EPA-600/2-76-135a
May 1976
Environmental Protection Technology Series
DEVELOPMENT OF THE WESTYACO
ACTIVATED CARBON PROCESS FOR
SOX RECOVERY AS ELEMENTAL SULFUR
Volume!
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research v
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-135a
May 1976
DEVELOPMENT OF THE WESTVACO
ACTIVATED CARBON PROCESS
FOR SOX RECOVERY AS ELEMENTAL SULFUR
VOLUME I
by
G. Nelson Brown, Carl M. Reed, Albert J. Repik,
Robert L. Stallings, and Samuel L. Torrence
Westvaco
Box 5207
North Charleston, South Carolina 29406
Contract No. 68-02-0003
ROAPNo. 21ACX-085
Program Element No. 1AB013
EPA Project Officer: Douglas A. Kemnitz
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
TABLE OF CONTENTS
Table of Contents iii
List of Figures v
List of Tables ix
Acknowledgements xiii
Management Summary xiv
Sections
1 CONCLUSIONS 1
1.1 Overall Process 1
1.2 Integral Pilot Plant 1
1.3 Bench Scale 2
1.4 Pilot Scale 3
1.5 Prototype 4
1.6 Commercial (1,000 MW) 4
2 RECOMMENDATIONS 5
3 INTRODUCTION _ 6
3.1 Process Concept 7
3.2 Methodology of Contract 9
3.3 Chronological Sequence of Development 10
3.3.1 S02 Sorption 11
3.3.2 Sulfur Generation 11
3.3.3 Sulfur Recovery and H2S Generation 11
3.3.4 Integration 12
4 INTEGRATED PILOT PLANT EQUIPMENT AND RESULTS 13
4.1 Pilot Plant Description 13
4.1.1 Introduction 13
4.1.2 Detailed Pilot Plant Description 16
4.2 Integral Pilot Plant Results 26
4.2.1 Overall Integral Results 28
4.2.2 Detailed Integral Results 36
4.2.3 Material Balances 53
4.2.4 Process Control 56
4.2.5 Process Concept Modifications 59
5 PRE-INTEGRAL PROCESS DEVELOPMENT 61
5.1 Apparatus and Procedure 61
5.1.1 Thermogravimetric Reactor 61
5.1.2 Fixed Bed 66
5.1.3 Moving Bed 66
5.1.4 Batch Fluid Bed 69
5.1.5 Multistage Fluid Bed Reactor 71
ill
-------
TABLE OF CONTENTS (Continued)
5.1.6 Sulfur-Carbon Thermal Equilibrium 71
5.1.7 Solvent Extraction of Sulfur Procedures 73
5.1.8 Procedures in Bench Scale H2S 77
Generation Studies
5.2 Pre-Integral Results 81
5.2.1 S02 Sorption 81
5.2.2 Sulfuric Acid Conversion to Sulfur 114
5.2.3 Sulfur Removal 154
5.2.4 H2S Generation 184
5.2.5 Combined S Stripping/H2S Generation 206
5.2.6 Elemental Sulfur Recovery 220
5.2.7 Fluidizing Mechanics 226
6 1,000 MW UTILITY BOILER FLUE GAS CLEAN-UP 236
6.1 Introduction 236
6.2 General Design Basis 238
6.2.1 Scope 238
6.2.2 Boiler Operating Characteristics 239
6.2.3 Product 240
6.2.4 Process Conditions 240
6.2.5 Activated Carbon Characteristics 242
6.3 Conceptual Design 242
6.3.1 Process Description 242
6.4 Heat and Material Balances 245
6.5 Costs of 1,000 MW Conceptual Design 257
Installation
6.5.1 Cost Summary 257
6.5.2 Capital Costs 257
6.5.3 Equipment Costs 257
6.5.4 Indirect Costs 258
6.6 Operating Costs of 1,000 MW Conceptual 258
Design Installation
7 15 MW DESIGN AND COST 261
7.1 Introduction 261
7.2 Scope of the Prototype Program 261
7.3 Description of Prototype Plant and 262
Operation
7.3.1 General 262
7.3.2 General Design Basis 263
7.3.3 Process Description - Prototype Plant 265
7.3.4 Heat and Material Balance 267
7.3.5 Start-up and Initial Operation 268
7.3.6 Demonstration Operation 268
7.3.7 Technical and Economic Review of 268
Operation
iv
-------
TABLE OF CONTENTS (Continued)
Page
7.4 Technical Approach 269
7.4.1 General 269
7.4.2 Description of Program Elements 269
8 Bibliography ' 273
9 Nomenclature 275
-------
LIST OF FIGURES
No. page
1 Chemistry of Westvaco S02 Process with Reactants 8
and Products Shown
2 Westvaco Process Integral Pilot Plant . 14
3 Mechanical Integration - 20,000 CFH S02 Pilot 17-A
Plant - Process Flowsheet
4 Continuous 18" Dia., 5 Stage S02 and S03 19
Adsorber Operating on Flue Gas from a 50 MW
Oil Fired Boiler
5 Schematic of Integral Westvaco S0£ Removal Pilot 27
Plant
6 S02 Removal Efficiency During Integral Pilot < 31
Tests
7 Carbon Burn-off During Integral Pilot Tests 32
8 Activated Carbon Attrition Rate During Integral 33
Pilot Tests
9 Activated Carbon Performance During Westvaco S02 41
Recovery Integral Pilot Runs
10 S02 Activity as a Function of Carbon Cycle Time 42
as Determined by Bench Scale Apparatus
11 Carbon Attrition, Mean Particle Diameter, and Ash 44
Content as a Function of Carbon Cycle Time
12 Carbon Dioxide Evolution as a Function of Carbon 46
Cycle Time
13 Pore Volume and Surface Area of Recycled Carbon 48
14 Sulfur Generator Performance 49
15 H2S Generator/Sulfur Stripper Performance 51
16 Sulfur Condenser Performance 52
17 Sulfur Balance for IR-2 Run During Operation 57
under Process H2S
18 Thermogravimetric Apparatus 62
19 Detail of the Thermogravimetric Reactor Sample 64
Bucket Envelope
20 Fixed Bed Reactor System 67
21 Moving Bed Reactor System 68
22 • Batch Fluid Bed Reactor 70
23 Multistage Fluidized Bed Reactor 72
24 Sulfur Adsorption Apparatus 74
25 Flow Schematic of Recycle Extraction Apparatus 76
26 H2S Generation Kinetics Apparatus 78
27 Sulfur Vapor Generator 79
28 Comparison of Westvaco Model to Sorption Data at 85
200°F
29 Effect of 02 on S02 sorption at 200°F with NO 88
Present
30 Effect of H20 Concentration on S02 Sorption at 90
200°F with NO Present
31 Effect of NO Concentration on S02 Sorption at 91
200°F
-------
LIST OF FIGURES (Continued)
No.
32 Effect of Temperature on the Westvaco Model with
Constant Order of Reaction for S02
33 Rate Constant as a Function of Temperature for
S02 Sorption
34 Comparison of the Westvaco Model A to Experi-
mental S02 Sorption on Activated Carbon in
Differential Rate Apparatus
35 Comparison of the Westvaco Model B to Experi- 103
mental S02 Sorption on Activated Carbon in
Differential Rate Apparatus
36 Comparison of the Westvaco Model C to Experi- 104
mental S02 Sorption on Activated Carbon in
Differential Rate Apparatus
37 Differential S02 Sorption Rate vs. H2S04 Loading 106
for a S02 Concentration of 2500 ppm at 150°,
200°, and 300°F
38 6" Sorber Data - Plot of Corrected Sorber Rate 109
using Stagewise Westvaco Model A vs. S02
Cone. Showing Curve Predicted from
Differential Bed Studies
39 Summary for Flue Gas Run (Run SA-34) - 18" Dia. 113
S02 Sorber - Water Sprays To Control Temp.
40 Effect of Linear Gas Velocity on Rate of 116
Sulfuric Acid Decomposition
41 Effect of Temperature on the Rate of Conversion 120
of Sorbed Sulfuric Acid to Elemental Sulfur
42 Comparison of the Sulfur Generation Rate Model 124
to the Experimental Data for 250° to 325°F
43 Effect of H20 Cone, on Rate of Sulfur Generation 125
44 Effect of Inlet H2S Cone, on Per Cent Conversion 130
to Sulfur in Simulation Experiments Using a
6" Diameter Fluid Bed Unit for Integrated
Operation with an 18" Diameter S02 Sorber
45 Effect of H2S Cone, and Carbon Residue Time on - . 131
Acid Evolved as S02 in Simulation Experiments
Using a 6" Dia. Fluid Bed Unit for Integrated
Operation with an 18" Dia. S02 Sorber
46 Moving Bed Sulfur Generator 139
47 Effect of Inlet Carbon Temp, on the Evaluation 144
of H2S04 as S02 in an 8" Dia. Moving Bed
Reactor
48 Effect of- Inlet Carbon Temp, on H2S Utilization 145
in an 8" Dia. Moving Bed Reactor
49 Effect of the Inlet Carbon Temp, on the Per Cent 146
Conversion to Sulfur
50 Effect of Steam Heater on Improved Carbon 148
Heating Capabilities
vii
-------
LIST OF FIGURES (Continued)
No. Page
51 Effect of Heat Exchanger System (1-1/2" Pipe 149
x 4") on Carbon Flow in an 8" Dia. Moving
Bed Reactor - Tracer Feed Composition)
52 Experimental Adsorption Isotherm Points for 157
Sulfur on Activated Carbon at 800° and 1000°F
53 Polanyi-Dubinin Plot of Sulfur Adsorption Data 158
54 Equilibrium Lines for Concentration vs. 160
Temperature at Various Loadings
55 4" Dia. Batch Fluidized Bed S Stripping and 163
Hydrogen Desulfurization Runs
56 Extraction of Sulfur Loaded Activated Carbon 166
with 15 Wt. % (NH4)2S Solution at 40°C
57 Extraction of Sulfur Loaded Activated Carbon 168
with CS2 at 25°C
58 Extraction of Sulfur Loaded Activated Carbon 169
with Xylene at 105°C
59 Effect of Percent Sulfur on Carbon on the S02 175
Activity
60 Effect of Recycle on S02 Ability for Isothermal 178
and Thermal/Reductive Regenerations
61 Effect of Treatment Time Using Hydrogen Post 182
Treatment of (NH4)2S Extracted Sample of
Sixth Cycle
62 Effect of Temperature of Hydrogen Post Treatment 183
of (NH4)2S Extracted Sample of Sixth Cycle
63 Equipment Schematic for H2 Chemisorption 188
Experiments on Virgin Carbon
64 Arrhenius Plots for Experimental and Literature 197
Data
65 Test of Integral Rate Equation, Z as Function of 203
Tube Volume
66 Test of Integral Rate Equation, Z as Function of 204
Tube Area
67 Variation of Conversion with Residence Time for 207
Three Bed Volumes
68 Effect of Flow Rate on Conversion at Constant 208
Residence Time
69 Effect of Temperature on Sulfur Stripping with 214
H2 Percent
70 Effect of H2 Concentration and Gas/Solid Contact 215
Time on Sulfur Stripping at 1200°F
71 Sulfur Removal from Activated Carbon in an 8 217
Stage, 4" Dia. Regenerator
72 Effect of Temperature on the Conversion of Sulfur 218
to Hydrogen Sulfide
viii
-------
LIST OF FIGURES (Continued)
No. Page
73 Operating Conditions for Sulfur Condenser 225
74 Experimental Determination of Minimum Fluidizing 228
Velocity for Westvaco Granular Carbon
75 Pressure Drop Characteristics of Distributor 230
Plates To Be Used in an 18" Dia. S02 Sorber
76 Westvaco S02 Recovery Process Schematic 237
Flowsheet (Dwg. 2563)
77 Westvaco S02, Process Flowsheet for 1,000 MW Unit 243-A
(250 MW Typical Module Shown) (Dwg. 2572)
78 Westvaco S02 Process Flowsheet for 15 MW 266
Prototype Unit (Dwg. 2573)
79 Prototype Program Schedule 270
ix
-------
LIST OF TABLES
No.
1 Range of Operating Conditions for Integral Pilot 29
Plant Run
2 Properties of Sulfur Product 31
3 Effect of Hydrogen Input on By-product Recovery 35
4 Integrated Operating Conditions and Results 37
5 Process Operating Performance 38
6 Carbon Balance for IR-2 Integral Run 53
7 Hydrogen Balance for H2S Generator/S Stripper 54
8 Sulfur Balance for Integral Runs 55
9 Rate Expressions To Approximate S02 Sorption 82
Data
10 Experimental Conditions for S02 Sorption in a 83
Differential Rate Apparatus
11 Standard Deviation for Westvaco Equation for 84
Sorption Data at 200°F with NO Present for
Acid Loading above 0.01 gm. Acid/gm. Carbon
12 Deviation of Westvaco Equation from Experimental 86
S02 Sorption Rate Data at 200°F with NO for
Loadings above 0.01 gm. Acid/gm. Carbon
13 Standard Deviation for Modified Westvaco Equation 86
for Sorption Data at 200°F with NO for Acid
Loading above 0.01 gm. Acid/gm. Carbon
14 Experiments To Determine 02 Dependency 87
15 Experiments To Determine H20 Dependency 89
16 Experiments To Determine NO Dependency 89
17 Experiments To Determine Effect of Temperature on 93
S02 Sorption
18 Rate Constants for the Westvaco Model 93
19 Deviation of Westvaco Model from Differential 97
Rate Data for Acid Loadings above 0.01 gm.
Acid/gm. Carbon
20 Deviation of Multiple Regression Models from 102
Differential Rate Data for Acid Loadings
above 0.01 gm. Acid/gm. Carbon
21 Comparison of Rates from 6" Dia. Sorber to Rates 107
Calculated from the Westvaco Model
22 Comparison of Predicted Number of Stages to 108
Actual Number for 6" Sorber Runs
23 Water Spray Cooling Tests Made in Pilot Fluid Bed 111
Reactors with Simulated and Actual Flue Gas
24 Experimental Conditions and Results for Sulfur 118
Generation Experiments in an 8 Stage, 4" Dia.
Fluidized Bed Regenerator
25 Overall Rates of Acid Decomposition and Conversion 119
to Sulfur for the Reaction
3 H2S + H2S04 -»• 4 S + 4 H20 in an 8 Stage, 4"
Dia. Regenerator
x
-------
LIST OF TABLES (Continued)
No. Page
26 Experimental Conditions for Differential Sulfur 122
Generation Runs
27 Summary of Sulfur Generation Results 129
28 Comparison of the Fluid Bed Design Model with 133
Experimental Sulfur Generation Fluid Bed Data
29 Fixed Bed Sulfur Generation Experiments 134
30 Comparison of Fluid Bed and Moving Bed Sulfur 136
Generation Tests
31 Data Summary - 1-1/2" Dia. Moving Bed Sulfur 137
Generation Tests
32 Design Conditions - Moving Bed Sulfur Generator 138
33 Summary of Sulfur Generation Results 141
34 Effect of Vol. % H20 and Temperature on the Cone. 151
of Acid Solution Sorbed on Carbon for Carbon
Preheater
35 Comparison of the Moving Bed Design Model with 153
Experimental Sulfur Generator Moving Bed Data
36 Experimental Results of Equilibrium Sulfur 156
Adsorption Measurements
37 Isosteric Heats of Adsorption of Sulfur Vapor on 161
Carbon
38 Sulfur Stripping in a Continuous 8-Stage Fluid 162
Bed
39 Effect of Solvent on Virgin Carbon 165
40 Comparison of S02 Activity and Surface Area and 171
Pore Volume Measurements
41 Effect of Recycle on S02 Activity for Isothermal 176
and Thermal/Reductive Regenerations
42 Pore Volume Distribution Results Using Engelhard 179
Isorpta Apparatus
43 S02 Activities Integral Rate Determined Using 179
Differential Rate Apparatus vs. Using Fixed
Bed
44 Effect of Post Treatments of (NH4)2S Extracted 181
Sixth Cycle Carbon on S02 Activity
45 Planned Experimental Program for Studying Hydrogen 186
Hydrogen Chemisorption on Activated Carbon
During Regeneration of C
46 H2 Chemisorption on Virgin Carbon 187
47 Experimental Conditions for H2S Generation Rate 190
Experiments
48 Comparison of H2S Formation Rates from Literature 191
and Experimental Data (Homogeneous)
49 Comparison of Homogeneous and Heterogeneous 192
Reaction Rates for Similar Inlet Concentrations
50 Experimental Conditions for Series HS-2 194
XI
-------
LIST OF TABLES (Continued)
No. Page
51 Comparison of Rate Constants at Different 195
Temperatures
52 Comparison of Rates and Rate Constants Based on 198
Reactor Surface Area from Series HS-2
53 Experimental Conditions for Runs HS-4 to HS-7 205
54 Experimental Conditions for Evaluation of 211
Combined S Stripping/H2S Generation
55 Experimental Conditions for Evaluation of 212
Combined S Stripping/H2S Generation
56 Experimental Results from Evaluation of Combined 213
S Stripping/Has Generation
57 Operating Conditions for Sulfur Condenser 221
Testing System
58 Sulfur Condenser Test Runs 222
59 Sulfur Condenser Operation 223
60 Calculated Values for Minimum Fluidizing 227
Velocity and Entrainment Velocity for Westvaco
Granular Activated Carbon
61 Operating Characteristics Distributor Plates To 232
Be Used in an 18" Dia. S02 Sorber
62 Gas Distributor Plate Characteristics Evaluated 233
for Carbon Weeepage During Fluidization
63 Gas Distributor Plate Specifications Designed for 235
Minimizing Carbon Attrition in the 18"0 S02
Sorber
64 Overall Sulfur Balance for 1,000 MW Power Plant 245
65 Overall Energy Balance for 1,000 MW Power Plant 246
66 Stream Conditions 247
66 Stream Conditions 248
68 Stream Conditions 249
69 Stream Conditions 250
70 Stream Conditions 251
71 Stream Conditions 252
72 Stream Conditions 253
73 Stream Conditions 254
74 Stream Conditions 255
75 Stream Conditions 256
76 Cost Summary 257
77 Capital Cost Summary 258
78 Annual Operating Costs 260
79 Cost of Prototype Program 272
xii
-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge the assistance of Doug Kemnitz
and Leon Stankus, Project Officers, whose guidance has been a
significant factor in carrying out this demonstration program.
We also would like to thank all the other individuals at EPA
who have contributed their time and effort to this project.
We would like to acknowledge the contributions made by
Dr. Frank J. Ball, Associate Corporate Research Director for
Westvaco. His guidance, suggestions and enthusiasm have
played a significant role in the development of the process.
We would like to also acknowledge the numerous contributions
made by many individuals at Westvaco. Listed alphabetically:
E. A. Ankersen, E. C. Arms, N. L. Davis, R. Deleon, R. C. Flowe,
B. J. Gooch, C. E. Grooms, W. D. Heape, H. R. Johnson,
E. B. Lipscomb, G. F. McAllister, C. C. Matthews, M. Nelson,
W. C. Rhodes, R. L. Roquemore, C. B. Rowe, C. Smith,
H. E. Sparks, R. A. Stanton, I. A. Stine, S. R. Thompson,
E. D. Tolles, B. H. Van Dyke, W. A. Wier, R. M. Wise. Many
other individuals at Westvaco were involved and thanks are
also extended to them.
Lastly the authors wish to acknowledge the efforts of
L. K. Hallex who was primarily responsible for typing and
editing this report. Her continuing effort was also signifi-
cant in the many reports and pieces of communication prepared
during the project.
xiii
-------
MANAGEMENT SUMMARY
An all dry fluidized bed S02 recovery process using activated
carbon has been developed by the Westvaco Corporation to
recover sulfur oxides from waste gases with elemental sulfur
as a product. The process was developed for five years by
Westvaco before joint development with the Environmental
Protection Agency (EPA) began in January 1971. During a
pause in the contract from December 1971 to September 1972,
Westvaco continued development of the process, and then from
September 1972 to June 1974 joint development continued with
EPA. The process has been shown to be technically feasible
and economically attractive compared to other processes cur-
rently being developed. Future plans are, therefore, to
demonstrate the process at the next stage of development,
presently anticipated as the equivalent of a 10 to 20 mega-
watt coal fired power plant.
In the process the activated carbon catalyzes the oxidation
of the sulfur oxides in the flue gas and adsorbs these con-
stituents as sulfuric acid. The carbon is regenerated and
the sulfuric acid is reduced to elemental sulfur in unit
operations separated from the flue gas clean-up. These two
steps of regeneration use a hydrogen containing gas to regene-
rate the carbon and reduce the sulfuric acid to elemental
sulfur by way of an intermediate reductant, hydrogen sulfide,
which is produced and recycled within the regeneration system.
The regenerated carbon is reused for flue gas clean-up and the
elemental sulfur is recovered to be stored or sold.
The demonstration of the activated carbon process for recover-
ing sulfur oxides as sulfur has been completed at an integral
pilot plant stage during which a slipstream of flue gas from
a 50 MW oil fired boiler was treated. The sulfur oxides were
recovered as elemental sulfur of over 99.77o purity for over
300 hours. During the integral run, the activated carbon was
continuously regenerated and reused more than 20 times for
flue gas clean-up and the carbon maintained its activity with
no decrease observed, as indicated by the average removal of
94% of the 2000 ppm sulfur oxides.
The hydrogen requirement of the process was found to be 3.9
moles of hydrogen per mole of sulfur oxide recovered. This
hydrogen requirement is 30% above that required by stoichi-
ometry for the reduction of sulfuric acid to elemental sulfur.
The recovery process was operated successfully as a closed
loop system, as anticipated in commercial operation.
xiv
-------
A considerable amount of process development preceded
the integral demonstration of the process. The major
technical developments were that a granular activated
carbon was chosen, the process chemistry established,
and the variables affecting the process specified. In
addition, the SOo removal process step was demonstrated
in a multistage fluid bed reactor treating a simulated
flue gas and an actual flue gas stream from an oil-fired
boiler. A satisfactory design procedure was developed
for this process step. The acid conversion to elemental
sulfur was also demonstrated in multistage fluid bed
equipment and a design procedure was developed.
Mechanical constraint, however, necessitated that a
moving bed reactor be used in the integral pilot plant
run. Several methods of sulfur recovery were assessed,
but the stripping of the sulfur with a ^-containing gas
appeared to be most suitable to sustain a high level of
S02 activity of the regenerated carbon. The sulfur re-
moval and internal H2S generation steps which finally
evolved were demonstrated as a combined operation in one
multistage fluid bed reactor. Sufficient information was
developed on the carbon regeneration step to, hopefully,
insure satisfactory scale-up to the next level of
anticipated development.
All of this process and design information which was
developed has permitted scale-up to a prototype unit of 15
MW. The preliminary 15 MW demonstration program includes
installation on a coal fired boiler and the use of a coal
fed gas producer to supply the necessary reducing gas.
The installation is anticipated to cost about $2.4 million
to install. The demonstration of the process at the proto-
type scale is anticipated to require 3 years at a total
operating cost of $1.4 million, or about $0.5 million/
year.
An economic assessment was also made for the recovery
process installed at a 1000 MW power plant. The capital
investment was estimated to be $35/KW and the annual
operating cost was estimated to be 2.0 mills/KWH. Even
though the costs of the process may increase after
additional process development information is obtained
at the prototype scale, these costs indicate that the
process is worthy of consideration for continued process
demonstration.
It is concluded that the technical viability and economic
competitiveness of the sulfur oxide process developed by
Westvaco with partial funding from EPA has been demonstrated.
It is recommended that the process be demonstrated at a
prototype stage on a coal fired boiler.
xv
-------
SECTION 1
CONCLUSIONS
The general objectives of the contract were to develop
further information on each process step of the
Westvaco S02 Recovery Process, to demonstrate the tech-
nical feasibility of the entire process, to evaluate
the performance of the carbon under extended recycling
conditions in an integrated pilot plant using flue gas
from an oil fired boiler, and to scale up the process
to a larger power plant installation.
1.1 CONCLUSIONS - OVERALL PROCESS
1. The process has shown to be technically feasible
and economically competitive.
2. The process can use activated carbon to
effectively remove flue gas S02 and 803 and
recover elemental sulfur by-product.
3. Sufficient information has been generated on the
performance of the activated carbon, process
chemistry and pilot operation to proceed to the
next stage of development.
1.2 CONCLUSIONS - INTEGRAL PILOT PLANT
The basic conclusions reached in the integral pilot
plant are:
1. Granular activated carbon of the type used can
effectively remove S02 from flue gas and can be
regenerated satisfactorily over a repeated number
of cycles without reduction in activity or an
unacceptable physical loss through chemical reac-
'tion or mechanical attrition.
2. Information has been developed on' each of the
three unit process steps, S02 sorption, sulfur
generation, and S stripping/H£S generation, to
define the principal variables affecting the
process chemistry and their correlations in regard
to rate of reaction.
3. An acceptable sulfur product can be produced by the
process using hydrogen as a reducing gas with H2S
as an internally generated intermediate reductant.
-------
4. Carbon burn-off is minimized, in fact almost elimi-
nated, in the S02 recovery process as proposed.
5. Mechanical attrition of the carbon was acceptable
with the improved carbons used in these tests and
further reduction by a factor of 3 or more is
indicated.
6. Complete reduction of sulfuric acid is not required
in the sulfur generator, but part of the acid can
be effectively reduced to elemental sulfur in the
sulfur stripper/H2S generator without undesirable
side effects, such as an increase in carbon burn-off.
7. The hydrogen requirement of the process was 3.9
moles of H£ per mole of sulfur oxide recovered for
the pilot plant conditions tested.
8. Use of fluidized beds present a viable and attractive
method of gas-solids contacting, although other
contacting means are also applicable.
9. Operation of the integral pilot plant over the
limited time did not appear to present any problems
in regard to control of the process.
1.3 CONCLUSIONS - BENCH SCALE
1. The S02 removal kinetics are a function of tempera-
ture, of gas concentrations of 02, H20, S02 and NO
and of the acid loading on the carbon.
2. The S02 kinetics developed from a bench scale differ-
ential reactor could be modeled by an empirical
expression and incorporated into a procedure for
reactor design.
3. Sulfuric acid adsorbed on the carbon can be con-
verted to elemental sulfur by reaction with H2S.
4. The kinetics of sulfur generation are functions of
acid concentration adsorbed on carbon, of tempera-
ture, and of gas concentrations of H2S and H20.
5. The kinetic data of sulfur generation can be repre-
sented by an empirical expression and can be incor-
porated into a procedure for reactor design.
-------
6. Sulfur adsorbed on carbon can be recovered by
solvent extraction, but the S02 activity of the
regenerated carbon can only be maintained at a high
level if it is treated further by thermal means.
7. Sulfur adsorbed on carbon can be removed by vapori-
zation and the S02 activity is maintained, especially
when exposed also to a hydrogen containing gas at
the temperature the sulfur is vaporized from the
carbon.
8. Equilibrium data of sulfur adsorbed on carbon can
be represented by an empirical expression known as
the Polanyi-Dubinin adsorption equation.
*
9. The kinetics of H2S generation over an activated
carbon catalyst can be represented by an empirical
expression.
1.4 CONCLUSIONS - PILOT SCALE
1. Multistage fluid bed S02 sorber can be effectively
used with activated carbon to recover S02 and 803
from flue gas of an oil fired boiler.
2. Direct flue gas cooling by water spray injection
after S03 removal can be accomplished in a fluid
bed reactor to improve S02 removal efficiencies and
minimize gas cooling costs.
3. A moving bed reactor for acid conversion to sulfur
used in the integral pilot plant because of mechani-
cal limitations associated with a fluid bed reactor
was satisfactory for acid conversion.
4. The process steps of sulfur stripping and H2S genera-
tion can be combined into a single operation.
5. Operating temperatures for the existing sulfur
stripper/H2S generator near 1000 to 1200°F are indi-
cated based on sulfur removal from the carbon.
6. A sulfur condenser utilizing recirculating sulfur
for a scrubbing fluid is suitable for use in inte-
gral operation to recover sulfur from the H2S
recycle gas to the acid converter.
7. A shell and tube cooler is suitable to cool regene-
rated carbon to be reused in the S02 sorber.
-------
1.5 CONCLUSIONS - PROTOTYPE
1. The equivalent of a 15 MW power plant appears to be
a suitable size to which to scale up the process.
2. The process information developed to date was suf-
ficient to design an installation of the S02 recovery
process at a 15 MW power plant.
1.6 CONCLUSIONS - COMMERCIAL (1,000 MW)
1. The S02 Recovery Process appears to be economically
competitive with other processes now being developed
for S02 recovery.
-------
SECTION 2
RECOMMENDATIONS
1. Based on conclusions from the integral pilot run
and process development to date, it is recommended
that scale-up to a larger prototype plant be
pursued as the next step toward a commercial
plant.
2. The prototype unit should be installed on a coal
fired boiler.
3. Provisions should be made to test the prototype
installation for use both upstream and downstream
of the precipitator.
-------
SECTION 3
INTRODUCTION
The use of activated carbon in dry regenerable S02
recovery processes avoids the critical control of chemi-
cal reactions necessary in wet processes and the costs
and problems involved in separating water from
by-product, either for recovery or disposal. Addition-
ally, in the wet processes, flue gas reheating may be
necessary for fan protection and plume control. Carbon
processes which have or are being used all depend upon
the catalytic and sorptive character of the carbon for
conversion of the S02 to sulfuric acid within the carbon
granules. These'processes generally differ in the mode
of removal and recovery of the sulfuric acid from the
carbon. In thermal regeneration, the acid reacts
chemically with the carbon to produce a S02 rich
by-product off-gas. In wet regeneration, the acid
loaded carbon is washed with water to produce a weak
sulfuric acid. Further differences exist in the addi-
tional methods of upgrading the by-product streams
through add-on steps for conversion of the S02 gas
stream to elemental sulfur or concentrated sulfuric
acid. The method of contacting flue gas with granular
carbon also varies in that fixed beds or moving beds
with an upflow or crossflow gas pattern are used.
Particle size and characteristics of the carbon granules
with respect to the rate of S02 removal may differ,
affecting pressure requirements and equipment size.
Westvaco, as a major producer of activated carbon,
embarked on a program in which carbon, with a high S02
pickup rate capability, is recycled with regeneration
of the carbon achieved by reducing the sulfuric acid
chemically within the process to elemental sulfur and
without the carbon being consumed. Furthermore, the
fluidized bed was selected for gas-solid contacting as
a stagewise approach to permit handling relatively
large volume rates of gases in contact with recirculat-
ing carbon solids. The effectiveness of fluidized
carbon bed systems has been demonstrated in large
commercial units in existence, handling gas rates up
to 540,000 cfm. The feasibility of using such a carbon
system was confirmed in bench scale and small pilot
equipment whereby H2S in contact with the sulfuric acid
on the carbon resulted in conversion to elemental
sulfur which was then stripped off the carbon by heating
6
-------
and then recovered as a 99.770 pure product. An outside
source of hydrogen was reacted with sulfur within the
regeneration system to produce the needed H2S.
The selection of a granular activated carbon and
identifying the major variables in the process chemistry
served as a basis for the joint work under this EPA con-
tract which essentially involved scaling up the SC>2 re-
moval sorption and the regeneration-sulfur recovery
steps to a 20,000 cfh pilot plant. The objectives of the
contract were initially to develop further information
on each process step and finally to demonstrate the
technical feasibility of the entire process and to
evaluate the performance of the carbon under extended
recycling conditions in an integrated pilot plant using
flue gas from an oil fired boiler.
3.1 PROCESS CONCEPT
In the Westvaco Process, dry granular activated carbon is
contacted with flue gas at stack gas temperatures. The
SO? is removed through catalyzed oxidation to 803 and
subsequent hydrolysis to sulfuric acid which remains
sorbed in the carbon granules, i.e.
S02 + 1/2 02 + H20 . arko H2S04 (Sorbed) (1)
Sufficient water vapor and oxygen are present normally
in the flue gas for the reaction. This reaction takes
place in a staged fluidized bed vessel with provisions
for adjusting the temperature for optimum S02 removal
rates .
The sulfuric acid loaded carbon is transported mechanically
to a second fluidized bed reactor wherein the acid comes
in contact with hydrogen sulfide to produce elemental
sulfur, which remains in the carbon granules, and water
vapor which is exhausted. Temperatures near 300°F are
required for the reaction, i.e.
H2S04 + 3 H2S Activated^ 4 s + 4 H 0 (2)
*• * *• Carbon /
-------
Generation of the required hydrogen sulfide and the
removal of the elemental sulfur for recovery are accom-
plished in a third fluidized bed reactor according to:
3 H2 + 4 S Activated^ 3 H2S + S (Product)
L* 3. IT DOXl
(3)
The thermal stripping of the sulfur and the reaction to
produce H2S requires temperatures near 1000°F.
S02 + 1/2 02 + H20 + 3 H2
4 H2°
(4)
S02
S02
SORBER
1/2 02, H20
3 H2
H2S04
S
GEN.
4 H20
3 S
SSTKEV
;S
3 H2S
Figure 1. Chemistry of Westvaco S02 Process with
reactants and products shown
8
-------
That is, the sulfur dioxide or sulfur trioxide from the
flue gas is reduced with hydrogen in the form of
recycled H2S to form water and elemental sulfur.
The hydrogen may be supplied through a number of commer-
cially available gasifiers utilizing coal or other
fossil fuels. Heating of the regenerating reactors may
be provided by conventional fuel burning units.
The carbon stream, not shown in Figure 1, serves as a
carrier and catalyst for promoting the reactions effi-
ciently, but does not directly take part. It is
recycled from the S02 removal vessel to the regeneration
vessels where its activity is restored before returning
to the S02 sorber.
3.2 METHODOLOGY OF CONTRACT
The main objective of this contract was to achieve inte-
grated operation for a sufficient time to determine the
effect on carbon and process performance while treating
flue gas from an oil fired boiler. Work was conducted
in two parts: from January to December 1971 and from
September 1972 to June 1974.
In the first part of the contract, the technical feasi-
bility and design information for the integral pilot
plant were to be determined, operating the S02 sorber on
real flue gas. These objectives were basically com-
pleted, with particular emphasis on the flue gas desul-
furization step. The kinetics of S02 removal were deter-
mined with bench scale equipment. The kinetic data were
modeled by an empirical expression, which was used in
the development of a design procedure for the S02 sorber.
Concurrent with the bench scale work on S02 sorption, a
variable study was made with actual flue gas from an
oil fired boiler. It was shown that the design proce-
dure developed from bench scale work satisfactorily
represented the pilot results. Acid conversion to
elemental sulfur was studied extensively in multistage
pilot fluid bed equipment. Sulfur recovery from acti-
vated carbon was assessed by both thermal and isothermal
(solvent extraction) schemes. The evaluation of
solvent extraction included testing of a number of
solvents. Ammonium sulfide was judged to be the most
suitable. The thermal process of sulfur vaporization
from the carbon was then compared to the solvent extrac-
tion process. The thermal process evolved as the most
-------
suitable since the carbon regenerated by extraction
required additional thermal treatment to maintain the
S02 activity, whereas in thermal regeneration, the S02
activity was maintained without additional carbon
treatment. This process development in the first
phase of the contract provided grounds for extended
development to be culminated by operation of an inte-
gral pilot plant run for an extended period of time.
