures was performed with Viatran  transducers.  Trans-
ducers were also used in conjunction  with digital  display  equipment to provide
local data readout.
     Three principal types of temperature measurement equipment were utilized
during RTU operation:  (1) thermocouples,  (2)  dial  thermometers and (3) bulb
thermometers.  Shielded type K (chromel-alumel)  thermocouples provided by
Heat Technology were used to monitor  mixer  and reactor  slurry temperatures
(TE-19, 20, 21 and TE-52 through 56).  Bi-metal  dial  thermometers  by Weston/
Disco were used for local display  purposes  such  as  monitoring wash water
temperatures and reagent storage temperatures.  Taylor  bulb thermometers in
thermowells were utilized in conjunction with  temperature  controllers such as
that controlling steam feed rates  to  heat exchanger E-2 (TIC-32).
     A variety of level gauges were employed to  monitor liquid, slurry and
coal depths in the RTU tanks and reactors.   Foam levels in the  mixer could be
determined by Jacoby-Tarbox sight  glasses  installed in  the T-2  vent lines.  A
Masoneilan buoyant level gauge was utilized to monitor  slurry levels in the
fifth stage of the pressurized primary reactor while  differential  (bubble tube)
level indicators were used in the  mixer  and secondary reactor.  Varec float-
type indicators were utilized for  monitoring reagent  and liquid waste levels
in  tank farm units.  Sonic level probes  by  National Sonics were utilized to
activate high and low level alarms in critical process  vessels  such as the
coal feed  tank and the primary reactor.
     Steam flow rates were monitored  with  Brooks rotameters.  Water flow rates
were monitored with both Brooks and Wallace and  Tiernan rotameters.

                                      39

-------
Additionally, Wallace and Tiernan rotameters were used in some gas flow  appli-
cations.  Fischer and Porter magnetic flow meters were used for all  slurry  flow
rate measurements (FE-29, 83-87) as well as for reagent feed rate measurements
(FE-31).  Oxygen feed and vent rates were monitored with Ramapo target flow
meters  FE-44 and FE-61.
     Outputs of all primary instrumentation and alarm instrumentation were
translated into engineering units and recorded by a Doric data processor.
The data processor is of modular construction with a solid-state channel
scanner, digitizer, microprocessor-based control circuits, channel display  and
printer.  A total of 99 channels for data storage are available and one addi-
tional  channel serves a self-check function.  The system contains a real-time
digital clock providing time data printout in days, hours, minutes and seconds.
Channel scan rates may be varied from 2 to 20 channels per second.  The sys-
tem is  capable of continuous scanning and logging or periodic logging at inter-
vals of 1, 5, 10, 15, 30 and 60 minutes.  Alternately, the system may scan
continuously but output only those channels which are beyond preset tolerance
limits.  When scanning at 2 channels per second, an integral digital  strip
printer may be utilized to provide hard copy printout; all data are recorded
on seven-track magnetic tape, regardless of scan speed or frequency.
4.2.17  Control Laboratory
     Sampling and analyses of all priocess solid and liquid streams were per-
formed by personnel from the control, laboratory at the Capistrano test site.
The analytical capabilities of the 760 square foot control laboratory include
coal moisture analyses, total sulfur analyses, sulfur forms analyses (pyritic
and sulfate sulfur), slurry solids determination, reagent ferrous iron and
total  iron analyses, reagent sulfate, reagent pH and coal  size distribution
analyses.   Selected coal samples were submitted to Warner Laboratories,
Cresson, Pennsylvania, for coal  short proximate analyses (moisture, ash, heat
content and total sulfur analyses).  Toluene extract residues were also sub-
mitted to  Warner for total sulfur analyses.  All analyses are performed in
accordance with standard ASTM methods.
     To ensure uniform sampling, sample preparation and sample analyses,
standard procedures were prepared and documented.  The procedural format is
                                     40

-------
essentially the same for sampling  and  analyses and includes the  following
elements:
        •   Procedural  scope  describing the operation or operations  to
            be defined.
        •   Procedural  conditions  specifying personnel and safety require-
            ments  and  listing all  relevant reference documentation.
        •   Test requirements including equipment requirements and appa-
            ratus  configuration specifications.
        •   Detailed analytical and sampling procedures and operational
            flow diagrams.
         •   Data recording format.
 A tabulation  of the standard  analytical and sampling procedures  is presented
 in Volume  II, Appendix B.
 4.2.18  Capistrano Test Site  Support Facilities
      In  addition to the control laboratory which was utilized on a continuous
 basis, the Capistrano  test site also provided a variety of services  on  an
 as-needed  basis  in support of the RTU.  The CIS weld, machine, and valve shops
 were all utilized  periodically for quick in situ and remote repairs, uncoded
 fabrications  and equipment maintenance.  All RTU instrumentation was cali-
 brated upon receipt by TRW metrology which also provided trouble shooting
 assistance during  plant shakedown and operation.  CTS procurement, shipping
 and receiving, and quality assurance assisted RTU operation by locating
 replacement parts  as needed,  identifying specialty fabrication shops and pro-
 curing all required chemicals.  Site power, water and sanitation facilities
 utilized during RTU operation were maintained by the CTS maintenance
 department.
                                       41

-------
                       5.  REACTOR TEST UNIT OPERATION

     On May 1, 1977, RTU shakedown activities were initiated.  The overall
shakedown effort was comprised of a system readiness and safety review, oper-
ator training, operating procedure verification, laboratory procedure verifi-
cation, data acquisition system verification, RTU operability evaluation, and
Operational Test Plan revision.  By late September, 1977, all objectives of
the RTU shakedown were met and the RTU was deemed ready to start the planned
operational phase of the project.
     Operational testing of the RTU was started on October 1, 1977, and was
terminated on January 26, 1978.  During that 4-month period, the unit processed
49,700 Ibs of coal over approximately 250 hours of operation in increments up
to 32 hours each.  Sufficient data was acquired during operational periods to
allow verification of process chemistry (discussed in Section 6), to paramet-
rically evaluate process variables (Section 6), to determine equipment char-
acteristics, and to generate coal quantities sufficient for vendor testing
(Section 9).
                   i
     The remainder of Section 5 presents a summary of achievements and obser-
vations relating to RTU shakedown and operation.  Sections 5.1 and 5.2 discuss
shakedown and operation activities respectively.  Additionally, Section 5.2
presents a detailed description of typical RTU start-up, operations, and shut-
down sequences.  Equipment operating experience is also delineated.  Conclusions
which were drawn from the first four months of RTU operational experience are
presented in Section 5.3.
5.1  RTU SHAKEDOWN ACTIVITIES
     The primary objectives of the RTU shakedown activities were to verify unit
operability, train and familiarize personnel with process operations, facilities,
and procedures, and acquire preliminary system data.  Necessary modifications
to the operational, smapling, analysis and calibration procedures and to the
reactor system itself were made during the shakedown period.
                                     42

-------
5.1.1  RTU Safety Review and Training
     A thorough pre-start-up safety review of the dual  reactor test  system was
initiated following completion of  plant  construction.   The Safety  Review Team
(SRT) was comprised of TRW personnel representing the  TRW Health and  Safety,
Structural Design and Fabrication, and Chemical  Engineering Departments,
Capistrano Test Site management and other upper  management representatives from
within TRW operations.  Most SRT members were individuals not previously inti-
mately involved in the RTU design  and operational decisions, thus  limiting
possible bias.  The review was conducted through study of prepared materials
(memos, reports, manuals, etc.) and visual site  inspections.
     The SRT reviewed all start-up, shutdown and operational procedures and
evaluated the unit operations from chemical (corrosion, flammability,  acidity,
etc.) and toxicological standpoints.  On-site safety measures such as  fire
extinguishers, breathing equipment, protective clothing, etc., were  reviewed
for  adequacy.  Additionally, all emergency procedures, process safety  control,
emergency bypass controls ar.d monitoring equipment (both for personnel and
environmental safety) were evaluated.  At the culmination of the safety review,
the  team prepared a report detailing all findings and  recommendations.  The
safety review report was evaluated by RTU project management and all  necessary
changes to plant operation, facilities,  and procedures were made.
      In parallel with the safety review, personnel training was conducted.
This pre-start-up training effort  had as its primary goal the safe and orderly
operation of the RTU during start-up, run and shutdown sequences.  To  facili-
tate the achievement of that goal, all  operating, laboratory and engineering
personnel were familiarized with the appropriate procedures, both  normal and
emergency.  The training consisted of both operating procedure study and
on-site experience with the processing  equipment and area to familiarize per-
sonnel with types and locations of processing equipment, valves, service facil-
ities, instrumentation and control hardware, effluent  monitoring equipment and
emergency control points.
5.1.2  Shakedown Operations
                                                                  (9)
     A detailed shakedown plan was submitted to  EPA in June 1977,      In
general, that document described in detail the planned objectives, approach,

                                      43

-------
sequences of operations, and expected results of the plant shakedown effort.
The broad objectives of the RTU equipment shakedown period were as follows:
        •   Demonstration of equipment operability using coal and iron
            sulfate leach solution.
        t   Demonstration of the operability of the RTU at steady-state
            condition.
        •   Demonstration that controls, indicators and data acquisition
            systems were functional  and suitable for repeated start-up,
            operation and shutdown sequences.
        t   Verification that sampling and analysis procedures were
            reproducible, safe and effective for the locations proposed,
            and that acceptable time delays between sampling and final
            result determinations were obtained,
        •   Estimation of the reproducibility of a single set of process
            conditions and the associated product characteristics.
        t   Standardization of settings of instruments, control devices
            and equipment to obtain desired ranges of pressure, tempera-
            ture, volume, flowrate,  weight, agitation rate, transfer
            rate, etc.
        t   Determination of the practical operational equipment limits
            for preparation and feeding of coal slurry, slurry flow and
            mixing and filtration operations.
        •   Demonstration that all RTU mechanical components and by-passes
            operated without unscheduled interruption or mechanical fail-
            ure over the proposed operational ranges of flowrate, pressure,
            temperature and chemical environment.
     The general approach to the shakedown of the RTU consisted of the sequen-
tial imposition of increasing numbers of stress-producing environments on the
unit operations of the process.  In this manner, identification and correction
of any problems were expected.  Also the gradual approach enabled personnel
development of familiarity with process controls, checkout procedures, etc.,
in an easily assimilated manner.  Each set of stress-producing environments
was studied under continuous steady-state operation of the RTU over a period
of time Chours), followed by inspection of the equipment and an assessment of
its performance.  The testing sequence which was followed during shakedown and
the resultant accomplishments are summarized in Table 2.  The  table indicates
the reactant media (i.e., coal type and size, liquid phase component, and gas-
eous phase component) and the primary activities carried out during each test run.
                                      44

-------
                                       TABLE  2.   RTU  SHAKEDOWN SEQUENCE
  Run                               Reactant Media
  No.       Coal Type     Liquid Phase           Gas  Phase
                                                                           Major
                                                                      Accomplishments
  001
              None
               Process Water
                        Nitrogen       System  operated warm.  System dynamically  leak checked
                                      under nitrogen gas pressure.  Foreign material flushed
                                      from system  components.
  002
   Coal 1
 14 mesh x 0
100 mesh x 0
Distilled water
(no added iron
 sulfate)
Nitrogen      System operated hot.   Initial  coal  feed operations  and
and oxygen    sampling procedures  evaluated  at two feed coal  sizes.
              Effluent .gas analysis  equipment and procedures  checked.
              Initial data acquisition system evaluated.  Pumping
              capabilities verified.
003/004      Coal 1      Water (no added
           14 mesh x 0    iron sulfate)
                                       Nitrogen      Coal feed system operation and sampling verified.  Reactor
                                                    operated at maximum temperatures and gas pressures.
                                                    Filter settings evaluated and filter cake samples obtained.
                                                    Two shift operation attained.  Initial mass balances
                                                    obtained.  Reactor wall and pump-around loop samples
                                                    evaluated.  Process sampling procedures and control lab
                                                    analysis procedures evaluated.
  005        Coal 1      Water (no added
           14 mesh 1 0    iron sulfate)
                                       Oxygen        Obtained baseline (no added iron salt) reaction data.
                                                    Verified operability of instrumentation needed for mass
                                                    balance.  Evaluated mixer and reactor sampling (wall  vs.
                                                    loop) results and procedures.  Verified coal  analysis
                                                    reproducibility.
  006
   Coal 1
 14 mesh x 0
Leach Solution
(water plus added
sulfuric acid and
iron sulfate)
Oxygen        Verified  proper mixer settings to effect efficient coal/
              leach  solution wetting without foaming.  Verified operat-
              ing,  sampling, and analysis procedures using leach solution.
              Obtained  initial coal desulfurization data with leach
              solution.   Regeneration rate evaluated.  Filter operation
              evaluated with leach solution.
   Test runs 003 and 004 were accomplished during  one  testing sequence.

-------
     As was expected, during shakedown operations,  some equipment related dif-

ficulties were encountered.  Those which resulted in significant equipment

modifications and/or lengthy duration rework are summarized below,

        •   Turbine meters were initially installed in liquid flow moni-
            toring service.  During shakedown operations,  most liquid
            circulating lines tended to build up in particulate loading
            (up to a few hundredths of a percent solids) due to the vari-
            ous recycle loops and the closed loops  water system.   The
            turbine meters were found to be inoperable when subjected to
            any degree of particulate loading.   They were, therefore,
            replaced by magnetic flow meters which  were capable of han-
            dling virtually any stream, from particulate free liquids to
            concentrated slurries.

        •   The Autoweigh weigh belt would not operate to  the required
            ±0.5 percent accuracy as initially configured.  This  degree
            of accuracy was required in order to achieve a reasonably
            accurate mass balance around the system.   Both hardware and
            electronic difficulties were encountered  during the early
            shakedown period.  Weigh belt internals and electronics were
            exchanged several times (by the manufacturer)  for redesigned/
            modified replacements.  Following supplier modifications, the
            unit was found to operate within the specified tolerance
            levels.

        •   Filter modifications were required  after  it was determined
            that back flooding of the belt (slurry  back-up and overflow
            from the rear of the unit) occurred when  trying to achieve
            maximum cake thicknesses or when belt slippage and momentary
            belt slowing occurred.  This problem was  solved by modifying
            the slurry feed distributor and by adding a dam at the rear
            of the belt.

        •   During the initial shakedown test,  it was determined  that the
            reactor feed pump P-l (and its potential  spare P-7) would not
            pump cold solution at the required  pressure.  This was found
            to be the result of internal pump rotor to stator tolerance
            levels.   The close tolerance required to enable the pump to
            achieve high delivery pressures at the  low flow rates and high
            temperatures required for reactor operations grew to  an inoper-
            able tolerance level when pump parts ran  cool.  This  occur-
            rence was the direct result of thermal  expansion/contraction
            effects.   The problem of cold feed  pump inoperability was cir-
            cumvented by revising procedures to include initial heat-up of
            the unit.  This was accomplished via closed loop recirculation
            of heated (by direct steam injection) mix tank T-2 contents
            through P-l and back to the mix tank.  P-l was usable in recir-
            culating  service because of the low head required to pump from
            P-l back to T-2 as opposed to the high  head required to pump
            against reactor R-l pressures.

                                      46

-------
       t   Several  hours into the initial leach solution run (Run 006)
           a  major  leach solution leak (acid spray) occurred which    '
           forced the system into an emergency shut-down.  The plant
           was  shut down and personnel evacuated the test stand within
           minutes  of the incident.  There were no injuries or equip-
           ment damage as the result of the rapid emergency procedure.
           The  cause of the acid spary was found to be an improperly
           installed pipe flange on one of the leach solution supply
           lines.   The flange was determined to be of carbon steel con-
           struction instead of the specified 316 Stainless Steel
           required for acid service.  As a result of the incident, a
           thorough reinspection of the RTU was carried out to assure
           that all piping, flanges, and valves had been installed as
           specified.  No other improperly installed equipment was
           discovered.

        •   Following evaluation of initial leach solution regeneration
           rate data, obtained during the first leach solution run, it
           was  determined that insufficient solution  regeneration in
           the  reactor was occurring.  This occurrence was thought to
           be the  result of low turbulence experienced in the reactor
           pump-around loops where regenerant oxygen is blended with
           circulating slurry.  To enhance regeneration and thereby
            increase coal desulfurization potential of the system, the
           oxygen  blending portion of the recirculation loops were mod-
            ified.   The original 1-1/4 inch Schedule 80 pipe was replaced
           with 3/4-inch Schedule 40 pipe to increase flow velocities
           and therefore increase turbulence.  It was determined in
            subsequent runs that reducing the pipe diameter did indeed
            increase regeneration rates to acceptable levels.

     In summary, the RTU was operating successfully during the final phase  of
shakedown operations and most of the planned objectives of the overall effort
were met.   The various subsystems of the RTU were found to operate in an  inte-
grated fashion as anticipated.  For instance, the three T-2 mixer stages  were
found to be maintainable at uniform temperature or with increasing temperature
for each succeeding stage within the planned slurry flow rate ranges.  Plant
startup was accomplished over an approximate 3-4 hour period, which was deemed
adequate to meet the objectives of the operational test plan.  Reactor tempera-

tures were demonstrated controllable up to 275°F using the available steam
source, and a fully burdened system, including heat exchangers, belt filter,
mixer and reactor.   The reactor pressurization and control system performed as
expected.   The nitrogen, air, and water supply systems were of adequate capacity
and controllability for all purposes intended.  The coal and slurry transfer

systems functioned  satisfactorily, as desired, following equipment modifica-

tions.   The belt filter performed as intended up to a maximum slurry feed rate

                                      47

-------
of 1500 pounds per hour.  Beyond that limit, belt flooding and incomplete  fil-
tration was found to occur.  An acceptable leach solution regeneration rate was
attained following modifications.
     The objectives of the concurrent support activities were also met during
shakedown activities.  All data generated during shakedown was evaluated with
acquisition hardware and software being modified and/or generated where appro-
priate.  Suitable data recording formats were developed which allowed for  fast
and accurate data retrieval.  The data itself was evaluated primarily to deter-
mine baseline process chemistry.  In the control laboratory, procedures were
verified and techniques developed which proved to yield accurate and reproduc-
ible laboratory analysis data.  At the conclusion of shakedown, a revised
Operational Test Plan^10^ was generated and submitted to EPA.  The plan was
revised in the context of the newly acquired data on a working knowledge of the
processing equipment operating limitations and capabilities.  The submission
of that document signaled the end of RTU shakedown operations.
5.2  RTU OPERATION
     The RTU was brought into operational status immediately following shake-
down activities.  During the course of operations, it was demonstrated that the
RTU could be operated continuously to test Meyers Process units for coal-leach
solution mixing, simultaneous coal depyritization and leach solution rege'nera-
tion, filtration of leach solution from treated coal, and product coal water
washing.  No difficulties were encountered in changing process variables such
as coal feed rates, leach solution feed rates, steam flow rates, reactor pres-
sure, oxygen flow rates, agitation and/or pump-around rates, or filter condi-
tions during RTU operations.  Sixteen distinct sets of operational variables
were systematically studied while treating 49,700 pounds of coal over a total
operating period of 254 hours.  Many of the variable changes were made during
continuous elongated run sequences, up to 32 hours in duration, demonstrating
the overall flexibility and controllability of the RTU.
5.2.1  Qperating Procedures
     Prior to initiation of the planned experimental testing effort, the RTU
operational procedures were reviewed and revised in light of the previously
acquired shakedown experience.  Those procedures are tabulated by name and >

                                      48

-------
identification number in Volume  U, Appendix  B of this report,  The following
paragraphs describe the nominal  start-up,  operating and shut-down  sequences
followed during operation of the RTU.
     Prior to RTU start-up, all  auxiliary  supply systems are activated.  It is
the auxiliary supply systems which  provide nitrogen, oxygen, utility/instrument
air, leach solution and process/cooling  water to the RTU.  Nitrogen is used
primarily as a purge gas in the  coal  feed  and reaction systems.  Oxygen is
employed for reagent regeneration in  the primary reactor and, with minor piping
modifications, could be utilized in the  secondary reactor also.  Air  is used to
drive equipment  (i.e., pumps P-8 and  P-ll) and to actuate various  plant instru-
mentation.  Start-up of auxiliary supply systems consists primarily of verifying
all associated valve and regulator settings,  verifying the adequacy of material
supplies, and subsequent supply  initiation.  Once these supporting systems have
been activated,  RTU start-up procedures  are initiated.
     RTU  start-up procedures may be summarized by processing unit.  The first
step in start-up of all units  is verification of associated instrumentation.
Prior to  start-up of the coal  feed system, the coal level in storage  tank T-l
is  verified to be sufficient for test initiation or, if needed, additional coal
is  transferred from tote bins  to T-l.  The storage tank T-l, weigh belt A-3 and
rotory valve A-4 are purged with nitrogen  and then activated.  Start-up of the
mixer 1-2 includes verification  of weir  heights and initiation of  mixing with
M-l through M-3.  Steam flow to  T-2 is initiated and slurry pump P-l  is acti-
vated in  a recirculation mode  to stage three of T-2; slurry recirculation is
maintained until the P-l effluent temperature exceeds 200°F. Start-up of the
primary reactor  R-l requires verification  of seal flush water flow, initiation
of  slurry recirculation through  pumps P-2  through P-6, and activation and
regulation of the R-l  pressure relief valve PIC-43.  Mixers M-4 through M-8 are
turned on to augment agitation provided  by the slurry recirculation loops.
Steam and gas  (either  nitrogen,  oxygen or  a mixture thereof) flows to R-l are
initiated after  verification of  PIC-43.   Activation of the belt filter S-l is
preceded  by verification of all  filter valving and activation of the  vacuum
pump K-l  and reagent filtrate  and wash filtrate pumps P-9 and P-10.   S-l belt
speeds are adjusted to a nominal setting and coal cake steaming and wash streams
are activated.   Processed  coal catch  bins  are positioned prior  to  initiation of
filtration,
                                       49

-------
     Upon pressurization of R-l and heat up of P-l, slurry flow through  the  RTU
may be initiated.  The coal discharge vibrator A-2 is activated and coal  is  fed
to A-3 and, subsequently, to 1-2,  Reagent feed pump P-13 is turned on providing
hot reagent to T-2 through V-l and T-3.  Reactor effluent slurry flow rates
through P-l and KV-241 (R-l slurry vent valve) are adjusted to maintain  desired
slurry levels in T-2 and R-l.  Filter belt speeds are adjusted at this time  to
yield a properly dewatered processed coal filter cake.  Procedures for plant
operation after steady slurry and gas flows have been obtained consist largely
of control adjustments to maintain specified reaction temperatures, pressures,
flow rates and slurry levels.  The complete start-up operation requires  approx-
imately 4 to 6 hours, depending primarily on the steady state reactor temperature
desired.
     Plant shut-down is initiated by terminating all  RTU feed and transfer
streams including coal, reagent, slurry, gas and steam.  Recirculation
pumps are shut off and the R-l unit is depressurized,   Belt filter operation
proceeds until coal is no longer discharged into the  catch bins and then S-l
operation is terminated.  All vacuum and filtrate pumps are subsequently shut
off.  The T-2 and R-l mixers are left operating to prohibit retained slurry
settling.  After the RTU has been secured, all slurry transfer pumps (including
recirculation pumps) and piping are drained and flushed with clear water.  The
final step in the RTU shut-down sequence is shut-down, venting and securing of
the auxiliary supply systems.  The entire shut-down sequence requires approxi-
mately 2 hours to accomplish.  The heated slurries contained within the
mixer T-2 and reactor R-l require an additional eight to ten hours to cool to
ambient temperatures.
5.2.2  Equipment Experience
     During the 4-month operational test phase of the project, 16 discrete tests
were conducted.   The specific test sequences and related processing conditions
over which the RTU was operated are delineated in Section 8.  A summary of the
major equipment operational ranges utilized while carrying out the experimental
test sequences is presented in Table 3.  In general,  the equipment performed
well within the limits studied.  There are, of course, several observations
relating to equipment operating experience worth noting.
                                      50

-------
                     TABLE 3.   EQUIPMENT RANGES UTILIZED
 Equipment
     Sen/fee
Weigh Belt A-3

Mixer T-2


Reactor R-l
Coal  Feed
Coal  -  Leach Solution
Coal  -  Leach  Solution - 0,
Pumps P-l, P-7    Slurry  Feed  Pumps
      P-2/6
      P-8
      P-9
      P-10
      P-ll
      P-12
      P-l 3

      P-14
      P-l 5
      P-16

Mixers M-l/3
       M-4/8
Slurry  Circulation  Pumps
Belt Wash  Pump
Filtrate Pump
Wash Pump
T-6 Discharge Pump
Leach Solution  Pump
Leach Solution  Feed  Pump

Waste Disposal  Pump
Cooling Water Pump
Seal Water Pump

Mix Tank Mixers
Reactor Mixers
  Range  Utilized During
   Operational Testing

 199  - 300 Ibs/hr

 156°- 215 °F
 1.1  - 1.8 hrs residence time

 199°- 273 °F
 23   -  80 PSIG
 5.1  - 8.2 hrs residence time
 58   - 141 SCFH 02

 .8   - 4.2 gpm
 23   -  80 PSI AP
 18   - 25 gpm
 0  -- 2  gpm
 .6   - 4.6 gpm
 1.1  - 2.0 gpm
 Not  Used
 1    - 20 gpm
 1.2  - 3.0 gpm
 23   -  80 PSI AP
 200  gpm
 100  gpm
 12  gpm

150 - 200 RPM
150 - 200 RPM
                                   51

-------
     Coal Feed Equipment.  The coal transfer equipment performed well for coal
feed rates up to 300 pounds per hour.   A minor modification to the weigh belt
unit was required to accommodate accidental coal spillage and dusting over belt
edges.  The modification consisted of the addition of a dust collection chute
and flanged discharge port underneath the belt compartment.  This modification
allowed quick periodic removal of coal  particulate buildup.  A second intermit-
tant situation requiring equipment modification occurred in this section of the
RTU.  The dry coal delivery chute to the mixer T-2 occasionally plugged, requir-
ing short duration recirculation mode operation of T-2 while the plug was dis-
sipated either by external chute tamping or chute disassembly and unplugging.
The cause of this occurrence was found to be steam backflow from the first
stage of the mixer (which was normally operated at 185° to 190°F) which resulted
in coal wetting and agglomeration.  The situation was circumvented by reducing
nominal first-stage mixer temperatures to 170°F, by increasing nitrogen down-
flow through the chute", and by installing a cooling coil to the outside of the
chute.
     Leach Solution Feed Equipment.  The leach solution feed equipment, includ-
ing the solution storage tanks T-7 and T-8, heat exchanger E-2, knock-out
drum V-l, associated piping, valves, control and instrumentation, performed
reliably during operation.  There was,  however, some difficulty experienced with
the leach solution feed pump P-13.  Midway through operational testing, several
outer shaft seal failures were experienced in rapid succession.  Following
thorough disassembly and inspection of pump internals, it was determined that
the pump contained a small inner connecting shaft of carbon steel construction.
It had been installed in the unit during pump manufacture.  The inner shaft had
corroded and resulted in an inner shaft seal failure which allowed the acidic
leach solution to attack the main outer seals.  A replacement of inner shaft 316L
was fabricated and installed in P-13.   Also, near the end of operations a
P-13 rotor related problem was experienced.  The problem was manifested as an
inability of the pump to meet desired throughputs at operating pressures.  It
was determined upon inspection of pump internals, that the chrome plated
316L rotor had crazed and that the chrome plating had flaked off, thus result-
ing in too great a clearance between the pump rotor and stator.  This oversize
clearance allowed backflow of the leach solution and therefore limited pumping
                                      52

-------
ability.   In order to minimize  plant down time, the P-13 rotor was  replated
with chrome and placed back  into service.  The rechromed rotor was  still in
service at the conclusion  of operations.   A new 316L stainless steel  rotor
(not chrome plated) was  fabricated and is currently warehoused as a future
replacement spare.

     Mixing Equipment.   The  coal mixing equipment, consisting of mix tank T-2,
mixers M-l through M-3,  foam knock-out unit T-3, and all associated hardware
performed as expected with the  exception  of the steam inlet lines to T-2.
Early in operations, the lines  were found to clog with coal particulate and
leach solution which had backflowed from  T-2 during unit shut-down  and cooling.
The apparent cause of the  backflow was steam collapse in the lines  upon cooling
combined with faulty check valves in the  inlet lines.  The check valves were
replaced and a steam venting sequence was added to the shut-down procedures.
Another observation worth  noting is that  there were no instances of foaming in
T-2 during the entire operational phase of the project.
     Reactor Equipment.   For the most part, the equipment associated with the
reactor R-l performed as expected.  The slurry circulating pumps P-2 through
P-6, the reactor mixers  M-4  through M-8f  the knock-out drum T-5, the oxygen
and nitrogen systems, the  temperature, pressure and level control systems» and
associated instrumentation performed satisfactorily.  The slurry feed pump P-l
(and its spare P-7) experienced the same  difficulties as experienced by the
leach solution feed pump P-13 (discussed  in preceding paragraphs) at the con-
clusion of operations.   In addition to the flaking and seal problems, the
316L pump rotor also experienced corrosion and/or errosion.  The unit lost
pumping capability as was  the case with P-13.  The additional corrosion/erosion
factor was believed to be  the result of the abrasive nature of the  coal contain-
ing slurry as feed to R-l  as opposed to the leach solution liquid handled by
P-13.  As in the case of P-13,  the reactor feed pump P-l (and P-7)  was replated
with chrome to the specified tolerance limit and put back into service to mini-
mize RTU down time.  The pump ran satisfactorily for several hours  prior to
chrome plate flaking followed by pump failure.  Following the second failure
it was determined that the thicker chrome plating (required to make up for
corrosi on/erosion loss)  was  unable to withstand the thermal expansion experi-
enced during plant start-up.  A new Hastelloy C-276 rotor was fabricated and

                                       53

-------
installed in P-l,  The selection of Hastelloy C-276 was based on RTU corrosion
studies (discussed in Section 10) and pump manufacturer recommendations.
     The R-l ball valve slurry sampling system was found to operate satisfac-
torily.  Both wall and pump-around loop sampling equipment obtained representa-
tive and reproducible R-l slurry samples (discussed in Section 8).  Therefore,
early  in the testing sequence, circulation loop sampling was terminated due to
a tendency for those sample outlets to clog during use.  The direct introduction
of steam into R-l to maintain desired reactor temperatures was found to dilute
the recycled leach solution over prolonged running periods.  An appreciable
part of the dilution was due to repeated plant start-ups (the plant was not
operated over weekends, and was shut down during planned maintenance and repairs).
This dilution factor would be less, were the RTU in a continuous operation mode.
However, reactor leach solution regeneration capability was sufficient to main-
tain adequately concentrated ferric reagent supply throughout operation,
despite the dilution effect.
     The reactor associated equipment which was subjected to elevated tempera-
ture (>215°F) and oxygen partial pressure operation did experience a systematic
corrosion related problem (discussed in detail in Section 10).   During the
course of operations, the reactor R-l, reactor internals, and slurry circulation
loop piping experienced progressively severe corrosion,  The reactor and its
internals experienced pitting and crevice corrosion.   The pump-around loops
experienced crevice corrosion at flange interfaces and pitting.   The damage to
the reactor and loop piping/flanges was sufficiently severe by mid-operations
to warrant reactor repair (filling of the deeper pits with weld rod material)
and replacement of reactor associated piping with new 316L material.  At con-
clusion of operations, following unit inspection, the reactor and its associated
piping were again found to have experienced severe pitting corrosion.  Based on
the RTU materials study effort (detailed in Section 10), it has been determined
that a new reactor and associated piping, flanges and valves should be con-
structed and put into service prior to further RTU operation.  Based on the
corrosion study, the new reactor and piping, flanges and valves should be of
titanium construction.  This construction material is predicated not only, by
material study results but also by the size of the equipment (the reactor and
piping are too small in size to be applicable to metallic and/or nonmetallic

                                      54

-------
liners and/or acid brick lining) and  the  relatively low cost of titanium
replacement parts.  For instance, a quote of $49,300 was obtained in January
1978 for a replacement reactor of titanium construction, as compared to $25,900
which was paid for the original 316L  reactor in  early 1976.
     Filtration Equipment.  The filtration equipment, including the belt fil-
ter S-l, the belt wash return pump  P-8,  the filtrate receiver V-2, the filtrate
pump P-9, the wash receiver V-3, the  wash pump P-10, and the filter vacuum pump
performed as expected.  The belt filter  control  variables (wash water tempera-
ture, flow rate, belt speed, wash location, vacuum level, etc.) were such that
adequate filtration and washing of  14 mesh x 0 coal was attained up to a total
slurry feed limit of 1500  pounds per  hour.  The  belt filter therefore became
the capacity limiting unit in the RTU.  Belt blinding did not occur to an appre-
ciable extent.  However, the filter belt did fail  (it was literally ripped
apart) at one point in the test program.   Belt failure was attributed to a
softening of the belt rubber facing primarily as a result of prolonged elevated
temperature  (180° to 190°F) exposure  to  acidic slurry feed.  The belt, as a
result of the softening, experienced  increased frictional forces at the belt
interface to the unit vacuum pans.   Following belt failure, the belt was
repaired and placed back into service.  The vacuum pan edges over which the
rubber belt  travels were fitted with  teflon overlays.  The teflon-belt interface
resulted  in  less  system frictional  forces, thus allowing proper belt operation
for  the duration of operations.
5.3   RTU OPERATION CONCLUSIONS
        \,   Pre-start-up RTU  safety reviews and personnel training were
             successfully completed.
        2.   Shakedown activities  demonstrated equipment operability, and
             equipment functionality over the desired test ranges.
        3.   Shakedown activities  verified RTU sampling and analysis pro-
             cedures and resulted  in the  acquisition of baseline process
             chemistry data.
        4,   RTU operating  procedures  were modified prior to RTU opera-
             tions in the context  of shakedown derived experience.
        5.  .A revised RTU  Operational Test Plan was generated  based on
             information obtained  during  shakedown.
                                       55

-------
 6.  During RTU operations, 49,700 pounds of coal were treated over
     a total of 254 hours.

 7.  Continuous RTU'operation was demonstrated for periods up to
     32 hours in duration.

 8.  Sixteen distinct sets of operational variables were studied.

 9.  RTU controllability and flexibility was demonstrated.

10.  Sufficient data  was obtained during operations to allow proc-
     ess chemistry evaluation and verification.

11.  Most of the RTU  equipment performed as expected without
     incident.
12.  Some RTU equipment experienced operating difficulties but in
     general modification to the affected equipment and/or pro-
     cedural changes  resulted in acceptable unit operation.

13.  Elevated temperature operation as experienced in the reactor
     section of the RTU resulted in unexpectedly severe corrosion
     to equipment and piping internals.

14.  The reactor R-l  and its associated  piping, flanges and  valves
     need to be replaced prior to further RTU operation.
                              56

-------
                         6.  REACTOR TEST  UNIT  DATA

6.1  SUMMARY OF PREVIOUS BENCH SCALE EXPERIMENTATION
     Meyers Process chemistry has been  characterized  through  bench scale
experimentation consisting of more than 200 mass  balanced  experiments.  These
tests were performed primarily with Appalachian coals^4'5'^1^ although coals
from the Interior Basin and Western regions have  also been tested^'4'.  The
resultant data compilation shows that the  pyrite  leaching  and reagent regen-
eration steps of the Meyers Process may be expressed  as  follows:

       FeS2 + 4.6 Fe2(S04)3 + 4.8 H20 -> 10.2 FeS04 +  4.8 H2S04 + 0.8 S

and

          2.3 02 + 9.2 FeS04 + 4.6 H2S04 -* 4.6  Fe2(S04)3 + 4.6 H20.

     Pyrite leach reaction products were initially determined by direct meas-
urement of sulfate sulfur and elemental sulfur  generated during ferric chloride
leaching.  Multi-stage toluene extractions were utilized to ensure complete
elemental sulfur recovery from processed coals.  Sulfur  balances generally
indicated better than 95% sulfur recovery  and further verified the measured
product sulfur ratios.  Later experimentation performed  with  ferric sulfate
reagent also verified preliminary findings through measurement of pyrite
removals and recovered elemental sulfur.  Thus  it was determined that each
mole of pyrite reacted yields 1.2 moles of reagent soluble sulfate sulfur
(FeS04 and H2S04) and 0.8 moles of elemental sulfur.   This product sulfate
sulfur-to-elemental sulfur ratio of 1.5 was found to  be  independent of the
following processing variables:  1) reaction temperature up to 270 F, 2) reagent
anion (sulfate or chloride), 3) reagent acidity down  to  a  pH  of 1,  4) reagent
total from concentration to approximately  10% w/w and 5) oxygen partial pres-
sures to 85 psia.                     _,
                                      57

-------
     It should be noted that losses or incomplete extraction of elemental
sulfur will result in an apparent product ratio which is greater  than  1.5.
For this reason, bench scale data obtained using single stage toluene  extrac-
tions tend to indicate slightly high product ratios.
     Pyrite leaching and reagent regeneration reactions have been found  to
proceed at rates which are expressed by the following empirical equations:
                                -dW
                                     = KL Y  Wp
where
     KL
     Y

     W_
    is  the  pyrite  leach  reaction  rate  constant
    is  the  ratio of reagent  ferric  ion-to-total  iron concentration
    is  the  weight  percent of pyrite in coal
 and
                 r  -
                 rR
                               dt
 where
      KF
      P0
      Fe
2
+2
is the reagent regeneration  rate  constant
is the oxygen partial  pressure

is the reagent ferrous ion concentration.
      Rate  constants  for  the pyrite leaching and reagent  regeneration  reactions,
 KL and KR, have an Arrhenius type exponential temperature dependence  which is
 expressed  as:

                             K = A exp [-E/RT]
 where
      A     is  the  pre-exponential or frequency factor
      E     is  the  apparent overall activation energy
                                       58

-------
     R      is the gas constant and
     T      is the reaction temperature.
     The pyrite leach reaction frequency  factor AL is a  function of coal type
and coal particle size distribution.   The value of AL increases with decreas-
ing coal particle size distribution  for a specified coal.   For Lower Kittanning
coals such as that from the Martinka  mine,  the value of  A,  obtained from 14
mesh top-size coal samples ranges from approximately 3 x 105  W ~* hour"1 to
7 x 10  Wp"  hour" .  The value of AL may be increased either by size reduc-
tion of the coal or by physical coal  cleaning.  Enhancement of A. by physical
coal cleaning may be attributed to the removal of coal particles containing
pyrite having a low specific  surface  area or low accessibility (isolation by
the matrix material).  Thus,  pyrite  removal  rates may be increased by coal
size reduction and/or physical cleaning.
     An apparent activation energy for the  pyrite leach  reaction E,  of approxi-
mately 11.1 kilocalories per  mole of pyrite reacted was  determined from bench
scale experimentation.  Hence, toe pyrite leach reaction is strongly tempera-
ture dependent with reaction  rates nearly doubling for each 36 F temperature
increment in the temperature  range investigated.   However,  the reaction rate
temperature dependence has been observed  to vanish after 80-90% pyrite removal
from some high pyrite ROM Lower Kittanning  coals.  This  observation appeared
to indicate that pyrite leach rates  can become diffusion limited and thus,
rather insensitive to temperature changes.
     The reagent regeneration rate constant KR is a function  of the regenera-
tion pre-exponential factor AR, apparent  activation energy  ER and reaction
temperature.  The value of AR is dependent  on the efficiency  with which oxygen
and reagent are contacted.  That is  AR contains an efficiency factor (i.e.,
AR = £A') which relates to the effective  reagent-oxygen  contact area and
contact time.  The efficiency factor  e of a specified regeneration apparatus
may have a slight temperature dependence  since solution  viscosity and specific
volume are temperature dependent.  However,  within the temperature range
investigated temperature effects upon e are considered to be  small.  During
bench scale experimentation,  values  of AR up to 6.7 x 10 liters/mole-
atmosphere-minute were obtained.  Since larger values of AR were not attainable
                                      59

-------
at bench scale, this value is assumed to correspond to an efficiency factor
of unity.  The value of ED was found to be a constant 13.2 ki local on" es per
          +2
mole of Fe   oxidized.
     Pyrite leach rates are proportional to the square of reagent Y.  Thus a
high ratio of ferric ion-to-total iron must be maintained during processing
to maximize pyrite leach rates.   In order to maintain high reagent Y values
during processing with low iron reagent or during processing with high solids
content slurries, reagent regeneration may be performed simultaneously with
the leach reaction.  Bench scale tests have demonstrated that reagent Y values
and pyrite leach rates are increased by simultaneous leach-regeneration (L-R)
processing with no apparent adverse affect on regeneration efficiency.
Further, L-R processing has not been found to affect the product sulfate
sulfur-to-elemental sulfur ratio.
     Selectivity of ferric sulfate toward the oxidation of coal pyrite has
been verified by measurements of excess ion consumption and by measurements
of mineral matter free (mmf) coal heat content changes.  The extent of both
excess ferric ion consumption and mmf heat content degradation have been
found to decrease with increasing coal rank.  Bituminous coals of Appalachia
may show approximately 1% decrease in mmf coal heat content (very nearly the
limit of analytical precision) as a result of Meyers processing at 212 F while
the mmf heat content of Western sub-bituminous coals has shown an average
decrease of nearly 7% under identical processing conditions.  Bench scale
experimentation performed with 14 mesh x 0 mine cleaned Martinka coal and
acidified iron sulfate reagent indicated an mmf heat content decrease of less
than 1% for up to 48 hours of 212°F processing, although L-R processing at
250 F for 6 hours or more resulted in an mmf heat content decrease of approxi-
mately 2%(''), Thus, it appears that selectivity of the iron sulfate reaction
with pyrite may decrease somewhat with increasing reaction temperature and/or
increasing oxygen partial pressure.
     Further verification of leach reaction selectivity was provided by mate-
rial balance data which, for Lower Kittanning coals processed at temperatures
of 212 F to 250 F with 0 to 5% w/w iron reagents, indicated virtually complete
solids recovery with a balance precision of approximately 2%.
                                      60

-------
     Product sulfate sulfur  and  elemental  sulfur are generally recovered from
Meyers Process leached coal  by water washing and subsequent solvent  extraction
or vaporization.  The sulfate  sulfur product is readily dissolved in the
reagent solution and may  be  recovered from recycled reagent slip-streams either
by liming or by a combination  of liming and evaporative concentration of
reagent to precipitate ferrous  sulfate.  Elemental sulfur recovery can be
effected by solvent extraction  with numerous solvents which are non-reactive
with sulfur at moderate temperatures including toluene, acetone and  hexane.
Toluene has been most commonly  used for routine laboratory extractions.  Bench
scale extraction experimentation performed with 14 mesh x 0 mine cleaned
Martinka coal shows approximately 97% elemental sulfur recovery with a single
stage toluene extraction.   Acetone has also been demonstrated to be  an effec-
tive sulfur extraction solvent which, unlike toluene, is completely  miscible
with water and can therefore be utilized in conjunction with coal washing.
As such, acetone may  be preferable to toluene for large scale extraction appli-
cations.  Alternate approaches to elemental sulfur recovery include  techniques
involving vaporization into an inert gas stream and sulfur recovery  by vacuum
distillation.  However, vaporization techniques were not employed during the
RTU program.
     Specific details of  previous bench scale experimentation as they apply
to the  RTU program are presented in the ensuing data reduction sections.
6.2  DATA FROM THE REACTOR TEST UNIT
     Operation of the  RTU was continuously monitored and recorded either by
the Doric data logger or  manually.  Primary data sources, those which required
monitoring for plant  operation or evaluation of the Meyers Process kinetics,
were recorded by the  data logger on magnetic tapes.  These tapes were shipped
to Space Park where they  were converted by computer program from the Doric
format  to engineering formats.   Secondary data from the RTU (i.e., flow rates
not critical to plant operation or kinetic evaluation) were monitored and
recorded manually.  Analytical  data relating to slurry and coal analyses were
also recorded manually although the raw data was processed through computer
programs to yield the various analytical results in selected formats.  Computer
programs developed for data treatment during this program are presented in
Volume  II, Appendix A.

