United States Energy
Research and Development
Administration
Office of Fossil Energy
Washington, D.Ci 20545
United States Environmental      Industrial Environmental Res ;arch
Protection Agency          Laboratory
Office of Research and Development  Research Triangle Park, N.C. 27711
                  EPA-600/7-76-019
                  October 1976
     A DEVELOPMENT  PROGRAM
     ON PRESSURIZED
     FLUIDIZED-BED COMBUSTION
     Interagency
     Energy-Environment
                         j
     Research and Development
     Program Report

-------
                                       RESEARCH REPORTING SERIES


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

                     1.   Environmental Health Effects Research
|      •               2.   Environmental Protection  Technology
                     3.   Ecological  Research
                     4.   Environmental Monitoring
                     5.   Socioeconomic Environmental Studies
                     6.   Scientific  and Technical  Assessment Reports  (STAR)
                     7.   Interagency Energy-Environment  Research  and  Development

                This  report has  been assigned to the INTERAGENCY  ENERGY-ENVIRONMENT
                RESEARCH AND DEVELOPMENT series.   Reports  in this series result from
                the effort  funded under the 17-agency Federal Energy/Environment
                Research and Development Program.   These studies  relate  to EPA's
                mission  to  protect the public health and welfare  from adverse effects
                of  pollutants associated with energy systems.   The  goal  of the Program
                is  to  assure the rapid development of domestic  energy supplies in  an
                environmentally—compatible manner by providing the necessary
                environmental data and control  technology.  Investigations include
                analyses of the  transport of  energy-related pollutants and their health
                and ecological effects;  assessments  of,  and development  of,  control
                technologies for energy systems; and integrated assessments  of a wide
                range  of energy-related environmental issues.
               This document is available  to the public through the National Technical
               Information Service, Springfield, Virginia  22161.

-------
                                                   EPA-600/7-76-019

                                                   October 1976
                     A  DEVELOPMENT  PROGRAM

                           ON  PRESSURIZED

                    FLUIDIZED-BED COMBUSTION
                                     by

              G.J. Vogel, I. Johnson, P.Cunningham,  B. Hubble,
              S. Lee, J.  Lenc,  J. Montagna, F. Nunes, S.  Siegel,
              G. Smith, R. Snyder, S.  Saxena, W. Swift, G. Teats,
                           I. Wilson, and A. Jonke

                         Argonne National Laboratory
                            9700 South Cass Avenue
                           Argonne, Illinois 60439
                            EPA No.  IAG-D5-E681
                          ERDA No.  14-32-0001-1780
                        Program Element No. EHE623A

                               Project Officers:

            Walter B. Steen                        John F. Geffken
EPA/Indus trial Environmental Research Lab   ERDA/Office of Fossil Energy
    Research Triangle Park,  NC 27711            Washington, DC 20545
                                 Prepared for
                                               U.S. Energy Research
  U.S. Environmental Protection Agency     and Devel    ent Administration
    Office of Research and Development           ^^ Qf Fosgil Energy
          Washington, DC 20460                 Washington, DC  20546

-------
                           TABLE OF CONTENTS

                                                                    Page

Abstract	      1

Summary	      1

Introduction	•	     13

Regeneration of S02-Accepting Additive,
Bench-Scale Studies  	     13

     Materials	     14
     Equipment	     14
     Regeneration of Sulfated Tymochtee Dolomite by the
       Incomplete Combustion of Methane 	     16
     Regeneration of Sulfated Greer Limestone by the
       Incomplete Combustion of Coal	     32
     Regeneration of Sulfated Tymochtee Dolomite by the
       Incomplete Combustion of Coal	     45
     Carbon Utilization and Carbon Balance Calculations 	     48
     Effect of Coal Ash on the Agglomeration of Sulfated
       Additive During Regeneration 	     48
     Mass and Energy Constrained Model for the
       Regeneration Process  	     57

Cyclic Combustion/Regeneration Experiments	     68

     Preliminary Experiment.  Sulfation of Regenerated
       Tymochtee Dolomite 	     69
     Combustion Step, Cycles 1, 2, and 3	     72
     Regeneration Step, Cycles 1 and 2	     79
     Coal Ash Buildup During Sulfation and Regeneration
       Steps	     80

Pressurized, Fluidized-Bed Combustion:  Bench-Scale Studies ...     84

     Materials	     85
     Bench-Scale Equipment.  . 	     85
     Experimental Procedure	     88
     Replicate of VAR-Series Experiment 	     88
     Effect of Coal and Sorbent Particle Size on Sulfur
       Retention,  Nitrogen Oxide Level in the Flue Gas,
       and Combustion Efficiency	     89
     Combustion of Lignite in a Fluidized Bed of Alumina	     94
     Effect of Fluidizing-Gas Velocity on Decrepitation Rate.  .  .     95
     Binary Salt of Magnesium and Calcium Sulfate ...-..,..     95

Synthetic S02  Sorbents	,	,  .  ,  .     99

     Introduction	     99
     Preparation of Synthetic Sorbents	  ...,,....     99
     Sulfation Studies	,  .    100
                                     111

-------
                          TABLE OF CONTENTS (Contd'.i)
     Metal Oxides in a-A!203	    Ill
     Regeneration Studies 	 ............    114
     Cyclic Sulfation-Regeneration Studies. . , .  .  ,	    121
     Methods of Preparing Supports with Optimum
       Physical Properties	    124

Sorbent Attrition Studies	    131

     Introduction	    131
     Attrition of Dolomite.	    131
     Attrition of Supports and Synthetic Sorbents 	    133

Combustion-Regeneration Chemistry	,  .  . ,	    134

     Half-Calcination Reaction	    134
     Regeneration by the CaSO^-CaS Reaction ....  	    139

Coal Combustion Reactions	    145

     Determination of Inorganic Constituents in the Effluent
       Gas from Coal Combustion	«    145
     Systematic Study of the Volatility of Trace Elements
       in Coal	    148

Equipment Changes 	    156

Miscellaneous Studies 	.......    158

     Preparation and Testing of Supported Additives
       (Dow Subcontract)	    158
     Limestone and Dolomite for the Fluidized-Bed
       Combustion of Coal:  Procurement and Disposal	    159
     The Properties of a Dolomite Bed of a Range of Particle
       Sizes and Shapes at Minimum Fluidization	    162
     Mathematical Modeling:  Noncatalytic Gas-Solid Reaction
       with Changing Particle Size, Unsteady State Heat
       Transfer	.  .  .	    164

Acknowledgments	    172

References	    172

Appendix A.   Characteristics of Raw Materials Used in Fluidized-
Bed Combustion Experiments	    177

Appendix B.   Plots of Operating Data and Experimental Results of
Combustion Experiments	
                                      IV

-------
                            LIST OF FIGURES

No.                              Title                              Page

 1.  Schematic Diagram of Present Regeneration System 	     15

 2.  Particle Size Distribution of the Fines Leaving the
     Cyclone in FAC-5	     19

 3.  The Effect of Fluidizing-Gas Velocity on Sulfur
     Regeneration for Sulfated Tymochtee Dolomite 	     20

 4.  Effect of Temperature on Sulfur Regeneration for
     Sulfated Tymochtee Dolomite	     21

 5.  Pore Distributions of Dolomite Samples from Different
     Process Stages	     23

 6.  Sulfation Reaction Data Obtained with a Thermogravimetric
     Analyzer at 900°C, 0.3% S02, and 5% 02	     24

 7.  Electron Microprobe Analyses of Sulfated Tymochtee
     Dolomite Particles 	     26

 8.  Electron Microprobe Analyses of Regenerated Tymochtee
     Dolomite Particles from Experiment FAC-1R2 	     27

 9.  Electron Microprobe Analyses of Regenerated Tymochtee
     Dolomite Particles from Experiment FAC-4 	     28

10.  FAC-1, Fractional Feed and Product Particle Size
     Distributions	     30

11.  Bed Temperature and Gas Concentrations in Off-Gas,
     Experiment LCS-2	     35

12.  Bed Temperature and Gas Concentrations in Off^Gas,
     Experiment LCS-3	«...	     37

13.  Bed Temperature and Gas Concentrations in Off-Gas,
     Experiment LCS-4D	     40

14.  Geometry of Oxidizing and Reducing Zones in Relation
     to Position of Coal Injection Line	 .     41

15.  Bed Temperature and Gas Concentrations in Off^-Gas,
     Experiment LCS-7	     42

16.  Regeneration of CaO and S02 Concentration in the Dry
     Flue Gas for Sulfated Limestone and Tymochtee Dolomite
     as a Function of Solids Residence Time , .  .  ,	     44

17.  Regeneration of Calcium Oxide as a Function of  Solids
     Residence Time 	 .............     46

-------
                       LIST OF FIGURES (Contd.)

No.                              Title                              Page

18,  Predicted and Experimental Sulfur Dioxide Concentrations
     at Three Regeneration Temperatures	 .      43

19.  Schematic of DTA Apparatus	,	      53

20.  Flow Diagram for the One-Step Regeneration Process .  , . . ,      58

21.  Experimental Solids Regeneration and Predicted Increase
     in Gas Volume During Regeneration Versus Solids Residence
     Time 	 ,..,..,...      61

22.  Predicted Required Coal Feed Rate and Oxygen Concentration
     in the Feed Gas as Functions of Solids Residence Time. ....      62

23.  Predicted Off-Gas Constituent Concentrations as Functions
     of Solids Residence Time ..,..,..,.., 	 .      63

24.  Predicted Individual Heat Requirements as a Function of
     Solids Residence Time	      64

25.  Predicted Fuel Cost for Regeneration per Electric Power
     Unit Produced when Burning 3% Sulfur Coal as a Function
     of Solids Residence Time	      65

26.  Bed Temperature and Flue-Gas Composition,  Segment of
     Experiment RC-1A	 . . ,	, .      70

27.  Bed Temperature and Flue-Gas Composition,  Segment of
     Experiment REC-1 (REC-1K and -1L)..............      74

28.  Bed Temperature and Flue-Gas Composition,  Segment of
     Experiment REC-2	      75

29.  Bed Temperature and Flue-Gas Composition,  Segments of
     Experiment REC-3 (REC-3A, REC-3B)	      76

30.  Bed Temperature and Gas Concentrations in Off-Gas for
     Experiment CCS-1, the Regeneration Step of the First
     Tymochtee Dolomite Utilization Cycle 	 .      82

31.  Bed Temperature and Gas Concentrations in Off-Gas
     for Experiment CCS-2 .  	 ,.....,.,....      83

32.  Simplified Equipment Flowsheet of Bench^Scale Fluidized-^Bed
     Combustor and Associated  Equipment .  .  ............      86

33.  Detail Drawing of 6-in.-dia, Pressurized,  Fluidized-Bed
     Combustor	      87
                                      VI

-------
                       LIST OF FIGURES (Contd.)

No.                              Title                              Page

34.  CaO Concentration in a-A!203 as a Function of Calcium
     Nitrate Concentration in Reactant Solution from which
     Prepared ................  • ......  ...    100

35.  Comparison of Calculated (Eq. 1) and Experimental Sulfation
     Rates at 900°C in 5% 02-N2 of 6.6% CaO in a-A!203 (Heat-
     Treated at 800°C) as a Function of S02 Concentration ....    102

36.  Rate of Sulfation at 900°C of 1100°C Heat-Treated 6.6%
     CaO in a-A!203 as a Function of S02 Concentration ......    103

37.  Effect of Heat-Treatment Temperature on Rate of Sulfation
     of 6.6% CaO in a-Al203 at 900°C ...............    103

38.  Effect of Oxygen Concentration on the Rate of Sulfation
     of 6.6% CaO in a-A!203 at 900°C ...... , .....  . • •    104

39.  Calculated and Experimental Rates of Sulfation of 6.6%
     CaO in a-Al2Os (800°C H.T.) with 0.3% S02-5% 02<-N2 as a
     Function of Sulfation Temperature. ....,..,.••••    1°5

40.  Calculated and Experimental Rates of Sulfation of 6.6%
     CaO in «-Al203 (1100°C H.T.) with 0.3% S02-5% 02-N2 as  a
     Function of Sulfation Temperature ....... .  .....  >    106

41.  Sulfation of Sorbents Having Various CaO Loadings. ,,...,    109

42.  Sorbent Weight Gain at 900°C as a Function of Calcium
     Loading in Sorbent .............. •  •, .....
43.  Comparison of the Rates of Sulfation of Tymochtee Dolomite
     and 6.6% CaO in a-A!203 ......  ......  ........    112

44.  Comparison of Sulfation Rates of Various Metal  Oxides
     at 900°C ...................  .......    113

45.  Regeneration of Sulfated  6.6% CaO-a-Al203 Pellets, Using
     Hydrogen at 1100°C  ...........  ,..,..,...    114

46.  Regeneration of Sulfated  6.6% CaO-ci-Al203 Pellets, Using
     Carbon Monoxide at  1100°C ..................
47.  Effect of C02 Concentration  in  the  Reducing  Gas  on the
     Rate of , Regeneration  of  Sulfated  6.6%  CaO  in a~A!203  at
     1100°C  .......  ..........  ...» ......    117

48.  Regeneration of  Sulfated 6.6% CaO in a-Al203 Sorbent,
     Using 1% H2-N2  .............  •  .....  «...    H8

49.  Comparison of the Rate of Regeneration of  CaSOit  in a-r-Al203
     and Sulfated Dolomite using  3%  H2 or 3% CO at 1100°C  . . ,  . .    120
                                      vii

-------
                       LIST OF FIGURES  (Contd.)

No.                              Title                              Page

50.  Regeneration Rate of Various Metal Sulfates in a-Al203
     at 1100°C using 3% H2	    122

51.  Cyclic Sulfation of 1100°C Heat-Treated Pellets (6.6%
     CaO in ct-A!203) Using 3% S02-5% 02-N2 at 900°C	    123

52.  Cyclic Sulfation of 11.1% CaO in 1500°C H.T. a-A!203
     Granular Sorbent Using 0.3% S02-5% 02-N2 at 900°C. ......    125

53.  Cyclic Sulfation of 12.5% CaO in 1350°C H.T. a-A!203
     Granular Sorbent Using 0.3% S02-5% 02-N2 at 900°C	    126

54.  Relationship of Cumulative Pore Volume to Pore Diameter. .  .    127

55.  Sulfation Rate at 900°C of CaO in Granular Supports	    127

56.  Relation of Pore Volume to Pore Diameter for the Original
     a-A!203 Support, the CaO in a-A!203 Sorbent, and the
     Sulfated Sorbent 	    129

57.  Relation of Pore Volume to Pore Diameter as a Function
     of Indicated CaO Concentrations in the Support	    130

58.  Percent Conversion VS Time for Half-Calcination Reaction
     under 100% C02 Environment	    136

59.  Percent Conversion vs Time for Half-Calcination Reaction
     under 40% C02-60% He Environment	    137

60.  CaO Content as a Function of Reaction Time .........    144

61.  Schematic Diagram of Batch Fixed-Bed Combustor System. . .  .    147

62.  Effect of Heat-Treatment Conditions on Weight Loss of
     340°C Ash	, .  . ,	    154

63.  Overall View of New Regeneration Facility.  . . .  ,	    157

64.  Schematic of TGA Apparatus	    160

65.  Gas-Mixing Apparatus for TGA Sulfation Experiments 	    161

66.  Qualitative Dependence of the Pressure Drop Across the
     Bed,  AP, on the Fluidizing-Gas Velocity	,  .    163

67.  Gas-Solid Reaction of a Growing Particle:   The Concen-
     tration and Temperature Profiles	 •  •    166
                                    Vlll

-------
                       LIST OF FIGURES (Contd.)

No.                              Title                              Page

B-l  Bed Temperature and Flue-Gas Composition,
     Experiment PSI-1R ............... ......      184

B-2  Bed Temperature and Flue-Gas Composition,
     Experiment PSI-2 ..... .......... .  .....
B-3  Bed Temperature and Flue-Gas Composition,
     Experiment PSI-3 .  .  ...... .............       186

B-4  Bed Temperature and Flue-Gas Composition,
     Experiment PSI-4 .........  .  ...........       187

B-5  Bed Temperature and Flue-Gas Composition,
     Experiment LIG-2D .....................       I88

B-6  Bed Temperature and Flue-Gas Composition,
     Experiment LIG-2R .....................       189
                                      ix

-------
                            LIST OF TABLES

No.                              Title                              Page

 1.  Design Experimental Conditions and Extent of Regeneration
     Results for the FAC-Series	     17

 2.  Chemical Analysis of Regenerated Product, Calculated
     Material Balances, and Regeneration Results from the
     FAC-Series	     18

 3.  Qualitative Chemical Compositions of Regenerated and
     Unregenerated Samples of Additive	     31

 4.  Qualitative Chemical Composition (Obtained by X-ray
     Diffraction Analysis) of Reacted Samples from DTA
     Experiments	     33

 5.  Experimental Conditions and Results for Regeneration
     Experiments in which Arkwright and Triangle Coals were
     Incompletely Combusted in a Fluidized Bed	     34

 6.  Experimental Conditions and Results for Regeneration
     Experiments Designed to Test the Effectiveness of the
     Two-Zone Reaction in Minimizing CaS Buildup	     38

 7.  Experimental Conditions and Results for Regeneration
     Experiments with Combustion of Arkwright and Triangle
     Coal in a Fluidized Bed	     43

 8.  Experimental Conditions and Results for Regeneration
     Experiments with Sulfated Tymochtee Dolomite Using
     Triangle Coal	     47

 9.  Experimental Conditions, Carbon Balances, and Carbon
     Utilizations for Regeneration Experiments	     49

10.  Fusion Temperatures, Under Reducing Conditions, of Ash
     from Arkwright No. 2 coal, Sewickley Coal, and Triangle
     Coal	     50

11.  Compositions of Greer Limestone and Sewickley Coal Ash ...     51

12.  Coal Ash Level in Sulfated Greer Limestone	     52

13.  Differential Thermal Analysis Results and X-ray Diffraction
     Analysis of the Reaction Products of Unsulfated Tymochtee
     Dolomite	     54

14.  Differential Thermal Analysis Results and X-ray Diffraction
     Analysis of the Reaction Products of PER Sulfated Greer
     Limestone and Residual Coal Ash	     55

15.  Regeneration of Tymochtee Dolomite 	     67

-------
                         LIST OF TABLES  (Contd.)

No.                              Title                              Page

16.  Operating Conditions and Flue-Gas Analysis for Segment
     of Combustion Experiment RC-1A	    69

17.  Utilization of Calcium in Overflow, Primary Cyclone,
     and Secondary Cyclone Samples from Experiment RC-1A 	    71

18.  Screen Analysis Results for Combustion Experiment RC-1A ...    72

19.  Operating Conditions and Flue-Gas Composition for Cyclic
     Combustion Experiments	    73

20.  Steady-State Flow of Calcium Through the ANL, 6-in.-
     dia Combustor during Combustion Experiment REG-IK 	    78

21.  Experimental Conditions and Results for Regeneration Segments
     of the Regeneration Step of the First Two Utilization Cycles.    81

22.  Calculated Ash Buildup During Sulfation and Regeneration
     of Tymochtee Dolomite, Based on Silicon, Iron, and
     Aluminum Concentration Changes	    84

23.  Concentrations of Ash Constituents in Sulfated Dolomite
     and Sulfated Greer Limestone	    84

24.  Operating Conditions and Flue-Gas Compositions for VAR-6
     Replicate Experiments 	    89

25.  Sieve Analyses of Arkwright Coal Size Fractions Used in
     PSI-Series Combustion Experiments 	    90

26.  Sieve Analyses of Tymochtee Dolomite Size Fractions Used
     In PSI-Series Combustion Experiments	    90

27.  Operating Conditions and Flue-Gas Analyses for PSI-Series
     of Combustion Experiments 	    92

28.  Sulfur Retention, Combustion Efficiency, and Sorbent
     Utilization for PSI-Series of Experiments	, .  .    93

29.  Operating Conditions and Flue-Gas Analyses for Combustion
     Experiments LIG-2D and LIG-2-R.  .	    96

30.  Operating Conditions and Flue-Gas Compositions for Experiments
     REG-IK and VEL-1 to Test the Effect of Gas Velocity on
     Decrepitation 	    97

31.  Analysis of Decrepitation in Experiment VEL-1 	  ,    98

32.  Calcium Content and Percent Sulfation of 6.6% CaO in
     a-A!203 Sorbent Sulfated at 900°C 	   107
                                     XI

-------
                         LIST OF TABLES (Contd.)

No.                              Title                              Page

33.  Sorbent Weight Gain during Sulfation for Various Calcium
     Oxide Concentrations. .	   110

34.  Most Promising S02 Sorbents Based on Thermodynamic
     Screening Results 	 	   112

35.  Regeneration of CaS03 to CaO, Calculated from Weight Loss .  .   116

36.  Carbon Content of Pellets after Regeneration with Methane .  .   117

37.  Product of the Regeneration Reaction at Various Temperatures.   118

38.  Calcium Utilization and Regeneration during Sulfation-
     Regeneration Cyclic Experiments, Using 1100°C H.T. Pellets.  .   124

39.  Sulfation of Supported Sorbents 	 .....   128

40.  Pore Volume for Various Synthetic Sorbents	   130

41.  Fluidized-Bed Attrition Experiments 	   132

42.  Half-Calcination Experiments on 1337 Dolomite 	   138

43.  Summary of TGA and X-Ray Diffraction Results for Solid-
     Solid Experiments at 950°C. .	   141

44.  Results of CaS04-CaS Reaction Kinetic Measurements at 945°C .   143

45.  Elemental Concentration of High-Temperature Ash Corrected
     for Weight Losses at the Stated Temperatures	   150

46.  Elemental Concentrations in High-Temperature Ash Calculated
     on the Original 340°C-Ash Basis 	   151

47.  Elemental Concentration in High-Temperature Ash as a
     Function of Oxygen Concentration in Gas Flow	   152

48.  Effect of Heating on Weight Loss of 340°C Ash	   153

49.  Elemental Concentrations in High-Temperature Ash Calculated
     on the Original 340°C-Ash Basis	   155

A-l  Particle-Size Distribution and Chemical and Physical
     Characteristics of Arkwright Coal 	   178

A-2  Particle-Size Distribution and Chemical and Physical
     Characteristics of Triangle Coal	   179

A-3  Particle-Size Distribution and Chemical Characteristics
     of Glenharold Lignite 	 .......  	   180
                                      xii

-------
                         LIST OF TABLES (Contd.)

No.                              Title                              Page

A-4  Particle-Size Distribution and Chemical Characteristics
     of Tymochtee Dolomite. .	       181

A-5  Particle-Size Distribution and Chemical Characteristics
     of Type 38 Alundum Grain Obtained from the Norton Company.       182
                                    xiii

-------
                       A DEVELOPMENT  PROGRAM ON
                 PRESSURIZED FLUIDIZED-BED  COMBUSTION

                             Annual Report
                         July 1975—June  1976


                                  by


       G.  J.  Vogel,  I.  Johnson,  P. T.  Cunningham,  B. R.  Hubble,
      S.  H.  Lee,  J.  F.  Lenc, J.  Montagna, F. F.  Nunes,  S.  Siegel,
          G.  W.  Smith,  R. B. Snyder,  S.  Saxena,  W.  M.  Swift,
       G.  F.  Teats,  C.  B. Turner, W.  I. Wilson,  and A.  A.  Jonke
                               ABSTRACT


          Information is presented on a continuing research and
     development program in which the concept of fluidized-bed
     combustion for ultimate use in power stations and steam-
     raising applications is being investigated.  Laboratory-
     scale equipment and process development units are being used
     in investigating the effect of combustion operating variables
     on release of pollutants to the atmosphere, on process
     efficiency, and on release of compounds injurious to turbine
     materials of construction.  Methods are being investigated
     for regenerating CaSOit, a product of reaction of CaO additive
     with sulfur compounds in the combustion process.  Significant
     progress has been made in demonstrating that (1) reductive
     decomposition of CaSOi* is a viable process for recovering CaO
     for reuse in the combustion step, (2) cyclic use of limestone
     in the combustion/regeneration processes appears feasible,
     and (3) synthetic additives can be prepared that have good
     sulfur retention and sulfur release properties.


                                SUMMARY

     Ongoing studies are in progress in support of the national program on
fluidized-bed combustion for electric power and industrial/commercial
applications.  The concept involves burning coal in a fluid bed of calcium-
containing stone such as limestone.  The sulfated stone is removed from
the combustor and can be regenerated for resse in the combustor.  Both
atmospheric pressure and pressurized concepts are under investigation.  In
the pressurized combustion concept, the hot flue gas from the combustor
would be expanded through a turbine to generate electricity.

     Current studies are aimed at (1) demonstrating a regeneration process,
(2) determining the effect of cyclic (combustion-regeneration) operation on
limestone quality, (3) continuing the study of the effect of combustion
operating variables on the release of pollutants and corrosive compounds, (4)
developing and testing a synthetic stone, and (5) elucidating the chemistry
of reactions in the combustion and regeneration processes.

-------
Regeneration of SO?- Accepting Additive, Bench-Scale Studies

     The feasibility of reductive decomposition of sulfated S02-accepting
additive in a fluidized bed is being investigated.  The reduction of the
CaSQi+ constituent of the sulfated stones to CaO is favored by high temper-
atures  (>1040°C) and mildly reducing conditions.  The solid-gas reactions
by which regeneration occurs can be summarized as follows :

          CaS04 + CO -> CaO + C02 + S02                                  (1)

                  H2 -> CaO + H20 + S02                                  (2)
     The present experimental system consists of a 10.8-cm-ID refractory-
lined reactor, an off-gas cleanup system, and a continuous gas sampling
and analyzing system.  The coal and sulfated sorbent are fed separately
to a common transport feed line.

     Sulfated Tymochtee dolomite and Greer limestone have been regenerated.
Methane and coal (in separate studies) have been combusted to provide the
required heat and reductants for the reactions.  The effects of fluidizing-
gas velocity, temperature, and solids residence time have been evaluated.

     Regeneration of Sulfated Tymochtee Dolomite by the Incomplete
Combustion of Methane.  Additional results are presented on an earlier
reported FAC-series of regeneration experiments in which sulfated dolomite
with 10.2 wt % S was regenerated by the incomplete combustion of methane
in a 7.6-cm-dia fluidized bed.  Total CaO regeneration values calculated
from chemical analyses of the regenerated samples were within analytical
agreement of fehe values calculated from flue-gas analyses.  Most of the
sulfur and calcium balances ranged from 90 to 110%, which is an acceptable
variation.  When the regeneration temperature was increased from 1010 to
1095°C, the extent of CaO regeneration increased from 21 to 89% and the
S02 concentration in the wet effluent gas increased from 0.7 to 5.3%.

     Dolomite that had been regenerated at a higher temperature (1095°C
as compared with 1040°C) contained a larger amount of large pores (>0.4 urn).
These large pores have been shown by others to be beneficial to sulfation.
However, when the reactivity of these samples as S02 acceptors was
evaluated in TGA experiments, the sample that had been regenerated at the
lower temperature (1040°C) was found to be more reactive.  Although higher
regeneration rates are obtained at high regeneration temperatures, the
effect of regeneration temperature on the reactivity of the additive in
subsequent cycles also has to be considered.

     Electron microprobe analyses of sulfated and regenerated Tymochtee
dolomite particles showed no sulfur concentration irregularities that
might lead to poor utilization "in subsequent sulfation cycles.

     At steady state, sulfide (S ) levels were £0.1% in the product
obtained in experiments performed at low reducing gas concentrations (3%
reducing gas in the effluent), and sulfide levels of 0.3 and 0.7% were
obtained in experiments performed at high reducing gas concentrations

-------
 (15% reducing gas in the effluent).  At these reducing conditions, it was
predicted that only CaS should exist at equilibrium.  The very low CaS
concentrations found were attributed to the formation in the regeneration
reactor of two zones (oxidizing zone at the bottom of the fluidized bed
and reducing zone at the top).

     The extent of decrepitation ranged from 5 to 15% of the total feed
 (calcium basis).  These particles, which were elutriated, were basically
smaller than the smallest sulfated particles fed to the regenerator.  For
all experiments considered, a decrease in the number of larger particles
 (>800 ym) and an increase in the number of medium-size particles  (>300
and <800 ym) were found in the bed.

     Agglomeration of sulfated additive occurred during some regeneration
experiments.  X-ray diffraction analyses showed that compounds of calcium-
magnesium silicates had formed during the agglomeration (melting) process.
In DTA (differential thermal analysis) experiments, sulfated dolomite
melted at 1200°C in atmospheres of nitrogen and air.  The X-ray diffraction
analysis of the DTA samples revealed the possible formation of a calcium
silicate compound, but no calcium-magnesium silicate trace was found.

     Regeneration of Sulfated Greer Limestone by the Incomplete Combustion
of Coal.   In a preliminary experiment, sulfated Greer limestone from Pope,
Evans and Robbins was regenerated.  The results were comparable with those
obtained when using sulfated Tymochtee dolomite.

     The next experiments made with Greer limestone explored the effect
of relative lengths of the oxidizing and reducing zones in the bed.  At
relatively high reducing gas concentrations (^4% in the effluent gas),
increasing the length of the oxidizing zone relative to the length of
the reducing zone resulted in a drastically reduced CaS concentration in
the regenerated product and an increase in the extent of regeneration.
All subsequent experiments were made with the larger oxidizing zone.

     In the main experimental studies, the Greer limestone was regenerated
at two temperatures, 1050 and 1100°C.  In experiments at ^1050°C,  decreasing
the solids residence time (which is inversely proportional to the sulfated
limestone feed rate) from 31 min to 12 min decreased the extent of CaO
regeneration from 89% to 70%.  At 1050°C, increasing the fluidizing-gas
velocity from 1.2 to 1.37 m/sec had no significant effect on the extent
of regeneration.  In the earlier regeneration study (FAC-series), it had
been found that increasing the fluidizing-gas velocity adversely affected
the extent of regeneration of sulfated additive.

     In one of the experiments performed at 1100°C, the solids residence
time in the regenerator was low, 7.5 min.  The S02 concentration in the
dry off-gas of 7.6% obtained at this solids residence time was the highest
in this series.   Although the extent of regeneration of CaO decreased with
decreasing solids residence time, it remained quite high (61%), even at a
relatively low solids residence time of 7.5 min.

-------
     Regeneration of Sulfated Tymochtee Dolomite by the Incomplete Combustion
of Coal.  Sulfated Tymochtee dolomite was regenerated at bed temperatures
of 1000, 1050, and 1100°C.  At 1000°C, the S02 concentration in the effluent
gas was not affected by reducing the solids residence time, but the extent
of regeneration decreased.  At 1050°C, when the solids residence time was
decreased from 34 min to 12 min, the S02 concentration in the dry effluent
increased from 3.0% to 4.8% and the extent of regeneration dropped from
90% to 56%.

     At 1100°C, the solids residence time was decreased from 13 to 9.4 min
and the SC-2 concentration in the dry effluent gas increased from 6.4% to
7.8%.  The extent of regeneration decreased from 85% to 77%.  Since the
extent of regeneration remained high at 1100°C at the above solids residence
times, it is expected that higher SC>2 off-gas concentrations.can be achieved
by further decreasing the solids residence time in the reactor (i.e., by
increasing the sulfated reactant throughput rate).

     Carbon Utilization and Carbon Balance Calculations.  Carbon balances
calculated for three of the earlier regeneration experiments were 97% and 99%,
and an unaccountably low 55%.  Carbon utilizations were ^80%.

     Effect of Coal Ash on the Agglomeration of Sulfated Additive during
Regeneration.  During sulfation of the sorbent (combustion step), some coal
ash is retained  in the fluidized bed and is removed with the sulfated sor-
bent.  Tymochtee dolomite has been sulfated during the combustion of Arkwright
coal at ANL and Greer limestone during combustion of Sewickley coal at Pope,
Evans and Robbins.  The ash fusion temperatures (initial deformation) under
reducing conditions for both of these coals are very close to the regeneration
temperature.  The presence of these coal ashes could contribute to initial
coalescing of particles in the fluidized bed and subsequent reactions with
the sulfated acceptor.  The coal ash level was found to be ^5% in once-sul-
fated Tymochtee dolomite and was M.0% in sulfated Greer limestone.

     Reactions in two different materials have been investigated in the
Differential Thermal Analyzer (DTA) to date:  (1) unsulfated Tymochtee
dolomite and (2) sulfated Greer limestone which contained ^10% coal ash.
The decomposition peak temperatures of MgC03 and CaC03 in the Tymochtee
dolomite samples were found to be 819 and 933°C, respectively, in good
agreement with literature values.  In the study of sulfated Greer
limestone, dehydration of Ca(OH)2 was observed at ^530°C, and decomposition
of the residual CaC03 occurred at 790-840°C.  The formation of 2Ca2Si1300°C), the samples fused and complex compounds of
calcium-aluminum silicates formed.  Since the fluidized-bed temperature
was ^1100°C, the formation of calcium silicates may be the first stage
in the agglomeration process, followed by loss of fluidity in the fluidized
bed and greater localized temperature increases.

     Mass and Energy Constrained Model for the One-Step Regeneration
Process.  A mass and energy constrained model for the one-step regeneration
process is being developed which can be used to perform sensitivity analyses

-------
for key variables such as regeneration temperature and pressure, fluidizing-
gas velocity, reactor size, solids residence time, and feed gas composition
and temperature.  The early objectives of this work are to predict the
effects of experimental variations and to use these results to guide further
experimental efforts in the investigation of the feasibility of the one-
step regeneration process.  The effects of some of these variables on
dependent variables such as off-gas composition (SC>2, C02, H^O), fuel and
oxygen requirements, and energy cost of regeneration have been evaluated.

Cyclic Combustion/Regeneration Experiments

     Preliminary Experiment.  A preliminary experiment was performed to
evaluate the effects of sorbent recycling on sorbent reactivity and decrepi-
tation.  Tymochtee dolomite which had been sulfated and regenerated under
various operating conditions was used as the sorbent.  Arkwright coal was
combusted at a bed temperature of 840°C, 810 kPa pressure, and ^17% excess
combustion air.  The Ca/S mole ratio, which was based on the unreacted
calcium in the regenerated dolomite, was 1.6.  Based on the flue-gas
analysis for SC>2 (250 ppm), sulfur retention was ^90%, indicating that
the sorbent retained its activity for sulfur retention into the second
combustion cycle.  Data showed that about a third of the dolomite fed to
the combustor was elutriated and recovered in the cyclones and filters.
Screen analyses of the feed, overflow, and cyclone materials indicated that
entrainment was not entirely the result of decrepitation of the dolomite
in the combustor.

     Combustion Step, Cycles 1, 2, and 3.  A ten-cycle series of combustion/
regeneration experiments was initiated using Arkwright coal and Tymochtee
dolomite.  Nominal operating conditions selected for the combustion portions
of each cycle were a 900°C bed temperature, 810 kPa system pressure, 1.5
Ca/S mole ratio, ^17% excess combustion air, 0.9 m/sec fluidizing-gas
velocity, and a 1.07-m bed height.

     The first 2 1/2 cycles (three combustion experiments) have been
completed.  The sulfur dioxide level in the flue gas increased from 290
ppm (86% retention) in the first combustion cycle to 400 and 490 ppm (81
and 77% retention) in cycles two and three.  Sulfur retention in cycle
three was sufficient to meet the EPA environmental emission standard for
sulfur dioxide.

     As the sorbent passed through the combustor in the first cycle, ^20%
of the +30 mesh additive feed was reduced in size to -30 mesh.  Based on
an approximate 5-hr solids residence time in the combustor, the decrepitation
rate can be expressed as ^4 wt %/hr.  For the same experiment, ^25% of
the additive entering the combustor was entrained with the flue gas and
the remaining ^75% was removed in the product overflow.  This corresponds
to an entrainment rate of ^5 wt %/hr.

     Quantitative evaluations of decrepitation and elutriation of sorbent
during the second and third cycles have not been completed, but estimated
losses are 2-3 wt %/hr for cycle two and 1-2 wt %/hr for cycle three,
indicating a significant reduction in the sorbent loss rate in succeeding
cycles.

-------
     Regeneration Step, Cycles 1 and 2.  The sulfated dolomite from the
combustion step was regenerated at a bed temperature of 1100°C, a pressure
of 153 kPa, and a solids residence time of ^7 min, with relatively good
results.  In the first cycle, the SC>2 concentration in the dry effluent
gas was 6.5% and the extent of CaO regeneration was 71%.  Solids losses
from the bed due to elutriation and/or attrition were found to be lower
than expected, 2%.  In the second cycle, CaO regeneration was 67%.  The
S02 concentration in the dry off-gas was 8.6%.  The solids losses due to
elutriation and/or attrition were negligible.  The low sorbent losses due
to attrition can be attributed to the short solids residence time in the
regenerator.

     Coal Ash Buildup during Sulfation and Regeneration Steps.  In the
sulfation (coal combustion) step, Arkwright coal is combusted under oxidizing
conditions.  In the regeneration step, Triangle coal is partially combusted
under reducing conditions.  A coal-ash level of 4-5% in the bed was
estimated for the combustion-sulfation step of the first cycle.  In the
regeneration step, no additional ash buildup was observed.

Pressurized, Fluidized-Bed Combustion;  Bench-Scale Studies

     The objectives of the experiments were to evaluate the following:  (1)
the effects of coal and additive particle size on sulfur retention, nitrogen
oxide flue-gas levels, and combustion efficiency; (2) the sulfur retention
capability of lignite ash, which has a high calcium content; and (3) the
relative effects of fluidizing-gas velocity and residence time on the
decrepitation of sorbent.

     Combustion experiments were performed, using a highly caking, high-
volatile bituminous, Pittsburgh seam coal from the Consolidation Coal
Company Arkwright mine and a lignite coal from the Consolidation Coal
Company Glenharold mine.  Tymochtee dolomite obtained from C. E. Duff and
Sons, Huntsville, Ohio, was air-dried and screened before being fed to the
combustor.

     The equipment, designed for operation at pressures up to 10 atm,
consists of a 6-in.-dia fluidized-bed combustor, a compressor for supplying
fluidizing-combustion air, a preheater for the fluidizing-combustion air,
coal and additive feeders, and an off-gas system (cyclones, filters, gas-
sampling equipment, and pressure let-down valve).  The system is thoroughly
instrumented and is equipped with an automatic data logging system.

     Replicate of VAR-Series Experiments.  A replicate of a previously
performed VAR-series experiment was made after the completion of equipment
modifications (separation of the combustion and regeneration systems).
In this experiment, the operation of the combustor and analytical instru-
mentation was checked and the validity of comparing current experiments
and past experiments was verified.  Except for a slight discrepancy in
the level of NO in the flue gas, the results of the experiment agreed very
well with the results of the previously performed experiments.

-------
      Effect  of  Coal  and  Sorbent  Particle  Size on  Sulfur Retention,
 Nitrogen  Oxide  Level in  the  Flue Gas, and Combustion Efficiency.  Four
 combustion experiments  (PSI-series) were  completed  to measure the effects
 of  coal and  sorbent  particle size on sulfur retention, NOX emission, and
 combustion efficiency.   Arkwright coal with mass-mean particle size levels
 of  ^150 ym and  ^640  ym and Tymochtee dolomite with  mass-mean particle
 size  levels  of  ^370  ym and ^740  ym were used in the experiments  (22 factorial
 design).  The experiments were performed  at 843°C (1550°F) , 8 atm, and
      excess  combustion air.
     The  level of S02 in  the effluent gas ranged from a low 160 ppm  (^93%
 sulfur retention) in experiment PSI-3 (fine coal, fine sorbent) to 240 ppm
 0^89% sulfur retention) in experiment PSI-4 (fine coal, coarse sorbent).
 The results indicate a slight increase in sulfur retention with decreasing
 dolomite  particle size.   No significant effect of coal particle size on
 sulfur retention was observed.

     There was no noticeable effect of coal or dolomite particle size on
 NO levels in the flue gas (120 to 150 ppm).  N02 levels were ^40 to ^60
 ppm, which are considerably higher than the anticipated N02 levels of 5
 to 10 ppm.

     Combustion efficiencies ranging from 89 to 93% were determined, but
 no consistent effect of particle size of coal or sorbent on combustion
 efficiency was indicated.  Utilization of sorbent material ranged from
 58 to 71% and increased with decreasing dolomite particle size.

     Combustion of Lignite in a Fluidized Bed of Alumina.  Glenharold
 lignite was combusted in  a fluidized bed of alumina in two replicate experi-
 ments to test the sulfur-retention capability of the high calcium lignite
 coal.  Sulfur retentions  were 86 and 89%, which compared favorably with
 a sulfur retention of 85% when lignite was combusted in a bed of dolomite
 under similar operating conditions.  Hence for this lignite, sulfur
 emissions can be adequately controlled without the addition of limestone.

     Effect of Fluid iz ing-Gas Velocity on Decrepitation Rate.   An experi~
 ment was performed to duplicate the conditions of the first-cycle combustion
 experiment discussed above, except that the velocity was reduced to 0.6
 m/s from 0.9 m/sec.   At the lower gas velocity, the coal and thus the
 additive feed rate is less, resulting in a higher solids mean residence
 time.  The data show that the rate of decrepitation was essentially
 unaffected (it increased  slightly) by the decrease in fluidiz ing-gas
velocity.  Although decrepitation increased slightly due to the increased
 residence time of the additive in the combustor, the decreased velocity
was apparently effective  in retaining a greater percentage of the additive
 feed in the product overflow stream.   Thus, the entrainment was actually
 less than the decrepitation rate.

 Synthetic SO? Sorbents

     Calcium oxide impregnated in a-A!203 is being investigated as an
 alternative to dolomite and limestone as an an additive for lowering the

-------
                                    8


sulfur dioxide level in the off-gas from fluidized-bed coal combustors.
In this program, the capability of the supported calcium oxide to react
with sulfur dioxide and of the resulting calcium sulfate to be regenerated
is being studied experimentally, using thermogravimetric analysis.

     Alpha-alumina pellets containing 6.6% calcium oxide by weight were
sulfated at 900°C, using concentrations of sulfur dioxide in the gas stream
ranging from 0.05 to 3%.  Calcium utilization ranged from 67 to 98%, and
the reaction went to completion in 4 to 10 hr.

     The oxygen concentration in the gas phase had only a small effect on
the sulfation rate when oxygen was present in stoichiometric excess.  However,
when sulfur dioxide was in excess, the rate was first order in oxygen
concentration.

     Water concentration in the synthetic combustion gas had no effect on
the sulfation rate.

     The sulfation rate of the pellets increased with sulfation temperature
up to 900°C, where it became independent of temperature.  Above 900°C, the
reaction is diffusion controlled.

     Synthetic sorbents containing 2-16.5% CaO in a-A!203 were studied for
their ability to capture S02.  The sulfation rate decreased with increasing
CaO concentration when measured as a fraction of the maximum possible
sulfation; however, the amount of S02 captured for a given residence time
was usually higher for higher CaO concentrations.

     Tymochtee dolomite and calcium oxide in a-A!203 pellets were sulfated
under similar experimental conditions to allow comparison.  The calcium
oxide in a-A!203 pellets was 95% sulfated in 6 hr at 900°C, using 0.3%
sulfur dioxide and excess oxygen; the dolomite was only 60% sulfated in
19 hr.  However, the dolomite contained four times as much calcium as the
sorbent did.

     In addition to CaO and Li20, the metal oxides, Na20, K20,  SrO,  and
BaO,  have been impregnated into alumina and tested as S02 sorbents.   Lithium
sulfate was found to be unstable at sulfation conditions (900°C).  Sodium
oxide and potassium oxide sorbents have a higher rate of reactivity with
S02 than does CaO sorbent; however, their sulfates would have appreciable
volatility at regeneration conditions (1100°C).  Strontium oxide and barium
oxide sorbents have essentially the same reactivity as does calcium oxide.