During the second part of the contract, additional data
was to be developed on the regeneration process steps
and an integral pilot plant was to be operated. The
design information developed was to be used to design
and cost a prototype installation for 10 to 15 MW
boiler. Also, the cost of a 1,000 MW installation was
to be projected. To expedite the completion of these
objectives, detailed program plans were drawn,up. In
the pilot development leading to the integral pilot
plant, the two key points were 1) could existing equip-
ment be used for acid conversion to sulfur during
integral operation, and 2) could the process steps of
sulfur stripping and H2S generation be carried out in
one process unit. It was shown that existing fluid
bed equipment could not be used for acid conversion
during integral operation. Subsequently a moving bed
sulfur generator was designed, installed and operated
as part of the preparation for integral operation. It
was also shown that the two process steps of sulfur
stripping and H2S generation could be carried out in
one reactor. All of the pilot equipment was integrated
and operated mechanically on an oil fired boiler, then
operated for an extended period to demonstrate long
term effects on process and carbon performance.
All performance information was used to design and cost
a prototype unit for installation on a 15 MW coal fired
boiler. Also the economics of the process for a 1,000
MW installation were assessed.
3.3 CHRONOLOGICAL SEQUENCE OF DEVELOPMENT
Development of the main steps of the Westvaco S02
Recovery Process occurred in several phases using vari-
ous types of equipment. The chronological sequence of
development is described for the main steps of the
process.
10
-------
3.3.1 S02 Sorption
Pre-contract experimentation with S02 sorption was carried
out initially in a 1" diameter fixed bed, and later in 4"
and 6" diameter multistage fluidized beds. Under the con-
tract, additional work was done in the 6" diameter unit
operating on flue gas from an oil fired boiler, and an
extensive kinetic study was carried out in a bench scale
differential reactor to obtain SO? sorption rate data,
which were modeled by an empirical expression. An 18"
diameter sorber was built and used initially to produce
the large quantities of acid loaded carbon needed to study
the other steps of the process. The 18" unit eventually
was used as the sorber for the integral pilot plant.
3.3.2 Sulfur Generation
Sulfur generation work was begun in a 1" diameter fixed
bed reactor but this proved unsatisfactory due to certain
inherent disadvantages of fixed beds. Experiments were
then carried out in a 4" diameter, 8 stage fluid bed
glass reactor and information was obtained on the reaction
rate and effects of important variables. A more
accurate kinetic study was conducted on a batch differential
basis, and this yielded a rate model. After construction
of the 4" diameter, 8 stage multipurpose reactor at the
powerhouse location, additional sulfur generation tests were
made in that reactor to determine whether the 6" diameter
reactor could serve as the sulfur generator for the inte-
grated pilot plant. These tests showed that the 6" unit
was unsatisfactory for the intended application and efforts
were subsequently directed toward development of an alter-
native moving bed sulfur generator. Bench scale work was
conducted in a 1.5" diameter moving bed and the results were
used to design an 8" diameter unit, which was installed
at the powerhouse location. Sulfur generation studies
were made with the 8" diameter moving bed to determine whether
its performance was adequate and to optimize operating con-
ditions. Satisfactory performance was obtained from the
unit and it became a component of the integrated pilot plant.
3.3.3 Sulfur Recovery and H2S Generation
In the early stages of the project, two different approaches
were considered for the removal of sulfur from carbon
solvent extraction and thermal stripping. Solvent extrac-
tion was determined to be unsuitable, primarily because the
S02 sorption activity of the carbon was reduced excessively
by the operation. After abandonment of the solvent extrac-
tion approach, efforts were directed toward high temperature
thermal stripping. Laboratory data was obtained on the
11
-------
equilibrium adsorption of sulfur on carbon at high
temperatures. In addition, a laboratory rate study was
made of the H2S generation steps in a single reactor.
The combined operation was studied in the 4" diameter,
8 stage fluidized bed reactor at the powerhouse, and
feasibility was demonstrated. The 4" diameter unit
functioned as sulfur stripper/I^S generator in the
integral pilot plant runs.
Recovery of sulfur product from the off-gas of the 4"
unit required development of a sulfur condensing system.
A condenser was designed and fabricated, and tests were
conducted. Satisfactory performance eventually was
obtained and the sulfur condenser was installed in the
integral pilot plant.
3.3.4 Integration
After each of the pilot scale processing units had been
operated successfully on an individual basis, they were
tied together to form what was termed "the integrated
pilot plant". Integration required: 1) closing the
carbon flow loop to permit continuous recycling of carbon
through the three reactors in the system, and 2) con-
necting the off-gas line from the 4" diameter sulfur
stripper/H2S generator to the gas inlet of the 8" diameter
sulfur generator, permitting operation with internally
generated H2S.
12
-------
SECTION 4
INTEGRATED PILOT PLANT EQUIPMENT AND RESULTS
4.1 PILOT PLANT DESCRIPTION
4.1.1 'Introduction
The integrated pilot plant for continuous removal of S02
from flue gas and reduction of S02 to elemental sulfur
consists of three main reaction vessels plus various other
processing units and auxiliary equipment, as shown in
Figure 2. The three main reactors and the reactions that
occur in each are:
S02 Adsorber SO + 1/2 0, + EUO •• H?S04 (5)
18" Dia., Fluid Bed L
Sulfur Generator H?S04 + 3 H«S » 4 S + 4 H20 (6)
8" Dia., Moving Bed
S Stripper/H2S Generator 3 H2 + 4 S —*. 3H2S + S (7)
4" Dia., Fluid Bed
NET REACTION S02 + 1/2 02 + 3 H2 » S + 3 H20 (8)
Other processing units include a sulfur condenser, fluid
bed carbon preconditioner, carbon cooler, and carbon
preheater. Auxiliary equipment includes carbon con-
veyors and flow controls, gas flow controls, electrical
heaters, temperature controllers and recorders, gas
analyzers, gas sample lines, pressure lines, blowers,
and dust collection devices.
The pilot plant processes a 20,000 acfh flue gas slip-
stream from a 50 MW oil fired boiler. Granular activated
carbon circulates by gravity flow downward through each
reactor at a rate of about 30 lbs./hr., with reactant
gases passing countercurrently upward through the vessels.
The activated carbon has a catalytic effect in all three
reactions. S02 is removed from the flue gas in the
adsorber. Regeneration of the sulfuric acid laden carbon
13
-------
Figure 2. Westvaco Process integral pilot plant
SOg 2000 PPM
20,000 CFH
S0g SOUSE* (
H"OI* X 17.5 FT (
ITS *F
SULFUR GENERATOR
8" DIA. X SFT
300 "F
S. STRIPPER
H2 S OENERATOR
4" DIA. X 19 FT.
8 STA8E8
IOOO°F
CAMON
I M<
FIMST STAftE
FLUID BED
OVERFLOW
WKIN
KCONO STAGE
FLUID KO
OA>
OI3TRIIUTOR
FLUIDIZIN9 8AS
FLUID BED DETAIL
iULFUR
REGENERATED CARBON
RECYCLE, APROX 30Lb/HR
BOLIDS
NATE
CONTROLLER
14
-------
is accomplished in the other two reactors, In the sulfur
generator the H2S04 is reduced to elemental sulfur by reac-
tion with H2S. The sulfur remains sorbed on the carbon. In
the combined sulfur stripper/H2S generator, hydrogen gas
reacts with about 75% of the sulfur to generate the H^S
for the sulfur generator. The remaining sulfur is stripped
from the carbon thermally and leaves the reactor as a vapor
in the off-gas. The sulfur vapor is removed from the off-
gas in a condenser and recovered in molten form. The
sulfur free off-gas containing HoS then passes to the
inlet of the sulfur generator. The regenerated carbon is
recycled back to the adsorber. The hydrogen rich gas used
to generate H2S for the sulfur generation step is the only
raw material besides activated carbon that is required by
the process.
Instrumentation and Control -
Sufficient instrumentation is available to maintain the
desired operating conditions during steady state conditions
and to collect the data necessary for performance
evaluation. All input gas flow rates are monitored through
meters and checked by gas analysis instruments. Tempera-
tures and pressures at appropriate points within the system
and reactors are either indicated or recorded.
i
Sample Points and Analysis -
Ports were positioned on the inlet and outlet of each of
the three reactors for sampling the granular carbon to
determine the amount and form of sulfur and moisture content.
Gas sample ports were also positioned so that various inlet
and outlet points in the system were analyzed chromato-
graphically for H2, 02, H2S, S02, N2, C02, CO and H20 at the
desired time. Samples of the carbon were analyzed using
standard tests for measuring the physical and adsorption
properties.
Granular Carbon -
The carbon used in the integral pilot operation is a commer-
cially producible coal-based carbon with a nominal 12x40
mesh size and bulk density of about 40 Ibs./cu. ft. The
S02 number and attrition number were 60 minimum and 97 maxi-
mum, respectively, as determined by specially designed
tests. This carbon showed satisfactory attrition resistance
in pilot testing. Carbons with improved resistance are
being developed. v
15
-------
Flue Gas Characteristics -
Flue gas used in the integral testing was from the stack of
a 50 MW oil fired boiler having a mechanical dust collector.
The sulfur content of the oil was about 1.8-2.0%, which
produces about 1100 ppm S02 in the flue gas. In order to
avoid variability at this stage of operation, provisions
were made for injecting additional S02 into the flue gas to
maintain a uniform level to the pilot plant. The tempera-
ture of the flue gas was kept at stack temperature (300°F)
before introducing into the S02 sorber.
Reducing Gas Composition -
The reducing gas was a mixture of hydrogen and nitrogen
from gas cylinders. Hydrogen content was varied from 40 to
48% of the total flow to establish process requirements.
General Operating Procedure -
In starting up the integral system, a known quantity of
carbon, about 500 pounds, was placed in the system and
recirculated while preheating with a start-up heater to
approach the desired operating temperatures. The switch to
flue gas was then made. The temperatures, carbon flow rate,
and gas flows and compositions were adjusted to the desired
conditions. Manual adjustments were made to the S02 added
to the flue gas above the actual oil produced S02 to main-
tain a constant level. The amount of carbon placed in the
system was sufficient to minimize adding fresh carbon
during the integral run and represents about 407o above that
needed to fill the reactors and conveying system.
4.1.2 Detailed Pilot Plant Description
A detailed flowsheet of the pilot plant is shown in
Drawing 2507 (Figure 3). The following description of the
pilot plant equipment is organized to follow the path of
the activated carbon as it moves through the system, begin-
ning with the S02 sorber. The operation of each processing
unit is described also, including important operating condi-
tions and procedures. Equipment identifications are based
on the process flowsheet, Figure 3. Instrumentation iden-
tifications are based on the instrumentation flowsheet,
Drawing 2528A in Appendix B-3,
S02 Sorber [RV-10] -
The sorber [RV-10] is an 18" dia. x 17.5 ft. high vessel
with five fluidized beds of carbon, each bed having an
expanded depth of 12 inches except for the bottom bed which
16
-------
17
-------
17-A
-------
Figure 3. Mechanical integration - 20,000 cfh S02
pilot plant - process flowsheet
a
b
> "****• **"•
/«" */=>
lf....s.-.y*?.y yr_>_
t"* Doc T "
i^r^
bj
-
IT ettv f-i
* IT t't i*
i
'
fw
. M
t ^
'
,
'3w.y-to-.MV
• IftovjrjhrfT
T
» J«.
« $"«.•./«> (
•f* -6f> 5(Z* (Hi
+ J-* CrHSIOOe'r
+ '• MtO
A/I /z&fff*
Gfiuseat\
ft, I-Iff CfM
M^
, Jii «**. '/-Z,
- ^ j—-^ uoo'F
go *• 77 »»» W,
H/wen
Z'fft. iu> «i
y,|^r,(
.'ttfzr..*^/ ^ / '^^ >
Westvaco
CHARLESTON RESEARCH CENTER
I*. O. «0» M07 NORTH CHARLESTON. S C. 2t«>
rLeitn
omi, it /ICHKD o»n
---
car rararTS-B-iio-i^ ----- vr -------
••'* \: 4iTn\ I DWG No. 2507
17-B
-------
has 8 inches. A photograph of the sorber is shown in
Figure 4. Assuming a carbon density of 38 Ibs./ft.-^, the
weight of carbon in the vessel is about 113 Ibs. The down-
comers are 3" dia. and extend to within 1-1/2 inches of the
distributor plates below. The gas distributor plates have
8% open area with 0.125" dia. holes drilled on 0.42" tri-
angular centers.
_\
Carbon is fed to the sorber at a rate of about 30 Ibs./hr.
A gravimeter feeder [CFC-1] feeds the carbon from an
inventory hopper [RV-90] into a bucket elevator [CV-3],
which raises the carbon to the necessary elevation and
empties it onto the top stage of the sorber. Carbon flows
by gravity downward through the column and passes out of
the unit into a seal leg between the sorber and the 6" dia.
carbon preconditioner [RV-300]. The carbon level in the
seal leg is controlled automatically by a level probe type
controller [LIC-301] which actuates a vibrating feeder
[CV-5] to feed carbon into the 6" unit.
The 22,000 acfh flue gas slipstream passes through a booster
fan [E-l] prior to entering the sorber. Flue gas flow rate
is measured by the pressure drop across an orifice. The
flow rate is controlled by adjusting a trunnion valve
located at the entrance to the sorber. The trunnion valve
has a pneumatic positioner which is operated mannually from
the panelboard. The sorber off-gas passes through a cyclone
dust collector [C-l] followed by a bag filter [F-l], and
then it is ejected to the atmosphere through an exhaust
blower [E-2]. A small negative pressure, about -0.5" W.C.,
is maintained at the top of the reactor to facilitate carbon
feed into the unit from the bucket elevator. The pressure
is controlled manually by adjusting a slide valve in the
off-gas duct.
Flue gas enters the sorber at 300°F with a total sulfur
oxides concentration of about 2,000 ppm, of which 50 ppm
is 803 and the remainder is S0£. The bottom bed of the
reactor, operating at 300°F, is used to remove the 863 in
order to avoid corrosion problems. The temperature at
Stage 2 then is lowered to 175°F to increase the rate of
SC-2 removal. Cooling is accomplished by injection of water
into the fluidized bed at a rate of 10 to 50 #/hr.
The rate of water injection into Stage 2 is automatically
adjusted to control the temperature on Stage 2. A tempera-
ture controller [TIC-102], linked to a pneumatic flow con-
trol valve through an electro-pneumatic transducer, regu-
lates the water flow rate. Temperatures in the upper
stages of the reactor are allowed to seek their own level,
which is determined by the balance between heat losses and
heat generated by the exothermic adsorption reaction.
These temperatures typically fall in the 180-190°F range.
18
-------
Figure 4. Continuous 18" dia., 5 stage S02 and 863
adsorber operating on flue gas from a 50 mw
oil fired boiler
19
-------
Carbon Conditioner [RV-3QO] -
The carbon conditioner [RV-300] is a 6" dia. x 2 ft. high
single stage fluid bed unit which is used to adjust the
temperature and moisture content of the carbon as it passes
from the S02 sorber into the sulfur generator. The unit is
operated at a carbon bed temperature of 320°F, with pure
steam or an air-steam mixture as the fluidizing gas. With
pure steam, a carbon moisture loading of about 0.10 Ib.
H20/lb. carbon is obtained. The steam flow rate is
measured by a rotameter. A temperature controller [TIC-302]
controls an electrical heater to maintain the desired inlet
gas temperature.
Carbon flow into the 6" dia. conditioner is controlled by
the level probe controller [LIC-301] which maintains the
proper carbon level in the seal leg between the S(>2 sorber
and the conditioner. Carbon passes out of the unit by
gravity flow into a seal leg between the conditioner and
the 8" dia. moving bed sulfur generator [RV-200]. There
is no direct control of the carbon flow out of the
conditioner. Instead, the carbon level in the seal leg is
maintained by a system which controls the carbon flow out
of RV-200. This system is described in the following section
on the sulfur generator.
Sulfur Generator [RV-200] -
The sulfur generator [RV-200] is a moving bed reactor, 8"
dia. x 10 ft. high, with an actual carbon bed depth of 6
feet. Carbon enters the unit from an upper seal leg between
it and the carbon conditioner [RV-300] and flows downward by
gravity in essentially a plug flow distribution. A vibrat-
ing conveyor [CV-2] feeds the carbon out of the unit into a
bucket elevator [CV-1] which transfers it to the next
reactor. A pneumatic controller [LIC-201] controls the
vibrating feeder so that a constant carbon level is main-
tained in the upper seal between RV-200 and RV-300. The
input signal to LIC-201 is simply the AP across the purged
seal leg, which is proportional to the height of carbon in
the leg.
Reactant gas containing about 30% H2S enters at the bottom
of the vessel and passes upward through the bed, counter-
current to the carbon flow. Total gas flow rate is 185-
260 scfh. The source of the reactant gas mixture is either
cylinder gases or else the recycled off-gas stream from the
4" dia. S stripper/H2S generator [RV-100]. When the 4" unit
off-gas is used, the pilot plant is said to be operating
with "process I^S". Cylinder gas rates are controlled at
the panel board with rotameters. An electrical heater and
20
-------
temperature controller maintain the desired inlet gas
temperature of about 300°F when cylinder gases are used.
When the pilot plant is operating with process H2S, the gas
is hot already and requires no further heating.
The off-gas from the sulfur generator is exhausted to the
boiler house stack on the negative pressure side of an
ID fan. A valve in the off-gas line permits adjustment
of the pressure inside the sulfur generator at 1 - 5 inches
W.C. at the top of the unit to facilitate gas sampling for
chromatographic analysis.
The temperature of the carbon bed typically ranges from
260°F at the bottom to 300°F at the top. Ideally the
temperature would be near 300°F throughout, but as the
reaction approaches completion in the lower section of
the bed, less heat is generated by the reaction and the
temperature drops off. Temperature is a function of 1)
inlet carbon temperature and moisture level, 2) reactant
feed rates, and 3) heat input through the walls of the
vessel. Three electrical heating mantles, individually
controlled with power-stats, are used to regulate heat
input through the walls. Heat input by this means is
limited by the S(>2 evolution problem at temperatures over
300°F, so that the wall heaters are operated to maintain
a wall temperature only slightly over 300°F. Maintenance
of proper bed temperature is strongly dependent on the
uninterrupted feed of reactants to the unit. Bed tempera-
tures drop rapidly if the feed of either reactant is
interrupted.
Sulfur Stripper/H2S Generator -
The sulfur stripper/H2S generator [RV-1001 is a 4" dia. x
24 ft. high vessel with 8 fluidized beds of carbon, each bed
having an expanded depth of 5 inches. Electrical heaters
around the unit maintain the temperature at 1000-1200°F.
The heaters are controlled automatically by 5 temperature
controllers, with one controller for every two stages and
the fifth controller for the inlet plenum heater. A sixth
controller handles the inlet gas heater.
The unit operates at a linear gas velocity of 2 - 3 ft./sec.
with a total gas flow rate of 200-300 cfh at 70°F.
Hydrogen and nitrogen are supplied from cylinders and
metered through rotameters. Flow rates are set manually.
The inlet hydrogen concentration was typically 38-4870
(shown incorrectly as 27-30% in Figure 3). The inlet gas
is preheated before entering the reactor. The reactor
off-gas passes through a cyclone dust separator [C-2]
before passing on into the sulfur condenser system, which
is described in the next section.
21
-------
The carbon feed for RV-100 is the sulfur loaded carbon
product from RV-200. A bucket elevator [CV-1] raises the
carbon above RV-100 so that it can feed by gravity flow
into the unit,. The carbon flows downward through ,the
reactor and passes into a seal leg connecting RV-100 to
the carbon cooler [HX-2]. A level probe controller
[LIC-102] maintains the proper carbon level in the purged
seal leg by controlling a vibrating conveyor [CV-4] which
feeds carbon from HX-2 back into the inventory hopper
[RV-90].
There are two carbon feed systems for the sulfur stripper/
H2S generator. In one, the carbon passes from the bucket
elevator into the upper seal leg and then is fed into the
reactor through a valve controlled by a pneumatic control-
ler [LIC-101] which maintains a constant carbon level in
the seal leg. An 8" section of the seal leg is a carbon
preheater [HX-1] with electrical heating and a temperature
controller. The other system consists of a pair of ball
valves controlled by a set of electronic timers so that
the valves open and close alternately, with one valve
always closed to maintain a gas seal. This system
bypasses the carbon preheater and feeds the carbon
directly from the bucket elevator into the reactor. The
two feed systems provide a number of options in the opera-
tion of"the reactor. They can be used simultaneously or
separately, and carbon can be fed to various stages in
the reactor. During integral operation, the double ball
valve arrangement became the primary system and the feed
was placed on the fifth stage from the bottom during most
of the run.
Sulfur Condensing System [RN-400] -
The sulfur condensing system [RN-400] is shown in detail
in Drawing 2537 in Appendix B-3. The sulfur condenser is
a combination gas cooler and scrubber in which the hot
sulfur-laden off-gas from the HoS generator/S stripper is
quenched from 1000 to 300°F, and the sulfur vapor is
removed by scrubbing with liquid sulfur. The condenser
is a jacketed baffle tray column, 3" I.D. x 4' long, with
overlapping baffles (1/2" overlap) spaced 3/4" apart and
occupying a 2' long section of the column. Regenerator
off-gas enters below the baffle section and passes upward
through the baffles, counter-current to the flow of recir-
culating liquid sulfur. The gas exits near the top and
is passed through a 6" thick wire mesh mist eliminator to
remove sulfur mist. The sulfur-free off-gas, containing
about 30% H2S, then goes to the sulfur generator [RV-200].
22
-------
Liquid sulfur is pumped from the collector pot [C-402]
beneath the condenser and introduced above the baffle
section. The sulfur flows downward through the baffles by
gravity flow and collects in C-402. Recovered sulfur
condensed from the gas spills over into a collection pot
[C-403] so that a constant inventory is maintained in C-402
Cooling is provided by water at 2506F parsed through" the
condenser jacket under 40 psig at a flow rate of 1.5-2.2
gpm. The cooling water is pumped through a heat exchanger
to heat it to the desired temperature.
Pilot Plant Instrumentation -
Pilot plant instrumentation falls into three categories:
control, general, and analytical instrumentation. All of
the instrumentation is indicated on Figure 3, or on
Drawing 2528A in Appendix B-3. Each category of instru-
mentation is discussed below.
Control Instrumentation -
Instrumentation for control of the pilot plant falls into
three main categories which are temperature control, carbon
flow control, and gas flow control. Time proportional
temperature controllers are used with the wall heaters and
gas heaters on the H2S generator/S stripper and the 6"
carbon preconditioner [RV-300]t and with the gas heater
for the sulfur generator. The wall heaters for the sulfur
generator are controlled manually with powerstats.
Temperature control instrumentation in the S0£ sorber con-
sists of a time proportional temperature controller linked
to an electro-pneumatic transducer control valve, which in
turn regulates the water spray rate.
Carbon flow instrumentation consists of a gravimetric
feeder, two sets of electronic level probes which operate
vibrating feeders, a set of electronic timers which operate
a double ball valve feeder, and two pneumatic controllers.
One of these operates a carbon flow metering valve, and the
other controls a vibrating feeder, through an improvised
mechanical linkage of pneumatic and electrical signals.
Gas flow rates are set manually by means of rotameters and
metering valves, except in the S02 sorber where an orifice
is used to measure the flow rate, and a piston-actuated
trunnion valve equipped with a positioner is used to
control the flow rate by means of a manual loading station.
The rate of liquid sulfur recirculation in the sulfur -
condenser system is controlled by manual adjustment of a
23
-------
variable speed SCR drive on the positive displacement
sulfur pump. Temperature in the condenser is controlled
by varying either the cooling water flow rate or its
temperature. Flow rate is set manually with a valve, and
temperature is controlled by adjustment of the steam pres-
sure on the heat exchanger used to heat the water to the
desired inlet temperature.
General Instrumentation -
The primary functions of non-control type instrumentation
are to provide temperature, pressure, and flow rate data.
Three instruments are used for the temperature data,
including an Acromag 10 point digital indicator, a
Honeywell 24-point strip chart recorder, and a Doric 40-
point data logger. The temperatures considered most
indicative of pilot plant operation and most important
to follow appear on the Acromag digital indicator. The
same ten temperatures and a few others appear on the
Honeywell strip-chart recorder, which provides a graph
showing variation with time. The same ten and thirty
more, representing all of the important process tempera-
tures, are recorded by the Doric-data logger in numerical
form, either automatically at 30 or 60 minute intervals
or else manually whenever a printed record is desired.
The indicated, logged and recorded temperatures are
depicted by TI, TL and TR, respectively, on Drawing 2528A
in Appendix B-3.
Pressure measurements are useful in diagnosing the operating
condition of the pilot plant. Differential pressure gauges
were used to measure pressure drop across each unit and
also across each stage within the fluid bed units. This
provided information on whether or not carbon flow is
satisfactory and reveals any buildup in the system. Pressure
drop across an orifice is used to measure gas flow rates
at four places in the system.
Analytical Instrumentation -
Analytical instrumentation provides gas composition data
for the inlet and outlet gas streams of the S02 sorber
and regenerator units. The S02 sorber gas streams are
analyzed for S02 concentration using an EnviroMetrics,
Inc., Series NS-200 S02 analyzer from which output is
recorded on a strip chart recorder. Gas streams in the
regenerator units are analyzed by a Bendix process
chromatograph for seven components: H2S, H20, S02, CO, H2,
C02 and N2- The chromatograph can analyze five samples
per hour. Peak heights for each component are read from
recorder, modified for bar chart display.
24
-------
Carbon Dust Collection -
Carbon dust is removed from the off-gases of the H2S
generator/sulfur stripper and the S02 sorber. Equipment
also was installed for dust removal from the 6" carbon
preconditioner off-gas, but it was necessary to bypass
this equipment due to operational difficulties. The S02
sorber is responsible for the major portion of the carbon
dust generated in the process. The sorber off-gas first
passes through a cyclone and then through a bag filter
to remove the finer particles. The combined dust removal
efficiency is greater than 997o by weight.
Dust is removed from the off-gas of the H£S generator/
sulfur stripper by a cyclone followed by a specially
designed filter containing a 4" depth of Fiberfrax Long
Staple Fine ceramic fiber material. The Fiberfrax filter
was bypassed, however, after it was found that the small
amount of carbon dust did not create problems in sulfur
condenser operation.
Sulfur Collection -
Sulfur condensed from the regenerator off-gas is recovered
as a liquid by draining the collection pot into a 6-liter
metal beaker at 8-hour intervals or more frequently if
desired. The sulfur is then weighed and transferred to a
larger container for storage.
The recovered sulfur is analyzed for purity either by a
combustion analysis in which the carbon impurity is measured
as C02, or by a sulfur vaporization analysis in which the
impurity is measured after vaporizing away the sulfur. Ash
content is also determined by a standard ash analysis.
25
-------
4.2 INTEGRAL PILOT PLANT RESULTS
The broad goals of the Westvaco Process are as follows:
1) High S02 removal
2) Minimum hydrogen use
3) Maximum S02 recovery as elemental sulfur
4) Minimum carbon burn-off.
With respect to these broad goals for process operation,
an original set of goals were defined based on studies of
each of the process steps separately. Another set of
goal bases which are directed to these broad goals and the
main process streams shown in Figure 5 are:
1) S02 Sorber
a. S02 removal
2) Sulfur Generator
a. Sulfur compounds evolved/S02 removed
3) H2S Generator/S Stripper
a. H2 utilized/S02 removed
b. Carbon burn-off/S02 removed
4) Sulfur Condenser
a. Sulfur recovery/S02 sorbed.
The basic difference compared to the original set of goals
is that in the latter case only streams that cross the
integral process boundary are chosen and in the former case
some of the goals were based on internal recycle streams.
The integral run, which totalled approximately 20 days,
was made with pilot equipment installed at an oil fired
boiler. A mechanical problem, which is described in
Appendix B-2, was developed after 11 days or about 18
carbon cycles; it was corrected and did not reoccur
during subsequent integral operation. The second part
of the integral run was voluntarily stopped after an addi-
tional 8 days or about 11 carbon cycles.
To allow continued evaluation of the carbon performance
for additional carbon cycles, the same carbon was used in
the entire run. In total, the carbon was circulated
integrally for approximately 29 cycles to allow evalu-
ation of carbon attrition for this particular carbon. Of
these 29 cycles, the carbon was exposed to flue gas/regenera-
tion gas conditions for 21 cycles of which 14 were with
process H2S. Since the source of the H2S, i.e. process
or cylinder, had no apparent effect on S02 sorption, 21
cycles were taken as the process exposure time in analyzing
the S02 removal performance of the carbon.
26
-------
Figure 5. Schematic of integral Westvaco S02
removal pilot plant
FLUE GAS-
S02
HYDROGEN-
S02
SORBER
S02 SORBER
OUTLET GAS
SULFUR
GENERATOR
SULFUR GENERATOR
OUTLET GAS
PROCESS H2S
SULFUR
CONDENSER
H2S
GEN./S
feTRIPPER
ELEMENTAL SULFUR
PRODUCT
T
REGENERATED
RECYCLE
CARBON
I
27
-------
The development of the Westvaco S02 Recovery Process
encompassed bench scale studies, pre-integral pilot scale
testing, and integral pilot plant operation. Many experi-
ments were performed in each of the phases of process
development. For clarity these results have been presented
here in two sections — integral and pre-integral work. In
Section 5 both the bench scale studies and the individual
pilot unit studies are discussed. Since the principal goal
of this program was to design and operate an integral pilot
plant of the process, the integral work is presented first
in the remainder of Section 4.
4.2.1 Overall Integral Results
Operating Conditions -
The general intent of the integral run was to maintain
constant conditions over an extended period in which the
granular carbon would be exposed to repeated sorption and
regeneration with H2S produced in the process. In earlier
studies, carbon had been exposed to flue gas during sorp-
tion, but in the sulfur generator step only cylinder H2S
had been used. An arbitrary time of 30 cycles was initi-
ally selected for the integral run, during which time any
trend would be detectable and indicative of longer term
effects. In addition, a 90% S02 removal efficiency was
to be maintained with a sulfuric acid loading on the carbon
of at least 18 Ibs. acid/100 Ibs. carbon. Other limits on
sulfur generation (acid conversion), sulfur recovery and
operating conditions were selected based on pre-integral
pilot and bench scale test results.
The operating conditions in Table 1 were selected to meet
the target goal of 90% S02 removal with an acid loading
of at least 18 Ibs. acid/100 Ibs. carbon. The inlet flue
gas was controlled at the rate of 22,000 scfh to the S02
sorber. The S02 content was adjusted as necessary to
maintain 1900 to 2000 ppm, and the inlet temperature was
maintained at 300°F on the first stage for S03 removal,
with the next stage cooled by water spray to 175°F for S02
removal. The temperatures of the remaining sorber stages
were not controlled and were allowed to rise due to heat
of reaction during S02 removal. Carbon bed depths of 3.5
inches on the bottom stage and 6 inches on each of the
remaining stages were set for a total of 56 inches in an
expanded state.
The recycle rate of carbon was set to achieve the desired
acid loading based on the previous relationships developed
between the operating parameters in the sorber.
28
-------
Table 1. RANGE OF OPERATING CONDITIONS FOR
INTEGRAL PILOT PLANT RUN
INLET FLUE GAS
Gas Rate:
Temperature:
Composition,
S03:
NO:
02:
H20:
Inert Gas (C02,4 N2):
CONDITION
22,000 SCFH
300°F
'1900-2000 PPM
50 PPM
150 PPM
4.5 Vol.
13 Vol.
Balance
S02 SORBER
Temperature, Stage 1 (Bottom):
Stage 2 (H20 Spray);
Carbon Bed Depth (Expanded):
Fluidizing Velocity:
Space Velocity:
300°F
175°F
Stage 1 - 8"
Stages 2 to 5 - 12"
3.5 Ft./Sec. (<> 300°F
3400 SCF Gas/CF Carbon-Hr.
REGENERATORS
Acid Converter
Temperature:
H2S Inlet Rate:
•
Space Velocity:
S Stripper/H2S Generator
Temperature:
H2 Inlet Rate:
Space Velocity, Stripper:
H2S Generator:
Fluidizing Gas Velocity:
290°F (Avg.)
Output from H2S Gen. Range =
2.5-2.9 mol/moles
100 SCF Gas/CF Carbon-Hr.
1000-1100°F
3.4, 3.9, 4.3 moles H2/mole
S02 Sorbed
2000-3100 SCF Gas/CF Carbon-Hr.
6200-9300 SCF Gas/CF Carbon-Hr.
1.8-2.7 Ft./Sec. @ 1000°F
CARBON RECYCLE RATE:
29-30 Lbs. C/Hr.
29
-------
Temperature and space velocity conditions for the sulfur
generator and sulfur stripper/H2S generator were based
on earlier process unit test results. The hydrogen flow
to the generator was varied above the stoichiometric
requirement of 3 moles/mole of S02 sorbed on the carbon.
The amount of H2S entering the sulfur generator was pre-
determined by the hydrogen input, with no further attempt
to control this rate.
The pilot plant was started and operated for 315 hours
under these conditions using previously described
procedures.
Summary of Results -
The overall results of the integral run were as follows:
No. of Process Cycles: 21
No. of Operating Hours: 315
Outlet S02: 120 ppm
Avg. S02 Removal: 94%
Product Sulfur: 99.7% Pure
Carbon Attrition: 0.26 Ib./hr.
Discussion of Major Response Parameters -
The main factors observed in the pilot operation were the
degree of S02 removal during cycling of the carbon, the
requirements for hydrogen for H2S generation, and the con-
version of S02 to elemental sulfur. Other factors were
the loss of carbon by chemical and/or mechanical means
and the disposition of any excess hydrogen.
S02 Removal Efficiencies -
The removal of S02 during the 300 hour period of the
integral run is given in Figure 6. The flue gas, contain-
ing 1900 to 2000 ppm S02, was desulfurized to well above
90% with a maximum of 97%, corresponding to 60 ppm remain-
ing in the effluent gas. By inspection of the plot,
there does not appear to be any trend toward reduction in
S02 removal efficiency. This has also been substantiated
by laboratory analysis of the recycled carbon. During
the integral run, the carbon was cycled through the
system some 21 times, based on a calculated carbon resi-
dence time of 15 hours in the integral system. The amount
of S02 picked up by the carbon in terms of sulfuric acid
averaged 24 Ibs./lOO Ibs. carbon, substantially above
target.
30
-------
Figure 6. S02 removal efficiency during
integral pilot tests
100 .-
95 ..
g 90
.-GOAL - 90T
S
* J_ INLET S02 ' 1900-2000 PPM
RUN TIME, HOURS
30 60 90 120 150 180 210 240 270 300 330 360
-i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 4 6 8 10 12 14 16 18 20 ?? 24
NUMBER Of CARBON CYCLES
No corrosion or dew point problems were noted in operation
at 175°F, since the 30 - 50 ppm gaseous 863 in the flue gas
is adsorbed on the carbon. This removal of 803 with
carbon was demonstrated in previous studies.
The 150 ppm NO in the flue gas is not directly affected by
the carbon and as such remains in the flue gas. The
initial effect of the NO is to suppress the 802 pickup.
This effect appears up to a NO concentration of about 150
ppm. This aspect is covered more fully in Section 5.21.
Sulfur Product -
It is important that the sulfur by-product from the regene-
ration system be a salable commodity. The elemental
sulfur recovered from the pilot tests had characteristics
as shown in Table 2.
Table 2. PROPERTIES OF SULFUR PRODUCT
Sulfur 99.7%
Ash 380 ppm
Carbon 2500 ppm
Acidity . 2 ppm
Chloride <2 ppm
31
-------
These properties, measured for Westvaco by a sulfur
producer, classify the sulfur collected as a commercial
grade.
The small amounts of carbon in the sulfur, a result of fines
carryover from the regenerator, gave the sulfur a greenish
cast. It was demonstrated that these fines could be readily
filtered out to give a bright sulfur product of 99.9% purity.