                                      61

-------
     Magnetic tapes generated by the Doric data logger are capable of  storing
data from 99 channels plus one self-check channel.  At the conclusion  of the
program, 49 channels were in use to record numerical data and 28 channels were
in service as indicators (alarm channels).  Remaining channels were available
for allocation as required.  Data stored on these tapes were converted from
the Doric format to two engineering formats which are presented in Tables 4
and 5.  The format presented in Table 4 represents a scan of all channels in
use at the time indicated by the heading (Julian day, hour:minute:second) and
is essentially self-explanatory.  Only active channels are displayed.  The
Ramapo excitation monitor (Channel 40) is used as a panel-mounted voltmeter
which facilitates calibration checks of the R-l feed and vent oxygen streams,
Fe-61 and Fe-44.  The transducer pressure sensor monitor (Channel  99) serves
a similar function with regard to pressure measurement equipment.   The format
presented in Table 5 provides all the data taken for each channel  during a
specified experimental period.  The latter format is somewhat more abbreviated
than that presented in Table 4 and improves the ease of data evaluation and
interpretation in terms of identifying trends.  Thus, the format presented in
Table 5 was used most frequently.
     After conversion of the magnetic tape contents to engineering units, the
data were available in storage for use in any of various computer programs
prepared during this project to facilitate data reduction (i.e., plotting,
averaging, mass balances, rate determinations, etc.).  Similarly,  all control
lab analytical data in storage was available for recall  into the data reduc-
tion programs.  Manually recorded pilot plant data required as input for data
reduction programs were input to central  data storage through the  computer
terminals or, for large inputs, were submitted as a keypunch card  deck.  Thus,
nearly all data obtained from the RTU was centrally stored and readily avail-
able for recall.  Only those manually recorded data relating solely to plant
operation were not included in the centralized data compilation.
     A complete file of the raw data obtained from the RTU facility is stored
at TRW.   Data which are pertinent to process evaluation are presented in the
ensuing subsections of Section 6.
                                      62

-------
TABLE 4.  RTU CHANNEL-SCAN AT FIXED TIME
RUN 03-03-00 DAY 10 513*100
CHANNEL
0
1
i
3
4
5
6
7
e
9
10
11
12
13
15
16
17
18
19
20
.21
22
23
24
25
26
27
28
29
30
31
32
11
34
35
34
37
38
39
40
41
42
43
44
.45.
46
47
48
90
91
99
PAPAMlTfR
SELF CHECK
T-Z CELL 1 TEW«>
T-2 CEIL 2 TE*P
T-2 CELL 3 TEMP
R-l CELL 1 TFMP
R-J Cf LL ? TEMP
R-l CELL 3 TEMP
R-l CELL 4 TEM»
R-l CELL 5 TEMP
OXYGEN TFMP TO R-l
V-l DISCH GAS TPMP
LEACH TO V-l TP.MP
LEACH FROM V-l TFMP
R-2 TEMP
QMGEN FLOW TO R-l
V-l DISCH GAS FLOW
02 ANAL REFERENCE
02 ANAL SAMPLE
AUTOWEIGH TOTAL
AUTOWEIGH RATE
R-i PRESSURE
OXYGEN TO R-I
V-l DISCH GAS PPFSS
P-l DISCHARGE FLOW
P-2 DISCHARGE FLOW
P-3 DISCHARGE FLOW
P-4 DISCHARGE FLOW
P-5 DISCHARGE FLOW
P-6 DISCHARGE FLnw
T-3 LEACH FLOW
P-9 DISCHARGE FLOW
P-10 DISCHARGE FLOW
?-7 DI5CHAR6E FLOW
T-2 SLURRY LEVFL
R-l SLURRY LEVFL
pyMP p-i JJISCH ol»FSs
N2 SUPPLY TO R-l
S-l VACUUM PAN A
S-l VACUUM PAN 9
RAMAPO EX. MONITOR
LEACH SL TEMP TO T-3
T-4 TANK TE*P
v-i TANK TEMP
T-4 H20 IN TEMP
T-JO FLUSH H20 TPMP
P-l DISCH TEMP
SI INLET TEMP SLURRY
T-4 H20 OUT TE*»
PUMP P-7 DISCH PRESS
if-2 LIQUID LEVFL
XOUCFR P»S. MONITOR
TAG
UJR-170
T*-l<>
TF-20
TP-?1
TF-5?
TP-53
TF-54
TF-55
TF-S6
TP-173
TF-149
TF-10Q
TF-?00
TF-90
FF-fcl
PP-44
A P-l 71
AF-171
WT-9
WT-9A
PT-173
PT-17'
PT-150
FP-?«»
FP-33
FF-»»4
F*-«5
FF-«6
FP-ft?
FP-31
FF-157
FP-15P
FF-101
PT-26
PT-5B
PT-?<»
<»T-107
PT-249
PT-?50

TP-3?
TF.-?Ol
TF-202
TF-203
TF-195
TP-?«
TP-
TP-204
pT-ibd
PT-95

DATA
4.999
135.7
134,3
1?7.Q
'1?.3
?02,5
20"5.4
?06.2
205.7
•55.0
55.0
50.5
71.6
56.1
*.l
o.o
OR. 61
*°*15
5661
-.2
-57.50
0,0
?1.8
.920
.869
.674
1*842
.465
-.077
-.014
.006
.005
-1.250
13.3
11.0
-50,00
60.10
14.560
-3,750
-6.606
5^.5
65.1
60.0
6ft. 6
69.2
77.3
5". 7
IS". 2
-7.96
2?, 3
14.781
UNITS
VOLTS
DEC- F
•^fc *• f
PEG F
9* & w r
DEC F
OEG F
DEG F
DEG F
DEG F
DEG F
OEG F
OEG F
DEG F
OEG F
DEG F
SCFH
SCFH
PERCENT
PERCENT
POUNDS
POUNDS/MR
PSIG
PSIG
PSIG
GPU
GPP
6PM
GPf
GPM
GPU
GPP
GPM
GPP
GPM
INCHES
INCHES
PSIG
PSIG
PSIA
PSIA
VOLTS
DFG F
DEG F
DEG F
DEG F
PEG F
DEIS F
DtG F
DEG F
PSIG
INCHES
ny
ALARM

































LOWER













LOWER

                  63

-------
                                      TABLE 5.  RTU TIME-SCAN BY CHANNEL
en

RUN
DAY
10
10
10
_m_
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10

03-03-00
TIME
1)23)00
1(26(00
1(29)00
1(35)00
1(38)00
K4KOO
1(44(00
1(47(00
1)50(00
1(53(00
1)56(00
1)59)00
2102(00
2(05(00
2(08(00
2(11(00
2d7(00
2)20100
2(23(00
2(26(00
2(29(00
2(32(00
2(35(00
2(38100
2(41(00
2144(00
2(47(00
2(50(00
2(53)00
2(56:00
2(59100
3(02(00
3(05(00
3)08(00
3(11(00
3114(00
CHANNEL
0
UJR-170
4.999
C_HAMEJL_
1
TE-19
156.5
CHANNEL CHANNEL
2 3
TE-20 TE-21
207.9 213.9
4.999 159.8 205.1
4.999 159.5 200.7
	 4.999 158.1 	 JO.4.,4
4.999 157.8 209.4
4.999 159.3 211.2
4.999 161.3 210.1
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4,999
4.999
4,999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
4.999
163.9
166.2
166.8
166,2
166.0
165.0
164.8
164.4
163.9
163.0
162.6
163.0
164.6
165.4
164.9
163.4
162.2
163.0
164.4
166.2
167.8
170.0
171.8
174.4
177,4
173,6
176.6
171,6
166. B
161.3
159.6
211.2
209. A
207,2
205.5
206.5
204.7
203.6
203. S
202.9
203.0
208.8
210.1
208.9
204.3
202.4
205.7
20B.1
210.8
211.6
212.9
213.2
213.8
233.9
212.9
213,3
211.7
207.4
202.7
199.4
193.9
190,1
213.8
213.8
	 2JL3.J>_
813.8
214.0
213.9
213.9
213.7
213.8
213.8
213.8
ai3.7
213.7
213.7
213.4
213.4
213.6
213.9
213.3
213.8
213.6
213.5
213.8
213.8
Z14.0
213.9
£14.0
214.0
214.1
214.0
214.0
214.0
213.8
213.9
213.6)
212.9
211.6
CHANNEL C
4
TE-52 T
235.9
234.5
233.5
zYi. 7 "~
231.5
230.3
228.7
226.6
£26.1
226.3
223.4
223.2
223.9
224.6
223.5
222.0
221.5
219.7
E19.7
219.5
219.9
219.9
219.2
215.5
217.3
219.2
219.7
216.5
219.0
219.3
217.5
217,3
217.2
217.9
216.7
216.4
;HANNE_L a
5
'£-51 Tl
230.0
229.1
229.3
"228"". 3
227.7
227.2
226.0
225.9
225.5
224.9
224.2
223.8
223.0
222.1
221.6
221.0
220.8
221.1
221.0
220.5
221.2
221.4
221.1
221.0
221.5
221.2
220.3
220.4
219.9
219.1
219.5
219.5
219.4
219.3
219.4
219.4
218.9
HANNEL
6
E-54
236.7
235.4
235.0
234,3
233.7
232.9
232.1
231.1
230.4
229.7
229.6
226.5
226.4
227.2
226.6
226.7
226.8
230.2
233.5
236.3
236.7
240.6
241.9
241.6
238.7
236.9
234.5
233.3
232.2
233.3
235.1
235.7
236.6
237.5
238.5
239.3
239.4
239.9
CHANNEL
7
TE-55
237.5
237.5
236.8
2.36. 4_
236.0
235.5
234.9
234.5
233.1
232.5
231.3
231.7
230.8
230.0
229.6
229.1
229.1
228.4
228.1
228.5
228.9
229.3
230.3
230.9
231.5
231.6
232.4
232.0
232.1
231.3
231.4
231.5
231.3
231.5
231.6
231.8
232.0
232.1
CHA N tiii. -CHAUHEi _ £iLAJ4N£L Jl H AJJ N E L
6 9 10 11
TE-56 TE-173 TE-148 TE-199
240.4 62.6 57.6 59.3
240.1
236.9
. 23,8.. 2
237.4
236.6
235.9
235.6
235.0
234.5
233.2
233.0
232.0
231.2
230.9
229.8
234.3
229.0
226.6
228.2
227.3
228.7
236.1
238.5
237.2
236.1
235.7
234.4
233.7
233.6
232.8
232.3
231.8
231.2
230.6
230.2
229.8
229.8
62.2
62.1
61.9
61.9
61,9
61.8
62.1
61,7
61.8
61.6
61.8
61.6
61.4
61.4
61.4
61.3
62.1
61.2
60.6
60.6
60.4
60.3
60.2
60.3
60.2
60.2
60.1
60.4
60.4
60.2
60.4
60.2
60.5
60.7
60.3
59.5
57.8
57.3
5ii«L_
57.4
57.3
51*4
57.9
57.5
97.9
57.7
57.6
57.6
57.7
58.2
57.8
58.2
58.2
58.1
57.6
57.7
58.0
58.0
57.4
57.4
57.6
57.6
57.8
57.3
57.5
58.0
58.0
57.6
57.5
57.4
57.4
57.4
57.8
59.7
59.3
59.2
59.2
59.4
5A.3
59.5
59.4
~59.4~
59.4
59.2
59.6
59.1
59.7
59.2
59.0
59.4
58. 8
59.4
59.5
59.2
59.4
58.7
59.4
59.4
59.3
59.4
59.6
59.6
59.7
59.7
59.1
59.2
58.4
59.4

-------
6.2.1  Mine-cleaned Martinka  Coal  Characterization
     Samples of mine-cleaned  Martinka coal  processed in the RTU were provided
by American Electric Power  (AEP)  at three different times:   (l) coal No. 1
was received in November 1976 and was utilized for RTU shakedown  operations,
(2) coal Nos. 2 and 3 were  received in June and July of 1977,  respectively,
and were utilized during the  RTU  operational phase of the current program.
Samples of each coal were obtained for analysis from the A-3 weigh belt at
frequent intervals during plant operation.   All samples were analyzed at the
TRW control lab for sulfate and pyritic sulfur (sulfur forms)  and a portion
of these were additionally  analyzed by Warner Laboratories for moisture, ash,
heat content, and total  sulfur (short proximate analysis), as  well  as for
sulfur forms.  Warner Laboratories also performed coal iron analyses on all
samples submitted.
     Starting coal analyses of the mine-cleaned Martinka coals are presented
in Table 6.  Results from each laboratory are presented separately and then
averaged for each coal.  The first column of the Table lists the  identifying
number of the coal lot  from which analytical samples were drawn.   Columns two
and three list the analytical laboratory and number of samples analyzed by
each.  Short proximate,  sulfur forms, and coal iron analyses are  listed in the
remainder of the columns.
     Analyses of coal No. 2 are presented in two parts since approximately
half of coal No. 2 (Coal 2A)  was  stored under a partially enclosed shelter
prior to processing while the other half (Coal 2B) was fully exposed to the
atmospheric environment.  Coal Nos. 2A and 2B were treated as  distinct RTU feed
coals prior to complete  analysis  since the extent of weathering  in these two
coals, and consequently  their starting pyritic sulfur content, was expected
to differ.
     To compare coal Nos.  1,  2, and 3, consider first their respective dry
short proximate analyses  (ash, heat content, and total sulfur).   Coal No. 1
differs significantly from  coal Nos. 2A, 2B, and 3 with respect  to ash content;
this difference exceeds  three times the largest indicated standard deviation
in ash measurement.  On  the other hand, coals 2A, 2B, and 3 are  statistically
identical with respect  to ash analyses.  As would be expected  for coals from
the same mine, the same  set of observations holds true with respect to coal

                                       65

-------
TABLE 6.  AS RECEIVED MINE-CLEANED MART IN KA COAL ANALYSES
Coal ID Lab "^
1 Warner 3
CTS 14
Average
2A Warner 8
CTS 10
Average
2B Warner 7
CTS 16
Average
2 (2A Warner plus
2B Warner & CTS)
3 Warner 10
CTS 12
Average
Moisture
1.32
±0.173

1.32
±0.173
0.93
±0.285

0.93
±0.286
0.79
±0.183

0.79
±0.183
0.86
±0.170
1.01
±0.204

1.01
±0.204
Ash
16.11
±0.261

16.11
±0.261
14.05
±0.339

14.05
±0.339
14.24
±0.354

14.24
±0.354
14.15
±0.245
14.32
±0.469

14.32
±0.469
Heat
Content
12508
± 77

12508
± 77
12928
± 103

12928 '
± 103
12919
± 33

12919
± 33
12923
± 54
12907
± 58

12907
± 58
ST
1.73
±0.045

1.73
±0.045
1.61
±0.089

1.61
±0.089
1.57
±0.059

1.57
±0.059
1.59
±0.053
1.49
±0.081

1.49
±0.081
sp
0.63
±0.063
0.70
±0.069
0.68
±0.071
0.62
±0.062
0.79
±0.061
0.71
±0.107
0.56
±0.065
0.57
±0.072
0.57
±0.069
0.58
±0.038
0.60
±0.023
0.67
±0.043
0.64
±0.051
Ss
0.47
±0.035
0.43
±0.076
0.43
±0.071
0.32
±0.058
0.30
±0.041
0.31
±0.049
0.35
±0.054
0.35
±0.073
0.35
±0.067
0.34
±0.036
0.31
±0.029
0.31
±0.026
0.31
±0.027
So
0.63
±0.011

0.62
0.67
±0.080

0.59
0.66
±0.054

0.65
0.67
0.59
±0.067

0.54
Fe
1 .33
±0.010

1 .33
±0.010
1.14
±0.063

1.14
±0.063
1 .11
±0.059

1.11
±0.059
1 .13
±0.043
1.02
±0.071

1 .02
±0.071

-------
heat  content.   Considering coal  total  sulfur analyses, coals 2A and 2B are
statistically  identical,  yet differ significantly from coals 1 and 3 which
differ significantly  from each other.   Hence, based on the short proximate
analysis,  coal  Nos.  1,  2, and 3 differ in total sulfur while 2A and 2B are
identical.   These results are not surprising since the three coal samples
represent  output of the Martinka mine cleaning facility at three different
times.  Note that while the short proximate analyses of coal Nos, 1, 2, and
3 are significantly different in a statistical sense, they may not represent
significant variance in output for a commercial cleaning facility.
     Best  estimates of the starting coal sulfur forms were generally obtained
by averaging mean results from both analytical laboratories.  This approach
was applied to obtain sulfate sulfur estimates for all starting coals and
pyritic sulfur estimates for coal Nos. 1 and 3.  However, considerable bias
was found between CTS and Warner analyses of S  in coal 2A; CTS and Warner
analyses of 0.79 and 0.62, respectively, differ by approximately three standard
deviations.  Since in all other respects coal Nos. 2A and 2B appear to be
identical, they should have identical S  contents.  Considering that Warner
analyzed coal No. 2A to have 0.62% w/w S  and that both labs agree on a
0.57% w/w S  content for coal No. 2B, the CTS analyses showing coal No. 2A
to contain 0.79% w/w S  are assumed to be Incorrect.  Hence, the best estimate
of coal No. 2 S  content was taken to be the average of available mean values,
disregarding the CTS analysis of coal No. 2A.  Average coal analyses thus
obtained for coal Nos. 1,2, and 3 were utilized in all ensuing leach rate
evaluations.
      Measured bias in sulfur forms analyses performed by CTS and Warner
Laboratories is summarized graphically in Figures 15 and 16 for starting and
 RTU  processed mine-cleaned Martinka coals.  Warner analyses are plotted as a
 function of CTS analyses in these Figures and the zero bias line is provided
for  reference.  Pyritic sulfur analyses are seen to be only slightly biased
with  most deviations from zero bias falling within the limits of combined
analytical and sampling uncertainty (approximately 0.06% w/w during this
experimentation as indicated by starting coal analyses).  The one data group-
 ing with bias which significantly exceeds analytical and sampling uncertainty
is the previously mentioned starting analyses of coal No. 2A.   Indeed, analyses

                                      67

-------
                    COAL 1
                    COAL 2A
                    COAL 2B
                 o  COAL 3
.1     .2     .3     .4    .5
              CIS ANALYSES
                                             .6
.8    .9
        Figure 15.   Comparison  of Capistrano  and  Warner Laboratories
                    Pyritic  Sulfur Analyses
of both starting and processed  coal  No.  2A indicate that CIS S  analyses are
biased somewhat high with respect to Warner analyses.   Whether this is due to
the analytical  apparatus  or personnel  variance is not known.  The generally
slight bias observed is substantially smaller than the ASTM acceptable-differ-
ence between two laboratories analyzing  the same coal  with less than 2% S ,
namely 0.30% w/w.  Hence, obtaining best estimates of the true coal S  content
by averaging results from both  laboratories appears justifiable on the basis
of both starting and processed  coal  analyses.
     Bias associated with S$ analyses is negligible as indicated by Figure  16.
Deviations from zero bias are generally  less than 0.04% w/w which corresponds

                                      68

-------
                        A  COAL  1
                        •  COAL  2
                        o  COAL  3
                0      o.l    0.2   0.3   0.4   0.5    0.6   0.7  0.8
                                    CTS ANALYSES

        Figure  16.  Comparison of Capistrano  and Warner Laboratories
                    Sulfate Sulfur Analyses

to the ASTM acceptable difference between  analyses for S  by two different
                                                        o
laboratories.
     Coal samples were provided by AEP  in  70 to  90 ton lots at a 1.25-inch
top size.  These coal samples were received by Ultrasysterns, Irvine,  California,
where they were  size reduced to approximate the  nominal 14 mesh x 0 size dis-
tribution previously tested at bench  scale.  Size distribution data from coal
Nos.  1, 2, and 3 are presented in Table 7.  These data indicate good coal
preparation reproducibility.  A graphical  comparison of these data and pre-
viously obtained bench scale size distribution data is presented in Figure 17.
Coals prepared for the RTU are seen to  be  only slightly finer than the nominal
14 mesh x 0 mine-cleaned Martinka coal  utilized  during bench scale testing.
                                     69

-------
TABLE 7.  SIZE DISTRIBUTION DATA FOR MINE-CLEANED MARTINKA COAL PROCESSED IN THE RTU
Screen
(Tyler Equiv.)
35
48
100
150
200
Opening
(Microns)
420
297
149
105
74
Cumulative Percentages by Weight
Coal 1 (6 samples)
7.9
18.1
43.4
53.0
63.5
± 0.49
± 0.68
± 1.13
± 1.10
± 1.13
Coal 2 (
6.5
14.7
38.5
52.0
62.9
8 samples)
± 1.14
± 1.66
± 2.41
± 1.68
± 1.42
Coal 3 (1
8.6
16.6
40.9
54.3
68.6
2 samples)
±0.63
±0.89
± 2.06
± 3.20
± 4.66

-------
              CUMULATIVE  PERCENTAGE BY WEIGHT
       o
o
OJ
o
CO
o
in
o
vo
o
00
LU
o.
o
.ObO

742


.525
071
. 31 I
.263
IOC
.131

• uz/o

nee



.0328
.0232
.0164
.0116
nno9
.UUo£

0058

.0041
no?q




















KB





*





















>DO



























*




























w



p

























»
*
JO












i— C
•J M
4 K

^ r


n -








'
cm








o
X
DC
LU
LU
sj CO i
J — 1 O
£ < CO
3 0
Jr > -T-

3 =3 Z
- |— LU
c D: CQ

a n ^









^








X
n:
CO
s:
o
0
LU
3;
C_3
•sz
LU '
CQ
1





















A
— i.
s
o
p m

i— «
in o
1 U j"
1 A ^<
_Li_ ;=
rn
20 3
m
on
48
cc
_a2._

100

150
200

   Figure 17.  Mine-Cleaned  Martinka Coal  Size Distribution Data
                                 71

-------
6.2.2  Coal  Processing Data
     RTU experimentation was performed under a variety of operating  conditions
which, at bench scale, were indicated to be suitable for demonstration  of both
plant operability and viability of the Meyers Process at large scale.   Param-
eters of primary interest were varied as follows:
        •   Reaction temperature - 230°F to 270°F
        •   Coal residence time - 5 to 10 hours
        •   Reagent total iron concentration - 0.0 to 4.5% w/w
        t   Reagent acid concentration - 0 to 4% w/w
        •   Oxygen partial pressure — 20 to 50 psia.
     Reaction times were varied by varying the number of stages utilized  in
the primary reactor R-l and/or varying the slurry feed rate.  Coal feed rates
of 200 to 300 pounds per hour were utilized with reagent feed rates  of  1  to
2 gpm.  This corresponds to a range of slurry solids concentrations  from  25-
33% which was tested during the RTU operational phase.
     Process evaluation was based primarily on analyses of product coal samples
taken from the RTU filter belt S-l.  Product coal samples were generally  taken
from the filter belt every hour during transient plant operation and more
frequently during steady state operation.  Slurry samples drawn from various
stages of the mixer T-2 and the reactor R-l provided data on reagent total
                                 +3
iron concentration, reagent Y (Fe  /total Fe ratio) and slurry solids.
     Coal samples obtained from S-l received a water rinse on the filter  belt
and were essentially free of bulk reagent.  Final coal washing was performed
at the CTS control lab and consisted of a slurry wash and subsequent cake wash.
Slurry samples were filtered and the solids were washed; slurry filtrates were
analyzed for total iron, iron forms, and sulfate content.
     The water wet coal was vacuum dried at 212°F for approximately  4 hours
prior to toluene extraction of product elemental sulfur.  The vacuum drying
process was repeated after toluene extraction of the coal.  Product  elemental
sulfur was recovered from the toluene extract by solvent distillation.  Fully
processed coal samples were then subjected to short proximate, sulfur forms,
and iron analyses.  The specific detailed procedures followed in accomplishing
                                       72

-------
these operations were identical  to those developed during previous bench-
scale programs  f4'5'11).
    Bench scale testing  of vacuum distillation^   ' indicates  that under  the
conditions of vacuum drying employed during the RTU program, little or no loss
of elemental sulfur is  incurred.   Hence, vacuum drying prior to toluene extrac-
tion of processed  coals was not expected to significantly affect measured
product sulfate sulfur-to-elemental sulfur ratios.
     Raw data from the  workup and analyses of coal and reagent samples were
reduced and  tabulated  in  the desired format by computer.  Computer programs
written during  this project to perform these computations are  included in
Volume II, Appendix A.
6.2.2.1  Pyrite Leach  Rate Data--
     Experimentation performed with the RTU consisted primarily of two types:
1) tests performed with acidified iron sulfate reagent and 2)  tests performed
with initially  iron  free aqueous dilute ^SO^.  In the latter  case, reagent
iron was derived  entirely from the coal being processed.
     During testing with acidified iron reagent, the leach solution was
recycled back to  the  reaction system via the leach solution storage tank T-7.
This mode of operation resulted in reagent dilution from the live steam used
to maintain reactor  temperature since salt was not continuously added to the
 reagent recycle stream.  Limited salt addition was employed to reduce this
 dilution effect.   However, since reagent total iron concentration does not
 directly affect pyrite leach rates during simultaneous Teaching-regeneration
 operation, reagent iron concentrations were allowed to range from 4.5% w/w
 to  2.7% w/w.

         Results  from Processing with Acidified Iron Sulfate Reagent

     Eleven sets  of  experimental conditions were tested (Experiment 01) with
 acidified iron  sulfate reagent during the operational phase of the program.
The initial  reagent  feed was composed of 4.5% w/w  iron (as ferric sulfate)
with 4.2% w/w H2S04.   Owing to dilution, the reagent total iron concentration
 ranged down to  2.7% w/w.   However, a sulfate-to-iron mole ratio of approxi-
mately 2 was maintained for all tests.

                                      73

-------
     A summary of the principal parameters associated with  that set of experi-
mental conditions is presented in Table 8.  The experiment  numbers are listed
in the first column.  Columns two and three list the experimental  start and
stop times, specified in terms of Julian day and hour.   Reactor temperature,
pressure and oxygen partial pressure are presented  in columns  four, five and
six.  Feed rates of coal, leach reagent and oxygen  are  listed  in columns seven
through nine.  Listed in column ten are the flow rates  of slurry through the
R-l circulation/oxygenation loop.  Columns eleven and twelve contain the esti-
mated mean slurry residence times in T-2 and R-l based  on the  ratio of reactor
volume-to-volumetric slurry flow rate.  All experiments  were performed utilizing
only three of the available five stages of R-l, thereby  maintaining relatively
short residence times.
     Evaluation of this experimentation was performed on the basis  of S  analy-
ses of steady state coal samples taken from the RTU filter  belt.   The minimum
processing time required to achieve steady state operation  under constant
operating conditions was estimated as the ratio of  utilized vessel  volume (T-2
and R-l) to slurry volumetric flow rate.  Verification of this  steady state
estimate through product coal analyses is typified  by Experiment 01-03, the
data from which are presented in Figure 18.  Steady state for  Experiment 01-03
was estimated to be attained after 7.7 hours of plant operation  at  the speci-
fied test conditions.  Note that the product coal S  content reached a value
which is constant, to within analytical precision,  after only  5  hours of
operation.  These data are typical of all experiments performed  to  date with
the RTU.  Hence, a plug flow type steady state estimation was  found to be a
suitable guideline for processed coal data acquisition  from the  RTU.
     Steady state processed coal analyses from RTU  processing  with  iron reagent
are summarized in Table 9.  All steady state data are included  in  the presented
averages.  The number of processed coal samples submitted for  short proximate
and sulfur forms analyses are designated by xP and yF,  respectively, and listed
in  column two of the Table.  Mean R-l reaction temperatures are  listed in col-
umn three.  And columns four through ten contain the average short proximate,
sulfur forms and coal iron analyses.  Experiments 01-04  and 01-08 are considered
to  be merely extensions of Experiments 01-03 and 01-07,  respectively, since no
effect was observed by increasing slurry recirculation  rates from 20 gpm to
25  gpm.  Hence, these experiments are not listed separately in Table 9.
                                      74

-------
             TABLE 8.  OPERATING PARAMETERS FOR ACIDIFIED IRON  SULFATE  REAGENT LEACH  EXPERIMENTATION
-J
en
Experiment Run time
No. ' (day/hour)
Start Stop
01-01
01-02
01-03
01-04
01-06
01-05
01-07
01-08
01-09
01-10
01-11
299/16:15
300/03:15
305/10:00
305/21:30
305/00:00
314/15:45
321/17:00
322/02:50
322/05:03
322/13.54
325/15:30
300/02:00
300/10:10
305/21:30
305/23:30
306/10:45
315/02:00
322/02:27
322/04:50
322/13:33
323/02:30
326/03:00
R-l*
Temp
(°F)
230
231
251
251
251
250
245
245
270
270
251
R-l
Pressure
(psig)
28
55
54
53
67
40-54
25-68
65
77
51
67
02 Coal
Pressure Feed
Rate
(psla) (Ib/hr)
22
49
39
38
52
25-39
12-55
52
50
24
52
299
299
299
299
299
298
298
298
298
298
298
Leach
Feed
Rate
(gpm)
1.0
1.0
0.8
0.8
0.8
1.0
1.0
1.0
1.0
1.0
1.0
02
Feed
Rate
(SCFH)
56
57
54
59
123
165
59
57
56
56
128
Reel re.
Rates
(gpm)
20
20
20
25
20
20
20
25
20
20
20
Residence time**
T-2 R-l
(hrs) (hrs)
1.6
1.6
1.83
1.83
1.83
1.6
1.6
1.6
1.6
1.6
1.6
5.1
5.1
5.9
5.9
5.9
5.1
5.1
5.1
5.1
5.1
5.1
      * The temperature of the primary reactor  was  varied as required for parametric investigations.  T-2
        mixer temperatures were the same  for  all  experiments and ranged from approximately 170°F in the first
        stage to 214°F 1n the third stage.

     ** Slurry residence time 1s computed as  the  ratio of reactor volume-to-volumetric slurry flow rate.

-------
0.20
01 ft
. 1 o
01 r
, 1 b
01 /i
. 1 *\
•3.
^ 0 12
Q.
00 o in
_, U-IU
0
-_. n np
tJ U , Uo
LU
oo
UO
LU
S^ n nfi
o U .UD
OL.
On/i
. U'l
n n?
0
























































.























•

m
-H
i — i
—1
- n
0
CO
—I
m
a
-H
m
—I
t— <
m
I

•











































AVE. =










0.165










t o.c










129 %










W/W









0123
456           7     8     9     10     11     12



           EXPERIMENTAL RUN  TIME,  HOURS
               Figure 18.   Pyritic Sulfur Analyses  of Processed  Coal  from  Experiment  01-03

-------
      TABLE  9.  SUMMARY  OF PROCESSED  COAL ANALYSES OBTAINED  DURING
                 RTU EXPERIMENTATION WITH ACIDIFIED IRON REAGENT
Exp. No. No. of* Processing
Samples Temp.,
(°F)
Starting 15P; 31 F
Coal 2
01-01 2P; 4F 230
01-02 IP; 7F 231
01-03 3P; 6F 251
01-05 IP; 4F 250
01-06 2P; 4F 251
01-07 2P; 7F 245
01-09 IP; 3F 270
01-10 2P; 12F 270
01-11 2P; 11F 251
Coal Composition, % w/w (Except Heat Content), Dry Basis
Ash
14.15
±0.245
12.28
±0.212
13.14
12.73
+0.085
13.96
12.69
±2.14
13.32
±0.863
10.65
13.56
±0.636
13.80
±0.636
Heat
Content,
Btu/lb
12923
± 54
13397
± 52
13300
13286
+ 65
12888
13163
± 299
13123
± 52
13487
13158
± 119
13071
± 96
Total
Sulfur,
St
1.59
±0.053
0.94
±0.000
0.85
1.00
+0.006
0.85
0.90
±0.014
0.87
±0.035
0.93
0.91
±0.092
0.83
±0.021
Pyri ti c
Sulfur,
SP
0.58
±0.038
0.23
±0.083
0.25
±0.055
0.16
+0.033
0.16
±0.068
0.15
±0.027
0.19
±0.075
0.13
±0.006
0.17
±0.037
0.16
±0.044
Sulfate Organic
Sulfur, Sulfur,
Ss So
0.34 0.67
±0.036
0.09 0.62
±0.033
0.13 0.47
±0.120
0.09 0.75
+ .040
0.11 0.58
±0.025
0.11 0.64
±0.015
0.12 0.56
±0.023
0.18 0.62
±0.015
0.12 0.62
±0.029
0.13 0.54
±0.025
Iron,
Fe
1.13
±0.043
0.52
±0.071
0.46
0.45
±0.007
0.53
0.42
±0.029
0.47
±0.021
0.50
0.49
±0.035
0.55
±0.007
The number of short proximate analyses  (ash, heat content,  total sulfur, and coal  iron) and  sulfur
forms  analyses (sulfate and pyritic sulfur) will be designated by xP and yF, respectively, where x
and y  are the number of analyses included in the indicated  average.   The total  number of coal samples
analyzed is y minus x.

-------
     Data presented in Table 9 indicate that all experiments, except  for  those
run at the nominal  230°F temperature, resulted in product coal Sp contents
between 0.13% w/w and 0.19% w/w.   While the 0.06% w/w spread is very  nearly
the limit of pyritic sulfur analytical  precision, the indicated Sp values are
means of distinct populations obtained under different operating conditions
(i.e., temperature, pressure, reagent Y) and must be treated as discrete data.
Experiments 01-01 and 01-02, which were performed at nominally 230 F, resulted
in product coal S  values of 0.23% w/w and 0.25% w/w, respectively.
                 P
     Product coal S  contents show an average reduction of 64% corresponding
to a reduction of S  from 0.34% w/w to approximately 0.12% w/w.  The  slightly
higher product coal S  content obtained from Experiment 01-09 may be  the result
of combined high temperature and high oxygen partial pressure (49 psia com-
pared to 23 psia for Experiment 01-10)  since the solubilities of basic ferric
sulfates, ferrous sulfate and gypsum all decrease with increasing temperature
above about 212°F.   However, the S  data presented in Table 9 do not  generally
indicate a temperature effect upon product coal  sulfate contents.
     The total sulfur content of the mine-cleaned Martinka coal was reduced
from 1.59% w/w to 0.83-1.01% w/w under the conditions tested.  Additional  water
washing of these product coals to reduce their S  contents to the design level
                                                o
of 0.02% w/w would enable products of the majority of these experiments to meet
federal standards of 1.2 pounds of S02 per 106 Btu.
     Ash removals above those anticipated on the basis of pyrite and  iron
sulfate removed from ROM Martinka coals have previously been found to exceed
            (1)
four percentv  .  However, the data presented in Table 9 shows a mean ash
decrease of approximately 1.25% w/w which may be compared to the mean theo-
retical ash removal (based on AS  and AS ) of 1.04% w/w.  Thus, only  0.2% w/w
of excess ash was removed.  This is not surprising since the concentrations of
reagent soluble non-pyritic mineral matter (i.e., iron oxides, iron sulfate and
gypsum) are expected to be reduced by physical cleaning.  That this is the case
for the Martinka coals provided for this program is indicated by the  fact that
these coals contain little or no iron other than that associated with pyrite
and iron sulfate.  Although little excess ash was removed from mine cleaned
Martinka coal, the  removal of 1.25% w/w ash orginating from pyrite and sulfate
represents an upgrading of the coal with respect to both the coal  sulfur content
and the coal heat value.
                                     78

-------
    Correlation of these data was obtained with  the previously determined
pyrite leaching rate expression,  namely:
                                 -dW
                                 	-E =
                                  dt
                                            2 U2
                                                                         (1)
where
     r.      is  the  pyrite  leaching  rate,  expressed in weight of pyrite
            removed per  100  weights of coal  per hour (rate of coal
            pyrite  cone, reduction),

     W      is  the  pyrite  concentration in coal at time t in wt.  percent,

     t      is  the  reaction  (leaching) time in hours,

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

     K.      is  the  pyrite  leaching  rate constant (a function of temperature
            and coal  particle  size) expressed in (hours)-l (wt.  percent
            pyrite  in coal)"1.
with
                            KL = AL exp (-EL/RT).
                                                                         (2)
where
     R

     T
           is the Arrhenius frequency factor  in  the  units of K. ,

           is the apparent activation energy  in  calories/mole,

           is the gas constant in calories/mole°K, and
           is the absolute temperature,  in   K.

     This rate expression was used in conjunction with  a model of the reaction
system as a series of continuous-flow stirred-tank  reactors.  Measured slurry
temperatures and reagent Y values for each reaction  stage were input for each

experimental condition and the value of K, was determined at the various

reaction temperatures tested.
                                      79

-------
     Equation (2)  indicates  that a  plot of In KL vs. 1/T should yield  a
straight line having a slope of -EL/R and an intercept of In A,_.  An Arrhenius
plot of the data from Experiments 01-01 through 01-11 is presented  in  Fig-
ure 19.  A very good linear  correlation was obtained.  The apparent activation
energy and frequency factor  indicated by these data are:
     E.  = 26.9 Kcalories/mole,  and
                  14  -1    -1
     AL - 9.2 x 101  W *  hr  .
     These values may be  compared with bench scale coal  processing results
which yielded an E  (Bench)  of  11.1 Kcalorie/mole and AL (Bench) of 2.95 x
105 W"1 hr"1.  Hence, K,  (RTU)  was  measured to be 3 to 8 times greater than
K  (Bench) for 14 mesh top-size coal  at temperatures between 230 F and 270 F.
     The deviations of A.  and E,  from the values obtained during batch reactor
(bench scale) operation were only partially anticipated.  The apparent value
of A.  was expected to increase  during stagewise operation as the mixing in
reaction stages deviated  from ideal and the average particle residence time
per stage increased.  Also,  under non-ideal mixing, pyrite rich particles
might be expected to have longer residence times than pyrite lean particles
which have a lower density.   Thus,  coal processing in a  continuous flow
stirred tank reactor is expected to result in a higher apparent A.  value than
would identical coal processing in  a batch reactor.  However, the apparent
doubling of the activation energy E.  above E.  (bench) was not anticipated.. Jo
effect such an increase in EL requires that deviations from ideal mixing or
segregation of pyrite rich coal  particles increase with  increasing temperature.
A decrease in reagent density could increase coal segregation despite mechan-
ical mixing and, perhaps, selectively cause the pyrite rich coal particles to
remain longer in each reaction  stage.  While density changes due to tempera-
ture alone are small, this effect coupled with reagent dilution can be sig-
nificant.  Note that the  conditions of Experiment 01 were performed in the
order of increasing temperature (Table 8) and that, as previously stated, the
reagent total iron was reduced  from 4.5% w/w to 2.7% w/w by dilution with
process steam.  The combined temperature and dilution effects are estimated
to have resulted in mean  reagent densities of 1.15 gm/cc during 230°F tests,
1.08 gm/cc during 250°F tests and 1.06 gm/cc during 270°F tests.  Thus, sig-
nificant reagent density  differences occurred between the 230°F and 270°F
                                     80

-------
   3.
   2.
   1.0
                                        o xo
                                         Q
,«•  .8
                                                 \
{_>


UJ
a:
                      270°F
         250°F
   230°F
     2.4
2.5
2.6 x 10~a
                                    1/T, '
        Figure 19.   Arrhenius Plot of Data from RTU Experiments

                     01-01  to  01-11
                                       81

-------
experimentation which may have resulted in the apparent increase  in  the  pyrite
leaching activation energy.   Whether or not this density difference  is in
itself sufficient to cause the observed factor of 2 increase in activation
energy cannot be determined from available data.  However, if the observed
increase is the result of a density effect then reagent density would become
an important reaction design parameter for systems utilizing stirred tank
reactors.

             Results from RTU Processing with Low Iron Reagent

     Two groups of experiments were performed in the RTU with low iron reagent:
(1) shakedown experimentation with coal 1 using pure water feed reagent and
(2) processing of coal 3 with acidified reagent during the RTU operational
phase (Experiment 03).  Both groups of experiments were performed utilizing
all five stages of R-l rather than the three stages used in Experiment 01.
Thus, longer reaction times were obtained for similar slurry feed rates.
     Operating parameters for low iron reagent experimentation are listed in
Table 10.  Shakedown experiments 8 and 9 were performed with pure water feed
and coal derived iron and sulfate.  Experiment 03 was performed with a nominal
1% w/w H2S04 feed solution.   However, due to inadequate reagent mixing .in
tank T-8, acid segregation in the reagent solution resulted in a 2% w/w acid
feed reagent for experimental condition 03-03.  The difference between 1% and
2% starting reagent acid concentration is not significant since acid addition
serves only to prevent iron oxide deposition from the reagent and possible sub-
sequent sulfate occlusion on the coal or blinding of the pyrite surfaces.  The
pH of these solutions ranges from 1.0 to 1.2 and either pH is sufficiently low
to suppress precipitation of iron oxide from the reagent.
     A summary of processed coal analyses from low iron reagent processing is
presented in Table 11.  Replicate shakedown experiments 8 and 9 show an S
reduction of 57* from 0.68% w/w to approximately 0.27% w/w during 265°F proc-
essing.  The sulfur forms analyses problem observed during Experiment 8
(i.e., the high SQ and low Sp) is most likely attributable to inefficient
pyrite extraction; the expected SQ based on ash removal is 0.63% w/w which
would indicate a true Sp value of 0.26% w/w.  Experimentation performed with
acidified reagent at 230°F indicates S  removals from 0.64% w/w to approxi-
mately 0.16% w/w or 75%.
                                      82

-------
                 TABLE  10.  OPERATING PARAMETERS FOR LOW  IRON  REAGENT LEACH EXPERIMENTATION
00
CO
Experiment
No.
8
(Shakedown)
9
(Shakedown)
03-01
03-02
03-03
Run time Reagent R-1 R-l 02 Coal Reagent 02 Recirc. Residence time*
(day/hour) H2S04 TemP Pressure Pressure Feed Feed Feed Rates T-2 R-l
Concentration Rate Rate Rate
Start Stop (* w/w) (°F) (psig) (psia) (Ib/hr) (gpm) (SCFH) (gpm) (hrs) (hrs)
238/12:00
249/11:15
6/17:04
9/12:00
10/17:07
238/22:41 0
249/22:31 0
7/03:00 1
10/03:20 1
11/07:28 2
265 50 26 198 1.0 94 23** 1.9 9.5
265 50 26 198 1.0 103 16** 1.9 9.5
222 31 28 298 1.7 94 20 1.1 5.7
232 41 34 199 1.2 88 20 1.6 8.2
234 41 33 199 1.2 150 20 1.6 8.2
           * Slurry residence time is computed as the ratio of reactor volume-to-volumetric slurry flow rate.

          ** Redrculation piping used during shakedown processing was 1.25  inch sch.  80 pipe while that used during Phase III
             processing was 0.75 inch sch. 40 pipe.

-------
                      TABLE 11.   SUMMARY OF PROCESSED COAL ANALYSES OBTAINED  DURING
                                 RTU EXPERIMENTATION WITH LOW IRON REAGENT
00
Exp. No. No. of* Processing
Samples Temp.,
(°F)
Starting 3P; 14F
Coal 1
8 IP; 5F 265
(Shakedown}
9 IP; 5F 265
(Shakedown)
Starting 10P; 22F
Coal 3
03-01** 4P; 8F 222
03-02** 4P; 11F 232
03-03** 4P; 9F 234
Coal Composition, % w/w
Ash
16.11
±0.261
13.83
12.33
14.32
±0.469
12.97
±0.894
12.02
±0.636
12.51
±0.953
Heat
Content,
Btu/lb
12508
± 77
12985
13213
12907
± 58
13258
± 162
13388
± 91
13265
± 155
Total
Sulfur,
St
1.73
±0.045
1.03
1.07
1.49
±0.081
0.68
±0.044
0.78
±0.065
0.75
±0.033
(Except Heat Content), Dry Basis
Pyri ti c
Sulfur,
SP
0.68
±0.071
0.17?
±0.061
0.29
±0.021
0.64
±0.051
0.18
±0.064
0.17
±0.042
0.14
±0.031
Sulfate
Sulfur,
Ss
0.43
±0.071
0.14
±0.025
0.15
±0.007
0.31
±0.027
0.01
±0.010
0.01
±0.006
0.02
±0.007
Organic Iron,
Sulfur,
So Fe
0.62 1.33
±0.010
0.72? 0.88
0.63 0.71
0.54 1.02
±0.071
0.49 0.23
±0.029
0.60 0.26
±0.049
0.59 0.25
±0.060
**
        The number of short proximate analyses  (ash,  heat  content, total sulfur and coal iron) and sulfur
        forms analyses (sulfate and pyritic sulfur) will be  designated by xP and yF, respectively, where
        x and y are the number of analyses included in  the indicated average.  The number of coal samples
        analyzed is y minus x.

        These experiments were performed with 1  to 2% w/w  H2SO» reagent.

