DOC, E 1.28: FE-1734- 42/V.9/ bk.1 BOOKSTACKS. DOCUMENTS FE-1734-42(Vol.9)(Bk.1 SUPPORT STUDIES BY SOUTH DAKOTA SCHOOL OF MINES AND TECHNOLOGY: C0 2 ACCEPTOR PROCESS GASIFICATION PILOT PLANT Final Report, Volume 9, Book 1 of 2 February 1971 -January 1978 Work Performed Under Contract No. EX-76-C-01-1734 Research Division Conoco Coal Development Company Library, Pennsylvania and South Dakota School of Mines and Technology Rapid City, South Dakota #0* ,tf° i#> ^ \^ oyS U. S. DEPARTMENT OF ENERGY UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPA1GN STACKS NOTICE This report was prepared as an account of work sponsored by the United States Government. Ne.ther the United States nor the United States Departm M of Energy no any of their employees, nor any of their contractors, subcontractors, or he/emXees makes any warranty, express or implied, or assumes any legal liability or relonsSv for' fte accuracy, completeness or usefulness of any information, appa^^T'or proce" d 1S closed, or represents that its use would not infringe privately owned righ s ? This report has been reproduced directly from the best available copy. Available from the National Technical Information Service, U. S. Department of Commerce, Springfield, Virginia 22161. Price: Paper Copy $13.25 Microfiche $3.00 FE-1734-42(Vol.9)(Bk.1) Distribution Category UC-90c SUPPORT STUDIES BY SOUTH DAKOTA SCHOOL OF MINES AND TECHNOLOGY C0 2 ACCEPTOR PROCESS GASIFICATION PILOT PLANT FINAL REPORT, VOLUME 9 BOOK 1 OF 2, REPORTS PERIOD: FEBRUARY 1971 - JANUARY 1978 CONOCO COAL DEVELOPMENT COMPANY RESEARCH DIVISION LIBRARY, PENNSYLVANIA 15129 AND SOUTH DAKOTA SCHOOL OF MINES AND TECHNOLOGY RAPID CITY, SOUTH DAKOTA 57701 PREPARED FOR UNITED STATES DEPARTMENT OF ENERGY AND AMERICAN GAS ASSOCIATION UNDER CONTRACT EX 76-C-01-1734 TABLE OF CONTENTS Page Abstract .... 1 SECTION 1 SUMMARY 1.1 Introduction 1.2 Acceptor Reconstitution Studies. . . . ... 3 1.3 Studies in the CaC0 3 -CaS0 4 -CaS and CaC0 3 -CaCOH) 2 -CaO Systems 1.4 Corrosion of Alloys in Contact with CaCOx-CafOHU Melts z 1.5 Carbonate Rock Resources Studies ...... ' ' l 1.6 Grinding Properties of Carbonate Acceptor ' in 1.7 Pregasification Beneficiation. ... 7, 1.8 Trace Element Studies ' \ ? 1.9 Waste Monitoring 1.10 Special Projects ]t SECTION 2 ACCEPTOR RECONSTITUTION STUDIES 2.1 A Study of the Reconstitution of CaC07-Ca(0H) 9 Acceptors, Final Report, June 30, 1975 . 17 2.2 Extension of Acceptor Life by Addition of' Impurities, Interim Report, July 15, 1974. 45 2.3 Acceptor Reconstitution, Interim Report July 31, 1972 . '. SECTION 3 STUDIES IN THE CaC0 3 -CaS0 4 -CaS AND CaC0 3 -Ca(OH) 2 -CaO SYSTEMS 3.1 Determination of Liquidus Temperatures in the CaC0 3 -CaS0 4 -CaS and CaC0 3 -Ca(OH) 2 -CaO Systems 71 SECTION 4 CORROSION OF MATERIALS IN CONTACT WITH CaC0 3 -Ca(OH) 2 MELTS 4.1 The Corrosion of Materials in Contact with CaCO.-CafOH) 9 Melts, Final Report, June 30, 1977 2 4 ' 2 M h ?. COr T° Si0n ° f Materials in Contact with CaC0,-Ca(0H) 7 Melts, Interim Report, June 30, 1975 .... 2 .91 125 11 Page SECTION 5 CARBONATE ROCK RESOURCES STUDIES 5.1 Investigation of Carbonate Rock Resources in the Logan, Montana, Area 148 5.2 Preliminary Resource Study of Carbonate Rocks Available for Lignite Gasification in Central Montana, Northern Wyoming, and Western South Dakota 161 SECTION 6 GRINDING PROPERTIES OF CARBONATE ACCEPTOR 6.1 Study of Physical Characteristics of Carbonate Rocks to Determine Wear Properties and Crushing Properties .... 211 SECTION 7 PREGASIFICATION BENEFICIATION 7.1 Sodium Removal from High-Sodium Lignite by Ion Exchange 222 7.2 Sodium Removal from Lignite by Ion Exchange, Interim Report, March 31, 1973 251 SECTION 8 TRACE ELEMENT STUDIES 8.1 Trace Elements in the Lignite CO2 Acceptor Process Gasification Pilot Plant Runs 320 8.2 Trace Elements in Lignite 338 8.3 Sulfur Studies 347 SECTION 9 WASTE MONITORING 9.1 Waste Monitoring at the CO2 Acceptor Process Gasification Pilot Plant, Final Report, June 21, 1977. . . . 357 9.2 Waste Monitoring at the CO2 Acceptor Process Gasifica- tion Pilot Plant, Interim Report, October 1, 1974 376 References 390 ill LIST OF FIGURES AND TABLES Page Figure 2-1. Cooling Curve Obtained with a Mixture Containing Calcium Carbonate, Calcium Hydroxide, and Magnesium Oxide at 350 PSIG 19 Figure 2-2. Regions of Fusion in the Calcium Hydroxide- Calcium Carbonate System in the Absence and Presence of Magnesium Oxide at 1400°F and 600 PSIG 21 Figure 2-3. Activity of Eutectic-Composition Melts Containing 7.5 Mole Percent Calcium Phosphate Prepared at Three Different Times as a Func- tion of Cycling Between Calcination and Carbcnation 2 x Figure 2-4. Activity of Eutectic-Composition Melts Con- taining Various Impurities at One Mole Per- cent _. 24 Figure 2-5. Activity of Eutectic Composition Melts Containing Various Additions of Aluminum 0xide 25 Figure 2-6. Activity of Eutectic-Composition Melts Containing Various Additions of Ferric Oxide ~, zo Figure 2-7. Activity of Eutectic-Composition Melts Containing Various Additions of Phosphorous Pentoxide 27 Figure 2-8. Activity of Eutectic-Composition Melts Containing Various Additions of Calcium Phosphate 2 e Figure 2-9. Activity of Eutectic-Composition Melts Containing Various Additions of Basic Calcium Phosphate 29 Figure 2-10. Activity of Eutectic-Composition Melts Containing Various Additions of Calcium Molybdate 31 Figure 2-11. Activity of Eutectic-Composition Melts Containing Various Additions of Calcium Silicate , 2 Figure 2-12. Activity of Eutectic-Composition Melts Containing Various Additions of Calcium Sulfate 33 IV LIST OF FIGURES AND TABLES (continued) Page Figure 2-13. Activity of Eutectic-Composition Melts in the Absence and Presence of 7.5 Mole Percent Calcium Phosphate 34 Figure 2-14. Activity of Eutectic-Composition Melts Con- taining 7.5 Mole Percent Calcium Phosphate at Two Calcination Temperatures 36 Figure 2-15. Activity of Eutectic-Composition Melts Con- taining 6.12 Mole Percent Ferric Oxide at Two Calcination Temperatures 37 Figure 2-16. Activity of Eutectic-Composition Melts Pro- duced from Reagent Grade Chemicals and Spent Acceptor 39 Figure 2-17. Activity of Eutectic-Composition Melts Con- taining Various Impurities at 1 Mole Percent as a Function of Cycling 47 Figure 2-18. Activity of Eutectic-Composition Melts Containing Various Additions of SrC03 48 Figure 2-19. Activity of Eutectic-Composition Melts Containing Additions of Fe203 and Si02 49 Figure 2-20. Activity of Eutectic-Composition Melts Containing Various Additions of AI2O3 50 Figure 2-21. Activity of Eutectic-Composition Melts Containing Various Additions of P2O5 51 Figure 2-22. Activity of Eutectic-Composition Melts Containing Various Additions of Ca3(P04)2 52 Figure 2-23. Activity of Eutectic-Composition Melts Containing Various Additions of Ca 10 (OH) 2 (P0 4 ) 6 53 Figure 2-24. Activity of Eutectic-Composition Melts Containing Various Additions of Calcium Sulfate 55 Figure 2-25. Activity of Eutectic-Composition Melts Containing Various Additions of CaSi03 56 Figure 2-26. Activity of Eutectic-Composition Melts Containing Various Additions of Calcium Molybdate 57 Figure 2-27. Cooling Curve Obtained with a Mixture Containing Calcium Carbonate, Calcium Hydroxide, and Magnesium Oxide at 350 PSIG 62 LIST OF FIGURES AND TABLES (continued) Page Figure 2-28. Regions of Fusion in the CaCOH^-CaCO^ System in Absence and Presence of MgO at 1400°F and 600 PSIG 64 Figure 2-29. Normalized Acceptor Activity as a Function of Number of Calcination-Carbonation Cycles .... 67 Figure 3-1. Apparatus in Which Melt Formation was Effected. . . 74 Figure 3-2. Electrical Circuit Utilized for Temperature Measurement 75 Figure 3-3. Apparatus Utilized for Carbonate and Hydroxide Analysis of Frozen Melts 77 Figure 3-4. Quartz Tube Reactor Utilized for the Decomposition of CaCC^ and Ca(0H)2 in Frozen Melts 79 Figure 3-5. Cooling Curve for a System Containing 30 Mole Percent CaCC>3 and 70 Mole Percent Ca(0H) 2 Initially at a Pressure of 660 PSIA 80 Figure 3-6. Liquidus and Solidus Temperatures in the CaS04-CaC0 3 Binary System 82 Figure 3-7. Areas of Melt Formation in the CaCC^-CaSG^- CaS System at 2100°F 83 Figure 3-8. Liquidus Temperatures in the CaCC^-CaSC^- CaS Ternary System 84 Figure 3-9. Liquidus Temperatures and Pressures in the CaC03-Ca(OH)2-CaO Ternary System 86 Figure 4-1. Corrosion of Type 304 Stainless Steel in a Stainless Steel Crucible in the Presence of Eutectic Composition Melt 99 Figure 4-2. Corrosion of Type 304L Stainless Steel in a Stainless Steel Crucible 100 Figure 4-3. Corrosion of Type 309 Stainless Steel in a Stainless Steel Crucible 101 Figure 4-4. Corrosion of Type 310 Stainless Steel in a Stainless Steel Crucible 102 Figure 4-5. Corrosion of Types 316 and 316L Stainless Steel in a Stainless Steel Crucible 103 Figure 4-6. Corrosion of Type 446 Stainless Steel in a Stainless Steel Crucible 104 Figure 4-7. Corrosion of Inconel 601 in a Stainless Steel Crucible 105 VI LIST OF FIGURES AND TABLES (continued) Page Figure 4-8. Corrosion of Nickel 200 in a Stainless Steel Crucible 106 Figure 4-9. Corrosion of Types 304 and 304L Stainless Steel in an Alumina Crucible 107 Figure 4-10. Corrosion of Types 309 and 310 Stainless Steel in an Alumina Crucible 108 Figure 4-11. Corrosion of Types 316 and 316L Stainless Steel in an Alumina Crucible 110 Figure 4-12. Corrosion of Type 446 Stainless Steel in an Alumina Crucible HI Figure 4-13. Corrosion of Inconel 601 and Nickel 200 in an Alumina Crucible 112 Figure 4-14. Microphotograph of Inconel 600 After Exposure to a Melt in an Alumina Crucible for 100 Hours 113 Figure 4-15. Corrosion of Several Alloys in Alumina Crucibles as a Function of Melt Composi- tion for a 100 Hour Contact Time 115 Figure 4-16. Corrosion of Several Alloys in Alumina Crucibles in the Presence of Melts Made From Spent Acceptor 117 Figure 4-17. Weight Loss of Type 304 Stainless Steel as a Function of Time 130 Figure 4-18. Weight Loss of Type 304L Stainless Steel as a Function of Time 131 Figure 4-19. Weight Loss of Type 309 Stainless Steel as a Function of Time 132 Figure 4-20. Weight Loss of Type 310 Stainless Steel as a Function of Time 133 Figure 4-21. Weight Loss of Type 316L Stainless Steel as a Function of Time 134 Figure 4-22. Weight Loss of Type 446 Stainless Steel as a Function of Time 135 Figure 4-23. Weight Loss of Inconel 601 as a Function of Time 136 Figure 4-24. Weight Loss of Nickel 200 as a Function of Time 137 Figure 4-25. Microstructure of Inconel 600 138 VII LIST OF FIGURES AND TABLES (continued) Figure 4-26. Figure 4-27. Figure 4-28. Figure 5-1. Figure 5-2. Figure 5-3. Figure 5-4. Figure 5-5. Figure 6-1. Figure 6-2. Figure 7-1. Figure 7-2. Figure 7-3. Figure 7-4. Figure 7-5. Figure 7-6. Figure 7-7. Figure 7-8. Figure 7-9. Figure 7-10. Figure 7-11. Figure 7-12. Figure 7-13. Figure 7-14. Figure 7-15. Figure 7-16. Page Micro-structure of Inconel 601 139 Microstructure of Nickel 200 140 Oxidation of Iron-Nickel-Chromium Alloys in Contact with a CaC0 3 -Ca(OH) 2 Melt 141 Percentage of Fines Produced for Jefferson Formation 152 Percentage of Fines Produced for Four Type Samples 153 Percentage of Fines Produced for Madison Limestone 154 General Index Map 162 Carbonate Distribution, Transportation, and Ownership Map (2 sheets) 163 Histograms of Crushed Rock Size Distribution Following One Blow of the Drop Hammer 215 Comparison of Wear Tests 217 Sodium Removal from Lignite I as a Function of Time and Particle Size 225 Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Function of Time and Aspect Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite Sodium Removal from Lignite as a Ratio 232 ,226 ,227 ,228 230 231 .233 .234 .235 .236 .237 ,238 ,239 ,240 241 vm LIST OF FIGURES AND TABLES (continued) Page Figure 7-17. Sodium Removal from Lignite II 242 Figure 7-18. Sodium Removal from Lignite I as a Function of Time and Solution Recycling 243 Figure 7-19. Sodium Removal from Lignite I -244 Figure 7-20. Sodium Removal from Lignite I 245 Figure 7-21. Sodium Removal from Lignite II 246 Figure 7-22. Sodium Removal from Lignite II 247 Figure 7-23. Sodium Removal from Lignite II 248 Figure 7-24. Sodium Removal from Beulah, North Dakota Lignite by Ion Exchange in Agitated Slurries. . . .254 Figure 7-25. Sodium Removal from Beulah, North Dakota Lignite 255 Figure 7-26. Sodium Removal from Beulah, North Dakota Lignite 256 Figure 7-27. Low Temperature Sodium Removal from Beulah, North Dakota Lignite by Ion Exchange in Agitated Slurries 257 Figure 7-28. Sodium Removal from Beulah, North Dakota Lignite by Ion Exchange in Agitated Slurries. . . .258 Figure 7-29. Sodium Removal from Glenharold Mine Lignite by Ion Exchange in Agitated Slurries 260 Figure 7-30. Sodium Removal from Glenharold Mine Lignite ■ -261 Figure 7-31. Sodium Removal from Glenharold Mine Lignite . . . .262 Figure 7-32. Low Temperature Sodium Removal from Glenharold Mine Lignite by Ion Exchange in Agitated Slurries 263 Figure 7-33. Low Temperature Sodium Removal from Glenharold Mine Lignite 264 Figure 7-34. The Effect of Percent Solids in Agitated Slurries on the Sodium Removal by Ion Exchange from Glenharold Mine Lignite 265 Figure 7-35. Sodium Removal from Burke County, North Dakota Lignite by Ion Exchange in Agitated Slurries. . . .267 IX LIST OF FIGURES AND TABLES (continued) Page Figure 7-36. Sodium Removal from Burke County, North Dakota Lignite 268 Figure 7-37. Sodium Removal from Burke County, North Dakota Lignite 269 Figure 7-38. Sodium Removal from Burke County, North Dakota Lignite 270 Figure 7-39. Low Temperature Sodium Removal from Burke County, North Dakota Lignite by Ion Exchange in Agitated Slurries 271 Figure 7-40. The Effect of Percent Solids in Agitated Slurries on the Sodium Removal by Ion Ex- change from Burke County, North Dakota Lignite. . .272 Figure 7-41. Sodium Removal from Bowman County, North Dakota Lignite by Ion Exchange in Agitated Slurries 274 Figure 7-42. Sodium Removal from Bowman County, North Dakota Lignite 275 Figure 7-43. Sodium Removal from Bowman County, North Dakota Lignite 276 Figure 7-44. Sodium Removal from Bowman County, North Dakota Lignite 277 Figure 7-45. Low Temperature Sodium Removal from Bowman County, North Dakota Lignite by Ion Exchange in Agitated Slurries 278 Figure 7-46. The Effect of Percent Solids in Agitated Slurries on the Sodium Removal by Ion Exchange from Bowman County, North Dakota Lignite 279 Figure 7-47. Sodium Removal from Ward County, North Dakota Lignite by Ion Exchange in Agitated Slurries. . . .281 Figure 7-48. Sodium Removal from Ward County, North Dakota Lignite 282 Figure 7-49. Sodium Removal from Ward County, North Dakota Lignite 283 Figure 7-50. Sodium Removal from Ward County, North Dakota Lignite 284 Figure 7-51. Low Temperature Sodium Removal from Ward County, North Dakota Lignite by Ion Exchange in Agitated Slurries 285 LIST OF FIGURES AND TABLES (continued) Page Figure 7-52. The Effect of Percent Solids in Agitated Slurries on the Sodium Removal by Ion Ex- change from Ward County, North Dakota Lignite . . .286 Figure 7-53. Experimental Arrangement for Sodium Removal from Lignite Using Column Ion Exchange 289 Figure 7-54. Sodium Removal from Lignite as a Function of Time and Flow Rate at 316 PPM Ca ++ Average Wash Solution Concentration 290 Figure 7-55. Sodium Removal from Lignite as a Function of Time and Flow Rate 291 Figure 7-56. Sodium Removal from Lignite as a Function of Time and Flow Rate 292 Figure 7-57. Sodium Removal from Lignite as a Function of Time and Flow Rate 293 Figure 7-58. Sodium Removal from Lignite as a Function of Time and Calcium Concentration in the Wash Solution 295 Figure 7-59. Sodium Removal from Lignite as a Function of Time and Flow Rate Using a Distilled Water Wash 296 Figure 7-60. Sodium Removal from Lignite as a Function of Time and Flow Rate 297 Figure 7-61. Sodium Removal from Lignite as a Function of Time and Flow Rate at 303 PPM Ca ++ Average Wash Solution Concentration 298 Figure 7-62. Sodium Removal from Lignite as a Function of Time and Flow Rate 299 Figure 7-63. Sodium Removal from Lignite as a Function of Time and Flow Rate 301 Figure 7-64. Sodium Removal from Lignite as a Function of Time and Flow Rate 302 Figure 7-65. Sodium Removal from Lignite as a Function of Time and Column Geometry 303 Figure 7-66. Low Temperature Sodium Removal from Lignite as a Function of Time and Column Geometry 304 Figure 7-67. Sodium Removal from Lignite as a Function of Time and Column Geometry 305 Figure 7-68. Low Temperature Sodium Removal from Lignite as a Function of Time and Column Geometry 306 XI LIST OF FIGURES AND TABLES (continued) Page Figure 7-69. Sodium Removal from Lignite as a Function of Time and Particle Size 307 Figure 7-70. Low Temperature Sodium Removal from Lignite as a Function of Time and Particle Size 308 Figure 7-71. Rate of Calcite Dissolution When Several Methods are Used to Introduce C0 2 into Solution 310 Figure 7-72. Rate of Calcite Dissolution in Unstirred Slurries as a Function of Particle Size 311 Figure 7-73. Rate of Calcite Dissolution in Unstirred Slurries as a Function of Particle Size 312 Figure 7-74. Rate of Calcite Dissolution in Stirred Slurries as a Function of Particle Size 313 Figure 7-75. Rate of Calcite Dissolution in Stirred Slurries as a Function of Particle Size 314 Figure 7-76. Rate of Low Temperature Calcite Dissolution in Stirred Slurries as a Function of Particle Size 315 Figure 7-77. Rate of Low Temperature Calcite Dissolution in Stirred Slurries as a Function of Particle Size 316 Figure 8-1. Schematic Arrangement of Light-Limiting Aperture 351 Figure 8-2. Calibration Curves 352 Figure 8-3. Permeation Tube Calibration Apparatus 354 Figure 8-4. Calibration Data into Parts-per-Billion Range . . .355 Figure 9-1. Stack Adapter Collar Assembly 384 Figure 9-2. Optical System 385 Figure 9-3. Optical Microscope Photograph of Stack Flow Particulate Matter Sample 387 xn LIST OF FIGURES AND TABLES (continued) Page Table 2-1. Eutectic and Liquidus Temperatures in the Binary System, CaC03-Ca(OH)2, in the Absence and Presence of MgO 20 Table 2-2. Additions of Various Impurities with Which Melts Could Not Be Obtained 30 Table 2-3. Composition of Melts Prepared in the Presence of Various Impurities 35 Table 2-4. Chemical Analysis and Types of Melts Obtained from Dolomites of Different Localities 40 Table 2-5. Fines Production from 48 X 100 Mesh Material in a Tumbling Mill 41 Table 2-6. Results of Injection of C0 2 on Solid Ca(0H) 2 . ... 42 Table 2-7. Fines Production from 30 Grams -48+100 Mesh Material in a Tumbling Mill 58 Table 2-8. Eutectic and Liquidus Temperatures in the Binary System, CaC03-Ca(OH) 2 , in the Absence and Presence of MgO 61 Table 2-9. Physical Properties of CaC0 3 -Ca(OH) 2 Melts in the Absence and Presence of MgO 65 Table 2-10. Results of Injection of C0 2 on Solid Ca(OH) 2 . ... 66 Table 4-1. Composition of the Iron-Chromium-Nickel Alloys Involved in this Investigation 93 Table 4-2. Elemental Analysis of Purge Acceptor Obtained from Run 33B 95 Table 4-3. The Calcium Carbonate and Hydroxide Contents of Carbonated and Hydrated Purge Acceptor Obtained from Run 33B 95 Table 4-4. Elemental Analysis of the CaC0 3 and Ca(0H) 2 Prepared from Run 33B Purge Acceptor 96 Table 4-5. The Calcium Carbonate and Hydroxide Contents of Seived Carbonated and Hydrated Purge Acceptor Obtained from Run 33B 97 Table 4-6. Weight Change of Several Alloys for a 100-Hour Contact Time at 1400°F in the Presence of 5 Mole Percent of the Salts Listed 114 Table 4-7. Oxide Thickness as a Function of Contact Time . . .116 Table 4-8. Corrosion Rate of the Alloys Involved in This Study 119 Xlll Table 4-9. Table 4-10. Table 4-11, Table 5-1. Table 5-2. Table 5-3. Table 5-4. Table 5-5. Table 5-6. Table 5-7. Table 5-8. Table 5-9. Table 5-10. Table 5-11. Table 5-12. Table 6-1. Table 7-1. Table 7-2. Table 8-1. LIST OF FIGURES AND TABLES (continued) Page Composition of Alloys Involved in This Investigation 126 Oxide Thickness as a Function of Time 128 Corrosion Rate of Several Alloys in Contact with Eutectic Composition CaC03-Ca(OH) 2 Melts and Cost of 2-Inch Plate 145 Chemical Analyses of the Jefferson Formation and Madison Group, Logan, Montana 156 X-Ray Analyses of Acid-Insoluble Residues 158 Drop Hammer Test Results (11 sheets) 170 Averaged Drop Hammer Results for Rocks on Which Multiple Samples were Run 181 Rock Mechanics Test Results for the Pahasapa Limestone 183 Rock Mechanics Test Results for the Tymochtee Dolomite 184 Rock Mechanics Test Results for the Minnekahta Limestone 185 Chemical Analysis Data (9 sheets) 187 Burlington Northern System Rates 201 Milwaukee Road Rates 202 Chicago and Northwestern Rates 203 Summary of Carbonate Rock Parameters (3 sheets) 206 Wear Factor Results of Los Angeles Abrasion Tests 213 Chemical Analyses of the Ash of Two Samples of Lignite Involved in the Investigation 223 Screen Analysis of the Lignite Used in Column Ion Exchange .287 Summary of Results for Coals Analyzed and Reported in Interim Report, December 1972 (Subsection 8.2) 32i XIV LIST OF FIGURES AND TABLES (continued) Page Table 8-2. Trace Elements in Samples Taken During Operation of Preheater, December 12, 1974 325 Table 8-3. Trace Elements in Samples Taken During Pilot Plant Run 26B .326 Table 8-4. Trace Elements in Samples Taken During Pilot Plant Run 27C (3 sheets) 327 Table 8-5. Trace Elements in Samples Taken During Pilot Plant Run 28B (2 sheets) 330 Table 8-6. Trace Elements in Samples Taken During Pilot Plant Run 33B (4 sheets) 332 Table 8-7. Summary of Results for Coals Analyzed 341 Table 8-8. Data for Determination of Mercury by Flameless AA after Ashing with HN0 3 and KCIO3 . . .342 Table 8-9. Data on Analysis and Recovery of Arsenic After Dry Ashing 342 Table 8-10. Data on Analysis and Recovery of Arsenic After Wet Ashing 343 Table 8-11. Comparison of Arsenic Determinations by Dry Ashing and by Wet Ashing Methods 343 Table 8-12. Data on Recovery of Selenium Added to Sample After Digestion 344 Table 8-13. Data on Recovery of Selenium Added to Sample Before Digestion 344 Table 8-14. Data on Recovery of Lead and Cadmium by Extraction into MIBK and Back-Extraction into 10 Percent Nitric Acid 345 Table 8-15. Data on Analysis and Recovery of Lead 345 Table 8-16. Data on Analysis and Recovery of Cadmium 346 Table 8-17. Calibration Data for COS 350 Table 9-1. Waste Water Analyses 360 Table 9-2. Trace Element Content of Outlet Water Compared with EPA Standards 359 Table 9-3. Composition of Suspended Solids 361 Table 9-4. Biochemical Oxygen Demand Analyses of Pond Waste Water 362 xv LIST OF FIGURES AND TABLES (continued) Page Table 9-5. Sulfite in Lignite Gasification Plant Samples . . .366 Table 9-6. Sulfide in Lignite Gasification Plant Samples . . .367 Table 9-7. Ammonia in Lignite Gasification Plant Samples . . .367 Table 9-8. Loss of Ammonia, Sulfide, and Sulfite with Time for Selected Plant Process Stream Samples Obtained During Run 39 368 Table 9-9. Loss of Ammonia, Sulfide, and Sulfite with Time for Selected Plant Process Stream Samples Obtained During Run 40B 369 Table 9-10. Comparison of Sulfide Single Electrode Measurements with Standard Method Values for Selected Samples from Plant Run 33B 370 Table 9-11. Comparison of Ammonia Single Electrode Measurements with Standard Method Values for Selected Samples from Plant Run 33B 370 Table 9-12. Particle Size Analysis Results for Selected Samples from Runs 27C and 28B 370 Table 9-13. Analyses Results of Samples Submitted by Radian Corporation. , 371 Table 9-14. C0 2 and CO Content of Stack Gas, July 12, 1974. . .373 Table 9-15. Waste Water Analyses 378 Table 9-16. Trace Element Content of Outlet Water Compared with EPA Standards 379 Table 9-17. Composition of Suspended Solids 379 Table 9-18. Calibration of Columns for Various Compositional Ranges 381 Table 9-19. C0 2 and CO Content of Stack Gas, July 12, 1974. . .381 XVI ABSTRACT The South Dakota School of Mines and Technology performed support studies related to the activities of the C0 2 Acceptor Process Gasi- fication Pilot Plant in Rapid City, South Dakota, under subcontract to Conoco Coal Development Company during the period 19 71 to 19 77. These studies investigated certain long-range problems associated with the possible design of a full-scale commercial gasification plant using the CO2 Acceptor Process. The areas of study include: acceptor reconstitution, studies in the CaC03-CaS04-CaS and CaC0 3 -Ca(OH)2-CaO systems, corrosion of alloys in contact with CaC0 3 -Ca(OH)2 melts, carbonate rock resources in the High Plains region, grinding properties of carbonate acceptor, lignite beneficiation by sodium removal through ion exchange, trace elements in lignite, and monitoring of waste from the Rapid City pilot plant. During the course of the subcontract, 16 interim and final reports were prepared covering various aspects of the investi- gations. These reports are included in their entirity in this volume in addition to an executive summary which describes the major conclusions of the studies and their significance. SECTION 1 SUMMARY During the period 1971-1977, operation of the 40 T/D C0 2 Acceptor Process Gasification Pilot Plant was carried out at Rapid City, South Dakota, under the joint sponsorship of the American Gas Association (now the Gas Research Institute) and the U. S. Government (successively through the Office of Coal Research, the Energy Research and Development Administration, and the Department of Energy) . The prime contractor in this development program was Conoco Coal Development Company (formerly the Research Department of Consolidation Coal Company) . A number of studies in support of this program were carried out under sub- contract by the South Dakota School of Mines and Technology. The purpose of the present volume (Volume 9, Books 1 and 2) is to consolidate, in a single reference, the several reports prepared by the SDSMT investigators. 1 . 1 INTRODUCTION In 1970 the South Dakota School of Mines and Technology entered into a subcontract with Conoco Coal Development Company (CCDC) to investigate a number of problems associated with pilot plant activities and to provide information on raw material problems which might arise in connection with commercial gasification operations. The initial study areas and principal investigators included the following: (1) Acceptor Reconstitution: Dr. M. C. Fuerstenau (2) Lignite Benef iciation by Sodium Removal through Ion Exchange: Dr. J. F. Clarkson (3) Trace Elements in Lignite: Dr. A. L. Lingard (4) Carbonate Rock Resources in the High Plains: Dr. J. C. Mickelson In July, 1971, a Waste Monitoring subproject under Dr. T. K. Oliver was added to the original contract. The basic reasons for choosing these areas of study were several unknowns. For example, the cost analysis of the original gasification project was based on the availability of large tonnages of natural carbonate rock at reasonable prices which could be used for the acceptor. The most successful previous testing had been done on the Tymochtee dolomite which was available from a quarry in Huntsville, Ohio, but obviously an equivalent rock much nearer to the actual site of a gasification plant was required. Hence it was decided to investigate all possible sources of natural carbonate rock which could possibly be competitive if a gasification site were chosen in the main lignite fields of western North Dakota. Inasmuch as railroad routes were limited in number and direct routes were not always feasible, it was also decided to investigate fully the possibilities of reconstituting the acceptor if costs of the natural acceptor were excessive. Researchers had been able to reconstitute CaC03 acceptor on a bench scale, but considerable additional data was needed on the possible reconstitution of dolomite and also on the prop- erties of reconstituted acceptor. At this preliminary stage it was also believed necessary to see if it were feasible to perform any pregasification beneficiation of the high- sodium lignites characteristic of many North Dakota lignite deposits. Excess sodium proves to be a problem in many conventional boiler-fired plants, and it was anticipated that it would also prove to be a problem in gasification utilizing the C02 Acceptor Process because of the use of a fluidized bed and the temperatures required. On a long-range scale it was believed necessary to obtain data on the trace element content of the lignites to be used in the pilot testing as well as on other potential lignite sources and, at the same time, to use and develop reliable analytic techniques for these elements so that data could be obtained for the necessary materials balance studies. The waste monitoring project was developed as a corollary need to have all available data pertinent to the pilot plant operation. As this work developed and various reports were completed or data obtained, the results dictated in many instances a shift in emphasis or in some instances a completely new tack. For example, with the comple- tion of an interim report on the availability and transportation costs of the natural carbonate rock acceptors, it seemed very likely that reconstitution of the acceptor was the probable route that a full-scale plant would have to follow. With this in view, an intensive study of the effect of impurities on the reconstitution properties of artificially produced acceptors was begun in October 1972. At the same time, an extensive study of the corrosion of alloys in contact with CaCO,-Ca(0H) melts at high temperatures and steam pressures was undertaken. The latter study was quite necessary because no data was available on the behavior of various alloys in the rigorous environment required for any large-scale commercial acceptor reconstitution. This study continued until the cessation of the subcontract. After completion of the effect of impurities on acceptor reconstitution it was believed that the basic data on the CaCO,-CaSO.-CaS ternary system should be completed; and in August 1975, work was begun on that system and later the related CaCO -Ca(OH) 2 -CaO system. During the entire duration of the contract, some 20,000 hours of runs were made in three autoclaves at temperatures ranging from 1100°F to 2300°F and pressures ranging from 100 to 1000 PSI in order to complete data related to accept- or studies and alloy corrosion. Based on data obtained on the physical properties of both natural and artificial acceptor, a brief investigation of the grinding properties was also initiated in October 1972. At the same time an analytic study on detection of different sulfur species at low levels of con- centration was begun. Although this work was discontinued in favor of concentrating such effort elsewhere, the successful calibration of equipment and techniques permitted use of the equipment at the pilot plant in order to monitor sulfur species from several flow streams. During this same period, analytic techniques concentrated on determining minor elements in pilot plant raw materials, and in late 1974 a con- centrated effort was initiated on the analysis of all different effluent streams for solid, liquid, and gaseous species required in final materials balance calculations. This effort was coordinated in part with a duplicate sampling and analysis program under the direction of Radian Corporation. Earlier attempts at particulate characterization had been discontinued due to the atypical particulate discharge of the pilot plant as compared to the probable full-scale commercial plant. During the extent of the subcontract, 16 interim and final reports were prepared covering various aspects of the subproject investigations. These are included in their entirety in this report, but the signifi- cance and major conclusions of the various subprojects are summarized in the following sections of this executive summary. 1.2 ACCEPTOR RECONSTITUTION STUDIES 1.2.1 BACKGROUND Initial bench-scale tests established by Conoco Coal Development Company established that the Tymochtee dolomite was a particular natural carbonate rock whose physical properties were especially favorable as an acceptor in the C0 2 Acceptor Gasification Process. Although the acceptor characteristics of this stone were favorable, it was not known exactly why and whether or not the best acceptor would have to be of dolomite, CaMg(C03) 2 , composition. It was known, however, through preliminary work U) on the CaC0 3 -Ca(OH)2 system, that limestone could readily be reconstituted. It was also unknown whether or not a final plant design would be depend- ent on natural rock acceptor or reconstituted acceptor, but it was known that large tonnages of makeup acceptor would be required. Further- more, a large percentage of carbonate fines would be generated during grinding and if it were possible to use these fines through reconsti- tution, a considerable cost saving might result. Consequently the decision was made to obtain additional data on possible acceptor recon- stitution and the work was undertaken. This work ultimately extended into modifications of acceptor via impurities and melt studies of ternary systems associated with acceptor re constitution and pilot plant operation. 1.2.2 RESULTS The work demonstrated conclusively that reconstituting dolomite was impractical at temperatures and pressures realistic with a commercial operation. Only about 22 mole percent MgO is permissible in a recon- stitution process limited by temperatures of about 1400° F and gas pressures to 600 PSI . Reconstituted acceptor prepared from eutectic mixtures of reagent materials was also shown to have better abrasion resistance than natural carbonate rocks that were tested. The work also demonstrated that the thermodynamically favorable carbonation reaction of Ca(OH) 2 +C0 2 produced no additional exothermic heat, pre- sumably because of the formation of a thin CaC03 film which resulted in unfavorable kinetics. As to the heat required for eutectic melt forma- tion, calculations indicate approximately 500 BTU would be required per pound of eutectic mixture CaC03-Ca(OH) 2 . Possibly one of the more significant aspects of the acceptor reconsti- tution studies involved the effect of impurity additions on acceptor reactivity. In the C0 2 gasification process it was well known that the continued recycling leads to a continuous reduction in the acceptor activity or the ability of the acceptor to take on C0 2 . This necessi- tates a continuing addition of fresh acceptor and any process whereby the decrease in activity could be reduced would be very desirable. Because it had been shown that growth of CaO crystallities accompanied activity loss, C 1 ) it was assumed that addition of the right impurity could possibly increase activity through decrease in crystallite formation. Impurities added were common carbonates (Ba, k, Mn, Na,Sr) , oxides (AI2O3, Fe203, Si02) , and various calcium compounds. The addition of alumina, silica, ferric oxide, manganese carbonate, calcium phosphate, basic calcium phosphate, and phosphorous pentoxide all had positive effects on acceptor activity when a calcination temperature of 1730° F and a carbonation temperature of 1375°F were used. For example, a eutectic- composition melt containing 7.5 mole percent Ca3(PC>4)2 had an activity of 0.36 as compared to 0.14 for a eutectic-composition melt containing no impurity after cycling between calcination and carbonation for twenty cycles. Similar increases in activity were found with addi- tions of P 2 5 and Cai (OH) 2 (P0 4 ) 6 ; and when Fe2C>3 was added at 6.12 mole percent, the activity of the eutectic-composition melt after six cycles of calcination- carbonation at these same temperatures was approximately double that of the iron- free melt. Importantly, the abrasion resistant qualities of eutectic-composition melts are not reduced significantly when impurity additions are made. Further, essentially all of the melts are harder than natural carbonate stone. The major drawback in the addition of impurities is that the beneficial effects on acceptor life decrease as the calcination and carbonation temperatures are raised. For example, at temperatures of calcination and carbonation of 1835°F and 1545° F respectively, there are no improvements in acceptor life. The kinetics of calcination and carbona- tion are, of course, greater at the higher temperatures. Any decision to modify these temperatures at a commercial plant as a trade-off on extended acceptor life through impurity addition would involve many other parameters beyond the scope of this study. 1.3 STUDIES IN THE CaC0 3 -CaS0 4 -CaS and CaC0 3 -Ca(OH) 2 -CaO SYSTEMS 1.3.1 BACKGROUND Although the CO2 Gasification Process relies largely on the retention of sulfur as CaS , it is known that some formation of CaS04 also occurs and that a low-temperature eutectic exists in the CaC03-CaSC>4 binary. If fusion occurs during the gasification process, the resulting agglo- meration can only lead to collapse of the fluidized bed in the regenera- tor and termination of the gasification process. Hence it was deemed desirable to know the full parameters of the CaC03-CaS0 4 -CaS ternary system relative to melt formation. A similar need was evident in the CaC03-Ca(OH)2-CaO system in order to establish optional process condi- tions for acceptor reconstitution. The necessary studies were completed to adequately delineate these systems as an adjunct to the possible design of a commercial gasification plant. 1.3.2 RESULTS The complete experimental results are presented in Section 3 of this report. However, the salient feature of this work was the definition in the CaC0 3 -CaS0 4 binary of the eutectic at 42 mole percent CaS0 4 at 1850°F. This is a significant difference as compared to results of earlier investigations which are believed in error because of inadequate C0 2 pressure to prevent decomposition of CaC0 3 . In the present autoclave studies a 500 PSIA was required in order to prevent CaCC>3 decomposition. If the binary eutectic composition ratio is used and CaS added, the increase in temperature in the ternary system is not large until approxi- mately 20 mole percent of CaS is added. Thus a relatively large area of potential melt formation exists in the ternary system at a tempera- ture of 2000° F. At that temperature, the composition range on the CaCC>3-CaS04 edge of the ternary ranges from 32 to 63 mole percent CaS04. The CaS-CaS0 4 binary eutectic could not be determined but is probably in the range of 55 to 65 mole percent CaSC>4 at temperatures greater than 2100° F. With the available equipment, melts also could not be obtained on the CaS-CaC0 3 binary. Data on the CaC0 3-Ca(OH)2-CaO ternary system at steam pressures ranging from 200 to 1000 PSI and temperatures up to 1480°F delineate the full compositional ranges which can lead to the formation of melts. The lowest possible fusion temperature in the system is 1180°F. It was also found that the CO2 is retained in these melts, and it is possible to determine the liquidus temperature of any melt feedstock with the data obtained. Such data should be of considerable use in the design of any acceptor reconstitution process. In summary, the basic data developed corroborates the necessary sulfur controls required to prevent melt formation, i.e., either remove nearly all of the sulfur or maintain adequate reducing conditions to prevent oxidation of the sulfide ion to sulfate. Relative to the CaC0 3 -Ca(0II)2- CaO system, the data permits the delineation of conditions necessary for the formation of melts required during acceptor reconstitution 1.4 CORROSION OF ALLOYS IN CONTACT WITH CaCO^Ca (OH) MELTS 1.4.1 BACKGROUND During the various work on subprojects related to the C0 ? gasification process pilot plant, it became apparent that because of natural stone acceptor costs there would be a strong possibility that the most economi- cally feasible design for a commercial gasification plant would entail reconstitution of the acceptor. Inasmuch as this would require prepa- ration of large quantities of molten CaCO -Ca(OH) at temperatures of approximately 1200°F and steam pressures on the order of 600 PSI, it was necessary to investigate the corrosion rates of various allovs under such conditions in order to design such a process. To our knowledge, no large-scale similar process involving such rigorous conditions is in operation today, and little work had been done on the corrosion rates of available alloys under similar conditions. Inasmuch as some corrosion phenomena are clearly related to the formation of intermediate compounds, galvanic action, etc., the work had to include adequate contact times and diverse conditions to amply quantify the rates under expected condi- tions. The required extensive studies were designed and carried out using a 1-liter autoclave. Alloys which exhibited the most favorable initial results were involved in runs as long as 400 hours to insure that corrosion rates remained constant . The corrosion studies involved the following alloys: AISI Types 304, 304L, 309. 310, 316, 316L, 446; Inconel 600 and 601; Incoloy 800; and Nickel 200. Both reagent -grade material and reconstituted spent acceptor were used in the preparation of the contact melt. Also, an extensive study was made of the effect on corrosion rates of various possible impurities that might occur in the melts, and both stainless steel and alumina crucibles were used to test for galvanic effects. In addition to weight loss of the alloys as a measure of corrosion, studies were made of the microstructure of oxidized materials and the rate of growth of oxide layers . 1.4.2 RESULTS The complete results of the corrosion studies presented in Section 4 of this report show, as might be expected, that the alloys react differently to the corrosive conditions. Some alloys such as Types 316, 316L, 304, 304L; Inconel 600 and 601; and Incoloy 800 had too high rates of corro- sion in the eutectic melt compositions, whereas others such as Type 309 had a low corrosion rate under all conditions except when exposed to certain impurities such as CaSO and KOH. Nickel 200 showed little corrosion at the eutectic composition in a steel crucible as measured by actual weight loss, but the intergranular disintegration precludes its usage. Based on the overall evaluation of corrosion rate at the eutectic composition, in both steel and alumina crucibles, and in the presence of impurities, the use of either Type 310 or 446 for a melt container is recommended. These high-Cr steels seem to have acceptable corrosion rates and fortunately are among the least expensive of the different alloys which were tested. 1.5 CARBONATE ROCK RESOURCES STUDIES 1.5.1 BACKGROUND At the beginning of the pilot plant studies, all previous bench-scale testing had been done using natural rock. The Tymochtee dolomite, obtained from a commercial quarry in Ohio, had proved to be the best C0 2 acceptor, presumably more because of its physical rather than unique chemical properties; and it was planned to use this material in the pilot plant runs. Lacking firm knowledge of why the Tymochtee was the "best" rock, it was decided to investigate all possible sources of natural carbonate which might be available to the lignite fields of western North Dakota and eastern Montana. Inasmuch as the best estimates indicate that a 2-percent makeup of acceptor would be required, the total annual tonnage for a commercial gasification plant would be quite large. Hence the reserves would have to be large, not too far away from a potential plant, easily quarried, and of a rock type presumably compa- rable to the Tymochtee. 1.5.2 RESULTS In the study of possible dolomite-limestone sources, extensive field investigations were made in the following areas: (1) Black Hills of South Dakota (2) Big Horn Mountains in Wyoming and Montana (3) Pryor Mountains, Montana (4) Little Rocky Mountains, Montana (5) Big Snowy Mountains, Montana (6) Three Forks area, Montana Samples were collected of possible quarry sections and the physical and chemical properties of the rocks were thoroughly evaluated by comparison with the Tymochtee characteristics. In addition a study was made of transportation routes and costs to an arbitrary point --in this case Dickinson, North Dakota. Large reserves of carbonate rock were found in all of the different areas but the physical character- istics as indicated by abrasion tests, etc., of the rock in most of the localities is inferior to the Tymochtee. The Big Horn dolomite on the northeast flank of the Big Horn Mountains is the best dolomite in terms of its physical properties and the Minnekahta Limestone in the Black Hills is the best limestone (this rock was subsequently successfully used in a gasification run at the pilot plant). Both rocks are comparable in toughness to the Tymochtee. However, both localities are plagued by impractical transportation routes. Although the Black Hills is only about 220 miles from Dickinson, the rail route is some 650 miles. The Big Horn site route is over 400 miles. Although the 1972 cost estimates on transportation were on the order of 10 dollars per ton, it is apparent that the present total costs for natural acceptor are far above those originally anticipated. Even possible underground mine sites in the lignite fields were investigated, but the mining costs would be prohibitive. 1.6 GRINDING PROPERTIES OF CARBONATE ACCEPTOR As an adjunct to any final plant design, the necessary comminution characteristics of natural carbonate rocks were investigated. This work included studies on abrasion wear and determination of other physical characteristics of samples collected from many of the areas studied under the carbonate resources work. Los Angeles abrasion tests were correlated with a much simpler drop hammer test which was specifically designed to establish the rock toughness. Additional studies included rod-and-ball mill grinding characteristics. The natural rock acceptor typically has a fine grain size and abrasion types of grinding tend to produce excessive fines. It would seem that a combination of impact crushing and rolls material will produce the least loss. Inasmuch as it is very possible that reconstituted acceptor will be used in any commercial plant, it would have been important to have run tests on reconstituted acceptor. Unfortunately no adequate samples were available. However, based on abrasion studies of the reconstituted acceptor, it is indicated that a combination of primary crushing (jaw crusher), secondary crushing (cone crushers), and tertiary crushing (rolls crusher) would generate the least amount of fines for the pure reconstituted acceptor. Virtually all industrial grinding processes in use on somewhat similar material are aimed at producing either a much finer sized product or coarser sized product. It is recommended that if it is possible to obtain enough reconstituted acceptor it should be subjected to the recommended crushing procedure. If this is not feasible, the grinding practices used for the production of coarse barite in the glass industry might be the closest approximation of how reconstituted acceptor would behave. 10 1.7 PREGASIFICATION BENEFICIATION 1.7.1 BACKGROUND Some western lignites are known to have a high sodium content which makes it impossible to use them for typical boiler feed in power plants unless they are diluted with low-sodium lignites. Although the latter technique is presently being used in some mines, an inexpensive treatment for lower- ing the sodium content is desirable. The deleterious effects of the high sodium content is manifested by the formation of extensive boiler deposits. The effect of sodium in promoting low-temperature melting is well known and would probably seriously affect any fluidized bed gasification process. Although work had been done on batch-process ion exchange using largely chloride solutions, it was considered essential to investigate ion exchange mechanisms using Ca ++ solutions which could be regenerated by using materials readily available at a commercial C0 9 gasification plant i.e., Ca and CO Accordingly, a study involving thl exchange of aqueous ta ions for Na m lignites from North Dakota was conducted The investigations covered such parameters as the effects of lignite particle size, P H, flow rate, Ca- concentration, time, and column cSfiguration 1.7.2 RESULTS Sodium removal by column ion exchange with Ca is quite easy and as expected, the efficiency and rate of removal depend largely on the particle size and flow rate of the Ca ion solutions. The smaller !m!% U f It* r ^ adil y ach ieved almost complete exchange in a reason- able time if the flow rate is maintained at 60 milliliters per minute but the time required for sodium removal from coarse sizes of lignite such as 3/8 x 1/4 inch is much greater. Although higher flow rates increased the rate of sodium removal for the finer sizes, in the coarsest samples there was virtually no change in the rate of removal The temperature of the wash solutions is also important in that there' is a decrease m the rate of exchange by a factor of about 1.2 when the + temperature is dropped from 25°C to 6°C. Variations in P H, the- J ^ C °f ent of the solutions, and the column design have little effect on the Na removal although it was found that a column having a 4 5-1 aspect (bed height to diameter) was the most efficient for the smaller lignite sizes. smdUtr tW Ct °V f f m P° rtanCe in any P° ssible commercial usage was the finding that wash solutions could be reused until they reached an equilibrium concentration of about 440 PPM Na* (initial solution 320 PPM fc^ ScO anVcO soluti °^ could be regenerated by contacting with ■j 2 Although there are many cost factors which enter in the possible beneficiation of high-sodium lignite, the work demonstrates that minus 35 mesh lignite containing 15 percent Na,0 in its ash could have the sodium content reduced to about 5 percent oy treatmenTwith r« ion solutions at a flow rate of 60 milliliters per *""£ for a foAT h C °? taCt time ' ThG TQady availabilit/of the material needed PC o^'an 1 ? cave" a C °2 ** Sifi -^ Pl-t might make .J h" 11 1.8 TRACE ELEMENT STUDIES 1.8.1 BACKGROUND Accurate and reproducible analytic data of all materials and flow streams involved in as complicated an industrial process as the C0 2 acceptor gasification process are absolutely necessary. Not only is the data required for a materials balance but some of the contained trace elements could result in an environmental problem considering the large tonnages of raw materials involved. Initially, analytic work and the testing of analytic methods were largely restricted to lignite which was potential feedstock and the elements determined were Cd, Pb, Se, As, and Hg. In later work, the additional elements Be, Cr, Ni, Se, Te, and V were included in the analyses of the raw lignite, individual pilot plant run feedstock, and effluent streams. Beginning in late 1974, analytic work was concentrated on the various effluent streams at the pilot plant during operational runs. Determi- nations of cyanide, isocyanate, nitrate, phosphate, sulfates, and phenols as well as other routine determinations were performed on liquid samples. At the same time, a separate effort performed determination of sulfite, sulfide, and ammonia from the same samples. Analytic techniques for the detection of very low concentrations of the different sulfur species COS, H S, and S0 2 were also investigated using a chromatographic separation of the sulfur compounds and analysis by a flame photometric detector. However, sulfur problems at the pilot plant necessitated installation of a zinc oxide scrubber and, after calibration, the analytic setup was used for the duration of the pilot runs at the pilot plant. In the meantime the decision was made to concentrate work on sulfur species at Consol's Library, Pennsylvania, and Ponca City, Oklahoma, laboratories. In addition to the specialized analytic services on feedstock and effluents, routine analytic data was provided to the pilot plant and the waste monitoring subproject when required. 1.8.2 RESULTS The individual analytic reports and descriptions of analytic techniques are included in full in Section 8 of this report. The analytic data on trace elements and other effluent compounds cannot be correlated into a materials balance because flow data is not available and the responsi- bility for that determination lies with the major contractor. Inasmuch as many analyses were made on a number of lignite samples of various sources, one can make some generalizations about the expected trace element composition of western lignites which might possibly be gasified. Based on the data obtained, it would seem that the following general concentrations (in PPM) would be expected: As -2, 12 Be *1, Cd = or <1, Cr =10, Pb <10, Hg = or <0.10, Ni -10, Se =0.5, Te <0.1, and V 5 to 20. However, it is also possible that certain lignite seams might deviate considerably from these expected values inasmuch as ietailed anlaytic data is not available on all possible lignites. 13 1.9 WASTE MONITORING The waste monitoring work was directed primarily at obtaining continuous data on various effluent streams at the plant. Initially, work plans included continuous monitoring of pH and dissolved oxygen in waste water and also monitoring of particulate material and gas discharge from the stack. To characterize the particulate discharge an optical nephelameter was designed and installed. Particulate matter was far greater than had been estimated and the decision was made in 19 74 to discontinue particulate monitoring because of the very atypical opera- tion of the pilot plant as compared to a commercial gasification plant. In late 1974, personnel began analytic work on samples of effluent streams collected during pilot runs. This work is summarized in Section 8, Trace Element Studies, and Section 9, Waste Monitoring. In early 1975, the work included determination of biochemical oxygen demand (BOD], suspended solids content, and flow rate of waste water which was entering the Rapid City sewer system. These reports were submitted to municipal authorities in order to meet the required regu- lations . Except for copies of the monthly determinations on BOD and suspended solids destined for municipal authorities, all monitoring data was included in the monthly reports throughout the duration of the contract. Data include variations in pH, DO, solids, flow rate of waste water, and size, composition, and other characteristics of particulate matter during the period that the latter studies were in effect. Inasmuch as most of the effort was involved in data collection and measurements, only generalized statements are possible and these are included in the interim and final reports which comprise Section 9. 