Carbon Chemical Consumption -
In passing through the regeneration sequence, the activated
carbon is exposed to temperatures progressively increasing
from 300°F to 1000°F. To prevent chemical consumption at
1000°F, the sulfuric acid is reduced to elemental sulfur at
300°F. In addition to production of elemental sulfur, a
goal of the Westvaco Process is to minimize the amount of
carbon reaction to produce C02. Measurements were made on
the C02 content of the regeneration off-gases to estimate
the amount of chemical consumption or "burn-off" that could
be occurring by this means.
As shown in Figure 7, the carbon burn-off, calculated
from CC-2 evolution reached a stable value of about 10 - 12
Ibs. C/ton of S02 sorbed from the flue gas. As shown by
the dotted line this compares to a "burn-off" of 187
Ibs./ton if the carbon were consumed by reacting with all
the sorbed acid under thermal regeneration conditions. This
Figure 7. Carbon burn-off during integral
pilot tests
190 ••
8 '85
1 180
MAXIMUM THEORETICAL BURN-OFF WITH THERMAL REGENERATION
& 25.
20-
15.
10-
0.
°\° \
N® Hi
X«2_ o o0 oo o ^
uw O Q0
»
1 II 1 1 1 H — l-4—l — 1— 1 — 1 1 I 1 I
*? ° °n
v o Q O t>
1 — ! — U- 1 — 1 — 1 — I— i
10
14
16
18
20
22
NUMBER OF CARBON CYCLES
24
32
-------
reduction in burn-off of about 95% shows that the original
objectives were achieved. By inspection of the data there
was no apparent effect on burn-off when the hydrogen input
was varied in the range of 3.9-4.6 moles/mole acid dis-
cussed earlier.
Complete conversion of the acid to sulfur was not required
to prevent burn-off. Earlier experiments on the bench
scale verified this fact, in that the addition of sulfur
by various means considerably reduced the chemical consump-
tion of the activated carbon during regeneration.
As discussed earlier there was some thermal decomposition
in the sulfur generator which would probably explain the
small amount of burn-off measured.
If all of the C02 measured is a result of burn-off, the low
values measured here would correspond to a complete replace-
ment of the inventory only about once every two years.
Carbon Mechanical Consumption -
The attrition rate experienced in the integral pilot opera-
tion, Figure 8, showed an initial decrease, probably due
to a rounding off of rough edges, and then a stabilization
at a rate of 0.26 Ibs./hr.
Figure 8. Activated carbon attrition rate
during integral pilot tests
8 12 16 20
NUMBER Of CARBON CYCLES
28
32
33
-------
The data indicated nearly all of this attrition occurred
in the fluidized beds of the S02 sorber. Additional work
has shown that the combination of larger particle sizes and
carbons with improved hardness will reduce the attrition
rates to about 10% of the values measured here. These
improvements should be incorporated in future scale-up work.
It is significant that there is no apparent increase in
the attrition rate as the carbon was recycled thermally
and chemically, as has been observed with other solid
adsorbents. The nature of S02 recovery with carbon,
providing a surface for catalysis and adsorption
rather than actually chemically participating in the reac-
tions as is done with metal oxides, probably results in
the maintenance of structural integrity and strength of
the carbon.
Regeneration Results -
In the integral runs, the intent was to demonstrate that
the carbon could be repeatedly regenerated for reuse and
to maximize the amount of elemental sulfur produced within
the limitations of the present pilot equipment. The only
deliberate change in regeneration conditions was in the
hydrogen input. Other conditions were pre-set based on
prior work.
As discussed in the preceding section, the activated carbon
retained its adsorptive capabilities throughout the run,
attesting to the suitability of regeneration under all
hydrogen input conditions.
•
Three levels of hydrogen input were evaluated during the
integral runs. Analyses of all the process streams were
used in preparing the material balance presented in
Table 3.
34
-------
Table 3. EFFECT OF HYDROGEN INPUT ON
BY-PRODUCT RECOVERY
Condition
A
TOTAL HYDROGEN INPUT 4.6
(moles/mole available acid)
HYDROGEN USAGE
(moles/mole available acid)
1. Formation of by-product 0 Qn 0 00 , _
sulfur z'90 2'88 3-°
2. Reaction with by-product • or ft ." .
sulfur to form H2S °'Zj °'n °
3. Reaction with chemisorbed n Qn , n
oxygen to form H20 u-au '•"_ u-a
TOTAL MEASURED HZ OUTPUT 4.05 3.99 3.90
Condition C essentially represents the process hydrogen
input necessary for conversion of the available acid to
elemental sulfur product. The hydrogen input above the
stoichiometric ratio of 3 reacted with chemisorbed oxygen
to form water and H2S did not appear in the sulfur
generator vent gas. The reaction of a part of the inlet
hydrogen with chemisorbed oxygen had been observed in
previous work and is apparently instrumental in retaining
the activated carbon's activity upon cycling. As the
hydrogen was increased from 4.3 to 4.6 a part of the
product sulfur reacted to form H2S which appeared in the
vent gas, while formation of water essentially remained
constant. The difference between the measured hydrogen
input and output amounts to about 1270 and could be the
result of analysis error or, possibly, chemisorption of
these small amounts of hydrogen on the carbon itself.
If the hydrogen ratio were lowered below that of Condition
C, S02 formation would be expected at the expense of part
of the sulfur product. This would be the desired
direction if the process is slightly out of balance since
S02 can be readily recycled to the sorber.
During these integral tests and with prior work, there was
limited temperature control in the moving bed sulfur
generator. As a result of higher than desired temperatures,
a part of the sorbed acid decomposed to S02 in the upper
35
-------
part of the unit and was not readily available for
conversion to sulfur. Thus a maximum of 857o conversion of
the sorbed acid to sulfur was obtained with this equipment.
Prior testing had shown that with proper temperature con-
trol essentially 100% conversion to sulfur is possible and
this should be readily attainable in larger equipment where
fluid beds will be used.
The gas residence time in the regenerators is only about 15
seconds; therefore, response of the system to hydrogen
input is very rapid. Thus, control should be readily
achieved by monitoring regeneration off-gases and adjust-
ing the hydrogen input accordingly.
4.2.2 Detailed Integral Results
Detailed Operating Conditions and Results -
A detailed summary of the integral run conditions and
results for both IR-1 and IR-2 is presented in Table 4.
As can be seen, five steady state operating periods
occurred during IR-1 and three during IR-2. All these
periods represent changes in the regeneration conditions
only. The S02 sorber conditions were maintained constant
throughout. In periods 4, 5, 7 and 8, the ratio of
hydrogen input to S02 sorbed (moles H2/mole S02 sorbed)
was deliberately changed to optimize the process for
hydrogen consumption. Although the ratio of H2/S02
sorbed is the same for Periods 7 and 8, Period 8 simu-
lated the use of shifted producer gas for sulfur stripper/
H2S generator feed to evaluate its effect on carbon
regeneration.
To evaluate the operating performance of the integral
pilot plant, five performance variables--S02 removal, H2$
utilization, acid conversion to sulfur, H2 utilization,
H2 utilization to H2S-- were initially chosen as indicators.
In the analysis of Runs 6, 7 and 8 where some of the acid
was converted to sulfur in the sulfur stripper/H2S genera-
tor, it became apparent that several other process responses
might be more suitable for assessing integral operation.
Since the overall process response time to changes in
process variables was not known, the attainment of steady
state was taken to be when no variations about a statisti-
cal average were measured in the process streams. The
data indicated that approximately 30 hours were required
while using process H2S before the total integral pilot
plant reaches steady state after a change. In view of
this, only Periods 3, 4 and 7 in Table 4 represent steady
state conditions. Period 5 was very close to steady state,
36
-------
Table 4. INTEGRATED OPERATING CONDITIONS AND RESULTS
STEADY STAIR 1'V.RIOD
CYL1XEW.R OR I'RiXESS H'S
TIME PERIOD (Hrs.)
CARBON RECYCLE RATE (Lb» . /llr . ) •
SOJ SORSER*
Condi t ions
Cas Rate. CFH
If rap. (Avg.). 0F
Outlet Gat Tcr.p.. °F
Inlet SO' Cone., ppm (Dry Bos Is)
Rate. CFH
Temp. (Range), °F
Results
Outlet S02 Cone., ppn (Dry Basil)
Rate. CFH
Outlet Acid Loading. Ib./lb. C
Pressure Drop, in. 1120
S02 Removal, "i Inlet
Carbon Attrition Rate, Ib./hr.
CARBON CONDITIONER
Conditions
**Steam Cone., vol. *
Carbon Bed Temp. (Avg.), °F
Results
Carbon Hoist. Load.. Ib./lb. C
SULFUR GENERATOR
Conditions
Temperature C Bed (Range) , °F
(Avg.). op
*—1 - *. M.. /»T"»*
*tt*e*. »*£ , W* (1
Inlet C02, CFH
Inlet Acid Load., Ib./lb. C
Results
Outlet Acid Load., Ib./lb. C
Outlet S Load.. Ib./lb. C
Outlet H2S, CFH
Outlet SO;. CFH
OutioL C02. c~!
Outlet HiO. CFH
H2S Utilization. %
Acid Conversion to S, 7.
S Evol., Ratio to 502 Sorbed
~ , tas H2S)
(as S02)
H2S GENERATOR/SULFUR STRIPPER
Conditions
Temp- C &ed (Range), °F
(AVR.). OF
**Inlet N2, CFH
"Inlet H2, CFH
**Inlct C02. CFH
Inlet S Load. . f S as S/# C
* S as H2SO&/* C
Results
Outlet S Load., t/f C
Outlet H2. CFH
Outlet H2S, CFH
Outlet S02. CFH
Outlet HjO. CFH
Outlet COj. CHI
H2 Utll., 7. of Inlet
Coin, to H;iS. 7.
Ratio to S<>2 Sorbed
SULFUR COMDr.llSER
Outlet Can Tcrpcrature (Avp, .), °F
S Kecov. , Ibn./lir.
1 of Stripped
T, of SO? Sorted
(by S Kt-cov.)
rhv S Kvrj !•/..
II 4
8
I'l OCOSS
H
29
,
15,400
185
184
?,030
;9.o
16S-197
80
1.1
0.23
26.3
96
0.28
100
324
0.09
262-306
287
•-Of
*.R
4o'g
0.23
0.04
0.21
1.3
4.4
40.9
126
98
62
0.05
0.16
500-1-350
1027-
66
95
40
0.21
0.04
0.047
0
58
0
45
40.9
100
61
•J.4
2/0
'/.O
'.J5
»li
M)
H 1
*A11 sorber conditions steady for whole run period of about 8 days or 13 cycles.
**0nly condition changes during run.
***Gas sampling error found and rectified.
****At end of period H2 leak was found.
37
-------
and Period 8 was close enough to steady state to yield
data indicative of overall process operation at the
corresponding process conditions.
The data in Table 5 indicate a number of results obtained
during the integral run. The S02 removal goal of 90% was
exceeded with 93 to 97% obtained for all the carbon cycles
Table 5. PROCESS OPERATING PERFORMANCE
Performance of Process Unit
S0£ Sorber
S02 Removal , % of Input
Sulfur Generator
H2S Utilization, % of input
H2S Evolution, % of SOz removed
in S02 sorber
S02 Evolution, % of S0£ removed
in S02 sorber
H2S04 Conversion to Sulfur, %
H2S Generator/S Stripper
Ratio H2/S02
H2 Utilization, % of input
S02 Evolution, % of S02 removed
in S02 sorber
Carbon Burn-off, Ibs. C/T S02
sorbed
Sulfur Condenser
Sulfur Recovery, % -of S02
removed in SOg sorber
Goal
90
95
99
90
IR-1 Run
4
94
92
21
7
87
4.3
100
0
10
67
5
94
96
11
10
90
3.9
100
0
10
77
IR-2 Run
7
95
100
0.3
14
70
3.4
100
0
13
84
8
95
98
5
16
60
3.4
100
0
11
83
*CTose approach to steady state in these regeneration periods. In
Task IV-A, Period 3 simulated producer gas used.
in both tasks. The H2S utilization was increased to 1007<,
by lowering the H£ to S02 sorbed ratio to 3.4, but as a
result, the sulfuric acid conversion to sulfur was decreased
in the sulfur generator. The overall sulfuric acid conver-
sion to elemental sulfur, however, increased as the desired
effect, as shown by the sulfur recovery increase. This
means that the effective acid conversion was completed in the
H2S generator/S stripper. The H2 utilization was 100%, and
38
-------
H2 utilization to H2S decreased which would result from the
increased acid conversion in the H2S generator/S stripper
or decreased recycle process H2S. The decreased regenerant
H2S gas recycle is viewed as a benefit, since the H2 utiliza-
tion is maximized. In summary then, the overall intended
goals were markedly improved in the IR-2-C run as reflected
by the increased flue gas S02 recovery as elemental sulfur.
Most of the additional S02 was evolved in the sulfur
generator outlet and could be recycled back to sorber to
effectively close the loop, with a resultant increase in
S02 recovery as elemental sulfur.
It is felt that S02 evolution is a thermal effect
which can be reduced by better temperature control in fluid
bed sulfur generators planned in future development. In
any case, to put this S02 evolution in perspective, the
sulfur generator off-gas contains S02, H20, C02 and N2, all
typical components of flue gas. The total gas flow rate
of this stream is about 250 to 300 cfh @ 70°F in this pilot
equipment. The'stream would be recycled back to the S02
sorber resulting in an increase of less than 2% in the gas
volume.
An unexpected results, mentioned above, that was found
during the integral runs was the distribution of the acid
conversion to elemental sulfur reaction between jthe
sulfur generator and sulfur stripper/H2S generator. With
the decrease in H2 input, the acid conversion to sulfur
decreased in the sulfur generator as shown in Table 5.
But, the overall acid conversion to sulfur, as indicated by
sulfur recovery, increased as a desired effect. This meant
that the effective acid conversion was completed in the
sulfur stripper/H2S generator. More importantly, this
acid conversion in the sulfur stripper/H2S generator was
accomplished without an attendant increase in carbon
burn-off as might be expected at the 1000°F reactor
temperature and no S02 was evolved indicating conversion of
the unconverted sulfuric acid to elemental sulfur. Also
accompanying this shift in acid conversion was a decrease
in H2 utilization to H2S while maintaining 100% H2
utilization. Such a decrease in the H2S recycle is con-
sidered beneficial since the H2S utilization is maximized.
A short run was made, Period 8, in which the effect of a
simulated shifted producer gas or reformer gas on regenera-
tion was evaluated. Although the run lasted only 8 hours
and "true steady state" results were not obtained, the
apparent general effects were:
1) H2S production in sulfur stripper/H2S
generator decreased
39
-------
2) Acid conversion to elemental sulfur in the
sulfur generator was further decreased.
3) No S02 evolution from sulfur stripper/H2S
generator
4) No effect on H£ reduction of flue gas S0£
to elemental sulfur.
5) Carbon burn-off constant.
Overall, the shifted producer gas had no deleterious
effects on process performance.
Activated Carbon Performance -
One of the main intents of the integral runs was to evalu-
ate the activated carbon performance over extended carbon
recycle conditions. The carbon was recycled for
about 29 cycles, of which 21 were under flue gas/regenera-
tion gas conditions. Some of the indicators of carbon
performance are S0£ activity, carbon attrition, carbon
burn-off, mean particle diameter, pore volume, surface area,
and ash content. Throughout the 29 cycles, there was no
indication of any adverse effects on the carbon performance.
S0£ Removal -
Some of the indicators of S02 activity of the carbon are
S02 removal efficiency, acid loading of the carbon, sulfur
loading of the regenerated carbon, and relative S02
activity as determined by an independent bench scale S02
sorption apparatus. The S02 removal efficiency, acid load-
ing of the sorber product and residual sulfur loading of
the regenerated carbon are shown in Figure 9 for the
IR-1 and IR-2 runs. For all 21 cycles,the SOo removal was
above 90% and actually ranged from 93 to 9770 for an inlet
S02 concentration of 1850 to 2050 ppm S02 to yield a carbon
loading of 22 to 26 Ibs. acid/100 Ibs. carbon from the
sorber. In IR-2 the S02 removal average was higher but
the inlet S02 concentration was slightly lower. The
residual sulfur loading of the regenerated carbon was about
0.035 Ib. sulfur/lb. carbon for IR-1 run and about 0.04 to
0.045 Ib. S/lb. C for the IR-2 run.
It would be expected that this slightly higher residual
sulfur loading might reduce the relative S02 activity. As
seen by Figure 10, in which the relative activity is given,
the activity was decreased slightly because of the slightly
higher residual loading, but was relatively constant for
the duration of IR-2 run. For the virgin precursor,the S02
40
-------
Figure 9. Activated carbon performance during Westvaco S02 recovery integral pilot runs
1
« 100-
i
1 „„.
ce 90
80
(
I 0.3-
g
O
o a n 9-
< __..f «•«.
t- -5T
LU O
1- CM
Z> 1C
o
5 3 0.1
ca
o
CM
co 0
1
O
z
| o.io-
CO <_>
5 ffl 0.05'
T ^ o-
CQ
a — J
LU
m
c
CO
CM
O r
CO v
IR-1 -»
AVERAGE INLET S02 CONC. • 2000 PPM
fv
START-UP .,„ • \ .J > v -J^
C RECYCLE V-* *r* **
NUMBER OF CARBON CYCLES
2 46 8 10 12 14 16
] 40 8*0 120 160 200 240
ELAPSED TIME,
y^^^^v^vj
•
.
NUMBER OF CARBON CYCLES
2 4 6 8 10 12 "14 16
3 40 80 120 160 200 240
ELAPSED TIME, F
>
NUMBER OF CARBON CYCLES
2 4 6 8 10 12 14 16
40 80 120 160 200 2'!Q
ELAPSED TIME, H
p»-IR-2
AVERAGE INLET S02 CONC. - 1900 PPM
V\^/V-^\
»^ RE-START.^. •yw> •
C RECYCLE - «,„,
•
18 20 22 24 - 26 28 30
280 320 360 400 440 480
k>URS
"^^ ?*v*^» / *y
18 20 22 24 26 28 30
280 320 360 400 440 480
OURS
1
1
18 20 22 24 26 28 30
280 320 350 400 440 480^
OURS
-------
Figure 10. S02 activity as a function of carbon cycle time as determined
by bench scale apparatus
N5
K
UJ
cc
J
*s_
£
•*
i-
E
£
CO
'
• AM
100-
t\f\
yj
80-
70-
60
A •
• e=-=^T.ri^
AVERAGE FOR FLUE GAS PERIOD
NUMBER OF
2 1 6 8 10 12
.1 t t I I 1
,
IR-l — ».|« — IR-2
|
•^-^-C/1^9
CARBON CYCLES ~
11 15 18 20
_! j . . , i , :
M P* ^* * ^ .
us FLUE GAS *•*
— ~~°~-<^_j: — • — "*^*^T^-* AVERAGE FOR FLUE
22 24 26 28 30 GAS PERIOD
.—; 1 ! — , 1 . — : > f
80
120 * 160
2CO 210 280
ELAPSED TIME/ HOURS
320
350
400
180
-------
number was 85 for comparative purposes. As shown in
Table 4, the average temperature of the H2S generator/
sulfur stripper was about 100 to 200°F lower than in the
IR-1 run even though the intention was to operate under
the same conditions as in IR-1. The reason was that the
temperature controllers were set at a slightly lower tempera-
ture to prevent possible temperature overshoots experienced
in the IR-1 run. It is expected, therefore, that the
residual loading might be increased or the relative
activity decrease, as was the case.
The switch from cylinder hydrogen to a simulated shifted
producer gas had no apparent effect on the carbon's per-
formance, as determined by the above measures of S02
activity.
Attrition -
The carbon attrition rate was measured as a function of
time, as shown in Figure. About 3070 of the carbon charge
for the IR-2 run was virgin carbon on the same original
batch of virgin carbon used in IR-1. The need for this
added virgin carbon was due to accidental carbon spillage
at the end of the IR-1 run and explains why the carbon
attrition rate was slightly higher at the start of the IR-2
run, than 0.27 Ib. C/hr. at the end of IR-1.
Problems in IR-2 with the S02 sorber dust filter precluded
use of that equipment, The rate of carbon collection was
relatively constant in the IR-1 run at 0.12 Ib. C/hr, so
that rate of collection is also taken for the IR-2 run.
This leads to a final attrition rate of about 0.25 Ib. C/
hr., compared to 0.3 Ib. C/hr. predicted by bench scale
measurements. This data should be sufficient to allow
estimates of the carbon attrition on economics of the SC-2
removal process.
Also a direct result of carbon attrition, the mean particle
diameter would be expected to decrease with no carbon
make-up. As shown in Figure 11, the mean particle diameter
did decrease from about 1.16 mm to about 1.0!) mm. This
decrease caused no apparent problems in process control or
other operations of the pilot plant. In an actual con-
tinuous operation with carbon make-up,the mean particle
diameter would be expected to stabilize.
43
-------
Figure 11. Carbon attrition, mean particle diameter, and ash content
as a function of carbon cycle time
IR-l
TOTAL AVERAGE CARBON
ATTRITION RATE
IR-2
TOTAL AVERAGE CARBON
ATTRITION RATE
8 10
NUMBER OF CARBON CYCLES
12 11 16 <:i8
10 80 120 160- 200 210 280
ELAPSED TIME, Houas
20 22 21 26
180
DUST COLLECTED WITH
S02 SORBER CYCLONE
DUST COLLECTED «:TH
S02 SORSER DUST
FILTER
NUMBER OF CARBON CYCLES I
J 0.9
c f
2.
) 10
f P
80
8
120
10 12
160
11 •
1
200
16
210
"i
280
20
1
22
320
21
360
26
100
28
1
30
110
180
ELAPSED TIME, HOURS
6-
5'
1-
3-
2
C
•
-^
2168
) 10 80 120
i
•
1
1
i
i
DUMBER OF CARBON CYCLES'
10 12 11 16 18- 20 22
i i i | i i i
160 200 210 280 320
»- TT tt
-«
21
1_ -
360
-, •-— '
26 28 30
i , .. -i— i
100 110 llo
ELAPSED TIME/ HCUFIS
-------
Ash Content -
The carbon is exposed to flue gas each carbon cycle during
which the carbon could pick up fly ash. Exposure for 21
cycles did not affect the ash content as shown by the
relatively constant value in Figure 11. The ash content
varied continuously between 4.4 to 5.2%, indicating no
significant fly ash pickup.
Carbon Burn-off -
Throughout the runs in both IR-1 and IR-2, the inlet and
outlet gas streams of the sulfur generator and H2S generator/
sulfur stripper are analyzed for H20, C02, CO, H2S, H2, N2
and S02 with a process gas chromatograph. Presumably the
difference between the C02 and CO content of the inlet and
outlet gas streams is a measure of the carbon burn-off in
these vessels. The detailed gas analyses are given in
Appendix A-12. No carbon monoxide was detected in the
off-gas of either vessel. Carbon dioxide, however, was
detected in both off-gas streams during process H2S gas
recycle as shown in Figure 12 for the IR-1 and IR-2 runs.
As seen from Figure 12, the C02 determined in the sulfur
generator was about equal to that determined from the
H2S generator/S stripper, which necessarily implies that
the burn-off occurs in the H2S generator/S stripper. The
carbon dioxide evolved in the H2S generator/S stripper
leveled off at about the same rate in the IR-2 run as the
IR-1 run. The C02 evolution rate of about 0.026 Ib. C as
C02/hr. (or about 10 Ibs. C/Ton S02) is apparently indica-
tive of what might be typically expected. This burn-off
rate agrees with that predicted from previous bench scale
results of about 4 to 10 Ibs. C/Ton S02 sorbed.
The maximum carbon burn-off that would be expected from
reaction with sorbed sulfuric acid, Equation (9), is
about 187 Ibs. C/Ton S02 sorbed.
H2S04 + 1/2 C »• 1/2 C02 + S02 + H£0 (9)
The decrease in burn-off from 187 to 10 Ibs. C/Ton S02
sorbed corresponds to a reduction of about 95%.
45
-------
Figure 12. Carbon dioxide evolution as a function of carbon
cycle time
=i
PsJ
S
V!
£ 1
_j
^
3
-j
o
UJ
CM
CD
f >
1
0.05-
0.05 •
0.04 '
0.03 •
0.02 •
0.01-
0
1
m-i -*l— iR-2
• I f
V . i \
^""v^ \ o A
^"Ssv»^ 1 vW *
^^S*~—— *t ** 1 ^^ * 0*
** ** *
1
NUMBER OF CARBON CYCLES'
2 4' 6 8 10 12 14 16 18 20 22 24 26 28 30
• t1ifttltlill.ll —
0 40 80 120 160 200 240 280 320 360 400 440 480
As DETERMINES FROM ^S
GENERATOR/S STRJPPES
As DETERMINED FROM
SULFUR GENERATOR
ELAPSED TIME, HOURS
-------
Pore Volume and Surface Area -
The effect of repeated cycling of the carbon under process
conditions on total pore volume and surface area was
measured. Total pore volume and surface area which were
determined for samples of regenerated carbon taken periodic-
ally during the two integral runs are shown in Figure 13.
Because of the scatter in the data, which is within
experimental error limits, the apparent slightly increasing
nature of both the pore volume and surface area plots may
not be considered statistically significant. Therefore,
it appears that both the carbon's total pore volume and
surface area were not appreciably changed with repeated
exposure to sorption/regeneration conditions.
Regeneration -
Sulfur Generator -
The major indicators of the performance of the moving bed
sulfur- generator are the acid conversion to sulfur and H2S
utilization. The results obtained for these responses in
this run are included in Figure 14. In comparing the two
runs, IR-1 and IR-2, the H2S utilization improved toward
100% utilization as the sulfur recovery increased. On
the other hand, the acid conversion decreased to a steady
value of about 70% since the increased H£S utilization
corresponded to a decrease in the stoichiometric quantity
of H£S recycled back to the sulfur generator. This '
decrease in acid conversion was accompanied by a slight
increase in S0£ evolution from the sulfur generator.
However, it is felt that the S02 evolution could be
reduced in fluid bed sulfur generators planned in future
development since the fluid bed would offer more uniform
temperature control. Moreover, the evolved S02 could be
recycled to the S02 sorber with a minimal increase in
load on the sorber. More importantly for the IR-2 run,
the increased amount of acid entering the H2S generator
did not result in any S02 evolution from that reactor,
indicating that the effective conversion of acid to
elemental sulfur was completed'in the H2S generator.
As pointed out earlier in this report,the various perform-
ance goals were based partially on past performance of
the various units operated separately. These goals may
not be particularly representative of the performance of
the integral process. Therefore, it is necessary to view
the effect of these recent H2S utilization and acid conver-
sion to sulfur responses in terms of the overall process
objectives, i.e. S02 removal, sulfur recovery, and H2
47
-------
Figure 13. Pore volume and surface area of recycled
carbon
0.07
100 200 300
Elapsed Time, hours
400
620
100 200 300
Elapsed Time, hours
400
48
-------
Figure 14. Sulfur generator performance
100-
s-e
2 9°'
i—
<
ISI
_ 1
= 80-
GO
cvj
re
70
C
& 100 i
Q
u
CO
o:
o
CO
t 80 •
CO
o
1—
z
i 60-
LLJ
>
z
o
CJ
o
0 40
0
1
•
m
CYLINDER H2S ~*
2468
) W 80 120
m
m
'— 1
'
.
2468
^0 80 120
IK-l — *•
_
1 '
1
"*~ PROCESS H2S
NUMBER OF CARBON CYCLES
110 12 I'l 16
160 200 240
ELAPSEH TIME, H
'
1 . i' '
NUMBER OF CARBON CYCLES
10 12 1H 16
till
"•-— 1K-Z
1
CYLINDER _».
H2S
18 20 22
III
280 320
OURS
1,
1
IS 20 22
i i i
160 200 210 28H 320
ELAPSED TIME/ HOURS
l— i
— <
— P1gSS
24 26 28 30
350 400 440 480
-
1 »
i — I
24 26 28 30
360 400 440~" 480
-------
input/S02 sorbed ratio. As can be seen from Table 5, the
reduction in acid conversion to sulfur in the sulfur
generator has not reduced the overall process performance.
In fact, the three above mentioned responses have changed
in the desired direction of increased S02 recovery as
elemental sulfur.
Sulfur Stripper/H£S Generator -
As shown by Figure 9 (inlet sulfur loading vs. elapsed
time), there was successful regeneration of the carbon to
the same conditions throughout the runs, Although the
inlet sulfur loading on the carbon was slightly higher in
the IR-2 run due to lower average temperatures in the H2S
generator/S stripper, the S02 removal remained steady and
well above the target goal of 90% throughout the run.
As shown in Figure 15, the lower H2 feed in IR-2, i.e. a
H2/S02 sorbed ratio of 3.4 compared to 3.9 for IR-1,
yielded 100% H2 utilization in the H2S generator. The H2
utilization to H2S, however, steadily decreased during the
IR-2-A run after the switch from cylinder H2S from 81%, to a
steady value of approximately 61% which was about the same
in the IR-2-C run. This decrease in the production of H2S
during the run was probably due to the shift in the distri-
bution of the acid conversion reaction between the sulfur
generator and H2S generator where the acid entering the H2S
generator was effectively converted to sulfur. As the
distribution of this conversion between the reactors stabi-
lizes, the H2S production reached steady state.
Sulfur Condenser -
Previously, sulfur recovery was reported on the basis
of the percent of stripped sulfur that was recovered. But
in the IR-2 run with more acid being converted to sulfur
in the H2S generator and the subsequent reduction in H2S
production, the percent of the stripped sulfur recovered
exceeded the 25% goal. To provide a more representative
evaluation of the sulfur condenser performance, the percent
sulfur recovered based on the amount sorbed as S02 is pre-
sented in Figure 16 for both the IR-1 and IR-2 runs. The
lower H2S production or lower H2S breakthrough from the
sulfur generator resulted in a higher sulfur recovery. For
IR-2, this response variable fluctuated, but steadied out
to approximately 88% as the H2S production stabilized com-
pared to a goal of 10070. This was significantly higher
than the 75% achieved for the IR-1 run.
50
-------
Figure 15. H2S generator/sulfur stripper performance
IR-2
Ul
100 T
90-
CM
70
<
rsl
100 T
80 ••
60 ••
1
CYLINDER H2S ""**
•
2 1 6 8 .
0 40 80 120
1KB
' 1 ,
t
2468
0 40 SO 120
— - • i 1 i
1
|
- PROCESS H2S 1 ^P -
1
1
1
1
NUMBER OF CARBON CYCLES
10 12 14 16 18 20 22
160 200 240 280 320
ELAPSED TIME, HOURS
1
|
I , ,
1
1
""I
NUMBER OF CARBON CYCLES |
10 12 14 16 18 20 22
II 1 f I t i
160 200 240 280 320
— pT2sss
24 26 28 30
360 400 440 480
^^^^j
r^™""i •
, a
24 2(6 23 50
360 400 440 tsn
ELAPSED TIME, HOURS
-------
Figure 16. Sulfur condenser performance
m
o
UJ
CQ
o:
o
oo
CM
O
0
u
100-
90 •
80-
70-
60 •
50 •
40 •
30
IR-1 »!•» IR-2 , ,
i -H!-_r-' "*"'
•""i
1- §
— * , _ _j
i
i
i
NUMBER OF CARBCN CYCLES
2 4 6 8 10 12 14 16 18 23 2.7 24 26 28 30
iti Tiii i I t i ifit
40 80 120 160 200 2*;0 280
ELAPSE! TIME/ HOURS
320
360
403
440
480
-------
4.2.3 Material Balances
Carbon Balance -
An attempt was made to perform a carbon balance in each of
the integral runs; but, due to excess carbon spillage in
the IR-1 run, only an overall carbon balance was made for
the IR-2 run.
For the carbon balance that was determined for Run IR-2,
the carbon loop was closed. At the start of the run the
carbon fed to the reactors and inventory hopper was weighed
carefully to determine the initial input to the system.
Carbon leaving the system during the run as dust, leakage,
or C02 was measured; and after completion of the run all
material was removed from the system and weighed. Material
weight was corrected for sulfur and moisture content to
obtain the true weight of carbon. The carbon balance is
presented in Table 6.
Table 6. CARBON BALANCE FOR
IR-2 INTEGRAL RUN
IN
SYSTEM CHARGE = 428 IBS. C
.
TOTAL IN =428 Lbs. C
OUT
CARBON REMOVED FROM
INTEGRAL PILOT PLANT
S02 SORBER CYCLONE
S02 SORBER DUST FILTER* =
CARBON SAMPLES
C02 EVOLUTION
LEAKAGE AT BUCKET
ELEVATOR
TOTAL OUT
359
31
20
3.
3.
6
423
Lbs
Lbs
Lbs
5 Lbs
6 Lbs
Lbs
Lbs.
. C
. C
. C
. C
. C
. C
C
*Based on constant rate of collection for the last steady
state period in IR-1 run of 0.12 Ib. C/hr.
53
-------
Hydrogen Balance -
The hydrogen balance around the H2S generator/S stripper
is given in Table 7. The sources of hydrogen input were
hydrogen in the inlet gas and hydrogen as adsorbed water
and sulfuric acid on the incoming carbon. Hydrogen output
consisted of H2S, H2 and H20 in the outlet gas.
Table 7. HYDROGEN BALANCE FOR H2S
GENERATOR/SULFUR STRIPPER
Run
IR-1
AVG.
IR-2
AVG.
Period
1
2
3
4
5
158 Mrs.
6
7
8
113 Mrs.
H2 Input,
Lbs./Hr.
Gas - H2
0.451
0.591
0.591
0.663
0.601
0.587
0.492
0.492
0.492
0.492
Carbon - H20
0.043
0.048
0.040
0.015
0.014
0.029
0.094
0.140
0.163
0.132
H2 Output,
Lbs./Hr.
Gas - H2S
0.321
0.409
0.409
0.480
0.417
0.416
0.397
0.332
0.300
0.343
Gas - H2
0
0.002
0.005
0.01
0
0.005
0
0
0
0
Gas - 1120
0.143
0.214
0.229
0.194
0.217
0.196
0.103
0.233
0.234
0.190
Total ,
Lbs./Hr.
In
0.494
0.639
0.631
0.678
0.615
0.616
0.586
0.632
0.655
0.624
Out
0.464
0.625
0.643
0.684
0.634
0.617
0.500
0.565
0.534
0.533
For the IR-1 run, the difference between input and output
averaged less than 170. For IR-2, the hydrogen output
averages 11% below the input. This 11/4 discrepancy could
reasonably be attributed to experimental error in gas
analyses, in flow rate measurements and in the inlet carbon
moisture analysis. The measured carbon moisture content for
IR-2 is higher than the theoretical value predicted by
assuming that the water associated with the sulfuric acid
on the carbon is at equilibrium between the gas and carbon
phases on discharge of the carbon from the sulfur generator.
Therefore, the 11% error in the H2 material balance is
considered within experimental error.
Sulfur Balance -
During the integral runs, the carbon and gas streams into
and out of each vessel were monitored to obtain a total
sulfur balance for each unit and for the total pilot plant.
For the IR-1 integral run, there were five steady state
periods. Two of the periods with cylinder H2S gave an
54
-------
overall sulfur balance of IN/OUT of 8.65/8.93 and 8.31/
6.41. Some recirculated sulfur was carried over into the
sulfur product due to flooding of the mist eliminator in the
sulfur condenser causing uncertainties in these balances.