-------
    These results are  somewhat  surprising since higher S  removals are
expected at higher processing  temperatures for coals from the same mine having
nearly identical starting  S  contents  and processed under similar oxygen par-
tial pressures and throughputs.   The observed difference could be due to sev-
eral things, including:   (1) differences between the pyrite particle sizes
and distributions in  coal  Nos,  1 and 3;  (2) differences in oxygen availability
and (3) the effect of reagent  acidification.   That coal Nos. 1 and 3 differ
with respect to  pyrite  particle  size and distribution appears feasible since
they are known to differ  in  other respects (i.e. , ash and organic sulfur con-
tents).  The higher ash coal No. 1 may contain pyrite which is difficult to
remove owing to  its low specific surface area.  Oxygen availability is impor-
tant to pyrite leach  rates obtained during low iron reagent processing.
Because the bulk reagent  contains very little iron, continuous and efficient
regeneration of  reagent ferric ion is  necessary to maintain high pyrite leach
rates.  The recirculation/oxygenation  loop used during shakedown was 1.25-inch
schedule 80 pipe while  that  used during Phase III operation was 0.75-inch
schedule 40 pipe.  Thus,  flow  velocities in the regenerator during shakedown
were approximately 40%  of those  obtained during Phase III operation (the asso-
ciated  shakedown Reynolds  number was  approximately 66% of that in Phase III).
Since  gas-reagent mixing  efficiency, and therefore oxygenation, may relate to
recirculation  flow velocities, slurry  oxygenation obtained during shakedown
experimentation  may have  been  less efficient than that obtained during Phase III
operation.
     Thus, the specific cause  of differences in pyrite leach rates obtained
during  low iron  processing of  coal Nos.  1 and 3 cannot be accurately defined
on the  basis of  available data.   Further experimentation, including leaching
of coal Nos. 1 and 3  with high iron reagent, would be required before confi-
dent conclusions can  be drawn.
     Differences in product  coal S contents indicated in Table 11 are due to
the fact that  samples from Experiments 8 and 9 were not washed in the labora-
tory prior to analysis.  However, bench scale experimentation performed to
date with acidified and non-acidified  water reagents has not indicated an
appreciable difference  in  processed coal S  contents.  It, therefore, appears
that further washing  would effect essentially complete sulfate removal from

                                      85

-------
Experiments 8 and 9 as was the case for Experiments 03-01 to 03-03.   Note that
all processed coals from Experiment 03-03 contain less than 1.2 pounds  S02
per 10  Btu.
     Data from low iron reagent processing may be compared with 4%  iron  reagent
processing data obtained at similar temperatures and 02 partial pressures.
Experiments 01-10 and 01-01 were performed at 270°F and 230 F, respectively,
with 02 pressures of approximately 23 psia.  A residence time of 5.1  hours at
elevated temperature and pressure was used for both Experiment 01 conditions.
A comparison of data from Experiment 01-10 (Table 9) and Experiments  8 and 9
indicates that similar AS  values were obtained from both groups of experiments
although only half as much residence time was required with iron reagent.
Thus, higher pyrite leach rates were obtained during processing with  iron
reagent.  On the other hand, extrapolation of the results from Experiment 01-01
(residence time of 5.1 hours) to the 5.7 and 8.2 hours of residence time uti-
lized in Experiments 03-01 through 03-03 yields essentially the same  results,
which were obtained in Experiment 03.  That is, assuming that the effective
pore reagent Y value for low iron reagent processing experiments was  similar
to that of Experiment 01-01, processed coal analyses from Experiment  03 appear
to indicate pyrite leach rates similar to those obtained during the higher
iron reagent Experiment 01-01.
     As mentioned previously, Coal No.  1 reacted at a lower rate than did
Coal No. 3 under low iron reagent processing conditions.   The comparison of
Experiments 8 and 9 with Experiment 01-10 indicates that Coal  No.  1 also
reacted slower than did Coal No. 2 which was processed with a higher  iron con-
tent reagent.  The comparison of Experiment 03 with Experiment 01-01  indicates
that Coal No. 3 processed with acidified low iron reagent reacted at  a rate
comparable to that obtained from higher iron reagent processing of Coal  No. 2.
The conclusion drawn from these observations depends on the assumed reaction
rate of Coal No. 2 relative to Coal Nos. 1 and 3.  Since all  three coals are
products of the same mine and physical  cleaning facility and were processed
at the same top-size, it seems reasonable to expect that these coals  would
react at similar rates under similar processing conditions.  If this  expecta-
tion is correct then the lower rates obtained during Experiments 8 and 9 are
most probably due to inefficient oxygenation.  Further, Experiments 01-01 and
03 indicate that reaction rates measured in the RTU were not affected by
reagent total iron content up to a concentration of 4.5% w/w.

                                      86

-------
    Coal processed in the RTU with acidified low iron reagent was found to
contain approximately 0.1% w/w less $s  after the control  lab washing stages
than did coals processed with 4%  iron reagent.   This observation is not con-
sidered to be reflective of differences in  process chemistry but, rather,
indicates the necessity of effective pore solution displacement during the
washing of coals processed with high iron reagent.
6.2.2.2  Reagent Regeneration Data--
    Each stage of the R-l reactor  is provided with a slurry recirculation/
oxygenation loop which promotes reagent regeneration concurrently with pyrite
leaching.  During RTU operation the oxygen  feed was proportionated among the
reaction stages according to estimated  regeneration requirements occurring at
each stage.  An excess of oxygen  was fed to each stage of the RTU which cor-
responded to between 4 and 7 times  the  actual stoichiometric requirement.
     Bench scale experimentation  has shown  that under such operation, the
reagent regeneration rate may be  expressed  as follows:
 where
     rR     is  the  moles  of ferric ion regenerated per unit time,
     P02    is  the  partial  pressure of oxygen in atmospheres,
     Fe+2   is  the  ferrous  ion concentration in moles per liter,
     KR     is  the  rate constant, a function of temperature only, in liters/
            mole-atm-unit time,
 and
  '                          KR = AR exp (-ER/RT)                          (4)
 with
     AR = 6.7 x 10  liters/mole-atm-min., and
     ER = 13.2 Kcal/mole.

                                      87

-------
     The bench scale determined ER value should apply to  RTU  regeneration since
reagent, unlike coal particles, is not prone to segregation and  ideal  flow
through the staged reactor system should be closely approximated.   On  the other
hand, values of AD computed from RTU data may differ from that determined at
                 K
bench scale since the extent or efficiency of regeneration may be  different.
Hence, the ER value of 13.2 Kcal/mole was assumed to be correct  for RTU  proc-
essing and the value of AD (RTU) was computed from process data.
                         K
     Computation of reagent regeneration rates in R-l requires data on inlet
and outlet reagent total iron concentrations and Y values, steady  state  Y data
                                                +2               '
for each reactor stage, (L pressure data, and Fe   generation rates.  Of these
                          +2
required data, only the Fe   generation rates are not directly measured  during
                                   +2
processing.  The major source of Fe   generation is the pyrite leach reaction
                                                                 +2
which was discussed in Section 6.2.2.1.  A secondary source of Fe    is reac-
tion between ferric ion and the coal matrix.  Although the extent  of this  oxi-
dation reaction is too small to be detected by heat content analysis of  the
                   +2
coal, sufficient Fe   is produced to warrant accounting during regeneration
                                     +2
rate computation.  The quantity of Fe   generated above that expected from
                                   +2
pyrite leaching is termed excess Fe  .
     Bench scale data from processing 14 mesh x 0 mine cleaned Martinka  coal
with 5% w/w iron reagent at 212°F are presented in Table  12 (Reference 8).
                                +2
These data indicate an excess Fe   generation rate of approximately 0.005
             +2
weights of Fe   per weight of coal per hour during the first 5 hours of  proc-
essing.  Subsequent processing resulted in little or no additional  excess  Fe+2
generation.  These data provided the basis for estimating the rate  of excess
  +2
Fe   generation during RTU coal processing.
     Equations (1) through (4) were used with steady state data  from Experi-
ment 01 to determine regeneration efficiency in the RTU.  Since  attainment of
steady state with respect to reagent Y  required long processing  times for con-
ditions of Experiment 01, not all conditions of this experiment  provided  true
steady state reagent Y data.  Conditions 02, 03, 06, and  10 all  provided  steady
state or near steady state Y data while conditions 01, 05, 07, 09,  and 11
yielded only transient data.  Experiments 01-04 and 01-08 were short term
tests involving pulse response,measurements only in the second stage of  reac-
tion and cannot be used for overall regeneration rate evaluation.

-------
    TABLE 12.   EXCESS Fe+2 GENERATION DATA  FROM PROCESSING 14 MESH x 0
               MINE CLEANED MARTINKA COAL WITH  5% w/w IRON REAGENT
=., 	 _____ 	 =
+2
Excess Fe Generation
Exp.
B.S.
B.S.
B.S.
No.
35 '
27
28
Temp.
(°F)
212
212
212
Process
Time
(hrs)
5
24
48
+2
weight Fe
weight coal
0.0258
0.0197
0.0294
+2
weight Fe
weight coal-hr
0.0052
0.0008
0.0006
     It should be noted that rapid  attainment of steady state processed coal
S  contents relative to times required  to  obtain steady state reagent Y values
is indicative of the fact that product  coal  analyses are rather insensitive to
small changes in reagent Y.  This is  particularly true when coal  S  contents
are reduced to 0.20% w/w or less.
     Examples of transient and near steady state reagent Y data are illus-
trated in Figure 20.  Reagent Y  data  from  Experiments 01-01 and 01-02 are
presented in the Figure as a function of process time.  Data from Experiment
01-01 indicate that the reagent  Y is  decreasing with time at a substantial rate
in all stages.  Increasing the partial  pressure of oxygen from 22 psia to
49 psia  (Experiment 01-02) is seen  to have essentially stabilized the reagent
Y value  in each stage.  Although data from Experiment 01-02 indicate that the
first and second stage Y values  may have still  been changing at the test con-
clusion, the rate of change appears to  be  rather small.  Note that the stage
two Y value should not exceed that  of stage three and that the first stage Y
                                                  +2
value should not exceed that of  stage two  since Fe   generation due to S
removal  decreases with each successive  stage.  Thus, the true steady state
Y values of the first two stages probably  do not differ from the measured
values by more than about 0.02 which  will  not substantially affect regenera-
tion rate estimation.
     Regeneration rate estimates computed  from Experiments 01-02, 01-03, 01-06,
and 01-10 are presented in Table 13.  Reaction temperature, oxygen partial
pressure and oxygen feed rate data  are  listed in columns two, three, and four.

                                     89

-------
    o
    o
VO
en

o


CM
 •
o
    00
    00
    CO

    o


    o
     •
    o
I-
5  vo

ii I   O
C£

    CM
     •
    O

    CO


    o
                 O  1ST STAGE R-l

                 0  2ND STAGE R-l

                 A  3RD STAGE R-l
EXP.
                               01-01
                                                                                  EXP. 01-02
   o
   to
                               22 psia
                                                                                  po.
                                                                   49  p
la
     17.00 18.00 19.00 20.00 21.00 22.00 23.00 24.00 01.00 02.00 03.00 04.00  05.00 06.00 07.00 08.00 09.00 10.00 11.00

                                                      CLOCK TIME (HOURS)
                       Figure  20.   Reagent Y Data From RTU Processing Experiments 01-01 And 01-02

-------
TABLE 13.  RTU REAGENT REGENERATION DATA
Exp. Temp. Pn
No. (°F) °2
(psia)

01-02 231 49


01-03 251 39


01-06 251 52


01-10 270 23



02 XS Fe2
Feed Generation
Rate / wt. Fe2 \
(SCFH) Vrft. coal-hr;

57 0.0
0.005 (bench scale)
0.010
54 0.0
0.005 (bench scale)
0.010
123 0.0
0.005 (bench scale)
0.010
56 0.0
0.005 (bench scale)
0.010
0.020
ARxlO"5
, liter _>
Stage
1
1.33
2.70
4.08
0.72
1.50
2.29
0.58
1.16
1.72
0.01
0.60
1.18
2.34
vmole-atm-min
Stage Stage
2
1.16
2.51
3.85
0.66
1.44
2.21
0.79
1.47
2.14
0.41
0.95
1.47
2.52
3
0.78
2.09
3.70
1.10
2.18
3.25
1.04
2.06
3.07
0.29
0.81
1.33
2.35
I
Average

1.09
2.43
3.88
0.83
1.71
2.58
0.80
1.56
2.31
0.24
0.79
1.33
2.40
% of
Bench
Scale
AR

16
36
57
12
25
39
12
23
35
4
12
20
36
Average KpxlO
, liter V ,
vmole-atm-min'

3.28
7.32
11.68
4.07
8.39
12.65
3.92
7.65
11.33
1.82
6.01
10.11
18.24

-------
                                                                           +2
The value of AR was estimated for each reaction stage based on  an  excess  Fe
generation rate of from zero to four times the value measured at bench  scale,
and an ER value of 13.2 Kcal/mole; the assumed excess Fe   generation and cor-
responding AR values are listed in columns five through nine.   The average AR
values obtained in the RTU are listed in column ten as a percentage of  the
bench scale AD value.  The average measured KR values from this testing are
             R                               "
listed for reference in the last column.
                                                                            +2
     Rate constant data presented in Table 13 for a fixed level of excess Fe
generation indicate little temperature dependence (Experiments  01-02, 01-03,
and 01-06) or show a rate decrease with increasing temperature  (Experiments
01-02, 01-03, and 01-06 compared to Experiment 01-10).  Since the temperature
effect on reagent regeneration is known with confidence, these  data must indi-
                   +?
cate that excess Fe   generation is not a constant, but rather  a function of
temperature which increases with increasing temperature.  As such, higher
         +2
excess Fe   generation rates and, therefore, higher KD values apply for higher
                                                      +2
processing temperatures.  If the bench scale excess Fe   generation rate is
assumed to be correct up to 230°F, then the excess rate must double with each
additional 20°F increment for the data to support an ED value of 13.2 Kcal/
                                                     o
mole.  Twice the bench scale excess must apply at 250 F and four times the
bench scale excess must apply at 270°F.  This observation is in agreement with
bench scale data indicating that excess ferric ion consumption  increases by at
least a factor of two between 212°F and 250°F when processing 14 mesh x 0 mine
cleaned Martinka coal ^   '.  Thus, these data appear to  indicate that regenera-
tion in the RTU proceeded with an AR (RTU) of approximately 2.4 x 105 liters/
mole-atm-min or at 37% of the rate obtained at bench scale.  This represents
an estimate of the minimum average regeneration rate obtained in the RTU since
excess Fe+  generation rates at 230°F may actually be higher than those meas-
ured at 212°F.
     These data also indicate that increasing the oxygen throughput from a
four-fold to a fourteen-fold stoichiometric excess did not improve regenera-
tion rates (compare Experiments 01-03 and 01-06).
     If indeed the regeneration efficiency in the RTU is 37% with respect to
bench scale regeneration, then a comparison of the two regeneration systems is
appropriate.  The primary parameters of concern are the oxygen-reagent  blending

                                     92

-------
path length, Reynolds number and gas-reagent  contact frequency (stage turn-

around time through the regenerator  loop).  These parameters compare as

follows:

       •   Both the RTU and bench scale  systems have gas-slurry mixing
           path lengths on the order of  10 feet.  Both systems also
           have bends up to about 90° at some point.  The bench unit,
           however, has a valved constriction in the mixing path which
           may double the fluid velocity at  the valve.  The RTU unit
           contains no flow constrictions.

       •   The RTU system circulates slurry  at a rate of 20 gpm through
           0.82-inch I.D. pipe while the bench unit circulates slurry at
           a rate of 1 gpm through  a 0.75-inch I.D. tube.  A ratio of the
           Reynolds numbers obtained in  the  RTU and bench unit mixing
           regions may be computed  to avoid  estimating viscosities of
           slurries at the reaction temperature.  The ratio Re(RTU)/
           Re(Bench) based on an estimated 50% constriction in the bench
           scale tubing is approximately nine.

       •   Each stage of the  RTU contained approximately 150 gallons of
           slurry which was circulated through the regeneration loop at
           a rate of 20 gpm giving  a slurry  turnaround time of approxi-
           mately 7.5 minutes.  The bench scale reactor generally con-
           tained about 1.6 gallons which was circulated at 1 gpm for a
           turnaround time of approximately  1.5 minutes.

     From  these facts, it appears that oxygen-slurry mixing was at least as
 vigorous  in the RTU  system as  it was in the bench scale apparatus.  Any addi-
 tional gas-slurry blending due to the restriction in the bench circulation
 line should be compensated for,  at  least  in part, by the higher overall
 Reynolds  number obtained in the  RTU  circulation lines.  Hence, the principal
 difference between the two systems  resides in their respective stage turnaround
 times through slurry circulation/oxygenation  loops; bench scale reactor turn-

 around times through the oxygenation loop were only one-fifth as long as RTU

 turnaround times.

     That  the regeneration efficiency should  relate to slurry turnaround time

 through the regeneration loop  appears reasonable if the oxygen bubbles coalesce
 and disentrain from  the slurry at a  reasonably rapid rate (on the order of
 minutes) after slurry reentry  into  the reactor.  The regeneration rate expres-

 sion (Equation 3) is expressed in terms of oxygen partial pressure but, more

 fundamentally, the regeneration  rate is proportional to the concentration of
                                      93

-------
dissolved oxygen.  The concentration of dissolved oxygen depends on  the  rate
of pyrite leaching and the rate of oxygen dissolution into the reagent.  The
latter depends on the partial  pressure of oxygen and the gas-reagent contact
area.  Since, under conditions of rapid oxygen disentrainment the gas-reagent
contact area outside the circulation loop decreases rapidly with time, regen-
eration efficiency would be expected to increase with decreasing stage turn-
around time through the recirculation loop.  Assuming that this hypothesis is
correct, it would appear that to increase regeneration rates in the RTU would
require an increase in the slurry recirculation rate (though not necessarily
the  Reynolds number) and/or the partial pressure of oxygen.
6.2.2.3  Elemental Sulfur Recovery-
     All coal samples drawn from the RTU were extracted with toluene to recover
product elemental sulfur (S ) prior to sulfur forms analyses.  The quantity of
S  generated during processing is determined by analyses of toluene extracts
from filter belt coal samples.  Additionally, S  recovery has been determined
for  a limited number of samples from the primary reactor.
     Results of S  extractions performed during RTU shakedown and Experiments
01 and 03 are presented in Table 14.  Included in this table are the quantity
of residue composed of sulfur and tar, sulfur content of the residue (based on
one  analysis per sample) and S  content of the coal based on extraction data.
Results from double toluene extractions are indicated by two tar and sulfur
content values.  Sample coal analysis data used in computation of the product
ratio S /S  are also presented.  The last column of the table lists the product
sulfur ratio Ss/Sn which is estimated as (AS  - S )/S .  These data were obtained
identically for all experimentation with the exception that twice as much
toluene extract was distilled to generate the sulfur-tar residue for Experi-
ment 03 than was used for the preceding experimentation.  Also, balance pre-
cision was increased from 0.005 gram to 0.00005 gram for Experiment 03.  The
resultant improved precision is evidenced by the relative S /S  data scatter
obtained before and after initiation of Experiment 03, namely ±1.66 compared
to ±0.71.  Thus, the substantial scatter in S /S  data obtained prior to
Experiment 03 is principally due to extraction residue weighing inaccuracies.
That these inaccuracies were essentially random is verified by the fact that
the mean Ss/Sn ratio from combined shakedown and Experiment 01 testing and from

                                      94

-------
                      TABLE 14.  ELEMENTAL  SULFUR  PRODUCT  GENERATION  DATA  FROM  RTU  PROCESSING
VO
en
Experiment # Sample ID
8 (Shakedown) SI 1900
SI 2000
SI 2200
9 (Shakedown) SI 1800
SI 2000
SI 2101
SI 2331
01-01 M7 0130
M7 2105
SI 2345
SI 0157
01-02 SI 0745
SI 0942
01-03 M7 2103
SI 2125
01-05 M7 1940
SI 1910
SI 1950
01-06 SI 1050
01-07 SI 0144
SI 0235
Sulfur + Tar
Residue
(% w/w)
0.67
0.92
0.76
0.76
0.88
0.97
0.73
0.41
0.32
0.52
0.31
0.55
0.55
0.78
0.65
0.75
0.45
0.52
0.32
0.17
0.15
Sulfur
Content
11.31
13.28
25.62
12.61
12.95
9.11
22.33
26.2
44.8
39.2
41.1
54.8
56.8
42.6
11.9
57.7
44.7
48.8
50.9
31.1
48.6
Elemental
Sulfur, S
(% w/w) "
0.08
0.12
0.19
0.10
0.11
0.09
0.16
0.11
0.14
0.20
0.13
0.30
0.31
0.33
0.08
0.43
0.20
0.25
0.16
0.05
0.07
Coal Analysis,
S°
0.68
0.68
0.68
0.68
0.68
0.68
0.68
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
0.58
Sl
0.21
0.17
0.07
0.32
0.32
0.23
0.33
0.24
0.21
0.23
0.29
0.33
0.22
0.30
0.17
0.13
0.17
0.16
0.13
0.20
0.33
% w/w
ASP
0.47
0.51
0.61
0.36
0.36
0.44
0.35
0.34
0.37
0.35
0.29
0.25
0.36
0.28
0.41
0.45
0.36
0.37
0.45
0.38
0.25
*
VSn
5.18
3.19
2.12
2.60
2.27
4.00
1.15
2.09
1.64
0.75
1.23
0
0.13
0
4.13
0.05
0.80
0.48
1.81
6.60
2.57
                                                    - Continued -

-------
                                              TABLE 14. (Continued)
CTv
Experiment 1 Sample ID
01-09 M8
SI
01-10 SI
01-11 SI
SI

03-01 SI
SI
si
03-02 SI
SI
SI
SI
03-03 ' SI
SI
SI
1445
1255
0108
0224
0301

2258
2356
0200
2213
0015
0119
0335
0431
0527
0727
Sulfur + Tar
Residue
(% w/w)




0.
0.
0.
0.
0.
0.
0.
0.
0.
0.33
0.36
0.38
0.23
0.35

47, 0
48, 0
60, 0
42, 0
0.45
45, 0
52, 0
50, 0
47, 0
38, 0




.23
.13
.22
.21
.21
.18
.14
.22
.31
Sulfur
Content
33
28
36
51
46

26.6
26.2
20.1
41.4
41
36.8
22.5
36.0
33.0
27.5
.0
.1
.2
.5
.9

, 0.8
, 3.7
, 1.1
, 6.3
.3
, 4.1
, 3.8
, 7.7
, 4.8
, 4.6
Elemental
Sulfur, S
(% w/w) n
0.11
0.10
0.14
0.13
0.16

0.13
0.13
0.13
0.19
0.19
0.18
0.13
0.19
0.16
0.12
Coal Analysis , 3
S°P
0.58
0.58
0.58
0.58
0.58

0.64
0.64
0.64
0.64
0.64
0.64
0.64
0.64
0.64
0.64
SJ
0.11
0.13
0.08
0.07
0.14
VSn
0.25
0.16
0.17
0.19
0.22
0.19
0.15
0.17
0.15
0.13
} W/W
ASP
0.47
0.45
0.50
0.51
0.44
Ave =
0.39
0.48
0.47
0.45
0.42
0.45
0.49
0.47
0.49
0.51
*
VSn
3.27
3.5
2.57
3.25
1.75
2.2 ± 1.66
2.00
2.69
2.61
1.37
1.21
1.50
2.77
1.47
2.06
3.25
                                                                                         VSn Ave * 2J * °'71
                                                                                     Ss/Sn Cum Ave = 2.2 t 1.45
The ratio Sg/Sn
                            is  computed  as  (AS  - Sn)/Sn,

-------
Experiment 03 are 2.2 ±1.66  and 2.1 ±0.71, respectively; these values are sta-
tistically indistinguishable.
    Collectively, the  data  presented in Table 14 indicate a product S /S
                                                             ^        s  n
ratio of 2.2 which may  be  compared to the previously determined S /S  ratio
      /c 11 \                                                      <;  n
of 1.5^  '.   The difference between these two values may be  the result of
incomplete $n recovery  from  the processed coal (leaving 0.04% S  on the coal
or losing it during  recovery would suffice for this experimentation), loss of
S  during the recovery  process (i.e., during vacuum drying), or a systematic
error  in any of the  numerous analyses and measurements used in computation of
the Ss/Sn ratio such as original and final S , residue weight, purity, etc.
Therefore, the  higher product Ss/$n ratio obtained from S  recovery data is
not considered  to be reflective of process chemistry but of the increasing
difficulty  in obtaining an S  balance from low S  coal processing experiments.
    A comparison of experiments performed with iron reagent (Experiment 01)
and experiments performed  with initially iron free reagent (Experiments 8, 9,
and 03)  indicates that  reagent iron concentration does not effect the product
s^/Sm  ratio.  This finding is in agreement with previously obtained bench
          (f\\
scale  data*  '.
     Most coal  samples  obtained during Experiment 03 were treated with two
stages of toluene extraction rather than the nominal single stage extraction.
Data from this  experiment  shows that 94 percent of the extracted $n was
recovered in  the first  stage of extraction.  Thus, approximately 0.01% w/w Sn
was recovered during the second toluene extraction of processed mine cleaned
Martinka coal.
6.2.2.4 Process Effects on Coal Heat Content-
     Mine cleaned Martinka feed coals and processed coals obtained from the
RTU may be  compared  on  the basis of mineral matter free (mmf) heat content
to determine  effects of the Meyers Process on the coal matrix.  An mmf heat
content was  computed based on starting coal analyses and average processed
coal analyses for each  experiment using the Parr formula (ASTM designation
D388-66).   Results of these computations are presented in Table 15. - These
data indicate a mean heat  content increase of 352 Btu per pound based on dry
coal analyses;  this  corresponds to a 2.7% increase in coal heating value.  On
                                      97

-------
                         TABLE  15.   SUMMARY  OF  HEAT  CONTENT  CHANGES  IN  RTU  PROCESSED COALS
•£>
CO
Experiment
No.
8 (Shakedown)
9 (Shakedown)
01-01
01-02
01-03
01-05
01-06
01-07
01-09
01-10
oi-n
03-01
03-02
03-03
Average Heat
Content Change
Heat Content,
Dry Basis, Btu/lb
Initial Final
12508
12508
12923
12923
12923
12923
12923
12923
12923
12923
12923
12907
12907
12907

12985
13213
13397
13300
13268
12888
13163
13123
13487
13158
13071
13258
13388
13265

Changes
Btu/lb Percent
477
705
474
377
345
-35
240
200
566
235
148
351
481
358
352
±188
3.8
5.6
3.7
2.9
2.7
-0.3
1.9
1.5
4.4
1.8
1 .1
2.7
3.7
2.8
2.7
±1.5
Heat Content, Dry Mineral
Matter Free Basis, Btu/lb
Initial Final
15172
15172
15283
15283
15283
15283
15283
15283
15283
15283
15283
15293
15293
15293

15279
15260
15459
15508
15396
15185
15264
15337
15255
15426
15367
15420
15393
15342

Changes
Btu/lb Percent
107
88
176
225
113
-98
-19
54
-28
143
84
127
100
49
80
±85
0.7
0.6
1.1
1 .5
0.7
-0.6
-0.1
0.3
0.2
0.9
0.5
0.8
0.7
0.3
0.5
±0.5

-------
a dry mmf basis a mean  heat  content increase of 80 Btu per pound is indicated
(0.5% of the mmf heat content).   The observed rnrnf heat content increase is not
believed to be meaningful  in the sense of coal upgrading but, rather, is con-
sidered to reflect  inaccuracies  associated with analyses and with the Parr
formula itself.  However,  on the basis of these data, it appears reasonable
to conclude  that no measurable degradation of the coal matrix occurred
during  RTU L-R  processing at temperatures of 230°F and 270°F with reagent
iron concentrations of 0-4.5% w/w.
     These data  are in  agreement with bench scale results indicating that the
Meyers  Process  has  little or no effect on the organic matrix of Appalachian
bituminous coals.   However,  bench scale L-R experimentation performed with
14 mesh x 0  mine cleaned Martinka coal resulted in an mmf heat content loss
of approximately 2% during 250 F processing for six hours or more.  Thus,
while no matrix oxidation was observed during L-R processing in the RTU, slight
matrix oxidation was observed during bench scale L-R operation.
     This discrepancy may be associated with coal weathering occurring during
the 8-month  period  between obtaining a coal sample for bench scale processing
and initiation  of RTU coal processing.  The unweathered coal (as indicated by
the lack of  sulfate sulfur)  utilized during bench scale processing had an mmf
heat content of 15,448 Btu per pound while that of the RTU feed coal was
 15,249 Btu per  pound (average of all three RTU feed coals).  The 239 Btu per
pound difference between bench scale and RTU feed coals represents 1.5% of the
unweathered  coals'  mmf heat content.  Thus, it is feasible that 1-2% of the
coal organic matrix was oxidized during weathering and that this portion of the
matrix is  readily oxidizable by any oxygen bearing process.  Data from the RTU
would then  indicate that no measurable coal oxidation can be further incurred
during short periods of time under moderate conditions.  It should be noted
that if the  preceding assumptions are correct, oxidation of 1-2% of the mmf
coal is inevitable  and will  occur either during coal storage at the respective
point of utilization or during coal processing.
6.2.3  Slurry Sampling
     Slurry  samples were drawn from each stage of T-2 and R-l during RTU
operation  for the purpose of obtaining reagent Y data and coal analyses.
Sampling ports  in the RTU are located at two positions:  (1) in the vessel
wall below the  slurry level  and (2) in the slurry recirculation loop (R-l
only).   Wall  samples are taken by allowing slurry to pass through a ball
                                    99

-------
valve into an evacuated sample bomb.  One bomb is used to clear out stagnant
coal which may accumulate in the sampling lines.  The contents  of this  bomb
are discarded and a second sample is immediately drawn for analysis.  Loop
samples are obtained from R-l stages by full stream diversion of circulating
slurry through a sampling tube which can be isolated and subsequently drained
of slurry.  Each sample drawn contained approximately one liter of slurry.
     Coal analysis data obtained from slurry samples have been  evaluated in
conjunction with reagent Y data and the reaction model presented  in Sec-
tion 6.2.2.  This treatment is limited to results from Experiment  01 since a
pyrite leach rate expression is required.  A plot of the difference between
computed and analyzed S  values (AS ) versus sample location is presented in
Figure 21.  The sampling locations are spaced along the abscissa according to
the theoretical time required to reach steady state compositions at each point.
Data are presented at two times for most stages since two slightly different
slurry residence times were utilized during Experiment 01.   Data from wall
samples and recirculation loop samples are indicated in the Figure by open
and closed symbols, respectively.
     Data from the three stages of mixing (T2-1, 2, and 3)  indicate that coal
samples from the first two stages contain significantly more pyrite than com-
puted, while stage three samples are lower in pyrite than computed.  Samples
from the first two stages of T-2 were found to consistently contain more
pyrite than the feed coal (1.4 ±0.66% w/w compared to a starting coal S  con-
tent of 0.58% w/w).  Stage three mixer samples can be shown to be non-
representative and low in S  by using measured ferrous iron generation data.
Measured S  values from R-l wall samples are also seen to be low with respect
to computed S  values although the AS  approximated analytical precision after
the first stage of pressurized processing.  The limited data from R-l loop
samples are generally consistent with wall sampling data although high meas-
ured pyrite values were obtained from Rl-3 samples at 270°F,
     These data suggest major segregation of the feed coal  prior to reagent
reflux in the third stage of mixing.  Bench scale experimentation performed
to determine causes of coal slurry foaming has indicated that reagent reflux
is key to effective coal wetting at atmospheric pressure ^ ^.   Hence, high
pyrite coal appears to segregate in the lower portion of T-2 (the vicinity

                                     100

-------
o
LU
M
1 ?

. 1
On


Or
Oc
0 4
0.3
•0.2
•0.1
0
0.1
0.2
0.3
0.4
o
a
a

A
o

o
A
0 °
A
o
A





O
a
o













8
£
A *
A @



o , * - 230°F TEST SA

MPLES (WALL, LOOP)
A, A - 250°F TEST SAMPLES (WALL, LOOP)
o,« - 270°F TEST SAMPLES (WALL, LOOP)







1








1
i
i


















A
A
1




(100) (200) (300) (400 MINUTES)
T2-1 T2-2 T2-3 Rl-3 R1-* Rl-5
                                                SAMPLE LOCATION (TIME)
                                Figure 21.  Summary of Estimated Sampling Bias

-------
of the sampling ports) prior to effective coal wetting.  The  reverse  segrega-
tion effect (drawing low pyrite samples rather than high pyrite  samples)
observed in the third stage of mixing is not understood, but  may be associated
with a froth flotation effect caused by reagent boiling and the  injection of
steam.  This phenomena is also observed in R-l where both oxygen and  steam are
injected into the slurry.
     The apparent reduction of sampling bias to the level of  analytical pre-
cision during the last two stages of R-l processing probably  relates  to the
fact that the S  contents of all size-gravity coal fractions  are very low.
That is, non-representative sampling of a depyritized coal is not likely to
yield samples having substantially different S  contents from the mean S
content of the whole coal.
     Coal segregation indicated by AS  data is verified by slurry solids con-
tent data which are summarized in Figure 22.  Slurry solids data from each
sampling location are presented in the Figure as a function of AS .   The nom-
inal solids content of the feed slurry was 33% for all conditions of  Experi-
ment 01 and dilution down to approximately 29% was effected by the addition of
live steam for heating.  Slurry solids data indicate that samples which were
high in S  (i.e., samples from T2-1, T2-2, and the Rl-3 loop) also tended to
be solids rich with respect to the feed slurry.  On the other hand, samples
which were low in S  tended to have reasonable or slightly low slurry solids
contents.  Thus, as would be expected, whether or not a slurry sample contains
coal which is representative of the whole coal is reflected in the solids con-
tent of the sample.  If the sample slurry solids content is not  representative
of the system as a whole, the coal contained in the sample probably is not
representative of the whole coal either.
     Despite the fact that slurry samples drawn from the RTU  were biased with
respect to S  and solids contents, reasonably good sampling precision was
obtained.  Replicate sampling generally yielded S  agreement  to  within
±0.07% w/w, Ss agreement to within ±0.03% w/w and slurry solids  agreement to
within ±3%.  Such precision coupled with sampling bias data would tend to
indicate that only specific slurry streams or eddys are accessible through
specific sampling ports and that the composition of the accessible slurry did
not change significantly with time.  Thus, improvement of slurry sampling most

                                     102

-------
O
CO
0.2
o
-0 2
n A
3
3
** -n fi
•
e?
M _n 8
>- u-°
_i
<
— -10
CL
' -1 2
0
LJ
^ 11
=s -1 .*
<_)
_J
^ -1 6
Q.

-1 ft
-2 Q



U










AA
V


.








I
X A 9b
^^j
A^










L
i . '
D-



i—
LU
1-
O
0
OL
in
UJ


A
&
D

+ *•
+
+
+






ZERO SAMP












LING BIAS

X^ X
y
X















.







X




• - T2-1
+ - T2-2
A - T2-3
O - Rl-3
n - Rl-4
O - Rl-5
X - Rl-3
A - Rl-4
* - Rl-5










(WALL)
(WALL)
(WALL)
(LOOP)
(LOOP)
(LOOP)

                             10
30         40        50


       SLURRY SOLIDS, PERCENT
                                                                             60
70
80
                                       Figure 22.  Slurry Concentration  Variation as  a
                                                   Function of Sample  Location

-------
likely requires more efficient slurry agitation within the various  stages  and/
or relocation of the sampling ports away from localized flow patterns  (i.e.,
wall effects).
6.2.4  RTU Mass Balance Data
     Flow rates of RTU process streams were monitored wherever possible t,o
enable mass balance determinations to be made around each of the major unit
operations, namely coal-reagent mixing, pyrite reaction and slurry  filtration.
Primary flow measurement equipment utilized in the RTU to facilitate plant
operation and provide mass balance data were:  (1) a weigh belt for metering
ground coal to the system, (2) magnetic flow meters for measuring liquid and
slurry flow rates, (3) rotameters for measuring steam, oxygen and wash water
flow rates and (4) target meters for measuring oxygen feed and exhaust flow
rates.  Output from the primary flow measurement equipment was generally
dependent upon the input of one or more auxiliary devices which provide such
data as stream temperature, pressure, gas purity, etc.  The majority of this
data was recorded at 3-minute intervals by the Doric data logger although
some data were recorded manually (i.e., steam and wash water feed rates) at
intervals commensurate with the data reduction needs.
     With the exception of the magnetic flow meters, all  primary flow measure-
ment devices used in the RTU measured mass flow directly.   The weight belt
meters the coal feed utilizing a load cell which measures the weight of coal
per unit area of belt.  Output from the weigh belt is transmitted as total  coal
weight and also as a feed rate.  Calibration can be obtained through direct
weighing of the belt effluent.  Rotameters and target meters are essentially
momentum measurement devices which may be calibrated at either standard con-
ditions or operating conditions; the former requires the application of tem-
perature and pressure corrections.  However, magnetic flow meters are true
velocity meters and, as such, yield accurate volumetric flow rates.  In order
to convert from the magnetic flow meter volumetric flow to mass flow, a gen-
eral slurry density expression was formulated in terms of temperature, total
iron sulfate concentration, acid concentration and slurry solids content.
Considering the liquid reagent components as ideal solutions and neglecting
the specific volume of dissolved salt, the slurry density was expressed as:
                                     104

-------
  p  "  riQO -  CA + 3.578 TFe)  ^     A   1 |\  .  » ~nM',Z   ZTT
      L      62.248            113366 J [l  + 0-0003(T - 77) j T 84(l . Q>Qls)

where
                                       3
    p     is the slurry  density,  Ib/ft ,
    A     is the sulfuric  acid concentration of the reagent, wt.  %,
    TFe   is the total iron  concentration  of the reagent, wt. %,
    S     is the slurry  solids content, wt.  %, and
    T     is the slurry  temperature,  °F.
    The  densities  of  water and sulfuric acid at 77°F (25°C) were  taken to
be 62.25  and  113.97 Ib/ft3, respectively.   A value of 84 Ib/ft3 was used for
the density of the  coal.
    This expression yields density estimates which agree with measured values
to within about 5%  or  less  for  the solutions used in the RTU.  Hence, the pre-
cision  of mass flow rates obtained from magnetic flow meters is approximately
5% rather than the  0.5-1% precision associated with volumetric flow readings.
    Mass flow rates which  are  of importance to mass balance efforts, but are
not measurable by primary flow  measurement  equipment, include the  steam flash
rate from R-l pressure let-down, slurry accumulation or depletion  in T-2 and
R-l, and  the  flow rate of wet product  coal  cake from the S-l filter.   These
quantities were either inferred from the other data normally obtained from
the RTU or obtained by special  measurement  during specified experimentation.
The quantity  of steam  flashed during R-l pressure let-down can be  estimated
from the  increase in the  temperature of vent gas scrubber water in T-4.   Thus,
assuming  perfect scrubber efficiency,  the mass flow rate of flash  steam was
calculated from the flow  rate of scrubber water and its temperature change
resulting from steam condensation.   Slurry  accumulation or depletion in T-2
and R-l is computed from  slurry level  data  generated by sensors located in
the last  stage of each vessel.   Special  measurements are required  to obtain
the mass  flow rate  of  wet coal  cake from S-l.  This involves collecting the
S-l solid effluent  in  a   tared dumpster and determining  weight increases  as a
                                     105

-------
function of time.  This procedure is both time consuming and  cumbersome  and
was performed only for special mass balance experimentation.  Thus,  mass  bal-
ances around S-l were not generally performed.
     Computer programs which are presented in Volume II, Appendix A  were
written to perform mass balances around T-2 and R-l.  These programs account
for all process streams and utilize both Doric data and manually recorded
data.  Data from Experiment 10 (a special shakedown mass balance test),
Experiment 01 and Experiment 03 were evaluated with the aid of these programs
to identify inconsistencies among the RTU flow measurement outputs.
     A low temperature and pressure test (Experiment No. 10) was performed
during RTU shakedown to obtain a complete system mass balance and verify the
major flow measurement devices.  This experiment was performed with  a water-
coal slurry using flow rates .of 3 gpm and 500 Ib. of coal per hour.  These
flow rates correspond to residence times of approximately 0.6 hour in the
mixer and 3.2 hours in the reactor (all five stages of R-l were utilized).
Mixer heating was utilized to maintain the slurry at 190°F (sufficiently hot
for proper P-l operation but below the solution boiling point).  Slurry tem-
peratures in R-l were allowed to drop in order to minimize evaporative losses
during pressure let-down.  The plant was operated for 10 hours under the nom-
inal conditions with the last 3 hours being considered to be at steady state.
     The solids balance between the A-3 plant coal feed and S-l filter belt
output indicates that 160 Ib. of dry coal was "lost".  This represents 11% of
the feed coal.  However, filter belt flooding conditions prevailed during the
entire steady state operational period and solids removed during the belt wash
were discarded to T-9 rather than being returned to the filter cake  as a fil-
ter belt wash water slurry per normal operation.  Post-run observation indi-
cated that a "large build-up of coal" was found in the filter belt drip pan.
Thus, the 160 Ib. coal "loss" appears to be a reasonable and explainable
occurrence.
     The overall liquid balance included units A-3, T-2, R-l, T-5 and S-l.
The vent steam could not be computed since the AT of T-4 scrubber water was
not measured.  The vent steam from T-2 (operating at 190°F) was assumed to
be negligible while the steam from R-l (operating at ^160°^) was computed to
to be negligible on the basis of inlet and outlet reagent temperature values
                                    106

-------
from the knockout drum V-l.  This  balance indicated an 830 Ib. liquid gain
(17% of input liquids).  This  result may be partially or totally due to the
following:  (1) error in reagent  filtrate flow rates, which, for this experi-
ment, were both FE-157 and  FE-158, or (2) leakage from water flushed pump and
mixer seals.  The possibility  of  incorrect reagent feed flow rates has been
discarded since a good mass balance was obtained around the 1-2 mixer.  These
data combine to yield the total system mass balance which includes units A-3,
T-2, R-l, T-5 and S-l.  The net result is a 662 Ib. gain which represents 11%
of the  input material.
    A  mass balance  around  T-2 based on coal, reagent and steam feed rates and
slurry  flow calibrations of FE-29 (slurry out) resulted in a total mass balance
of better than  93%.  This result  appeared to be reasonable since the accuracy
of each of the  two magnetic flow  meters is estimated to be approximately 5%.
A heat  balance  around T-2 essentially verified the validity of FI-16/18 steam
flow indicators.  A  mass balance  around the combined units T-2 and R-l indi-
cated  a loss of 1755 Ib.  (28%  of  the input material).  This result was found
to be  caused by improper operation of the R-l outlet slurry flow meter (FE-60).
Subsequent experimentation  revealed that signal response times of some mag-
netic  flow meter  components were  sufficiently slow compared to slurry pulse
flow times  (pulse lengths are  generally less than two seconds) that entire
pulses  were sometimes not detected.  Attempts were made to resolve this
response problem  throughout Experiment 01.  However, satisfactory performance
of FE-60 was never obtained during pulse flow operation and the flow meter
was ultimately  removed  from service.
     Data from  the Experiment  10  mass balance test gave rise to the following
observations:
        t   Stream flow measurement equipment associated with T-2 was
           operational and accurate.  Subsequent experimentation indi-
           cated that  steam flow meters FI-16 through 18 were somewhat
           undersized which necessitated frequent use of flow meter
           bypass lines to maintain desired T-2 temperatures.  Thus,
           under normal operating conditions, steam feed rates to T-2
           were  not accurately  known at all times.
        •   Magnetic flow meters  could not be utilized under conditions
           of  pulse flow.   Thus, the quantity of slurry leaving R-l and
           entering S-l could not be directly measured with available
                                     107

-------
            flow measurement equipment.   Although the mass flow rate of
            slurry from R-l can in principle be estimated from the
            increase in temperature of T-4 scrubber water, adequate
            monitoring of scrubber water temperatures was never achieved
            during RTU operation.   As such, mass balances could not be
            performed on either R-l or S-l with available data.
        •   An accurate accounting of system solids cannot be made unless
            the S-l filter belt is operating properly.  Filter belt opera-
            tion under flooding or near flooding conditions necessitates
            discard of the belt wash solution and its entrained coal
            component.
     Experimentation performed in  the RTU did provide meaningful  mass balance
data relating to the T-2 mixing unit.  Results of the T-2 mass balances from
RTU experimentation are presented  in Table 16.  The first two columns of the
table list the experiment number and day.  The second two columns list the
run times included in the mass balance.   These times do not necessarily reflect
actual experimental run times since periods of time which included data noise
on one or more of the required data channels have been eliminated.  The remain-
ing columns in Table 16 list mass  balance data for each experiment.   The accu-
mulation column lists material "gains" or "losses" associated with slurry level
changes within T-2.  Slurry level  data from Experiment 01 was recorded man-
ually and, therefore, in limited quantity; available data is insufficient to
permit computation of material accumulation.  During Experiment 03 the slurry
level of T-2 was recorded by the Doric Data Logger and included in the mass
balance.  The accumulation term is generally seen to be small as  would be
expected since every effort was made to maintain a constant slurry level
throughout each experiment.
     Data presented in Table 16 indicate that a good material balance was
obtained for the T-2 unit.  An average recovery of 100% was obtained and most
experiments balanced to within 7%.  Thus, the precision obtained in mass bal-
ancing the T-2 unit is comparable  to that of the flow meters monitoring the
liquid flow streams.  The trend from slightly low material recovery to slightly
high material recovery during the  period of RTU operation may relate to instru-
ment calibration drift or an interaction between magnetic flow meter electrodes
and the leach reagent during the course of experimentation (i.e., scaling or
corrosion).  However, iron concentration, in this case, should have no effect
on the magnetic flow meter calibration; that this is correct is evidenced by

                                     108

-------
                            TABLE  16.  T-2 MASS BALANCE  DATA FROM RTU PROCESSING EXPERIMENTATION
o
<£>
Experiment
No.
10 (Shakedown)
01-01
01-02
01-03

01-04
01-06
01-05


01-07

01-08
01-09
01-10
01-11
03-01
03-02


03-02
Day
250
299
300
305

305
306
314


319
321
322
322
322
325
6
9


10
Mass Balance Time
Start
20:59
16:36
03:03
10:02
16:26
21:32
00:02
12:31
15:52
17:22
22:18
17:00
02:51
04:51
13:51
17:31
17:40
12:02
20:05
21:38
17:10
Stop
23:56
02:57
09:00
15:29
21:29
23:29
10:38
13:48
16:46
01:13
00:48
02:18
04:48
13:48
01:48
02:58
01:58
19:14
21:14
03:23
07:28
Coal
1468
3080
1770
1621
1506
580
3139
388
272
2327
745
2718
580
2663
3557
2820
2469
1430
228
1145
2852
Material
Reagent
4317
6195
3565
3089
2871
1108
5976
732
519
4519
1394
4974
1056
4797
6366
7739
7651
4174
666
3304
8313
In, Ibs.
Steam
150
288
165
151
140
54
293
40
27
238
70
129
27
124
106
131
569
500
80
399
1012

Total
5935
9563
5500
4861
4517
1742
9408
1160
818
7084
2209
7821
1663
7584
10029
10690
10689
6104
974
4848
12177
Material Accum. Total
out, Ibs out, Ibs
5609 -74 5535
9369
5185
4656
4252
1781
ti_
9369
5185
4656
4252
1781
j
9307 a 9307
1019 d 1019
807 3 807
6892 fe 6892
2215 < 2215
7840 § 7840
1663
7741
10492
8535
1663
7741
10492
8535
11267 279 11546
6533 -64 6469
1123 -20 1103
5160 55 5215
13350 62 13412
Sain,
Ibs
-400
-194
-315
-205
-265
39
-101
-141
- 11
-192
6
19
0
157
463
-2155
857
365
129
367
1235
Percent
Recovery
93
98
94
96
94
102
99
88
99
97
100
100
100
102
105
80
108
106
113
107
110
               Material accumulated in the mixer.