     Sulfated sorbents were regenerated using various reducing gases (CO,
H2, and CH^) .  In each case, the reaction is 0.8 order in reducing gas
concentration.  The rate was the same for hydrogen and methane and the rate
for carbon monoxide was one-third the former rate, requiring only 4 min
when using  2_&% reducing gas.  The addition of carbon dioxide (15% or more)
to the reducing gas lowered the regeneration rate since carbon dioxide
reacted with hydrogen to form carbon monoxide.

-------
     The reaction product for the CaO in a-Al203 sorbent was dependent
upon the regeneration temperature; above 1050°C, the product was CaO;
below 900°C, it was CaS.  Between 900 and 1050°C, the product was a mixture
of CaO and CaS, with CaO concentration increasing with temperature.

     Other data showed that dolomite was regenerated at a slightly lower
rate than was the supported additive; moreover, the product of the reaction
when dolomite was regenerated in the TGA apparatus contained 50% CaS and
not 100% CaO as with the supported additive.

     A cyclic sulfation-regeneration experiment (five cycles) was performed
on 6.6% CaO in a-Al203 additive that had been heat-treated at 1100°C.
The rates of sulfation and regeneration were the same in each cycle.

     Porosity measurements have been made on granular supports that had
been heat treated at 1100, 1200, and 1500°C.  The higher heat-treatment
temperatures produced supports containing larger pores.  The supports that
contained larger pores produced sorbents having a higher reactivity with
S02 and a greater calcium utilization.

Sorbent Attrition Studies

     Preliminary attrition studies on dolomite have shown that sulfated
dolomite is more attrition-resistant than is fresh half-calcined dolomite.
The attrition rate of regenerated dolomite is high; however, due to the
short residence times in regenerators, the material loss per cycle is
low.  Supports for synthetic sorbents have approximately the same attrition
rate as does sulfated dolomite.
                                 *
Combustion-Regeneration Chemistry

     The fundamental aspects of the chemical reactions associated with
the cyclic use of BCR 1337 dolomite, chosen as a model sorbent system, to
control sulfur emissions are being investigated.  Kinetic studies are based
on the use of a thermogravimentric analysis (TGA) technique, and structural
studies are based on X-ray diffraction and microscopy techniques.

     The half calcination reaction

          [CaC03'MgC03] ->- [CaC03-MgO] + C02

has been studied at 1 atm over the temperature range 640-800°C in two
environments:  (a) 100% C02 and (b) 40% C02-60% He.  The reaction rate
increases rapidly with temperature for both 100% CO  and 40% C02 environ-
ments, particularly above a temperature near 700°C.  The effect of C02
concentration is subtle; in general, the rate is higher in the 40% C02
environment.   X-ray diffraction studies show that size and degree of
preferred orientation of the CaC03 (calcite) pseudo crystals formed by
 Sponsored in part by ERDA, Division of Physical Research.

-------
                                   10
 this reaction are dependent on the kinetics (temperature) of the reaction,
 i. e. j larger sizes and greater degrees of preferred orientation result
 when the reaction proceeds slowly (at lower temperatures).

     Optical microscopic examination of the half-calcined dolomite sample
 shows that dolomite crystals are transformed to calcite along grain
 boundaries, as well as within dolomite crystals and that the half-calcined
 grain structure formed as a result of the half-calcination reaction is
 completely destroyed by heat treatment.

     Experiments have been performed to establish whether the solid-solid
 reaction between CaSOi^ and CaS would be practical as a regeneration scheme.
 The yield of the reaction is shown to depend on the percentages of CaS
 and CaSOij in the dolomite starting material for the reaction.  In addition,
 the kinetics of the reaction are shown to be such that at 950°C, the
 reaction reaches 80% completion in less than 6 hr.

 Coal Combustion Reactions

     Determination of Inorganic Constituents in the Effluent Gas from
 Coal Combustion.  Some chemical elements carried by the combustion gas
 from coal combustion are known to cause severe metal corrosion, for example,
 to turbine blades.  A study is under way to determine quantitatively which
 elements are present in the hot combustion gas of coal, in either volatile
 or particulate form, and to differentiate between volatile and particulate
 species.  Of interest are (1) identification of the compound forms and
 amounts of particulate species and (2) determination of the amount and
 form of condensable species.

     The detailed design of the laboratory-scale batch fixed-bed combustor
 for this study has been completed, the combustor has been fabricated in
 the central shop of ANL, and the combustor system is now being assembled.
A schematic flow diagram of the combustor system is shown and described.

     Prior to fabrication of the combustor, detailed engineering drawings,
 specifications for fabrication, and stress calculations to support the
 safety of the design were reviewed and approved by a design/preliminary
 safety review committee.  Modifications to the design are described in
 this report.  The combustor is now considered to have a safe design.

     Systematic Study of the Volatility of Trace Elements in Coal.  The
 effluent gases from coal combustion are known to contain trace elements,
 some of which cause severe metal corrosion in coal-fired boilers and gas
turbines.  Knowledge of the temperature at which trace elements in coal
 start to volatilize and of the rate of volatilization is important for
 the utilization of coal.  The objective of this study was to obtain data
 on the volatility of these trace elements under practical coal combustion
 conditions.  The experimental phase of this study has been completed.

     In all experiments, ash samples prepared from Illinois Herrin No. 6
Montgomery County coal by ashing at 340°C were heat-treated in a tubular

-------
                                    11
 furnace  to various temperatures  in  an air flow of 0.6 scfh for 20 hr or
 more, and the ash residue was analyzed  for trace elements of interest.

     From the experimental results, it  is concluded that under dry oxidizing
 conditions, most of the chlorine in the coal evolves at temperatures below
 640°C, although trace amounts of chlorine still remain in the high-temperature
 ashes.   However, the metallic elements, Na, K, Fe, Al, Mg, Ca, Ti, Zn, Mn,
 Ni, Co,  Cu, Cr, and Li are generally retained in the fused ash up to 1250°C
 under both dry and wet oxidizing environments, and the retention of these
 metallic elements in the fused ash  is not affected by the oxygen concen-
 tration  in dry flowing gas.  Water  vapor in the flowing gas did not affect
 the evolution of these metallic  elements; however, it caused a greater weight
 loss of  the ash, indicating that water vapor in the flowing gas affects
 the evolution of certain substances from the 340°C ash during heat treatment.

 Equipment Changes

     As  originally installed, the pressurized, fluidized-bed combustor and
 the regenerator utilized several common components and therefore could not
 be operated simultaneously.  Additional equipment and instruments were
 installed to permit concurrent investigations.  Equipment is also being
 installed so that solids can be  transported between the combustion and
 regenerator reactors.

 Miscellaneous Studies

     Preparation and Testing of  Supported Additives (Dow Subcontract).  A
 thermogravimetric analysis apparatus to be used in evaluating various
 candidate S02 sorbents has been  constructed as part of a research program
 subcontracted to Dow Chemical Company.  Three synthetic sorbents (CaO,
 BaO, and SrO) have been impregnated into low-surface-area and intermediate-
 surf ace-area Al2C>3 support material.  The work plan includes screening
 six kinds of metal oxide-support combinations in both the sulfation and
 regeneration modes.

     Limestone and Dolomite for  the Fluidized-Bed Combustion of Coal:
 Procurement and Disposal.  The potential demand, supply, and cost as well
 as the disposal aspects associated with the use of limestone or dolomite
 as sulfur-accepting additives in the fluidized-bed combustion (FBC) of
 coal are assessed.  Also, the market for the regeneration by-products
 (sulfur  and sulfuric acid) is discussed.

     The Properties of a Dolomite Bed of a Range of Particle Sizes and
 Shapes at Minimum Fluidization.   Experiments have been performed to determine
minimum  fluidization velocity as a function of temperature and pressure
 for a dolomite bed having a wide size range of nonspherical particles.
An improved correlation has been developed in the form of Ergun's correlation.
Experiments revealed that partial segregation of the bed occurred if the
ratio of the diameter of the largest particle to the diameter of the
smallest particle exceeded at least 16.   As a result, the prediction of
minimum  fluidization velocity for the entire bed becomes complicated, making

-------
                                   12
precise prediction uncertain.  A procedure is suggested for the prediction
of minimum fluidization velocity of the entire bed.

     Mathematical Modeling;  Noncatalytic Gas-Solid Reaction with Changing
Particle Size,  Unsteady State Heat Transfer.  A mathematical expression
for the unsteady state heat balance is derived for a nonisothermal, non-
catalytic, first order gas-solid reaction.  The formulation is based on
a shrinking core model and takes into account the changing size of the
spherical particle during reaction.  Thus, the model has two moving
boundaries viz.3 the reaction front and external particle diameter which
grows or shrinks due to reaction.  This formulation constitutes the first
basic step in the modeling of combustion efficiency of coal in a fluidized
bed.

-------
                                    13
                              INTRODUCTION

      In a national program,  the  feasibility of an  environmentally
acceptable  fluidized-bed  combustion process for use  in power and steam
plant applications is being  studied.  The concept  involves burning  fuels
such as coal  in a fluidized  bed  of particulate lime  solids which react
with the sulfur oxides released  during combustion.   The sulfated lime
solids will be either discarded  or regenerated, the  choice depending on
environmental and economic considerations.  One concept is pressurized
combustion, in which the  hot flue gas would be expanded through a gas
turbine.

     The current program  for demonstration of regeneration process
feasibility consists of studying:  (1) the reductive decomposition  of
CaSOtf at VL100°C, using coal as  a source of reducing agent, (2) the
conditions  at which sintering of particles in the  bed can occur, (3)
the modeling  of the process,  and (4) other possible  methods of regener-
ation.

     To demonstrate that  sulfur  retention by the lime solids during
combustion  and sulfur release during regeneration  are possible without
excessive breakup of the  solids  during cycling, batch cyclic studies
(alternate  combustion and regeneration of the lime solids) have been
started.  Combustion studies  have been performed to  determine (1) the
effects of  stone and coal particle sizes on sulfur retention, NO level
in flue gas,  and combustion  efficiency, (2) whether  the concentration of
calcium in  lignite is sufficient that enough sulfur  is retained in  the
bed and the EPA specification for S02 in the flue  gas is met, (3) the
effect of fluidzing-gas velocity on decrepitation  rate, and (4) the fates
of the biologically hazardous trace elements and compounds which can
cause hot corrosion of turbine metal blades.  Miscellaneous peripheral
studies include (1) a survey  on  procurement and disposal of sulfated
limestone,  (2) properties at  minimum fluidization  of a bed having a range
of particle sizes and shapes, and (3) a mathematical modeling of a
particle undergoing a noncatalytic reaction and changing size; unsteady
state heat  transfer effects  are  taken into consideration.

         REGENERATION OF  S02-ACCEPTING ADDITIVE, BENCH-SCALE STUDIES
[J. Montagna  (Principal Investigator),. R. Mowry, C.  Sehoffstoll, G. Smith,
 G. Teats]

     The technical feasibility of a process is being investigated for
the one-step  reductive regeneration of additive in a fluidized bed  in
which the required heat and reductants are provided by partial combustion
of a fuel in  the regeneration reactor.  The reduction of CaSO^ to CaO is
favored by high temperatures  (>1040°C/1900°F)  and mildly reducing
conditions.

                 + CO -+ CaO + S02 + C02                               (1)

                 + H2 -* CaO + S02 + H20                               (2)

-------
                                    14


 Other  reductants  such  as  CH4  can  play the  same  role  as  CO  and  H2.  At
 lower  temperatures  (<1000°C)  and  more highly  reducing conditions  the
 formation  of  CaS  is  favored:

            CaSOi4  + 4CO ->  CaS  + 4C02                                    (3)

            CaS04  + 4H2 ->  CaS  + 4H20                                    (4)

 Since  these are competing reactions,  the reaction conditions must be
 chosen to  minimize the buildup of CaS, maximize the  conversion of CaSOi+
 to CaO, and maximize the  S02  concentration  in the off-gas  from the
 reactor for economical reprocessing of the  off-gas from the regeneration
 process.

     Sulfated dolomite (Tymochtee) and limestone (Greer) have  been regen-
 erated.  Methane  and coal (in separate studies)  have been  partially
 combusted  during  the regeneration experiments.   The  effects of temperature,
 fluidizing-gas velocity,  and  solids residence time on the  regeneration
 of CaO from CaSO^ have been evaluated at different experimental conditions.

 Materials

     Tymochtee dolomite obtained  from C. E. Duff and Sons, Huntsville,
 Ohio,  which had been sulfated during  coal combustion experiments  (using
 Arkwright  coal) at ^900°C and 810 kPa, was  regenerated.  Also  regenerated
 was Greer  limestone which had been sulfated by  Pope, Evans and Robbins
 at 840-900°C and atmospheric  pressure during  the combustion of Sewickley
 coal.

     Initially high-volatile  Arkwright bituminous coal  and later Triangle coal
 (ash fusion temperatures  of VL100°C and 1380°C,  respectively).were combusted
 during these regenerations.   Chemical and physical characteristics of these
 coals  are  presented in Appendix A, Table A-l.

 Equipment

     The bench-scale experimental system used for regeneration experiments
 (in which methane was  combusted) has been previously described.1  .Recently,
 the experimental regeneration system was separated from the combustion
 system and  the regenerator was rebuilt and modified.  Figure 1 is a
 schematic diagram of the new regeneration system.  The  inside diameter
of the reactor vessel has been enlarged from  7.62 cm (3.0  in.)  to 10.8
 cm (4.25 in.), and the interior metal overflow  pipe formerly used to
control the fluidized-bed height has been replaced with an overflow pipe
that is external to the fluidized bed.  The coal and the sulfated sorbent
are fed separately (for independent control)  to a common pneumatic transport
line which enters the bottom of the reactor.  In addition  to the regeneration
reactor, the experimental system consists of an L-valve for metering
sulfated additive, a peripherally sealed rotary feeder  for metering coal,
and the flue-gas solids-cleanup system.  Another component is an electrically
heated pipe heat exchanger for preheating some of the fluidizing gas and
preheating air (used in startup only)  to ^400°C.

-------
      TO GAS
      ANALYZERS
      COAL
ROTARY VALVE
   AIR
                        SAMPLE
                   GAS CONDITIONER
SULFATED-ADDITIVE
     HOPPER
                 PULSED
                  AIR
                                      Fl LTER
                                                                                   •EXHAUST
                                                               FILTER
                                                   PRESSURE
                                                   CONTROL
                                                    VALVE
                                               REGENERATOR
                                               (10.8-cmID)
                                  LINE PREHEATER
                                                   SURGE
                                                    TANK
                                        PRODUCT
                                       COLLECTOR
                                  -C
AIR

N2
           Fig. 1.  Schematic Diagram of  Present Regeneration System.

-------
                                   16


     The 10.8-cm-dia pressurized, fluidlzed-bed reactor is lined with a
^4.8-cm-thick Plibrico castable refractory encased in an 8-in. (20.3-cm)
dia Schedule 40 pipe (Type 316 SS), approximately 2.29 m (7 1/2 ft) long,
with its entire length contained within a 12-in.-dia Schedule 20 carbon
steel pipe.  Differential thermal expansion of the inner and outer pipes
is accommodated by the use of packing glands on lines entering the bottom
flange of the unit.

     The solids transport air constitutes ^40% of the total fluidizing
gas added to the bed in the reactor.   The remaining fluidizing gas is a
mixture of pure % and G£.   The gases are added separately to permit
better temperature control in the fluidized bed (if there should be a
mild temperature excursion, combustion could be stopped, without de-
fluidizing the bed, by shutting off the oxygen).  Also, the concentration
of oxygen in the entering gas can be higher than that in air, allowing
more coal to be combusted without significantly changing the fluidization
gas velocity.  A large amount of heat is required to compensate for (1)
the heat losses in the comparatively small experimental system and (2)
the heat load imposed by feeding cold sulfated additive to the system.

Regeneration of Sulfated Tymochtee Dolomite by the Incomplete Combustion
of Methane

     The FAC-series of regeneration experiments was performed in the 7.6-
cm-ID regenerator using the -in situ  combustion of methane to evaluate
the effects of temperature, residence time, height of fluidized bed,
and total reducing gas conditions on the regeneration of CaO in sulfated
Tymochtee dolomite.  Earlier results, which were based only on flue-gas
analyses, have been reported and discussed.1  To aid in the discussion
of additional results obtained by chemical analysis of particulate
product samples, the experimental conditions and the results based on
flue-gas analyses from the FAC-series are re-presented in Table 1.

     Sampling was initiated when over 90% of the fluidized bed had been
replaced while the regenerator was at the design experimental conditions.
Steady-state samples, each composed of the overflow product taken over a
30-min period, were analyzed for total sulfur, sulfide, and calcium (Table
2).  The accuracy of the analytical techniques ranged from 5 to 10%.

     Regeneration of CaO.  Regeneration of CaO was calculated from chemical
analyses of the steady-state products.  It was based on the sulfur to
calcium ratios in (1) the sulfated-dolomite feed and (2) the steady-state
product after regeneration.  These calculated regeneration ratios are
compared in Table 2 with ratios that are based on off-gas analyses.  The
latter values are the ratios of the total sulfur released into the off-gas
stream to the total sulfur contained in the sulfated dolomite feed.  Since
the off-gas flow rate was not measured, it had been assumed to be equal
to the feed gas flow rate.   This has caused a discrepancy in the two calcul-
ated regeneration results.   As discussed in the "Mass and Energy Constrained
Model for the Regeneration Process" section of this report, the gas flow rate
increases during regeneration due to the coal combustion and the decomposition
reactions that occur.  The predicted gas volumetric increase  was included in

-------
           Table 1.   Design Experimental  Conditions and Extent of Regeneration Results
                     for the FAC-Series.

                     Reactor Diameter,  in.:   3
                     Pressure,  psig:   12
                     Additive:   Sulfated  Tymochtee dolomite (10.2% S)
Design Conditions



Exp.
FAC-1
FAC-1R1
FAC-1R2
FAC-2
FAC-3
FAC-4
FAC-5
FAC-9A
FAC-7
FAC-8
FAC-9


Temp
(°C)
1040
1040
1040
1095
1040
1095
1040
1040
1040
1010
1040

Additive
Feed Rate
(Ib/hr)
6
6
6
6
10
10
10
6
10
6
6

Residence
Time
(min)
30
30
30
30
18
18
30
30
30
30
30

Bed
Height
(ft)
1.5
1.5
1.5
1.5
1.5
1.5
2.5
1.5
2.5
1.5
1.5
Total
Reducing Gas
in Effluent3
(vol %)
3
3
3
15
15
3
15
3
3
3
3


Fluidizing- Measured S02,
Gas Velocity in Effluent
(ft/sec) (%
2.2
2.4
2.5
Experiment
2.6
2.5
2.7
3.0
2.5
2.3
2.5
wet)/(% dry)
3.0 / 4.0
2.6 / 3.6
2.3 / 2.9
could not be
4.1 / 5.7
5.3 / 7.3
3.5 / 4.4
1.3 / 1.5
2.7 / 3.3
0.65/ 0.78
1.3 / 1.7

CaO
Regeneration
(%)
70
65
60
completed
70
83
61
40
43
15
58
 The actual reducing gas  concentrations  were within 15% of the design levels.
^Determined by infrared 862  analysis  of  the dry flue gas.
"Based  on flue gas analysis.

-------
          Table 2.  Chemical Analysis of Regenerated Product, Calculated Material
                    Balances, and Regeneration Results from the FAC-Series.
Chemical Analysis of Regenerated
Samples Taken at Steady State
Temp Total Sulfur
Exp.
FAC-1
FAC-1R1
FAC-1R2
FAC-2
FAC-3
FAC-4
FAC-5
FAC-9A
FAC-7
FAC-8
FAC-9
( c)
1040
1040
1040
1095
1040
1095
1040
1040
1040
1010
1040
(%)
3.30
3.73
5.08

2.89
1.53
2.79
7.34
7.41
8.38
2.74
Sulfide
(%)
0.1
<0.1
<0.1
5.2
0.3
<0.1
0.7
0.1
<0.1
<0.1
<0.1
Calcium
(%)
29.8
28.7
26.2

30.2
30.0
29.0
26.8
25.3
23.1
29.1
Elutriated3
Material Balances Calcium x 100
CaO
Sulfur Calcium Feed Calcium Regeneration
(%) (%)
94 109
99 109
106 113
This experiment
91 102
88 83
Production
— -
109 110
71 72
95 91
(%) 0
5
15
13
could not be completed
12
9
rate was not measured
-
15
14
11
0 /(%)
70/76
65/72
60/58

70/79
83/89
61/79
40/40
43/36
15/21
58/55
                                                                                                         00
 Only particles larger than 15 ym were monitored.
 Based on flue gas analysis.
GBased on chemical analysis of particulate dolomite samples.

-------
                                    19
the recalculation of regeneration based on off-gas analysis.  The sulfur
regeneration values obtained by chemical analysis of the regenerated products
are generally within analytical agreement of the values calculated from off-
gas analyses.

     Material Balances.  Material balances for sulfur and calcium were
not performed for the entire experiments, as was the case for balances
in other experiments that are reported later.  Instead, the balances were
each made over a 30-min steady-state period.

     Because the experimental system did not contain a final filtration
system, particles smaller than -vL5 ym escaped from the system unmonitored.
The total mass rate of these fines was measured only in FAC-5, in which
the flue gas exiting from the last cyclone was passed through a sintered-
metal filter.  A particle pad ^0.3 cm thick was built up on the filter,
ensuring good particle collection efficiency for the fine particles.  A
mass collection rate of 4.6 g/min was obtained in this sintered-metal
filter, in comparison with a 8.3 g/min captured in the other particle
collection devices.  The particle size distribution for the FAC-5 sample
of fines was obtained with a Coulter counter particle analyzer and is
given in Fig. 2.  None of the calculated mass balances contained the
contribution from these fines and thus they are biased low.
                   99.9
                    O.I
                                      10
                                    DIAMETER,
           Fig.  2.   Particle Size Distribution of the Fines
                    Leaving the Cyclone in FAC-5.

-------
                                     20.


     The calcium balances  ranges  from 72%  in FAC-8  to 113% in FAC-1R2.
These balances were based  on  the  mass rates  and  chemical analyses of the
feed, product, and elutriated solids.   The calcium  analyses are reliable
within +5%.  Errors greater than  ^5%  are due mainly to unreliability of
the measurements of the mass  rates.

     The sulfur balances were based on the flue-gas flow rates, the con-
centrations of sulfur  (as  S02 only) in the flue  gas,  and the flow rates
and chemical analyses  of the  solid streams.   The sulfur balances ranged
from 71% in FAC-8 to 109%  in  FAC-7.   Acceptable  material balances should
range from 90 to 110%; most sulfur balances  were acceptable.

     The lowest sulfur and calcium balances  obtained  (FAC-8)  were probably
caused by incomplete removal  of the 30-min steady-state sample from the
receiver.  Thus the measured  mass rate of  regenerated solid was probably
low.  (The pre-steady-state removal rate was ^50% larger than the calculated
steady-state rate.)  Generally, sulfur and calcium  balances agreed within
10%.

     Effect of Fluidizing-Gas Velocity.  The effect of fluidizing-gas
velocity on extent of  regeneration was evaluated and  reported earlier.1
Regeneration data were calculated using S02  concentrations  in the off-gas.
Now these data with additional CaO regeneration  values based  on chemical
analyses of the regenerated product are plotted  in  Fig.  3.  As the gas
velocity was increased from 0.67  m/sec to  0.91 m/sec,  CaO regeneration
decreased from 76 to 40% and  the  SC>2  concentration  in the wet effluent
gas decreased from 3%  to 1.3%.  In a  study by Martin  et a£. 2  on the
reductive decomposition of gypsum in  a 25.4-cm-dia  (lO-in.-dia)  fluidized
bed, the optimum fluidizing-gas velocity (for decomposition)  was found
to be about 0.6 m/sec.  However,  the  effect  of fluidizing-gas velocity
was reported to be small.
                                      a  BASED ON SOLID
                                          ANALYSIS
                                      o  BASED ON FLUE
                                          GAS ANALYSIS"
                     0.60       075       0.90
                         FLUIDIZING-GAS VELOCITY, m/sec
          Fig. 3.  The Effect of Fluidizing-Gas Velocity on Sulfur
                   Regeneration for Sulfated Tymochtee Dolomite.

-------
                                     21
     In this study,  fluidizing-gas velocity had a meaningful effect in the
velocity range  investigated.   The relatively small  internal  diameter (7.62 cm)
of the fluidized-bed reactor  probably enhanced the  detrimental effect of
gas velocity, lowering fluidization quality.

     Effect of  Temperature on Regeneration and Resulfation.   The effect
of regeneration temperature (ranging from 1010 to 1095°C)  on the regener-
ation of dolomite was reported earlier.1  The extent  of regeneration was
based on the S02 content of the off-gas.  Now, regeneration  based on
solid product analyses are given in Pig. 4 with the earlier  reported data.
(Regeneration temperature was the only variable in  all  experiments,
except that in  the  1095°C experiment the solids residence  time was 18 min
instead of 30 min.)   As the temperature was increased from 1010°C to 1095°C
and the solids  residence time remained constant or  decreased, CaO regen-
eration (based  on analysis of the regenerated product)  increased from 21%
to 89%.  With increasing regeneration temperature,  the  SC>2 concentration
in the wet off-gas  increased  from 0.7% to 5.3%  (from  0.8%  to 7.3% in the
dry off-gas).'

     The temperature dependence of the regeneration of  sulfated Tymochtee
dolomite in these experiments was compared  (see Fig.  4) with similar
results obtained by Martin et al. 2 for the decomposition of  gypsum in a
25.4-cm-dia fluidized bed with a solids residence time  of  ^90 min.  In
the latter investigation, carbon was the reductant.  The extent of regen-
eration was based on chemical anlaysis of the gypsum  before  and after
decomposition.
                      100
 o CaO REGENERATION
   (Based on flue gas analysis)
 A CaO REGENERATION
   (Based on product analysis!
 a S02 CONCENTRATION
-"•--DECOMPOSITION OF GYPSUM,
   CaO REGENERATION (MARTIN et al.)
   (Based on product analysis)
                                                        s?
                                                        to"
                                                        
-------
                                     22


     To evaluate the effect of regeneration temperature on the quality of
the regenerated dolomite as a S02~acceptor, the pore size distribution of
samples (-25 +35 mesh particles) of calcined, sulfated and regenerated
dolomite were obtained (Fig. 5).  Porosity was measured as the extent of
penetration of mercury as a function of pressure.  The cumulative pore
volume (mercury penetration) per given mass is plotted (Fig. 5) as a
function of pore diameter.  It was reported by Hartman and Coughlin3 that
most sulfation takes place in the larger pores (>_0.4 ym).  Pores smaller
than MD.4 ym are relatively small and easy to plug and do not contribute
much to the extent of sulfation.  The pores shrink during sulfation of
CaO as a result of changes in molecular volume.

     Curve A represents the pore size distribution of virgin dolomite
that had been completely calcined at 900°C for 2 hr in a furnace (preheated
to 900°C) with a 20% C02-80% air environment.  The pore volume for pores
0.4 ym and larger was found to be 0.1 cm3/0.5 g.   This pore distribution
was compared with that of sulfated and regenerated dolomite because
dolomite is fully calcined during regeneration.

     Curve B represents the pore structure of dolomite that had been
sulfated (10.2 wt % S) in coal combustion experiments at 900°C and was
used as the feed for the FAC regeneration experiments.  The pore volume
for pores 0.4 ym and larger was found to be 0.045 cm3/Q.5 g.   In contrast
to the pore volume distribution of the calcined material, the pore structure
of the sulfated material was very much plugged by sulfation.

     Curves C and D illustrate the pore structures of the sulfated dolomite
that had been regenerated at 1040°C and 1095°C.  At the higher regeneration
temperature, the volume of the larger pores (>0.4 ym) increased.  On the
basis of this result, the sulfation reactivity of dolomite regenerated
at the higher temperature was expected to be greater.

     Sulfation experiments were performed in a thermogravimetric analyzer
with the regenerated samples and the calcined dolomite sample.  The data
show (see Fig. 6) that the higher regeneration temperature did not improve
the reactivity of the dolomite with S02-  The dolomite that had been
regenerated at 1095°C had a lower reactivity than did the dolomite regen-
erated at 1040°C and also the dolomite precalcined at 900°C.

     Although high regeneration reaction rates are obtained at higher
temperatures, the potential detrimental effect of high temperatures on the
reactivity of the dolomite in subsequent sulfation cycles requires further
study.  A decrease in the reactivity of regenerated additive during cyclic
exposures to high temperature was reported by Skopp e~k at.   and the reason
was presumed to be loss in porosity.  In the FAC experiments, the porosity
increased after one and a half cycles.

     Electron Microprobe Analysis of Sulfated and Regenerated Particles.
Regenerated Tymochtee dolomite particles from FAC-1R2 and FAC-4, and
unregenerated, sulfated Tymochtee dolomite particles were analyzed with
an electron microprobe to determine the effect of regeneration on the

-------
          A= CALCINED TYMOCHTEE
            DOLOMITE  900°C
          B = SULFATED  TYMOCHTEE
            DOLOMITE  900°C
          C=REGENERATED TYMOCHTEE
            DOLOMITE  I040°C
          D=REGENERATED TYMOCHTEE
            DOLOMITE  I095°C
Q_
     0
10           1.0          O.I
   D,  PORE DIAMETER,
                                                                              N>
                                                                              OJ
   Fig. 5.  Pore Distributions of Dolomite Samples from Different Process Stages.

-------
                                   24
sulfur concentration profiles.   The radial distributions  of  calcium,
magnesium,  and  sulfur were measured.  To prepare the sample  for analysis
a random sample of particles was screened, and approximately twenty
-20 +25 mesh (707-841 ym) particles were mounted in epoxy and machined to
remove the  equivalent of one-half the diameter of a typical  mounted
particle (^390  ym).  A thin gold layer  (^50 angstroms)  was sputtered on
the machined surface to enhance the conductivity of the mounts.  Apatite
(39.9 wt %  Ca),  FeS2 (53.4 wt % S), and MgO (60.3 wt %  Mg) were used
as standards to calibrate the probe and to obtain a rough quantitative
estimate of the local component concentrations.  One-half of the particles
were analyzed.
             n  VIRGIN DOLOMITE PRECALCINED  AT
                900°C FOR 2 hr IN 20% C02
             o  REGENERATED AT 1040°C
                REGENERATED AT 1095 °C
                                     2            3
                             SULFATION  TIME, hr
         Fig. 6.  Sulfation Reaction Data Obtained with a
                 Thermogravimetric Analyzer at 900°C, 0.3%
                 S02, and 5%  02.

-------
                                     25
      Microprobe analyses of sulfated  Tymochtee  dolomite particles  were
 compared with chemical  anlayses  that  indicated  average sulfur  and  calcium
 concentrations of 10.2  wt % and  22.2  wt  %,  respectively.   The  component
 radial concentration distributions  for calcium  and  sulfur  for  three
 typically sulfated particles are given in Fig.  7  (magnesium concentration
 profiles are not presented because  they  are similar to those for calcium).
 The extent of sulfation in these particles  varied.   These  particles had
 been in the continuously fed,  fluidized-bed coal  combustor for different
 periods of time (solids are backmixed).   The extent of sulfation was least
 for the particle represented at  the top  of  the  figure  and  greatest for
 the bottom particle.  As sulfation  progressed,  the  edge of the sulfated
 shell moved towards the center of the particle.   The local sulfur  and
 calcium concentrations  obtained  were  lower  than the true local concen-
 trations because irregularities  on  the surfaces cause  scattering of the
 characteristic emitted  by X-ray  (the  sample surfaces could not be  machined
 as  smoothly as the calibration sample surfaces).

      Regenerated particles from  experiment  FAC-1R2  (performed  at 1040°C)
 were chemically analyzed and found  to contain 5.1 wt % S and 26.2  wt %
 Ca.   The electron probe analyses of two  sample  particles from  this experi-
 ment are given in Fig.  8.   Regeneration  appears to  take place  in two
 stages.   In the particle whose analysis  is  given  at the top of Fig.  8,
 regeneration apparently is in  an early stage, with  the primary regener-
 ation-reaction zone moving radially into  the uniformly sulfated particle.
 Some residual sulfur  (sulfur that is  less available, probably  due  to
 diffusion limitation) is left  behind  the  advancing  primary reaction zone.
 The  bulk of the sulfur  is  removed in  the  first  stage of regeneration.
 The  particle represented by the  other electron  probe analysis  is in a
 later  (slower)  stage  of regeneration,  and the residual concentration of
 sulfur  in the entire  particle  is relatively low (note  difference in  scales
 for  the  two diagrams).   In an  industrial process  in which  the  solids
 residence time would  be low (high solids  throughput rate),  probably only
 the  first (rapid)  stage of regeneration would occur.   Hence, most  but  not
 all  of  the  sulfur  would be removed  in an industrial process.

     Electron microprobe analyses were performed  on regenerated particles
 from experiment  FAC-4,  which was conducted  at the higher temperature of
 1095°C.   The  calcium  and sulfur  concentrations  in regenerated  Tymochtee
 dolomite  from this  experiment were  30 wt %  and  1.5  wt  %, respectively.
 The  electron  probe  analyses  for  calcium and  sulfur  in  three particles
 from this experiment  are shown in Fig. 9.   The  particle whose  analysis is
 given at  the  top of Fig.  9  appears  to be near the end  of the first  regen-
 eration  stage;  the  primary  desulfurization  reaction zone has almost  reached
 the  center  of  the  particle.  Regeneration of  the other  two  particles is
more nearly complete  (note  difference in scale  of sulfur concentrations).
A greater extent of regeneration was  observed in particles  from FAC-4
 than in particles  from  FAC-1R2 which were regenerated  at a  lower temperature.
 Generally,  the regenerated  particles  showed no  sulfur  concentration
 irregularities that might  lead to poor utilization  in  subsequent sulfation
 cycles.

-------
                                    26.
                30
                10
           "   0
           I"   12
           c •>"  8
                         PARTICLE, 1060
           
-------
                                     27
                 CO
                                PARTICLE, 1170/im
                                 PARTICLE, 1500 /an
         Fig. 8.  Electron Microprobe Analyses of Regenerated
                  Tymochtee Dolomite Particles from Experiment
                  FAC-1R2.
with 'v.S vol % total reducing gas in the effluent).  With vL5% reducing
gas in the effluent, the formation of CaS was enhanced—for example,  in
two experiments, sulfide concentrations in the products were 0.3 and  0.7%,
respectively.

     The experimental sulfide concentrations obtained are much lower
than those predicted for the equilibrium equation  (below) at the experi-
mental off-gas conditions.
1/4
                   + CO •* 1/4 CaS + C02,   - =
PCO

PC02
(3a)
At 1040°C, the maximum partial pressure ratio of CO to C02(?CO/PC02)  for
which CaS and CaSO^ coexist is ^0.017.  At higher partial pressure ratios,
only CaS should exist at equilibrium.5  Since the experimental values  of
partial pressure ratios of CO to C02 in the effluent for the FAC-experi-
ments were all above 0.02, higher sulfide concentrations would be expected.

-------
                                    28
                 s   30
                 -8s 20
                    -

                 o
CO
 0
10
 5
 0
                              JV/v
                               PARTICLE, 1200
^ ou
§5? 20
«*'8
o 0
of 1.0
=>s«
i£0.5
<" n
A^/-^*^-*^^. /^-^xx-v A\


-
                              •PARTICLE,  MOOfim-
                  s   30
                  Ess 20

                  ^'°0
                 00
                             PARTICLE,  900/im
         Fig. 9.   Electron Microprobe Analyses of Regenerated
                  Tymochtee Dolomite Particles from Experiment
                  FAC-4.
The experimental results may be explained on the basis that (1)  the
reaction in the f luidized-bed reactor is not at equilibrium and  (2) the
conditions in the fluidized bed are not reducing throughout.   Rather,
solid reactants and products are ciculated between reducing and  oxidizing
zones , allowing CaS to be oxidized to
     Elutriation and Decrepitation of Additive.   The extent of decrepitation
of additive is of major concern in the development of one-step reductive
regeneration.  The physical integrity of additive particles may be affected
by thermal stresses at the high process temperature (^1100°C)  and by
molecular rearrangements within the particles as a results of  regeneration
reactions.

-------
                                     29
     Losses of regenerated product due to decrepitation  (including decrepi-
tation during pneumatic transport) at the various experimental conditions
were measured.  Since the sulfated feed dolomite was nominally -14 +45 mesh
and the elutriated particle samples in the flue gas were all nominally
-45 mesh  (<350 ym), it was assumed that the elutriated particles were
decrepitation fragments of the feed Tymochtee dolomite.  Table 2 shows the
fraction  of additive lost during regeneration due to decrepitation and
subsequent elutriation expressed as a percentage of feed calcium.  The
loss ranged from 5 to 15%.

     In a more direct approach to the evaluation of decrepitation, the
fractional distributions of the sulfated feed and regenerated product for
experiment FAC-1 were plotted (Fig. 10, upper plot).  The fractional
distribution.curves suggest a particle population shift to smaller diameters.
This was  confirmed by plotting the mass ratios (i.e., the calcium mass
content of the product to the calcium mass content of the feed additive)
as a function of particle diameter (Fig. 10, lower plot).   Generally, all
plots of  product-to-feed calcium ratio as a function of fractional particle
diameter  showed that during regeneration, the number of larger particles
(>800 ym) decreased and the number of medium-size particles (>300 and <800 ym)
increased due to the decrepitation of the larger particles.

     Composition and Agglomeration of Sulfated Additive.  Partial agglom-
eration of the sulfated dolomite bed occurred in experiment FAC-2 which
was attempted using a bed temperature of 1095°C and a 15% reducing gas
concentration in the off-gas.  It is believed that upsets in the fluidization
of the additive bed caused local temperatures to rise higher than the
measured  operating temperature which, in turn, caused partial melting of
the additive followed by agglomeration.  (It has been previously reported
by Wheelock and Boylan5 that the surface of gypsum becomes glassy during
reductive decomposition at 1250°C.)

     Fresh, sulfated, regenerated (FAC-4), and agglomerated (FAC-2)
Tymochtee dolomite were analyzed by X-ray diffraction (Table 3).  In unsul-
fated dolomite, traces of a-quartz were found as expected, since the dolomite
contains ^2% Si02.  Sulfated dolomite contained the expected components,
including traces of a-quartz and weak traces of other phases, possibly
merwinite, Ca3Mg(SiOit)2.  The regenerated dolomite from FAC-4 which was not
agglomerated contained mostly CaO and MgO with traces of a-quartz, CaSO^
and weak traces of other phases (possibly merwinite).  During a regeneration
experiment (FAC-2),  partial agglomeration of the sulfated additive occurred.
The loosely bound particles around the agglomerate were found to contain
approximately the same major constitutents and the same minor constituents
of CaSO^ and a-quartz as were found in the regenerated sample from FAC-4.
In addition, traces of CaMg(Si03)2 were found.  In the cores of the
agglomerates,  the additive particles were fused together and contained
Ca3Mg(SiOt).)2 as a medium constituent and CaO as a medium rather than a
major constituent.

     Samples of Pope, Evans and Robbins (PER) sulfated Greer limestone were
also analyzed by X-ray diffraction.  An agglomerate taken from a regeneration

-------
                                    30
100
8s ,,-
% FRACTIONAL WEIGHT,
r>o 01 -
O CJl O CJ
	 1 	 1 	 r
o = SULFATED S02
°=REGENERATED





	 W"


^
i
ACCEPTOR
S02 ACCEPTOR


/

(
I \





1




)
s


\
\
\

= 200"
<
cc.
in 150"
CO
<£
^E i /\/\
PRODUCT/FEED
cji C
-° P . ^



•^
00


/


/


^x




•^



\



s








N



\
\
1000
DIAMETER.pn
         Fig. 10.  FAC-1, Fractional Feed and Product Particle Size
                   Distributions.  Decrepitation is characterized
                   by the fractional product to feed mass ratios
                   at different particle diameters.
experiment (Coal Test 3)7 was analyzed at two locations in the agglomerate.
The loosely bound limestone particles contained Ca3Mg(SiOit)2 as a minor
constitutent.  The fused limestone particles contained CaaMg (810^)2 as a
major constituent.  Again, as the additive softened, the formation of the
calcium-magnesium silicate compound progressed.

     On the basis of the X-ray diffraction analysis described, it appears
that as the particles soften, calcium-magnesium silicate compounds form.
The melting point of this class of materials is %1250°C which indicates
that the localized temperature at which agglomeration occurred was much
higher than the design experimental temperature
     Preliminary differential thermal analyses (DTA) were performed on
unsulfated and sulfated Tymcohtee dolomite in several gas environments.

-------
     Table 3.  Qualitative Chemical Compositions of Regenerated and Unregenerated
               Samples of Additive.
               Analysis Method:  X-ray diffraction
Material
o
Regenerated
Pertinent Constituents
Source
fused particles
Major
MgO
Medium
CaO and possible
Minor
CaSOi+
Tymochtee dolomite
from FAC-2
           Q
Regenerated
Tymochtee dolomite
from FAC-2
           f\
Regenerated
Tymochtee dolomite
from FAC-4
Sulfated Tymochtee
dolomite
Virgin
Tymochtee dolomite

Regenerated
PERC limestone
from Coal Test
No. 3

Regenerated
PER limestone
from Coal Test
No. 3
from regenerator
bed

Loosely bound
particles from
regenerator bed

Unbound regen-
erated additive
from regenerator
bed

Combustor product
              Ca3Mg(S10^)
Fused particles
from regenerator
bed

Loosely bound
particles from
regenerator bed
MgO, CaO
CaO, MgO
     , CaC03
CaMg(C03)2


MgO and
probable
Ca3Mg(Si04)2


CaS04, MgO
Possible
a-quartz, and
CaMg(Si03)2
a-quartz, possible
CaS04, CaMg(C03)2,,
Ca(OH)2, and other
phases

CaO, possible
a-quartz, and other
phases

a-quartz and possible
Possible CaS04,
MgAl2Oit; very minor:
possible CaO


Possible Ca3Mg(SiOtt)2
  Regeneration temperature was
  Regeneration temperature was 1040°C.
  Pope, Evans and Robbins.

-------
                                    32


The X-ray diffraction analyses of the reacted samples are given in Table 4.
Only one reaction was observed in unsulfated dolomite between ambient temper-
ature and 1300°C in either air or nitrogen.  The calcination reaction occurred
at 550-850°C.  The sulfated dolomite (which contained a small amount of
carbonate) calcined at 600-750°C, and melted at 1200°C in both air and
nitrogen.  This is lower than the melting point of any of the major
consitutents in sulfated dolomite.  In addition to the weight loss attributed
to calcination of these sulfated samples, a weight loss was noted as their
temperature was raised above 1100°C.  This might have been caused by decom-
position of CaSO^.  During cooling of the sample from 1300°C, an exothermic
reaction (solidification) occurred at 1150°C.  X-ray diffraction analysis
showed that CaSO^ and MgO were the major constituents and CaO the minor
constituent.  CaSOij was present as larger crystals (indicative of post-
melting) instead of the small crystals normally present in sulfated dolomite.
Also, in both of these samples, calcium silicate was present as a possible
minor constituent, instead of the calcium-magnesium silicate compounds
found in agglomerated dolomite samples.