This was rectified and the overall sulfur balances with
process H2S were very good. The sulfur balances for the
other periods are given in Table 8. Prior to the last
two steady state periods, a problem with the sulfur
Table 8. SULFUR BALANCE FOR INTEGRAL RUNS
Integral
Run
IR-1
IR-2
Overall
Period
3
4
5
7
8
87 Mrs.
Input
Sulfur
S02
Sorber,
Ibs./hr.
2.54
2.68
2.68
2.43
2.40 '
2.42
Sulfur Outputs,
Ibs./hr.
S02
Sorber
Off -Gas
0.19
0.20
0.20
0.13
0.16
0.14
Sulfur
Gen.
Off-Gas
0.92
0.55
0.46
0.38
0.31
Sulfur
Cond.
Product
1.58
1.61
1.86
1.94
1.94
1.85
Total
Output,
Ibs./hr.
2.73
2.61
2.53
2.48
2.30
Input/Output
Ratio
--
0.96
0.99
1.05
analysis of the sulfur generator off-gas was rectified so
the last two sulfur material balances should be very
reliable. The sulfur balance IN/OUT ratios were 2.54/2.43,
2.70/2.64, and 2.78/2.57 as given in Table 8 for opera-
tion with process H2S. This gave an average overall sulfur
balance while on process H2S of 2.64 Ibs. S/hr. in and
55
-------
2.56 Ibs. S/hr. out tor an average discrepancy of about 37o,
or 9770 of the sulfur in or out was accounted for. This
indicates that especially for the Periods A and B, the
conversion calculations and indicators of carbon perform-
ance are accurate.
An overall sulfur balance around the pilot plant system was
calculated for the total period of process H2S utilization
during the IR-2 runs. The results of the sulfur balance
are given in Figure 17. The amount of sulfur entering
the sorber was 211 pounds and the amount leaving the sorber
was 12 pounds, so that the amount sorbed which entered the
regeneration system was 199 pounds. Of this, 27 pounds
passed out of the system in the sulfur generator off-gas
and 161 pounds was recovered in the sulfur condenser. This
accounts for all but 11 pounds, or 5%, of the input sulfur,
and this is probably within experimental error.
Twenty-eight pounds of sulfur which leaked out of the sulfur
pump were collected after the run was over and included in
the material balance as part of the 161 pounds of recovered
sulfur, but the 28 pounds may not have accounted for the
entire leak.
Sulfur balances for periods of operation with process H2S
during both the IR-1 and IR-2 runs are shown in Table 8.
The results are presented as rates of sulfur input and
output during each interval. Input consists of S02
adsorbed from the flue gas, and output is divided into
three categories: ,1) sulfur contained in the S02 sorber
off-gas as S02 or sorbed on the carbon dust, 2) sulfur in
the sulfur generator off-gas as S02 or H2S, and 3) sulfur
product collected in the condenser. The values for sulfur
collection rate in the condenser have not been corrected
for the sulfur leak, because of uncertainty in distribution
the leak over the various periods. It is clear from
Table 8, however, that the leak was worst during the
periods from 81 to 126 hours, when the sulfur collection
rate fell significantly.
4.2.4 Process Control
Overall process control of the integral pilot plant was
achieved by a combination of automatic and manual control
of critical process variables. Of these variables, only
the temperatures of the S02 sorber's second stage and the
sulfur stripper/H2S generator, the overall carbon recycle
rate, and the carbon levels in the seal legs of the
reactors were under automatic control. The other vari-
ables, mainly gas rates and compositions, were controlled
56
-------
Figure 17. Sulfur balance for IR-2 run during operation
under process H2S
211 IBS.
REGENERATED
CARBON
DUST
CYCLONE FILTER
OFF-GAS
S02 SORBER
OFF-GAS
SULFUR
GENERATOR
-------
manually on the basis of material balances and the process
performance responses.
Manual control of the integral plant was accomplished by
maintaining the flue gas rate and S02 concentration con-
stant to the S02 sorber and adjusting the gas rates and
compositions to the other reactors. During operation with
process H2S, the hydrogen/nitrogen rate to the sulfur
stripper/H2S generator was the only gas flow varied.
The overall carbon recirculation rate was controlled auto-
matically at 30 Ibs./hr. by a gravimetric solids rate feeder
to the S02 sorber. However, at several points in the
carbon loop, the carbon flow was under local automatic con-
trol to maintain the desired levels in the carbon seal legs
to and from the reactor vessels. The control variable
for these seal legs was the pressure drop of N2 purge gas
in the legs.
In the S02 sorber, the flue gas entered at 300°F and was
cooled to about 175°F at the second stage by direct water
injection into the second stage carbon bed. This tempera-
ture was maintained by automatic control of the water
injection rate. The temperatures of the remaining three
stages were not controlled, but allowed to reach their
equilibrium temperatures.
The temperatures of the sulfur stripper/H2S generator,
sulfur generator, and sulfur stripper were maintained
by electrical heating. In the sulfur stripper/
H2S generator, the temperature was automatically controlled
by regulating the current to electrical resistance heaters
on the reactor walls. Temperature in the acid converter on
the other hand were controlled primarily by manually setting
the voltage rheostats to the electrical resistance heaters.
Response to changes in temperature in this reactor was slow
because of the heat transfer characteristics of moving bed
reactors. Also, due to the exothermic reaction occurring
in this moving bed reactor, localized over-temperatures were
somewhat compensated for by adjusting the water content on
the carbon before it entered the reactor. The temperature
in the sulfur condenser was also controlled manually be
regulating steam pressures to the exchanger and tracing.
58
-------
i.2.5 Process Concept Modifications
Based on findings which arose at various stages in this
development program, several modifications of the original
process concept have evolved that not only have simplified
process design and control requirements but also have
reduced the capital and operating costs of the system.
Briefly, these modifications involve:
1) The method in which the flue gas temperature
is controlled within the S02 sorber
2) The combination of sulfur stripping and H2S
generation into one process step
3) The fact that some of the acid can be reduced
in the sulfur stripper/H2S generator without
deleterious effects on the process.
In the S02 sorber, the first stage was operated at the
stack gas temperature of approximately 300°F to remove the
30 to 50 ppm S03, and the second stage was operated at
175°F to facilitate SC>2 removal. The temperatures of the
remaining fluid bed stages were not controlled, but were
allowed to reach their equilibrium temperatures. To cool
the flue gas between the first and second stages, an
external shell and tube heat exchanger was installed and
evaluated initially. The off-gas from the first stage was
withdrawn from the sorber, circulated through the exchanger
and returned to the sorber below the second carbon bed.
Since certain design and economic constraints made this
heat exchanger method of gas cooling particularly unattrac-
tive for large process units, direct water injection into
the second stage fluid bed was installed. After the posi-
tion of this water spray nozzle and the proper nozzle
pressure were determined, this latter method was used
satisfactorily during the integral runs and has been
incorporated into the designs for the larger systems. As
a result, the capital investment was reduced.
In the original process concept, the thermal stripping of
the elemental sulfur from the carbon and subsequent partial
reaction of this sulfur with hydrogen to H2S were to be
performed in separate process vessels. However, bench
scale and pre-integral pilot plant testing indicated that
these two process steps could be carried out satisfactorily
in one reaction vessel, thereby appreciably simplifying
the overall process. This process change was subsequently
incorporated into the integral pilot plant as a fluid bed
sulfur stripper/H2S generator.
59
-------
The design and operating conditions for the sulfur generator
were selected initially to achieve 997o conversion of acid
to elemental sulfur with maximum H2S utilization. It was
felt that this high level of acid conversion was necessary
because the introduction of acid loaded carbon into the
high-temperature H£S generator could increase carbon burn-
off and S02 evolution. During the integral runs, however,
it was found that part of the acid on the carbon was reduced
to sulfur in the sulfur stripper/H2S generator without
deleterious effects of S02 evolution or enhanced carbon
burn-off. As a result the integral process can be operated
with lower H2S recycle and, consequently, without H2S break-
through from the sulfur generator. Moreover, reducing the
amount of acid conversion in the acid converter should
allow greater flexibility in process control.
60
-------
SECTION 5
PRE-INTEGRAL PROCESS DEVELOPMENT
5.1 APPARATUS AND PROCEDURE
There were numerous types of reactors and procedures that
were used in the process development. In general the
studies can be classified by the type of reactor which
included thermogravimetric, fixed bed, moving bed, batch
fluid bed, and multistage fluid bed reactors. For each
specific reactor type, the procedures are similar for
studying the separate process steps and many of the
procedures will thus be grouped as to reactor type.
5.1.1 Thermogravimetric Reactor
In many adsorption or desorption processes using activated
carbon, the change in carbon weight is directly proportional
to the amount of a constituent adsorbed. Therefore, by sus-
pending a small container with a carbon sample from a gravi
metric balance and placing the suspended sample in a therm-
ally controlled environment, the carbon can be exposed to
a gas containing some constituent to be removed and the
amount of pickup be followed directly as a function of time.
Such an apparatus was assembled as described below, but
since it was only used under this contract for studying the
kinetics of S02 sorption, frequent mention will be made to
S02 even though the equipment can be used more generally.
The apparatus is shown schematically in Figure 18.
Simulated flue gas or some other suitable gas mixture is
mixed by metering sulfur dioxide, nitric oxide, and nitrogen
through individually calibrated Brooks "Sho-Rate" rotameters
equipped with dual floats (steel and glass) and highly
stable "ELF" nonrising stem needle valves. Use of these
valves together with ordinary cylinder gas pressure
regulators allows accurately reproducible setting of flow
rates. In clean rotameter tubes, drift in flow indication
has never exceeded 0.5% of the tube scale length over the
course of any experiment. Flow rates in a convenient range
are obtained by using cylinder gases containing approximately
1% S0£ in N£ or He and 0.3% NO in N2 or He. Actual cylinder
concentrations for the S02 tanks are determined by wet
chemical analysis. The NO tanks are calibrated by the
manufacturer.
Water vapor is added by means of a bubbler apparatus con-
sisting of two 14" x 2-1/2" OD gas scrubbers connected in
series and completely submerged in a thermostat bath. A
head of about 6-10 inches of water was maintained above
61
-------
Figure 18. Thermogravimetric apparatus
to
A - Magnehelic Pressure Gauge
B - Humidifier Thermostat
C - Sample Thermostat
0 - Bucket Envelope
E - Fiber Envelope
F - Electrobalance
G - Oven Temperature Regulator
H'- Vacuum Pump
J - Thermocouple
K - Vent
L - Inlet Toggle Valve
M - Exit Toggle Valve
-------
the fritted dispersion tube in the first bubbler and about
2 inches in the second. Analysis of the gas effluent from
the humidifier assembly using a Cambridge Systems dew point
hygrometer has shown that saturation is achieved at all
thermostat temperatures and carrier gas flow rates tested
in the S02 sorption rate measurements.
The dry and humidified gas streams are mixed and then pass
either to the sample or to a vent. This enables the
simulated flue gas mixture to be completely mixed before
introducing to the sample.
A needle valve in the vent line allows the pressure drop
in the vent and sorption lines to be equalized so that flow
may be switched at the start of a run without disturbance
of pressures in the humidifier or the rotameters.
In the main part of the reactor, shown in detail in Figure
19, the carbon sample is held in a cylindrical fused silica
bucket suspended from the beam of a Cahn RG electrobalance.
The gas mixture is admitted to the sample through a nozzle
positioned inside the bucket about 3 to 5 mm above the
surface of the carbon layer. The nozzle was made of Teflon
stock drilled to give dispersion of gas over the whole
sample. Gas is conducted to the nozzle through a 1/16"
stainless steel tube which passes through a Teflon sealed
thermocouple gland attached by Swagelok fittings to a
standard glass ball and socket joint which is sealed to the
envelope in which the sample is suspended. Use of the gland
and joint allows the lateral, rotary and angular motion
required to position the dispersion nozzle within the bucket.
The gas injection tube is connected to the gas mixing
system by a coil of 15 feet of 1/16" SS tubing which allows
preheating of the influent gas and provides sufficient
flexibility to avoid undue strain on the glass-to-tnetal
connections.
Also connected to the sample bucket envelope is a separate
nitrogen purge line which is used to prevent air from
entering the sorption apparatus after outgassing the sample
and during the time that gas mixtures are being set up
prior to a run.
The sample bucket envelope is removable so that samples may
be changed and it is attached by a ball and socket joint
to an upper tube which surrounds the bucket suspension
fiber. The fiber envelope is about 14 inches long by 25 mm
in diameter and contains a 3" section of 10 mm tubing at
the upper end in order to prevent back diffusion of air
into the sample section. Just above this construction, a
side arm is attached which is connected to a Cast vacuum
63
-------
Figure 19. Detail of the thermogravimetric reactor
sample bucket envelope
SJ 35/25 Pyrex Rail Joint
SJ 18/9 Pyrex Ball Joint
Swagelok Fittings and Thermocouple Gland
Gas'Injection Tube
Purge Gas Line
Sample Bucket
Dispersion Nozzle
Thermocouple
Suspension Fiber
Swagelok Connection to Preheat Coil
64
-------
pump through a needle valve and rotameter. During a run,
the flow through the vacuum system is maintained at such
a rate that no corrosive or toxic gases are allowed to
enter the atmosphere or the electrobalance mechanism case.
Mixture with sufficient room air entering through the open
top of the fiber envelope also prevents condensation of
water from the gas mixture on the upper end of the fiber
or in the vacuum system.
Reaction temperature is maintained by an air bath oven
which is placed up around the reactor assembly and encloses
the sample bucket envelope and the preheat coils attached
to the gas injection and purge lines. The oven, which is
very well insulated but weighs only about 10 pounds, was
constructed in two parts. The bottom section was made
from a 12" x 6" x 7" aluminum chassis box, insulated on the
inside with 1/2" Fiberfrax ceramic fiber blanket. This
section contains two 660 watt exposed coil heating elements
and a centrifugal blower with an external fan cooled motor,
The upper section, which formed the air bath, was constructed
from 1/2" Fiberfrax blanket sandwiched between 9 mil
corrugated aluminum sheeting. In operation hot air is blown
out of the lower section, circulates through the air bath
and returns through a 2-1/2" hole in the top of the heater
chamber. Adequate temperature control is obtained using a
Honeywell time proportioning controller with a thermistor
sensor mounted in the outlet of the blower. Reaction tempe-
rature is measured*by a thermocouple placed just above the
sample bucket.
All lines coming into contact with water vapor are traced
with electrical heating tapes to prevent condensation.
Weight pickup of SC>2 as sulfuric acid is measured by a Cahn
RG electrobalance which is connected to a 10" Texas
Instruments recorder to give a continuous readout of weight
versus time. The sensitivity of the balance system is such
that a full chart width deflection corresponds to a pickup
of about 5 Ibs. S02/100 Ibs. carbon. In operation the flow
of gases into the sample bucket causes a weight deflection
corresponding to the flow rate. At the rate used in these
experiments (1,000 cc/min.) the deflection is about 10
milligrams. Because the flow rates are very stable and
reproducible, the deflection is always constant over the
course of any experiment and the flow deflection can simply
be subtracted from the observed weight changes due to S02
sorption. Gas flow around the bucket and suspension fiber
also introduces a small amount of vibration into the weighing
system. However the resultant noise band is less than 0.1 mg
wide in most cases and had no effect on the accuracy of the
measurements.
65
-------
The weight pickup vs. time curves are differentiated by
finding straight line slopes across closely spaced time
intervals and plotting these rates vs. the mean loading
in the corresponding interval.
5.1.2 Fixed Bed
The basic components of a typical fixed bed apparatus are
shown schematically in Figure 20, The reactor, containing
a known weight of carbon vertically oriented, typically
has been 1" diameter. The reactor is equipped with a
heating system consisting of electrical heaters wrapped
around the pipe and a rheostat to control the heat input.
Carbon temperature is measured with a thermocouple inserted
into the carbon bed and gas temperature is also measured
with a thermocouple. A gas mixing system provides control
of reactant gas flows. Rotameters are used to measure the
gas flow rates. The direction of gas flow is generally
downward through the bed to avoid any possibility of
fluidizing the bed. A gas chromatograph was used to moni-
tor inlet and outlet concentrations of the gas constituents.
In operating a fixed bed system, the general procedure is
to establish the desired inlet gas conditions with the gas
bypassing the reactor, and to establish the desired carbon
bed temperature with the heaters. When all conditions are
set, the gas flow is switched to enter the reactor. The
outlet gas is analyzed to give concentration vs. time data.
The data from this type of experiment is transient rather
than steady state. Outlet gas concentrations change con-
tinuously and eventually approach the inlet concentrations.
Following the run the carbon is removed and analyzed for
sulfur compounds.
The data can be analyzed to provide screening information
on overall degree and rates of conversion and type type of
reactor was used to study each of the process reaction
steps.
5.1.3 Moving Bed
The main features of a moving bed system are shown schematic-
ally in Figure 21. There are many similarities to the fixed
bed, including the systems for gas mixing, reactor heating,
and gas analysis, the vertical reactor orientation, and
the fact that the carbon beds are not fluidized. The main
differences are that the moving bed reactor is a continuous
flow system which operates under steady state conditions
and the direction of gas flow is upward through the bed in
66
-------
Figure 20. Fixed bed reactor system
Thermocouple
Gas
Mixing
System
RheostaT Heater
..r
Vent
or
Gas
Anal.
1"0 Fixed Bed
Carbon Sample
67
-------
Figure 21. Moving bed reactor system
tarbon
Feeder
\
Vent
Tae
OdS
Mixing
System
Heater C
or (
Heating /
System ?
i
i
iMHMI
••••
*
/
TGas
Moving Bed
Reactor
Thermocouples (Typical)
__i
— Carbon
_. .Cone Discharge Angled To
Insure Plug Flow of
Carbon and Carbon Forms
Gas Distributor
£1
Carbon
Discharge
Feeder
Carbon
68
-------
order to provide countercurrent contact with the carbon
phase. In addition, the moving bed requires a carbon
flow control system to maintain the desired carbon flow
rate through the reactor. Carbon feeders are used at both
the inlet and outlet of the reactor. The carbon flow
distribution through the reactor is close to theoretical
plug flow.
The moving bed reactor is subject to temperature control
problems as the vessel size increases due to poor heat
transfer characteristics. Its main advantage is the
capability for a long carbon residence time and low gas space
velocity, two parameters which are important in obtaining
high conversion of reactants.
Low gas space velocities can be obtained in a moving bed
because the minimum fluidizing velocity constraint which
applies to a fluid bed reactor is not applicable in a moving
bed.
A bench scale (1-1/2" dia.) and pilot (8" dia.) moving bed
reactor were used to study the conversion of sulfuria acid
to sulfur by reaction with H2S. Other reaction steps were
not studied in the moving bed equipment.
5.1.4 Batch Fluid Bed
The batch fluid bed reactor shown in Figure 22 is a differ-
ential reactor useful primarily for studying kinetics of
a reaction. The approach is to fluidize a batch carbon
sample with a reactant gas at a space velocity sufficiently
high that the concentration of the gas reactant remains
essentially unchanged as it passes through the carbon bed.
Carbon samples are taken from the bed at closely spaced
time intervals and are analyzed to generate data that is
convenient for a kinetic analysis.
The equipment consists of a cylindrical vessel with a gas
distributor plate to support the carbon bed. Electrical
heaters provide heat input through the walls and are con-
trolled manually with a rheostat. Temperature can be con-
trolled accurately and is uniform throughout the bed
because of the excellent heat transfer and mixing charac-
teristics of the fluidized bed. A gas mixing system pro-
vides the desired gas flow rate and constituent concentra-
tions, and the gas is introduced below the distributor plate.
s
A 4 inch diameter reactor of this type was used to develop
a model for the reaction of acid with H2S to form elemental
sulfur and to study the thermal stripping of sulfur.
69
-------
Figure 22. Batch fluid bed reactor
Thermocouple
Gas
Mixing
System
Heater
To Vent or -
i
Gas
Anal.
_ Manual Carbon Sampler
4"0 Batch Fluid
Bed Reactor
Fluid Bed of Carbon
Drilled Gas Distributor
Plate
Manual
Carbon
Discharge
70
-------
5.1.5 Multistage Fluid Bed Reactor
A multistage fluid bed reactor system is shown in Figure 23.
Carbon is fed at a constant rate onto the top stage of the
reactor and flows by gravity from one stage to the next.
Gas distributor plates support the carbon beds. The plates
are drilled to a specified percent open area, typically 8%
with 0.125" dia. holes. The expanded bed depth, determined
by the height of the overflow weirs, is generally limited
to a value equal to the vessel diameter in order to
minimize slugging of the bed. Downcomers extend into the
carbon beds to within about 1-1/2" of the distributor
plates. The downcomers contain baffling to prevent
slugging and improve the carbon flow characteristics.
Carbon passes out of the reactor into a seal leg designed
to prevent the escape of any reactant gas. The carbon level
in the seal leg is controlled automatically.
Reactant gas enters below the bottom distributor plate and
passes upward through the reactor, countercurrent to the
carbon flow. The gas is sampled for analysis at the inlet
and outlet of the reactor, and may also be sampled at each
stage of the reactor.
Temperature is controlled by electrical heaters wrapped on
the outside of the reactor. Separate temperature control
is frequently provided for each stage or for every two
stages.
The multistage fluid bed reactor operates at steady state
conditions. The time needed to reach steady state after
start-up or after a change in operating conditions may be
fairly long; however, so this type1of reactor is not
fenerally used in extensive kinetic studies. Multistage
luid bed units ranging from 4-18 inches in diameter were
used to study the separate reaction steps and for thermal
sulfur stripping.
5.1.6 Sulfur-Carbon Thermal Equilibrium
Adsorption isotherm points were obtained by saturating an
inert gas stream with sulfur vapor at particular temper-
atures in order to produce a known partial pressure of
sulfur, and passing this stream over carbon held at the
desired adsorption temperature for a length of time suffi-
cient to establish equilibrium. Sulfur loading on the
carbon was determined for each set of equilibrium conditions
by means of combustion analysis of the individual samples.
71
-------
Figure 23. Multistage fluidized bed reactor
Gas Anal.
(Typical)
Carbon Anal.
(Typical)
Carbon
Feed System
Heater;
(Typical)
Thermocouple (Typi
Pressure
Drop ©
(Typical)
Gas Source
or Mixing
System
I
Gas
Analysis
1
Gas
Analysis
Multistage Fluidized
Bed Reactor
Overflow Weirs
Baffling
•Downcomer
.Fluidized Carbon Bed
Drilled Gas
Distributor Plate
Level Control
72
-------
The apparatus used is shown diagramatically in Figure 24.
In operation about 0.5 gm. of carbon sample was introduced,
and after purging with nitrogen, the vapor generator and
adsorption tubes were heated to the desired temperatures by
manual control of Variac transformers. Nitrogen flow
through the apparatus was maintained at a rate (10 cc/min.)
low enough to attain saturation in the relatively inefficient
generator. Connecting tubing between the generator and
adsorber was heated to a temperature above that of the
sulfur in order to prevent condensation. The exit tube
at the bottom of the adsorber was also heated down to a
condensation trap which was used to prevent sulfur from
plugging the exit line. Nitrogen was injected into the
line at this point in order to seal against entry of air
from the exit.
After an equilibrium time of 9-1/2 hours which was set based
on observations during the first experiment in which the
highest sulfur pickup was achieved, gas flow through the
generator was shut off. The stopcock between generator
and adsorber was closed, and the entire apparatus was
removed from the furnaces and allowed to cool. The carbon
was then weighed, avoiding contact with atmospheric moisture,
and the sulfur loading of the entire sample determined using
a Dietert Sulfur Analyzer.
The carbon starting material used in these experiments had
initially been loaded with about 24% sulfur by complete
reduction of HoS04 in sulfur generation experiments. Carbon
having an initial sulfur load was used in order to reproduce
adsorbent properties due to S02 pickup and reduction in
accordance with our intent to use these data in connection
with thermal stripping work.
Since the existence of chemisorbed sulfur on carbon has been
suggested by previous thermal stripping studies in which a
residual loading remained after extended purging with inert
gas, additional runs were made in which samples were purged
for 9.5 hours with pure nitrogen at several temperatures
and then analyzed as noted above.
5.1.7 Solvent Extraction of Sulfur Procedures
Screening evaluations were made of sulfur extraction with
carbon disulfide, ammonium sulfide and xylene by successive
stages of slurry contact in stirred beakers. Since the
procedures differed slightly for each solvent they will be
described separately.
73
-------
Figure 24. Sulfur adsorption apparatus
Thermocouples
i 1
Tapered Joint LJ
"^-C/^
Variac < — l //
Vycor
Adsorption
Tube
' -^^
Carbon __^_^^
^ M •* ^ -4» A j4 n^r-i^
Fn uted DISK — -
Variac < — °
r
VL
i
A
^
,i
i*v
!U*
Vr/
i
Hoa-Hnn Tanp 1 ...
i
;
j^^
_,
<
Heating Tape — ^
^j f
i / ) ' ' "'
Stopcock '
Tube Furnace ~^_^
^^^«
"" Thenrocouple
Shield
t/\
L
—
i i
— »_
_^
' */
\ ' Taoered \\
\ \ '°H Variac
Sulfur
Variac
Vent <
Condenser
=:—NZ
N2
-------
Carbon Bisulfide Extraction -
Twenty grams of a sulfur loaded carbon were slurried in a
vertical stoppered flask for about 0.5 hr. at room temper-
ature. The solvent was then decanted and evaporated to
recover the dissolved sulfur for weighing. A fresh
quantity of solvent was added and the procedure repeated
to the desired number of extraction stages. After the
final extraction the carbon was freed of excess solvent on
a steam bath and then oven dried. Samples of the carbon
were also taken at the end of each extraction stage, dried
and analyzed for sulfur.
Ammonium Sulfide Extraction -
Forty grams of sulfur loaded carbon were slurried in a
beaker at room temperature with 100 ml. of 20% ammonium
sulfide for 15 minutes. The ammonium sulfide solution was
then filtered off and the slurrying repeated with 100 ml. of
fresh solution. After each stage of extraction a carbon
sample was taken, slurry washed with water to remove excess
ammonium sulfide, oven dried and analyzed for sulfur. The
extraction solutions after each stage were acidified and
boiled to liberate recovered sulfur which was then reclaimed
through carbon disulfide extraction and dried.
Xylene Extraction -
Twenty grams of a loaded carbon were slurried with 50 ml. of
xylene, in a flask with an extended neck and thermometer,
at a temperature of 105-110°C (220-230°F) for 20 to 30
minutes. The solvent was decanted and a fresh quantity of
solvent added and the procedure repeated. The carbon was
sampled between extraction stages. The final carbon was
freed of xylene by boiling in water and then oven dried
overnight. The xylene extracts after each stage were
evaporated on a steam bath and residues oven dried for an
hour before weighing to determine sulfur extracted.
Bench Scale Extractor -
A schematic of a recirculating bench scale apparatus used
in a study of sulfur extraction from carbon is shown in
Figure 25. The carbon to be extracted was placed in a
quartz tube inside the Hoskins furnace. The furnace was
then heated to the extraction temperature and the system
was purged with pure nitrogen or helium metered from a gas
cylinder to expel air from the system. The extractant
was then added to the system from a separatory funnel and
recirculated through the system by a stainless steel
centrifugal pump at a rate of 35 ml./min. for 30 minutes.
75
-------
Figure 25. Flow schematic of recycle
extraction apparatus
Syringe
Pump
Distilled
Flash
Vaporizer
N2
Cylinder
Flowmeter
Furnace with
Quartz Tube
Carbon
Sample
Pump
Drain Exhaust Drain
76
-------
The solvent was then drained from the system, fresh solvent
added and the procedure repeated for the desired number of
stages. Samples of carbon were taken after each extraction
stage, dried and analyzed for sulfur. The extractant was
saved for sulfur analysis. After the final extraction the
carbon was steamed for eight hours at a known rate by
metering water with a syringe pump to a flash vaporizer to
produce the steam. The carbon was then cooled under a
nitrogen purge, oven dried and analyzed for sulfur. The
system contained a water leg for pressure relief, a
manometer for pressure measurement and a trap for solvent
surges.
5.1.8 Procedures in Bench Scale H2S Generation Studies
Apparatus for investigating reaction rates of hydrogen
and sulfur vapor is shown schematically in Figure 26.
Component testing was performed as follows:
1. Sulfur Vapor Generator - The generator, shown in
Figure 27, operates by saturation of a stream of
nitrogen which is bubbled through molten sulfur.
This stream is mixed with the main reaction gas
stream and manipulation of the sulfur temperature and
the relative flow rates allows adjustment of the final
sulfur vapor concentration. Heat is provided by a
beaded heater wound on the body of the vessel in such
a way as to provide a greater watt density in the
upper portions above the liquid level to prevent con-
densation in this zone. Splashing of the liquid
against the higher temperature walls in this zone was
to be prevented by an internal baffle. Visual
observations through the range of expected operating
conditions confirmed that the baffle was operating
properly. Tests were also made to determine that
the unit could be brought back to room temperature
and reheated without difficulties. Maintenance
of a low flow of nitrogen during cooling prevents loss
of liquid sulfur into the gas dispersion tube which
would lead to plugging. It was also found that
channeling in the solidifying melt occurs so as to
maintain free gas passage through the vessel after
cooling, thus allowing inert purging during the next
heating cycle.
2. Leak tests showed that the 2" pipe union cap on the
sulfur generator would not seal properly under the
closing torque which could be applied to the assembled
apparatus. Lapping the sealing surfaces allowed tight
closure but if another such unit were constructed, the
cap should probably be flanged and gasketed.
77
-------
Figure 26. H2S generation kinetics apparatus
Vent
00
GC Valve
GC
Vent
Heaters
A - Sulfur Generator
B - Gas Preheat
C - Manifold Tracing
D - Tube Furnace
E - Tube Furnace
F - Manifold Tracing
G - Condenser Inlet Heat
Sulfur
Condenser
and Filter
r
N2 H2
Low Press.
Regulator
-------
Figure 27. Sulfur vapor generator
Vapor Out
Normal Sulfur Liquid Level
Internal Baffle
Sintered SS Gas Dispersion Tube
N2 In
79
-------
3. Heating tests were made to establish power loads
required for heatup and maintenance of design tempera
tures in sevel independent zones. It was found that
good temperature control could be obtained in all
zones by manual adjustment of auto-transf ormers .
4. All gas rotameters were calibrated.
5. Chromatograph system tests using H2$ at concentrations
in the 1-10% range showed good response and linearity.
80
-------
5.2 PRE-INTEGRAL RESULTS
5.2.1 S02 Sorption
Development of S02 Sorption Rate Model -
A large part of the pre-integral S02 sorption work was
directed toward measurement of the S02 kinetics to be used
in designing fluid bed adsorbers. A rate model was found
to mathematically describe the reaction kinetics and was
useful in developing the reactor design procedure. S0£
sorption is shown in Equation (5):
S02 + 1/2 02 + H20 Activated H2S04 (Sorbed) (5)
Carbon
The reaction rate is defined as the time rate change of
acid loading, dXv/dt,
Rate = L = f(T, Xv, Xg, X(), y, y, y, y^y* (10)
If all variables except acid loading and S02 concentration
are held constant, the rate equation reduces to the form
= K g(Xv) h(ySQ2) (11)
Several forms of this equation are given in Table 9.
These were evaluated by determining how well each could
fit the experimental data. The model which best fit
the data was then expanded to include the effects of
the other important variables.
*Terms are defined in Section 9, Nomenclature.
81
-------
Table 9. RATE EXPRESSIONS TO APPROXIMATE
S02 SORPTION DATA
Designation
Westvaco
Modified Westvaco
1*
Amundson
Jost
AVCO L
Differential Equation
Tracer
dXy_
dt
dXv
dt
dXy
dt
dXv
~dT
dt
dXv
"IT
___
2 aXv + b
AxvsYS02
The term Xvs in Table 9 is the saturation acid loading,
for which a value of 0.38 gin. H2S04/gm. carbon was obtained
by extrapolation of experimental rate data.
Kinetic data on S02 sorption was generated in 21 differen-
tial rate experiments in which the weight gain of a carbon
sample was monitored as a function of time while the sample
was exposed continuously to a constant concentration of
S02- The apparatus and procedure were described in
Section 5.1.1 and the experimental conditions are shown in
Table 10.
Raw data in the form of total sample weight versus time
was converted to a curve of acid loading versus time. This
curve was differentiated by calculating its slope at
numerous points to obtain reaction rate values at different
acid loadings. The data in this form is included in
Appendix A-l for the 21 differential rate experiments.
*References listed in Section 8, Bibliography.
82
-------
Table 10. EXPERIMENTAL CONDITIONS FOR S02 SORPTION
IN A DIFFERENTIAL RATE APPARATUS
Run
Number
1W, 153
ll*9
150
11*0
162
163
138
160
139
157
lM*
161
I1* 5
11*1
ll*2, l6k
' 1U3
151
152
158
Temperature
"F
150
150
150
200
200
200
200
200
200
200
200
200
200*
200
200
200
300
300
300
Gas Composition
S02
PPM
2500
1500
500
2500
2000
2000
2000
2000
2000
2000
2000
2000
2000
1500
1000
500
2500
1500
500
NO
PPM
150
150
150
150
0
50
150
150
150
150
150
300
150
150
150
150
150
150
150
°2
2
2
2
2
2
2
0.7
2
2
2
3.5
2
•<>
2
2
2
2
2
2
H20
10
10
10
10
10
10
10
5
10
15
10
10
JO
10
10
10
10
•10
10
C02
i
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
0
11.3
11.3
11.3
11.3
11.3
11.3
Inerts
I
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
Balance
and He were used as the inert gas. Experiments using He versus
and Ng showed a small effect on the rate which was within the
experimental error.
The initial rate model evaluation was based on Runs 139,
140, 141, 143 and 164. These five runs were conducted
at 200°F with all conditions held constant, except for
concentration, which was varied from 500 to 2500 ppm.
The reaction rate vs . acid loading data given in Appendix
A-l was applied to each rate equation by the following
procedure. The Westvaco equation in Table 9, which eventu-
ally was selected as the best model, is used as an example.
The Westvaco equation
dXy
dt
(12)
83
-------
can be rearranged to the form
In [
dXy/dt
1 -
In K + mln jr = Z
S02
(13)
At a constant S02 concentration, the right side of the equa-
tion is constant, and the values of In [^—_ ^ yx—] or Z,
calculated from the data, should also be constant if the
model is valid. A graph of Z versus In 7302 snould give a
straight line with a slope of m and an intercept of In K.
Values of Z are tabulated for each run in Appendix A-2.
The falues are reasonably constant for acid loadings
above 0.01 gm H2S04/gm carbon. The average Z value for
each run, omitting data at acid loadings below the 0.01
level, is given in Table 11 with the standard deviation
Table 11. STANDARD DEVIATION FOR WESTVACO
EQUATION FOR SORPTION DATA AT
200°F WITH NITRIC OXIDE PRESENT
FOR ACID LOADING ABOVE 0.01 GM
ACID/GM CARBON
Run
140
139
141
164
143
S02 Cone. ,
ppm
2500
2000
1500
1000
500
Average z
-6.99
-7.06
-7.18
-7.30
-7.60
Standard
Deviation
0.056
0.069
0.023
0.019
0.038
of the data. A plot of Z versus In Ygo? ^s snown in
Figure 28. After determination of the constants from the
slope and intercept, the resultant rate equation is:
0.0X05 (1 -
(14)
84
-------
Figure 28. Comparison of Westvaco model to
sorption data at 200°F
-6.0
Represents Error of Two
Standard Deviations
Ln
Equation (14) was used to calculate reaction rate values
for comparison with the experimental data. The experi-
mental and calculated sets of data are tabulated in
Appendix A-3 and the average percent difference between
them for each run is given in Table 12. The percentage
varies from 1.7 to 6.3% with an overall percent deviation
of 4.3%, which is considered very good agreement.