-------
the fact that Experiment 10 was performed with pure water and  still  indicates
a slightly low material recovery as do nearly all early experiments.
     An oxygen mass balance around R-l for an actual coal processing  experi-
ment requires the following data:  (1) oxygen flow rates into  R-l,  (2)  V-l
vent gas flow rates, (3) the oxygen content of V-l vent gas, (4)  Fe   genera-
tion rates from both pyrite reaction and matrix oxidation,  (5) starting and
        • O
final Fe   concentrations for each reaction stage, and (6)  inlet  and  outlet
reagent Fe   concentrations.  Since the oxygen mass balance is reflective of
both the instrument and analytical precisions as well as either inferred or
extrapolated data (i.e., excess Fe+2 generation rates), an accurate oxygen
balance of a coal processing experiment is not readily obtained.  However, an
effort was made to perform oxygen balances for all experimentation performed
in the RTU.
     Estimates of the average oxygen consumption rates during processing were
                                               +2
made based on pyrite removal data and excess Fe   generation rates presented
in Section 6.2.2 and on reagent Y data.  These values are compared to oxygen
consumption values indicated from oxygen flow data and R-l pressure data in
Table 17.   Experimental conditions and run times are reiterated in the  first
five columns of the table.  Values of total input and vent oxygen obtained by
integrating the output of FE-61 and FE-44 are listed in columns six and seven.
Quantities of oxygen accumulated or depleted in R-l as indicated  by starting
and final  reactor pressures are listed in column eight.  The difference between
oxygen input (column six) and the sum of vent oxygen and oxygen accumulated in
the reactor (columns seven and eight) is the oxygen consumption indicated by
RTU flow and pressure equipment and is listed in column nine.  The average
measured oxygen consumption rate for each experiment (listed in column  ten)
may be compared to the average oxygen consumption rates computed  from AS  and
         +2                                                             P
excess Fe   data (listed in column eleven) to indicate the validity of  acquired
mass balance data.
     Data presented in Table 17 show little correlation between measured and
computed oxygen consumption values.  The observed discrepancies do not  appear
to be associated with the estimated excess Fe+2 generation rates  since  esti-
mation errors of this type should result in comparable differences between
measured and computed oxygen consumptions for experiments performed at  similar

                                     110

-------
TABLE 17.  OXYGEN MASS BALANCE DATA OBTAINED DURING RTU  EXPERIMENTATION
Experiment
No.
01-01
01-02
01-03
01-04
01-05
01-06
01-07
01-08
01-09
01-10
01-11
03-01
03-02
03-03
Temp.
°F
230
231
251
251
250
251
245
245
270
270
251
222
232
234
V
psia
22
49
39
39
25-39
52
10-53
50
49
23
52
28
34
33
0? Feed
Rate
SCFH
56
57
54
59
NA
123
59
57
56
56
128
94
88
150
Run
Time,
Mrs
9.7
5.9
11.5
1.9
10.2
10.7
9.5
1.9
8.5
12.3
11.5
10.1
15.3
14.3
Oxygen, SCF Average Q£
Rate,
In
541
395
21
113
1318
557
109
479
693
1478
909
1200
2254
Out
543
128
178
28
911
172
9
0
594
1133
953
1085
3189
Accum
27
2
11
-2
-3
36
-3
7
3
-4
-42
5
9
Consumed
-29
265
432
87
DATA NOT AVAILABLE
410
349
103
472
96
349
-2
no
-944
Measured
-3
38
38
43
38
37
51
55
8
30
0
7
-66
Consumption
SCFH
Computed
9
9
18
18
23
14
14
27
20
21
15
15
22

-------
temperatures.  Results from Experiments 01-01 and 01-02, 01-09  and  01-10,  and
03-02 and 03-03 do not indicate that this is the case.
     Oxygen accumulation quantities are not sufficiently large  to have  a sig-
nificant impact on the measured consumption rates and therefore do  not  repre-
sent a major source of error.  Oxygen feed rates measured with  FE-61  should
be accurate since meter calibration was performed under conditions  identical
to those used during plant operation and rough flow rate verification is pro-
vided by downstream rotameters used to proportion oxygen among  the  reactor
stages (Fl-62 through Fl-66).  Thus, the most probable source of the observed
discrepancy appears to be the vent oxygen flow rate measurement.
     The rate at which oxygen is vented from V-l is computed from the vent gas
oxygen content indicated by AE-171 and flow rate indicated by FE-44.  The
oxygen analyzer provides continuous analysis of a dried slipstream  of the V-l
discharge gas.  Analyses from AE-171 cannot be wholly responsible for observed
discrepancies since to bring Experiments 01-02 through 01-09 and 01-11 into
agreement with computed consumption rates would require stream  oxygen purities
greater than 100%.  However, there are reasons why FE-44 may be singly respon-
sible for observed data discrepancies:
        •   Water saturated vent gas was cooled or in the process of cool-
            ing when its flow rate was being measured.  Thus, entrained
            water droplets or condensate may have contacted the flow tar-
            get and thereby altered the meter performance.   This possi-
            bility can be eliminated during future experimentation  by
            locating the flow meter in a heated region with a large cross
            section to prevent condensation and to decrease the likeli-
            hood of entrainment.
        •   Owing to the pressure regulator used in the RTU, reactor
            venting was often pulsed rather than continuous.  Pulses
            occurring during venting were up to approximately 9 minutes
            in duration and had amplitudes ranging to nearly 400 SCFH.
            Since the data logger was recording at 3-minute intervals,
            true pulse shapes are not known and the possibility exists
            that entire pulses were not detected.  It should be noted
            however that venting during Experiments 01-03, 01-09, 01-10,
            01-11 and 03-03 was reasonably smooth and pulse free.   Yet,
            data from these experiments do not correspond to computed
            oxygen consumption estimates.
     It therefore appears that an additional design effort is required in
conjunction with further elevated temperature processing experimentation

                                     112

-------
performed  in  the absence of coal to resolve the problems associated with

obtaining  a balance of oxygen in the RTU.  Additional bench scale tests are
                                 +2
also required to define excess Fe   generation rates at temperatures above
212°F.

6.3  DATA  ANALYSIS CONCLUSIONS

        1. Three distinct coal samples  (coal Nos. 1, 2 and 3) were sup-
           plied by American Electric and Power for coal processing
           experimentation in the RTU.  These coals differed with respect
           to coal ash, heat content, total sulfur, and sulfur forms
           analyses.

        2. RTU processing times required to attain steady state operation
           with respect to reagent Y can be substantially larger than
           would be estimated by this ratio of reactor volume-to-volu-
           metric slurry flow rate.  However, since product coal Sp con-
            tents are relatively insensitive to small changes in reagent Y
            (particularly when Sp is reduced to 0.2% w/w or less), a plug
            flow type estimate of time required to achieve steady state
           appears to be effective with respect to product coal Sp
            contents.

        3. All RTU processing experiments performed with 14 mesh top-size
           mine cleaned Martinka coal and acidified iron reagent at
            temperatures above 230°F yielded product coals having Sp con-
            tents of 0.13% w/w to 0.19% w/w after nominal residence times
           of 5 to 6 hours.  Similar processing at 230°F yielded product
           coals with Sp contents of 0.24% w/w.

        4.  RTU coal processing experiments performed with acidified iron
            sulfate reagent reduced coal sulfate sulfur from 0.34% w/w to
           approximately 0.12% w/w after application of a preliminary
            filter belt wash followed by the standard laboratory washing
           and work-up.  Coal processing with acidified iron free reagent
            reduced coal sulfate sulfur from 0.31% w/w to approximately
           0.02% w/w after a filter belt wash and standard laboratory
           work-up.

        5. Little or no excess ash removal was observed during RTU proc-
           essing of mine cleaned Martinka coal.  Ash depletion from the
           coal generally corresponded directly to the quantity of pyrite
           and sulfate removed during processing.

        6. Coal processing data obtained from the RTU was correctable
           with the previously determined leach rate expression, namely:


                                        2   2
                               PL - KL Y2 Wp2
                                     113

-------
     where

          K     is  the  leach reaction rate constant

          Y     is  the  ratio of reagent ferric ion-to-total iron
                concentration,  and

          W     is  the  weight percent of pyrite in the coal.


 7.   Coal  desulfurization  rates in  the RTU exhibited an Arrhenius
     type  exponential temperature dependence as did previous bench
     scale experimentation.   That is,
                      KL =  AL exp [-EL/RT]



     where

          A.     is the pre-exponential  factor or frequency factor

          E.     is the activation energy

          R     is the gas  constant,  and

          T     is the reaction temperature.

     However, rate constants K|_ observed during RTU processing were
     3 to 8 times greater than those  obtained during bench-scale
     experimentation.   This is attributed in  part to non-ideal
     mixing of the reaction system which is a series of continu-
     ous flow stirred  tank  reactors.

 8.   Desulfurization rates  obtained during coal processing with
     acidified water reagent appear to  be essentially the same
     as those obtained by processing  with acidified iron sulfate
     reagent.  Thus, reagent iron concentration does not affect
     pyrite leach rates with acidified  reagents.

 9.   Mine cleaned Marti nka  coal can be  processed through the RTU
     to yield a product coal having less than 0.6 pound of sulfur
     per 10b Btu.

10.   Reagent regeneration in the RTU  appeared to be adequately
     described by the  previously determined regeneration rate
     expression:
                              \
                               114

-------
    where

          KR      is an Arrhenius type rate constant

          PQ2     is the oxygen partial pressure
            +2
          Fe      is the concentration of ferrous ion in reagent.

     The reagent regeneration rate constant obtained in the RTU
     appeared to be at least 37% of the rate constant obtained at
     bench-scale.   This possible disparity may be the result of
     relatively rapid disentrainment of oxygen bubbles from the
     slurry and the fact that the regenerator volume (slurry recir-
     culation loop volume) per unit volume slurry is lower in the
     RTU than in the bench-scale unit.

11.  Reagent regeneration rates could not be improved by increas-
     ing the oxygen throughput from a fourfold to a fourteenfold
     stoichiometric excess.

12.  Excess ferric ion consumption appears to be a temperature
     dependent phenomenon.  The degree to which this dependence is
     quantified will determine the precision with which reagent
     regeneration rates can be measured during L-R processing of
     low pyrite coals (i.e., the mine cleaned Martinka coal).

13.  An average dry nmf heat content increase of 80 Btu per pound
     (±85 Btu per pound) was obtained during L-R coal processing
     of 14 mesh top-size mine cleaned Martinka coal in the RTU at
     temperatures of 230° to 270°F with 0 to 4.5% w/w iron reagent.
     Thus, no coal matrix degradation was observed during RTU coal
     processing.

14.  RTU coal processing experimentation performed with acidified
     reagents having 0 to 4.5% w/w total iron concentrations indi-
     cated a product sulfate sulfur-to-elemental sulfur ratio of
     2.2.  Because of inefficiencies associated with the recovery
     of small quantities of elemental sulfur (less than 0.3% w/w)
     and physical  losses which may be incurred, this data is con-
     sidered to be essentially in agreement with previous bench-
     scale data indicating a 1.5 product sulfur ratio.  Thus,
     reagent total iron concentration did not appear to affect
     the ratio of sulfur products during RTU operation.

15.  Slurry sampling during RTU operation was generally of high
     precision but was not representative of the bulk slurry.
     It, therefore, appears that meaningful slurry sampling will
     require more efficient slurry agitation and/or relocation of
     the sampling ports away from areas subject to localized flow
     patterns, such as wall effects and eddys.

16.  Magnetic flow meters are not suitable for measurement of
     slurry flow rates under pulse flow conditions encountered

                              115

-------
     in the R-l  effluent.   To obtain a material  balance of the
     R-l unit will  require an alternate flow measurement device
     with faster response  times  or an estimation of slurry effluent
     from the quantity  of  steam  flashed during R-l pressure let-down.

17.   An excellent material  balance was obtained  for the mixer 1-2
     with an average  material  recovery of 100% being obtained.
     Material balance precision  was approximately 7% which is
     comparable  to  the  precision of flow meters  used to monitor
     the liquid  flow  streams.

18.   The acquisition  of an oxygen mass balance for the  R-l unit
     appears to  require modification of the  vent gas flow measure-
     ment system.
                              116

-------
                  7.   SUPPORTING BENCH-SCALE EXPERIMENTATION

7.1  INTRODUCTION
     Supporting bench-scale experimentation has been conducted concurrent with
the RTU program to expand coal cleaning technology utilizing the Meyers Process
and to assist on-line RTU operations by supplying technical processing data to
assist in the overall program.
     This section principally contains experimental work involving the use of
the Meyers Process reagent for gravity separation of coal prior to chemical
cleaning as well as a review of related studies performed earlier, in the U.S.S.R.
     Included in the section are experimental studies examining gravity separa-
tion of coal under equilibrium and non-equilibrium conditions and their atten-
dant effect upon recovery of clean coal fractions.
     Additional experimental studies are presented herein which include:  (1)
chemical processing (Meyers Process) of sink fractions of gravity separated
coal; (2) comparisons with studies made in the U.S.S.R.; and (3) general exper-
imentation to investigate processing changes, weathering of coal, density
determination of coal slurries, and elemental sulfur recovery by acetone.
7.2  GRAVICHEM SEPARATION
7.2.1  Background
     A practical method for actual float-sink cleaning of coal in a dense liquid
has long been sought as an alternative to mechanical cleaning.  Heavy liquids,
such as zinc chloride-water, or chlorinated, brominated, or fluorinated hydro-
carbons are useful for prediction of yields which can theoretically be obtained
in mechanical washing plants, but are impractical for actual production because
they are expensive and add pollutants to the coal, atmosphere or water table.
We find that the density of aqueous ferric sulfate leach solution, as utilized
in the Meyers Process, is ideal for accomplishing a practical gravity separation

                                     117

-------
of coal for specific gravities between 1.2 and 1.4.  For example,  for  about  40%
of Appalachian coal, the float coal (which is often 40-60% by weight of  the
total) averages 0.8-1.2 Ibs S02/106 Btu and needs no further processing  to meet
standards, while the sink coal slurry contains most of the coal  pyrite.  This
sink slurry can then be processed through the Meyers Process to  produce  coal
containing 1.2-1.5 S02/106 Btu,  This approach allows production of coal which
will meet SO  emissions standards for both new and existing stationary
            J\
sources, and reduces processing costs relative to the straight through Meyers
Process by allowing 40-60% of the coal to bypass reactor, elemental sulfur
extraction and dryer units.  This float-sink approach was tested many years ago
by A.Z. Yurovskii in the U.S.S.R. but only published in the open literature very
recently^2).  Yurovskii subsequently treated the sink coal with a mixture of
nitric acid and ferric sulfate to remove pyritic sulfur, not realizing that the
separation medium itself was sufficiently active to accomplish near total pyritic
sulfur removal.
7.2.1.1  Literature Background--
     The cited work of Yurovskii reports laboratory (bench-scale) work for the
removal of iron pyrite (FeS2) from coal  by reacting it with ferric sulfate solu-
tions under a variety of conditions, and scale-up studies at pilot plant level
which demonstrates ferric sulfate solution gravity separation using centrifugal
separators with simultaneous sulfur and ash reduction.   Principally, he inves-
tigated the effects of temperature, multiple ferric sulfate treatments and ash
content on the degree of pyrite removal.  His data support the observations
from Gravichem processing (Sections 7.2.2 and 7.2.3) that significant reduction
in ash content and pyrite is observed upon preliminary or sustained treatment
of coal with ferric sulfate solution.  Yurovskii, however, only  reported 43%
removal as his highest value for pyrite depletion in his bench-scale studies
after twofold one-hour desulfurizations, which is low when compared to the
Meyers Process.  Yurovskii's work is discussed in more detail in Section 7.2.2
and 7.2.3.
7.2.1.2  Process Application--
     The Gravichem variation of the Meyers Process involves utilization of
ferric sulfate leach solution to gravity separate input coal into a very low
                                     118

-------
pyrite float material  and  a high pyrite sink material.  The sink material is
subsequently heated  to reaction temperature and processed through the Meyers
Process.  The Gravichem process modification is shown schematically in Fig-
ure 23.  Run-of-mine coal, often containing 5-10 Ibs S02/MM Btu, is crushed and
screened to remove rock and easily removed pyrite.  This rough-cleaned coal is
mixed with leach  solution  and pumped to a tank, where coal lighter than the
selected specific gravity  of the leach solution floats and the heavier coal
sinks.  The float fraction is filtered and washed but not dried, while the sink
fraction is treated  by the Meyers Process to give a dry coal product.  The
processed-sink  fraction may be hot-briquetted without binder (see Figure 24)
to give a  near  0% moisture product suitable for shipping in an open hopper-car.
For stoker-boiler (industrial-type boiler) applications, it may be recombined
with the wet  float coal filter cake (about 20% moisture) to give a combined
product of a  shippable 1"  x 0 grind with a combined moisture content of 10%.
7.2.2  Equilibrium Gravichem Separation
     A series of float-sink separations of Martinka mine coal (14 mesh x 0) in
ferric sulfate leach solution of varying specific gravity were performed
 (Table 18) which demonstrate the feasibility and the physical and chemical basis
of the Gravichem concept.
     These tests were of an equilibrium type in which coal was mixed with hot
 leach solution (80°C) in an agitator for 15 minutes to ensure wetting, then
 transferred  to a float sink apparatus maintained at 80°C (Figure 25) and allowed
 to separate  without  stirring.  After the specified times, the float product was
 skimmed from the top of the vessel and the sink product was separated from the
 remaining  leach solution.   The results (Table 18) show that float coal is
 obtained  in  23-48%  w/w yield, in each case meeting the NSPS standard of 1.2 Ibs
 StyiO6 Btu  - i.e.,  0.94-1.16 Ibs S02/106 Btu, while the sink fraction treated
 by the Meyers Process gave a product of about 1.2-1.3 Ibs S02/10  Btu.
     The  first separation  (Experiment 1 of Table 18) was performed utilizing a
 standard mixed organic solvent of specific gravity 1.30 for purposes of com-
 parison with leach  solution gravity separations (Experiments 2-7).  It can be
 seen that  the yield  of leach-solution float at 1.28 specific gravity was equiv-
alent to  that obtained with the organic solvent, indicating that the flotation
utilizing  ferric sulfate leach solution is essentially a physical separation
                                      119

-------
ro
o
                                  CRUSH/SCREEN
                                                   GRAVI-
                                                   SEPARATOR
                                                   1.3 SP.  GR.
                                                     LEACH
                                                   .SOLUTION,
                                                                             FILTER/ WASH
                                                                            MEYERS PROCESS
                                                                                             CaSO.
                           ROCK
                                              Figure  23.   Gravichem Process

-------
ro
                  MEYERS   PROCESSED   COAL  VIHICH   HAS
                  BEEN    BRIQUETTED  WITHOUT   BINDER
                                     Figure  24.  Hot-Briquetted Coal  (1" -Size)

-------
   TABLE 18.   EQUILIBRIUM  GRAVICHEM TESTS ON  14 MESH  x  0 MARTINKA MINE COAL WITH  VARIATION
                OF MEDIUM AND SEPARATION  TIME
Fraction, % w/w
of Coal***
Sp Gr
1. 1.30*
2. 1.28**
3. 1.28**
4. 1.28**
!\3 5. 1.33**
6. 1.33**
7. 1.43**

*Mixture of
Separation
Time, hrs
0.25
3
16
46
3
16
3

Float
27.6
23.0
31.0
24.0
33.0
40.0
48.0

Sink
72.4
77.0
69.0
76.0
67.0
60.0
52.0

toluene and perchloroethylene, wt
**Aqueous ferric sulfate at
80°C with
specific
Total Sulfur;
Float
0.75
0.80
0.72
0.69
0.77
0.76
0.83
0.80
ratio 1:1
gravity
Sink
1.89
1.45
1.39
--
1.56
1.56
2.11
2.26
.35.
obtained as
, % w/w Heat
Processed
Sink-* Float
14711
0.79 15057
0.86 14628
14829
14604
14748
14336
14413

Fe+++
Fe2(S(
(wt 5
follows: 1.28 = 7.5
1.33 = 5.0
1.43 = 5.0
Content, Btu/lb
Sink
12041
12917
12810
--
12665
12435
12108
11956

as
I) (wt %
4
21
30
Processed
Sink+
—
13105
12907
--
--
--
—
--

H20
) (wt %)
88.5
74
65
Sulfur Content, Ibs S/106 Btu
Float
0.51
0.53
0.49
0.47
0.53
0.52
0.58
0.56


Processed
Sink Sink*
1.57
1.12 0.60
1.09 0.67
-.
1.23
1.25
1.74
1.89


fHeated at 100°C for 24 hrs in aqueous ferric sulfate separation media, washed with water, then extracted with toluene and  dried.
  *Fiitered from separation medium, washed of leached solution treated and dried to constant weight in a vacuum oven at 100°C.

-------
                 LEACH
         COAL  SOLUTION
 AGITATOR
   SETTLER -(4 INCH DIAMETER X
           24 INCH LENGTH)
                                                    FLOAT
                                                    ZONE 1
                                                    ZONE 2
                                                 -»- SINK
Figure  25.   Bench-Scale Gravichem Test  Apparatus
             Flow  Diagram
                         123

-------
equivalent to that obtained with organic solvents of the same density.  However,
the sulfur content of the Gravichem float products, at 1.28 specific gravity,
decreases with increasing time of gravity separation (Experiments 2-4) indicat-
ing that leaching of pyritic sulfur is taking place, as expected, during the
gravity separation.  Similarly, the sink fraction is undergoing significant
pyrite leaching - e.g., a sulfur content of 1.89% w/w .is seen for the sink coal
obtained from organic solvent but a sulfur content of 1.39-1.45% w/w is seen
for the Gravichem sink.  The Gravichem sink material has greatly enhanced heat
content relative to the organic solvent sink coal because of dissolution of
coal mineral matter in addition to pyrite (Experiments 2-6).
     Increasing the specific gravity of the iron sulfate leach solution gives
increased yields of float coal with only slightly higher sulfur content and
slightly less heat content (Experiments 5-7).  No definite trend is seen for
the relative amounts of float material as separation is varied from 3 to 46 hours
(Experiments 2-4), while a possible trend is seen at 1.33 specific gravity
between 3 hours and 16 hours.
     Coal from the Kentucky No. 9 Mine, supplied by the TVA Cumberland Power
Plant, was likewise processed by the Gravichem method.   Ground coal at 3/8-inch
topsize was mixed with iron sulfate leach solutions of 1.28-1.34 specific grav-
ity (Figure 26), heated to 80°C and allowed to gravity separate in a holding
tank to give a 38-52% w/w yield of float product, after removal of residual
iron sulfate leach solution.  The float product (Table 19) from all of the sep-
arations is a power plant fuel containing 3.09-3.50 Ibs S02/106 Btu, having a
heat content range between 13774 and 14354 Btu/lb.  The sink fractions
(48-62% w/w) contain most of the coal pyrite.
     The sink coal was size-reduced while still in leach solution to a 14 mesh x
0 coal/Ieach solution slurry (Table 20), then treated at 102°C according to
Meyers Process procedures.  The product contained less than 4 Ibs S0?/106 Btu.
Thus, both float and processed sink coals meet the Tennessee State Implementa-
tion Standard requirements of 4 Ibs S02/106 Btu.  A substantial decrease in
coal pyrite actually has occurred.
     Gravity separation and chemical processing of Kentucky Mine No. 9 coal
produced a significant reduction in ash and total sulfur and an increase in
heat content as noted in Table 19.  A more extensive analysis summary is
                                      124

-------
         3/8 INCH X 0   LEACH
                COAL  SOLUTION
         MIX
         TANK
ro
en
                                             FLOAT
                                                       FILTERED, WASHED
                                                   GRAVITY
                                                   SEPARATION
                                                   TANK
                                            AGITATOR
                                            TANK
3/8" X 0 FLOAT PRODUCT
                                                                            MEYERS PROCESS
                14 MESH X 0
                PROCESSED SINK PRODUCT
                                      Figure 26.  Gravichem  Processing of TVA Coal

-------
               TABLE  19.    EQUILIBRIUM GRAVICHEM TESTS AT 80°C  ON  3/8"  TOP-SIZE  KENTUCKY  NO.  9 MINE  COAL

                              WITH VARIATION OF MEDIUM
ro
o>
Sample
As Received





Fraction, %
w/w of Coal***
Exp
No.
1
2
3
4
5
6
7
*
**
Sp. Gr. Separation Float
Time, hrs.
1.30
1.28
1 .28
1 .28
1.33-1
1.30-1
1.30-1
Mixture
Aqueous
0.5
0.5
3.0
**
3.0
.34 3.0
.34 3.0
.34 3.0
of toluene and
ferric sulfate
40
35
38
41
45
48
52
Sink
60
65
62
59
54
52
48
Total Pyritic Ash
Sulfur Sulfur X w/w
ST sp
#1 4.40
12 4.49
13 3.99

Total Sulfur,
X w/w
2.13 12.79
1.81 12.79
*1.93 10.78

Heat
Content
Btu/lb.
12408
12414
12687

Heat Content,
Btu/lb
Sulfur Content
Ibs SO^/
106 Btu
7.09
7.24
6.29

Ash,
% w/w
Float Sink Processed Float Sink ; Processed Float Sink Processed
Sink* Sink+ Sink+
2.40 5.61
2.33 	 2
2.22 4.92
2.41 ,, 4.78 (2
(2.29)f
2.27 	 2
2.35
2.44 	 2
perchloroethylene, wt ratio 1:1.35.
at 80°C with specific gravity obtained
13889 11369
.42 14128 	
14354 12295
.54) 14046 . 12004
(14322)"
.57 13774 	
14205 	
.51 13960 	
- +++
Fe as
as follows: Fe2(SO^)3
(wt %)
	 3.
12827 3
	 3.
	 3.
(2.
12973 3.
	 3.
12335 4.
H2S04
(wt X)
93 17.50 	
48 	 9.51
19 14.83 	
22 , 14.78 	
43)*
38 	 10.40
52 	
13 	 13.04
H20
(wt X)



Sulfur Content,
Ibs S02/105 Btu
Float Sink
3
3
3
3
(3
3
3
.46 9.87
.30 	
.09 8.00
.30 j, 7.96
.20)*
.30 	
.31 	
3.50 	


Processed
Sink+
	
3.77
	
	
3.96
	
4.0

                                                                                 •>. 7.5        4       88.5


              ***  Filtered from separation medium, washed if leached solution treated and dried to constant weight in a vacuum over at 100°C.


                +  Processed in 1.3 S.G.  leach solution (above) for 48 hours at 102°C subsequent to size reduction in Waring blender for 5 minutes @ 15,000 rpm.


                I  Processed float.

-------
                  TABLE  20.  PARTICLE SIZE DISTRIBUTION OF
                             SIZE-REDUCED KENTUCKY NO. 9*
                             SINK COAL
Screen Size
14
35
48
100
150
200
Pan
Retained,
% w/w
1.75
3.73
9.25
23.72
14.10
12.92
34.32
                   *
                    3/8  inch x 0 sink coal in Gravichem
                    leach solution, size-reduced in a Waring
                    blender for 5 minutes at 15,000 rpm.

presented in Table  21  which, in addition to the above noted parameters,  also
contains sulfur forms  analyses, including elemental sulfur (Sn) on three float
samples and four  processed sink fractions.
     Referring to Table 21 the reduction in total sulfur is related to both
pyritic and sulfate sulfur removal during gravity separation, as was noted for
the Martinka coal  (above), e.g., compare Experiment 3 with the organic sink in
Experiment 1.  Additionally, it is seen that elemental sulfur is formed  during
the separation process.  In Experiment 3, correcting for approximately 0.10% w/w
elemental sulfur* on the  starting coal (Section 7.3.3) due to weathering, the
amount of S  formed is greater than would be expected (0.40% w/w ASp) based on
the amount of pyrite  (S ) removed.  Sp formed in Experiments 4 and 5 is  as
expected.  For AS  of  2.38 and 2.32% w/w during Gravichem processing, 0.95 and
0.93% w/w S  should be produced.  Based on the two analyses (1.0% w/w) and cor-
recting for 0.1%  w/w Sn on the starting coal, it is seen that these data are
 'Assumption  is that  sink fraction contents the majority of the elemental
 sulfur formed during  weathering.
                                      127

-------
                                TABLE 21.  GRAVICHEM  PROCESSED KENTUCKY MINE  NO.  9 COAL
ro
oo
Exp
No.
Coal Gravi ty
Fraction
%
Control Total coal
100
Sample
(As Received)
1*
2
3

4



5

6
7

Organic Float**
Organic Sink**
Organic Float**
Organic Sink***
Gravi chem Float**
Gravi chem Sink**
Gravi chem Float**
Gravi chem Float***
Gravi chem Sink**
Gravi chem Sink***
Gravi chem Float**
Gravi chem Sink***
Gravi chem Float**
Gravi chem Sink***
Gravi chem Float**
Gravi chem Sink**
Gravi chem Sink***
40
60
35
65
38
62
41
41
59
59
46
54
48
52
52
48
48
Ash
% w/w
12.12
±1.16
3.93
17.50
3.48
9.51
3.19
14.83
3.22
2.43
14.78
—
3.38
10.40
3.52
11.11
4.13
16.20
13.04
Heat
Content,
Btu/lb
12503
± 159
13889
11369
14128
12827
14354
12295
14046
14322
12004
—
13774
12973
14205
12627
13960
11767
12335
Total
Sulfur,
% w/w
4.27
±0.22
2.40
5.61
2.33
2.42
2.22
4.92
2.41
2.29
4.92
2.54
2.27
2.57
2.35
2.73
2.44
4.99
2.51
Pyri ti c
Sulfur,
% w/w
1.96
±0.13
0.54
2.59
0.33
0.11
0.39
2.30
0.38
0.12
2.45
0.21
0.52
0.27
0.35
0.48
0.50
2.20
0.33
Sulfate
Sulfur,
% w/w
0.38
±0.15
0.18
0.92
0.02
0.33
0.01
0.10
0.05
0.20
0.26
0.37
0.06
0.41
0.01
0.39
0.04
0.34
0.31
Organic
Sulfur,
% w/w
1.92
±0.20
1.68
2.10
1.98
1.98
1.82
2.52
1.98
1.96
2.21
1.97
1.69
1.89
1.99
1.86
1.91
2.45
1.89
Elemental
Sulfur,
% w/w



.03
.02
.35



1.0

1.0

0.05
1.40
           *  Refer to Table 5-3.
          **  Gravity Separation only.
         ***  Gravichem processed.
            +  Elemental sulfur removed  with  toluene  before coal analysis

-------
in agreement with postulated Meyers  Process chemistry.  Experiment 7, however,
indicates a much higher formation  of Sn (1.30% w/w)* than would be expected from
the ASn (2.26% w/w), i.e., 0.90% w/w S  would be expected based on 0.4% w/w AS
     p                                "                                       p"
    With respect to the formation of Sn during oxidation of pyrite by ferric
ion, Yurovskii reports that  it  is  only formed at low Fe2(S04)3 concentrations.
Many of his experiments were conducted using 16% Fe2(S04)3 or 4.48% Fe+++, which
he considered high, and the  presence of Sn was not confirmed.  The experiments
on Kentucky Mine No. 9 coal, shown in Table 21  were conducted in leach solution
containing 7.5% Fe   .  Thus, Yurovskii's error in not observing S  leads to an
erroneous conclusion with  respect  to the efficiency of Fe2(S04)3 as an oxidizing
agent for removal of pyrite  from coal.  He based his conclusions for determining
the degree of pyritic sulfur reduction on total sulfur (Eschka method) differ-
ence before and after processing.
    Therefore  the  elemental sulfur, which was undoubtedly formed, increased
the  total  sulfur of the coal and was assumed to be unreacted pyrite, thus lower-
ing  the  observed efficiency  of  the reaction.  Table 22 is a summary of
Yurovskii's work on three  coals of varying sulfur ash content.  Also shown in
the  Table  are data  obtained  on  Kentucky Mine No. 9 coal for various reaction
times  at 102°C.  Referring to  Table 22, the degree of oxidation of coal pyrite
(FeS2),  as  reported by  Yurovskii,  based on total sulfur reduction, has been cor-
rected for  Sn,  i.e., the  theoretical amount of Sn formed  (0.4% w/w AS ) has
also been  converted to  pyrite  loss and reported in the table as corrected data.
Corrections were made as  follows:

                     Yurovskii assumed S^

where
         =  original coal  total  sulfur, %,
         =  treated  coal  total  sulfur, %,
         =  pyritic  sulfur reduction, %,
 Corrected  for  initial  sulfur on coal.
                                      129

-------
                      TABLE 22.   SULFUR REMOVAL FROM COAL  WITH  FERRIC  SULFATE  SOLUTIONS
Ash
Coal Content Treatment
of Coal Time
% w/w hrs.
Lutuginugol 14.0 1
1
1
2
7.1 2
Gorlovskugol 8.3 2
4.2 2
o Krasnodonugol 21.0 2
9.0 2
Kentucky No. 9 15.5 2
(1.3 sink) 4
8
14
32
48
Temp
°C
80
90
100
100
100
100
100
100
100
102
.102
102
102
102
102
Solution
Fe
Content
%
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
7.5
7.5
7.5
7.5
7.. 5
7.5
Sulfur Content of Coal , % w/w
Original Treated
Total
S
4.47
4.47
4.47
4.47
5.11
3.60
4.23
4.20
5.10
4.92
4.92
4.92
4.92
4.92
4.92
Pyri ti c
S
2.97
2.97
2.97
2.97
3.38
2.34
2.74
2.72
3.37
2.45
2.45
2.45
2.45
2.45
2.45
Total
S
4.03
3.70
3.61
3.45
3.80
2.63
2.73
3.90
3.80
3.16
2.91
2.65
2.43
2.45
2.54
Pyri ti c
S
2.53
2.20
2.11
1.96
2.07
1.33
1.24
2.42
2.07
1.28
1.01
0.62
0.37
0.25
0.21
Degree of Oxidation
of Coal FeS2, % w/w
Yurovskii
*
15
26
29
34
40
43
54
11
40






Corrected
**
25
43
48
57
67
72
90
18
67






Mpyers
+









48
59
75
85
90
91
*   Yurovskii's data as presented.
**  Yurovskii's data corrected for  elemental sulfur.
f  Experimental Meyers Process.

-------
However, from theory (above)  sti  - (Stf - Sn) = ASp (Assume Sn =0.4 ASp)

                          sti - Stf » ASp - 0.4 ASp

                                Stl - Stf
                          ASP — o —

                                  AS
Degree of oxidation, % = original Pyritic S. %
     It is also seen in Table 22  that Kentucky Mine No. 9 sink coal has a com-
parable, although lower, degree of oxidation of coal pyrite (compare with 2  hours
@ 100°C - Lutuginugol, 14% ash).  Comparing the reaction rate constant  (KL)  in
 Figure 27 of Kentucky Mine No. 9 coal (KL = 0.14 hours"1 W ~l) with Yurovskii's
 Lutuginugol coal (0.21 hours ~1 W "-1), it is seen that the pyrite oxidation rate
 is 67% of the Lutuginugol coal KL>  Y = 0.96 was estimated for Yurovskii's work
 since he used 12 grams of coal in 100 ml of 16% Fe2(S04)3 leach  solution.
     The effect of ash content on coal pyrite oxidation rate was noted  by
 Yurovskii.  Plotting the corrected data from Table 22 (2 hours 0 100°C) for
 each coal  (degree of oxidation versus ash content),  it is seen in Figure 28
 that a correlation does exist.  It may be noted that Kentucky Mine No.  9 coal
 supports the correlation.  Based on these data, the  KL for Kentucky Mine No.  9
 coal (15.5% ash) should be lower than 14% Lutuginugol, although  not as  low as
 was observed.
     A final comparison of the work of Yurovskii and data obtained on Kentucky
 Mine No. 9 coal is seen in Table 23 .  As was noted earlier for Kentucky Mine
 No. 9 coal, Yurovskii also observed a large reduction in ash content of the coal
 fraction separated by his chemicogravitational process.  It may  be noted that
 a comparable degree of cleaning is noted for the two methods although Kentucky
 No. 9 was cleaned at a larger  particle size  (0-9 mm).
 7.2.3  Non-Equilibrium Separation
     Following  the equilibrium Gravichem separation  studies  (Section 7.2.2)  it
 was desirable to investigate non-equilibrium Gravichem float-sink  separation
 which would more nearly be the case in a process plant.
                                      131

-------
to
ro
                                                                      KENTUCKY MINE NO. 9
                               48%*LUTUGINUGOL
                             /K  =0.21
                               50%* (2.4 HOURS)
                                                                       •REDUCTION IN
                                                                        PYRITE
                                               15         20         25

                                                 PROCESS TIME, HOURS
                          Figure 27.  Puritic Sulfur Leaching Data for  1.3  Specific Gravity
                                      Sink  Coal  at 102°C (Kentucky Mine No.  9)

-------
IUU
# 90
tb
^-*
X
O 80
u
UJ
E 70
t
£ 60
u
u_
O 50
O
i *
X
O 30
u.
O
g 20
0
UJ
Q 10
n
\.

X
• X

% CQX = 103-3.88 ASH (%)
CORRELATION COEFFICIENT





• YUROVSKII DATA
0 KENTUCKY MINE NO. 9

V
>*. •

\
N









\
• S

                     ASH, %
Figure 28.   Coal  Oxidation After  Two Hours at 100°C
            as  a  Function of Ash  Content
                        133

-------
                  TABLE 23.  COMPARATIVE RESULTS OF CLEANING
Coal
Size,
Method mm
Chemicograv itational 0-3
(Yurovskii)
Gravichem 0-9
(Kentucky No. 9)

Original Coal,
% w/w


Ash
25.79

12.12
12.12
12.12

Total
Sulfur
5.82

4.27
4.27
4.27
Float, % w/w


Yield
60

52
48
46


Ash
4.50

4.13
3.52
3.38


Sulfur
2.17

2.44
2.35
2.27
     Sampling ports were added to the bench scale Gravichem separator (Fig-
ure 25) and the float fraction was skimmed off the top of the leach solution
as before, while coal particles and leach solution in zones 1 and 2 were removed
through valves and finally the sink fraction was obtained by separating from
the filtrate.  This allows isolation of coal "in transit" for a first prelimi-
nary look at the behavior of float and sink particles in a dynamic situation.
     Two series of non-equilibrium Gravichem float/sink separations were con-
ducted with Martinka coal at 80°C using 1.2 and 1.3 specific gravity (S.G.)
leach solutions.  Other experimental parameters include:
        •   Weight of coal:  900 g (nominal).
        t   Weight of leach solution:  3600 g (nominal).
        §   Coal size:  14 mesh.
     The results for 1.3 S.G. leach solution (Table 24) show that the samples
taken were indeed non-equilibrium, in that times greater than 2 hours would be
required for the slurry to have attained equilibrium.
     It may be noted that the combined float and zone 1 fractions (1.3 S.G.)
are representative of the float portion taken in equilibrium Gravichem separa-
tion (Btu, ash, total sulfur, pyritic sulfur,  etc.) as shown in Table 18.
Likewise, the combined zone 2 and sink portions are comparable to equilibrium
sink percent recovery (coal % w/w) in Table 18;  however, the zone 2 fraction
in Table 24 is more nearly like the combined float and zone 1 fractions in the
same study.  Thus, non-equilibrium Gravichem shows a greater yield of clean

                                     134

-------
                 TABLE 24.  ANALYSES OF FRACTIONS OF MARTINKA COAL OBTAINED FROM
                            NONEQUILIBRIUM  FLOAT/SINK EXPERIMENTS
Fraction
Float


Zone 1


Zone 2


Sink


Total Coal
(Weighted
Average)
Starting
Coal
*
Trace;
*+
Time S.G. Coal
Hrs. of Leach % w/w
Solution
0.5
1.0
2.0
0.5
1.0
2.0
0.5
1.0
2.0
0.5
1.0
2.0
0.5
1.0
2.0
...
values for S
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
—
were
10.0
14.0
14.0
21.3
21.0
23.9
30.3
23.0
17.0
38.4
42.0
46.0

— -
< 0.01; 0.01
Sulfur Analysis - !
St
0.79
0.87
1.01
0.91
0.79
0.81
0.96
0.92
0.84
1.76
1.72
1.66
1.24
1.24
1.24
1.61
was used
SP
0.13
0.13
0.18
0.16
0.12
0.14
0.26
0.27
0.16
1.15
0.98
0.92
0.57
0.54
0.51
0.72
forS0
Ss
Tr*
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
Tr
0.18
I w/w
So
0.65
0.73
0.82
0.74
0.66
0.66
0.69
0.65
0.67
0.60
0.73
0.73
0.66
0.69
0.72
0.71
calculation
Heat
Content
Btu/lb.
14851
14650
14454
14604
14642
14635
14069
14105
14136
11357
11132
11732
13220
12951
13321
12950
(starting coal
Ash
% w/w
4.22
5.35
6.25
6.39
5.41
5.31
8.46
8.27
7.84
23.95
23.85
21.87
13.54
14.30
13.54
14.70
excepted).
Sulfur Content - Ibs
Individual 04.7!**
Fraction hL i]
0.53
0.59
0.70
0.62 0.59
0.54 0.56
0.55 0.61
0.68
0.65
0.59
1.55
1.55
1.41
0.94
0.96
0.93
1.24

S/106 Btu
FL+Z1+Z2






0.63
0.60
0.60






Weighted average of Float (FL) and Zone 1  (Zl) or Float and Zone 1  and Zone 2 (Z2).

-------
coal may be attained at 1.3 S.G. than was noted in the preceding equilibrium
studies (Section 7.2.2).
     Table 25 includes the non-equilibrium data obtained at 1,2 S.G.  Forty-
three percent of the coal is retained in zone 1 and zone 2 at 0.2 hours, which
has Btu, ash and total sulfur content comparable to zone 1 and zone 2 fractions
obtained at 1.3 S.G. after 0.5 hours.  The recovered coal, however, is less for
the 1.2 S.G. separation (43% versus 52%), plus an additional 10% of float was
also recovered from the 1.3 S.G. 0.5 hours separation.  With respect to the
1.2 S.G. separation, although more coal was recovered in zone 1 and zone 2 (51%)
for 0.1 hours at 1.2 S.G. than for 0.2 hours (43%), the total sulfur and heat
content, however, indicate a less-clean coal fraction compared to the 0.2 hours
1.2 S.G. (0.70 and 0.63 Ibs S/106 Btu, respectively).
     Particle size distributions for each coal density fraction listed in
Table 24 are shown in Figure 29.  It is seen that although the percentage dis-
tribution of particles is comparable between 1 and 2 hours for each mean parti-
cle size (mesh) within the float and zone 1 and 2 fractions, the decrease in
60 mesh and increase in pan (-200 mesh), within the sink fraction, suggests that
equilibrium has not been achieved.
     Based on the work of Yurovskii (op cit) and the application of Stokes Law,
it is not to be expected that small particles will attain equilibrium within
the system due to gravity alone over short periods of time.  It is well  known
that the velocity at which fine particles settle in heavy liquids is given by
Stokes Law:

                                 2 r2
                             v =    -(p  - P
where
      r = mean particle radius, cm,
     PI = density of the liquid, g/cc
     P2 = density of the settling particle, g/cc,
      n = viscosity of liquid, poise, and
      g = acceleration due to gravity.
                                     136

-------
                              TABLE 25.   ANALYSES  OF FRACTIONS  OF MARTINKA  COAL  OBTAINED FROM
                                           NONEQUILIBRIUM  FLOAT/SINK EXPERIMENTS
CO
Fraction
Float
Zone 1

Zone 2

Sink

Total Coal
(Weighted
Average)
Starting
Coal
Time
Hrs.
0.2
0.1
0.2
0.1
0.2
0.1
0.2
0.1
0.2
-..
S.G.
of Leach
Solution
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2

Coal
% w/w
0.2
22.5 0
14.0 0
28.5 0
28.6 0
49.0 1
57.2 1
1
1
1
Sulfur Analysis - °,
St
**
.98
.90
.99
.88
.55
.49
.26
.23
.61


0
0
0
0
0
0
0
0
0
SP Ss
Tr*
.22 Tr
.18 Tr
.23
.22
.89
.86
.55
.58
.72 0.18
£ w/w
So

0.75
0.71
0.75
0.65
0.65
0.62
0.70
0.64
0.71
Heat
Content
Btu/lb.