     The last DTA experiment in this series was performed with sulfated
dolomite in a reducing atmosphere (100% H2).  A calcination weight loss
occurred at 600-700°C, followed by a further weight loss which continued
as the temperature rose to ^1000°C.  The latter weight loss was probably
due to reduction of CaSO^. to CaS which, with MgO, was found to be the
major constituents in the final product.  No melting was observed up
to 1300°C.

     The samples, which were small, ^30 mg , may not have been representative
of nonhomogeneous Tymochtee dolomite.  Larger samples are being used in
experiments currently being performed in another DTA apparatus.

Regeneration of Sulfated Greer Limestone by the Incomplete Combustion
of Coal

     Exploratory Experiments.  Sulfated Greer limestone from Pope, Evans
and Robbins  (Test 620) has been regenerated, with the reducing gas and
the required heat generated by incomplete combustion in the bed by either
Arkwright or Triangle coal.

     For experiment LCS-1 the experimental conditions and results are
given in Table 5.  A S02 concentration of 2.5% in the wet effluent gas and
a CaO regeneration of 88% were obtained.  The extents of regeneration
based on chemical analysis of the regenerated products from these experi-
ments were  almost identical.  The steady-state concentrations of COS and
CS2 in the  flue gas, obtained by mass spectrometric analysis, were 0.02%
and £0.01%,  respectively.

     Experiment LCS-2 was performed at a higher bed temperature, 1100°C,
than LCS-1.  The operating temperature and the concentrations of gases
in the effluent gas are  given in Fig. 11, and the results are given  in
Table 5.  The feed rate  in this experiment was 5.4 kg/hr, in comparison
to 4.9 kg/hr in LCS-1.   The S02 concentration in the wet off-gas was 3,1%

-------
    Table 4.  Qualitative Chemical Composition (Obtained by X-Ray
              Diffraction Analysis) of Reacted Samples from DTA
              Experiments.a
Reaction
Atmosphere
during DTA
Material Experiment
Tymochtee N2
dolomite
Tymochtee Air
dolomite
Sulfated N2
Tymochtee
dolomite
Sulfated Air
Tymochtee
dolomite
Sulfated H2
Tymochtee
dolomite
Pertinent Constituents
Major Medium
MgO CaO
CaO, MgO
MgO,' CaSOiLj
(very crystalline)
MgO, CaSOi^
(very crystalline)
CaS, MgO
Minor
Ca(OH)2, possible
CaC03 and other
weak phases
very minor :
possible CaAl2Ott
CaO, possible
calcium silicate
CaO, possible
calcium silicate

Final DTA furnace temperature was 1300°C in each experiment.

-------
                 Table  5.
Experimental Conditions and Results for Regeneration Experiments  in
which Arkwright and Triangle Coals were Incompletely Combusted  in a
Fluidized Bed.
                            Nominal  Particle Residence  Time:   15-30 min
                            Nominal  Fluidized-Bed Height:   46 cm
                            Reactor  ID:   10.8  cm
                            Coal:  Arkwright coal  (2.82 wt  %  S)
                                  reduction conditions,  1105°C  (LCS-1, -2)
                                  Triangle coal  (0.98  wt % S),   Ash fusion  temperature under
                                  reduction conditions,  1390°C  (LCS-3)
                            Additive:  Sulfated Greer limestone
                                       (LCS-1, -2,  -3).
                                      Ash fusion temperature under
                                            -2)
                                            iion temperature under
                                            (initial deformation)
                                    (9.48 wt % S) -10 +50 mesh
Exp. Pressure
No. (kPa)
LCS-1
LCS-2
LCS-3°
Based
Based
184
153
153
Bed
Temperature
1040
1100
1060
on flue gas analysis,
on chemical analysis
Feed
02 Cone.
18.9
18.6
23.6
Fluidiz ing-
Gas
Velocity
(m/sec)
0.91
1.1
1.0
Additive
Feed
Rate
(kg/hr)
4.9
5.4
8.6
Reducing Gas
Concentration
in Effluent (%)
1.3
1.3
1.6
Measured S02/H2S
in Wet
Effluent Gas
2.5/96
3.1/235
3.5/76
CaO
Regeneration
88/88
94/94
72/84
•
of dolomite and limestone samples.
                                                                                                                 u>
                                                                                                                 -P-
^Triangle coal was used;  in the other two experiments,  Arkwright coal was used.

-------
                                 35
    ^1200
    E
    QJ
    O)
    CD
      1000
*Srv*


.AW-C7

      2000
    E
    o.
    o.

    o"
.2  c^
 O)
 o
 cz
 o
o  gS


 %  £
CD  o
GO

 I
 OJ

|  5«



-------
                                    36
in LCS-2, and the CaO regeneration based on off-gas analysis was 94%.  In
comparison with LCS-1, the extent of regeneration and the S02 content of
the effluent gas were improved by increasing the temperature in LCS-2.
The concentrations of COS and CS2 in the flue gas for LCS-2 were 0.02%
and 0.01%, respectively.

     An additional experiment was performed with sulfated Greer limestone,
using Triangle coal instead of. Arkwright coal (a bituminous coal with
a higher ash fusion temperature than Arkwright coal).  The experimental
conditions for this regeneration experiment (LCS-3) were an additive feed
rate of 8.6 kg/hr (which is higher than in earlier experiments at a bed
temperature of 1060°C), a fluidizing-gas velocity of 1.0 m/sec, and a
reducing gas concentration of 1.6% in the effluent gas.  The operating
temperature and the concentrations of gases in the effluent gas are given
in Fig. 12; the experimental results are given in Table 5.  From the
average SC>2 concentration of 3.5% in the wet effluent gas, a sulfur (or
CaO) regeneration of 72% was calculated (based on off-gas analysis).  A
greater sulfated additive feed rate (i.e., reduced solids residence time)
increased the S02 concentration in the ^effluent gas without greatly
decreasing the extent of regeneration.

     Effect of the Relative Lengths of the Oxidizing and Reducing Zones
during Regeneration of Greer Limestone.  When fuel (coal) is combusted
under reducing conditions in the fluidized bed, both an oxidizing zone
and a reducing zone are present.  An oxidizing zone is established at
the bottom of the fluidized bed (mainly below the fuel injection point)
where combustion air enters; a reducing zone is established in the upper,
fuel-rich portion of the fluidized bed.  Calcium sulfide is formed in
the reducing zone by the following typical reaction,

          CaS04 + 4 CO -»- CaS + 4 C02                                  (3)

and its formation is favored by higher reducing gas concentrations.  The
presence of large amounts of CaS is detrimental to the regeneration of
CaSOif to CaO since CaO does not form directly from CaS at the conditions
used.  However, since the additive is continually circulated in the bed,
the CaS can be reoxidized in the oxidizing zone of the bed by the
reaction:

          CaS + 2 02 -> CaS04                                          (5)

Thus an effective oxidizing zone is important for preventing the buildup
of CaS.

     To test the effect of the relative lengths of the oxidizing and
reducing zones on the steady-state level of CaS in the bed, two experiments
have been performed using sulfated Greer limestone feed and sufficient
coal to give a reducing gas concentration of ^4% in the effluent gas
(approximately twice as high a concentration as in other regeneration
experiments with coal).  The results for experiments LCS-4D and LCS-7, in
which the ratios of the nominal heights of the reducing zone to the oxidizing
zone were 5 and VL.3,  respectively, a're given in Table 6.

-------
  H-
  OQ
O H
0) (D
  i-!
  (0
t-1 O
O (B
U) O
  o
  (D
  3
  rt
  H

  rt
  H-
  O
              	Above Bed  Gas Concentrations	

               S02,  %     CH4, %      CO,  %      02, %      C02, %   NO,  ppm Bed  Temp, °C
             0        50        10        20         10        30 0       100  1000     1200
            o I •  -  ' o I   to '  '  '—i—|   I—*--—i—•—|  1—•—'—•—)i—(  I  '  *  * x-
                                      t-
                                                                                                            oo

-------
              Table 6.
Experimental Conditions and Results for Regeneration Experiments Designed
to Test the Effectiveness of the Two-Zone Reactor in Minimizing CaS Buildup.
                                                               15-30
                                              Temperature, °C:  1060
                                              Pressure, kPa:  153
Nominal Particle Residence Time, min:
Nominal Fluidized-Bed Height, cm:  46
Reactor ID, cm:  10.8
Coal:  Triangle coal (0.98 wt % S).  Ash fusion temp, under red.
       conditions:  1390°C (initial deformation)
Additive:  Sulfated Greer limestone (9.48 wt % S) -10 +30 mesh
Sulfur/Sulfide Measured S02/

Exp.
No.
LCS-4D
LCS-7

Red. Zone Length
Oxid. Zone Length
5
^1.3
Fluidizing
Gas Velocity
(m/sec)
0.97
0.95
Feed
Rate
(kg/hr)
5.3
5.9
Reducing Gas in Regenerated H2S in Wet
Concentration
in Effluent (%)
3.8
4.1
Additive
(%)/(%)
7.5/4.9
2.3/1.1
Effluent Gas
(%)/(ppm)
0.9/625
1.9/1900
CaO
Regeneration
(%)a/(%)b
20/41
67/72
                                                                                                                 U)
                                                                                                                 oo
a
 Based on flue-gas analysis.
 Based on chemical analysis of limestone samples before and after regeneration.

-------
                                     39
     In the first experiment, LCS-4D, regeneration  (conversion of CaSO^ to
CaO) was very poor, 41%; sulfide  (S ) concentration was high, 4.9%.  The
intermittent off-gas analysis for this experiment is given in Fig. 13.
The total fluidized-bed height was 46 cm  (illustrated in Fig. 14).

     In experiment LCS-7 (and in all subsequent regeneration experiments)
the coal injection line was 13 cm longer  than that  formerly used (Fig. 14),
decreasing the ratio of the heights of the reducing and oxidizing zones.
The intermittent flue-gas analyses for LCS-7 are given in Fig. 15.  Solids
regeneration was 71%, and the regenerated additive contained 1.1% sulfide.
By increasing the length of the oxidizing zone relative to the length of
the reducing zone, the buildup of CaS was reduced and greater regeneration
was obtained.

     Since 40% of the total fluidizing gas is used to inject the coal and
additive into the fluidized bed, a fluidizing-gas velocity discontinuity
would have existed in LCS-7 between the portion of the fluidized bed below
and above the injection line.  To compensate for this discontinuity, a
20-cm-long ceramic insert with an ID of 8.9 cm (3.5 in.) was installed
above the gas distributor, as illustrated in Fig. 14.  This insert was
retained in all future experiments.

     Effect of Solids Residence Time and  Temperature on Regeneration.
Additional regeneration experiments were performed with sulfated limestone
from Pope, Evans and Robbins (PER) to study the effects of solids residence
time and bed temperature on regeneration. .In five of these experiments
(LCS-8 through LCS-12), Greer limestone which had been sulfated 0\J.l wt
% S) by PER in Test 621 was regenerated.  In the remaining two experiments
(LCS-16, -17), sulfated limestone (^9.3 wt % S) from PER which had an
uncertain experimental origin, but was probably a mixture of Greer and
Germany Valley limestones, was regenerated.

     The first three experiments, LCS-8,  -9, and -10, were performed at
the same fluidizing-gas velocity  (^1.2 m/sec) and at nearly the same bed
temperature (^1050°C).  The results and experimental conditions for these
and the other regeneration experiments are given in Table 7.  The sulfated
Greer limestone feed rate, was increased  from 5.9 kg/hr to 15.4 kg/hr in
this series of runs, and as a result,  the extent of regeneration (based on
chemical analysis of solids) decreased from 89% to 70%.  This decrease
was expected since increasing the solids  feed rate decreased the particle
residence time in the regenerator from 31 min to 12 min.  Nevertheless, the
extent of regeneration for LCS-10 (70%) was higher than that obtained for
a similar experiment using sulfated Tymochtee dolomite (9.4 wt % S) in
which 58% of the sulfur was regenerated (LCS-10, discussed below).  The
reason for the higher extent of regeneration in these experiments was
probably the lower sulfur content (7.1 wt % S) of the Greer limestone.

     Experiment LCS-11 was performed at the same temperature and sulfated
additive feed rate as was LCS-10.  It was found that increasing the
fluidizing-gas velocity from 1.2 to 1.4 m/sec caused no significant change
in the extent of regeneration.  In contrast, in the FAC-series of regeneration

-------
  U>
O W
Ml (D
l-h CU

O H
PJ (D
•d rt
ft) C
i-i i-i
H- n>
3
(t> fu
3 0
rt CL

f Q
O S3
CO CO

-P- O
  o
  (D
  0
  s
  CO
    S02,  %
  0        10 0
o4e-- — • — i — i
                             Above Bed Gas Concentrations ----------

                             %      CO, %       02,  %      C02, %    NO,  ppm  Bed  Temp,  °C
                               20         40         20        30 0        200  1000     1200
        8
        §
        i-

-------
                                    41
FLUIDIZED
BED 	
COAL AND
ADDITIVE
INJECTION
LINE 	
GAS
DISTRIBL


-*•
IA A A
ITOR-^
i- DISTANCE ABOVE GAS DISTRIBUTOR, cm
o —
co — ro<-M.c>encr>->joou3o
1 O OOOOOOOOOOO


FLUIDIZEO
BED 	
COAL AND
ADDITIVE
INJECTION
LINE 	


al »»i V
) 10 20 30
(YGEN CONC.,%
4D
                                                    "100
                                                      90
                                                      80
                                                      70
                                                      60
                                                      50
                                                      40
                                                      30
                                                      20

                                                       0
                                                    5  "0  10 20 30
                                                      OXYGEN CONC.,%
                                                     LCS-7
          Fig. 14.  Geometry of Oxidizing and Reducing Zones  in
                    Relation to Position of Coal Injection Line.
experiments (discussed above), the extent of regeneration  decreased  at
higher fluidizing-gas velocities.  However, the FAC-experiments were per-
formed in a smaller reactor  (7»6-cm ID), at lower fluidizing-gas
velocities (0.6-0.9 m/sec),  and with CH^ instead of coal.  Additional
experiments will be performed to test the effects of  fluidizing-gas
velocity on regeneration.

     The other experiments listed in Table 7 were performed  at 1100°C.
LCS-16 was performed with  the same solids feed rate as was used in LCS-12.
The S02 content of the dry flue gas was 4.8% in LCS-16,  in comparison with
3.4% in LCS-12.  The higher  S02 content in the flue gas  in LCS-16 was due
mainly to the lower fluidizing gas velocity (less dilution)  used in  LCS-16.

     Experiment LCS-17 was performed with a much higher  sulfated-additive
feed rate, 25.4 kg/hr.  This corresponds to a solids  residence time  in  the
regenerator of 7.5 min.  The SC-2 concentration in the dry  flue gas was
7.6%, and the extent of CaO  regeneration was 61%.

-------
  CW
O W
l-h fD
O H
(u n>
X pi
"rj ri*
n> e

H- (D

(D P

rt CL

  O
O
  o

  §
  3
  s
  CO
             	Above Bed Gas  Concentrations	

              S02, %    CH4,  %     CO, %      02, %      C02, %   NO,  ppm  Bed Temp,  °C
             0        10 020         40         20       300       290  1000      1200
                          \


1

:
^




                                                                                                         S3

-------
   Table 7.
Experimental Conditions and Results for Regeneration Experiments with
Combustion of Arkwright and Triangle Coal in a Fluidized Bed.
                                                      Pressure, kPa:  153
Nominal fluidized-bed height:   46 cm
Reactor ID:  10.8 cm
Coal:  Triangle coal (0.98 wt % S)
       Ash fusion temp under reducing conditions, 1390°C (initial
       deformation)
Additive:  Sulfated Greer limestone
           a.   -10 +30 mesh, 7.12 wt % S (LCS-8 through -12)
           b.   -14 +30 mesh, 9.26 wt % S (LCS-16 through -17)
Bed
Exp. Temperature
No. (°C)
LCS-8
LCS-9
LCS-10
LCS-11
LCS-12
LCS-16
LCS-17
1043
1065
1050
1050
1100
1100
1100
Fluidizing- Feed
Gas Velocity Rate
(m/sec) (kg/hr)
1.22
1.21
1.20
1.37
1.48
1.14
1.20
5.9
10.4
15.4
15.0
14.3
15.4
24.5
Particle
Residence
Time (min)
31
18
12
12
13
12
7.5
Reducing Gas Measured S02/H2S
Concentration in Dry Effluent
in Effluent (%) Gas (%)/(ppm)
2.1
1.9
2.2
^2.0
^2.0
2.4
2.5
2.3/300
3.4/0
3.2/400
3.1/400
3.4/300
4.8/1100
7.6/400
CaO
Regeneration
(%)a/(%)b
97/89
81/80
52/70
59/72
70/75
57/76
59/61
Based on flue-gas analysis.
Based on chemical analyses of limestone samples before and after regeneration.

-------
                                   44
     The experimental results obtained with sulfated PER limestone at
1100°C,  including experiment LCS-2  (Table 5), have been plotted  in Fig. 16
as a function of solids residence time.  The SC>2 concentration obtained in
LCS-12 was  adjusted to compensate for dilution due to the higher fluidizing-
gas velocity used in that experiment.  The extent of regeneration of CaO
from sulfated Tymochtee dolomite at  1100°C (described in the next section
of this  report) is also included on  Fig. 16.  The S02 concentration of the
dry off-gas from the sulfated Tymochtee dolomite experiments was higher
than that for the sulfated limestone experiments.  At low solids residence
time, the 862 concentrations in the  off-gas were almost equal because the
sulfur content of the sulfated stones no longer affected the rate of SC>2
evolved.  The extent of regeneration was vLO% lower for limestone than for
dolomite.
    100
           I—I—I
        TYMOCHTEE
        DOLOMITE
                                      EXTENT  OF LIMESTONE
                                          REGENERATION
                                       O  SOLID ANALYSIS
                                       D  GAS ANALYSIS
               LIMESTONE
               A S02 CONCENTRATION
               A ADJUSTED SO, CONCENTRATION  FOR  LC5-I2
                      10     15     20     25     30
                      SOLIDS  RESIDENCE TIME, min
        Fig.  16.  Regeneration of CaO and S02 Concentration in the
                 Dry Flue Gas for Sulfated Limestone and Tymochtee
                 Dolomite as a Function of Solids Residence Time.
                 T=1100°C, Fluidizing Gas Velocity =1.0-1.5 m/sec.

-------
                                     45


Regeneration of Sulfated Tymochtee Dolomite by the Incomplete Combustion of
Coal

     Effect of Solids Residence Time and Temperature on Regeneration.  A
series of regeneration experiments with sulfated Tymochtee dolomite has
been initiated to further investigate  the effects of regeneration temper-
ature and solids residence time on the extent of regeneration and the
concentration of S02 in the dry effluent gas.  The experimental conditions
and results are given in Table 8.  The extents of regeneration as a
function of solids residence time, based on solids analysis, are plotted
in Fig. 17 for the three bed temperature levels 1000, 1050, and 1100°C.
At 1000°C, when the solids residence time was decreased from 37 min to 18
min, the extent of regeneration decreased drastically from 77% to 30%
(Table 8), but the S02 concentration in the dry effluent gas, 2.5%, did
not change.

     At 1050°C, as the solids residence time was decreased from 34 min to
12 min, the extent of regeneration decreased from 90% to 56% and the S02
concentration in the dry effluent gas increased from 3.0% to 4.8%.  At the
highest temperature level, 1100°C, it was found that as the solids residence
time was decreased from 13 min to 9.4 min, the extent of regeneration
decreased from 85% to 77% and the S02 concentration in the dry effluent gas
increased from 6.4% to 7.8%.  Relatively high S02 concentrations in regen-
erator off-gases have also been reported by Pope, Evans and Robbins.8

     Since the extent of regeneration remained quite high at 1100°C when
the solids residence time was lowered to 9.4 min, it is expected that S02
concentrations in excess of 10% in the effluent gas can be obtained by
further increasing the solids rate through the reactor.

     Effect of Solids Residence Time and Temperature on S02 Concentrations
in the Off-Gas.  The curves representing regeneration as a function of
solids residence time in Fig. 17 were extrapolated through the experimentally
obtained solids regeneration values.  By use of the regeneration curves
and the predicted gas volume increases during regeneration (see section on
Mass and Energy Constrained Model for the Regeneration Process),  the S02
concentrations in the regenerator off-gas were predicted at the same three
temperature levels, a pressure of 153 kPa, and a fluidizing-gas velocity
of 1.06 m/sec.  The predicted S02 concentrations in the effluent gas are
given in Fig. 18.  At 1000°C, the effluent gas S02 concentration cannot be
increased by decreasing the solids residence time (increasing the solids
throughput rate) because the regeneration rate would be too low.

     At 1100°C, higher S02 concentrations can be obtained by decreasing the
solids residence time.  Additional experiments are being performed to further
quantify these results.   The regeneration results with limestone and dolomite
show that as the solids residence time is decreased (-i.e.,  at a higher solids
throughput rate), the extent of regeneration is sacrificed but the S02 con-
centration in the dry off-gas is enriched.

     High temperatures improve both the extent of regeneration and the
S02 concentration in the off-gas from the regeneration process.  However,

-------
                       REGENERATION  OF CaO, %
 H-
era
era


n>
H
cu
rt


8

o
n
cu
O
X
co

P
§
o
s
en
o
H-
CL
CO
CO

H-
CU
o
ro

-------
          Table 8.   Experimental Conditions and Results for Regeneration Experiments
                    with Sulfated Tymochtee Dolomite Using Triangle Coal.

                    Nominal Fluidized-Bed Height:  46  cm    Pressure:   153  kPa
                    Reactor ID:   10.8 cm
                    Coal:  Triangle coal (0.98 wt %  S)   Ash fusion  temperature
                          under  reducine conditions.  1390°C (initial deformation)
                    Additive:   (1) -14 +50 mesh, 9.0 wt  %  S (CS-6,  -7,  -8)
                               (2) -14 +50 mesh, 9.4 wt  %  S (CS-10,  -11,  -12)
Bed
Temperature
Exp. (°C)
CS-6
CS-8
CS-7
CS-10
CS-11
CS-12
1000
1000
1050
1050
1100
1100
Fluidizing- Feed
Gas Velocity Rate
(m/sec) (kg/hr)
0.98
0.92
0.92
0.98
1.07
1.16
5.0
10.0
5.4
15.0
14.3
19.5
Solids
Residence
Time (min)
37
18
34
12
13
9.4
Reducing Gas Measured SO in CaO
Concentration Dry Effluent Regeneration
in Effluent (%) Gas (%) (%)a/(%)5
1.4
2.2
1.9
2.5
2.9
2.9
2.5
2.5
3.0
4.8
6.4
7.8
83/77
39/30
84/90
53/56
80/85
79/77
Based on flue-gas analysis.

Based on chemical analysis of dolomite samples.

-------
                                    48
                                      PREDICTED S02 CONCENTRATION
                                           i  I   I	I	I	I	1
                            10    15    20    25    30    35    40
                             SOLIDS RESIDENCE TIME , min
          Fig.  18.   Predicted and Experimental Sulfur Dioxide
                    Concentrations at Three Regeneration Temperatures.
                    Pressure:  153 kPa, fluidizing-gas velocity:
                    1.06 m/sec.
at regeneration temperatures above 1100°C, agglomeration of the sulfated
additive has been found to be aggravated.

Carbon Utilization and Carbon Balance Calculations

     In regeneration of additive, coal is combusted under reducing
conditions to provide the heat and the reductants required for reaction.
Under such reducing conditions, sufficient oxygen is not available to
ensure complete combustion of the coal.  Thus, relatively low carbon
utilization may be expected.

     To determine the utilization of the coal, carbon material balances
and carbon utilizations were calculated for three earlier reported
experiments listed in Table 9.  The carbon balances were good for
experiments CS-5 (97%) and LCS-2  (99%) and low for LCS-1 (55%).  The
carbon utilizations were found to be ^80%.  (The carbon was assumed to
have been utilized if it was oxidized to CO or C02.)  Additional data
are being obtained.

Effect of Coal Ash on the Agglomeration of Sulfated Additive during
Regeneration

     Ash Buildup in Sulfated Additive.  Partial agglomeration of sulfated

-------
        Table 9.   Experimental Conditions,  Carbon Balances, and Carbon
                  Utilizations for Regeneration Experiments.

                  Nominal Particle Residence Time:  30 min
                  Nominal Fluidized-Bed Height:  46 cm
                  Reactor ID:   10.8 cm
                  Coal:   Arkwright
                  Additive:   (a)  Sulfated Tymochtee dolomite
                                 -14 +50 mesh (CS-5)
                             (b)  Sulfated Greer limestone
                                 -10 +50 mesh (LCS-1, -2)
Exp.
No.
CS-5
LCS-1
LCS-2
Pressure
(kPa)
184
184
153
Bed
Temperature
(°C)
1040
1040
1100
Fluid iz ing-
Gas Velocity
(m/sec)
0.93
0.91
1.1
Additive
Feed Rate
(kg/hr)
5.0
4.9
5.4
Coal
Feed Rate
(kg/hr)
1.59
1.47
1.47
Carbon
Balance
(%)
97
55
99
Carbon
Utilization
(%)
80
72
82
                                                                                                        -e-
                                                                                                        VO
Carbon was assumed to have utilized when oxidized to C02 or CO.

-------
                                    50


additive (specifically, sulfated Greer limestone from Pope, Evans and
Robbins) has occurred, even during mild temperature upsets near 1100°C.
Observations of agglomerated interior surfaces with an optical microscope
revealed ash cenospheres, which appeared as tiny glass-like beads.  These
beads are formed as tiny bubbles as gases evolve from the ash.  X-ray
diffraction analysis showed them to be amorphous.

     The agglomerating tendency of molten ash (the Godel phenomenon) is
utilized as a means of removing the ash from a fluidized bed of bituminous
coke in Ignifluid gasification boilers,9 which are operated at high
fluidizing-gas velocities (10-15 m/sec) and at temperatures of 1200-1400°C.

     A suggested mechanism for one type of agglomeration in the presence
of coal ash is that at temperatures where these spheres and/or the ash
are molten, they could serve as coalescing agents between adjacent additive
particles in the regenerator.  Although it has been reported that molten
ash has a high surface tension,10 its interfacial tension with sulfated
additive at high temperature might be low enough to wet the additive,
causing the additive particles to become sticky and agglomerate.  This
could occur at temperatures above the ash fusion point and below the fusion
point of the additive.  Other types of agglomeration initiated by the
fusion of sulfated additive or by the formation of coal-ash additive com-
pounds are also possible.

     The sulfated additive feed to the regenerator contains coal ash because
during the preceding sulfation (combustion) ;;step, some coal ash is retained in
the fluidized bed.   Tymochtee dolomite has been sulfated during the combustion
of Arkwright No. 2 coal at ANL, and Greer limestone has been sulfated
during the combustion of Sewickley coal.  The ash fusion temperatures
(determined by ATSM D-1857-74) for these coals and for Triangle coal are
given in Table 10.   The ash fusion temperature under reducing conditions
for both Arkwright and Sewickley coals are very close to the regeneration
temperatures used.   The presence of ash in the additive might have con-
tributed to the above coalescing mechanism and disrupted the fluidization
in the regenerator.  Once the fluidity of the bed was disrupted, localized
temperature excursions provide the atmosphere for chemical reactions of
the coal ash with additive.
          Table 10.  Fusion Temperatures, Under Reducing Conditions,
                     of Ash from Arkwright No. 2 Coal, Sewickley
                     Coal, and Triangle Coal.

                            Arkwright Ash   Sewickley Ash   Triangle Ash
      Initial Deformation       1104             1100           1383
      Softening (H = W)         1177             1188           1444
      Softening (H = 1/2 W)     1193             1215           1485
      Fluid                     1232             1288           1510

-------
                                    51
     During some regeneration experiments, agglomeration of additive and
coal ash occurred even during mild temperature fluctuations caused by
upsets in the regenerating environment at ^1100°C.  Molten ash is believed
to be responsible for initiating some agglomeration at these temperatures;
analyses of sulfated additive have been performed in an attempt to further
understand this postulated mechanism.

     From a chemical analysis of once-sulfated Greter limestone (-8 mesh)
provided by Pope, Evans and Robbins  (Test 620), its composition was cal-
culated and compared (Table 11) with the composition of unreacted Greer
limestone.11  The concentrations of Fe203, Si02, and A1203 increased
markedly during sulfation.  These compounds are the major constituents of
the ash of Sewickley coal (Table 11) used in Test 620.  The temperature
of initial deformation of the ash from this coal (Table 10) under reducing
conditions is
     Calcium concentration was used as a basis (ash contains only a small
amount of calcium) for estimating the amount of ash in the sulfated
limestone.  Increases in silicon, iron, and aluminum were calculated from
the analyses.  The corresponding calculated ash contents of the sulfated
limestone are given in Table 12.  Since SiC>2  is  the most abundant ash
.constituent, about 50% of the ash, its chemical analysis is expected
to given the most reliable result.  The ash content of the sulfated lime-
stone based on the silicon analyses was 10%;  This high level of ash
may have been responsible for the formation of agglomerates during
relatively mild temperature upsets.
      Table 11.  Compositions of Greer Limestone and Sewickley Coal Ash.
Unsulfated Greer
Limestone

CaO
CaSO^
MgO
Fe203
Si02
A1203
S
Others
C02 Loss
Total
(wt %)
44.80
1.90
0.80
10.50
3.60
0.17
0.71
37.52
100
Sulfated Greer
Limestone3
(wt %)
36.3
30.4
1.94
2.53
16.63
5.75
7.16b
—
3.54
100
Sewickley
Coal Ash
(wt %)
5.0
1.0
20.3
49.8
19.5
4.4
—
100
  Sewickley coal was combusted in the Pope, Evans and Robbins experiment
  in which this limestone was sulfated.
  Also included in CaSOij value.

-------
                                    52
          Table 12.  Coal Ash Level in Sulfated Greer Limestone.
                     Mass basis:  100 g unsulfated limestone

                                            Equivalent
                                            Ash Content

Ca
Fe
Si
Al
a.
0
8.1
10.4
9.1
          3.
           It was assumed that the ash contributed no calcium
           (Sewickley coal ash contains only 3.6 wt % calcium).


     The effect of the presence of these ash constituents on the fusion
temperature of the limestone is being investigated with a differential
thermal analyzer combined with X-ray diffraction analysis.

     Differential Thermal Analysis Study of Additive and Coal Ash
Reactions.  A differential thermal analyzer (DTA) has been constructed to
investigate the problem of agglomeration of sulfated additive with residual
coal ash at elevated temperatures (>1100°C) during the regeneration of
sulfated additive.  In future and ongoing work with the DTA apparatus, the
objectives are to study the following:

     1.  Chemical reactions and their temperatures, as well as fusion
         temperatures, involved in the agglomeration process.
     2.  The effects of key ash constituents (Si02, A1203, Fe203).
     3.  The effects of residual coal ash on chemical reactions and
         physical changes.

     The DTA experimental apparatus (Fig. 19) consists of a platinum-
wound tube furnace which contains two crucibles, one for holding the sample
material and the other containing a reference material, alumina powder.
Thermocouples imbedded in the crucibles are connected to temperature
recorders or a potentiometer.  The sample temperature and the temperature
difference between the sample and the reference material are recorded as
a function of time.  The latter plot is of particular interest since it
permits observation of pertinent heat changes with a rise in sample
temperature (a heating rate of 15°C per minute is employed).  Heat of
reaction, heat of fusion, and other energy changes can be observed.  They
appear as peaks  (deviations from the baseline) on the plot of temperature
difference (AT) versus time.  The gas environment in the furnace is air.

     The experimental approach consists of first heating  (at 15°C/min) a
sample to a high temperature—1400°C, for example.  All detectable chemical
and physical changes which occur up to this temperature are surveyed.  In
the next phase,  fresh samples of the same material are heated just beyond
each of the reaction peaks noted in the original AT versus time plot.  The
composition of a sample from each stage of the reaction is determined by

-------
                                    53
                                AMPLIFIER
                 SAMPLE
              TEMPERATURE
                RECORDER
                 SAMPLE
              TEMPERATURE
              POTENTIOMETER
                 FURNACE
                 AMBIENT
               TEMPERATURE
                RECORDER
AT RECORDER
 PROGRAMED
TEMPERATURE
 CONTROLLER
                          0 °C REFERENCE JUNCTION
                           FURNACE CHAMBER


                "EXPERIMENTAL SAMPLE
                                       REFERENCE SAMPLE'
     Fig.  19.   Schematic of DTA Apparatus.

X-ray diffraction.   Relatively large ground  samples  (>JL  g)  are used,  to
compensate for the nonhomogeneity of the  materials.

     The reactions in two different materials have been investigated to
date:  (1) unsulfated Tymochtee dolomite and (2) sulfated Greer limestone
from Pope, Evans and Robbins, Test 620.  The reaction temperatures and
product analyses from these two series of DTA experiments are given in
Tables 13 and 14.  In each of these tables,  the column headed "maximum
temperature" is the temperature at which heating of the sample was
terminated; "peak temperatures" are the observed (AT) peak temperatures;
"product condition" represents visual evaluations of the post-reaction
samples at room temperature; and the remaining columns list the product
constituents in decreasing order of importance.

     To assist in the calibration of the newly constructed differential
thermaly analyzer, the decomposition temperatures of unsulfated Tymochtee
dolomite [CaMg(C03)2] were determined.  Literature values for the main
peak temperatures for tite two dolomite decomposition peaks are 790°C
(for the decomposition of MgCOs to MgO) and  940°C (for the decomposition
of CaC03 to CaO).12  The temperatures obtained of 819°C and 933°G (see

-------
Table 13.  Differential Thermal Analysis Results and X-Ray Diffraction
           Analysis of the Reaction Products of Unsulfated Tymochtee Dolomite.

Maximum
Temperature
Sample No. (°C)
DOL-3A
DOL-3B

836
990

Peak
Temperatures Product
(°C) Condition
819 Unchanged
819, 933 Unchanged



Product Constituents
Major Medium
CaC03 MgO
CaO, MgO

Minor Very Minor
CaO a-quartz
possible
Ca2SiOi+,
a-quartz
                                                                                                  Ln

-------
         Table 14.  Differential Thermal Analysis Results and X-Ray Diffraction Analysis of the
                    Reaction Products of PER Sulfated Greer Limestone and Residual Coal Ash.a
Maximum
Temperature
Sample No. (°C)
PER-620
PER-620-1
PER-620-2
PER- 62 0-3
PER-620-4
PER- 62 0-5
PER-620-6B
PER-620-6A
PER-620-6A°
not
heated
625
790
960
1180
1218
1297
1409

Peak
Temperatures
not
heated
516
510, 770
523, 787,
842
523, 785,
843,1093
506, 789,
841,1098
1208
520, 786,
846,1086,
1212,1277
534, 790,
843,1101,
1208,1264

_ , Product Constituents
Product
Condition Major Medium
-
unchanged CaSOif , CaO
unchanged CaO CaSOi^
CaSOtf CaO
unchanged
b
sintered, 2Ca2SiOtt'CaSOtt CaSOif
very brittle
b
fused, 2Ca2Si04'CaSOit CaSO^
slightly
brittle
fused, very 2CaO-Al203-Si02,
slightly CaSOi^.
brittle
completely possible Ca^AlgO^SO^
fused, hard, or sulfo-aluminous
dark-colored clinker, CaSOi^
Cae.iMg^sia.eO!^
Minor Very Minor
CaO, Ca(OH)2, a-quartz
CaC03
CaC03 a-quartz,
Ca(OH)2
a-quartz
a-quartz
CaO, possible
Ca2Al2SiOy ,
CaO, possible
$-Ca2S±Oi+ possible

possible CaSOi^.,
2CaO-Al203-Si02,
High ash content (^22 weight percent).
Composition reported to be uncertain.
Sample taken from product surface, instead of from the core  (other PER-620-6A sample).
                                                                                                               Ln
                                                                                                               Ui

-------
                                    56


Table 13) are in good agreement with these literature values.  The X-ray
diffraction analysis confirmed that the expected decomposition products
were obtained.  In a thesis by Beck,13 it was sugggested that the first
dolomite peak (decomposition of MgC03) occurs at a higher temperature than
does that for magnesite alone because energy must be supplied to break
down the crystal structure of dolomite, as well as to decompose the magnesium
carbonate.

     The sulfated  Greer  limestone  from Pope, Evans  and Robbins  (PER-620)
 contained 36.3%  CaO,  2.53% Fe203,  16.6%  Si02,  5.75% A1203, and  3.5%  C02.
 Based on this analysis,  it was  estimated that  the sulfated additive  con-
 tained  10%  coal  ash.  The X-ray diffraction analysis of a sample also
 revealed the presence of Ca(OH)2 (Table  14).

     In this series of DTA experiments,  experiment  PER-620-6A was performed
 first (a sample was heated to 1409°C)  to survey all detectable  chemical
 and physical changes.  Next, individual  samples were heated  past each of
 the temperature peaks observed  in  the  first experiment.  The maximum heating
 temperatures, peak temperatures, and product conditions for  this series of
 experiments are given in Table  14.

     Comparison of the X-ray diffraction analysis of the product sample
 from experiment PER-620-1 (a sample that  was heated to 625°C) with that
 of the  unheated material (PER-620) indicates that Ca(OH)2 dehydrated to
 form CaO.  This dehydration occurred at  temperatures ranging from 506 to
 534°C where an endothermic peak was observed.  The  handbook value for
 the dehydration temperature of  pure Ca(OH)2 is 580°C.  Our samples are
not pure, however, and the possible effects of other constituents on the
dehydration peak temperature of Ca(OH)2 must be considered.  There is
evidence14 of decreasing peak temperatures of Ca(OH)2 with increasing
dilution of the original sample.   In other words, pure Ca(OH)2 yielded
the highest dehydration peak temperature, while the most dilute samples
yielded  the lowest values.

     The results from the next  two experiments, PER-620-2 and -3 (in which
endothermic peak temperatures of ^790°C and ^840°C that represented
areaetionshwere observed)-were compared with results for magnesian limestone
that have been discussed by MacKenzie.15  MacKenzie observed three peaks
for a magnesian limestone.  The first peak at 780°C was attributed to
the decomposition of MgC03.  The last two peaks were attributed to the
decomposition of the dolomitic CaC03 (at  880°C) and the free calcium
carbonate (at 940°C).  It is unlikely that any MgC03 is present in the
PER sulfated limestone samples since the  sulfation temperature of the
material was ^900°C.  It was concluded that the peaks at 790°C and 840°C
are CaCOs calcination peaks.   X-ray diffraction analysis confirmed the
disappearance of CaC03.  The decomposition peak temperatures obtained
 (Table 14) are about 100°C below the referenced values.   This may have the
same explanation as the discrepancy for the Ca(OH)2 peak, i.e.,  the dilute
nature of the sample may tend to decrease the peak temperatures of individual
components.

     In  experiment PER-620-4, the sample was sintered and the peak temper-
ature ranged from 1086-110l"C (the definition of the exothermic peak was

-------
                                    57


not clear) .   This is near the temperature at which agglomeration of PER-620
material has occurred in the regenerator.  X-ray diffraction analysis of the
product from PER-620-4 indicates that 2Ca2SiCVCaSOit is the major constituent.
The high silica content of the sulfated limestone contributes to the for-
mation of this compound.  Since the CaO level was reduced from a medium
constituent to a minor constituent in the PER-620-4 experiment when the
sample  was heated to 1180°C, it is suspected that the principal reaction
occurring may have been:

          2CaO + Si02 -»• Ca^SiOit                                       (6)
     Sample PER-620-5 (heated to 1218°C) showed no change in the analysis
from that for the sample heated to 1180°C although an endothermic peak
was observed at 1208 °C.  This peak may represent the transition of g-CaSOtt
to a-CaS04.  Previous DTA work16 by West and Sutton on pure CaS04 showed
that this transition occurred at 1225 °C and that on cooling, an exothermic
peak appeared 40 to 50 degrees lower than in the heating thermogram.  This
agrees with the results in Table 14 and with unreported cooling peak
temperature results.

     Analyses of samples taken from the products of experiments PER-620-6B
and PER-620-6A, showed that complex molecules of calcium-aluminum silicates
formed in the sample interior.  Gag . jMgj . iS±3. 6°14. 4 was found at the
surface of the sample from PER-620-6A (which had been heated to 1409°C).
West and Sutton16 had similar results, demonstrating that significant
reaction between anhydrite (a-CaSO^) and Si02 and Fe20s occurs above the
3-a transition temperature (1225°C) for anhydrite.  The results obtained
indicate that ash contributes further to the agglomeration problem since
aluminum (as well as silicon) is a constituent of the complex molecules.

     Agglomerates from regeneration experiments (at ^1100°C) with sulfated
Greer limestone (PER-620) have also been analyzed by X-ray diffraction;
y- and 3-Ca2SiOi+ and Ca2Al2SiC>7 have been found in different locations in
the agglomerates.  The agglomerates consists of large particles of sulfated
limestone and residual coal ash (-10 +30 mesh) .  Since the one-step regen-
eration process is operated at VL100°C, the formation of calcium silicates
at this temperature may be the first stage in agglomeration of the fluidized-
bed material, which may be followed by a loss of fluidity and greater
localized temperature increases.  The objectives are to find the tolerable
limits of silicates and other compounds for the regeneration process.

Mass and Energy Constrained Model for the Regeneration Process

     In the investigation of the feasibility of the one-step regeneration
process for sulfated S02-acceptors, one objective is to optimize the process
conditions from technical and economic points of view.  The one-step
reductive decomposition is carried out in a fluidized bed at elevated
temperature (M.100°C).  The heat and the reducing gases are provided by
the combustion of coal (or other fuels) in the fluidized bed under reducing
conditions.

-------
                                    58
     A mass and energy constrained model for the one-step regeneration
process is being developed which  is being used to predict the effects of
experimental variables on:

     1.  Off-gas composition  (SC>2, C02, H2<), etc.)
     2.  Volumetric gas changes
     3.  Fuel and oxygen requirements
     4.  Individual energy terms
     5.  Energy cost of regeneration

Sensitivity analyses can be performed for key variables such as:
     1.  Regeneration temperature, T
     2.  Regeneration pressure  (near atmospheric), P
     3.  Fluidizing-gas velocity, V
     4.  Reactor size
     5.  Feed solids and gas  temperatures and compositions

     The first phase of this  work is concentrated only on what occurs
in the regeneration reactor,  as illustrated in the flow diagram (Fig. 20).
The early objectives are to predict the effects of experimental variations
and to use these results to guide future experimental efforts.

     In the one-step reductive regeneration process, the assumed combustion
reactions for coal (CH  + nH^O) are as follows:
oxidizing conditions:

     (CH  + nH20) (s) + (y + 1)02 + CQ2

                                                  n)H20
                                         (7)
                              OFF -GAS
                     ELUTRIATED  COAL PARTICLES
         SULFATED
          ADDITIVE
                                  t
TEMPERATURE,

  PRESSURE,

 FLUIDIZING-

 GAS VELOCITY
                            i            r
                                                   HEAT  LOSS
                                                    REGENERATED
                                                      ADDITIVE
                           COAL      °2'N2

          Fig.  20.   Flow Diagram for the One-Step Regeneration Process,

-------
                                    59


     reducing condition:

           (CHm + nH20) (s) +  1/2 02 ->  CO + y H2 + nH20                   (8)

     The main decomposition  reactions that  the S02-accepting  additive
 (natural stones) undergo are the calcination reaction:

           CaC03(s) -> CaO(s)  + C02                                       (9)

 and the reductive decomposition reactions:

                 ) + CO ->- CaO(s) + S02 + C02                            (1)

                 ) + H2 -> CaO(s) + S02 + H20                            (2)
All of these decomposition reactions are endothermic, and  their  thermal
requirements are provided by  the coal combustion reactions.  From  the
above combustion and decomposition reactions, it is apparent that  gas
volume throughput increases as a result of reactions in the regeneration
reactor.