85
-------
Table 12. DEVIATION OF WESTVACO EQUATION FROM
EXPERIMENTAL S02 SORPTION RATE DATA
AT 200°F WITH NITRIC OXIDE FOR
LOADINGS ABOVE 0.01 GM ACID/GM CARBON
S02 Concentration,
ppm
2500
2000
1500
1000
500
Average Per Cent
Deviation
5.0
6.3
1.7
4.8
3.7
A second Westvaco rate equation was evaluated in a
similar fashion, but with poorer results. The equation was
arranged in the form
(4-7)
dXy/dt
(l-Xv/Xvs)2
] = In K + m In ygo = Z
(15)
and values of Z were calculated as before. Table 13
shows the average Z values and the standard deviation
values, which are much greater than for the first Westvaco
model. The second model clearly does not represent the
Table 13. STANDARD DEVIATION FOR MODIFIED WESTVACO
EQUATION FOR SORPTION DATA AT 200°F WITH
NITRIC OXIDE FOR ACID LOADING ABOVE
0.01 GM ACID/GM CARBON
S02 Concentration,
ppm
2500
2000
1500
1000
500
Average z
-6.77
-6.82
-6.94
-7.09
-7.39
Standard
Deviation
0.200
0.238
0.176
0.161
0.162
86
-------
experimental data as well as the first model does. The
other models were also'evaluated by similar techniques,
but none represented the data as well as the first
Westvaco model. The evaluation of these other models is
included in Appendix J-l.
After selecting the Westvaco model as the basic form of
the rate expression and determining the order of the reac-
tion with respect to S02 concentration, further experiments
were conducted in order to expand the equation to include
effects of 02, H20, and NO concentrations and of temperature
The simplifying assumption was made that there is no first
order interaction of variables. This allowed a stepwise
determination of each variable's effect.
For example, to determine the dependency on 02 concentra-
tion, Equation (14) was modified to yield
dt
= K
°'40 v n
y02
(16)
vs
which reduces to Equation (14) if K
Equation (16) is arranged in the form
= 0.0105. If
ln
dXy/dt
1 -
_
] = In [K
0.40
+ n
= Z
(17)
it is seen that K and n can be determined by analyzing the
experimental data in the same manner described before.
Data from the three experiments used to find the 02
concentration dependency are given in Table 14 and the
graph of Z versus In yQ is shown in Figure 29.
Table 14. EXPERIMENTS TO DETERMINE 02 DEPENDENCY
Run
138
139
144
Temp.
°F
200
200
200
S02
ppm
2000
2000
2000
NO
ppm
150
150
150
H20
%
10
10
10
02
%
0.7
2.0
3.7
Average
Z
-7.67
-7.06
-6.67
Standard
Deviation
0,053
0.080
0.072
87
-------
T3
00 ^
00
Figure 29. Effect of C>2 on SC>2 sorption at 200°F with NO present
-1
-2
-3
-4
-5
-------
The values of K and n are obtained graphically, and the
order of the reaction with respect to oxygen is found to
be 0.63. The new K value is 0.1245 so that the rate
equation now becomes
dX
= 0.1245 y °'40 y °'63
dt S02 02
XVs
for
T = 200°F
NO = 150 ppm (18)
H20 = 10%
The H£0 dependency is found in the same manner. Data from
the pertinent experiments is given in Table 15, and the
graph of Z versus In ys02 ^s shown in Figure 30.
Table 15. EXPERIMENTS TO DETERMINE H20 DEPENDENCY
Run
160
139
144
Temp.
oF
200
200
200
S02
ppm
2000
2000
2000
NO
ppm
150
150
150
02
%
2
2
2
H20
%
5
10
15
Average
Z
-7.56
-7.06
-6.76
Standard
Deviation
0.164
0.080
0.031
The graph is a straight line, and the order of the reac-
tion with respect to H20 is 0.73. The new K is 0.667 and
the rate equation becomes
0.63 0.73
S02
Xvs
-) for
[~T •
[NO =
= 200°F
150 ppm
(19)
The NO dependency was treated in a similar manner, but in
this case the results showed a zero order dependency at
NO concentrations above 100 ppm. The data is presented
in Table 16 and Figure 31.
Table 16. EXPERIMENTS TO DETERMINE
NITRIC OXIDE DEPENDENCY
Run
162
163
139
161
Temp.
°F
200
200
200
200
S02
ppm
2000
2000
2000
2000
02
%
2
2
2
2
H20
%
10
10
10
10
NO
ppm
0
50
150
300
Average
Z
-5.97
-6.88
-r.oe
-7.07
Standard
Deviation
0.214
0.048
0.080
0.086
89
-------
Figure 30. Effect of H20 concentration on S02 sorption at 200°F with NO present
VO
X
•o
-1.5
-2.0
JL
-2.5
Ln
-3.0
-3.5
-------
Figure 31. Effect of NO concentration on S0£ sorption at 200°F
0.25
I-1 >
100
200
yNf) x.!0°, Vol. Fraction
300
400
-------
Because flue gas NO concentrations are typically higher
than 100 ppm, modification of the rate equation to extend
the valid range below 100 ppm was considered unnecessary,
so that the rate equation now is expressed properly as
°'40 °'63 °'73 /i XVx f r =200°F
ys02 y02 yH20 (1 -j-g f°rrc> > 100 pp»
Temperature Dependence -
In determining the temperature dependence of the reaction ,
the rate constant was assumed to be the only temperature
dependent term, and the Arrhenius equation was assumed to
give a satisfactory representation of the temperature
effect. The rate constant K is defined by the Arrhenius
equation as
K = k0 e"E/RT (21)
where ko = frequency factor
E = activation energy
RT = product of gas law constant and
temperature.
By taking the logarithm of both sides, one obtains
InK = In k0 -E/RT (22)
from which it is seen that a graph of In k versus 1/T
should yield a straight line with a slope of -E/RT and an
intercept of In ko.
The value of the rate constant was already specified at
200°F by rate equation (20). The rate donstant was also deter-
mined at 150 and 300°F from the experiments listed in
Table 17.
92
-------
Table 17. EXPERIMENTS TO DETERMINE THE
EFFECT OF TEMPERATURE ON S02
SORPTION
Run
148-G
153-G
149-G
150-G
140-G
139-G
141-G
142-G
164-G
143-G
151-G
152-G
158-G
Temperature
°F
150
150
150
150
200
200
200
200
200
200
300
300
300
1
Gas Composition
S02
ppm
2500
2500
1500
500
2500
2000
1500
1000
1000
500
2500
1500
500
NO
ppm
150
150
150
150
150
150
150
150
150
150
150
150
150
02
o/
JO
2
2
2
2
2
2
2
2
2
2
2
2
2
HgO
%
10
10
10
10
10
10
10
10
10
10
10
10
10
Average Z
-6.01
-6.25
-6.22
-6.76
-6.99
-7.06
-7.18
-7.39
-7.30
-7.60
-7.49
-8.08
-8.76
Standard
Deviation
0.688
0.569
0.659
0.469
0.056
0.080
0.025
0.086
0.019
0.038
0.149
0.211
0.323
Values of Z and the standard deviation of the data are
also included in the table. To calculate the rate con-
stants at 150 and 300°F, the Z values were plotted versus
In yso? as seen i-n Figure 32. Straight lines were drawn
through the points subject to the constraint that the order
of the reaction with respect to S02 was not a function of
temperature so that the slope was constant. Once the lines
were drawn, the rate constants could be found by inserting
the Z and In YS02 coordinates of any point on a line into
equation
In
(23)
and then solving for k. Table 18 gives the rate con-
stants at three different temperatures. Figure 33 is
Table 18. RATE CONSTANTS FOR THE WESTVACO MODEL
Temperature ,
op
150
200
300
Rate Constant,
gm acid/gms carbon/min.
1.476
0.667
0.244
93
-------
Figure 32. Effect of temperature on the Westvaco model with constant order
of reaction for S02
-9
-------
Figure 33. Rate constant as a function of temperature
for S02 sorption
2.0
1.0
0..8
0.6
0.4
0.2
0.1
1.1
1.2
1.3
1.4
1(103).
1.5
'R"
00-1
1.6
1.7
1.8
95
-------
the graph of In K versus 1/T from which the frequency
factor and the activation energy were obtained. The
results are
k0 = 1.59xlO~4 (24)
E = -1.079xl04 BTU/lb. mole (25)
or
E/R = -5520. (26)
The final form of the Westvaco rate equation is then
(27)
. e ^0.40 ^0.63 ^0.73
One additional qualification is placed on Equation (27) ;
it is strictly valid only for acid loadings above
0.01 gm H2S04/gm carbon, because rate data at lower acid
loadings was excluded from the calculations. Equation (27)
predicts a lower reaction rate than measured experimentally
at loadings below 0.01 gm H2S04/gm C.
Success of Westvaco Equation in Fitting
Differential Rate Data
Analysis of the Westvaco rate equation showed that it
represents the experimental data very well at 200°F but
less satisfactorily at 150 and 300°F. The results of cal-
culations comparing the Westvaco equation predictions with
the differential rate data are tabulated in Appendix A-3.
Table 19 is a summary of these results showing the
average percent difference between the predicted and
experimental values. The overall average difference for
the twelve 200°F runs is only 7%, but the deviation is 3770
at 150°F and 23% at 300°F.
ro
96
-------
Table 19. DEVIATION OF WESTVACO MODEL FROM
DIFFERENTIAL RATE DATA FOR ACID
LOADINGS ABOVE 0.01 GM ACID/
GM CARBON
Run
138
139
140
141
142
143
144
145
157
160
161
164
Temperature
OF
200
ii
ti
ii
ii
ii
it
ti
ii
ii
ii
ti
AVERAGE 200
148
149
150
153
AVER
151
152
158
AVER
150
ii
ii •
ii
AGE 150
300
ii
it
AGE 300
Average 70
Deviation
4.5
"1.7
8.1
5.6
1.9
3.9
6.6
5.0
4.8
13.9
8.0
12.3
6.9
47.7
40.0
20.6
39.2
36.9
36.8
15.9
17^4
23.4
Another comparison between predicted and experimental
values is presented in Figure 34, which is a plot of the
average Z value for each differential rate experiment
versus the term
5520/T + 0.40 In ys02 + 0.63 In y02 +0.73 In yH20
(28)
This term is the natural logarithm of the variable part
of the right side of the Westvaco equation in the form
dXv/dt
l-Xv/0.38
P-40 0.63
- 1.59(10-4) e5520/T y -0 ^
.73
(29)
97
-------
Figure 34.
-4.5
-5.0-
-5.5-
-6.0-
-6.5-
,-7.0-
-7.5-
-8.0-
Comparison of the Westvaco Model A to experimental
S0£ sorption on activated carbon in differential
rate apparatus
• 150°F
• 200°F
A 300°F
Represents
Error or
Two Standard
Deviations
-8.5-
Westvaco Equation, Model A;
5520 •
^ » (1.59)(10-4)e
NO Concentration i 100 PPM
dXdt
-9.5-
-9.5
0.4
0.8
1.2
98
1.6
2.0
2.4
2.8
-------
The straight line in Figure 34 represents the Westvaco
equation as the natural log of both sides of the above
expression. The experimental Z values are shown with an
error band representing twice the standard deviation of
the data. The Z values and standard deviations are in
Table 17.
The standard deviation provides a useful measure of the
data's probable accuracy. There is a 95% probability
that the true Z value lies within an error band'that is
two standard deviations wide, so that a small standard
deviation indicates high accuracy of the data. It is
seen from Table 17 and Figure 34 that the standard
deviation of the data is much smaller at 200°F than at
150 or 300°F. As a percentage of the average Z value,
the standard deviation averages 0.7% at 200°F, 2.8% at
300°F, and 9.4% at 150°F, so that statistically, the
200°F data is excellent but the 150 and 300°F data is
not as good.
Sorbet Design Study - Effects on Size -
Based on the Westvaco S02 sorption rate expression and two
important reactor model assumptions, a procedure was
developed for calculating the required size of a multistage
fluid bed S02 sorption reactor. Using this procedure a
sorber design study was made encompassing the effects of
seven important variables. The complete derivation of the
design equations is in Appendix C.
The two important reactor model assumptions concerned the
flow characteristics of the gas and carbon phases within
the reactor. It was assumed that the carbon phase was well-
mixed on each stage, so that the carbon beds represented a
series of backmix reactors. Plug flow was assumed for the
gas phase.
The design equations used in the reactor size calculation
procedure are:
(30)
n.6 ... .0.6 . ,, ^ 0^8' X02 1H20
' (YS02)j'6 - 3.687(10-4)
" XVi 0.63 0.73 2
(l-TT^o") Y09 YHOO hD
5520/T
qT
XVj+1 = XVj - (3^) (g) l(Yqno). - (YSn,)<4.il <31)
99
-------
The first equation (30) was the acid loading and S02 concen-
tration on a given stage to calculate the S02 concentration
on the next stage. The second equation (31) is a material
balance relationship which is used to calculate the acid
loading on the new stage from the other variables. Using
these equations, plate-to-plate calculations are made
beginning with the bottom stage of the reactor and pro-
ceeding until the end conditions at the top of the reactor
are reached. This determines the required number of stages
and hence the reactor size.
A computer program was written to facilitate the calcula-
tional procedure. A listing of the program is an Appendix
E. The effects of the following parameters on reactor
size were studied:
1) Inlet S02 concentration (2,000 ppm)
2) Inlet 02 concentration (0.046 mole fraction)
3) Inlet H20 concentration (0.13 mole fraction)
4) Outlet acid loading on carbon (0.184 Ib. acid/
Ib. carbon)
5) Temperature (200°F)
6) Linear Gas Velocity (4 ft./sec.)
7) Carbon bed depth/stage (9 inches).
Each parameter was varied separately while holding all
others constant at the base conditions given in parentheses
in the preceding list. The results of the design study are
in Appendix A-9.
Attempts To Further Improve Model
by Multiple Regression Analyses
Multiple regression analyses were carried out in an attempt
to improve the Westvaco model's representation of the data
at 150 and 300°F. Based on the following form of the rate
equation,
dt
= k e
-E/RT m
100
-------
The multiple regression analyses were used to determine
the best values of the constants k, E/R, m, p, and q for
the two cases where the order of'reaction relative to the
acid concentration in the carbon phase, constant A, was
allowed to vary in one case and equal to 1.0 in the other
case. Both cases were processed by computer using a
program written by IBM. A listing of the program is in
Appendix E-l. The Westvaco equation, (27), was designated
Model A, and the multiple regression variations were desig-
nated Models B and C. For Model B with A = 1.0 the multi-
ple regression analysis yielded the following expression:
- 3.32(10-*) e»"/T y°-53 y°-62 ,°'73 U-> (33)
Allowing the constant A to seek its statistical value
resulted in the following expression for Model C:
. 3.42(10-4) e5732/T ^0.53 ^0.62 ^0.73 ^_J^I.B3 (34)
U * jo
Calculations were carried out on the computer to compare
Models B and C with the differential rate data and the
results are included in Appendix A-7. Table 20 is a
summary of these results showing the average percent
difference between the predicted and experimental values.
Figures 34, 35, and 36 provide a graphical comparison
of the three models.- The results in Table 20 show that
Model B fits the data slightly better than Model A at
300°F but not as well at 150°F, while Model C provides a
significant improvement at both 150 and 300°F. At 200°F
Models B and C are both worse than Model A, as expected.
Clearly, Model B does not offer any real improvement at
150°F to warrant its substitution for Model A near this
temperature and perhaps over the 150 to 175°F range, but
Model A remains the best choice over the remainder of the
temperature-range from 175 to 300°F, because of its
superior performance at 200°F.
101
-------
Table 20. DEVIATION OF MULTIPLE REGRESSION MODELS
FROM DIFFERENTIAL RATE DATA FOR ACID
LOADINGS ABOVE 0.01 GM ACID/GM CARBON
Run
138
139
140
141
142
143
144
145
157
160
163
164
AVER
148
149
150
153
' AVER/
151
152
158
AVER;
Temperature
'oF
200
it
ii
ii
it
ii
n
ii
n
n
n
n
\GE 200
150
n
u
n
\GE 150
300
u
n
\GE 300
Averayo Percent: Dew 1. 'it ion
Model A
4.5
7.7
8.1
5.6
1.9
3.9
6.6
5.0'
4.8
13.9
8.0
12.3
6.9
47.7
40.0
20.6
39.2
36.9
36.8
15.9
17.4
23.4
Model H | Model C
13.6
17.6
23.1
16.3
8.1
4.6
13.1
8.0
18.1
24.1
20.5 :
18.4
15.5
51.3
41.0
20.7
47.6
40.2
25.9
16.1
19 . 4
20.5
22.8
25.9
30.7 •
19.5
13.9
11.1
24.5
14.6
21.3
26.8
24.5
26.1
21.8
33.5
25.8
13.2
22.1
23 . 6
27.6
8.2
5.0
13.6
102
-------
Figure 35
-4.5
Comparison of the Westvaco Model B to experimental
S02 sorption on activated carbon in differential
rate apparatus
-5.0-
-5.5-
-6.0-
-6.5-
-7.0-
-7.5-
-8.0-
-8.5-
-9.0-
-9.5
150°F
200°F
300° F
Represents
Error of .
Two Standard
Deviations
.0 I
-0.4
dX
Westvaco Equation, Model B:
5416
= (3.32)(10'4)e "~ (1
Xv v ., 0.48 .. 0.62 .. 0.73
M
H
NO Concentration > 100 PPM
dXv/dt
0
I
0.4
0.8
i
1.2
i
1.6
2.f
103
-------
Figure 36.
Comparison of the Westvaco Model C to experimental
S02 sorption on activated carbon in differential
rate apparatus
-4.5-
-5.0 -
-5.5 -
-6.0 -
-6.5 -
-7.0 -
-7.5 -
-8.0 -
-8.5 -
-9.0-
150°F
200°F
300°F
Represents
Error of
Two Standard
Deviations
(3.42) (10-
Westvaco Equation, Model C:
5732
1.83 „ 0.53 „ 0.62 „ 0.73
NO Concentration i 100 PPM
Y = Ln
dXy/dt
"
.83
5732 0.53 0.62 0.73
Z = -j- * Ln [YS02 Y02 YH20 ]
I
-0.4
0.4
i
0.8
Z
1.2
i
1.6
2.0
104
-------
In the Westvaco equation (Model A) the rate is proportional
to (l-Xv/0.38) or equivalently, to (0.38-XV) where
0 < Xv < 0.38. This model is valid if the plot of dXv/dt
versus Xv is a straight line. Examination of the rate
curves in Figure 28 reveals that dXv/dt versus Xv
approaches a straight line at both 200 and 300°F, but not
at 150°F. Model A should, therefore, give satisfactory
results at the higher temperatures, but a different model
may necessarily be needed at 150°F. In Model C the rate
is proportional to (1 - Xv/0.38)1-83 which provides a better
fit at 150°F.
From Figure 37 it is seen that the 200 and 300°F curves
are not linear until the acid loading is above 1 gin/100
gms carbon. This explains why the data at acid loadings
less than 1.0 gm/100 gms carbon was excluded from the rate
model computations. It also indicates that application of
the model at low acid loadings would cause a significant
underestimation of the sorption rate.
Fluid Bed S02 Sorption Experiments -
This section covers the pre-integral fluid bed sorption
work that was done in the 6" and 18" diameter S02 sorbers.
Objectives included demonstration of satisfactory mechani-
cal operation and process performance, gathering of addi-
tional information on reaction rate and process character-
istics, and completion of necessary developmental work in
preparation for integral pilot plant operation.
6" -Diameter Sorber -
Operation of a 6" diameter, eight stage fluid bed sorber
provided the first test of the S02 sorption step under
actual flue gas conditions of a slipstream from a 50 MW
oil fired boiler. Successful demonstration of the 6"
unit was an important .achievement in development of the
process. Operation of the 6" sorber provided valuable
reaction rate data for comparison with the Westvaco rate
model.
The 6" sorber runs offered the first opportunities to test
the rate model in predicting the performance of an actual
flue gas system. The data from six runs was processed to
give average values of S02 concentration, acid loading and
reaction rate on each stage of the reactor. These average
values are given with the run conditions in Appendix A-5.
The yso2 an<* xv values were then used in the Westvaco rate
equation to calculate predicted rates for comparison with
the experimental rates. The results of the comparison are
presented in Table 21. The predicted rates average
about 147o below the experimental rates.
105
-------
Figure 37
Differential S02 sorption rate versus H2S04 loading for an
S02 concentration of 2500 ppm at 150, 200, and 300°F
4.0
o
cr>
e
o
60
o
CO
CM
EC
C
O
o
cs
<4-l
O
V
3.0
2.0
1.0'
Sulfuric Acid Loading, gms. H2S04/100 gms. C
-------
Table 21. COMPARISON OF RATES FROM 6" DIAMETER SORBER
TO RATES CALCULATED FROM THE WESTVACO MODEL
Run*
SA-21
SA-23
SA-26
SA-22
„
SA-25
SA-24
Stage
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3,
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
Temperature,
°F
200
200
200
200
200
200
gms Acid/gm Carbon
0.0213
0.0426
0.0662
0.0920
0.1178
0.1480
0.1817
0.0087
0.0260
0.4777
0.0703
0.0954
0.1241
0.1562
0.0170
0.0377
0.0583
0.0817
0.1068
0.1346
0.1660
0.0308
0.0607
0.0888
0.1214
0.1531
0.1875
0.2274
0.0140
0.0307
0.0464
0.0657
0.0832
0.1043
0.1279
0.0241
0.0482
0.0738
0.1002
0.1282
0.1530
0.1981
S02
Concentration,
ppm
251
425
607
808
1,018
1,247
1,507
46
183
388
621
872
1,155
1,475
132
324
534
758
1 ,005 •
1,274
1,576
1,306
1,612
1,904
2,210
2,535
2,868
3,242
484
644
813
996
1,187
1,388
1,621
1,219
1,502
1,795
2,101
2,420
2,731
3,142
Sorption Rate x 10J,
#Acid/#Carbon-Min.
6 Inch**
0.76
0.76
0.84
0.92
, 0.92
1.08
1.20
0.35
0.71
0.89
0.92
1.03
1.17
1.31
0.72
0.88
0.88
0.99
1.07
1.18
1.33
1.24
1.21
1.13
1.32
1.28
1.39
1.61
0.56
0.67
0.63
0.77
0.70
0.84
0.95
0.96
0.96
1.02
1.05
1.11
0.99
1.79
Model ***
0.72
0.84
0.90
0.93
0.93
0.89
0.82
0.38
0.63
0.80
0.90
0.95
0.95
0.92
0.58
0.78
0.89
0.95
0.98
0.96
0.91
1.36
1.36
1.32
1.25
1.16
1.03
0.86
0.52
0.56
0.59
0.60
0.61
0.60
0.59
0.79
0.80
0.80
0.78
0.74
0.70
0.59
*Inlet gas compositions for the runs are given in Appendix B.
**Rate data from 6" diameter sorber.
***Rate data calculated from Westvaco Model, Equation (27).
107
-------
A graphical comparison of the rate model to the fluid bed
rate data is given in Figure 38. Average experimental
reaction rates are plotted versus S02 concentration, with
the rate equation superimposed for comparison. The reac-
tion rates in Figure 38 were obtained from the actual
rates by dividing by all variable terms in the equation
except yso?- This mathematical comparison assumes that
all differences then arise with the S02 concentration
term in the model, which is not necessarily true but if
large discrepancies had occurred, then additional analysis
would have been required. The graph shows a reasonably
good fit of the data by the model.
The rate model was also tested by using it in a design
procedure to calculate the theoretical number of stages
required for'each run. The detailed derivation of the
design procedure is given in Appendix C-l. The results
of this comparison are shown in Table 22. The average
predicted number of stages is 9.9 compared to the actual
8, which is an over-estimation of 2570. The rate model,
therefore, can be expected to yield conservative esti-
mates of reactor size.
Table 22. COMPARISON OF PREDICTED NUMBER OF
STAGES TO ACTUAL NUMBER FOR 6"
SORBER RUNS
Run
SA-21
SA-23
SA-26
SA-22
SA-24
SA-25
Temperature ,
°F
200
200
200
200
250
250
Inlet S02
Cone . ,
ppm
2000
2000
2000
4000
2000
4000
Actual
No. of
Stages
8
8
8
8
8
8
AVERAGE : 8
Predicted
No. of Stages
Model A
9.2
8.5
10.1
9.3
11.6
10.4
9.9
\
108
-------
Figure 38
6" sorber data - plot of corrected sorber rate
using stagewise Westvaco Model A vs. S02
concentration showing curve predicted from
differential bed studies
100+
: .10-
CM
X
U-l
LO
•
r~i
O
M
e>
f>
ai
Prediction by Stagewise
Westvaco Model A
O 200°F, 2000 PPM
O 250°F, 2000 PPM
• 200°F, 4000 PPM S02
• 250°F, 4000 PPM S02
100
1,000
S02 Concentration, Volume Fraction
100,000
109
-------
18" Diameter Sorber -
Operation of an 18" diameter sorber was the next step in
development and provided additional process information.
In the operation of an S02 sorber, the flue gas typically
enters at300°Fto350°F. Before the gas was cooled, the
863 at a concentration of about 50 ppm was removed to pre-
vent corrosion which would result if the 803 were allowed
to condense. 803 removal is accomplished in the first
stage of the reactor, and then the gas is cooled to
increase the rate of 802 removal in the remaining stages.
Based on prior development work of using water spray cool-
ing in the production of carbon in fluidized bed reactors,
a process change was made in 802 sorption, namely, the
substitution of direct flue gas cooling by means of water
injection in place of indirect heat exchange. .In the 6"
diameter sorber and in the initial design of the 18"
diameter unit, the flue gas after 803 removal in the
bottom stage of the reactor was cooled by indirect heat
exchange before carbon/flue gas contact in the second and
subsequent stages. The heat exchanger presented tempera-
variation problems and was also an expensive item in the
capital investment estimates of the Westvaco Process. These
reasons provide^ the impetus for development of a direct
cooling method. The success of the new cooling method,
water spray injection, was a significant break-
through which simplified operation and substantially
reduced capital cost estimates. The initial direct cool-
ing tests, Table 23 were made in a one-stage, 18"
diameter unit. The gas temperature was lowered from 300
to 150°F with no apparent operating problems.
After demonstrating satisfactory water spray operation
in the one stage unit, modifications to the existing 5
stage, 18" diameter fluid bed sorber at No. 6 power boiler
were made to replace the indirect heat exchanger with a
direct water spray cooling system. Prior to operation
with flue gas, test runs SC-7, -8 and -11 were made to
check out the equipment using air and an air-steam mixture
which simulated the flue gas moisture conditions. No
.operating problems were encountered in these tests with
cooling down to 150°F.
Direct water spray cooling of actual flue gas was first
demonstrated for use in the Westvaco Process in Runs '
SC-14, -16, -16A and -16B. These runs lasted 6 to 10
hours each with no difficulties in equipment operation.
In an extended demonstration run, SA-32, mechanical
problems were encountered that were unrelated to the
water spray system, but despite the problems a total
110
-------
Table 23. WATER SPRAY COOLING TESTS MADE IN PILOT FLUID BED REACTORS
WITH SIMULATED AND ACTUAL FLUE GAS
RUN
NUMBER
SC-1
SC-2
SC-3
SC-4
SC-5
SC-7
SC-8
SC-11
SC-14
SC-15
SA-16A
SC-16B
SA-32
UNIT USED
18" Dia.
Fluid Bed
One-Stage
Unit
18" Dia.
Fluid Bed
S02 Sorber
NUMBER
OF
STAGES
1
1
1
1
1
4
4
4
4
4
4
4
4
TOTAL
CARBON
BED
HEIGHT
(SETTLED)
INCHES
6
6
6
7.5
9
15.5**
15.5**
15.5**
15.5**
15.5**
15.5**
15.5**
15.5**
INLET
GAS
TEMP.
»F
300
it
II
380
330
327
310
300
310
310
315
5TBGT
TEMPERATURE
°F
11
200
178
149
152
149
350
317
310
295
280
290
290
295
n
...
242
140
190
175
140
150
150
175
#3
---
250
150
183
180
155
160
160
180
#4
...
240
150
175
175
170
160
160
180
LINEAR
GAS
VELOC.
FT/SEC
2.9
3.1
3.1
3.1
3.1
3.1
2.6
2.9
3.0*
3.1*
3.1*
3.1*
3.1*
H20
SPRAY
RATE
I/HR.
23.4
31.2
43.0
33.2
27.4
10 .
37.2
44.7
29.8
45
40.2
40.2
34
TOTAL
GAS
FLOW
RAT£,
CFH ?
70°F
14,070
14,500
16,500
16,500
16,500
14,830
14,430
1 5,2'80
15,250
16,230
16,200
16,200
16,767
INLET
02
ANAL,
VOL.
%
10
It
II
Air
II
H
3.1
3.1
3.1
3.1
3.1
INLET
H20
ANAI
VOL.
%
10
II
it
~4
-4
-12
12.7
12.7
12.7
12.7
12.9
SULFUR DIOXIDE
ANALYSIS, PPM
INLET
995
880
1,125
1,100
2,200
STAGE
1
925
1,070
2,088
STAGE
2
540
560
1,511
STAGE
3
360
460
1,008
OUTLET
70
170
160
55
522
CARBON
PRECURSOR
SKID NO.
AND
CHAS. NO.
WV-W
II
II
II
It
96257, C-71-98
II
II
H
II
96324, C-71-23
96283, C-71 -100
CARBON ACID
10ADING
LBS. SOz/
100 LBS. C
INLET
...
0
0
0
0
4.1
8.7
0
0
OUTLET
...
0
0
0
10.3
13.7
20.9
14.2
11.9
CARBON
DATT
TC/HR.
0
0
0
0
0
23
23
23
23.2
19.2
20.2
19.9
32.9
*Based on average temperature of Stages 2, 3 and 4.
"Excluding carbon on bottom stage at about 300°F (about 3 inches carbon).
-------
operating time of 29 hours was achieved, and the water
spray cooling system subsequently was declared a success.
The advantages of using water spray cooling instead of
indirect heat exchange are summarized as follows:
1) A savings on the order of 2570 in the capital
investment estimated for the Westvaco Process
f
2) Temperatures as low as 150°F can be achieved
which would be difficult using a heat exchanger
because of moisture condensation at cooling
surfaces or very large surface areas for heat
transfer.
3) The small amount of moisture from the water
spray increases the rate of S02 removal.
4) Better control of the column temperature has
been realized with water spray cooling.
Additional runs were made in the 18" diameter sorber in
order to produce acid loaded carbon needed for the sulfur
generation and sulfur stripping experiments. These runs
contributed valuable information about the unit's operat-
ing characteristics and they also led to early recognition
of problem areas, thereby providing a basis for necessary
corrective measures and improvements to increase opera-
tional reliability.
The summary for a typical sorption run, SA-32, is given
in Figure 39. Temperature and S02 concentration pro-
files through the column are shown. Average column
temperature for Stages 2 to 5 was 168°F. The S02 removal
was 94%, and the material balance agreed within 67o.
Effect of Fly Ash on Carbon S02 Activity -
Application of the Westvaco S02 Process in the flue gas
desulfurization area would involve both coal and oil
fired boilers. One difference between the two is the
significantly higher fly ash concentration in the flue
gas from coal fired boilers. Pilot plant S02 sorption
work has been conducted using flue gas from an oil fired
boiler only, so that the results are not necessarily
applicable under coal fired boiler conditions.
Of greatest concern is the possibility that prolonged
exposure of activated carbon to high fly ash concentra-
tions may reduce the carbon's S02 activity. This was
112
-------
Figure 39. Summary for flue gas run (run SA-34);
18" diameter S02 sorber; water sprays
to control temperature
38.1 Ibs. C/hr.
165°F
174°F
171°F
H20 Pxate,
(AUTO Control) 162op
263°F
290°F
Flue Gas + S02
2,050 PPM of S02
14,800 CFH 9 70° F
I
4'.^:+&-:ili:&tl:
*
^^(fS^^ri
^Tfj:-!-':^' '••• '•• \ ' Jv?-' '•.
_ .*•
J^-iv^^^i^
m^m
"
i
j3^
Stage No. 5
Stage No. 4
Stage No 3
Stage No. 2
Stage No. 1
^•Outlet S02 Cone. = 125 PPM
Cone.
= 725 PPM
-^ = 1,125 PPM
Ccr.c.
S02
Cone.
Cone.
1,550 PPM
1,950 PPM
0.135 Ib. S02/lb. C
37.3 Ibs. C/hr.
AVG. TEMP.: 168°F
% S02 REMOVAL: 94
MAT'L. BALANCE
S02 IN: 5.0 Ibs./hr.
S02 OUT: 5.3 Ibs./hr.
113
-------
investigated in a bench scale experiment in which recycled
carbon was exposed to fly ash laden air for a period of
5 days. The ash content of the carbon was measured each
day and the S02 activity was measured before and after.
The ash content did not increase above the initial level
corresponding to the inherent ash content of the carbon,
which is typically in the range of 4-1/2 to 5% by weight.
S02 activity measurements showed only a 6% drop in
activity which is within experimental error and, therefore,
not really significant. This result is encouraging and
indicates that the carbon should perform satisfactorily
under coal fired boiler conditions, Additional details
of the fly ash exposure experiment are included in
Appendix J-2.
5.2.2 Sulfuric Acid Conversion to Sulfur
Sulfuric acid conversion to sulfur is the second step in
the Westvaco Process, and involves the reaction of sorbed
H2S04 with H2S to form sorbed elemental sulfur and water
vapor. The overall reaction is:
H2S04 + 3 H2S Activated ^ 4 S + 4 H20 + ( 6 )
Carbon
The main objective of sulfur generation studies were:
1) to obtain enough information to properly design and
construct an acid conversion reactor and optimize its
operating conditions, and 2) to demonstrate satisfactory
operation of a pilot acid converter incorporated into
the integrated S02 pilot plant. An important requirement
in satisfactory demonstration of a pilot reactor was high
degree of conversion of both reactants.
The reaction was studied extensively in various bench and
pilot scale equipment, including:
1) Fixed bed reactors - 1" and 1.6" diameters
2) One stage, 4" diameter fluid bed batch reactor
.3) 8 stage, 4" diameter fluid bed glass column
counter current reactor
4) 8 stage, 4" diameter fluid bed reactor (eventu-
ally used as sulfur stripper/H2S generator in
integral pilot plant)
5) Moving bed reactors - 1.5" and 8" diameters
(8"0 eventually used as sulfur generator in
integral pilot plant).
114
-------
A rate expression was found to represent kinetic data
measured in a one stage fluid bed reactor. The kinetic
data^was incorporated into a reactor design procedure.
The important variables in the reaction rate are
temperature, acid concentration on carbon, and the gas
concentrations of H2S and H20.
Experimental studies indicated that the required high
conversions of both H2S and H2S04 were not possible at the
space velocities obtainable in a 6"0 fluid bed reactor
that initially was considered for use in the integrated
pilot plant. Subsequent tests in a 1.5" diameter bench
scale moving bed reactor indicated that a moving bed
was better suited for integral operation of the pilot
plant. A pilot scale 8" diameter moving bed unit was then
designed and tested; and satisfactory performance was
demonstrated.