14050
14234
13992
14189
12250
12446
13151
13157
12950
Ash
% w/w

7.97
7.61
8.45
8.03
18.83
18.06
13.43
13.67
14.70
Sulfur Content - Ibs S/106 Btu
individual FL+Z1*** FL+Z1+Z2—
Fraction

0.
0.
0.
0.
1.
1.
0.
0.
1.

70 0.70
63 0.63
71 0.71
62 0.62
26
20
95
93
24
             Trace; values for S  were «  0.01;  0.01 was  used for S  calculation  (starting coal  excepted).

             Insufficient coal to process for analysis.

             Weighted average of Float (FL) and Zone 1  (Zl) or Float and Zone 1 and Zone  2 (Z2).

-------
00
       70
       60
       50
    S  40
    Z
      30
      20
       10
                  	 FLOAT  	         	 ZONE 1 	J        I	ZONE 2 	H        	  SINK  	1
                 1 HR   2 HR
                  60  100 200 PAN
60  100  200  PAN

              MESH
60  100  200  PAN
60  100 200  PAN
                          Figure 29.  Particle Size  Distribution in Fractions of RTU 3  Coal

-------
It was pointed out  by  Yurovskii  that for particles of a given diameter and
specific gravity, an increase in the speed of particle movement is possible
only when the speed will  depend  not on acceleration due to gravity, g, but on
the  acceleration due to centrifugal force, i.e.,
                                    2
                            Vc = 97T(p2 ' pl} Cg
where
        =  _ _
      g    90
     rc =  rotational  radius of centrifuge, cm,
      n =  number of revolutions of the centrifuge/min (rpm).
Table 26 illustrates the relative sensitivity of small particles of Martinka
coal and pyrite to gravitational and centrifugal forces with respect to veloc-
ity of settling.  It may be noted that the interaction effects (particle-
particle)  are not included in this comparison; thus, the velocities are in
reality lower than those given.
     It is seen that the centrifugal method is particularly effective and would
enhance coal  separation in the Gravichem process.  The effectiveness of the
centrifugal method is demonstrated by Yurovskii 's data in Table 27, wherein he
compares a flotation and centrifugal separation of a coal containing originally
24.5% w/w ash and 3.21% w/w sulfur.  It is seen that for essentially equal con-
centrate yield from 1.5 S.G. solution, the centrifugal method gives a much
cleaner concentrate (float).
     It was desirable to investigate the non-equilibrium float/sink process
further by assessing the effect upon coal separation (yield) when discrete par-
ticle size fractions of Martinka coal are used, i.e., the standard input coal
(14 mesh x 0) used in the preceding experiments (Tables 24,  25) was separated
into four  particle size fractions (+60, -60 x 100, -100 x 200, -200 mesh).
Non-equilibrium float sink experiments were conducted at 80°C on each of the
four fractions for 30 minutes in 1.3 S.G. leach solution, as described above.
It is seen in Table 28  that the combined yield for float, zone 1 and zone 2
fractions  is  similar to the previous experiment (Table 24) using the whole coal.
                                      139

-------
TABLE 26.  RELATIVE SETTLING RATES OF COAL AND PYRITE SUBJECTED TO
           GRAVITATIONAL AND CENTRIFUGAL FORCES
Particle Particle
(mesh)
Coal 100
200
Pyri te 1 00
200
V - fl_(p2-Pl) g
** 9 K-2 r
V - 2 r fr, n }
V ' 9 n Ip2 pl'
Size
(cm)
0.015
0.0074
0.015
0.0074
c»z
90
Density of
Particle
(g/cc)
1.45
1.45
5.0
5.0

*
Gravitational
Vel oci ty
cm/sec
9 x 10"3
2 x 10"5
0.22
5.6 x 10"4

**
Centrifugal Velocity
(1000 rpm)
cm/sec
r = 10 r = 300
c c
1.02 30
2 x 10"3 6 x 10"2
24.9 747
6.3 x 10"2 1.89


-------
                   TABLE 27.   FLOTATION  AND CENTRIFUGAL  SEPARATION  ON  COALS*
Original Coal, % w/w
Sample Source Ash Sulfur
Irminskaya 31.0 5.30
TsOF


Kalmiusskaya 24.5 3.21
TsOF

Cleaning Method
Flotation proper
Cleaning flotation (with a
depressor)
Centrifugal separation in
sulfate solution
Centrifugal separation after
preliminary chemical
treatment
Flotation proper
Cleaning flotation (with a
depressor)
Centrifugal separation in
sulfate solution
Concentrate, %
Yield
71.9
62.0
62.8
62.4
79.8
69.5
68.4
Ash
14.7
9.9
6.2
3.7
11.0
8.2
4.4
w/w
Sulfur
3.70
3.55
2.60
1.84
2.60
2.61
1.86
* Table 80, Yurovskii"2'

-------
-p.
ro
                    TABLE 28.  YIELD OF MARTINKA COAL SEPARATED IN NONEQUILIBRIUM FLOAT/SINK

                               EXPERIMENTS AFTER 0.5 HOURS @ 80°C (S.G.  1.3)
Feed Coal
Size Fraction
(Mesh)
-200
-100 x +200
- 60 x +100
- 14 x +60
Cumulative Fraction
Recovered
Fraction Recovered
from 14 x 0 mesh Coal
Feed Coal
Weight
(gm)
537
725
776
792

*

Float
5
3
160
194
13%
10%
Recovered
Zone 1
177
302
194
126
29%
21%
Coal (gm)
Zone 2
200
161
76
32
17%
30%

Sink
130
235
344
425
41%
38%
Recovery
%
95
97
99
98


          * 0.5 hour, 1.3 S.G, from Table 23.

-------
Likewise,  the sink fraction is comparable.  It is noted, however, that the yield
of zone 1  and zone 2 is approximately reversed which indicates that some lower
specific gravity coal  may be contained in zone 2 in the previous experiment.
     Referring again to Figures 29 and 30, one notes that the total sulfur,
based on the weighted average for total coal, decreases ^0.35% w/w.  This is
seen at either 0.1 hours (S.G. 1.2) or 2 hours (S.G. 1.3).  Such a decrease
cannot be  accounted for (including a correction for removal of S ) on the basis
of the rate constant (KL) for removal of pyritic sulfur (S ) from Martinka coal.
However, the data show the decrease to be consistent for all of the experiments
in both tables.  A series of experiments were conducted at 80°C with 14 mesh
Martinka coal wherein samples were wetted by 30 minutes prefoaming, leached
separately in water and 4% w/w aqueous H2S04 solution for 1 hour, and for 0.5,
1 and 3 hours in 4% w/w aqueous H2S04 containing 7.5% iron by weight.
     Subsequent coal analyses from these experiments are presented in Table 29.
The following observations are made from these data.
        t    The leached coal total sulfur decreases primarily as a result
            of reductions in sulfate and organic sulfur forms.
        t    Since organic sulfur is not water soluble and since the organic
            sulfur value of the feed coal analyzed abnormally high, it can
            can be concluded that the reduction in total sulfur by water
            and by dilute H2S04 represents sulfate leaching (the reduction
            in organic sulfur is attributed to errors by sulfur forms
            analyses).
        t    Elemental sulfur formed indicates a maximum of 0.12% w/w reduc-
            tion in pyrite during treatment with acid-Fe reagent (assuming
            0.03% present as a result of weathering).
7.3  PROCESSING OF RTU COAL
     Experimental studies were conducted at bench-scale level to evaluate proc-
essing variables and their attendant effect on sulfur removal from Rartinka
mine cleaned coals in support of the RTU.  A series of experiments were com-
pleted which include pressurized bench-scale processing at 120°C and 135°C,
and unpressurized processing at 102°C.  Additional tasks included determination
of the density of leach solution/coal  (slurry) and evaluation of the weathering
of RTU candidate coals.
                                      143

-------
                    Pressure
                            Wet
                    Control  Test
                            Meter
                                 Gas
                               Analysis
[ PT3  L
      MIXER

      AND

    DEFOAMER
                                                       N,
                                                       Pump Seal
                                                      Purge System
        lC_OAL._j
         SLURRY
                                                                              I   FINAL PROCESSING*
                                                             Reactor-Settler   Stirrer Reactor
STEP  #1 (* 1.0 HOUR)
                  STEP #2 (1-8 HOURS)**
STEP  #3 (18-24 HOURS)
     * Final processing  includes elemental sulfur recovery, coal washing and drying.

    ** Reactor volume i>  13 liters.
                 Figure 30.  Bench-Scale Coal  Leaching and Reagent Regeneration Apparatus

-------
                      TABLE 29.   LEACHING  OF MARTINKA COAL  AT  80°C
Reagent




Feed Coal

Water
4% H2S04 in
Water
4% H2S04 +7.5%
Fe in H20

Time
Hrs.





1.0
1.0
1 0.5
1 1.0
( 3.0

Total
Sulfur
s.
t
1.62

1.35
1.27
1.30
1.15
1.23
Coal
Pyri ti c
Sulfur
S
P
0.72

0.84
0.76
0.75
0.68
0.64
Analyses
Sulfate
Sulfur
S
s
0.18
**
Tr
Tr
Tr
Tr
Tr
X, w/w
Organic
Sulfur
S
0
0.72

0.50
0.50
0.55
0.47
0.59

Elemental
Sulfur
S
n
0.03***

0.03
0.04
0.06
0.06
0.08

 **
***
Coal analyses performed after removal  of elemental  sulfur (except feed coal)
<0.01%, reported as trace
Included in the organic sulfur column

-------
7.3.1  Pressurized Bench-Scale Processing
     Experimental studies were conducted in the pressurized bench-scale  reactor
(Figure 30) as described previously'5', using 14 mesh Martinka coal  processed
at 135°C and 100 psig 02 pressure.  The object of the study was to  investigate
the effect of reduced Fe+++ and acid (H2S04) concentrations on the  removal of
sulfur from this coal under simulated RTU conditions.  It is seen in Table 30
(Experiments 1 and 2) that the amount of sulfur removal is not affected  by a
               _i __ ii
reduction in Fe    and acid concentrations when compared to coal processed in
      -
5% Fe    and 4% H2S04 standard RTU leach solution [Reference Experiment 01-10
(2P; 12F)].
7.3.2  Ambient Pressure Bench Scale Processing
     Gravichem separation of coal using leach solution with various densities
achieved by increasing the acid or Fe content, required the investigation of
these solutions with respect to their suitability as reagents for the Meyers
Process.*  It was demonstrated in Section 7.2 that 7.5% Fe    and 4% acid is a
suitable leach solution as shown by the reduction of inorganic sulfur in Ken-
tucky Mine No. 9 coal.  It has also been shown (Table 30, Experiment 3) that
high (20%) I^SO^ concentration does not change the response of inorganic sulfur
to the oxidizing reaction of the Meyers Process.  A significant reduction in
ash and increase in heat content is noted, however, with the higher acid
concentrations.
7.3.3  Weathering of RTU Coal
     A study was conducted to determine the effect of coal particle size on
weathering of Martinka coal.  RTU coal No. 1 was used for these tests since
high sulfate values (MD.40% w/w sulfate sulfur) were found on this coal which
was stored for sometime after grinding.  The objective was to determine if the
sulfate was due to weathering of the as-received 1-1/2" x 0 coal or the result
of post-grinding storage.  A representative sample of the coal pile was taken
at the storage area and particle size fractions were obtained from the sample
to determine the extent of weathering for three fractions after grinding to
3/8" topsize (3/8" x 14 mesh; 14 x 100 mesh; 100 mesh x 0).
 Standard starting leach solution contains 18% Ferric Sulfate (5% Fe),
 4% acid, 78% H20.
                                      146

-------
                       TABLE 30.  BENCH-SCALE PROCESSING OF MARTINKA COAL
Experiment
No.
Control
RTU Processed
Exp. 01-10
*
*
**
3
*
Processed
**
Processed
***
Temp Tine 02
Pressure
(°C) (Mrs) (psi)
132
135 12 72
135 12 72
***
102 48 0
at 100 psig.
at ambient pressure.

Fe H0SO,, Coal Analysis,
" Ash Total Pyritic
(%) (%) Sulfur Sulfur
St SP
5 4 — - 0.91 0.17
0 2 15.76 0.91 0.23
0 0 16.15 0.90 0.20
5 20 11.80 0.96 0.29



% w/w
Sulfate Organic
Sulfur Sulfur
Ss So
0.12 0.65
0.17 0.51
0.01 0.69
0.06 0.61



Meat
Content
(Btu/lb)
___
12360
12509
13456



Equivalent to ^ 14 hours 0 135°C.

-------
     Analytical data are summarized in Table 31,  Included in the Table  are
similar data for the two additional RTU coals (designated RTU #2 and  RTU #3)
which were received from the mine in July 1977.  Comparing RTU 1* with RTU 2*
and RTU 3* in Table 31, it is seen that a significant decrease in pyritic sulfur
and an increase in sulfate sulfur occurs for RTU 1.
     The data indicate that the majority of the observed weathering in the RTU
ground coal occurred prior to the grinding operation.  Additionally,  the weath-
ering is not a pile surface phenomenon since it is seen that the RTU  1 (Interior)
sample likewise shows decreased pyritic sulfur and increased sulfate  sulfur.
     Elemental sulfur analyses of coals processed in the program studies have
been included  in Table 32.  Samples of nonprocessed input coal were ground to
14 mesh and extracted with toluene to remove elemental sulfur.  It is seen that
uniformally sampled RTU coals have 0.03% w/w elemental sulfur whereas Kentucky
Mine No. 9 coal (a Gravichem coal) has 0.10% w/w elemental sulfur.  These data
show that elemental sulfur is produced during weathering, however, not to the
extent that is is produced (M3.4% w/w AS ) during Meyers Processing.
     During the RTU coal sampling studies it was noted (Table 32) that random
(yellow) deposits were present throughout the coal pile.  Sampling of these
deposits with  subsequent processing for removal of elemental sulfur revealed
the presence of higher than normal elemental sulfur.  Additionally, qualitatively,
high presence  of sulfate was determined.
7.3.4  Density of Coal Slurry
     A series  of density determinations were made on 100 mesh x 0 and 14 mesh x 0
Martinka mine  cleaned coal in slurry concentrations ranging  from 19%  to  42% coal
by weight in distilled water.  Densities of the slurry combinations are  tabu-
lated in Table 33.  It is seen, in addition, that the average density of the
14 mesh x 0 coal is 1.4241 gm/cc ±0.0052 gm/cc.  Similar values for the
100 mesh x 0 coal  (1.4221 gm/cc ±0.0053 gm/cc)  indicate that comparable  wetting
of the coal is achieved for both size distributions.
 *Ground  to  3/8"  x 0 before  separation into fractions.
                                      148

-------
                        TABLE 31.   DISTRIBUTION  OF  INORGANIC SULFUR IN RTU COALS  (%)
vo
Coal
RTU 1***
RTU 1
(Interior)
RTU 2***
RTU 2
(Interior)
RTU 3***

RTU 1
RTU 1A
RTU 2
RTU 3
Combined 3/8" x 14 Mesh TOO Mesh x 14 100 Mesh x 0
Coal Fractions Sn Sc S_ S S S
s*s** p s p ps
P s
0.92 0.59 0.78 0.44 0.66 0.50 1.88 1.3
0.70 0.58
1.16 0.01 1.05 0.01 1.00 0.11 2.18 0.36
1.16 0.01
1.02 0.05 0.82 0 0.92 0 2.00 0.11
(COAL SIZE FRACTIONS)
58 26 16
51 18 6
63 25 13
67 22 11

            S   =  Pyritic Sulfur
        **   P
            S   =  Sulfate Sulfur

            Ground to 3/8" x 0 before separating  into  fractions

-------
                    TABLE 32.   ELEMENTAL SULFUR CONTENT OF
                               NONPROCESSED COAL
                                               Elemental Sulfur,
                       Coal*                         % w/w
               RTU 1**                               0.01
               RTU 2**                               0.02
               RTU 3**                               0.03
               Kentucky Mine No. 9**                 0.10
               RTU 1 (Yellow)***                     0.07
               RTU 2 (Yellow)***                     0.10
               *Toluene extracted to remove elemental sulfur
               **Uniform sampling
               ***Speciany sampled yellow deposits
7.3.5  Elemental Sulfur Recovery
     The product elemental sulfur from the oxidation of pyrite in coal by the
Meyers Process has been routinely recovered by extraction with toluene and sub-
sequent distillation of the solvent.  Sulfur recovery is virtually complete by
a single stage extraction of leached coal with toluene provided the coal is
predried or azeotropically dewatered.  Experimentation has demonstrated that
acetone is preferable to toluene for product sulfur recovery because  (a) it
does not require that the coal be dewatered prior to recovery, (b) it extracts
iron sulfates as well as elemental sulfur from wet coal, and (c) it is easily
(56°C) and completely (>99.5%) recoverable from coal.  Acetone has been selected
as the most cost effective technique for elemental sulfur recovery from Meyers
Process leached coal.
     Investigations on acetone efficiency in recovery of product elemental
sulfur, Sn, from the oxidation of pyrite in coal  (Meyers Process) was performed
at bench-scale with bench-scale leached and RTU leached Martinka coal samples.
Parameters examined for appreciable effect on recovery efficiency were:  (a) the
moisture content of leached coal and feed solvent (from 1% to 72% water based
                                     150

-------
                   TABLE 33.   DENSITY OF COAL/WATER SLURRIES
Coal Slurry Temp
Size Concentration °C
(Mesh) %
14 21 40
60
80
28 40
60
80
42 40
60
80

100 23 40
60
80

Density
of Slurry
(gm/cc)
1.0594
1.0510
1.0418
1.0849
1.0768
1.0685
1.1363
1.1286
1.1215

1.0920
1.0835
1.0760

Density
of Coal
(gm/cc)
1.4265
1.4187
1.4235
1.4325
1.4256
1.4301
1.4243
1.4171
1.4190
x 1.4241
a 0.0052
1.4260
1.4161
1.4241
x 1.4221
a 0.0053
Slurry
               EQUATIONS  FOR  CALCULATING  DENSITY  (D), gm/cc
          Function of Temp, °C

  21     D = 1.07713 - 0.00044 Temp

  28     D = 1.10133 - 0.00041 Temp

  42     D = 1.15100 - 0.00037 Temp
Temp
 °C
  40

  60

  80
    Function of Slurry
     Concentration, %
D = 0.9824 + 0.00366 Slurry

D = 0.9734 + 0.00370 Slurry

D = 0.9622 + 0.00379 Slurry
                                    151

-------
on dry coal weight*), (b) coal particle size  (14 and 100 mesh top-size), (c)
extraction residence time per stage  (0,5, 1.0, and  2.0  hours), (d)  number of
extraction stages (1 to 3), and (e)  sample prehistory.  The  latter  included
leached coal samples extracted immediately after taken  from  the  RTU belt filter
(water rinse only) and samples that  were stored for as  long  as 30 days;  the
study also included thoroughly washed coal samples and  oven  dried samples.   The
majority of experiments were performed with 14 mesh top-size,  RTU leached coal
slurried in twice its dry weight acetone.  This coal contained 0.30 grams ele-
mental sulfur per 100 grams of dry coal; one  of the bench-scale  leached  Martinka
coals contained 0.40% elemental sulfur.  Acetone extractions  were performed at
the reflux temperature of the slurry (56°C).
     The data generated in elemental sulfur recovery by acetone  investigations
are summarized in Table 34 and Figures 31 and 32.  Table 34  presents data from
19 experiments involving two-and three-stage  extractions of  leached coal  with
acetone.  Data are also presented from toluene extractions of  dry leached coal
for comparison purposes.  The table  shows the estimated weight of elemental
sulfur and of moisture on 100 weights (dry basis) of leached  coal fed  to  the
first extraction stage (columns 2 and 3), the weight of the  input sulfur
recovered in each of the three stages, the total S  recovered, and  the percent
recovery (recovery efficiency).  The elemental sulfur on the  feed coal (leached
coal) was estimated by three independent techniques:  (1) the  sulfur recovered
by three toluene extractions, (2) forty percent of the  chemically leached pyrite,
and (3) from the increase in organic sulfur of leached  but not extracted  coal;
the three techniques agreed within 10%.  The  asterisked experiments were  performed
with leached coal which was not completely washed from  the iron  sulfate  reagent;
thus, it is conceivable that a small quantity of sulfate sulfur  may have  been
analyzed as elemental sulfur which could have exaggerated slightly  the quoted
efficiencies.
     Figure 31 depicts the expected  elemental sulfur recovery efficiencies, as
a function of acetone extraction stages, from up to 14  mesh  top-size chemically
leached Martinka coal.  The data were generated at reflux temperatures with
 Water was present on the feed coal and in a limited number of  experiments
 in the acetone feed.

                                     152

-------
                             TABLE 34.   PRODUCT  ELEMENTAL SULFUR RECOVERY  FROM LEACHED MARTINKA COAL
01
to
Sulfur (Sn), Grams, Per TOO Grams of Coal
Experiment
No.
A. Acetone
1**
2**
3
4**
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Estimated Sn
on Coal
(% w/w)
Extractions
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.30
0.40
0.40
0.40
0.30
0.40
0.30
0.30
0.30
0.30
Moisture on
Feed Coal
(% w/w)

1
1
1
1
1
20
20
20
20
20
24
28
42
42
52
72
72
72
72
1st Stage
Extraction
Sn Recovered

0.23
0.29
0.21
0.27
0.27
0.27
0.22
0.20
0.28
0.22
0.26
0.27
0.20
0.18
0.20
0.13
0.17
0.14
0.14
2nd Stage
Extraction
S Recovered

0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.05
0.03
-
0.08
0.07
0.12
-
0.13
_
_
_
-
3rd Stage
Extraction
S Recovered

_
0.02
_
.
0.01
.
0.001
_
0.01
-
0.05
0.05
0.05
-
0.08
.
_
_
-
Total Sn
Recovery
(% w/w)

0.27
0.35
.25
0.31
0.32
0.31
0.26
.25
0.32
-
0.39
0.39
0.37
-
0.41
_
_
_
-
%
Recovery

90
117
83
103
107
103
87
83
107
-
98
98
93
-
103
_
_
_
-
B. Toluene Extractions
20
21
0.30
0.30
11
1
0.28
0.26
0.01
0.01
_
~
0.29
0.27
97
90
            One  hour extractions of 33% coal slurries at reflux  temperatures  (56°C for acetone
            slurries, 110°C for toluene).


            Filter belt water  rinse only on process coal.

-------
                                         WATER CONTENT  OF  COAL
                                        t FED TO FIRST STAGE
                                      • FfH-f H H t Hi 11 -H- H-H-H -ftt-f ttt-H+H-H
                                            20 (23 SAMPLES)

                                            50 (10 SAMPLES)
                                            72 ( 7 SAMPLES)
               EXTRACTION STAGES  EMPLOYED

Figure 31.  Elemental  Sulfur Recovery as a Function
            of Acetone Extraction Stages

                         154

-------
                                                     -tr rr!T tft;
                                                  1st STAGE
                       ON COAL FED TO  FIRST  STAGE
Figure 32.  Effect of
            Recovered
Water on Elemental Sulfur
by Successive Acetone Stages
                         155

-------
coal which was leached chemically under a variety of conditions  and  up to

30 days prior to acetone extraction.  Acetone extraction  time  was  varied between

30 minutes and 2 hours per stage.  The ±10 percent band about  the  first stage

S  recovery value indicated for the 0-20 percent moisture coal samples applies

also to the higher moisture coal samples also.  This band represents the maxi-

mum uncertainty expected in the quoted average value and  it  is due partly to

experimental and analysis uncertainties and partly to minor  parametric effects

(including effects due to small differences in water content).   Similar or even

larger relative uncertainties in the second and third extraction stages are

masked by the small percentage of original elemental sulfur  recovered  in these

stages.

     Figure 32 summarizes the effect of water in the coal-acetone slurry on

elemental sulfur recovery from chemically leached coal.   The moisture  effect

is depicted for each successive extraction stage.

     The data presented here complemented with chemical analyses on  the  recovered

elemental sulfur residues and determinations on acetone retention  by extracted

and dried coal led to the following observations:

        •   Elemental sulfur recovery from Meyers Process  leached  coal  by
            a single stage acetone extraction ranged from  49% to 85%
            depending on the moisture content of the leached coal.

        •   The average elemental sulfur recovery by a single stage acetone
            extraction of processed coal containing 1-20% moisture was  83%.
            An additional 13% (approximately 80% of the residual Sn) was
            recovered in the second acetone extraction stage and another
            2-3% in the third stage.  The corresponding elemental  sulfur
            recoveries from wet coal containing 50-70% moisture were 50,
            30, and 20%.

        •   The water content of the chemically leached coal or of the  feed
            (recycled) acetone had a pronounced effect on  sulfur recovery
            efficiency during the first extraction stage.  The moisture
            effect diminished in subsequent stages.

        t   Coal particle size up to 14 mesh top-size and  extraction res-
            idence time beyond 30 minutes did not exhibit measurable effect
            on elemental sulfur recovery efficiencies by  acetone.

        •   The purity of the sulfur ranged recovered from the first stage
            ranged between 50-80% and that of the second  stage was between
            40-60%; the sulfur content of the residue recovered  in the
            third stage was approximately 10%.

                                     156

-------
       •   Approximately 0.5% of the coal matrix was dissolved in the
           three acetone extraction stages; at least 60% was dissolved
           in the first stage.  The quantity of coal matrix dissolved
           in acetone appeared to be independent of the amount of sulfur
           recovered.  Thus, product sulfur purity should increase with
           increasing pyrite concentration in the feed coal (higher
           concentration of product sulfur in the coal).

    The above observations lead to the following conclusions:

       t   Elemental sulfur can be readily recovered from wet leached
           coal by acetone.

       t   Two 1-hour acetone extraction stages should be adequate for
           better than 85% sulfur recovery provided the water content
           of the acetone-coal slurry does not exceed 40% of the weight
           of the coal in the slurry.  (A third stage may be required
           if the water concentration of the slurry is appreciably
           higher).

       t   Water concentration is the only parameter of those examined
           and in the ranges investigated that had a pronounced effect
           on elemental sulfur recovery by acetone.

       •   Acetone retention by the coal is expected to be less than
           0.5% upon drying in commercial driers (extrapolation of
           data generated at TRW and at Wyssmont Co., Inc. laboratories).

       •   Acetone extraction is selective to elemental sulfur.  Less
           than 0.5% of the coal matrix was dissolved in three stages
           of extraction (most of it was dissolved in the first stage).

7.4   LABORATORY STUDY CONCLUSIONS

       1.  Equilibrium float-sink separations of Martinka mine coal
           (14 mesh x 0) in ferric sulfate leach solution of varying
           specific gravity demonstrate the feasibility and the physical
           and chemical basis of the Gravichem method.

       2.  The sulfur content of Gravichem float and sink coal fractions
           is reduced during gravity separation because of leaching of
           pyritic sulfur.

       3.  Gravichem sink material has greatly enhanced heat content
           because of dissolution of coal mineral matter in addition
           to pyrite.

       4.  Equilibrium Gravichem processing of 3/8-inch top-size coal
           from Kentucky No. 9 Mine (TVA coal) generated a float product
           (38-52% w/w yield) containing 3.09-3.50 Ibs SOp/HP Btu, hav-
           ing a heat content range between 13774 and 14354 Btu/lb.
           This coal meets Tennessee State sulfur standards.

                                    157

-------
 5.   The  sink fraction of the TVA coal  when treated at 102°C^accord-
     ing  to Meyers process procedures,  yields a coal containing less
     than 4 Ibs S02/1Q6 Btu.

 6.   Non-equilibrium Gravichem separation of coal  into four frac-
     tions revealed that a greater yield of clean  coal may be
     attained than was noted  for simple float/sink fractions (two
     fractions).

 7.   Data from coal cleaning  tests in the U.S.S.R. using ferric sul-
     fate solution as the separation medium corroborates the TRW
     data and verifies the applicability of Gravichem as a Meyers
     process modification.

 8.   The  amount of pyritic sulfur removal by chemical leaching is
     not  affected by Fe+++ and acid concentration  under simultaneous
     coal leaching-reagent regeneration conditions.

 9.   High (20%) H2S04 concentration does not change the response of
     inorganic sulfur to the  oxidizing  reaction of the Meyers process.

10.   Weathering of RTU coals  during storage prior  to processing was
     observed.  Weathering was manifested by reduction in the pyritic
     sulfur content of the coal, by increase in sulfate sulfur con-
     tent, and by formation of elemental  sulfur.              <

11.   Elemental sulfur can be  readily recovered from wet leached coal
     by acetone.
                              158

-------
                 8.   ENGINEERING DESIGN AND COST ESTIMATION

     This  section of the report documents the results of a process engineering
study performed in support of the current RTU Project.  The primary objectives
of the process engineering effort were twofold:
        1)  To prepare conceptual full-scale process designs specific to
            several  coal feed stocks incorporating updated process data
            from the RTU.
        2)  To evaluate the new Gravichem process option on a conceptual
            full-scale basis combining data obtained at lab or bench-
            scale with RTU plant data.
The detailed results of the current process engineering task are presented on
the following pages.
     Section 8.1 contains a brief introduction and the historical background
which lead to the current chemical coal desulfurization process scheme.  Also
described  in this section is the generalized Gravichem Process concept.  The
balance of Section 8.1 deals with the approach which was utilized in generat-
ing conceptual full-scale process designs, determining battery limits process
capital and operating costs, and evaluating integrated grass roots process
economics.
     Section 8.2 details the specific base case designs, flow sheets, equip-
ment lists, battery limits costs, and integrated grass roots facility costs
and economics evaluations.  Section 8.3 presents the data analysis conclusions
drawn from the engineering effort.
8.1  INTRODUCTION AND BACKGROUND
     Throughout bench-scale development, processing techniques and their asso-
ciated costs have frequently been reviewed with an objective of focusing experi
mental effort in the areas of greatest cost sensitivity.  The capital cost of
equipment  required to perform the pyrite leaching must be carefully controlled
                                    159

-------
to maintain a low processing cost per ton of  coal  product.   A considerable
effort has been made to reduce costs in the core processing steps.   While some
improvements in cost have been made in the basic Meyers  process  compared to
previous designs, the major emphasis has centered  on  obtaining the  large cost
advantages possible by gravity separation of  coal  into clean (float)  and high
pyrite (sink) fractions and chemically processing  only the  high  pyrite  frac-
tion.  The goal has therefore been to obtain  maximum  benefit from the new
Gravichem float/sink technology presently under development.   During  the
course of the current engineering task, two coal specific process designs were
generated and evaluated.  The designs incorporate  Meyers process  technology
into recently demonstrated gravity separation techniques (which  utilize  Meyers
process leach solution as a separation media) yielding improved  desulfuriza-
tion schemes defined as the "Gravichem Process".   The two designs presented
herein are specific to bench-scale and RTU tested  mine cleaned Martinka  coal
and bench-scale tested TVA Kentucky Number 9 coal.
8.1.1  Historical Background
     Over the last six years, there have been several process concepts devel-
oped by TRW which involve Meyers' Coal Desulfurization Process Technology.
The earliest of the conceptual process designs was generated  in  1972  and has
been reported in detail previously'4).  That design was based on  processing
a generalized 3.2% w/w pyritic sulfur ROM Lower Kittanning  coal  (pulverized
to -100 mesh) in a batch mode to remove 95% w/w of the contained  pyritic sul-
fur.  The specific design was based on early  laboratory-scale test data and
the processing scheme represented a direct Meyers  process treatment of coal.
     The next series of designs were generated in  1975 and  1976(5'.   Those
designs also related to the desulfurization of a generalized  ROM  Lower Kit-
tanning coal.  However, the designs generated in that time  period pertained
to continuous mode coarse coal processing (0.25-inch top-size) and continuous
mode fine coal  processing (16 mesh x 0).  Of the four design  cases which were
developed, one represented a straight-through Meyers Process, one involved a
physical  cleaning pretreatment and two of the schemes coupled physical clean-
ing, conventional gravity separation and Meyers process treatment.

     The most recent of the reported bench-scale based design studies^11' was
completed in late 1977.  That design was specific  to mine cleaned Martinka
coal containing 1.2% w/w pyritic sulfur.  The design was based on continuous mode
                                    160

-------
operation  utilizing 14 mesh x 0 coal.  The overall process design incorporated
Meyers  process treatment with the newly developed float/sink gravity separation
technology which utilized a 1.3 specific gravity Meyers  process leach solution
as the  liquid media and takes advantage of concurrent  pyrite leaching.  The
overall process is termed the Gravichem Process.  This technology differs from
the previous float/sink Meyers Process option^ in that the conventional sep-
aration technique utilizes stabilized magnetite water  solutions as the liquid
media thus requiring a separate processing circuit.    A  series of block dia-
grams is presented in Figure 33 which indicate the genesis of the current
process technology scheme.
8.1.2  Current Generalized Process Concept
     As mentioned in the preceding paragraphs, the most  recent generalized
coal desulfurization processing technique, Gravichem,  utilizes the previously
demonstrated Meyers Process in conjunction with a newly  developed leach solu-
tion gravity separation approach.  A block diagram of  the Gravichem Process
is presented as Figure 34.  It should be noted that the  block diagram presents
a processing concept which is generally applicable to  any high pyrite contain-
ing suspendable coal (nominally 14 mesh top  size).  The  scheme presented in
Figure 34 represents the current state of Gravichem Process technology.  It is
based on past developmental experience^  '  '   ' as well as information obtained
during the current RTU program and its associated laboratory/bench scale sup-
port activities (discussed in Sections 6 and 7 of this report).  The following
paragraphs describe the Gravichem Process (Figure 34)  and discuss the intercon-
nections between each of the main process operations.
     Ground coal, with a nominal top-size of 14 mesh,  is mixed with hot recy-
cled leach solution.  Process required make-up sulfuric  acid is also added to
the mix tank such that the final leach solution specific gravity is maintained
at the desired level.  After coal wetting and mixing is  complete at the solu-
tion boiling temperature, the now partially  reacted coal slurry is cooled and
introduced into a float/sink separator.  The lighter low pyrite, low ash, coal
particles float to the top of the separator  in a relatively dilute slurry
while the more dense high pyrite, high ash,  coal particles in a much thicker
slurry are pumped from the bottom of the unit.
                                    161

-------
                 1972 - MEYERS PROCESS (100 MESH TOP SIZE)

                 (REF. 4)
ROM UK. COAL'
100 T/HR COAL
3.2% PYRITIC SULFUR
20% ASH.
19.7 X 10° MM BTUAR"
MIXING







WASH



SULFUR
EXTRACTION






94.2 T/HR
16% PYRITIC SULFUR
15% ASH
19.4 X 10° MM BTUAR
                 1976 - CLEANED FINE COAL (IX MESH TOP SIZE)
(REF. 7)
ROML.K. COAL
120 T/HR COAL
20% ASH
»- 4% npmc SULFUR ^

PHYSICAL
CLEANING
DISCARD

MM BTUAR 20 T/HR COAL
75% ASH
10 - 14% .PYRITIC SUL
i.4x IO'MMBTUA

100 T/HR COAL
9% ASH
MEYERS PROCESS
FINE COAL
CONFIGURATION
1.5 -2% PYRITIC SULFUR
22.2X10° MM BTUAR
FUR
!
COAL PRODUCT
93 T/HR COAL
6% ASH
. 1% PYRITIC SULFUR.
21. 3 X 10° MM BTUAR
                 1976 - RUN-OF-MINE COARSE COAL (1/4 IN. TOP SIZE)
(REF. 7)
ROM L. K. COAL
100 T/HR COAL
20% ASH
3-4% PYRITIC SULFUR
19.7 X 10*MM BTUAR

MEYERS PROCESS
COARSE COAL
CONFIGURATION

COAL PRODUCT
85 T/HR COAL
15% ASH
.2% PYRITIC SULFUR
17.5 X 10* MM BTU/YR
no
                 1976 - DEEP CLEANED FINE AND COARSE COAL WITH 50% MEYERS PROCESS BYPASS (14 MESH TOP SIZE)
(REF. 7)
ROMl.K. COAL
240 T/HR COAL
20% ASH
3 - 4% PYRITIC SULFUR
A AST

PHYSICAL
CLEANING
DISCARD
47.2X10" MM BTU/YR jJT/MRCQ),
75% ASH
10 - 14% PYR
2.8 X 10° M(i

200 T/HR COAL
1.5-2% PYRITIC SULFUR
L 44.4 X 10* MM BTUAR
1TIC SULFUR
* BTUAR

CONVENTIOh
GRAVITY
SEPARATION
1

ML
100 T/HR COAL
MEYERS PROCESS
CONFIGURATION
3 - 4% PYRITIC
SULFUR
COAL PRODUCT _
85-90 T/HR COAL
10- 15% ASH
.2% PYRITIC SULFUR
BYPASS 100 T/HR COAL
LOW ASH
LOW SULFUR
185 - 190 T/HR COAL
6% ASH
.2% PYRITIC SULFUR
FINE ,
43.4 X 10° MM BTUAR
COARSE .
42.2X10° MM BTUAR
                  1977 GRAVICHEM PROCESSING WITH 55% FLOAT PRODUCT (14 MESH TOP SIZE)

                  (REF. 8)
                 MINE CLEANED
                 MARTI NKA COAL
                 225 T/HR COAL
                 14% ASH
                 1% PYRITIC SULFUR

                 46.4 X 10* MM BTUAR
                  *ROML.K. COAL - RUN-OF-MINE LOWER KITTANNING COAL
                 "8000 OFHMING HOURS PER YEAR, DRY COAL BASIS.
FLOAT _ 1 WATER 1
123 T/HR CO;
6% ASH
.26% PYRfTIC
28.1 X10*M
SINK
01 T/HR COAL
3.7% ASH
.53% PYRITIC
8.2 X 10° MH
^L | WASH [
SULFUR
M BTUAR

EXCEPT MIXING
SULFUR
K BTUAR

123 T/HR COAL
6% ASH
.26% PYRITIC SULFUR
28.1 X 10* MM BTUAR
98 T/HR COAL 223 T/HR COAL
22.2% ASH 13.2% ASH
. 15% PYRITIC SULFUR .21% PYRITIC SULFUR
IB. 1 X 10° MM BTUAR 46.2 X 10° MM BTUAR
                                             Figure  33.    History of  the  Gravichem  Process

-------
01
to
                   COAL/IEACH
                   SOLUTION
                   MIXING
                     Figure 34.   Gravtchem  Process Block  Diagram Suspendable Coal
                                  (14 mesh x 0)  Approach

-------
     The clean (low pyrite, low ash) coal float fraction  is  next  sent  through
a series of filtrations and water contacting operations to wash the  iron
sulfate-sulfuric acid leach solution from the coal.  The  washed coal cake  is
then ready for final blending into the chemically desulfurized sink  fraction.
The recombined fractions comprise the product coal stream.   The float  cake
wash water is supplied from internally process generated  steam condensate and
make-up water.  The concentrated leach solution filtrate  is  sent  through a
regenerator where the ratio of Fe+3 to total Fe is adjusted  via oxygen induced
oxidation.  The regenerated reagent stream is then recycled  back  to  the mixing
unit.
     The high pyrite, high ash sink fraction from the separator is introduced
into a reaction vessel.  It is here that the majority of  the exothermic pyrite
leach reaction takes place at elevated temperatures and pressures in the pres-
ence of oxygen.  The oxygen is simultaneously added to exothermally  regenerate
(adjust the Fe   to total iron ratio) the leach solution.
     The regenerated, fully reacted slurry is then subjected to a filtration
and cake wash operation.  The water used for sink cake washing is obtained
from internally generated process steam condensate and make-up water.  The
concentrated leach solution filtrate is recycled directly to the mix tank.
The wash filtrate (very dilute leach solution) is split with the majority of
the steam going to the neutralization operation.  It is during the neutraliza-
tion operation that lime is added to the dilute leach solution stream to effect
recovery of product iron and sulfate as a rusty gypsum.  The remainder of the
wash filtrate is concentrated via evaporation, mixed with the concentrated
leach solution, and recycled to the mix tank.  The steam generated during
evaporation/concentration is utilized within the battery limits process in
various process heating applications with the condensate being recycled as
cake wash water.  Utility purchased high pressure steam is used as the heat
source for evaporation.
     The water washed sink cake is next contacted with an appropriate organic
solvent, namely acetone, to effect elemental sulfur extraction from  the cake.
The solvent rich coal slurry is then subjected to either filtration  or centrif-
ugation, depending on specific process requirements.  The solvent -miscible
water -elemental  sulfur-sulfate sulfur solution is then separated from the

                                    164

-------
acetone wet cake.  The multi-component solution is then distilled under ele-
vated pressure and temperature such that molten elemental sulfur is recovered
for disposal.  The recovered acetone  is recycled  to  contacting while the sul-
fate laden water stream is sent to neutralization for sulfate removal and water
recycle to cake wash operations.
     The solvent wet sink cake is next sent  through  a drying operation where
all but very small quantities of the  acetone are  flashed from the cake, con-
densed and recycled to solvent contacting.   High  pressure utility steam is
used to heat the circulation inert gas (N2)  stripping media utilized in the
dryer.  The hot dry sink cake is next blended with the cooler water wet float
cake to produce a low pyrite, reduced ash, Gravichem Process clean coal
product.
8.1.3  Base Case Design^ Cost Estimate, and  Economics Approach
     During the course of the current RTU project, two base case conceptual
full scale process designs were generated.   Base  Case 1 is a process designed
to treat mine cleaned Martinka coal,  containing 1.51% w/w total sulfur (1.0%
w/w  pyritic sulfur) containing 2.3 Ibs S02/MM Btu to a sulfur content level
equivalent to 1.2 Ibs S02/MM Btu.  Base Case 2 is specific to the treatment
of TVA  Kentucky No. 9 coal.  The TVA  coal process objective was to decrease
the  initial 4.3% w/w total sulfur  (2.4% w/w  pyritic  sulfur) containing
6.94 Ibs S02/MM Btu coal down to an equivalent sulfur content of 4 Ibs S02/MM
Btu.
8.1.3.1  Design Bases-
     Presented in this section are the design bases  and related assumptions
which were utilized in generating  the two base case  conceptual full scale
Gravichem Process designs.  Unless otherwise noted,  the information deline-
ated below is common to both base cases.
     Feed Coal .
         Base Case  1
        •    14 mesh x 0 coal from  the Martinka Mine, Lower Kittanning seam.
        •   225 TPH total coal feed rate, chosen  such that 100 TPH is
            processed in leach reactor

                                    165

-------
   •   Coal  contains 1.51% w/w total  sulfur with 1% w/w pyritic
       sulfur
   t   12,900 Btu/lb HHV
   •   3% w/w moisture
   Base Case 2
   •   14 mesh x 0 TVA coal  from Hopkins County, Kentucky Number 9
       seam
   t   200 TPH total coal  feed rate such that 100 TPH is processed
       in each leach reactor
   •   Coal  contains 4.3%  w/w total sulfur with 2.4% w/w pyritic
       sulfur
   •   12,400 Btu/lb HHV
   •   10% w/w moisture
Product Coal.
   t   Base Case 1 combined float and sink coal to meet 1.2 Ib
       SOp/MM Btu specification
   •   Base Case 2 combined float and sink coal to meet 4.0 Ib
       S02/MM Btu specification
Mixing.
   •   Coal  slurry mixed to 25% w/w coal based on requirements for
       float/sink separation
   •   Coal  mixed at 215°F for 45-minute residence time in agitated
       vessel.  Based on RTU experience for adequate wetting and
       deaeration
   •   3 stage mixer to ensure residence time
   •   Leach solution contains 7.5% w/w Fe and 4% H2$04 resulting
       in a 1.3 specific gravity required for float/sink separation
Float/sink separation.
   •   One hour residence  time in separator at 176°F based on bench-
       scale studies discussed in Section 5
   •   Coal  fractons based on bench-scale experimentation (Section 5)
            Base Case 1 - 125 TPH float/100 TPH sink
            Base Case 2 - 100 TPH float/100 TPH sink
                               166

-------
       •    Sink  slurry is 33% w/w coal, float slurry is about 20% w/w
            coal  based on information from suppliers of coal cleaning
            equipment and bench-scale studies
     Reaction.  The net overall reaction between pyrite and ferric sulfate
leach solution  is represented by:

        FeS2 +  4.6 Fe2(S04)3 + 4.8 H20 + 10.2 FeS04 + 4.8 H2S04 + 0.8 S

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

The reaction rate, having second order dependence on the pyrite concentration
and fraction c
rate equation
and fraction  of total  iron as Fe+3, is represented by the following empirical
                              -d[W ]
                         rL - -1^ ' KL
where
     W  = wt % pyrite in coal  at time t
     Y  = fraction of total  iron as ferric ion at time t
     K|_ = leach rate constant  (function of temperature and  coal  top size)
     KR = reactor constant (function of reactor size,  configuration, etc.)
        Base Case 1
        •   6-hour residence time in reactor at 250°F  and 35  psig based on
            RTU experience and bench-scale determination of rate constants
            (Section 5 and Reference 11).   Reaction rates were found to be
            higher in the RTU  than those found at bench-scale thus requir-
            ing the use of KR  in the rate expression.
        •   Extent of reaction carried out in the mixer and separator  is
            based on rate constants determined at bench-scale adjusted
            for the respective temperatures.
        t   Extent of ash removal  on the leached coal  is based on the
            amount of pyrite reacted.
                                    167

-------
        Base Case 2

        t   6-hour residence time at 250°F and 35 psig based on bench-
            scale studies of TVA coal resulting in equal or faster reac-
            tion rates than Martinka coal.  Reaction rate constants were
            adjusted by KR assuming the same increase in rate would be
            observed at larger scale on the TVA coal as was observed on
            Martinka coal at RTU scale.