14 1.10 SPECIAL PROJECTS In addition to specialized analytic work and X-ray studies on items such as low-melting solid deposits in the gasifier, etc., at the pilot plant, other specialized studies were conducted on problems related to the pilot plant operation. For example, a particle impact tester was designed and used to test relative degree of abrasiveness at various impact angles of dolomite and spent acceptor on INCO 800. This data was needed when the plant encountered problems with abrasive corrosion of piping in the solids circuits in the pilot plant. Additional assistance was provided in the calibration of various equipment which was used at the pilot plant. 15 SECTION 2 ACCEPTOR RECONSTITUTION STUDIES This section contains the following studies on acceptor reconstitution performed by the South Dakota School of Mines and Technology: the final report on this project, "A Study of the Reconstition of CaC0 3 -Ca(OH) 2 ,' dated June 30, 19 75, an interim report, "Extension of Acceptor Life by Addition of Impurities," dated July 15, 19 74, and an interim report, "Acceptor Reconstitution," dated July 31, 1972. 16 2.1 A STUDY OF THE RECONSTITUTION OF CaC0 3 -Ca(OH) 2 ACCEPTORS FINAL REPORT, JUNE 30, 1975 2.1.1 INTRODUCTION In the course of gasification of lignite using the C0 2 Acceptor Process, a significant decrease in the activity of the carbonate acceptor occurs! Additionally, a portion of the acceptor is lost due to physical attri- tion, and fresh carbonate acceptor must be added for process makeup. As a result large quantities of natural stone will have to be trans- ported to the plant site, and since only a narrow size range of acceptor can be used, considerable quantities of carbonate fines will be produced upon acceptor preparation. Furthermore, spent acceptor will also be generated during operation of the process. In view of these facts, the possibility of reconstituting spent acceptor and carbonate fines by melting at elevated pressure is a most attactive alternate to using natural material as process makeup. Due to the po- tential economic benefits of acceptor reconstitution, the technical feasibility of this process was examined by Curran and Gorin^ with limestone-base acceptor in a systematic study of the CaC0 3 -Ca(0H) 2 system. This work did not include a detailed study of the dolomite- base system, however. One of the objectives of this investigation was to establish whether dolomite-base acceptors can be reconstituted. Other objectives were to establish whether acceptor life can be ex- tended with the addition of impurities and to measure the mechanical properties of these various acceptors. 2.1.2 EXPERIMENTAL MATERIALS AND PROCEDURES Melts were formed in an autoclave of 1- liter capacity by heating mix- tures of reagent-grade Ca(0H) 2 and CaC0 3 in the presence of excess water to the desired temperature in the absence and presence of various im- purities. The reagent-grade chemicals added as impurities were: A1 2 3 Ca 3 (P0 l+ ) 2 , Ca 10 (OH) 2 (PO 4 ) 6 , CaSO^, CuC0 3 , Fe 2 3 , K 2 C0 3 , MgO, MnC0 3 , Na 2 C0 3 , P 2 5 , and Si0 2 . In some experiments natural limestone was used in place of reagent-grade CaC0 3 , and, additionally, in a portion of the study natural dolomites were also involved. In those cases where liquidus and eutectic temperatures were measured, a calibrated ch rome 1 - al ume 1 thermocouple was immersed into the melt. Vessels for holding samples were made from Type 310 stainless steel.' Water was vented during the heating cycle until the required steam pressure was obtained. 17 Acceptor activity was established as follows: (1) 10 grams of 28 X 48 mesh material were dried at 200°F and were then placed in a fluidized reactor which had been dried and weighed previous ly . (2) The fluidized reactor was raised into a tube furnace and held for 30 minutes at 1825°F (bed temperature) with nitrogen passing through the reactor. (3) At the end of this period, the nitrogen flow was terminated, and the reactor was removed from the furnace. The ends of the reactor were plugged to prevent entry of water, and the furnace was allowed to cool to 1600°F. The weight of the reactor together with the calcined sample was established. (4) The reactor was then raised into the furnace, and C0 2 was passed through the reactor for 30 minutes to carbonate the acceptor. (5) After 30 minutes the C0 2 was shut off, the reactor was removed from the furnace, the ends were plugged, and the apparatus was allowed to cool to room temperature. The reactor and sample were then weighed, and the activity of the acceptor was calcu- lated by the weight gain of the calcined sample. This procedure constituted one cycle of the acceptor. The process was repeated to obtain the activity as a function of the number of cycles. Acceptor activity is expressed as the following molar ratio: CaC0 3 Activity = CaC0 3 + CaO Abrasion resistance of the acceptors was established by dry-grinding 30 grams of 48 X 100 mesh acceptor for 10 minutes in a 6-1/2 X 6-1/2- inch porcelain mill containing 254 grams of porcelain media. The quantity of minus 100 mesh material produced in this test was weighed after grinding. 2.1.3 EXPERIMENTAL RESULTS AND DISCUSSION 2.1.3.1 Temperature Measurements in Ca(C0 3 )-Ca(0H) 2 -MgO Systems The first portion of this investigation involved checking the liquidus and eutectic temperatures of the CaCO 3 -Ca(0H) 2 system and also involved establishing solubility of MgO in these melts. A typical cooling curve is presented in Figure 2-1. In this case a mixture containing 36 mole percent CaC0 3 , 54 mole percent Ca(0H) 2 , and 10 mole percent MgO was in- volved at 350 PSI steam pressure. A eutectic ratio of CaC0 3 to Ca(0H) 2 was involved in this instance, and, as can be seen, the eutectic temper- ature is 1170°F. The value agrees well with the value reported by Curran and Gorin.' J Excellent agreement was also noted in liquidus temperatures. These values together with temperature measurements on other eutectic systems are listed in Table 2-1. Since the eutectic temperature did not change in the presence of MgO, the solubility of MgO in this melt is negligible In view of this fact, magnesium oxide will be present in the melt as occlusions and, as a result, there is a limit as to the amount of MgO that can be contained physically. 18 °°c > I i i I 1 o o 34 _ o O SYSTEM O O 36 mole % CaC0 3 33 O O 54 mole% Ca(0H) 2 — u_ o 10 mole % MgO o 32 — _ — LU O ° 350 psig O LU O 31 O U_ G LU 0£ O *> O ^e 30 O — _j O < H O Z LJ 29 O 1- O G Q. G G 28 G G G G 27 GO% G G 26 ■ 1 ' 1 0_L_ _]_ c 1 10 20 30 40 50 TIME (min) 6C Figure 2-1. COOLING CURVE OBTAINED WITH A MIXTURE CONTAINING CALCIUM CARBONATE, CALCIUM HYDROXIDE, AND MAGNE- SIUM OXIDE AT 350 PSIG 19 MOLAR COMPOSITION (%) CaC0 3 Ca(0H) 2 MgO 30 70 40 60 45 55 32 48 20 36 54 10 EUTECTIC LIQUIDUS TEMPERATURE (°F) TEMPERATURE (°F) 1170 1283 1175 * 1175 1270 1175 * 1170 * *Eutectic Composition Table 2-1. EUTECTTC AND LIQUIDUS TEMPERATURES IN THE BINARY SYSTEM, CaC0 3 -Ca(0H) 2 , IN THE ABSENCE AND PRESENCE OF MgO The next series of experiments were conducted to establish the maxi- mal amount of MgO that the CaC0 3 -Ca(OH) 2 melts can hold. This infor- mation is necessary to establish the feasibility of reconstituting dolomite-base acceptors. The results of these experiments are presented on the triangular diagram in Figure 2-2. As shown, melts are formed in systems containing up to 22 mole percent MgO. When 27, 30, and 35 mole percent MgO was present, mixtures of fused and unfused material were obtained. In the presence of 50 mole percent MgO, however, no fusion whatsoever was noted. This observation is especially signifi- cant, since this is the composition that a dolomite-base acceptor would have. This fact may be seen from the following reasoning. On a 2-mole basis, when dolomite is calcined and carbonated, the resulting mixture contains 1 mole of CaC0 3 and 1 mole of MgO. The MgO is inactive and does not recarbonate. To obtain the eutectic composition of 0.6 mole of CaO and 0.4 mole of CaC0 3 ,0.6 mole of carbonated dolomite would have to be calcined. With this 0.6 mole of CaC0 3 ,0.6 mole of MgO would also be present. When the balance of the two original moles, 0.4 mole CaC0 3 and 0.4 mole MgO, is combined with the calcine, the resulting mixture would contain 0.6 mole CaO, 0.4 mole CaC0 3 , and 1.0 mole MgO. Material of this composition could not be fused (Point A, Figure 2-2). The data indicate that if dolomite is to be used as acceptor, reconstituion of spent acceptor will not be possible. 2.1.3.2 Impurity Additions Upon cycling between calcination and carbonation, both the artificially- produced acceptor and the natural carbonate stone lose their ability to accept C0 2 at a reasonable rate. The mechanism of activity loss has been shown to involve the growth of CaO crystallites which result in particle shrinkage and loss of pore volume. C 1 ) Producing acceptor that is resistant to loss of activity with cycling would be most de- sirable from a processing standpoint. Since the presence of small 20 MgO O NO FUSION FUSED MIXTURE • MELT FORMED Ca(OH) CaCO. Figure 2-2. REGIONS OF FUSION IN THE CALCIUM HYDROXIDE-CALCIUM CARBONATE SYSTEM IN THE ABSENCE AND PRESENCE OF MAGNESIUM OXIDE AT 1400°F AND 600 PSIG 21 amounts of impurities is known to retard crystal growth, an investi- gation of the effect that small amounts of non- lime components have on the life of the acceptor was undertaken. Three types of impurities were added, namely, (1) carbonate salts of various metals, (2) anionic salts of various calcium compounds, and (3) various oxides. After specific additions of these compounds to lime-base melts, acceptor activity was measured as a function of cycling in a fluo-solids bed reactor. The first series of experiments in this portion of the investigation involved establishing the reproducibility of acceptor activity as a function of cycling between calcination at 1730°F and carbonation at 1375°F (bed temperatures) . Eutectic-composition melts containing 7.5 mole percent Ca 3 (POi + ) 2 were prepared at three different times, and activities were established as a function of cycling. These re- sults are presented in Figure 2-3 and show that the reproducibility of results is good. The effect of additions of various carbonate salts of barium, manganese, potassium, sodium and strontium was investigated next. The presence of separate additions of 10 mole percent barium carbonate or manganese carbonate and 5 mole percent strontium carbonate exhibited no signif- icant effect on the activity of eutectic-composition melts. When either potassium or sodium carbonate was added, however, acceptor activity was reduced. (See Figure 2-4.) Oxides of aluminum, iron, phosphorous and silicon were also added as impurities to eutectic-composition melts. Additions of 1.0 and 7.5 mole percent A1 2 3 were made. The addition of the lower value had no effect on acceptor activity, whereas the addition of 7.5 mole percent increased activity by about 25 percent (Figure 2-5). Two additions of Fe 2 3 were also made to eutectic-composition melts. As shown in Figure 2-6, the activity of the melt containing 6.12 mole percent Fe 2 3 was double that with no impurity addition after six cycles of calcination-carbonation. When Si0 2 was added as impurity with similar conditions of calcination- carbonation, no significant increase in acceptor activity was observed. Of the oxides added, the addition of phosphorous pentoxide exhibited the most beneficial effect on acceptor activity. After six cycles of calcination-carbonation, activity was more than doubled when 10 mole percent P 2 5 was added. (See Figure 2-7.) Phosphate salts of calcium were also added to these melts. As shown in Figures 2-8 and 2-9, acceptor activity was increased substantially with the additions of both calcium phosphate and basic calcium phos- phate . Three additional calcium salts were examined, namely calcium molybdate, 22 L CO 6 CD to LU _l - to OJ ro 1— 1- 1- -J _l _J LU LU LU 2 S 2 O • <] CO ^t. CM o 6 A1IAI13V U01d300V E- woz U »-i w os h h a, uj o S K co 2 =5 O U, l-H E- < < CO l-H U < uj u CJ3 S 2 l-H l-H E-H 2 l-H H < 2 E- UJ 2 OS O UJ U U, CO l-H E- Q UJ tu 2 UJ OS 2 I O H 2 O l-H E-> < 2 O CQ OS < E- < Q UJ os < UJ U OS i-H a, u uj CO o a, o u I UJ H UJ U-. o >H E- > 1-1 H U U < u 2 O l-H < 2 l-H U J < u 2 UJ uj a. 3= o t- o UJ bO CQ 1^ l-H U 2 - H-l UJ J os u 3 >- E- u < a, uj O Cu t-O I CM 2 o I— I E-> < 2 CO U < 2 w UJ H UJ ex < > u, o o CQ fj 2 < H 2 O J UJ U U Di >- D co U H < O UJ a. . 2 O UJ HH H E- U 2 2 O 3 hh Uh E- < < 2 E- UJ 2 O CO o S O U CO u < J I u H 2 U UJ UJ u < u E- UJ tu UJ O J o >- H E- U < < 2 O oa < Q 2 < A1IAI10V dOlcGOOV I u 3 •rH 24 ^ ro _a> a> o O a> o e E lO UJ _l o >- o ro CVJ CVJ 6 a Q < CO -■> c- 2 u w 2 W 2 O HH E- < 2 t— i CJ -J < U l-H co u E- >- H co o o U to 2 r- 2 O w u OS O < A1IAI10V d01d300V E- HH i— i H co u O 2 a, 3 o u < U CO z< u w W Q E- t-n 3 X W O U, 2 O 3 2 > J >-> < < o LO I CM < u u 2 2 HH W 2 W M S: H, H J W J U OS W >- D Ld 2 a, aS -I o o o w >- o E- 2 S t-H O CU W H h O H (X U 2 2 2 O O 3 M U tu E-* i < U - M E- H OS < M 2 2 > W O h-t UL, CQ E- OS U fc < < o u \0 rsj tU h 3 W> A1IAI10V d01d300V CL, 26 —i N ^ o^ $5 O o E E O O m O □ #

ro PJ o w HH 3= Uh C* E- O < w o > CO IT) CJ u r-i 2 2 i— i i— i * 2 -J w h- 1 u cr: < >■ X E- u H 2 O PL,' £ u o co 2 s E- o w J h- 1 H w E- s U 2 2 o 2 3 1— I o EX, H 1— 1 < E- < 2 h- 1 h- 1 CO CO U o < J a, < S w u o Q u i— i 1 X • u o 2 1— 1 E~ O E- 2 KH U eg H W ex, < H 2 •=> CO O w D CQ tu o % o o 33 CJ > Dh Q E- CO 2 t-H o < > X 1— 1 Cu 2 E- O U CL, h- 1 < o E-i r^ - O E- w E- < Z u ►J < < w > w z < H z o u CO E- .J W z o I— I E- i— i co o a, E- u- wo oa o to Z iH J - U PJ >- 2 u u, o z o I— I E-H u z 3 H co o u I U PJ H E- U w H w o E- W Ph w E- z o l-H E- < z I— I u < u D O M CQ U OS J < U E- U u PL, o Q Z < CO I 00 A1IAI10V d01d300V 28 Ixl _l O >- (J 1 CO < 2 2 C i— i t— u E- i— : HH < Q l; Q < 2 CO w 3 p O H M W tu, os 03 O < O > u K) 2 r-» u t— 1 i— i 2 J h- 1 u *\ 2 >- cu i— i u OS < 3 H Lt, E- 2 O o 2 u 2 w O a, CO i— i S E- H cq >-J U E- w 2 S D 2 Uh O 2 i— i O < E- i— i < E- CO 2 h- 1 < i— i CO U o IX] J IX H < o s u u 1 a, CO , u o 2 h- 1 X O E- Cu i— i U E- w s < E- 3 2 3 1— 1 O m u CQ -J OS U-, < < o u U >H u Q H n 2 i— i CO < > < HH 33 2 E- O u Jh — i < 3 '" CTi r^j a; U 3 00 A1IAI13V d01d330V p., calcium silicate and calcium sulfate. Each of these salts was found to be detrimental to acceptor activity. (See Figures 2-10, 2-11, and 2-12.) Interestingly, the higher the addition of both calcium silicate and sulfate, the less is the effect of the impurity. 2.1.3.3 Limits of Impurity Additions Upper limits of addition of the impurities which improved acceptor activity of eutectic-composition melts were established. These values are listed in Table 2-2. MATERIAL RESULT Eutectic No Melt (10.0 mole % A1 2 3 ) Eutectic No Melt (10.0 mole % Fe 2 3 ) Eutectic No Melt (10.0 mole % Si0 2 ) Eutectic No Melt (12.5 mole % P 2 5 ) Eutectic No Melt (10.0 mole % Ca 10 (OH) 2 (PO 1+ ) 6 Table 2-2. ADDITIONS OF VARIOUS IMPURITIES WITH WHICH MELTS COULD NOT BE OBTAINED 2.1.3.4 Activity and Extended Cycling Periods Six cycles of calcination-carbonation were used so that a number of impurity additions could be investigated. Experiments were also con- ducted to establish activity of eutectic-composition melts in the ab- sence and presence of Ca3(P0i+) 2 after extended cycling. As shown in Figure 2-13, activity of both melts remained essentially constant after about 13 cycles. The activity of the melt containing 7.5 mole percent calcium phosphate was approximately 2-1/2 times that which contained no impurity. 2.1.3.5 X-Ray Analysis of Melts The composition of melts prepared with the eutectic ratio of calcium carbonate to calcium hydroxide in the presence of various impurities was established with X-ray diffraction. The results of this series of experiments are presented in Table 2-3. 30 ->8 >P o o 0) 0> o o - o oo 2 o E- m 2 00 O l-H OS < u w w 3: E-> U-, 2UO i-h CQ O 2 to HH O 1^ < 2 rH H 2 O u CO E- -J ►J •> u tu >- OS u [I, o 2 o u 2 PL, 2 O i— i E- i— i co O a, O CO u < i U PJ l-H E- E- < U Q W CQ 0J a. cu E- 2 O HH E- < 3 O HH E-h < 2 O CQ U OS ►J < < U u Q U P- 2 < O < O >- E- i— ( > t-H E- A1IAI10V U01d300V i CD u •H 51 - iD - lO - *- UJ _l o >- o - ro - CVJ Q co 2 2 < O M 2 H O Q E- Q < < 2 i— i CO U 3 J B5 > <: u 2 PJ w S£ H UJ ca P- uo 2 O < H 2 O U U co u -J UJ 2 O P. o 2 O l-H H U 2 P. LO O a, S o u I u I— I E- u w u 3 -J UJ i-h CO o s UJ UJ 2 o I— I E- < 2 i— i U < 2 o E- A1IAI1DV U01d330V H U < J 2 < O u oa H oi u u, < < o u I CM 0) • H 32 v8 s 55 a> a> o o - o ro - OJ C\J 6 A1IAI13V d01d300V 2 o a c 00 O 2 O 1— < < 2 CJ < < > z w U UJ Ei, 2 S O H h O 2 uj to HCQN < E- 2 O u u CO >- E- OS E- 2 2 W O O H E- E- 00 2 O 3 a, tu O CJ I u I— I H u OJ E- PJ tu O 2 O 1— 1 H < 2 < hH u < E-h < < CJ O 2 O 1 — 1 E- < 2 O CQ OS < cj a 2 < CM 1— I I CM + >- *t H o tr c Q_ Z> ro Ql o o 2 z O o UJ _J >■ o A1IAI10V U01d300V w u 2 w CO CJ W 2 OS i-h a. o o to OS u H >- u w _ CO t— I < u 2 H 2 co E- W CO < w UJ a. cu E- 2 O i— i 2 i— i U J < co o c ■=> u < u E- 2 U W w u E- 04 3 PJ PJ 0- P- PJ O J a >- E- 2 O i— i < 2 O aa as u Q 2 < 2 O i— i E- < 2 h-l U < > • e i— i r-- S E- H U U- PJ U- z o w U 2 OS O (JJ hH 0- H c_> w z J 3 O tu s < LO • CO r- < U CO z w 1-4 OS Z 3 HH E- < < • E- c2 z z w o OftH u 2 H w < CO E- Z H O -j z oa PJ O OS S w < Ill E- U _J Z < O Z Q o l-H HH Z E- cj < CO < z o u o Oh hh S O H o 3= < U E~ Z 1 t-H U E- U Kh < J E- < uwu W E- E- < Z 3IW W Oh U CO S d-OH OIU Oh CQ >H E- 2 O t-i 3 Z > l-H M M U hJ E-i J U U < >- < u u A1IAI10V d01d300V 0) •H 36 z u. o H- o < ro 2 f^ O "■" _J < o o IxJ _J o 2 H i-h W U y >* OS u w a, u, o tn -J 2 9 ° s i— i E- (N U rH 2 • 3 vO tU g< l-H CO 2 < i— i < CO E-< W 2 OS O 3 U H CO E- W 2 O E- M CO o Oh o u I u i-i o E- S U E- W E- 2 O i— i E- < 2 O OQ OS < u a 2 < 2 O rH < 2 O E- A1IAI10V dOld3D0V a j o 2 l-H CJ PJ > l-H UJ l-H OS ^ E-i OS E- u w w < P-, cq LO rH I - o ro C\J C\l 6 w a Si u H 2 UJ -2 Uh UhO w o < u UJ w < UJ OS s o OS Uh Q OJ u Q O OS ft, u p-J uj U OS E- CO u H 2 w O UJ Oh s w H O i— i < 2 CO U < -J < OS u o E-> Oh • UJ 2 _ u o U U M I < E- u i-h E- E- 2 _ U UJ CQ W Oh OS co o 0, o < 2 O H CO PJ Uh O >h H a 2 < •J < u < 2 O > HH E- U < OJ 2 U U A1IAI10V d01d300V 1-4 I OJ •H Uh 39 CaC0 3 WT % MgC0 3 WT % RESULTS 59.2 39.1 Partially fused melt 49.5 33.3 Melt 60.2 33.1 Fused melt 56.2 26.0 No melt 62.2 33.4 Partially fused melt 58.0 38.0 Fused melt 47.5 36.0 No melt SAMPLE Bighorn (WY) Pahasapa (Pringle, SD) Pahasapa (Whitewood, SD) Pahasapa (Rimrock, SD) Pahasapa (Spearfish, SD) JCTMH (MT) Guernsey (WY) Table 2-4. CHEMICAL ANALYSIS AND TYPES OF MELTS OBTAINED FROM DOLO- MITES OF DIFFERENT LOCALITIES 1.3.9 Physical Properties of Acceptors In addition to the chemical properties of acceptors, the abrasion characteristics of these materials are also important. During cal- cination and carbonation as well as transport between the two pro- cesses, acceptor particles will necessarily rub against each other and against the walls of the transfer lines and reactors. Accordingly, abrasion experiments were conducted with acceptors in the absence and presence of various impurities and compared with that of natural stone. The results of these experiments are listed in Table 2-5. Some of these values are different from those reported in Subsection 2.2, "Extension of Acceptor Life by Addition of Impurities." These later figures are average values of additional experiments. As can be noted, artificially prepared acceptors are hard materials which in general are harder than natural stone. 40 MATERIAL Natural Stone (Tymochtee) Natural Stone (Pahasapa) Eutectic (No Impurity) Eutectic (from Spent Acceptor) Eutectic (1.0 mole % P 2 5 ) (3.0 mole % P 2 5 ) (5.0 mole % P 2 5 ) (10.0 mole % P 2 5 ) Eutectic (5.0 mole (7.5 mole Eutectic (3. 33 mole (6 .67 mole Eutectic (6.12 mole Eutectic (5.0 mole \ (7.5 mole \ Eutectic (5.0 mole % (6. 25 mole i Ca 3 (PO tt ) 2 ) i Ca 3 (P0O 2 ) % Ca 10 (OH) 2 (P0 1+ ) 6 ) % Ca 10 (OPl) 2 (PO 4 ) 5 ) % Fe 2 3 ) A1 2 3 ) A1 2 3 ) Si0 2 ) % Si0 2 ) Eutectic (5.0 mole Eutectic (5.0 mole Eutectic (5.0 mole (10.0 mol Eutectic (3.0 mole (5.0 mole Eutectic (3.0 mole Table 2-5, % K 2 C0 3 ) % Na 2 C0 3 ) % MnC0 3 ) e % MnC0 3 ) % CaMoO^) % CaMoO^) % CaSi0 3 ) INITIAL WEIGHT (GM) 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 IVT -100 MESH PRODUCED (GM) 14.9 16.3 10.0 7.8 7.6 5.5 7.1 8.4 12.7 15.9 5.0 12.9 7.8 8 8 .5 .9 7 9 .5 .1 12, ,2 15. 8 8. 9. 5 5. 6. 2 8 12.5 FINES PRODUCTION FROM 48 X 100 MESH MATERIAL IN A TUMBLING MILL. 41 800 420 705 600 660 600 2.1.3.10 C0 2 Injection as Source of Heat A final series of experiments were conducted to evaluate the feasi- bility of heating solid Ca(01I) 2 by C0 2 injection, since this carbona- tion reaction is highly exothermic. Thermodynamic calculations show that the injection of 1 mole of C0 2 at room temperature into 2 moles of Ca(0H) 2 at 800°F would produce an unmelted eutectic mixture (40 percent CaC0 3 -- 60 percent Ca(0H) 2 ) at 1070°F. This is close to the melting point of the eutectic mixture, 1175°F. Even more thermal energy would be released if the C0 2 were injected into a mixture of 40 percent CaO and 60 percent Ca(0H) 2 . In this case the resulting eutectic mixture would be approximately 50 percent molten, assuming a stoichiometric quantity of C0 2 were injected. Experiments were conducted by injecting C0 2 into Ca(0H) 2 at this temp- erature. Three separate additions of C0 2 were made, and the results are presented in Table 2-6. INJECTION CONDITIONS PRIOR TO INJECTION CONDITIONS AFTER INJECTION ' CYCLES TEMPERATURE PRESSURE TEMPERATURE PRESSURE (°F) (PSIG) (°F) (PSIG) 1 865 200 2 740 300 3 700 300 TABLE 2-6. RESULTS OF INJECTION OF C0 2 ON SOLID Ca(0H) 2 The heat that was expected from the carbonation reaction was not gener- ated and the injection of C0 2 into the reaction chamber actually cooled the system. Samples from the system showed the apparent cause for the absence of generated heat to be the presence of a layer of calcium car- bonate on the particles. The incompleteness of carbonation is attri- buted to slow diffusion of C0 2 into the Ca(0H) 2 particles. Although attractive from a thermodynamic point of view, kinetic limitations apparently are such that use of this technique will not be possible. 2.1.3.11 Preliminary Analysis of Cost of Acceptor Reconstitution Attempts to measure the heat of fusion of the acceptor by making a heat balance on the autoclave were not made. The uncertainty from this type of experiment would be excessive due to the lack of control of heat loss from the autoclave. As a result, an estimate of the heat required to melt eutectic-composition acceptor (40 mole percent CaC0 3 and 60 mole percent Ca(0H) 2 has been made on the following basis. The heat^ required to raise the temperature of CaC0 3 from room temperature to 1175 F is 42 15,832 CAL/mole, while the heat required to raise the temperature of Ca(OH) 2 to the same value is 15,524 CAL/mole. (2) The heat of fusion of CaC0 3 is 12,700 CAL/mole, (3) and an estimated heat of fusion of Ca(0H) 2 is 5,000 CAL/mole. The heat of fusion of Ca(0H) 2 has appar- ently never been measured. The value of 5 KCAL has been assumed, since the only value for a divalent metal hydroxide that could be found is that of Ba(0H) 2 , which is 4,590 CAL/mole. (3) The heat of fusion of BaO is 13.800 CAL/mole, while the heat of fusion of CaO is 12 000 CAL/mole. The heats of fusion of univalent metal hydroxides are generally about 2,000 CAL/mole. On this basis the heat required to melt one GM-mole of eutectic mixture would be 23,730 CAL or 505 BTU/LB In view of the uncertainty of the heat of fusion of Ca(0H) 2 the value' to two significant figures is 500 BTU/LB. This type of estimate is made on the assumption that ideal mixing of these components takes place and neglects the effects of solid phase transformations and temperature on the heat of fusion. However these effects are usually secondary with respect to the heat of fusion and so it is felt that this value is a reasonable estimate of the heat re- quired to melt eutectic-composition acceptor in the reconstitution process . 2.1.4 SUMMARY AND CONCLUSIONS The following conclusions can be drawn from this investigation: (1) The eutectic temperature of the CaC0 3 -Ca(OH) 2 system is 1175°F which agrees well with the value reported by Curran and Gorin (2) MgO exhibits no solubility in CaC0 3 -Ca(0H) 2 melts. (3) Acceptor melts are obtained when the MgO content is less than about 22 mole percent. (4) Dolomite-base acceptor which contains 50 mole percent MgO cannot be reconstituted under the conditions used. (5) Dolomite-based acceptor cannot be fused. (6) Acceptor activity decreases upon cycling between calcination and carbonation. C7) Additions of A1 2 3 , Fe 2 3 , P 2 5 , Ca 3 (POi + ) 2 , and Ca 10 (OH) 2 (P 0tt ) 6 to eutectic-composition melts increase the activity of acceptor upon cycling. These reagents result in the formation of new phases in the acceptor. (8) Additions of CaMoO,, CaSi0 3 , CaS 0l+ , K 2 C0 3 , and Na 2 C0 3 are detri- mental to acceptor activity upon cycling (9) An increase in the calcination temperature decreases the acceptor activity upon cycling. (10) An increase in the calcination temperature offsets the increase fill Svn?h^ ?t ° r aCtlVitX im P arted with the addition of impurities. (11) Synthetic acceptors are hard materials. In general, the arti- ficially produced acceptors are harder than natural stone. 43 (12) Acceptor, reconstituted from spent acceptor, loses its ability to accept C0 2 less rapidly with cycling than does synthetic acceptor produced from chemicals. (13) Kinetic limitations appear to rule out the possibility of using C0 2 injection on Ca(0H) 2 as a source of heat. (14) The heat required to melt eutectic mixture of CaC0 3 -Ca(0H) 2 is estimated to be 500 BTU/LB. 44 2.2 EXTENSION OF ACCEPTOR LIFE BY ADDITION OF IMPURITIES INTERIM REPORT, JULY 15, 19 74 2.2.1 INTRODUCTION In the C0 2 Acceptor Process, the activity of the carbonate acceptor decreases significantly upon cycling and, also, some of the acceptor is lost due to physical attrition. As a result, fresh carbonate accep- tor must be added for process makeup. Should this process be imple- mented large quantities of carbonate rock would have to be transported to the site. Since a narrow size range of acceptor is used in the process, considerable quantities of fine carbonate rock would be produced. Furthermore, spent acceptor would also be produced. In view of these facts, the possibility of reconstituting spent acceptor and carbonate fines by melting at elevated pressure is a most attrac- tive alternate to using natural material as makeup. The technical feasibility of such a process has been demonstrated by Goran and Currant 1) with limes tone- base acceptor in a systematic study of the CaC0 3 -Ca(OH) 2 system. Subsequent work (refer to Subsection 2.3) on this system has shown that the reconstituted acceptor exhibits about the same activity as natural carbonate stone and also exhibits greater abrasion resistance than the natural stone. Upon cycling between calcination and carbonation, both the artificially produced acceptor and the natural carbonate stone lose their ability to accept C0 2 . The mechanism of activity loss has been shown to in- volve the growth of CaO crystallites which result in particle shrinkage and loss of pore volume. C 1 ) Producing acceptor that is resistant to loss of activity with cycling would be most desirable from a pro- cessing standpoint. Since the presence of small amounts of impurities is known to retard crystal growth, an investigation of the effect, that small amounts of non-lime components have on the life of the acceptor was undertaken. Three types of impurities were added, namely (1) carbonate salts of various metals, (2) anionic salts of various cal- cium compounds, and (3) various oxides. After specific additions of these compounds to lime-base melts, acceptor activity was measured as a function of cycling in a fluo-solids bed reactor. 2.2.2 EXPERIMENTAL MATERIALS AND PROCEDURES Reagent-grade chemicals were used in the investigation. They were- A1 2 3 , CaC0 3 , CaO, Ca 10 (OH) 2 (P0 4 ) 6 , Ca 3 (PO l+ ) 2 , CaMoO^, Ca(N0 3 ) 2 , CaSi0 3 , CaSO^, Fe 2 3 , MnC0 3 , K 2 C0 3 , Na 2 C0 3 , P 2 5 , SrC0 3 , and calcium tartrate. 45 2.2.3 Melts were formed in an autoclave of 1-liter capacity by heating mix- tures of reagent-grade CaO and CaC0 3 in the presence of excess water to 1400° F (furnace temperature) in the absence and presence of impuri- ties. In every case the ratio of CaO/CaC0 3 was the eutectic ratio. Acceptor activity was established by calcining 10 grams of 28 X 48 mesh material for 30 minutes at 1800°F (sample temperature). Caten- ation was effected by passing C0 2 through the bed for 30 minutes at 1360°F (sample temperature). Change of weight was measured after each cycle. Activity is defined as the molar amount of C0 2 accepted as compared to that which could be accepted theoretically. Abrasion resistance was established by placing 30 grams of 48 X 100 mesh material into a 6-1/2 X 6-1/2-inch porcelain mill containing 200 grams of porcelain balls and tumbling the mill for 10 minutes. EXPERIMENTAL RESULTS AND DISCUSSION 2.2.3.1 Activity Determinations The first portion of the investigation involved establishing the in- fluence that various metal carbonate impurities exhibit on acceptor activity. Carbonate salts of manganese, potassium, sodium, and stron- tium were added to eutectic- composition melts. The effect that each of these salts had on acceptor activity is shown in Figures 2-17 and 2-18 Manganese and strontium carbonates had no effect on activity, while the presence of both potassium and sodium resulted in a reduction in acceptor activity. When A1 2 3 , Fe 2 3 , and Si0 2 were added to eutectic- composition melts, no measureable difference in activity was noted. See Figures 2-19 and 2-20 However, when P 2 5 was added to these melts, significant improve- ment'in chemical life of the acceptor was effected. As shown in Figure 2-21 activity of acceptor containing no phosphate was 0.24 after six cycles. With additions of 3.0 and 10.0 mole percent P 2 5 , activities of 0.38 and 0.48, respectively, were measured. Melts could not be achieved with an addition of 11.0 mole percent P 2 5 . The beneficial effect that P 2 5 additions have on the ability of the acceptor to accept C0 2 on cycling was thought to be due to the reaction of P 2 5 with water forming phosphate ion. In this view, experiments were run with two calcium phosphate salts, Ca 3 (P0 1+ ) 2 and Ca 10 (OH) 2 (PO l4 ) 6 . The results obtained with 5.0 and 7.5 mole percent Ca 3 (P0 1+ ) 2 are pre- sented in Figure 2-22. Activities of 0.38 and 0.45 were measured with melts containing 5.0 and 7.5 mole percent impurity after six cycles. The addition of basic calcium phosphate, Ca 10 (0H) 2 (P0 4 ) 6 , also resulted in the extension of the chemical life of the acceptor. (See Figure 2-23.) The activity of the melt containing 6.67 mole percent impurity was 0.52 after six cycles. Melts could not be achieved with an addition of 7.5 mole percent Ca 10 (OH) 2 (P0 4 ) 6 . 46 4-> •i— S- 3 9- CO CO o CO o o o c_> C\J o c «o C\J O 3 Q ID ir> u C_5 CO C\J CO (U i— i f-H 1— 1 CC D o* S l-H CO 3 O HH oi < > CJ 2 i— i 2 HH < U E- 2 2 M O J U U >H co u E- J b tq o S 2 2 O O HH l— I H E- u i— i 2 CO :=> o Ci, CU S < o u CO 1 < u l-H H E- 2 U W UJ U E- OS 3 W W a, tu UJ O — 1 o >, S E- i— i jj > -2 l-H o E- U — 1 < < A';lal;d\/ i CM 4 7 s^ O) 1 — CO o o fc <_> at s_ c o oo o • O 00 o to oo CM CO 2 O i~4 H h-l Q a < CO ^ o 1— 1 Pi < > u 2 1— 1 2 h— i < H 2 O u CO H J PJ 2 CD o n— 2 2 <_> O M >> M J C_3 E- U M >- co u o Oh a. S O o U 2 1 o U M hH h E- U U 2 W 3 E-i pu D UJ < CU CO O < >H tO E- O l-H U > ^ i-i co E- U CU < o CM O 00 1-4 I CM CD u A'4LAL^D\/ CO o CM C\J o O) •r™ u. to >1 +-» s-s s-s •r- s_ c o o o • • z •— •— o <3 Q <£> IT) — .3- (J CO C\J 00 o CM o tu O CO 2 o 1— 1 H i— i Q Q < U 2 i-t 2 i— i < H 2 O U u co 2 H KH J hJ PJ u S >- u 2 O u. i— i o H i— i 2 CO O O h-l (X H s U o 2 u 3 1 IX u 1— 1 < H u 00 UJ < H b CM UJ o •H u. CO O "0 >-< c H rt h- 1 > to (— 1 o H CM U u 2 K- 1 2 HH < H 2 O u co H ►J w 2 u 2 2 O !— 1 i— i J E- u h-t >- co u o IX a. S O o u 2 O u i— i 1— 1 E-< H U u 2 UJ 3 E- u- 3 eg < CL, co o < > to E-> o 1— 1 CM > i— I 1— 1 < H U tu < o o CM 1 CM 0) h 3 M U- 50 o w 2 o 5 CO B i— i OS 2 O u CO E- W S 2 O i— i CO O 0, o u I u E- U tu E- W U, O >~ H U U o 2 O i— i U 2 PL, CO < o o. A'}lal:pv I (N •H PU 51 r-^ O z o Q 00 o t£> — ID CD O >> CO v? ^ cu CD r— i — o o E E O) c o LT) o • z: IT) r-» c\j O 3 C\J o Q Q < C/) O i— i OS u z < E- 1 z o u CO H PJ to o 0- o u I u t— I H u PJ E- w PL. o >- O z I— I hJ u >- u PL, o H u z PL, C/3 < O 0- /^;ial;d\/ H to < U CM CN I CN 9) bO •H PL. 52 o o in 2 o Q Q < O i— i OS > 2 < O u u 2 E- -J j u 2 u o o I— t E- 2 M O c/3 t-i O E- Oh U 2 2 O 3 U tu u < E- U UJ E- PJ o >- H U < / V o a, i — \ X o ^lal;dv I rsi <4 ^ o o O s: 51 oo (D ITJ c o LO O o • • Z21 >d- r^ o D UD LO 00 O o CO 2 o Q Q < CO o u 2 i— i 2 i— i < H 2 O u C3 2 to hH H J hJ u W >H 2 u 2 PL, O o hH H 2 l-H o co l-H o H a. u ^ 2 o 3 u Pi, u < l-H H CO u < w H W 3 E-i w < Uh U, J o 3 CO >- H ^ i— t 5 > i— i l-H u H -J U < < u ^3- CN CM a> fH 3 W) tu 55 o> u a;lal;d\/ fc o co 2 o KH E-" t— i Q Q < CO 3 O i— i OS < > u 2 i— i 2 I— i < H 2 O u CO H J UJ S 2 O u i—i 2 H i— i i— i J CO u O >- a, u 2 o u. u o u 2 1—4 O H i— i U H UJ U H 2 3 3 uj U. Uh - «a H t— i tn > O i— i ■H H CO U aj < u LO CM CN CD U 3 00 cu 56 CD CJ o O Q Q < CO O h- 1 OS s 2 < H Z CJ O 2 U h* i-J CO u E- >-> _) UJ O HH E- co cj O 2 Cu 3 S u. o cj < I cj to £« CJ UJ W H H < 3 Q UJ CO o > l-H CJ < CJ A'U Alloy CN I CN 0) ■H Uh 57 MATERIAL Natural Dolomite (Tymochtee) Eutectic (No impurity) Eutectic (1.0 mole % P 2 5 ) (3.0 mole % P 2 5 ) (5.0 mole % P 2 5 ) (10.0 mole % P 2 5 ) Eutectic (5.0 mole % Ca 3 (P0 4 ) 2 ) (7.5 mole % Ca 3 (PO lt ) 2 ) Eutectic (3.33 mole (6.67 mole Ca 10 (OH) 2 (P0 1+ ) 6 ) Ca 10 (OH) 2 (PO 1+ ) 6 ) Eutectic (3.0 mole % CaMoO^) (5.0 mole % CaMo0 4 ) Eutectic (3.0 mole % CaSi0 3 ) INITIAL WEIGHT (GM) 30 30 30 30 30 30 30 30 30 30 30 30 30 WT -100 MESH PRODUCED (GM) 14.9 4.7 5.7 5.5 5.4 8.4 12.7 17.7 5.0 22.0 5.2 6.8 12.5 Table 2-7. FINES PRODUCTION FROM 30 GRAMS IN A TUMBLING MILL 48 +100 MESH MATERIAL These data shown that the eutectic melt is harder than natural Tymochtee dolomite. Importantly the presence of P 2 5 does not alter the physical properties appreciably. Relatively large additions of Ca 3 (P0 4 ) 2 resulted in producing material with about the same abrasion characteristics as natural stone, whereas the addition of 6.67 mole percent of the basic slat resulted in making the melt softer than Tymochtee dolomite. Of all of the salts added as impurity, only the phosphorous -bearing compounds were noted to exhibit a beneficial effect on the chemical life of the acceptor. The reasons for this phenomenon are obviously complex and still need to be delineated. For example, additions of P 2 5 and of the phosphate salts resulted in approximately the same in- crease in activity of the acceptor. With additions of P 2 5 , however, the acceptor did not suffer a loss in physical properties as contrasted to phosphate salt additions (see Table 2-7). 58 Future research will involve X-ray diffraction and electron microscopy studies of the acceptor containing phosphate impurities. Further, the acceptor will be analyzed chemically for phosphorous as a function of cycling to establish whether phosphorous oxide is being volatilized, leaving pores in the acceptor. All of the experimental work thus far has involved reagent-grade chemi- cals . The next phase of the investigation will also involve reconsti- tuting spent acceptor and natural stone that exhibits poor physical properties both in the absence and presence of phosphate. 59 2.3.2 2.3 ACCEPTOR RECONSTITUTION INTERIM REPORT, JULY 31, 19 72 2.3.1 INTRODUCTION In the course of gasification of lignite using the C0 2 Acceptor Process, the activity of the carbonate acceptor decreases significantly, and also some of the acceptor is lost due to physical attrition As a result fresh carbonate acceptor will have to be added for process makeup Large quan- tities of dolomite or limestone will have to be transported to the plant site and since a narrow size range of acceptor can only be used, con- siderable quantities of fine dolomite will be produced. Furthermore, pent acceptor will also be produced. In view of these facts, the possi- bility of reconstituting spent acceptor and carbonate fines by melting at elevated pressure is a most attractive alternate to using natural material as makeup. The technical feasibility of such a process has been demon strategy Curran and GorinO) with limestone-base acceptor in a systematic study of the CaCO 3 -Ca(OH) 2 system. This work did not include a detailed study of the dolomite-base system, however The objective of this investigation is to establish whether dolomite-base acceptors can be reconstituted and, secondly, to measure the mechanical properties and activities of both lime and dolomite-base acceptors. Specific areas that have been investigated are: CI) Liquidus and eutectic temperatures of the CaC0 3 -Ca(OH) 2 system in the absence and presence of MgO (2) Conditions required for melt formation (3) Physical properties of reconstituted acceptors f41 The effect of recycling on acceptor activity (5) The feasibility of using C0 2 acceptance as a source of heat in melt formation *.;„-, (6) A preliminary analysis of the cost of acceptor ^constitution EXPERIMENTAL MATERIALS AND PROCEDURES Melts were formed in an autoclave of 1- liter capacity by heating fixtures of reagent-grade CaO and CaC0 3 in the presence of excess water to the de- sired Temper at ure in the absence and presence of MgO. none experiment a natural limestone was used in place of reagent-grade ^3. Water was vented during the heating cycle until the required steam pressure IZ obtained. In those cases where liquidus and eutectic temperatures were measured, a calibrated chrome 1-alumel thermocouple was immersed into the melt. In all experiments the melt was agitated with an over- head stirrer. Vessels for holding samples and stirrer were made from stainless steel, type 310. Physical properties evaluated were abrasion resistance and activity. Abrasion resistance was established by dry-grinding a given quantity of 48 X 100 mesh acceptor for 10 minutes in a 6-1/2 X 6-1/2-inch 60 2.3.3.1 porcelain mill containing 254 grams of porcelain media. The quantity of minus 100 mesh material produced in this test was weighed after grinding. Acceptor activity was established by calcining 3 grams of material for 2 hours at 2400° F. Two grams of calcined material were then re- carbonated at 1500°F for 15 minutes in a tube furnace. Acceptor activi ty is expressed as the following molar ratio: . . . CaCO Activity = CaCO 3 + CaO 2.3.3 EXPERIMENTAL RESULTS AND DISCUSSION Temperature Measurements in CaCO 3 -Ca(OH) 2 System The first portion of this investigation involved checking the liquidus and eutectic temperatures of the CaCO 3 -Ca(0H) 2 system and also involved establishing solubility of MgO in these melts. A typical cooling curve is presented in Figure 2-27. In this case a mixture containing 36 mole percent CaC0 3 , 54 mole percent Ca(0H) 2 , and 10 mole percent MgO was in- volved at 350 PSI steam pressure. This ratio of CaC0 3 to Ca(0H) 2 is the eutectic ratio, and as can be seen, the eutectic temperature is 1170°F. The value agrees well with the value reported by Curran and Gorin.U) Excellent agreement was also noted in liquidus temperatures. These values together with temperature measurements on other eutectic systems are listed in Table 2-8. MOLAR COMPOSITION (%) EUTECTIC LIQUIDUS CaC0 3 Ca(0H) 2 MgO TEMPERATURE (°F) TEMPERATURE (°F) 30 70 1170 1283 40 60 1175 * 1270 45 55 1175 32 48 20 1175 * 36 54 10 H70 * *Eutectic Composition Table 2-8. EUTECTIC AND LIQUIDUS TEMPERATURES IN THE BINARY SYSTEM CaC0 3 -Ca(0H) 2 , IN THE ABSENCE AND PRESENCE OF MgO 61 35 ^r o o 34 33 32 o o U c 9) S- O <*- a: •t— *» c a> o 31 30 29 28 27 26 O O O O O O O O O O O O O © © © © _L 10 20 36% CaC0 3 54% Ca(0H) 2 10% MgO 350 psig © G i 30 Time (min.) 40 1 j 50 60 Figure 2-27. COOLING CURVE OBTAINED WITH A MIXTURE CONTAINING CALCIUM CARBONATE, CALCIUM HYDROXIDE, AND MAGNESIUM OXIDE AT 350 PSIG 62 Since the eutectic temperature did not change in the presence of MgO, the solubility of MgO in this melt is negligible. In view of this fact, magnesium oxide will be present in the melt as occlusions and as a re- sult, there is a limit as to the amount of MgO that can be contained physically. The next series of experiments were conducted to establish the maximal amount of MgO that can be present and still have melt formation. This information is necessary to establish the feasibility of reconstituting dolomite-base acceptors. The results of these experiments are presented on a triangular diagram in Figure 2-28. As shown, melts are formed in systems containing up to 22 mole percent MgO. When 2 7, 30, and 35 mole percent MgO was present, mixtures of fused and unfused material were ob- tained. In the presence of 50 mole percent MgO, however, no fusion whatsoever was noted. This observation is especially significant, since this is the composition that a dolomite-base acceptor would'have. This fact may be seen from the following reasoning. On a 2-mole basis \ when dolomite is calcined and carbonated, the resulting mixture contains 1 mole of CaC0 3 and 1 mole of MgO. The MgO is inactive and does not re- carbonate. To obtain the eutectic composition of 0.6 mole of CaO and 0.4 mole of CaC0 3 , 0.6 mole of carbonated dolomite would have to be calcined. With this 0.6 mole of CaC0 3 , 0.6 mole of MgO would also be present. When the balance of the 2 original moles, 0.4 mole CaC0 3 and 0.4 mole MgO is combined with the calcine, the resulting mixture would contain 0.6 mole CaO, 0.4 mole CaC0 3 , and 1.0 mole MgO. This is the composition of the mixture that could not be fused (Point A, Figure 2-28). If dolomite is to be used as acceptor, re constitution of spent acceptor will not be possible. 2.3.3.2 Physical Properties of Reconstituted Acceptors In those systems in which the MgO content was less than about 22 percent, melts were formed. The physical properties and the acceptor activities of the melts were not affected by the presence of MgO. (See Table 2-9 ) In general the abrasion resistances of all of the synthetic acceptors are all about the same, and all are greater than that of a natural dolo- mite. Included in this data is a melt made with natural limestone; its properties are similar to other synthetic melts. With regard to acceptor activity, the activities measured were lower than those reported previously by Curran and Gorin. C 1 ) This fact is probably due to differences in technique, in that those investigators calcined in a fluid-bed reactor at 1600°F in a stream of nitrogen. The data listed in Table 2-9 were obtained by calcination at 2400° F in a muffle furnace No calcination was effected at 1600° F in a horizontal tube furnace while nitrogen was passed through it. Presumably the passage of gas through lt\™%l ™ °f part i cles enables effective calcination to take place at 1600 F The elevated temperature required to obtain complete calci- nation without a fluidized bed apparently resulted in a reduction of the ability of particles to accept C0 2 upon carbonation. 63 O No Fusion © Fused Mixture Melt Formed Ca(OH), CaCO. Figure 2-28. REGIONS OF FUSION IN THE Ca(OH) 2 -CaC0 3 SYSTEM IN ABSENCE AND PRESENCE OF MgO AT 1400°F AND 600 PSIG 64 > vO i—i l—i o to cm LO tO vO CM oo vO LO to CM 00 i— 1 i— t o vO CM i— i b < o o o O O o o o o o O o o o O bfl S LL, ►J o < o z VO 1— 1 LO CM 0} o f- CTl VO vO vO vO r^ t— i CD tu 7— 1 t—t 1 LL, to ^d- vO to LO CM tO to i—( CM Tf to CM to 1— 1 u z: w CO CO s od o o r— I r^ to LO LO vO • I—I 1-1 oc + < VO to to vO LO r-- 03 00 \£. <* Tt ^r -vf ^r rt < "* 1—4 v 1 | LL, <■""» s Q * — / o J o o o o o o H o < LO LO LO LO LO LO frj ■— < t— 1 o + (- I— I 00 1— 1 W "tf 2 LO o LO LO o LO 00 to o en vo vo o LO o LO vO vO o LO o LO ■^ « o LO LU u 2 lu co « < co ►J CM CO O H o h-4 o o rf r-4 O CM O <* i— 1 o vO "St t— i o ■—4 o o •St- I— I o o 1— 1 8 i-H o o <* 1—4 o o 1— 1 E- i— i Q 2 UJ O O o o o o o LO o 8 VO O •— i u co a. cl, LO LO LO LO o vO o LO LO vO o LO o vO CM LO o vO o tu <— > ►J X is cti u o u u o vO § o o LO o CM O to o LO to a> vO LO £ O tu e O -H O vO CD C O S o o to vO CM O o LO vO to o o o <=> o o o o •<* ^ Tt o 8 o CM 00 LO oo LO CM to LO LO to CM o O o o vO vO CM to 1—1 to CM to to CD ■(-> 'U O ■—I o a o3 •s 5 U to 8 01 CJ LL, o CO LU t— ( LU a, s a, < u i— i co >-" X a. CM CD i—i •8 E-> 65 Monies have been budgeted in the extension of work on this project for a fluo-solids reactor. All of the samples have been saved, and the acti- vities of these samples will be remeasured with this unit. 2.3.3.3 Acceptor Activity as a Function of Recycling Activity loss upon acceptor recycling is most important from a processing standpoint It\as been known that acceptor activity decreases with re- cycling" and experiments were run with a synthetically-produced eutec ic- composition mel? and with a natural dolomite. Normalized acceptor acti- vity is depicted as a function of number of calcination-carbonation cycles in Figure 2-29. Both materials were processed identically, and as a result a comparison of the data can be made. With regard to degeneration of activity upon recycling, eutectic- compo- sition acceptor exhibits more favorable properties than does Pah as apa dolomite After two cycles, for example, dolomite has lost 80 percent of its initial activity whereas synthetic acceptor has lost approximately 45 percent of its initial activity. It is felt that the unusually large rate of activity decrease is again due to the high calcining temperature used in this work. 2.3.3.4 C0 2 Injection as Source of Heat A final series of experiments were conducted to evaluate the feasibility of heating solid Ca^H) 2 by C0 Z injection, since the carbonat ion reaction L hi hypothermic, ^hemodynamic calculations show that the injection of 1 mole of C0 2 at room temperature into 2 moles of Ca(0H) 2 at 800 F would produce an unmelted eutectic mixture (40 percent CaC0 3 - 60 percent rlrnm wt 1070°F This is close to the melting point of the eutectic mixture 3 117 °F Even more thermal energy would be released if the CO 2 ere injected into a mixture of 40 percent CaO and 60 percent Ca OH 2 . irhis^asr^re^ult^g eutectic mixture would be ^Proximately SO percent molten, assuming a stoichiometric quantity of C0 2 were injected. Experiments were conducted by injecting C0 2 into Ca(0H) 2 at this > temp- erature. Three separate additions of C0 2 were made, and the results are presented in Table 2-10. INJECTION CONDITIONS PRIOR TO INJECTION ^^^^^p^fj^' CYCLES TEMPERATURE PRESSURE TEMPERATURE PRESSURE op psiG F Hblb 1 865 200 2 740 30 3 700 300 Table 2-10. RESULTS OF INJECTION OF C0 2 ON SOLID Ca(0H) 2 66 800 420 705 600 660 600 <£> fO s_ Q. o to Q. fO «■-> ■r— (J E OJ O 4-> n— ZJ o Ul Q 6 G 0- >> C\J CJ o LT> CD LO LO CM o o Alal;d\/ pezLieiujoN I 2 o I— I E- < 2 I— I u < o, o os PJ DQ o o I— I E- cj 2 Oh CO < E- U < OS O CO E- OJ CU -I w u u u < 2 o I— I H < 2 O co OS < en CM I (Nl 0) h txo •H Oh 67 The heat that was expected from the carbonation reaction was not gene- rated, and the injection of C0 2 into the reaction chamber actually cooled the system. Samples from the system showed the apparent cause for the absence of generated heat to be the presence of a layer of calcium car- bonate on the particles. The incompleteness of carbonation is attributed to slow diffusion of C0 2 into the Ca(0H) 2 particles. Although attractive from a thermodynamic point of view, kinetic limitations apparently are such that use of this technique will not be possible. 2.3.3.5 Preliminary Analysis of Cost of Acceptor Re constitution Attempts to measure the heat of fusion of the acceptor by making a heat balance on the autoclave were not made. The uncertainty from this type of experiment would be excessive due to the lack of control of heat loss from the autoclave by convection. As a result, an estimate of the heat required to melt eutectic composition acceptor (40 mole percent CaCO 3 and 60 mole percent Ca(0H) 2 has been made on the following basis. The heat required to raise the temperature of CaCO 3 from room temperature to 1175°F is 15,832 CAL/mole, while the heat required to raise the temperature of Ca(0H) 2 to the same value is 15,524 CAL/mole . ^ ) The heat of fusion of CaCO 3 is 12,700 CAL/mole, L 3 -' and an estimated heat of fusion of Ca(0H) 2 is 5,000 CAL/mole. The heat of fusion of Ca(0H) 2 has apparently never been measured. The value of 5 CAL/mole has been assumed, since the only value for a divalent metal hydroxide that could be found is that of Ba(0H) 2 , which is 4,590 CAL/mole. ^ The heat of fusion of BaO is 13,800 CAL/mole, while the heat of fusion of CaO is 12 000 CAL/mole. ^ The heats of fusion of univalent metal hydroxides are generally about 2,000 CAL/mole. On this basis the heat required to melt 1 GM-mole of eutectic mixture would be 23,730 CAL or 505 BTU/LB. In view of the uncertainty of the heat of fusion of Ca(0H) 2 , the value to two significant figures is 500 BTU/LB. This type of estimate is made on the assumption that ideal mixing of these components takes place and neglects the effects of solid phase transformations and temperature on the heat of fusion. However, these effects are usually secondary with respect to the heat of fusion, and so it is felt that this value is a reasonable estimate of the heat re- quired to melt eutectic composition acceptor in the reconstitution process . 2.3.4 SUMMARY AND CONCLUSIONS The following conclusions can be drawn from this investigation: (1) Dolomite-base acceptor which contains 50 mole percent MgO cannot be reconstituted under the conditions used. o (2) The eutectic temperature of the CaCO 3 -Ca(OH) 2 system is 1175 F which agrees well with the value reported by Curran and Gorin. 68 (3) MgO exhibits no solubility in CaC0 3 -Ca(OH) 2 melts. (4) Acceptor melts are obtained when the MgO content is less than about 22 mole percent. (5) Synthetic acceptors are hard materials. All of the acceptors pro- duced exhibited greater resistance to abrasion than one naturally occurring dolomite (Pahasapa) . (6) Acceptor activity decreases upon cycling between calcination and carbonation. Synthetic acceptor (limestone-base, eutectic compo- sition) retains its ability to accept C0 2 longer than a natural dolomite. (7) Kinetic limitations appear to rule out the possibility of using C0 2 injection on Ca(OH) 2 as a source of heat. (8) The heat required to melt eutectic mixture of CaC0 2 -Ca(0H) 3 is estimated to be 500 BTU/LB. 69 SECTION 3 STUDIES IN THE CaCOj-CaSC^-CaS AND CaCO -Ca(OH) 2 -CaO SYSTEMS This section contains a single report, "Determination of Liquidus Temperatures in the CaCO -CaSO.-CaS and CaCO -Ca(OH) 2 -CaO Systems," dated June 30, 1977, submitted by the South Dakota School of Mines and Technology. 70 3.1 DETERMINATION OF LIQUTDUS TEMPERATURES IN THE CaCO.-CaSO^-CaS AND CaC0 3 -Ca(OH) 2 -CaO SYSTEMS 3.1.1 INTRODUCTION Gasification of coal has been the subject of considerable investigation during the past decade in the search for alternate sources of fuel In this view the Consolidation Coal Company has developed an attractive technique for the conversion of western lignite coal to pipeline-quality fuel gas. Extensive testing of this process has been carried out on a pilot plant scale to obtain engineering data for subsequent process S CcU.GUp # Coal conversion in this instance is effected by contacting hydro-devola- tilized lignite char with steam at 1500° F and 150 PSIG to yield CO CO? and H 2 . The CO and H 2 , thus produced, are converted subsequently to CHl in the presence of a nickel catalyst. An essential feature of this gasification process is the removal of CO trom the reaction products with CaO which is termed a C0 2 acceptor. (5,6) Ca °(s) + C °2 (g) * CaC0 3(s) (1) This reaction is effected by the introduction of calcined dolomite or limestone into the fluidized bed reactor in which gasification takes place. In addition to removing an undesirable product from this system this reaction provides a large fraction of the heat required to sustain' the formation of CO and H 2 from lignite char. The CaO, thus consumed, is regenerated by calcination of the CaCO, reaction product. Following calcination, the regenerator acceptor is reintroduced back into the gasification process. The presence of sulfur in the coal feedstock, which reports as H 2 S upon gasification, can cause a considerable number of operating problems re- lated to the formation of CaS in the acceptor, Ca °(s) * H ^(g) * CaS (s) * »20 Cg) W ?i f r^n Ul S ^ ° CCUr When a P ortiOT of the CaS oxidized subsequently to Labu^, that is J CaS (s) + 2 %g) * CaS % s ) (3) because melts can form in the CaCOa-CaSO^-CaS ternary system causing acceptor particles to bond through surface cohesive forces. As a result a th C e eP rln\TJ tiCl t S agglomerate causi "g collapse of the fluidized bed in the CaO regenerator terminating operation of the gasification process 71 If this problem is to be avoided, it is imperative that the behavior of the CaCOq-CaSOu-CaS ternary system at elevated temperatures be known. In this view a preliminary investigation of the CaCOa-CaSO^ binary system was undertaken by Zielke.C 7 ) This investigator found that a binary eu- tectic exists at 55 mole percent CaSO^ which exhibits a fusion temperature of 1850°F. In studies of this type, it is important to maintain the C0 2 pressure at a value which is sufficiently high that decomposition of CaC0 3 does not occur, CaO_, + COo, , (4) CaC0 3(l) - CaD (1) ♦ C0 2(g) on For this reason, Zielke worked at a CO. pressure of 100 PSIA. CaSO^ the other hand, possesses sufficient thermodynamic stability tnat no such precautions need to be taken to prevent decomposition of this compound. A study of the CaS-CaSC^ system was also undertaken by this investigator. ( Partial fusion of mixtures containing 55 to 70 mole percent CaSO^ was re- ported, but complete melting could not be obtained. The results of these two investigations indicate that materials containing CaCO CaSO u , and CaS can fuse under conditions which exist in the gasi- fication of coal by the C0 2 Acceptor Process. However, these studies involved only binary systems, and additional work is, therefore, necessary in order tc obtain a better understanding of the conditions under which fusion can occur. One of the objectives of this investigation was to establish the liquidus temperatures in the CaC0 3 -CaSO lt -CaS system. The CaC0.