It should be recognized that the use of a moving bed
sulfur generator in the integrated pilot plant did not
alter plans to specify fluid bed reactors in larger scale
applications. The rate expression was developed from fluid
bed data obtained in a set of batch differential experi-
ments, and to the extent that the model accurately repre-
sents the reaction, it is entirely valid for design of
either reactor type.
Fixed Bed Acid Conversion Experiments -
Fixed bed studies were carried out in 1" and 1.6" diameter
reactors. The more important results are presented here,
and the complete results are given in Appendix J-3. The
purpose of experiments in the 1" fixed bed was to determine
the effect of linear gas velocity on the reaction rate.
Runs were made at linear velocities from 0.03 to 0.14
ft./sec., as shown in Figure 40. It was found that the
reaction rate increased with increasing linear velocity
up to a velocity of about 0.12 ft./sec, Further
increases above 0.12 ft./sec. did not raise the reaction
rate. This result indicates that external diffusion does
not limit the reaction at linear velocities above 0.12
ft./sec. The data point shown above the line on Figure
40 represents a temperature of 395°F, about 75°F above
the other data points.
The objective of experiments in the 1.6" diameter fixed bed
was to study the effects of space velocity on reactor
performance. Runs were made at constant reactor volume at
space velocities from 100 to 1,000 hr". ~i. The inlet gas
composition was held constant throughout each run, and
the outlet H2S concentration was monitored continuously to
determine the time at which H2S breakthrough occurred.
The results showed almost immediate H2S breakthrough at
115
-------
Figure 40. Effect of linear gas velocity on rate of
sulfuric acid decomposition
-1.2
I
o
§>
X
•o
I
•>
o
in
O
o
TJ
to
a:
0.02 0.04 0.06 0.08 0.10
Linear Gas Velocity, ft./sec.
0.12
0.14
116
-------
space velocities above 300 hr.'1, but at 100 hr."1 break-
through did not occur until 130 minutes. Overshadowing
these results, however, were the implications of the
temperature rise that occurred in the bed due to the
exothermic (-65 kcal/mole acid) heat of reaction. The
same problem occurred to a lesser degree in the 1" fixed
bed runs. A temperature rise of 135°F was recorded in 3
out of 8 runs in the 1.6" diameter bed. The most import-
ant result of these experiments, therefore, was to
clearly demonstrate the unsuitability of a fixed bed
reactor in kinetic studies of the sulfur generation
reaction. The temperature control which is necessary"to
obtain useful data could not be achieved in a fixed bed
unit.
Fluid Bed Rate Studies -
The effects of temperature and H2S concentration on the
rate of reaction were studied in an 8 stage, 4" diameter
fluidized bed reactor constructed from flanged sections of
glass pipe. Each stage had 4 inch overflow weirs for a
typical carbon bed depth of 2.5 inches/stage, so that the
total settled bed depth was typically 20 inches of carbon.
Experiments were run at temperatures of 250, 275, 300, 325
and 350°F, and at inlet H2S concentrations from 127o to
42%. Average steady state conditions for each run are
shown in Table 24.
The overall rate of sulfuric acid decomposition was calcu-
lated for each run based on both gas and carbon analyses
for sulfur compounds. A comparison of the results given
in Table 25 shows that reaction rates calculated from
carbon analyses are about 2070 lower than those determined
from gas analyses. In the following discussion of results,
the rates based on carbon analyses are used because they
are considered more reliable and are more conservative.
The rate of acid decomposition increases steadily with
increasing temperature, rising from 0.007 Ib. acid/lb.
C-min. at 250°F to 0.20 at 325°F. The rate levels out
around 325°F and is about the same at 350°F as at 325°F.
Figure 41 shows a plot of rate versus temperature.
An important temperature dependent effect begins to appear
at 300°F which effectively places an upper limit of 325 to
350°F on the practical operating temperature range. This
refers to the evolution of reactant acid in the form of
S02 in the outlet gas. Although the exact mechanism is
uncertain, the net result is that acid converts to S02 at
the expense of complete acid reduction to elemental sulfur.
117
-------
Table 24. EXPERIMENTAL CONDITIONS AND RESULTS FOR SULFUR GENERATION EXPERIMENTS
IN AN EIGHT STAGE, 4" DIAMETER FLUIDIZED BED REGENERATOR
Run Number
and Purpose
Estimation of
Requirements
f cr Total Acid
Conversion
Effect of 4
Temperature
Effect of
TT ft
HgS i
Cone.
fsG-27
SG-28
/SG-29
1 SG-30
[sG-31****
/^G-33
1 SG-3^
) SG-35
{ SG-36
JSG-37
(SG-3&
SG-39
SG-^0
SQ-kl
Temperature
°P
300
300
300
300
300
300
250
275
325
350
350
•325
325
325
Inlet Gas
Flow Rate,
N2
285
302
302
33^
320
283
SOU
29!+
2?6
267
267
. 276
236
362
H2S
96
108
108
72
108
130
131
126
118
Ilk
Ilk
119
158
37
Gas Composition*,
Volume %
Inlet
H2S
2k
27
.27.2
17.0
25.6
27-7
30. k
31.5
30.8
30
32.0
31.8
42.0
12.0
Outlet
H2S
10
17.8
11.2
7.6
13.5
15.7
2k. k
20
6.0
8.8
6.0
12.0
3.1
SO?
0
0
0.12
0.25
0.12
0
0
0
0.35
2.6
1.3
2.0
1.6
1.6
Total Solid Flow**
Rate, RTI
Ibs./hr.
53
61
58
56
61
61
6k
6k
56
5k
53
55
76
18
Sulfur Analysis
on Inlet Solid, Wt. -":
Inlet***
Psi
5.6
5.6
5.0
5.1
14.5
6.1
6.0
6.1
6.1
6.2
6.k
6.0
5.5
5.7
Outlet
Pso
13.0
11.5
12.7
11.0
22.4
14.2
9-k
11.3
18.7
18.0
18.3
18.0
17.0
ik.k
*As determined by gas chromatograph.
**Total solid flow rate of the carbon plus its sorbed sulfur compounds.
***Except for SG-31 the per cent sulfur is present only as sorbed acid.
****Inlet material for this run was a blend of the partially converted loaded carbon from
Runs SG-28 and -29 which originally had an average acid loading of approximately .19 # HgSOl^ C.
-------
Table 25. OVERALL RATES OF ACID DECOMPOSITION AND
CONVERSION TO SULFUR FOR THE REACTION
3 H2S + H2S04 —* 4 S + 4 H20 IN AN
EIGHT STAGE, 4" DIAMETER REGENERATOR
Run
No.
SG-27
S6-28
SG-29
SG-30
SG-31
SG-33
SG-34
SG-35
SG-36
SG-37
SG-38
SG-39
SG-40
SG-41
Temp.
°F
300
300
300
300
300
300
250
275
325
350 '
350
325
325
325
Inlet
Acid
Loading
#H2SOzi/#C
0.21
0.21
0.18
0.19
****
0.24
0.23
0.24
0.24
0.24
0.25
0.23
0.21
0.22
Inlet
H2S
%
24
27
27.2
17
25.6
27.7
30.4
31.5
30.8
30
32
31.8
42
12
Outlet
S02
%
-
0
0
0.12
0.25
0.12
0
0
0
0.35
2.6
1.3
2.0
1.6
1.6
Conv.
to
S***,
%
I • r- 1 i ii
43
34
51
39
115
46
18
24
62
66
61
66
65
50
Overall Rate of Reac.,
# H2S04/#C-min.
Gas
Analysis*
•••^" [••^•••^•^^•^•^•^^^^^^.^^^^
1. 52(10-2)
1-05(10-2)
1.87(10-2}
1.08(10-2)
1.40(10-2)
1.39(10-2)
0.71(10-2)
1.33(10-2)
2.83(10-2)
2.54(10-2)
2.95(10-2)
3.24(10 )
1.04(10 )
Carbon
Analysis**
1.2(10-2)
1.1(10-2)
1.34(10-2)
1.06(10-2)
1.62(10-2)
0.66(10-2)
0.88(10-2)
2.0(10-2)
2.08(10-2)
1.96(10-2)
2.04(10-2)
2.56(10"2)
0.48(10-2)
*Based on gas analysis using a chromatograph.
**Based on carbon analysis using a combustion analysis technique.
***Calculated from experimental data.
****Inlet material for this run was a blend of partially converted
loaded carbon from Runs SG-28 and -29 which originally had an
average acid loading of approximately .19 Ib. acid/lb. C.
119
-------
Figure 41. Effect of temperature on the rate of conversion
of sorbed sulfuric acid to elemental sulfur
Inlet HgS = 31 vol % (AVG)
Inlet Acid = 0.24 Ib. acid/lb.
Loading (AVG)
250
275 300 325
Reactor Tempera tvre, °F
120
-------
The following two-step reaction sequence is believed to be
the most probable mechanism for conversion of acid to
elemental sulfur:
H2S04 + H2S *» S02 + S + 2 H20 (35)
S02 + 2 H2S +. 3 S + 2 H20 (36)
H2S04 + 3 H2S +• 4 S + 2 H20 (37)
Because SC>2 is an intermediate in these reactions, its
presence in the outlet gas could be expected under at
least some reaction conditions.
The results in Table 25 show that the outlet gas S02
concentration increases above 300°F, rising from about
0.15% at 300°F to 27, at 350°F. The 2% figure represents
about a 20% conversion of acid to S02- The results indi-
cate that operating temperature above 325°F should be
avoided in order to minimize S02 formation.
The reaction rate also increases with increasing H2S
concentration, as seen most clearly by inspecting the
results of SG-39, -40 and -41. Further discussion of
the H2S concentration effect is postponed to the next
section on rate model development.
These experiments also provided an average reaction rate
for anticipated process conditions to be used in prelimi-
nary design work. To obtain the average rate over a
wide set of operating conditions it was necessary to
obtain complete acid conversion, and this was accom-
plished by passing the carbon through the reactor twice.
The partially reacted carbon product from Runs SG-28 and
-29 was combined and fed to the reactor again in SG-31.
The results of these experiments indicate an overall
average reaction rate to complete conversion under antici-
pated conditions of about 0.01 Ib. acid/lb. C-min.
Rate Model Development -
A series of differential batch experiments was carried
out in a 4" diameter fluidized bed reactor in order to
obtain data to develop a rate expression. In each run a
small batch of acid loaded carbon was fluidized with
reactant gas at constant conditions of temperature and
121
-------
gas concentration. Samples of carbon were removed at
intervals for analysis, thereby providing reaction data
as a function of time for total sulfur content and acid
loading. Space velocity was sufficient that the inlet
and outlet reactant gas concentrations were essentially
identical. This justified the differential reactor
assumption.
The variables studied were temperature, acid loading,
sulfur loading, and the gas concentrations of H2S and H20
The conditions for each run are presented in Table 26.
Temperatures of 250, 300 and 325°F were tested, and H2S
Table 26. EXPERIMENTAL CONDITIONS FOR DIFFERENTIAL
SULFUR GENERATION RUNS
Carbon Used:
Virgin Procursor:
Residual S on Virgin Carbon:
Carbon Density:
Linear Gas Velocity:
Diluent Gas:
SA-27-A (335-365 min.)
Skid 96298
0.6 wt. $
1*9.7 lbs./ft.3
2 ft./sec.
Nitrogen
Run
Humber
DSG-2U
DSG--25
DSG-26
DSG-lil
DSG-2T
BSG-23
DSC -29
DSG-30
DSG-31
DSG-32-
DSG-33
DfiG-3'i
DSG-35
DSG--36
DRG-3Y
DSG-30
DSG-39
Gas Concentrations ,
Volume %
H_2S
J».»i
lit.l
25. U
lull
5.3
10.0
15.0
20.8
31.6
5.3
5.3
5.3
5.0
9.1
18.2
26. 1>
37.6
H?0
0
0
0
10
0
0
0
0
0
30
20
10
0
0
0
0
0
Nominal Bed
Temperature ,
°F
250
250
250
250
300
300
300
300
300
300
300
300
325
3?5
325
325
325 *
122
-------
concentration was varied over a range of 4 to 37%. The
H20 concentration was varied from 0 to 30%. The data
from each run is given in Appendix A-ll.
To evaluate the data the rate expression was assumed to
be of the form:
g2(Xv) g2(XS) g3(y) g(y) (38)
The data was analyzed by stepwise and multiple regression
methods and the following expression was obtained for
experiments of an inlet water concentration of zero.
23.6 a'2644/1 v °'58 X?-67 Ibs. S/lb. C-min. (39)
The sulfur loading (Xs) was found to have no observable
effect on the rate, or at least the effect could not be
isolated from the effect of the acid loading (Xv), so
that g(XS) = 1. The function of water, gCyuorp' was not
determined explicitly due to the nature of this variable's
effect on the reaction, and consequently Equation (39)
applies only for yjjoQ = ^- Equation (39) fits the differ-
ential data reasonably well as shown in Figure 42. A
statistical analysis showed the average difference between
the model and the differential data to be 17%, with a
maximum difference of 50%.
Extension of the rate model to situations where water
vapor is present required some provision for the effect
of this variable. Figure 43 shows the effect that water
vapor has on the reaction rate. The data indicate that
the rate decreases sharply as yH2o increases from 0 to 20%,
but that there is no further decrease above 20%, There is
a factor of about 6 between the rates at 0 and
20% water vapor.
The rate model was tested against the results of sulfur
generation runs in the 4" diameter fluid bed pilot regene-
rator (subsequently described in this section) by using it
in a design equation to calculate predicted reactor volumes
In conducting the evaluation,various representations of
123
-------
Figure 42
Comparison of the sulfur generation rate model
to the experimental data for 250 to 325°F
6 8 10 20
HgS Concentration, Volume %
40
60 80
124
-------
Figure 43. Effect of H20 concentration on rate of sulfur generation
ID
O
oo
IT)
OJ
4J
to
*
*1-
UD
10 20
H20 Cone., volume %
-------
g(yH20) were tried in an attempt to obtain the best corre
lation between actual and predicted reactor volumes. The
approach which proved most successful was to take a con-*
servative rate constant adjusted by a factor of 6 for H20
concentrations above 2070. More sophisticated methods did
not significantly improve the correlation. The simple
approach is considered justifiable with respect to the
available data on the effect of y^O' anc* it: yie^s con~
servative design estimates. The adjusted rate equation
proposed as the best available expression for designing
sulfur generators is, therefore,
(40)
3.8 e- 0-58 ^0.67
dt H2S
Design equations were derived from this rate expression
based on the same multistage fluid bed reactor model used
for the S02 sorber design equations. The important assump-
tions in the reactor model were that the carbon phase was
well-mixed and the gas phase was plug flow. The following
design equations were derived. The complete derivation
is in Appendix A-9.
N0 42 , 0-42 -2644/T 9 0.67
(YH2s)j+l *
-------
Fluid Bed Studies in 4" Diameter Pilot
Regenerator To Determine Adequacy of Existing
6" Diameter Reactor as Pilot Sulfur Generator -
Experiments were carried out in the 8 stage, 4" diameter
fluid bed pilot regenerator to determine whether the
existing 6" diameter S02 sorber would make a satis-
factory sulfur generator for the integrated pilot plant.
The primary question was whether high conversion of
both reactants could be obtained in the 6" diameter reactor.
The goal was set at 99% acid conversion to sulfur at the
maximum H2S utilization at the space velocities possible
in the existing 6" diameter reactor. The goal of 997o conver-
sion of acid to sulfur was considered necessary during this
period in development in order to avoid disruptions caused
by unreacted acid entering the next step of regeneration.
Later results from integrated operation showed that
unreacted acid could be tolerated, but this was not known
at the time decisions were made concerning adequacy of the
6" unit as a sulfur generator.
Achievement of 99% conversion of acid to sulfur was expected
to be a difficult goal because the inlet M2S concentration
would have to be kept fairly low due to fluidization
requirements in the 6" diameter unit and in order to
obtain acceptable H2S utilization. In fact, the stoichi-
ometric H2S concentration required for complete acid con-
version at 1007o H2S utilization was only about 6% for the
6" diameter unit. This concentration was based on the i
anticipated sulfuric acid feed rate for integral operation,
which determined the stoichiometric H2S feed requirement,
and on the volumetric gas flow rate necessary for proper
fluidization in the 6" reactor, which determined the
degree of dilution.
The experiments in the 4" diameter unit were designed to
simulate the anticipated conditions in the 6" unit as
closely as possible, but the conditions could not be simu-
lated with complete accuracy due to a difference in total
carbon bed depth between the two reactors. The 6" unit's
planned bed depth was greater by a factor of 1.7 to 2.3
depending on the length of the overflow weirs, so that
simulation of carbon residence time in the 4" unit required
a reduction in the carbon feed rate per unit area to
attempt to compensate for lower bed depth. As a conse-
quence of the reduced carbon feed rate per unit area, the
stoichiometric H2S concentration was lower for the simu-
lation runs. The lower bed depth in the 4" unit also
caused a higher space velocity, since linear gas velocity
was held constant. These two inaccuracies in simulation--
higher space velocity and lower stoichiometric H2S
concentration--had a negative effect on conversion of
127
-------
reactants. Therefore, the results of the simulation experi-
ments were conservative with respect to the likely per-
formance of the 6" unit.
The conditions and results of the experiments are presented
in Table 27. The runs were made at carbon feed rates of
9.8 and 6.3 Ibs./hr. to provide carbon residence times of
32 and 50 minutes, respectively. Inlet acid loading was
about 0.21 Ib. acid/lb. C. Average column temperature
was about 290°F in most experiments. Lower temperatures,
about 260°F, were tried in a few runs. Linear gas velocity
was 2.0 ft./sec. The inlet H2S concentration ranged from
3.6% to 29.8% and the H2S feed rate ranged from 1 to 5 times
the stoichiometric requirement for complete acid conversion.
Space velocity was 2800 to 3000 hr.'1.
The main conclusion of the experiments was that the goal
of 99% acid conversion to sulfur with high utilization of
H2S was not attainable under the 6" diameter unit
conditions. The highest H2S utilization, 72%, was achieved
in Run SG-55 at a H2S feed rate of 1.05 times the stoichi-
ometric requirement, with an acid conversion to sulfur of
63%. At higher H2S feed rates the H2S utilization was
lower, falling to about 50% at a ratio of 1.6, 40% at 2.2,
and 30% at 4 with acid conversions to sulfur of 73, 85,
and 90%, respectively.
In discussing the conversion of H2S04, it is important to
note that in addition to forming the desired elemental
sulfur product, a fraction of the acid can react to form
S02 that is evolved in the outlet gas of the acid conver
converter. This reaction was observed in the experiments.
The results show that the percent acid conversion was
about the same for all runs, falling between 92% and 9770.
The conversion to sulfur, however, ranged from 61 to 9070;
and the amount of acid converted to S02 varied from 0 to
43%. Inlet H2S concentration appeared to be the main
determinant of the product distribution between sulfur
and S02- As H2S concentration was increased, conversion
to sulfur also increased, and evolution of S02 decreased.
The effects of inlet H2S concentration on acid conversion
to sulfur and S02 evolution are shown in Figures 44 and
45, respectively.
In order to obtain 90% acid conversion to sulfur, an inlet
H2S concentration of 20 to 30% was required, which repre-
sented a H2S excess of 3 to 5 time's the stoichiometric
requirement and yielded a H2S utilization below 30%.
Experiments in which higher H2S utilization was sought
128
-------
Table 27. SUMMARY OF SULFUR GENERATION RESULTS
RUN
HO.
SG-55
SG-57
SG-64
SG-56
SG-62
SG-65
SG-58
SG-59
SG-60
SG-61
SG-6fi
SG-67
SG-68
SG-69
SG-70
COLUMN
TEMP.,
"f
AVG
294
29E
2S5
294
291
2S3
286
290
280
284
280
289
290
260
262
RANGE
275-312
276-315
270-303
275-324
271-316
242-270
270-313
270-311
250-303
263-297
236-297
237-303
242-320
223-273
225-281
CARBON
• INLET STREAM
C
RATE,
*C/HR
9.8
It
n
ii
it
M
6.3
if
N
«
9.8
M
n
ii
u
RES.
TIME
KIN.
32
II
"
M
(1
H
50
II
y
H
32
U
II
H
II
ACID
LOAD.,
*AC1D/
*C
0.215
• •
t
Ii
•>
II
II
•
II
»
0.200
0.209
0.208
0.207
0.211
GAS INLET STREAM
GAS*
FLOW
CFH?
70°F
438
439
439
436
435
435
438
438
437
435
464
443
435
456
461
H2S CONC.
VOL.
6.1
9.6
9.6
13.6
12.8
13.4
3.6
6.3
10.4
13.1
10.7
21.7
29.8
22.6
8.4
TIMES
STOICI!
1.05
1.66
1.66
2.36
2.22
2.32
0.92
1.62
2.67
3.45
2.15
3.74
5.13
4.29
1.58
CARBON OUTLET STRFAH
ACID
LOAD.,
#ACID/
K
0.015
•0.012
0.011
0.011
0.011
0.017
0.013
0.012
0.013
0.012
0.009
0.005
0.008
0.010
0.014
S LOAD.,
K/K
THEO.
0.282
"
•
II
11
11
II
II
II
0.261
0.272
0.271
0.269
0.275
ACT.
0.178
0.206
0.222
0.217
0.239
0.236
0.171
0.197
0.230
0.240
0.207
0.240
0.245
0.239
0.205
WT.
FINES
"C/HR
...
—
...
0.007
—
—
0.006
—
...
...
...
—
...
—
GAS OUTLET STREAM
COMPOSITION**,
VOLUME I
H?S
1.8
4.4
4.5
8.0
7.3
8.0
1.5
2.6
5.0
8.0
4.8
15.2
22.0
15.3
3.8
SO?
0.79
0.49
0.44
0.36
0.30
0.16
0.50
0.35
0.17
0
0.46
0.09
0
0.07
0.17
H?0
7.0
7.3
8.6
7.9
8.9
6.6
6.2
7.8
10.0
8.0
6.4
8.0
7.6
8.6
5.5
N?
90.6
87.7
86.5
83.7
83,5
85.2
91.8
89.2
84.8
84.0
88.3
76.7
70.4
76.0
90.5
~so7
EVOL.
",. IN
43.0
27.5
23.9
19.1
16.4
8.7
43.0
30.0
14.6
0
30.9
5.2
0
4.2
9.6
H2S
UTIL.
% OF
72
52
52
40
41
39
57
56
49
37
54
30
26
29
55
S BALANCE,
LBS.
IN
2.96
4.28
4.25
5.
5.37
5.62
1.78
2.79
4.25
5.20
4.83
8.90
11.5
9.26
8.94
OUT
2.87
4.12
4.26
5.45
5.22
5.49
2.00
2.50
3.60
4.63
4.34
8.4
10.5
6.61
3.36
ACID
DECOMP
X
98.0
94.4
95.0
95.0
95.0
92.0
93.0
95.0
94.0
94.0
95.5
97.6
96.3
95'. o
93.2
CONV
TO S
V
«B
63
73
79
77
85
84
62
70
82
85
79
88
90
89
75
ACID
DECOMP.
RATE,
#ACID/
#C-MIN.
*. 27 (ID"3)
4. 94(10-3)
5.33(10-3]
5. 2K10-3)
5.74(10-3)
5.66(10-3)
2.64(10-3)
3.04(10-3)
3.55(10-3)
3.70(10-3)
4.97(10-3)
5.76(10-3)
5.88(10-3)
5.73(10-3)
4.92(10-3)
SPACE
YELOC.
KR.'1
2,ss:
2.S60
2,860
2.840
2,530
2,830
2,830
2,850
2,840
2.83Q
3,020
2,880
2,830
2,970
3,000
*Linear gas velocity is 1.9-2 ft./sec. at average column temperature.
**C02 and CO not detected in outlet gas.
-------
Figure 44,
Effect of inlet H2S concentration on the
per cent conversion to sulfur in simulation
experiments using a 6" diameter fluid bed
unit for integrated operation with an 18"
diameter S02 sorber
o
-------
Figure 45
Effect of H2S concentration and carbon residue
time on acid evolved as S02 in simulation
experiments using a 6" diameter fluid bed unit
for integrated operation with an 18" diameter
S02 sorber
-a
0)
•o
•r—
O
C Residence Time, 32 Min
Avg. Temp. Near 285°F
C Residence Time, 50 Min.
Avg. Temp. Near 285°F
C Residence Time, 32 Min.
Low Avg. Temp. Near 255
10 15 20 25
Inlet H2$ Concentration, Volume %
131
-------
yielded lower acid conversions, in the 60 - 707o range. The
results indicated convincingly, therefore, that reactant
conversion goals could not be met in the 6" unit. It was
estimated ba.sed on the experimental data that an increase
in carbon residence time in the 6"0 unit to about 90
minutes would be required to achieve the goal of 99% acid
conversion to sulfu*. An inlet H2S concentration of 30%
would also be necessary, resulting in a H2S utilization of
about 20%. To obtain the 90 minute residence time would
require a carbon bed depth 3.25 times that employed in the
4"0 unit simulation experiments. This would mean an
increase in the number of stages in the 6"0 unit from 9
to 13, with 8" overflow weirs. The increased bed depth
would reduce the space velocity to about 1,000 hr.~l.
Comparison of Fluid Bed Design
Model with Experimental Results -
The fluid bed acid converter design model was compared to
the 4" diameter fluid bed runs by using the model to pre-
dict the theoretical reactor size and number of stages
under the conditions for each run. The results of the
comparison are presented in Table 28. The predicted
reactor volume averages about 307o higher than the actual
volume. The predicted number of stages ranges from 8 to
13 with an average of 10.5, compared to the actual 8 stages
in the reactor. The model, therefore, conservatively pre-
dicts reactor volume, and it is considered adequate for
prediction and scale-up to the next development stage.
Fixed Bed Experiments -
After finding that acid conversion is limited to a maximum
of 90% in the simulation experiments, tests were made in
a 1" diameter fixed bed apparatus to ascertain whether
the 9970 conversion goal was obtainable for the particular
batch of carbon in question. The tests were made at a
H2S concentration of 30%. Runs were conducted using fresh
acid loaded carbon and partially reacted product material
from Runs SG-65 and -68. The results are presented in
Table 29.
For the acid loaded sample, acid decomposition was measured
at 97.5% and conversion to sulfur was 119%. The SG-65
product material yielded 98.1%, acid decomposition and 9970
conversion to sulfur, while the yields for the SG-68 sample
were 97.3% acid decomposition and 94%, conversion to sulfur.
These results indicate that 99% acid conversion to sulfur
definitely is possible with the carbon in question.
132
-------
Table 28. COMPARISON OF THE FLUID BED DESIGN MODEL WITH EXPERIMENTAL
SULFUR GENERATION FLUID BED DATA
u>
RUN
NO.
S6-27
SS-55
SG-56
SS-57
SG-58
SG-59
SG-60
S6-61
S6-62
S6-64
S6-65
S6-66
S6-67
SG-6S
SG-69
AVG.
TEMP.
°F
300
294
294
295
286
290
280
284
291
286
253
280
289
299
260
CARBON
RATE,
*/HR.
42.8
9.8
9.8
9.8
6.3
6.3
6.3
6.3
9.8
9.9
9.8
9.8
9.8
9.8
9.8
INLET
GAS
RATE,
CFH 1? 70°F
410
438
436
439
438
.438-
437
435
435
439
435
464
443
435
456
INLET
*H2s
0.24
0.061
0.136
0.096
0.036
0.063
0.104
0.131
0.128
0.096
0.134
0.107
0.217
0.298
0.226
OUTLET yH2$
ACTUAL
0.100
0.016
0.080
0.044
0.015
0.026
0.050
0.080
0.073
0.045
0.080
0.048
0.152
0.220
0.153
CALC.*
0.128
0.032
0.092
0.057
0.017
0.039
0.075
0.096
0.082
0.055
0.086
0.073
0.166
0.245
0.178
INLET ACID
LOADING,
#/#C
ACTUAL
0.210
0.216
0.216
0.216
0.216
0.216
0.216
0.216
0.216
0.216
0.216
0.200
0.209
0.208
0.207
CALC.**
0.210
0.125
0.175
0.159
0.127
0.153
0.186
0.216
0.182
0.165
0.198
0.144
0.198
0.208
0.199
S02
EVOL.,
% INLET
ACID
0
42.1
19.2
26.2
41.1
29.0
14.1
0
15.9
23.6
8.5
28.2
5.0
0
4.1
rOTJTtET
ACID
LOAD.,
*/#C
0.1197
0.0150
0.0110
0.0120
0.0130
0.0120
0.0130
0.0120
0.0110
0.0110
0.0170
0.0090
0.0050
0.0080
0.0100
' MATERIAL
BALANCE ,
LBS. SULFUR
IN
11.10
2.91
5.61
4.18
1.75
2.73
4.21
5.17
5.31
4.18
5.52
4.76
8.64
11.4
• 9.21
OUT
10.30
2.67
5.23
3.86
1.83
2.35
3.38
4.46
5.16
4.03
5.35
4.11
8.08
10.50
8.30
NUHBER
OF
STAGES
ACTUAL
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
CALC.
9.5
12.6
10.3
11.3
12.4
9.8
8.3
8.0
11.1
12.8
12.2
11.7
9.9
7.9
9.7
SPACE
VELOCITY,
HR.'1
ACTUAL
2,725
2,850
2,840
2,860
2,850
2,850
2,840
2,830
2,830
2,860
2,830
3,020
2,880
2,830
2,970
CALC.
2,290
1,840
2,250
1,980
1,880
2,330,
2,800
2,900
2,080
1,820
1,900
2,110
2,370
2,940
2,490
TOTAL
VOLUME,
CU. FT.
ACTUAL
0.1 39C
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
0.1396
CALC.
0.1663
0.2203
0.1796
0.2057
0.2165
0.1704
0.1448
0.1390
0.1937
0.2240
0.2123
0.2045
0.1735
0.1372
0.1698
*Higher outlet H2S concentration used by computer to
**Inlet acid loading corrected for S02 evolution.
obtain correct material balance.
-------
Table 29. FIXED BED SULFUR GENERATION EXPERIMENTS
CO
Run
Nunber
2
3***
4***-*
5*****
Gas Flow
cc/mi n .
N2
125
tl
It
II
H2S
55
n
it
M
Average
Temp.
°F
286
285
289
282
Space
Velocity
hr.-1
1,000
II
«
(1
Carbon Anal
Inlet
% S
5.56
18.1
0.56
19.3
Acid Load, 1 acid/# C
From
% S
0.194**
___
—
—
SO?
Anal .
0.218
__.
—
Acid
Titr.
0.181
0.020
0
0.011
/ses
Outlet
# S/*-C*
0.30
0.28
0.016
0.25
* acid/I C
0.005
0.004
0.001
0.006
Acid
Decomposition
%
97.5
98.1
97.3
Conversion
to Sulfur
%
119
99
...
94
*Less residual sulfur on virgin carbon.
**Used as inlet acid loading, since have standard for total sulfur analysis.
***Feed sample to fixed bed from product of Run SG-65 (18.1% S, 0.0164 Ib. acid/lb. material).
****Feed is virgin carbon.
*****peecl sample to fixed bed from product of Run SG-68 (19.3% S, 0.00882 Ib. acid/lb. material).
-------
The 11970 conversion to sulfur for the acid loaded sample
was an indication of a separate phenomenon, believed to be
reaction of H2S with chemisorbed oxygen. This would cause
additional formation of elemental sulfur on the carbon. To
investigate this possibility, a sample of virgin carbon was
tested in the fixed bed apparatus, and the results showed
enough sulfur formation to account for TL of the 19%
excess conversion. This was considered valid qualitative
evidence that reaction of H2S with chemisorbed oxygen does
occur. The other 12% was attributed to experimental and
analytical errors.
Development of Moving Bed Reactor
for Integrated Pilot Plant-
The results of the simulation experiments in the 4" diameter
fluid bed reactor indicated that the 6" diameter unit was
unsuitable for use as a sulfur generator in the integrated
pilot plant because it was oversized for the application.
A sulfur generator properly sized to match the existing
18" diameter S02 sorber would be only 2.5" diameter, which
would not be feasible from a mechanical standpoint.
Fixed bed tests showed that the desired degree of acid con-
version was obtainable. This led to the suggestion of a
moving bed reactor as an attractive alternative for the
fluid bed reactor initially intended for the pilot plant.
Exploratory experiments were then carried out in a 1.5"
diameter moving bed reactor in order to evaluate the
potential of this approach. The results were favorable,
indicating acid conversion of 98-100% and H2S utilization
over 99%, at space velocities in the 100-330 hr.-l range
and.carbon residence times between 70 and 185 minutes.
The conditions and results of the moving bed runs are
compared to fluid bed data in Tables 30 and 31.
The success of the 1.5" diameter moving bed experiments led
to the decision to proceed with development of a pilot scale
moving bed sulfur generator. Design calculations were made
for reactors of 6" and 8" diameters. The 8" diameter size
was chosen because in a 6" diameter unit there was a possi-
bility that the minimum fluidizing velocity could be
approached in the reactor and seal leg. (A moving bed of
this size does not have the fluidization requirements, but
in fact fluidization should be avoided.) The design speci-
fications for the 8" diameter moving bed sulfur generator
are given in Table 32, and a drawing of the reactor system
is shown in Figure 46. Although the overall reactor
length is 10 feet, the carbon bed depth is 6 feet in the
135
-------
TABLE 30. COMPARISON OF FLUID BED AND MOVING BED SULFUR GENERATION TESTS
Equipment
4" Dia.
Fluid
Bed
Fluid
Run
No.
SG-68
Bed*
Modified
6" Dia.
Moving
Bed**
SG-73
Column
Ten?., °F
Avg.
290
290
285
Range
242-320
270-300
Carbon
Inlet Acid
Loading
# Acid/* C
0.208
0.184
0.165
Residence
Time
minutes'
32
~90 -
-180
Rate
!/hr.
12.1
35
1.1
1
Hydrogen
Cone.
29.8
Stoich.
Ratio
5.1
I
30 i 5
,
32.3
1.!
Sulf-;ce
Outlet! Util.
1
22.0 ! 26
!
24
-0
~20
>90
i
Acid Lest as
SOz in Off Gas
Inlet
0
0
0
Cone.
in Gas
—0
0
~0
Acid
Decorap.
96
98
94
Sulfur
8aUr.ce
In/Out
1.1
—
Outlet Sulfur
Loea, ? S/- C
Theor.
0.271
0.240
!
—
0.212
Actual
0.248
0.238
i
Conversion j y*?»«
Sulfur ; hr-
90
99
0.216 : 102
2.850
1,000
100
' 1
OJ
*These are projected results ir. 6" diameter unit if codified based on all results to operate at a space velocity
of 1,000 hr.~l, i.e. increase niniber of stages from 9 to 13 and bet height to 8 inches.
**Actua1 moving bed data.
-------
Table 31. DATA SUMMARY - 1-1/2" DIAMETER MOVING BED SULFUR GENERATION TESTS
Run
No.
SG-74
SG-79
Column
Temp.
°f
Avg.
270
268
Range
260 -
280
255 -
'280
Carbon Inlet Stream
Reac.
Inv.
# C
3.09
0.92
Carbon
Rate
?C/hr.
1.02
0.81
Res.
Time
min.
185
68
Acid
Load
#Acid/#C
0.19
0.19
Gas Inlet Stream
Gas
Flow
CFH ?
70° F
9.14
6.2
H2S
Cone.
Vol.
a
IB
29.5
30.6
Times
Stoich.
1.1
1.1
Linear
Gas
Velocity
ft. /sec.
0.3
0.2
Carbon Outlet Stream
Sulfur Load
Ib. S/lb. C
Theoreti cal
0.248
0.252
Actual
0.248
0.247
Gas Outlet Stream
Composition,
Vol. % (Dry)
12$
~0
.18
S02
-0
.10
C02
—
CO
—
H20
N.D.