        •   Extent of reaction in mixer and separator are based on
            bench-scale determined rate constants adjusted for the
            operating temperatures of the mixer and separator.

        •   Extent of ash removed is based on the amount of pyrite
            reacted although bench scale data (Section 7) indicates
            a removal of other ash on the processed sink coal.  The
            implications of this difference are discussed in Section 8.3.1,

     Regeneration.  The regeneration of the leach solution is represented by
the following reaction:
            FeS04 + 0.5 H2S04 + 0.25 02 -> 0.5 Fe2(S02)3 + 0.5 H20


        AH = -18.6 Kcal/g-mole FeS04 = 0.0335 MM Btu/lb-mole FeS04


        •   Regeneration in the reactor at 250°F and 35 psig maintains  a
            Y = 0.95 based on previously determined bench-scale data(4>5),

        •   0.5-hour residence time in regenerator to maintain Y at 0.90
            for recycled leach solution at the mixer based on bench-scale
            studies(4,5).

     Filtration and cake washing.   Based on RTU and bench-scale experimentation
(Sections 7 and 8 and Reference 11)   the filter/washing operation is  modeled as

follows:

        •   At the interface on the filter drum between filtration and
            cake washing the coal  cake contains 50 pounds of water plus
            its associated salt per 100 pounds of coal.

        •   90% of the salt remaining on the cake at the interface is sur-
            face salt which is removed by washing.  The .other 10% of the
            salt is in the coal  pores and is not removed by cake washing,

        •   The minimum wash water required is equal to 1.4 times the
            amount of water on the coal  cake at the interface.
                                   168

-------
   •   The "dry" coal cake contains 50 pounds water plus the residual
       salt per 100 pounds of coal.

   •   Subsequent reslurrying of the coal cake with water dilutes the
       salt contained in the coal pores to a homogeneous solution of
       salt and water which can be removed by additional filtration
       and washing.

Evaporation.

   •   Water removal carried out to the extent that recycled leach
       solution contains 7.5% w/w iron.

   •   Evaporator operated at 290°F and 35 psig such the steam pro-
       duced can be utilized for process heating requirements.

Solvent contacting.

   •   The water washed coal cake is contacted with acetone to make
       up a slurry which is 40% by weight coal.

   •   Pore moisture and salt comes to equilibrium with the acetone,
       based on bench-scale experimentation (Section 7).

Solvent separation.

   Base Case 1

   •   Elemental sulfur on coal is reduced to 0.08% w/w based on
       bench-scale studies (Section 1).

   •   Final separation of coal by centrifuge reduces acetone on
       coal cake to 12% w/w based on vendor supplied information
       (Bird Machine Co.).

   Base Case 2

   •   Elemental sulfur on coal reduced to 0.2% w/w based on bench-
       scale studies (Section 1).

   •   Final separation of acetone from coal by rotary pan filter
       resulting in a coal cake containing 27% w/w acetone (Bird
       Machine Co.).

Drying.

   •   Requires 3 driers each with 54 trays and 60-minute residence
       time based on vendor testing and experience (Section 9).

   •   Inert gas temperature of 400°F heating coal to 225 F.
                                169

-------
        •    Residual  dry coal  acetone levels of 0.1% w/w were based on
            vendor testing  (discussed in Section 9) and additional manu-
            facturer  supplied  information.   While the vendor tests indi-
            cated residual  acetone levels of 0.5% w/w, Wyssmont Inc.
            personnel  indicated that a residual level of 0.1% w/w acetone
            would be  attainable through proper adjustment of dryer gas
            recirculating rate and temperature, and solids residence time.

     Solvent distillation.

        •    Column operated at 15 psig and 250°F to keep elemental sul-
            fur molten in bottoms.

        •    Column diameter sized to maintain 5 ft/sec vapor flow.

        •    Column height is 65 feet using 20 trays based on suppliers
            information (Artisan Industries).

     Neutralization.

        •    Stoichiometric  conversion of H-SO.  and iron sulfates to
            gypsum and iron oxides.

     Materials of Construction.

     Based on RTU experience,  materials studies and supplier recommendations

the following materials were determined suitable for the various leach solution-

coal services:

        t    Fiber reinforced plastics and elastomers up to 180°F.

        •    316 L stainless up to atmospheric boiling temperatures (215°F).

        •    Hastelloy C and rubber/acid brick lined carbon steel for ser-
            vice up to 265°F.

        •    Titanium for service up  to 300°F.

8.1.3.2  Battery Limits Process Costs--

     For each of the  base case designs generated, a battery limits capital and
operating  cost was determined.  Equipment lists, tabulating the required major
equipment  were generated.  The associated equipment numbers presented on the
lists are  keyed to the process flow  diagrams.  The selected equipment was sized
to approach the optimum cost for processing the specific coal  to the desired
product sulfur content.

     Battery limits capital cost.  Capital  equipment costs adjusted to mid-

1978, were obtained from various sources:  technical literature, equipment
suppliers  and internal (TRW) costing data.   The specific sources of data for
                                    170

-------
the various classes of equipment are presented in Table 35.  When cost data
were obtained from literature or other noncurrent sources, appropriate cost
escalation factors, based on the Marshall and Swift Equipment Cost Index (to
escalate costs from date of publication to mid-1978), were applied.  The
capital  equipment cost is also tabulated on the equipment list.  The costs
are presented in terms of FOB equipment cost and installed equipment cost.
The FOB equipment cost is the base, uninstalled cost at point of manufacture
or point of shipment.  The installed equipment cost includes the following
elements:
        t   FOB Equipment Cost
        •   Field Materials
            -  Equipment
            -  Piping
            -  Concrete
            -  Steel
            -  Instruments
            -  Electrical
            -  Insulation
            -  Paint
        •   Material Erection
        •   Direct Field Labor
        •   Indirect Costs
            -  Freight
            -  Taxes
            -  Construction Overhead
            -  Fringe Benefits
            -  Labor Burden
            -  Field Supervision
            -  Temporary Facilities
            -  Construction Equipment
            -  Small Tools
            -  Miscellaneous Field Costs
            -  Contractor Engineering
The installed equipment cost does not include a contingency factor.
     Battery limits process operating cost.  The process operating costs have
also been estimated for each of the base cost designs.  The basis for these
estimates was technical literature and informal supplier quotes.  Specific
sources of information are presented in Table 36.  The total estimated battery
limits processing cost was determined as presented in Table 3Z.
                                     171

-------
              TABLE 35.  SOURCES OF EQUIPMENT COST INFORMATION
                Equipment Type
 Information  Source
       Hoppers
       Conveyors
       Mixers
       Pumps
       Reactors
       Vessels >40,000 gal
       Vessels <40,000 gal
       Tanks >40,000 gal
       Tanks <40,000 gal
       Drums
       Centrifuges and Support Equipment
       Filters and Support Equipment
       Heat Exchangers
       Evaporators
       Dryers and Support Equipment
       Rotary Valves
       Separators
TRW Data
Reference 13
Reference 14
Reference 15
TRW Data
TRW Data
Reference 15
TRW Data
Reference 15
Reference 15
Bird Machine Company
Ametec Company
Reference 15
Reference 16
Wyssmont Incorporated
Reference 15
TRW Data
8.1.3.3  Overall  Process Economics--
     The overall  Gravichem Process economics were evaluated for each of the
base cases.  The battery limit plants were incorporated with their associated
off-sites so that integrated grass roots processing facilities could be defined,
The following paragraphs describe the overall integrated Gravichem process
model as well as the bases for the economic analyses performed.
                                    172

-------
              TABLE 36.   SOURCES  OF OPERATING COST INFORMATION
.	:	Cost  Element	Information Source

    Maintenance,  Insurance, Taxes  and  Interest         Reference  14
    Labor Requirement  (Number  of Positions)             Reference  14
    Labor Cost                                          TRW  Data
    Utilities
         Electrical Power                               TRW  Data
         Cooling  Water                                 TRW  Data
         Process  Water                                 Reference  14
    Materials
         Oxygen                                         Linde, Division of
                                                        Union Carbide
         Nitrogen                                      Linde, Division of
                                                        Union Carbide
         Sulfuric Acid                                 Reference  17
         Lime                                           Reference  17
         Acetone                                        Reference  17
         Waste Disposal                                 TRW  Data

    Integrated facility  description.   The  integrated desulfurization facility
includes all battery limit  Gravichem equipment  plus the  required off-sites.
The  off-sites include such  items as:
       •    Feed and product coal storage,  handling and  transport
            equipment.
       t    Shallow physical coal cleaning  facilities and size separation
            equipment.
       •    By-product  handling  and storage facilities.
       t   Waste  treatment (physical cleaning  and process generated)
           and storage facilities.
       •    Process water treatment, storage and pumping facilities.
       •   Cooling water treatment and pumping equipment.	

                                    173

-------
           TABLE 37.  BATTERY LIMITS OPERATING COSTS FORMAT
Capital Related Costs:                              Annual Cost, $1000
     Depreciation - IQ% straight line                      XXXX
     Maintenance, insurance, taxes                         XXXX
     interest - 15% of capital
Labor:
     Labor, XX operating positions                         XXXX
            at $160,000/position
Utilities:
     Electric power, XXX Kw                                XXXX
          at 5^/Kw-hr
     Cooling water, XXXX gpm,                               XXX
          30°F rise at 5<£/1000 gal
     Process water, XX gpm                                   XX
          at 25^/1000 gal
     Heating, XX MM Btu/hr, fired duty,
          XX TPH coal product equivalent
Materials:
     Oxygen, XX TPH at $50/ton                              XXX
     Nitrogen, XX TPH (included in oxygen cost)
     Sulfuric acid, XX TPH at $55/ton                         X
     Lime, XX TPH at $35/ton                                  X
     Acetone, X TPH at $320/ton                               X
     Waste disposal, X TPH at $6/ton                    	X
TOTAL COST                                               XXXXXX

    •   Power and steam generation facilities
    •   Site office buildings and shop structures.
    t   Other site improvements such as roads, fences, railroad spurs,
        etc.
                                  174

-------
It should  be  noted that the economic evaluations presented herein do not include
land costs and assume that oxygen and nitrogen is purchased as an over-the-fence
utility  item  (i.e., neither battery limit nor off-site equipment include an oxy-
gen plant).   For an integrated process facility of this type, it has been esti-
mated  that the installed off-site capital cost will be equivalent to approximately
50% of the installed battery limits
     Process economics model .  The technique used in capitalizing the integrated
Gravichem process is the primary determinant in calculating the required unit
price of the product.  Due to the nature of the Gravichem process (clean energy
production), only utility financing methods were considered in this evaluation.
The specific method utilized in this analysis  is based on the technique used by
the Federal Power Commission Synthetic Gas-Coal Task Force in their report on
             MO)
synthetic gasv   .  The recommended economic evaluation criteria are given in
Table 38.
     The annual i zed revenue expressions for utility financing which were devel-
oped by the FPC Task Force can be combined to  give a simplified expression of
the following form:
    ^
                       P =  (0.140X + 0.121Y +  1.006Z)/E

where ,
     P = is the required sales price for processed coal, $/MM Btu
     X = working capital for raw materials and supplies, $
     Y - sum of the total plant investment and start-up cost, $
     Z = annual total operating cost,  $/year
     E = annual energy output, MM Btu/yr
When this expression is utilized in conjunction with the criteria presented in
Table 38, the required sales price for the processed coal is obtained.  From
the required sales price and the cost  of the feed coal, an equivalent upgrading
cost can be determined.
                                      175

-------
         TABLE 38.   ECONOMIC EVALUATION CRITERIA UTILITY FINANCING*
Operating Cost Criteria
     Raw material  - coal  at $15/ton, $20/ton, $30/ton
     Utilities
        Electricity at 5^/kw-hr
        Oxygen at $50/ton (nitrogen included in cost)
        Cooling water at 5£/1000 gal.
        Process water at 25
-------
                             TABLE  38.   (Continued)
Capital Cost  Criteria
     Battery  limits  capital as discussed in text
     Off-site capital at 50% of battery limits capital
     Overhead and profit at 22% of battery limits + off-sites
     Engineering and design at 10% of battery limits + off-sites
     Contingency at  15% of battery limits + off-sites + overhead and profit +
     engineering and design
     Total  plant investment = sum of battery limits + off-sites + overhead and
     profit + engineering and design + contingency
     Interest for construction at 9% of total plant investment x 1.875
     Start-up cost at 20% of total operating costs
     Working  capital = sum of raw materials inventory of 60 days at full rate
     + materials and supplies at 0.9% total plant investment + 1/24 annual
     product  revenue

8.2  DESIGN AND COST ESTIMATE BASE CASES
     The  following pages present the results of two process engineering studies
relating  to Gravichem application to two specific coal feedstocks.  The coal
feedstocks  were mine cleaned Martinka Coal and TVA Kentucky No. 9 coal.  Each
of the base case designs are presented in terms of a brief process description,
process flow  diagrams, a major equipment list, battery limits process costs,
and overall process  economics.  The detailed mass balance and their associated
flow diagrams as well as steam balances for each of the design cases are presented
in Appendix C.
8.2.1  Base Case 1 -Mine Cleaned Marti nka Coal
     Base Case 1 deals with a conceptual full-scale Gravichem Process design
based on a processing facility capable of treating coal fed at a rate of 225
TPH from the  Martinka Mine, Lower Kittanning seam.  The coal has an "as
received" heating value of 12,900 Btu/lb.  The coal contains 1.51% w/w total
sulfur with a ayritic sulfur level of 1.0% w/w.  Process flow diagrams  (Dwg.
Nos. 2473-3 and 4) for the design are presented  as Figures  35a and 35b.   Pre-
sented on the flovf diagrams are equipment, key process  temperatures and pres-
sures, flow rates and composition of selected  coal streams, and flow rates for
steam and raw materials.
                                     177

-------
00
                                                                                                      WASH FltJWrE   CONTACIO»    CONT*CIO«   FILTIATE  fILTM    WMH FILl
                                                                                                      MCEIVH     M)XCK                (ECEIVIR         «CEIVt«
                    3NS/HK
                    3TAL SULFUR, %


                     SULFATt SULFUR. %
                     ELEMENt L SIAfUR, 1i
                   MOISTURE, %
                   HEM VALUE. BT
                   LB SO./W Wl
                                 M COMKSITlONS (DitY aAilSl_
                                                                              _P-Lr-z_p-a_p-i_      p^iLfiiS.     F-IO
                                                                              SLUBRV   RfACTOFL   REACTOR   LEACH    LEACH    WASH    FILTER   FILTRATE   WASH W,
                                                                              FEED    RED     DISCHARGE  SOLUTION   FILTRATE   FILTRATE  FEED    PUMP     FEED PUT
                                                                              PUMP    PUMP     PUMP     FEED PUMP   PUMP    PUMP    PUMP
                                                                                                                                                                  WCQAL DESLRFUKiZATIOl
                                                                                                                                                                   GRAV10HEM PBOCESS
                                                                                                                                                                              NUMBM, W3-0)
                                                         Figure  35a.    TRW   Coal   Desulfurization  Gravichem  Process

-------
10
                              —   -  -blS.       £3!      «->    £d   T-IL    T-ia       coi      T-U
                              ntJMK  WASHFILIRATE  CONTACTOR   SLUWIV   flLTIR  TfLHATf  CONTACTOR   CENTRIFUGE   CENTW1
                              MCIIVIR  naiwt              COOLER       utxiv»                   RECBVEI


                              JErf     JS-i     t-\^    T-W^      T-17     Ma_      T;W_        MU      .._      ..._,       ^^^          ___    ^^

                              IIRimR   ACETONE  SIWP«»  STRIPKR     STRIfft*   NEUTULIICR  AGO NEUTkAUZK  CONTACTOR  CONTACTOR  NEUTRALIZE!!  CARE DM ADSORPTION  SOUBMK  STRIPPER
                              PtEIKAIER  JTRIfPER   REBDILER CONDENSATE  BOTTOMS  RED COOLER  TANK        MIXIR     MIXER     MIXED      OKUM                 OVHHIAO
5K5T  OK&  CVCIONI  S^RHWO   ^NDEHSAJE
                  COMPENSER  RECEIVE ft
SWAM NUMBED
TONS/™
TDTAlSUlfUlLIt
PVDTIC SUIPJR. %
OICANIC SULFUK, %
SIX FATE iltiFUS %

^o^feii™AI
i
:u

LIB
b
ITS

"ifS
7
'w
.08

"ijS
9
1
SKS
'^
                                                                                                                                   SS3"""
                                                                                                                                                                                          IKW COAl WSULFUHIIATIO.

                                                                                                                                                                                            GMVICHEM PftOCISi
                                                          Figure  35b.     TRW  Coal   Desulfurizatlon   Gravichem  Process

-------
8.2.1.1  Process Description--
     Feed and mixer.  Crushed coal, nominally  14 mesh  top-size,  is fed from
feed hopper A-l to the mix tank T-l by conveyor C-l.   The  coal,  containing
1.0% w/w pyritic sulfur and 3% w/w moisture is fed at  a  rate of  225 TPH on a
dry basis.  Recycled leach solution at its boiling point (215°F)  is introduced
to the first mixer stage after passing through the knock-out drum T-2.   Flash
steam from T-3 and low pressure steam are used to heat the coal  on the conveyor
to 215°F.  The mixer T-l was sized for three stages at 0.25-hour residence time
per stage to provide adequate wetting of the coal and  partial  reaction  of  the
pyri te.
     Separator.  The slurry from the mixer which is 25%  w/w coal  is pumped
through the slurry cooler to the separator S-l.  The slurry is cooled  to 176°F
in the slurry cooler E-l.  The separator provides 1-hour residence time where
the heavy pyrite rich coal will settle and the lighter low pyrite  coal will
float in the 1.3 specific gravity leach solution (see  Section  7).   Based on
current bench scale desulfurization studies, about 55% of  the  coal  feed will
leave the separator as the float portion in a slurry which  is  about 21% w/w
coal.  After washing to remove leach solution, this coal will meet the total
sulfur specification without further pyrite removal.   The  sink portion, con-
taining 45% of the coal feed, is pumped to the primary reactor, R-l, as a  33%
w/w coal slurry for further pyrite reduction.
     Float coal washing and dewaterinq.  The float slurry  from the separator
S-l containing about 125 TPH coal  is filtered, washed  and  recontacted with
water in a 3-stage countercurrent arrangement of .three filters (F-l, F-2 and
F-5) and two contactors (J-6 and T-22).  The coal cake from the last stage con-
tains 33% w/w moisture and residual sulfate sulfur has been reduced to less
than 0.01% w/w coal.  The coal cake from this filter is  the product coal from
the float portion which will be recombined with the sink coal  after it is  pro-
cessed for pyrite removal.  The float coal  contains 0.83% w/w  total  sulfur with
a heating value of 14,290 Btu/lb,  which is below the specification of 1.2  Ib
S02/106 Btu.
     Primary reactor.   The sink slurry from the separator  containing about
ICO TPH coal is pumped to the primary reactor R-l where  the leach  reaction for
the removal  of pyrite  takes place.  The reactor was sized  to provide 6-hour
                                     180

-------
residence-time  at 250 F and 35 psig oxygen  partial  pressure  which,  based on
RTUexperimentations, will adequately remove  pyrite to  a  level  of 0.15% w/w
pyritic  sulfur  in the coal (see Section 6).   The  reactor  is  maintained at
250°F by the generation of heat from heats  of reaction  and regeneration and the
injection  of 300 psig saturated steam into  the reactor.   The reactor  is divided
into 5 stages each with a mixer.  Simultaneous regeneration  of  the  leach solu-
                              +3
tion to  maintain a ratio of Fe   to total iron equal  to 0.95 is accomplished
by feeding about 2 TPH oxygen to the reactor.   The  inert  gas leaving  the
reactor is vented through a knock-out drum  T-4 where  it is contacted  with
recycled leach solution before venting to the atmosphere.
     Sink coal  filtration and washing.  The slurry  effluent  from the  reactor
is flashed to about 1 psig and then pumped  to the rotary  drum vacuum  filter
F-3 where the leach solution is separated and the coal  cake  washed  to remove
residual sulfate salts.  The wash water for the filter  is recycled  water from
the wash water hold tank T-8.  Make-up water  from the pond is required.  The
coal cake from F-3 contains 33% w/w moisture.
     Leach solution recycle loop.  The wash filtrates from filters  F-l and F-3
containing iron sulfates are collected and  fed to the evaporator EV-1 for water
removal.  In order to remove the excess iron  produced from the  leach  reaction
a portion of the wash filtrate F-3 is sent  to the lime  neutralizer  for disposal.
The evaporator feed stream is heated to 233°F in  the  feed/bottoms exchanger
E-3.  The evaporator is operated at 35 psig and about 290°F  such that the water
removed as steam may be used for process heating  service. The  leach  solution
.in  the evaporator bottoms is concentrated in  iron to  the  extent that  the
recycled leach solution will be 7.5% w/w total iron at  the mixer T-l.  The
filtrate from filter F-l is heated to 250°F in exchanger  E-2 and fed  to the
regenerator R-2.  The regenerator R-2, operating  at 250°F and 35 psig, is
designed for 0.5-hour residence time to regenerate  the  F-l filtrate such that
the combined recycle streams will have a ratio of Fe+3  to total iron  equal to
0.90 at the mixer T-l.  The oxygen requirement for  the  regenerator  is about
0.25 TPH.  The evaporator bottoms, regenerator effluent,  F-3 filtrate and the
makeup sulfuric acid are combined and fed to  the  knock-out drum T-4,  then
pumped to the mixer T-l through knock-out drum T-2, thus  completing the
recycle loop.

                                     181

-------
     Acetone washing and elemental sulfur removal.  The coal cake  from filter
F-3 containing elemental sulfur produced in the leach reaction  and residual
sulfate salts from the leach solution is contacted with about 100  TPH  of ace-
tone in the contactor T-ll.   The coal slurry is then filtered on a rotary pan
filter F-4 where the cake is washed with acetone.  The coal cake from  F-4 is
again recontacted with acetone (110 TPH) in T-13 and the slurry fed to a  centri-
fuge CG-1 where the resulting coal cake is reduced to about 12% w/w acetone.
Bench-scale experience with two-stage acetone washing shows that the elemental
sulfur on the coal is reduced to 0.08% w/w (see Section 7).  The sulfate  sul-
fur on the cake is reduced to about 0.02$ w/w after the filter  F-4.  All  of
the acetone contacting equipment are vented to the atmosphere through  a water
scrubber GS-1 and carbon bed T-20 for residual acetone removal.
     Acetone recovery and sulfur removal.  The acetone streams  from  filter F-4
and the centrifuge CG-1 which contain elemental sulfur, water,  and salts  are
preheated in E-9 and fed to the acetone stripper SS-1.  The overhead vapor
from the stripper, operating at 250°F and 15 psig, containing essentially all
of the acetone feed is condensed and recycled to the contactor  T-ll.   The
stripper bottoms containing water, elemental sulfur and salts is fed to the
stripper bottoms separator T-17 where the elemental sulfur is removed.  The
sulfate rich solution is cooled by E-12 and fed to the neutralizer.
     Lime Neutra1ization.  The sulfate rich streams from the acetone stripper
and the wash filtrate from F-3 are contacted with a lime slurry in the acid
neutralizer T-18 where the sulfuric acid is neutralized and the salts  reacted
to produce gypsum.  The gypsum slurry is dumped to a settling pond for disposal,
Following settling, the pond water is returned to filter F-3 as part of the
wash water following water treatment to remove any trace acetone.
     Coal drying.  The coal cake from the centrifuge CG-1 is fed through  a
rotary valve RF-2 to the coal drier D-l where the coal is heated to 225°F by
a 400°F nitrogen gas stream.  The drier gas system is a closed  loop with  a
condenser to remove the acetone, a recycle gas blower B-l, and  two steam
heaters E-7 and E-8 to heat the recycled nitrogen stream back to 400°F.   The
recovered acetone is returned to the system for further use.  The  dried coal
from the drier is discharged through a rotary valve RV-3 and combined  with  the
coal product from the float portion.  The coal from the drier contains 0.69%

                                     182

-------
w/w total  sulfur and has a heating value  of  11,570  Btu/lb.   The combined  coal
product contains 0.77% w/w total sulfur with a  heating  value of 13,080  Btu/lb
resulting  in 1.18 Ib S02/106 Btu material.
8.2.1.2  Battery Limit Costs—
     Equipment list.  The previous paragraphs of this document presented  a con-
ceptual process design and flow diagram (Figures 35a and 35b)  for a  battery
limits processing facility treating  225 TPH  of  feed coal (221.8 TPH  of  desul-
furized coal product).  An equipment list, tabulating the required major  equip-
ment, is presented in Table 39.  The selected equipment was  sized to process
the 1.5% total sulfur feed coal such that the combined  Gravichem product  would
contain no more than 1.2 Ibs S02/MM  Btu.  As may be seen in  Table 39, the total
battery limits FOB equipment cost  is 9.6  million dollars while the total
installed equipment cost is 17.4 million  dollars.
     Battery limits process operating cost.   Process operating costs have been
estimated as presented in Table 40.   This estimated total  annualized operating
cost results in an equivalent battery limit  production  cost  of $6.61/ton  of
coal product.  This production cost  is based on an  apparent  product  coal  rate
of 209.5 TPH (221.8 TPH of desulfurized product less 12.3 TPH  for process
steam generation).  The battery limit processing facility therefore  operates
with an overall coal yield of 93%  on a dry weight basis and  an overall  energy
yield of 94% on a Btu basis.  Also of interest  is the net overall  plant energy
efficiency when the coal required  for producing the electric power loads  for
the process and oxygen plant is accounted for along with the plant steam
requirements.  This efficiency, assuming  a 33%  efficiency for  electric  power
from coal, is equal to 93.5% where the coal  requirement for  power production
is 2.1 TPH.  Another number of interest is the  noncapitalized  battery limits
operating cost.  This is essentially the  total  process  operating cost less the
annual depreciation charge (10% of installed equipment).  The  noncapitalized
battery limit operating cost is $5.18/ton of coal feed  (20*/MM Btu equivalent)
or $5.57/ton of coal product (21tf/MM Btu  equivalent).   The noncapitalized oper-
ating cost represents a pure operating cost  of  the  battery limit plant.
                                     183

-------
                                   TABLE 39.  BASE CASE  1 - PROCESS  EQUIPMENT LIST
00
.£»
MAJOR
A-l
B-l
C-l
CG-1
CY-1
D-l
E-l
E-2
E-3
E-5
E-6
E-8
E-9
E-10
E-ll
E-12
EV-1
PROCESS EQUIPMENT $9.64 MM FOB, $17.39 MM INSTALLED
Feed Hopper - 11,000 ft3
Recycle Gas Blower (3) - CS, 1.7 Compression Ratio, 40 hp
Feed Conveyor - 40 in. wide x 20 ft x 10 hp, 200 ft/min.
Centrifuge Package (2) - 44" * x 132" Screen Bowl, CS, 200 ph
Cyclone (3) - 5 psig, 850 ft3, CS
Coal Dryer (3) - 30' $ 55', CS, 5 psig, 15 hp, 54 trays, 60 min.
Slurry Cooler - 3550 ft2, CS/SS
Regenerator Feed Heater - 1840 ft2, CS/Hastelloy C
Evaporator Feed/Bottoms Exchanger - 4120 ft2, Hastelloy C/Hastelloy C
Slurry Cooler - 3400 ft2, CS/SS
p
Dryer Overhead Condenser (3) - 700 ft , CS/CS
Recycle Gas Heater (3) - 602 ft2, CS/CS
Stripper Preheater - 125 ft2, CS/SS
Stripper Reboiler - 7,560 ft2, Hastelloy C
Stripper Overhead Condenser (2) - 9,420 ft , CS/CS
Neutral izer Feed Cooler - 800 ft2, CS/SS
Evaporator - 2000 ft , 35 psig, CS with titanium clad, 20 hp
$K
FOB
22

16
599

2100
80
50
112
78


13
80
260
31
900
$K
Installed
26
as D-l
27
1197
as D-l
2625
217
136
303
212
as D-l
as D-l
48
223
780
90
1440
(Continued)

-------
oo
in

F-l
F-2
F-3
F-4
F-5
GS-1
M-l A/C
M-2 A/E
M-3
M-4
M-5
M-6
M-7
P-l
P-2

Rotary Drum Vacuum Filter Package (2) -
Rotary Drum Vacuum Filter Package (2) -
Rotary Drum Vacuum Filter Package -
Rotary Pan Filter Package (2) - 24' ,
Rotary Drum Vacuum Filter Package (2) -
Scrubber - 5'  x 30' , 20 trays, 0 psig
Mix Tank Mixers (6) - 25 hp, SS
Reactor Mixers (5) - 190 hp, Hastelloy
Contactor Mixer (2) - 25 hp, SS
Contactor Mixer - 50 hp, SS
Contactor Mixer (2) - 20 hp, CS
Neutral izer Mixer - 15 hp, SS
Contactor Mixer (2) - 25 hp, SS
Slurry Feed Pump (2) -1500 gpm, 15 psi,

12' $ x 16', 608 ft2, SS,
130 hp
12' <(> x 16', 608 ft2, SS,
130 hp
12 <() x 24, 912 ft2, SS,
250 hp (inc. vac. sys.)
445 ft2, SS, 105 hp (inc. vac.
sys.)
12'  x 16', 608 ft2, SS,
130 hp
, CS

C





15 hp, SS
Reactor Feed Pump - 1000 gpm, 15 psi, 10 hp, SS
$K
FOB
247
247
184
941
247
13
57
194
20
13
8
7
20
15
5
$K
Installed
392
392
292
1491
392
21
92
315
35
21
14
11
35
45
15
                                                     (Continued)

-------
                                            TABLE  39.  (Continued)
                                                                                         $K            $K
                                                                                         FOB        Installed
       P-3      Reactor Discharge Pump - 1000 gpm, 15 psi, 10 hp, SS                        5           15
       P-4      Leach Solution Feed Pump (2) - 1100 gpm, 15 psi, 15 hp,  SS                  8           21
       P-5      Leach Filtrate Pump (2) - 600 gpm, 65 psi, 30 hp, SS                           as F-l
       P-6      Wash Filtrate Pump (2) - 300 gpm, 70 psi, 15 hp, SS                            as F-l
       P-7      Filter Feed Pump (2) - 550 gpm, 10 psi, 5 hp, SS                            4           13
       P-8      Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS                                 as F-2
£      P-9      CaS04 Slurry Pump - 320 gpm, 4 hp, SS, 5 psi                                3            8
       P-10     Wash Water Feed Pump - 825 gpm, 15 psi, 10 hp, CS                           4           13
       P-ll     Leach Filtrate Pump - 500 gpm, 5 psi, 1.5 hp, SS                               as F-3
       P-12     Wash Filtrate Pump - 250 gpm, 10 psi, 2 hp, SS                                 as F-3
       P-13     Contactor Slurry Pump - 1000 gpm, 15 psi, 15 hp, SS                         6           17
       P-14     Filtrate Pump (2) - 350 gpm, 7.5 psi, 3 hp, SS                                 as F-4
       P-15     Contactor Slurry Pump (2)  - 550 gpm, 7.5 psi, 6 hp, CS                      3           n
       P-16     Acetone  Centrate Pump (2)  - 450 gpm, 7.5 psi, 5 hp, CS                         as CG-1
       P-17     Acetone  Return Pump (3) -  20 gpm, 0.5 hp, CS, 15 psi                           as D-l
       P-18     Scrubber Water Return Pump - 150 gpm, 10 psi, 2 hp, CS                      2           16
                                                   (Continued)

-------
                                              TABLE 39.   (Continued)
CXI

P-19
P-20
P-21
P-22
P-23
R-l
R-2
RV-1
RV-2
RV-3
S-l
SS-1
T-l
T-2
T-3
T-4
T-5

Acetone Condensate Pump - 1400 gpm, 15 hp, CS5 15 psi
Wash Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS
Filter Feed Pump (2) - 550 gpm, 10 psi, 5 hp, SS
Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS
Wash Filtrate Pump (2) - 250 gpm, 7.5 psi, 2 hp, SS
Reactor - 23' $ x 115', CS with acid-brick lining, 45 psig
Regenerator - 10.5' 4. x 63', 50 psig, CS with acid-brick lining
Rotary Valve (2) - 1 hp, 24" x 24", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Separator - 0.5 hr Residence Time
Acetone Stripper - 13' <|> x 65', SS, 20 Trays, 15 psig
Mix Tank (2) - 21' 4 x 41', SS, 0 psig, 102,000 gal
Knock-Out Drum (2) - 3200 gal, SS, 0 psig
Flash Drum - 6100 gal, 7' 4 x 21', SS, 2 psig
Knock-Out Drum - 13500 gal, 81 4 x 24' , SS, 0 psig
Wash Filtrate Receiver (2) - 1600 gal., Vac, SS
$K
FOB
5

4


808
219
39
25
25
855
192
288
38
40
42

$K
Installed
15
as F-2
13
as F-5
as F-5
1975
534
43
27
27
1710
327
504
87
91
81
as F-l
(Continued)

-------
                                            TABLE 39.   (Continued)
                                                                                        $K            $K
                                                                                        FOB         Installed
       T-6      Contactor (2)  -  25,000 gal,  0 psig,  SS                                     130          260
       T-7      Filtrate Receiver (2)   -  1600 gal, Vac,  SS                                     as F-2
       T-8      Wash Water Hold  Tank  - 72,000 gal, 0 psig,  FRP                             50          100
       T-9      Filtrate Receiver - 2000  gal, Vac, SS                                         as F-3
       T-10     Wash Filtrate  Receiver -  2500 gal, Vac,  SS                                     as F-3
       T-ll     Contactor - 40,000 gal, 0 psi,  SS                                          81          185
       T-12     Filtrate Receiver (2)  - 1300 gal, Vac, SS                                      as F-4
       T-13     Contactor (2)  -  20,000 gal,  0 psig,  CS                                     33            66
£     T-14     Centrate Receiver (2)  - 1300 gal, Vac, CS                                      as CG-1
00
       T-15     Dryer Condensate Receiver (3) -  140  gal, CS,  0  psig                            as D-l
       T-16     Stripper Condensate Receiver -  9500  gal, CS,  15 psig                         9            17
       T-17     Stripper Bottoms Separator - 2700 gal, Hastelloy C,  15  psig                19            43
       T-18     Acid Neutralizer Tank  - 11,000  gal,  FRP, 0  psig                             7            15
       T-19     Filtrate Receiver (2)  - 3000 gal, Vac, SS                                      as F-l
       T-20     Carbon Adsorption Drum -  5'   x  10',  FRP, 0 psig                            3            10
       T-21     Wash Filtrate  Receiver (2) - 1600 gal, Vac, SS                                 as F-2
       T-22     Contactor (2)  -  25,000 gal,  0 psig,  SS                                     130           260
       T-23     Filtrate Receiver (2)  - 1600 gal, Vac, SS            ~                          as F-5
       T-24     Wash Filtrate  Receiver -  1600 gal, Vac,  SS                                     as F-5
       *CS/SS - Carbon Steel Sheet/Stainless Steel Tubes

-------
             TABLE 40.  BATTERY  LIMIT  PROCESS  COSTS - BASE CASE 1
   Capital Related Costs:                                Annual Cost, $1000
        Depreciation -  10%  straight  line                        Ij739
        Maintenance, insurance,  taxes,                          ? cnq
        interest - 15%  of capital                                '
   Labor:
        Labor, 13 operating positions                           2 nsn
               at $160,000/position
   Utilities:
        Electric power, 2850 Kw  at 5^/Kw-hr                     1,128
        Cooling water,  16,700 gpm, 30°F rise                      428
                        at 5^/1000 gal
        Process water,  216 gpm,  at 25^/1000 gal                    26
        Heating, 321 MM Btu/hr,  fired duty, 12.3 TPH
                 coal product equivalent
   Materials:
        Oxygen, 2.4 TPH at $50/ton                                940
        Nitrogen, 0.25  TPH (included in oxygen cost)
        Sulfuric acid,  1.4 TPH at $55/ton                        615
        Lime, 2.6 TPH at $35/ton                                  728
        Acetone, 0.1 TPH at $320/ton                             253
        Waste disposal,  9 TPH at $6/ton                          428
   TOTAL COST                                                  10,974

8.2.1.3  Process Economics--
     The equipment  list  (Table 39)  presented earlier shows  that a 225 TPH bat-
tery  limit Gravichem  process train  has an installed cost estimated to be 17.4
million dollars.  That battery limit plant when incorporated with its associ-
ated  off-sites  represents  an integrated grass  roots coal  desulfurization facil-
ity.   The  off-sites for  a  plant of  this  type are estimated  to  be approximately
50% of the installed  battery limits  cost^  or in  this  case 8.7 million dollars.
                                     189

-------
     The results of the integrated grass roots Gravichem  process  economic eval-
uation are presented in Table 41.  Calculations were  performed using three
assumed run-of-mine (ROM) coal costs; $15, $20 and  $30  per  ton.   These values
were selected since they represent the broad range  of currently reported ROM
coal costs ($15/ton at mine mouth to $30/ton reported delivery price at some
plant sites).
     As may be determined from the data (Table 41), the required  market value
of the treated coal product ranges from $1.01/MM Btu  with $15/ton ROM feed
coal to $1.66/MM Btu with $30/ton ROM feed material.  An equivalent upgrading
cost is found by deducting the cost of the ROM "dirty"  coal energy ($Q.58/MM
Btu, $0.78/MM Btu, and $1.16/MM Btu at $15, $20 and $30/ton ROM coal  respec-
tively).  The upgrading costs ranged from $0.43/MM  Btu  to $0.50/MM Btu when
ROM coal feed varied from $15/ton to $30/ton.
8.2.2  Base Case 2 - TVA Kentucky No. 9 Coal
     Base Case 2 involves the desulfurization of TVA  Kentucky  No.  9 coal  via
the Gravichem Process.  The conceptual design described is based  on a  coal feed
rate of 200 TPH for TVA coal from Hopkins County, Kentucky Number 9 seam, with
a heating value of 12,400 Btu/lb.  The coal contains  4.3% w/w  total  sulfur with
a pyritic sulfur level of 2.4% w/w.  Process flow diagrams (Dwg.  Nos.  2473-5
and 6) for the design are presented as Figures 36a  and  36b.  The  flow  diagrams
indicate equipment, key process temperatures and pressures, flow  rates and
compositions of key coal streams, and steam and raw material flow rates.
8.2.2.1  Process Description--
     Feed and mixer.  200 TPH (dry basis) coal are  fed  to the  mixer T-l from
the feed hopper A-l by conveyor C-l after heating to  215°F on  the conveyor by
flash steam and low pressure process steam.  At the mixer the  coal  is  contacted
with recycled leach solution at 215°F where the 0.75-hour residence time,  in
three stages, provides thorough wetting of the coal and partial reaction of the
pyrite.  The feed coal contains 10% moisture and 2.4% w/w pyritic sulfur on a
dry basis.
     Separator.  The mixer slurry containing 24% coal is pumped to the sepa-
rator S-l through the slurry cooler E-l where it is cooled to  176°F.   The
                                      190

-------
              TABLE 41.
BASE CASE 1 - Pi?nrFSS Frflf)n
 Annual Product, 1.66 MM Tons/yr
 Energy Value, 43.4 x 106 MM  Btu/yr
 Capital Related Requi rements. $MM
 Battery Limit Capital
 Off-site Capital
 Overhead and Profit
 Engineering and Design
 Contingency
 Total  Plant Investment**
 Interest for Construction
 Start-up Costs
 Working Capital  (Utility Financing)
 Total  Capital  Related Costs (Utility)
 Operating Costs. $MM/Yr
 Raw Material  (Coal)
 Chemicals (Lime, Sulfuric Acid,  Acetone)
 Supplies
 Disposal
 Utilities
 Labor  (19 Positions)
 Taxes and Insurance
 Total Operating  Costs
 Required Coal Market Price, $/MM Btu
 Utility Financing
Upgrading Cost, $/MM Btu
Utility Financing

$15/Ton
17.39
8.70
5.74
2.61
5.17
39.61
6.68
7.40
7.04
60.73
26.70
1.60
0.88
0.43
3.09
2.83
1.47
37.00
ROM Coal Cost
$20/Ton
17.39
8.70
5.74
2.61
5.17
39.61
6.68
9.18
9.05
64.52
35.60
1.60
0.88
0.43
3.09
2.83
1.47
45.90

$30/Ton
17.39
8.70
5.74
2.61
5.17
39.61
6.68
12.74
13.07
72.10
53.40
1.60
0.88
0.43
3.09
2.83
1.47
63.70
                    1.01
                    0.43
1.22
0.44
1.66
0.50
 * Assumes,  debt/equity = 75/25, interest on debt = 9%, return on
  equity =  15%.
**Equivalent to  a plant capital investment of $68/kw.
                                 191

-------
separator is sized for one-hour residence time where  the  heavier high pyrite
coal is separated as the sink fraction in the 1.3 specific  gravity leach solu-
tion and the light low pyrite coal goes to the float  fraction.   50% of the
coal separates as the sink portion in a slurry containing 33% w/w coal based
on current float/sink studies using TVA coal (Section  7).   The  sink slurry  is
pumped to the reactor R-l for pyrite removal via the  leach  reaction.   The float
coal slurry, containing about 20% w/w coal, meets the  total  sulfur specifica-
tion without further pyrite removal after washing and  dewatering.
     Float coal washing and dewatering.  The 100 TPH  coal in the float slurry
is fed to a 3-stage countercurrent arrangement of three filters  (F-l,  F-2 and
F-5) and two contactors (T-6 and T-22) where the leach solution  is separated
from the coal and the coal water washed such that the  residual  sulfate salts
on the coal cake are less than 0.01% w/w sulfate sulfur.  The coal  cake from
the last filtration stage, containing 33% moisture and 2.56% w/w total  sulfur
with a heating value of 13,830 Btu/lb, is well below  the  product specification
of 4.0 Ib S02/MM Btu.
     Primary reactor.  The 100 TPH coal in the sink slurry,  containing 3.13%
w/w pyritic sulfur, is pumped to the primary reactor  R-l  where  pyrite  removal
by the leach reaction takes place at 250°F and 35 psig.   Based  on  bench-scale
and RTU experimentation (Sections 6 and 7) the reactor is sized  for six-hour
residence time where the pyritic sulfur is reduced to  a level of 0.53% w/w
(dry basis).  Simultaneous regeneration of the leach  solution to maintain the
           +3
ratio of Fe   to total iron at 0.95 is accomplished by the  injection of 3.9
TPH oxygen.  The inert gas leaving the reactor contacts the  recycled  leach
solution in the knockout drum T-4 prior to venting to  the atmosphere.   The
heats of reaction and regeneration are removed in a pump-around  loop where
the reactor slurry is exchanged with cold recycle leach solution in exchanger
E-4, maintaining the reactor temperature at 250°F.
     Sink coal filtration and washing.  The reactor slurry  is cooled  to 215°F
by flashing to about 1 psig then pumped to a rotary drum  vacuum filter F-3
where the leach solution is separated for recycle and  the coal  cake washed
for the removal of sulfate salts.  Wash water is provided from  condensate of
low pressure steam and recycled pond water.  The coal  cake  from F-3 contains
33% w/w moisture.