-Ca(0H) 2 -Ca0 system is also of importance in the gasification of lignite coal by the C0 2 Acceptor Process because of its utility in acceptor reconstitution. It has been found that regeneration of spent CO acceptor can be effected by conversion of the CaO in this material to 2 Ca(0H) 9 and CaC0 3 in an aqueous solution. Following solid- liquid separation, the resulting hydroxide -carbon ate mixture can be melted readily The eutectic in this instance occurs at 42 mole percent CaCO. and exhibits a fusion temperature of 1180°F (refer to Subsection 2.3 of this report and Reference 1) . These melts also contain an appreciable quantity of CaO due to the ther- mal decomposition of Ca(0H) 2 as shown below, Ca(0H) 2(i) = CaO (i) ♦ H 2 (g) (5) This reaction occurs to an appreciable extent under the conditions required for fusion in the CaC0 3 -Ca(OH) 2 binary system. In order to gain a thorough understanding of the conditions under which acceptor reconstitution can be effected, the second phase of this research 72 involved determination of liquidus temperatures of materials containing CaC0 3 , Ca(0H) 2 , and CaO. Fusion temperatures were measured as a function of melt composition and steam pressure in order to obtain data from which optimal process conditions could be established. 3.1.2 EXPERIMENTAL PROCEDURES Reagent-grade materials were utilized in this investigation. Prior to use, CaS and CaC0 3 were dried for 4 hours at 220°F, and CaSO^ was de- hydrated at 450°F for 4 hours. Following this treatment, CaC0 3 , CaS0 4 and CaS were stored in a water- free C0 2 atmosphere. Deionized water was utilized for the generation of steam in the investigation of the carbon- ate, hydroxide, oxide system and was prepared by passing distilled water through a mixed-bed ion exchange column. Melts were made in a reactor constructed of Type RA 330 stainless steel which is shown in Figure 3-1. Because of the corrosive nature of sulfur- bearing melts, a 0.9-inch-diameter, 2-inch-high cylindrical platinum crucible was used to contain reactants in both of the systems studied The crucible was covered with a platinum disk to prevent contamination of the melt from oxide which spalled off the sides of the stainless steel reactor at elevated temperatures. The temperature in the melt container was monitored with a 1/8-inch-dia- meter chrome 1-alumel thermocouple sheathed in Type 316 stainless steel. Additionally, a 3-inch long platinum cover was placed over the thermo- ' couple to prevent interaction between the stainless steel sheath and the melt. , Liquidus temperatures were measured in these systems by generation of a cooling curve for the melts under investigation. Initial solidification of the melt was detected by monitoring the temperature difference between this material and the wall of the stainless steel container. This quan- tity exhibited a marked discontinuity when solidification initiated, and the temperature at which this phenomenon occurred was read from the con- tinuous record of melt temperature which was maintained. An electrical diagram of the system utilized to measure the melt tempera- ture and temperature difference across the reactor is shown in Figure 3-2. A high-input-impedance, two-channel, strip-chart recorder was used to maintain a record of the appropriate thermocouple outputs as a function of time. Experiments were initiated in the CaC0 3 -CaS0 4 -CaS ternary system by mixing a predetermined quantity of each of these materials in a C0 2 -filled glove box. A platinum crucible was packed with 14 grams of this mixture and the crucible was placed in the stainless steel reactor described previously The cover was placed on this vessel, and, after removal from the glove ° X k™ C ° Ve J and b ° dy Were j0ined ^ ^ectric arc welding with Type RA 330 stainless steel rod. 73 Sheathed Thermocouple Feed-through Pressure Gauge / / z i / y Vacuum Type RA330 Stainless Steel Reactor Covered Platinum Crucible Sheathed Thermocouple Figure 3-1. APPARATUS IN WHICH MELT FORMATION WAS EFFECTED 74 Temperature Across Reactor Sample Temperature Sample Reactors Ice- point Reference. ■*"*"■ "•"Wv' ~+~*~ V Figure 3-2. meas;reSent CIRCUIT UTILIZED F0R tempe Ra™re 75 After sealing was effected, the vessel was heated to 500 F and was pumped down to a pressure of 1 PSIA to remove any air or water vapor that may have entered the reactor during welding. The system was then backfilled to a pressure of 200 PSIA with C0 2 and was again evacuated to a pressure of 1 PSIA to further remove any gaseous contaminations. The system was cleansed in this manner five times. Following this treatment, the C0 2 pressure was raised to 200 PSIA and the system was heated in a vertically-mounted tub * ^™ace to 1 hour. Sample temperature and C0 2 pressure increased to 2100 F and 500 PSIA respectively, under these conditions. The reactors was held at 2100 F for 30 minutes after which time power was removed from the furnace, and a cooling curve for the system was generated. After cooling, the stain- less steel vessel was opened on a lathe, and the materials in the platinum crucible were examined to verify that a melt had formed. The platinum crucible was subsequently cleaned in an aqueous solution containing 50 percent HC1 by volume to remove the solidified melt. Melts in the CaC0 3 -Ca(0H) 2 -CaO system were made from reagent-grade CaC0 3 and Ca(0H) 2 . The CaO content of the liquid phase was adjusted by varying the pressure of steam in the reactor. Experiments were conducted in this system in much the same manner as in the CaS-CaSO.-CaCOs system. In this phase of the study, however mixing and loading of the reactor was not carried out in the presence of a pro- tective 00* atmosphere. Additionally, water was added for the generation of steam- the quantity involved varied from 1/2 to 5 grams depending on the final steam pressure desired. Cooling curves were generated by heating the reactor to 1500°F, and, after a 30 minute holding time, power was re- moved from the furnace, and the reactor was allowed to cool. Analysis of melts in the carbonate-hydroxide-oxide system was initiated by thermal decomposition of CaC0 3 and Ca(0H) 2 to yield C0 2 and H 2 0, re- spectively, as shown below: CaC0 3[s) - CaD (s) ♦ C0 2(g) (6) Ca C 0H) 2(s) - CaO (s) + H 2 (g) (?) The gaseous products which were liberated in this manner were collected and w™d to establish the quantities of CaCO 3 and Ca(OH) 2 contained in the material which was subjected to analysis. Nitrogen was used to transport the C0 2 and H 2 liberated by thermal de- composition to the gas analysis apparatus. In order to insure that contamination of the gas analysis apparatus did not occur, the nitrogen which was involved was purified prior to use in this system as shown in Figure 3-3 Water was removed from this material by passing the gas through a silica-gel-packed drier which was followed by a U- tube con- taining ascarite to absorb C0 2 . Ascarite, which is comprised of NaOH 76 Quartz Tube Reactor COo Removal V X HoO Removal z SG D 2 Tube' Furnace I a / Exit Bubbler A-Ascarite COp Absorption ^ D HpO Absorption D C0 2 -H 2 Trap D-Drierite SG-Silica Gel Figure 3-3. APPARATUS UTILIZED FOR CARBONATE AND HYDROXIDE ANALYSIS OF FROZEN MELTS 77 contained in an asbestos matrix, removes C0 2 from the gas stream by the reaction shown below: C0 2(g) + 2NaOH (s) Na 2 C0 3(s) ♦ H 2 (1) (8) In order to insure that the water which is liberated is not introduced into the analytical system, this substance was absorbed subsequently on drierite which is anhydrous calcium sulfate, CaS0 4(s) = 2H 20 C1) CaS0 4 .2H 2 (s) (9) Drierite was, therefore, packed into the exit portion of the U-tube used for C0 2 removal to retain the water liberated by the ascante. Analysis of melts was initiated by transferring a representative sample of material to a 0. 75-inch-diameter, 1.6-inch-high platinum crucible which was placed in a 2. 0-inch- diameter, 3.5-inch-high cylindrical quartz tube reactor which is shown in Figure 3-4. The temperature of reactor was raised subsequently to 1600°F, and the carbon dioxide and water vapor evolved were swept from the reactor with high purity nitrogen and collected for analysis in the apparatus shown in Figure 3-3. A nitrogen flowrate of 30 CU CM/MIN was involved in this work. The quantity of water liberated upon calcination of the solidified melt was established by removing this substance from the nitrogen gas stream with drierite contained in a Nesbitt tube. Water analysis was effected by establishing the weight gain of this container during analysis. The quantity of C0 2 liberated was established in a similar manner using a Nesbitt tube containing ascarite followed by drierite. This apparatus was protected against entry of atmospheric C0 2 and H 2 through the gas exit by placing a U-tube containing ascarite and drierite at this point. The quantity of CaO contained in the melt was established by subtracting the weights of CaC0 3 and Ca(0H) 2 present from the total weight of sub- stance involved. Analysis of CaC0 3 -CaS0 4 -CaS melts for CaC0 3 content was conducted in a similar manner. In this instance the temperature at which thermal de- composition was effected was also maintained at 1600°F to prevent the decomposition of CaSCV This phenomenon results in the evolution ot S0 3 which, in addition to C0 2 , reports to the ascarite. 3.1.3 EXPERIMENTAL RESULTS The data obtained in the generation of a typical cooling curve in the CaC0 3 -Ca(OH) 2 -CaO system are shown in Figure 3-5. The lower plot shows the difference between temperatures of the melt and reactor wall as a function of time for a system initially containing 30 mole percent CaCO 3 78 48' r S Transport Gas 150 mesh Fritted Disc h Ground Joint ■2" Diameter 150 mesh Fritted Disc Figure 3-4. QUARTZ TUBE REACTOR UTILIZED FOR THE DECOMPOSITION OF CaC0 3 AND Ca(OH) 2 IN FROZEN MELTS 79 Sample Temperature (5mv/in.) u. ~ co CO U o co rt u to o u aS U w 3d H o o a W u. o co < UJ OS < I to u 3 bO •H U. to O o o o 83 V) -o o 8 8 ■5 2 w to en w en rt U *t O co u I to o u a} u H4 F CO w OS H 2 w H co Q i— 1 00 I to W5 O O 84 The retention of C0 2 in these melts was checked for a system containing 41 mole percent CaC0 3 and 59 mole percent CaSO^. These materials were fused at a temperature of 2000° F. Upon analysis of this melt with the apparatus discussed previously, it was found that 99.85 mole percent of the C0 2 originally present remained in the melt under these conditions. The next phase of this work involved determination of liquidus tempera- tures in the CaC0 3 -Ca(OH) 2 -CaO system. Melts were made from reagent-grade CaC0 3 and Ca(0H) 2 ; the CaO content of this system was controlled by varying the steam pressure in the reactor. The results of this work are presented in Figure 3-9. Isothermal and isobaric lines are shown as a function of the CaC0 3 and CaO content of these melts. As can be noted, melts can be obtained over a relatively broad range in CaC0 3 composition, i.e. 10 to 65 mole percent. The maxi- mum CaO content of these melts is observed at a CaC0 3 composition of 20 mole percent, and this value is 5.2 mole percent. The minimum pressure required for melt formation is 200 PSIA, and the minimum temperature re- quired is 1180°F under the conditions investigated. Analysis of these melts indicates that all of the C0 2 present initially in this system is retained in the liquid phase. 3.1.4 DISCUSSION OF RESULTS The first phase of this research involved a study of the CaS0 4 -CaC0 3 binary system, and the phase diagram obtained is shown in Figure 3-6. The composition of the eutectic observed in this system is 42 mole per- cent CaSO^. Interestingly, this eutectic composition is signifcantly different from that reported by Zielke, which is 55 mole percent, ob- tained at a C0 2 pressure of 100 PSIA. The difference in these results can probably be attributed to the fact that significant decomposition of CaC0 3 occurred when the C0 2 pressure was maintained at 100 PSIA. This premise is supported by the thermodynamic data of Gaskell for the decomposition of CaC0 3 (refer to Subsection 2.3), CaC °3 (s) t Ca0 ( s ) + C0 2 (g) AG° = 31,471 - 19.1 T (10) AG and T are expressed in units of CAL/GM-mole and degrees Farenheit, respectively, and the standard states involved are the pure solids and gas at a pressure of 1 atmosphere. These data indicate that at a temp- erature of 1900°F and a C0 2 pressure of 94 PSIA, the activities of CaO and CaC0 3 in a melt would be equal. Work must, therefore, be carried out at pressures in great excess of this value to insure that significant decomposition of CaC0 3 in the melt does not occur, and for this reason C0 2 pressures of 500 PSIA were utilized in this study. Subsequently analysis of a CaCO^CaSO,, melt indicated that C0 2 retention was 99 85 percent, demonstrating that significant decomposition of CaCOo did not occur under these conditions. o o o 5 • Melt O No Melt Formed Isothermal Line (°F) Isobaric Line (psi) 5- -°a l l 10 20 30 40 50 60 70 Mol %CaC0 3 Figure 3-9. LIQUIDUS TEMPERATURES AND PRESSURES IN THE CaC0 3 -Ca(OH) 2 - CaO TERNARY SYSTEM 86 The data presented in Figure 3-7 summarize the areas of fusion, partial melting, and sintering which were observed in the CaC0 3 -CaSO l+ -CaS ternary- system. These data indicate that agglomeration of materials can occur over a relatively wide range of compositions when heated to 2000° F. Melt formation can be especially troublesome in the maintenance of a fluidized bed because this phenomenon can give rise to relatively strong interaction of materials in the particulate state. Fusion of materials is very pronounced near the CaCOs-CaSO^ binary eutectic and can occur for systems containing up to 21 mole percent CaS at 2000° F as shown by the data presented in Figure 3-8. Materials with a eutectic ratio of CaC0 3 to CaSO^ in the ternary system exhibit only a slight increase in liquidus temperature with CaS addition until the level of this material reaches 20 mole percent. The data presented in the ternary phase diagram suggest two techniques which could be applied to prevent the agglomeration of acceptor particles due to the formation of molten materials. One solution to this problem would involve removal of sulfur from the system as is currently prac- ticed at the Pilot Plant. Secondly, by maintaining conditions suffi- ciently reducing, the oxidation of sulfide ion to sulfate can be pre- vented. In the absence of CaSO^, melting will not occur because fusion is not observed in the CaC0 3 -CaS binary system as indicated by the data presented in Figure 3-7. The results of the study of the liquidus temperatures in the CaC0 3 - Ca(0H) 2 -Ca0 system are shown in Figure 3-9. As shown, the maximum level of CaO found in CaCO 3 -Ca(0H) 2 melts is 5.2 mole percent which occurs in melts that contain 20 mole percent CaC0 3 . Fusion can be obtained over a broad range of CaC0 3 contents, i.e. from 10 to 65 mole percent CaC0 3 , when the CaO level is relatively low. The calcium oxide contents of these melts increases with decreasing steam pressure due to decomposition of Ca(0H) 2 as shown by Ca(0H) 2(1) + CaO (1) + H 2 (g) (11) This phenomenon is reflected in the data presented in Figure 3-9 in that those melts prepared at the highest steam pressures have the lowest CaO contents . The CaO content of the melt must also increase with Ca(0H) 2 level, as shown by the reaction presented in Equation 11 and the data presented in Figure 3-9. For example, at a steam pressure of 800 PSIA, the CaO content of the melt rises from 2.5 mole percent when the Ca(0H) 2 level is 40 mole percent to 4.2 mole percent at a Ca(0H) 2 level of 79.8 percent. The results of calcium carbonate and hydroxide analysis of solidified melts indicate that C0 2 is retained in the molten phase upon fusion whereas water is not. The fact that C0 2 retention occurs is important 87 in the prediction of fusion temperatures in this ternary system from the phase diagram presented in Figure 3-9. The importance of this fact can be illustrated by consideration of a mix- ture containing CaO, Ca(0H) 2 , and CaC0 3 which is to be melted. Although water is exchanged in this system upon heating in the presence of steam through the reaction shown in Equation 11, the sum of the number of moles of calcium oxide, hydroxide, and carbonate present remains constant during this process. Since CaC0 3 does not decompose in this system, as indicated by the fact that all of the C0 2 is retained in the melt, the number of moles of this material in the melt is also constant. The mole fraction of CaC0 3 , the molar quantity of this material divided by the sum of the number of moles of calcium oxide, hydroxide, and carbonate, likewise, remains fixed at that fraction of CaC0 3 present initially in that mater- ial which is being subjected to fusion. This information is essential to determination of the liquidus temperature of a melt feedstock by consideration of the ternary diagram presented in Figure 3-9. Since the mole fraction of CaC0 3 in this material remains constant upon heating and fusion, the composition of the melt must fall on a vertical line that represents the CaC0 3 content of the melt feed- stock. Clearly, by specifying the pressure under which melting is to be effected, the liquidus temperature of this material is given by the intersection of the line of constant CaC0 3 content and the isobar which represents the steam pressure involved. Alternatively, if the desired fusion temperature is known, the pressure of steam required to effect melting can be obtained from these data using the appropriate isothermal line In either case the CaO and Ca(0H) 2 contents of the materials which will'be obtained under solidification of these melts are also specified by the point of intersection described above. By way of example, a mixture of materials containing 30 mole percent CaC0 3 would exhibit a liquidus temperature of 1280° F if melted in the presence of a steam pressure of 400 PSIA. Alternatively materials con- taining 40 mole percent CaC0 3 would require a steam pressure of 500 PSIA to effect fusion at 1200° F. 3.1.5 CONCLUSIONS The following conclusions can be drawn from this investigation: (1) The eutectic in the CaC0 3 -CaS0 4 system occurs at 42 mole percent CaSOit and exhibits a fusion temperature of 1850°F. (2) The eutectic in the CaS-CaSO^ system appears to be in the range of 55 to 65 mole percent CaS0 4 and exhibits a fusion temperature over 2100°F. 88 (3) Melts cannot be obtained in the CaS-CaC0 3 system at a temperature of 2100°F. (4) Fusion occurs in the CaC0 3 -CaS(VCaS system near the CaC0 3 -CaS0 4 binary eutectic. Melts can be obtained for CaS levels up to 21 mole percent at 2000° F when the CaCC^-CaSQi, ratio is that of the binary eutectic. (5) Melts made from CaC0 3 and Ca(OH) 2 can contain up to 5.2 mole per- cent CaO. (6) The CaO level of melts in this system increases with Ca(0H) 2 content and decreases with steam pressure. (7) The lowest fusion temperature observed in the CaC0 3 -Ca(OH) 2 -CaO system is 1180°F. (8) C0 2 is retained by CaC0 3 -Ca(OH) 2 -CaO melts under the conditions investigated whereas water is not. 89 SECTION 4 CORROSION OF MATERIALS IN CONTACT WITH CaC0 3 -Ca(OH) 2 MELTS This section includes two reports, both of which are titled "The Corrosion of Materials in Contact with CaC03-Ca(0H) 2 Melts." The first is the final report, dated June 30, 1977, and the second is an interim report, dated June 30, 1975. 90 4.1 THE CORROSION OF MATERIALS IN CONTACT WITH CaCO 3 -Ca(OH) 2 MELTS FINAL REPORT, JUNE 30, 19 77 4.1.1 INTRODUCTION With the increasing level of interest in the gasification of lignite coals, a considerable amount of effort has been expended in recent years to develop processes by which efficient coal conversion can be effected. One such technique which has been investigated extensively is the C0 2 Acceptor Pmcess. This technique, which is described by Fink, Sudbury, and Curran 1 - J , involves gasification of lignite with steam in the pre- sence of CaO to provide process heat through the C0 2 acceptor reaction, CaO (s) ♦ C0 2(g) + CaC0 3 (D The CaO consumed in this reaction is regenerated by calcination of the CaCO 3 reaction product, and the regenerated acceptor is reintroduced into the gasification process. Upon recycling at elevated temperature in this manner, the CaO crystals in the acceptor grow, and this increase in crystal size is accompanied by a decrease in the rate at which this material will react with C0 2 . The acceptor becomes unusable under these conditions and must be regenerated. This is effected by forming a melt from the spent material in the CaC0 3 -Ca(0H) 2 binary system. Reprocessing of CaO in this instance is initiated by the conversion of a major portion of this substance to Ca(0H) 2 in an aqueous solution. Ca °(s) + H 2°(aq) + Ca C° H )2 (s) (2) Approximately one-half of the Ca(0H) 2 thus produced is converted subsequently to CaCO 3 with aqueous C0 2 , Ca(0H) 2(s) ♦ C0 2(aq) * CaC0 3(5) + H 2 (aq) (3) The Ca(0H) 2 -CaC0 3 mixture which is obtained can be fused readily. Laboratory-scale investigations by Curran and Gorin C 1 ) and Fuerstenauu (refer to Subsection 2.2 of this report) indicate that the CaO produced upon solidification, cooling, and calcination of this product would exhibit favorable kinetic properties upon reintroduction into the CO Acceptor Process. Fusion of these materials has been the subject of numerous investiga- tions. Curran and Gorin (- 1 ) and Fuerstenau (refer to Subsection 2.3) demonstrated that melts can be formed readily when an atmosphere of high-pressure steam is maintained over the system to prevent significant decomposition of Ca(0H) 2 , Ca(0H) 2(i) - CaO (i) + H 2 (g) (4 ) 91 These investigators established that the binary hydroxide- carbonate system exhibits a eutectic at 41 mole percent CaC0 3 . The fusion temperature of this material is 1180°F. The subsequent work of Fuerstenau (refer to Section 3) indicates that a eutectic melt prepared at a steam pressure of 600 PSIG contains 3.5 mole percent CaO which forms at the expense of Ca(0H) 2 as shown by Equation 4. This investigator likewise demonstrated that MgO, which would be present if reconstitution of a dolomite acceptor were attempted, does not exhibit solubility in these melts, and, hence, prevents melt formation (refer to Subsection 2.3). Although the physio chemical properties of these lime-base C0 2 acceptors have been well quantified, very little work has been done to establish the degree to which fused CaC0 3 -Ca(0H) 2 mixtures interact with potential melt container materials. This area of investigation is of importance because large quantities of spent acceptor and comminuted fines will have to be reconstituted in the large-scale production of fuel gas by the C0 2 Acceptor Process. Containers used for reconstitution will be in almost continuous contact with CaC0 3 -Ca(0H) 2 melts and, thus, must be constructed of a material which can withstand a relatively rigorous environment. Little work has been done to quantify the rate of corrosion of steels in molten salts, such as those involved in this system, and studies that have been done were limited primarily to corrosion by molten univalent metal hydroxides and carbonates C 9 » 10,11 , 12) m n is difficult to predict the behavior of materials in contact with CaC0 3 -Ca(OH) 2 melts from these studies, however, because of the high steam pressures and temperatures required to form melts in the lime-base system. Due to the absence of corrosion data for molten CaC0 3 -Ca(OH) 2 systems, a study was initiated to identify those materials which would function well as melt containers. Since the high steam pressures involved pre- clude the use of mild steels in this application, corrosion- resist ant nickel and chromium alloys were involved. Specific areas that were investigated are: (1) measurement of the weight loss of metals in contact with CaC0 3 -Ca(0H) 2 melts, (2) examination of the micros tructure of oxidized alloys and oxidation products, and (3) determination of the rate of growth of the oxide layer on alloy steels. 4.1.2 EXPERIMENTAL PROCEDURES The following alloys were involved in this work: AISI Types 304, 304L, 309, 310, 316, 316L, 446: Inconel 600, and 601; Incoloy 800; and Nickel 200. Cold drawn, centerless ground, and annealed 1/2-inch- diameter rounds were utilized as the alloy source in each instance. The surface of this material exhibited a bright metallic luster and was completely free of mill scale. The composition of the alloys in- volved in this work is presented in Table 4-1. Preparation of samples for weight-loss experiments was initiated by cutting 0.3 inch of material from the appropriate alloy stock. Each face of the resulting cylinder was ground with 240-grit abrasive paper to remove cutting marks after which the sample was washed in water, dried, and measured with a micrometer. Determination of the sample 92 weight was initiated by washing in acetone to remove trace amounts of cutting oil present on the metal surface. After weighing, the sample was washed subsequently in acetone and was reweighed. This cycle was continued until a weight change of less than 0.1 MG was observed; the last weight obtained was utilized in the subsequent calculation of weight change in the corrosion experiment. ALLOY TYPE Cr Ni Mn Co AISI 304 30 4L* 309 310 316** 316L* 446 Inconel 600 601 Nickel 200 - 99.1 0.2 0.1 *Low carbon Types 304 and 316, respectively. The maximum carbon content of these materials is specified as 0.03%. Additionally, this alloy contains 2.6% Molybdenum. Analyses are expressed in weight percent. The balance of material in each alloy is iron. 18.6 9.0 1.6 - 18.9 9.3 1.7 _ 22.0 12.0 1.7 _ 24.0 20.0 1.7 _ 18.6 12.0 1.7 _ 18.6 13.0 1.7 _ 25.4 - 0.9 - 15.6 72.0 1.0 2.3 20.4 56.0 1.0 2.0 ** Table 4-1. COMPOSITION OF THE IRON-CHROMIUM-NICKEL ALLOYS INVOLVED IN THIS INVESTIGATION In those cases in which the importance of galvanic phenomena was to be established, the weight loss experiment was initiated by placing the weighed sample into a 5-inch-diameter, 7-inch-high, cylindrical Type 310 stainless steel crucible. When new crucibles were received, they were heated in the presence of air at 2000° F for 5 hours to ensure that a uniform oxide layer was present on the surface of the stainless steel container. The sample was placed in the crucible and was covered with 200 grams of a mixture containing a predetermined quantity of reagent- grade CaC0 3 and Ca(0H) 2 , 100 ML of distilled-deionized water was added, and the crucible was placed in a 1-liter autoclave. The autoclave was' sealed subsequently and heated for approximately 8 hours to yield a sample temperature of 1400°F. Excess water was bled from the system during heating to maintain the steam pressure at 600 PSIG. Measurement of sample exposure time was initiated when the autoclave attained the desired temperature. 93 In subsequent weight- loss experiments, samples were contained in dense alumina crucibles, which were obtained from Alfa Division of the Ventron Corporation. The crucibles were 1.6 inches high with an inside diameter of 1.0 inch. In this instance weight- loss experiments were initiated by placing an alloy sample in the crucible after which 17 grams of a mixture containing reagent-grade CaC0 3 and Ca(0H) 2 were packed into the container. The dense alumina container was subsequently placed in an autoclave liner; excess water was added, and the system was heated to yield a sample temperature of 1400°F. The pressure of the system was again maintained at 600 PSIG. After the samples had been contacted with melt in a stainless steel or alumina crucible for a predetermined length of time, the autoclave was allowed to cool for 12 hours after which time the crucibles were removed. The sample was broken out of the solidified melt, and adhering solid was removed by contacting the sample for 10 hours with a solution containing 10 G/L anhydrous sodium acetate adjusted to pH 5.0 with acetic acid. Previous work indicated that this solution would not corrode the alloys which were involved in this work. After the solidified melt had been dissolved from the surface of the sample, loosely-bound oxide was removed by abrading the materials involved for 10 hours in an 8-inch -diameter pebble mill rotated at 50 RPM and filled to 50 volume percent with 20 X 35 mesh silica sand. Following abrasion, the samples were cleaned in acetone and weighed as described previously. This cyclic abrasion-weighing treatment was continued until the sample weight remained constant. Previous work indicated that the alloys involved in this work do not lose weight during this abrasion step, indicating that conditions are not sufficiently severe to remove metal from the surface of the sample. The thickness of the oxide layer present on the surface of these steels was also established. In order to prevent change in the dimension of the oxide layer, these samples were not contacted with the acidic solu- tion of sodium acetate, nor were they abraded in the pebble mill. Thick- ness measurements were initiated by mounting samples in a metal lographic bakelite holder inside of a . 75-inch-diameter ring to ensure good edge retention upon grinding and polishing. The sample was then cut so that the edges could be observed along a plane parallel to the axis of the cylindrical metal sample. The exposed surface was polished with 3, 2, 1, 0, 2/0, 3/0, and 4/0 grit a-Al 2 3 on a polishing wheel. Oxide layer thickness was established by measuring this quantity optically at a magnification of 500X using a metallograph . Fifty thickness measure- ments were made for each sample. The latter phase of this work involved contacting alloy samples with melts made from purge acceptor obtained from Run 33B of the C0 2 Acceptor Process Gasification Pilot Plant. The feed stock for this acceptor was Minnekahta limestone. An elemental analysis of this spent acceptor is presented in Table 4-2. 94 A1 2 3 2.81 Total CI 0.025 CaO 85.77 Fe^ 1.66 K2 0.03 MgO 3.35 Na 2 . 14 Si0 2 5.76 SO 3 0.0 7 Muffle loss 0.26 Analyses are expressed in weight percent. Table 4-2. ELEMENTAL ANALYSIS OF PURGE ACCEPTOR OBTAINED FROM RUN 33B In order to make a melt from this spent acceptor, it was necessary to convert a major fraction of the CaO which was present to Ca(0H) 2 or to CaC0 3. Ca(0H)2 was produced by hydrating the spent acceptor in a slurry containing 50 percent solids by weight. CaC0 3 was produced by initially hydrating an equal quantity of spent acceptor to Ca(0H) 2 which was carbo- nated subsequently by passing CO2 through the slurry for 4 hours. The resulting materials were dried at 110°C. An analysis of the hydroxide and carbonate contents of these materials is presented in Table 4-3. CARBONATED STOCK HYDROXIDE STOCK CaC03 80.3 4.9 Ca(OH) 2 6.2 81.7 CaO plus inert material 13.5 13. 4 Analyses are expressed in weight percent. Table 4-3. THE CALCIUM CARBONATE AND HYDROXIDE CONTENTS OF CARBONATED AND HYDRATE D PURGE ACCEPTOR OBTAINED FROM RUN 33B 95 Experiments involving this material were initiated by making a mixture containing 59.5 weight percent of the carbonated acceptor; the balance of material was the hydrated acceptor. The required composition was calculated assuming that the CaO present in this mixture would hydrate upon fusion to yield a melt of equimolar CaCO 3 -Ca(OH) 2 content. It was found that melts formed in this system exhibited excessive poro- sity and were therefore unusable in the corrosion experiments. Melting temperatures in the range of 1400 to 1500°F and pressures in the range of 500 to 1000 PSIG were investigated, but this problem persisted. Melt formation was achieved finally in this system by removal of a portion of the siliceous material present; this work was performed by Dr Michael Lancet of the Conoco Coal Development Research Laboratory, Library, Pennsylvania. Hydration was effected by placing 100 grams of spent acceptor into a quartz tube reactor, and sufficient water was added to make slurry of 25 percent solids. This solution was heated and held at 140°F for 30 minutes and, after cooling, was passed through a 200-mesh screen to remove particles of siliceous material. The sieve undersize which comprised 65 percent of the material processed was dewatered on a Buchner funnel and was dried in a vacuum to yield the calcium hydroxide stock material. Carbonated acceptor was made by combining 50 grams of the hydrated acceptor with 100 grams of water and passing C0 2 through the resulting slurry. Carbonation was continued until the slurry temperature, which rises initially, returned to ambient. The slurry remained in contact with C0 2 in this system for approximately 1 hour. The carbonate melt stock was recovered subsequently by filtration and vacuum drying. Analyses of these carbonate and hydroxide melt stocks are presented in Tables 4-4 and 4-5. CARBONATE STOCK HYDROXIDE STOCK AI2O3 L34 n 2 A\ CaO 91-98 90 - 53 Fe 2 3 0.81 lAS K 2 0.01 0.01 MgO 1-67 1.13 Na 2 0-06 0.67 p5 0.05 0.05 Sl ° 3 ' 53 n'no SO, 0.07 0.09 Ti0 2 0.06 0.07 Muffle loss 22.31 41.99 Analyses are expressed in weight percent. Table 4-4. ELEMENTAL AN/ iSIS OF THE CaCO 3 AND Ca(0H) 2 PREPARED FROM RUN 33B PURGE ACCEPTOR AT THE CONOCO COAL DEVELOP- MENT RESEARCH LABORATORY The carbonate and hydroxide analyses of each material are presented in Table 4-5. 96 CARBONATE STOCK HYDROXIDE STOCK CaC0 3 90.6 4.2 Ca(OH) 2 0.0 83.2 CaO plus inert material 9.4 12.6 Analyses are presented in weight percent. Table 4-5. THE CALCIUM CARBONATE AND HYDROXIDE CONTENTS OF SEIVED CARBONATED AND HYDRATED PURGE ACCEPTOR OBTAINED FROM RUN 33B Melt formation was initiated in this system by making a mixture contain- ing 58.8 and 41.2 weight percent of the carbonate and hydroxide stock materials, respectively. The required composition of this mixture was calculated assuming that the CaO present in materials would form Ca(0H) 2 upon fusion. Subsequent analysis of material melted at 1400°F and 500 " PSIG steam pressure established that the actual content of the fused material was 54.1 mole percent CaC0 3 and 45.9 mole percent Ca(0H) 2 . This composition is very nearly that which would be predicted assuming the CaO in the carbonate and hydroxide stocks does not hydrate upon fusion. These data indicate that hydration of CaO does not occur completely in this melt system which is consistent with the ternary diagram reported by Fuerstenau (refer to Section 3) . Samples which had been contacted with CaC0 3 -Ca(0H) 2 melts were also subjected to microscopic examination to establish if intergranular attack or pitting had occurred. Materials to be examined were polished in the same manner as the steels which were involved in the oxide layer thickness measurements. The alloys were etched subsequently for 1 second in a solution of concentrated HC1 saturated with CuCl 2 2H 2 0. The sample was then examined with a metallograph at a magnification of approximately 200 times. The alloys involved in this work were also electropolished and sub- jected to metallographic examination. Stainless steel samples which had been polished with 4/0 emery paper were electropolished at a current of 0.8 amperes for 3 minutes in a solution containing two parts by volume methanol and one part concentrated nitric acid. The high nickel alloys, i.e., Nickel 200, Inconcel 600, and Inconel 601 were electropolished in a solution containing 50 percent concentrated phosphoric acid, 30 percent concentrated sulfuric acid, and 20 percent water by volume. Identification of corrosion products was initiated by mechanically removing approximately 10 MG of oxide from the metal surface. Subsequent examination was effected by X-ray diffraction with copper K-a radiation. EXPERIMENTAL RESULTS Initial work involved following the corrosion of several iron- chromium- nickel alloys as a function of contact time with eutectic composition melts made from pure materials. The weight loss of alloys contained 97 in a Type 310 stainless steel crucible and exposed to melts at 1400°F and a steam pressure of 600 PSIG was the subject of the first phase of this investigation. The corrosion data for two 18-8 stainless steels, Types 304 and 304L, are shown in Figures 4-1 and 4-2. Linear weight loss curves are observed, and the low-carbon alloy, Type 304L, exhibits a higher corrosion rate than that of the material which has a relatively high carbon content. The corrosion of two alloys of relatively high chromium content, Types 309 and 310 stainless steel, is shown in Figures 4-3 and 4-4. These materials also exhibit linear weight loss curves. The rate at which weight loss occurs is significantly lower than that observed for the Type 304 alloys described previously. The next phase of this investigation involved following the weight loss of two additional 18-8 stainless steels, Types 316 and 316L, as a function of time, and the results of this phase of the investi- gation have been presented in Figure 4-5. These data indicate that the corrosion rates of these alloys are very nearly equal. However, Type 316L stainless steel loses weight at a higher rate initially than does the high-carbon alloy. Corrosion of the alloy of highest chromium content, Type 446 stainless steel, is shown in Figure 4-6. As demonstrated by these data, a rela- tively low corrosion rate is observed in this instance. The behavior of two alloys of relatively high nickel content, Inconel 601 and Nickel 200, is shown in Figures 4-7 and 4-8 respectively. Inconel 601 exhibits a relatively low weight loss for a 50-hour contact time. However, weight loss increases dramatically for longer reaction times. Conversely, Nickel 200 exhibits a weight-loss curve which is typical of the other alloys studied, i.e., weight loss occurs at a relatively high rate initially which approaches a steady state value for relatively long contact times. The second portion of the weight -loss study involved following the corrosion of these alloys when alumina crucibles were used to contain the metal samples. Steam pressure and temperature were maintained at 600 PSIG and 1400°F, respectively. The weight loss of two 18-8 stainless steels, Types 304 and 304L, is shown as a function of time in Figure 4-9. These data indicate that weight loss occurs at a constant rate in this instance. Additionally, Type 304L exhibits a higher corrosion rate than does Type 304. The corrosion of two alloys of higher chromium content, Types 309 and 310, is shown in Figure 4-10. The alloy with the highest chromium and nickel contents, Type 310, exhibits the lowest corrosion rate of the two stainless steels. Weight loss occurs at a constant rate in both cases. 98 C\J o CD < s- a. on o en CD 50 100 150 Contact Time (hrs) 200 Figure 4-1. CORROSION OF TYPE 304 STAINLESS STEEL IN A STAINLESS STEEL CRUCIBLE IN THE PRESENCE OF EUTECTIC COMPOSITION MELT 99 C\J CD E s- < +-> •r— C ZD S- •r- 150 100 - 50 - 50 100 150 Contact Time (hrs) 200 Figure 4-7. CORROSION OF INCONEL 601 IN A STAINLESS STEEL CRUCIBLE IN THE PRESENCE OF A EUTECTIC COM- POSITION MELT 105 CM O $- s- ^: •r— CD 200 300 Contact Time (hrs) 400 500 Figure 4-12. CORROSION OF TYPE 446 STAINLESS STEEL IN AN ALUMINA CRUCIBLE IN THE PRESENCE OF A EUTECTIC COMPOSITION MELT 111 20 C\J 15 _ E Q- to O x: CD 10 - 100 200 300 Contact Time (hrs) 400 500 Figure 4-13. CORROSION OF INCONEL 601 AND NICKEL 200 IN AN ALUMINA CRUCIBLE IN THE PRESENCE OF A EUTECTIC COMPOSITION MELT 112 Figure 4-14. MICROPHOTOGRAPH OF THE SURFACE OF INCONEL 600 AFTER CONTACT WITH A EUTECTIC COMPOSITION MELT FOR 100 HOURS IN AN ALUMINA CRUCIBLE. THE METAL SUBSTRATE IS AT THE BOTTOM OF THE PHOTOGRAPH. MAGNIFICATION 170X 113 The next phase of this work involved following corrosion rate as a function of the composition of the CaC0 3 -Ca(0H) 2 melt for a 100-hour contact time. Alumina crucibles were utilized, and temperature was maintained at 1400° F. Six alloys of a wide range in composition were chosen for this work, and these were Types 304, 309, 316, and 446 stain- less steel, Inconel 600, and Inconel 601. The results of this phase of the investigation are shown in Figure 4-15. These data indicate that corrosion rate is not a function of the carbonate content of the melt. All alloys, with the exception of Inconel 600, exhibit very low weight loss under the conditions examined. Conversely, Inconel 600 gained weight in all of these melts. Rate was found to be independent of composition in all cases except that of Type 316 stainless steel; weight loss was found to decrease slightly with carbonate content in the case of this alloy. The next phase of this work involved following the weight loss of two of the 18-8 stainless steels, Types 304 and 316; three alloys of relatively high chromium content, Types 309, 310, and 446; and a high nickel alloy, Inconel 601, in the presence of additions of compounds containing chloride, sulfide, sulfate, potassium and sodium ions. Samples were contained in alumina crucibles in this portion of the investigation. Initially, additions of CaCl 2 , CaS , CaSO^ , KOH, and NaOH were made to eutectic-ratio CaC0 3 -Ca(0H) 2 melts, in separate experiments, at a level of 1 mole percent. Under these conditions the weight loss of these alloys could not be distinguished from those observed in the presence of the eutectic-composition melts. The addition of these salts was, therefore, increased to 5 mole percent in order that potential problems with these compounds could be identi- fied. The weight changes that were observed are presented in Table 4-6 ALLOY 304 309 310 316 446 1601 No Addition -2.2 -1.7 -0.6 -1.7 -0.7 -0.7 CaCl 2 -2.8 -2.0 -0.4 -1.9 +0.1 -1.5 CaS -4.5 -2.2 -0.7 -4.8 -0.2 -4.1 CaSO^ -1.7 -3.1 -0.3 -0.6 + 0.5 -3.1 KOH -2.2 -5.4 -0.1 -3.1 +0.1 -6.1 NaOH -0.5 -1.8 -0.5 -2.2 -0.7 -1.8 *This value is the average weight loss measured in two experiments. Samples were contained in alumina crucibles, and the molar ratio of CaC03 to Ca(0H) 2 was 41 to 59, i.e. that of a eutectic composition melt. Weight change is expressed in units of MG/SQ CM. Table 4-6. WEIGHT CHANGE OF SEVERAL ALLOYS FOR A 100-HOUR CONTACT TIME AT 1400°F IN THE PRESENCE OF 5 MOLE PERCENT OF THE SALTS LISTED 114 304 t 30 -8- i 35 -o- L_ 40 o 45 ~6 I t < 0) ex. 01 C7> c s: o> •t— 0) 100 200 Contact Time (hrs) 300 400 Figure 4-16. CORROSION OF SEVERAL ALLOYS IN ALUMINA CRUCIBLES IN THE PRESENCE OF MELTS MADE FROM SPENT ACCEPTOR, EACH MELT CONTAINED EQUIMOLAR QUANTITIES OF CaC0 3 AND Ca(0H) 2 117 crucible. In the case of Nickel 200, NiO and NiOOH-Ni0 2 were found when corrosion was effected in an alumina crucible, and NiO was found when a stainless steel container was utilized. 4.1.4 DISCUSSION OF RESULTS The corrosion of the materials involved in this work is presented as a function of time in Figures 4-1 through 4-13. As indicated by these data, the alloys involved in this work lose weight at a rela- tively high rate when initial melt-metal contact is made. After the first few hours of contact, however, rate becomes constant as indicated by the linear relationship between weight loss and time. This pheno- menon can probably be attributed to the fact that the oxide layer which forms initially under a high rate of growth contains more defects, and, hence, is less protective than that material which forms relatively slowly in the later stages of oxidation. Conversely, in the case of Inconel 601 contained in a stainless steel crucible, a relatively low weight loss is observed for a contact time of 50 hours. For longer times, however, a very large change in weight is observed. These data most likely reflect the formation of a product layer which spalls off of the surface of the alloy after an initial period of growth. The long-term corrosion rate in this system can be assessed from the slopes of the weight-loss curves, since rate assumes a steady-state value relatively quickly. The rate of weight loss for these alloys was obtained by linear regression from the data presented in Figures 4-1 through 4-13. The rate of reduction in the thickness of these materials, T, commonly expressed in units of mils per year, can be obtained from these data by application of T - 2. (5) P where r is the rate of weight loss per unit area which is obtained from the weight-loss time data, and p is density. The quantities shown in Equation 5 must be expressed in consistent units. The rates of corrosion calculated in this manner have been presented in Table 4-8 for alloys contained in both stainless steel and alumina crucibles. Additionally, the cost of 1-ton quantities of 2-inch plate fabricated from these materials, as of July 1977, has also been presented 118 ALLOY CORROSION RATE (MILS/YR) STAINLESS STEEL CRUCIBLE AISI 304 15 304L 17 309 6 310 6 316 25 316L 20 446 6 Inconel 600 * Inconel 601 390 Incoloy 800 ** Nickel 200 11 ALUMI NA COST PER CRUCI BLE 100 LBS($) 5 106.56 10 113.55 3 145.84 2 1 82 . 5 1 5 136.54 12 143.82 3 95.00 ' 414.43 1 379.00 11 459.57 * Gains weight due to the formation of NiOCH-Ni02 on the metal surface ** Due to the high rate of corrosion, work with this material was not undertaken. The cost of 2-inch plate is presented for 1-ton quantities of material. Table 4-8. CORROSION RATE OF THE ALLOYS INVOLVED IN THIS STUDY IN CONTACT WITH EUTECTIC-COMPOSITION MELTS MADE FROM PURE MATERIALS . These data indicate that Types 309, 310, and 446 stainless steel exhibit the best behavior of the alloys studied. The relatively high corrosion rates observed when these alloys are corroded in a stainless steel crucible indicate that galvanic phenomena occur in this system, as would be expected. The highest corrosion rate observed for these alloys is 6 mils per year. Uhlig (* 3 ) states that in the case of materials utilized for handling chemical media, rates of 5 mils per year or less are indicative of good corrosion resistance when attack is uniform. These materials very nearly meet this criterion when exposed to melt in a stainless steel crucible and perform very well when exposed in an alumina crucible, indicating that they can be utilized to contain melts in this system. The stainless steels of lower chromium content, Types 304, 304L, 316, and 316L exhibit higher corrosion rates in the presence of both alumina and stainless steel crucibles. Moderate corrosion rates are observed when the alloys are contained in alumina crucibles, in the range of 5 to 12 mils per year. When stainless steel crucibles are 119 utilized, rates increase to 15 and 25 mils per year. These data indi- cate that galvanic phenomena again assume importance and cause the 316 series of steels to degrade at a rate which makes their appli- cation in this system questionable. I.iconel 600 exhibits a very high rate of corrosion. The phenomena which are observed can be described best in terms of other data, and the performance of this alloy will be reviewed, therefore, in subsequent discussion. Inconel 601 exhibits a very low rate of corrosion when contacted with melts in an alumina container. However, the corrosion rate observed when corrosion is effected in a stainless steel crucible is almost eight times greater than the limit established by Uhlig for materials used in handling chemical media, which value is 50 mils per year. The extreme sensitivity of this alloy to galvanic phenomena make its use in this system undesirable. Nickel 200 exhibits a corrosion rate of 11 mils per year when contained in both alumina and stainless steel crucibles. In view of the intended application, this corrosion rate is acceptable, but other considerations make the use of this material undesirable as will be discussed subse- quently. The materials that were involved in the weight-loss experiments were also subjected to metallographic investigation. The results of this work are most significant in that substantial quantities of oxide reaction products, identified subsequently by X-ray diffraction, are observed on the surface of Inconel 600 as shown in Figure 4-14. The tendency to oxidize at a high rate is observed when corrosion occurs in both alumina and stainless steel crucibles although the rate is ten times higher in the latter case. The excessively high corrosion rate that is observed in conjunction with the porous nature of the oxide indicates that the product layer is not protective in this instance. Because of this phenomenon, Inconel 600 must be ruled out for use in the construction of a melt container. The metallographic examination also indicates that extensive inter- granular attack of Nickel 200 occurs when this material is corroded in a stainless steel container. (Refer to Subsection 4.2.) This phenomenon, is of such a nature that the use of this alloy cannot be recommended. Metallographic examination also confirmed that extensive attack of Inconel 601 occurs as shown by the uneven nature of the metal surface. (Refer to Subsection 4.2.) Following this work, the performance of Types 304, 309, 316, and 446 stainless steel, Inconel 600, and Inconel 601, alloys of a rela- tively wide range of nickel and chromium content, was studied as a function of melt composition. The results of this work, shown in Figure 4-15 indicate that melt composition is not an important factor in the selection of an alloy for the construction of a melt container. 120 With the exception of Type 316 stainless steel, weight loss for a 100-hour contact time was independent of the composition of the melt. Weight loss for Type 316 stainless steel, on the other hand, decreases as a function of the CaCO content of the melt. All alloys, with the exception of Inconel 600, exhibit a relatively small weight loss, less than 2.5 MG/SQ CM. The large weight gains observed in the case of Inconel 600 occur at all compositions investigated, indicating that this material would not function as a melt container in any system of practical significance. Following this work the corrosion of six alloys, Types 304, 309, 310, 316, and 446 stainless steel, and Inconel 601, was established in the presence of melts containing separate additions of 5 mole percent CaCl 2 , CaS, CaSO^, K0H, and NaOH. Relatively high additions of these salts were involved in order to obtain weight losses that could be differentiated statistically from those observed in the presence of only the eutectic composition melt. The results of this work, which are presented in Table 4-6, indicate that Types 304 and 316 stainless steel exhibit sensitivity towards the presence of sulfide ion. In the presence of CaS, weight loss is about twice as high as when this material is absent. Additionally, the corrosion rate of Type 316 stainless steel accelerates slightly in the presence of KOH. The slight increase in weight loss for Types 304 and 316 stainless steel in the presence of CaCl2 is not thought to be statistically significant. Conversely, these data indicate that the corrosion of Type 309 stain- less steel is enhanced in the presence of CaSO^, NaOH, and KOH. In- conel 601, likewise, exhibits a tendency to lose weight at an accele- rated rate in the presence of these materials and CaS. Increased weight loss observed in the presence of CaS and CaSO^ is probably due to the formation of nonprotective sulfur compounds. The accelerated rate observed in the presence of NaOH and KOH may indicate that alkali metal ions enhance the dissolution of protective oxides or promote the spalling of this corrosion product from the metal surface. It is interesting to note that Inconel 601 also exhibits enhanced weight loss in the presence of CaCl2- This is the only alloy of those involved in this phase of this investigation that was found to exhibit sensitivity toward the presence of chloride ion; the other materials involved in this work performed very well in the presence of relatively high additions of this species. Two alloys of relatively high chromium content, Types 310 and 446 stain- less steel, perform well in the presence of these melts in spite of the relatively high concentrations of sulfur, chlorine and alkali metals involved. These data confirm that these materials would function well in the construction of a melt container. The behavior of six alloys, Types 304, 309, 310, 316, and 446 stain- less steel and Inconel 601, in contact with spent acceptor for 100 hours at a steam pressure of 600 PSIG and a temperature of 1400° F was also 121 examined. The results of this work, shown in Figure 4-16, indicate that these alloys gain, rather than lose, weight when contacted with a melt made from spent acceptor. The weight gains which are observed are relatively small, on the order of 2.0 MG/SQ CM. These results are interesting and probably indicate that the corrosion products which form on alloys exposed to this melt do not detach readily from the metal surface. The protective nature of the oxide layer in this system is confirmed by the fact that weight losses are not observed in this system for contact times as long as 300 hours. The products found on the surface of stainless steels contacted with melts made from spent acceptor are identical to those present on steels which had been in contact with melts made from pure materials. These observations indicate that similar reactions occur in the corrosion of the stainless steels in the two melts investigated. The melts made from spent acceptor, therefore, do not appear to impart any destructive reactions to the alloys involved in this work. Additionally, microscopic examination of a cross section of the surface of each alloy indicated that these materials do not exhibit any undesirable tendency to pit or undergo intergranular attack. The favorable behavior of these alloys in contact with spent acceptor can probably be attributed to the relatively low levels of harmful impurities present in the melt materials. For example, the levels of sulfide, sulfate, and chloride present in the spent acceptor are many orders of magnitude less than those involved in experiments in which 5 mole percent CaS, CaS0t+, or CaCl2 were added to eutectic-composition melts. The fact that relatively low corrosion rates are observed and that attack by the impurities present does not occur is encouraging and indicates that melts can be contained during reconstitution with- out undue difficulty. The growth of oxide on the surface of alloys in contact with acceptors made from pure materials was also examined. As shown by the data pre- sented in Table 4-7, the rate of growth of this product layer is rela- tively low during the initial stage of oxidation and decreases for longer contact times. These data simply reflect the fact that the oxide which forms on the surface of these alloys spalls off of the metal surface for relatively long contact times. Oxide layer thick- ness becomes constant when the thickness of this product layer becomes such that the rate of spalling and growth of new oxide become equal. Importantly, observation of these oxidation products indicated that the oxide did not exhibit a tendency to form fissures or deep cracks which would be detrimental to the protective properties of these surface reaction products. 