N.D.
N2
N.D.
N.D.
S02
Evol.
% In.
Acid
0
<1%
H2S
Util.
% of
Inlet
>99
>99
S Balance
S
In
Ib.
0.28
0.22
S
Out
Ib.
0.27
0.18
%
Conv.
to
S
100
98
Acid
Decompositior
Rate,
#Acid/#C/min.
t
IxlO'3
3xlO"3
Space
Veloc.
hr.'1
100
335
-------
Table 32. DESIGN CONDITIONS - MOVING
BED SULFUR GENERATOR
Unit Size
Nominal, NPS 8
I.D., inches 8.33
Temperature, °F 300
Carbon Rate, Ibs./hr. 27.7
Acid Load, Ib./hr. 0.184
Gas Rate, SCFH 220
Gas Velocity, ft./sec. at 300°F 0.25
Carbon Residence Time at S = 100 hr."1, hrs. 2.85
Carbon Inventory at S = 100 hr."1, Ibs. 79
Carbon Bed Height, ft.
S = 100 hr.'1 6
S = 200 hr."] 3
S = 500 hr."1 1.2
Pressure Drop* at S = 100 hr."1, in. H20 31
Reactor Length, ft. 10
Seal Leg Length, ft.
Total Length 8
Purged Length 6
Seal Leg Velocity, ft./sec.
S = 100 hr."1 0.25
S = 200 hr.'1 <.l
S = 500 hr.'1 <.l
*Based on previously measured pressure drop.
138
-------
Figure 46. Moving bed sulfur generator
VO
DIA « I STA5E FLUID 6EO
e"c« » " ic' oss
T--2SO-5ZS *F
NOTES
I. ALL OIANAETER OMANMlEi BO
CARBOK1 Uie 2S% CONJB.S
2.THCRivioce>up'..ss Ti-3 4 Ti-4
£AS SAMPLE PROBE 5.S-1
ARE RETRACTABLE.
TO BUCKET E.I.E.VA70R.
COKjVt'N'OK
-------
8" diameter unit to give a space velocity of about 100 hr."-*-
for a gas flow rate of 220 SCFH. Carbon residence time is
about 2.85 hours for a carbon feed rate of 27.7 Ibs./hr.
An experimental program was carried out to evaluate the
performnace of the 8" diameter moving bed reactor and to
determine the optimum operating conditions. The experi-
ments were divided chronologically into three series of
runs. An initial series of runs was made primarily to
study the effect of temperature. The results indicated a
need for additional heating in the lower section of the
reactor. The reactor was modified by installing a
special finned steam heater, and a second series of runs L
was carried out. These runs were not too useful because
the finned heater caused a carbon flow problem which led
to poorer reactor performance. The finned heater was
removed, and a final series of runs was made to study
other methods of improving reactor performance. The
experimental conditions and results of all three series of
runs are summarized in Table 33.
A carbon feed rate of 28 Ibs./hr. was used in most runs,
with an acid loading between 0.14 and 0.24 Ib. H2S04/lb.
carbon. The H2S feed rate was 0.94 to 1.4 times the
stoichiometric requirement for complete acid conversion.
The H2S concentration was 27.6 to 34.67o by volume, and
total gas flow rate was 200 to 280 cfh at 70°F. Tempera-
ture was an important variable with average bed tempera-
tures ranging from 233 to 319°F. An average of around 270-
280°F was typical in many runs. Temperatures at the top
of the reactor were higher than at the bottom, with vari-
ations ranging from 20 to 100°F.
Initial Runs -
The initial eight runs were primarily intended to study the
effect of temperature. These runs were made at average
carbon bed temperatures from 230 to 320°F and at inlet H2S
concentrations near 30 volume 7o. The total gas flow rate
was adjusted as necessary to provide the 110% of stoichi-
ometric H2S reactant needed to convert the sorbed sulfuric
acid which varied from about 0.17 to 0.23 Ib. acid/lb. C.
The results generally indicate that the higher temperatures
favor higher H2S utilization, but lower temperatures favor
lower amounts of sorbed acid evolved as S02 in the off-gas.
Effect of Temperature on Evolution of S02 -
The effect of the inlet carbon temperature, which is the
maximum measured carbon temperature in the present experi-
140
-------
Table 33. SUMMARY OF SULFUR GENERATION RESULTS
RUN
NO.
SG-86
SG-87
56-88
SG-88R
SG-89
SG-90
SR-91
SG-92
COLUMN
TEMP.,
AVG
319
290
253
262
243
242
?33
244
RANGE
326
260-
305
91Q
265
236-
301
206-
296
198-
300
192-
236
203-
289
CARBON INLET STREAM
REAC
INV.
75
H
H
n
tt
n
,
it
c
RATE
28
28
27
28
28
28
?8
28
RES.
TIME
KIN.
162
162
166
162
162
162
16?
162
ACID
LOAD.
#ACIO/
#C -
0.167
0.185
0.176
0.227
0.204
0.227
0.220
0.212
GAS INLET STREAM
*
GAS
FLOW
CFH@
70°F
204
II
280
280
280
280
280
H2S
CONC.**
VOL.
30.5
28.8
27.6
30.4
32.0
29.5
31.8
32.1
TIMES
STOICH
1.1
0.94
0.96
1.1
1.3
1.1
1.2
1.3
LINEAR
GAS
VELOC.
R/SEC
0.24
0.23
0.22
0.30
0.30
0.30
0.29
0.30
CARBON OUTLET
STREAM
ACID
LOAD.
fACID/
1C
0.019
0.023
0.015
0.018
0.010
0.007
0.012
0.007
SULIUR
LOADING,
THEO.
0.218
0.241
0.230
0.296
0.26E
0.29f,
0.287
0.277
ACT.
0.204
0.221
0.205
0.273
0.264
0.285
0.281
0.284
GAS OUTLET STREAM
COMPOSITION, VOL. %
H2S
1.2
1.6
5.9
2.0
4.3
1.3
4.7
3.5
S02
2.2
0.8
0.09
0.90
0.8
0.80
0.43
0.61
C02
0.3
0.18
0.09
0.0
0.1
0.0
0.09
0.09
CO
0
0
0
0
0
0
0
0
H20
37
35,5
34.8
39.4
37.1
34.8
38.1
40.6
N2
61.5
61.6
62.2
56.8
57.0
59.7
59.3
59.8
S02
EVOL.
% OF
INLET
ACID
27
6
1
12
12
10
6
11
H2S
UTIL.
% OF
INLET
95
97
74
92
84
95
82
83
SULFUR
BALANCE
IN
6.9
6.7
6.3
9.3
9.7
9.0
9.8
9.8
OUT
6.5
7.0
7.0
8.9
9.4
8.5
9.7
9.6
ACID
DECOMP
88
88
91
92
95
97
94
97
CONV
TO S
at
90
92
***
90
92
99
94
98
102
SPACE
VELOC
HR.-l
90
90
90
123
123
123
123
123
PURPOSE
OF RUN
Effect of
Temp. -
320° F
Effect of
Temp. -
290°F
Effect of
Tercp. -
250DF
C Prehtr.
& lO'.R.H.
C Prehtr.
(3 230CF
i 5SR-.H.
C Prehtr.
g 300" F
I KK.H.
C Prehtr.
9 230°F
f. 10IR.H.
C Prehtr.
P 28CCF i
Add'l r.tg
from Wai 1
at Bottom
of Reactor
*Frora rotameters
**As determined ay gas chromatograph.
***An error analysis indicated the maximum uncertainty in this numter is ±10%.
calculated to be tO.3 for inlet t S in.
The error in the sulfur material balance was
(continued)
-------
Table 33 (continued). SUMMARY OF SULFUR GENERATION RESULTS
t"0
RUN
HO.
S6-94
SS-95
SG-93
SG-96
SG-99
SG-100
SG-101
SG-103
SG-102
COLUWi
TEMP. ,
°F
AVG
280
283
270
276
274
282
277
255
PA'IGE
250-
305
247-
312
238-
?i n
2J7-
311
254-
285
256-
230
259-
232
251-
273
280-
316
CARBON INLET STREAI!
9EAC
ICV.
i C
70
II
tl
75
»
I
•
«
C
RATE
aC/HR
28
24
24
24
28
23
23
28
28
RES.
TIME
MIN.
151
176
176
175
162
162
162
162
162
ACID
LOAD.
*ACID/
#C
0.239
0.229
0.224
0.222
0.163
0.166
0.168
0.143
0.227
GAS INLET STREAM
GAS
FLOW
CFH 0
70°F
214
231
214
214
206
206
206
200
206
H2S
CCNC."
VOL.
w
A
32.4
34.6
32.6
32.2
32.8
33.0
33.0
31.0
32.9
TIMES
STOICH
1.0
1.3
1.3
1.3
1.3
1.3
1.3
1.4
1.0
LINEAR
GAS
VELOC.
FT/SEC
0.25
0.27
0.25
0.25
0.26
0.26
0.26
0.25
0.26
CARBwmrfLET
STREAM
ACID
LOAD.
SAC ID/
*C
0.046
0.039
0.047
0.052
0.007
0.009
0.007
0.011
0.014
SULFUR
LOADING,
«S/IC
THEO.
0.287
0.299
0.293
0.290
0.212
0.217
0.220
0.187
0.296
ACT.
0.226
0.251
0.227
0.221
0.154
0.2CO
0.202
0.199
0.190
GAS OUTLET STREAM
COMPOSITION, VOL. 1.
H2S
3.0
s.s
7.0
7.2
4.4
5.0
5.4
4.0
4.1
S02
0.95
1.0?
0.37
0.31
0.35
0.38
0.33
0.22
1.34
CO?
0.12
0.07
0.08
0.07
0
0
0
0
0
CO
0
0
0
0
0
0
0
0
0
H20
37.4
36.3
34.6
33.7
36.3
36.6
34.1
36.2
33.9
N2
56.0
54 6
53.5
52.7
59.1
59.0
59.2
59.8
60.7
532
EVOL.
; OF
INLET
ACID
10
14
5
4
5
5
4
3
13
H2S
UTIL.
? OF
INLET
89
82
80
74
85
83
81
85
86
STFlTut)
BALANCE
IN
' S
8.8
9,1
3.6
8.7
7.5
7.5
7.6
6.9
8.0
OUT
*S
8.1
8,3
7.7
8.0
6.7
6.9
7.1
6.7
6.8
ACID
DECOMP
V
81
83
79
77
96
95
96
92
94
CONV
TO S
tt
/a
74
84
77
76
92
92
92
106
64
SPACE
VELOC
HR.-l
100
no
100
100
9C
90
90
90
90
PURPOSE
OF RUN
Cir. 18"
to 8'
Dir. 18"
to S"
(Higher
H->S
Stoich.
Ratio)
C Pre"itr
r: 2SO F
& lO'-R.H.
CFrehtr.
P 270'F
& 1C K.I-'.
C Preitr.
C27C:~r
•: IJTR.a.
Increas.
C Tero.
with
InlstGas
Ircreas.
C lenp.
witn C
Frecore.
Decreas.
i2S Cone
Dir. 18"
to S";
Moisture
Le»el
*From rotameters
**As determined by gas chromatograph.
-------
in Figure 47. As can be seen, as the inlet carbon tempera-
ture was decreased from 325 to 265°F, the evolution of S02
decreased from about 30% to 1% of the inlet acid. The
data indicates that the "best" temperature for the inlet
carbon would be in the range of 250 to 270°F.
Effect of Temperature on H2S Utilization-
The effect of the inlet carbon temperature (maximum
reactor temperature) on the utilization of H2S is given
in Figure 48. The data indicate that as the temperature
increased from 260°F to 325°F, H2S utilization increased
from 75% to 97%. The condition favoring highest H2S
utilization is the higher reactor temperature.
Effect of Temperature on Per Cent Conversion to Sulfur -
The effect of the inlet carbon temperature on the per cent
conversion to sulfur is shown in Figure 49. As can be
seen the percent conversion passes through a maximum at
about 285 to 295°F due to a decreased acid evolution as
S02 and then decreases after 285°F because of a decreased
utilization of H2S.
Vertical Temperature Profile -
In all of the runs, the temperature was higher at the top
of the carbon bed and decreased toward the bottom of the
reactor. For Run SG-90, the carbon temperature varied from
about 190°F at the bottom of the reactor to about 300°F at
the top, for an average of about 240°F. For Run SG-91, a
decrease of about 15°F at the top of the reactor resulted
in an average decrease in temperature of about 10°F, but
cut the S02 evolution almost in half (from 10 to 6% of
inlet acid). On the other hand, the utilization of inlet
H2S decreased from 95% for SG-90 to about 82% for SG-91
for the 15 degree decrease in temperature at the top of
the reactor.
Gas Concentration Profiles for H2S and S02 -
The H2S concentration typically decreased uniformly from
the bottom to the top of the bed. The S02 concentration,
however, was essentially zero until near the top of the
reactor. The formation of S02 apparently occurred almost
entirely in the top six inches of the carbon bed. This is
believed to be attributable mainly to the significantly
higher temperature in the upper part of the bed.
143
-------
Figure 47. Effect of inlet carbon temperature* on the evolution of H2S04
as S02 in an 8" diameter moving bed reactor
32 ••
-------
GO
CVJ
70 -
65 -
60 -
Figure 48
Effect of inlet carbon temperature on H£S utilization
in an 8" diameter moving bed reactor
100
-P-
Ui
95 -
90 -•
85 -•
- 80 -•
10
N
Carbon:
Rate
Type
Gas:
Rate •
Type •
28 # C/hr.
Acid Loaded
204 - 280 CFH @ 70°F
30% H2S/Bal. N2
75 -•
55
250
260
270
280 290 300
Inlet Carbon Temperature, °F
310
320
330
-------
104
oo
o
1C2--
100
98--
96-
- Od --
5 " • ^
'.o
i.
o
a
92--
= 90-
Q- 83 -r
86
250
Figure 49. Effect of the inlet carbon temperature on the per cent
conversion to sulfur
260
270
280 250 SCO
inlet Carbon Temperature, CF
310
320
330
-------
Strategy for Improving Performance
of 8 Diameter Moving Bed
The results of the initial eight runs indicated that a sub-
stantial improvement in performance could be obtained if
the vertical temperature profile was modified so as to
raise the temperature in the bottom of the reactor and
lower it at the top. The anticipated effects were a
reduction in formation of S02 and an increase in H2S
utilization. Modifying the temperature profile in this
way required some means of providing additional heat input
to the bed. To obtain this heat, a finned steam-operated
heat exchanger was installed vertically inside the
reactor.
Finned Steam Heater Experiments -
A series of four runs, SG-93, -94, -95 and -96, was made
with the finned steam heater installed. The finned heater
was successful in raising the temperature, as
shown in Figure 50. Reactor performance did not
improve, however, due to carbon flow stagnation on one side
of the reactor, which was caused by the presence of the
heater. A residence time distribution study clearly showed
flow stagnation and a large deviation from plug flow. This
is shown in Figure 51.
In these runs the 18" diameter 862 sorber was run con-
currently with the 8" diameter moving bed acid converter.
The four runs included two runs with the carbon from
the sorber routed directly to the 8" diameter unit, and
two runs with the carbon passing from the sorber to the 6"
diameter carbon conditioner and then into the 8" diameter
moving bed.
As seen in Table 33, the results of these runs are signifi-
cantly poorer than the results of the first series of runs.
The H2S utilization is consistently less than 90% and acid
conversion to sulfur averages only 8070. Formation of S02
is about the same as before. Further discussion of the
results would not be meaningful in view of the overall
performance reduction which resulted from installation of
the finned heater.
Final Series of Runs -
After determining that the flow problem associated with the
finned heater could not be corrected, the heater was removed
and a final series of-runs was made. The objective in these
runs was to study the effects of 1) additional heat input
from sources other than an internal heat exchanger, 2) H2S
concentration, and 3) moisture content of the entering
carbon.
147
-------
Figure 50. Effect of steam heater on improved carbon heating capabilities
00
350
-------
Figure 51.
Effect of heat exchanger system (1-1/2" pipe x 4") on carbon
flow in an 8" diameter moving bed reactor - tracer feed
composition
vo
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
Predicted Profile for
Theoretical Plug Flow
Virgin Carbon
Actual Measured Profile
with Heat Exchanger and
Exit Pipe Present
*_O
••« mmm ^^^ ^^ . •
i i i i
0
12 18 22 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84
Time, minutes
92 96
-------
Heat Input through Inlet Gas -
In Run SG-100, an attempt was made to raise the temperature
in the bottom of the reactor by increasing the inlet gas
temperature from 300 to 400°F. This produced a temperature
rise of about 10 to 20°F in the lower part of the carbon
bed, but there was no noticeable improvement in acid con-
version or H2S utilization.
Heat Input through Carbon Conditioner -
In Run SG-101,the carbon conditioner was operated at 325°F
in order to raise the temperature of the carbon entering
the 8" diameter unit. This did not produce the expected
increase in carbon bed temperature, however, apparently
because of heat losses in the transfer piping between the'
carbon conditioner and acid converter. Operating the
carbon conditioner at 325°F required the use of pure steam
as the fluidizing gas in order to obtain the desired 1570
relative humidity. Higher carbon conditioner temperatures
would not be practical because it would then be necessary
to operate above atmospheric pressure in order to obtain
the desired relative humidity.
Effect of Inlet Carbon Moisture Content -
In all the experiments in the 8" diameter moving bed, the
inlet carbon moisture content was found to be an important
parameter in acid conversion. Variation of the moisture
level had a large effect on the temperature at the top of
the carbon bed, which apparently was directly related to
the extent of S02 formation. At a low moisture level,
the temperature at the top of the bed was high and forma-
tion of S02 was also high. By increasing the moisture
level, temperature was lowered and formation of S02 was
reduced.
The probable explanation is that adsorption of water vapor
from the gas phase onto the carbon occurs unless the carbon
moisture content is above a certain level. Adsorption
of water is exothermic and causes the observed temperature
rise. Formation of SC-2 is the end result.
The results in Tables 33 and 34 support this explanation.
In Table 34, the carbon conditioner operating conditions
are shown for all the moving bed runs. Carbon moisture
level is presented in terms of the sulfuric acid solution
concentration, Ibs. acid/(lbs. acid + Ibs. H20). A
high acid concentration means a low moisture content.
The effect of moisture level on temperature in the top
of the bed and on S0£ formation is seen by comparing
Runs SG-94 and -95 with Runs SG-100 and -101. These
runs are chosen for comparison because the average bed
150
-------
Table 34. EFFECT OF VOLUME % H20 AND TEMPERATURE
ON THE CONCENTRATION OF ACID SOLUTION
SORBED ON CARBON FOR CARBON PREHEATER
Total
Gas Flow
Rate,
CFH @ 70°F
cid Cone.,
# Acid/(# Acid + #
Steam
Cone.,
Vol. %
Relative
Humidity,
Predicted Meas
*Carbon direct from 18" diameter S02 sorber to 8" diameter sulfur
generator.
151
-------
temperatures are all in the 277-283°F range. In SG-94
and -95 the relative humidity in the carbon conditioner was
3.3% compared to about 14.5% in SG-100 and -101. These
conditions produced a low moisture level in SG-94 and -95,
and a high moisture level in SG-100 and -101. The results
in Table 33 show an average S02 formation of 12% in
SG-94 and -95 compared to 5% in SG-100 and -101, which is
a significant difference.
Overall Reactor Performance Summary -
The results of the final moving bed runs showed that perform-
ance of the 8" diameter reactor was adequate for the intended
application as the acid converter for the integrated
pilot plant. Conservatively, the results indicated that
the reactor could be expected to perform at the following
levels:
Acid Decomposition >95%
Acid Conversion to Sulfur >92%
Acid Conversion to S02 < 5%
H2S Utilization >85%
Although better results for each individual response were
obtained in some runs, the above performance levels repre-
sent conditions that can be expected with a fair degree
of certainty.
Comparison of Moving Bed Runs with Design Model -
Based on the rate expression for the acid decomposition
reaction, a design model was derived for a moving bed
reactor, assuming plug flow of both the carbon and gas
phases. The design model was used to predict teactor
volume for the conditions in the moving bed experiments.
The results of the comparison are shown in Table 35.
It is seen that the predicted reactor volume in all of
the 8" diameter moving bed runs is less than the actual
volume by a factor of 6. In view of the possible gas
channelling and solids flow problems inherent in the
moving bed used, these discrepancies might be expected.
The design model comparison demonstrates, therefore, that
the results of the 8" diameter moving bed experiments
unsuitable for purposes of reactor modeling. This had
no effect, however, on the adequacy of the 8" diameter
moving bed reactor to perform the acid conversion step
in the integral pilot plant runs.
152
-------
Table 35. COMPARISON OF THE MOVING BED DESIGN MODEL WITH EXPERIMENTAL
SULFUR GENERATOR MOVING BED DATA
RUN
NO.
SG-74
SS-79
SG-99
SG-100
SG-101
SG-102
SG-103
1NT-7
AVG.
TEMP.
°F
270
270
274
282
277
300
265
289
CARBON
RATE,
*/HR.
1.02
0.81
28.0
28.0
28.0
28.0
28.0
29.0
INLET
GAS
RATE,
CFH 8 70°F
9.14
6.20
206
206
206
206
200
228
INLET
yH2S
0.295
0.306
0.328
0.330
0.330
0.329
0.310'
0.272
OUTLET yH2S
ACTUAL
0.0005
0.0018
0.044
0.050
0.054
0.041
0.040
0.0003
CALC.*
0.0565
0.028
0.090
0.091
0.083
0.041
0.098
0.039
INLET ACID
LOADING, f/tC
ACTUAL
0.190
0.190
0.163
0.166
0,163
0.227
0.143
0.220
CALC.**
0.190
0.190
0.155
0.158
0.161
0.198
0.139
0.189
S02
EVOL.,
% INLET
ACID
0
0
4.9
4.8
4.2
12.8
2.8
14.1
OUTLET
ACID
LOAD.,
#/*C
0.010
0.010
0.007
0.009
0.007
0.018
0.011
0.034
MAT'L. BAL.,
LBS. SULFUR
IN
0.286
0.208
7.092
7.153
7.172
7.694
6.447
7.225
OUT
0.257
0.204
6.381
6.683
6.784
6.238
6.432
6.719
BED HEIGHT.
INCHES
ACTUAL
82.9
16.8
72.0
72.0
72.0
72.0
72.0
72.0
CALC.
15.46
13.03
12.39
11.67
12.76
11.54
11.53
11.11
TOTAL VOLUME,
CU. FT.
ACTUAL
O.C848
0.0172
2.094
2.094
2.094
2.094
2.094
2.094
CALC.
0.0158
0.0133
0.366
0.339
0.371
0.336
0.335
0.323
*Higher outlet H2S concentration used by computer to obtain correct material balance.
**Inlet acid loading corrected for 502 evolution.
-------
5.2.3 Sulfur Removal
The sulfur sorbed on the activated carbon has to be
removed to recover the sulfur values and to regenerate the
carbon for reuse. Two basic methods were used to effect
this sulfur removal. The first was to vaporize the
sulfur from the carbon, and other was to extract the
sulfur. The bench scale results of each of these methods
are discussed, followed by a comparison of the two methods.
Under thermal sulfur recovery, equilibrium data of sulfur
adsorbed on activated carbon is presented. The sulfur
stripping runs made in fluid bed reactors are presented,
but since it was demonstrated that the unit operations of
sulfur stripping and of H2S generation could be combined
into one reactor the operational data on sulfur stripping
only is more limited than combined operation discussed in
Section 5.2.4 (H2S Generation).
Sulfur recovery by solvent extraction is also presented.
The two methods of sulfur recovery were compared by
recycle experiments of six carbon cycles. The data indi-
cated that sulfur vaporization from carbon was preferred
because the S02 activity of the carbon was maintained,
whereas the solvent extracted carbon required further
treatment.
As back-up information, a pilot plant was designed for
sulfur recovery from carbon by solvent extraction. The
design and economics are presented.
Thermal Stripping Studies -
Equilibrium adsorption data for sulfur vapor on activated
carbons were obtained by contacting carbon with a known
partial pressure of sulfur in a stream of nitrogen followed
by combustion analysis of the carbon to determine sulfur
loading.
The sorbed sulfur was assumed to exist in both physically
adsorbed and chemisorbed states. The amount chemisorbed
was taken as the residual loading after extended purging
with inert gas, and assumed to be constant below 1000°F.
The remaining physically adsorbed portion was found to be
characterized by a form of the Polanyi-Dubinin adsorption
equation with respect to equilibrium vapor pressure and
temperature:
154
-------
In(L-Lc) - A - K(T log ^)2 (43)
where L = total equilibrium sulfur loading
Lc = the amount chemisorbed
T = adsorption temperature
PS = saturation vapor pressure of sulfur at T
P = equilibrium sulfur pressure
A,K = constants.
Data which fit this equation were obtained over ranges of
temperature, pressure, and loading which are of interest
in the analysis of thermal stripping operations associated
with S02 recovery.
A search of the literature shows that previous activities
in the field have been primarily concerned with sulfur
chemisorption on carbon. An exception is the work of
Juza and Blanke2 who measured sulfur vapor isotherms
manometrically under static conditions. Data obtained for
two activated carbons showed evidence of chemisorption,
physical adsorption, and capillary condensation over various
ranges of sulfur loading at temperatures of about 700°F.
In the present work, adsorption measurements were made
dynamically rather than under static conditions in order
to avoid complications due to the production of gaseous
reaction products.
Experimental Results for Equilibrium of Sulfur over Carbon -
Table 36 lists the equilibrium sulfur loadings found on
carbon at the various experimental carbon temperatures
and sulfur vapor pressures. These pressures were taken
from the corresponding sulfur temperatures according to
data presented in the SULFUR DATA BOOK3.
Inspection of Table 36 shows that at the higher sulfur
loadings, the amount sorbed depends upon both temperature
and vapor pressure as would be expected in physical
adsorption. It is seen from the purge data, however, that
155
-------
Table 36. EXPERIMENTAL RESULTS OF EQUILIBRIUM
SULFUR ADSORPTION MEASUREMENTS
Carbon
Adsorption
Temperature ,
op-
Sulfur
Generator
. Temperature,
»F
Sulfur
Vapor
Pressure,
torr
Equilibrium
Sulfur
Loading,
gms S/100 gins C
Adsorption Data
650
650
800
800
1000
800
1000
800
550
450
550
450
550
380
450
325
38
7.1
38
7.1
38
1.55
7.1
0.36
63.0
47.7
46.6
31.8
24.3
16.2
12.7
10.6
Purge Data
1000
1200
1400
0
0
0
7.3
7.1
6.6
there exists at a given temperature a minimum sulfur loading
Figure 52,showing experimental adsorption isotherm points
for 1000° and 800°F, illustrates this further. As the
equilibrium pressure approaches zero, a residual loading is
retained which probably represents chemisorbed material.
Attempts to describe equilibrium adsorption of sulfur in
terms of sulfur loading, vapor- concentration and carbon
temperatures must, therefore, consider both physically and
chemically bound sulfur.
It has now been found that the temperature dependency of
the experimental sulfur sorption data is well represented
by a form of the Polanyi-Dubinin adsorption, Equation 43.
Figure 53 shows the data of Table 36 plotted in this
form as ln(L - Lc)"vs. (T log PS/P)Z where Lc was assumed
to be constant below 1000°F, the maximum temperature of
the equilibrium data. Taking Slope K and intercept A
from the straight line plot, the adsorption equation
becomes:
ln(L-7.3) = 4.10-0.179(1 log ) x 10
~6
(44)
where L = gm S/100 gms C
T = °R.
156
-------
Figure 52
Experimental adsorption isotherm points for
sulfur on activated carbon at 800° and 1000°F
.p
103
157
-------
Figure 53. Polanyi-Dubinin plot of sulfur adsorption data
0.8
158
-------
Using^this equation the partial pressure of sulfur in
equilibrium with carbon having a given loading can be
calculated at any temperature up to about 1000°F. At
higher temperatures, up to 1400° for example, better
results might be obtained by substituting correct values
for Lc as determined by extended purging. Such values are
noted in Table 36. In this case it would be necessary to
recalculate the proper values for constants A and K.
For the purposes of analyzing thermal stripping data, it
may be more convenient to express vapor pressures in terms
of sulfur concentration. Since sulfur exists in the vapor
phase as polyatomic molecules in which molecular weight
depends on temperature, the relationship between pressure
and concentration is:
where % S = concentration expressed as monatomic
sulfur at 1 atm. total pressure
P = vapor pressure in torr
= average number of sulfur atoms per
molecule at the temperature of
T
interest.
A table of values for this latter term vs. temperature is
given in the SULFUR DATA BOOK3.
Figure 54 shows a plot of sulfur concentrations as Si vs.
temperature for various sulfur loadings as calculated from
Equations (45) and (46). Such equilibrium lines may be
used, for example, to calculate the minimum number of stages
required to reach a certain residual sulfur loading on
carbon by means of thermal stripping.
From the temperature and pressure dependency of sulfur
adsorption it is possible to calculate the isosteric heat
of adsorption according to the relation:
159
-------
Figure 54. Equilibrium lines for concentration vs.
temperature at various loadings
0.02
0.01
30 gms S/100 gms C
24 gms S/100 gms C
18 gms S/100 gms C
12 gms S/100 gms C
600
700 800 900
•
Temperature, °F
1000
160
-------
where PI
P2
q
equilibrium pressure at temperature TI
pressure at T2
differential heat of adsorption evaluated
at a particular loading.
The calculated heats of physical adsorption are given in
Table 37 for various total loadings.
Table 37. ISOSTERIC HEATS OF ADSORPTION
OF SULFUR VAPOR ON CARBON
Total Load
^ms S/100 gms C
12
18
24
q
Kcal/mol
24.7
23.0
21.9
q
BTU/lb.
223
208
198
These heats of adsorption which must be supplied during
stripping may be compared to the heat of vaporization of
bulk sulfur of 134 BTU/lb. at 1000°F. The relative magni-
tudes of these heats are reasonable for systems involving
physical adsorption.
Comparison of the experimental sulfur adsorption isotherm
results obtained here with those of Juza and Blanke noted
previously was made by means of Polanyi-Dubinin type plots
Their data obtained near 700°F showed similarities to the
present data within the limits which might be expected to
result from variations in carbon type. However, the
temperature dependency was not properly described by the
P-D relation since the few data reported for higher
temperatures did not fall on a single curve with those at
700°F. The reason for disagreement is not clear although
it is possible that small amounts of gaseous reaction
products formed at the higher temperatures could have
caused errors in the static sulfur pressure measurements
used in this work.
161
-------
Fluid Bed Sulfur Stripping -
A trial sulfur stripping run was made at an average column
temperature of approximately 1050°F in an 8 stage 4"0
fluid bed regenerator. The results are" summarized in
Table 38 below.
Table 38. SULFUR STRIPPING IN A CONTINUOUS
8-STAGE FLUID BED
Run
No.
SIS-1
Carbon
Rate
Ibs./hr.
- 31
Temp.
OF
(Avg.)
1048
S Loading
Ib./lb. C
In
0.307
Out
.094
Sulfur
Removal
%
66*
*The removal of 66% of the adsorbed sulfur
in the 11 minute residence time compared to
60% obtained previously in a three stage
fluid bed unit.
A McCabe-Thiele analysis, assuming 100% stage efficiency,
indicated six and a fraction theoretical stages were
required to achieve the observed results from the trial
run in the 8 stage column. To facilitate the stage effici-
ency calculation a computer program was written. The
program executes an iterative search procedure which con-
verges to an average value of the Murphree tray efficiency
over the entire column. For the trial stripping run an
efficiency of 79% was calculated. Assuming the data from
future sulfur stripping runs yield similar values for the
stage efficiency, then the value that is obtained should
be a valuable piece of information in designing a sulfur
stripper for a particular application.
Bench Scale Sulfur Stripping/H2S Generation -
Two stripping runs were made at 1000°F in the batch 4"
electrically heated fluid bed to determine the effect of
contact time (space velocity) on the approach to equilib-
rium in sulfur stripping. Following the stripping of the
physically sorbed sulfur, removal of chemisorbed sulfur
was investigated as a function of the concentration of
the hydrogen reductant. The results of these runs are
shown graphically in Figure 55. It appears that the
162
-------
Figure 55,
4" diameter batch fluidized bed sulfur stripping and hydrogen
desulfurization runs (Runs JFC-1 and -2)
24
u>
20
16
c
o
£
3 12
c
o
s-
3
CO
JFC-2
"26,500 hr.'1 S. V,
-------
stripping is most likely equilibrium limited and
independent of space velocity over the range investigated.
Conditions run were 2 ft./sec. with 4 inch and 6 inch
expanded bed depths at 1000°F. The space velocities corre-
sponding to the corresponding settled bed depths were
38,400 v/v/hr. (4" expanded bed) and 26,500 v/v/hr. (6"
expanded bed). If the stripping were equilibrium limited,
the rate of sulfur removal at the lower space velocity
would be equal to that at the higher space velocity.
Since total sulfur removed is equal to the product of the
bed weight and the change in sulfur content, sulfur removal
as a function of time can be calculated at each space
velocity. Over the straight line portion of the curve,
the total amount of sulfur removed is the same for the two
runs.
The rate of removal of chemisorbed sulfur appeared to be
zero order with respect to hydrogen concentration within
the accuracy of the analyses. Run JFC-1 was made with 33%
H2 and Run JFC-2 was made with 17% H2. A constant sulfur
content of 3.5% was attained in both cases after 45
minutes' exposure to hydrogen.
Solvent Extraction Studies -
An experimental program was completed to obtain design
information for the evaluation of extraction systems.
Measurements of sorption rate, pore volume, surface area,
and residual sulfur content were made on carbons previously
loaded with 14 wt. % sulfur and extracted with (NH4)2C, /
CS2, xylene or ether. The dependence of sulfur loading on
sulfur removal was investigated by batch extractions of
carbon loaded with 14 wt. % sulfur. This work progressed
to the point of carrying out ten-stage extractions with
CS2 at 25°C, 15 wt. % (NH4)£S at 40°C, and xylene'at 105°C.
These experiments were to determine the extraction
behavior of the sulfur deposited on the carbon.
Physically sorbed sulfur would be expected to be removed
quantitatively but with increasing difficulty for sulfur
in smaller pores. Chemisorbed sulfur removal would not
be expected under extraction conditions.
It was also necessary to measure the S02 sorption rates
of these extracted carbons and of hydrogen treated carbons.
It might be expected that relative activity will be more
dependent on the removal of chemisorbed sulfur than physic-
ally adsorbed sulfur. Several pairs of extracted samples
have been subjected to high temperature (1000°F) purges,
one sample with pure helium, the other with 30% hydrogen
in the helium. The high temperature purge was an attempt
164
-------
to restore the activity of extracted carbon to the level
of virgin carbon. Ether extraction was performed to
determine if a sulfur-free extraction solvent, which could
be removed from the carbon at low temperatures, would
produce an extracted carbon retaining more of its initial
S02 sorption activity than the other solvents tested.
This was primarily aimed at gaining insight into tempera-
ture and solvent deactivation effects, not at any potential
commercial use of ether as an extraction solvent.
A literature survey on the removal of sulfur by extraction
was made. The results of the survey are given in detail
in Appendix A-17. Included are sulfur-solvent equilibrium
data for ammonium sulfide, carbon disulfide, xylene,
benzene, toluene, and pitch oil.