                                     192

-------
     Leach  solution recycle loop.  The leach  solution  filtrate  stream  from both
the float and sink coal filters are recycled  to  the  mixer  T-l.   The wash  fil-
trates  from F-l and F-3 are dilute in iron  sulfates  and must  be dewatered in
the evaporator EV-1.  A portion of the F-3  wash  stream is  pumped to the neutral-
izer in order to remove from the system  the excess iron produced in the leach
reaction.  The feed to the evaporator is  preheated to  264°F by  the feed/bottoms
exchanger E-3 and the evaporator feed preheater  E-9.   The  evaporator is operated
at 35 psig  and 290 F such that the steam produced can  be utilized for  process
heating service and the condensate recycled as filter  wash water.  The evapo-
rator bottoms are concentrated in salt to the extent that  the total leach solu-
tion recycle stream at the mixer T-l is  7.5%  w/w iron.   The F-l filtrate  stream
is heated to 237°F in exchangers E-4 and E-2  and fed to the regenerator R-2,
operating at 35 psig and 250°F, where 0.5-hour residence time is provided to
regenerate  the leach solution to the extent that the combined recycle  streams
have a ratio of Fe   to total iron equal  to 0.90 at  the mixer T-l.  The oxygen
requirement for regeneration is 0.92 TPH.   The evaporator  bottoms, regenerator
effluent, F-3 filtrate and 2.8 TPH makeup sulfuric acid are combined and  fed
to the knockout drum T-4, then pumped to the  mixer T-l through  knockout drum
T-2 thus completing the recycle loop.
     Acetone washing and elemental sulfur removal.   The sink  coal cake from
filter F-3 is fed to the contactor T-ll  where it is  slurried  with acetone to
dissolve the elemental sulfur produced in the leach  reaction.   The resulting
402 coal slurry is then filtered on a rotary  pan filter F-4 where the  cake is
washed with acetone.  The elemental sulfur  on the coal  cake is  reduced to
about 0.2% w/w based on bench-scale studies (Section 7) and the sulfate sulfur
further reduced to about 0.02% w/w.  The acetone contacting equipment  are
vented to the atmosphere through a water scrubber GS-1 and a  carbon bed T-20
for removal of acetone vapors.
     Acetone recovery and sulfur removal.   The acetone-water  filtrate  from fil-
ter F-4 containing elemental sulfur and  sulfate  salts  are  fed to the acetone
stripper SS-1 operating at 250°F and 15  psig. The stripper bottoms consisting
of water, molten elemental sulfur and sulfate salts  are fed to  the bottoms sep-
arator where the molten sulfur is removed and the water phase decanted.   The
sulfate laden water phase is cooled by E-12 and  fed  to the neutral izer.   The

                                     193

-------
stripper overhead vapor containing essentially all of  the  acetone feed is con-
densed and recycled to the contactor T-ll.
     Lime neutralization.  The bottoms stream from the acetone  stripper and a
portion of the F-3 wash filtrate are removed from the  system  at the neutralizer
in order to maintain the iron concentration in the leach solution at balance.
The neutralizer feed streams are contacted with lime forming  gypsum and iron
oxides and pumped to the settling pond for solids removal.  The solids-free
pond water is treated and recycled to the system as wash water  for the filters.
     Coal drying.  The acetone wet coal cake from filter F-4  containing about
27% acetone is fed to the coal drier F-l where the coal is heated to 225°F and
the acetone flashed by a 400°F stream of nitrogen gas.  The acetone is condensed
from the gas stream, which is a closed loop, and returned  to  the  system for cake
washing.  The gas stream is compressed and reheated to 400°F  by blower B-l and
recycle gas heaters E-7 and E-8.  The dried coal cake  is discharged from the
drier and recombined with the float coal cake.  The dried  sink  coal  portion
contains 2.5% w/w total sulfur and has a heating value of  11,990  Btu/lb.  The
combined coal product contains 2.53% total sulfur with a heating  value of
12,930 Btu/lb resulting in a coal with 3.91 Ib S02/MM  Btu.
8.2.2.2  Battery Limit Process Costs—
     Equipment list.  The previous section of this document presented  a  concep-
tual process design and flow diagram (Figures 36a and  36b) for  a  battery  limits
processing facility treating 200 TPH of feed coal (193.3 TPH  of desulfurized
coal product).  The process was designed to reduce the 4.3% w/w total  sulfur
feed coal to a combined Gravichem product containing no more  than  4.0  Ibs
S02/MM Btu.  An equipment list, tabulating the required major equipment,  is
presented as Table 42.  The total battery limits FOB equipment  cost was  deter-
mined to be 8.7 million dollars while the total  installed  equipment cost  is
15.3 million dollars.
     Battery limits process operating costs.  The total estimated  battery limit
processing cost is presented in Table 43.  As may be determined from Table 43,
the estimated total annualized operating cost results  in an equivalent battery
limit production cost of $8.63/ton of coal product.   This  production cost is
based on cleaned coal production 185.5 TPH (193.3 TPH of desulfurized  product

                                     194

-------
10
in
                      Figure 36a.  TRW Coal Desulfurization Gravichem  Process

-------
a*
                                 Figure 36b.   TRW Coal  Desulfurizatlon Gravichem Process

-------
TABLE 42.  BASE CASE 2 - PROCESS EQUIPMENT LIST
MAJOR PROCESS EQUIPMENT $8.67 MM FOB, $15.31 MM INSTALLED
A-l
B-l
C-l
CY-1
D-l
E-l
E-2
E-3
E-4
E-5
E-6
E-7
E-8
E-9
E-10
E-ll
Feed Hopper - 10,000 ft3
Recycle Gas Blower (3) - CS, 1.7 Compression Ratio, 85 hp
Feed Conveyor - 40 in. wide x 20 ft x 10 hp, 200 ft/min.
Cyclone (3) - 5 psig, 850 ft3, CS
Coal Dryer (3) - 30' <(> 55', CS, 5 psig, 15 hp, 54 trays, 60 min.
Slurry Cooler - 3185 ft2, CS/SS
Regenerator Feed Heater - 1040 ft , CS/Hastelloy C
Evaporator Feed/Bottoms Exchanger - 2890 ft , Hastelloy C/
Hastell oy C
2
Reactor Pumparound/Regenerator Feed Exchanger - 1360 ft , Hastelloy
C/ Hastelloy C
Slurry Cooler - 3400 ft2, CS/SS
Dryer Overhead Condenser (3) - 1770 ft2, CS/CS
Recycle Gas Preheater (3) - 1500 ft2, CS/SS
Recycle Gas Heater (3) - 1050 ft2, CS/CS
2
Evaporator Feed Preheater - 1140 ft , CS/Hastelloy C
Stripper Reboiler - 3120 ft2, Hastelloy C
Stripper Overhead Condenser - 7421 ft , CS/CS
$K
FOB
22
as
16
as
2450
76
37
95
65
78
as
as
as
37
51
114
$K
Installed
26
D-l
27
D-l
3062
205
102
258
177
212
D-l
D-l
D-l
102
142
362
                  (Continued)

-------
TABLE 42.  (Continued)

E-12
E-13
EV-1
F-l

F-2

-, F"3
00
F-4

F-5

GS-1
M-l A/C
M-2 A/E
M-3
M-4
M-6
M-7

Neutral izer Feed Cooler - 800 ft2, CS/SS
Pondwater Heater - 50 ft2, CS/SS


2
Evaporator - 1400 ft , 35 psig, CS with titanium clad
Rotary Drum Vacuum Filter Package - 12'  x 24'
250 hp (inc
Rotary Drum Vacuum Filter Package - 12'  x 24'
250 hp (inc
Rotary Drum Vacuum Filter Package - 12' $ x 24'
250 hp (inc
Rotary Pan Filter Package (2) - 24' 4, 445 ft2,
(inc. vac. sys.
Rotary Drum Vacuum Filter Package - 12' $ x 24'
250 hp (inc
Scrubber - 5' 4 x 30', 20 trays, 0 psig, CS
Mix Tank Mixers (6) - 25 hp, SS
Reactor Mixers (5) - 190 hp, Hastelloy C
Contactor Mixer - 50 hp, SS
Contactor Mixer - 50 hp, SS
Neutral izer Mixer - 15 hp, SS
Contactor Mixer - 50 hp, SS
, 912 ft2, SS,
. vac . sys . )
, 912 ft2, SS,
. vac. sys.)
, 912 ft2, SS,
. vac. sys.)
SS, 105 hp
)
, 912 ft2, SS,
. vac. sys.)







$K
FOB
31
8
722
184

184

184

941

184

13
57
194
13
13
7
13
$K
Installed
90
25
1155
292

292

292

1491

292

21
92
315
21
21
11
21
       (Continued)

-------
TABLE 42.  (Continued)
$K $K
FOB Installed
P-l
P-2
P-3
P-4
P-5
P-6
P-7
P-8
P-9
P-10
P-ll
P-12
P-13
P-14
P-17
P-18
P-19
Slurry Feed Pump (2) - 1350 gpm, 15 psi, 15 hp, SS
Reactor Feed Pump - 1000 gpm, 15 psi, 10 hp, SS
Reactor Discharge Pump - 960 gpm, 15 psi, 10 hp, SS
Leach Solution Feed Pump (2) - 960 gpm, 15 psi, 15 hp, SS
Leach Filtrate Pump - 1120 gpm, 75 psi, 75 hp, SS
Wash Filtrate Pump - 400 gpm, 70 psi, 20 hp, SS
Filter Feed Pump - 850 gpm, 10 psi, 7.5 hp, SS
Filtrate Pump - 340 gpm, 7.5 psi, 3 hp, SS
CaSO^ Slurry Pump - 320 gpm, 2 hp, SS, 5 psi
Wash Water Feed Pump - 730 gpm, 15 psi, 10 hp, CS
Leach Filtrate Pump - 500 gpm, 5 psi, 1.5 hp, SS
Wash Filtrate Pump - 250 gpm, 10 psi, 2 hp, SS
Contactor Slurry Pump - 1000 gpm, 15 psi, 15 hp, SS
Filtrate Pump (2) - 350 gpm, 7.5 psi, 3 hp, SS
Acetone Return Pump (3) - 65 gpm, 1.5 hp, CS, 15 psi
Scrubber Water Return Pump - 150 gpm, 10 psi, 2 hp, CS
Acetone Condensate Pump - 540 gpm, 7.5 hp, CS, 15 psi
15 45
5 15
5 15
8 21
as F-l
as F-l
5 18
as F-2
3 8
4 13
as F-3
as F-3
6 17
as F-4
as D-l
2 16
3 9
       (Continued)

-------
                                       TABLE 42.   (Continued)
o
o

P-20
P-21
P-22
P-23
P-24
R-l
R-2
RV-1
RV-2
RV-3
S-l
SS-1
T-l
T-2
T-3
T-4
T-5
T-6

Wash Filtrate Pump - 370 gpm, 7.5 psi, 5 hp, SS
Filter Feed Pump - 850 gpm, 10 psi, 7.5 hp, SS
Filtrate Pump - 340 gpm, 7.5 psi, 3 hp, SS
Wash Filtrate Pump - 370 gpm, 7.5 psi, 5 hp, SS
Reactor Pumparound Pump - 325 gpm, 15 psi, 5 hp, Hastelloy C
Reactor - 23' $ x 115', CS with acid-brick lining, 45 psig
Regenerator - 10' $ x 61', 50 psig, CS with acid-brick lining
Rotary Valve (2) - 1 hp, 24" x 24", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Rotary Valve (3) - 0.5 hp, 18" x 18", 20 RPM
Separator - 0.5 hr Residence Time (Acid Brick)
Acetone Stripper - 8'  x 54' , SS, 20 Trays, 15 psig
Mix Tank (2) - 20' $ x 40', SS, 0 psig, 91,330 gal
Knock-Out Drum (2) - 3200 gal, SS, 0 psig
Flash Drum - 6100 gal, 7'  x 21', SS, 2 psig
Knock-Out Drum - 13,500 gal, 8' 4, x 24', SS, 0 psig
Wash Filtrate Receiver - 2000 gal, Vac, SS
Contactor - 40,000 gal, 0 psig, SS
$K
FOB
as
5
as
as
4
808
219
39
25
25
855
139
274
38
40
42
as
81
$K
Installed
F-2
18
F-5
F-5
12
1975
534
43
27
27
1710
236
480
87
91
81
F-l
185
                                              (Continued)

-------
                                        TABLE 42.   (Continued)
                                                                                        $K          $K
                                                                                        FOB      Installed
     T-7        Filtrate  Receiver -  2000  gal,  Vac,  SS                                         as F-2
     T-8        Wash Water  Hold  Tank -  40,000  gal,  0 psig,  FRP                             34         67
     T-9        Filtrate  Receiver -  2000  gal,  Vac,  SS                                         as F-3
     T-10       Wash Filtrate  Receiver  -  2500  gal,  Vac,  SS                                    as F-3
     T-ll       Contactor - 40,000 gal, 0 psig,  SS                                         81        185
     T-12       Filtrate  Receiver (2) - 1300 gal, Vac, SS                                     as F-4
     T-15       Dryer  Condensate Receiver (3)  -  450 gal, CS,  0  psig                           as D-l
2    T-16       Stripper  Condensate  Receiver - 3400 gal, CS,  15 psig                        5          9
     T-17       Stripper  Bottoms Separator - 2700 gal, Hastelloy C, 15  psig                19         43
     T-18       Acid Neutralizer Tank - 15,000 gal, FRP, 0  psig                            9         19
     T-19       Filtrate  Receiver -  5500  gal,  Vac,  SS                                         as  F-l
     T-20       Carbon Adsorption Drum  -  5' <|>  x  10', FRP, 0 psig                           3         10
     T-21       Wash  Filtrate  Receiver  -  2500  gal,  Vac,  SS                                    as  F-2
     T-22       Contactor - 40,000 gal, 0 psig,  SS                                        81        185
      T-23      Filtrate  Receiver- 2000 gal, Vac, SS                                          as  F-5
      T-24      Wash Filtrate  Receiver  -  2500  gal,  Vac,  SS                                   as F-5

       CS/SS - Carbon Steel Sheet/Stainless Steel Tubes

-------
            TABLE 43.  BATTERY LIMITS PROCESS COSTS  -  BASE  CASE 2

   Capital Related Costs:                                Annual  Cost,  $1000
        Depreciation - 10% straight line                         1,531
        Maintenance, insurance, taxes,                           2,296
        interest - 15% of capital
   Labor:
        Labor, 12 operating positions at                         1,920
               $160,000/position
   Utilities:
        Electric power, 2280 Kw at 5<£/Kw-hr                       902
        Cooling water, 11,500 gpm, 30°F rise at                   273
               5£/1000 gal
        Process water, 100 gpm at 25<£/1000 gal                     12
        Heating, 201 MM Btu/hr, fired duty, 7.8 TPH
               coal product equivalent
   Materials:
        Oxygen, 4.8 TPH at $507ton                              1,909
        Nitrogen, 0.53 TPH tincluded in oxygen cost)
        Sulfuric acid, 2.7 TPH at $55/ton                       1,194
        Lime, 5.5 TPH at $35/ton                                1,530
        Acetone, 0.1 TPH at $3207ton                              253
        Waste disposal, 18 TPH at $6/ton                          846
   TOTAL COST	^	12,666

less 7.8 TPH for process steam generation).   The battery limit processing facil-
ity therefore operates with an overall  coal  yield of 93% on a dry weight basis
and an overall energy yield of 97% on a Btu  basis.  The coal required for pro-
cess and oxygen plant electric power consumption  is 2.1 TPH resulting in a
95.6% net overall  energy efficiency when included with the coal usage for plant
steam generation.   The noncapitalized battery limit operating cost, the total
                                     202

-------
process operating cost less the annual  depreciation  charge  (10%  of installed
equipment),  is $7.03/ton of coal  feed  (28<£/MM Btu  equivalent)  or $7.58/ton of
coal  product (29<£/MM Btu equivalent).   This  noncapitalized  operating  cost essen-
tially represents a pure battery  limit operating cost.
8.2.2.3  Process Economics—
     The Base Case 2 design, as discussed  earlier, resulted in an  installed
battery limit plant costing 15.3  million dollars,  Using  rational  presented
in Section 8.2.1.3, it is therefore  expected that  the  associated off-sites
would cost an additional 7.7 million dollars.   The resultant integrated grass
roots facility economics are then determined to be as  shown in Table  44.
     As may be seen, the required market value of  the  coal  product ranges from
$1.09/MM Btu with $15/ton ROM  feed coal to $1.75/MM  Btu with $30/ton  ROM feed
material.  The equivalent upgrading  cost,  found by deducting the cost of the
ROM "dirty" coal energy ($0.60/MM Btu, $0.81/MM Btu, and  $1.21/MM Btu at $15,
$20 and $307ton ROM coal respectively) from the required  market  value, ranges
from $0.49/MM Btu to $0.54/MM  Btu.
8.3   ENGINEERING ANALYSIS CONCLUSIONS
     As discussed earlier, two conceptual  full-scale process designs  and cost
estimates were generated based on the  most current technical data  available.
Each of the two design packages,  termed "Base Cases",  were  aimed at evaluating
the recently conceived Gravichem  Process as  applied  to two  differing  coals.
Furthermore, each Base Case deals with desulfurizing coal to different residual
sulfur levels.  Base Case 1 is representative of Gravichem  application to a
relatively low sulfur coal (1.51% w/w  total  sulfur with 1.035 w/w pyritic sulfur)
when processed to a residual total sulfur  content  equivalent to  1.2 Ibs S02/MM
Btu.  Base Case 2 is typical of Gravichem  application  to  a  relatively high sul-
fur coal (4.3% w/w total sulfur with 2.4%  w/w pyritic  sulfur)  when processed
to the less stringent equivalent  sulfur limit of 4 Ibs S02/MM  Btu.
8.3.1  Base Case Results
     Table 45 presents a summary  of  the Gravichem  process economics as they
relate to processing low (Base Case  1) and high (Base  Case  2)  sulfur  content
coals.  It should be noted that the  economics picture, as presented,  is specific
                                       203

-------
              TABLE 44.  BASE CASE 2 - PROCESS ECONOMICS'
Annual Product, 1.46 MM Tons/yr
Energy Value, 38.0 x 106 MM Btu/yr

Capital Related Requirements, $MM
Battery Limit Capital
Off-site Capital
Overhead and Profit
Engineering and Design
Contingency
Total Plant Investment**
Interest for Construction
Start-up Costs
Working Capital (Utility Financing)
Total Capital Related Costs (Utility)
Operating Costs, $MM/Yr
Raw Material (Coal)
Chemicals (Lime, Sulfuric Acid, Acetone)
Supplies
Disposal
Utilities
Labor (18 Positions)
Taxes and Insurance
Total Operating Costs
Required Coal Market Price, $/MM Btu
Utility Financing
Upgrading Cost, $/MM Btu
Utility Financing


$15/Ton
15.31
7.66
5.05
2.30
4.55
34.87
5.88
7.08
6.35
54.18
23.76
2.98
0.79
0.85
3.10
2.61
1.29
35.38
1.09

0.49

ROM Coal Cost
$20/Ton
15.31
7.66
5.05
2.30
4.55
34.87
5.88
8.66
8.13
57.54
31.68
2.98
0.79
0.85
3.10
2.61
1.29
43.30
1.31

0.50


$30/Ton
15.31
7.66
5.05
2.30
4.55
34.87
5.88
11.83
11.71
64.29
47.52
2.98
0.79
0.85
3.10
2.61
1.29
59.14
1.75

0.54
 *Assumes, debt/equity =  75/25,  interest on  debt = 9%,  return on
  equity = 15%.
**
  Equivalent to a plant capital  investment of $69/kw.

                                  204

-------
                                          TABLE 45.  SUMMARY OF BASE  CASE ECONOMICS RESULTS
ro
o
Base Case 1
Cost -Marti nka Coal-
225 TPH Coal Feed Basis
ROM Coal Cost 15 20 30
$/Ton
Battery Limit 17.4 17.4 17.4
Capital - $ MM
Total Plant 39.6 39.6 39.6
Investment - $ MM
Total Operating 37.0 45.9 63.7
Costs - $ MM/Yr
Required Product 1.01 1.22 1.66
Market Price - $/MM Btu
Upgrading Cost .43 .44 .50
$/MM Btu
Base Case 2** Adjusted Base Case 2***
-Kentucky No. 9 Coal- -Kentucky No. 9 Coal-
200 TPH Coal Feed Basis 225 TPH Coal Feed Basis
15 20 30 15 20

15.3 15.3 15.3 17.2 17.2

34.9 34.9 34.9 39.2 39.2

35.4 43.3 59.1 39.8 48.7

1.09 1.31 1.75 1.09 1.31

.49 .50 .54 .49 .50

30

17.2

39.2

66.5

1.75

.54

        **
Mine cleaned Martinka  coal,  1.51% w/w total sulfur as  fed,  1.2 Ibs S02/MM Btu as processed,  1.66 MM TPY product,

43.4 X 106 MM Btu/yr energy  value.


TVA supplied Kentucky  No.  9  coal, 4.3% w/w total  sulfur as  fed, 4 Ibs S02/MM Btu as processed, 1.46 MM TPY product,
3ft n Y ir>6 MM Rtu/vr *>nerav  value.
           38.0 X 105 MM Btu/yr energy value.


       *** Base Case 2 adjusted to  an equivalent 1.64 MM TPY product, 42.8 X 106 MM Btu/yr energy  value.

-------
to utility financing under the ground rules as detailed in Section  8.1.3.3.  As
may be seen from the table, the resultant upgrading costs range from  $0.43/MMBtu
to $0.50/MM Btu for the low sulfur coal case while the range is $0.49/MM Btu
to $0.54/MM Btu for the higher sulfur Kentucky coal case.  The low  sulfur Base
Case 1, which requires 0.79% w/w pyritic sulfur removal, has essentially the
same capital requirement as Base Case 2 which effects a 1.88% w/w pyritic sul-
fur removal.  This finding is substantiated when the Base Case 2 economics are
adjusted to reflect comparable plant throughputs.  The adjusted Base  Case 2
economics are also presented on Table 45.
     The apparent higher processing costs (upgrading cost) associated with
Base Case 2 are generally the direct result of reacting and removing  greater
quantities of pyrite.  The additional pyrite removal required by the  Kentucky
No. 9 coal (1.88% w/w as compared to 0.79% w/w for the Martinka coal) requires
a directly proportional increase in certain operating costs.  Specific costs
which increased in Base Case 2 are associated with oxygen consumption as a
regenerant, lime utilization for neutralization of sulfate, sulfuric  acid
makeup, and sulfur and gypsum disposal.
     Bench-scale processing of Kentucky No. 9 sink (Section 7), performed dur-
ing the latter stages of the design effort, indicate that ash other than pyrite
is also leached from the coal during processing.  The bench-scale data shows
that the sink processed material will have an ash content ranging from 9.5%
to 10.5% w/w while the process design material balance, which considers only
pyrite removal, shows an ash content of 16.3% w/w.  Incorporating this inform-
ation into the overall process design and assuming a 10% ash content  in the
processed sink would result in a heating value of 12,890 Btu/lb for the sink
coal.  The combined float/sink product would therefore have an ash content of
6.5% w/w with a heating value of 13,390 Btu/lb.  The net energy output of the
plant would remain unchanged although processing costs would increase some-
what owing to increased neutralization and disposal costs for the removal of
the additional ash.  This new information would, however, indicate  that a sig-
nificantly cleaner (less ash) and therefore more desirable, desulfurized coal
product would be generated than is indicated by the Base Case 2 process design.
8.3.2  Cost Sensitivity Analysis
     In addition to each of the specific Base Case studies (Sections  8.2.1
and 8.2.2), a parametric analysis relating to process economics was performed
                                     206

-------
to evaluate cost sensitivity.  As  part of the sensitivity study,  three  process
cost related elements were  varied.   For each Base Case,  installed battery  limit
capital  cost was varied to  evaluate  the dependence of the bottom  line upgrading
cost on equipment cost estimates.  The assumed battery limits  cost was  allowed
to vary by approximately  ±50%  while  all operating costs  were held constant.
Figures 37 and 38 present the  effects of battery limit capital  on upgrading
costs as a function of ROM  coal  cost for each of the Base Cases.   Figures  39
and 40 are cross plots showing the effects of ROM coal  cost on upgrading cost
as a function of battery  limit capital.  As may be seen from the  Figures,
upgrading costs can vary  from  approximately -9i/m Btu to +10tf/MM Btu over the
range of ±50% of the Base Case battery limit capital for any given ROM  coal
cost.  This holds true for  both  high and low sulfur coal  processing.
     Another variable evaluated  was  that of percent pyrite or  pyritic sulfur
removed.  In this study,  the feed  coal for each Base Case was  varied with
respect to starting pyrite  content.   The resultant pyritic sulfur removal
requirements were then determined  in order to meet the appropriate residual
sulfur levels for each of the  cases.  In this evaluation, the  capital costs
were assumed to remain constant and  the appropriate operating  costs  (oxygen,
lime, sulfuric acid, and  disposal) were modified.  The results of that  eval-
uation are presented in  Figures  41,  42, 43 and 44.  Figures 41 and 42 present
the upgrading costs for  each case  as a function of percent pyritic sulfur
removed for each of the  three  ROM  coal costs.  Figures 43 and  44  indicate  the
upgrading costs as a function  of ROM coal cost at several pyritic sulfur
removal levels.  The sulfur removal  levels ranged from 0.5% w/w to 4% w/w
for each case.  As may be seen from  the Figures, the resultant upgrading costs
vary significantly with  the amount of pyritic sulfur removed.   In fact, over
the range evaluated for  both the high and low sulfur cases,  the upgrading
costs increase by 65% to  77% of the  0.5% w/w removal values  as the amount
of pyritic sulfur removed increases  to 4% w/w.
     The third variable evaluated  was the percent float which  is  allowed to,
in essence, bypass the conventional  Meyers Process.  For each  Base Case, the
float/sink ratio was allowed to  vary from 33/66 to 66/33 (Base Case  1  is a
55/45 split while Base Case 2  is a 50/50 split).  For this analysis, essen-
tially all operating costs  were  held constant while the capital related costs

                                      207

-------
  .60
   55
   .50
0.45

O
z
Q
o
Q.
D
   .40
   ,35
   ,30
                                         ROM COAL
                                           $AON
                                         BASE CASE 1 $17.4 MM
                10    12    14    16     18    20   ,22

                      BATTERY LIMIT CAPITAL $/MM
24    26
              Figure 37.   Base Case 1.   Upgrading Cost
                         vs. Battery Limit Capital

                                208

-------
                                 ROM COAL
                                  $/TON
                           BASE CASE 2 $15.31 MM
        10    12    14    16    18    20

            BATTERY LIMIT CAPITAL $MM
22   24
Figure  38.  Base Case 2.  Upgrading Cost
           vs. Battery Limit Capital
                   209

-------
   ,60
   ,55
fe
  .50
<->  .45
O
Z
  .40
  .35
  .30
               BATTERY LIMIT
               CAPITAL COST
                  $25 MM
BASE CASE 1
$17.4 MM
                 $17 MM

                 $10 MM
     0          10         20         30
           ROM COAL COST $/TON
 Figure 39.  Base Case 1.   Upgrading Cost
            vs. ROM Coal  Cost
                   210

-------
   .65,
               BATTERY LIMIT
               CAPITAL COST
   .60
   .55
VJ-
to
o
U
O
z
5 .50
o
Q_
   .45
  .401
                   $23 MM
$15.3 MM
(BASE
CASE 2)
                  $8 MM
      0         10          20
            ROM COAL COST - $/TON
                     30
   Figure 40.   Base Case 2.   Upgrading Cost
               vs. ROM Coal  Cost
                      211

-------
                                      ROM COAL
                                      COST $/TON
              BASE CASE 1
              .78% SP REMOVAL
      .40
                    1          2
                         % SP REMOVED
F-fgure 41.  Base Casel.  Upgrading Cost vs.  %  Pyritic Sulfur Removed
                             212

-------
                                   BASE CASE 2
                                   1.88% SP REMOVAL
                                                      ROM COAL
                                                      COST - $AON

                                                      30
                    1           2
                         %SP REMOVED
Figure 42.  Base Case 2.  Upgrading Cost  vs. % Pyritic Sulfur Removed

                             213

-------
                     .80
                     .70

                  8 -60
                  o
                  z
                  o
                  o
                  ^  =50
                     ,40
                                       %SP
                                     REMOVED
                                       4.0-
                                       2.0-
                                        1.0,
                         0.79 (BASE CASE 1).
                              i      I   0.5i
                        0     5     10     15     20    25
                                ROM COAL COST $/TON
30
                     Figure 43.  Base Case 1.  Upgrading
                                 Cost vs. ROM Coal Cost
                                 at Different Levels of
                                 Pyritic Sulfur Removal

 (i.e.,  battery limit costs, off-sites, etc.) were adjusted.  The capital  costs
 were adjusted to reflect the appropriate percent treatment in the more costly
 Meyers  Process portion of the plant, and the less costly bypass float washing.
 In all  cases, the costs associated with mixing, float/sink separation, and
 wash water generator (evaporator) were held constant.
     Figures 45 and 46 show the effect of float processing on the upgrading
 costs as a function of ROM coal cost.   Figures 47 and 48 present the cross
 plots indicating upgrading costs as a  function of ROM coal costs at several
float/sink ratios.   As may be seen from the Figures,  the overall processing
costs do indeed drop with increasing float percentage.  The costs drop by
approximately 5<£/MM Btu over the float/sink ratio range of 0.5 to 2 Ci.e.,
                                    214

-------
33/66  to  66/33)  for  each of the ROM coal costs evaluated.  This  finding holds
true for  both  the  high  and low sulfur coal base cases.
                .30
                         5     10    15    20    25
                          ROM COAL COST $/TON
30
               Figure 44,   Base Case  2.   Upgrading Cost vs,
                           ROM Coal Cost  at Different
                           Levels  of  Pyritic Sulfur
                           Removal
                                    215

-------
                                ROM COAL
                                COST $/TON
                                  BASE CASE 1
                                  55% FLOAT
     10   20    30    40   50    60    70    80
    % OF TOTAL COAL TO FLOAT PROCESSING
Figure 45.  Base Case 1.  Upgrading Cost vs,
           Coal to Float Processing
% of Total
                      216

-------
             BASE CASE 2 - 50%
                                ROM
                            I COAL COST
                            I— $/TON  -
 0     10    20   30    40   50    60    70
  % OF TOTAL COAL TO FLOAT PROCESSING
Figure 46.,
Base  Case 2.  Upgrading Cost vs.
Total Coal to Float Processing
                    217

-------
   .55
   -5°
to
o
u
o
        RATIO OF FLOAT COAL
           TO SINK COAL
                   1/2
 o
 0.
   (BASE CASE 1)
       55/45   '
                   2/1
   .40
                                 Figure 47.
Base Case 1.
Upgrading Cost
vs. ROM Coal
Cost
           5    10    15    20    25
             ROM COAL COST $/FON
                            30
   .60
   .55
        RATIO OF FLOAT COAL
        TO SINK COAL
O
u
o
z
   -50
2
   .45
(BASE CASE 2)
         1/1


         2/1
                                 Figure 48.
Base Case 2.
Upgrading Cost
vs. ROM Coal
Cost
           5    10    15    20    25
            ROM COAL COST $AON
                            30
                         218

-------
                                 9.  VENDOR TESTING
   9.1  GENERAL APPROACH
        The  current RTU equipment is intended to emphasize study of the desulfur-
   ization reaction portion of the Meyers Process.  Operations which may be eval-
   uated are coal  transfer and leach solution contacting, elevated temperature-
   pressure  depyritization, and solution regeneration.  A filtration operation is
   also included in the RTU to facilitate coal and solution separation and handling
   as are solution storage facilities.  The specific equipment which would be
   required  to evaluate operations relating to efficient coal-leach solution sep-
   aration,  elemental  sulfur removal via solvent extraction, leach solution concen-
   tration and neutralization, and coal drying is not included in the RTU.  These
   operations would be required to demonstrate an integrated plant facility  uti-
   lizing the Meyers Process.
        The  vendor testing task was intended to provide the necessary information
   on those  unit operations not in the RTU through utilization of equipment supplier
   testing services.  It was reasoned that basic data and scale-up information
   could be  obtained in a timely cost-effective manner by allowing equipment vendors
   to apply  their unit operation experience in testing coal prepared in the RTU.
   Thus, depyritized coal from the RTU would be packaged and transferred to each
   vendor.  The vendor, following an approved test plan, would treat and return the
•   coal for  detailed analysis by TRW.  The success of the vendor study was thought
   to be closely associated with the initial vendor selection and the detailed
   coordination, planning, and review of activities both in preparation and testing
   of the coal.
        A vendor survey was completed and a vendor coordination and test plan was
   generated.   An  initial vendor coal supply was also obtained.  Actual tests have
   been provided by one vendor to date.  A processed coal supply is currently
   stored at the RTU which can be used for further vendor testing.  The details of
   the above mentioned activities are provided in the following sections.
                                        219

-------
9.1.1  Equipment Suppliers Survey and Selection
     A critical part of the vendor testing task was the selection of suitable
vendors to be utilized.  The general approach taken was to initially select
vendor candidates for filtration, centrifugation, solvent stripping, coal dry-
ing and solution neutralization unit operations based on previous contacts.
Then, the most promising vendor's, (based on scale-up capability, cost, sched-
ule, ability to duplicate actual process conditions, sample sizes required and
versatility for different tests at the same facility) would be determined.
Finally a single vendor would be selected to test each unit operation and a
specific cost, test plan, schedule etc., would be generated.  Table 46  provides
a list of the potential vendors considered for this program and the findings
obtained from telephone inquiries and written correspondence.  In general, each
vendor was supplied with essentially the same type of information consisting of:
        •   Description of feed (composition, physical and chemical
            properties.
        •   Description of product (desired composition, physical and
            chemical properties).
        •   Description of tests desired (a matrix was provided showing
            the desired variables and parameters to be measured).
        •   Description of possible acceptable approaches (types of
            unit operations, sequences of operations, possible
            coreactant/contacting solution compositions, etc.).
     By the conclusion of Phase 3 of the current program, initial vendor selec-
tions were completed.  The selected vendors and the associated services which
were anticipated are as follows:
        Envirotech Corporation
        •   Filtration of coal slurry in leach or water solution, 4 tests,
            40 gallons of slurry required (bench scale).
        •   Acetone slurry make-up, contacting and filtration, 4 tests
            (bench scale).
        •   Solvent stripping of acetone filter cake, 8 tests  (bench
            scale).
        •   Neutralization of coal slurry leach solution filtrate,
            8 tests (bench scale).

                                      220

-------
                                             TABLE  46.  VENDOR SELECTION PROGRAM
               Vendor
    Address
  Unit Operation
 Test Requirement
    (Per Test)

 Time    Quantity
(Days)   (Gallon)
Comments
         Envirotech Corp.
         Eimco  BSP
         Division
Salt Lake City,
Utah
         Bird Machine Co.
ro
ro
So. Wai pole,
Mass.
Filtration and
Solvent Stripping
                                                 Neutralization
Centrifuge and
Solvent Stripping
  15        10      Small  pilot scale equipment
          Slurry    for leach solution and ace-
                    tone slurry filtration and
                    stripping.

   2                Will  use material  provided
                    for filtration study for
                    neutralization tests.

   3       300      Able  to  do  testing including
(est.)     Slurry    acetone  slurry repulping at
          (est.)    pilot  scale on 18"  solid
                    bowl centrifuge.   Would  run
                    bench  scale tests  prior  to
                    pilot  tests to determine
                    operating parameters.
          Artisan  Industries,    Waltham,  Mass.    Solvent Distillation      4
          Inc.
          WyssmontCo.,  Inc.     Ft.  Lee,  N.J.     Drying

                                                 Drying
                                                            Able to run tests on 4"
                                                            glass column.  Artisan will
                                                            provide acetone and makeup
                                                            feed at their facilities.
                                            5      6000      Test can be performed at the
                                                            facilities in Ft. Lee on a
                                            4         5      small pilot scale dryer or
                                                            on a large pilot dryer at
                                                            another location.  Either
                                                            drier will provide design
                                                            information for scale-up to
                                                            full size.
                                                             (Continued)

-------
                                                    TABLE  46.  (Continued)
               Vendor
    Address
         Ametek Process
Temecula, Ca.
ro
l\3
ro
         Dorr-Oliver
Oakland, Ca.
  Unit Operation
 Test Requirement
    (Per  Test)
 Time    Quantity
(Days)    (Gallon)
Comments
Filtration

Centrifuge




Drying



Filtration

Centrifuge
         Joy Manufacturing
         Co.
         Denver  Equipment
         Division
Denver, Col,
Filtration

Solvent Stripping
   2         5      Can only perform bench scale
                   tests with acetone since
   2         5      pilot equipment is not
                   explosion proof.  Able to do
                   leach solution tests at
                   pilot scale.

   2         5      Test could be done using a
                   small pilot dryer with
                   solvent recovery capability.

   5       300      Cannot handle acetone slurry
                   due to inadequate ventila-
                   tion.  Leach solution tests
                   could be done at pilot scale.
                   Testing could be done at
                   bench scale for acetone
                   slurry.

                   Testing can only be done at
                   bench scale since the pilot
                   equipment is not corrosion
                   resistant to leach solution.

-------
       Bird Machine Company
       •   Centrifugation of leach  solution  or water coal  slurry,
           4 tests, 1200 gallons  of slurry required (pilot scale).
       t   Centrifugation of acetone slurry,  4 tests,  4 gallons
           acetone slurry required  (bench scale).
       Myssmont Corporation
       •   Drying of acetone wet  cake,  6 tests, 30 gallons acetone  wet
           cake required  (bench scale).
       Artisan Industries
       •   Acetone stripping,  6 tests (bench scale).
Test plans  and implementation  strategy needed to develop firm  agreements for
the studies were also developed and  are discussed in the following paragraphs.
9.1.2  Vendor Test  Implementation  Plan
     A detailed test plan  was  developed to define the test  and data  require-
ments for the RTU vendor testing effort^10).   Briefly,  the  approach  presented in
the plan was to initially  contact  vendors with testing  capabilities, select the
most promising vendor,  prepare  specific test  objectives, prepare and package
RTU processed coal  for  testing, process the coal at the vendor test  facility
under several parametric sets  of conditions,  and to observe and evaluate the
tests through on-site test monitoring and by  analysis of feed  and product
materials.
     At the conclusion  of  each test  series, the vendor was  to  provide documen-
tation of the following:
        •   Run date, duration.
        •   Type and  size  of  coal  used.
        •   Equipment model,  operating conditions,  and controls.
        •   Feed rate,  mass closure, energy consumption where  obtainable.
        •   Observations during the  run,  such as uniformity of product,
           thickness,  apparent wetness,  and  problems.
        t   Conclusions and recommendations in terms of feed rates,
           effectiveness  of  the unit operation for purpose intended,
           and limiting factors for equipment usage.
                                      223

-------
It was also planned to obtain full-scale process equipment costs and  utilities
requirements based on the vendor test data for use in conceptual full-scale
design work.
9.1.3  Vendor Testing Results
     During RTU operation (Phase 3 of the current program), initial small-scale
laboratory drying tests were conducted by Wyssmont.  These studies were per-
formed to obtain preliminary drying characteristics of acetone wet Martinka
coal in tray dryer type equipment.  It was anticipated that this preliminary
data would minimize the chances of nonproductive testing at the more costly
larger pilot scale.  For purposes of this preliminary study, one gallon of
14 mesh x 0 coal product (Martinka, Coal #1), was obtained from the RTU on
30 June 1977.  The RTU processing consisted of coal-water contacting and mixing
at 212°F for approximately 6 hours followed by filtration.  Since there was no
added iron sulfate in solution and since oxygen was not utilized in the reactor,
it was anticipated that the coal would experience negligable reaction.  The RTU
processed coal was further processed in TRW laboratory facilities to simulate
downstream Meyers Process operations prior to the final drying (i.e., acetone
extraction) step.  The acetone wet cake was then sent to the Wyssmont labora-
tories for testing.  The specific procedure followed in preparing the cake for
shipment to Wyssmont was as outlined below.
        1.  Water wet cake from the RTU was dried in vacuum overnight.
        2.  To 5.43 Ibs. of the dry coal, 0.81 Ibs. of acetone was added,
            resulting in a 14.9% w/w acetone wet cake.
        3.  The cake was placed in a 1-gallon can in an inert N2 atmo-
            sphere and was then shipped to Wyssmont.
     The Wyssmont processed coal samples were returned to TRW for chemical anal-
ysis.  Those samples were evacuated under hard vacuum for 24 hours at 212°F.
During the 24 hours all condensables were captured in cold traps.  The condensed
volatiles represented 0.59 percent of the starting material by weight.  The
condensed volatiles were subjected to mass spectrometric analysis and were found
to be 83% w/w acetone and 17% w/w water.  These results indicate that the orig-
inal 14,9% w/w acetone wet cake was reduced to 0.49% w/w acetone in the small-
scale test apparatus.
                                     224

-------
     In  addition to the input/output analysis and  results  described  above,
samples  of the cake were removed from  the  drying apparatus at Wyssmont as a
function of drying time and examined for weight  loss  (presumed to  be acetone-
water volatilization).  Samples were removed from  the 250°F drying apparatus
after 15, 30, 40, 50 and 60 minutes of total drying  time.   They were found to
contain  approximately 2.0% w/w, 1.8% w/w,  0.9% w/w,  1.1% w/w and 0.6% w/w
moisture respectively.  It should  be noted that  these results are  somewhat
crude in nature due to the lack of sophistication  in  equipment at  the Wyssmont
facility and the inherent nature of the test apparatus (dust evolution could
not be detected if minor and  therefore may interfere  with  results).   However,
the final 60-minute weight loss result agrees very well  with the more precise
mass  spectrometric analytical  results.
      With regard to full scale applicability, Wyssmont personnel indicated that
 although the small scale vendor tests  indicated  residual acetone levels of
 0.5%  w/w, a residual  level of 0.1% w/w acetone would  be attainable through
 proper  adjustment of  dryer gas recirculating rate, and temperature,  and solids
 residence time.
 9.1.4  Coal  Supply for Vendor Studies
      In preparation for the  vendor studies plan  discussed  in 9.1.2,  coal was
 collected from  the RTU filter belt during  experiments 03-01, 03-02,  and 03-03,
 after steady state operation had  been  attained.   Table 47  describes  the coal
 compositions and  the  coal  quantities  obtained  for future vendor testing.
 9.2  VENDOR TESTING CONCLUSIONS
      An obvious need  currently exists  to demonstrate total Meyers  Process oper-
 ability beyond  that of  the limited RTU capability.  The vendor testing approach
 still appears  to  be a very advantageous method of obtaining the desired proc-
 essing  information and  scale-up  data,   Based on the enthusiastic response of
 the vendors contacted and  the degree of flexibility found in several of the
 selected test facilities,  it is  anticipated that a great deal of meaningful
 information could be  obtained in  an efficient, cost effective manner.  A system
 encompassing RTU  sample.generation, sample collection,  sample documentation,
 sample  analysis before and after  testing,  and TRW-vendor  communication and tech-
 nical support has been  demonstrated.   Samples of RTU  processed coal   are stored

                                       225

-------
ro
ro
                                      TABLE 47.  COAL  CAKE COLLECTED  FOR VENDOR TESTING
                                                 (14 MESH X 0 MARTINKA,  RTU COAL #3)
Average*Coal Cake Composition, % w/w (Except Heat Content), Dry Basis
Container No. of Cake Wt.
*
Designation Containers Ibs.
0301-1/6 6 2700
0302-1/6 6 2800
0303-1/8 8 3900
Heat
Ash Content,
Btu/lb
13.0 13260
12.0 13390
12.5 13265
Total
Sulfur
ST
.91
1.03
1.04
Pyrite
Sulfur
SP
.18
.17
.14
Sulfate
Sulfur
SS
.06
.07
.11
Organic
Sulfur
S0
.49
.60
.59
Elemental
Sulfur
SE
.18
.19
.20
Cake
Iron Moisture
Fe
.23 15.5
.26 19.8
.25 19.1
          55 gallon  steel open head drums with insertable polyethylene liners and  header.

          Exp.  03-01,  average of 12 samples and their  analyses.
          Exp.  03-02,  average of 15 samples and their  analyses.
          Exp.  03~03,  average of 13 samples and their  analyses.