4.1.5 RECOMMENDATIONS Two alloys of relatively high chromium content, Types 310 and 446 stainless steel, exhibit the most favorable performance of those 122 materials studied. The weight loss of these materials when exposed to melts in both alumina and stainless steel crucibles is relatively low. Likewise, these materials do not exhibit sensitivity toward the presence of high levels of sulfur, chlorine, or alkali metals. In no case was undesirable pitting or intergranular attack of these materials observed. These materials are also of relatively low cost as indicated by the data shown in Table 4-8. Importantly, an analysis of price trends for these alloys indicates that their cost has not increased appreciably over the last 2 years; this is not the case with the alloys of high nickel content. It is therefore recommended that Type 310 or 446 stainless steel be utilized in the construction of a melt container. 4.1.6 CONCLUSIONS (1) Materials of relatively high chromium content, i.e. Types 309, 310, and 446 stainless steel exhibit good corrosion resistance when con- tacted with eutectic-composition melts made from pure materials in both alumina and stainless steel crucibles. Additionally, Type 304 stainless steel and Inconel 601 exhibit good performance in this system when they are contained in an alumina crucible. (2) Types 304L and 316L stainless steel and Nickel 200 exhibit satisfac- tory performance when contacted with eutectic-composition melts in alumina crucibles. (3) Although the corrosion rate of Nickel 200 in the presence of eutectic- composition melt contained in a stainless steel crucible is rela- tively low, extensive intergranular attack of this material under these conditions prohibits the use of this alloy. (4) Types 304 and 304L stainless steel exhibit marginal corrosion resistance when exposed to eutectic-composition melts in a stainless steel crucible. (5) The performance of Types 316 and 316L stainless steel in the presence of eutectic-composition melt in a stainless steel crucible is marginal. (6) The performance of Inconel 601 in the presence of eutectic-composition melt contained in a stainless steel crucible is unsatisfactory. (7) The performances of Inconel 600 and Incoloy 800 in this system are unsatisfactory. (8) Types 310, and 446 stainless steel exhibit good performance in the presence of eutectic ratio CaC0 3 -Ca(OH) 2 melts containing separate additions of 5 mole percent CaCl 2 , CaS, CaSO^, KOH, and NaOH in alumina crucibles. (9) Type 304 stainless steel exhibits sensitivity toward the presence of CaS at a level of 5 mole percent. (10) Type 309 stainless steel exhibits an accelerated rate of corrosion when exposed to melts containing 5 mole percent CaSO^ and KOH. (11) An increase in the rate of corrosion of Inconel 601 is observed when this alloy is exposed to separate melts containing CaCl 2 , CaS, CaSOtt, KOH, and NaOH. This effect is most significant in the cases of CaS, CaSO^, and KOH. 123 (12) The corrosion rates of Types 304, 309, and 446 stainless steel and Inconel 600 and 601 are independent of melt composition. (13) Types 304, 309, 310, 316, and 446 stainless steel and Inconel 601 exhibit good performance in contact with spent acceptor in alumina crucibles . (14) With the exception of Inconel 600, the oxide that forms in this system is protective. In the case of Inconel 600, porous non- protective Ni00H-Ni0 2 is observed on the metal surface. 124 4.2 THE CORROSION OF MATERIALS IN CONTACT WITH CaC0 3 -Ca(OH) 2 MELTS INTERIM REPORT, JUNE 30, 1975 4.2.1 INTRODUCTION The technical feasibility of reconstititing lime-base C0 2 acceptors has been well demonstrated. Initial work in this area by Curran and Gorin demonstrated that reconstitution could be effected through fusion of CaC0 3 -Ca(OH) 2 mixtures. C 1 ) Further investigation by Fuerstenau estab- lished the low solubility of MgO in CaC0 3 -Ca(OH) 2 melts, the physical properties of synthetic acceptors, and the effect of impurities on the loss of activity of acceptor obtained through reconstitition. (Refer to Subsections 2.2 and 2.3.) Although the physiochemical properties of synthetic lime-base C0 2 accep- tors have been well quantified, very little work has been done to es- tablish the degree to which fused CaC0 3 -Ca(0H) 2 mixtures interact with melt container materials. This area of investigation is of utmost im- portance because large quantities of spent acceptor and comminution fines will have to be reconstituted in the large-scale production of fuel gas by the C0 2 acceptor process. Melt containers used for recon- stitition will be in almost continuous contact with CaC0 3 -Ca(0H) 2 melts and, thus, must be constructed of material which can withstand a very rigorous environment. Relatively little work has been done in the area of corrosion by molten salts, and studies that have been done were limited primarily to cor- rosion by molten univalent metal hydroxides and carbonates . *• 4 » » b > - 1 It is difficult to predict the behavior of materials in contact with CaC0 3 -Ca(0H) 2 melts from these studies, however, because of the higher pressures and temperatures required to form melts in lime-base systems. Due to the absence of corrosion data for molten CaC0 3 -Ca(0H) 2 systems, a study was initiated to identify those materials which would function as melt containers. Since the high steam pressures involved preclude the use of mild steels in the application, corrosion- resistant nickel and chromium alloys were involved. Specific areas that were investi- gated are: (1) growth of the oxide layer on alloy steels, (2) weight loss of metal samples in contact with melts, and (3) microstructure of oxidized alloys and oxidation products. 4.2.2 EXPERIMENTAL MATERIALS AND PROCEDURES The following alloys were involved in this work, AISI Types 304, 304L, 309, 310, 316, 316L, 446, Inconel 600, Inconel 601, and Nickel 200. The alloys which were obtained for this work were cold-drawn, centerless- ground, and annealed. The composition of the alloys is presented in Table 4-9. 125 18.6 9.0 1.6 - 18.9 9.3 1.7 _ 22.0 12.0 1.7 _ 24.0 20.0 1.7 - 18.6 12.0 1.7 _ 18.6 13.0 1.7 _ 25.4 - 0.9 - 15.6 72.0 1.0 2.3 20.4 56.0 1.0 2.0 ALLOY TYPE Cr Ni Mn Co AISI 304 30 4L* 309 310 316 316L* 446 Inconel 600 601 Nickel 200 99.1 0.2 0.1 *Low carbon Types 304 and 316, respectively. The balance of material in each alloy is iron. Table 4-9. COMPOSITION OF ALLOYS, EXPRESSED IN WEIGHT PERCENT, INVOLVED IN THIS INVESTIGATION The initial work which was undertaken in this study involved measuring the rate of corrosion of stainless steel samples by establishing the rate at which metal oxide reported to the melt. These experiments were initiated by placing 10 grams of a eutectic mixture of CaC0 3 and Ca(0H) 2 into a 2-inch-diameter cylindrical copper crucible. The cruci- ble was placed into a 1- liter autoclave, and a sample of the material to be tested, fastened to the cover of the autoclave, was lowered into the copper crucible. The autoclave cover was then fastened to the reactor, and the system was heated with an electric furnace. Excess water was added prior to initiation of each corrosion experiment, and excess steam was vented during heating to maintain the pressure at 600 PSI . The sample was maintained in contact with melt at 1200°F for 2 hours after which time the system was cooled, and the sample was removed from the melt. Analysis of the frozen melt was initiated by dissolution of this material in an aqueous solution containing 25 volume percent hydrochloric acid. The acidic solution was then analyzed for iron, nickel, and chromium using atomic absorption or colorimetric techniques. It was found that these analytical approaches did not work well. Although reagent-grade chemicals were used in this investigation, the levels of nickel, iron, and chromium in the CaC0 3 and Ca(0H) 2 used to prepare melts were ex- cessively high. Experimental difficulties were therefore encountered in detecting the small quantities of iron, nickel, and chromium which reported to the melt during each corrosion experiment. Use of this technique was therefore discontinued. 126 Following this work the oxidation of materials was followed by observing the thickness of the oxide on the metal surface as a function of time. Sample preparation was initiated by cutting a 0.75-inch sample from a • 1/2-inch-diameter rod with an Aloxide cutoff wheel using generous amounts of water as coolant. The ends of the samples were then ground to a smooth finish with a 240-grit abrasive belt. Experiments were initiated by placing five samples into a Type 310 stain- less steel container. Oxidized stainless steel containers were used to contain the melt to preclude electrical contact between the samples and the crucible which could cause samples to become anodic and oxidize rapidly. When newly constructed containers were to be used, the cruci- ble was heated to 2000°F for 5 hours in the presence of air to ensure that the surface of the container was oxidized completely. After the samples were placed in the container, 2Q0 grams of a eutectic mixture of CaC0 3 -Ca(0H) 2 were added, and the crucible was placed in a 1- liter autoclave. Excess water was added to the system, the autoclave was covered, and the system was heated to an internal temperature of 1200° F. Excess water was again vented during heating to maintain the pressure at 600 PSI. The samples were maintained under these conditions for a predetermined length of time after which the system was cooled, and samples were removed from the melt. Oxide thickness measurement was effected by mounting samples in a metal- lographic bakelite holder. The sample was then cut so that the edges could be observed along a plane parallel to the major axis of the cylindrical metal sample. The exposed survace was polished with 3, 2, 1, 0, 2/0, 3/0, and 4/0 grit aAl 2 3 polishing papers. Final polishing was effected with 0.05-micron a-Al 2 3 on a polishing wheel. Oxide- layer thickness was established by measuring this quantity optically at a magnification of 500X using a metallograph. Fifty thickness measure- ments were made for each sample. Corrosion rate was also measured by following the weight loss of alloys which had been contacted with CaC0 3-Ca(OH)2 melts. Cylindrical samples were cut from 1/2-inch-diameter rod as described previously. Samples were prepared for weighing by washing the metal thoroughly in distilled water followed by contacting the sample with 200 ML of reagent-grade acetone for 6 hours. The sample was then dried, weighed, recontacted with acetone, and reweighed. This treatment was continued until the sample weight remained constant to within 0.1 milligram. Corrosion experiments were effected by contacting five samples with a eutectic composition melt in an oxidized crucible as discussed previously, Following contact for a predetermined length of time, the autoclave was cooled, and the metal was removed from the container. Contact time was measured from the point at which the internal temperature of the auto- clave reached 1200°F. A heating time of approximately 4 hours was re- quired to reach this temperature. 127 4.2.3 After the samples had been removed from the melt, CaC0 3 and Ca(0H) 2 which remained on the surface of the sample was removed by contacting the alloy for 5 hours with 200 milliliters of solution which contained 0.13 M sodium acetate adjusted to pH 5 with acetic acid. The metal sample's were then washed in distilled water followed by acetone, dried at 150°F for 2 minutes, and reweighed. To establish that corrosion of alloys in the acetate cleaning solution was insignificant, several unoxidized metal samples were contacted with solution for 6 hours and reweighed. It was found that metal samples did not lose weight while in contact with sodium acetate solution. Samples which had been contacted with CaC0 3 -Ca(0H) 2 melts were also sub- jected to microscopic examination to establish if intergranular attack had occurred and to determine the physical soundness of the corrosion products. Samples to be examined in this manner were mounted in bakelite and cut and polished as discussed previously, and the region near the surface was examined. Oxidation products were also removed from the metal surfaces and analyzed by X-ray diffraction to identify the products of corrosion which formed in this system. EXPERIMENTAL RESULTS The results of oxide thickness measurements have been presented in Table 4-10 for contact times of 10, 50, and 150 hours. These results indicate that the oxide layer on these alloys grows very rapidly during the first 5 hours of contact. For longer times the oxide thickness remains rela- tively constant as can be seen by comparison of the 50- and 150-hour data. The oxide layer on Nickel 200 is also significantly thinner than that observed on other alloys. Samples of this alloy which have been exposed to the melt for 200 hours, for example, still exhibit a bright luster. OXIDE THICKNESS (MICRONS) ALLOY TYPE 10 HR 50 HR 150 HR AISI 304 30 4L 309 310 316 316L 446 In cone 1 600 601 27 60 62 31 64 68 18 41 45 11 31 34 27 46 50 35 62 67 18 42 45 16 53 59 23 67 70 Nickel 200 -- -- 30 Table 4-10. OXIDE THICKNESS AS A FUNCTION OF CONTACT TIME 128 The weight loss per unit for several alloys is shown as a function of time in Figures 4-17 through 4-24. In all cases weight loss is a linear function of contact time. Corrosion rates obtained from these data have been summarized in Subsection 4.2.4 of this report. In the case of Type 316 stainless steel the data are not reproducible, and additional work will be required to establish the corrosion rate in this system. Oxidation products of the stainless steels, Types 304 through 446, were also identified by X-ray diffraction and were found to consist of Cr 2 3 and Fe 2 3 . In the case of the Inconel series of alloys, the afore- mentioned oxides and NiCr 2 4 were found. NiO was found on the surface of Nickel 200. The microstructures of Inconel 600, Inconel 601, and Nickel 200 are shown in Figures 4-25 through 4-27, respectively. In the case of Inconel 600, a porous structure is observed on the surface of the alloy after contact with a eutectic composition CaC0 3 -Ca(0H) 2 melt. Fissures are observed in this reaction product as shown by Figure 4-25, and it was found further that the product layer could be detached from the metal surface by application of slight mechanical shock to the metal sample. The reaction product layer is 0.12 CM thick after a contact time of 100 hours . In the case of Inconel 601, the porous structure observed on the surface of Inconel 600 is absent as shown by Figure 4-26. The surface of Inconel 601 is very irregular, however, after the metal has been in contact with melt as shown by this figure. The micros tructure of Nickel 200 after contact with a eutectic-composition melt for 150 hours is shown in Figure 4-27. The surface of this alloy has undergone extensive intergranular attack which causes the grain boundaries to become very visible in this photograph. The depth of penetration of intergranular attack after a 150-hour contact time is approximately 0.05 CM. Microscopic investigation of AISI Types 304, 304L, 309, 310, 316, 316L, and 446 stainless steels demonstrated that intergranular corrosion or development of porosity does not occur when these alloys were contacted with CaC0 3 -Ca(0H) 2 melts for the maximum contact times shown in Figures 4-17 through 4-24. 4.2.4 DISCUSSION OF RESULTS Corrosion of alloys in contact with CaC0 3 -Ca(0H) 2 melts can be understood best in terms of the electrochemical model of oxidation developed by Wagner^ 18 ) and shown schematically in Figure 4-28. Metal oxidation is an electrochemical phenomenon because metal ions consumed in the for- mation of oxide are generated at the anodic metal-oxide interface and 129 20 n3 0) S- C\J CD £ O 4-J en CD 15 - 10 - 5 - 50 100 150 Contact Time (hr) 200 Figure 4-17. WEIGHT LOSS OF TYPE 304 STAINLESS STEEL AS A FUNCTION OF TIME 130 20 CD 5- E CD o> 3 - 2 - 1 - 50 100 150 Contact Time (hr) 200 Figure 4-20. WEIGHT LOSS OF TYPE 310 STAINLESS STEEL AS A FUNCTION OF TIME 133 20 (T3 CD J- < S- C\J CU E Ql a oo E o •r— 15 - 10 50 100 150 Contact Time (hr) 200 Figure 4-21. WEIGHT LOSS OF TYPE 316L STAINLESS STEEL AS A FUNCTION OF TIME 134 (T3 0) t-CSJ d) E Q. O oo E O •— ' 0J 3 " 2 " 1 - 50 100 150 Contact Time (hr) 200 Figure 4-22. WEIGHT LOSS OF TYPE 446 STAINLESS STEEL AS A FUNCTION OF TIME 135 200 < s- cm a> e a. a to E o --> +-> en •r- 150 " 100 50 - 50 100 150 Contact Time (hr) 200 Figure 4-23. WEIGHT LOSS OF INCONEL 601 AS A FUNCTION OF TIME 136 20 CD J- S- CNJ •r- 01 15 " 10 100 150 Contact Time (hr) 200 Figure 4-24. WEIGHT LOSS OF NICKEL 200 AS A FUNCTION OF TIME 137 , » Figure 4-25. FISSURED POROUS REACTION PRODUCT PRESENT ON THE SURFACE OF INCONEL 600 AFTER CONTACT WITH A EUTECTIC COMPOSITION CaCO. Ca(OH) 2 FOR 100 HOURS. UNETCHED SAMPLE AT A MAGNIFICATION* OF 55 TIMES 138 Figure 4-26. IRREGULAR SURFACE OF INCONEL 601 AFTER CONTACT WITH A EUTECTIC COMPOSITION MELT FOR 100 HOURS. UNETCHED SAMPLE AT A MAGNIFICATION OF 165 TIMES 139 Figure 4-27. INTERGRANULAR ATTACK OF NICKEL 200 AFTER CONTACT WITH A EUTECTIC COMPOSITION CaC0 3 -Ca(OH) 2 MELT FOR 150 HOURS. THE SAMPLE WAS POLISHED, ETCHED IN HC1 SATURATED WITH CuCl 2 FOR ONE SECOND, AND REPOLISHED TO REMOVE STRUCTURAL FEATURES WITHIN THE METAL GRAINS. THE MAGNIFICATION IS 135 TIMES 140 Figure 4-28. OXIDATION OF IRON-NICKEL-CHROMIUM ALLOYS IN CONTACT WITH A CaC0 3 -Ca(OH) 2 MELT 141 diffuse with liberated electrons to the outer edge of the oxide layer. The oxide-melt interface is cathodic because electrons are consumed in this region by reaction with the oxidant, probably 0H~ in this system, to form . Oxide ion produced at the oxide-melt interface combines with metal ions in this area to form the new oxide structure. The important role which diffusion plays in the oxidation of metal alloys can be seen from the Wagner model of oxidation. The rate at which re- action occurs in systems when the oxide layer is protective is governed primarily by the diffusion of metal ions to the outer region of the oxide layer. Diffusion of metal ions in the oxide is a relatively slow process, and therefore little degradation of the alloy occurs when the oxide reaction product is sufficiently sound to prevent oxi- dant from coming in direct contact with the metal. This mechanism can be related to the oxidation of alloys in contact with CaC0 3-Ca(0H) 2 melts through the oxide thickness data. Oxide growth in systems in which diffusion through the oxide is important occurs in a very characteristic manner. In the initial stages of oxi- dation when the layer is relatively thin, the rate of growth of the reaction product is high because the diffusion path, i.e., the distance across the oxide, is relatively short. As oxidation continues, the oxide layer become thicker, and the rate of reaction decreases because the diffusion path for metal ions is longer under these conditions. This phenomenon is reflected in the oxide thickness data presented in Table 4-10. As would be expected, the average rate of growth, simply oxide thickness divided by contact time, over the initial 10 hours of contact is much greater than that observed for 50 hours. This indicates that diffusion processes play an important role in this system and that the oxide layer which is formed is protective. This hypothesis is fur- ther substantiated by the fact that microscopic examinations of oxidized stainless steels, AISI Types 304 through 446, indicate that this layer is physically sound, I.e., the oxide is not penetrated by large fissures. In the case of the high-nickel-content alloys, i.e., Inconel 600, Inconel 601, and Nickel 200, some physical degradation of the product layer is observed, and this phenomenon is discussed subsequently in this section of this report. The weight loss data indicate that another phenomenon also assumes im- portance in these systems. Weight loss must be due to removal of oxide from the surface of the alloy, and this occurs through spalling of the oxide at the melt-oxide interface. This phenomenon is called Kofstad breakaway and is confirmed by the fact that minute metal oxide chips are found in frozen melts which have been in contact with metal samples. Spalling of the oxide at the melt-oxide interface has also been shown schematically in Figure 4-28. 142 Weight loss is a linear function of time under these conditions because the rate of weight loss through spalling is simply dependent on the mechanical properties of the oxide layer near the oxide-melt interface. When this layer is relatively thick the mechanical properties are inde- pendent of oxide layer thickness and this phenomenon therefore occurs at a cons tant rate . Interestingly, the initial oxide formed on the metal surface does not appear to be as physically stable as that material which is formed in the later stages of oxidation. This observation is based on the fact that the rate of weight loss is very high in the first few hours of the corrosion experiment for all alloys except Inconel 601. Apparently, the new oxide which forms on the metal surface when the rate is high, i.e., in the initial stages of corrosion, is under more stress than that material which is formed when the rate of oxide growth is lower. Importantly, the fact that discrete particles of oxide are observed in the melt indicates that the oxides exhibit little solubility in the CaC03-Ca(0H)2 melt. Lack of interaction between the melt and oxides which form on the metal surfaces is also confirmed by the fact that the oxidation products on the metal surfaces are not compounds of any of the elements contained in the melt except oxygen. This fact is very impor- tant because in systems in which low melting point oxides are formed, catastrophic corrosion rates are frequently observed. This undesirable phenomenon is not observed in this system, however, but significant dis- solution could occur with the addition of impurities to acceptor melts. With regard to the weight loss data and photomicrographs of the alloys involved in this work, the materials which contain relatively large quantities of nickel, i.e., Nickel 200, Inconel 600, and Inconel 601 exhibit unfavorable corrosion resistance. In the case of Inconel 600, severe porosity and cracking of surface developed after contact with CrC0 3 -Ca(0H) 2 melts for a relatively short length of time as shown by the photomicrograph presented in Figure 4-25. Although no work was undertaken to identify the material which remained on the surface, simi- lar investigations in fused NaOH systems have demonstrated that this phenomenon is caused by the diffusion of chromium and iron from Inconel alloys which leaves a nickel-rich porous matrix on the alloy surface. (16) Microscopic examination demonstrated that the layer of porous material on the surface after contact with melt for 100 hours was 0.12 CM thick which is a rate of penetration of 4 inches per year. Since this rate is excessively high, Inconel 600 is not a satisfactory material for use as a melt container. In the case of Inconel 601, a porous layer of reaction product is not found on the alloy surface. The rate of loss of Inconel 601, however, is 0.39 inches per year as shown by the summary of the corrosion data presented in Table 4-11. This rate of degradation, of course, is ex- tremely high. Likewise, the photomicrograph of this material presented in Figure 4-26 shown that the metal-oxide interface is very irregular which also indicates that the material has been attacked extensively by the melt. 143 The fact that a porous layer of metallic reaction product is not observed on the surface of Inconel 601 is probably due to the fact that this alloy contains 16 percent less nickel than Inconel 600. Because of the lower nickel content of Inconel 601, it would be expected that the nickel-rich residue which would form would not be as dense as in the case of Inconel 600. Lower density of the porous reaction product would certainly make detachment of the product layer more likely. The weight loss data for Inconel 601 are also somewhat unusual. Little corrosion takes place during the first 50 hours of contact after which time the rate of loss increases dramatically. This phenomenon may also be related to the development of the unstable metallic reaction product. Significant loss of material probably does not occur until porosity be- comes fairly well developed; hence, a high rate of weight loss is not observed until oxidation of the alloy has occurred for approximately 50 hours . In the case of Nickel 200, the rate of weight loss was not unusually high, but microscopic examination revealed extensive intergranular cor- rosion of this material as shown in Figure 4-27. This phenomenon has been observed by other investigators for relatively pure nickel in contact with fused NaOH. ( 13 ) Because intergranular attack could result in ex- tensive degradation of Nickel 200 for prolonged contact times, the use of this material as a melt container cannot be recommended. The stainless steels, Types 304 through 446, exhibit much more favorable corrosion resistance as shown by the corrosion data presented in Table 4-11. Good correlation is observed between chromium content and corros- ion rate for these materials. Types 304, 304L, and 316L, for example, contain approximately 19 percent chromium and the average corrosion rate for these materials is 0.017 inches per year. The alloys which contain 2 2 to 25 percent chromium, Types 309, 310, and 446, corrode at a rate which is 65 percent lower, or 0.006 inches per year. The effect of chromium content on corrosion rate has, of course, been demonstrated previously in the oxidation of alloys in contact with air. 1 - - 1 Chromium has been shown to improve oxidation resistance by decreasing the rate of ion and electron diffusion in the protective oxide layer. 144 CORROSION RATE COST OF ALLOY (inches per year) (doL lars /pound) 0.015 1.07 0.017 1.14 0.006 1.47 0.006 1.84 CD 1.37 0.020 1.44 0.006 (3) (2) 3.48 0.39 3.18 ALLOY AISI 304 304L 309 310 316 316L 446 Inconel 600 601 Nickel 200 0.011 3.97 (1) Additional data required to establish corrosion rate. (2) Extreme porosity of metal surface developed. (3) Not currently produced as plate. Table 4-11. CORROSION RATE OF SEVERAL ALLOYS IN CONTACT WITH EUTECTIC COMPOSITION CaC0 3 -Ca(0H) 2 MELTS AT 1200° F, 600 -PSI STEAM PRESSURE, AND THE COST OF 2- INCH PLATE IN 1-TON QUANTITIES (MARCH 19 75 PRICES) It is interesting to note that no appreciable difference exists between the corrosion rates of Type 304 stainless steel and its low carbon equi- valent, Type 304L. The fact that essentially equal corrosion rates are observed for these alloys is because low carbon content imparts corrosion resistance by preventing intergranular attack. ( 2 ] Since this phenomenon is not observed in the corrosion of Type 304 stainless steel in this system, a large difference in the rates of corrosion of Type 304 and its low carbon equivalent would not be expected. Likewise, Types 316 and 316L probably also corrode at similar rates. Additional data are required to evaulate the performance of Type 316 stainless steel before accurate comparision of the corrosion rates of these materials can be made. From the data collected to this point, the most suitable alloys for use as melt containers, therefore, are the high- chromium alloys, Types 309, 310, and 446. As shown by the cost of 2-inch plate presented in Table 4-11, Type 309 is the mose economical alloy which is currently in pro- duction. At the corrosion rates measured, a loss of 0.1 inch of material over a 20-year service life of the container could be expected. 145 The corrosion resistance of relatively high -chromium alloys is likewise reflected in experience which has been gained in the use of Type 310 stainless steel melt containers in this work. Excessive degradation of containers which have undergone extensive thermal cycling in the presence of melt has not been observed. 4.2.5 CONCLUSIONS (1) Weight-loss measurements indicate that the stainless steels corrode primarily through spalling of the protective oxide layer. (2) Because corrosion rate decreases as the chromium content of the stainless steels increases, high chromium alloys, i.e., Types 30 9, 310, and 446, are best suited as melt containers. (3) The alloys which contain relatively large amounts of nickel, i.e., Inconel 600, Inconel 601, and Nickel 200, would not function well as melt containers because of development of surface porosity or because of high corrosion rates. 146 SECTION 5 CARBONATE ROCK RESOURCES STUDIES This section contains two reports: "Investigations of Carbonate Rock Resources in the Logan, Montana, Area," dated January 6, 1978, and "Preliminary Resource Study of Carbonate Rocks Available for Lignite Gasification in Central Montana, Northern Wyoming, and Western South Dakota," dated December 15, 19 72, submitted by the South Dakota School of Mines and Technology. 147 5.1 INVESTIGATION OF CARBONATE ROCK RESOURCES IN THE LOGAN, MONTANA, AREA 5.1.1 INTRODUCTION This study was undertaken with the purpose of determining the occurrence and availability of carbonate rocks in the Three Forks-Logan, Montana, area. Carbonate rocks of Devonian and Mississippian age were measured, described, and sampled to determine if these rocks offer suitable raw material for the carbonate acceptor process being developed for use in lignite gasification. The Three Forks-Logan area was selected for study because exposures are close to the main line of the Burlington Northern Railroad and Interstate 90 for highway travel, and they offer quarry sites with large, easily quarriable limestone and dolomite. The large Ideal Cement Company quarry is a few miles west of the town of Logan at Trident, Montana. The study area is approximately 500 rail miles from Williston, North Dakota, the eastern edge of the lignite area. The study concentrated on the Lodgepole and Mission Canyon Formations of the Madison Group and the Jefferson Formation of Devonian age. The Madison Group is almost entirely limestone and the Jefferson Formation is dolomite or dolomitic limestone. Either formation appears to offer the potentially large reserves of rock necessary to supply a large lig- nite gasification acceptor rock need. 5.1.2 METHOD OF STUDY The initial work was to review the literature pertaining to carbonate rock available in the area in close proximity to the existing railroad lines. In addition, the personal knowledge of the staff of the South Dakota School of Mines and Technology Department of Geology and Geolo- gical Engineering was utilized in planning field work. Because there are relatively massive carbonate sections, a sample was collected from every 5-foot thickness interval. Detailed field descriptions were made for every 5 feet and any distinct lithologic change within the 5- foot interval was noted. At 25-foot intervals, or when a major lithologic change was apparent, a larger sample was collected for additional lab- oratory tests. Ninety-nine 5-foot samples of Jefferson dolomite were collected. Seventy-four samples of Lodgepole limestone were collected at 5-foot intervals and 160 samples of Mission Canyon limestone were collected at 5- foot intervals. In the laboratory all samples were examined with the binocular micro- scope and detailed sections showing thickness, composition, and texture were constructed based on field and laboratory studies. A number of thin sections were made of representative samples of the Jefferson dolomite. These thin sections made the differentiation of calcite from dolomite much easier and resulted in more reliable Ca:Mg 148 ratio estimates. The dolomite contents presented in the sections are based on petrographic or petrologic work plus chemical analyses made to confirm petrographic or petrologic analysis. The large samples collected from the outcrop were tested for crushing characteristics or toughness in a drop hammer device designed for an earlier phase of the project. Los Angeles abrasion tests were made to compare results of the drop hammer test. The standard for comparison was the Tymochtee dolomite which has been used in tests of the CO2 Gasification Process by Consol (now Conoco Coal Development Company) . 5.1.3 GEOLOGY OF THE THREE FORKS AREA, MONTANA The Three Forks of the Missouri River mark the junction of the Madison, Jefferson, and Gallatin Rivers. The townsite of Logan, Montana, serves as a division point for the main line of the Burlington Northern Railroad as well as the Milwaukee Railroad which goes north via the Missouri River Canyon. The towns of Three Forks, Logan, and Trident are situated in a basin formed by the junction of the three rivers. This basin is bounded on the east by the Bridger Range, to the south by the Madison Range, and to the southwest by the Tobacco Root Mountains, also known as the Jeffer- son Range, and to the northwest and north by the much more subdued Horse Shoe Hills. Drainage is north via the Missouri River. The stratigraphic succession consists of rocks of Late Precambrian to Recent in age. All systems except Ordovician, Silurian, and Triassic are represented. The late Precambrian age arkoses are overlain discon- formably by about 1500 feet of Middle and Late Cambrian marine strata. The Cambrian is dis conformably overlain by the marine Maywood Formation of Devonian age. Conformably overlying the Maywood is the marine Jeffer- son Formation, 515 feet thick, of late Devonian age. Conformably above the Jefferson is the marine Three Forks Shale, of Late Devonian and Mississippian age. Succeeding the Three Forks are the rocks of the Mississippian Madison Group, The Lodgepole Formation, and the Mission Canyon Formation. The Madison Group is overlain by the late Mississippian Amsden Formation and the Pennsylvanian age Quadrant Formation. The Permian is represented by the marine Phosphoria Formation. Mesozoic, Tertiary, and Recent age rocks are poorly exposed and were not studied or examined in this eval- uation. The structural history is complex. Most of this folding and faulting was in response to compression along north-south axes, very late in the Cretaceous to early Tertiary. 149 5.1.4.2.2 Mission Canyon Formation. The Mission Canyon limestone conformably over- lies the Lodgepole Formation. The contact of the Lodgepole Formation with the overlying Mission Canyon Formation occurs at the base of the first massive limestone cliff above the base of the Mississippian section. Mission Canyon beds range from 2 feet to 6 feet in thickness in contrast to the Lodgepole beds which are 2 to 8 inches thick. At the contact the beds strike N 39 E , dip 33° NW. The limestones range from dark gray-brown near the base, to gray-black 200 feet above the base, to gray or light brown near the top of the mea- sured section. A total of 450 feet of Mission Canyon limestone was measured, described, and sampled. This did not reach to Mission Canyon- Amsden contact. This portion of the section was covered. The Mission Canyon limestone is reported to be the most important source of high-purity limestone in Montana. It characteristically contains more than 95 percent CaC0 3 . Our chemical analysis supports this statement, for the Logan, Montana, area. The textures of the Mission Canyon are quite varied, ranging from very fine microcrystalline, coarse bio-sparites; with local breccias, oolitic pellets. Though fossil debris is not as common as in the Lodgepole For- mation, relicts of brachiopods , corals, bryozoan, and unidentified mater- ials were noted. Sparry calcine was common as a replacement material associated with mega-fossi Is . 5.1.5 DROP HAMMER TESTS Several samples were selected for testing by the drop hammer method. The selection was based on observable texture and composition, with the aim being to test each of the common textures of carbonate present. Where the texture and composition of the carbonates was not observed to be sig- nificantly different, the samples were selected at intervals. The test consists of applying a standard force to a standard sample and measuring the size distribution of the resulting fragments. This is ac- complished with the aid of a specially designed and built apparatus com- prised of a heavy-gauge steel tube 3 inches inside diameter, 6 feet long, resting in a short receiver (this is the "guide" or "barrel") and the hammer itself which is a steel cylinder slightly less than 3 inches in diameter, weighing 25 pounds, which is free to slide inside the barrel. The test procedure is to cut the sample into four or more 1-inch cubes, place them one at a time in the receiver, and drop the hammer from a height of 4 feet. This is repeated for each of the cubes in the sample; then the crushed cubes are combined in a 6 sieve-plus pan nest and sieved for 10 minutes in a mechanical siever. The contents of each pan are then weighed to the nearest 0.01 gram and converted to a percentage of the total. This data is presented in histograph form in Figures 5-1 through 5-3. 5.1.4 CARBONATE ROCKS OF DEVONIAN AND MISSISSIPPIAN AGE 5.1.4.1 Devonian - Jefferson Dolomite The measured and described section is located in a gulch and on bluffs on the north site of the Gallatin River, northeast of Logan, in SE 1/4, sec. 25, T.2N., R.2E., Gallatin County, Montana. Beds strike N 35° to 40° E and dip 40° to 44° NW. The Jefferson Formation is largely a mottled dolomite, with lesser amounts of dolomitic limestone, and small amounts of limestone and claystone. It is dark colored, varying from black to brown and gray. Textures vary from very fine microcrystalline carbonate muds to fine crystalline to small amounts of medium- to- coarse crystalline materials. The dolomites quite characteristically show a sugary or sucrosic texture. Fossil debris de- rived from crinoid stems, stromatoparoids , brachiopods, and locally abun- dant Amphipora. Sparry calcine commonly is present in areas of abundant mega- fossils . Very small amounts of rusty hematitic material was observed on joint or fracture faces. As previously noted, the Jefferson directly overlies the Maywood Formation and in turn is overlain by the Three Forks Formation. A total of 515 feet was measured at the Logan, Montana outcrop. A detailed description log at 5 feet per inch is included as Appendix A in Book 2 of this volume. 5.1.4.2 Mississippian - Madison Group 5.1.4.2.1 Lodgepole Formation. The Lodgepole limestone consists of thin- bedded limestone interbedded with small amounts of chert and shale. Its com- position is suitable for cement manufacture, and the Ideal Cement Company is quarrying limesone for that purpose at Trident, Montana. The orange calcareous siltstone and sandstone at the top of the Three Forks were separated as the Sappington Formation of Mississippian age by Berry. The orange siltstone is gradational with the underlying green shale and is in sharp contact with the overlying Lodgepole. A total of 730 feet of Lodgepole limestone was measured, sampled, and described from rocks cropping out in the SE 1/4 sec. 25, T. 2 N., R. 2 E., Gallatin County, Montana. This is on the north side of the Gallatin River just northeast of the town of Logan, Montana. The beds strike N 41° E and dip 46° NW. The limestones are dark gray to black grading to gray brown and gray near the top. Beds are 2 to 6 inches thick throughout much of the lower part of the section with a few beds 4 feet thick near the top. The limestones vary in texture with most of the lower section being microcrystalline to very fine grained near the base, with argillaceous streaks showing as laminae on weathered surfaces. The textures become fine- to-medium cry- stalline near the top. Sparry calcine is common along fractures and replacing fossil materials. Brachiopods, crinoids, corals, bryozoan, and other fossil debris is common throughout the Lodgepole. Occasional oolite beds are noted. 150 50 40 30 % 20 10 . OS 6* CO fe5 CO ££ N 3 h 9 fc~ *H SO 2 46 m % 50 40 _ 30 20 10 Sample 240-248 feet 50 40 30 > 20 10 fcjN o 5^ r-i r- CM r>° fc p CO n CO CO 50 40 — ._ 30 CM _E& _^. V 4.7 20 .83 4 9 20 .246 mm. 60 mesh Sample 270-275 feet 20 10 m & CO CM 94 2 4.7 4 2.0 .83 .50 .246 mm. 9 20 32 60 mesh 9.4 2 47 4 2.0 .83 9 20 .50 .246 mm. 32 60 mesh Sample 275-280 feet Sample 285-290 feet Jefferson Formation Logan, Montana Figure 5-2. PERCENTAGE OF FINES PRODUCED FOR FOUR TYPE SAMPLES 153 40 30 20 10 rfi fe^ CM in i—i tfi U5 c: ft 50 40 • 30 20 10 fey 6^ CO 6^ oo ^ tf ■5f "* tt 9.4 4.7 2.0 .83 .50 .246 mm. 2 4 9 20 32 60 mesh Lodgepol e samp le 3 feet 9.4 4,7 2.0 .83 4 9 20 f0 .246 mm. 60 mesh Lodgepole sample 760 feet 50 40 JO 20 10 fe§ in feS CM fe^ m in CO &3 IT5 fcP n CO 50 40 . 30 e 9 a • 0, 20 10 r> 9.4 4.7 2.0 .83 .50 .246 mm. 2 4 9 20 32 60 mesh Mission Canyon Sample 1160 feet 9.4 4.7 2.0 .83 .50 2 4 9 20 32 .246 mm. 60 mesh Madison Limestone, Lodgepole member, sample from 3 feet. Logan, Montana Madison Limestone, Mission Canyon member, sample from 1160 feet. Logan, Montana Madison Limestone, Lodgepole member, sample from 760 to 765 feet. Logan, Montana Figure 5-3. PERCENTAGE OF FINES PRODUCED FOR MADISON LIMESTONE 154 Three features are used to compare one type of carbonate with another; they are: (1) The percentage of "fines" produced, which is the total percentage of material passing a 20 mesh sieve (2) The peak size range, which in this test is usually the 2X4 mesh fraction (3) The percentage of material in the 9 X 20 mesh fraction Evaluation of the histograms is somewhat subjective as no totally satis- factory numerical method of weighting the three fractors has been devised. However, the values so far determined have not been too close to each other, so evaluation has been fairly easy and the system does allow for changes in criteria. A method of comparison using the Tymochtee Dolomite as the standard has been used. This has been called "toughness" for lack of a better term. "Toughness" is the percentage of material passing the 20 mesh screen as compared to the Tymochtee. This system is of value as a method of comparing carbonates at a gross level only, and is not reliable for the final evaulation of a carbonate. For final evaluation, all aspects of the drop hammer results as well as the Los Angeles abra- sion tests and compression tests must be considered. For the purposes of eliminating totally unsuitable rocks the following system has been devised: CI) "Toughness" less than 10 percent is tougher than Tymochtee Dolomite (2) "Toughness" 10 to 10 percent is comparable with Tymochtee Dolomite (3) "Toughness" 15 to 20 percent is weaker than the Tymochtee Dolomite (4) "Toughness" greater than 20 percent is much weaker than Tymochtee Dolomite 5.1.6 CHEMICAL AND MINERALOGIC DETERMINATIONS 5.1.6.1 Calcium-Magnesium Ratios and Acid Insolubles Calcium-magnesium ratios were determined for 100-foot intervals of the Jefferson and Madison Formations using the EDTA wet chemical method. Acid insolubles were a byproduct of putting the carbonates into solution. The results of this work are presented in Table 5-1. The procedure followed was to crush samples taken at 5- foot intervals (where possible, refer to stratigraphic sections) and combine 1 gram of each sample to form a 20-gram sample representing 100 feet of section. Acid insolubles are those minerals which were insoluble in warm concen- trated hydrochloric acid. The insoluble residue was analyzed by X-ray diffraction. CaCC>3 and MgCC>3 were calculated as weight percent and are shown as adding up to 100 percent. This is not actually the case, as other elements, such as iron, manganese, potassium, and sodium, are dissolved in the acid. Cal- cium and magnesium are by far the most abundant cations and the significant ones for this project. Samples are available should analysis for other elements be required. 155 Tests show that, with the exception of 100 to 200 feet, the Jefferson Formation is around two-thirds dolomite, while the Madison Formation is almost pure limestone. CALCIUM- MAGNESIUM RATIOS DEPTH (Feet) CALCIUM (Percent) Devonian Jefferson Formation to 100 72 100 to 200 85 200 to 300 68 300 to 400 57 400 to 495 61 MAGNESIUM (Percent) 28 15 32 43 39 Mississippian Madison Group to 100 9 8 100 to 200 95 200 to 300 100 300 to 400 99.5 400 to 500 100 500 to 600 97 600 to 700 98 700 to 800 100 800 to 900 94 900 to 1000 96 1000 to 1100 - 98 1100 to 1180 98 2 5 0.5 3 2 6 4 2 2 Values are accurate to within one percentage point. Table 5-1. CHEMICAL ANALYSES OF THE JEFFERSON FORMATION AND MADISON GROUP, LOGAN, MONTANA 156 5.1.6.2 X-Ray Analysis of Acid Insoluble Residues 5.1.6.2.1 Devonian Jefferson Formation. X-ray diffraction analysis of oriented silt sized samples indicates that the composition of the residue is dominantly illite/mica with about half as much potassium feldspar. Quartz is a relatively minor constituent, being present but barely noticeable in the X-ray readout. Chlorite is present in approximately the same abundance as quartz. Illite and mica have been lumped together because the diffraction patterns of illite, muscovite, biotite, and phlogopite are virtually indistinguishable by this technique, and the difference is not important enough to this project to warrant further tests. Mica flakes, probably muscovite, were observed visually in the residue, although the conditions under which deposition of the forma- tion took place suggest that mica is subordinate to illite. There was no significant change in composition of the five insoluble residue samples noted and the results of all determinations are given in Table 5-2. 5.1.6.2.2 Mississippian Madison Formation. X-ray diffraction analysis of oriented silt sized samples indicates that the composition of the residue is dom- inantly illite/mica with about half as much potassium feldspar. Quartz is a relatively minor constituent, being present but barely noticeable in the X-ray readout. Chlorite is present in approximately the same abundance as quartz. Illite and mica have been lumped together because the diffraction patterns of illite, muscovite, biotite, and phlogopite are virtually indistinguishable by this technique, and the difference is not important enough to this project to warrant further tests. Mica flakes, probably muscovite, were observed visually in the residue, al- though the geologic conditions under which deposition of the formation took place suggest that mica is subordinate to illite. The 900-to-1000-foot and 1000-to- 1100- foot samples had one unidentified peak each. There was no significant change in composition of the 12 insoluble resi- due samples noted. 157 RESULTS OF ACID INSOLUBLE STUDIES DEPTH (Feet) INSOLUBLE RESIDUE (Percent) Devonian Jefferson Formation to 100 4.95 100 to 200 1.40 200 to 300 5.05 300 to 400 6.20 400 to 495 2.15 Mississippian Madison Formation to 100 11.25 100 to 200 8.29 200 to 300 7.85 300 to 400 15 400 to 500 8.43 500 to 600 4 600 to 700 4.71 700 to 800 1.9 800 to 900 4.08 900 to 1000 3.94 1000 to 1100 0.81 1100 to 1180 0.92 Table 5-2. X-RAY ANALYSES OF ACID-INSOLUBLE RESIDUES 158 5.1.7 SUMMARY AND GENERAL CONCLUSIONS (1) Two types of carbonate rock are present in the Logan-Three Forks area. One a pure limestone and the other a 60 -percent -magnesium, 40-percent-calcium carbonate rock. Both are available in large quantity. (2) If the pure limestone is selected, fine material produced during crushing could possibly be sold for cement production to Ideal Cement Company at Trident, Montana. It might be possible to pur- chase limestone from Ideal Cement Company. (3) Transportation by rail to a gasification plant site, if located in the eastern Montana/western North Dakota area would be via Burlington Northern Railroad direct to the area of intended use. (4) Acquisition of quarry sites would be from private landholders rather than state or federal agencies. (5) The area is well suited, topographically, for quarry development. (6) Adequate water is available for washing, dust abatement, and other uses from the Gallatin River. 5.1.8 BIBLIOGRAPHY Further information regarding the geology of the Logan, Montana, area may be found in the following sources: (1) Baars, D. L., "Devonian System," Geologic Atlas of the Rocky Moun- tain Area , Mallory, W. W. (ed.) Rocky Mountain Association of Geologists, 1972. (2) Berry, G. W. , "Stratigraphy and Structure at Three Forks, Montana," Bulletin of the GSA , Volume 54, 1943, pp. 1-30. (3) Craig, L. G. , "Mississippian System," Geologic Atlas of the Rocky Mountain Area , Mallory, W. W. (ed.) Rocky Mountain Association of Geologists, 19 72. (4) McMannis, W. J., "Devonian Stratigraphy between Three Forks, Montana and Yellowstone Park," 13th Annual Field Conference Guidebook , Hansen and McKeever (eds.), Billings Geological Society, 1962. (5) Peale, A. C. , "The Paleozoic Section in the Vicinity of Three Forks, Montana," USGS Bulletin , 110, 1893. (6) Robinson, G. D. , "Origin and Development of the Three Forks Basin, Montana," GSA Bulletin , Volume 72, 1961, pp. 1003-1014. (7) Robinson, G. D. , "Geology of the Three Forks Quadrangle, Montana," U.S.G.S. Professional Paper 370, 143 pages, 1963. 159 (8) Sandberg, C. A., "Stratigraphic section of the type Three Forks and Jefferson Formations at Logan, Montana," 13th Annual Field Confer- ence Guidebook , Hansen and McKeever (eds.), Billings Geological Society, 1962. (9) Sandberg, C. A., "Nomenclature and Correlation of Lithologic Subr divisions of the Jefferson and Three Forks Formations of Southern Montana and Northern Wyoming," U.S.G.S. Bulletin , 1194-N, 1965, pp. N1-N18. (10) Sandberg, C. A. and Maple, W. J., "Devonian of the Northern Rocky Mountains and Plains," International Symposium on the Devonian System, Calgary , Oswald, D. H. (ed.), Alberta Society of Petroleum Geologists, 1967, pp. 843-877. (11) Sloss, L. L. and Larid, W. M. , "Devonian System in Central and Northwestern Montana," AAPG Bulletin , Volume 31, 1974, pp. 1404- 14 30. (12) Smith, Donald L., "Stratigraphy and Carbonate Petroloty of the Mississippian Lodgepole Formation in Central Montana," Univ. of Montana, Ph.D. thesis, 1972. 160 LlbNllb bA ^ d WESTERN SOUTH DAKOTA 5 2 1 INTRODUCTION This study was undertaken with the purpose of making J a pre U-ln^ ger- mination of the distribution, quantity, physica jur engt of characteristics, -<> "Ui™^"- n< Scutl f Dakota, Wyoming, and carbonate rocks exposed in certain P«" Acceptor Process Montana which are potential ""»* £* "°*J d cit£ South Dakota. A ^rdSn^rr^estrnTortrStfrd e-astem Montana. Potential carbonate-producing areas closest to *e lignite production are in the Little Rocky Mountains £« Sn^fcuem of northern Wyoming; Mountains of central Montana; the Big Horn J*™*^ ted on the and the Black ""^"/^tssisfip^i^ ag Madison Formation and the massive carbonates of the Mississippi^ g MU i va i e nts in these mountain Ordovician age Big Horn Dolomite and their equivalents ranges. A limited amount of work was done on the Pennsy * Amsden Formation in the Big Snowy Momt ^ m ™y°* r es 5 4 and 5-5) 7 •„ «** Riark Hills The accompanying maps (^ig ureb ° H ^ tZ Thelutcrop distribution of these carbonates in the study area. 5 2 2 METHODS OF STUDY The first step in this study « -u > assemble a ; ^*££™«TZlr possible about areas of P°"ntia carbonate r P nformation came fr0 m proximity to the lignite Production areas ■ lh ^ s m ub and the South Dakota School of Mines and '««^* f J the Ge / logy personal knowledge of the authors and other ambers o s Department. Approximately 200 man-days «e« P »' n ™ rela . reconnaissance work and measuring ^Vh^i^ measured a sample was tively massive carbonate .sections were b « "asured. ^ ^^ £ ^foot^er^ rw n'a^or ;-o r gicaian f e a Was apparent, labeled and stored for later work. sr=MsrJK=»." -^ SEES "" detailed sections showing thicknesses, composition and texture were constructed based on field and laboratory notes. Since most or 161 '-t_l_ >- ", /' \ i =^^-rr r ^^l «sp^--r-...r.. IL ]-— - -> s, ^ j K © n £CT^ : *t/sZ»r~\ i , * , ^i M "" M !: -T- H OT^jQCVI i A -^ 'I/O N * ° K*^ t'Ti-'A tl, "lK-V'"Ti AT"- A i CARBONATE DISTRIBUTION IN THE NORTHERN GREAT PLAINS AND ROCKY MOUNTAINS lllltmi 1 \ CASS "\ IIINII H i I / \ /- l i._fc s -*J r « • " » j-. — -'.