     The total internal heat  requirement (Qc) in the regeneration  reactor
is:

          Q  = Q,    +Qj   +Q-,+QJJ+Q    +Q                   (10)
          sc   xlost    des    cal    add    gas    ec                 v  '

where Q-.    = heat losses from the reactor
      xlost
      Q,    = heat of desulfation
      Q  .,  = heat of calcination
       cal
      Q     = sensible heat difference between regenerated and sulfated
              additive
      Q     = sensible heat difference between effluent and input  gas
       gas
      Q     = sensible heat for the unburned elutriated coal
       ec
It is assumed that this internal heat requirement will be satisfied by
the combustion coal.

      Q  = Q . + Q                                                     (11)
      xc   xci   xcc
where
      Q . = heat liberated during the required incomplete combustion of
            coal

If Q -  < Q ,
    ci    c

     Q   = additional heat requirement to be supplied by the complete
      cc
           combustion of coal

-------
                                    60
if Qci > Qc,

     Q   = heat which must be removed from the process

     In the model, no kinetic predictions are made; instead, experimental
results for the extent of solid (CaO) regeneration that have been correlated
as a function of solids residence time are used.  The above mass and energy
relations are interlocked and are solved at different reactor solids resi-
dence times (the solids residence time is equal to the equivalent mass of
the fluidized bed devided by the mass feed rate of sulfated additive).  The
major assumptions and properties of the model are summarized below.

     1.  Solids and the off-gas exit at the reactor temperature.
     2.  The additive is fully calcined.
     3.  The extent of regeneration is a function of solids residence
         time and temperature (the assumption is based on experimental
         data).
     4.  The chemical composition of the additive is a function of the
         extent of regeneration.
     5.  The heat capacities of solids and gas constituents are
         temperature-dependent.
     6.  Oxygen and nitrogen can be fed separately.

     Examples of plotted calculated output are given in Fig. 21 through
25.  The input regeneration conditions were:

          Temperature:  1094°C (2000°F)
          Fluidizing-gas velocity:  1.07 m/sec (3.5 ft/sec)
          Pressure:  153 kPa (22.5 psia)
          Reactor ID:  10.8 cm (4.25 in.)
          Nominal fluidized-bed height:  46 cm (19 in.)
          Feed gas temp:  344°C (650°F)
          Feed solids temp:  25°C (77°F)
          Solids residence time:  2-200 min (mass feed rates:
                                  90-0.9 kg/hr)
          Sulfated Tymochtee dolomite:  9.5% sulfur and 9.5 CO^

     Effects of Solids Residence Time on Volumentric Gas Change.  The input
functional dependence of extent of solids regeneration on solids residence
time (shown in Fig. 21) was obtained from experimental results.  This
graph also shows the calculated increase in gas volume during regeneration
as a function of solids residence time  (SRT).  At SRTs of less than 10 min,
large gas volumentric increases are predicted, caused by increased coal
combustion due to increased heat requirements (sensible heat differences
of solids and gases, and heat of decomposition reactions) and increased
solids decomposition for the process, equations 1, 2, 7, 8, 9. The volu-
metric increase dilutes the S02 ifl the effluent gas.

     Relation of Coal Feed Rate and Oxygen Concentration to Solids Residence
Time.  In Fig. 22, the predicted coal feed rate and the predicted required
oxygen concentration are given as functions of SRT.  The experimental

-------
                              61
o-
V

EXTENT OF REGENERATION , %
0 50 60 70 80 90 100 UO
i i i i i i i


\
\
\
\











\
\




/






\




'







\


1








\

;










\
/










\
'


i i i i i 1 1
o = EXTENT OF REGENERATION . %
o = INCREASE IN GAS VOLUME , %







\
/



EXPERIMENTAL CONDITIOf^
T=2000°F, V= 3.5 FT/SE
P = 22.5 PSIA
IS
^
FEED GAS 650 °F
FEED SOLID



\








\


'D






^



^

77 °F
•





»^














•fe






•^-,






•~~-






^^-^^
-&
-8

D 10 20 30 40 50 60 70
INCREASE IN GAS VOLUME , %
                      RESIDENCE  TIME .  MIN
                                                   100
Fig. 21.   Experimental Solids  Regeneration and Predicted Increase
          in Gas Volume During Regeneration Versus Solids Residence
          Time.

-------
                               62
    8
  K


  Q
  W
O -
*H
P
o
8
£


n
S

•-4
\









\
\






\

\
\





\
\

\
\




s
H
\


\




0
0
a
\






LI
L

\


\



D
L




>



III 1
a = COAL FEED RATE
o = OXYGEN IN FEED GAS ,
SYMBOLS : EXPERIMENTAL VALUES
OW SYMBOLS : CALCULATED VALUES

CS-12
\
•\cs-
CS-12
\ CS-
\





-II
\
-II
\






\
















tJ








--^






•v
--«,






3,
>-






-•







-—
•& — &—&-fif


                                                           •8
                                                           s
                                                           •8
                                                             x
                                                             o
                                                           •a
                        10

                     RESIDENCE TIME ,  MIN
100
Fig. 22.  Predicted Required Coal Feed Rate and Oxygen

          Concentration in the Feed Gas as  Functions  of

          Solids Residence Time.

-------
                             63
                                      i—i—r


                                      a = S02



                                      A = C02
               rm	


                DRY BASIS

                WET BASIS
                       SOLID SYMBOLS: EXPERIMENTAL VALUES

                       HOLLOW SYMBOLS:CALCULATED VALUES
       8

                           CS-12
     C/2


     O


     w
CS--II
     o


     £s
     w
     u
     K
     W
     (X,
                       \
                           CS-12

                               CS
  -•II
                          10
                                                  100
                       RESIDENCE TIME ,  MIN
Fig.  23.  Predicted Off-Gas Constituent Concentrations  as

         Functions of  Solids Residence Time.

-------
                              64
                              I     I   I  I  I  I I  I I
                            D = HEAT OF COMBUSTION
                            o = SENSIBLE HEAT OF SOLID
                            A = SENSIBLE HEAT OF GAS
                            + = HEAT OF DESULFATION
                            x = HEAT OF CALCINATION
                            o = HEAT LOST
                                               100
                    RESIDENCE  TIME .  MIN
Fig.  24.   Predicted Individual Heat  Requirements as a Function
          of Solids Residence Time.

-------
                               65
     (O

     CO
     CM

     CO
     co

     N
   o
   t— 1
   E-i
      CM
    53 «
    o -
    o
    W
    E«
7
                          10
                                                   100
                       RESIDENCE TIME .  MIN
Fig. 25.  Predicted Fuel Cost for Regeneration per Electric Power
          Unit Produced when Burning 3% Sulfur Coal as a Function
          of Solids Residence Time.   (These costs are not repre-
          sentative of those in a commercial plant since conditions
          have not been optimized.)

-------
                                     66


conditions and results  for the measured variables  in  experiments  CS-11
and-12 are also plotted for comparison with  the predictions.  Agreement
of predicted with experimentally obtained values was good.

     Relation of Off-Gas Composition to Solids Residence  Time.  The predicted
off-gas constituent concentrations as functions of solids residence time
are given in Fig. 23.   The concentrations of all pertinent constituents
increase with decreasing SRT  (increasing solids feed  rate).  At a SRT of
5 min, a S02 concentration of 9.4% in the dry off-gas is  predicted for
the experimental conditions described above.

     Relation of Heat Requirements to Solids Residence Time.  The predicted
individual heat requirements for the one-step regeneration process are
plotted in Fig. 24 as functions of solids residence time.  At low SRT, the
sensible heat requirements of the solids and gas constitute most  of the
required heat.

     Relation of Regeneration Fuel Cost to Solids Residence Time.  The fuel
cost for regeneration per electric power unit produced when burning 3%
sulfur coal has been predicted as a  function of SRT and is shown  in Fig. 25.
The experimental conditions for the  regeneration are given above.
The costs plotted are relative costs obtained to estimate the effect of
operating conditions such as SRT and feed solid and gas temperatures.  In
an industrial process,  these costs would be reduced by recovering  the
sensible heat from the  effluent regenerator streams.  (A coal cost of $26/ton
was used in predicting  the regeneration fuel costs.)  It has been  found that
for the experimental conditions outlined above, the energy cost of regen-
eration is optimum at a SRT of ^9 min.   At lower solids residence  times,
the extent of regeneration decreases rapidly and the sensible heat require-
ments for the solids and gas increase rapidly.

     Predicted Effects  of Input Conditions on Regeneration Results.  The
effects of solids and gas feed temperatures, pressure, carbonate  concentration
in the additive,  and reactor diameter on:  S02  concentration in the dry
effluent gas, oxygen requirement, increase in gas volume, process heat
requirement,  and energy cost of regeneration were predicted.   The results
at a SRT of 5 min are given in Table 15.  Case 1 represents the design
experimental condition  capabilities  of the existing bench-scale experimental
regeneration system at ANL.  In cases 2, 7,  8,  9, and 10, the effects of
preheating the feed solids and gas to temperatures as high as 1094°C (2000°F)
were predicted for a SRT of 5 min.   The SC>2 concentration in the wet effluent
gas (not shown) was predicted to increase from 7.8% in case 1 (gas containing
17% water) to 8.3% in case 10 (gas containing 2.9% water).  However, it is
predicted that if the temperatures of the feed solids and gas are  increased
as in case 2 and dases  7-10, the S02 concentration in the dry effluent gas
would decrease because  the water content decreases.  The required oxygen
concentration in the feed gas decreased from 53% in case 1 to 4.0% in case
10.  These data are useful in selecting operating conditions for a realistic
plant situation in which air rather  than oxygen-enriched gas would be utilized.

-------
                                       Table  15.  Regeneration of Tymochtee Dolomite.

                                                  Regeneration Temperature:  1094°C
                                                  Solids Residence Time:  5 min
                                                  Fluidizing Gas Velocity:  1.07 m/sec
Input Conditions
Case
1
2
3
4
5
6
7
8
9
10
Feed
Solids
Pressure Temp
(kPa) (°C)
153
153
153
153
153
102
153
153
153
153
25
650
650
650
'650
650
870
1094
870
1094
Feed
Gas
Temp
345
650
650
650
650
650
870
870
1094
1094
$
9.5
9.5
9.5
9.5
0
9.5
9.5
9.5
9.5
9.5
Reactor
Dia
(cm)
10.8
10.8
21.6
43.2
10.8
10.8
10.8
10.8
10.8
10.8
S02 in Dry
Effluent
Gas (%)
9.4
8.9
8.8
8.8
9.7
12.0
8.7
8.6
8.7
8.6
Req . 02
in Feed
52.9
22.8
19.1
17.8
17.6
30.9
10.8
4.0
9.8
4.0
Predicted Results
Gas
Volume
Increase
48.5
41.7
40.9
40.6
28.1
60.9
39.0
37.7
38.8
37.7
Process Heat
Requirement
(Btu/hr)
+92 K
+34 K
+109K
+395K
+24 K
+30 K
+11 K
-9 K
+9 K
-11 K
Fuel Cost for
Regeneration3
(raill/kWh)
0.60
0.30
0.27
0.25
0.25
0.28
0.18
0.10
0.17
0.10
Tons of Coal for
Regeneration/Ton of
Coal for Combustion
.069
.035
.031
.029
.029
.032
.021
.012
.020
.012
aThese costs are not representative of those in a commercial plant since conditions have  not  been optimized.
 Coal cost, $26/ton.
 A 3 wt % S coal was assumed to be burned during the combustion step.

-------
                                     68


     The process heat requirement (Qcc) is the heat that must be added
to or removed from the regeneration process in addition to the heat liberated
by incomplete combustion of coal.  In ciases 8 and 10 in which temperatures
of feed solids and gas are high, heat must be removed from the reactor to
maintain the design experimental conditions.

     The energy cost of regeneration per unit of electric power generated
with the combustion of 3% sulfur coal decreased from 0.6 (case 1) to 0.1
mill/kWh (case 10).  It was assumed in these calculations that no energy
was recovered from the hot (1095°C) regenerated additive.  In terms of
coal usage, VL-3% additional coal to that combusted in the power generating
system would be required for regeneration of the sulfated sorbent.

     In cases 2, 3, and 4, the effects of increasing the reactor ID from
10.8 cm to 43.2 cm were predicted.  The effects were not dramatic.   For
larger reactors, the percent heat losses are reduced and thus the energy
cost of regeneration is reduced.  The SC-2 concentration in the dry
effluent gas is predicted to remain almost the same.

     The effect of carbonate content of the sulfated additive was predicted
by comparing cases 2 and 5.  By decreasing the concentration of CO^ from
9.5% to 0, the S02 concentration in the dry effluent gas was predicted
to be increased from 8.9% to 9.7%.  The energy cost of regeneration would
be decreased from 0.3 to 0.25 mill/kWh.

     The parameter that was most effective for increasing the S02 concen-
tration in the dry effluent gas was found to be regeneration pressure.  It
was predicted that by decreasing the pressure from 153 kPa (22.5 psig) in
case 2 to 102 kPa (15 psia) in case 6, the S02 concentration would increase
from 8.9% (case 2) to 12% (case 6) because the total gas feed rate would
decrease by 50%.

     These predicted effects of experimental variations on dependent
variables (such as 50% concentration in the effluent gas) will be used
to guide future experimental efforts.

                 CYCLIC COMBUSTION/REGENERATION EXPERIMENTS
 J. Montagna and W. Swift (Principal Investigators), H. Lautermilch,
R. Mowry, F. Nunes, C. Schoffstoll, G. Smith, S. Smith, J. Stockbar, G. Teats

     In the utilization of regeneration technology, recycling of the
sorbent a sufficient number of times without loss of its reactivity for
either sulfation or regeneration and without severe decrepitation is import-
ant.  Unless both of these requirements are met, the sorbent makeup rate will
be so high that regeneration will not be justified.  An experimental effort
is being made, therefore, to evaluate the effects of cyclic operation on
the resistance to decrepitation and the reactivity of Tymochtee dolomite in
ten combustion/regeneration cycles, 2 1/2 of which have been completed.
Prior to the cyclic experiments, Tymochtee dolomite which had been sulfated
and regenerated was again sulfated in the combustor to obtain preliminary
operating data.

-------
                                     69
Preliminary Experiment.  Sulfation of Regenerated Tymochtee Dolomite

     Before the 10-cycle series of experiments was begun, experiment RC-1A
was performed as the initial experiment* in the planned evaluation of the
effects of sorbent recycling on sorbent reactivity and decrepitation.  The
sorbent used was Tymochtee dolomite (^50% CaCO$, ^40% MgC03. as received)
which had been sulfated and regenerated.  Sulfation was done in the
bench-scale combustor in the PSI-series experiments (combustion with
Arkwright coal.; see Pressurized, Fluidized-Bed Combustion Studies).  After
sulfation, it was found that utilization of the additive had been ^62% and
that the sorbent contained VLO-11% sulfur by weight.  The sulfated dolomite
was then regenerated in the 4 l/4-in.-dia regenerator under various operating
conditions; it was then thoroughly mixed and resulfated in experiment RC-1A.
It contained 4.3 wt % sulfur, which corresponded to an additive utilization
of 18%.

     Experiment RC-1A was performed over a period of three days; 0;54.5 kg
of regenerated dolomite was processed.  Operating conditions and flue-gas
analysis results are given in Table 16 for the steady-state conditions
existing during the longest (^6 1/2-hr), most stable operating period of
the experiment.  Bed temperature and flue-gas composition data for that
period of operation are plotted in Fig. 26.

     Based on the flue-gas analysis for S02 (250 ppm), sulfur retention
(based on sulfur in the coal and not including the sulfur in the original
sorbent) was determined to be ^90%.  This S02 level agrees very well with
the expected level in the flue gas of 240 ppm S02 calculated from correlations
of previous experiments.17  Thus, the results of the preliminary recycle
experiment were very encouraging regarding the capability of the sorbent to
retain sulfur in the second combustion cycle.
     Table 16.  Operating Conditions and Flue-Gas Analysis for Segment
                of Combustion Experiment RC-1A.
Combustor: ANL, 6-in.
Bed Temp: 840 °C
Pressure: 810 kPa
Fluidizing-
Gas
Feed Rate (kg/hr) Ca/S Velocity
Coal Sorbent Ratioa (m/sec)
14.1 3.2 1.6 1.0
-dia Excess Air: 17%
Coal: Arkwright
Sorbent : Regenerated
Tymochtee Dolomite
Flue-Gas Analysis
S02 NO NOX CO C02 02
(ppm) (ppm) (ppm) (ppm) (%) (%)
250 120 150 40 17 3.1
 Based on unreacted calcium (CaO) in regenerated dolomite and sulfur in coal.
 Equipment, materials, and procedures used in the combustion steps of
 the cyclic experiments are described in a following section, Pressurized,
 Fluidized-Bed Combustion;  Bench-Scale Studies.

-------
                                                 Gas Concentrations	
  H-
  00
w w
X (D
T3 CU
(D
H H
(B XI
is ro
£
  rt
  c
  Pi

  *i

  E1
  ro

  o
  P3


  n


  I
  o
  en
   CO

   ro
   §
   rt


   O
               S02, ppm CH4, ppm  CO, ppm    O2, /<

               0        500 0         50 0        150 0
                               -<
                               rvi
                               m
                               73
     C02,  %    NO, ppm Bed Temp, CC

10 0         50        150   750      1000
•-4
c
t
1
t
1
t
1

1
t
1
t
1
•1
c
1
1
t
t


,

,


\
,

f








                                                                  o
                                                                  o
                                                                 ro


                                                                  I
                                                                  Nl
                                                                  m
                                                                  73
                                                                  r>
                                                                  —I
                                                                  m
                                                                  o

                                                                  —I
                                                                  O

-------
                                    71


     Calcium utilizations were calculated from steady-state samples of the
overflow, the primary cyclone, and the secondary cyclone (Table 17).   Utili-
zations based on the sulfur and calcium contents of the respective samples
ranged from 56% in overflow to 67% in secondary cyclone samples.   Also given
in Table 17 are utilization values adjusted to reflect the 18% calcium
utilization of the dolomite feed to the combustor.   The adjusted utilizations
ranged from 47% in-overflow to 60% in secondary cyclone samples.   These
utilization values provide evidence that finer dolomite particles are more
active in S02 retention.

     A calcium material balance for experiment RC-1A showed that of the
sorbent fed to the combustor, ^62% was removed in the product overflow,
v30% was recovered in the primary cyclone, ^2% was recovered in the
secondary cyclone, and ^6% was unaccounted for.

     Screen analyses of samples of regenerated dolomite.feed, overflow,
primary cyclone, and secondary cyclone materials are presented in  Table 18.
Although conclusive evidence  in the form of chemical determinations is
lacking, it appears from the  screen analyses in Table 18 that decrepitation
of dolomite particles in the  fluidized bed was not the sole source of ash
products recovered in the cyclone.  Comparison of the steady-state overflow
with the regenerated dolomite feed shows a larger loss by entrainment  (in
percent) for the smaller particle sizes.  Examination of the primary
cyclone screen analysis shows a significant amount of +80 mesh material,
which is the size of the smallest particles in the regenerated dolomite
feed.  This indicated that at least some of the sorbent carryover  into the
cyclones may have occurred without decrepitation of the sorbent in the
fluidized bed.  Based on these results, more rigorous determinations  of
decrepitation during selected cycles  of  the ten-cycle series of combustion/
regeneration experiments are  planned.


     Table 17.  Utilization of Calcium  in Overflow, Primary Cyclone,
                and Secondary Cyclone Samples  from Experiment RC-1A.

                              Calcium  Utilization (mol S/mol Ca) x  100
                              Based on  Sulfur  and   Adjusted  for Feed
                              Calcium Content  of    Having 18% Calcium
Sample
Overflow
Primary Cyclone
Secondary Cyclone
Sample
56%
62%
67%
Utilization
47%
54%
60%
        Weighted Average
         (based  on product
        distribution  of
        calcium)

-------
                                     72
     Table 18.  Screen Analysis Results for Combustion Experiment RC-1A.
Wt Fraction in Size Range
U.S.
Sieve No.
+14
-14 +25
-25 +35
-35 +45
-45 +80
-80 +170
-170
Regenerated
Dolomite Feed
0.00
0.17
0.18
0.43
0.22
0.00
0.00
Steady-State
Overflow
0.00
0.36
0.28
0.30
0.05
0.00
0.00
U.S.
Sieve No.
+45
-45 +80
-80 +100
-100+170
-170+230
-230+325
-325
Wt Fraction
Primary
Cyclone
0.30
0.25
0.04
0.10
0.06
0.08
0.16
in Size Range
Secondary
Cyclone
0.20
0.12
0.08
0.10
0.13
0.14
0.24
Total
1.00
0.99
0.99
1.01
Combustion Step, Cycles 1, 2, and 3

     After the initial recycle experiment (RC-1A), a ten-cycle series of
combustion/regeneration experiments was initiated using Arkwright coal and
Tymochtee dolomite.  Nominal operating conditions selected for the combustion
portions of each cycle are a 900°C bed temperature, 810 kPa pressure, 1.5
Ca/S mole ratio, ^17% excess combustion air, 0.91 m/sec fluidizing-gas
velocity, and a 1.07 m bed height.  Chemical and physical characteristics
of the coal and dolomite are presented in Tables A-l and A-4,  respectively.
The coal was fed to the combustor as received; the dolomite was prescreened
to -14 +30 mesh.

     Three combustion steps (2 1/2 cycles) have been completed.  During the
first combustion test, which required VL70 hr of operating time, 620 kg
of dolomite (^124 kg on a calcium basis) was sulfated.  The second and third
combustion tests were made with 270 kg and 250 kg of sorbent (^80 and 75 kg
on a calcium basis) and required approximately 100 and 95 hr of operating
time, respectively.  As experience is gained regarding sorbent losses which
can reasonably be expected per cycle, the quantity of sorbent  processed
will be reduced to shorten the operating time required to complete the full
ten-cycle test.  Operating conditions and flue-gas composition data for
representative steady-state segments of the first three combustion cycle
experiments are presented in Table 19.  Bed temperature and flue-gas data
for representative segments of the cyclic combustion experiments are plotted
in Fig. 27-29.

     Sulfur Retention.  The sulfur dioxide level in the flue gas increased
from 290 ppm in the first combustion cycle to 400 ppm and 490  ppm in
combustion cycles two and three.  Sulfur retention, based on flue-gas analysis
has correspondingly decreased from 86% in cycle one to 81% and 77% in cycles
two and three.  The sulfur retention in cycle three, which corresponds to an
emission of 0.95 Ib S02/106 Btu, meets the EPA environmental emission
standard of 1.2 Ib S02/106 Btu.

-------
                  Table 19.
Operating Conditions and Flue-Gas Composition for
Cyclic Combustion Experiments.
                             Combustor:  ANL, 6-in.-dia
                             Coal:  Arkwright, -14 mesh, 2.8 wt % S
                             Sorbent:  Cycle 1, Tymochtee dolomite,
                                       -14 +30 mesh
                                       Cycles 2 and 3, Regenerated
                                       Tymochtee dolomite
                                                       900°C
Temperature:
Pressure:  810 kPa
Excess Air: M.7%
Additive
Feed
Analysis Feed Rate
(wt %) (kg/hr)
Cycle
REC-1
REC-2
REC-3
Ca S Coal
20 - 14
29.7 4.7 13
29.7 4.6 13
.6
.3
.5
Sorbent
4.1
2.8
2.9
Fluidizing-
Ca/S Gas
Mole Velocity
Ratio3- (m/sec)
1.6
1.4
1.5
0.94
0.91
0.91
Bed
Height
(m)
1.1
0.9
0.9
Flue-Gas Analysis (avg values)
S02
(ppm)
290
400
490
NO
(ppm)
200
120
130
(ppm)
32
30
N.D.
CO C02 02 I
(PPm) (%) (%)
90 16.0 3.4
40 15.5 3.1
° 20 16.0 3.2
Sulfur
letention^
86
81
77
kRatio of unsulfated calcium in dolomite feed to sulfur in coal.
 Based on flue-gas analysis.
 No data.

-------
                                74
 o
 ^^
 a
  waX^Si!
"^75s^>»Wsf^


<&&&<






  cvz
 O
 O








S i
o
T3
j_i
3
X!
t/3
,ri — i





Analyzer discon
from d
ita logge


nee ted
r
x



             240      480      720      960
                               Time, min
1200
                                                          1440
                  1680
Eig. 27.  Bed Temperature and Flue-Gas Composition,  Segment
          of Experiment   REC-1 (REG-IK and -1L).

-------
                                        Ga.s  Concentrations
   H-
  TO
   00
o  w
Hi (D

W
O  I "1
1^1  r~3
•O  (D
n>  3
h^ *o
H- (D
B  H
(D  fB
0  rt
rt e
   H
?0 (D
   
   CO

   o
   O
   o
   CO
   H-
   rt-
   H-
   s
   en
   (D
  cw
   0
   rt
S02> ppm  CH4> ppm CO,  ppm     02,  %      C02, %    NO, ppm  Bed Temp,  °C
0       10000        -tO 0         12-50         50        20  0        200  750      1000
                                                                                                                Ul

-------
   	GAS CONCENTRATIONS	
         PH                                       BED
S02, ppm  ppj*' N CO, ppm   02, %   C02 , %  NO, ppm  TEMp> «c
H-
TO
•

£ 0
• o
o w
rti (D
X 1-3 |\J
y I-* ^^\
(D 5 ^^
^™i *o {_j
m P

rt C ^
5d (D ^}
Pd ^\
0 (B *->
1 9

/•N *^d
I If H 0
.' X ^ O
wo s.
^S f
SS1 3 oo
OS -• O
wo 0
Cd co
v_/ H-
• rt
h*
B o
0
en O
(D
H
5 ro
n 0
0
£
.1 y


7 ,
:
i
- /
- i
- \
- \

—



~ c

r
-
L^

—
-
"
r
-
"1 1
o
O
KK"1 — -
O 0
o o ro
OO OO cno cno
y i
1





-

i
\
\
y

1
?
i
?



\
6
r
p

i i



























i 1 1 i
• ' §
• .5" •.
\
r~
n
tn
JO
1.
^-4

'
r>
s
-1
o
|
r—
§
m
30





. .-r\


1 1 1 1



























1 1 1 1
o
o
N>
§
s
m
S3
O
O
O
z
z
m
o
-I
rri
o
0
o
5 .
r-
i?
m








1 1 1 1



























-/


b
\
i
\
\
\
\
\
\
\
I

i




[

c


ij
>




*




*





























!









s
l





L
P
'

s

!>

>

i
i
1
)




























1 1 i/i
l
j
I
i

i
1
L
^C
o
1
1
1
Y
\
f\
T
<>
r
6
1
0
1
Y
£
1 lYl



























1 l l





4



<




-
™
—
-


_
—
-
-
_
_
—
-
i
A-
1—
1 -
sJ ~
vl ~
r




. <
1 1
j -
~
-


-
-



-------
                                    77
     The sulfur retention capability of the dolomite would be expected to
decrease during the first few cycles after less than 100% regeneration of
the sorbent has been achieved.  For the first combustion cycle, the Ca/S
mole ratio of 1.6 is based on the total calcium content of the sorbent.
If a shrinking-core sulfation model is assumed, the sorbent from the first
combustion cycle consists of inner unreacted calcium, a layer of unregen-
erated calcium sulfate, and a surface layer of regenerated calcium oxide.
For subsequent combustion cycles, the Ca/S mole ratio of 1.5 is selectively
based on regenerated calcium oxide at the surface and less available
calcium (due to diffusional mass transport limitations) at the core of
the particle.  If this were the only factor affecting sulfur retention
during cyclic operation, however, a balance between sulfation and regen-
eration would be expected, after which there would be no further decrease
in the reactivity of the dolomite.

     Decrepitation Rate.  Analysis of samples of solids from a representative
segment of the first combustion cycle experiment, REG-IK, has been completed,
permitting evaluation of the decrepitation of Tymochtee dolomite during the
experiment.  (The analysis does not distinguish decrepitation in the feeding
operation, which might be appreciable, from decrepitation which occurs
during dolomite residence in the combustor.)  The approach taken in the
analysis was to determine the degree to which +30 mesh dolomite (^99.2% of
feed) was reduced to -30 mesh additive, elutriated from the bed, and collected
in the cyclones.  The primary cyclone material, a mixture of ash and dolomite,
was separated into +30 mesh and -30 mesh fractions which were analyzed for
both calcium and magnesium to determine the dolomitic calcium in each size
fraction.

     The results of the calculations are presented in Table 20.  The overall
calcium material balance is 103%.  Decrepitation of the Tymochtee dolomite,
defined as the amount of +30 mesh additive reduced to -30 mesh, was ^20%
for the segment REG-IK of the first combustion cycle or ^4%/hr (solids mean
residence time in the combustor was ^5 hr).  Entrainment, defined as the
percentage of calcium entering the combustor which subsequently leaves the
combustor in the flue gas, was ^25% or ^5%/hr.

     The high level of decrepitation measured during the first combustion
cycle does not agree with measurements of decrepitation of Tymochtee dolomite
reported previously.17  In the earlier study, an inventory was made of +45
mesh Tymochtee dolomite during eleven combustion experiments with an average
loss for all experiments of only 4% (solids mean residence times were
between 2 and 16 hr).  The +45 mesh fraction accounted for approximately
87 wt % of the -14 +100 mesh additive feed used in the previous investigation,
as compared with experiment REC-lK in which the +30 mesh additive accounted
for ^99% of the -14 +30 additive feed.  The previous decrepitation measure-
ments were based on balances made around the entire experiments; the current
determination is based on a steady-state material balance.  However, neither
of these differences in the determination of decrepitation seems sufficient
to account for the large difference in decrepitation values reported.

-------
                                     78
  Stream
            Table 20.  Steady-State Flow of Calcium Through the
                       ANL, o-in.-dia Combustor during combustion
                       Experiment REG-IK
                           Composition      Dolomite Calcium
 Rate
(kg/hr)
  Ca        Mg     +30 mesh  -30 mesh   Coal Calcium
(wt %)     (wt %)     (g/hr)     (g/hr)        (g/hr)
    IN

Additive
Feed

Coal
Feed
  4.08     20.0
 14.6
   OUT

Product
Overflow

Primary
Cyc, +30 mesh

Primary
Cyc , -30 mesh

Secondary
Cyclone, all
-30 mesh
  2.4
  0.4
  2.0
  0.15
 N.D.
 19.3
  8.'5
           11.3
          810
          6.7
  0.27      negl.

          TOTAL
                                 39

          810        6.7         39

        GRAND TOTAL        ^860 g/hr
 N.D.     600
         32
10.6
 4.19
72
        150
  6.1       2.22      -         5.9

          TOTAL      670      190


                    GRAND TOTAL
      negl.


       2.1


      22



       3.2

      27


^890  g/hr
  No  data.

-------
                                    79
     The most likely explanation for the higher decrepitation rate in exper-
iment REG-IK than those in previous combustion experiments at Argonne is
that experiment REC-1 was performed using a different batch of dolomite
from the supplier, C. E. Duff and Sons, Huntsville, Ohio.  Exxon has
reported similar increases in attrition of Tymochtee dolomite received in
later shipments from the same quarry.  Exxon tested (in a batch combustion
experiment) a sample of Tymochtee dolomite used in the earlier Argonne
experiments and reported a significant reduction in the attrition rate
for the sample.

     There is a discrepancy in the method of reporting attrition data by
the various contractors investigating FBC which could make the comparison
of results difficult.  In results reported by Argonne, a distinction has
been made between the amount of sorbent entrained in the flue gas and the
amount of decrepitation (i.e.3 the quantity of additive particles reduced
below a certain size).  Such a distinction is important since it has been
found that some sorbent particles that are larger than the smallest particle
size in the sorbent feed can be carried over into the cyclones.  Other
investigators of fluidized-bed combustion have reported as attrition the
amount of additive which leaves in the fluidized-bed overhead and is
recovered from the flue gas.

     Entrainment and decrepitation for an experiment can differ greatly.
For example, in earlier work at Argonne in which the average attrition of
Tymochtee dolomite was reportedly ^4% of the dolomite fed to the combustor*,
carryover of sorbent material into the cyclone was reported to be as high
as ^80% at fluidizing-gas velocities of 1.5 m/sec.

     Exxon has reported attrition rates of 20-25 wt %/hr (expressed as
percent of the bed weight lost per hour) for Tymochtee dolomite in batch
combustion tests.18  If this basis were used for reporting, the 25%
entrainment reported here for experiment REG-IK would correspond to an
"attrition" rate of ^5 wt %/hr.  However, on the basis of the solids mean
residence time of 5 hr for experiment REG-IK, the rate at which +30 mesh
additive feed material was reduced to -30 mesh material in the combustor
was ^4 wt %/hr.

     Quantitative evaluations are being made of decrepitation and
elutriation of sorbent from the combustor during the second and third
combustion cycles.  Preliminary indications are that the sorbent losses
are significantly less than the 4 to 5%/hr (based on the weight of the
bed) measured in the first combustion cycle.  Losses were ^2-3%/hr during
the second combustion cycle and VL-2% during the third combustion cycle.

Regeneration Step, Cycles 1 and 2

     In the regeneration steps, Triangle coal (for which the ash fusion
temperature is high) is being partially combusted at a system pressure of
153 kPa and a bed temperature of 1100°C.  Chemical and physical character-
istics of Triangle coal are presented in Appendix A.  The results, up to a
third regeneration step of the ten-cycle experiment, are reported.

-------
                                    80
     Extent of Regeneration of CaO.  The experimental conditions and results
for the regeneration steps (CCS-1 and CCS-2) of the first two cycles are
presented in Table 21, and the intermittent off-gas analyses are given in
Figs. 30 and 31.  In the regeneration step of the first sorbent utilization
cycle, the S02 concentration in the dry flue gas was 6.5% and the extent
of CaO regeneration was 71% (based on analyses of solid products).

     The extent of CaO regeneration in cycle two (67%)  was slightly lower than
that in the first cycle (71%).   This difference was probably due to the higher
sulfur content of the feed sulfated dolomite (10.7 wt % as compared with 9.1
wt % in cycle 1).  The S02 concentration in the dry off-gas increased from
6.5% in the first cycle to 8.6% in the second cycle.  There were two reasons
for the increase:  the fluidizing-gas velocity was lower in the second cycle
and hence the extent of S02 dilution was lower.  Secondly, the sulfur content
of the sulfated additive fed was higher in the second cycle and hence more
S02 was evolved per unit time during regeneration.

     Attrition Rate.  A first estimate of the attrition rate was obtained
by calculating the difference between the calcium in the feed and the
product streams per unit time at steady-state conditions.  The calcium
feed rate for CCS-1 was 7.11 kg/hr, and calcium was removed in the regen-
erated product at a rate of 6.94 kg/hr.  On the basis of these rates,
0.17 kg/hr or approximately 2% of the sulfated feed dolomite was attrited
and elutriated.  The attrition rate during the second cycle has been found
 (within analytical accuracy) to be negligible based on the calcium feed rate
 (5.50 kg/hr) and the calcium removal rate (5.52 kg/hr) in the regenerated
sorbent.  The low sorbent losses from attrition and elutriation can be
attributed to the short solids residence time in the regenerator reactor.

Coal Ash Buildup during Sulfation and Regeneration Steps

     Of concern in developing a regenerative fluidized-bed coal combustion
process is the extent of coal ash buildup in the fluidized bed and the
effect of ash on (1) the S02-accepting capability of the additive bed and
 (2) ash-additive reactions during regeneration.  The ash buildups during
the first-cycle sulfation and regeneration steps have been calculated
from wet-chemical analyses of the samples and are given in Table 22.  As
a basis for calculation, 100 g of unsulfated Tymochtee dolomite was used.
Some of the analytical methods used for unsulfated Tymochtee dolomite were
obtained from the literature.19 (Analysis of the virgin dolomite is
presently being performed at ANL, and ash builup will be recalculated on
the basis of these results.)  An ash buildup of 4-5% during sulfation was
calculated, based on the enrichment of Si, Fe, and Al.

     The concentrations of ash constituents in sulfated dolomite are
compared to the concentrations in Greer limestone in Table 23.  As expected,
the extent of ash buildup was smaller in the experiments with Tymochtee
dolomite (in which a relatively low ash Arkwright coal was used).

-------
                Table 21.  Experimental Conditions and Results for Regenerative Segments
                           of the Regeneration Step of the First Two Utilization Cycles.
                           Nominal Fluidized-Bed Height:  46 cm
                           Reactor ID:  10.8 cm
                           Pressure:  153 kPa
                           Temperature:  1100°C
                           Coal:  Triangle (0.98 wt %• S), Ash fusion temperature under
                                  reducing conditions:  1390°C (initital deformation)
                           Additive:  -14 +30 mesh, sulfated Tymochtee dolomite



Fluidizing- Solids

Cycle Exp.
No.
1 CCS-1
2 CCS-2
Gas
Velocity
(m/sec)
1.43
1.26
Residence
Time
(min)
7
7.5


Reducing Gas
Concentration
in Off-Gas
(
2
3
y\
i°)
.8
.0
Sulfur in
Feed
Sulfated
Sorbent
(%)
9.1
10.7


CaO
Regeneration
(%)a/(%)
73/71
67/67




Major

S02
6.5
8.6
in

0
0






Sulfur Compounds
Dry
H2S
.04
.02
Off-Gas
COS
0.06
0.1
(%)
CS2
0.04
0.1
                                                                                                               00
Based on off-gas analysis.
Based on chemical analysis of dolomite samples.

-------
                                 82

-------
            S0, %
  H-
  CW
O fD
i-i CL

fd H
!* (D
t3 g
(D "d
i-i (D
H- H
3 fu
(0 ft-
3 C
rf H
  (D
O
O
CL

o

CO

O
o
   (B
   rt
   l-|-

   §
   co
   O
   Hi
   H)


   O
   (11
   CO
- Above Bed Gas Concentrations	

,  %      CO, %       02,  %      C02, %   NO, ppm Bed Temp, °C
40
                                                       19  0
                                                  1600 1000
                                                                                           °
                                                                                         1200
                                                                                                            00

-------
                                     84
      Table 22.  Calculated Ash Buildup During Sulfation and Regeneration
                 of Tymochtee Dolomite Based on Silicon, Iron, and
                 Aluminum Concentration Changes.

                 (Mass basis:  100 g unsulfated Tymochtee dolomite)

                                  Equivalent Ash in    Equivalent Ash
                                 Grams per 100 Grams       Content
                                   Dolomite Basis            (%)

Si
Fe
Al

Si
Fe
Al
Sulfated Dolomite
4.6 3.7
6.8 5.6
5.5 4.5
Regenerated Dolomite
4.6 4.8
6.7 7.1
4.4 4.6
      Table 23.  Concentrations of Ash Constituents in Sulfated
                 Dolomite and Sulfated Greer Limestone.

                          Tymochtee Dolomite (%)     Sulfated Greer
                         Unsulfated      Sulfated    Limestone (%)
Si02
A1203
Fe203
4.96
1.64
0.41
5.73
4.72
2.91
16.63
5.75
2.53
       PRESSURIZED, FLUIDIZED-BED COMBUSTION:  BENCH-SCALE STUDIES
[W. Swift, (Principal Investigator), H. Lautermilch, F.  Nunes, S. Smith,
J. Stockbar]

      Experiments were performed in the ANL, 6-in.-dia,  pressurized,
 fluidized-bed combustor as part of a continuing program to evaluate and
 demonstrate the feasibility and potential of fluidized-bed combustion at
 elevated pressures (<1014 kPa).  In previous pressurized, fluidized-bed
 combustion studies, the effects of bed temperature (788-95$°C), gas
 velocity (0.6-1.5 m/sec), and Ca/S mole ratio (1-3) on the sulfur retention,
 NO level in the flue gas, combustion efficiency, and particulate loading
 in the flue gas were evaluated.  Tymochtee dolomite sorbent, Arkwright
 coal (2.8 wt % S) and a pressure of 811 kPa were used.   The effects of
 pressure, different sorbents  (limestone vs. dolomite),  precalcination of
 sorbent, amount of excess air, and burning various ranks of coal have
 also been reported. '17

-------
                                     85
     The objectives of the experiments reported here were:  to evaluate the
effects of coal and additive particle size on sulfur retention, nitrogen
oxide flue-gas levels, and combustion efficiency; to evaluate the sulfur
retention capability of lignite ash, which has a high calcium content; and
to evaluate the relative effects of gas velocity and residence time on the
decrepitation of the sorbent.

Materials

     Coal.  The principal coal tested was a high caking, high-volatile
bituminous, Pittsburgh seam coal obtained from the Consolidation Coal
Company Arkwright mine.  As received, the coal contained ^2.8 wt % sulfur,
7.7 wt % ash, and 2.9 wt % moisture, had a heating value of 7610 kcal/kg,
and had an average particle size of 320 urn.  Lignite coal from the Consoli-
dation Coal Company Glenharold mine in North Dakota was also tested.  As
received, the lignite contained 0.53 wt % S, 6.1 wt % ash, and 30.9 wt %
moisture, had a heating value of  4240 kcal/kg, and had an average particle
size of 350 pm.  The coals were fed to the combustor either as received or
after screening to the desired range of particle size for studies on the
effect of particle size.  Chemical and physical characteristics of the
Arkwright and Glenharold coals are presented in Tables A-l and A-3, Appendix
A.

     Sorbent.  Tymochtee dolomite (viO wt % CaC03 and ^40 wt % MgC03) obtained
from C. E. Duff and Sons, Huntsville, Ohio was used in all combustion experi-
ments reported here.  The dolomite was air dried and screened to the desired
range of particle size prior to its use in the combustor.  Data on the
chemical characteristics of the Tymochtee dolomite are presented in Table A-4,
Appendix A.

Bench-Scale Equipment

     The experimental equipment and instrumentation consist of a 6-in.-dia,
fluidized-bed combustor that can be operated at pressures up to 1014 kPa, a
compressor to provide fluidizing-combustion air, a preheater for the fluid-
izing-combustion air, peripheral-sealed rotary feeders for metering solids
into an air stream fed into the combustor, two cyclone separators and two
filters in series for solids removal from the flue gas, associated heating
and cooling arrangements and controls, and temperature- and pressure-sensing  .
and display devices.  A simplified schematic flowsheet of the combustion
equipment is presented in Fig. 32.

     Details of the bench-scale combustor are presented in Fig. 33.  The
reactor vessel consists of a 6-in.-dia, Schedule 40 pipe (Type 316 SS),
approximately 11 ft long.  The reactor is centrally contained inside a 9-ft
section of 12-in.-dia, Schedule 10 pipe (Type 304 SS).   A bubble-type gas
distributor is flanged to the bottom of the inner vessel.  Fluidizing-air
inlets, thermocouples for monitoring bed temperatures,  solids feed lines,
and solids removal lines are accommodated by the bubble cap gas distributor.
The coal and additive feed lines extend 2 in. above the top surface of the
distributor plate and are angled 20° from the vertical.  A constant bed height

-------
                                     86
                                                   TO GAS ANALYSIS
                                                       SYSTEM
 AIR
                                              uu
                                                                  VENTILATION
PRESSURE
CONTROL
 VALVE
         EXHAUST
                                                SINTERED
                                                STAINLESS
                                                 STEEL
                                                FILTERS
                                             SECONDARY
                                             CYCLONE
                                           PRIMARY
                                           CYCLONE
    PREHEATER
     Fig. 32.  Simplified Equipment Flowsheet of Bench-Scale  Fluidized-
               Bed Combustor and Associated Equipment.
is maintained in the combustor by use of a  36-in.-high  standpipe.
The 6-in.-dia pipe is alternately wrapped with resistance-type heating
elements and cooling coils onto which a layer of heat-conducting copper
and then an overlay of oxidation-resistant  stainless  steel  have been
applied.  Additional cooling capacity is provided by  three  internal,
hairpin-shaped coils that extend down from  the flanged  top  of the combustor
to within 12 in. of the top surface of the  gas distributor.   The coolant
is water entrained in air.