-------
and awaiting possible future shipment to the selected vendors.  In summary,
the availability of vendors, vendor facilities, plans, and material is such
that the vendor test program could be started at any time.  It is therefore
recommended that vendor testing of RTU generated coal product be reinstituted
at the earliest possible time to obtain the desired process scale-up data.
                                      227

-------
                 10.  MATERIALS OF CONSTRUCTION EVALUATION

     Tests were performed on the TRW Coal Desulfurization Process RTU to
evaluate the effects of the process reactant environment on selected materials
of construction.  Following initial shakedown tests in which acidified leach
solution and/or coal-water solutions were used, the plant was put into
operation using the full operational cycle.   Experimental test sequences
were conducted over the periods 12 October to 23 November 1977 (Experiment
01) and 5 January to 11 January 1978 (Experiment 03).
     The corrosive medium was an abrasive slurry/leach solution of variable
composition containing coal, water, 1 to 5 percent sulfuric acid, 0.1 to
5 percent ferric and ferrous sulfate, and blended oxygen.  Oxygen partial
pressure ranged from zero up to about 45 psig.  Chemical analysis of the
leach solution also showed the presence of traces of chloride ions (28 to
107 ppm) and fluoride ions (15 to 32 ppm).  The reactor was operated at
temperatures up to 275°F and total pressures up to about 80 psig.
     The RTU is generally constructed of 316L Cres (stainless steel).  There-
fore the plant equipment itself serves as an extensive metallurgical specimen.
Other materials of construction were evaluated through the use of metallur-
gical test speciments which were installed inside the R-l pressure vessel
(the primary reactor) just prior to Experiments 01 and 03.  The test specimens
included coupon test samples and stress-corrosion test samples made of a
variety of metals, non-metals and non-metallic coatings and liner materials.
Metallic and elastomeric erosion test orifice plates were place in the
reactor pump-around loops.
     Experiment 01 ran in product test environment for approximately 170 hours
until leaks developed in the piping/valve system, forcing shutdown.  The RTU was
inspected and the metallurgical test specimens were removed for analysis in late
                                     228

-------
November 1977.  During shutdown,  the  R-l  reactor  was  repaired.   Deeply corroded
areas were filled with 316L weld  rod  material  and smoothed  by grinding.  Shallow
pits were removed via grinding only.   New test specimens were installed and the
plant was restarted on 5 January  1978.   Experiment 03 ran for approximately
80 hours before an unexpected pump  failure forced shutdown.  After draining
and cleaning, the internals of the  R-l reactor and the metallurgical test
specimens were again inspected.   The  results  of the inspections  and analyses
of the R-l reactor and associated piping, and of  the  metallurgical test
specimens installed in the RTU are  reported in Section 10.1 and  10.2
respectively.
10.1  INSPECTION AND ANALYSIS OF  RTU  EQUIPMENT
     After each shutdown, an inspection and analysis  of corroded or failed
RTU components was carried out.   Of particular interest were parts of the
piping system, the R-l reactor, and the R-2 mix tank.
     The first inspection, in November 1977,  showed the effects of approxi-
mately 250 hours of operation at  elevated temperatures.  The first 80 hours
were shakedown runs.  Some of these runs were without coal or oxygen.  In all
but pne, a very low concentration (less than  0.2%)  of iron existed in the
system.  Sulfuric acid concentrations during  shakedown ranged from zero to
4 percent.   During Experiment 01, the plant ran for about 170 hours at normal
operating temperature  (230°F to 275°F) under  conditions of high iron and acid
concentrations.  The specific operating conditions  are delineated in Section 6
of this report.
     The second inspection, in January 1978,  showed the effects of about 80
hours of operation during Experiment  03 at 230°F.   A  slurry containing only
a small amount of iron (less than 0.2%) was used.   The types of corrosion
noted after  Experiment 03 were different than those found after  Experiment 01,
suggesting that different methods of  attack were  present.
10.1.1  Piping System
     After Experiment 01, various components  from the discharge  lines and
return loops  for Cells 3, 4, and  5  of R-l were inspected where leaks were
observed.  These lines were exposed to leach  solution at 230 F to 270 F.  The
flush port fittings in the discharge  lines all contained leaks,  some obviously

                                     229

-------
associated with welds.  The Cell 4 return line had a leak at  the  reducer
weld while the Cell 5 line had a leak in a formed bend area.
     Figure 49 shows the Cell 5 flush port fitting.  Salts were visible at
leaks in the fitting.  Large pits were present on the inside  surface.  In
addition, many smaller pits formed.  Cross sectioning the sample  at one of
the smaller pits revealed a porous "spongy" structure, Figure 50.  The porous
structure penetrated about 25% of the way through the wall.   Little evidence
of weld sensitization was found while some degree of incipient attack was
noted on machined inside surfaces.
     The Cell 5 return line was sectioned and a large pitted  area was found
(Figure 51) which was associated with the pipe weld.  No sensitization was
noted and the metallurgy of the parent metal and weld appeared normal.  The
attack in the weld area appeared to follow grain boundaries.  While the inside
surface appeared to be shal lowly pitted when viewed normal to the surface,
cross sectioning revealed that the porous material extended the full thickness
of the pipe and caused perforation.
     Pitting was observed in reducers where a crevice condition existed,
Figure 52.  These pits were not associated with welds.  Portions of pipe
which fit into the reducers were badly attacked due to crevice corrosion also.
Flanges throughout the system showed evidence of crevice corrosion beneath
their gaskets as shown in Figure 53.   In contrast, inspection after Experi-
ment 03 showed no crevice corrosion on the flanges.
     Inspection of the failed P-l feed pump after Experiment  03 indicated that
the chrome plated 316L rotor had been severely attacked by the slurry.  The
plating had flaked away, exposing the 316L, which then eroded and/or corroded
away until pump tolerances grew and pumping capability was lost.
     Electron microprobe analysis (EMP) was performed on several  failed 316L
components.  The flange in Figure 53 was probed to determine  the  chemistry
of an area pitted by crevice corrosion.  Both the pitted area and a deposited
crystal were analyzed.  The EMP indicated high Si and 0 and significant P,
Fe, and Cr in the corroded areas.   Analysis of a pit from a  pipe end,
 High        = >10%
 Significant = 1-10%
 Trace       = <1%                  23Q

-------
OJ
                                            Figure 49.  Flush Port Fitting

-------
                        Figure 50.  Porous Spongy Structure

indicative of pitting corrosion, showed a different chemistry.   High  Ca,  K,
Fe, S, and 0 and significant amounts of Cl,  Mn,  Cr, Si,  and Na  were found,   A
common feature of all the pits analyzed was  that Mo and  Ni, if  present, were
only found in trace amounts, even though they comprised  2 and 12 percent
respectively of the starting metal (316L).
     R-l  Reactor
     The R-l  reactor is constructed of 316L  stainless  steel  and is  approxi-
mately 3 feet in diameter and 15 feet long.   It  is divided into five  cells
Connections between the vessel's many nozzles and ports  and the plant's
piping/valve system are welded and/or bolted at  gasketed flanges.
     Visual inspection of the reactor after  Experiment 01  showed crevice
corrosion at the manways and at weir/ring interfaces,  and pitting  corrosion
near the "salt line" (water line) in the reactor.  The area below  the water
line was  green indicating a ferrous sulfate  deposit while that  above  the
water line was orange-red indicating ferric  sulfate.  A  cell  by cell  summary
of the observations is given below.

                                     232

-------
CO
CO
                                                 Figure  51.   Pipe  Weld  Corrosion

-------
              Figure 52.  Crevice Corrosion in Reducer Fitting

     Cell 1.  There was no evidence of pitting or crevice attack.   Wall
deposits especially on the manway side, made inspection difficult.   Some small
pits were detected at the mixer port flange.
     Cell 2.  No evidence of pitting.
     Cell 3.  Several large pits were  noted at or near the water line.   A
.38-inch x .12-inch x .015-inch deep*  pit was found at the weld near the #2
bulkhead on the opposite side from the port.  Other small pits were visible
in this area.   The metal around these  pits crumbled when probed with a  knife
blade revealing a porous structure similar to that noted on the piping/valve
components.   There were about 30 of these small  pits ranging from .015-inch
to .030-inch diameter.  After probing, the dimensions of the larger exposed
areas were .188-inch x .125-inch x .060-inch deep, .125-inch diameter x
.125-inch deep and .100-inch diameter  x .040-inch deep.  Other pits were
noted near the center of Cell  3 at a small port.   One of these pits was
.250-inch x .188-inch x .030-inch deep.  There were also six interconnected
 Note:  Depths were estimated.
                                      234

-------
ro
CO
en
                                            Figure 53.   Flange  Crevice Corrosion

-------
pits in this area all approximately .030-inch deep.  The largest of these was
.125-inch diameter.  These various pits were determined to be typical of those
found in numerous other regions within Cell 3.
     Cell 4.  Heavy deposits were found in Cell 4 making inspection difficult.
Two pits were found near the #3 bulkhead on the far side.  Their sizes were
.188-inch x .125-inch x .090-inch deep and .250 inch x .090 inch x .030-inch
deep.  On the port side, a .125-inch diameter x .030-inch deep pit was found
near the #5 bulkhead.  The bulkheads (weir rings) between Cells 3 and 4 and
between Cells 4 and 5 were attacked on the edges.  Large areas could be
scraped away with the knife blade revealing a porous, crumbling, structure.
The weir nuts were attacked, especially where they were welded to the bulkheads,
The baffle nuts and bolts were severely attacked and could be separated by
hand, see Figure 54.  The weir baffle was attacked at the edge, Figure 55.
     Cell 5.  Porosity and pitted structure which could be scraped away was
found near the #4 bulkhead on the far side from the manway.   Exposed areas
were .188-inch x .125-inch x .060-inch deep, .125-inch diameter x .030-inch
deep, .250-inch x .125-inch x .030-inch deep and .250-inch diameter x .030-inch
deep.  The bolts holding the baffles crumbled away as did the nuts.   The lower
part of the bracket was pitted and the bracket-to-wall  weld had a .125-inch
diameter x .030-inch deep pit.  Two pits were found on  the port side near the
dome weld.  These were .250-inch diameter x .030-inch deep and .125-inch diam-
eter x .020-inch deep.
     The TE 56 temperature probe, although appearing solid and shiny,  crumbled
when scraped with the knife blade exposing a large area of porous, heavily
attacked material (Figure 56),  The corrosion apparently originated at the
tip-to-probe weld.  EMP analysis showed the probe base  metal  to be 316L.  High
amounts of Si and 0 were found in the corroded area.  Iron,  Cr, and Ni were
either undetected or present in only trace amounts.
     Inspection of the R-l reactor internals after Experiment 03 showed only
minor signs of the spongy pitting that was observed after Experiment 01.  This
type of attack was visually evident only at the water line and on a welded
attachment.  However, many smooth bottom pits were observed.   General  pitting
occurred where grinders had been operated tn repairing  the reactor (Figure  57).
                                    236

-------
INJ
CO
                                    Figure  54.   Baffle Bolts and Nuts from R-l Cell  4

-------
Figure 55.   Weir Baffle

-------
ro
CJ
10
                                            Figure 56.  TE-56  Thermocouple Probe

-------
Figure 57.   R-l  Reactor Internals AFter Experiment 03 Showing Pits Where
            Grinders Had Operated

-------
Pitting was evident at welds  and  on free surfaces (Figure 58).  No crevice
corrosion was visible.
10.1.3  T-2 Mix Tank
     The T-2 mix tank was  inspected on 9 December 1977.   Some crevice  corrosion
was noted on the tnanway  flange  and door in the seal  area.  The mixer port
flanges were unattacked.   In  general,  the internal surfaces of the mix tank
were not attacked by pitting  or crevice corrosion.  No crevice corrosion was
noted on the baffles, bolts or  nuts.   The temperature probes were  also not
attacked.  The only attack noted  was  under a piece of masking tape which had
been left adhered to the wall in  Cell  Number 3.   A shallow  pit had formed in
the crevice caused by the  tape.   The  pit was .250-inch x  .188-inch x .015-inch
deep.  Two or three pits (about .100-inch diameter)  were  found on  a blade from
one of the mixers.  Incipient pitting  was noted  on the bore  of the propeller.
The weir actuator shaft  was attacked  in the area of  attachment to  the weir
plate (crevice condition)  and the corresponding  weir plate guide showed a
possible pit.  It should be noted that while the T-2 mix  tank experienced the
identical coal/slurry stream  that the  R-l reactor saw, the temperatures in T-2
never exceeded 212°F since it is  an atmospheric  pressure  vessel.  Also, there
is no oxygen injection into T-2.   These differences  are believed to relate to
the much lesser degree of  corrosion experienced  in T-2 as compared with the
reactor R-l.
10.2  ANALYSIS OF TEST SAMPLES
     The primary types of  attack  which RTU materials of construction must
resist are erosion-corrosion  in high flowrate areas,  stress  corrosion crack-
ing (SCC) in high stress areas, pitting corrosion, crevice corrosion, and
general corrosion.  The  tests described in this  section were  designed to
predict the performance  of a  variety of materials used in the  corrosive
environment of the R-l reactor.
     Materials samples in  the form of  metal  and  alloy corrosion coupons,
stress-corrosion coupons,  coated  and lined coupons,  and .non-metallic coupons
were mounted in the R-l  reactor in Cell  3.   A 316L and Teflon  rack was  con-
structed and the specimens were mounted using Ti-6Al-4V bolts,  316 Cres bolts,
and 304 Cres nuts.  The  bolts and nuts were sealed (overrated) using a
Hypalon elastomer coating. In addition,  metallic and elastomeric orifice
                                     241

-------
rvi
-f^
ro
                     Figure 58.  R-l Reactor Internals After  Experiment  03,  Showing Smooth

                                     Bottom Pits at Welds  and  on  Free  Surfaces

-------
plates were installed  in the  cell  discharge and return lines to assess erosion
characteristics of these materials.   All  coupon and orifice plates had areas
where crevice conditions existed.  A summary of the results of the inspection
and analysis after removal  on 29 November 1977 are given in Table  48.   The
observations noted for samples removed during both the November 1977 and
January 1978 inspections are  presented below.
10.2.1  Corrosion Coupons  and SCC Specimens
     Figure 59  shows corrosion coupons and SCC samples attached to the mount-
ing rack prior  to installation in the R-l reactor.  After removal,  the samples
were cleaned, weighed, and inspected for signs of corrosive attack.
     Metal Corrosion Coupons.  Inspections of samples  tested in Experiment 01
suggest there are at least two distinct corrosion mechanisms operating, namely
corrosion  in crevices  and  pit formation on free surfaces.   Once corrosion has
initiated, a spongy porous structure forms which can tunnel  under  the  surface
to honeycomb the  substructure.  The  crevice attack appear to be caused by a
concentration celP   ', with  oxygen  the most probable  species experiencing the
concentration gradient.  The  pitting corrosion appears to be triggered by CSL~
ion and sustained by the presence of Fe + ion and possibly complicated by
localized  deposits  causing crevice conditions.
      A third type of pitting  was observed during Experiment 03. Pits  which
formed did not  have a  spongy, porous structure but were smooth with  rounded
bottoms.   Test  conditions  for Experiment 03 were less  severe than  Experiment 01
 (lower temperature,  lower iron and halide ion concentration, shorter duration).
However,  it  is  unclear what mechanism of attack caused the round bottomed pits.
      The  effects  of the corrosive environment on specific coupon materials is
 summarized below.
        •    316L  Cres  Stainless Steel.  A 316 Cres sample was put  in as
             a  control.  Crevice pitting occured under the bolt head and
             the nut during Experiment 01 resulting in a weight loss of
             0  14  percent.   Inspection of the coupons after Experiment 03
             revealed  smooth  bottomed pits but no crevice corrosion
             (see  Figure 60).
        t    304 Stainless Steel.  The samples exhibited smooth bottom
             pits, but no crevice corrosion when inspected after
             Experiment 03  (see Figure 61).

                                      243

-------
TABLE 48.   COUPON  CORROSION  TEST  RESULTS  -  29  NOVEMBER  INSPECTION
                        Material
          Result
          Metals:
                    316  Cres

                    Pure Nickel

                    Inconel 601


                    Inconel 617


                    Inconel 625

                    Incoloy 825


                    Titanium

                    Lead
          Non-Metal lies:

                    FRP
                    Ashland  197/3 + AT-8-4
                    Ashland  800/801L-68-26
                    Furfuryl Alcohol Resin
                    Ashland 800-10-19


                    Ashland 7240-4-37
                    Unfilled Polyester Resin

                    Ashland 7241-6-27
                    Unfilled Polyester Resin


                    Ashland 72L + 5b-l-23


                    FRP  Pultrusion
                    Plasites 4020, 4030, 4092
                    Vinyl  Ester Coatings on Steel

                    Ceil cote 252
          Coatings:
                    Kynar  202/205



                    EA919  Epoxy

                    Viton  A

                    AF-E-332(EPDM)
Crevice Corrosion

Completely Dissolved

Crevice Corrosion
Noncrevice Pitting

Crevice Corrosion
Noncrevice Pitting

Crevice Corrosion

Crevice Corrosion
Slight Non Crevice Pitting

No Attack

No Attack
No Pitting
Yellow - Red Coating
Edge Cracking
Severe Weight Loss
Resin Bond Attack

Blistered
Embrittled
Cracked

Surface Etched
Crazed
Pitted

Severe Weight Loss
Surface Etching

Blistered
Discolored

Blistered
Edge Cracking
Discolored

Severe Weight Loss
Etched

Embrittled
Cracked

Some Etching
Discoloration

Etched
Cracked
Delaminated
Peeled
Coating Undamaged

Badly Attacked

Unaffected

Unaffected
                                           244

-------
ro
                                Figure  59.   Coupon  Mounting  Rack  and  Test  Specimens

-------
fNJ
-P>
CTl
                      Figure 60.  31
                                  r „
316L Cres Coupons Showing Signs of Crevice  Attack
Experiment 01 (L) and Smooth Bottom  Pitting During
Experiment 03 (R)

-------
ro
                    Figure 61.  304 Stainless Steel Coupons from Experiment 03 at and
                                Above Water Line (L) and Below Water Line (R)

-------
        •    Nickel Base  Metals  and  Alloys.   Coupons  of pure nickel  and
            of  various  Inconel  and  Incoloy  alloys  were tested.   The pure
            nickel coupon  dissolved completely  during  Experiment 01.
            Inconel  601  showed  a  weight  loss in excess of 4%.   It pitted
            badly  around the  bolts  which attached  it to the mounting
            rack,  and also showed significant attack in noncrevice  areas.
            The Inconel  617 coupon  was subject  to  similar crevice cor-
            rosion and  noncrevice pitting,  but  experienced only a
            0.18%  weight loss.   Inconel  625 and Incoloy 825 samples
            were not as  prone to  noncrevice attack,  but did show
            significant  crevice corrosion (Figure  62).  Weight  losses
            of  0.055% and  0.31%,  respectively,  were  recovered.   Coupons
            of  Hastelloy G and  Hastelloy C-276  mounted in the R-l
            reactor  during Experiment 03 showed no visible signs of
            corrosion when removed  and examined (Figure 63).

        t   Titanium.   Samples  of commercially  pure  titanium (Ti-6&A)
            and of Ti-12 alloy  were tested  in Experiment 01.  Neither
            coupon was  visibly  attacked  (Figure 64).   Weight losses were
            less than 0.01% for both coupons.  Ti-50A  coupons tested
            in  Experiment  03  are  also shown in  Figure 64).  They showed
            no  visible  signs  of corrosion. -Normally,  titanium  can
            be  attacked  by low  concentrations of sulfuric acid.
            However, ferric ion suppresses  this attack and acts  as
            an  effective corrosion  inhibitor(21).  The corrosive
            effects  of  fluorides  are not known, however,  the pres-
            ence of  calcium is  known to  tie up  fluorides  and thus
            act as a corrosion  inhibitor in titanium applications^2) _
            It  is  expected that enough calcium  is  present in the
            desulfurization environment  to  prevent fluoride attack.

        •   Lead.  The  lead samples had  a yellow and red  coating
            which  was adherent.  This was probably lead sulfate.  The
            surfaces of the lead  samples were not  pitted  (Figure 65).

     Non-Metal  Coupons.  The  nonmetallic coupons were  removed from  the R-l

reactor on 29 November  1977.  The specimens were examined and evaluated  for
evidence of degradation  and the results  are as  follows:

        •   Ashland  197/3-400-7,  FRP High Temperature  Polyester
            Resin.  The  fiberglass  specimen underwent  a 1.17%
            weight loss.  The coated edges  of the  sample  were
            cracked  as  was the  surface gel  coat.   The  fiberglass-
            to-resin bond  was obviously  attacked as  evidenced by
            the pronounced fiber  prominence at  the specimen
            surface. The  portion of the sample covered by the
            Teflon mounting plate was not discolored whereas the
            exposed  portion was red-brown in color.
                                     248

-------
Figure 62.  Inconel and Incoloy Alloy Coupons, 7.5X-'
            Inconel 601 (Upper L), Inconel 617 (Upper
            R), Inconel 625 (Lower L), Incoloy 825
            (Lower R)
                           249

-------
tNJ
cn
O
                           Figure 63.  Hastelloy C-276 Coupons from Experiment  03 at  and  Above
                                       Water Line  (L) and Below Water Line  (R)

-------

Figure 64.
Titanium Coupons:  Ti-50A (L)  and
Ti-12 (R), from Experiment 01  (Top)
and Experiment 03 (Bottom)
                         251

-------
Figure 65.   Lead  Coupons  from  Experiment  01.
            Surface Marked  by  Cleaning  Operation.
            No  Pitting  or Crevice  Corrosion.  1.5X
                 252

-------
       •   Ashland  197/3 + AT-8-4 (Flame Retarded Version of 197/3).
           The  surface of the specimen was blistered, but otherwise'
           showed no visible signs of surface attack.  The weight
           loss was almost negligible (0.50%).  The sample, however,
           was  drastically embrittled and the surface gel coat
           cracked  during handling.

       •   Ashland  800/801L-68-26, Furfuryl Alcohol (Furan) Resin.
           The  weight loss on this sample was low, 0.95%.  The sur-
           face however, was badly etched except where covered by the
           Teflon mounting plate.  Close examination of the surface
           showed it to be badly crazed and pitted.  Surface veil
           fibers were exposed in the etched areas.

       •   Ashland  800FR-10-19 (Flame Retarded Version of 800/801).
           The  weight loss on this sample was much more severe (2.50%).
           The  surface etching and veil fiber exposure was similar
           to that  of the 800/801 specimen.  Again, the area under
           the  mounting plate was largely untouched.

       •   Ashland  7240-4-37, Unfilled Polyester Resin.  The weight
            gain on  this sample was almost negligible (0.01%).   How-
            ever the sample was badly degraded.  The surface was
            blistered and the glass reinforcement was very prominent
            due to degradation of the glass fiber-to-resin bond.  The
            surface, except for the portion under the mounting plate,
           was  a dark brown in color.

        •    Ashland 7241-6-27, Unfilled Polyester Resin.  This sample
            showed degradation similar to that of the 7240 resin
            sample.   The edge coating was cracked and in many places
            nonexistent.  Additionally, the sample was warped.

        •    Ashland 72L + Sb-1-23, (Flame Retarded Version of 7240
            and 7241 Resin).  This specimen showed a 2.50% weight
            loss.  The surface was badly cracked and etched except
            under the mounting plate.  Additionally, small blisters
            were evident over all the sample surface and the edge
            resin coating was cracked.

        •    Fiberglass Reinforced Polyester Pultrusion.  The specimen
            was badly embrittled and cracked during  handling.  The
            area around the mounting hole ripped during removal from
            the mounting plate.

Reinforced Coatings

        •    Plasites 4020, 4030, and 4092 Filled Vinyl  Ester Coatings
            on Steel.  These specimens, although showing  some surface
            etching and discoloration,  seem to survive.   However, their
            long term resistance is impossible to  predict on the basis
            of this short exposure.

                                     253

-------
Coatings
            Ceil cote 252, Flakeglass Filled Polyester on Carbon Steel.
            The surface of the coating appeared to be etched and the
            flakeglass filler exposed.  The coating was cracked in one
            corner and the steel  substrate exposed.  This was probably
            the result of exposure to high temperature.
        t   Kynar 202/205 Coating on Carbon Steel (Approximately .006-
            inch thick).  The coating delaminated and cracked, probably
            because of differential thermal expansion and loss of
            adhesion due to moisture permeation.  The steel almost
            completely corroded away, but the coating film looked like
            new.

        •   Kynar 202/205 Coating on 304 Stainless Steel (Approximately
            .006-inch thick).  The coating was also cracked and peeled
            from this specimen although some was still adhering well to
            the surface.  Some crevice corrosion was exhibited under
            the attachment hole after the coating was peeled back,
            however the 304 was largely protected by the Kynar.

        t   Kynar Sheet Stock Heat Sealed Coating on Carbon Steel
            (Pennwalt Sample).  This specimen was thicker than the
            other Kynar samples.  It showed no signs of degradation
            or substrate corrosion.  The greater thickness of the
            sample enabled it to withstand thermally induced stresses.

        t   EA 919 Epoxy Coating On 304 Stainless Steel.  This specimen
            was badly attacked.  The coating could not be found and
            the stainless steel substrate crumbled into powder.  The
            epoxy coating probably failed by hydrolytic breakdown of
            the resin in the hot oxidizing acid solution as well as
            thermal cracking and moisture permeation, which caused
            loss of adhesion to the steel substrate.

        •   Viton A Sheet Stock Bonded with RTV 732 Silicon Adhesive
            (Dow-Corning) to Carbon Steel.  The Viton A was unaffected
            with no change in hardness and the steel substrate suffered
            minor corrosion because of a small leak in the RTV peri-
            pheral bond seal.  The silicone showed signs of edge
            attack - small cracks and chalking.  Viton A may be consid-
            ered as a liner material; however, a suitable bonding and
            sealing agent must be found.

        •   AF-E-332 Sheet Stock (EPDM), Bonded to Carbon Steel with
            RTV 731 (Dow-Corning).  The silicone rubber bond failed
            in adhesion to the EPDM (Ethylene Propylene Diene Monomer)
            surface and the carbon steel substrate completely dis-
            solved.  The EPDM showed no signs of degradation and change
            in hardness.  The silicone seal suffered edge attack
            similar to the RTV 732; that is, extensive edge cracking

                                    254

-------
            and chalking.  Although  the  EPDM itself was  unaffected  by
            process conditions,  bonding  and  sealing of the material
            to the metal substrate presents  a very difficult  problem.
     Stress Corrosion Cracking Specimens.  None  of the five SCC  specimens
(titanium, welded titanium,  Incoloy  825,  316L Cres, Inconel 625)  showed attack.
However, the 304 Cres nuts failed, thus  relieving  the  stress  at  some point
during the run so that  the test  results  are  inconclusive.  Some  crevice attack
occurred on the 316L Cres and the  Incoloy 825 samples  in  the  attachment area.
     Coupon Mounting Rack.   The  316L Cres support  rack exhibited  crevice
corrosion at the mounting holes  and  localized pitting  attack  on  its edges.
This attack was not associated with  welds.   The  Teflon portion of the rack
was unattacked although cold flow  had occurred under bolt heads which had
been torqued to a high  value.  The PTFE  Teflon mounting plates showed no
signs of degradation, either from  chemical attack  or abrasion.
     The Ti-6Al-4V rack mounting bolts appeared  not to be attacked during the
test but the 304 Cres nuts and 316 Cres  bolts were severely attacked and, in
some cases, dissolved.
     The Hystl 6793-175 (glass reinforced, alumina trihydrate filler polybuta-
diene resin) mounting plate  was  softened and discolored.  The plate cracked
during handling and a cross  section  showed almost  complete penetration of the
leach solution into the material.  Again,  the hot  oxidizing acid  conditions
appear to be the cause  of the failure.
10.2.2  Orifice Plate Specimens
     Five metallic and  two elastomeric orifice plate specimens were tested
during Experiment 01.   The metallic  plates were  mounted  using gaskets while
the elastomeric plates  acted as  gaskets.   The metallic materials  tested were
Inconels 601, 617, and  625,  Incoloy  825, and Titanium  50A.  The  Inconels
exhibited crevice corrosion  under  the gasket, although only a few pits were
noted in the Inconel 625 plate.  The corrosion was found  on both sides of the
Inconel 601 plate and on the downstream  side of  the Inconel 617  and 625
plates.  Pitting on the exposed  surface  was  noted  for  Inconel 601 and incipient
pitting (mottled surfaces) for  Inconel 617.   Incoloy 825 was  severely attacked
in the crevice area on  both  sides  of the orifice plate.   The  Titanium 50A plate
was unaffected by the exposure and no corrosive  attack was  detected.
                                     255

-------
     The Hypalon (chlorosulfonated polyethylene) orifice plate showed no
signs of abrasion or tearing.  The elastomer showed no decrease in cut resist-
ance and in Shore A hardness.  The Neoprene (chloroisoprene) orifice plate
showed no signs of abrasion.  However the Shore A hardness jumped ten points
and material subjected to the hot sulfuric acid solution in the orifice area
was torn due to a dramatic decrease in its localized tear resistance.
10.3  DISCUSSION
     The observed corrosion in the coal desulfurization reactor and pump
circuit components was extremely severe.  The short operating time under full
process conditions indicates that very rapid 316L corrosion rates are being
experienced.  The spongy, porous nature of some corroded areas suggests that
serious damage can result even though visual inspection of the surface
indicates little or no attack.  In many cases, scraping what appeared to be
a slightly attacked region resulted in material crumbling away to reveal
large, deep pits.  In one instance, a pit was detected in the R-l reactor
which penetrated about 25% through the wall.
     The porous nature of the observed pitting is of critical  concern.   It
appears that once a pit has been initiated, either by crevice conditions  or
by chloride attack on a free surface, that the corrosion follows metallurgical
paths offering the least resistance, such as grain boundaries, stringers, cold
worked areas, etc.  This could result in multiple very small but relatively
deep subpits within the main pit.  At the root of these subpits, the local
conditions would be much like a crevice so that an  oxygen concentration  cell
can exist.  This is, in essence, an electrolytic cell where the driving force
to cause corrosion is an oxygen concentration gradient between the root of
the subpit (low concentration) and the bulk of the solution (high concentra-
tion).  These conditions would cause the subpit to rapidly extend producing
the sponge-like structure.  This possible mechanism could be worsened by
high oxygen contents in the main slurry.  It is interesting to note that
neither crevice corrosion nor spongy structure was prevalent in the T-2 mixer
which is oxygen free.  While some crevice initiated pits were found in T-2,
such as under the masking tape, the pits appeared to be shallow and smooth
bottomed.
                                    256

-------
     The coal desulfurization  environment is extremely agressive  to  stainless
steel, especially for  the  type of attack observed.   The relatively high tem-
peratures, low PH (PH  of 1 to  2), crevice conditions,  and  high  oxygen content
of the slurry greatly  accelerate crevice (concentration cell) attack.  Pitting
attack is greatly accelerated  by the presence of halide ions and  the reaction
sustained by the presence  of a metal ion, such as ferric or cupric ion, which
can be reduced to provide  the  cathode reaction.   Pitting rates  are also
greatly accelerated  by high temperature, low pH, high  velocity, and high
oxygen content.  Rough surfaces or surface deposits, internal or  external
stresses on metal parts, and metallurgical  defects,  such as inclusions, pores,
and stringers, are especially  susceptable to pitting attack.  All of these
factors play a part  in the corrosion behavior of the desulfurization unit  to
varying degrees in different parts of the system.
     During the course of  the  RTU corrosion study, it  was discovered that
similar material problems  had  been encountered in a Sherrit-Cominco copper
ore processing pilot plant in  Fort Saskatchewan,  Alberta^18',   In this  process,
aggressive slurries  must be handled, which  are similar to those encountered
in the Meyers Coal Desulfurization Process.   Significant corrosion problems
were experienced before final  material  selections were made.  Results from the
Sherrit-Cominco pilot  plant indicate that near ambient temperature storage
tanks and thickeners constructed of fiberglass reinforced plastic have  per-
formed well.  Where  temperatures do not exceed 190°F,  rubber lined mild steel
has been used satisfactorily.   At temperatures up to 230°F and  1% to 5%
sulfuric acid concentrations,  316 Cres  can  be used if  a high (50  to 100 g/i)
concentration of CuSCL is  maintained.   (Areas which do not contain significant
copper sulfate concentration experience rapid failure  of 316 Cres components.)
Where temperatures reach 350°F to 400°F, titanium lines, valves,  and heat
exchangers are used  and perform well.   Pumps and lines subjected  to severe
abrasion and aggressive slurry compositions are constructed of  Hastelloy
C-276, selected as a result of a development program during which several
other materials were eliminated.  The above mentioned  information appears  to
be very consistent with RTU related observations.
                                      257

-------
10.4  CORROSION STUDY CONCLUSIONS
     The obvious conclusion drawn from the inspection and analysis of the
system components and test samples is that unprotected 316L Cres is an
unacceptable material for the R-l reactor and pump around loops when they are
operated at nominal process conditions (02 present, .2% to 5% iron, .2% to
4% sulfuric acid, and temperatures greater than 215°F).
     Commercially pure titanium (Ti-50A) and titanium alloy Ti-12 appear to
be suitable materials of construction for the R-l reactor.  The Ti-12 alloy
may be required at flanges, mounting rings, etc., since it is more resistant
to crevice attack than Ti-50A.  The cost of pure titanium and Ti-12 is about
2 to 2-1/2 times the cost of 316 stainless but less expensive than Hastelloy
 (23)
Cv  '.  This cost differential has been verified by a comparison of the
existing RTU reactor cost and a 1978 vendor quote for a replacement reactor
constructed of Ti-12.  It should be noted that while titanium has performed
well in short term test runs on the RTU, long term tests are needed to verify
its acceptability.
     Based on tests conducted during Experiment 03 and similar experience at
Sherrit-Cominco, Hastelloy C-276 may also be able to withstand the hostile
chemical environment of the Meyers Process reactor system.  Hastelloy alloys
are prime candidates for pump-around loop components.
     Various metallic and non-metallic materials may be considered for compon-
ent liners.  Lead is a possibility as a lining material but long term corrosion
rates and erosion resistance must be further evaluated.  Most of the non-
metallic samples were attacked to some degree by the high temperature sulfuric
acid/iron sulfate solution.  Notable exceptions were Viton A, EPDM, and Teflon.
An effective means of bonding and sealing these materials to a metal substrate
must be found before they can be considered practical as liners.
     For applications in Meyers process equipment operating at or near leach
solution normal boiling points (up to 213°F) in the absence of oxygen, results
indicate that 316L and Fiberglass reinforced plastics (FRP) may be suitable
materials of construction.  There are also a number of coatings which also
appear to be satisfactory in this application as discussed in Section 10.2.1.
It should be stressed that all results are based on relatively short test

                                     258

-------
durations.  Therefore, an organized long term, materials testing activity
would be desirable to conclusively determine the merits of the studied
materials in Meyers Process applications.
                                    259

-------
                                11.   REFERENCES

1.  Hamersma, J. W.  and M.  L.  Kraft.   Applicability  of the  Meyers  Process  for
    Chemical  Desulfurization of Coal:   Survey of Thirty-Five Coals.   Report  No.
    EPA-650/2-74-025a prepared by TRW  Systems and Energy  for the U.S.  Environ-
    mental Protection Agency under Contract  No.  68-02-0647, Washington, D.C.,
    1975.

2.  Hamersma, 0. W., M. L.  Kraft, R. A.  Meyers,  C. A.  Flegal and A. A. Lee.
    Applicability of the Meyers Process  for  Chemical Desulfurization  of Coal:
    Initial Survey of Fifteen Coals.   Report No.  EPA-650/2-74-025  prepared by
    TRW Systems and Energy for the U.S.  Environmental  Protection Agency under
    Contract No. 68-02-0647, Washington, D.C., 1974.

3.  Hamersma, J. W., M. L.  Kraft, E.  P.  Koutsoukos and R. A. Meyers.   Chemical
    Removal of Pyritic Sulfur from Coal.  In:  Advances in  Chemistry,  Series
    No. 127, American Chemical Society,  Washington,  D.C., 1973.

4.  Hamersma. J. W,, E. P.  Koutsoukos, M.  L.  Kraft,  R. A. Meyers,  G.  J. Ogle and
    L. J. Van Nice.  Program for Processes for the Selective Chemical  Extraction
    of Organic and Pyritic Sulfur from Fossil  Fuels.   Report No. 17270-6011-
    RO-00, Vol. I and II, prepared by  TRW  Systems and  Energy for the  U.S.
    Environmental Protection Agency,  Research Triangle Park, under Contract
    No. EHSD 71-7, No. Carolina, 1973.

5.  Meyers, R. A., E. P. Koutsoukos, M.  L. Kraft, R. A. Orsini, M. J.  Santy and
    L. J. Van Nice.  Meyers Process  Development  for  Chemical Desulfurization of
    Coal.  Report No. EPA-600/2-76-143a, Vol.  I  and  II, prepared by TRW Systems
    and Energy for the U.S. Environmental  Protection Agency, Research  Triangle
    Park, No. Carolina, 1976.

6.  Van Nice, L. J., and M. J. Santy.   Pilot Plant Design for Chemical Desulfuriza-
    tion of Coal.  Environmental Protection  Technology Series, EPA-600/2-77-080,
    1977.
                                      260

-------
                            11.  (Continued)

 7.   Nekervis,  W.  F., and E. F. Hensley.  Conceptual  Design  of a Commercial Scale
     Plant for  Chemical  Desulfurization of Coal.   Environmental Protection Technology
     Series,  EPA-600/2-75-051, 1975.

 8.   McGee, E.  M.   Evaluation of Pollution Control  in Fossil  Fuel Conversion
     Processes, Coal Treatment:  Section 1, Meyers  Process.   Environmental
     Protection Technology Series, EPA-650/2-74-009K, 1975.

 9.   Meyers,  R. A., M. J. Santy, E. P. Koutsoukos,  L. J.  Van  Nice and R. A. Crsini.
     Reactor Test Project for Chemical Removal  of Pyritic Sulfur from Coal.
     Report No. 25305-6023-TU-OO prepared by TRW  Systems  and  Energy for U.S.
     Environmental Protection Agency, under Contract  No.  68-02-1880, Research
     Triangle Park, No.  Carolina, 1977.

10.   Meyers, R. A.  Reactor Test Project for Chemical Removal of Pyritic Sulfur
     from Coal.  Report No. 25305-6038-TU-OO prepared by  TRW  Systems and Energy
     for the U.S. Environmental Protection Agency under Contract No. 68-02-1880,
     Research Triangle Park, No. Carolina, 1977.

11.   Meyers. R. A., L. J. Van Nice, E. P. Koutsoukos, M.  J.  Santy and R. A. Orsini.
     Bench-Scale Development of Meyers Process for  Coal Desulfurization.  Report
     No. 28456-6025-TU-OO prepared by TRW Systems and Energy for the U.S.
     Environmental Protection Agency under Contract No. 68-02-2121, Research
     Triangle Park, No.  Carolina, 1977.

12.   Yurovskii, A. J.  Sulfur in Coals.  U.S. Department  of  Commerce, National
     Technical  Information Service, Springfield,  Virginia, 1974.

13.   Mills, H.  E.  Costs of Process Equipment.  In:  Modern  Cost Engineering
     Techniques, H. Popper, eds. McGraw-Hill, 1970. pp 111-134.
                                      261

-------
                           11.   (Continued)

14.  Happel, J.  and D-  Jordan.   Chemical  Process  Economics,  2nd  ed.,  Marcel
     Dekker, Inc., 1975.

15.  Guthrie, K. M.  Capital  Cost Estimating.  In:  Modern Cost  Engineering
     Techniques, H. Popper,  eds.  McGraw-Hill,  1970.   pp 80-108.

16.  Perry, R. and C. Chilton, Chemical  Engineer's Handbook,  5th ed., McGraw-
     Hill, 1973.

17.  Chemical Marketing Reporter, Schnell  Publishing  Co.,  1977.

18.  TRW Correspondence with  Sherrit-Cominco,  Fort Saskatchewan, Alberta, Canada.

19.  The Synthetic Gas-Coal Task  Force for the Supply-Technical Advisory Com-
     mittee, National Gas Survey, Federal  Power Commission.   Final Report -
     The Supply - Technical Advisory  Task  Force - Synthetic Gas - Coal. 1973.

20.  Graver, D.  L.  Pitting Corrosion of Stainless Steel.  In:  Technical Horizons,
     Allegheney Ludlum  Steel  Corp., Pittsburgh, Pennsylvania, 1961.

21.  Covington,  L. C.  The Role  of Multi-Valent Metal Ions in Suppressing Crevice
     Corrosion of Titanium.   In:   Titanium Science and Technology, R. I. Jaffee
     and H. M. Burte, eds. Plenum Press, New  York, 1973.   pp  2395-2403.

22.  TRW Correspondence with  L.  C. Covington,  Timet,  Henderson, Nevada.

23.  Outlook for Titanium Brightens with CPI  Gains.   In:   Chemical Engineering,
     December 19, 1977.  pp 40-42.
                                    262

-------
                   12.  GLOSSARY OF ABBREVIATIONS AND SYMBOLS
Abbreviations
    Abs
    ASTM
    Btu
    cal
    eq
    Exp.
    Kcal
    No.
    wt
Symbol s

    \

    AR
    C
    A
    E
    u
    p
absolute
American Society of Testing Materials
British Thermal Unit
calories
equation
experiment
kilocalories
number
weight
Arrhenius constant in leach reaction (hours)"1
(wt % pyrite in coal)"1
Arrhenius constant in regeneration reaction
(minutes)'1 (atm)-1 (liters/mole)
concentration
difference in quantity following delta
activation energy for pyritic sulfur leaching
reaction, Kcal/mole
activation energy for ferric ion regeneration
reaction, Kcal/mole
pyritic sulfur leaching rate constant (units
same as A. )
ferric ion regeneration rate constant (units
same as AR)
micron
total pressure, atmospheres
oxygen partial pressure
gas constant, cal/mole - K
                                     263

-------
Symbols (cont'd)
    r.                   pyritic sulfur leaching rate, weight of pyrite
                         removed per 100 wts of coal  per hour
    rR                   ferric ion regeneration rate, moles per liter
                         per minute
    S                    elemental  sulfur
    S                    organic sulfur
    S                    pyritic sulfur
    S                    sulfate sulfur
    S                    total  sulfur
      U                                         -_
    T                    absolute temperature,   K
    t                    time,  hours (leaching)-minutes (regeneration)
    W                    pyrite concentration in coal, wt %
    Y                    ferric ion to total  iron ratio
                                     264

-------
1. REPORT NO.
 EPA-600/7-79-013a
                                TECHNICAL REPORT DATA
                         iriease read Instructions on the reverse before completing}
           3. RECIPIENT'S ACCESSION NO.
Reactor Test Project for Chemical Removal of
  Pyritic Sulfur from Coal; Volume I.  Final Report
           5. REPORT DATE
            January 1979
           6. PERFORMING ORGANIZATION CODE
R.A.Meyers, M.J.Santy, W.D.Hart,
   L.C.McClanathan, and R. A.Orsini
                                                     8. PERFORMING ORGANIZATION REPORT NO.
  PERFORMING ORGANIZATION NAME AND ADDRESS
TRW, Defense and Space Systems Group
One Space Park
Redondo Beach,  California  90278
           10. PROGRAM ELEMENT NO.
           EHB527
           11. CONTRACT/GRANT NO.

           68-02-1880
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC 27711
           13. TYPE OF REPORT AND PERIOD COVERED
            Final: 6/75 - 6/78   	
           14. SPONSORING AGENCY CODE
             EPA/600/13
^SUPPLEMENTARY NOTES ffiRL-RTP project officer is Lewis D. Tamny, Mail Drop 61, 919/
& 109.
is. ABSTRACT The report giveg results of an evaluation of the initial performance of the
Reactor Test Unit (RTU) for chemical removal of pyritic sulfur from an Appalachian
coal. Operation of the plant--from its checkout and shakedown in September 1977
through January 1978--demonstrated that the RTU could be run continuously in three-
shift operation to reduce th$ coal from 2.4 Ib SO2/million   Btu to a level of 1.0 to
1. 2 Ib SO2/million Btu, after rinsing and extraction of generated elemental sulfur.
There was no measurable coal oxidation during processing and leach rates in the
plant were greatly improved over bench-scale values.  The leach solution/coal/oxy-
gen environment was found to be corrosive to the installed stainless steel reactor,
necessitating future  upgrading to support additional testing. Bench-scale experiments
showed that the leach solution can be used as a homogeneous dense-media to effi-
ciently gravity-separate coal prior to processing. Beneficial engineering cost impro-
vements are obtained based on using this approach, resulting in capital cost estima-
tes of $68-69/kW and with #0.44-0. 50/million Btu processing costs, including amor-
tization of capital, for input coal costing $0. 78-0. 81/million Btu. Overall energy ef-
ficiency was 93-96%.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                          b.IDENTIFIERS/OPEN ENDED TERMS
                          COSATI Field/Group
 Pollution
 Coal
 Desulfurization
 Pyrite
 Chemical Cleaning
 Rinsing
 Extract
Pollution Control
Stationary Sources
Pyritic Sulfur
13B
08G,21D
07A,07D

13H
18. DISTRIBUTION STATEMENT
                                           Unclassified
                            277
 Unlimited
EPA Form 2220-1 (9-73)
20. SECURITY CLASS (Thispage)
Unclassified
                                                                  22. PRICE
                                        265

-------