^ v, i BC,NT os " ° IC " Y !.««« ;' ."•••¥ *v.„j ( "•■•"•Jl „„,„ ■--.- i i i IHICMLAWON ICAUPMui " C '"'■MM | \ r 1 J o e w c r ValwortiJ [Diiiiti .i . ( i 4 "T « : I IIIKC* ;«»MST»ON9 v I J . I / POTTCA ■ FIIIU ! 1 | , | N „ I icoomoToit ; < ,'r.r I 8 « " S O H ' ' c ' o o «■ I -J" C A D C I $ U L L T I "rot 1 i r HAMLIN ] O /% ii ---^— . t ;""" fc l HAND i^ S y T | H V-D-iA! K ;.9 A0 T A j " « » « N | " ^ «INQSBU« ._.,- — -t I 1.1 ! ' " • ! ''. N!s" « y" — OJ lUfFALOl JC«»ULO. fCHULI |A«"0«» tv t- i i - --" ,. . « f ;" oot " j .M L >- « I f t B I IIIMII . S »IN«lii , T p»W»»l KCCOOK »IMNtM»M» ; i r~* __i _i \ |MUTCMINSO«; T U»Mtl< ' ,J ' aoiaoni 'sHtm.it itnC^ limcoi" 1 , """\ H».11C CLATl _ .. W ■ \. v ss "•^\ V Outllntd oreos show corbonoit outcrop Arias of thick, continuous hgmtt b«ds mors than 30 tnchAs thick EXPLANATION Slatt boundary County boundary Railroads jfi?) Locotion of Surfocs Stction ScoU I 2.500,000 25 jo 75 100 128 180 178 200 MilAS LEGEND Miction Canyon - MC Brown ■ Canyon ' B C Alntha Bsnch • A B Oayton Stction Day Spr.ng CrstS • SC Dor« Canyon • OC «h.l»»ood Crssk • WhC ■iisa'fish Canyon • SpC *hit««ood Dolo Typs • WT Rnoods \ Ranch • R R I ij-urc 5-4. CilINL-UAL INDEX MAP 162 Figure 5-5 CARBONATE DISTRIBUTION, TRANS- PORTATION, AND OWNERSHIP MAP (Sheet 1 of 2, North Half) 163 Figure 5-5. CARBONATE DISTRIBUTION, TRANS- PORTATION, AND OWNERSHIP MAP (Sheet 1 of 2, North Half) 163 Figure 5-5. CARBONATE DISTRIBUTION, TRANS- PORTATION, AND OWNERSHIP MAP (Sheet 2 of 2, South Half) 164 -T3^ Outliood orsos »ho» outcrop* of tht Moditor. Limmton, and ,f. tqu.voknt. In .om, orw. othor rock typ*. art mclutM du« to lack of dataikd mopping \&^\ *'•« of rhieh. cont 30 inch*, thick i.grm* btdt mora thai Figure 5-5. CARBONATE DISTRIBUTION, TRANS- PORTATION, AND OWNERSHIP MAP (Sheet 2 of 2, South Half) 164 Black Hills sections were measured and sampled in the earliest part of the field work, there was time to have about 300 thin sections made and stained in order to distinguish calcite from dolomite. Results of the petrographic study of these thin sections are incorporated in the sections constructed for the Black Hills area. The dolomite contents presented in all of the sections are based on petrographic or petrologic work. Selected samples largely from the Black Hills sections were prepared for analysis of calcium, magnesium, and acid insoluble content by wet chemi- cal methods by the SDSMT Engineering and Mining Experiment Station. These results are presented in Subsection 5.2.5. Large samples collected from surface sections were tested for crushing strength or toughness in a drop hammer device designed especially for this project when it became evident that it would be impossible to hand carry the large specimens required for conventional analytical techniques from the more inaccessible outcrops. Also a limited suite of standard rock mechanics tests were run on three selected rock types in order to compare with the results of the drop hammer tests and to give a general idea of the elastic parameters of the carbonate rocks being studies. Comparable tests were also conducted on the Tymochtee Dolomite which has been successfully used in bench-scale tests of the C0 2 Acceptor Process by Consol (now Conoco Coal Development Company) . Detailed descriptions of laboratory procedures followed are given in the individual subsections of this report. Ownership, economic, and transport data are based on maps, rate schedules, and inquiries and are as current as possible as of the date of this report (December 19 72). As research progressed, an extensive bibliography of carbonate rocks was assembled and is included as Appendix B in Book 2 of this volume. 5.2.3 GENERAL DISTRIBUTION OF THE CARBONATE ROCKS 5.2.3.1 Central Montana Thickness of the Madison Group (Mississippi an) carbonate rocks in the central Montana area ranges from 1600 + feet in southern Fergus County to 700 feet in northern Blaine County. General thinning is south to north, and nowhere is the thickness less than 650 feet. The Madison Group in the study area consists essentially of the Lodgepole Formation overlain by the Mission Canyon Formation. The strata overlie Devonian beds throughout the north half of Montana and central Montana except in local areas along the axis of the Cedar Creek anticline, where a pro- nounced unconformity exists and the Mississippi an rocks rest on Silurian or Ordovician. The Madison Group is overlain by Upper Carboniferous Big Snowy-Amsden Groups. Thickness pattern of the Mississippian is strongly influenced by pre-Middle Jurassic erosion throughout northern Montana. 165 Best outcrops are the upturned flanks of the isolated mountain groups- - the Little Rocky Mountains around Landusky, Montana - - and the Big Snowy Mountains about 20 miles south of Lewistown, Montana. Outcrops in the Little Rocky Mountains stand as arcuate hogbacks circling the domal structure of the central core. Dips on the lower flanks range from 20 to 30 degrees, gradually steepening, and then flattening out on the crest. Numerous deep canyons cutting the Madison Group radiate from the central core. Some of these canyons are accessible by road for quarry purposes. Logistics are, however, hampered by the fact that many of the outcrops are located on the Fort Belknap Indian Reservation; and that Landusky is 45 miles from Roy, Montana, the nearest railhead on the Milwaukee Road. The nearest railhead on the Northern Pacific is at Harlem, Montana, 6 miles to the north of Landusky. (See Figure 5-5.) In the Big Snowy Mountains Madison Group strata crop out west of Becket, Montana along the South Fork of McDonald Creek. The strata, with a general north dip, are exposed on the steep walls of the canyon. Additional outcrops are exposed on the south flank of the Big Snowy Mountains, especially in Swimming Woman Canyon where a creek cuts through a 20- to- 30 -degree dip slope section of hogback ridges. Both of these sections are accessible by public and forest service roads; and are near facilities of the Milwaukee and Great Northern Railroads. Ordovician age rocks are not present in the Big Snowy Mountain area. But a section of Red River Dolomite is exposed in the Little Rockies area. This section consists of some 275 feet of shaly, cherty dolomite but the area is also relatively inaccessible. Another section of the Red River was studied on a high ridge in the Brown's Canyon area in Sections 11 and 14, T26N, R24E, Blaine County, Montana. To construct an access road into the area would probably cost upwards to $100,000. This seems prohibitive for such a project. A large portion of the section is covered. 5.2.3.2 Big Horn-Pryor Mountains, Wyoming-Montana The Madison Group (Mississippi an) carbonates generally thicken from 650 feet in the central Big Horn Mountains of north central Wyoming to more than 800 feet in the Pryor Mountains in south central Montana. The strata conformably overlie rocks of Devonian age throughout the area. The Madison Group is overlain by the Upper Carboniferous Sacajawea-Amsden formations or equivalents. In northern Wyoming pre-Pennsylvanian erosion has affected the Mississippian surface. Channels (5 to 35 feet thick) commonly contain large angular fragments of limestone and chert in a quartz sand matrix. Steep to gently dipping strata of the Madison Group crop out along the entire flank of the northern Big Horn Mountains. Numerous canyons cut the Mississippian in the area around and to the north of Sheridan, 166 Wyoming, all within easy access of the Burlington Railroad which follows the east margin of the mountain front north from Sheridan (See Figure 5-5) Madison Group strata, gently to steeply dipping, crop out on the flanks and over the crest of the Pryor Mountains. Access is hampered in a number of places by poor roads and at the present time access is pro- hibited by the Crow Indians. Hence, the area was not studied in any detail. However, there is a good outcrop and quarry site on the west flank of the Mountains at an abandoned lime kiln located on a country road 10 miles due north of Warren, Montana, the nearest railhead on the Burlington Railroad. On the east flank of the Pryors the closest rail spur is at St. Xavier, Montana, about 25 miles from the outcrops. 5.2.3.3 Black Hills Area Thickness of the Madison Group outcrops in the Black Hills increases from about 250 feet in the south to about 600 feet in the north. The carbonates crop out in a band around the flanks of the Hills, which widens from about 2 miles in the southeast to 5 miles in the northeast. On the west flank the band is considerably wider, averaging about 5 miles, and in places, increasing to 11 or 12 miles. In the Black Hills the Devonian Englewood Formation underlies and grades into the Madison locally called Pahasapa Limestone in the surface sections. The Paha- sapa is in turn overlain by the Minnelusa Formation. The gently dipping strata of the Madison, arched by the central dome of the Black Hills, are exposed in a number of canyons around the periphery of the Hills. Notable among these is the exposure in Spearfish Canyon on the northwest flank of the Hills, where the most complete section is exposed. Additional good exposures include Whitewood Creek Canyon, and Boxelder Canyon. All of these are within easy access of the high- way, although steepness of the canyon walls in some instances may pose a problem in the selection of quarry sites. However, with the number of sections and the abundance of outcrops available, no problem should be encountered in finding suitable material at a convenient place to quarry . The western plateau area of the Hills offers many sites for relatively easy quarrying, but transportation would be a real problem. Ordovician age Whitewood Dolomite crops out in the northern Black Hills area. This unit averages 50 feet in thickness and dips northward into the Williston Basin. The Whitewood Dolomite is a buff-colored dolomitic unit 50 to 60 feet thick which is unconformably overlain by the Engle- wood Dolomitic shale of Denovian and Mississippian age. The basal portion of the Whitewood is gradational in lithology with the underlying Roughlock Siltstone. The dolomites are thin to massive bedded, tend to be argillaceous and impure, and hence probably not entirely suitable for acceptor rock. 167 5.2.4 PHYSICAL CHARACTERISTICS OF CARBONATE ROCKS Physical tests have been conducted on the Tymochtee Dolomite from Ohio which is currently being used by the Rapid City Pilot Plant; the Minnekahta Limestone from quarries near Rapid City; and the Madison Limestone (and its Black Hills equivalent the Pahasapa Limestone) from several outcrop areas in the study area. The physical testing was done to provide some basis of comparison between rock types with which Consol has had some operating experience and rocks encountered during the present study. Because a minimal generation of fines and a maxi- mum "toughness" are important for a suitable acceptor rock, parameters were chosen for testing that relate directly to these characteristics. The Tymochtee and Minnekahta were quarry-fresh samples, whereas the Madison-Pahasapa samples in most cases were collected from weathered outcrops because there were no quarries available. The differences in the physical characteristics between the weathered and unweathered samples is still under study but is not believed to be of major impor- tance in most cases. The physical tests were of two main types: drop hammer tests and stan- dard rock mechanics analysis. The drop hammer tests were run in a device designed for this project when it became apparent that in rela- tively inaccessible areas it would be impractical to collect largo samples required by more standard testing methods such as the Los Angeles Abrasion tests. 5.2.4.1 Drop Hammer Tests Drop hammer tests were run on the larger (2 to 5 kilogram) samples from the surface sections every 20 to 30 feet or at each important change in rock character. The drop hammer tester consists of a barrel made of a piece of heavy 3-inch-ID pipe 54 inches long with 1/2-inch holes bored through the wall at 6-inch intervals. The holes serve as drop distance references and also as air escape vents when the hammer is dropped. The barrel fits into a base consisting of a section of plate steel fitted with a short sleeve that receives the barrel and supports it in a vertical posi- tion. The hammer is a 25 pound section of cylindrical bar stock that slides freely in the barrel. A uniform method of testing was used for all test.s for this project. Three or four roughly cubical chunks of rock approximately 1 inch on a side were prepared from a given sample. These three or four chunks consituted one sample for drop hammer test purposes. The barrel was removed from the base and a cube placed in the sleeve. The barrel was repositioned in the base and the 25-pound hammer was then dropped from 4 feet. The barrel was removed and the rock debris emptied into a container. This procedure was repeated until all cubes from each sample had been crushed. Thus each cube received one blow from a hammer of standard weight dropped from a standard distance. The total debris for each sample was then rotap sieved through a 6 sieve- plus pan nest for 10 minutes. The size fractions were weighed to 168 0.01 gram and converted to percentages. These percentages were then plotted in histogram form with percent versus size. Dual sample analyses have confirmed good repeatability of results using this procedure. A special effort was made to include a 9 X 20 mesh frac- tion in the analysis which is the desired acceptor size range presently used. A visual comparison can be made between histograms to note relative differences in generation of fines and other size ranges. Direct comparison may be made with the Tymochtee which was also tested. A large block (approximately 100 pounds) of Tymochtee was obtained from the quarry in Ohio. It could be generally divided into a massi- vely bedded and a banded or fine bedded portion. Three samples from each of these portions were run. Three samples of the basal Pahasapa collected near Highway 40, 10 miles west of Rapid City, were also run so that drop hammer data would be available for all rocks upon which rock mechanics tests were run. This sample site was selected because the rock has about the same Ca:Mg ratio as the Tymochtee. For the remainder of the Pahasapa limited sample size allowed only one test to be run on each selected horizon. Three samples of the Minne- kahta Limestone from the same block of raw stone from the Hills Materials Quarry, Rapid City, South Dakota were run. The histograms from all drop hammer tests are included in Appendix C, Book 2 of this volume, and a tabulation of the percent of 9 X 20 mesh size particle production and the percent of less than 20 mesh production is included as Table 5-3. For ready reference purposes Table 5-4 has been prepared to present certain averaged results for rocks on which more than one test was run. These average values provide a relative index for comparison with the tabulated data in Table 5-3 and the histograms. In addition a separate column has been provided in Table 5-3 which gives a compa- rative "toughness" value for the tested sample relative to the fines generation value of 11.50 percent for the overall Tymochtee. A value of 10 percent or less was rated tougher; 10 to 15 percent, comparable; 15 to 20 percent, weaker; and over 20 percent, much weaker. It is readily noted that most of the rocks tested, except for the Big Horn Dolomite, seem to produce a relatively high amount of fines as compared to the Tymochtee. Again it should be remembered that the Tymochtee tested was fresh from the quarry whereas all of the rocks from the study area except for the Minnekahta from the Black Hills, and the rocks from the Dayton, Wyoming, Section were from relatively weathered outcrops. The influence of weathering on fines production and other properties of these carbonate rock remains to be determined, but it is felt that reasonably valid general comparisons can be made.' The cost of obtaining unweathered samples from outcrops would be prohibitive. 169 Weight Percentage of 2.0 to 0. 83 MM (9 X 20) Mesh and Less Than 0.83 MM (20 Mesh) Size Grade From Drop Hammer Tests LOCATION SAMPLE NO, INTERVAL WT % WT % "TOUGHNESS" (FT) 2.0 to FINER AS COMPARED 0.83MM THAN TO Tymochtee (9 X 20 0.83 MM BASED ON FINES MESH) (20 MESH) GENERATED Mission Canyon (Madison-Lodgepole) Little Rocky Mountains, Montana Mlp 1 0-5 12.77 18.34 weaker Mlp 3-4 15-30 25.92 18.97 weaker Mlp 12-14 100-120 16.42 21.95 much weaker Mlp 15-16 125-135 18.37 26.22 much weaker Mlp 18-21 145-170 17.54 25.57 much weaker Mlp 26-27 210-220 16.89 22.70 much weaker Mlp 40-44 280-305 15.54 21.84 much weaker Mission Canyon (Madison-Mission Canyon) Little Rocky Mountains, Montana Mmc 1-3 0.15 Mmc 4-6 15-30 Mmc 7-15 30-75 Mmc 18-19 85-95 Mmc 21-23 100-115 14.46 15.80 17.92 15.08 7.00 27.10 20.84 25.06 22.45 28.48 much weaker much weaker much weaker much weaker much weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 1 of 11) 170 LOCATION 1 SAMPLE NO. INTERVAL (FT) WT % 2.0 TO 0.83 MM (9 X 20 MESH) WT % "TOUGHNESS" FINER AS COMPARED THAN TO Tymochtee 0.83 MM BASED ON FINES (20 MESH) GENERATED Mission Canyon (Continued) Mmc 24- ■28 115 -140 14. ,18 26 .82 much weaker Mmc 30- ■33 145 -165 10. 96 23 .76 much weaker Mmc 35- -38 170 -190 27. ,56 24 .07 much weaker Mmc 39- ■43 190 -215 14. ,13 38 .64 much weaker Mmc 44- -46 215 -230 17. ,07 36 .09 much weaker Mmc 48- -51 235 -255 16. ,09 34 .20 much weaker Mmc 53- •57 260 -285 18. ,96 32 .53 much weaker Mmc 60- -67 295 -335 14, .25 29 .88 much weaker Mmc 68- -74 335 -370 15. ,48 27 .65 much weaker Brown' s Canyon (Madison-Lodgepole and Mission Canyon) Little Rocky Mountains Montana Mlp 81 55- 60 14. ,72 28 .99 much weaker Mlp- -mc 61-68 120- 160 14. ,12 26 .15 much weaker Mmc 56- ■60 160- 185 12. 65 19, .82 weaker Mmc 53- ■55 185- 200 15. 33 24, ,02 much weaker Mmc 47- ■52 200- 230 12. 50 28, .41 much weaker Mmc 33 295- 300 16. 79 22, ,47 much weaker Mmc 29- ■32 300-. 320 10. 03 9. ,70 tougl ler Table 5-3. DROP HAMMER TEST RESULTS (Sheet 2 of 11) 171 LOCATION SAMPLE NO INTERVAL WT % WT % "TOUGHNESS" (FT) . 2.0 TO FINER AS COMPARED 0.83 MM THAN TO Tymochtee (9 X 20 0.83 MM BASED ON FINES MESH) (20 MESH) GENERATED Brown's Canyon (Continued) Mmc 19-24 340-370 Mmc 14-18 370-395 Mmc 10-13 395-415 Mmc 6-9 415-435 Mmc 1-5 435-460 Dayton (Big Horn Dolomite) Big Horn Mountains , Wyoming Day Ob 3 Day Ob 8 Day Ob 14 Day Ob 15 Day Ob 18 Day Ob 22 Day Ob 24 Day Ob 29 Day Ob 35 Day Ob 37 Day Ob 38 10-15 35-40 65-70 70-75 85-90 105-110 115-120 140-145 170-175 180-185 185-190 15.98 18.21 weaker 14.94 17.48 weaker 15.03 19.81 weaker 13.98 28.32 much weaker 16.75 26.95 much weaker 5.36 6.72 8.11 8.06 6.73 6.03 10.68 7.65 5.95 3.96 2.56 6.69 7.90 13.42 11.50 9.16 7.71 13.81 9.15 7.90 14.42 8.43 tougher tougher comparable comparable tougher tougher comparable tougher tougher comparable tougher Table 5-3. DROP HAMMER TEST RESULTS (Sheet 3 of 11) 172 LOCATION SAMPLE NO. INTERVAL WT % WT % "TOUGHNESS" (FT) 2.0 TO FINER AS COMPARED 0.83 MM THAN TO Tymochtee (9 X 20 0.83 MM BASED ON FINES MESH) (20 MESH) GENERATED Dayton (Continued) Day Ob 43 210-215 9.66 16.81 weaker Day Ob 44 215-220 8.32 13.54 comp arab le Day Ob 46 225-230 10.46 16.75 weaker Day Ob 48 235-240 10.68 33.06 much weaker Day Ob 53 260-265 8.90 11.36 comparab le Day Ob 55 2 70-275 12.62 16.13 weaker Day Ob 60 295-300 5.22 6.85 tougher Day Ob 6 3 310-315 10.20 14.57 comparable Day Ob 66 325-330 6.05 10.09 comparable Day Ob 69 340-345 4.95 7.10 tougher Day Ob 74 365-370 7.51 15.40 weaker Day Ob 77 380-385 7.26 8.83 tougher Day Ob 80 395-400 6.59 8.33 tougher Day Ob 82 405-410 7.73 9.51 tougher Dayton (Madison) Big Horn Wyoming Mountains , Day Mm 1 0-5 17.94 29.99 much weaker Day Mm 3 10-15 15.77 29.80 much weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 4 of 11) 173 LOCATION SAMPLE NO INTERVAL (FT) Dayton (Continued) Day Mm 6 Day Mm 7 Day Mm 10 Day Mm 13 Day Mm 16 Day Mm 17 Day Mm 20 Day Mm 21 Day Mm 22 Day Mm 23 Day Mm 24 Day Mm 38 Day Mm 39 Day Mm 40 Day Mm 41 Day Mm 43 Day Mm 44 Day Mm 45 25-30 30-35 45-50 60-65 75-80 80-85 95-100 100-105 105-110 110-115 115-120 185-190 190-195 195-200 200-205 210-215 215-220 220-225 WT % 2.0 TO 0.83 MM (9 X 20 MESH) 14.95 14.12 13.41 14.23 13.58 13.59 15.18 10.90 12.46 14.46 15.93 17.68 14.78 15.87 16.96 14.84 15.15 13.13 WT % FINER THAN 0.83 MM (20 MESH) ■'TOUGHNESS" AS COMPARED TO Tymochtee BASED ON FINES GENERATED 29.25 34.36 20.78 21.49 20.42 26.46 19.85 41.43 24.16 22.43 26.21 24.81 16.70 25.09 24.12 25.90 23.31 21.91 much weaker much weaker much weaker much weaker much weaker much weaker weaker much weaker much weaker much weaker much weaker much weaker weaker much weaker much weaker much weaker much weaker much weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 5 of 11) 174 LOCATION SAMPLE NO. INTERVAL WT % WT % "TOUGHNESS" (FT) 2.0 TO FINER AS COMPARED 0.83 MM THAN TO Tymochtee (9 X 20 0.83 MM BASED ON FINES MESH) (20 MESH) GENERATED Dayton (Continued) Day Mm 46 225-230 13.33 18.53 weaker Day Mm 48 235-240 14.58 21.84 much weaker Day Mm 49 240-245 11.97 16.56 weaker Day Mm 50 245-250 14.01 20.34 much weaker Day Mm 51 250-255 16.82 25.66 much weaker Day Mm 52 255-260 14.26 12.90 comparab le Day Mm 53 260-265 16.49 29.79 much weaker Day Mm 54 265-270 14.02 18.09 weaker Day Mm 56 2 75-280 13.41 20.82 much weaker Day Mm 57 280-285 16.10 36.91 much weaker Day Mm 58 285-290 19.60 24.68 much weaker Day Mm 59 290-295 15.33 20.98 much weaker Day Mm 60 295-300 19.47 26.12 much weaker Day Mm 61 300-305 16.97 23.20 much weaker Repeat Day Mm 61 300-305 20.67 22.67 much weaker Day Mm 63 310-315 15.71 18.43 weaker Day Mm 65 320-325 17.61 20.24 much weaker Day Mm 66 325-330 16.85 20.60 much weaker Day Mm 67 330-335 15.13 20.78 much weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 6 of 11) 175 LOCATION SAMPLE NO INTERVAL WT % WT % (FT) 2.0 TO FINER 0.83 MM THAN (9 X 20 0.83 MM MESH) (20 MESH) "TOUGHNESS" AS COMPARED TO Tymochtee BASED ON FINES GENERATED Spring Creek (Pahasapa) Black Hills, SD SC 3 20-25 6.34 8.66 tougher SC 11 60-65 9.18 13.28 comparable SC 28 165-170 6.99 15.58 weaker SC 32 185-190 8.80 19.79 weaker SC 38 215-220 7.35 9.79 tougher SC 40 225-230 6.16 11.71 comparable SC 47 335-340 8.55 14.78 comparable SC 56 380-385 7.64 7.47 tougher SC 64 420-425 7.60 30.25 much weaker Spearfish Canyon (Pahasapa) Black Hills, SD SpC 102 0-5 13.39 19.94 weaker SpC 96 30-35 12.82 42.43 much weaker SpC 92 50-55 12.56 26.28 much weaker SpC 87 75-80 13.66 18.84 weaker SpC 81 105-110 13.09 28.92 much weaker SpC 73 145-150 12.35 21.84 much weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 7 of 11) 176 LOCATION SAMPLE NO. INTERVAL (FT) WT % 2.0 TO 0.83 MM (9 X 20 MESH) WT % FINER THAN 0.83 MM (20 MESH) "TOUGHNESS" AS COMPARED TO Tymochtee BASED ON FINES GENERATED Spearfish Canyon (Continued) SpC 66 180-185 12.60 24.91 much weaker SpC 61 205-210 12.50 24.39 much weaker SpC 52 250-255 12.36 20.10 much weaker SpC 48 270-275 11.42 21.86 much weaker SpC 43 295-300 11.96 20.15 much weaker SpC 39 39 0-395 15.43 25.16 much weaker SpC 32 425-430 12.70 24.42 much weaker SpC 25 46 0-465 8.82 12.06 comparable SpC 18 505-510 9.32 15.27 weaker SpC 10 545-550 10.13 12.93 comparable SpC 6 565-570 11.55 15.92 weaker Whitewood Creek (Pahasapa) Black Hills, SD WhC 3 5-10 14.89 24.52 much weaker WhC 7 25-30 15.23 34.24 much weaker WhC 9 35-40 15.98 31.99 much weaker WhC 16 70-75 13.17 26.65 much weaker WhC 23 105-110 13.69 26.49 much weaker WhC 25 125-130 14.90 31.39 much weaker WhC 27 137-140 14.38 18.59 weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 8 of 11) 177 LOCATION SAMPLE NO. INTERVAL CFT) WT % 2.0 TO 0.83 MM (9 X 20 MESH) WT % FINER THAN 0.83 MM (20 MESH) "TOUGHNESS" AS COMPARED TO Tymochtee BASED ON FINES GENERATED Whitewood Creek (Cont: Lnued) WhC 33 165-170 12.48 17.53 weaker WhC 37 185-190 10.51 13.82 comparable WhC 41 205-210 10.82 14.9 comp ar ab le WhC 45 225-230 12.09 23.94 much weaker WhC 51 255-260 7.43 12.60 comparable WhC 56 280-285 11.82 32.79 much weaker WhC 60 300-305 12.98 21.66 much weaker WhC 6 7 335-340 10.90 15.62 weaker WhC 73 365-370 8.75 10.23 comp arab le WhC 77 390-395 8.30 10.99 comparable WhC 85 440-445 13.93 43.59 much weaker Rhoades Ranch (Englewood and Pahasapa) Black Hills, SD RR Ce 1 -5-0 14.57 16.66 weaker RR 2 0-5 11.82 16.28 weaker RR 5 15-20 13.77 25.87 much weaker RR 7 25-30 12.66 29.67 much weaker RR 11 45-50 13.24 27.61 much weaker RR 16 70-75 11.12 14.19 comparable Table 5-3. DROP HAMMER TEST RESULTS (Sheet 9 of 11) 178 LOCATION SAMPLE NO. INTERVAL WT % WT % "TOUGHNESS" (FT) 2.0 TO FINER AS COMPARED 0.83 MM THAN TO Tymochtee (9 X 20 0.83 MM BASED ON FINES MESH) (20 MESH) GENERATED Rhoades Ranch (Continued) RR 18 80-85 RR 24 170-175 RR 31 29 0-295 Whitewood Type Section (Whitewood Dolomite) Black Hills, SD Ow 16 5-10 Ow 17 10-15 Basal Pahasapa Highway 40 W. (Black Hills, SD) 1 2 3 Tymochtee - Banded (Ohio) # 1A # 2A # 3A Table 5-3. 0-5 0-5 0-5 9.51 12.20 9.63 10.90 8.44 9.53 10.11 9.67 11.27 12.33 11.20 11.64 33.39 53.24 21.91 12.91 15.93 15.41 15.40 12.61 13.87 12.94 comparable much weaker much weaker much weaker comparable weaker weaker weaker DROP HAMMER TEST RESULTS (Sheet 10 of 11) 179 LOCATION SAMPLE NO, Tymochtee-Mas (Ohio) sive # 1A # 2A # 3A Minnekahta Limestone (Black Hills, SD) # 1 # 2 # 3 INTERVAL (FT) WT % 2.0 TO 0.83 MM (9 X 20 MESH) FINER THAN 0.83 MM (20 MESH) 11.45 7.48 11.64 11.73 7.46 10.39 Hills Materials Quarry 11.85 13.67 10.60 13.74 13.21 15.19 "TOUGHNESS" AS COMPARED TO Tymochtee BASED ON FINES GENERATED comparable comparable weaker Table 5-3. DROP HAMMER TEST RESULTS (Sheet 11 of 11) 180 ROCK NO. OF AVG WT AVG WT "TOUGHNESS" SAMPLES % OF 2.0 % OF LESS AS COMPARED AVERAGED TO 0.83 THAN 0.83 TO Tymochtee [9 X 20 MM (20 BASED ON FINES MESH) MESH) GENERATED YIELD YIELD Tymochtee Dolomite Ohio, banded Tymochtee Dolomite massive 11.60 10.19 13.14 9.86 Overall Tymochtee Dolomite Basal Pahasapa, 10 miles West of Rapid City on Highway 40 Minnekahta-Hills Materials Quarry, Rapid City 10.90 8.77 11.89 11.50 15.58 14.20 weaker comparable Table 5-4. AVERAGED DROP HAMMER RESULTS FOR ROCKS ON WHICH MULTIPLE SAMPLES WERE RUN 181 Work currently in progress indicates that the drop hammer tests may have a linear and consistent relationship to more costly and time-consuming tests requiring large samples such as the Los Angeles Abrasion Test. A Los Angeles percent wear of about 20 appears to correlate with a drop hammer percent fines of about 12.5, and a Los Angeles percent wear of about 40 with a drop hammer percent fines of about 16. Since drop hammer tests have so many advantages (sample size, time of test, equip- ment cost, mobility, etc.) over Los Angeles tests, the work to date is encouraging but much remains to be done before any definitive statements can be made . Work is also continuing in an attempt to establish relationships between textural and chemical parameters and the physical characteristics of these carbonate rocks. No general conclusions can yet be reached but some further comments will be made in the discussion of individual surface sections. 5.2.4.2 Rock Mechanics Analyses It was decided to run a suite of standard rock mechanics tests on three carbonate rocks to gain some idea of the elastic parameters of these rocks and to provide some basis for comparison with the Tymochtee Dolomite which was used by Consol in its bench-scale gasification. The work shows that it is possible in a very general way to relate the results of the drop hammer tests with the elastic parameters determined by standard rock mechanics tests. Rock mechanics tests were limited to rocks where large specimens were available. Tests were run on the Tymochtee, the basal Pahasapa from Highway 40, 10 miles west of Rapid City, and the Minnekahta from Rapid City. In all cases 7/8-inch diameter cores were cut perpendicular to apparent bedding planes. The ends of the test core specimens were finished in a centerless grinder to minimize stress concentration. Finished specimens were between 2.00 and 2.30 inches in length. The main apparatus used in testing was the Tinius-Olsen testing machine. The rock mechanics analyses were conducted in the laboratories of the Dempartment of Mining Engineering at SDSMT. Although the most complete suite of tests was run on the basal Pahasapa, the results of these tests are approximate for this formation because considerable variation is to be expected from place to place depending on mineralogy, fabric, weathering, and other rock parameters. In Table 5-5 tests 1 through 4 were unconfined compression tests, 15 through 18 indirect, tension or Brazilian tests, and test 19 an unconfined compression test with longitudinal and transverse strain gauges. In the Brazilian test a short specimen was stressed along a diametrical plane causing a tension failure on that plane. Tensile strength can then be calculated from the load at failure. The strain gauge test was run in order to determine Young's Modulus and Poisson's Ratio. 182 The general elastic parameters for the Pahasapa are as follows: C = un confined compressive strength = 13,304 PSI T = un confined tensile strentgh = -771 PSI o S = intrinsic shear strength = 2800 PSI o E = Young's Modulus = 6.34 X 10 PSI y = Poisson's Ratio = 0.235 Unconfined Compression Tests TEST FAILURE STRESS 1 8,731 PSI (discarded) 2 6,984 PSI (discarded) 3 15,300 PSI 4 11,308 PSI Average of tests not discarded = C = 13,304 PSI Unconfined Tension Tests TEST SPECIMEN LOAD AT a 3 LENGTH FAILURE TENSIONAL STRESS (IN) AT FAILURE 15 0.5 550 LB -800.0 PSI 16 0.63 17 0.5 300 LB -727.3 PSI 18 0.6 650 LB -788.1 PSI Average 03 at failure = T = -771.7 PSI Table 5-5. ROCK MECHANICS TEST RESULTS FOR THE PAHASAPA LIMESTONE 183 Tests that were run on the Tymochtee Dolomite included six unconfined compression tests, five Brazilian or tension tests, and one strain gauge test. The Tymochtee is a brittle highly fractured rock and it is diffi- cult to obtain suitable test specimens. Specimens suitable for testing are probably not representative of the total rock. Results from the unconfined compression and tension tests are given in Table 5-6. The general elastic parameters for the Tymochtee are as follows: C q = unconfined compressive strength = 48,771 PSI T q = unconfined tensile strength = -2578 PSI E = Yound'g Modulus = 17.77 X 10 PSI y = Poisson's Ratio = 0.333 Unconfined Compression Tests TEST 1 2 3 4 5 6 FAILURE STRESS 58,620 PSI 41,575.1 PSI 53,299.3 PSI 19,124.5 PSI (discarded) 35,920.9 PSI (discarded) 41,590.3 PSI Average of tests not discarded = C Q = 48,771.4 PSI Unconfined Tension Tests •ST SPECIMEN LOAD AT LENGTH FAILURE (IN) (LB) 1 0.428 925 2 0.487 2255 3 0.425 680 4 0.496 2620 5 0.477 1930 °3 TENSIONAL STRESS AT FAILURE (PSI) -1572.4 -3368.9 -1164.1 -3843.1 -2943.8 Average at failure ao = -2578.4 = T o Table 5-6. ROCK MECHANICS TEST RESULTS FOR TYMOCHTEE DOLOMITE 184 Tests run on the Minnekahta Limestone included two unconfined compression and five Brazilian tests. The results from these tests are given in Table 5-7. The general elastic parameters for the Minnekahta as deter- mined are: C = unconfined compressive strength = 21,951 PSI T = unconfined tensile strength = -769 PSI o The generally greater strength of the Tymochtee appears to correlate with its lower generation of fines from the drop hammer tests. However, this is a very tenuous assumption and other factors could be important . Unconfined Compression Tests TEST FAILURE STRESS 1 23,282.1 PSI 2 20,621.3 PSI 3 (discarded) Average of tests not discarded = C = 21,951.6 PSI Unconfined Tension Tests TENSIO^AL STRESS AT FAILURE (PSI) -261.7 -1072.1 -858.4 -1119.3 -535.2 Average a_ at failure = T = -769.3 PSI Table 5-7. ROCK MECHANICS TEST RESULTS FOR THE MINNEKAHTA LIMESTONE EST SPECIMEN LOAD AT LENGTH FAILURE (IN) (LB) 1 0.556 200 2 0.509 750 3 0.534 630 4 0.468 720 5 0.435 320 5.2.5 CHEMICAL ANALYSIS Samples from the Tymochtee Dolomite (Ohio) , the basal Pahasapa Limestone (10 miles west of Rapid City on Highway 40), and the Pahasapa Limestone 185 from many levels in the measured surface sections in the Black Hills were analyzed for weight percent of CaC0 3 , MgC0 3 , and acid insolubles. These data are presented in tabular form in Table 5-8. The totals in many samples are less than 100 percent because some elements in addition to Ca and Mg are present in the carbonates and these go into solution in the hot HCI-HNO3 mixture used to digest the rock. These additional elements are largely Fe, Mn, K, Na, and other minor constituents and thus they are represented by the difference between the total percent reported and 100 percent. It should be pointed out that the heading "% ACID INSOLUBLE" in Table 5-8 is different from the term which is commonly used in many geologic reports for the material remaining after digestion in dilute HC1. Based on petrographic work with the whole rocks the total acid insoluble percentage can be assumed to be silica and various clays in most cases. For each sample a theoretical weight percent dolomite was calculated and is presented in the last column of Table 5-8. Data for Dark Canyon are from a thesis by Ellis (1960) and were originally presented in a different form. His data have been recalculated and the theoretical weight percent dolomite computed so as to be conformable with the other data. The percent of dolomite determined by petrographic work tends to be several percent higher than those arrived at by wet chemical analysis. This is commonly noted and is due in large part to two factors. First, t he insoluble material and very fine grained intercrystalline amount is often not identified in petrographic work and is not adequately considered. Second, the theoretical chemical formula for dolomite is used in the calculations whereas the actual composition undoubtedly departs from the theoretical. In addition, there is some question as to the action of staining reagents relative to high-Mg calcite. They may not stain them. So, in general, the dolomite percentages arrived at by petrographic means tend to be somewhat high and those from chemi- cal analysis may be a little too low. The descrepancy is not regarded as critical. It is, however, an interesting problem and further work in this area is planned. There seems to be no obvious relationship between chemical composition and the percent of fines developed by impact tests. Weathering seems to be a more important factor. Lack of any relationship may be related to the reaction of the different samples to weathering. Further re- search into relating the various textural, "toughness" and chemical parameters is underway. Some of the purer dolomites are quite tough and some are weak; the same holds for limestone samples. So many factors may relate to the "toughness" of a rock (e.g. chemical compo- sition, crystal size, porosity, cementation, weathering, fracturing and jointing, bedding, etc.) that any definitive answer would require a tremendous amount of work. 186 WT % WT % % ACID CALCULATED CaC0 3 MgC0 3 INSOLUBLE THEORETICAL WT % DOLOMITE TO + 0.5% SPEARFISH CANYON (Pahasapa) 50.3 36.0 5.4 79.0 2.5 84.5 1.8 82.5 0.3 86 . 5 0.4 84 . 0.3 84.5 6.3 75 . 5 0.5 83.5 0.4 86 . 5 1.0 82.5 7.9 76.0 3.9 81.0 1.2 82.5 1.4 70.0 2.0 58.0 1.8 68.0 0.3 57.0 0.5 70.0 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 1 of 9) SAMPLE NO. INTERVAL (FT FROM BOTTOM) SpC 102 S 0-5 SpC 99 15-20 SpC 96 30-35 SpC 92 50-55 SpC 89 65-70 SpC 87 75-80 SpC 84 90-95 SpC 81 105-110 SpC 76 130-135 SpC 73 145-150 SpC 70 160-165 SpC 66 180-185 SpC 63 195-200 SpC 61 200-205 SpC 58 220-225 SpC 55 235-240 SpC 52 250-255 SpC 50 260-265 52.8 38.5 53.8 37.6 53.3 39.3 55.1 38.2 55.9 38.3 53.1 34.3 54.7 37.8 53.2 39.4 55.2 37.4 51.7 34.4 52.2 36.9 54.4 37.6 60.6 31.9 66.0 26.3 61.4 31.2 68.5 26.0 62.6 31.9 187 SAMPLE NO. INTERVAL WT % WT % % ACID CALCULATED (FT FROM CaC0 3 MgC0 3 INSOLUBLE THEORETICAL BOTTOM) WT % DOLOMITE TO + 0.5% SpC 48 270-275 71.7 24.4 0.3 53.5 SpC 47 275-280 65.4 28.4 1.2 62.5 SpC 46 280-285 56.0 37.4 0.5 82.5 SpC 43 295-300 55.3 38.1 0.3 84.0 SpC 39 390-395 55.5 38.6 0.0 85.0 SpC 36 405-410 54.0 39.1 0.8 86.0 SpC 32 425-430 54.7 37.7 0.4 83.0 SpC 28 445-450 53.5 39.9 0.5 87.5 SpC 25 460-465 48.3 31.8 13.9 70.0 SpC 22 485-490 53.3 38.6 2.2 85.0 SpC 20 495-500 57.1 30.8 5.8 67.5 SpC 19 500-505 98.1 -- 0.9 SpC 18 505-510 95.7 0.7 1.5 1.5 SpC 17 510-515 97.2 -- 1.5 SpC 16 515-520 97.3 -- 1.7 SpC 13 530-535 98.1 -- 0.8 SpC 10 545-550 62.7 21.2 2.1 46.5 SpC 8 555-560 59.5 33.8 0.7 74.0 SpC 6 565-570 51.6 29.9 12.6 65.5 SpC 3 585-590 98.4 -- 0.8 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 2 of 9) 188 SAMPLE NO. INTERVAL (FT FROM BOTTOM) WT % CaC0 3 WT % MgC0 3 % ACID INSOLUBLE CALCULATED THEORETICAL WT % DOLOMITE TO +_ . 5% WHITEWOOD CREEK (Pah; isapa) WhC 3 5-10 53.6 33.0 6.4 70.5 WhC 7 25-30 53.5 37.9 1.0 83.0 WhC 9 35-40 56.5 36.5 1.0 80.0 WhC 14 60-65 71.3 24.2 0.6 53.0 WhC 16 70-75 67.2 29.2 0.2 64.5 WhC 19 85-90 90.1 7.2 0.5 16.0 WhC 23 105-111 59.7 33.4 0.3 73.5 WhC 25 125-130 64.5 29.2 0.4 64.0 WhC 27 135-140 65.3 28.0 1.4 61.5 WhC 30 150-155 51.1 31.7 9.9 69.5 WhC 33 165-170 60.5 28.4 5.4 62.5 WhC 35 175-180 74.5 20.9 0.7 50.0 WhC 37 185-190 79.1 9.0 8.6 20.0 WhC 39 195-200 76.9 17.9 1.5 39.5 WhC 41 205-210 57.3 14.3 4.1 32.0 WhC 45 225-230 58.1 35.0 0.7 77.0 WhC 48 240-245 55.8 37.3 0.6 82.0 WhC 51 255-260 64.4 29.5 1.3 64.5 WhC 53 265-2 70 53.9 38.6 0.7 85.0 WhC 56 280-285 54.8 36.5 2.3 80.0 WhC 60 300-305 53.7 37.1 3.3 81.5 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 3 of 9) 189 SAMPLE NO. INTERVAL (FT FROM BOTTOM) WT % CaC0 3 WT % MgC0 3 % ACID INSOLUBLE CALCULATED THEORETICAL WT % DOLOMITE TO + 0.5% WhC 62 310-315 55.0 37.7 0.3 83.0 WhC 65 325-330 97.1 0.8 0.5 1.5 WhC 66 330-335 98.2 0.1 0.6 -- WhC 67 335-340 97.8 0.9 0.7 1.0 WhC 68 340-345 69.1 25.3 0.7 55.5 WhC 69 345-350 56.0 36.6 0.5 80.0 WhC 71 355-360 53.8 39.0 0.6 86.0 WhC 73 365-370 45.1 31.6 17.5 69.0 WhC 77 390-395 42.6 29.1 22.5 63.5 WhC 81 405-410 54.3 38.6 0.6 85.0 WhC 82 425-430 67.1 26.9 0.6 59.0 WhC 85 440-445 65.2 SPRING CREEK 27.9 (Pah as 1.6 apa) 61.5 SC 3 20-25 55.3 37.1 0.4 81.5 SC 7 40-45 54.4 38.1 0.6 84.0 SC 11 60-65 54.1 38.2 0.2 84.0 SC 14 75-80 53.2 38.8 0.5 85.0 SC 18 95-100 53.7 40.0 0.2 88.0 SC 22 115-120 53.7 39.2 0.5 86.0 SC 23 140-145 48.3 37.3 8.7 82.0 SC 28 165-170 55.5 37.0 1.3 81.0 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 4 of 9) 190 SAMPLE NO. INTERVAL (FT FROM BOTTOM) WT % CaC0 3 WT % MgC0 3 % ACID INSOLUBLE CALCULATED THEORETICAL WT % DOLOMITE TO + 0.5% SC 32 185-190 55.2 37.8 0.5 83.5 SC 35 200-205 60.3 33.4 0.6 73.5 SC 38 215-220 54.3 38.2 1.1 84.0 SC 40 225-230 54.5 38.2 0.3 86.0 SC 43 240-245 91.8 3.5 2.2 8.0 SC 46(2) (breccia) 255-260) ) 50.4 91.8 32.9 8.3 7.9 72.0 SC 47 335- 340 99.0 0.0 0.6 0.0 SC 48 340-345 -- -- 81.8 -- SC 49 345-350 98.2 -- 0.7 -- SC 50 350-355 98.6 -- 0.8 -- SC 53 365-370 -- -- 93.9 -- SC 56 380-385 54.2 37.7 3.3 83.0 SC 60 400-405 54.7 37.4 0.8 82.5 SC 64 420-425 53.8 43.8 3.0 96.5 RHOADES RANCH (Englewood-Pahasapa) RR 1 -0 87.7 3.0 6.4 6.5 RR 2 0-5 53.7 39.4 0.5 86.5 RR 5 15-20 54.3 39.6 1.6 87.0 RR 7 25-30 55.7 38.2 0.5 84.0 RR 11 45-50 54.7 38.4 0.1 84.5 RR 16 70-75 55.6 38.9 0.2 85.5 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 5 of 9) 191 SAMPLE NO. INTERVAL (FT FROM BOTTOM) WT % CaCO- IVT % MgC0 3 % ACID INSOLUBLE CALCULATED THEORETICAL WT % DOLOMITE TO ± 0.5% RR 18 80-85 56.0 36.6 0.3 80.0 RR 22 100-105 55.7 38.0 0.2 84.0 RR 24 170-175 59.2 33.6 1.1 73.5 RR 29 190-195 54.3 38.0 0.4 84.0 RR 31 290-295 56.5 36.4 0.2 80.0 BASAL PAHASAPA (Highway 40 Ten Miles West of Rapid City) 0-5 # 1 # 2 # 3 49.5 33.2 9.1 72.5 TYMOCHTEE (Ohio) 52.3 37.2 1.1 81.5 53.7 37.3 1.5 82.0 53.2 37.7 0.4 83.0 DARK CANYON (From Ellis, 1960) (Pahasapa) CALC. CALC. WT % WT % CaC0 3 MgC0 3 1964 0-5 62.9 34.0 1.36 75.0 1965 5-10 61.1 37.4 0.89 82.5 1966 10-15 60.1 38.5 0.71 84.5 1967 15-20 59.1 39.6 0.78 87.0 1968 20-25 59.0 39.6 0.70 87.0 1969 25-30 58.4 40.3 0.70 88.5 1970 30-35 60.0 39.6 0.35 87.0 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 6 of 9) 192 SAMPLE NO 1971 1972 1973 1974 1975 1976 1977 1978 19 79 19 80 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 INTERVAL (FT FROM BOTTOM) 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 90-95 9 5-100 100-105 105-110 110-115 115-120 135-140 140-145 145-150 150-155 155-160 CALC, WT \ CaCO. CALC. WT % MgC0 3 63.0 36.0 58.3 40.8 59.3 39.8 59.6 39.3 59.4 39.8 59.2 40.3 59.6 40.0 60.4 39.2 64.2 35.3 59.9 39.5 60.2 39.2 60.3 39.2 61.4 38.1 63.3 36.1 63.0 36.5 63.3 36.0 64.4 34.8 60.3 38.6 61.3 37.4 64.2 33.4 60.5 35.6 59.70 38.1 % ACID INSOLUBLE CALCULATED THEORETICAL WT % DOLOMITE TO +0.5% 0.57 79.0 0.57 89.5 0.75 87.5 0.52 86.5 0.51 87.5 0.30 88.5 0.18 88.0 0.10 86.0 0.12 77.5 0.15 87.0 0.27 86.0 0.11 86.0 0.10 84.0 0.20 79.0 0.19 80.0 0.37 79.0 0.57 76.0 0.84 85.0 1. 10 82.5 2.13 73.5 3.44 78.0 2.10 84.0 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 7 of 9) 193 SAMPLE NO. INTERVAL CALC. CALC. % ACID CALCULATED (FT FROM WT % WT % INSOLUBLE THEORETICAL BOTTOM) CaC0 3 MgC0 3 WT % DOLOMITE TO + 0.5% 1993 160-165 65.7 33.4 0.62 73.5 1994 165-170 64.6 31.8 3.28 70.0 1995 170-175 77.9 18.7 3.26 41.5 1996 175-180 59.8 38.9 1.20 85.5 1997 180-185 75.8 21.3 2.65 46.5 1998 185-190 62.1 31.7 4.49 69.5 1999 190-195 63.8 33.9 1.56 74.5 2000 195-200 83.8 15.1 0.62 33.5 2001 200-205 88.7 10.7 0.28 24.0 2002 205-210 77.2 21.6 0.62 47.5 2003 210-215 81.6 17.1 0.64 38.0 2004 215-220 87.2 11.4 1.24 25.5 2005 220-225 72.6 0.0 27.14 0.0 2006 225-230 76.9 0.0 22.41 0.0 2007 230-235 84.8 0.0 14.55 0.0 2008 235-240 99.3 0.0 0.22 0.0 2009 240-245 99.8 0.0 0.09 0.0 2010 245-250 99.6 0.0 0.17 0.0 2011 250-255 99.4 0.0 0.48 0.0 2012 255-260 99.5 0.0 0.42 0.0 2013 260-265 98.8 0.0 0.99 0.0 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 8 of 9) 194 SAMPLE NO, 2014 2015 2016 INTERVAL (FT FROM BOTTOM) 265-270 270-275 275-280 CALC. CALC. % ACID CALCULATED WT % WT % INSOLUBLE THEORETICAL CaC0 3 MgC0 3 WT TO % DOLOMITE + 0.5% 99.2 0.0 0.64 0.0 98.5 0.0 1.13 0.0 98.9 0.0 0.75 0.0 Table 5-8. CHEMICAL ANALYSIS DATA (Sheet 9 of 9) 195 5.2.6 DESCRIPTION OF SURFACE SECTIONS The surface sections described herein are presented graphically in Appendix D, Book 2, of this volume. 5.2.6.1 Central Montana 5.2.6.1.1 Mission Canyon Section. One of the classical sections of Mississippian age rocks is exposed along Mission Canyon in the northwest portion of the Little Rocky Mountains of north central Montana. The exposed sec- tion consists of 335 feet of Lodgepole Limestone and 370 feet of Mis- sion Canyon Limestone. This is the type section of the Mission Canyon Formation of the Madison Group. (Designated by Collier and Cathcart in 1922.) The Mission Canyon consists of gray to pink-tan limestones ranging from very fine-grained micritic limestone to coarse crystalline limestone. Fossils are abundant and sheared and brecciated zones are scattered throughout the upper two-thirs of the formation. The lime- stone was quarried some time ago for use at a now-abandoned sugar re- finery at Chinook, Montana. The Mission Canyon section is marked by massive, cavernous bedding. Only the upper half of the Lodgepole Formation of the Madison Group is exposed in Mission Canyon. The Lodgepole is thin bedded, cherty and more argillaceous than the over- lying Mission Canyon. 5.2.6.1.2 Brown's Canyon. The Brown's Canyon section is located a few miles east of the Mission Canyon section. The Mission Canyon Formation and the upper 140 feet of the Lodgepole Formation were measured. Again the Lodgepole is thinner bedded, more cherty and argillaceous than the overlying Mission Canyon Formation. The rocks are largely gray very fossiliferous limestone ranging from micro to coarse crystalline biosparites . As at Mission Canyon the contact is marked by a reddish zone, probably indicating a short period of exposure and weathering on the upper Lodgepole before deposition of the overlying Mission Canyon Formation. The Mission Canyon Formation is a buff-gray limestone ranging from very fine crystalline to coarse crystalline-brecciated zones. Evidence of post-depositional solution is abundant in the breccias of the upper one-third of the formation. The carbonates exposed at both Mission Canyon and Brown's Canyon are in general much weaker than the Tymochtee based on drop hammer tests. 5.2.6.1.3 Alaska Bench - Amsden Formation. The Pennsylvanian age Amsden Formation crops out in a narrow band on the flanks of the Big Snowy Mountain up- lift. The north slope dips gently 5 to 15 degrees whereas on the south flank of the uplift the dip is much steeper, averaging about 45 degrees. The Amsden. is a light gray-tan cherty, argillaceous, limy dolomite, with thin interbeds of sandstone and shale. The rock does not appear to have many of the attributes desired for an acceptor rock. 5.2.6.2 Big Horn Mountains, Wyoming (Dayton Section) 5.2.6.2.1 Big Horn Formation. The Ordovician rocks exposed along Highway 14, 5 miles west of Dayton, consist of 420 feet of dolomites and limestones. The lower 250 feet of the Big Horn Formation consists of limestone, limy 196 dolomites, and dolomites. This lower unit is a massive cliff former. The bedding is poorly defined by bands of chert nodules . The unit weathers gray but on fresh surface shows a yellowish cast. The very base of the unit may contain quartz grains which disappear near the top of the massive cliff forming beds. The insoluble residue contains clay and quartz. The upper unit is less cherty and at the very top consists of thin beds of nearly white, porcelain appearing dolomites. The Big Horn here is generally tougher than the Tymochtee based on drop hammer tests. 5.2.6.2.2 Madison Formation. The lower 50 feet of the section is a tan dolomite that varies from very fine crystalline to sucrosic. The top of the lower 50 feet is marked by a thin bed of limestone. The next 70 feet consists of alternating tan dolomites and light gray limestones. This is overlain by 30 feet of gray, medium-grained limestone that is fossi- liferous and oolitic. The overlying 100 feet consists of purple to gray limestone alternating with fine crystalline lining dolomite. The next 45 feet consists of dolomites and limy dolomites. The top 40 feet consists of gray, very fine to medium crystalline limestone with few beds of limy dolomite. The Madison is, according to drop hammer tests, much weaker in general at this section than the Tymochtee. 5.2.6.3 Black Hills, South Dakota 5.2.6.3.1 Spring Creek. This section was measured just east of the Stratobowl along Spring Creek about 8 miles southwest of Rapid City. The section is within 2 miles of Highway 16. The Pahasapa here is 427 feet thick and the section has a fairly good basal contact with the underlying Englewood Formation and its upper contact with the Minnelusa is well exposed. Morphologically the section is characterized by steep, almost vertical cliffs and gently sloped grassy ledges. The lower cliff con- sists of about 120 feet of gray-tan, fine crystalline dolomite with crinoid fossils and some spongy to vuggy porosity. A 20-foot grassy slope is followed by a 100-foot cliff of gray-tan-pink, fine crystalline dolomite with some calcite cement and fossils. There is up to 5 percent quartz silt in the top 20 feet. This section is followed by a 100-foot cliff of massive limestone, dolomite, chert, and clay breccia. Above the breccia is about 40 feet of gray- tan, micro to fine crystalline limestone and dolomite with many chert nodules and bands. The top 40 feet of the section is varicolored, fine to coarse crystalline dolo- mite with many fossil corals and brachiopods and some quartz silt. Samples were tested via the drop hammer test as shown in Table 5-3. In general these rocks are comparable to the Tymochtee in terms of amount of fines . 5.2.6.3.2 Dark Canyon. This section was measured immediately west of Rapid City along Rapid Creek. The thickness of the Pahasapa here is 275 feet and the section has good upper and lower contacts. The lower 130 feet of the section is buff-tan, fine crystalline dolomite with fossils and porosity in some zones. Some calcite is found as intercrystalline cement. Above this is about 65 feet of buff, fine to medium, crystalline, more calcareous 197 dolomite with recrystallized pellets and oolites and many fossils in certain zones. Next comes 80 feet of buff-gray, clastic limestone with zones of chert breccia, oolites and fossils. No samples were available for drop hammer analysis of this section, but the rock appears to be generally weaker than the Spring Creek rocks. 5.2.6.3.3 Whitewood Creek. This section was measured at the inner mouth of Whitewood Canyon about 3 miles northeast of Deadwood. The Pahasapa is 465 feet thick and the lower 165 feet is brownish, fine crystalline dolomite with fossils and zones of spongy porosity. The next 200 feet is gray-buff, fine cry- stalline dolomite with fairly abundant calcite cement, zones of oolites and abundant fossils. The remaining 200 feet is essentially gray-brown- buff, fine crystalline dolomite with zones of chert and cherty breccia near the top. Fossils, oolites, vugs and spongy porosity still occur in zones as does intercrystalline calcite cement. The contact with the underlying Englewood is quite good but the upper contact is less reliable. According to drop hammer tests these rocks are in general much weaker than the Tymochtee. This particular site does have fair highway access and the cliff could be quarried. 5.2.6.3.4 Spearfish Canyon. This section was measured on the west wall of Spearfish Canyon about 6 miles south of Spearfish. The upper and lower contacts are both present and the Pahasapa is 600 feet thick at the site. The bottom 310 feet is gray-brown-buff, fine crystalline dolomite with many fossils and zones of vuggy to spongy porosity. This unit becomes moderately calcareous in the upper one- third and is followed by 75 feet of covered interval in a grassy, rubbly slope. This is followed by 95 feet of brown-gray, fine crystalline dolomite with some bands of chert nodules. The next unit is 65 feet of gray-buff, fine crystalline dolomitic limestone with abundant calcite cement. It is very fossili- ferous and corals, brachiopods, crinoids , and algae were noted. There is about 20 feet of limestone-dolomite breccia at the base of the unit. The top 55 feet of the section consists of gray to tan, fine crystalline, interbedded limy dolomites and dolomitic limestones with chert nodules. These rocks are in general much weaker than the Tymochtee according to drop hammer tests. 5 2 6 3 5 Whitewood Dolomite. This section is the type section for the Ordovician Whitewood Dolomite of the Black Hills. It is located immediately east of Deadwood on Whitewood Creek. The carbonate part of the Whitewood is 48 feet thick at the site and consists of yellow-brown to pinkish mottled, fine to very fine crystalline dolomite. The section becomes progressively siltier towards the base and in fact becomes a siltstone (The Roughlock) in the next lower unit. The Whitewood has zones of vuggy porosity and intercrystalline calcite cement. The dolomite rhombs are well developed and the texture can be described as sucrosic or sugary. The strength of these rocks is variable but they are gener- ally somewhat weaker than the Tymochtee, based on drop hammer tests. The Whitewood would be a difficult rock to quarry because it usually has considerable overburden for its thickness. The silica content is also rather high. 198 5.2.6.3.6 Rhoades Ranch. This section was measured in the west central Black Hills about 25 miles west of Rapid City along the inner edge of the great Pahasapa Limestone plateau that flanks the western Hills. The Pahasapa is at least 307 feet thick here and the upper contact is not exposed. The lower 105 feet of this section forms a near vertical cliff made up of light yellow-brown, fine crystalline dolomite with fossils and zones of vuggy-spongy porosity. Above a grassy 60-foot covered interval is a 35-foot section of pink-gray, fine to medium crystalline laminated dolomite with zones of vuggy porosity. Next above is another covered interval of 85 feet followed by 22 feet of brown to yellow-white, fine to coarse crystalline dolomite with some intercrystalline calcite cement and clay material. Good intercrystalline and vuggy porosity is evident and some fossils are found. These rocks are generally weaker than the Tymochtee on the basis of drop hammer tests. 5.2.7 TRANSPORTATION FACILITIES AND COST ESTIMATES 5.2.7.1 Railroad 5.2.7.1.1 Montana. Principal railroads in eastern Montana are the Burlington Northern and the Milwaukee Road. The Burlington Northern was formed by a merger of the Northern Pacific, Great Northern, and Chicago, Burlington and Quincy Railroads. Both the Burlington Northern and the Milwaukee Road have main lines running east-west through Montana as is shown on the maps. (See Figure 5-5). Branch lines serve most areas quite thoroughly, but parts of the southeastern and east central sections of the state have no trackage. These areas contain some large lignite reserves. Railroad service in the limestone areas is adequate . 5.2.7.1.2 North Dakota. Western North Dakota is well covered by the Burlington Northern system, with some coverage by the Milwaukee Road in the south- west corner of the state. The rail coverage of this area, which con- tains very large reserves of lignite, should encourage further develop- ment. Lignite is presently being shipped from North Dakota points to power plants east of the producing area, and some lignite used in the pilot plant at Rapid City comes from North Dakota. Only a small part of western North Dakota is more than 30 miles from a rail line. 5.2.7.1.3 South Dakota. Rail coverage of southwestern South Dakota is adequate, though some of the road beds are in poor condition. Large limestone reserves in the Black Hills are close to the railroads but there is no direct route to North Dakota. The lignite reserves of South Dakota, located in the northwest quarter, are not very near to rail lines, and spurs would have to be built if rail transportation were to be required. The lignite reserves in South Dakota are small compared with surrounding states, but the South Dakota material contains uranium, a valuable byproduct . 5.2.7.1.4 Wyoming. Rail coverage in northern Wyoming is relatively good, with two primary lines of the Burlington Northern traversing the area from southeast to northwest. Two areas that could be critical lack desirable 199 trackage. The first is the Powder River Basin in northeast Wyoming between the Black Hills and the Big Horns. This basin contains very- large reserves of subbituminous coal, an estimated 62 billion tons in Campbell County alone. The Burlington Northern has announced tentative plans to construct a connecting line between Gillette and Douglas, Wyoming. The second area that may require rail coverage is in north- eastern Wyoming between the Burlington main line at Moorcroft and the Chicago and Northwestern branch line east of Spearfish, South Dakota. At present, subbituminous coal is shipped from near Gillette to Rapid City, but it has to travel south to Chadron, Nebraska, before coming back north to Rapid City. Shipping costs (see also the subsection on rates) are now too high for large movements over the present route. 5.2.7.1.5 Track Conditions. It appears that the condition of the railroads is largely a function of the operating companies. The Burlington Northern system probably has the best track and roadbed conditions of the three railroads considered here. Trains moving over the Burlington main lines travel at speeds o c 50 or more miles per hour. The Chicago and North- western tracks, on the other hand, are generally in poor condition, with derailments fairly common. Although faster speeds can possibly be expected on more level runs, it takes 6 to 8 hours for freight to move the 100 miles from Rapid City, South Dakota, to Chadron, Nebraska. Only rarely does the train's speed reach 30 miles per hour on that run. Faster service cannot now be expected from the Chicago and Northwestern in the area covered by the accompanying map (Figure 5-5) . Milwaukee road service is not particularly fast in western South Dakota. Their east-west main line across North Dakota and Montana can be expected to give relatively fast and reliable service. 5.2.7.1.6 Rates Charged for Transportation of Bulk Materials. The different rail- roads have slightly differing price quotations and the rates charged are shown on the following tables. (Refer to Tables 5-9, 5-10, and 5-11.) Three things should be kept in mind: (1) rates can change quickly; updating must be continuous, (2) unit train transportation costs can be much less and may approach 3 mills' per ton mile, depending on the length of the route, (3) special rates would be negotiated for a large volume user. 5.2.7.2 Truck 5.2.7.2.1 Montana. Eastern Montana is fairly well covered by its primary road system, consisting of the state and federal highways shown on the accompanying map (Figure 5-5). In addition, there are numerous second- ary roads, not shown on the map, which make road coverage of eastern Montana quite good. One possible problem is that much of the state's interstate system is not yet complete as of this writing (1972) . 5.2.7.2.2 North Dakota. Western North Dakota is also well covered by good quality primary roads. It too has a well developed secondary road network, which provides even better coverage than the secondary system in adjacent Montana. North Dakota has another distinct advantage in that its east- west interstate link (1-94) is complete across the entire state. 200 ORIGIN Gillette, Wyoming Gillette, Wyoming Zap $ Buelah, ND Baukol-Noonan Siding, ND All Montana Origins DESTINATION Rapid City, SD Osage, Wyoming Rapid City, SD Rapid City, SD Rapid City, SD COMMODITY CENTS /TON OF 2000 LB Coal 360 Coal 146 Lignite 496 Lignite 605 Limestone 1841 N.B.: In all but the Gillette-Osage transfer, more than one company is involved. Table 5-9. BURLINGTON NORTHERN SYSTEM RATES 201 o o LO o ■>* 00 to to tO 00 LO to LO Tl- ■sD I 1 f^ tO r-^ i i CM to to i | to to o O i— i LO cm o r,l to to 8 to o vO O o 00 CM CM CM to "J- to I/) LO LO h> fs r^ a> i— 1 CM CM o to o eu o o o CM O e o p p o LO t^ to o O o ^O ^H I— ( CM to 00 m H i— i Z ^-n LO r^ vO r^| r» r^ C (/> -H