     Fluidizing-combustion air is supplied  by a 75-hp,  screw-type compressor
capable of delivering 100 cfm at 150 psig.  The air can be  heated to
approximately 540°C in a 6-in.-dia, 10-ft-tall preheater containing eight
2700-W, clamshell-type heaters.

     Coal and either dolomite or limestone  are pneumatically fed from
hoppers to the combustor, using two  6-in.-dia rotary valve feeders.  The
feeders and hoppers are mounted on platform-type scales.

     The flue gas is sampled continuously and is analyzed for the components
of primary importance.  Nitrogen oxide and  total NOX  are analyzed using
a chemiluminescent analyzer; sulfur dioxide, methane, carbon monoxide, and
carbon dioxide determinations are made using infrared analyzers; oxygen is
monitored using a paramagnetic analyzer; and total hydrocarbons are analyzed
by flame ionization.  Prior to and during each experiment,  the response of

-------
                               87
 60 AND 85 in FREEBOARD
 THERMOCOUPLES
 INTERNAL COOLING
   COIL LEADS
PURGE GAS OUTLET
HEATER CONTROL
THERMOCOUPLES
            RUPTURE DISK

               FLUE GAS TO
           CYCLONE AND FILTERS
                                             EXPANSION BELLOWS
                                             RUPTURE DISK
                                             12-in. JACKET
 SHELL PURGE
 GAS  INLET
EXTERNAL COOLING
COIL LEADS-
        36 OR 48 in.
   SOLIDS OVERFLOW
   6, 12,AND 44 in.
 BED THERMOCOUPLE WELL
                                        EXTERNAL COOLING COIL
                                       (WRAPPED ON 6-in:DIA WALL)
       CALROD TYPE HEATER

       INTERNAL COOLING COILS


       BUBBLE CAP DISTRIBUTOR
          ELECTRICAL HEATER
              LEADS
                                         PLIBRICO FILLED VOLUME
       SOLIDS FEED LINES
FLUIDIZING AIR
 Fig.  33.  Detail Drawing of 6-in.-Dia,  Pressurized, Fluidized-
          Bed Combustor.

-------
                                    88


each analytical instrument is checked,using standard gas mixtures of flue-
gas constituents in nitrogen.  Batch samples of flue gas can be taken and
analyzed for constituents of secondary importance.

     The combustion system is equipped with a Hewlitt-Packard 2010C data
acquisition system to monitor and record the temperature, pressure, gas
flow, and flue-gas concentration for subsequent data handling and analysis.

Experimental Procedure

     Although the experimental procedure was subject to minor variations,
it was basically as follows:  a preweighed amount (^15 kg)  of either
partially sulfated sorbent from a previous experiment or fresh unsulfated
sorbent was charged to the reactor to provide an initial bed of material.
The starting bed temperature was then raised to about 430°C by passing
fluidizing air (preheated to 430-480°C) through the combustor and simul-
taneously employing the resistance heaters on the combustor wall.  Once
the bed temperature reached 430°C, the system was brought to the desired
operating pressure, and coal (entrained in a transport air stream) was
injected into the bed.  To prevent carbon accumulation in the fluidized
bed during startup, coal was initially injected in small amounts inter-
mittently until a rapidly increasing temperature and a changing flue-gas
composition confirmed ignition and sustained combustion.  Continuous
injection of coal was then initiated, and the bed temperature was raised
to a selected combustion temperature.  The desired temperature was maintained
by the use of external and internal cooling coils.

     Injection of the sulfur-accepting sorbent was begun when the bed
reached operating temperatures.  The air, coal, and sorbent feed rates
were adjusted to give a specified mole ratio of calcium in the sorbent to
sulfur in the coal, a specified superficial gas velocity, and a specified
level of oxygen in the flue gas leaving the combustor.  Sulfated sorbent
was removed from the combustor by means of a standpipe to maintain a
constant fluidized-bed level.

Replicate of VAR-Series Experiment

     After the combustion and regeneration systems were separated (discussed
in a following section of this report), an additional replicate experiment
(VAR-6-3R) of one of the VAR-series experiments (VAR-6) was made to check
the operation of the combustor and analytical instrumentation and to verify
the validity of comparing results of current experiments with those of
experiments performed previously.  Operating conditions for experiment
VAR-6-3R duplicated the operating conditions of replicate experiments VAR-6,
VAR-6-R, and VAR-6-2R, the results of which have been reported previously.17
Operating conditions and results for the four experiments are given in
Table 24.

     The results of experiment VAR-6-3R are very encouraging.  With the
exception of the concentration of NO in the flue gas, the results of VAR-6-3R
agree very well with the previously performed experiments in the VAR series.

-------
                                     89
     Table 24.  Operating Conditions and Flue-Gas Compositions
                for VAR-6 Replicate Experiments.

                Combustor:  ANL, 6-in.-dia
                Bed Temperature:  850°C (1544°F)
                System Pressure:  810 kPa (8 atm)
                Bed Height:  0.9 m (3 ft)
                Excess Air:  ^17%
                Additive:  Tymochtee dolomite
                Coal:  Arkwright (2.8 wt % sulfur)
Exp . No .
VAR-6
VAR-6-R
VAR-6-2R
VAR-6-3R
Feed
Coal
13.5
13.3
13.3
13.4
Rate (kg/hr)
Dolomite
4.7
4.8
4.7
4.9
Ca/S
Mole
Ratio
2.0
2.0
2.0
2.1
Fluidizing-
Gas Dry Flue-Gas Compositions
Velocity
(m/sec)
1.10
1.10
1.07
0.97
S02
(ppm)
170
210
190
170
02
(%)
3.0
2.9
3.0
3.2
C02
(%)
16
16
16
16
NO
(ppm)
190
180
160
135
CO
(ppm)
32
33
33
20
     A combustion efficiency of ^94% and an additive utilization of ^54%
were calculated for experiment VAR-6-3R on the basis of analyses of solid
samples from the experiment.  An average combustion efficiency of 94% was
reported previously for the other VAR-6 replicate experiments.17  The average
additive utilization reported for the other VAR-6 replicates was 46.5% ,
somewhat lower than in VAR-6-3R.

     The generally good reproducibility of results obtained in experiment
VAR-6-3R helps to establish the VAR-series of experiments as a basis for
comparison for future experiments in the ANL 6-in.-dia combustor.

Effect of Coal and Sorbent Particle Size on Sulfur Retention, Nitrogen
Oxide Level in the Flue Gas, and Combustion Efficiency

     Four fluidized-bed combustion (FBC) experiments (PSI-series) were
completed to measure the effects of coal and sorbent particle sizes on
combustor response variables such as sulfur retention and NOX emissions.
The experiments were made in a 22 factorial design at two levels each of
coal and additive mean particle sizes.  Arkwright coal with mass-mean
particle diameter levels of VL50 ym (-50 mesh) and ^640 ym (+50 mesh) and
Tymochtee dolomite with mass-mean particle diameter levels of <370 ym
(-30 +50 mesh) and ^740 ym (-14 +30 mesh) were used in the series of
experiments.

     The as-received Arkwright coal was sieved at a 50-mesh breakpoint,
which resulted in approximately a 50-50 split by weight and a factor of
^4 difference in the mass-mean diameters of the two fractions.  The Tymochtee
dolomite size ranges also resulted in a 50-50 split by weight of the as-
received material and a factor of ^2 difference in the mass-mean diameters

-------
                                     90
of  the two fractions.  The -50 mesh particles were removed from the finer
fraction of dolomite to reduce elutriation of bed material from the combustor.
Sieve analyses of the coal and additive feed materials are given in Tables
25 and 26, respectively.
     Table 25.   Sieve Analyses of Arkwright Coal Size Fractions Used
                in PSI-Series Combustion Experiments.
Coarse
U.S. Sieve No.
+14
-14 +20
-20 +30
-30 +45
-45 +50
-50 +80
-80

Mass-Mean Diameter, ym
Surface-Mean Diameter,
Material
% on Sieve
0.0
15.2
37.3
37.2
6.4
3.8
0.0
99.9
^640
ym ^560
Fine
U.S. Sieve No
-30 +45
-45 +50
-50 +80
-80 +170
-170+230
-230+325
-325



Material
% on Sieve
2.4
3.8
27.1
34.0
10.0
10.0
12.6
99.9
^150
^ 78
     Table 26.   Sieve Analyses of Tymochtee Dolomite Size Fractions
                Used in PSI-Series Combustion Experiments.
Coarse
U.S. Sieve No.
+14
-14 +20
-20 +25
-25 +30
-30 +35
-35 +45
-45

Material
% on Sieve
0.0
28.7
18.6
18.2
12.1
19.6
2.8
100.0
Fine
U.S. Sieve No.
-20 +30
-30 +35
-35 +45
-45 +50
-50 +60
-60 +80
-80

Mass-Mean Diameter, ym ^740
Surface-Mean Diameter
, ym ^620

Material
% on Sieve
1.5
2.0
51.6
23.0
11.7
7.0
3.2
100.0
^370
^320

-------
                                     91


     The nominal operating conditions chosen for the series of experiments
were a bed temperature of 840°C, 810 kPa pressure, ^17% excess combustion
air (3% 02 in dry flue gas), 1.07 m/sec fluidizing-gas velocity, and 0.9 m
fluidized-bed height.  However, the fluidizing-gas velocity is not a
directly controlled operating variable.   Rather, the conditions of coal
feed rate (in this case, 12.8 kg/hr) and oxygen level in the flue gas (3%)
are specified as operator-controlled variables.  The design velocity of
1.07 m/sec is derived theoretically from the controlled variables (at 100%
combustion efficiency).  The value of 1.07 m/sec was selected for this
series of experiments to prevent elutriation of large amounts of bed
material, particularly in the experiments with the finer size fractions of
dolomite.  The actual operating conditions and flue-gas compositions for
the four combustion experiments are summarized in Table 27.  The bed
temperature and flue-gas analysis data for the four experiments are plotted
in Figs. B-l to B-4, Appendix B.

     The level of SC>2 in the flue gas ranged from a low of 160 ppm in
experiment PSI-3 (fine coal, fine additive) to 240 ppm in experiment PSI-4
(fine coal, coarse additive).  These S02 levels correspond to sulfur
retentions of ^93 and ^89%, respectively.  The observed levels of S02 for
the four experiments (Table 27) indicate slight increases in sulfur retention
when the additive particle mass-mean diameter is reduced from 740 urn to
370 ym.  Sulfur retentions were 92 and 93% with the finer dolomite fraction
(PSI-2 and -3) as compared with 90 and 89% with the coarser fraction (PSI-1R
and -4).

     In terms of the observed S02 levels for the PSI-series of combustion
experiments, the effects of increasing the additive particle size at the
low and high levels of coal particle size were +80 ppm and +30 ppm,
respectively.  This represents an average effect of increasing the S02 level
in the off-gas by 55 ppm (an average percent increase of 33%).  Similarly,
the effects on S02 level of increasing the coal particle size at the low
and high levels of additive particle size were +20 and -30 ppm, respectively.
This represents an average effect of decreasing the S02 level in the off-
gas by only 5 ppm, an insignificant amount.

     With the exception of the S02 level of 240 ppm for experiment PSI-4
(fine coal, coarse additive), the recorded S02 levels for the PSI experiments
are considerably lower than the S02 levels that would be predicted using
the correlation of S02 levels based on the VAR-series of combustion experi-
ments.17  The Arkwright coal used in the VAR-series experiments had a mass-
mean diameter of 320 ym (surface-mean diameter of 120 ym) and the Tymochtee
dolomite a mass-mean diameter of 750 ym (surface-mean diameter of 560 ym).
Thus, in terms of particle size, experiment PSI-4 corresponds most closely
with the VAR-series conditions (on the basis of surface-mean diameters).
The remaining PSI-series experiments were performed using finer additive
and/or coarser coal as compared with the VAR-series experiments.  This is
in keeping with the effect indicated above that with small additive particle
size the S02 levels in the off-gas are lower.  Although not demonstrated by
the PSI-series experiments, there is (in the comparison with the VAR-series
experiments) an indication that larger coal particle sizes may also reduce
the S02 level in the flue gas.

-------
                 Table 27.
Operating Conditions and Flue-Gas Analyses for PSI-Series
of Combustion Experiments.
                            Combustor:  ANL, 6-in.-dia
                            Bed Temp:  840°C
                            Pressure:  810 kPa
                              Fluidized-Bed Height:  0.9 m
                              Excess Combustion Air:
Arkwright Coal

Exp.
No.
PSI-1R

PSI-2

PSI-3

PSI-4


dp3
(ym)
640

640

150

150


Feed Rate
(kg/hr)
13.9

12.6

12.3

13.2

Tymochtee Dolomite

dma
(ym)
740

370

370

740


Feed Rate
(kg/hr)
3.3

3.1

2.8

3.2

Ca/S
Mole
Ratio
1.3

1.4

1.3

1.4

Gas
Velocity
(m/sec)
0.94

0.85

0.73

0.82

Avg Flue^Gas Composition, Dry Basis
02
(%)
2.8

3.1

3.1

3.0

S02
(ppm)
210
(90%)°
180
(92%)c
160
(93%)°
240
(89%)c
NO
(ppm)
120

120

130

150

NOX
(ppm)
160

170

180

210

CO
(ppm)
*b

A

62

50

fMass-mean particle diameter.
 Analyzer inoperative.
 Sulfur retention.

-------
                                    93


     In combustion experiments (SA-series) made previously at ANL at
atmospheric pressure,20 no significant effect of additive particle size on
sulfur retention was observed.  The earlier experiments were done with
Illinois No. 6 coal and limestone No.  1359 having average particle sizes
of 25 and 100 urn.  Since the observed effect of additive particle size during
the PSI-series of combustion experiments was quite small (an average difference
of only 55 ppm SC>2), it is not unreasonable to expect that the effect at the
even finer sizes of additive used in the SA-series would be insignificant.

     The levels of NO were quite low for all four combustion experiments,
ranging from 120 to 150 ppm.  Thus, particle size does not appear to affect
NO emissions significantly.

     By use of a recently installed, on-line chemiluminescence analyzer, it
was also possible to obtain values for total NOX emissions during the four
combustion experiments.  Values for NOX (see Table 27) ranged from 160 to
210 ppm, indicating that N02 levels (NOX level - NO level) were ^40 to
^60 ppm.  These values of N02 levels are considerably higher than the anti-
cipated levels of 5 to 10 ppm.

     Combustion efficiency for the four experiments ranged from 89 to 93% ,
with no consistent effect of either coal or additive particle size indicated
(see Table 28).  For the larger coal particles and the smaller additive
particles, however, the combustion efficiency was considerably lower (89%)
than in the other three experiments.
     Table 28.  Sulfur Retention, Combustion Efficiency, and
                Sorbent Utilization for PSI-Series of Experiments.
Response Variable

Exp.
No.
PSI-1R
PSI-2
PSI-3
PSI-4
Mass-Mean Particle
Diameter (ym)
Coal Dolomite
640 740
640 370
150 370
150 740
Ca/S
Mole
Ratio
1.3
1.4
1.3
1.4
Sulfur a
Retention
(%)
90
92
93
90
Combustion
Efficiency
(%)
94
89
93
92
Sorbent ,
Utilization
(%)
58(65)°
71(59)
65(66)
54(58)
 Based on level of sulfur dioxide in flue gas.
 Utilization based on chemical analysis of sulfated sorbent.  Utilization (%) =
 [(wt % S) (40/32) (100)]/(wt % Ca).
Utilization calculated from sulfur retention and combustion efficiency.
 Utilization (%) = [I/(Ca/S Mole Ratio)][Sulfur Retention (%) + Combustion
 Efficiency (%) - 100%].

-------
                                    94


     Sorbent utilizations for final bed material and bed overflow material
are also given in Table 28.  The utilizations based on chemical analyses of
these sulfated sorbents are consistent (with the exception of experiment
PSI-2) with utilization values predicted on the basis of sulfur retention
(derived from sulfur dioxide flue-gas analysis), combustion efficiency
(derived from carbon balance), and Ca/S mole feed ratio.

     One aspect of the data reported in Table 27 that is incongruous with
the design experimental conditions (and with past experience with the
combustor) is the low gas velocities in the bed reported for the four
combustion experiments.  As indicated above, the design operating conditions
for this series of experiments (consistent with a coal feed rate of 12.8
kg/hr and 17% excess combustion air) was a gas velocity of 1.07 m/sec
(3.5 ft/sec).  Gas velocities calculated from metered air flows ranged from
0.73 m/sec for experiment PSI-3 to 0.94 m/sec for experiment PSI-1R.
These calculated velocities ranged from 'WO to ^80%, respectively, of the
expected velocities of .1.02 to 1.17 m/sec (calculation based on air required
for coal combustion).

     A possible explanation for this discrepancy in the data pertains to
the rotary valve coal feeder.  The rotary valve has peripheral seals, and
an external air pressure is applied to them that is approximately equivalent
to the system pressure.  During maintenance of the valve after completion
of the experiments, it was found that the peripheral seals were leaking
badly.  Possibly, a large supplemental air flow (unmetered) leaked through
the peripheral seals of the rotary valve into the coal transport air
stream.  Thus, actual velocities may have been near the expected velocities
of 1.02 to 1.17 m/sec.  A rotameter has been installed to measure the air
inleakage and to maintain it at a negligible value.

Combustion of Lignite in a Fluidized Bed of Alumina

     Combustion experiments LIG-2D and LIG-2-R with Glenharold lignite
were made to duplicate all operating conditions of a previous combustion
experiment, LIG-1, except that combustion in LIG-2D and LIG-2-R was carried
out in a fluidized bed of alumina (chemical and physical characteristics
of the alumina are given in Table A-5, Appendix A).  Combustion in LIG-1
had been carried out in a fluidized bed of Tymochtee dolomite as part of
a study investigating the burning of various ranks of coal in the com-
bustor.17   The Ca/S mole ratio for LIG-1, based on the calcium in the
dolomite to sulfur in the lignite, was 1.1.  Sulfur retention for that experi-
ment was reported to be 85%, which was excellent for such a low Ca/S mole ratio.

     It has since been proposed that when lignite is burned, sulfur is
retained in the ash due to the relatively high calcium content of most
lignite coals, even at combustion temperatures as high as 1200°C.    If
the coal calcium content were to be included in the Ca/S ratio for experi-
ment LIG-1, the potentially effective Ca/S mole ratio for that experiment
would be 3.0.

-------
                                    95


     Operating conditions and results for experiments LIG-2D and LIG-Z-R
are given in Table 29.  Bed temperatures and flue-gas compositions for the
two experiments are plotted in Figs. B-5 and B-6, Appendix B.  Sulfur
retentions (calculated on the basis of flue-gas analyses) for experiments
LIG-2D and LIG-2-R were 89 and 86%, respectively.  These values compare
extremely well with the value of 85% calculated for experiment LIG-1 made
with a dolomite bed.  These experiments confirm the premise that sulfur
is retained by the ash during the combustion of lignite.

Effect of Fluidizing-Gas Velocity on Decrepitation Rate

     To test the effect of gas velocity (hence solids mean residence time)
on the attrition and entrainment rates observed during experiment REG-IK,
an experiment (VEL-1) was performed duplicating the conditions of REG-IK
except that a lower fluidizing-gas velocity (0.6 m/sec) was used.  The
experimental conditions and flue-gas analyses for both experiments are
presented in Table 30.  The solids mean residence time in the combustor
for experiment VEL-1 was 'W hr as compared with ^5 hr for experiment REG-IK.

     Decrepitation of Tymochtee dolomite during experiment VEL-1 was analyzed
in the same manner as was used in experiment REG-IK.  The data are summarized
in Table 31.   The steady-state calcium balance was 90%, which is acceptable
within experimental error.  Reduction of +30 mesh additive to -30 mesh
additive was ^35% which, on the basis of the 7-hr residence time in the
combustor, results in a decrepitation rate of ^5 wt %/hr.  This is similar
to the ^4 wt %/hr decrepitation rate of experiment REG-IK.

     The significance of the results is that the rate of decrepitation (4-5 wt
%/hr) was unaffected by the change in gas velocity from 0.94 to 0.60 m/sec.
Due to the increased residence time of the additive in the combustor when
the gas velocity was decreased, however, the percentage of additive feed
decrepitated increased from ^20% to ^35%.

     The second significant observation is that, although the percentage of
additive feed decrepitated increased as the velocity was reduced, there was
slightly less entrainment (or "attrition") of material from the combustor—
from 25% (^5 wt %/hr) to ^22% (^3 wt %/hr).  Thus, in experiment VEL-1,
the rate of decrepitation was actually greater than the rate at which
sorbent was carried over into the cyclones with the flue gas.  It would be
expected, therefore, that if the gas velocity had been decreased at a
constant residence time (by lowering the bed height), the decrease in
entrainment would have been greater since there would have been no increase
in the amount decrepitated.

Binary Salt of Magnesium and Calcium Sulfate

     In thermogravimetric experiments performed at Argonne National Laboratory,
the presence of Mg3Ca(SOit)i+ (formula not definitely established) was detected
by X-ray diffraction analysis in samples of dolomite simultaneously sulfated
and half-calcined.22  Samples of sulfated dolomite from earlier bench-scale
combustion experiments performed at 843, 899, and 954°C (1550, 1650, and
1750°F) were subsequently analyzed to determine if the binary salt might be
forming in the bench-scale combustor.   X-ray diffraction, however, failed to
detect any binary salt formation in any of the samples.

-------
            Table 29..  Operating Conditions and Flue-Gas Analyses  for
                       Combustion Experiments LIG-2D and L1G-2-R.

                        Combustor:  ANL, 6-in.-dia   Excess Air:   ^17%
                        Bed Temp:  840°C             Coal:  Glenharold
                        Pressure:  810 kPa                  lignite  (0.56 wt
                        Bed Height:  0.9 m           Bed:  30 mesh alumina

Exp.
No.
LIG-2D
LIG-2-R
Coal Feed
Rate
(kg/hr)
24.2
24.5
Fluid iz ing-Gas
Velocity
(m/sec)
0.91
1.00
S02
(ppm)
80
102
Flue- Gas
NO
(ppm)
90
140
NOX

150
170
Analysis
CO
(ppm)
80
35

C02
(%)
19
b

°2

3.1
3.8
Sulfur
Retention3
(%)
89
86
.Based on flue-gas analysis.
 No data.

-------
                            97
   Table 30.  Operating Conditions and Flue-Gas Compositions
              for Experiments REG-IK and VEL-1 to Test the
              Effect of Gas Velocity on Decrepitation.

              Combustor:  ANL, 6-in.-dia
              Coal:  Arkwright, -14 mesh
              Additive:  Tymochtee dolomite, -14 +30 mesh
              Temperature:  900°C
              Pressure:  810 kPa
              Excess Air:
Operating Conditions
Coal feed rate, kg/hr
Additive feed rate, kg/hr
Ca/S Mole ratio
Gas velocity, m/sec
Bed height, m
Run duration, hr
REC-1K
14.6
4.1
1.6
0.94
1.1
11.2
VEL-1
9.2
2.6
1.6
0.60
0.82
14.8
    Flue-Gas Analysis

S02, ppm                                   290        130
NO, ppm                                    200        150

CH4, ppm                                    32         20
CO, ppm                                     90         20

C02, %                                      16         16
02, %                                        3.4        3.5


    Sulfur Retention, %                     86         94

-------
                         98
Table 31.  Analysis of Decrepitation in Experiment VEL-1,

                    Calcium In, kg/hr

                    Dolomite Feed
                      +30 mesh    0.55
                      -30 mesh    0.04
                           TOTAL  0.59

                 Calcium Out, kg/hr
                  Product Overflow
                      +30 mesh    0.33
                      -30 mesh    0.10
                   Primary Cyclone
                      -30 mesh    0.10
                  Secondary Cyclone
                      —30 mesh    neg.
                           TOTAL  0.53
Calcium Balance:  90%
Recovery of +30 mesh additive material:  60% ('x-AO % loss)
Increase in -30 mesh as percentage of +30 mesh feed: ^30%
Entrainment:  18%

-------
                                    99
                           SYNTHETIC S02 SORBENTS
R.  Snyder  (Principal  Investigator), I. Wilson

Introduction

     In fluidized-bed coal combustors, naturally occurring limestone
and dolomite have been the only materials investigated extensively as
additives for removing sulfur dioxide from the combustion gas.  This is
primarily due to the low cost of these calcium-bearing materials.  However,
limestone and dolomite have some disadvantages.  For example, decrepitation
and agglomeration can occur in the regenerator.  Also, during cyclic
sulfation and regeneration, the reactivity of dolomite or limestone additive
with S02 in the combustor may decrease as the number of cycles increases.
Finally, if it should be determined that regeneration of these sulfated
additives is not economical, large quantities of dolomite and limestone
must be mined and disposed of.  Due to these potential disadvantages,
synthetic S02-sorbent materials are being studied as alternatives to
dolomite and limestone.  Methods of supporting a metal oxide in a highly
dispersed state in a matrix of high-strength inert material are being
studied, and the sorbents prepared are being evaluated.

      Various .sorbents (metal oxides in a-Al2C>3) have been prepared and
tested with a thermal gravimetric analyzer (TGA) for their ability to
capture S02.  All kinetic studies for both the sulfation and the regen-
eration of the synthetic sorbents were performed using a TGA apparatus.
The operation and a schematic diagram of this apparatus are given in the
preceding annual report.

     Since CaO is presently the most promising metal oxide, the kinetics
of the reaction of CaO in a-Al203 with S02 and 02 was extensively studied
under a variety of conditions.  Also,  the regeneration kinetics of the
sulfated sorbent, CaSO^. in a-A!203, was determined for various reducing
gases.  Cyclic sulfation-regeneration experiments were performed, and
optimization of the support material was studied.  This involved determining
the reactivity during sulfation and attrition resistance of supports having
various pore size distributions.

Preparation of Synthetic Sorbents

     A method of preparing metal oxides in a-Al203 support at various oxide
concentrations has been developed and is described below.

     The alumina support pellets are placed in an aqueous metal nitrate
solution, refluxed for 8 hr, and then cooled to 25°C for 4 hr.  The pellets
are removed from the solution and slowly heated to a heat-treating (H.T.)
temperature of either 800 or 1100°C, where they are held for one hour.
Heating must proceed very slowly at the temperatures where phase changes
in the various chemical compounds occur.   For example, in the preparation
of CaO in a-A!203, heating is allowed to proceed very slowly near 100°C,
132°C, and 561°C since at these temperatures there occur, respectively,
the evolution of H20, decomposition of Ca(N03)2'4H20, and decomposition
of Ca(N03)2 to CaO.

-------
                                   100
     The quantity of metal oxide in the support depends on the metal
nitrate concentration in the aqueous solution.  This is illustrated in
Fig. 34, which shows the percent CaO in the support for various concen-
trations of Ca(N03)2-4H20 in aqueous solution.

     The nature of the calcium compounds formed in the support depends
on the heat-treatment temperature.  X-ray diffraction results indicate that
when the sorbent was heat-treated at 800°C for 4 hr, CaO, CaO-6Al20'3,
and 3CaO'5Al203 were the calcium products formed.  Heat-treating the
sorbent at 1100°C produced CaO-Al203 and CaO'2Al203.

Sulfation Studies

     The sulfation reaction (reaction of CaO in a-A!203 with S02 and 02)
was performed under a variety of conditions.  The effects on the reaction
kinetics of the synthetic gas composition (S02, 02, H20), temperature,
and initial calcium oxide concentration in the sorbent were determined.
The effect of the heat-treatment temperature that had been used in sorbent
preparation on the kinetics was also studied.  The sulfation rates of
various metal oxides (Na20, K20, SrO, BaO) in a-A!203 were compared with
those for CaO in a-A!203.  Finally, the sulfation rate for CaO in a-A!203
was compared with that for Tymochtee dolomite.
            0
             0
 2       3       4       56
Ca(N03)2-4H20/H20 WEIGHT RATIO
      Fig.  34.   CaO Concentration in a-Al2Os  as  a Function of Calcium
                Nitrate Concentration in Reactant Solution from which
                Prepared.

-------
                                    101

                                                                        *
     Effect of Gas Composition on Sulfation.Rate.  A 6.6% CaO in a-A!203
sorbent was used to detemrine the effect of the various conditions on the
sulfation rate.  Sorbents heat-treated (H.T.) at both 800°C and 1100°C
were studied.

     Sulfation experiments with a S02 gas concentration range of from 0.05
to 3% were performed at 900°C on 800°C H.T. sorbent.  Preparation of the
gas mixtures for these reactions required blending of 02, S02, and N2 to
the various specified concentrations.  Mass spectrometric analysis were
performed on two samples of the blended gas mixtures to confirm the
concentration of each constituent.  For gas mixtures prepared to contain
(1) 0.3% S02 and 5% 02 and (2) 0.05% S02, 5% 02, with the balance N2,
the mass spectrometric results were (1) 0.3% S02, 5.3% 02 and (2) 0.047%
S02, 5.2% 02, respectively.

     The sulfation results for 800°C H.T. sorbent are shown in Fig. 35,
where the percent conversion of CaO in the a-alumina pellets to CaSO^ is
given as a function of time and S02 concentration in the gas stream.
Samples were analyzed by wet chemical analysis to determine the extent
of reaction in each experiment (discussed below).  The time required for
the reaction to go to completion was 4 to 10 hr, depending on the S02
concentration in the gas stream.  The residence times for the S02-sorbent
in commercial fluidized beds will probably be several hours; therefore,
the rate of sulfation appears to be adequate.

     The order of reaction as a function of S02 concentration in the gas
mixture was found to be 0.7.   This is in good agreement with the result
of 0.76 reported by Yang et at. 23  They also reported that the rate was
first order in S02 when H20 was present.

     The results for 1100°C H.T. 6.6% CaO in a-A!203 are shown in Fig. 36.
The sulfation rate for the 1100°C H.T. sorbent is approximately 65% of
that for 800°C H.T. sorbent.   Also, the functional dependence on the S02
concentration of the rate changes from 0.6 initially to 0.45 at 50%
completion.  The sulfation rates at 900°C of the 800°C and 1100°C H.T.
sorbents, using 0.1, 0.3, and 1% S02 with 5% 02, are compared in Fig. 37.
The sulfation rate for the 800°C H.T. pellets was about 1.5 times higher
than for the 1100°C H.T. pellets for all S02 concentrations in the gas
stream.  It is speculated that the lower rate for the 1100°C H.T. pellets
is due to the formation of more stable aluminate complexes (CaC"Al203 and
CaO-2Al203) at 1100°C than at 800°C (CaO-6A1203 and 3CaO-5Al203).

     Experiments have been performed to determine the effect of the oxygen
concentration on the sulfation rate of the sorbent.  Results for runs at
900°C using 0.3% S02 mixed with 5%, 0.5%, and 26 ppm 02 (balance is nitrogen)
*
 The a-A!203 (T708) was obtained from the Girdler Catalyst Co. and
 consisted of 1/8-in. x 1/8-in. cylindrical pellets.

-------
                                     102
                                      CALCULATED
                                   A  3%S02
                                   a  1% S02
                                     0.3%S02
                                   v O.I%S02
                                   o 0.05% SO
                                      160     200
                                      TIME, min
320
     Fig. 35.  Comparison of Calculated  (Eq.  1) and  Experimental
               Sulfation Rates at 900°C  in  5% 02-N2  of  6.6%  CaO
               in a-Al203 (Heat-Treated  at  800°C)  as a  Function
               of S02 Concentration.
are shown in Fig. 38.  When oxygen is in excess,  the  rate  is nearly
independent of oxygen concentration and is approximately 0.1 order.
However, when S02 is in excess, the rate is  first order in oxygen con-
centration.  This is consistent with the assumption that S02 reacts with
02 to form SOs before reacting with CaO.

          S02 + 1/2 02 -> S03

          SO 3 + CaO -> CaSOij

Yang et al. 3 reported a reaction order of 0.22 in oxygen  when in excess
for the sulfation of dolomite, in good agreement  with the  above results.
Yang et aZ.23 found a beneficial effect of water  vapor on  the  sulfation
rate of dolomite, the concentration of H20 being  immaterial.   The rate of
sulfation of the sorbents at 900°C was determined with reactant gases
containing 0.3% S02 - 5% 02 and 0, 0.1, 0.47, and 1.27% H20 (est).   The
sulfation rate was independent of the H20 concentration in the feed gas.

-------
                  PERCENT CONVERSION OF CoO TO CoS04
cw
•


U)
Cfl H
C Mi
H It)
Mi CD
PJ O
rt rt
H-
O O
3 Mi


O EC
ON rt
•  I
O\ H
n PJ
PJ rt
H- 3
3  rt

P  HI
N>  (D
O H
co  pi
rt  H
   n>
VO
O  O
O  3
 o
o  fa
   o
   Ml
                                                                        CONVERSION OF CaO  TO  CaS04, %
                                                                     .00000000000
                                                              H-
                                                              00
                                                          n fd
                                                          P pi
                                                          O rt
                                                            ro
p
I
                                                            M)


                                                            en
                                                            c
                                                            h-1
                                                            Ml
                                                         ho
                                                          o
                                                         oo  rt

                                                          PI  O
                                                          co  3

                                                          P>  pi
                                                            rt
                                                          P O
                                                          O O
                                                          rt  o
                                                          H. O


                                                          8
                                                          I!
                                                          i!
                                                          P8.

-------
                                     104
               Effect of Oxygen Concentration on the Rate of Sulfation
               of 6.6% CaO in a-Al203 at 900°C.  Sulfating gas mixture;
               0.3% S02 in N2 plus indicated concentration of 02.
     Effect of Sulfation Temperature on Sulfation Rate.  As shown in
Fig. 39, the rate of Sulfation of the 800°C H.T. sorbent increases with
temperature up to 900°C, where it becomes independent of temperature
(within experimental error).   The results were reproducible for various
CaO/S02 ratios, indicating that the sulfation rates were not S02-limited.
The independence of the rate at temperatures above 900°C indicates that
the reaction is diffusion-controlled at and above that temperature.

     The same type of results (Fig. 40) was found for the 1100°C H.T.
sorbent.  However, the activation energy below 900°C was higher for the
1100°C H.T. sorbent—65 kcal/mole compared with 27 kcal/mole for the
800°C H.T. sorbent.

-------
                                    105
                                 	CALCULATED, EQ. I (SEE TEXT)
                                     I050°C
                                      900°C
                                    160      200
                                     TIME, min
                                                     320
     Fig. 39.
Calculated and Experimental Rates of Sulfation of 6.6%
CaO in a-Al203 (800°C H.T.) with 0.3% S02 - 5% 02-N2cuas
a Function of Sulfation Temperature.
     Calcium Utilization.  Wet chemical analyses for calcium and sulfur
have been performed on sulfated sorbents which initially contained about
6.6% CaO in a-A!203.  These results, along with TGA weight change data
obtained in sulfation experiments, helped determine the extent of sul-
fation of the sorbent under different conditions.  Table 32 compares the
percent calcium utilization (percent sulfation) computed from the chemical
analyses with the utilization estimated from the weight change which
occurred during sulfation.  The calcium content of the sorbent used in
various runs, determined by chemical analysis, is shown in column 2, the
average calcium concentration is 4.27%, which is less than the value of
4.71% Ca (6.6% CaO) estimated from the gain in weight of the support
after loading with CaO.

     The percent sulfation (or calcium utilization) based on chemical
analyses is given in column 3.  An average of 84.4 percent was obtained.
Column 4 gives the percent calcium utilizations calculated from the
weight change which occurred due to CaO converting to CaSO^.  Column 5
presents the percent deviation between the two methods of calculating
the percent calcium utilization.  Fair agreement (a variance of 6.1%)
was found for the two methods of calculating the percent calcium utili-
zation.  The variance was only 4.6% when Run W40 was excluded from the

-------
                                   106
             SUL RATION
             • -1050^1
             D -  900
               -  850   EXPERIMENTAL
             A -  800
             O-  750J
                 CALCULATED
TEMPERATURE,
                                   TIME, hr

     Fig.  40.   Calculated and Experimental Rates  of Sulfation of 6.6%
               CaO in a-A!203 (110Q0C H.T.) with  0.3% S02 = 5% 02-N2
               as a Function of Sulfation Temperature.
calculations.   The 80-90% calcium utilization at  900°C is only slightly
dependent upon S02 concentration.  This result contrasts with the experi-
mental TGA results for dolomite, which indicated  that the percent
utilization is strongly dependent on S02 concentration.

     The percent  calcium utilizations at 750 and  800°C were 63 and 72%,
respectively.   These values are somewhat lower than the 80 to 100% con-
versions found in the 850 to 1050°C temperature range.

     Mathematical Analysis of Sulfation Rate.  An equation was developed
that correlates the rate of Sulfation of 6.6% CaO in a-Al203  (original
composition) with decreasing CaO concentration in the sorbent during
reaction, S02 concentration in the feed gas, and  temperature.  This
equation is:
                                  -0.025       \
                                                      0.7
                            1 +
                                  2.9 x
                                exp (-27000/RT)
                        [S02]
                                             (D

-------
             Table 32.  Calcium Content and Percent Sulfation of 6.6% CaO
                        in a-Al203 Sorbent Sulfated at 900°C
1



Run
W51
W50
W49
W23
W23 (repeat)
W39
W40
W90
W89
W73
W23
W70
W72
Av.
2

Ca Cone,
Chem Analysis
(wt %)
4.22
4.39
'4.29
4.27
4.06
4.26
4.19
4.01
4.47
4.07
4.06
4.63
4.52
4.27
3
Sulfation
Computed from
Chem Analysis
(%)
82.3
91.4
86.5
98.1
94.2
75.4
88.3
67.6
71.3
92.8
94.2
85.3
79.4
84.4
4
Sulfation
Computed from
Weight Change
(%)
82.4
91.2
88.4
94.1
98.9
82.3
100.6
62.6
72.1
93.0
99.0
87.0
75.8
85.7
5
Deviation,
Col. 3 vs
Col. 4
(%)
0.0%
-0.2%
+2 . 1%
-4.1%
+5 . 0%
+9.2%
+13.9%
-7 . 4%
+1.1%
+0.3%
+5.0%
+2.0%
-4.6%
variance 6.1%
6


Operating
Conditions
0.05% S02 900°C
0.1% S02 900°C
0.2% S02 900°C
0.3% S02 900°C
0.3% S02 900°C
1% S02 900°C
3% S02 900°C
0.3% S02 750°C
0.3% S02 800°C
0.3% S02 850°C
0.3% S02 900°C
0.3% S02 950°C
0.3% S02 1050°C
(4.6%)a
Excluding Run @40.

-------
                                     108
where  t =  time,  sec
       T =  temperature,  °K
       [S02] -  cone, of  S02, mole %
       [CaO]t=t
       rrani - = fraction of CaO, remaining at time t
       1    Jt=0

The reaction was found  to be 0.7 order in S02, first order in CaO con-
centration, independent of 02 concentration when oxygen is present in
stoichiometfic excess, and independent of H20 concentration.

     The temperature dependence of the reaction was determined by measuring
the rate over  a  temperature range of 750 to 1050°C, using a 0.3% S02 -
5% 02  gas mixture.  A gas-solids reaction can be either diffusion-controlled
or chemical reaction-controlled; for some such reactions, diffusion control
occurs at the  higher temperatures.  In these experiments, the sulfation
rate increased with temperature up to 900°C; at higher temperatures, it
was independent  of temperature.  The expression for the overall rate
constant, kOverall» which is shown below, takes into account both diffusion
control and chemical reaction control.24  When DS02 » k, koveran = kr;
when kr » Dso2, koverall ^ DSo2


                         D   /6
          k        =
           overall
where koveran = experimentally measured rate constant, sec"
          DS02 = s°2 diffusion coefficient, cm2 /sec
             
-------
                                    109
     Sulfation Rate as a Function of Calcium Loading in Support.   In  Fig.  41,
the percent conversion of CaO to CaS(\ is given as  a function  of  time for
2-20.9% CaO in a-Al203 heat-treated at 1100°C.   The rate of sulfation, when
measured as a fraction of the maximum possible  sulfation,  decreases with
increasing CaO concentration.   However,  as shown in Table 33,  the sorbent
weight gain at any given time during sulfation  is approximately the same
for all sorbents provided that 100% sulfation has not occurred.

     The 803  weight gain for Tymochtee dolomite is  also given  to  allow
comparison.  It captures approximately twice as much S02 as does  16.5%
CaO in a-Al203 sorbent for any given residence  time.
     Figure 42 shows the sorbent weight gain at 900 °C as a function of
CaO concentration in a-Al203 at the end of 1, 3,  and 6 hr; Fig.  42  is a
plot of some of the data in Table 33.
                                                      Sulfation, 900°C
                                                    0.3%S02-5%02;
                                                       IIOO°CH.T.
                                                      CaO in a-AI203
                               60      80     100
                                    Time, min
     Fig. 41.  Sulfation of Sorbents Having Various CaO Loadings.

-------
     Table 33.
                                  110
Sorbent Weight Gain during Sulfation for  Various
Calcium Oxide Concentrations.
               Sulfation Conditions:
                      Feed Gas, 0.3% S02,  5% 02
                      Temp, 900°C
Hours
into
Run
0.5
1
2
3
4
5
6
Max.
possible
wt gain

2%
Ca.O
18.0
21.0
22.5
23.7
—
—



28.6

3.3%
CaO
30.2
38.1
42.9
44.6
45.4
—
—


47.1

5.4%
CaO
27.5
40.6
53.2
60.7
64.5
67.0
—


77.1
Weight
6.6%
CaO
34.9
48.1
76.7
84.4
88.2
—
—


94.3
gain, g/kg of 1100°C H.T. sorbent
8.3%
CaO
34.4
51.8
74.7
88.6
96.7
100.7
102.9


119.0
10.5%
CaO
31.5
49.3
77.7
95.3
109.2
119.4
123.6


150
14.8%
CaO
32.6
49.0
74.0
94.7
112.7
126.4
136.3


211.4
16.5% Tymochtee
CaO Dolomite
24.3
43.9
72.6
96.0
112.0
126.4
140


235.8
93.4

161.2
191.2
213.1
230.0
243.0


546.0
   240
 « 220
;! 200
5 180
 ° 160
 ^ 140
~ 120
 ^ 100
•5  80
    40
en
      0
                              I  I   I   !   I   i  I   I   I   I
       0    2     4     6    8     10    12    14    16    18    20   22
                            CaO, wt % in a-AI203
    Fig. 42.  Sorbent  Weight Gain at 900°C as a Function of Calcium
              Loading  of  Sorbent.

-------
                                    Ill
The fact that all sorbents studied capture S02 at the same rate is an
indication that the reaction is diffusion-controlled, which is in agreement
with the observed temperature independence of the reaction rate above
900°C (which also implies diffusion control).

      Comparison of Sulfation Rates of Tymochtee Dolomite and Synthetic
Sorbents.  Tymochtee dolomite was sulfated in the TGA unit at 900°C, using
0.3% S02 - 5% 02 in N2, for comparison with 6.6% CaO in a-A!203 (Fig. 43).
The information obtained will help in determining the relative effectiveness
of the synthetic sorbent in reducing S02 concentrations in the effluent
gas from a fluidized-bed combustor.  Sulfation of the synthetic sorbent
was complete in 6 hr; Sulfation of the Tymochtee dolomite was approximately
60% complete in 19 hr.