- x S to ■ S ■ o Z ■5 a 3 o 2 S o 22-ga 5 2 S - 5 §2 g S8.3 eg O CD eo d ;1 E-> 2 f3 S2 B -2 h 7 lie If 5-g fe2 III 3 C9 -O It a « •2 5 1 E E * o 3 r 2 o a II is 3 • ® n -5 ii i it I < O • (J Her" mc£ a- ^ 2 c 8* *3 t 2 2k2 s 4> 9 ■£ .2 S5So •" O ~ £ 2 c C CJ 3 00 00 00 Q ^ ^ 93 V a 2 E - — aj "• c e o oo 5 g = S 8 8.3 « 4) 3 g « - 41 -1 * s 2% © a « X J* ■Sf Si Is >• z ^ e 5 i o. a fl -g in o ™ a z m a III 0< a to o G* - ^ CO *. • p: « o » ii £>S § a S S o 3" Z — CO s s a •8 J, -o 3-s & _, o o *1 2 » 6 S u, u. _* k * -2 I 5 ■■J CD > .g m fa O CD O t* 3 o 2 o -J k 2 2 3 41 1 si ■ S 2 5 B — 4) > o CD t— I E- 206 « - i - •* s * e ° ° S — <~ 3 o Z S B S ls|3 • S g 5 fi E o I* a. 5 » E lllll SS C it: -o .— s.' o i-s 0. O « 3 3 § ass 2 8 „ g t 3 ^ •T 3 2 £ a o h 1 * *" -a T5 .3 ■8.8 & B«S«S P ~ — si ri i—x-OB-caoi "■BO* 3 « O ™ . o S 2 « . o p a « ■ S C e i a 3 C > QeaoCu o> > z h. z. a >" S5 Pn £> a. 3S» 0. o o bo 3 3 5 £ 1 5 2 „ 2 S a 2 •o -2 g • a s|* s ° I s - 8 « E a « o o 2 < t- o c >- a J o 1! o o 0. * ** "S Si £ s s 5 »■ s * o Q „, — [•.■£ £ » 8 S g^« «• « tt ^ -^ 5 E s i I H Q fi 1 >, « £ o £ * « -s ~ § § £ 2 a » e o 9 a o a 3 c " S o I g Si a " a o till!! 25 - ■CJS m a S I S — TJ S l " i- c a o I a S 3 £ S 8? alii 207 2 £ « j z - « 8 o> 2 a o h ■ -o o •- 3 a o z 2 s 5 i 1 S3 ^ S § S £ * III. Bs 5 i . ? T5 8 C ■ JS ■ U JO t < u C D. O f §| 81 58. M » 2 . — = " I +l -8 . 3 i • C a a, cs a - - + 1 "35 C K ^ 5 S I? e§5 o it I < o « ?S e 8 M 5- -1&J „ a o t o8 3 60 ? O > 5 ° o i c « 2. o ElS .C J3 9fi id is 8 *u 0.3 tjtfl 2 il a J - Is — -n - "- E g a o 0. * Z fc. :f -a 0t • C 2 a 1 v- a O at _ 1& ^E •8 ■ 01 ~ 3 x 3 -. • 2 2 5 - * 11 B — • © > 2 « 6 a s 3 §1 H 5- So §3 3 s o I 208 (2) Transportation will be one of the most important factors in the cost of an acceptor and it is one of the most difficult to analyze until a plant site is chosen and detailed acceptor specifications established. Assuming large volumes of material to be moved, trucks would be uneconomical for long hauls but feasible for short hauls. However there are no surface sources of possible carbonate acceptor rocks closer than 230 miles to the major lignite-producing areas. Rail transportation is available within reasonable distances from most potential quarry sites (except the Black Hills) to eastern North Dakota. Rail transport is by far the cheapest method for movement of bulk materials with existing transportation methods. (3) There have been substantial increases in truck and railroad freight rates in the past 18 months. Costs have increased by 50 to 100 percent. It is unlikely that rates will decrease and more likely they will continue to increase. (4) There is a strong trend toward increasing natural gas prices both because of inflationary trends and a general realization that gas prices are way underpriced when viewed against finding, process- ing, and delivery costs. This should offset or more than offset increasing freight rates. (5) The problem of access could be involved inasmuch as the Little Rockies fall on the Fort Belknap Indian Reservation and National Forest. Much of the Big Snowys are National Forest land as are the Big Horn Mountains of Wyoming. The Crow Indian Reservation of southern Montana (a potential source) is off limits to prospecting at the present time but it is possible that reasonable royalty rates could be negotiated. Once a suitable acceptor rock is found access to outcrop area could be gained by purchase of fee land, possibly by negotiations, or in some cases by dealing with established quarries. (6) Although quarrying is not generally a serious problem, requirements of the EPA must be given consideration. These can be adjudicated but usually at an increase in cost. There is no way in estimating how much this cost increase will be at this time. Some estimates are as high as 25 percent. Ten percent seems reasonable. When decisions as to plant location and acceptor specifications and quantity are made it will be possible to make a more detailed analysis . 209 SECTION 6 GRINDING PROPERTIES OF CARBONATE ACCEPTOR This section includes a single report, "Study of Physical Character- istics of Carbonate Rocks to Determine Wear Properties and Crushing Properties," submitted by the South Dakota School of Mines and Tech- nology on January 6, 1978. Related work, which determined abrasion characteristics of carbonate acceptor materials, was performed in the acceptor reconstitution studies covered in Section 2 of this volume. 210 6.1 STUDY OF PHYSICAL CHARACTERISTICS OF CARBONATE ROCKS TO DETERMINE WEAR PROPERTIES AND CRUSHING PROPERTIES ' .o 6.1.1 INTRODUCTION The purpose of this study was to determine the physical characteristics of a large number of carbonate samples in order to determine the optimum type of material and method of crushing to yield the lowest cost acceptor material delivered to the gasification plant site, and to determine this information at the lowest cost. Four methods of testing physical properties have been attempted, each having advantages and disadvantages. Preliminary evaluation of a particular carbonate consisted of the drop hammer test using a simple impact-type crusher and a small sample. This, and the other tests are described in detail in this section. This type of test is of value for two principal reasons, first, the sample requirement is on the order of 1000 grams, an amount which can easily be carried out of the least accessible outcrops with ease; and second, little time and equipment cost is required to make the determinations once the sample is obtained. The drop hammer is useful for eliminating those carbo- nates which are definitely not suitable. The so-called "hammer test" used to evaluate acceptor in the original bench scale test, is not objective enough to use for repeatable results. If an area showed promise, a larger sample was collected for a Los Angeles abrasion test. This requires a sample of at least 40,000 grams to assure that enough material can be crushed to the size range required for the test. The test measures basically the resistance of the rock to abrasion which is to some extent a measure of its grinding character- istics. The size of the sample limited the number tests, but enough information was obtained to show certain trends. Since size, power requirements, and capacity of a crushing system depend to a large extent on the compressive strength of the rock being used, a series of standard compressive tests were run. Once again, a sample of 30,000 to 40,000 grams was needed to be reasonably certain of getting the flawless cores used in the test. The ultimate test, of course, is to take a sample of the rock under con- sideration and crush it in a variety of types of crushing equipment . This was attempted using our laboratory-scale equipment, but with the problems of scaling and the fact that the equipment was overpowered for this type rock, the experiments were not repeatable enough to be considered valid so the attempts were dropped. Research into the literature led to the data sheets in Subsection 6.1.4 and to the con- clusion that the difference in crushing properties is not nearly as important a factor as the carbonates' suitability to the acceptor process. In other words, once the most suitable carbonate acceptor is found, the cheapest way to crush it should be designed, but the ease of crushing should not be a criteria in the selection of a carbonate. With the accompanying data sheets completed, it should be possible 211 6.1.2 for an engineering firm specializing in crushing systems to design the plant layout without the need of a pilot plant for crushing or other studies of that type. MECHANICAL TESTS 6.1.2.1 The Los Angeles Abrasion Tests The Los Angeles abrasion test measures resistance to self-abrasion and impact abrasion utilizing a ball mill. The test simulates the wear- resistant characteristics of a material in an environment similar to the wear on a gravel road. 6.1.2.1.1 6.1.2.1.2 Method. The tests were carried out as per ASTM Designation C131-69 which is the appropriate standard method for rock anticipated as a tertiary crusher product. Both size range B and size range C tests were run. The sample and steel ball size parameters for these tests are as follows: SAMPLE SIZE (RANGE C) 3/8 X 1/4 IN 2500 grams 1/4 IN X #4 (4.76MM) 2500 grams Total sample 5000 grams SAMPLE SIZE (RANGE B) 3/4 X 1/2 IN 2500 grams 1/2 X 3/8 IN 2500 grams Total sample 5000 grams ABRASIVE SIZE (RANGE C) 8 steel balls 27/32 IN (46.8MM) diameter and 390 to 445 GM each. Total weight 3330 grams +20 grams ABRASIVE SIZE (RANGE B) 11 spheres as above. Total weight 4584 grams +25 grams The sample is subjected to 500 revolutions in a standard -si zed ball mill and then sieved using a #12 sieve. The weight percent of the sample which will pass the #12 sieve is the wear factor. The lower the wear factor, the better the material is for abrasion usage. The only deviation from the standard test procedure was that samples were not washed after grinding because experience indicated only a minimal effort on the wear factor was due to adherence of fines . Results and Evaluation. The resulting wear factors for the different samples on which Los Angeles Abrasion tests were completed are shown in Table 6-1. The wear factors listed are presumably a reasonably good measure of the abrasion resistance of the different samples. The size range B tests are probably not exactly comparable to the size range C tests but the greater ball charge compensates in part for the size difference. The samples were chosen to be as representative as possible but it should be emphasized that there will be some deviation in the rock characteristics at any one locality. For example, the very high wear factor for the lower Bighorn Dolomite sample can probably be explained by an unexpected oriented crystal structure which was only 212 revealed by microscopic examination. Furthermore, that sample was from a weathered outcrop whereas other samples such as the basal Pahasapa, Tymochtee, and Minnekahta samples were from fresh-quarried rock. Nevertheless, it is believed valid to list the rocks in terms of compa- rable abrasion resistance based on the Los Angeles Abrasion tests and general field observations as follows: EQUIVALENT TO THE TYMOCHTEE DOLOMITE Jefferson Dolomite (Logan, Montana) Madison Limestone (Logan, Montana) SOMEWHAT WEAKER Minnekahta Limestone (Rapid City, South Dakota) Pahasapa Dolomite (Pringle, South Dakota) Bighorn Dolomite, upper sample (Dayton, Wyoming) Whitewood Limestone (Black Hills, South Dakota) MUCH WEAKER Pahasapa Dolomite, basal sample (Highway 40, South Dakota) Bighorn Dolomite, lower sample (Dayton, Wyoming) SIZE RANGE "C" WEAR FACTOR (%) Pahasapa (Pringle, SD) 29 Whitewood (Black Hills.SD) 33 Bighorn (Dayton, WY) -Lower sample 60 Bighorn (Dayton, WY) -Upper sample 32 Madison (Logan, MT) 26 Jefferson (Logan, MT) 24 SIZE RANGE "B" Minnekahta (Rapid City, SD) 27 Tymochtee (OH) 19 Pahasapa (Basal Section, Highway 40, Black Hills, SD) 40 TABLE 6-1. WEAR FACTOR RESULTS OF LOS ANGELES ABRASION TESTS 6.1.2.2 Drop Hammer Tests The Drop Hammer test is a test designed for this project in order to evaluate carbonate materials with small samples and a minimum of time and equipment. The parameters the test determines are the rocks' resis- tance to impact, the manner in which it breaks when impacted, and the size distribution of the fragments. The test is more semiquantitative, but is of great value for making first -order selections of carbonate sources. Based on the drop hammer test results and field observations, 213 three-fourths of the potential carbonate rocks have been eliminated. This has made the development of the procedure and equipment well worth- while. The drop hammer test results have been used in conjunction with the standard tests to help select the best material, but as a guide and not as the primary criteria; its best use remains in making primary selections. 6.1.2.2.1 Method. The test itself consist of applying a standard force to a standard sample and measuring the size distribution of the resulting fragments. This is accomplished with the aid of a specially designed and built apparatus comprised of a heavy guage steel tube 3 inches inside diameter, 6 feet long, resting in a short receiver (this is the "guide" or "barrel") and the hammer itself, which is free to slide inside the barrel. The test procedure is to cut the sample into four or more 1-IN cubes, place them one at a time in the receiver and drop the hammer from a height of 4 feet. This is repeated for each of the cubes in the sample, then the crushed cubes are combined and sieved in a 6 sieve-plus pan nest for 10 minutes in a mechanical siever. The contents of each sieve and of the pan are then weighed to the nearest hundredth of a gram and converted to a percentage of the total. This data is then presented as a histogram for easy comparison in Figure 6-1. 6.1.2.2.2 Results and Evaluation. A complete list of all the drop hammer tests except those from the Three Forks area is given in Subsection 5.2, "Preliminary Resource Study of Carbonate Rock Resources in Cental Montana, Northern Wyoming, and Western South Dakota" submitted in December 1972. In this report typical or average drop hammer results were selected for comparison with other tests. Those carbonates which did not yield good results in the drop hammer tests were excluded from further consideration. The "toughness" of a particular rock is the -percentage of material finer than 20 mesh. This has in the past been compared with the Tymochtee Dolomite whose fines averaged from 9 to 12.5 percent depending on the part of the formation used (massive or banded material). Less than 10 percent fines was rated as tougher, 10 to 15 percent was rated as comparable, 15 to 20 percent weaker, and more than 20 percent was much weaker than the Tymochtee Dolomite. For the purposes of this study, the actual percentages are somewhat easier to use. It should be noted that the peak size fraction for the unique drop hammer tests almost invariably falls in the 2X4 mesh range. An attempt was made to shift the peak down to the 9 X 20 mesh range by adding more energy to the hammer. Drops were made at 5 and 6 feet, which did, indeed, move the peak closer to the 9 X 20 range, but the extra fines increased disproportionately. This and observa- tions during the preparation of samples in jaw and roller crushers for the Los Angeles abrasion tests indicated, and subsequent experi- ments showed, that the most efficient (suitable sized rock per ton of raw material) method of crushing was to sieve off the correct and the undersized material between crushing stages. 214 V a uu 40 30 20 10 n fe9 CO co CO CO fe^ o 6? Cm i— 1 &5 CM eg feS 1 50 40 30 % 20 10 50 40 ► 30 20 10 9.4 4.7 2.0 .83 .50 .246 mm 2 4 9 20 32 60 mesh Upper Bighorn 6* r-i fe^ o 6^ r-l fei CO f^ 6^ ££ ■* CO 6? CM CM fet CO fe« CD CO CO 9.4 4.7 2.0 .83 .50 .246 mm 2 4 9 20 32 60 mesh Lower Bighorn 9,4 4.7 2.0 .83 2 4 9 20 White wood .246 mm. 60 mesh 3U 40 .- 30 e • • OL 20 10 fe? CO o ° t-° &e CO co 9.4 47 2.0 .83 .50 .246 mm. 2 4 9 20 32 60 mesh Paha Sapa (Pringle) Bighorn Dolomite, Dayton Wyoming Whitewood Formation, Black Hills, S. D. Bighorn Dolomite, Dayton Wyoming Paha Sapa Dolomite, Pringle, S.D. Figure 6-1. HISTOGRAMS OF CRUSHED ROCK SIZE DISTRIBUTION FOLLOWING ONE BLOW OF THE DROP HAMMER 215 The results of the drop hammer tests which largely measure impact resistance show that the most favorable rocks according to that criteria are as follows: MOST IMPACT RESISTANT Weight \ finer than 20 mesh Dayton, WY , Bighorn Dolomite 10 to 15 FT 5 Logan, MT, Jefferson Dolomite 240 to 248 FT 5 Logan, MT, Jefferson Dolomite 270 to 275 FT 9 Tymochtee dolomite 9 AVERAGE IMPACT RESISTANT Logan, MT, Lodgepole 760 FT 11 Minnekahta Limestone 15 LEAST IMPACT RESISTANT Little Rockies, Lodgepole 15 to 30 FT 26 Little Rockies, Mission Canyon 100 to 115 FT 28 6.1.2.3 Comparison of the Drop Hammer with the Los Angeles Abrasion Test In order to compare the results of the drop hammer tests versus Los Angeles abrasion tests, Figure 6-2 was prepared as a plot of the percentage of fines produced by one method against the other. Figure 6-2 shows that the data from the Los Angeles abrasion tests correlate only in a general way. The line running diagonally across the page is the line of best fit and it should have its origin near zero on both scales. That is, if it were found that a certain rock was not damaged by the drop hammer, we would expect little, if any, abrasion in a Los Angeles test. The distribution of the points indicates that the results of the drop hammer test of a particular rock is not too reliable an indicator of how it will perform in a Los Angeles test. However, this is not to say that either test is valid because they do not determine the same physical characteristic. What the graph does indicate is that certain rocks may represent the best compro- mise of resistance to impact and resistance to abrasion. Those rocks are situated in the lower left hand corner of the graph, near the theoretical line. The least resistant of the rocks selected for these tests are those located in the upper right hand corner, as they were neither resistant to abrasion nor impact. Those rocks which lay in the upper left or lower right were found to be resistant to one but not both types of destructive force. They could still be suitable for acceptor use but we would expect problems in comminutions. 6.1.3 PROBABLE CRUSHER SYSTEMS REQUIRED Before the crusher system can be selected, criteria must be set for the type product required, and the physical properties of the raw 216 o o o c s_ o c in cu CJ en c u o C r- c o C 1/) o ■r- l/l ■o s~ re cu M- **~ <♦- O "3 o CM to E- CO w E- OS < (X) tu O O to i— i < O U CN I u SOU I J JOLJl'JPII tlOJQ 10 o < c o o o> Ll 217 material must be supplied. To design the correct size unit, naturally, the desired capacity must be given. In accordance with this, the information sheets in Subsection 6.1.4 have been prepared. They should offer the information required for an engineering firm specializing in crusher design and construction to do the preliminary studies on setting up a quarry. Prior to the actual construction of a quarry and plant, a complete survey of the chosen site will be required to determine such things as quarry shape, transportation to the crusher plant, exact size of the primary feed and so on. The basic size range desired eliminates most types of crusher systems, however. Some systems such as hammermills and some types of roller mills are designed for the production of large amounts of fines, usually where the product is road surface material or cement. Other types of roller mills are designed for a minimum of fine material. Jaw and gyratory systems can be used to break rock down from quarry feed to a size small enough for double-roller crushers to handle without excep- tionally many fines. A survey of the literature indicates that the best system would use a single-roller crusher (not an attrition type) followed by as many stages of double-roller as is most economical to attain the desired sizing. Screening to remove fine material and the correct size fraction would be carried out between each crusher stage. Any system which involves "attrition" as part of the comminution process should be avoided, as fine material is the only product of attrition. Some examples of attrition or partially attrition systems are all rod or ball mills, hammermills, and some roller mills. Great care should be taken to select the system which produces the most suitable material for the dollar, especially in this project, because few crushers are used to produce this size range of carbonate, as a certain percentage of fine material is acceptable or desired in most carbonate quarry operations or else a much coarser product is required. About the only industries requiring material in the size range of the desired acceptor are some of the ceramic and glass industries but they typically are not grinding dolomite or limestone so their circuits are not comparable. Coarse sand-sized barite is needed in some spe- cialized glass products and since barite and calcite are reasonably similar in their physical characteristics, the required grinding circuits should probably be comparable. Because of the high transportation costs of the natural carbonate rocks, work on other aspects of the CC- Gasification Process indicates there is a good probability that the acceptor will be artifically re- constituted from CaCO--Ca(OH) 2 melts. It would have been desirable to have run grinding tests on the reconstituted acceptor but there was not an adequate quantity of material available. Based on the abrasion studies on small quantities of the reconstituted acceptor by Dr. M. C. Fuerstenau, a combination of primary jaw crushing, secondary cone crushing, and tertiary rolls crushing would generate the least amount of fines from the pure reconstituted acceptor. 218 6.1.4 INFORMATION NEEDED FOR THE SELECTION OF ROCK CRUSHING EQUIPMENT 6.1.4.1 Bighorn Dolomite Source: Dayton, Wyoming Condition (Wet, Dry, Frozen): Dry to Wet Moisture Percent: 5 to 20 (Temperature): -30 to +9 0°F Hardness (MOH scale) : 3.5 Physical Characteristics (Bedding, Porosity, etc.): Massive, Micritic texture Percentage of fines in feed: <1% through 9 mesh Will fines be removed by prescreening? yes Size product desired: cubic to rounded (Shape) 9 X 20 mesh (Tyler size) Fines desired in product: (MAX) Product removed by: conveyor to truck 6.1.4.2 Minnekahta Limestone Source: Rapid City, South Dakota Compressive strength (LBS/SQ IN): 22,000 6.1.4.3 Pahasapa Dolomitic Limestone Source: Pringle, South Dakota Condition (Wet, Dry, Frozen): Dry to wet Moisture Percent: 5 to 15 (Temperature): -15 to +100°F Hardness (MOH scale): 3.5 Compressive strength (LB/SQ IN): 13,000 (basal, Highway 40) Physical Characteristics (Bedding, Porosity, etc.): thick bedded to massive, uniform texture Percentage of fines in feed: <1% through 9 mesh Will fines be removed by prescreening? yes Size product desired: cubic to rounded (Shape) 9 X 20 mesh (Tyler size) Fines desired in product: (MAX) Product removed by: conveyor to car or stockpile 6.1.4.4 Madison Limestone Source: Logan, Montana Condition (Wet, Dry, Frozen): Dry Moisture Percent: <5 (Temperature): -30 to + 120°F Hardness (MOH scale): 3 Physical Characteristics (Bedding, Porosity, etc.): 6 IN to massive bedded, average 1 FT Percentage of fines in feed: <1% through 9 mesh Will fines be removed by prescreening? yes Size product desired: cubic to rounded (Shape) 9 X 20 mesh (Tyler size) Fines desired in product: (MAX) Product removed by: conveyor to car or stockpile 219 6.1.4.5 Jefferson Dolomitic Limestone Source: Logan, Montana Condition (Wet, Dry, Frozen): Dry Moisture Percent: <5 (Temperature) -30 to + 120°F Hardness (MOH scale): 3.5 Compressive strength (LB/SQ IN) : too fractured to cut 7/8 IN X 2 IN sample Physical Characteristics (Bedding, Porosity, etc.): Bedding 1 IN to 3 FT average 4 to 6 IN, sucrosic texture Percentage of fines in feed: <1% through 9 mesh Will fines be removed by prescreening? yes Size product desired: cubic to rounded (Shape) 9 X 20 mesh (Tyler size) Fines desired in product : (MAX) Product removed by: conveyor to car or stockpile 220 SECTION 7 PREGASIFI CATION BENEFICIATION This section contains the following South Dakota School of Mines and Technology studies concerning pregasifi cation beneficiation of lignite the final report, dated May 1, 19 74, "Sodium Removal from High -Sodium Lignite by Ion Exchange" and an interim report, dated March 31, 19 73, "Sodium Removal from Lignite by Ion Exchange." 221 7.1 SODIUM REMOVAL FROM HIGH-SODIUM LIGNITE BY ION EXCHANGE 7.1.1 INTRODUCTION The presence of high contents of sodium in coals used as fuel for power plants or for feed to gasification plants is deleterious to these pro- cesses. In the past, selective mining of coals containing relatively low levels of sodium has been used to circumvent this problem. This practice cannot be used indefinitely, and techniques will have to be developed to reduce the sodium content of such coals to acceptable levels . In an earlier study, Clarkson (refer to Subsection 7.2), demonstrated that sodium can be effectively removed from low-sodium coals utilizing ion exchange techniques described by the Bureau of Mines. ^^ The ob- jective of the present investigation was to establish whether the sodium content of high-sodium coals can be lowered by ion exchange to levels acceptable for use in gasification and power plants. Specific parameters investigated were: lignite particle size, solution flow rate, column aspect, and temperature. 7.1.2 EXPERIMENTAL MATERIALS AND TECHNIQUES 7.1.2.1 Materials Two samples of coal were supplied by the Consolidation Coal Company [now Conoco Coal Development Company). Each of these coals was sam- pled and analyzed for sodium, silica, ash, and moisture content. The analyses were conducted by Stearns -Roger, Inc., and are presented on a moisture-free basis in Table 7-1. All of the chemicals used in this investigation were reagent grade in quality. Water was prepared by passing distilled water through an ion exchange column. 222 LIGNITE I SIZE FRACTION 3/8 X 1/4 inch 6X8 mesh 35 X 48 mesh % Na 2 % Si0 2 % Ash % H 2 14.9 9.7 7.2 35.9 16.9 10.1 6.1 33.8 13.7 21.0 8.5 33.4 LIGNITE II % Na 2 % Si0 2 % Ash % H 2 17.3 8.5 6.0 41.3 15.3 9.5 6.3 41.8 15.2 12.5 7.2 38.2 SIZE FRACTION 3/8 X 1/4 inch 6X8 mesh 35 X 48 mesh Table 7-1. CHEMICAL ANALYSES OF THE ASH OF TWO SAMPLES OF LIGNITE INVOLVED IN THE INVESTIGATION 7.1.2.2 Techniques 7.1.2.2.1 Sample Preparation. Samples of each of the lignites were crushed and sized; three fractions were collected and used for study. There were the 3/8 X 1/4 inch, the 6 X 8 mesh, and the 35 X 48 mesh fractions Each of the sized fractions was thoroughly mixed, split into 100-gram portions, and stored in plastic bags to prevent dehydration of the coal. 7.1.2.2.2 Solution Preparation. The exchange solution containing Ca ++ was prepared by combining 60 grams of calcium carbonate and 8 liters of water and bubbling carbon dioxide into the suspension for 45 minutes. The maximum Ca concentration that was obtained with this technique was 320 + 5 PPM The pH of the solution was 6.5. ~ 7.1.2.2.3 Ion Exchange Experiments. Columnar ion exchange experiments were con- ducted utilizing columns of 1-1/4, 1-1/2, or 1-7/8-inch ID. One hundred grams of sized coal were used in each experiment. Solution was perco- lated through the bed at flow rates of 25 ML/MIN and 60 ML/MIN at two temperatures, 6° and 35° C. Temperature was maintained within + 0.1°C with a constant temperature bath. When 6 C was involved, an ice water solution was utilized. 223 Two other parameters were investigated; these were aspect ratio of the column and the effect of recycling exchange solution. In all of the experiments, solution was collected after 1, 2, 4, 8, 16, and 32 minutes of exchange. These solutions were analyzed for sodium with a flame photometer. 7.1.2.2.4 Experimental Reliability. One experiment was repeated eight times to establish the reliability of the experimental results obtained with this procedure. Lignite I was used in the study, and the amounts of solium removed after 32 minutes of exchange in the presence of 320 PPM Ca were: EXPERIMENT 1 2 3 4 5 6 7 8 Na REMOVED 12.1 13.1 12.5 12.4 12.5 12.4 12.7 12.6 The average percent of sodium removed was 12.5 ^ 0.7 at a level of con- fidence of 95 percent. An additional measure of experimental reliability can be obtained by com- paring the data obtained with the 35 X 48 mesh fraction and aspect ratio, 2.5:1, in Figures 7-1 and 7-7. These experiments were run separately and sodium removal as a function of time was as follows: EXCHANGE TIME (MIN) SODIUM REMOVAL (%) EXPERIMENT 1 EXPERIMENT 2 (Figure 7-1) (Figure 7-2) 2 4 8 16 32 5.3 8.0 15.6 27.6 50.7 5.3 8.9 16.0 29.4 52.5 7.1.3 EXPERIMENTAL RESULTS The first series of experiments involved establishing the extent of sodium removal as a function of particle size and exchange time. The results obtained with both samples of lignite are + presented in Fig 7-1 through 7-4. As can be seen, exchange of Ca + for Na ures is increased 224 100 80 O 35 x 48 mesh A 6 x 8 mesh 3 1 . u a 8 X 4 1nc ^? > o E QJ a: -a o 60 10 20 Time (min) 30 40 Figure 7-1. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND PARTICLE SIZE. T = 25QC, FLOW RATE = 25 ML/MIN BED DIAMETER = 1-7/8 IN, Ca ++ = 320 PPM. 225 100 80 ^? «T3 > O E (U en E =3 •i— "O o en 60 O 35 x 48 mesh A 6 x 8 mesh □ g- x j inch 10 20 Time (min) 30 40 Figure 7-2. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND PARTICLE SIZE. T = 25°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-7/8 IN, Ca ++ = 320 PPM 226 100 80 s< rt3 > O E CD Cxi O 60 40 20 O 35 x 48 iiiesh A 6 < 8 mesh 3 1 . . u g < ^ inch 40 Figure 7-3. SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND PARTICLE SIZE. T = 25°C, FLOW RATE = 25 ML/MIN, BED DIAMETER = 1-7/8 IN, Ca ++ = 320 PPM 227 10 20 Time (min) 30 Figure 7-4 SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND PARTICLE SIZE. T = 25°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-7/8 IN, Ca ++ = 320 PPM 223 as the particle size is decreased and as the flow rate is increased. In the case of Lignite II , the amount of Na exchanged increased by a fac- tor of 1.5 for the finer size fraction when the flow rate was increased from 25 to 60 ML/MIN. With the higher flow rate, 47 percent of the sodium was removed from the 6X8 mesh fraction of Lignite II, while 83 percent of the sodium was removed from the 35 X 48 mesh fraction after 32 minutes of exchange. Temperature of the exchange solution was investigated in the next series of experiments. (See Figures 7-5 and 7-6.) Exchange of Ca for Na was decreased by a factor of about 1.2 in the case of the 35 X 48-mesh frac- tions of both lignites when the temperature was decreased from 25° to 6°C. In the case of Lignite II, 73 percent of the total sodium was still removed after 32 minutes of exchange with a flow rate of 60 ML/MIN. Bed size was investigated for both samples of lignite at two flow rates of exchange solution. Aspect ratios of 7.8, 4.5, and 2.5, bed height to bed diameter, were employed. The results obtained with the three size fractions of Lignite I at 25 ML/MIN and 60 ML/MIN are shown in Figures 7-7 through 7-12. In the case of the 35 X 48 mesh fraction, aspect ratio affected the rate of exchange significantly. The optimum ratio of the three column aspects employed was 4.5:1. As the particle size of lignite was increased and as the solution flow was reduced, the rate of exchange of Ca for Na became less sensitive to column geometry. (See Figures 7-10 through 7-13.) The results obtained with Lignite I when the solution temperature was 6°C are presented in Figures 7-13 through 7-16. In the cases of the finer size ranges, an aspect ratio of 4.5:1 was again found to be the most favorable. Similar experiments were conducted with Lignite II, and similar results were obtained with the exception of the 35 X 48 mesh fraction with a flow rate of 60 ML/MIN. As can be noted in Figure 7-17, exchange rate was independent of aspect ratio under these conditions. The final series of experiments involved establishing the extent to which exchange solutions can be recycled. In these experiments the solution collected from one pass through a bed of lignite was recycled and used as the exchange solution for a fresh bed of coal. This oper- ation was repeated twice, and the results obtained with the three size fractions of Lignite I are shown in Figures 7-18, 7-19, and 7-20. With the coarser size fractions, the same amount of Na was removed with each of the three passes through the beds. However, when the 35 X 48 mesh fraction was involved, the ability of the exchange solution to remove Na upon recycling was drastically impaired. Similar observations were made with Lignite II. (See Figures 7-21, 7-22, and 7-23.) 229 100 80 O 35 x 48 mesh A 6 x 8 mesh B I* I inch <*« > o E o 60 40 20 O 35 x 48 mesh A 6 x 8 mesh a g x j inch 10 20 Time (min) 30 40 Figure 7-6. SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND PARTICLE SIZE. T = 6°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-7/8 IN, Ca ++ = 320 PPM 231 > o Od o 10 20 Time (min) 30 40 Figure 7-7. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 35 x 48 MESH, T = 25°C, FLOW RATE = 25 ML/MIN, Ca ++ = 320 PPM 232 a* > o E oc E ■o O V) 20 Time (min) 40 Figure 7-8. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 35 x 48 MESH, T = 250C, FLOW RATE = 60 ML/MIN, Ca ++ = 320 PPM 233 >« «3 > O E cu en o 10 20 Time (min) 30 Figure 7-9. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE =6x8 MESH, T = 25°C, FLOW RATE = 25 ML/MIN, Ca ++ = 320 PPM 234 100 80 yj — 60 > -o O 00 40 20 Aspect Ratio O 7.8 A 4.5 □ 2.5 20 Time (min) 40 Figure 7-10. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE =6x8 MESH, T = 25°C, FLOW RATE = 60 ML/MIN, Ca ++ = 320 PPM 235 »* > o ■o o CO 20 Time (min) 40 Figure 7-11. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 3/8 x 1/4 IN, T = 25°C, FLOW RATE = 25 ML/MIN, Ca ++ = 320 PPM 236 100 80 - ^ > o a: E -D O CO 60 40 - 20 - Aspect Ratio O 7.8 — A 4.5 - □ 2.5 fl Ti i | | 1 1 10 20 Time (min) 30 40 Figure 7-12. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 3/8 x 1/4 IN, T = 25°C, FLOW RATE = 60 ML/MIN, Ca ++ = 320 PPM 237 <* o o 20 Time (min) 40 Figure 7-13. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 35 x 48 MESH, T = 6°C, FLOW RATE = 60 ML/MIN, Ca ++ = 320 PPM 238 *« > o E on T3 O to uu Aspect Ratio 7.8 • 80 A 4.5 □ 2.5 60 - 40 20 y^ i I i l I 1 1 10 20 Time (min) 30 40 Figure 7-14. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 35 x 48 MESH, T = 6°C, FLOW RATE = 25 ML/MIN, Ca ++ = 320 PPM 239 &« > o E •f— -a o m 10 20 Time (min) 30 Figure 7-15. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE =6x8 MESH, T = 6°C, FLOW RATE = 25 ML/MIN, Ca ++ = 320 PPM 240 100 80- »« > o § a: E =J •r- T3 O 60- 40- 20- - Aspect Ratio 7.8 wm A 4.5 - □ 2.5 - 91 i^ i i i i i 10 20 Time (min) 30 40 Figure 7-16. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 3/8 x 1/4 IN, T = 6°C, FLOW RATE = 25 ML/MIN, Ca + + 320 PPM 241 100 80 ~ &« — 60 - «3 > O E O) q: E 3 XJ O 00 40 - 20 10 20 Time (min) 30 40 Figure 7-17. SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND ASPECT RATIO. SIZE = 35 x 48 MESH, T = 25°C, FLOW RATE = 60 ML/MIN, Ca ++ = 320 PPM 242 100 80 *« ~ 60 O Wash A Recycle 1 □ Recycle 2 > o I E •r— -a o 40 10 20 Time (min) 40 Figure 7-18. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND SOLUTION RECYCLING. SIZE = 35 x 48 MESH, T = 25°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-1/4 IN, Ca ++ = 320 PPM 243 100 80 *? > o E o 60 40 20 O Wash A Recycle 1 B Recycle 2 20 Time (min) 30 40 Figure 7-19. SODIUM REMOVAL FROM LIGNITE I AS A FUNCTION OF TIME AND SOLUTION RECYCLING. SIZE =6x8 MESH, T = 25°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-1/4 IN, Ca ++ 320 PPM 244 TOO 80 &« to > o E o i E 3 •r— •a o CO 20 Time (min) Figure 7-21. SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND SOLUTION RECYCLING. SIZE = 35 x 48 MESH, T = 25°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-1/4 IN, Ca ++ = 320 PPM 246 TOO 80 »« — 60 <0 > o E a: T3 o O Wash A Recycle 1 b Recycle 2 40 10 20 Time (min) 30 40 Figure 7-22. SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND SOLUTION RECYCLING. SIZE =6x8 MESH, T = 25QC, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-1/4 IN, Ca ++ '= 320 PPM 247 100 80 a* ~ 60 CO > o cc o 40 20 O Wash A Recycle 1 □ Recycle 2 20 Time (min) 40 Figure 7-23. SODIUM REMOVAL FROM LIGNITE II AS A FUNCTION OF TIME AND SOLUTION RECYCLING. SIZE = 3/8 x 1/4 IN, T = 25°C, FLOW RATE = 60 ML/MIN, BED DIAMETER = 1-1/4 IN, Ca ++ = 320 PPM 248 SUMMARY OF RESULTS Sodium is contained in lignite as water-soluble sodium silicate or as a salt of humic acid. That which is present as sodium silicate can be removed by water washing whereas the remainder must be removed by ion exchange. Exchange with Ca can be written as follows: 2 Na + + Ca ++ -*• Ca ++ + 2 Na + where underscoring denotes that the ion is contained in the lignite. Ion exchange in this system is probably controlled by diffusion of ions into and out of the solid. This fact is borne out by the exchange data pre- sented in Figures 7-1 through 7-4. That is, as the particle size is decreased, the diffusion path length is reduced and exchange rate is increased. The experimental results also show that sodium removal is a function of solution flow rate. Since the kinetics are probably controlled by dif- fusion of Na within the solid, it seems likely that the rate dependence on flow rate may be due to the lower Na concentration that would be pre- sent in the solution surrounding the coal particles with the higher flow rates . Assuming that a level of 5 percent sodium oxide in the ash is acceptable, grinding the coal to -35 mesh and conducting exchange for approximately 30 minutes with a solution containing 320 PPM Ca ++ at a flow rate of 60 ML/MIN will accomplish this. With coal eight times coarser than this size (6X8 mesh), only about one-third to one-half of the sodium can be removed under these conditions. With the 3/8 X 1/4 inch material, only about 15 percent of the sodium is removed. Since the exchange reactions are diffusion controlled, the rates will be temperature dependent. In the case of Lignite I, for example, the total removal of sodium from the 35 X 48 mesh fraction was reduced from 73 to 62 percent when the temperature was reduced from 25° to 6°C, while in the case of Lignite II, sodium removal was reduced from 85 to 73 per- cent. From a processing standpoint, recycling exchange solutions would be ad- vantageous. In this view, experiments were run to establish the amount of sodium that could be removed utilizing solution that had been contacted previously with coal. As shown by the data in Figure 7-21, equilibrium was reached when the Na concentration was about 440 PPM. Once this concentration is attained, additional Ca + would have to be introduced into the solution for its reuse through contact with calcium carbonate and carbon dioxide. Column aspect ratio was the last parameter investigated. With coarse lignite (3/8 X 1/4 inch), column geometry had no effect on exchange rate. In the case of finer material, however, an aspect ratio of 4.5:1, 249 height to diameter, resulted in faster exchange rates than when the aspect ratio was 7.8:1 and 2.5:1. When the rate of exchange is rela- tively fast, as with the 35 X 48 mesh fraction of Lignite II (Figure 7-17), exchange is also found to be independent of column geometry. 7.1.5 CONCLUSIONS The following conclusions can be drawn from this investigation: (1) Assuming that a content of 5 percent Na 2 in the ash is acceptable, contacting coal that contains about 15 percent Na 2 in the ash and that has been reduced in size to approximately minus 35 mesh with an exchange solution containing 320 PPM Ca at a flow rate of 60 ML/MIN for 32 minutes will lower the sodium content to this level. (2) Lowering the temperature of the exchange solution reduces the rate of exchange of Ca for Na . In the case of the 35 X 48 mesh frac- tion, exchange of Ca for Na is decreased by a factor of about 1.2 when the temperature is reduced from 25° to 6°C. (3) Acceptable levels of Na are obtained with the 35 X 48 mesh fraction of Lignite II even at 6°C when the exchange solution flow rate is 60 ML/MIN. (4) Increasing the flow rate of exchange solution increases the rate of exchange of Ca for Na . In the case of the 35 X 48 mesh and 6X8 mesh fractions, exchange of Ca for Na is increased by a factor of about 1.4 when the flow rate is increased from 25 ML/MIN to 60 ML/MIN. Essentially no increase in exchange of Ca ' for Na + in the 3/8 X 1/4 inch fraction is obtained with this same increase in solution flow rate. (5) Aspect ratio affects the rate of exchange of Ca ' for Na + in the 35 X 48 mesh fraction significantly. An aspect ratio of 4.5:1, bed height to bed diameter, is optimum. (6) As the particle size of lignite is increased and as the solution flow rate is reduced, the rate of exchange of Ca for Na becomes less sensitive to column geometry. (7) The extent to which exchange solutions may be recycled is a function of particle size. Equilibrium is reached in this system when the Na + concentration is about 440 PPM. (8) Exchange solution may be recycled after equilibrium has been attained by contacting the solutions with calcium carbonate and carbon dioxide again. 250 7.2 SODIUM REMOVAL FROM LIGNITE BY ION EXCHANGE INTERIM REPORT, MARCH 31, 1973 7.2.1 INTRODUCTION The major contaminant of lignite is sodium. Where the lignites are used in high- temperature applications, whether it be in a gasification process or as a fuel in an electric power plant, the contained sodium is oxidized to sodium oxide (Na 2 0) . Because the Na 2 melts at 1682°F and has a fairly high vapor pressure, it passes into the offgases and is carried along until it deposits on cooler surfaces. The result is a coating of reactor vessel walls in the gasification process or a foul- ing of the boiler tubes in power plant usage. The deposits of condensed Na 2 continue to build up until the loss of efficiency in heat transfer forces a shutdown of the unit so that it can be cleaned. Not only is this an unpleasant task but it is also expensive. On boiler tubes the Na 2 crusts can grow to several inches in thickness and can be very tenacious. In addition to the selective mining of low-sodium lignites, two additional approaches have been used in dealing with the problems caused by the sodium in lignite. One route has been to develop additives which when added to the lignite prior to combustion are supposed to change the physical characteristics of the Na 2 crust. The goal of this technique is either a self-cleaning through spalling or, at the minimum, to form a crust which breaks easily from the surfaces when they are vibrated by knocking. These additives have given mixed results and are successful only part of the time. They have not been generally accepted. A more promising approach to controlling the sodium problem has been developed by the Bureau of Mines . ( 22 ) They have demonstrated that the sodium can be removed from lignite by ion-exchange techniques. When the sodium lignite is placed in contact with solutions containing other ions, the sodium passes from the lignite into solution and is replaced with the ions originally in solution. The process can be represented by the following reaction: lignite-2 NA + Ca (water) = lignite-Ca ++ + 2 Na + (water) In their ion-exchange experiments, a mixture of lignite particles in water solutions containing various cations (added as chloride salts) resulted in a slurry which was stirred for various periods of time. At the end of each run the lignite was filtered out of suspension and the solution was analyzed for the original cation remaining and for sodium. In these batch studies, solutions of KC1, CaCl 2 , MgCl 2 , A1C1 3 , FeCl 2 , FeCL 3 , Fe 2 (S0 4 ) 3 , and A1 2 (S0 4 ) 3 were used. In addition to the various cations, the variables studied were ionic strength, pH, particle size, liquid-to-solid ratio, and time. 251 The studies demonstrated that the sodium contained in lignite is present in an exchangeable form with the organic fraction and that the ion- exchange technique is effective in removing a major portion of the sodium. In essence, the lignite is a cation exchange resin which is in the sodium form. Loading the exchanger with other cations drives the sodium into solution. Our initial efforts on sodium removal from samples of lignite from the Glenharold Mine were directed along lines similar to those followed by the Bureau of Mines. The results of this work are given under the stirred batch sections of this report and in general confirm the find- ings of the work done by the Bureau of Mines. Because chloride ions in the lignite would make it unacceptable for gasification by the acceptor process, a new series of experiments was designed which would accomplish the removal of sodium from lignite without using chlorides. In addition, it was desirable to keep the projected process as inexpensive as possible. This was accomplished by using reagents which simulated those raw materials which are used or produced in the gasification process. Specifically, there were two changes involved in the revised experiments. First, the cation containing exchange solutions was prepared by bubbling CO2 through water in the presence of calcium carbonate, the primary constituent of limestone. This produces an effective calcium solution (about 300 PPM) which is almost neutral (pH = 6.4) and contains no chlorides. Secondly, the equipment was changed from a stirred batch system to a column ion-exchange arrangement to allow for a more complete removal of sodium. The results from the modified experiments using column ion-exchange and nonchloride calcium solutions to remove sodium from lignite as well as the preparation of calcium wash solutions are presented in this subsection. 7.2.2 PROCEDURES FOR STIRRED BATCH ION EXCHANGE To remove sodium from lignite by stirred batch methods, a slurry is made of lignite particles in water solutions containing various cations added as chloride salts. The slurry is placed in a beaker and stirred to keep the solids in suspension while the ion-exchange process is taking place . In the ion-exchange process, the ions initially in solution penetrate the lignite particles and replace the sodium originally in the lignite. The displaced sodium then comes out of the lignite particle and into the water solution. At the end of each run, the lignite was filtered out of suspension and the solution analyzed for sodium. Experiments were run using solutions of CaCl2, MgCl£, and distilled water to remove sodium. The variables studied were time, solution concentration, pH, particle size, temperature, and solidrliquid ratio. The experimental 252 data and results are tabulated in Appendix E, Book 2 of this vol ume, The lignite samples used are not all high-sodium lignites which would require beneficiation before using. The samples were collected or supplied from active mines which presumably are shipping lignite having a satisfactorily low sodium. 7.2.3 RESULTS OF STIRRED BATCH ION EXCHANGE 7.2.3.1 Lignite from Beulah, North Dakota The first lignite studied was from Beulah, North Dakota, and had 10.8 percent ash and 0.51 percent sodium (Na determined on an air-dry basis). The sodium removal data is given in Figures 7-24 through 7-28. Magne- sium chloride solutions were not used to remove sodium from the Beulah lignite nor was the effect of the solid: liquid ratio studied. The figures show that sodium removal from Beulah lignite is favored by high temperatures and concentrated calcium solutions. Over half of the sodium is removed in 15 minutes using 10" 1 molar calcium solutions at 23°C. Sodium is removed more effectively from the finer size frac- tions. The Beulah lignite appears to contain a significant amount of water-soluble sodium because distilled water removes almost as much sodium as does 10" 3 molar solutions of calcium. The sodium removal is only slightly pH dependent with acid solutions being somewhat more effective in sodium removal by ion exchange. 253 100 90 80 T 1 1 1 J 1 1 1 1 1 1 1 r 50 qm lignite : 150 ml soln pH = 6.2, -14+20 mesh CaCl 2 = 10" 3 molar > O E 70 60 50- Q \° 40- 30 J I L 50 T = 23 T r T = 3-6 °C j L 100 Time (minutes) Figure 7-24. SODIUM REMOVAL FROM BEULAH, NORTH DAKOTA LIGNITE BY ION EX- CHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF TEMPERATURE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 254 t 1 1 r t r t — T > o E O Z TOO 90 50 gm lignite : 150 ml soln pH = 6.5, -14+20 mesh CaCl 2 = 10 molar 100 Time (minutes) Figure 7-25. SODIUM REMOVAL FROM BEULAH, NORTH DAKOTA LIGNITE BY ION EX- CHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF TEMPERATURE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 255 100 90 Two hours contact time 50 gm lignite : 300 ml soln T = 23 °C, -10+14 mesh 80 70 CaCl 2 = 10 molar > o E 60 50 n\° 40 30 20 10- CaCl 2 = 10" molar ■• Distilled H 2 6 pH Figure 7-26 SODIUM REMOVAL FROM BEULAH, NORTH DAKOTA LIGNITE BY ION EX- CHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF pH AND CALCIUM ION CONCENTRATION. 256 100 > o E 90 80- 70 60- 5C- 4C- 3C- 2C- 1C- CaCl 2 = 10" molar Two hours contact time 50 gm lignite : 150 ml soln T = 3-6 °C, -14+20 mesh -L 6 PH Figure 7-27. LOW TEMPERATURE SODIUM REMOVAL FROM BEULAH, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF pH 257 50 40 > O E 0) 30 20 10- 1 Two hours contact time 50 gm lignite : 300 ml soln T = 23 °C A CaCl 2 = 10" 3 molar, pH = 3.1 ■ CaCl 2 = 10" 3 molar, pH = 5.7 ▼ Distilled H 2 0, pH = 3.1 • Distilled H 2 0, pH = 5.7 J. -6+10 -14+20 Mesh Size ■20 Figure 7-28 SODIUM REMOVAL FROM BEULAH, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE, pH, AND CALCIUM ION CONCENTRATION 258 Lignite from Stanton, North Dakota The lignite from Stanton, North Dakota, was supplied by the Consolidation Coal Company from the Glenharold Mine. This lignite had 6.7 percent ash and 0.47 percent sodium (Na determined on an air-dry basis). The sodium removal data is given in Figures 7-29 through 7-34. The figures show that more sodium is removed faster from the finer size fractions. The Stanton, North Dakota, lignites do not contain much water-soluble sodium and the 10-3 molar solutions of calcium and magnesium remove less than 15 percent of the sodium in 2 hours The ifalZ re TK° Va i- in these J li g nites ^ not pH dependent to any significant degree. The figures indicate that sodium is best removed in 10-1 molar solutions of calcium or magnesium at 23°C. Under these conditions a ' slurry of 10 percent solids results in complete sodium removal from Glenharold Mine lignite in 2 hours. 259 10 "X5 O E to at D \° -28+48 mesh 50 gm lignite : 150 ml soln T = 24 °C, pH = 8.0 Distilled H 2 Time (hoursj Figure 7-29. SODIUM REMOVAL FROM GLENHAROLD MINE LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 260 14 12 50 gm lignite : 150 ml sol n T = 24 °C, pH = 8.0 MgCl 2 = 10' 3 molar 28+48 mesh > o E & Time (hours) Figure 7-30. SODIUM REMOVAL FROM GLENHAROLD MINE LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 261 100 90 50 gm lignite : 150 ml soln Two hours contact time T = 24 °C, -8+14 mesh 80 no > o E 70 60 50 CaClo = 10" molar MgClp = 10 molar <* 40 30 20 MgCl 2 = 10" J molar 10 CaCl ? = 10" molar 6 pH Figure 7-31. SODIUM REMOVAL FROM GLENHAROLD MINE LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 262 > o E o o Z 100- 90- 80- 70- 6C- 5C- 4C- 3C- 2C- 1C- T Two hours contact time 50 gm lignite : 150 ml soln T = 3-6 °C, -14+20 mesh CaCl 2 = 10" 1 molar CaCl 2 = 10" 3 molar Distilled H 9 6 PH Figure 7-32. LOW TEMPERATURE SODIUM REMOVAL FROM GLENHAROLD MINE LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF pH AND CALCIUM ION CONCENTRATION 263 0) > o E n\° ■ l 1 1 1 f 1 i 100 Two hours contact time " 90 50 qm lignite : 300 ml soln T = 3-6 °C, -14+20 mesh 80 - 70 - - 60 - 50 - 40 - - 30 - 20 - 10 A . . „., „ ▲ _3 CaCU = 10 molar 1 ▲ * — . Distilled H o pH Figure 7-33. LOW TEMPERATURE SODIUM REMOVAL FROM GLENHAROLD MINE LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF pH AND TWO CALCIUM CONCENTRATIONS 264 > O E (D D \° " nr 1 T ~r i i 1 1 r- 00 - - 90 - • ^^ - 80 - • - 70 - /% - 60 - 50 - 40 •/ - 30 Two hours contact time 25 gm lignite, -8+14 mesh ™ 20 T = 24 °C, pH = 5.8 MgCl 2 = 10" molar - 10 -/ - • ' • .... 1 1 1 — i i i 1:1 1:3 1:6 1:9 Solid: Liquid Ratio 1:12 Figure 7-34. THE EFFECT OF PERCENT SOLIDS IN AGITATED SLURRIES ON THE SODIUM REMOVAL BY ION EXCHANGE FROM GLENHAROLD MINE LIGNITE 265 7.2.3.3 Lignite from Burke County, North Dakota A lignite sample from Burke County, North Dakota, contained 8.7 percent ash and 1.33 percent sodium (Na determined on an air-dry basis). The sodium removal data is given in Figures 7-35 through 7-40. The lignite contains more sodium than any of the others which were tested. The sodium is more effectively removed from the finer size fractions. In these finer size fractions, about 10 percent of the sodium is removed by either distilled water or 10"^ molar solutions of calcium or magnesium. The higher concentration magnesium or calcium solutions, 10" * molar, more effectively removed sodium than the dilute solutions . The sodium removal from this higher sodium lignite is not as strongly affected by temperature as are the lower sodium lignites. Only slightly more sodium is removed at 2 4°C than at 3 to 6°C. The higher sodium lignites respond differently to ion exchange and Figure 7-40 shows that only three -fourths of the sodium can be removed. A good portion of the sodium is evidently present in the high-sodium lignite in a form which does not readily respond to a simple ion-exchange procedure. While higher temperatures and more concentrated solutions might have removed more sodium, it did not seem practical to extend the data into these regions. Other variables being equal, the calcium and magnesium solutions had about equal sodium removing ability. The sodium removal in these lignites is not pH dependent to any significant degree. 266 > O E O E © O Z >° 6 s 14 12- 1C_ 50 qm lignite : 150 ml soln T = 24 °C, pH = 8.0 CaCl 2 = 10" 3 molar •28+48 mesh ■14+28 mesh -8+14 mesh Time (hours) Figure 7-36. SODIUM REMOVAL FROM BURKE COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 268 14- 50 gm lignite : 150 ml soln T = 24 °C, pH = 8.0 -3 MqCl 2 = " molar 12- > o E Time(hoursj Figure 7-37. SODIUM REMOVAL FROM BURKE COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 269 T 70 Two hours contact time 50 gm lignite : 150 ml soln T = 24 °C, -8+14 mesh > O E a> O Z 60 50 40 30 20 MgCl 2 = 10 molar CaClp = 10 mola 10 _3 CaClp =10 molar — • • MgCl 2 =10 molar 6 pH Figure 7-38. SODIUM REMOVAL FROM BURKE COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 270 0) > O E o & 70 60 Two hours contact time 50 gm lignite : 150 ml soln T = 3-6 °C, -8+14 mesh 50 40 30 20 CaClp = 10" molar MgCl 2 = 10 molar 10 CaCl ? = 10" molar = .3 • A MgCl 9 = 10 molar, 6 PH Figure 7-39. LOW TEMPERATURE SODIUM REMOVAL FROM BURKE COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 271 > O E Z 10C- 9C- 80 70- 60 50 40- 30 20 10 Two hours contact time 25 gm lignite T = 24 °C, pH 5.9 CaCl 2 = 10" 1 molar -8+14 mesh 1:1 i : 3 1:6 1:9 1:12 1:15 1:18 1:21 1:24 Solid : Liquid Ratio Figure 7-40. THE EFFECT OF PERCENT SOLIDS IN AGITATED SLURRIES ON THE SODIUM REMOVAL BY ION EXCHANGE FROM BURKE COUNTY, NORTH DAKOTA LIGNITE 272 7.2.3.4 Lignite from Bowman County, North Dakota Lignite from Bowman County, North Dakota, contained 13.0 percent ash but only 0.16 percent sodium (Na determined on an air -dry basis). The sodium removal data is given in Figures 7-41 through 7-46. This low-sodium lignite contains a large fraction of water-soluble sodium. The distilled water treatment removes about 45 percent of the contained sodium over 8 hours. The dilute 10" 3 molar calcium or magnesium solutions remove the same 45 percent of the sodium in about 15 minutes. The effect of particle size on sodium removal is not as critical in this low-sodium lignite. The spread of values in sodium removed from different particle sizes using dilute ion solutions is smaller than when treating higher sodium lignites. The sodium removal from this lignite is not very temperature dependent. Only slightly more sodium is removed at 24°C than is removed at 3 to 6°C. The sodium removed has little pH dependence and calcium and magnesium are about equally effective in causing sodium removal. The exception to this occurs at 24°C where the 10~1 molar calcium removes more sodium than does an equal concentration of magnesium. The sodium can be almost totally removed from this low-sodium lignite by ion exchange in a single contact-stirred slurry batch. Figure 7-46 shows that a slurry of 5 percent solids in a 10" * molar calcium solu- tion at 24°C will accomplish this in 2 hours and similar slurry of 25 percent solids will result in 90 percent removal in the same time. 273 -a o > o E # 70 60 50 qm lignite : 150 ml soln T = 24 °C, pH = 7.8 Distil le H 2 -28+48 mesh Time (hours) Figure 7-41 SODIUM REMOVAL FROM BOWMAN COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 274 70 60 > o E 50 qm lignite : 150 ml soln T = 24 °C, pH = 8.1 CaClp =10 molar •28+48 mesh -14+28 mesh ■8+14 mesh Time (hours) Figure 7-42. SODIUM REMOVAL FROM BOWMAN COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 275 70 60- 50 gm lignite : 150 ml soln T = 24 °C, pH = 8.0 -3 MaCl 2 = ' molar > o E a: o Z >P 2 3 Time (hours) Figure 7-43, SODIUM REMOVAL FROM BOWMAN COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 276 7C" -A- CaCU = 10" 1 molar A— 6C" MqCl 2 = 10 molar > o E 50. 4C - -3 MqCl ? =10 molar CaCl 2 = 10 molar 3C- 2C- 1C- Two hours contact time 50 am liqnite : 150 ml soln T = 24 °C, -8+14 mesh PH Figure 7-44. SODIUM REMOVAL FROM BOWMAN COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 277 "0 > O E a" 70 60 50 40 CaClp = IP" 1 molar ^r MqCl 2 = 10 molar CaCl 2 = 10" molar MgCl 2 = 10" 3 molar 30" Two hours contact time 50 gm lignite : 150 ml sol 'n T = 3-6 °C, -8+14 mesh H 6 pH Figure 7-45. LOW TEMPERATURE SODIUM REMOVAL FROM BOWMAN COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 278 10C- o > o E 0) o\° Two hours contact time -8+14 mesh CaCU = 10" molar 25 gm of lignite pH = 6.0 1:1 1:3 1:6 1:9 1:12 1:15 1:18 1:21 Solid: Liquid Ratio 1:24 Figure 7-46. THE EFFECT OF PERCENT SOLIDS IN AGITATED SLURRIES ON THE SODIUM REMOVAL BY ION EXCHANGE FROM BOWMAN COUNTY, NORTH DAKOTA LIGNITE 279 7.2.3.5 Lignite from Ward County, North Dakota The lignite from Ward County, North Dakota, had 5.6 percent ash and a very low 0.05 percent sodium (Na determined on an air-dry basis). The sodium removal data is given in Figures 7-47 through 7-52. The sodium removal from this low-sodium lignite is very similar to the low-sodium lignite from Bowman County, North Dakota. Again, there is a large fraction of water-soluble sodium present and the sodium removal is not very dependent on particle size. The higher concentration 10"! molar solutions are more effective in removing sodium than are the 10~3 molar solutions. The sodium removal from this low-sodium lignite is only slightly temperature dependent. Sodium removal at 24° C is not much more than that at 3 to 6°C. The sodium removed has little pH dependence and calcium and magnesium are about equally effective in causing sodium removal. The exception to this occurs at 24° C where the 10" 1 molar calcium removes more sodium than does an equal concentration of magne- sium. The sodium can be almost totally removed from this low-sodium lignite by ion exchange in a single contact-stirred slurry batch. A slurry of 5 to 10 percent solids in a 10~1 molar calcium solution at 24°C will remove more than 90 percent of the contained sodium in 2 hours. 280 70 r 1 1 50 gm lignite : T = 24 °C, pH Distilled H 2 1 ! 150 ml soln = 8.2 60 - > o E o 28+48 mesh \° Time (hours) Figure 7-47. SODIUM REMOVAL FROM WARD COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 281 70 60 50 gm lignite : 150 ml soln T = 24 °C, pH = 8.0 _3 CaCU = 10 molar 28+48 mesh ■o O E \° -8+14 mesh Time (hours) Figure 7-48. SODIUM REMOVAL FROM WARD COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF PARTICLE SIZE AND CONTACT TIME BETWEEN LIGNITE AND SOLUTION 282 70 50 gm lignite : 150 ml soln T = 24 °C, pH = 8.0 MgCl 2 = 10" 3 molar 60 > o E an Q \° ■ -28+48 mesh -14+28 mesh -8+14 mesh Time (hours) Figure 7-49, SODIUM REMOVAL FROM WARD COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A SOLUTION ° F PARTICLE SIZE AN ° C0NTACT TIME BETWEEN LIGNITE AND 283 no o > o E 100 90 80 70 60 50 T Two hours contact time 50 gm lignite : 150 ml soln T = 24 °C, -8+14 mesh CaCl = 10" molar 2 4 MqCl 2 = 10 molar O z 40 30 MgCl 2 = 10 molar CaClp = 10 molar 20 10 pH Figure 7-50. SODIUM REMOVAL FROM WARD COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 284 > O E © o z 70 60 50 - 40 30 20 10 Two hour contact time 50 gm lignite : 150 ml soln T = 3-6 °C, -8+14 mesh CaCl 2 = 10 molar A MgCl 2 =10"' molar CaCl 2 = 10" molar MgClp = 10 molar • PH Figure 7-51. LOW TEMPERATURE SODIUM REMOVAL FROM WARD COUNTY, NORTH DAKOTA LIGNITE BY ION EXCHANGE IN AGITATED SLURRIES. THE REMOVAL WAS STUDIED AS A FUNCTION OF ION TYPE, pH, AND ION CONCENTRATION IN SOLUTION 28S 0) > o E o Z ■ T - 1 ■ 1 1 II i " — T— 100 - - 90 - 80 - - 70 - /• - 60 - • / - 50 - - 40 - 30 Two hours contact time T = 24 °C, -8+14 mesh - 20 - / CaCl 2 = 10" molar 25 gm lianite pH = 6.0 - 10 "I ~ /, 1 ' 1 1 i i i i __ -4— 1:1 1:3 1:6 1:9 1:12 1:15 1:18 1:21 Solids Liquid Ratio 1:24 Figure 7-52. THE EFFECT OF PERCENT SOLIDS IN AGITATED SLURRIES ON THE SODIUM REMOVAL BY ION EXCHANGE FROM WARD COUNTY, NORTH DAKOTA LIGNITE 286 7.2.4 PROCEDURES FOR COLUMN ION EXCHANGE 7.2.4.1 Materials The calcium solution was prepared by placing 30 grams of CaC0 3 in 4 liters of distilled water and then bubbling C0 2 through the mixture for 30 minutes at room temperature. The result, after filtering out the excess CaCO 3 , was a solution containing about 300 PPM Ca ++ at pH = 6.4. The batch of lignite used in these experiments was from the Glenharold Mine, Stanton, North Dakota. The lignite was from a lot used in a crushing study at the Rapid City C0 2 Acceptor Process Gasification Pilot Plant. The batch was carefully sampled and analyzed. The results, 12.2 percent ash and 0.36 percent Na (determined on an air- dry basis) in the lignite, were used as the initial sodium concen- tration for each run. The sample used in each run was all lignite passing through an 8 mesh screen. A detailed size analysis of the minus 8 mesh lignite is given in Table 7-2. This size material was used in all experiments except those specifically designated as to size. The other materials used in this investigation were standard reagent- grade chemicals and distilled water. MESH SIZE SAMPLE 1 WEIGHT PERCENT SAMPLE 2 SAMPLE 3 AVERAGE +6 -6+8 -8+10 -10+14 -14+20 -20+28 -28+35 -35+48 -48+65 -65+100 -100+150 ■150+200 ■200+325 -325 1.75 2.27 3.45 6.67 9.48 11.97 7.60 17.74 10.17 7.40 4.89 5.45 6.87 4.29 2.16 2.44 3.91 7.19 9.75 12.17 7.03 17.27 10.00 7.28 4.74 5.23 6.94 3.86 1.73 2.24 3.76 7.14 9.83 11.97 7.65 17.28 10.04 7.48 4.87 5.22 6.80 4.01 1.88 2.32 3.71 7.00 9.69 12.04 7.43 17.43 10.07 7.39 4.83 5.30 6.87 4.05 Table 7-2. SCREEN ANALYSIS OF THE LIGNITE USED IN COLUMN ION EXCHANGE 287 7.2.4.2 Experimental Design A schematic representation of the experimental arrangement is shown in Figure 7-53. For all experiments, 50 grams of lignite (-8 mesh) were charged into a 1- 1/2-inch-ID glass column where It was supported on a bed of glass wool. Height of the lignite column for this charge was 3 inches. In a few runs, changes in column diameter and height were made in order to determine the effects of these parameters. The calcium solution, at a known pH value, was then passed through the lignite bed by gravity flow at a controlled flow rate. Periodically the volume of solution issuing from the bottom of the column was measured and the solution analyzed for sodium. The results are tabulated in Appendix E, Book 2 of this volume. To make the low- temperature runs, both the lignite samples and the calcium solution were first chilled in a refrigerator. The apparatus was then assembled with ice packed around the container of calcium solution. This maintained the column at a fairly constant low temperature for the duration of the run. 7.2.4.3 Parameters Investigated Variables tested for the nominal minus 8 mesh lignite sodium removal were: (1) Flow rate of wash solution (2) Time (3) pH (4) Temperature (5) Calcium content of wash solutions In addition, runs were made using different solumn diameters and heights in order to determine the effects of these parameters. A limited amount of work was done on specially sized lignite to see the effects of particle size. 7.2.5 RESULTS AND DISCUSSION OF COLUMN ION EXCHANGE The results of sodium removal using calcium solutions at room tempe- rature on - 8 mesh lignite are presented in Figures 7-54 through 7-57. The Ca ++ concentration was maintained at approximately 320 PPM and the low pH values were achieved through small sulfuric acid additions, As can be seen, the sodium removal is highly dependent on the wash solution flow rate through the column of lignite. The highest sodium removal occurs at the fastest flow rates at all pH values. The figures show that the sodium removal is relatively insensitive 288 / flllllll Wash Solution Plastic Tube 500 ml Buret Lignite Sample Glass Wool Sample Beaker Figure 7-53. EXPERIMENTAL ARRANGEMENT FOR SODIUM REMOVAL FROM LIGNITE USING COLUMN ION-EXCHANGE 289 50 100 Time (minutes) 150 Figure 7-54. SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND FLOW RATE AT 24°C, pH 3.13 AND 316 PPM Ca ++ AVERAGE WASH SOLUTION CONCENTRATION 290 50 100 Time (minutes) 150 Figure 7-55, RA?E U AT R 74Sr AL hTA IGNITE AS A FUNCTI0N 0F TIME AND FLOW mmTioN P 4 '° AND 325 PPM Ca++ AVERAGE WASH s °^t.on 291 ■ 51 ml/min a 35 ml/min • 24 ml/min i I i i— i — i — i — j- 50 100 Time (minu tes) Figure 7-56. SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND FLOW RATE AT 24°C, pH 5.6 AND 322 PPM Ca ++ AVERAGE WASH SOLUTION CONCENTRATION 292 50 100 Time (minutes) 150 Figure 7-57. SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND FLOW AVERAGE WASH SOLUTION RATE AT 240C, p H 6.5 AND 321 PPM Ca CONCENTRATION 293 + + to the pH of the calcium wash solutions. The low-pH calcium solutions are no more effective in sodium removal than those solutions close to a neutral pH value. At the highest flow rates used, all of the sodium was removed from the lignite in 2 hours or less. Also, the figures indicate that sodium removal occurs very fast initially and that the reactions slow down with time. At the highest flow rates, the calcium wash solutions remove at least 70 percent of the contained sodium in the first 30 minutes at all pH values. Figure 7-58 represents the completeness of sodium removal when varying the calcium concentration in the wash solutions. No significance is attached to the small separation between the two curves. A compari- son of the curves shows that a relatively concentrated calcium wash solution is no more effective in sodium removal than a more dilute one. The chemical form of sodium as it is contained in lignite is not known precisely. It is thought that most of the sodium is present as salts of humic acids or bound to an organic framework as it would be in an organic ion-exchange resin. However, there is a portion of the sodium which can be removed by distilled water and this is speculated to be in the form of water-soluble sodium minerals. Figure 7-59 shows the amount of sodium removed from Glenharold Mine Lignite by a distilled water wash and Figure 7-60 gives the results determined at a lower pH. Apparently 40 percent of the contained sodium is removed by water. While the above amount of sodium may be present as sodium minerals, there is an ion-exchange mechanism which also can account for the sodium removal by water. Using X to present the organic framework, the relationship is given by: X-Na + +H 2 ■* X-H + + NaOH This reaction places a hydrogen ion on a site formerly occupied by a sodium ion and has been found to be operative in inorganic ion exchangers. C 23 ) Thus, the question of the total chemical forms of sodium in lignite has not been resolved. The effect of low temperatures on sodium removal from lignites by calcium wash solutions was examined and the results are given in Figures 7-61 and 7-62. Again, the higher flow rates result in more sodium being removed. Also, the low-pH calcium wash solutions are no more effective than those near neutrality. The principal result of low temperatures is slower rates of reaction. Although the 294 150 Time (minutes) Figure 7-58. SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND CAL- CIUM CONCENTRATION IN THE WASH SOLUTION AT 24°C, pH 6.4 AND AN AVERAGE FLOW RATE OF 25 ML/MIN 295 # 100 90 80 70 -U 60 > o E 50 O E o ■ 60 ml/min • 25 ml/min 50 100 Time (minutes) 150 Figure 7-64 SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND FLOW RATE AT 1 - 70C, pH 2.9 AND USING A DISTILLED WATER WASH 302 100 T5 > o E o Z # -i 1 1 1 1 i 1 » i i i r h = 1-1/2", d ? 1-7/8" " h = 6-1/2", d = 1" T = 24 W C 50 gm lignite Ca ++ = 320 ppm 25 ml /min sol 'n flow l ' i u 1 20 1 50 Time (minutes) 210 Figure 7-65. SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND COLUMN GEOMETRY 303 t 1 1 r "i i 1 1 1 1 1 r 0) > o e D z

O E O Z s 100 90 80 I I h = 3", d = 1-1/2" h = 1-5/8", d = 1-7/8" 50 gm lignite Ca = 337 ppm 59 ml/min sol 'n flow 90 120 Time (minutes) 180 210 Figure 7-67. SODIUM REMOVAL FROM LIGNITE *AS A FUNCTION OF TIME AND COLUMN GEOMETRY 305 T 1 T i 1 1 1 1 1 r -o > o E 4) <£ 100 90 80 h = 3", d = l-J/2' h = 1-1/2", d = 1-7/8" T = 2 - 8 °C 50 gm lignite Ca = 325 ppm 58 ml/min sol 'n flow 150 1 ■ 180 210 Time (minutes) Figure 7-68, LOW TEMPERATURE SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND COLUMN GEOMETRY 306 i i 1 1 r t— i 1 1 1 1 r > o £ D z 100- 90" 80" 70- 60- 50- 40- 30- 20- T = 24' °C 50 gm lignite Ca ++ = 330 ppm 60 ml/min sol 'n flow h = 3", d = 1-1/2" i. 60 90 120 Time (minutes) 150 180 210 Figure 7-69. SODIUM REMOVAL FROM LIGNITE AS A FUNCTION OF TIME AND PARTICLE SIZE 307 t r t r t r 100 > o E o 50 gm lignite Ca =315 ppm 60 ml/min sol 'n flow h = 3", d = 1-1/2" 60 90 120 150 Time (minutes) 180 210 Figure 7-70. LOW TEMPERATURE SODIUM REMOVAL AS A FUNCTION OF TIME AND PARTICLE SIZE 308 EXPERIMENTS ON PREPARATION OF CALCIUM WASH SOLUTIONS The calcium wash solutions used to remove sodium by column ion exchange from the lignite samples were prepared by bubbling C0 2 through water in the presence of CaC0 3 . The dissolving CaC0 3 places calcium in solution. Industrially, this would probably be accomplished using limestone instead of reagent-grade CaC0 3 . Since calcite (CaC0 3 ) is the principle mineral in many limestones, we studied the rate of calcite dissolution to find out the best way to make wash solutions. The parameters studied—which affect the rate of calcite dissolution-- were particle size, temperature, agitation, CO2 flow rate, and method of bubbling C0 2 through the solution. The agitated calcite slurries were stirred to keep the particles in suspension. The results are given in Figure 7-71 through 7-77. The calcite dissolution experiments began by bubbling CO2 through 4 liters of distilled water for 5 minutes. A styrofoam cover was fitted to the beaker to maintain a C0 2 atmosphere over the solution. Thirty grams of calcite were then placed in the solution. The calcite was dissolved by continuing to bubble C0 2 through the slurry for 75 minutes. Fifty -milliliter samples were taken at specific time intervals, filtered, and titrated to determine the calcium concentration in solution. Figure 7-71 shows that C0 2 introduced through a fritted glass bubbler more effectively dissolves calcite than does C0 2 bubbled through a 4-millimeter glass tube. The fritted glass causes many fine bubbles, having a large surface area, to be formed and the C0 2 gas goes into solution more quickly. Comparison of Figures 7-72 and 7-73 with the other graphs shows that agitation of the slurries is needed to get the maximum amount of calcium in solution in the shortest time. The agitation keeps the particles suspended in solution and the stirring aids in diffusion, both of which increase the rate of calcite dissolution. Figures 7-74 through 7-77 illustrate the effect of C0 2 flow rate, temperature, and particle size on the rate of calcite dissolution. Duplicate runs were made at the higher C0 2 flow rates and with the finer size fractions. The reproducibility is good in all cases except for the -325 mesh calcite and 24°C, and even here the final calcium concentration did not vary more than 10 percent in three runs. However, there is considerable scatter in the approach to equilibrium. This is probably because of difficulties in filtering the samples taken for analysis. 309 300 250 E a a c o 200 S 150 c o u o U ioo -1 1 — T = 24 °C Slurry stirred 1330 ml/min CO, -100+150 mesh Fritted Glass Open Tube 20 30 40 Time (minutes) 50 60 70 Figure 7-71. RATE OF CALCITE DISSOLUTION WHEN SEVERAL METHODS ARE USED TO INTRODUCE C0 2 INTO SOLUTION 310 400 350 1 1 1 1 T = 24 °C Slurry unstirred 60 ml/min C0 2 ▼ -325 mesh • -100+150 mesh ■ -28+35 mesh ▲ -8+10 mesh 30 40 Time (minutes) 50 60 70 Figure 7-72. RATE OF CALCITE DISSOLUTION IN UNSTIRRED SLURRIES AS A FUNCTION OF PARTICLE SIZE 311 400 T T 350 300 T = 24 °C Slurry unstirred 1330 ml /min C0 o ▼ -325 mesh • -100+150 mesh ▲ -28+35 mesh ■ -8+10 mesh 10 20 30 40 50 Time (minutes) 60 70 Figure 7-73. 0F T PARTICLE C SIZE DISS0LUTI0N IN UNSTIRRED SLURRIES AS A FUNCTION 312 E a <± c o c a> u c O u o U 20 30 40 50 Time (minutes) Figure 7-74. RATE OF CALCITE DISSOLUTION IN STIRRED SLURRIES AS A FUNCTION OF PARTICLE SIZE 313 400i 350 300 a a 250 c o c 33 ) The original purpose of this study was to develop satisfactory methods of sample preparation that would avoid loss of volatile elements and to develop or adapt analytic methods and techniques for determination of trace elements in lignite used in the pilot gasification plant as well as the determination of the fate of these elements during the gasification process. Later developments limited the amount of work done on trace ele- ment determinations in process-stream samples at the pilot plant but in- creased the responsibility of this laboratory for the determination of cyanide, isocyanate, nitrate phosphate, sulfates, and phenols -- as well as suspended solids, dissolved solids, pH, hardness, COD, chloride, and fluoride in a few samples -- in liquid-stream samples under equilibrium conditions during pilot plant runs. This change of emphasis came about when a private laboratory (Radian Corporation) was given the responsi- bility of making an overall materials -balance study at the pilot plant including determination of sulfur and nitrogen compounds and particulates in the gas streams as well as minor and trace elements and compounds in the solid and liquid wastes. 321 Sampling was done by personnel at the gasification pilot plant during pilot plant runs after equilibrium conditions were successfully estab- lished. 8.1.2 EXPERIMENTAL PROCEDURES 8.1.2.1 Sample Preparation 8.1.2.1.1 Perchloric Acid Digestion of Whole Coal Samples. In this procedure, 1.000 gram of pulverized (-200 mesh) whole coal is placed in a 250-ML Erlenmeyer flask with 10 ML of concentrated nitric acid. A reflux head is placed in the flask and the contents are warmed on a hot plate until any initial reaction has subsided. Next, 25 ML of concentrated perchloric acid is added and the digestion is continued at a tempera- ture that will produce mild fuming of the perchloric acid until the sample is completely oxidized. Additional acid is added if required. The residue is cooled and diluted with about 25 ML of deionized water, warmed slightly, and filtered through Whatman Number 541 paper into a 50-ML volumetric flask. Small volumes of water are used to wash the paper and residue before making the solution up to volume. This solution is used for the determination of antimony, arsenic, cadmium, lead, selenium, and tellurium by atomic absorption methods. 8.1.2.1.2 Aqua Regia Digestion of Whole Coal Samples. A 1.000-gram sample of pul- verized whole coal is placed in a 125-ML Erlenmeyer flask with 8 ML of concentrated hydrochloric acid and 2 ML of concentrated nitric acid. The flask is placed on a hot plate and the contents are digested at a temperature just below the boiling point for 1 hour. The solution is then cooled, diluted with about 25 ML of deionized water, filtered into a 50-ML volumetric flask and made up to volume. An aliquot of this solution is then used for determination of mercury by an atomic absorp- tion method. 8.1.2.1.3 Acid Digestion of Coal Ash. A 10-gram sample of coal is ashed at 500°C in a muffle furnace, then 0.500 gram of ash is placed in a platinum cru- cible with 1 ML of concentrated nitric acid and heated on a hot plate. When the brown fumes have disappeared, the sample is cooled and 3 ML of concentrated perchloric acid is added and the sample is heated to fumes of perchloric acid. After cooling, 10 ML of hydrofluroic acid is added and the contents are again heated to white fumes, then cooled and diluted with 10 ML of deionized water and heated to boiling. The solution is cooled and filtered through Whatman Number 541 paper into a 50-ML volu- metric flask, containing 50 MG of potassium (as potassium chloride). The pater and residue are washed with deionized water and the solution is made up to volume. Beryllium, chromium, nickel, and vanadium are determined by atomic absorption methods from this solution. 322 .1.2.1.4 Acid Digestion of Limestone or Dolomite. A 1.000 gram measure of -100 mesh sample is placed in a 250-ML beaker and wet with a few drops of dis- tilled water. Then 25 ML of 1:1 hydrochloric acid is added carefully to avoid spattering and the solution is warmed on a hot plate until the car- bonates have been taken into solution. After cooling, the solution is filtered into a 100-ML volumetric flask and made up to volume. This solution is used for the determination of antimony, aresnic, beryllium, cadmium, chromium, lead, mercury, nickel, selenium, tellurium, and vana- dium by atomic absorption methods . .1.2.2 Determination of Elements .1.2.2.1 Antimony, Arsenic, Selenium, and Tellerium by Atomic Absorption After Reduction to the Hydride with Sodium Borohydride. A 10 -ML aliquot of solution is placed in a 125-ML Erlenmeyer flask and 1 ML of concentrated hydrochloric acid is added. The flask is then attached to a train by which argon gas can be swept through the flask to the middle of a quartz tube which replaces the burner in the light path of the atomic absorption spectrophotometer. The quartz tube is heated electrically to a tempera- ture sufficient to ignite the hydrogen generated when the sample is in- troduced by dropping a pellet of sodium borohydride into the flask after all air has been swept out with the flowing argon gas. Hydrides of the elements are formed by reaction with the sodium borohydride and are carried along with the excess hydrogen into the quartz tube where they are burned and absorb specific resonant wavelengths which are monitored by the spectrophotometer. Elements lines used in the determinations of 217.6 nanometers (NM) for antimony, 193.7 NM for arsenic, 196.0 NM for selenium, and 214.3 NM for tellurium. Absorption peaks are recorded and compared with peak heights obtained with reagent blanks and with standards containing from 0.005 to 0.05 PPM of a given element. 1.2.2.2 Cadmium by Atomization of the Solution into an Air-Acetylene Flame. The sample solution is atomized directly into an air-acetylene flame using the cadmium line at 228.8 NM and the concentration of cadmium is calcu- lated from a standard curve obtained by atomizing standard solutions, under the same conditions, containing from 0.02 to 0.1 PPM cadmium. 1.2.2.3 Lead by Flame less Atomization. Five microliters of sample solution are introduced into the graphite furnace which is operated under the fol- lowing conditions: Dry 15 SEC at temperature setting of 3 Ash 20 SEC at temperature setting of 3 Atomize 6 SEC at temperature setting of 5 Nitrogen flow: 6.5 Hydrogen flow during atomization: 1.5 The lead line at 283.3 NM is used and the concentration of lead is cal- culated from a standard curve obtained from a series of standard solutions containing 0.02 to 0.2 PPM lead. 323 8.1.2.2.4 Mercury by the Cold Vapor Technique. A 10 -ML aliquot of sample solution is placed in a 40-ML round-bottom glass tube and 1 ML of a 10-percent solution of stannous chloride in 0.5 N 'hydrochloric acid is added to re- duce the mercury to metallic mercury. The tube is immediately attached to a train which will allow a stream of air to bubble through the solution and then across a gold foil which traps any mercury released from the solu- tion. The train is then connected to an absorption cell in the light path of the spectrophotometer, the gold foil is heated in a stream of hydrogen and the mercury is released and carried into the absorption cell. The mercury line at 253.7 NM is used for analysis with standard solutions containing 0.002 to 0.010 PPM mercury to establish a calibration curve. 8.1.2.2.5 Nickel and Chromium by Atomization of the Solution into an Air-Acetylene Flame. The sample solution is atomized directly into an air-acetylene flame with a 2X-scale expansion using the nickel line at 232.0 NM for nickel and the chromium line at 357.9 NM for chromium determinations. Calibration curves are established with standard solutions containing 0.1 to 2.0 PPM of the respective element. 8.1.2.2.6 Beryllium and Vanadium by Atomization into a Nitrous Oxide-Acetylene Flame. The sample soultion is atomized directly into a nitrous oxide- acetylene flame with a 5X-scale expansion using the beryllium line at 234.9 NM and the vanadium line at 318.4 NM for determination of the two elements, respectively. Standard solutions containing 0.05 to 1.0 PPM of each of the elements are used to establish calibration curves. 8.1.3 RESULTS AND DISCUSSION Analytic results for samples collected by plant personnel during various runs of the pilot plant are shown in Tables 8-2 through 8-6. Since Radian Corporation was given the responsibility of determining the dis- tribution of trace elements in the effluent streams of the pilot plant, and since flow rate figures at the various sampling points were not sub- mitted with the samples, no attempt was made to calculate depletions or enrichments of specific elements in the various streams. Even quali- tatively it is difficult to draw any definite conclusions concerning the fate of the various elements during processing since no spectacular differences for any of the elements are observed at the various sampling points. This may mean that volatilization losses of even the most vola- tile elements may be less than would be expected in the process. 324 ELEMENT Arsenic Beryllium Cadmium Chromium Lead Mercury- Nickel Selenium Tellurium Vanadium WATER FROM SCRUBBER SUSPENDED LIGNITE TO LIGNITE FROM LIQUID SOLIDS PREHEATER PREHEATER SR5584 SR5584 SR5585 SR5589 <0.01 4 5 6 <0.1 0.2 0.4 0.6 0.02 0.2 0.2 0.2 <0.05 6 5 4 0.001 2 1.5 1.5 <0.005 0.31 0.35 0.40 0.06 3.5 11 13 <0.01 1 1 1 0.03 4 8 9 <0.03 10 3 7 *A11 samples in PPM are averages of 2 to 10 determinations. Table 8-2. TRACE ELEMENTS* IN SAMPLES TAKEN DURING OPERATION OF PREHEATER, DECEMBER 12, 19 74 325 3 S3 o to CM i— ( o CJ Z J oo 00 o o OS o o as i-H V OQ. W co to to to v 2 Q OS 0J O E- H r-* i— I a. to OS W CM E- U 00 H <-> OS < < CO CN o o 00 o L0 LO o V o LO V Q OS QUO ■Z. H E~ <; h q, OS LLJ X H U < < < OS o h s - z OSu j o 3 8£S 2 o H OS w cu E- H < hH LU 2 S J a. oo to CM 00 OS co CM 00 OS CO CU OS E- CU 1— 1 l—l (O S u. to O S HH CM J O CO 00 Q OS < QU.U OS CO LO t> CM 00 OS CO to r-. CM oo OS CO o oo O "-H o oo ^3 R vO o l—l V 8 o CU E- H OS HWW MHh .— 1 Oh fc LO fcZH CM SOW oo i-H o OH< OS i-H a j u CO V Q O LO CM o o 03 to o vO V LO V C o •H +J cd c vO CM 2 OS E- a. i-H *3" rH •H o o 00 o o o ^ E- o ■"*■ CM V V LO V 4-> a> I-H a. o u ■-H S5 l-H o Tl- t-H o +-> CM a! D Q o o 00 o o o "3- oo CM V V LO V <4H o t/> a> cd U > 2 W i * iH V a. c •H 9 i-H > i-H pj U H to i oo o o 00 o O o r-H CM i-H V V LO V < * 0) --H cd E- c a; CD DO cd U 2 cd 3 O CD CO cd a; cd 326 lo Q -h os X, (N < H i ac i— i pj o 8 CO Q Q PJ oa OS PJ 1-1 PL, O M CO o Q H ^ 04 OS H W PJ < E-, _i P- 3 a, ju 00 DS CM 00 OS co LO K> 00 CTl OS co to 00 a> OS CO to to 00 CTl OS CO vO PJ .-H H « 00 i-i pj en Z E- OS C3 < CO 3£ Q 2 PJ 0- LO M o OS o OS Q E- CO LO o o rt tO o o 00 O V o LO V vO LO LO CN vO O V o V o LO V o o o CN o CN CN O V o V o LO V 00 o o o CN V V LO V ej) o CN o v o LO V t~- rH o o to i— I to o i— i V o I— ( V o o V o LO V (?) 1— 1 o o V o L0 V to <+-! o i-H +-> CD JS co v ' u r~~ CN (A § o OS •H 4-> E- rt s*[ C 5 •H J g 0- E- s -o 1— ( o- o r-i O z O 1— 1 4-> 65 CN Q CM Z O PJ s*s (/) < 0) E- M cd co u PJ CD J > p , aJ § CD CO J-i rt z 1— 1 Z 0- * 0- CO H e z •H § to PJ CD ■J 3 PJ i— 1 aJ PJ > CJ i— i 2 r-H E- < * -* i 00 o •H •H i— 1 | • H fr t— 1 •H f-l •H C i-H •H E 3 CD e d "0 CD fc g O T) o JM CD i-H rt t/) T) 5-i rt Fh o t— t i-H s Jh CD rt 6 CD CD •H CD ■a 327 OS < o in 2 3 5ga LO LO 00 01 OS CO LO o o tO o o * 2 Q> V cc: O as Q a, co o CN O V o V eg id c o •H ■p aJ C 0) T3 O CO LM o CN +-> CD CD co u CN os Cu O 1— 1 Oh U 2 cu CN LW o a LU H OS < 2 H Cu CO UJ D, U < U U < S o 00 r-. ,-h os CO 00 o t^> o LO o LO O O V LO V 10 CD a3 CD H rt S a 0- co 2 a 0; w 2 H 00 CO LO rH J u CT> V < u as u < CO O R o LO 00 to o V o V o LO V s rH > u 2 I 00 •H •H r— 1 | •H fc c rH •d e 3 CD fr 1 rH -a rH u CD rt X CD < 03 U CJ j S 1— 1 •H u CD c 3 ■s CD rH r-( r— 1 •H CD CD 2 co H 3 C > •s 328 ^s a i— < 01 CM J i— i tO U- j CT> 1 P-. o OS! o u-, uj CO co CN o o cm o .—I V o LO >- LO tO OS Q tO OS ^ M Z J I j r-l i-J CO Q rH o co UJ r-i ££ O hJ OS UJ £2 UJ z UJ E- oo CT> CT> OS CO 00 00 CT> o> OS CO LO 00 o\ OS CO 00 2) OS CO 00 CO o to 00 OS CO CTi to 00 OS CO 00 to 00 OS co 00 o o o CM o o V 00 rH O o o o I— I V rH O CM o o V o o 1—1 V 1—1 o o CM o o o V CM o CM o o o V to o t— I V o r— I V o V o o V L0 V CM o to o 1— 1 V o oo o o rH o V o V o LO V o to o o o 1— 1 o V o V o LO V c o •H p ctJ c •H 6 u a> p CD T3 O P CM o (/) cd CD rH aJ ft- ft- * M-i O to p o u CM OS E- ft- E- s z r- 1 OS a z UJ co UJ ►J co * CO UJ -J UJ UJ CJ < OS H 00 o 1—1 ■9 E- cd 8 6 CD fr i— i •H 3 ca c o r* CD rH O i— 1 i— i Cl, o U UJ i— i ►J u i— i < u OS to u < CO LO o to oo vO vO vO o V vO «4-l o 4-> (I) Q UJ H < 2 OS o o rf 0Q E- 00 QS Cl. o < UJ 1— I u u 1—1 S < OS ^r CO to "3- o o 00 to vO LO o v O OS CO vO O LO o f-t V o t— I V o 1— I V VO o o V VO t/1 M rt fn > < * w UJ UJ u < OS LO I 00 o •H •H | •H fr a ■—1 •H § 3 fr I 2 "0 ed o U < 0Q U 6 r— 1 •H •a (3 3 •^ d) i—i O i—i ,— i •H 2 CO E- •H •V •s 330 o os H 2 2 H r- 1 gSSS 00 O Ul Z J H f— 1 reg: QUE OUT SOL r— 1 PS CO f— 1 o o o LO o LO CN o os pj a 0- £ CO Q i— l J O CO PJ i— i 3 i-i O -J oS CO en o i-H i— I a: CO CN O o V to O o oo ■— < o V LO o o o V to CN o V CM o CO X CO 00 CN OS u X < OS o PJ 2S to o i-H i-H OS co o vO o CN o V to o • H c •H E H (D 4-> CD o •H E ■H i— ( p •H c i— 1 • H s 0) fc 1 o J-i cu a3 X < CQ U u fr !— 1 •H 3 a> bo rt H CO > PJ CN E— E- < r-. Q hH PJ CN uj z X . T— 1 1— 1 OS l-J CO LO VO LO o V LO o o V o tO to o V LO LO CD z O I-H <— I QS O Q Z CN PJ O H OS o p 5 2 5 UJ z UJ X Q X O UJ < H 2 o 00 OS CO LO o VO to* o LO LO o V LO o o V to to 00 o o V to 64 CO u CD > * UJ UJ UJ u 2 vO 00 CD c o E •H +-> c CD < £ % CD CQ E o T3 CD 3 o CD CJ c CD i-H CD CO t 3 CD E- 4 rt •8 332 t Z 2U uiou a. co oi uj. c/5 <; cl, oS LO i-H as CO tO to V Ln o o v o CTl tO O V 00 os o UJ Z 2W PJ X Q O a, co os uj m <£ tu 2 00 Ln oS CO vO 1^ 9 LO o o V to 00 o V *3- LO OS ~J PJ < UJ to 1— 1 U- 2 z c CO to 1— I UJ ^J UJ "3- CO H u z 1— 1 < X UJ S 1— 1 tin OS CO a> lo o o v 00 cm o V tO CM 1= O •H Q UJ OS E- o < J H z OS CQ H UJ CM oS a. Z to < UJ UJ o IO +-> CD CD -C co co G 03 O to •H to P cd G § •H OS g H CD P O- CN CJ3 2 Mh I-H O OS 3 w Q CD bo 2 cd W 1 > E- cd CO UJ J i-H U4 i-H < cu * U < OS E- v£> 1 00 CD i-H £> cd H 334 H yj Q CQ 5 l— : 3 J 3 CMC- tU C/ OUtH CO CO UJ J ci o UJ z UJ C3 Pi E- U UJ §£ O'O a UJ H •— i 0 J CO Q oo to I— I OS CO CM o o vO CM O o en ** ^ o o o o V ro .— i a to 1— 1 ■=t MJ t— i o os CO CO o CO UJ < b! Q UJ 2 J C/ 2 aj Z h-i a S> h J o to CM CM V O a> LO CM ro r-i a to 1— 1 ^ -J I— t o Cm CO CO CM V to CM o V to LO to .—I CO o to 1—1 C- CO o o 1— I to 1— t CM CM LO o 1—1 o o O O i— I O o o o o O O o o o o o o V V V V V V \D CT> LO to o V LO CT> O vO to o ■sl- o 1— I • o£ o CO LO 1—1 1— < CM CM CM LO o 1—1 o o O o 1— 1 o o o o o o o o o o o o o V V V V V V V o 6 o •H CO U < fr •§ CQ U CD mJ o 0) CD CD CO CD H '4 > cw o «t 4-> JCj CO > ' CQ to • to co o § •H OS •P rt H • H ^ fj a, ■P E- -a s 1—1 o O- i—i O o z +-> M Cm CM 3 a CH O z UJ 3 W) E- rt 5n CO UJ > J CtJ i: < u CO rt z S I-H Cm Cm * CO c E- •H z a co iS 92 UJ 3 mJ i— i UJ rt > UJ u i-H <£ i— 1 Cm < H * vO i oo i—i •9 H 335 Many of the elements are at concentration levels where sensitivity of the analytic methods fails to prove highly reliable data where small differ- ences may be significant. In spite of this uncertainty, the data should ^ of value for comparison with values obtained by Radian for trace ele- ment levels at the same sampling stations for samples which they collected, The analytic methods which are described are considered to be an improve ment and more reliable than methods which were used and descried in the" interim report "Trace Elements in Lignite" (refer to subsect on 8 2) which was issued in December 19 72. and tenurTnf eteC ^ ble C °? centration f °r antimony, arsenic, selenium, and tellurium m the sample solution is about 0.002 to 0.010 PPM Thus since a 1-gram sample is digested and made to 50 ML for analysis' the ' troTpPMT^ 16 COn r tration in s °^ samples is approximately 1 to 0.5 PPM for these elements. The relative standard deviation is about 6 percent for each of the elements at a concentration level of 02 PpT m the sample solution. Attempts to extend the detection limit for cadmium to lower concentration levels by use of the graphite furnace in the flameless atomizatin tech- fZe LTh ^ successful . because of too high reagent blanks. By the flame method the sensitivity and reproducibility for cadmium is about the same as for antimony, arsenic, selenium, and tellurium by the gaseous hydride method described above. g^eous Sufficient sensitivity for lead could not be obtained by direct aspir- ation of the sample solution. Some success in the determination of lead by complexmg with ammonium pyrrolidine dithio carbonate (APDC) and ' extraction into methyl-isobutyl ketone (MIBK) prior to aspiration was obtained but the method was more time consuming and did not appear to be as reproducible as determinations by the graphite furnace method for which a relative standard deviation of 6.5 percent was obtained. In determining mercury digestion of the sample with aqua regia followed by dilution to a definite volume and trapping of the mercury from an aliquot portion of the solution, after reduction with stannous chloride, onto gold foil proved to be much more reproducible than the method of combustion of the sample in a stream of air or oxygen and trapping the mercury on gold foil from the effluent combustion stream where hydro carbons m the combustion products interfere with the efficient trapping of the mercury. Liquid samples were run for mercury by release of the mercury directly from, solution into the absorption cell without the intermediate goldtrap but this was not possible with the aqua regia ^n thi 0nS i°5- S ol \ d .\ am P les because ^ interference from nitrogen oxides ion cell" A°def ^ *? ^f* al ° ng WUh the merCUI ^ int0 the Por- tion cell A definite figure for a relative standard deviation was never established but analyses of the NBS coal standard re fere" ample indicated that a recovery of only about 60 to 70 percent of the mere Zl might be recovered by the digestion process mercury 536 Nickel and chromium are at relatively high concentration levels in most of the samples so the sample solutions could be atomized directly into ?he air-a^tylene flame ^ith good reproducibility and sufficient sensi- tivity for most of the sample. Some improvement in sensitivity could be obtained by use of a nitrous oxide- acetylene flame but with an ac- companying loss in reproducibility. Beryllium and vanadium are both determined by direct aspiration of the 3 solution into a nitrous oxide- acetylene flame but both lack good sets tivity. Attempts to improve sensitivity by pre concentration techniques such as chelation and extraction or ion adsorption and stripping and by use of the flameless atomizer met with limited success but need consider- ably more investigation before a satisfactory method can be described. 337 8.2 TRACE ELEMENTS IN LIGNITE Any commercial lignite gasification process will be concerned with utilization of several million tons of lignite per year. On this volume basis toxic elements which are present in even trace amounts in the lignite must be considered as potentially hazardous to the local environment in the area of the gasification plant until their fate during the gasification process is determined. Seemingly satisfactory methods for the determination of mercury lead cadmium, arsenic, and selenium in lignite have been developed. Addi- ' tional work to determine reliability of the methods is continuing and determinations of other elements will be investigated. Determines 5 ? / I ° . elements during the gasification process will be studied when the pilot plant is in operation. 8.2.1 INTRODUCTION Increasing public concern over environmental pollution has led numerous chemical companies and other industrial concerns to investigations of raw material. ; processes, and products as sources of toxic metal pol- lutants C34,3S) 0ne of the raw materials which is rec ^ m P potential source of toxic metals in the atmosphere is coal which has been shown to contain a large number of metals in trace amounts . ( 36 > 37, 38, 39) Noxious elements, even though present in only trace amounts in lignite must be considered potentially hazardous when lignite is consumed as a' fuel or converted to gaseous products. While the amounts of lignite to be utilized in the Rapid City pilot gasification plant pose no serious threat to the environment, a commercial gasification plant consuming ' several million tons of lignite per year could release several tons of toxic metal pollutants to the atmosphere. It becomes important then to know the concentrations of trace elements in the lignite and their dispo- sition during the gasification process. Most of the potentially hazardous elements present in coal possess properties that cause them to require special analytical techniques tor their detection and analytical determination at the concentration levels at which they exist in the coal. Numerous studies are underway to evaluate analytical techniques for their determination. (31, 32, 33) The purpose of this study is to develop satisfactory methods of sample preparation that will avoid loss of volatile elements and to develop or adapt analytical methods and techniques that will allow reliable deter- minations of trace elements in lignite used in the pilot gasificati. plant and determinations of the fate of these elements during the gasification process. .on 338 2.2 The original scope of this project included sampling of lignite beds Ind a correlation of noxioi elements concentrations with geographic l^caUons and coal seams, as well as determination of the fate of the undr simulated processing conditions. Subsequent discussions and correspondence with Dr. D. R. Bomberger, however, prompted us to Sndt our investigation to samples supplied to us by Consol and to ielay Z studies of the fate of the elements during the gasification process "until the pilot plant is in operation and real process streams are available. EXPERIMENTAL PROCEDURES 8.2.2.1 Sample Preparation It has been found that samples can be satisfactorily oxidized and put into solution by a wet oxidation process involving digestion with nitric and sulfuric acids followed by cautious addition of a mixture of percnToric and nitric acids. After oxidation is complete the con- tents of the digestion flask are heated to strong fumes of sulfuric acid, cooled, and diluted with water and hydrochloric acid. The solution is warmed slightly, filtered into a volumetric flask, and made to volume. 8.2.2.2 Mercury Determination Mercury can be determined by either of two methods involving flameless atomic absorption spectrometry: m An aliquot of the solution of the oxidized sample is placed in a special reaction vessel and treated with stannous chloride to reduce any mercury present to elemental mercury which is then carried into an absorption cell in the light path of the atomic absorption spectrophotometer by bubbling air through the solu- tion The absorption of the 2537 A resonance line of mercury is measured and compared with the absorption of standard mercury solutions. . -i^i u^o+ -; c (2) A weighed portion of raw lignite contained ma nickel boat is placed in a combustion tube which is connected to a small tube containing gold foil. Mercury present in the lignite is trapped on the gold foil when heat is applied to the combustion tube and the sample is burned in a current of air sweeping through the tube The gold foil trap is disconnected and washed with acetone to remove any water or hydrocarbons, then connected to the absorp- tion cell and heated. A stream of air carries the released mercury into the absorption cell where it is measured as above. 339 8.2.2.3 Cadmium and Lead Determination After wet oxidation of the lignite, cadmium and lead are chelated with ammonium pyrrolidine dithiocarbamate and extracted into methyl isobutyl ketone. The MIBK solution is then contacted with a small volume of 10- percent nitric acid solution and the cadmium and lead are back-extracted into the acid solution which is aspirated into the flame of the AA spectrometer. The absorption is compared with that of standard solutions. 8.2.2.4 Arsenic Determination Arsenic is determined by a colorimetric procedure in which arsenic compounds react with zinc to form arsine which then reacts with silver diethyldithiocarbamate to form a soluble red compound. The absorbance of the solution is measured spectrophotometrically and the arsenic is determined by comparison with similar measurements for standard arsenic solutions . 8.2.2.5 Selenium Determination A colorimetric method employing 3, 3'-diaminobenzidine is used for the determination of selenium after coprecipitation from solution with arsenic, The colored complex formed by reaction between 3, 3» di ami nob en zi dine and selenium is extracted into benzene and its absorbance at 400 nanometers (NM) is measured by means of a spectrophotometer and compared with s tandards . 8.2.3 RESULTS AND DISCUSSION Samples analyzed were supplied by Consol and results are reported on an "as received coal" basis. Initial studies were made on samples which were ground to -100 mesh. However, it was found that the coal could be wet-ashed much more readily if ground to -200 mesh and most of the results reported are for -200 mesh samples. In Table 8-7, the samples are identi- fied and a summary of results for five noxious elements in these coals is given. The results shown in Table 8-7 and subsequent tables represent only a small fraction of the total effort expended in seeking satisfactory methods of sample preparation and analysis. 340 Hg Pb Cd As Se SAMPLE SOURCE M ^ received coa i basis) NO. „ n 10 1.2 0.06 2.6 1.3 Lignite, Dewey County, South Dakota 2 Lignite Char, North Dakota 3 Lignite (lower seam) , Mercer County, North Dakota 4 Lignite (upper seam) , Mercer County, North Dakota 5 Rosebud Seam, Rosebud County, Montana •Selenium not determined because of insufficient sample Table 8-7. SUMMARY OF RESULTS FOR COALS ANALYZED Other methods of sample preparation which were tried and abandoned include: 0.10 0.05 3.4 0.06 4.2 0.7 0.06 3.8 0.05 2.1 0.06 3.6 0.06 1.0 0.06 10.3 0.10 0.8 1.1 (1) (2) Oxidation of the sample by burning in a Shonige r flask Oxidation by burning the sample in a Parr bomb _ . i _ . .1 -».v. ni tn n and (3) Digestion of the sample with nitric acid potassium chlorate (4) Wet-ashing of the sample utilizing nitric acid and potass The Shoniger flask method is ^^^^^T^t\^^ elements because of the very smal J^ S1 * e * \ * eaves incompletely the flask. Digestion of the samp ^^ ^causes problems in subse- oxidized organic material in solution which causes pro ^ qU ent manipulations /^--^^f^^wUh^le^nts from the bomb contamination ot tne resulting =>u potassium tent and very high blanks. Kesuius nakota using the nitric in the lignite sample from Dewey County South ^'^ J Table 8 . 8 acid and potassium chlorate method of oxidation ™ S ™ d by this me thod. and illustrate the high blank va ^l^^lyZ option tube Mercury values shown in Table 8-7 were ooidincu y method. 341 SA ^ LE ^ Hg yG Hg/G % RECOVERY N0 - FOUND LIGNITE Blank it it 1* it 1* + 0.2 yg Hg 1* + 0.4 yg Hg *2g sample yG Hg FOUND 0.14 0.14 0.25 0.35 0.22 0.35 0.55 0.55 0.48 0.65 0.85 AVG AVG 0.48 0.12 85 93 Table 8-8. DATA FOR DETERMINATION OF MERCURY BY FLAMELESS AA AFTER ASHING WITH HN0 3 AND KCIO3 A dry-ashing procedure was originally tried for the determination of arsenic. The method employed is that of HertzogC 40 ) in which the lig- nite is mixed with sodium carbonate, magnesium oxide, and potassium nitrate and burned slowly in a muffle furnace. The residue is dissolved in hydrochloric acid and the arsenic is determined by the silver diethyldithiocarbamate method after distillation as arsine. Results of arsenic determinations by this procedure are shown in Table 8-9. Results of similar determinations after wet-ashing are shown in Table 8-10, and a comparison of the two methods is given in Table 8-11. It can be seen that the results appear to be slightly higher after wet- ashing but the differences are probably of little significance. SAMPLE SAMPLE NO. 1VT (grams) 1 + 1 ^ As 1 3.2 3.2 70 1 + 1 yG As » 2 1 3 1 yG As FOUND 2 .4 2 .4 2 .5 2, .5 2, ,6 3. 2 3. 6 3. 8 2. 1 yG As/G % RECOVERY SAMPLE 1 2.4 2.4 2.4 2.5 2.5 2.6 3.6 110 3.8 2.1 Table 8-9. DATA ON ANALYSIS AND RECOVERY OF ARSENIC AFTER DRY ASHING 342 SAMPLE NO. 1 + 2.4 PG As 1 + 2.4 ^G As 2 I! tt 2 + 2 UG As 2 + 4 PG As 2 SAMPLE WT (grams) 1 ii 1 it u M II 2 yG As yG As/G % RECOVERY FOUND SAMPLE 2.4 2.4 2.8 2.8 4.1 4.1 63 5.3 5.3 113 4.6 4.6 3.9 3.9 3.8 3.8 6.0 6.0 90 7.3 7.3 80 8.5 4.3 Table 8-10. DATA ON ANALYSIS AND RECOVERY OF ARSENIC AFTER WET WASHING SAMPLE NO. 1 1 2 2 3 3 Table 8-11. METHOD OF ASHING Dry ashing Wet ashing Dry ashing Wet ashing Dry ashing Wet ashing AVERAGE VALUE yG As/G LIGNITE 2.5 2.6 3.8 4.2 2.1 1.6 NUMBER OF AVLUES AVERAGED 5 2 2 4 2 1 COMPARISON OF ARESNIC DETERMINATIONS BY DRY ASHING AND BY WET ASHING METHODS The method employed for selenium is a modification of that of Stanton and McDonald ( 41 J in which the selenium is separated from interfering elements by coprecipitation with added arsenic, filtered, and then re- dissolved after the filter paper is destroyed by oxidation with a nit ric- acid-perchloric-acid mixture. This oxidation step, as well as that of the original wet-ashing procedure, appears to be critical in that serious loss of selenium can occur if the heating is allowed to continue to dryness. 