      Dolomite contains approximately four times as much calcium as does
the synthetic sorbent.  Therefore, in 6 hr the dolomite utilized approxi-
mately 2.5 times the quantity of calcium utilized by the synthetic sorbent.
In an earlier EA-series experiment in the bench-scale combustor,
approximately 60% calcium utilization was obtained for Tymochtee dolomite
in the fluidized bed, which is in good agreement with the calcium utilization
obtained for dolomite in the TGA unit.

      TGA results can also be compared with an earlier series of combustion
experiments  (VAR-series) .  Calcium utilization in dolomite in the bench-scale
fluidized bed ranged from 33 to 83% and was inversely. proportional to the Ca/S
ratio.17  The maximum calcium utilization (83%) was obtained with low Ca/S
ratios in the VAR-series.  In comparison, the TGA runs were performed with
excess S02 and therefore low Ca/S ratios.  However, calcium utilization was
only 60% in  the TGA runs.  Thus, TGA results do not agree with results of
the VAR experiments.  Possibly, long residence times may account for the
high calcium utilization results obtained in the bench-scale fluidized bed.
Also, H20 (which should increase the reaction rate) was absent from  the TGA
sulfation gas but was present in the VAR-series.

Metal Oxides in ot-
      Metal oxides other  than CaO have been tested  for  their ability  to
react with S02 and 02 and for the ability of the sulfated sorbent  to  be
regenerated.  Phillips25'26 and Cusumano and Levy27 have made thermodynamic
calculations to determine the most promising compounds  for reaction with
S02 and for subsequent regeneration with a reducing gas.  Phillips suggested
that LiA102 and Li2Ti03 are the most promising candidates.  They were
chosen from a list of 16  shown in Table 34 (Cusumano and Levy have
suggested twelve candidate materials, all of which  were suggested  previously
by Phillips).

      Neither LiA102 nor  Li2Ti03 appears to be a good choice in view  of
the instability of Li2SOi+ at high temperatures.  Li2S04 has been reported
by Stern and Weise28 to decompose at a noticeable rate  not far above  its
melting point  (860°C) .  Ficalora et al. 29 report that Li2SOt+ decomposed
into gaseous Li, 02, and  S02 at temperatures above  780°C.  Nevertheless,
4.5% Li20 in a-A!203 was  prepared and tested.  X-ray diffraction results

-------
                                     112
  100
o
o
   0
                                           O  TYMOCHTEE DOLOMITE
                                           •  6.6% CaOIN AI203
                                            i   I   i    I   i
j	I
     02       4      6       8       10      12     14      16     18
                                     TIME, hr

       Fig.  43.   Comparison  of  the  Rates of Sulfation of Tymochtee
                 Dolomite and 6.6%  CaO in a-A!203.   Sulfation Gas
                 Mixture: 0.3% S02 and 5% 02 in N2.  Sulfation
                 Temperature:  900°C.
          20
            Table  34.  Most  Promising S02 Sorbents Based on
                      Thermodynamic Screening Results.3
                     Sorbent
                                        Temperature Range
Na20
CaO
SrO
BaO
LiA102
LiFe02
Li2Ti03
NaA102
NaFe02
CaAl2Oi+
Ca2Fe205
CaTi03
SrAl2Oit
SrTi03
BaAl2Oit
BaTi03
1100-1200
750-1090
950-1200
1080-1200
750-1200
750-950
750-1200
750-820
750-1020
750-950
800-1150
830-1150
750-1000
750-920
750-1000
750-1000
              Selected from references 5 and 6.

-------
                                    113
indicated that LiAlsOs was present in the sample.  The sample was found
to lose weight rather than gain weight under sulfation conditions.  The
weight loss was presumably due to decomposition of the Li^Q^ formed by
the sulfation reaction.  No further tests are planned with this material.

     The sorbents 7.9% Na20 in a-A!203, 5% K20 in a-Al203, 14.4% SrO in
a-A!203, and 5.2% BaO in a-A!203 have been tested.  X-ray diffraction
results indicated that the starting materials were B-NaA102, KAlnOi7,
BaAl2Oit, CaAl2Oit, and SrAl2Oit in a-A!203.  Their rates of sulfation at
900°C using 0.3% S02 - 5% 02 in N2 are compared with the rates for 6.6%
and 14.8% CaO in a-A!203 in Fig. 44.  Potassium and sodium sorbents
clearly have higher rates of sulfation than do CaO sorbents while the
rates of sulfation of the BaO,* SrO, and CaO sorbents are essentially the
same.  The products of sulfation were Na2SO^, K2SOt,., BaS04, SrSO^, and
CaSOit.  After regeneration, the original aluminates given above were
found.

      Since CaO, BaO, and SrO sorbents containing comparable amounts of
metal oxides have the same sulfation rate and since calcium is likely to
be the least expensive material, calcium is probably the most desirable.
Potassium and sodium sorbents have higher sulfation rates than CaO; how-
ever, due to their surface instabilities at regeneration'conditions  (see
Regeneration section), they do not seem promising.
 It is believed that if the BaO concentration had been
 sulfation would be comparable to that of CaO.
   the rate  of
                        0.5
2.5
                                     Time, hr
      Fig. 44.  Comparison of  Sulfation Rates  of  Various  Metal
                Oxides at 900°C.

-------
                                     114
Regeneration Studies

     Regeneration kinetics were extensively studied for sulfated 6.6%  CaO
and a-Al203 sorbent as a function of reducing gas concentrations and tem-
perature.  The other metal sulfates in a-A!203 were regenerated at only
one set of conditions and were compared with the rate of regenerating
      to CaO.
     Effect of Reducing Gas.  The rate of regeneration of CaSOi^ in a-Al203
was studied at 1100°C as a function of reducing gas concentration for CO, H2 ,
and CH4.  The results are given in Figs. 45 and 46 for H2 and CO.  The
regeneration rate for CHi^ was the same as that for H2 .  For each reducing
gas, the percent reduction of CaSO^ is given as a function of time, for
reducing gas concentrations ranging from 0.1% to 6%.  For all three gases,
the reaction is 0.8 order in reducing gas concentration, less than 4 min
being required for complete regeneration when using a 6% reducing gas
concentration.  X-ray diffraction results on the regenerated sorbents showed
that for the hydrogen and carbon monoxide runs, the product was a mixture
of CaO-Al203 and CaO-2Al203 and that no CaS or CaSOit was present.  However,
in runs (not plotted) in which methane was used as a reductant, SEM analysis,
using a scanning electron microscope, showed that trace amounts of sulfur
were present.
                                             EXPERIMENTAL POINTS
                                             CALCULATED, EQ.4 (SEE TEXT)
                                      TIME, min

     Fig. 45.  Regeneration of Sulfated 6.6% CaO-a-Al203 Pellets,
               Using Hydrogen at 1100°C.

-------
                                              6% CO
                                              3% CO
                                              I%CO
EXPERIMENTAL POINTS
                                             0.1% CO
                                              CALCULATED, EQ.5 (SEE TEXT)
                                                                                                  H-1
                                                                                                  M
                                                                                                  Ui
                                      TIME, min
Fig.  46.   Regeneration of Sulfated 6.6% CaO-a-Al203 Pellets,  Using
          Carbon Monoxide at 1100°C.

-------
                                     116
     Points plotted in Figs. 45 and 46 are based on the assumption of 100%
regeneration of CaSOi+ to CaO.  This seems reasonable on the basis of weight
loss information.  Table 35 presents the percent regeneration for each
run at the end of reaction.  In each regeneration run, the experimental
weight loss was obtained by weighing the sample before and after each
reaction.  The regeneration rate when CO was used was a factor of three
lower than the rates of regeneration when using CH^ or H2 which are
essentially the same.  An explanation is that thermodynamically, the de-
composition of CHij to C and H2 is favored at 1100°C.  Therefore, it is
thought that hydrogen is the actual regeneration gas when CH^ is used, and
that the decomposition of CHi). is not rate-limiting.  Further evidence in
support of this idea was found when sorbents reduced with methane, hydrogen,
or carbon monoxide were sectioned for SEM analysis.  The sorbents reduced
with CHfj were black on the interior while those regenerated with H2 or CO
were white.  Analysis verified that carbon deposition occurred when
was used.
     Calculations indicate that if 1.3 mg of C (from the reduction of
to C and H2> is present in each 35-mg pellet, enough H2 would have been
produced to reduce all CaSOt,. to CaO.  Table 36 gives the quantity of
carbon found in the sorbents after regeneration at various reducing gas
concentrations.  The amount of carbon deposition in the sorbents is less
than that needed to obtain enough hydrogen for complete regeneration of
the CaSOi^.   However, carbon deposition must also have occurred throughout
the TGA unit, making available the required amount of hydrogen for regen-
eration.

     The effect of.C02 concentration in the reducing gas on the regeneration
rate of sulfated sorbent was studied.  Carbon dioxide might decrease the rate
due to the thermodynamically favorable reaction of C02 with H2 to form H20
and CO.  The regeneration rate apparently decreased only when the C02 concen-
tration was larger than 15% (Fig. 47).  As reported above, when CO was used
as the reducing gas, the rate of regeneration was approximately one-third
that when hydrogen was used.
     Table 35.  Regeneration of CaSOt,. to CaO, Calculated from Weight Loss.

       Reducing                   Calculated Completeness of Regeneration,
     Gas Cone, %                      CO             H2             CH4

        6                            102             93              87
        3                            102             99             109
        1                             96            106              98
        0.5                           86            108              89
        0.1                          110            103              92

-------
                                      117
      Table 36.  Carbon Content of Pellets after Regeneration with Methane.
      Reducing Gas and
      its Concentration              Carbon
      in Experiment                  (ppm)
                                                         Carbon
                                                   (mg/35-mg pellet)
          6%
          1%
          0.5%
          3% H2
                                 2300
                                  850
                                  400
                                  200
                                                             0.41
                                                             0.15
                                                             0.07
                                                             0.04
      1.3 mg of C from the reduction of CH^ to H2 + C would be needed
      in a 35-mg pellet to obtain enough H2 to reduce all CaSO^ in  the
      pellet to CaO.
   100

    90

  3 80
sf70
3 so
U.
o 50
| 40
3
               at 1100° C.  Regeneration Gas:  1% CO, 1% H2, 1% CH4 in
               N2 + indicated C02.

     Regeneration Rate as a. Function of Temperature. 'Regeneration experi-
ments over a temperature range of 800 to 1200°C, using 1% H2. in the  gas
stream, have been completed.  The results for 1200, 1100, 1000, and  900°C
(Fig. 48) show that the regeneration rate increases with temperature.

     The composition of the regeneration reaction product is a function of
temperature, as shown in Table 37.  X-ray diffraction analyses indicated
that above 1100°C, the only product is CaO, and below 900°C, the entire

-------
                                     118
                                   — CALCULATED
                                   o - I200°C
                                   n-IIOO°C
                                     -IOOO°C
                                     - 900°C
                 3   4   5   6   7   8   9   10   II   12   13   14   15  16   17   18
     Fig. 48.  Regeneration of Sulfated 6.6% CaO  in a-Al203  Sorbent,
               Using 1% H2 - N2.
     Table 37.  Product3 of the Regeneration Reaction  at Various
                Temperatures.
                  Regeneration
               Temperatures  (°C)
     Product
                      1200
                      1150
                      1100
                      1050
                      1000

                       950
                       900
                       850
                       800
CaO
CaO
CaO
Mostly CaO-little CaS
Medium quantities of CaO
 and CaS
Mostly CaS-little CaO
Mostly CaS-little CaO
CaS
CaS
      Determined by X-ray diffraction analysis.

product is CaS.  In the intermediate temperature  range of  900 to 1050°C, the
product is a mixture of CaO and CaS, with CaO  concentration increasing with
temperature.  The kinetic information determined  as  a function of temperature
was used to determine the activation energy, which was approximately 15
kcal/mole.  It was found to be independent  of  the product  formed (CaO or CaS).

-------
                                     119
     Mathematical Analysis of Regeneration Kinetics.  The experimental
regeneration kinetic data were analyzed, using the following equation:

                              d[CaSOtJ
          regeneration rate = ——,  n
                                 dt                                     (3)
                            = -k[R.G.]X[CaS01+]y exp  (-E /RT)
                                                       3.
where  [CaSOi^] = CaSO^ cone, mole/m
       [R.G.]  = reducing gas cone, mole/m3
       t = time, sec
       Ea = activation energy, cal/mole
       T = temperature, °K
       R = gas constant
       x, y, k = constants

     A linear dependence of the log of the experimental regenerated rate
as a function of 1/T, [R.G.], and [CaSOiJ was found.  Therefore, Ea, k,
x, and y could be determined.  The results are shown below for each reducing
gas:
              dt
                   =  -3.36
                             H20'8
                             or
[CaSOtf]exp(-14,900/RT)             (4)
          dfCaSOiJ =  __1>08  [CO]0'8  [CaSOu]exp(-14,900/RT)              (5)
             dt

The order of the reaction with respect to the reducing gas concentration
is 0.8 for all reducing gases; however, the regeneration rate is three
times lower for CO than for H2 or CH4.  The activation energy was deter-
mined from the experimental data for hydrogen presented above; the activation
energy for CO was assumed to be the same as that for hydrogen.  Figures
46, 47, and 48 give a comparison of the experimental and calculated  (Eq.
4 and 5) results.  Agreement of the experimental and calculated results is
good for the regeneration rate as a function of temperature and hydrogen
concentration (shown in Figs. 48 and 45, respectively).  It should be noted
that the activation energy was found to be independent of the product
formed.  Figure 46 presents the regeneration rate as a function of CO
concentration; agreement of experimental and theoretical results was poor
only at the lowest CO concentration  (0.1%).

     Comparison of Regeneration Rates of Sulfated Synthetic Sorbents and Sul-
fated Tymochtee Dolomite.  In an attempt to generate  meaningful data applicable
to pilot plant fluidized-bed regeneration experiments, Tymochtee dolomite
that had been sulfated in the coal-fired, fluidized-bed combustor was
reduced with CO, H2, and CE^ at 3% and 1% concentrations for comparison
with regeneration of the synthetic sorbents.  Data are plotted in Fig. 49
for 3% reducing gas concentration of H2 or CO.  The dolomite contained
10.1% S and was expected to lose about three times more sulfur than the
synthetic sorbents during regeneration.  Also, the synthetic sorbent pellet
diameter was approximately three times larger than dolomite particles dia-
meters.

-------
                            REDUCTION OF  CaS04, %
  OQ


  •P-
l-h-d
pj pj
rt H
fl> H-
p. CO
  O
O 0
O
H-1 O
O Hi
rt
fD
s
H- rt
9 (D

09 o
CO l-h
S3 ft)
N) OQ
   (D
O  0
CO
   "•*    _
n o
o 3

p> o
rt i-h

(-• n
l_l (a
o en
o o
  o-fr
   s
  l-o
   O
  00

   I
            ooooooooo    o   o    o
                                     OZT

-------
                                    121
     The regeneration rates for dolomite were somewhat lower than those
for the synthetic sorbent; the decomposition rate of CaSOi^ in the dolomite
decreased rapdily near the end of the reaction.  This decrease in rate
was not observed for the synthetic sorbents.  However, the residence times
in a regenerator for dolomite and synthetic sorbent would not differ
significantly.

     The major difference was in the product found at the end of the
reaction.  As stated above, for the synthetic sorbents, the products were
CaOAl203 and CaO-2Al203; for dolomite, the products were CaO and CaS.  (In
neither case was CaSO^ found.)  Wet chemical analysis will be used to
quantify the amount of CaS in the dolomite.  X-ray diffraction data indi-
cate that CaS might constitute as much as 30-50% of the product.

     The SEM results showed that most of the sulfur in the dolomite was
near the surface of the particles.  This is not surprising since these
particles were only 50% sulfated in the combustor before being reduced
in the TGA.  It should also be noted that both iron and chlorine were
also found on the regeneration surface.

     Regeneration of Various Sulfated Metal Oxides.  Various sulfated
metal oxides in a-A!203 were regenerated for comparison with the regen-
eration rate of CaSOit in ot-Al203.  The four metal sulfates studied were
Na2S04, K2S04, SrSOtt, and BaS04 in ot-Al203.  Regeneration rates (using
3% H2 at 1100°C) are shown in Fig. 50, where they are compared with the
regeneration rates of CaSOi^. in a-A!203.  The regeneration rates for all
metal sulfates were extremely high compared with their sulfation rates.
A maximum of 5 min was required for regeneration.  The percent regeneration
ranged from 85 to 94%.  The reason for not attaining 100% regeneration
has not yet been determined.

     The regeneration results for K2SOtt are in disagreement with the
thermodynamic calculations of Phillips.25  In his report, potassium was
rejected on the premise that potassium sulfate is theoretically stable
at regeneration conditions.  As can be seen in Fig. 50, experimentally,
the regeneration of K2SOu was comparable to the regeneration rate of
However, the instability29 of both Na2S04 and K2SO^ at the regeneration
temperature of 1100°C must also be considered.  A part of the weight loss
observed during regeneration may have been due to decomposition of the
compounds.  This same observation has been made by Pierce.30

Cyclic Sulfation-Regeneration Studies

     Five cyclic sulfation-regeneration reactions using 6.6% CaO in a-A!203
(1100°C H.T.)* were performed to determine if a loss in reactivity would
be related to the number of cycles.  In the sulfation reaction, 3% S02 -
5% 02 in N2 was used at 900°C; in the regeneration reaction, 3% H2 was
used at 1100°C.
 The final step in impregnating the supports with calcium oxide is heat-
 treating the sorbent at a proper temperature to form stable calcium
 aluminates that react with S02 and can be re-formed during regeneration,

-------
  0
Fig. 50.
Regeneration Rate of Various Metal Sulfates in a-A!203 at  1100°C
Using 3% H2.
                                                                                                     NJ

-------
                                     123
     As shown in Fig. 51, the rate of sulfation of the sorbent was the same
for the five cycles (100% sulfation in each cycle was assumed).  In Table 38,
the percent calcium utilization during sulfation and percent conversion of
CaSOtt in a-A!203 to CaO during regeneration are given.  If it is assumed that
the sorbents contain 6.64% CaO, calcium utilization was greater than 96%.

     A minimum of approximately 50 cycles is probably needed to determine
if any structural degradation will occur due to phase changes occurring
because of internal chemical reactions.  The purpose of this cyclic test
was to show that reproducible and predictable sulfation rates could be
obtained when the synthetic additive was heat-treated at 1100°C.  This
reproducibility was not found for -the 800°C heat-treated sorbent.
     Additional cyclic sulfation-regeneration experiments have been
performed with two synthetic sorbents  containing 11.1 or 12.5% CaO.
supports for these sorbents were prepared here at Argonne.
                                                       The
 These synthetic sorbents were prepared by heating granular £-1 boehmite to
 above 1500°C for 8 hr to form a-A!203 and impregnating it with CaO.
 Heat-treating temperature determines the porosity of the support and
 affects the sorbent reactivity.
                                    CYCLE
                                    CYCLE 2
                                    CYCLE 3
                                    CYCLE 4
                                    CYCLE 5
              I
   I
I
I
         20   40  60   80  100
     Fig. 51.
                120  140  160  180  200  220  240  260  280  300
                    TIME, min
Cylic Sulfation of 1100°C Heat-Treated Pellets (6.6%
CaO in a-A!203) Using 3% S02 - 5% 02-N2 at 900°C.

-------
                                    124
           Table 38.   Calcium Utilization and  Regeneration
                      during Sulfation-Regeneration  Cyclic
                      Experiments,  Using  1100°C H.T.  Pellets

             Cycle        Sulfation  (%)       Regeneration  (%)
1
2
3
4
5
___
97
98
98
97

97
97
99
96
     Samples were sulfated at 900°C, using 0.3% S02 - 5% 02 in N2, and
were regenerated at 1100°C, using 3% H2 with the balance N2.  Sorbent
reactivity with S02 decreased, upon recycling, for the granular sorbents
containing 1.11 or 12.5% CaO.  The loss of reactivity upon recycling
of the 11.1% CaO in a-A!203 sorbent (support heat-treated at 1500°C)
and the 12.5% CaO in a-A!203 sorbent (support heat-treated at 1350°C)
can be seen in Fig. 52 and 53.  The amount of CaO that reacted with S02
and 02 to form CaSOi,. decreased from 81% to 31% in 10 cycles for synthetic
sorbent containing 11.1% CaO.  The CaO utilization for synthetic sorbent
containing 12.5% CaO was somewhat better; nevertheless, approximately 52%
of the CaO that was reactive in the first cycle was still reactive in
the tenth cycle.  Calcium analysis of the synthetic sorbents prepared
from granular supports showed no calcium loss that might explain the
loss of reactivity.

Methods of Preparing Supports with Optimum Physical Properties

     An investigation has been initiated to detemrine methods of producing
supports having the required properties (surface area, porosity, and
strength).  Porosity measurements have been performed on a number of
granular supports to determine what effect pore size distribution has
on the synthetic sorbent reactivity during sulfation.   All synthetic
supports tested to date were prepared from granular (-14 +30 mesh) boehmite,
y-AlO(OH), obtained from Alcoa (Aluminum Co.  of America).  Samples were
heat-treated at 1100°C, 1200°C,  and 1500°C for 8 hr.   (The rate of heating
to the heat-treatment temperature has no effect on the sample's pore size
distribution.)   The pore size distributions (cumulative pore volume versus
pore diameter)  of the 1100°C, 1200°C,  and 1500°C H.T.  supports are given
in Fig. 54.   The pore size distribution for as-received boehmite is also
shown.   As can be seen in Fig.  54,  the higher the heat-treatment temperature,
the larger the pore diameter.  Also,  it should be noted that 76% of the
pore volume of the 1500°C H.T.  support is due to pores having a diameter
larger than 0.3 pm and that this material contains essentially no pores
smaller than 0.13 ym.

     Synthetic sorbents containing 11.1 and 11.4% CaO were prepared from
the 1500°C H.T.  support;  a sorbent containing 12.5% CaO was prepared using
the 1200°C H.T.  support;  and a sorbent containing 12.5% CaO was prepared

-------
                                   125
                                                      O-7  O-9
                                                          -8  *-IO
     Fig. 52.  Cyclic Sulfation of 11.1% CaO in 1500°C H.T.  a-A!203
               Granular F-l Sorbent Using 0.3% S02 - 5% 02-N2  atr900°C.
using the 1100°C H.T. support.  These sorbents were sulfated in the TGA
to determine their reactivity with 0.3% S02 - 5% 02 in N2 at 900°C.  In
Fig. 55, the rates of reaction (or conversion) of calcium oxide with S02
to form calcium sulfate are shown for 8.8%, 11.1%, 11.4%, and 12.5% CaO
in a-A!203 sorbents.  Also, the sulfation rate for 10.5% CaO in a-Al203
(Girdler T-708) sorbent is shown.  The sorbents prepared from 1500°C H.T.
support has the highest rate of conversion of CaO to CaSO^i its reactivity
is even higher than that for 10.5% CaO in a-A!203 (T-708).   These results
suggest a correlation between pore size distribution in supports and
the reactivity with S02 of sorbents made from these supports;  the larger
the pore diameter, the greater the reactivity.  Synthetic sorbents prepared
from supports containing a large percentage of pores smaller than 0.1-0.2  ym
have low sulfation rates.

     The quantities of S02 captured by given quantities of synthetic
sorbents at various residence times are given in Table 39.   The sorbents
prepared from the 1500°C H.T. supports captured more S02 than did sorbents
prepared from either the 1100°C or 1200°C H.T. supports.  Also, this granular
sarbent performed better than did the 10.5% CaO in a-Al2Os  (T-708) sorbent.

-------
                                     126
     Fig. 53.  Cyclic Sulfation of 12.5% CaO in 1350°C H.T. a-A!203
               Granular F-l Sorbant Using 0.3% S02 - 5% 02-N2 at 900°C.
     To improve our understanding of the effect of pore £ize distribution
on reactivity, porosity measurements were made on:  (1) a support prepared
by heat treatment at 1500°C, (2) a support (from the same batch) after
loading with 11.4% CaO, and (3) the CaO-loaded support after sulfation.
The results are given in Fig.  56, where the cumulative pore volume, cm3/0.5 g
of material, is given as a function of pore diameter.   This information is
helpful in determining the calcium and sulfur distributions in the pores.
The porosity, as expected, is highest for the original support and lowest
for the sulfated sorbent.  A theoretical porosity curve for the sorbent
containing 11.4% CaO was calculated using the density for calcium aluminates,
CaSO^, and a-A!203 and assuming that 11.4% CaO was distributed evenly in
the pores.  Agreement of theoretical and experimental values is fair for
the large pores (>0.2 pm), but poor for the small pores.  These results
indicate that the CaO sealed off the entrance to the small pores during
preparation of the sorbent.

     If it is assumed that the experimental porosity curve is correct for
the 11.4% CaO in a-A!203 sorbent, a theoretical porosity curve (shown in
Fig. 56) can be calculated for the sulfated sorbent.  This theoretical
curve agrees well with experimental porosity measurements for large pores

-------
                Conversion  of CaO to CaS04,  %
                                                                Cumulative  Pore  Volume, crrr/0.5g
     Ui
rt O CO
O CD  C
  CO  I-1
H. ..  Hi
H- O H-
O  •  O
fU  U> 0
rt  S>9
fD    W
CL. CO (U
   O rt
rt N>  fD
fD

I  '  R
fD  UI
I-i  &-S VO
p    O
rtOO
C K>
H  I
fD  21
CO N>
     O
     O
     Hi
     n
   E« (U
   e o
   O 3
   H
   rt O
   TO H
   fD W
   P> C
   ft 13
   1 -O
   rt o
   i-i i-i
   fD rt
   P3 CO
   rt •
   CD
                        ro
                        o
                         CD
                         O
                                           OO
                                           O
O
o
               ro
            CO
O>
    cn
                                     -!- I    I   I   I
    pppoppopoopp
    OOOCD  —  —  —  '—  •—  roro'ro
O  ro -i^   cr>  oo  O  ro
CD  cx>   o  ro
                                                          H-
                                                          OQ
                                                          Ln
                                                          •P-
                                              fD
                                              h-1

                                              rt
                                              H-
                                              O
                                              0
                                              co

                                              H-


                                              O
                                              Hi

                                              O
                                                          Pi
                                                          rt
                                              fD    —
                                              O
                                              I-!
                                              fD
                                                          fD
                                                          rt
                                                          O
                                               fD

                                               G
                                               H-
                                               fD
                                               rt
                                               fD
                                               K
                                                                                                                       N3

-------
                       Table 39.  Sulfation of Supported Sorbents.
Residence

1/2
1
2
3
4
5

8.8% CaO in F-l
Granular ot-Al203
(1100°C H.T.)
13
18
24
29
33
35

12.5% CaO in F-l
Granular a-Al203
(1200°C H.T.)
31
41
52
59
64
—

11.1% CaO in F-l
Granular a-A!203
(1500°C H.T.)
63
91
117
127
129
—

11.4% CaO in F-l
Granular a-A!203
(1500°C H.T.)
57
84
109
117
123
—

10.5% CaO
in T-708a
ex-alumina
32
49
78
95
109
—


Tymochtee
Dolomite
93
—
161
191
213 ^
230 °°
T-708 is the support generally used in the supported additive experiments.

-------
                                 129
 in
  •
 o
10
   0.24
E
o
 -
LJ
   0.20
   0.16
 UJ
 gO.I2
 Q.

 > 0.08
 o
         Mill  I
                T
             Mill I
HIM I I   I
TTTTT
             I-I500°C HEAT-TREATED ORIGINAL SUPPORT
            •2
            -2
            •3
       11.4% CaO in a-AI203 SUPPORT (EXPERIMENTAL)'
       11.4% CaO ina-AI203 SUPPORT (THEORETICAL)
       SULFATED 11.4% CaO in a-AI 0
_     SUPPORT  (EXPERIMENTAL)
	3 SULFATED 11.4% CaO in a-AI203
       SUPPORT (THEORETICAL)
   0.04


     0
       100


    Fig.  56.
                     10           I             O.I
                        D,  PORE DIAMETER,
              Relation of Pore Volume to Pore Diameter for the Original
              a-A!203 Support, the CaO in a-A!203 Sorbent, and the
              Sulfated Sorbent.
 (>0.2 ym).  However, the theoretical result indicates  that there  should
 be a larger number of small pores than was experimentally found.  The
 observed lowering of the total pore volume below the theoretical  value
 might be due to complete blockage of small pores by the formation of
 These results indicate that possibly the small pores are not being fully
 utilized and that there might be better preparation techniques that would
 allow utilization of the small pores (<0.2 ym).

     Porosity measurements were also performed on sorbents containing
 various concentrations of CaO (0 to 16.5%).  The results are given in Fig.
 57.  As the CaO content increased, the total pore volume decreased, which
 was expected due to the pores being filled with CaO.  The total pore
 volume found experimentally for the original support (0.205 cm3/0.5 g) was
 used to calculate the theoretical pore volume for the  four sorbents (Table 40)

-------
                                  130
  C7»
  .  0.24
 o
 X
ro
  £0.20
 lit
 §0.16
      Illl I I  I  I    Mill I I  I  I    Illll I I   I  I    III I I |  | I   I   |I|I I I I
 O

 LJ
 (T
 O
 Q.
0.12
      —   1-0,  ORIGINAL SUPPORT
      —   2-3.3 % CaO
      —   3- 5.4 % CaO
      _   4-10.5 % CaO
            5-16.5 % CaO
                                      11  i  i  i    Inn i i  i  i   linn i
                      10           I             O.I
                       D, PORE DIAMETER ,
                                                         0.01
     Fig. 57.  Relation of Pore Volume to Pore Diameter as a Function
              of Indicated CaO Concentration in the  Support.
         Table 40.   Pore Volume for Various Synthetic Sorbents.


CaO in
Support
(%)
0
3.3
5.4
10.5
16.5


Experimental
Pore Volume
(cm3/0.5 g)
0.205
0.183
0.168
0.150
0.127


Theoretical
Pore Volume
(cm3/0.5 g)
0.205
0.196
0.189
0.172
0.151

Experimental Pore
Volume Loss by
CaO Addition
(cm3/ 0.5 g)
0
0.022
0.040
0.055
0.078
Theoretical
Volume Loss
By CaO
Addition
(cm3/0.5 g)
0
0.009
0.016
0.033
0.054
The theoretical decrease in pore volume due to CaO addition is substantially
less than that found experimentally for each sorbent.  This again indicates
that possibly, CaO plugs the entrances to small pores, lowering the observed
total porosity below the theoretical value.

-------
                                     131
                          SORBENT ATTRITION STUDIES
 R.  Synder (Principal  Investigator),  I.  Wilson

Introduction

     Recently, a research program has been initiated to study sorbent
attrition (natural and synthetic sorbents).  The objectives of this
program are to determine the mechanism of attrition in fluidized beds
and to determine a relatively simple method for determining the attrition
resistance of sorbents when used in fluidizied-bed coal combustors.

     In a series of experiments, Tymochtee dolomite, various support
materials for syntehtic sorbents, and one synthetic sorbent were tested
for attrition resistance in a cold fluidized bed under various fluidization
conditions.   The sorbents were fluidized with air in a 5.08-cm-dia laboratory-
scale fluidized bed at room temperature.

     The results are shown in Table 41.   The percentage loss was calculated
as the quantity of overhead material produced in 10 hr.  The overhead
particles were all smaller than 70 mesh; -the size range of the original
material was -14 +30 mesh.  The L/D ratio was 1.38 for most experiments.
(A larger value produces a slugging bed.)  Results of the various fluidi-
zation -^experiments are compared below.

Attrition of Dolomite

     The results of run 1 and run 2 with half-calcined Tymochtee dolomite
show that the larger the particles, the higher the attrition rate.

     In runs 3, 4, and 5, the same material (sulfated dolomite containing
6.3% S) was used.  Increasing the fluidization velocity produced greater
attrition rates (run 3 vs. run 4).  Increasing the L/D ratio from 1.25
to 3.5 (a slugging bed) reduced decrepitation drastically, i.e., by a
factor of 4 (run 4 vs. run 5).

     Dolomites which had been 33%, 48%,  and 63% sulfated (runs 3, 6, and 7)
were fluidized to determine attrition resistance as a function of the
extent of sulfation.  Run 3, with a 33% calcium utilization, showed an
unusually large attrition rate (16% loss) compared with runs 6 and 7
(2-3% loss of material).  At present, this cannot be explained.  All three
samples had been sulfated in the bench-scale combustor during FY 1974.17
The sulfated samples had been stored in containers that were not air-tight.
Because the effect of the 2-yr storage period on these samples is uncertain,
half-calcined dolomite will be sulfated to various sulfur levels and tested
to determine their attrition resistance.

     Runs 1 and 8 show the attrition rate determined in the laboratory-scale
fluidized bed for (1)  the feed to the bench-scale combustor and  (2) the
sulfated bed material from cycle 1 of the cyclic bench-scale sulfation-
regeneration experiments.  Run 8a shows the experimental bench-scale results.
The sulfated dolomite (run 8) had a low attrition rate (0.3%/hr) when tested
in the laboratory-scale fluidized bed.  This is not in agreement with the

-------
                                         132
             Table 41.   Fluidized-Bed Attrition  Experiments^

Run
No.
1


2



Sample
Tymochtee Dolomite
(half -calcined)
(14 +30)
Tymochtee Dolomite
(half calcined)


L/D
1.25


1.25

Fluidiz ing-Gas
Velocity
(ft/sec)
2.2


4

Particles in
Bed after Percent Loss3
10 hr (%) in 10 hr
52.6 47


0 100
(10.5/hr)
       (-4 +14)

 3     Sulfated Dolomite       1.25
       (from VAR 8; 6.3% S,
       24% Ca)

 4     Sulfated Dolomite       1.25
       (from VAR 8)

 5     Sulfated Dolomite       3.5
       (from VAR 8)

 6     Sulfated Dolomite       1.25
       (VAR 6; 9% S,
       23.2% Ca)

 7     Sulfated Dolomite       1.25
       (VAR 5; 11.2% S,
       22.4% Ca)

 1     Tymochtee Dolomite      1.25
       (Feed for bench-
       scale combustor
       Cycle 1)

 8     Sulfated Dolomite       1.25
       from Cycle 1

 8a    Sulfated Dolomite0      6

 9     Regenerated Dolomite     1.25
       (feed for combustor,
       Cycle 2)

 9a    Regenerated Dolomitec    4.26
2.2



3.0


2.2


2.2



2.2



2.2




2.2


2-3

2.2



4.69
84.1



75.2


96.0


97.8



97



52.6




96.7
                 16
25
 4.0
 2.2
47
                 4-5%/hr

                100
           (1.9% loss in
           7 min)

           2.4% loss in
           7-min residence
           time
10    1350°C H.T.  granular     1.25
      a-Al2C>3 support

11    1100'C H.T.  granular     1.25
      a-A!203 support

12    86.9% A1203, 11.6%      3.25
      SiC>2 support
      (Norton SA5203)

13    96% A1203,  4% SiO       1.25
      support (Carborundum
      SAHT-96)

14    7.8% CaO in SAHT-96      1.25

15    94% (Zr02 + Hf02),       1.25
      3.5% CaO, 1.6% Si02
      support (Norton 5264)
2.2


2.2
2.2



2.2

4.2
90.3


97.4


60.4
97.1

54.7
10


 3


40
                100
           (15.4 in 1 hr)
 2.9

45.3
 The percent loss  determined as grams of overhead material xlOO/gram original material.
.All overhead material was smaller than 70 mesh.
bRun
     1 repeated  to make comparison easier.
 Bench-scale results.

-------
                                    133


high loss rate found on a bench-scale (4-5%/hr, run 8a).   The fluidization
conditions in the laboratory-scale and bench-scale vessels are not the same;
the bench-scale results include  the loss due to attrition in the feed
line.  The bench-scale results agree more closely with laboratory results
for half-calcined dolomite (run 1) .   The regenerated dolomite had a high
materials loss rate (run 9), as was also observed in the bench-scale
operation (run 9a).  However, due to the short residence times needed for
regeneration (5-10 min) , the loss rate per cycle is only about 2%.

Attrition of Supports and Synthetic Sorbents

     Runs 10, 11, 12, 13, and 15 were tests of support material for synthetic
sorbents.  The a-Al203 support had the lowest materials loss rate in com-
parison to the other supports studied.

     The addition of Si02 was expected to strengthen the supports givea the
same physical properties (porosity, surface area).  Al203~Si02 supports were
tested (runs 12 and 13) and were expected to be more attrition-resistant, but
this was not the case.  It is speculated that the high attrition rates for
the Al203-Si02 supports are due to larger-diameter pores.  Silicon dioxide,
moreover, may not be desirable as a strengthening agent because it reacts
with CaO to form stable calcium silicates.

      It was shown that large pores in a-A!203 caused high attrition rates.
 This was the result in runs 10 and 11.   The 1350°C heat-treated a-A!203
 support had three times the attrition rate of the 1100°C H.T. a-A!203
 support.   Higher heat-treatment temperatures produce larger pores.

      Porosity measurements will be made on Norton SA5203 and Carborundum
 SAHT-96 Al203~Si02 supports.   If they are found to contain large pores,
 Si02-containing supports will be tested that have smaller pores than these
 two supports.

      The Norton 5264 support (run 15) had a high attrition rate,  probably
 due once again to large pores.   The compounds Zr02 and Hf02 is this support
 material are expensive and are not  likely to be considered as support
 materials.   However,  IV-B transition metal oxides in small concentrations
 may strengthen a support.   Calcium titanates react with S0225 and may
 lower the sorbent attrition rate.   A search for a commercial .source for
 a-A!203-Ti02 material has been unsuccessful.   The possibility of  preparing
 small quantities of this support at Argonne is being assessed.

      One synthetic sorbent,  7.8% CaO in 96% Al203-4% Si02 (run 14)  was
 tested for  decrepitation.   In this  test, attrition rate was substantially
 smaller  than that of the original support material (run 13); attrition
 was reduced by a factor of approximately 50 to a 3% loss of material in
 10 hr.   If  the same factor of 50 degrees in attrition rate could  be
 realized  for  the a-A!2C>3 supports (runs 10 and 11) by adding CaO, then
 synthetic  sorbents would have a substantially higher attrition resistance
 than do  natural sorbents.   Synthetic sorbents, CaO in a-A!203, are being
 prepared  for  attrition testing.

-------
                                     134


                     COMBUSTION-REGENERATION CHEMISTRY*
[P. Cunningham (Principal Investigator), B. Hubble, S. Siegel]

     The fundamental aspects of the chemical reactions associated with
the cyclic use of a sorbent material in sulfur removal processes, such
as fluidized-bed combustion, are being studied to provide information that
will suggest how these reactions might be optimized.  Dolomite No. 1337
has been chosen as a model sorbent system for these studies.  The program,
which emphasizes chemical kinetic measurements, also includes studies
of the structural changes that take place in the dolomite sorbent.
Kinetic studies are based on the use of a thermogravimetric (TGA)
technique, and structural studies are based on X-ray diffraction and micro-
scopy techniques.

     Previous reports have described results pertaining to the sulfation
reaction of half-calcined dolomite:17'1

          [CaC03-MgO] + S02 + 1/2 02 -> [CaSO^-MgO] + C02                (1)

the half-calcination reaction reaction:

          [CaC03-MgC03J —*• [CaC03-MgO] + C02                           (2)
                        C02

the formation of the Ca-Mg binary sulfate:

          [CaC03-MgC03] + S02 + 1/2 02T"*"
                                       C02
                            x[CaS04'MgO]  + y[CaMg3(SO^)^]  + C02         (3)

the two-step regeneration method

          [CaSO^-MgO] + 4H2 -> [CaS-MgO] + 4H20                          (4)

          [CaS-MgO] + C02 + H20 + [CaC03 MgO] + H2S                     (5)

and the feasibility of the solid-solid reaction between CaSO^ and CaS
as a regeneration method
          3[CaS04-MgO] +  [CaS-MgO] ->• 4[CaO-MgO] + 4S02                  (6)

Additional information is reported herein on the processes described by
reaction 2 and 6.

Half-Calcination Reaction

     In an earlier report,1 results were described of kinetic studies in
which this reaction was carried out at 1-atm pressure in an environment
of 100% C02 over the temperature range, 640 to 740°C.  This investigation
has been expanded to include:   (a) measurements under the 100% C02 environ-
ment at 800°C, and (b) measurements under an environment of 40% C02-60% He
 The work reported in this section is supported in part by the U.S. Energy
 Research and Development Administration, Division of Physical Research.

-------
                                    135
over the temperature range 640 to 800°C.  These kinetic results are summarized
in Fig. 58* and 59, where percent conversion is plotted against time.   From
Fig. 58 and 59, it is apparent that the kinetics of this reaction are
dominated by a temperature effect.  The reaction rate increases rapidly with
increasing temperature for both 100% and 40% C02 environments, particularly
above some temperature near 700°C.  The effect of C02 concentration is more
subtle.  In general, the rate is higher in the 40% C02 environment, but
this effect becomes less pronounced at higher temperatures.

     The results of X-ray diffraction analysis of stones from the 40% C02 -
60% He environment experiments were in agreement with the earlier reported
results for the 100% C02 environment experiments.   That is, the size and
degree of preferred orientation of the CaCOs pseudo crystals formed during
this reaction are dependent on the kinetics (temperature) of the reaction.
Larger crystals and greater degrees of preferred orientation result when
the reaction proceeds slowly (at lower temperatures); crystals are smaller
and there is less preferred orientation when the reaction proceeds rapidly
(at higher temperatures).

     Additional experiments in which conditions were selected on the basis
of the above results have been performed to provide a series of half-calcined
samples for optical microscopic examination.  These experiments are sum-
marized in Table 42.  Polished sections of these samples were examined in
reflected polarized light and were compared with sections of untreated
dolomite particles.  The results of this study are summarized as follows.

     Untreated Dolomite.  Crystals in any one particle are equigranular and
coarse-grained, but the average size varies, for different particles,
from 0.04 to 0.3 mm.  Grain boundaries are sharp, and optical extinction
is uniform across each grain.  Occasionally, a grain contains small well-
defined occluded grains whose crystal orientations differs.

     Sample from Experiment I.  Dolomite grain boundaries became slightly
diffuse, and the extinction across grains was observed to be interrupted
by patchy inclusions of irregularly shaped grains whose crystallographic
orientation differed from that of the surrounding grains.  The sample was
then mildly etched with 0.02N HCL for 5 s.  This treatment serves  to
distinguish calcite from dolomite (dolomite surfaces are unaffected, but
the calcite surfaces become etched).  Comparison of the section before and
after etching shows that the onset of calcite formation occurred both
along dolomite grain boundaries and within the dolomite grains as  patchy
areas.  The shapes of the etched areas correlate very well with those of
patch inclusions observed in polarized light on the unetched section.
Calcite growth may begin from the edge of a dolomite grain inward  but may
also begin within a grain.

     Sample from Experiment II.  This sample, heat-treated for 60  hr longer
than was the sample from experiment I, shows all of the above-described
*t                                            i
 Figure 58 includes data reported previously,  along with new data for
 800°C conversion.