343 Results of selenium determinations are shown in Tables 8-12 and 8-13. These data indicate that there may be a significant loss of selenium during digestion. It is probable that both selenium and arsenic could be determined by an atomic- absorption method in which the selenium or arsenic is converted to the hydride and carried into the flame of the AA spectrophotometer. However, this method was not available to us since the relatively weak signals from the arsenic and selenium tubes require wider spectrometer slits than are standard on our AA equipment. Test NO. 1 2 3 4 PPM Se added 6.0 6.0 3.0 3.0 PPM Se found 6.0 5.6 3.9 3.3 Sample NO. 1 1 1 1 % Recovery* 78 86 67 67 *Based on average value of 1.3 PPM Se in Sample NO. 1 Table 8-12. DATA ON RECOVERY OF SELENIUM ADDED TO SAMPLE AFTER DIGESTION Test NO. 123456789 10 PPM Se added 3.0 3.0 4.0 4.0 4.0 3.0 3.0 2.4 2.0 2.0 PPM Se found 2.6 2.8 2.8 3.5 2.8 2.5 2.5 2.6 2.2 2.2 Sample NO. 111112225 5 % Recovery* 43 50 38 55 38 60 60 79 55 55 ♦Based on average values of 1.3 PPM, 0.7 PPM, and 1.1 PPM Se for Samples NO. 1, NO. 2, and NO. 5, respectively Table 8-13. DATA ON RECOVERY OF SELENIUM ADDED TO SAMPLE BEFORE DIGESTION 344 Data for lead and cadmium determinations and »^*~ ™ *™^» Tables 8-14 8-15, and 8-16. The recoveries appear to be satisfactory Ifr'the spiied raw lignite samples but rather low recoveries .were obtained on the char sample. The explanation for the J™ ^eries fmm this sample may be that the char is more difficult to oxidize tnan the raw lignTte samples by the wet-ashing method resulting m a greater loss of volatile elements during the ashing procedure. SAMPLE PG Pb % RECOVERY VG Cd % RECOVERY ttZr* cniTKin FOUND DESIG FOUND A 9 A spiked* 18 0.16 90 0.42 65 B B spiked* 18 C 15 C spiked* 25 D 15 D spiked* 26 9 0.19 90 0.50 77 0.43 100 0.81 95 0.43 110 0.86 107 E E spiked* 20 °' 49 30 100 0.90 103 *Each sample spiked with 10 uG Pb and 0.4 yG Cd after digestion Table 8 14 DATA ON RECOVERY OF LEAD AND CADMIUM BY EXTRACTION Table 8-14. DAI A u^ ^ BACK . E fraction INTO 10 PERCENT NITRIC ACID SAMPLE AVG PPM Pb NO. OF % RECOVERY NO FOUND DETERMINATIONS 1 1.2 1 (spiked)* 5.4 2 3.4 2 (spiked)* 5.6 5 10-3 5 (spiked)* 15.5 *Sample spiked with 4 PPM Pb before digestion Table 8-15. DATA ON ANALYSIS AND RECOVERY OF LEAD 345 5 5 105 3 3 55 2 2 130 SAMPLE NO. AVG PPM Cd FOUND NO. OF DETERMINATIONS % RECOVERY 1 1 (spiked)* 0.06 0.20 4 6 70 2 2 (spiked) * 0.06 0.14 3 3 40 5 5 (spiked)* 0.10 0.27 3 3 85 'Sample spiked with 0.2 PPM Cd before digestion Table 8-16. DATA ON ANALYSIS AND RECOVERY OF CADMIUM It has recently come to our attention that a wet-ashing procedure for coal using perchloric acid mixed with periodic acid (42) appears to be more effective than our previously described technique using nitric sulfuric, and perchloric acids. Furthermore, a recent report (24) suggests that coals can be dry-ashed prior to the determination of lithium, beryllium, vanadium, chromium, manganese, nickel, copper zinc silver cadmium, and lead without danger of loss of any of these elements with the possible exception of beryllium, during the dry-ashing procedure! No attempts have been made to calculate standard deviations on the data presented since insufficient determinations have been made to make such calculations valid. Future work will include an investigation of the perchloric-periodic acids wet-ashing technique and the dry-ashing pro- cedure as well as additional determinations in such numbers that re- liability of the methods can be established. It is also anticipated that the determination of additional elements including molybdenum and beryllium will be investigated. And, finally, after the pilot plant is in operation, analyses will be performed on the raw lignite used and on process streams to determine the fate of elements in the gasification process. 346 8.3 SULFUR STUDIES A chromatographic system employing a flame photometric detector and a Deactigel column was obtained and calibrated to measure hydrogen sulfide, carbonyl sulfide, and sulfur dioxide in the product gas of the ignite Gasification CO, Acceptor Process. The equipment was moved to the Rapid Sty pilot plant and was used successfully to monitor the hydrogen sulfide content of the reycle gas in the very low parts-per-million range Preliminary results in the laboratory at South Dakota School oTmnes ana Technology indicated that measurements could be extended to the low parts-per-billion range but the project was terminated be- fore that capability could be explored fully in this laboratory. 3.1 INTRODUCTION Tr Tune of 1973 a study was initiated in the South Dakota School of Mines and Technology laboratory to assess the capability of determining the nature and concentration of sulfur compounds in the product gas of ly,l m Accentor Process. The objective of the study was to develop a ^etho/of Suorin the sulfur content of the gas in *e low parts-per billion range prior to the methanation step in the proce **• The p reject was terminated in our laboratory in June of 1974, after a decision Dy representatives of Consolidation Coal Company to continue further work [n the laboratory of the gasification pilot plant at Rapid City, South Dakota, and in tLir own laboratories at Library, Pennsylvania, and at Ponca City, Oklahoma. Analytic capability of determining the nature and concentration of sulfur compounds in the carbon monoxide-hydrogen pro duct of the COg Acceptor Process is vital in the methanation step to produc a high BTU pipeline gas from this product. This results from the fact that sulfur" compounds have a highly deleterious effect on the activity of the nickel catalyst used in the methanation process and must be reduced to a concentration level in the parts-per-billion range in the develop- ment of a commercial methanation process. Sulfur occurs in lignite in relatively small concentrations of less than 5 percent in most deposits and is probably present primarily m the sulfiae form as iron pyrite and organic sulfur compounds. During the gasification process the sulfur may be converted to gaseous compounds such as hydrogen sulfide, sufur dioxide, carbonyl sulfide methyl mercaptan! dimethyl sulfide, and other possible organic sulfur compounds The determination and characterization of these ^^fj 5 ^™: s . important from the standpoint of design of techniques from their elimi nation from the feed of the methanation process. 347 The analytic technique which was the subject of this study was a chromatographic separation of the sulfur compounds in conjunction with a special detector which allows the measurement of ambient sulfur species at the parts-per-billion range. The purpose of the study was to adapt existing techniques to the specific problem of determining sulfur compounds in the product gas of the C0 2 Acceptor Process before and after suitable scrubbers to remove the sulfur compounds. 8.3.2 SELECTION OF EQUIPMENT Numerous research papers (43,44,45,46,47) published within the past few years have been concerned with the determination of sulfur gases in the atmosphere and in natural gas. While most of these studies were initiated from an interest in the environmental effects of sulfur compounds, the analytic techniques are applicable to the determination of these compounds in the C0 2 Acceptor Process gas . Since the chromatograph was to be used specifically for sulfur compounds, a very simple instrument appeared to be adequate for this application. The only departure from ordinary chromatographic instrumentation was the requirement for a special detector which is much more sensitive to sulfur compounds than the common detectors. Such a detector is the flame photometric detector (FPD) . (48,49,50) ^g FPD utilizes a very small hydrogen flame to burn the effluent gases from the chromato- graphic column in a light-proof housing in conjunction with a photo- multiplier tube and filter system to isolate a particular wavelength band which is characteristic of a sulfur species formed when any sulfur compound is burned in the hydrogen flame. Since great flexibility in the gas chromatograph was not a requirement, a highly sophisticated instrument was not considered necessary for this specific application. The instrument which was obtained was an Antek*Model 40 gas chromatograph equipped with a flame photometric detector and an Antek Model 3300 recorder. 8.3.3. EXPERIMENTAL PROCEDURES 8.3.3.1 Electronic Problems Initial problems of high-background and high-noise levels were corrected by modification in the equipment which included wrapping the burner chimney of the FPD with heating tape to eliminate formation of water condensate in. the burner well and annular space surrounding the burner, and further grounding of electrometer and recorder leads to decrease stray electrical signals. *Antek Instruments, Inc., 6005 North Freeway, Houston, Texas 77022 348 It was found that the pair of transistors in the electronic circuit of the FPD were failing prematurely, presumably because of excessive heat- ing from the burner and heating tape on the burner chimney. This problem was overcome by reducing the temperature of the heating tape and by replacing the transistors with a matched set which seemed to increase their stability. After the instrument was moved to the laboratory in the gasification plant the heating tape was removed from the burner chimney without adversely affecting the noise level. Presumably, a drier atmosphere in the laboratory at the gasification plant allowed such operation without excessive liquid water formation and buildup in the enclosed burner space. Removal of this heat source together with the use of a matched pair of transistors resulted in operation of the instrument without further problems of transistor failure. 8.3.3.2 Chromatographic Column Selection of a chromatographic column was made on the basis of adequate separation of hydrogen sulfide (H 2 S), carbonyl sulfide (COS), and sulfur dioxide (S0 2 ) without escessive retention times on the column. These three gases were considered to be the most likely sulfur compounds to be found in the gaseous product. It has been shown ( 51 ) that a column packed with Porapak Q (a silica gel) will give a good separation of the three compounds H 2 S, COS, and S0 2 if it is continually conditioned to S0 2 with a carrier gas of helium containing 100 PPM S0 2 . However, since the objective of this project was to detect and measure these compounds in the parts -per- billion range, the quantity of S0 2 in the carrier gas precludes use for the sensitivities required for this project. The column eventually selected was a 1/8-inch-diameter by 1-1/2- foot- long Teflon column packed with Deactigel which is a chromic acid-washed Davison silica gel from Varian Aerograph. (52,53) This column operated with an oven temperature of 95° to 100° C and with a nitrogen carrier gas flow of approximately 40 ML/MIN; it produced a COS peak at slightly less than 1 minute, an H 2 S peak at nearly 1-1/2 minutes, and an S0 2 peak at about 5 minutes. The retention time for S0 2 on this column is longer than desirable and the peak tails rather badly, but since no S0 2 was found in the gas produced at the pilot plant the column was considered satisfactory for monitoring the gas produced there. 8.3.3.3 Calibration Severe corrosion problems at the pilot plant which were traced to a high level of H 2 S in the recycle gas prompted installation of a zinc oxide tower to remove the H 2 S. The need ther existed to use the chroma- tograph at the plant to monitor the fulfur content of the recycle gas before and after the zinc oxide tower. This called for calibration of 349 the instrument for H 2 S at concentration levels several orders of magni- tude greater than the parts -per-billion level intended in the methanation step. It was found that a 1/4-milliliter sample containing several hundred PPM H S, when introduced onto the column, would saturate the photo- multiplier tube resulting in a prolonged recovery time before operating effectively again. To reduce the amount of light reaching the photo- multiplier tube a light- limiting aperture was placed in the filter housing between the burner and the photomultiplier tube. The aperture was made from an aluminum "Spec-Cap"* by drilling a hole approximately 1.5 milli- meters in diameter in the center of the cap and then placing the cap in the filter housing on the photomultiplier tube side of the filter. It was found that the cap should be turned so that the bottom of the cap is away from the filter to avoid double peaks which are produced when the aperture is next to the filter. The position of the aperture is illustrated in Figure 8-1. Calibration data were obtained using an Exponential Dilution Flask (EDF) as a source of samples. Pure samples of H2S , COS, and SO2 were intro- duced into the flask with a syringe to give a known initial concentration, and a flow of nitrogen through the flask was then set to give a 1/10-fold dilution of the contents every 10 minutes. The gas stream was allowed to flow through a 1/4-milliliter sample loop and samples were intro- duced onto the column at specific intervals of time. The peak heights were recorded as the gases exited from the column. Typical data for such runs are shown in Table 8-17. A plot of peak heights times attenuation versus concentration in parts per million for each of the three gases is shown in Figure 8-2. ELAPSED TIME CONCENTRATION COS PEAK HEIGHT X ATTENUATION (minutes) (PPM) RUN 1 RUN 2 10 12 14 16 251 91138 81920 18 159 48128 44544 20 100 22528 21504 22 63 9472 8832 24 40 3584 3240 26 25 1216 1248 28 16 384 432 30 10 144 144 Table 8-17. CALIBRATION DATA FOR COS CONCENTRATION COS PEAK 1 (PPM) RUN 1 10,000 1,000 631 39 8 251 91138 159 48128 100 22528 63 9472 40 3584 25 1216 16 384 10 144 ''Spex Industries, Inc., Post Office Box 798, Metuchen, New Jersey 08840 350 CD -P ■H 4h CO N fH -P CD (1) Ph fl ca CD u 2 rH 3 ot fc pq o •h tm -P -H CO S 3 o o P trj O (X, O P CD O e cu CO -P i— I CD E- OS W < z E- 0-, o E- Z UJ UJ u z OS < u I— I H U LA I 00 3 •H 351 10 100 1000 o •H p co fi 0) p p M Q CO CD U B -p CD Ph CD £ CD H -P PI CO -p co PI o o CD ■§ EH PI O ■H -P co 0) Pl, fv.V,.V^ '■VN'.VyV^^V V^V I LO to o o o o o 00 H <* a w l»N ^f "t 1 "* 1 1 Z J \ Tl- O 1 o 1 1 2£ o o to I to 1 1 I— 1 i— i 1 o Ol E-l ■<* Q uj r~ o O I cn 00 1 Z 3 *\ vO 1— 1 1 LO 00 1 O E- o a> ■-H 1 CM 1 a. 3 i— i o £ E- rf o vO 1 , ^- O 1 Q PJ i>. CN vO 1 *t CN 1 Z J \, ■>* O 1 O t a. 5 00 CM CM o 00 oo O LO CM .© I Tj-OOCMLOfMvOi-HOO I \0 fO o *t CN I ^ oo I Ol HON N CO O O LO LO LO •n- (N (N n- rH o •<* CM to O O i-t vO LO V CN L0 o o o vo 00 HIONH to rr cm CN CM L0 to >-* O ( 00 o ^t rH o 00 8 LO CN Ol^-lrtiOO^fOlflCO tO r\J 00 i— I V V LO t— I rH oo lo o <-H •s o o -a -o i— I tfl 00 CO I 0) o LO f-l cd -a t) .-H c -s a UJ a> T3 Ifl a, 1— I h CO i! 7^ c O Cl> ■3 ■H 3 •H 4-> 03 4-1 cd S < z w a, -H a CO 03 to UJ .2 9 H i— I o o u g O E- O D O ■p X> "J 3 O. L0 to •H •8 r-H •H 49 > 03 C H a cd | •H CO rH i— 1 ^( tj •H 4-1 •rH .— < o U -I ■H CD UJ UJ rt rt R) ed ^H cd ^H 03 ■fi i cd o & a. h 4-> *-> +-> iH O r-t o ■»-> C 4-> .— i >- < O O O O L0 o LO o 03 o o 03 03 H Q H E- H > > H LJ Z E- u S o LO 8 O "* O oo vO CN Oi LO o CN CN Ol O 4) <1> H M t — * 1 8 Z c3 g >vXl xo cd 4-* _3 £3 o cd ^ c &o cd x cd cd o <4-i ■H ^ fj o C ^ D- M S '- 1 -i CO H > — ' t— 1 « — ' ,£ 3 r-l 4J o o Jh •H -H L0 CL, 1— 1 2 < < U CO C/) Z O CL, CO CD 4-> cd to &, O (0 c O CD Xi fS a, a. CD 4-i tSO-r-l • CU E «-> 0) 6% / — \ TO to 4-> 4-> O •H Ifl 14H O r-t o o ■6 CO cd 4-> o ■a Q X 4J 3 o CO c o % 4-> CO 4-> t3 C CD CD 4-> 6 cd •H O r-l -H CD "O C B 1 . UJ CD 00 tfl C -H +J -o o 3 to ifl 00 (D C rH t B C a. •h a. 00 c c LL) -H CD CD ^ b 4-> CO ^H to O 4-> •H u 3 cd OOrH C r-l •r-l < _3 G^ • o < r-l o • c CD «"° f ; tube so that no liquid will be siphoned back into the condenser when the heat is removed, rinse any liquid left in the condenser or discharge tube into the boric acid solution. Titrate the distillate with standard hydrochloric acid using bromcresol green as an indicator. Subtract the result from a distilled water blank treated as a sample for each determination. Calculate the NH 3 concentration using the following equation: 365 NH 3 ,MG/L = N (ML -ML HC1 Sample Blank) X 17.000 ML sample 9.1.3.2.4 Particulate Size Analysis. Size analyses were made by measuring particle size on photographs taken with a scanning electron microscope. Refe- rence-calibrated latex spheres were used in all cases to determine particle size. 9.1.3.2.5 Single-Ion Electrode Methods. The equipment required for using single- ion electrodes consists of a suitable pH meter or specific-ion meter (mi Hi volt meter) and the appropriate specific-ion electrode and reference electrodes. The procedures used are those published by Orion Research Incorporated and supplied with the electrodes. Compa- rison measurements were made for sulfite, sulfide, and ammonia 9.1.3.3 Results 1.3.3.1 Standard Method Sulfite, Sulfide, and Ammonia. During the period of August 1975 to February 1976 three sets of process water stream samples were received for analysis of sulfite, sulfide, and ammonia. The plant runs represented are 27C, 28B, and 33B. The sample locations are as follows: 1-A PREHEATER INLET 1-B PREHEATER OUTLET 2- A REGENERATOR INLET 2-B REGENERATOR OUTLET 3 S0 2 SCRUBBER EFFLUENT 4-B ASH SLURRY TANK 13 WASTE POND EFFLUENT 16 -A GAS I FIE R INLET 16-B GASIFIER OUTLET 18 CHEMICAL SEWER SUMP Tables 9-5, 9-6, and 9-7 show the results obtained for sulfite, sulfide and ammonia respectively. ' SAMPLE RUN 27C 1-A <0.01 1-B 0.68 2-A 0.84 2-B 5.4 3 57.9 4-B 102.5 13 <0.01 16-A No sample 16-B 0.30 18 <0.01 CONCENTRATION IN MG/ LITER RUN 28B RUN 33B 2.32 <0.01 0.40 4.12 0.01 <0.01 6.65 43.8 125 99.2 8.54 No sample 4.13 <0.01 <0.01 0.65 10.7 12.7 <0.01 1.52 Table 9-5. SULFITE IN LIGNITE GASIFICATION PLANT SAMPLES 366 SAMPLE RUN 27C 1-A <0.03 1-B 0.23 2-A 0.62 2-B <0.03 3 10.8 4-B 230 13 <0.03 16-A No sample 16-B <0.03 18 0.47 Blank 0.41 (Dist. , H 2 0) Tat )le 9- 6. SULFID CONCENTRATION IN MG/LITER RUN 28B RUN 33B 0.72 0.19 294 L96 0.46 0.42 0.45 3.24 3.96 2.17 18.8 No sample 0.48 <0.03 0.50 <0.03 1.60 84.1 0.57 0.49 0.41 0.14 SULFIDE IN LIGNITE GASIFICATION PLANT SAMPLES SAMPLE CONCENTRATION IN MG/LITER RUN 27C 1-A 0.20 1-B 1.8 2-A 1.7 2-B 37.9 3 10.0 4-B 18.0 13 242 16-A No sample 16-B 1180 18 1.38 Ta ble 9-7. AMMONI RUN 28B RUN 33B 0.43 257 3.01 162 0.18 <0.03 33.0 292 95.8 11.5 220 No sample 276 296 1.02 <0.03 1505 1250 2.08 4.95 AMMONIA IN LIGNITE GASIFICATION PLANT SAMPLES 367 Because of the time required to perform the standard analyses and because of the easily oxidized species of the ions involved (SO! and S ), it is necessary to take steps to minimize loss of these components prior to analysis. Ammonia also tends to be lost unless the sample is acidified to prevent evolution of NH 3 gas. Table 9-8 shows typical loss of components with time for three samples: Gasifier Quench (GO) Regenerator Quench (RQ) , and PreheaterC L-114). The samples were ' obtained during Run 39. Table 9-9 shows results obtained on four samples obtained on November 18, 1976, during Run 40B. The fourth sample was obtained from the S0 2 scrubber. CONCENTRATION IN MG/LITER DATE PRE HEATER L-114 GASIFIER QUENCH REGENERATOR QUENCH AMMONIA Sept : 28, 1976 Oct 20 SULFIDE Sept 28 Oct 20 Nov 4 Nov 10 SULFITE Sept 28 Oct 20 Nov 4 Nov 10 Nov 16 *S ample lost <0.03 <0.03 3.0 0.03 8.0 <0.03 29.5 30.4 19.5 3.4 47.0 49.9 5.0 3.8 <0.03 53.7 49.2 0.6 1.8 41.2 7.3 1.5 2.2 Table 9-8. LOSS OF AMMONIA, SULFIDE, AND SULFITE WITH TIME FOR SELECTED PLANT PROCESS STREAM SAMPLES OBTAINED DURING RUN 39 368 CONCENTRATIONS IN MG/LITER DATE PREHEATER GASIFIER REGENERATOR SO 2 L-114 QUENCH QUENCH SCRUBBER AMMONIA Nov 22, 1976 109.2 * 106.3 387.7 Dec 1, 1976 43.1 2141 114 84.0 Dec 8, 1976 - 2119 136 176 Jan 5, 1977 - 1704 114 ~z ~+ Jan 5, 1977 1983 99.9 SULFIDE Nov 22, 1976 2.2 69.2 2.4 18.6 Dec 1, 1976 1.0 28.0 2.0 39.0 Dec 8, 1976 6.0 2.0 3.0 41.0 Jan 5, , 1977 1.0 5.0 + 1.0 49. + Jan 5, , 1977 5.2 40. 5 + SULFITE Nov 22 : , 1976 18.3 42.6 1.1 37.4 Dec 1, , 1976 56.0 71.0 6.0 120.0 Dec 8; , 1976 88.0 5.0 0.03 59.0 Jan 5 , 1977 40.0 45.0 0.03 * *Sample lost +Sample stabilized on December 8, 1976 Table 9-9 LOSS OF AMMONIA, SULFIDE, AND SULFITE WITH TIME FOR SELECTED PLANT PROCESS STREAM SAMPLES OBTAINED DUR- ING RUN 40 B Tables 9-10 and 9-11 show the results of a comparison study of re- sults from standard methods and results obtained using single-ion electrode measuring devices for sulfide and ammonia respectively. No consistent results were obtained using the sulfite single-ion electrode. 369 SULFIDE, MG/LITER SAMPLE METHOD TITRATION 1-A 0.19 0.112 2-B 3.24 0.20 3 2.17 0.56 16 -A 0.03 0.02 16-B 84.1 134.7 18 0.49 0.02 Table 9-10. COMPARISON OF ! 5ULFIDR STNRI.F. PI SINGLE ELECTRODE STANDARD SUBTRACTION 0.056 0.25 0.062 66.1 0.009 WITH STANDARD METHOD VALUES FOR SELECTED SAMPLES FROM PLANT RUN 33B SAMPLE 3 13 Table 9-11 STANDARD METHOD 11.5 218.5 AMMONIA, MG/LITER SINGLE ELECTRODE KNOWN ADDITION METHOD 8.5 194 COMPARISON OF AMMONIA SINGLE ELECTRODE MEASUREMENTS WITH STANDARD METHOD VALUES FOR SELECTED SAMPLES FROM PLANT RUN 33B 9.1.3.3.2 Particulate Size Distribution. Selected solids collected from plant process stream samples from Runs 27C and 28B were analyzed using the SEM. The results are given in Table 9-12. Sample numbers refer to liquid samples taken for chemical analysis. SAMPLE IB H 2 wash HC1 wash 2B H 2 wash 4B H 2 wash HC1 wash 16B H 2 wash HC1 wash RUN 27C MEAN SIZE, X (ym) 1.29 - 1.08 1.94 - 1.58 3.01 - 2.33 2.07 - 1.6 3.11 i 5.41 0.86 - 0.62 2.51 ± 1.7 RUN 28B MEAN SIZE, X (ym) 2.54 i 1.62 1.84 - 1.32 1.67 - 1.35 2.79 - 1.04 Table 9-12. PARTICLE SIZE ANALYSIS RESULTS FOR SELECTED SAMPLES FROM RUNS 2 7C AND 28B. (MEAN SIZE - STANDARD DEVIATION) 370 The results are given in Table 9-13. 13 3 3 Radian Corporation Samples. During the past few months several samples collected by Radian Corporation have_been analyzed in-order to help^ evaluate their sampling procedure SAMPLE (Feb 8, 1977) Gasifier Quench Spiked Water Spiked Gasifier Quench Water (Feb 9, 1977) SSB(S0 2 Scrubber Bottoms) CSD (Chemical Sump Discharge) PVI (Preheater Venturi Inlet) PVI-2(Preheater Venturi Inlet) RQW (Regenerator Quench Water) GQW(Gasifier Quench Water) RW(Raw Water) (Mar 8, 1977) Gasifier Quench Spiked Water Spiked Gasifier Quench Water (Mar 11, 1977) Time Gasifier 0000 Quench 0300 0600 0930 1130 1230 1330 1430 1539 1630 SULFITE SULFIDE AMMONIA (MG/LITER) (MG/LITER) (MG/LITER) 55 68 2019 25 33 29 77 106 1972 141 mm — 255 _ _ 242 — 222 34 12 104 17 81 1776 <0.01 2 sample lost 2059 _ — 1121 3194 1478 1750 1841 1825 1827 1693 1823 628 254 162 (June 6, 1977) SSB(S0 2 Scrubber Bottoms) CSD (Chemical Sump Discharge) PVI (Preheater Venturi Inlet) PVO (Preheater Venturi Outlet) RQW (Regenerator Quench Water) GQW(Gasifier Quench Water) RW(Raw Water) Table 9-13. ANALYSES RESULTS OF SAMPLES SUBMITTED BY RADIAN CORPORATION Sample lost 0.4 8.0 14.0 65 12 0.2 14.0 2.6 2.5 5.0 14.0 52.0 1.4 231 13.0 249 224 83 1089 10 371 9.1.3.4 Summary Procedures for the determination of sulfite, sulfide, and ammonia have been adapted from standard waste water analysis methods. These methods have been shown to give consistent and reliable results. The use of single-ion electrodes to determine these quantities was shown to be unreliable. It might be possible to use the ammonia elec- trode under certain conditions but it will not be generally usable. It was also shown that unless the samples can be analyzed within 1 day they must be stabilized to prevent loss of all three anions. During the last few months several samples taken by Radian Corporation have been analyzed in order to help evaluate their sampling procedure. 9.1.4 STACK GAS MONITORING 9.1.4.1 Introduction The initial proposal for this part of the project was to analyze the gas vented from the lignite preheater using gas -liquid chromatographic methods. The object of the proposed work was twofold: (1) To assist operating engineers to determine the best operating conditions for the plant. (2) To predict, based on measured data, what air pollution threat, if any, would be posed by a full-scale plant. Other gaseous emissions from the pilot plant could be measured if it were found to be desirable. 9.1.4.2 Results A limited amount of data was gathered. Some of the time that the plant was in operation our chromatograph was not operable so no data was obtained. In the fall of 19 73 a more sensitive chromato- graph, the Carle 8515, was obtained in order to detect S0 2 at very low levels. This dual-column unit was installed in the shed on the pilot plant roof and was operable by June 19 74. Data were being obtained during the July 9 to 16 run when condensed water and carbon from the stack leaked into the shed via an electrical conduit on July 14. Only one record, July 12, was recovered after the equipment was completely covered with water and carbon. The results indicate that generally the C0 2 content was twice the CO content. There was some fluctuation in absolute C0 2 and CO percentage as shown in Table 9-14. 372 TIME C0o% co% 10:00 AM 9 J. 6 11:00 - b.^ L-J 12:00 ? ; 1:00 PM 5 2 - 7 2:00 5 2 ' 7 3:00 5 2 ' 7 6:00 5 2 - 7 Table 9-14. C0 2 AND CO CONTENT OF STACK GAS, JULY 12, 1974 9.1.4.3 Summary In the stack gas monitoring phase, only a limited amount of data was gathered. The results obtained indicate that the monitoring of the stack gas could possibly be useful in evaluating plant opera- tion The C0 2 to CO ratio appears to be constant at about 2. The gas chromatograph technique would also be suitable to measure N 2 , 2 , HpO NO H 2 S and S0 2 . Consol management personnel requested that the' stack gas monitoring effort be redirected into other areas be- cause of the atypical characteristics of the pilot plant in relation to a commercial plant. There has been no activity in stack gas monitoring since the interim report was written in October 1974. For more detail on this phase of the waste monitoring project, reter that report which is included as Subsection 9.2. 9.1.5 PARTICULATE MONITORING 9.1.5.1 Introduction Monitoring of preheater vent stack particulate emissions was, in the initial stages of plant operation, deemed to be an important consideration both from the viewpoint of potential environmental impact and as an indicator of plant operating conditions. The funda- mental objective of the particulate monitoring was to obtain con- tinuous monitoring of both particulate volume density (mass flow) and mean particle size. 9.1.5.2 Results Since the vent stack is preceded by a wet scrubber in the vent gas stream flow, the initial specifications given for the stack gas were a hot supersaturated C0 2 and air flow containing very low con- centrations of sub-10-micrometer particles, predominantly carbon Based on these specifications, an optical nephelometer was installed on the vent stack near its terminus. The system was required to yield mean particle size and size averaged density, so a system utilizing optical scattering at 45 degrees, and direct transmission of an optical beam transverse to the stack flow, was to be utilized. After initial difficulties in getting the system installed, it was quickly dis- covered that the particulate emission rate far exceeded that initially 373 anticipated. The fouling of the nephelometer optics was immediate and continuous. Various modifications of the optical system were tried, but were basically unsuccessful. Spot sampling methods were used to obtain particulate samples from the vent stack. These samples were then analyzed by optical and scanning electron microscopes. Individual particle sizes fall into the l-to-10-micrometer range. The size range does not seem to be limited by microscope resolution and a realistic range can be obtained by this method. 9.1.5.3 Summary Spot samples of particulate emissions have been obtained from the vent stack yielding particles in the l-to-10-micrometer range. Agglomeration produces substantially increased apparent particle size. Due to the agglomeration plus the large particulate load of the pilot plant stack, optical transmission measurement is deemed unfeasible. Because it is not expected that particulate matter produced by the pilot process will be representative of a commercial plant, Consol representatives made the decision in June 1974 to discontinue this part of the waste monitoring project. There has been no activity in particulate monitoring since June 1974, so for more detail on this phase of the waste monitoring project, refer to the interim report "Waste Monitoring at the C0 2 Gasification Pilot Plant, Rapid City, South Dakota," October 1, 1974, which comprises Subsection 9.2. 9.1.6 GENERAL SUMMARY The pH of the solids holding pond varies from about 7.0 to 9.5 during plant operation. Dead-burned dolomite can raise pH to 11 or 12 while char, lignite, and dolomite do not appreciably. affect the pH. The dissolved oxygen content of the waste water pond varies from a few tenths to about 7.0 parts per million during plant run periods. There seems to be some tendency for the dis- solved oxygen to decrease during plant runs that extend over a period of time. It is also common to see a good deal of fluctua- tion in the amount of dissolved oxygen of the pond discharge water during periods of plant operation. The waste water sample analyses apparently do not indicate any serious problem areas. The biochemical oxygen demand varies over a range from about 1.1 to 65 milligrams per liter. There appears to be an increase in the biochemical oxygen demand as plant operation extends over a period of time. This apparent trend may be related to a buildup of ammonia in the waste water. The average flow rate of waste water from the pond during periods of plant operation would probably be in the range of 110 to 130 gallons per minute. A typical average value of suspended solids in the waste water being dis- charged from the pond during periods of plant operation would be 130 MG/liter with a typical maximum value being in the 350-to-400- MG/liter range. Procedures for the determination of sulfite, sulfide, and ammonia have been adapted from standard waste water analysis methods. These methods have been shown to give consistent and reliable results. The use of single-ion electrodes to determine these quantities was shown to be unreliable. It might be possible to use the ammonia elec- trode under certain conditions, but it will not be generally usable. It was also shown that unless the samples can be analyzed within 1 day they must be stabilized to prevent loss of all three anions. In the stack gas monitoring phase, only a limited amount of data was gathered. The results obtained indicate that the monitoring of the stack gas could possibly be useful in evaluating plant opera- tion. The C0 2 to CO ratio appears to be constant at about 2. The gas chromatograph technique would also be suitable to measure N 2 , 2 , H 2 0, NO, H 2 S, and S0 2 . In late 1974 Consol management personnel requested that the stack gas monitoring effort be redirected into other areas because of the atypical characteristics of the pilot plant in relation to a commercial plant. Spot samples of particulate emissions have been obtained from the lignite preheater vent stack yielding particles in the l-to-10-micro- meter range. Agglomeration produces substantially increased apparent particle size. Due to the agglomeration plus the heavy particulate load of the pilot plant stack, optical transmission measurement is deemed unfeasible. Because it is not expected that particulate matter produced by the pilot process will be representative of a commercial plant, Consol representatives made the decision in June 1974 to dis- continue this part of the waste monitoring project. 375 9.2 WASTE MONITORING AT THE C0 2 ACCEPTOR PROCESS GASIFICATION PILOT PLANT INTERIM REPORT, OCTOBER 1, 1974 9.2.1 INTRODUCTION The waste monitoring project of the CO Acceptor Process Gasification Pilot Plant at Rapid City, South Dakota, was originally set up to moni- tor all waste products (gas, solids, and liquids) produced by the plant Accordingly, the project was divided into the following three parts: (1) waste water, (2) stack gas, and (3) particulate matter under the direction of Ronald J. Schmitz, J. Haworth Jonte, and Carl Gruber, respectively. Because of the difference in the monitoring problems involved, this subsection is divided into three parts covering these three different areas. 9.2.2 WASTE WATER MONITORING 9.2.2.1 Introduction The goals of the waste water monitoring were to determine the changes in pH, dissolved oxygen, and other chemical changes in the liquid waste discharged from the pilot plant and to relate these changes to possible design or hazard situations which would be pertinent in the design of commercial facilities utilizing the C0 2 Acceptor Process should the pilot phase be successful. 9.2.2.2 Procedures A monitoring station at the outlet end of the pond contains equipment for the continuous recording of pH and the dissolved oxygen in parts per million of the waste water. The pH is measured with a Chemtrix Type 40 pH meter. The dissolved oxygen is measured with a Beckman Model 735 dissolved oxygen analyzer. Esterline Angus Minigraph strip-chart recorders are used with both of the above pieces of equipment to give a continuous record of results. Some difficulties with continuous record- ing of pH have been encountered recently. This problem is presently being worked on. Water samples are ordinarily taken at the pond outlet in gallon quantities during the time the plant is in operation with time and date being re- corded. These samples are then analyzed by the South Dakota School of Mines and Technology (SDSMT) Engineering and Mining Experiment Station. For the most part, samples are analyzed in accordance with established procedures as outlined in the current edition of "Standard Methods for the Examination of Water and Wastewater," published by the American Public Health Association. 376 9.2.2.3 Results 9.2.2.3.1 Waste Water pH Variation Relative to Plant Operation. The pH of the waste water pond varied from 6.7 to 11.0, but most values ranged be- tween 7.5 and 9.0 during plant runs. Some variation in pH exists in the discharges from the two different inlets to the pond. The east inlet is basically blowdown from various systems and floor drains. It generally has a lower pH (6.5 to 7) than the west inlet (8 to 9) which is water from sources that can include ash, attrited acceptor, lignite or char fines, quench water, and water from the S0 2 scrubber. The values of pH for both inlets varied all the way from 6 to 10. Correlation of the changes in pH with various plant operations such as char grind, dolomite grind, etc., is not always possible. Work in the laboratory with saturated solutions of ground char, lignite, dolomite, or mixtures of these materials in tap water shows the pH starts out at about 7.5 and increases with time over several days to about 8.0. A saturated solution of dead-burned dolomite in tap water starts out at about 10.5 and increases to about 11.5 or 12.0 as it sets for several days. This suggests that char, lignite, or dolomite grinds will not cause much pH change in the pond. On the other hand, the pH of the pond may become quite high while the regenerator is being operated if a significant amount of dead-burned dolomite is transferred from the regenerator to the waste pond. 9.2.2.3.2 Waste Water Dissolved Oxygen Variation Relative to Plant Operation. The dissolved oxygen in the waste water does not seem to behave in any completely predictable fashion during periods of plant operation. During the plant run that included the period from July 4 through July 16, 1974, the dissolved oxygen went from 4.5 PPM on July 4 down to about 0.3 PPM by July 9. The level of DO stayed at about this level until the run ended July 16. It should also be noted that this coincides with the hottest part of the month. It then went back up to 4.5 PPM in 2 days. This type of trend has not been noticeable in any other run and may not be significant, but it will be watched for in the future. 9.2.2.3.3 Chemical Changes in Waste Water. A number of pond water samples have been analyzed and these results are summarized in Table 9-15. For comparison purposes, an analysis of Rapid City water, which makes up the input water to the plant, is included. Table 9-15 shows in general that the waste water increases in most of the categories analyzed. However, none of the changes seem to be significant on the scale of the pilot plant operation. Phenols in the samples analyzed range from 0.004 to 15 PPM, but we believe that these phenol figures may not be completely reliable. The U.S. Public Health Service 1962 drinking water standards are included in Table 9-15 solely for comparison purposes and have no signi- ficance in terms of waste water samples. Waste water standards vary from one treatment plant to another and also are presently changing at a fairly rapid rate so it is difficult to list any standard permissible 377 co 2 X rH a. ^ OS a OS CtJ UJ < 3= < H co c_> ft as r^ m \ f— 00 <: o o LO o I vO I CN I -* ft ft O o tO o \o cn vo ft o CN o O CTi LO ft ft O ft CN H M 00 00 LO to O o o O LO CM .O o o o o LO LO (N CN LO ft 00 ft ft o to O O 1 CN 00 vO i- 1 1 LO 00 G\ i-l 1 CM ft ft O CNI o CN ft 00 CN LO CN vO i— too \0 tO O ft CN i— I (N O —I ftOftlOvOLOLOOO ft * ^f o m« LO LO i— I LO CN O t-» CN O ft o O C7> LO CTl \Ov£5^OCN00tO i-H r-H (N r-H to CN LO \D o to o CN Ol CN f-» i-H i-H CN r-H to E- r- i-H CN Q uj ^ 2 -J ooolo^J"olooo N ^ Nh tO CN CN r-H -a -a -8 o o -a T3 CO L0 0) > X) -O 13 r-H C UJ O 0) J > XI in ex §! -H C JQ O iH UJ UJ cd ct3 cd rd i— t CO r-H a h +-> 4-> +-> r-H O rH O >- < O O O O '.0 O CO H O. Hf-H > > r-H V) -O o to fn CO 4) rt •-H fn a> to • rj CCJ +J rO CO 9 > c fH C +-> r-H g)^ O ccj O O rd rt o E- U 2 H U S co o LO o fi- eri o 00 ft ft CN CTi ft LO O CTi CN CN cr> o to % c +-> o O rH a hH o *-> (0 -H CU C TO r-H bo rt C^^H S < •H ,— N £ ft A — N CO cO c 2 O a C to +-> 4-> rH $ w 0) +-> CJ J= XO &0-H 3 Ch 4J to < — < o fc^ rH -H iH 1) 4-> O C XX) CO +-) JS i-H X u 3 O CO § •H +-> CO ■U CO +J "T3 c a> LO <+H to o +j • H u 3 cd bO rH •w < -J • o < rH o • c rH Xi a o - oa m CO < oS < o u «s w H a. < u CO C7> h •H IX, O ^ 384 o a v 4J ^ O > 0) — n o •H 0) Q Pi -4- ^1 o p 6 •H CO J^ C / <-^\ CD / <~> u r ^nr c ' 1 M T H J-i 4-> CD jC -. P 60 O 4J •H W ^| H-l rf l •n W o T J«l M P p c c CO O 4J O 60 •H «H C! c e ±> O •r-< CO ex. •H ^ C O P <1) CO o p U 4-1 c p H O •H ttj P u I X! CO W — 1 *5 i n * * UJ H CO >- CO < u I— I a. o CM I a> CD b0 385 average particle size (assuming carbon to be the particulate) is obtained as well as the particulate mass density with a lower cutoff limit of approximately 0.45 micrometer diameter. 9.2.4.3 Results 9.2.4.3.1 Optical System. After initial difficulties in getting the collar in- stalled, it was quickly discovered that the particulate emission rate far exceeded that initially anticipated. Fouling of the sapphire optics was immediate and continuous. During even an aborted run with char the windows were rendered opaque within a few hours by a substantial coating of char. In an attempt to alleviate this condition, new mountings for the windows were devised and installed. The windows were moved back from the stack well and a flow of inert gas was directed tangentially across the inside face of the window, this in effect providing a gentle flow of gas away from the window and into the stack reducing the back flow of "dirty" stack gas. This technique has proven partially successful since the foul rate was greatly reduced. After three partial runs, the windows were 60 percent transmissive. Obviously, the level of carbon in the stack is still great enough to render optical methods of monitoring extremely difficult, if workable at all, in the long term. The effects using lignite have not yet been determined since the collar was removed from the stack before lignite runs were made. 9.2.4.3.2 Secondary Methods. In an attempt to obtain some information on particulate matter in the stack flow, two other sampling techniques were tried. In the first, a large impinger was installed in the gas sampling line (1/4-inch stainless steel tubing) connected to the collar. This served two purposes: first, any particulate and liquid water condensate was removed from the gas flow prior to the gas chromatograph sampling line (this also had been repeatedly fouled by carbon despite inline filter- ing) ; and second, a sample of precipitated carbon could be removed from the impinger along with a water sample. After a run an approximate measure of total steam and suspended carbon carried by the stack flow can be obtained. The sample removed from the impinger is then analyzed by optical/SEM microscopy. This was done for at least two runs this summer with marginal success since very small samples were obtained. A second method of direct sampling was used when the preheater was operating but the rest of the plant was down. A glass slide was inserted directly into the stack through one of the optical ports. Within a few seconds it became liberally coated with a layer of carbon. Results of examination of samples obtained in this manner by optical microscope are shown in Figure 9-3. Note that the particles tend to 386 » - • •V ■*..-* ' » * 5 MAGNIFICATION: 10.0 jzm/mm MAGNIFICATION: 1 .k jzm/mm Figure 9-3. PHOTOMICROGRAPHS OF GLASS SLIDE GRAB SAMPLES FROM PREHEATER VENT STACK. NOTE AGGLOMERATION OF PARTICLES 387 agglomerate to a high degree. In fact, "blobs" as large as 1 to 3 CM in diameter have been observed being ejected from the top of the stack. The reason for this is as yet undetermined. Visual examination under the microscope shows a substantial fraction of the carbon particles to have an iridescent bronze-colored coating as yet of undetermined nature. Shortly we will have SEM pictures of the same samples used in the optical microscopy work, but from overt appearances the particles range in size from not less than 2 to 3 micrometers to approximately 20 micrometers (and probably larger uncollected sizes) and are a mixture of angular to slightly rounded fragments. The size range does not seem to be limited by microscope resolution and a realistic range can be obtained by this method. 9.2.4.4 Summary Spot samples of particulate emissions have been obtained from the vent stack yielding individual particle sizes in the l-to-10 micrometer range. Agglomeration produces substantially increased apparent particle size. Due to the above factors plus the extensive fouling of optics that occur, optical transmission measurement of particulate properties is deemed unfeasible. Many monitoring problems have been encountered due to the exceptionally heavy load of particulate matter in the pilot process. Since it is not expected that the particulate matter produced by the pilot process will be representative of a commercial plant, the decision was made in June 1974 by Consol officials to discontinue this part of the waste monitor- ing project. 9.2.5 GENERAL SUMMARY The pH of the pond waste water varies from about 7.5 to 9 during plant operation. Dead-burned dolomite can raise pH to 11 or 12 while char, lignite, and dolomite do not appreciably affect the pH. The dissolved oxygen in the waste water can become very low during some gasification pilot runs. However, on other runs there is little effect and the reason for the inconsistency is not known. Waste water samples taken at various times indicate that there is apparently no serious buildup of dissolved and suspended substances. However, this cannot be fully evaluated until projected to the scale of a commercial plant. Trace element analysis of one water sample also indicates no serious buildup. The average flow rate of waste water out of the pond during plant operation is about 100 GPM. In the stack gas monitoring phase, only a limited amount of data has been gathered to date. The results obtained indicate that the monitor- ing of the stack gas could possibly be useful in evaluating plant 388 operation. The C0 2 -to-C0 ratio appears to be constant at about 2. No data has been gathered since pilot plant personnel disconnected the sample line about August 10, 1974. Consol gasification plant personnel have requested that the stack gas monitoring effort be re- directed into other areas because of the atypical characteristics of the pilot plant in relation to a commercial plant. Future efforts will emphasize waste water chemical properties from various subsystems within the plant. Spot samples of particulate emissions have been obtained from the vent stack yielding particles in the l-to-10-micrometer range. Agglo- meration produces substantially increased apparent particle size. Due to the agglomeration plus the much greater particulate load of the pilot plant stack, optical transmission measurement is deemed un- feasible. Because it is not expected that particulate matter produced by the pilot process will be representative of a commercial plant, Consol representatives made the decision in June 1974 to discontinue this part of the waste monitoring project. 389 REFERENCES (1) G. P. Curran and E. Gorin, Phase II , Bench-Scale Research on CSG Research , R and D Report No. 16, Interim Report Number 3, Book 2, Consolidation Coal Company, March 1968. (2) K. K. Kelly, "Contributions to the Data on Theoretical Metallurgy," Bulletin 584, U.S. Bureau of Mines, 1960. (3) Handbook of Chemistry and Physics , 50th Edition, The Chemical Rubber Company, B-260, 1970. (4) C. E. Wicks and F. E. Block, "Thermodynamic Properties of 65 Elements — Their Oxides, Halides, Carbides and Nitrides," Bulletin 605, U.S. Bureau of Mines, 1963. (5) C. E. Fink, J. D. Sudbury, and G. P. Curran, "C0 2 Acceptor Gasification Process," presented at the 77th National Meeting AICHE, Pittsburgh, Pennsylvania, 1974. (6) C. E. Fink, G. P. Curran, and J. D. Sudbury, "C0 2 Acceptor Process Pilot Plant— 1974," presented at the Sixth Synthetic Pipeline Gas Symposium, Chicago, Illinois, 1974. (7) C. W. Zielke, Conoco Coal Development Company Laboratory, Library, Pennsylvania, personal communication to M. C. Fuerstenau, South Dakota School of Mines and Technology. (8) C. W. Zielke, "Exploratory Studies of Solid Compositions that Lead to Agglomeration of Melt Formation," Consolidation Coal Company Memorandum RM-13496, Project Number 537.40, 1974. (9) J. W. Koger, "Chromium Depletion and Void Formation in Fe-Cr-Ni Alloys During Molten Salt Corrosion and Related Processes," Advances in Corrosion Science and Technology , Volume 4, Edited by M. G. Fontana and R. W. Staehle, Plenum Press, New York, New York, 1974, pp. 245-313. (10) J. W. Koger, "Corrosion Product Deposition in Molten Flouride Salt Systems," Corrosion , Volume 30, Number 4, April 1974, pp. 125-130. (11) J. W. Koger, "Flouride Salt Corrosion and Mass Transfer in High Temperature Dynamic Systems," Corrosion , Volume 29, Number 3, March 1973, pp. 115-122. (12) "Development of a Molten Carbonate Process for Removal of Sulfur Dioxide from Power Plant Stack Gases," Progress Report Number 2, Part III, Materials Studies, April 1, 1968, to October 27, 1968, Atomics International, North American Rockwell, Canoga Park, California. 390 (13) H H Uhlig, Corrosion and Corrosion Control , John Wiley and Sons, Inc., New York, New York, 1963, pp. 13-14 and 312. (14) G. P. Smith, M. E. Steidlitz, and E. E. Hoffman, "Corrosion and Transport in Fused Sodium Hydroxide, Part I," Corrosion , September 1957, pp. 561t - 564t. (15) G P. Smith, M. E. Steidlitz, and E. E. Hoffman, "Corrosion and Transport in Fused Sodium Hydroxide, Part II," Corrosion, September 1957, pp. 627t-630t. (16) G P. Smith, M. E. Steidlitz, and E. E. Hoffman, "Corrosion and Transport in Fused Sodium Hydroxide, Part III," Corrosion, January 1958, pp. 47t-52t. (17) G. J. Janz, A. Conte, and E. Neuenschwander, "Corrosion of Pt Au, Ag and Refractories in Molten Carbonates," Corrosion, August 1963, pp. 292t-294t. (18) M. G. Fontana and N. D. Greene, Corrosion Engineering , McGraw-Hill Book Company, New York, New York, 1967, pp. 347-353. (19) M. G. Fontana and N. D. Greene, op. cit., pp. 355-356. (20) H. H. Uhlig, op. cit., pp. 168-169. (21) H. H. Uhlig, op. cit., pp. 266-270. (22) L E. Paulson and W. W. Fowkes, "Changes in Ash Composition of North Dakota Lignite Treated by Ion Exchange," U.S. Bureau of Mines, RI 7176, 1968. (23) J N Mukherjee, "On the Nature of Hydrolytic Adsorption with Reference to the Adsorption of Electrolytes and of Water," J- Indian Chem. Soc , II_ 1925, pp. 191-224. (24) E N Pollack, "Trace Impurities in Coal," Technical Report Number 325, Ledgemont Laboratory, Kennecot Copper Corporation, Lexington, Massachusetts, August 1972. (25) D. N. Baria, "A Survey of Trace Elements in North Dakota Lignite and Effluent Streams from Combustion and Gasification Facilities," The Engineering Experiment Station, University of North Dakota, Grand Forks, North Dakota, May 1975. (26) R. R. Ruch, H. J. Gluskoter, and E. J. Kennedy, "Mercury Content of Illinois Coals," Environmental Geology Notes , Number 43, Illinois State Geological Survey, February 1971. 391 (27) R. R. Ruch, H. J. Gluskoter, and N. F. Shimp, "Occurrence and Distribution of Potentially Volatile Trace Elements in Coal," Environmental Geology Notes , Number 61, Illinois State Geological Survey, April 1973. (28) R. C. Diehl, E. A. Hattman, H. Schultz, and R. J. Haren, "Fate of Trace Mercury in the Combustion of Coal," Technical Progress Report 54, Bureau of Mines Managing Coal Wastes and Pollution Program, U.S. Department of the Interior, May 1972. (29) A. Attari, "Fate of Trace Constituents of Coal During Gasification," PB-223001, National Technical Information Service, U.S. Department of Commerce, August 1973. (30) Radian Corporation, "Coal Fired Power Plant Trace Element Study," Environmental Protection Agency, Rocky Mountain Prairie Region, Region VIII, September 1975. (31) D. J. VonLehmden, R. H. Jungers, and R. E. Lee, Jr., "Evaluation of Analytic Techniques for the Determination of Trace Elements in Coal," Abstracts with Programs , Volume 4, Number 7, G.S.A., October 1972, pp. 698-699. (32) M. D. Schlesinger and H. Schultz, "An Evaluation of Methods for Detecting Mercury in Some U.S. Coals," Bureau of Mines Report of Investigation 7909, 1972. (33) J. V. O'Gorman, N. H. Suhr, and P. L. Walker, Jr., "The Determination of Mercury in Some American Coals," App. Spectros ., Volume 26, Number 1, January/February 1972. (34) "Trace Metals: Unknown, Unseen Pollution Threat," Chemical and Engineering News , Volume 49, Number 29, July 19, 1971, pp. 29, 30, 33. (35) J. R. Jones and E. N. Pollock, "The Determination of Trace Elements in Coal," Abstracts with Programs , Volume 4, Number 7, G.S.A., October 1972, p. 556. (36) R. M. Abernathy, M. J. Peterson, and F. H. Gibson, "Spectrochemical Analyses of Coal Ash for Trace Elements," Bureau of Mines Report of Investigation 7281, 1969, p. 30. (37) P. G. Freeman, W. Becker ing, and W. W. Fowkes, "Some Trace Elements in Western Coals and Their Combustion Products," Abstracts with Programs , Volume 4, Number 7, G.S.A., October 1972, p. 510. 392 (M , D. J. Swaine, "TVace Blunts in Australian Co als « ^"gcwwith Programs , Volume 4, Number 7, G.S.A., Octooer i»/ , pp ^ t n vinP "Composition of Coal, Southwestern (39) Lt^^l^^CtSkSk. " 4, Nuabe* 7, G.S.A.. October 1972, pp. 683-684. (40) E S. Hertzog, "Determination of Arsenic in Coal," Ind^U^ Chem., Volume 7, 1935, pp. 163-165. (41) R. E. Stanton and A. J. McDonald, "^^^^^/^f ™ " Soils and Sediments with 3, 3' -Diammobenzidine,' The Analyst , Volume 90, August 1965, pp. 497-499. (42) G I. Spielholtz and H. Diehl, "Wet Ashing of Coal with ^rchloric C } Acid Mixed with Periodic Acid for the Determination of Sulphur and Certain Other Constituents," Talanta, Volume 13, 1966, pp. 991 (43) R. A. Rasmussen, "Analysis of Trace Organic Sulfur Compounds in Air," American Laboratory , December 1972, pp. bb-bi. (44) G C B. Cave, "The Collection and Analysis of Odorous Gases from ( Kraft Pulp Mills," Tappi, Volume 46, Number 1, January 1963, pp. 1-20. r.5l R K Stevens A. B. O'Keefe, and G. C. Ortman, "Absolute Calibration ( } of a' Flame Photometric Detector to Volatile Sulfur Compounds at Sub-Par?-Per-Million Levels," Environmental Science and technology , Volume 3, Number 7, July 1969, pp. 652-655. (AM R K Stevens and A. E. O'Keefe, "Modern Aspects of Air Pollution (4 } Luorlng!" Analytical Chemistry , Volume 42, Number 2, February 1970, pp. 143A-148A. cat> r F Pecsar and C H. Hartman, "Automated Gas Chromatographic m Llysis of Sulfur Pollutants," Journal of Chromatogr aphic Science, Volume 11, September 1973, pp. 492-502. (48) S S. Brody and J. E. Chaney, "Flame Photometric Detector," Jpurnal_ of Gas Chromatography , February 1966, pp. 42-46. (49) J. F. O'Donnell, "Element-Selective Detectors for Gas Chromatography," American Laboratory , February 1969, pp. 31-36. (50) P. A. Gibbons and K. A. Goode, "Sulfur-Selective Detector in Natural Gas Chromatography," Gas Journal , October 9, 1968, pp. Z/-M. 393 (51) E. L. Obermiller and G. 0. Charlier, "Gas Chromatographic Separation of Nitrogen, .Oxygen, Argon, Carbon Monoxide, Carbon Dioxide, Hydrogen Sulfide, Carbonyl Sulfide, and Sulphur Dioxide," Journal of Chromatographic Science , Volume 7, September 1969, p. 580. (52) B. Mitchell and R. DeLew, "A New Column for Trace Analysis of Sulfur-Containing Gases," Application Tips , Number 90, Varian Aerograph, August 1971. (53) W. L. Thornsberry, Jr., "Isothermal Gas Chromatographic Separation of Carbon Dioxide, Carbon Oxysulfide, Hydrogen Sulfide, Carbon Disulfide, and Sulfur Dioxide," Analytical Chemistry , Volume 43, Number 3, March 1971, pp. 452-453. 394 ^.'^"S^VOFILLINOIS-URBANA 3 0112 084233433