-------
                           % CONVERSION  OF  CCA Me (00,^1   TO


                                                   8
  09
  oo
o o
rt (D
H- 3
O ft
3
  o
e o
0 3
cu 
H M
  co
I-1 H-
O O
O 3
5*

n co
o
  Hi
  O
3 M
rt HI
  1
  n
  3
  OJ
  rt
  H-
  O
  3

-------
                              % CONVERSION OF ICAM6(C03)e]  TO  CCAC08*
  H-
  00
(D fD
CD i-!
O O
rt CD
H- 0

§n
  o
e o
0 0
& 3
to n>
H H
  CO
*> H-
O O
?-S 0

o si
O to
ro
 I H
&•« CD


(D O



!g
H> H
l^ Ml
o i
0 n


§&
rt H-
•  0
H    -
m
H
PI

CD

-------
                                     138


     Table  42.  Half-Calcination Experiments on 1337 Dolomite.
Sample % Conversion to Calcite
Experiment History (from TGA results)
I
II
III
IV
V
100% C02,
60 hr
100% C02,
120 hr
40% C02,
0.3 hr
40% C02,
6 hr
40% C02,
26 hr
, 640°C,
, 640°C,
800°C,
800°C,
800°C,
50
75
100
100
100
effects, but to a greater degree.  The original dolomite grain boundaries
are still present, but they are more diffuse.  An etched section under
the microscope appears to be 75% converted to calcite, in agreement with
the TGA results.  The MgO is too fine grained to be visible in reflected
light.

     Sample from Experiment III.  Calcite crystals have grown to diameters
of 5 ym, and they form a mosaic-like texture.  Some preferred orientation
is present; the direction of orientation of the dominant component differs
from one area to another.  These well-defined areas correspond dimensionally
to the grain outlines of the original dolomite crystals, even though each
dolomite grain has been completely transformed to imany calcite crystallites
which assume an unknown, but preferred, orientation differing from that in
the original dolomite grains.  This preferred orientation was also observed
in X-ray powder patterns.

     As was observed for the sample from experiment II, MgO is too fine
grained to be seen in polished section, but its distribution might materially
affect subsequent sulfation reactions since it is more resistant to
sulfation than is CaO or CaCOa.  However, in some preliminary SEM studies
on this sample, it was found that MgO is uniformly distributed throughout
the half-calcined particles.  It neither migrates nor forms at grain
boundaries of calcite or relic grain boundaries of dolomite, but is occluded
in the calcite structure.

     Sample from Experiment IV.  This sample, like the sample from experi-
ment III, was completely recrystallized into calcite.  Unlike the sample
from experiment III, relic dolomite grain boundaries have been completely
obscured in this sample by recystallization into randomly oriented calcite
crystals which have an average diameter of 20 ym.  The X-ray results confirm
this loss in preferred orientation.

-------
                                    139


     Sample from Experiment V.  The 20 hr of additional heating (in comparison
with the preceding sample) had no major effect except a slight indication
of enchanced calcite crystal growth.

     Discussion.  The examination of the above samples by petrographic
methods has yielded significant results.  It has been shown that dolomite
crystals are transformed to calcite along grain boundaries, as well as
within dolomite crystals.  This suggests that greater efficiencies in the
half-calcination process may be achieved by using finer-grained starting
material from the quarry, if this is practical.  In addition, it has been
shown that the half-calcined grain structure is completely destroyed by
heat treatment.

Regeneration by the CaSO^-CaS Reaction

     Additional experiments have been performed to establish whether reaction
6 constitutes a practical regeneration scheme.  Work included:  (a) an
investigation of the yields obtained when the composition of the starting
material for this reaction is varied, and (b) a kinetic study of the reaction
as a basis for estimating the type of the kinetics associated with the
reaction system.

     Effect of Starting Material Composition.  A number of experimental
variables could affect the progress of reaction 6; for example, the composi-
tion of reactants (CaSO^/CaS ratio in the dolomite stones) and the reaction
temperature.  The initial experiments1 illustrated that the reaction proceeded
at a temperature as low as 950°C.  Accordingly, a number of experiments
were carried out at 950°C with different starting compositions of CaS04 and
CaS in the dolomite stones to see if the yield of CaO is affected.

     The experimental procedure employed was as follows:  A large stock
supply of sulfated dolomite was prepared by half-calcining the stones and
subsequently sulfating in a 4% S02-5% 02-N2 40% C02 simulated flue-gas mixture
until the gaining of weight halted.  Stones from this stock supply were used
in every experiment.

     The starting materials for experiments to study the solid-solid reaction,
Eq. 6, were always prepared by the same procedure, -I.e., aliquots of the
sulfated dolomite stock supply were reduced to the desired extent at a re-
action temperature of 880°C, using a gas mixture containing 3% H2 and the
balance helium.  Starting materials that had sulfide concentrations of 24
to 39% (TGA analysis) were prepared.

     In addition, all sulfate-sulfide reaction experiments were performed
at the same conditions, 'I.e., under 1-atm partial pressure of helium at
950°C.  Most sulfate-sulfide reaction experiments were run for 5 hr, but
some were run overnight.

     X-ray diffraction analyses31 were performed on samples of the starting
material and on samples of the products of all sulfate-sulfide reaction
experiments.  For each analysis, an aliquot of 30 to 50 stones was analyzed
to minimize sampling error problems.

-------
                                     140


      The results of the experiments are summarized in Table 43.  The first
 column identifies"the experiment and the sample analyzed by X-ray diffraction
 and the second column gives the sample history.  The third through sixth
 columns list TGA results.   The third column gives the percentage of the
 available material in the  stones that has been converted.   The fourth,  fifth,
 and sixth columns give the compositions of the sample as mole percentages.
 The seventh through ninth  columns list the X-ray diffraction results as mole
 percentages.

      Varying the composition of the starting material for the solid-solid
 reaction had the following effects:  In all cases,  the increase in CaO
 concentration as a  result  of solid-solid reaction,  was substantial—that is,
 in the 30 to 60% range.  However,  it is apparent from columns 3 and 7 that
 a  starting material containing greater than 25% sulfide (25% sulfide is
 suggested by the stoichiometry of Eq.  6)  is necessary to approach higher
 yields (i.e.,  consumption  of most of the CaSO^  and  CaS before reaction
 stops).   For example,  in Experiment II, in which the starting material
 contained 25% of the CaS form,  the CaS was completely consumed and a sub-
 stantial amount  of  CaSOn remained when reaction ceased.   From these
 preliminary results,  a starting material containing about  35% of the reduced
 form is  believed necessary to force the solid-solid reaction toward
 completion at  a  reaction temperature of 950°C.

      As  indicated in the preceding report  in this series,1  a method was
 developed for  quantitative measurements of various  phases  present  in the
 X-ray diffraction patterns.   A  description of the X-ray  technique  for this
 purpose  appears  elsewhere.31  The  reason  for developing  this technique  is
 related  to the fact  that a TGA  measures only weight  changes  and  use relatively
 small samples.   Accordingly,  the  TGA technique  requires  the  use  of an
 independent chemical analysis technique to identify  the  chemical reaction(s)
 associated with  the weight  changes  being monitored.   For this  reason, it  is
 of  interest to compare TGA and  X-ray results  for the  composition of the
 phases in the stones.  The sulfation and reduction  reactions  are understood
 and  therefore the weight changes observed  in these  two reactions can be
 correlated with  the  compositions of  the stones.  If  the X-ray  technique
 agrees with the  TGA  results  for these  two  reactions,  the X-ray technique
 can be used to test whether  the correct chemical reaction  is being assigned
 to TGA weight changes observed  in  solid-solid reaction.

      Comparison  of TGA and X-ray results for  the sulfated and reduced stones
 (Table 43, columns 4 and 7 and columns  5 and  8) in each of the experiments
 suggests  that the two methods are generally  in agreement.  The agreement for
 the amounts of CaO and CaSOi^ in the  samples  are extremely good.  However,
 the agreement between the  two methods  for  the amount of CaS present  (columns
 6 and 9)  is poorer than desired in each case, the X-ray result is  lower than
 the TGA result.  Any of the  following  factors might lead to  experimental
 errors in  the X-ray technique that could explain the lack of agreement  in
 the CaS results:  (a) the use of peak heights instead of integrated  inten-
 sities,  (b) the presence of amorphous materials, and  (c) complete masking of
 usable CaS lines by CaSOtj. lines, leading to substantial errors in  the CaS
measurements.  In general,  the agreement of the two techniques is  surprisingly

-------
              Table 43.   Summary of TGA and X-Ray Diffraction Results for
                         Solid-Solid Experiments at 950°C
Experiment
Sample
Number
I
II-A
II-B

II-C

III-A
III-B

IV-A
IV-B

V-A
V-B

VI-A

VI-B

Sample
History
Stock Material
Partical Reduction
Sulfate-Sulfide
Reaction, 5 hr
Sulfate-Sulfide
Reaction, Overnight
Partial Reduction
Sulfate-Sulfide
Reaction, 5 hr
Partial Reduction
Sulfate-Sulfide
Reaction, 5 hr
Partial Reduction
Sulfate-Sulfide
Reaction, Overnight
Partial Reduction

Sulfate-Sulfide
Reaction, 5 hr
Available
Material
Converted
92% sulfated
25% reduced
b

— b

32% reduced
— b

36% reduced
-b

36% reduced
— b

40% reduced
t
—

Composition — TGA Composition — X-ray
Rp.sults Diffraction Studies
CaO CaSOit CaS CaO
(mol %) (mol %) (mol %) (mol %)
15 85 0 11
15 64 21 13
40

— — 58

15 38 27 12
58

15 54 31 15
43

15 54 31 15
— 68

15 51 34 10
/.Q
— — — — — ~ f y

(mol %)
64
62
42

21

66
12

38
19

58
7

57



CaS
(mol %)
4
a
a

a

10
7

16
	 a

16
4

21
13


aCaS lines in X-ray patterns were  either not  detectable or, when detectable,  were not intense

.enough to measure.                                                                      .
bDue to the manner in which these  experiments were performed,  it was not  possible to monitor

 sulfate-sulfide reaction progress from TGA weight changes.

-------
                                     142


 good and it is concluded that the X-ray technique results can be used to
 check whether the correct reaction is being assigned to TGA weight changes.

      Solid-Solid Reaction Kinetics (Partially Reduced Starting Material).
 The above studies illutrate that  the yield  of the solid-solid reaction
 depends on the composition of the starting  material,  suggesting that  it
 should be possible to identify conditions which maximize thevyield.   If  the
 reaction is to be considered in a regeneration scheme,  the other important
 reaction property to  be considered is kinetics.   Even if the yield is
 acceptable,  slow kinetics would result in the reaction not being a practicable
 regeneration candidate.   To gain  an appreciation  for the salient nature  of
 the kinetics associated with this reaction,  a kinetic study was done  on
 sulfated dolomite which had been  36% reduced.

      The experimental procedure followed was similar  to that  described
 above.   A large stock supply of sulfated dolomite 1337  was prepared by
 half-calcining the stone and subsequently sulfating  at  750°C  in a 4%  S02-
 5%  02  simulated flue  gas mixture  until the weight gain  process  halted.
 The percentage of sulfation achieved was determined by  performing a reduction
 experiment on an aliquot of the sulfated stones at 880°C with a 3% hydrogen
 in  helium.

     A stock supply of  partially  reduced stones for use in the  kinetic
 experiments  was prepared by a  reduction reaction  at  880°C,  using a 3%
 hydrogen in  helium mixture.  This  material was used in  all kinetic experi-
 ments  which  were performed  at  945°C  under 1-atm partial pressure of helium
 flowing at a rate of  300 cm3/min  to  remove the S02 formed.  In  each kinetic
 experiment,  the starting material  was  placed  in the TGA apparatus, which
 was  at  945°C with the helium purge flowing,  and the weight  change was
 monitored  until the experiment was halted.   Reaction  times were  1/4, 1/2,
 1,  2,  5 1/2, and 18 hr.   X-ray  diffraction analyses were obtained  on aliquots
 (30  to  50  stones)  of material  samples  at the  end  of each kinetic  experiment.

     The results  of these experiments  are summarized  in Table 44.  The first
 column  identifies  the experiment,   and  the second  column gives the  samples
 history.   Columns  three  through seven  lists  the  results based on analysis
 of TGA  measurements.  The third column  lists  the  percentages of  the available
 material in  the  sample that have been  converted as a  result of the history
 of the  samples.  The fourth  through  sixth columns give  the compositions  of
 the  samples  as mole percentages of CaO, CaSO^, and CaS, respectively.

     The seventh column lists calcium  species material balances  as total
 percentages of  calcium species existing in the stones.  The X-ray diffraction
 results are given  in the eighth through eleventh  columns, with columns eight
 through ten listing the mole percentages of CaO,   CaSO^, and CaS, respectively,
 and  the last column giving the material balances  for  the calcium species.

     From Table 44, it is apparent that for each  experiment, extremely good
 agreement exists for TGA and X-ray analyses of CaO and CaSO^ concentrations
 in the material  (columns 4 and 8,  and columns 5 and 9, respectively).   The
values are in good agreement, and  values from both techniques show a similar

-------
Table 44.  Results of CaSOu-CaS Reaction Kinetic Measurements at 945°C.   Starting
           material was  prepared by partial reduction of sulfated dolomite.
Expt.
No.
I


II


III


IV


V


VI


VII


VIII


Sample
History
Sulfated 1337
Dolomite

Partially Re-
duced Stones
0 hr
0.25 hr
Reaction

0.50 hr
Reaction

1.0 hr
Reaction

2.0 hr
Reaction

5.5 hr
Reaction

18.0 hr
Reaction

Stone
% of
Available
Material
Converted
91% of
CaC03
Sulfated
36% of CaSOi<.
Reduced

20% of CaSOk
Converted
to CaO
26% of CaSO^
Converted
to CaO
29% of CaSO^
Converted
to CaO
32% of CaS04
Converted
to CaO
40% of CaS04
Converted
to CaO
38% of CaS04
Converted
to CaO
Composition - TGA
CaO CaSO^
mol % mol %
16 84


16 53


30 43


34 40


36 38


38 37


45 32


43 33


Results
CaS
mol %
0


31


27


26


26


25


23


24


Stone Composition
E % Ca CaO CaSO^
Species mol % mol %
100 17 90


100 23 60


100 29 39


100 29 47


100 31 46


100 27 50


100 35 38


100 46 30


- X-ray
CaS
mol %
0


20


16


8


10


7


5


7


Results
E % Ca
Species
107


103


94


84


87


84


78


83



-------
                                    144
trend of change as the reaction proceeds.  As for the earlier reported
results, the results of the two techniques for amount of CaS do not agree
as well as do the results for CaO and CaSOt,..  However, there is agreement
for the trend of change as the reaction proceeds.

     The chemical kinetics of this reaction are summarized in Fig. 60, where
mole percentage of CaO in the stones is plotted against reaction time.  Each
mole percentage CaO value in this figure is an average of the values from
the TGA and X-ray analyses listed in Table 44.  The CaO content of the
stones .increased from 20% to 45%.  The yield of the reaction (amount of
CaO formed as a result of reaction) was not as great as that obtained in
some earlier reported experiments, where it was shown that yield is
dependent on the composition of the starting material .  However, the point
to be emphasized is that the rate of the solid-solid reaction was sur-
prisingly high, that is, the CaO content increased from 20% to 40% in less
than six hours; stated in another manner, the reaction reached 80% completion
in less than six hours.  Such a reaction rate is comparable to sulfation
reaction rates which were reported earlier.17  Accordingly, these results
are encouraging from the point of view of considering the solid-solid
reaction as a candidate for a regeneration scheme.
    0      2


      Fig. 60.
12
14
  4       6      8      10
                 TIME, hr

CaO Content as a Function of Reaction Time.

-------
                                    145


                          COAL COMBUSTION REACTIONS
S. Lee (Principal Investigator), C. Turner

Determination of Inorganic Constituents in the Effluent Gas from Coal
Combustion

     In the operation of coal-fired boilers and gas turbines, it has been
observed that some chemical elements carried by the combustion gas cause
severe metallic corrosion, as of turbine blades.  A study has been initiated
to determine quantitatively which elements are present in the hot combustion
gas of coal, in either volatile .or particulate form, and to differentiate
between volatile and particulate species.  Identification of the compound
forms and  amounts of particulate species and determination of the amounts
and the forms of condensable species are of interest.

     An outline of the experimental program and a general  description
of the laboratory-scale batch fixed-bed combustor designed for this study
were presented in the preceding report1 in this series.  The combustor was
fabricated in the central shop at ANL and is now being assembled.

     Modification of Conceptual Design of Equipment.  The combustor was
fabricated according to a fabrication specification and work plan that
included a manufacturing plan, a test plan, and a quality control plan._
Prior to approval of the engineering drawings for fabrication, the drawings,
specifications  for  fabrication, and stress calculations  supporting the.
safety of the design were reviewed by a design/preliminary safety review
committee of Chemical Engineering Division.  The purpose of  this review
was  to determine if the design of  the combustor met safety requirements.
At the review meeting, several recommendations  pertaining to design  and
operational  safety were made.  Based on  these recommendations, the design
was  revised.  The revisions  are described below:

      The thermal stresses  from heating and axial thermal gradients at both
the  intersection of the preheating arid  combustion  sections and  the  inter-
section of  the  filtration and cold trap  sections have been considered.
The  thermal stresses at both intersections have been  reduced by  the  following
modifications of the design:

      1.  At the preheating section,  the  cooling coils have been removed.
          Instead, a 1/8-in.-thick  Fiberfrax insulation layer will be
          inserted between the internal heater and  the inner  wall of  the
          pipe.   This  insulation layer  will decrease heat flow radially.
         Heat  transfer  calculations indicate that  the wall temperature
          of the pipe at this section will be only  about 260°C (500  F)
          and that  the temperature  difference between the inner wall and
          the outer  wall will be about  28°C (50°F).   The thermal stresses
          due to this thermal gradient  have been computed and are shown
          to be insignificant when compared with the maximum allowable
          stress of   the material at that temperature.  In addition,
          since no restraint to thermal expansion is imposed on the pipe
          at this section, no thermal stress due to axial thermal gradients
          are expected.

-------
                                      146


      2.  At the cold trap section, the number of turns of the cooling
          coil has been reduced and there is now a 'W-in. transitional
          section between the end of the hot filtration zone and the
          point where the.cooling coils start.  Heat transfer calcul-
          ations show that the average wall temperature at the point
          where the cooling coils start will be about 150°C (300°F).
          The 310 stainless steel is strong enough at this temperature
          to withstand the thermal stresses caused by a thermal gradient
          through the wall as large as 50°C (90°F).   The thermal
          gradient will not each 50°C at this section of the pipe.

      3.  Two more metal supports, one at  each end,  have been added to
          hold the ends of the combustor,  where heavier loads are
          located.   There are total of five mechanical supports along
          this 6 1/2-ft-long combustor.  Because there is a rather
          short span between supports,  the mechanical support stresses
          are negligible in comparison with other types of stresses.
          For example,  the  filtration section has the greatest span;
          therefore,  the maximum flexural  stress due to a bending
          moment is expected at  this  section.   Stress calculations
          indicate  that  this maximum  flexural stress is only of the
          order of  a  few pounds  per square inch.

          The table on which the combustor will  be supported has  been
 constructed  so that  the axis of the  combustor body  is  tilted  5°  downward
 (from the horizontal) toward the  gas discharge  end;  connections  will so
 oriented  that  water  will drain  out if  there  is  a minimum accumulation.
 Air  (not  oxygen as was  stated in  the previous report)  will  be  the  gas
 introduced into the  combustor upstream  from  the hot  filter  to  complete the
 combustion of  organic volatiles.   The design of  the  combustor  is now
 considered adequately safe.  Except for one recorder and  one gas analyzer,
 all  of  the auxiliary components needed  to  complete  the assembly  of  the
 combustor system are on hand.

     ^A  schematic diagram of  the batch fixed-bed  combustor system is  shown
 as Fig. 61.  The safety aspects of the operation are described in  the
 following sections.

     Gas System.  The oxygen and air will be fed into  the system from high-
 pressure cylinders; nitrogen gas will be obtained from the house nitrogen
 supply.  For each gas line, a pressure-regulating and  shutoff valve, a
 silica-gel dryer, a  flowmeter, and a normally closed (N.C.) solenoid valve
 are installed.  The solenoid valve will be activated (i.e., opened) during
normal operating conditions, and will be deactivated (i.e., closed) whenever
 any of the following abnormal operating conditions occur:   (1) the temper-
ature of the combustor wall exceeds 900°C  (1650°F) or  the internal pressure
 exceeds 2 atm  (absolute) or  (2) the flow rate of cooling water to cold traps
 is less than 2.5 gal/min.  When a solenoid valve is  closed (deactivated),  it
not only stops gas input to the system but also prevents any backflow of
 combustion gas to gas supply lines.

-------
                          147
         GAS INPUT
                             RELIEF VALVE
                                 VENT     COOLING
                                          WATER IN
          N.C. SOLENOID
             .VALVE
                 OOOOO  x
                      I  (TI
                      TUBULAR FILTER
                      TO EFFLUENT
                     GAS ANALYZERS
             INDUCTION
            HEATING UNIT
                                        COOLING
                                      WATER OUT
         VENT
           SAMPLING
             PORT
                                     CONDENSOR
                           BACK PRESSURE
(PC) PRESSURE GAUGE         REGULATOR
§    PRESSURE SWITCH
    INTERLOCKING FLOW SWITCH
,-M- NEEDLE VALVE
-[XV CHECK VALVE
    ON-OFF  BALL VALVE
    .FLOW METER
    DRYER
                                   SEPARATOR
                                                f
                                              LIQUID
Fig. 61.  Schematic Diagram of Batch Fixed-Bed Combustion System.

-------
                                    148
     Combustor.  The conceptual design of the combustor unit was presented
in the preceding report1 in this series.  To maintain system safety, several
feastures are provided in the combustor.  Thermocouple burnout protection
and limit-control protection are installed for the preheater.  The wall
temperatures of the combustor in both the combustion and the filtration
sections are measured by thermocouples, and if the wall temperature in
either section should exceed 900°C (1650°F), all input power would be
shut off and the solenoid valve controlling the inlet gas for the section
would be deactivated.  In addition, a pressure switch installed on the
combustor is used to shut down the entire system if the system pressure
should exceed two atmospheres absolute (the design pressure of the
combustor).

     Cooling-Water System.  Cooling water from the laboratory cooling-
water supply is fed to the induction heating unit, cooling coils, and cold
traps in the system.  For the cold-trap feed line, a check valve is
connected to prevent water backflow into the supply lines, and for the
discharge lines, an interlocking flow switch (IFS) is installed to ensure
sufficient cooling-water flow into the cold traps.  The switch is so
connected that whenever the water flow is lower than the preset minimum
limit of 2.5 gal/min, all electrical power inputs to furnaces, the preheater,
and the induction heating unit will be shut off, and all solenoid valves
will be deactivated.  Thereby, this IFS protects the cold traps from
overheating that could cause a failure of the cold trap, resulting in
steam generation.

     Downstream System.  Downstream from the combustor, the combustion
gas is further cooled in a condenser and flushed into a separator.  The
gas stream leaves the separator and is reduced in pressure by passage
through a back-pressure regulator.  Then a small fraction of the low-
pressure combustion gas is directed to gas analyzers and the rest of the
combustion gas is vented to the building exhaust.  A sampling port is also
provided in this downstream section.

Systematic Study of the Volatility of Trace Elements in Coal

     The effluent gases from coal combustion are known to contain trace
elements, some of which cause severe metal corrosion in coal-fired boilers
and gas turbines.  The evolution of these trace elements has been generally
believed to be due to the volatilization of elements and compounds at high
temperatures.  Therefore, knowledge of the temperatures at which these trace
elements start to volatilize and of the rates of volatilization is important
for the utilization of coal.  The purpose of this study was to obtain data
on the volatility of these elements under coal combustion conditions.

     The experimental setup for this study was presented in  the preceding
report1 in this  series.   In this report, the experimental results obtained
along with some  conclusions are reported.

     In this  study,  340°C ash samples were heated  in a tubular furnace to
various temperature  ranges in a gas flow of 0.6  scfh for a specified period
of time.   The  340°C  ash samples were prepared by ashing -200 mesh Illinois

-------
                                    149


Herrin No. 6 Montgomery County coal in an air flow of 9 liters/min for
48 hr in a muffle furnace.  This step removed the carbon and concentrated
the trace elements.

     The resulting ash residue was analyzed for elements of interest.  In
most of the experiments, ^1-g ash samples were used.

     Three series of experiments were completed in this study.  In the first
series of experiments, each ash sample was heated in an air flow to a
temperature between 540°C and 990°C for 24 hr; in the second series of
experiments, each ash sample was heat-treated to temperatures between 850
and 1250°C for 20 hr in either air or an oxygen-enriched air flow.  The
experimental results obtained from these experiments have been reported  '
and are summarized in Tables 45-47.  These tables show that most of the
chlorine in coal evolves at temperatures below 640°C and that the
elemental concentrations of the metallic elements investigated are constant
up to 1250°C, indicating that these metallic elements generally remained
in the fused ash up to 1250°C under the oxidizing environments.  Also,
the retention of these elements in the fused ash is not affected by the
oxygen concentration in the flow gas, as is shown in Table 47.  Under the
oxidizing conditions present during heat treatment of these ash samples,
no loss of sodium or potassium was observed.

     The third series of experiments was conducted to investigate the
volatility characteristic of trace elements in coal under wet oxidizing
conditions (to simulate coal gasification conditions).  In the Series-3
experiments reported here (wet oxidizing conditions), each 340°C ash
sample was heat-treated at a temperature between 850 and 1250°C for 20 hr
in water vapor-bearing air at a flow rate of 0.6 scfh.  The water vapor
was introduced into the system by bubbling the air through a heated evapor-
ator.  The ratio of water vapor to air in this series of experiments was
0.13 by weight, which is in the H20/air ratio range of maximum efficiency
in the coal gasification processes.

     The weight losses of the 340°C ash due to heat treatments in Series-3
experiments are given in Table 48.  These data are also plotted in Fig. 62,
along with the weight-loss results obtained in Series-2 experiments
(presented in Table 18 of Ref. 33).  The weight loss was greater in humidified
air than in dry air, indicating that water vapor in the combustion gas
affects the evolution of certain substances in the 340°C ash.  The effect
at lower temperatures is greater than at high temperatures.

     The X-ray fluorescence (XRF) method was used to analyze elemental
concentrations of ash samples from Series-3 experiments.  In this method,
a 0.25-g ash sample was mixed with an equal amount of lithium tetraborate
binder in a plastic vial in the "Wig-L-Bug" grinder.  The mixture was then
pelletized under a pressure of 10,000 psi to form a 5/8-in.-dia pellet.
When this procedure was applied to hard fused ash samples, this material
was pulverized to -100 mesh powder in a tungsten carbide vial  (which
contained a ball) in the "Wig-L-Bug" grinder prior to being mixed with
the binder.  The pellet prepared in this way was found to be homogeneous,

-------
        Table 45.   Elemental  Concentration   of High-Temperature  Ash Corrected for Weight
                   Losses  at  the  Stated Temperatures.  Ash  prepared at  542  to 990°C by
                   heating for  24 hr  in air  flow of  0.6  scfh.
Heating Temperature (°C)





Be
Pb
V
Cr
Co
Ni
DDm
FP Cu





Zn
Mn
Hg
Li
Cl
r AI
Fe
1 Na
7 Mg
K
Ca
Ti
340
6.5 +0.6
15+2
210+50
90+9
17.0 +0.8
52+5
130+10
540+30
500+20
0.01
59+3
90+10
4.6 +0.2
10.5 +0.5
0.74+0.04
0.28+0.01
1.08+0.05
4.9 +0.2
0.6 +0.1
542
6.3 +0.6
26+3
200+50
91+9
16.9 +0.8
52+5
110+10
510+30
550+30

64+3
91+10
5.7 +0.3
10.7 +0.6
0.75+0.04
0.29+0.01
1.26+0.06
4.3 +0.2
0.6 +0.1
640
6.2 +0.6
29+3
230+60
87+8
17.1 +0.9
53+5
110+10
520+30
540+30

62+3
22+10
5.4 +0.3
11.1 +0.6
0.87+0.04
0.33+0.02
1.21+0.06
4.2 +0.2
0.5 +0.1
740
6.2 +0.6
29+3
150+40
100+10
15.4 +0.8
62+6
120+10
460+20
680+30

63+3
17+10
8.1 +0.4
11.2 +0.6
0.79+0.04
0.54+0.03
1.24+0.06
6.4 +0.3
0.4 +0.1
840
6.8 +0.7
54+5
130+30
94+9
14.0 +0.7
60+6
120+10
480+20
680+30

63+3
17+10
7.7 +0.4
11.3 +0.6
0.80+0.04
0.42+0.02
1.25+0.06
6.0 +0.3
0.4 +0.1
940
7.2 +0.7
10+1
150+40
93+9
14.4 +0.7
66+7
126+10
560+30
660+30

63+3
13+10
8.1 +0.4
11.5 +0.6
0.81+0.04
0.40+0.02
1.29+0.06
5.6 +0.3
0.4 +0.1
990
6.3 +0.6
21+2
130+30
170+20
13.9 +0.7
100+10
130+30
560+30
670+30
—
66+3
21+10
8.0 +0.4
11.5 +0.6
0.76+0.04
0.40+0.02
1.28+0.06
6.2 +0.3
0.5 +0.1
By atomic absorption.  Each precision is based on an estimate of the precision of
measurement obtainable with standard solution.

-------
        Table 46.  Elemental Concentrations in High-Temperature Ash Calculated on the
                   Original 340°C-Ash Basis.  Ashes prepared at 850 to 1250°C by heating
                   for 20 hr in an air flow of 0.6 scfh.
Heating Temperature (°C)

340
lement AAa
Fe
Al
Na
i K
Mg
0
Ca
Ti
- Zn
Mn
Ni
Co
pm Cu
r
I Cr
Li
I V
aAA
11.9+0.6
9.5 +0.5
1.00+0.05
1.55+0.08

2.5 +0.1
1.4 +0.3
380+20
360+20
70+4

110+10
150+20
75+4-
290+70

NAAb
11+2

0.8 +0.2
1.4 +0.3




380+80

17+3

120+25


- Atomic absorption; each
850
AA
11.8 +0.6
9.7 +0.5
1.14+0.06
1.62+0.08
0.36+0.02


390+20
340+20
70+4
20+1




precision is
950
AA
11.4 +0.6
9.6 +0.5
1.04+0.05
1.59+0.08
0.36+0.02
2.5 +0.1
1.3 +0.3
420+20
300+20
72+4
17+1
130+10
170+20
77+4
280+70
based on an
. 1100
AA
11.4 +0.6
10.0 +0.5
1.15+0.06
1.56+0.08
0.35+0.02


400+20
300+20
74+4
20+1




estimate of
1150
AA
11.2 +0.6
9.8 +0.5
0.97+0.05
1.5 +0.08
0.34+0.2


400+20
320+20
69+4
19+1




the precision
1200
AA
10.9 +0.5
8.6 +0.4
0.94+0.05
1.40+0.07
0.30+0.2


400+20
320+20
68+3
18+1




of
1250
AA
11.8 +0.6
9.9 +0.5
1.13+0.06
1.54+0.08
0.34+0.02
2.5 +0.1
1.4 +0.3
380+20
320+20
84+4
17+1
140+10
160+20
74+4
280+70

       measurement  obtainable with standard  solution.
3NAA - Neutron activation  analysis; precision  is  estimated  to  be less  than 20% at this
       stage.

-------
            Table 47.  Elemental Concentration  in High-Temperature Ash as a Function
                       of Oxygen Concentration in Gas Flow.  Ashes prepared by heating
                       for 20 hr in a gas flow of 0.6 scfh.
Element
TFe
Al
Na
% K
1 Mg
Ca
L Ti
r Zn
Mn
Ni
ppm Co
Cu
Cr




340

21
11.9 +0.6
9.5 +0.5
1.00+0.05
1.55+0.08

2.5 +0.1
1.4 +0.3
380+20
360+20
70+4

110+10
150+20


290+70
Heating Temperature (°C)

1150

1200
Oxygen Concentration
21
11.2 +0.6
9.8 +0.5
0.97+0.05
1.53+0.08



440+20
320+20
69+3






45
11.3 +0.6
9.6 +0.5
1.11+0.06
1.52+0.08



440+20
290+15
78+4






68
11.3 +0.6
9.5 +0.5
1.08+0.05
1.55+0.08



420+20
310+20
67+3






21
10.9 +0.5
8.6 +0.4
0.94+0.05
1.40+0.07



400+20
320+30
68+3






(vol %)
45
11.2 +0.6
8.5 +0.4
1.14+0.06
1.44+0.07



410+20
290+15
70+4







68
10.9 +0.5
9.0 +0.4
1.00+0.05
1.43+0.07
0.35+0.02
2.3 +0.1
1.3 +0.3
400+20
280+15

18+1
130+10
250+20

69+4
270+70
1250

21
11.8 +0.6
9.9 +0.5
1.13+0.6
1.54+0.08
0.34+0.02
2.5 +0.1
1.4 +0.3
380+20
320+20
83+4
17+1
140+10
160+20

74+4
280+70
 Calculated on the original 340°C  ash  basis;  obtained  by  an  atomic  absorption method;  each
 precision is based on an estimate of  the precision  of measurement  obtainable with standard
.solution.
 Balance was nitrogen.

-------
Table 48.   Effect of Heating on Weight Loss of 340°C Ash.
           Experimental Conditions:
0.6 scfh air flow,
H20/Air = 0.13 by wt,
20-hr heating time
Heating
Temperature
(°C)
850
950
1100
1200
1250
Wt of
Sample
(g)
1.1138
1.1112
1.1204
1.1095
1.1031
1.1142
1.1029
1.1085
1.0129
1.0985
Wt Loss as a
Result of Heating
(g)
0.1047
0.1055
0.1213
0.1180
0.1269
0.1283
0.1294
0.1299
0.1197
0.1299
Wt Loss
(%)
9.40
9.49
10.83
10.64
11.50
11.51
11.73
11.72
11.82
11.82
Average
Wt Loss
(%)
9.44
10.74
11.50
11.72
11.82

-------
                                    154
            co
            CO
            o
                               O HUMIDIFIED AIR
                                 H20/AIR = OI3BYwt
800
                         900     1000      MOO    1200
                            HEATING TEMPERATURE °C
                                         1300
     Fig. 62.  Effect of Heat-Treatment Conditions on Weight Loss
               of 340°C Ash.
as indicated by good agreement in analytical results for samples obtained
from two sides of each pellet.  This XRF technique was used to analyze
for the elements  Fe, Al, Ca, K, Ti, Zn, Cu, Ni, Mn, and Cr.  Because not
enough standard samples are available to calibrate the XRF instrument,
absolute elemental concentrations of the samples can not be calculated
at this stage.  Instead, the results compared are the relative values
(in counts/sec) of the intensities of the characteristic radiations
produced by the elements in a sample.

     Table 49 shows the results obtained by XRF.  Also included in this
table are AA results for the elements Na, Mg, Li, and V, which can not
be analyzed by XRF.  As can be seen in Table 49, within the experimental
and analytical errors, the XRF intensities and AA results of Series-3 ash
samples were found to be the same as those of the corresponding ashes of
Series-2 experiments.  Since, as indicated by AA results in Table 46.
950 and 1250°C ashes of Series-2 experiments have the same elemental
concentrations as those of 340°C ash, it can be concluded that the
elemental concentrations of 950 and 1250°C ashes of Series-3 experiments
which agree with Series-2 results within the limits of analytical and
experimental errors should also be the same as those of 340°C ash.  In
other words, the fourteen elements being analyzed generally remain in the
ash up to 1250°C in the humidified oxidizing conditions.

-------
             Table 49.   Elemental Concentrations  in High-Temperature  Ash  Calculated  on the
                        Original 340°C-Ash Basis.
                        Ashes  prepared by heating for  20  hr  in  either dry air (Series-2)  or
                        humidified  air  (Series-3)  flow of 0.6 scfh.   In Series-3,
                        H20/Air = 0.13 by weight.
Heating Temperature (°C)



Fe
Al
Na
K
Mg
Ca
Ti
Zn
Mn
Ni
Cu
Cr
Li
V
Series
950
..a XRFb
AA f ^
(cps)
3700+400
240+20
1.04+0.05
900+90
0.36+0.02
2400+200
1300+100





77+4
280+70
2
1250
AA XRF
(cps)
3500+400
150+20
1.13+0.06
700+70
0.34+0.02
1900+200
1000+100
46+5
100+10
69+7
100+10
130+10
74+4
280+70
Series 3
950 1250
AA , , AA , ,
3800+400 3600+400
290+30 150+20
1.01+0.05 1.00+0.05
1000+100 800+80
0.35+0.02 0.35+0.02
2500+200 2100+200
1300+100 1000+100
38+4
93+9
61+6
86+9
120+10
75+4 80+4
330+80 310+80
aAA  - Atomic absorption; each precision is based on an estimate of the precision of measurement
       obtainable with a standard solution.  For Na amd Mg, concentrations are in %; for Li and V,
       in ppm.
 XRF - X-ray fluorescence; the intensity of the characteristic radiation is proportional to the
       amount of the elements; precision is estimated to be less than 10% at this stage.

-------
                                     156


      It is also apparent from Table 49 that the XRF elemental radiation
 intensities of 950°C ash in both Series-2 and Series-3 experiments are
 greater than those of 1250°C ashes.  The 340°C ash has also been analyzed
 using XRF and was found to have much greater intensities for Al and Ca
 than those shown in Table 49.   This decrease in elemental radiation inten-
 sities in high-temperature ashes is probably due to maxtrix-related problems.
 As mentioned previously,33 the 340°C ash is a gray powder,  the 950°C ash
 a brown powder, and the 1250°C ash a fused black,  glassy, hard agglomerate.
 Heat treatment of the ash to a high temperature subjects it to severe
 thermal effects and chemical reactions;  it can reasonably be expected that
 its matrix differs from other  ash matrices due to  changes in composition.
 X-ray fluorescence in known to be sensitively affected by the matrix of
 the element in the sample being investigated.

      From the above experimental results and discussions, the following
 conclusions can be drawn:   under the dry oxidizing conditions present in
 the heat  treatemnt of 340°C ash,  most  of the chlorine  in coal evolves at
 temperatures below 640°C,  although trace amounts of chlorine still remain
 in the  high-temperature ashes.   In constrast to the nometallic chlorine,
 the metallic elements  Na,  K,  Fe,  Al,  Mg,  Ca,  Ti,  Zn,  Mn, Ni,  Co,  Cu,  Cr.
 and Li  are generally retained  in the fused ash up  to 1250°C  in both dry
 and wet oxidizing environments,  and the  retention  of these metallic
 elements  in the fused ash  is not  affected  by the oxygen  concentration in
 the dry flowing gas.   Water vapor  present  in the flowing gas shows no
 effect  on  the evolution of  elements being  studied;  however,  it causes  a
 greater weight loss  from the ash  residue,  indicating that water vapor in
 the flowing gas  affects the evolution  of certain substances  in the 340°C
 ash during heat  treatment.  The  preceding  conclusions  from this study
 appear  to  suggest  that  the  evolution of  sodium,  potassium, and some other
 metallic elements  during the combustion  of  coal  in  other  investigations1
 may probably  be  attributed  to  the  reducing  environments  present around
 the coal or ash  particles in the coal  combustion bed.  A reducing  environ-
 ment  can occur  in  a  region  of  the  bed where  oxygen  is  deficient.

                            EQUIPMENT CHANGES
 J.  Lenc (Principal Investigator)

     As originally installed,  the  6-in.-dia, pressurized, fluidized-bed
 combustor  and  the  3-in.-dia, pressurized, fluidized-bed regenerator
 utilized several components in common.  Due  to the dual  function of these
 components, the  two units could not be operated simultaneously.  The
 equipment  common to both units included the  inlet and outlet surge  tanks,
 the gas preheater, the additive solids feeder, the off-gas system  (cyclones,
 filters, pressure-control valve, etc.), and  the off-gas analysis system.
Modifications of the two systems and installation of additional equipment
were completed to physically separate the combustor from the regenerator.
The modified systems have been described earlier in this report.  An
overall view of  the new  regeneration facility is shown in Fig. 63.

     Several changes were made in the regenerator.   The diameter was increased
from 3 in. to 4 1/4 in., an in-bed product overflow pipe was replaced with
an external pipe, and separate feeding-weighing systems were installed for
coal and for sulfated additive feed.

-------
                       157
Fig. 63.  Overall View of New Regeneration Facility.

-------
                                     158


      Various pieces of equipment are being either fabricated or purchased
 that will permit continuous cycling, by means of pneumatic conveying, of
 sulfated additive from the combustor to the regenerator and of regenerated
 additive from the regenerator back to the combustor.  The equipment includes
 solids product receivers, solids feeders with attached hoppers, system-
 pressure control valves, and inertial separators and filters for removing
 particulate solids from off-gas.  At present, additive must be transferred
 manually between the two units in investigations related to the effect of
 additive recycling on such variables as decrepitation, reactivity for S02
 retention,  and buildup of coal ash in the additive.


                           MISCELLANEOUS STUDIES

 Preparation and Testing of Supported Additives (Dow  Subcontract)

      A research program,  subcontracted  to Dow Chemical company,  is concerned
 with assessing the technical and economic feasibility  of using candidate
 synthetic S02  sorbents in fluidized-bed combustors.  One task consists of
 the  preparation of synthetic SOZ sorbents from fluidizable  "catalyst  type"
 support materials.

      The other  task consists of  thermogravimetric  analysis  of these
 sorbents.   By  obtaining data on  sulfation and regeneration  rates  for  sorbents
 over a broad range  of  simulated  combustion conditions,  the  following  can
 be determined:

      1.  Which  sorbent  is  the most appropriate  for a fluidized-bed
         process  in which  high bed temperatures  are  not  of  paramount
         importance.
      2.  Which  sorbent  is  the most appropriate  for a fluidized-bed
         process  in which  hot gas turbines will  be incorporated into
         the design.
      3.  Which  sorbent  requires  the  least  severe operating  conditions
         in the regeneration mode to achieve  regeneration rates
         comparable to  those obtained with CaO.

     An economic evaluation will be made  on the  basis of data obtained
at Dow and other sites  and will be accompanied by recommendations for
future developmental work.

     Preparation of Synthetic SO? Sorbents.  Three sorbents have been
impregnated on low-surface-area  (<1 nrVg) and intermediate-surface-area
 (1-100 m2/g) fluidizable A1203 carriers supplied by Norton Company.  The
sorbents and their concentrations are CaO  (6.6,  6.9,  10, 15, 20, and 30%),
BaO  (6.9%),  and SrO (6.9%).  The preparation method was:  (1) placing an
aqueous solution of the metal nitrate in a Rinco flask along with support
material, (2) rotating  the contents for 30 min without heat or vacuum,
 (3) rotating the contents for ^1.5 hr with heat lamp or vacuum until dry,
 (4) placing the impregnated supports in a crucible, and  (5) firing the
supported additive in a muffle furnace at 1100°C for 2 hr.

-------
                                    159


     Thermogravimetric Analysis of Candidate Sorbents.   A thermogravimetric
analyzer (Fig.  64) utilizing quartz springs has been constructed.  A change
in mass of a sample is detected by using a cathetometer to observe the
deflection of cross hairs attached to the quartz springs.  The sample pan
is suspended by a quartz fiber into an electric furnace which can operate
at temperatures in excess of 1200°C.  There will be a nitrogen purge through
the insulation case to increase the lifetime of the furnace at high temper-
atures.

     The gases from the gas mixture apparatus enter a quartz coil wrapped
around the heating block to preheat the gases before they enter the quartz
tube and eventually contact the sample.  Laboratory jacks under the steel
cabinet lower the furnace so that there is room to suspend the sample on
the quartz fiber.  The desired temperature is obtained by use of a Variac.
The temperature is monitored by measuring the skin temperature.  The actual
temperature in the tube at the location of the sample is measured at several
power settings to obtain a correlation of power setting, skin temperature,
and sample temperature at different gas flow rates.

     A standard arrangement of the apparatus for preparing simulated stack
gas with reproducible  compositions has been assembled  (Fig. 65).  The use
of separate gas lines allows flexibility in selecting gas compositions.
A typical gas flow composition is 0.3% S02, 5% 02, and  the balance nitrogen.
Whitey microvalves are used to control gas flow.

     This work and the final report are scheduled to be  completed in
October 1976.

Limestone and Dolomite for the Fluidized-Bed Combustion of Coal:
Procurement and Disposal   (Dr. B.  S. Friedman, consultant for Argonne
National Laboratory)

     Supplies and possible markets  for sulfated limestone, sulfated
dolomite, sulfur, and sulfuric acid  (which may be produced in the fluidized-
bed combustion of coal) are discussed  in relation to the costs and avail-
ability of competing materials such as natural limestone; natural dolomite;
wet sludge  (spent limestone from  wet scrubbers); landplaster; wet gypsum
 (a byproduct of wet-process phosphoric acid plants); gypsum  sulfur recovered
from natural gas, petroleum, and  tar sands; and sulfur  produced  by the
Frasch process.   The agricultural use  of sulfated limestone  and  sulfated
dolomite  for soil amendment may be  economic in"localities where  shipping
costs  of  competing materials are  high.  However, the effects of  these
materials on various soil  types and crops  (i.e., peanuts) and  on noncrop
plants, animals,  birds, fish, eto. have not been thoroughly evaluated.

     A projected  rate  for  production of sulfur  from sulfated limestone
and dolomite is  90,000  long tons/year  in 1985.  It  is  expected that  such
a  supply would be absorbed by  the expected  increase in  world demand  for
sulfur, which  should also  absorb  the sulfur produced  from scrubber waste,
coal liquefaction and  gasification, and shale  oil recovery processes.

-------
                        160
                                 TRIPOD  BASE
                             CONTRAROTATING QUARTZ
                                    SPRINGS
                                  CATHETOMETER
                                        in. PLYWOOD
                                        TABLE TOP
 THERMOCOUPLE
FIBER FRAX
INSULATION
   110 V
  POWER
  SOURCE
N2 PURGE
LABORATORY JACKS
REMOVABLE QUARTZ
  TUBE SECTION
      QUARTZ
  HANGDOWN TUBE
  QUARTZ TUBING
    7.5 mm ID

  SAMPLE HOLDER
ELECTRICAL FURNACE
  7.3cm OD x I4.lcm

\
  STEEL CABINET
   37x37 x 46 cm
      Fig. 64.  Schematic of TGA Apparatus.

-------
  GAS TO TGA
0.3 % S02
5 % 02
BALANCE N2
(PLUS
MOISTURE)
/*
/
1


o
-T-'




T i





\ <





ROTAMETER
BYPASS
FOR N2


}
i
o
T








O
I








0
X

C
r
o
o
•
I
X
I
i
1
' 1
*
S02
L
°2
                                          H20
Fig.  65.  Gas-Mixing Apparatus for TGA Sulfation Experiments,

-------
                                     162
 Industrial  processes  for  using  sulfur  in  insulation,  asphalt mixes,  concrete
 mixes,  and  bonding materials  are under investigation.

      Landfill  or  open-pit disposal  of  sulfated  stones  is not expected  to
 cause serious  pollution problems.

      A  complete report has been published.33

 The Properties of a Dolomite  Bed of a  Range of  Particle Sizes and  Shapes
 at Minimum  Fluidization (Dr.  S. Saxena, consultant to  Argonne National
 Laboratory)

      In a previous report,.34  the results  of a series of seven batch
 fluidization experiments  performed without combustion  in the ANL 6-in.-dia,
 fluidized-bed reactor on  partially  sulfated dolomite particles of  a wide
 range of sizes (About 1410-88 pm) and  shapes were described.  Various  com-
 binations of temperature  (70-800°E) and pressure (26-121 psia) were used.
 These results were correlated on the basis of the Ergun relation35 in  which
 the mean shape factor, <|>s, was  computed from the data  for two experiments
 for which the Reynolds number at minimum  fluidization,_Repjinf, was less
 than 20 and the simplified Ergun relation was valid.   s has now been
 computed7 for seven experiments from the  complete Ergun relation, and  these
 suggest a mean value of 0.379.  Use of this value could successfully repro-
 duce all of the experimental values of minimum  fluidization velocities.

     This concept was further checked by performing a  series of eight
 batch experiments with a  fresh dolomite charge  of particles having about
 the same size range, the  results confirmed the  procedure.7  These experi-
ments and related calculations are described in detail in an earlier report
 and provide a reliable basis for the prediction and correlation, over  a
 range of temperatures and pressures of interest in coal combustion, of
minimum fluidization velocity of a bed of nonspherical particles of wide
 size range.   Knowledge of the minimum fluidization velocity facilitates
 explanation of the reaction kinetics of a fluidized-bed reactor since  it
permits the determination of the gas in the bubble phase (an excess over
 that required for fluidization).

     Additional similar experiments were conducted on two fresh unsulfated
dolomites, one with a particle size range of about 2000-88 ym (Series A)
and the other with a range of about 2000-44 pm  (Series B); the results
have been reported in detail.32  These experiments were done to gain a
better understanding of the quality of fluidization and to correlate
minimum fluidization velocity with (1) particle size distribution of the
solids making up the bed,   (2) bed temperature,  and (3) reactor pressure.

     In these experiments, the pressure drop across the bed, AP, increased
as the fluidizing-air velocity was increased until a value of the air
velocity was reached at which the pressure drop remained constant for a
range with increasing fluidizing-air velocity.   At this AP,  the bed was
partially fluidized.   The fluidizing-air velocity at which AP first
becomes constant is referred to as the minimum fluidization velocity for
the partial bed and is denoted by umf.  As the fluidizing-air velocity was

-------
                                    163
further increased, the pressure drop increased further and more and more
of the bed was fluidized.  At some value of the fluidizing-air velocity,
the pressure drop again became constant and did not change with further
increases in the air velocity.  At this stage, the entire bed was fluidized
and the fluidization factor, Q, (defined as the ratio of the pressure
drop across the bed to the weight of the bed per unit area), was unity.
The fluidizing-air velocity at which AP = (W/A) and consequently the
fluidization factor is unity is referred to here as the minimum fluidization
velocity for the total bed and is denoted by umf.  W is the total weight
of the bed, and A is its effective cross-sectional area.  Figure 66 shows
(qualitatively) how pressure drop across the bed, AP, varies with increasing
fluidizing-air velocity, and also changes the minimum fluidization velocities,
    and u".
u
 lmf
u ...
 mf
     The value of AP when the fluidization factor, Q, is unity and the
entire bed is fluidized is denoted by APmfeb.  Similarly, the pressure
drop value corresponding to partial fluidization of the bed is indicated
by APmfpb.  The degree of segregation, S, may be computed from the following
relation:
           S =
                
                    AP
                                                               (1)
                      mfeb
Alternatively, S may be defined in terms of the weight of bed fluidized,
W', such that
             _  (W - W*)
            S "    W
                                                                (2)

      Fig 66.   Qualitative Dependence of the Pressure Drop Across the
               Bed, AP,  on the Fluidizing-Gas Velocity.

-------
                                     164
and

          Q = AP/(W/A)                                                  (3)
Q and  S are related,  such that

          Q = 1 - S

For a  completely fluidized bed, the fluidization factor is unity and the
degree of segregation is zero.

     The experiments  mentioned above substantiated the procedure depicted
in Fig. 66 for determining umf.  It was found that the linear plot is a
good guide for determining umf at which AP = W/A.  The pressure drop
remains constant for  a range of u values beyond umf, but after a certain
value  is reached, more and more of the bed is fluidized and AP increases
with u at a higher rate than its initial rise in the range u < umf.  The
experiments suggest that the occurrence of partial fluidization of the
bed at u = umf only influences the pattern of approach to the fluidization
of the entire bed in  the region of umf < u < umf.  The final state, when
the entire bed is fluidized, is that which would be obtained by extrapolating
the rate of rise of AP versus u in the range u < umf.   We suggest
this as the basis of  a procedure for predicting minimum fluidization
velocities for beds for which the degree of segregation is not high, that
is to  say, up to about 0.3 (i.e., with a Q about 0.7 or larger).  Experi-
ments  further revealed that partial segregation of the bed occurred if
the ratio of the diameter of the largest particle to the diameter of the
smallest particle exceeded at least 16.  As a result, the prediction of
minimum fluidization  velocity for the entire bed becomes complicated,
making precise prediction uncertain.

     As discussed earler,^ our findings on quality of fluidization and
segregation are in conformity with the criterion of Geldarts,36 experiments
of Rowe and coworkers,37"39 and Knowlton,1*0 and correlations of Jolly and
Doig,41 and Wen and Yu.42

Mathematical Modeling:  Noncatalytic Gas-Solid Reaction with Changing
Particle Size:  Unsteady State Heat Transfer  (Dr.  S. C. Saxena, consultant
to Argonne National Laboratory)

     Knowledge of the interior (core)  temperature of a solid reacting
particle is important for catalytic as well as noncatalytic exothermic  reac-
tions inasmuch as this temperature may differ considerably from the  ambient
temperature.   In catalytic reactions,  such temperature differences may
lead to severe decreases in catalyst reactivity.  In noncatalytic reactions
such as the combustion of coal to produce low-Btu gas, the composition
of the product gas may differ considerably, depending on the temperature
of the reaction front within the particle and high  temperatures may cause
slagging of the solid reactant.

     For many gas-solid reactions, it is  reasonable to use the quasi-steady
state  approximation to obtain a first approximate description of the
reacting system.   However, solution of the heat balance equation under
quasi-steady state approximation can lead to errors in the estimates of

-------
                                     165


(1) the reacting particle temperatures and (2) the instantaneous transition
of the rate-controlling regime from kinetic to diffusion or vice versa.
The particle temperature has a significant influence on the effect dif-
fusivity and the reaction rate, since both of these are strongly dependent
on temperature.  For gas-solid reactions occurring in fluidized beds, the
assumption that these are governed by the constant bed temperature is also
not valid.  Accurate data on internal particle temperature and changes
in particle size must be considered to obtain reasonable predictions either
for design purposes or for comparison with the observed f luidized-bed
operations.

     For diffusion-controlled reactions, it has been shown43 that a change
in particle size as the' reaction progress has considerable influence on
the conversion- time relationship.  Generally, high temperatures are generated
in the diffusion-controlled regim of the gas-solid reaction, and it is
imperative that the changing size of the particle be included in the analysis
to obtain more realistic estimates of the internal particle temperatures.
The effects of changing particle size on unsteady state heat transfer has
not yet been analyzed in the literature.  The inclusion of this phenomenon
presents an unique problem of two moving boundaries — (1) the reaction front
and (2) the external particle diameter of the particle  (due to growth or
shrinkgage of  the particle with reaction) .  The temperature rise during
reaction, especially within a porous catalyst, has been investigated by
many authors.44-51  In these investigations, which have been reveiwed,   it
has been assumed that the particle size remains invariant throughout the
reaction.

     In formulating a mathematical expression for the unsteady state heat
transfer that  addresses the problem  of internal particle temperature, we
have introduced a parameter Z which  characterizes the change in particle
size as reaction proceeds.  The model governing such a  system is developed
below.

Model  of the System

     The analysis is based on a shrinking-core model for a spherical pellet
 (Fig.  67).  The derivation essentially involves the same steps as are
described by Rehmat and Saxena.43  The following  single reaction is  considered
in the present analysis:*

          aA(g) + S(s) ieE(g) + jSi(s)                               W

The rate of reaction is assumed  to be first  order with  respect  to gaseous
and solid reactants and is given by:

          -r   = kl(T ) CC                                             (5)

 *Symbols  are defined  in "Notation"  of  this  report  section.

-------
                                     166
                                         ASH  LAYER
               R0   r
r     R0   R
     Fig. 67.  Gas-Solid Reaction of a Growing Particle:
               Concentration and Temperature Profiles.
    The
where the temperature dependence of the rate constant  is  assumed  to be
of the Arrhenius type.  The reaction takes place at an interface  situated
at a distance rc from the center.   The solid reactant  is  assumed  to lie
within the interface (0 < r < rc),  and the solid product  forms  the
surrounding shell (rc < r < R).   The particle is assumed  to  retain its
spherical shape as it reacts.  As  the reaction proceeds,  the overall  size
of the particle may change from the initial radius  Ro.  The  solid particle
is immersed in a flowing gas mixture containing the gaseous  reactant  A

-------
                                     167
whose concentration in the bulk phase is represented by CAO-  Our aim is
to introduce the transient term in the energy balance equation and to
examine its effects on both the conversion-time relationship and the interior
particle temperature.  Analysis of the system requires simultaneous solution
of mass and heat balance equations in the particle.  In the following
analysis, we will employ the quasi-steady state approximation for the mass
balance equation since such an assumption is justified by Bischoff.

     Based on a shrinking-core model for a nonisothermal system under the
quasi-steady state assumption, a material balance for a gaseous reactant A
leads to the following relation:
                doo
                    = 0
                                                                        (6)
The boundary conditions are:

          do),
and
JA.N
 d£    Sho
          du.
                     U
                         exp
Here
                     Ej

                    RT
l_
U
                                              at
          NSho ' NSh  (Ro/2R)  < W
 The  solution  of Eq. 6  in  conjunction with the boundary conditions of
 Eq.  7 and 8 lead to the following relation for the concentration of A
 at the reaction surface:
  Ar-
  AC
        +
2
<|> 1 5 exp
c

"EI / i \
— — 1 "^ ~ u~ )
RT \ c/
o
•a., ,.,,., —*
                                         U
                                     •J^r—  exp
                                                               1 -
                                                                   U
                                                                        (7)
                                                             (8)
                                                             (9)
                                                             (10)
 The  simplifying  assumptions  implicit  in the above derivation are discussed
 in detail  by  Rehmat  and  Saxena.43
      Similarly,  the heat balance equation with transient term included
 in the  analysis  is  given by:
         82U .  2
                                                                        (11)
               80
 with the  following boundary conditions:
_ M = N   m  -
  8?    Nuok s

                        uR   s
                                      (Us -
                                                              (12)

-------
                                      168
 and
       -(—\
        \9g /
         RT _
                  u
                      exp
                           RT
                                  GE,  9U
                                  	c^	

                                   3  86
                                                    at £ = £
                                                         (13)
 The  initial  condition is  given by:
 U=U  =1 at
      c
                           =0
                                                                        (14)
In the above, we have  assumed  that  the  particle  core  maintains  a uniform

temperature and that the walls of the reactor  are maintained  at the ambient

temperature, T .




     In the case of the quasi-steady state  asumption,  the  above equations

will not contain the accumulation term,  and consequently  (9U/98)  and

(9Uc/96) will both be  zero.  Thus,  the  equations  will  simplify  to the

following:
                2._

                5 d?   U
with the boundary conditions given by
            dU
= NM   (U  -
   Nuo   s
                                 NNuR(Us -
and
                                                              (15)
                                                              (16)
""AC

iU  ,
                                 exp
                                      RT
                                              - "
                                           at  £ =  £,
                                                        (17)
The following symbols are used in the above equations:



          N,
           Nuo
               =h(2R)/k
          N
           NuR = aR T3/k
                   o o
     = CAODeA(To)(-AHl)R/keEl




     - CAODeA(To)CP(Sl)/aCSOke


G    = p(S)C (S)T/aC  (-AHi)
          V
                     "p	o'" SO
                                                        (18)



                                                        (19)




                                                        (20)



                                                        (21)



                                                        (22)



                                                        (23)

-------
                169


                                                   (24)

01oAOso
           T = P(S)R/k1(T)CCM(S)                                 (25)
and
           6 = t/T                                                     <26)

Further, the following relations are also valid and apply to the system
under analysis:

          Z = jM(S1)p(S)/M(S)p(S1)
                     ,                                                   (29)
           Sho ' 5
                  s
and                          -i/q      1/2
                             1/3
               _                                                        (30)
          NNuo " 5g          ?l/2

The effect of the varying size  of  the particle  is  contained  in  Z.

      Since NSc, NReo, and Npr refer  to  properties  at  the  constant  ambient
conditions and the initial  size of the  particle, Eq.  29 and  30  can be
simplified further so that

          N    = IL   K3                                                (3D
          NSho   5g     1/2
                     +
                                                                        (32)
 where

           K.Q = K.O  (N  /    \"r»
            0    ^    Sc      Keo

 and



 It can be seen from Eq.  31 and 32 that variations in K3 and K^ would
 result in the variation of NSho and NNuo.  Thus, in our subsequent analysis
 instead of varying Nsho and %UO, we will vary K3 and Kit to _investigate
 the effects of mass and heat transfer coefficients, respectively.

-------
                                      170
      Finally, we need an expression to relate conversion to time.  Such an
 expression has already been derived43 and has the following form:
                   "AC
de     u
        c
                       exp
                             RT        c
                               o
 with the boundary conditions given by
                                                                        (35)
            c
              = 1 at 8 = 0                                              (36)
 The conversion of solid reactant X is related to the system paramter
 for a spherical pellet, it is given by:
                                                                        (37)
      Equations 11,  12,  13,  and 14 in conjunction with Eq.  10,  35, and 36
 are to be solved simultaneously to obtain con ver si on- time  relationships.


 Notation

 a       Stoichiometric  coefficient of gaseous component, A
 A       Gas  reactant
 CA      Concentration of  A  at  the radial  distance r,  mol/ft3
 CAO     Concentration of  A  in  bulk phase,  mol/ft3
 CAC     Concentration of  A  at  the core of  the particle, mol/ft3
 CAS     Concentration of  A  at  the surface  of  the particle, mol/ft3
 CSO     Initial  concentration  of  the solid reactant,  mol/ft3
 Cp      Specific heat of  the bulk gas at  constant pressure, Btu/(lb)(°R)
 Cp(S)    Specific heat of  solid reactants,  Btu/(lb)(°R)
 CpCSjJ   Specific heat of  solid product, Sls Btu/(lb)(°R)
 DA      Molecular diffusivity  of  component A  in  the bulk gas phase,  ft2/hr
 DeA     Effective diffusivity  of  the component A in the ash layer, ft2/hr
 e        Stoichiometric  coefficient  for  the gaseous component, E
 E        Product  gas
 EI       Activation energy,  Btu/mol
 g        Signifies the gaseous  state
 G        Dimensionless quantity defined by  Eq. 23
 h        Convective heat transfer  coefficient, Btu/(hr) (f t2) (°R)
 AHj      Heat of  reaction  per mole of  reactant, Btu/mol
 j        Stoichiometric  coefficient of  the  solid  component, S^
 k        Thermal  conductivity of the bulk gas, Btu/(hr) (ft) (°R)
 ki       Reaction rate constant, f t4/ (mol) (hr)
 ke       Effective thermal conductivity of  the ash layer, Btu/(hr) (f t) (°R)
 K!      A numerical constant which occurs  in the correlation of Sherwood
         and. Nusselt numbers
Kf      K^/2, dimensionless
K2      A numerical constant which occurs  in the correlation of Sherwood
        and Nusselt numbers
K£      K2/2, dimensionless
K3      Defined by Eq.  33
K^      Defined by Eq.  34

-------
                                     171
M(S)    Molecular weight of solid, S
M(Si)   Molecular weight of solid, Si
NNU     Nusselt number, 2Rh/k, dimensionless
NNuo    Defined by Eq. 18, dimensionless
%UR    Defined by Eq. 20, dimensionless
Npr     Prandtl number Cpy/k, dimensionless
NRS     Reynolds number, 2urp/y
%eo    NRe(R0/R), dimensionless
No      Schmidt number, y/pDA
Nsh     Sherwood number, 2RkmA/DA, dimensionless
No,-     defined by Eq. 9, dimensionless
r       Radial distance from the center of the spherical particle, £t
rA      Rate of reaction of A, mol/(hr)(ft2)
rc      Radius of the unreacted core, ft
R       Particle radius, ft
R0      Initial particle'radius, ft

R       Gas constant, Btu/(mol)(°R)
s       Signifies solid state
S       Solid reactant
Si      Solid product
t       Time, hr
T       Temperature,  °R
T0      Ambient temperature,  °R
T       Temperature  of the unreacted core,  °R
Tg      Temperature  of the outer  surface of  the  particle,  °R
U       Reduced temperature  T/TO,  dimensionless
Uc      Reduced core temperature  TC/TQ, dimensionless
Us      Reduced particle  surface  temperature TS/TQ,  dimensionless
V       Defined by Eq. 22, dimensionless
XA      Mole  fraction of  component A, dimensionless
XAO    Value of  XA  in the bulk gas, dimensionless
XAr    Value of  XA  at the unreacted core  surface, dimensionless
X7c    Value of  XA at the center surface  of the particle, dimension!
X       Conversion of solid  reactant S defined by Eq.  37,  dimensionless
 Z       A parameter to characterize growth or shrinkage of the particle,
         defined by Eq. 27,  dimensionless

 Greek Letters

 B!      Defined by Eq. 21,  dimensionless
 0       Reduced time defined by Eq. 26, dimensionless
 y       Viscosity of bulk gas, lb/(ft)(hr)
 £       Reduced distance r/RQ, dimensionless
 5       Reduced core radius of the particle rc/R0, dimensionless
 5g      Reduced size of the particle R/RO, dimensionless
 pS      Density of the bulk gas, lb/ft3
 p(S)    Density of the solid S, lb/ft3
 p(Si)   Density of the solid, Si, lb/ft
 T       Characteristic time defined by Eq. 25, hr
 
-------
                                      172
 wA     Reduced value of XA,  XA/XAc,  dimensionless
 WAC    Reduced value of XAC, XAC/XAO,  dimensionless
 OJAS    Reduced value of XAS, X  /X    dimensionless
 a      Radiative heat transfer coefficient,  Btu/(hr)(ft2)
                              ACKNOWLEDGMENTS


     Many  people  have  contributed  to  the  progress made  in  the  studies
reported here.  We  gratefully acknowledge the help  given by Mr. L.  Burris

   'D'
  f v   P                F' CafaSS°  in directinS  the program,  the  assistance
of Mr. C. Schoffstoll, Mr. J. Stockbar, Mr. S. Smith, Mr. R.  Mowry, and
Mr. H. Lautermilch in operating and maintaining  the PDUs and  Mr.  E. Nielsen
for operating the TGA.

     We would also like to express our appreciation for the analytical services
recexved  from Ms. C. Blogg, Mr. M.  Homa, Mr. R.  Bane, Mr. R.  Telford, Mr. B.
Tani, Mr. K. Jensen, Mr. N. Johnson, and Ms. A.  Engelkemeir .

     The microscopy study and interpretation of the data were done by Dr  L
Fuchs.

     Design and drafting services were provided by Mr. R.  Stimac and Mr.  R.
Frank.  Finally, we acknowledge the typing and assembly of this report bv
Ms. M. Sobczak.                                                          y

                              REFERENCES


 1.  G.  J. Vogel et at.,  "A Development  Program on Pressurized Fluidized-
     Bed Combustion," Annual Report, July 1,  1974 - June 30  1975
     ANL/ES-CEN-1011  (July  1975).

 2.  D.  A. Martin,  F. E.  Brantley,  and D.  M.  Yergensen,  "Decomposition of
     Gypsum in  a Fluidized-Bed Reactor,"  Bureau of Mines Report of Investi-
     gation 6286 (1963).

 3.  M.  Hartman  and R. W. Coughlin,  "Reactions  of Sulfur Dioxide with
    Limestone and  the Influence  of  Pore  Structure," Ind.  Ens. Chem 13
     (3),  248  (1974).                                               ' —

4.  A.  Skopp, J. T.  Sears,  and R. R. Bertrant, "Fluid Bed  Studies of  the
    Limestone Based  Flue Gas  Desulf urization Process," Report No.  GR-9-FGS-69,
    Esso  Research and Engineering Company, Government Research Div. ,
    Linden, N.  J., August  1969.

5.  G. J. Vogel et al, , "Reduction  of Atmospheric  Pollution by the
    Application of Fluidized-Bed Combustion and  Regeneration  of Sulfur-
    Containing  Additives," Annual Report, July 1971-June 1972," ANL/ES-CEN-1005
    and EPA-R2-72-253.

-------
                                    173


 6   T   D   Wheelock and  D.  R.  Boylan,  "Reductive Decomposition  of  Gypsum
     by Carbon Monoxide,"  Ind.  Eng.  Chem.  .52  215  (March  1960).

 7.   G.  J.  Vogel et al.3 "A Development Program on Pressurized  Fluidized-
     Bed Combustion,"  Quarterly Report, July-Sept.  1975,  ANL/ES-CEN-1013.

 8.   J.  S.  Gordon et al. ,  "Study of  the Characterization and Control of
     Air Pollutants from a Fluidized-Bed Boiler -  The S02-Acceptor Process,"
     EPA-R2-72-021 (July 1972).

 9.   J. Yerushalmi et  al.,  "Agglomeration of  Ash  in Fluidized Beds
     Gasifying Coal:  The  Godel Phenomenon,"  Science 187, 646 (1975).

10.   P. Nicholls and W.  T.  Reid, Trans. Am.  Soc. Mech. Eng.  62_, 141 (1940).

11.   S. Ehrlich and W. A.  McCurdy,  "Developing a  Fluidized-Bed Boiler,"
     9th Intersociety  Energy Conversion Engineering Conference Proceedings,
     p. 1035 (1974).

12.   T. L.  Webb and H. Heystek, "The Differential Thermal Investigation
     of Clays" (R. C.  MacKenzie, Ed.), Mineralogical Society, London,
     pp. 329-63 (1957).

13.   C. W.  Beck, Thesis, Harvard University,  Cambridge Massachusetts
     (1946).

14.   R. C.  MacKenzie  (Ed.), Differential Thermal  Analysis, Vol. 1,
     Academic Press, London, pp. 254-56 (1970).

15.   ibid., p. 318.

16.   R. R.  West and W. J.  Sutton, "Thermography of Gypsum," J. Am. Ceramic
     Soc. 37_, 221-24  (May 1954).

17.   G. J.  Vogel et al.3 "Reduction of Atmospheric Pollution by the
     Application of Fluidized-Bed Combustion and  Regeneration  of Sulfur-
     Containing Additives, Annual Report, July 1973 - June 1974, Argonne
     National Laboratory,  ANL/ES-CEN-1007 (1974).

18.   R. C.  Hoke, L. A. Ruth, and H. Shaw, "Combustion and Desulfurization
     of Coal in a Fluidized Bed of Limestone," Combustion 4^(7), 6-12
     (1975).

19.   C. W.  Zielke et  dl.}  "Sulfur Removal During Combustion of Solid   Fuels
     in a Fluidized Bed of Dolomite,"  J. Air Pollution Contr.  Assoc.,  164
     (March 1970).

20.   A. A. Jonke et al., "Reduction of Atmospheric Pollution by the
     Application of Fluidized-Bed Combustion," Annual Report,  July 1969 -
     June  1970, Argonne National Laboratory, ANL/ES-CEN-1002  (1970).

-------
                                      174


 21.  H. D.  Levene and J.  W.  Hand,  "Sulfur Stays in the Ash When Lignite
      Burns," Coal Min.  Process.  12_(2),  46-48 (February 1975).

 22.  B. R.  Hubble, S. Siegel,  L. H.  Fuchs,  and P.  T.  Cunningham,  "Chemical,
      Structural,  and Morphological Studies  of Dolomite in Sulfation and
      Regeneration Reactions,"  Proceedings of the Fourth International
      Conference on Fluidized-Bed Combustion, Mitre Corporation, McLean
      Virginia, May,  1976, p. 367.

 23.   R.  T.  Yang,  P.  T.  Cunningham, W. I.  Wilson,  and  S.  A.  Johnson,  "Sulfur
      Removal and  Recovery from Industrial Processes," Advan. Chem.   Series
      137, pp.  149-157 (1974).

 24.   R.  B.  Bird,  W.  E.  Stewart, and  E.  N. Lightfoot,  Transport Phenomena,
      New York,  1962,  pp.  522-537.

 25.   N.  P.  Phillips,  Radian  Corporation Technical  Note 200-045-10-01
      (September 1974) .

 26.   N.  P.  Phillips,  Radian  Corporation Technical  Note.

 27.   J.  A.  Cusumano  and R. B.  Levy,  Catalytica Associates,  Inc., EPRI  TPS
      75-603 (October  1975).

 28.   K.  H.  Stern  and  E. L. Weise, "High Temperature Properties and
      Decomposition of Inorganic Salts,  Part  I.  Sulfates," NSRDS-NBS-7
      (October 1966).

 29.   P.  J.  Fiscalora, 0. M. Uy, D.  M. Meunow,  and  J. L. Margrave,
      J.  Am. Ceramic  Soc., 51,  574  (1968).

 30.   T.  Pierce, Dow Chemical Co., private communication.

 31.   P.  T.  Cunningham et al.,  "Chemical Engineering Division Environmental
      Chemistry Annual Report, July 1974 - June 1975, USERDA Report ANL-75-51,

 32.   G.  J. Vogel et at., "A Development Program on Pressurized Fluidized-
      Bed Combustion," Quarterly Report,  October 1, 1975 - December 31, 1975,
     Argonne National Laboratory,  ANL/ES-CEN-1014.

 33.  G. J. Vogel et al., "A Development Program on Pressurized Fluidized-Bed
     Combustion," Quarterly Report, January 1, 1976 - March 31,  1976,
     Argonne National Laboratory, ANL/ES-CEN-1015.

34.  S. C.  Saxena and G. J. Vogel,  "The Properties of a Dolomite Bed of a
     Range of Particle Sizes and  Shapes  at Minimum Fluidization," Argonne
     National Laboratory,  ANL/ES-CEN-1012 (1975).

35.  D. Kunni and D.  Levenspiel,  Fluidization Engineering, Ch.  3,  John Wiley
     & Sons, Inc., New York (1969).

-------
                                     175
36.  D. Geldart, Powder Tech. 1_, 285 (1973).

37.  P. N. Rowe, A. W. Nienow, and A. J. Agbin, Trans. Inst. Chem. Eng.
     _50, 310 (1972).

38.  P. N. Rowe, a. W. Nienow, and A. J. Agbin, Trans. Inst. Chem. Eng.
     _50, 324 (1972).

39.  A. W. Nienow, P. N. Rowe, and A. J. Agbin, Trans. Inst. Chem. Eng.
     J51, 260 (1973).

40.  T. M. Knowlton, High Pressure Fluidization Characteristics  of Several
     Particulate Solids:  Primarily Coal and Coal-Derived Materials,  Paper
     No. 9b, 67th Annual Meeting of the American  Institute  of  Chemical
     Engineers held in Washington, D.C. during December  1-5, 1975.

41.  R. D. Jolly and I. D. Doig, Chem. Eng. Sci.  28_,  971-973  (1973).

42.  C. Y. Wen and Y. H. Yu,  Chem. Eng. Prog. Sym.  Series  62_,  No.  62, 100
     (1966).

43.  A. Rehmat and  S. C. Saxena, Ind. Eng.  Chem.  Process Design  Develop.
     15(2),'343  (1976).

44.  C. D. Prater, Chem. Eng. Sci. 8_, 284  (1958).

45.  A. Bondi, R.  S. Miller,  and W. G.  Schlaffer, Ind. Eng. Chem.  Process
     Design Develop. l_, 196  (1962).

46.  J. Wei, Chem.  Eng. Sci., 21_,  1171  (1966).

47.  D. Luss and N. R. Amundson, A.I.Ch.E.  J.  15(2),  194 (1969).

48.  G. S. G. Beveridge and  P.  J.  Goldie,  Chem.  Eng.  Sci.  23.,  913 (1968).

49.  J. Shen and J. M.  Smith, Ind. Eng.  Chem.  Fundament. 4-, 293  (1965).

50.  C. Y. Wen  and S. G. Wang,  Ind.  Eng.  Chem.  62_(8), 30 (1970).

51.  R. H. Knapp and R. Aris, Arch.  Ration. Mech. Anal.  44_(37),  165  (1972)

52.  K. B. Bischoff, Chem. Eng.  Sci.  18.,  711  (1963).

-------

-------
                    177
              APPENDIX A.

 CHARACTERISTICS OF RAW MATERIALS USED
IN FLUIDIZED-BED COMBUSTION EXPERIMENTS

-------
                                178
Table A-l.  Particle-Size Distribution and Chemical and Physical
            Characteristics of Arkwright Coal.

                	Sieve Analysis
                U.S. Sieve No.      % on Sieve

                        +14             0.0
                    -14 +25             8.0
                    -25 +35            14.2
                    -35 +45            12.3
                    -45 +80            24.7
                    -80 +170           17.9
                    -170               23.0

                Mean Particle Dia:   323 ym
                              Proximate Analysis (wt %)
                              As Received      Dry Basis
        Moisture                   2.89            	
Volatile Matter
Fixed Carbon
Ash

Sulfur, wt %
Heating value,
Btu/lb
38.51
50.92
7.68
100.00
2.82

13,706
39.66
52.43
7.91
100.00
2.90

14,114
                                Ultimate Analysis  (wt  %)
       Carbon                               77.14
       Hydrogen                              5.23
       Sulfur                                2.90
       Nitrogen                              1.66
       Chlorine                              0.19
       Ash                                   7.91
       Oxygen  (by difference)                4.97
                                Fusion Temperature of Ash
                                Reducing          Oxidizing
                                Atm  (°C)           Atm  (*C)
       Initial deformation       1104               1160
       Softening (H = W)         1177               1216
       Softening (H = 1/2W)      1193               1243
       Fluid                     1232               1271

-------
                               179
Table A-2.  Particle-Size Distribution and Chemical and Physical
            Characteristics of Triangle Coal.

               	Sieve Analysis	
               U.S. Sieve No.        % on Sieve
+14
-14 +25
-25 +35
-35 +45
-45 +80
-80 +170
-170
Mean Particle Dia:
0.0
25.0
22.6
26.0
24.0
2.0
0.3
576 ym
                             Proximate Analysis  (wt  %)

                             As  Received      Dry  Basis
Moisture
Volatile Matter
Fixed Carbon
Ash

Sulfur, wt %
Heating value,
Btu/lb
3.46
31.47
55.69
9.38
100.00
0.98

13,053
—
32.60
57.68
9.72
100.00
1.02

13,521
                                   Ultimate Analysis (wt %)

           Carbon                          76.11
           Hydrogen                         4.99
           Sulfur                           I-02
           Nitrogen                         1.30
           Chlorine                         0.22
           Ash                              9.72
           Oxygen (by difference)           6.64

                                    Fusion Temperature of Ash
                                     Reducing        Oxidizing
                                     Atm (°C)         Atm (*C)
           Initial Deformation        1383            1430
           Softening (H = W)          1444            1480
           Softening (H = 1/2 W)      1485            1510+
           Fluid                      1510+           1510+

-------
                                180
  Table A-3.   Particle-Size Distribution and Chemical
              Characteristics of Glenharold Lignite.

                         Sieve Analysis
              U.S.  Sieve No.
  %  on  Sieve
+14
-14 +25
-25 +35
-35 +45
-45 +80
-80 +170
-170
Mean Particle Dia:
0.00
9.88
11.19
19.46
28.90
18.15
12.42
353 ym
                    Proximate Analysis, wt  %
                 As Received         Dry Basis
Moisture
Ash
Volatile Matter
Fixed Carbon

Sulfur, wt %
Heating value,
Btu/lb
30.90
6.11
30.00
32.99
100.00
0.53
7,625


8.84
43.42
47.74
100.00
0.77
11,035

                   Ultimate Analysis, wt
                As Received
Dry Basis
Moisture
Carbon
Hydrogen
Nitrogen
Chlorine
Sulfur
Ash
Oxygen
(by difference)

30.90
46.04
3.03
0.72
0.04
0.53
6.11
12.63


100.00
__
66.63
4.38
1.04
0.06
0.77
8.84
18.28


100.00
Average of two samples, as received.

-------
                               181
Table A-4.  Particle-Size Distribution and Chemical Characteristics
            of Tymochtee Dolomite.


                 	Sieve Analysis	
                U.S. Sieve No.      % on Sieve

                      +14               0.4
                  -14 +25              48.6
                  -25 +35              19.9 .
                  -35 +45              18.8
                  -45 +80              11.7
                  -80 +170              0.4
                  -170                  0.4

               Average Particle Dia:  750 ym
Component
Ca
Mg
C02
Si
Al
Fe
H20
Chemical Analysis (,wt /0,
20.0
11.3
38.5
2.3
0.87
0.29
0.2
            Derived Composition

               CaC03                          50.0
               MgCOs                          39.1

-------
                              182
Table A-5.  Particle-Size Distribution and Chemical Characteristcs
            of Type-38 Alundum Grain Obtained from the Norton
            Company.
        Sieve Analysis
U.S. Sieve No.   Wt % on Sieve
      +14
  -14 +25
  -25 +35
  -35 +45
  -40 +80
                   Typical Chemical Analysis
                   Component          Wt %
                                        A1203
                                        Si02
                                        Ti02
                                        Na20
                                      99.49
                                       0.05
                                       0.10
                                       0.01
                                       0.35
          Total
100.0

-------
                     183
               APPENDIX B.

PLOTS OF OPERATING DATA AND EXPERIMENTAL RESULTS
        OF COMBUSTION EXPERIMENTS

-------
                                      184
    Q.
                               XXX
   QQ
     8
    e "
    Q.
    0.
A \
-4-^-W
V-^4-
T^^H-^-^—


--*— * 1 1 1



	 1 	
,^| 	 L
\

     in
 r

 S
 o
 c.
1->
 c
 CD
 O
 C
 o
o
   °

Analyzer
Inoperat ive


1

CM
O
?  !^Vs-^rA
   Q.
   CL
                                     Time,  mln
E —
a. -
O.

-------
                                   185
                          Analyzer Inaperative
CL
CL
g 	

Analyzer In
jperative



                                                Additive Fe
                                                Increased t
                                                (Ca/S = 20)
              120
240
360        480
 Time,  mln
                                                        600
                                          720
                                                                             840
     Fig. B-2.  Bed Temperature and  Flue-Gas Composition,
                Experiment PSI-2.

-------
                                      186
  o
O

Analyzer I
loperative




                      ^-"Overnight shutdown
                120
240
360       480
 Time, mln
600
                                                                    720
                                                    840
    Fig. B-3.  Bed Temperature and Flue-Gas Composition,
               Experiment PSI-3.

-------
                                     187
   £
   CL
   CL


    «.
   O
     o


     in
 §s

"•S°o
    e"
    Q.
    O.
^H-H^S»44j*
l-VH^VH*4*^

Analyzer I
loperative





Analyzer
Inoperative




                 120
       2-10
         360        480

          Time,  mln
                                                        600
                                               720
                                                                            840
       Fig.  B-4.  Bed Temperature and Flue-Gas Composition,

                 Experiment PSI-4.

-------
                                 188
          60
120
130        2iO
 Time,  mln
                                                 300
Fig. B-5.  Bed Temperature and Flue-Gas Composition,
           Experiment LIG-2D.

-------
                                                Gas Concentrations-
so.
  H-
  00
fd W
x ro
T3 CL
ro
i-! H

II
3 (D
rt ^
 i m
NJ
 1 (U
  n>

  o
  P
  CO

  O
  o

  I
  o
  en
  H-
  rt
  H-

  §
                   ppm CH
                      300 0
                4., ppm  CO, ppm
                       50  0        39000
"7
So
o-
   20  0
C02, %
   NO, ppm  Bed Temp,  °C
5  0        200  750      1000
                                                     !» ro
                                                      I
                                                     I-*
                                                     O 3
                                                     
                                                     >-i
                                                      n
                                                      o
                                                      d
                                                      3
                                                      m
                                                      o
                                                      rr
                                                      (U
                                                    	0.
                                                                                                         oo
                                                                                                         VO

-------

-------
                                TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
 I. REPORT NO.
 EPA-500/7-76-019
                           2.
                                                      3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE A DEVELOPMENT PROGRAM ON
 PRESSURIZED FLUIDIZED-BED COMBUSTION
            5. REPORT DATE
             October 1976
                                                      6. PERFORMING ORGANIZATION CODE
7.AUTHORIS) GtVogel, I.Johnson, P.Cunningham, B.Hubble
S.Lee, J.Lenc, J.Montagna, F.Nunes, S.Siegel,
G.Smith. R.Snvder. S.Saxena,  W.Swift, et al.	
            8. PERFORMING ORGANIZATION REPORT NO,

            ANL/ES-CEN-1016
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Argonne National Laboratory
9700 South  Cass Avenue
Argonne, Illinois 60439
            10. PROGRAM ELEMENT NO.
            EHE623A
            11. CONTRACT/GRANT NO.
             (ERDA) 14-32-0001-1780
             (EPA) IAG-D5-E681
 12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development*
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC 27711
            13. TYPE OF REPORT AND PERIOD COVERED
            Annual: 7/75-6/76     	
            14. SPONSORING AGENCY CODE
             EPA-ORD
15. SUPPLEMENTARY NOTES (*) ERDA (Office of Fossil Energy) is cosponsor of this report.
Project officers are J. F. Geffken (ERDA) and W. B.Steen  (EPA).
16. ABSTRACT
          The document reports progress of an evaluation of the feasibility of using
fluidized-bed (FB) combustors in power and steam plants.  The concept involves bur-
ning fuels such as coal in a FB of either a limestone or a synthetically prepared Ca-
containing stone.  The Ca reacts with the S to form CaSO4, which remains in the bed,
thus decreasing the SO2  level in the flue gas.  NOx levels in the flue gas are low.  In
a separate step,  the CaSO4 is regenerated to CaOby reductive decomposition at about
1100 C for reuse in the combustor.   Progress is reported on: the effect of regener-
ation operating variables on extent of regeneration and SO2 concentration in the off-gas
using coal as the source of reducing agent and of heat; the alternate combustion and
regeneration behavior of stone; the rate and extent of sulfation of agents impregnated
on A12O3; the effect of variables  on sorption and release of S for CaO-impregnated
stone; attrition resistance of stone;  the kinetic and structural changes occurring
during half-calcination of dolomite;  the CaS-CaSO4 regeneration reaction; and the vola-
tility of trace elements when heating coal ash.   Procurement and disposal of regene-
rated stone, minimum fluidization studies, modeling of a gas-solid combustion reac-
tion and of the regeneration process, and combustion studies using different sizes of
coal and additive and  also using lignite  are  reported.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                          b.lDENTIFIERS/OPEN ENDED TERMS
                        c.  COS AT I Field/Group
Air Pollution, Fluidized Bed Processing
Sulfur Oxides, Electric Power Plants
Fossil Fuals, Combustion, Coal
Flue Gases,  Limestone, Nitrogen Oxides
Regeneration (Engineering), Sulfation
Dolomite (Rock), Trace Elements, Ashes
Air Pollution Control
Stationary Sources
Fluid Bed Combustion
Combined-Cycle Power
  Generation
13B
07B
2 ID
13H,07A
10A
21B
08G
07C
IB/DISTRIBUTION STATEMENT

 Unlimited
19. SECURITY CLASS (This Report)
Unclassified
                         21. NO. OF PAGES
                                          20. SECURITY CLASS (Thispage)
                                          Unclassified
                                                                  22. PRICE
EPA Form 2220-1 (9-73)

-------