Ammonium Sulfide Extractions -
After a series of bench scale extractions, it was decided
to run the ammonium sulfide extraction in the closed
circulating system (see Section 5.1.7) under an inert
atmosphere to avoid any possibility of oxidation of
(NH4)2S to sulfur. It was found that both carbon disul-
fide and ammonium sulfide decompose to some extent to pro-
duce sulfur when contacted with virgin activated carbon.
This effect was observed even when the carbon had been
pretreated with N2 at 1800°F to remove oxygen, Table 39.
Table 39. EFFECT OF SOLVENT ON VIRGIN CARBON
Virgin Carbon C-70-77
Inert
Gas
NO
NO
He
N2
Carbon
Degassed
No
No
No
1800°F
1 hr.
Solvent
Wash
IxCSz
30 min.
lx(NH4)2S
30 min.
lx(NH4)2S
30 min.
lx(NH4)2S
60 min.
Steamed
No
No
1 .5 hrs.
200°F
8 hrs.
250-275°F
0.4% S
2.2% S
4.0% S
1.7% S
1 .8% S
165
-------
A ten stage extraction was run in the closed, circulating
system. The sulfur level was reduced to 2.08% after the
final extraction, washing and steaming. The equilibrium
for this system is shown in Figure 56.
Figure 56.
Extraction of sulfur loaded activated
carbon with 15 weight % (NH4)2S
solution at 40°C
Numbers refer to
successive contacts with
fresh 15% (NH4)2S
.001
9 s/g Carbon (Solvent Free)
166
-------
Carbon Bisulfide Extraction -
Carbon disulfide extractions were carried out to determine
if carbon disulfide would be superior to ammonium sulfide
as an extractant. As shown in" Table 39, carbon disulfide
contact with virgin carbon was found to have a higher
level of residual sulfur (2.2% sulfur) on the carbon than
did ammonium sulfide (1.8% sulfur). The decomposition
of CS2 on activated carbon has been enountered previously
in the operating experience of textile companies.
These companies have, for many years, used large scale
activated carbon units for carbon disulfide recovery and air
purification. The adsorbed carbon disulfide tends to be
hydrolyzed to a small degree, releasing hydrogen sulfide
which is oxidized by the air to sulfur. Since this sulfur
builds up on the carbon and reduces sorption efficiency,
steps are taken to remove it. In the early carbon processes
recovering carbon disulfide, this sulfur contaminant was
removed by extraction with aqueous sodium sulfide; but
present fluid bed recovery plants remove this by continu-
ously stripping a slip stream of the circulating carbon
with hot inert gases.
The carbon disulfide extractions were all run on the bench
and closely paralleled the results of the ammonium sulfide
extraction. The ten stage extraction resulted in a
residual sulfur level of 2.6% on the carbon. The equilib-
rium data for this system are shown in Figure 57.
Xylene Extractions -
A series of bench scale extractions with xylene were
carried out in an attempt to obtain lower sulfur levels.
The extraction was carried out in five stages, the carbon
oven dried and the extraction continued for five addi-
tional stages. The sulfur level was found to be nearly
independent of sulfur concentration in the solvent. It
may be suspected that xylene would be strongly sorbed in
the smallest pores and that the residual sulfur would
also be concentrated there. A carbon loaded with 14%
sulfur was taken to 3.7 g. S/g. carbon after a single
contact with xylene at 105°C. Four subsequent contacts
for a total of five reduced it to 2,9 g. S/g. carbon. At
this point, the carbon was dried in an attempt to bring
to the carbon surface a portion of the sulfur contained
in the xylene trapped in the smaller pores. This was
indicated by both an increase in sulfur concentration
in the xylene and a sharp drop in sulfur concentration
on the carbon in the sixth extraction. Subsequent extrac-
tions followed the pattern of the first five stages.
After the tenth stage, the residual sulfur level was
reduced to 1.5 g. S/g. carbon. The equilibrium for this
system are shown in Figure 58.
167
-------
Figure 57.
Extraction of sulfur loaded activated
carbon with CS2 at 25°C
.03
.02
.0001
.04 .06 .08
g. Sulfur/g. Solvent Free Carbon
.10
.12
168
-------
Figure 58
Extraction of sulfur loaded activated
carbon with xylene at 105 °C
I
r"
O
to
cr.
O>
• vv
.04
.02
.01 •
.008
.006
.004
.003
/\f\O
• \)\)c*
.001
.0008
.0005
.0004
.0003
.000?
S
'£
to
c
•r-
C
*l —
r—
1
_
O 6
t
D°
8
1
i
9
J,.
c
?'
1
1
r
1
i
i
1
1
' ^
5
3
' NOTTS:
1. Numbers refer to succes-
siv
xyl
2. Car
and
3. Sta
FHS
Rof
e contacts w
ene.
bon dried be
6th stages.
rting carbon
-9 - [112 - 1
. I AS 3175:1
ith rresh
two en Gth
sample
«', Sulfur ~
0-11
'.01 .02 .03 .01
g. Sulfur/g. Solvent-Free Carbon
.05
169
-------
Consideration must be given also to the implications of
xylene extraction on the regeneration sequence in a stack
gas treating process. Prior applications of xylene extrac-
tion have been in applications where trace amounts of
xylene in the treated gas were not a serious drawback. Our
experience in solvent emission control indicates that a
final stripping of xylene extracted carbon at 500 to 600°F
would likely be required to minimize xylene loss into the
circulating carbon stream and subsequent transfer to the
flue gas.
Comparative S0£ Sorption Rate of
Solvent Extracted Carbon Samples -
Measurements of the relative S02 adsorption rates were
made on various extracted carbon samples and carbon samples
of related interest. These measurements were made in the
differential rate apparatus as described in Section 5.1.1
It appears that the activity of carbon for S02 adsorption
is not a simple function of residual sulfur. One would
conclude that there are differences between treat-
ments required to restore activity for H2S pickup and that
required for S02 pickup. Literature data indicate that
extractions with (NH4)2S and CS2 as performed here would
have restored most of the carbon's activity for H2S
oxidation. In H2S oxidation, the carbon may also be loaded
with an equal weight of sulfur and still pick up H2S at a
satisfactory rate.
Solvent extraction of sulfur from carbon decreases the
carbon's activity for S02 sorption. The relative activi-
ties of sulfur loaded carbons extracted in 10 stages with
xylene, (NH4)2S and CS2 and with ether in the Soxhlet
Extractor were measured on the sorption apparatus.
Further treatment of these samples included thermal strip-
ping under helium and under a mixture of 3070 hydrogen and
70% helium. The results of these measurements are given
in Table 40 along with surface area and pore volume
distribution comparisons.
The various solvent extracted samples were thermally
stripped with an inert gas when it became apparent that the
residual xylene on the carbon was interfering with the S02
activity. Treatment at 555°F did not remove all the xylene
and it was necessary to go to 1000 F to remove all the
solvent. With this increase in activity, the (NH4>2S and
CS2 extracted samples were similarly stripped and these
activities also improved.
170
-------
Table 40. COMPARISON OF S02 ACTIVITY AND SURFACE AREA
AND PORE VOLUME MEASUREMENTS
1
Sa=ple
3C7-1
ETH-1
SXA-1
iXA-2
ZZA-3
ZXA-1.
ZZA-5
2X3-3
EX^i
122-5
j ZXC-1
EXC-2
J3C-3
i
I 3C-1
i
Material
Virgin Carbon
Sulfur Loaded Carbon
Sulfur Loaded Carbon
Sulfur Loaded Careen
Sulfur Loaded Carbon
Sulfur Loaded Carbon
Sulfur Loaded Carbon
Sulfur Loaded Carbon
Sulfur Loaded Carbon
Sulfur Leaded Carbon
Sulfur Leaded Carbon
Sulfur Leaded Carbon
Sulfur Loaded Carbon
Sulfur Loaded Carbon
L_ .
Treatment
Solvent
None
Kone
IC-Stage Xylene
10-Stage Xylene
10-Stage Xylene
10-Stage Xylene
(2nd Sun)
10-Stage Xylene
(2nd Run)
10-Stage (NHj,)2S
10-Stage (Iffii^gS
10-Etage (1IH^)2S
10-Stage CS2
10-Stage CS2
10-Stage CS2
Staer (Soxhlet)
Stripping
Hone
None
—
555°F
1000°F
1000°F
1000°F
—
1000°F
1000°F
1000°F
1000°F
—
Stripping
Gas
Kone
Ncne
—
Helium
Eelium
EeliuB
30? Eydrogen
70? Eelium
—
Helium
30? Hydrogen
70? Helium
Helium
30? Eydrogen
70? Keliua
—
? Activity '•% Surface Area
Coapared to '• Compared to
Virgin Carbon Vireir Carbon
100.0
7.5
9.1*
1.8.0
1.6.0
1.1.9
56.1
78.6
77.6
1*5.1;
63.0
57.2
1*6.1
100.0
1<5.3
It It. 8*
20.8
80.9
98.9
93.0
96.6
81*. 5
90.9
101.9
91-6
93.6
19-7
% Pore Volume Coarared to Virgin Carbon
?otal Pores
r<500^- Radius
100.0
1.7.7
1*7.9
21.. 3
81.0
'99-7
91*. o
98.1
83.3
91.1
102.2
91.2
92.8
23.3
Pores
100
-------
From earlier work on hydrogen treatment of regenerated
samples, we found a return of activity for SO? when the
sample was stripped at 1000°F with 33/0 H2 in helium.
In addition, a similar sample treated at 1500°F increased
its activity 65% over the virgin material. In view of
these earlier results, we decided to run parallel extracted
samples in which 30 volume 70 H2 was added to the helium
purge gas at 1000°F. The activities remained essentially
unchanged with H2 added to the purge gas. However, the pore
volume, surface area and residual sulfur measurements indi-
cate that the bench scale stripping conditions chosen were
not sufficient to give good sulfur removal from the smallest
pores. The hydrogen flow was much lower than used in earlier
work on hydrogen regeneration. Replicate runs will be made
at higher purge rates to allow further evaluation of the
solvent extracted carbon's ability to be restored to its
original activity by hydrogen treatment.
Conclusions which may be drawn from the data at this
point are as follows. Xylene appears to be an inferior
extraction solvent for sulfur, due to the difficulty in
removing it from carbon and some residual solvent deacti-
vation effect which persists, even through a high tempera-
ture hydrogen treatment. Ammonium sulfide and carbon
disulfide appear to be close in effectiveness. However,
the ammonium sulfide appears to have a lesser residual
solvent deactivation effect when the carbon is given
either a high temperature treatment with inert gas or
hydrogen.
Indications are that ether has no less of a solvent deacti-
vation effect on carbon than does carbon disulfide, and
the solvent deactivation effect is not strongly influenced
by the presence of sulfur in the solvent structure.
While it appears that a return of surface area or pore
volume will not guarantee a return of activity for S02, a
significant loss of either will indicate a loss in activity
for S02- Similarly, high residual sulfur content will
tend to cause activity loss, but poor activity can persist
even at low residual sulfur levels.
The indication that iso-thermal regeneration is as effective
as high temperature regeneration is somewhat clouded by the
mild regeneration conditions. As one attempts to return
carbon to its original activity, chemisorbed oxygen prevents
this unless higher temperatures are used. Hydrogen is
believed to remove the chemisorbed oxygen at lower
temperatures. Chemisorbed oxygen also has a cumulative
effect under recycle conditions.
172
-------
Recycle Work - Extraction Regeneration
vs. Thermal/Reduction Regeneration -
Two samples of carbon were recycled through the adsorp-
tion and regeneration steps to allow a direct compari-
son of the isothermal (extraction) and thermal regeneration
sequences. The experiments also showed the effect of
recycle on the S02 activity of the carbon for each of the
two regeneration sequences. The results of the study
were: 1) that the S02 activity of thermal/reductive
regenerated carbon leveled off at about 92% of the activity
of the virgin carbon, and 2) that the S02 activity of the
isothermally [(NH4)2S] regenerated carbon decreased gradu-
ally with each cycle and after 6 cycles was at about 34%
of the activity of the virgin carbon. Pore volume meas-
urements showed a decrease in the total pore volume and
surface area of the isothermally regenerated sample and
an increase in total pore volume and surface area of the
thermally regenerated sample. The S02 sorption differen-
tial rate apparatus was shown to be useful in assessing
various treatments of the (NH4)2S extracted sample after
six cycles.
Details of the recycle experiments are given in Appendix
A-15.
Experimental Results and Discussion -
A recycle series was run to compare the isothermal
(extraction) and thermal/reductive regeneration sequences.
Sample A will be referred to as the isothermally regene-
rated carbon and Sample B referred to as the thermally
regenerated carbon. Samples A and B were loaded with acid,
reacted with H2S to form sulfur on the carbon, and then
regenerated by (NH4)2S extraction or thermally with H2,
respectively. The sequence in the cycling was as follows:
A. Isothermal B. Thermal/Reduct ive
1. S02 sorption 1. S02 sorption
2. Sulfur generation 2. Sulfur generation
3. Sulfur extraction 3. High temperature
with (NH4)2S followed H2~S reaction on
by steaming carbon
173
-------
Each sample was then passed through successive complete
cycles. The S02 sorption step on each cycle provided a
means of comparing S02 activities to the virgin material.
As a basis for comparing S02 activity, the acid loading at
210 minutes relative to the virgin material was defined
as the S02 activity.
SOo Activitv = Acid Load, at 210 Min. for Sample A or B for Cycle i
2 y Acid Load, at 210 Min. for Virgin Carbon
The rate of S02 sorption is a function of temperature.
Variation in the average sorption temperature for
any cycle was corrected back to average temperature tor
loading the virgin material with acid. The Westvaco equa-
tion was used as a guide to make this correction,
TF, Temperature Correction Factor, = 552°-(TY*rgin ~ Tcycle i) (43)
Tvirgin TCyCie i
where Tvirgin = sorption temperature at which
virgin material is loaded, °R
Tcycle i = sorption temperature at which
sample is being loaded in cycle i
The sulfur remaining on the carbon after isothermal or thermal
treatment varied between 2 and 470 S. In order to compare
S02 activities at comparable sulfur loadings,the S02
activity was corrected to a common sulfur level of 270.
Previous work, in which varying amounts of sulfur were
sorbed on activated carbon, was used to make this sulfur
correction. The results of the work are shown in Figure
59. The sulfur correction factor, Sp, is given as
SF, S Correction Factor to 2% S Basis, = Activity at % Si
Activity at 2% S
Activity at % Sj
0.909
The recycle work with the respective correction factors is
summarized in Table 41. The final S02 activity is calcu
lated by multiplying the uncorrected S02 activity by
174
-------
Figure 59
Effect of percent sulfur on carbon
on the S02 activity
1.0
0.95.
r:
o
o
c
0.85- •
•r-
4->
to
•£
•r-
2 0.75--
O
10
0.65
3 4
% S on Carbon
175
-------
Table 41. EFFECT OF RECYCLE ON S02 ACTIVITY FOR ISOTHERMAL
AND THERMAL/REDUCTIVE REGENERATIONS
•v-
'•~~er
Sul
A
~ r* o -*
210 :•:
g=s Hasc
furic
eid
ir.£ at
i-utes,
I;/— Car con
Un corrected
SCj
Act^v-ty
3XT-S-IA"
-2A
-3A**
,
-5A
-6A
i
0.155 i 1
0.1^-0
0.1^5
0.1^^
0.136
0.140
0.903
0-935
^
0.929
0.877
0.903
2C
C.'^.
1C
«. s
2C
2C
IS
Arr:oni^- S'^lfide Sxtre
5XT-7-1A***
-2^.
-3A
-UA
-5A
-6A
0.167
0.077
0.090- /
0.072
0.05^
0.062
J.
0.461
•0.539
0.431
0.323
0.371
i
2C
2C
2C
2C
2C
2C
Avera-o 5"-? • I
C|r--v— s- -
-
-.-
j. CIT."0 S T 2,'w "J^T S
,-^__=^^.-,_
'-0-s-
*-*^t_ A <4 • J
Factor
Sulfur
Correction
Factor to
£'" S
Bisis
SO
Activity
riegensrc-vxcns
^
o
* _>
.8
.2
.5
. 2
|
|
i (^ '.
-L j C.-
1.0233 i 2.6
- ^ _ ^
0.9^39
0-9766
0.5515
? •?
2.6
3.2
3.6
0-909
l.C^l
. 1.0-32
1.0~5
1.077
1.181
0.909
0-962
0.905
0.915
0.922
0.9^0
:ion Regenerations
.0
i
, Lf
^ T
^
* s *
.5
.2
1
0-9679
0-9535'
1.0063
0.9956
0.9775
0.4
U.I
3.1
2.1
2.9
2.0
0.909
1.143
1.072
1.053
1.057
1
0.909
0.510
0.563
0.457
0.339
0.363
*Virgin carbon is -!A;and -6- signifies therzal regeneraticr..
**H\sidifier teipsrature vas low, correction factor for lover moisture content using Westvaco
Bc-uation = 1.0712.
» "^
ss*'/ir£in carbon is -1A; and -7- signifies isothermal regeneration.
-------
the two correction factors for temperature and sulfur
level. The S02 activity is shown as a function of the
cycle number in Figure 60. As can be seen from Figure 60
and Table 41, for the thermal/reductive sequence the S02
activity of the carbon levels off somewhere near 92?0 of
the virgin carbon activity. The virgin carbon activity
was corrected to the 2% sulfur level. The S02 activity
of the carbon, which was regenerated isothermally using
(NH4)2S, decreased from about 6070 to 34% of the original
carbon activity after 6 cycles. It is obvious that this
rapid decrease in S02 activity is an undesirable aspect
of extraction. Therefore after 6 cycles the recycle
experiments were halted to attempt to find an appropriate
treatment of the solvent extracted carbon which would
return its activity.
Pore volume distributions were run on sixth cycle Samples
A and B to gain a possible insight into the deactivation
process. The results are given in Table 42. It can be
seen that there is a definite decrease in the total pore
volume and surface area of the sixth cycle isothermally
regenerated (extracted) carbon (RXT-7-6C) and an increase
in the total pore volume and surface area of the sixth
cycle thermally regenerated carbon (RXT-6-6C). The pore
volume and surface area measurements have seemed to indi-
cate only gross effects in S02 activity reduction. That
is, a significant drop in pore volume or surface area
definitely results in a decrease in S02 activity. Small
changes, however, do not necessarily indicate a decrease
in S02 activity.
As mentioned above, the recycle experiments were stopped
after 6 cycles to assess an appropriate treatment of the
(NH4)2S extracted carbon to restore its activity. Since
there was only about 30 grams of this sixth cycle carbon,
it was desirable to use as small a quantity of the sample
as possible, but a large enough sample to assess each
treatment experiment. It was decided to use the S02
sorption differential rate apparatus discussed in previous
sections to measure S02 activity, as only 0.1 g. samples
are required. As a check on the fixed bed S02 sorber,
two each of the thermally regenerated (KXT-6-) and iso-
thermally regenerated (RXT-7-) carbon cycle samples were
used. The results of the S02 activity found for each
of the corresponding runs in Table 41 are given in
Table 43. As can be ssen from the table, reasonably good
agreement of the S02 activity was obtained from the
integral rate determined in either apparatus.
177
-------
Figure 60. Effect of recycle on S02 ability for isothermal
and thermal/reductive regenerations*
c
o
CJ
c
>
4->
TO
>
/•i—
4->
U
-------
Table 42. PORE VOLUME DISTRIBUTION RESULTS USING ENGELHARD
ISORPTA APPARATUS
Sample
EXV-1
(Virgin Carbon)
RXT-6-6C
(Thermal)
RXT-7-6C
(Extracted)
7o Pore Vol. Compared to Virgin Carbon
Total Pores
<500A
100
118
93.8
Pores
100-500A
100
77.1
67.0
/
Pores
10-100A
100
114
95.4
Pores
<10A
100
121.0
94.2
70 Surface Area
Compared to
Virgin Carbon
100
119.0
92.6
Table 43. S02 ACTIVITIES INTEGRAL RATE DETERMINED
USING DIFFERENTIAL RATE APPARATUS VERSUS
USING FIXED BED
Run
^•^^^^^^•-^••••••••M—l^— •M*«v^«MMM«
RXT-6-4A
RXT-6-6A
RXT-7-4A
RXT-7-6A
S02 Activity for
Differential Bed*
^^^^B^^^mmrH**^^^^***—*^^^*^^'^^**^******'^*^^^"*******^*"1**'^^"^^*'^^***
0.834
0.818
0.426
0.397
S02 Activity**
Differential Bed
0.866
0.966
0.449
0.397
Fixed Bed
0.915
0.940
0.457
0.363
*Uncorrected to 2% S basis.
**Corrected to 2% S basis.
179
-------
Treatments To Restore S02 Activity
of Solvent Extracted Carbon -
The sixth cycle carbon sample from the isothermal regene-
ration recycle studies which had about 34% of its original
activity was put through a number of post treatments, which
were assessed as to their success in restoring the S02 sorp-
tion activity. The SC>2 sorption differential rate apparatus
was used to make the evaluation of the success of each of
the post treatments. The experiments included both iso-
thermal and thermal treatments. The runs, respective post
treatments, and activity measurements are summarized in
Table 44.
From the data given in Table 44 it can be seen that the
post treatment experiments included the determination of:
1) Isothermal NH40H treatment to remove the
sulfate or other possible deactivation
species
2) Effect of treatment time with H2
3) Effect of treatment temperature
4) Effect of using heat only
5) Effect of using CO as the reductant.
The isothermal treatment with NH40H was unsuccessful in
returning the S02 activity of the solvent extracted carbon.
In fact, some further deactivation was noted. For compari-
son purposes virgin carbon showed a drop in activity to
60% of its original value upon being treated with the
NH40H experimental step.
The effect of treatment time using H2 at 1000°F is also
shown in Figure 61. The increase in the SC>2 activity is
less rapid with increasing treatment time. It appears that
the optimum time in a fixed bed would be 4 to 6 hours.
The effect of temperature using !!£ for a 4 hour treatment
is also shown in Figure 62. The S02 activity of the
(NH4>2S extracted sample increases rapidly with treatment
temperatures from 800 to 1200°F. SO? activities of more
than 130% of the original value can be obtained at 1200°F
with a 4 hour treatment time. It appears that the lower
limit of temperature for H2 treatment might be near 800°F.
It is seen from the experiment using N2 at 1000°F that
although heating is giving some increase in S02 activity,
the H2 is playing a definite added role of reactivation.
The N2 treated sample at 1000°F had an S02 activity of
180
-------
Table 44. EFFECT OF POST TREATMENTS OF (NH4)2S EXTRACTED SIXTH CYCLE
CARBON ON S02 ACTIVITY
Eun
Kister
16C-G
197-G
191-G
189-G
193-G
192-G
191-G
195-G
196-G
198-G
Sample
Virgin
RXT-7-6C
HXT-7-6C-U
HXT-7-6C-1
RXT-7-6C-6
RXT-7-6C-5
RXT-7-6C-7
RXT-7-6C-3
KXT-7-6C-9
Virgin
Treatment
Gas
„
305? H2
305? H2
30/5 K2
30/1 K2
K2
30? CO
°F '
•800
1000
1000
1200
1000
1000
Tine,
hrs. .
U
2
U
h
h
k
2% NH1,OH;
Purged l.-hr. with N2 at UOO°F
. 2% KKj-OK;-
Purged 1 hr. vith N2 at UOO°F
H2£0!4
Loading**
nss. Acid/em . C
173-90
63-17
116.69
16C.37
I81t.95
23«..'*9
88.37
131.29
58.56
119-32
Uncorrected
SC2"
Activity
l.CCC
0.363
0.671
0.925
1.06U
1.3.3
0.508
0.755
0.337
0.636
^r
o.u
2-1
0.8
0.9
0.8
0.8
0.6
1.0
l.U
0-7
Sulfur
Correction
Factor*
0.988
1.076
1.000
1.005
1.000
1.000
0.988
1.005
,.03,
0.995
502 Activity
Corrected to
0.8:* S Basis
0.983
0.391
0,671
0-929
1.06U
1.31:3
0.502
0.759
..*.
0.655
00
*Using relationship developed of effect of sulfur on S02 activity.
**To 210 minutes sorption time.
-------
Figure 61
Effect of treatment time using hydrogen post
treatment of (NH4)2S extracted sample of
sixth cycle
Temperature:
Hydrogen Concentration:
Carbon:
iono°F
30 volume %
RXT-7-6C
0.2
Treatment Time, hours
182
-------
Figure 62
/; ?f temPerature of hydrogen post treatment
ot (NH4)2S extracted sample of sixth cycle
Treatment Time: 4 Hours
Hydrogen Concentration: 30 volume
Carbon: RXT-7-6C
I
600 800
Treatment Temperature, °F
1000
1200
183
-------
about 50%, whereas the H£ treated sample at 1000°F had
about 90%, of its original activity of the virgin carbon.
The CO treated sample had an activity of about 75% which
lies somewhere between the two extremes of thermal treat-
ment only and thermal/reductive H2- This gives even more
support to the definite beneficial effect H£ has on
returning the activity of activated carbon.
5.2.4 H2S Generation
Hydrogen is the basic reductant that is used to reduce the
sulfuric acid sorbed on the carbon. It was found, however,
that a secondary reductant produced from the hydrogen,
hydrogen sulfide, is a better reductant of the sulfuric
acid. Sulfuric acid reduction was found to occur in the
temperature range of 200 to 300°F, as discussed in a
previous section, so process development proceeded toward
producing this recycled reductant within the process. The
initial process concept was to vaporize the sulfur, origi-
nating from the S02 in the flue gas, from the carbon.
The remaining recycled sulfur on the carbon was then
reacted with hydrogen to form the hydrogen sulfide needed
to reduce the sulfuric acid to elemental sulfur. The
major deviation from that initial concept is that the two
unit operations of sulfur stripping and H2S generation
have been combined successfully into a single fluid bed
reactor as discussed in a later section.
In this section,the bench scale data that was developed on
the kinetics of H2S formation from hydrogen and sulfur vapor
over a catalyst of activated carbon is discussed. The pilot
work that was done on this unit operation is discussed in the
later section on combined sulfur stripping/H2S generation.
In addition to the use of hydrogen for sulfuric acid reduc-
tion by way of hydrogen sulfide, other possibilities of
hydrogen requirements are hydrogen to remove chemisorbed
oxygen from the reused activated carbon and, hydrogen
chemisorption by the regenerated activated carbon.
The chemisorbed 02 is discussed under the integral run
results. The studies of H2 chemisorption on activated carbon
are discussed in this section.
Fixed Bed Studies H2 Chemisorption -
The objectives of the study were to ascertain the extent of
H2 chemisorption on carbon and the effect of cycling the
carbon. The general approach was to evaluate H2 chemisorp-
tion by following the gas phase analysis of H2 during
exposure of virgin carbon in a fixed bed.
184
-------
The next step, not initiated, was to start a recycle bench
scale experiment in which virgin carbon is loaded with S02,
reacted to form S, then exposed to H2 to convert the
sulrur to H2S and to measure the H£ chemisorbed. During
the runs the off-gas from each process step would be
analyzed by gas chromatograph. The carbon at the end of
each cycle would be analyzed for total sulfur content and
tor S02 activity in a differential S02 sorber as a check
on the S02 sorption step. It was anticipated initially to
complete six cycles. This second part of the study was not
completed because of the delay that would follow in complet-
ing the integral run.
The major response was the amount of H2 chemisorbed either
on virgin carbon or on recycled carbon. Secondary responses
were the S02 activity of recycled carbon and the effect of
recycle on conversion of sorbed H2SC-4 to S and of the sulfur
to H2S, and on H2 chemisorption.
H2 Chemisorption on Virgin Carbon -
The virgin carbon was heated to about 1850°F in an inert
atmosphere to purge off any chemisorbed oxygen. The sample
was then cooled to the indicated temperature under an inert
purge and H2 introduced to the reactor at the indicated
concentration. During all phases of treatment the off-gas
was analyzed using a gas chromatograph.
Recycle Experiments - H2 Chemisorption on Recycled Carbon -
Virgin carbon was to be loaded with H2S04, reacted with
H2S to form elemental sulfur, and then treated at 1200°F
with 30 vol. 70 H2 to form H2S and measure chemisorbed H2-
During each process step the off-gas was to be analyzed
with the gas chromatograph or S02 analyzer. The carbon
was then to be analyzed for S02 activity and recycled
through each process step again.
Run Conditions -
The run conditions initially anticipated are listed in
Table 45.
The first H2 Chemisorption studies on the bench scale
were planned using virgin carbon, but the original equip-
ment and procedures used were unsatisfactory for detecting
H2 Chemisorption. Five experiments indicated that either
no H2 was chemisorbed or that the process was so fast that
the Chemisorption could not be detected with the restraint
that a gas analysis could only be made every 12 minutes.
185
-------
Table 45. PLANNED EXPERIMENTAL PROGRAM FOR STUDYING HYDROGEN
CHEMISORPTION ON ACTIVATED CARBON DURING
REGENERATION OF THE CARBON
Run
HCS-1
HCS-2
IICS- 3
HCS-ll
HCS-5
Run
Inlet
Hg Cone.
Vol. %
38
38
38
20
10
*
HCS-u-A-_
Run
HCS-6-B-_
Run
HCS-6-C-_
Grams
Carbon
15-2
16.2
17.3
15.2
15.2
Inlet
S02
PPM
2500
HO
PPM
150
Lin. Gas
Velocity
ft. /sec.
0.20
0.20
0.20
0.20
0.20
Gas Cone.
Vol. %
N2
Bal.
Inlet
HgS Cone.
Vol. %
30
Inlet
H2 Cone.
Vol. %
30
02 H20
3.5 10
Grams
Carbon
Space
Velocity
1300
1300
1300
1300
1300
Grams
Carbon
15
lit
Grams
Carbon
15
Temp.
1200
1000
800
1200
1200
Effect
Temperature
M
tt
Inlet H2 Cone.
11
Lin . Gas Space
Velocity Velocity
ft. /sec. hr."1
0.18 3,000
Lin. Gas
Velocity
ft. /sec.
O.Oll*
Lin. Gas
Velocity
ft. /sec.
0.20
Space
Velocity
hr.-1
200
Space
Velocity
hr.-1
1300
OF
200
Temp.
300
Temp.
op
1200
Purpose
H2
Chemisorption
on Virgin
Carbon
Purpose
Effect of
Recycle
on H2
Chemisorption
*Run HCS-6- -
Denotes cycle number
A denotes S02 sorption.
*• B denotes sulfur generation.
C denotes H2S generation.
186
-------
Therefore a different equipment setup was used which was
composed of a thermal conductivity cell as used on a gas
chromatograph. A schematic of the equipment used is shown
in Figure 63. A major difference in this setup is that
the off-gas from contact with the carbon is analyzed
continuously. The experimental procedure was as follows:
1) The virgin carbon was heat treated at 1800°F
with 120 scc/min. of N2,(R-l) for 3 hours.
2) The sample was cooled to room temperature at
the same N2 flow rate.
3) The detector was calibrated. The H2/N2 mixture
(about 30%) was passed through the reference
side of the detector and known mixtures were
passed through the other side of the detector.
The resultant responses were plotted and
served as a calibration curve.
4) At time zero, the N2 (R-l) was cut off and the
H2/N2 mixture (R-2, R-3) was passed thrqugh the
bed when the response from the detector was
zero, the sample was heated to 1200°F. The
resultant deflection is proportional to the
amount of H2 chemisorbed.
Two experiments were run by this experimental procedure.
The results of the experiments are summarized in Table 46.
The experiments differed in the heat-up rate from 100 to
1200°F, namely 20 and 75 minutes. At the inlet hydrogen
concentration of about 30 volume % at a total gas flow rate
Table 46. H2 CHEMISORPTION ON VIRGIN CARBON
Run
HCS-8
HCS-9
Heat-up
Time to
1200°F,
mi nutes
20
75
Initial
Carbon
Charge,
gms.
15.2
15.2
Total Gas
Flow
Rate,
scc/min.
49
50
Inlet
H2
Cone.,
Vol. %
30.6
31.7
Total
H2
Pick-up,
sec
85
101
Total H2
Chemisorbed,
moles H2/100 gms C
0.026
0.030
187
-------
Figure 63.
Equipment schematic for H2 chemisorption
experiments on virgin carbon
Recorder
Power Supply
Temperature
Potentiometer
Thermal
Conductivity
Detector
T T
Toggle Valves
188
-------
of 50 scc/min. the amount of H£ chemisorbed was about
0.028 moles H2/100 gms carbon. In earlier experiments,
the amount of H2 unaccounted for was of the order of 0.1
moles H2/100 gms carbon. Therefore only about 30% of
the necessary H2 is accounted for if the H2 chemisorption
phenomena is similar on virgin carbon and sulfur loaded
carbon.
There are a number of factors which make determination of
H2 chemisorption on sulfur loaded carbon difficult to
determine on the bench scale. These include the following:
1) The bench scale is at unsteady state.
2) The phenomena of H2 chemisorption is apparently
fast at the temperatures of H2S generation.
3) The system for continuous analysis (Figure 63)
is only suitable for a two component system
(inert of N2 + one other component such as H2),
but during H2S generation,H2S is also formed,as
well as a number of other possible compounds.
4) All of the problems mentioned above would prob-
ably lead to difficulties in setting up an ana-
lytical system with available equipment for
studying H2 chemisorption as a function of
cycling on the bench scale.
5) Because of the difficulties mentioned in Step 4
a delay in achieving integral operation was
anticipated.
Further analysis of the impact of the possible increased
H2 requirements due to H2 chemisorption was made as to the
effect on overall process economics. The process economics
were assessed on the basis of 20% additional H2 required.
The results based on a previous economic analysis by
Westvaco, which should at least indicate the relative cost
increases even if the economic bases may have changed
somewhat. There is a projected increase in costs of
about 1.7% in the capital costs and about 3% in the
operating costs for a 1,000 MW installation.
At this time a program that lead to integration and opera-
tion of the pilot plant for repetitive cycling of the carbon
was being pursued. It was felt and shown that this opera-
tion gave a good indication of the hydrogen requirements on
a long term basis. Preliminary bench scale tests indicated
189
-------
that complete definition of the hydrogen use in an
unsteady state system would require significant effort
and instrumentation not available at this time.
Any further effort on bench scale experiments would have
delayed the integration of the pilot plant and it was felt
that further consideration of bench scale work was not
justified. This was based on the minimal impact of the
increased hydrogen use on economics and anticipated data
that was obtained in the pilot plant.
H2S Generation Kinetic Studies -
The first runs were made to find the effects of H£ concen-
tration, and sulfur vapor concentration on the formation
of H2S with and without activated carbon as a catalyst.
Conditions for these runs are listed in Table 47.
Table 47. EXPERIMENTAL CONDITIONS FOR H2S
GENERATION RATE EXPERIMENTS
Run I
Run II
Inlet H£ Cone.
Inlet S Cone, as Si
Total Flow Rate
Reaction Temperature
Linear Veloc. @ 1000°F
Bed Depth
Space Veloc. @ 1000°F
Carbon Weight
Carbon Type
Empty Reactor Volume
Carbon Bed Volume
28.5% 18.9%
6.36% 7.13%
645 cc/min. STP 625 cc/min. STP
1000°F 1000°F
0.22 ft./sec. 0.21 ft./sec.
0.2 inch 0.2 inch
47,800 hr."1 46,200 hr.'1
1.65 gm
Virgin WV-W 12x40, Log C-70-30
0.00273 ft.3
0.0000908 ft.3
Homogeneous Reaction -
The reactor system was constructed with two identical
reactor tubes, one of which contains a carbon sample and
another which is empty. Reference to the homogeneous reac-
tion indicates reaction in the empty tube although, of
190
-------
course, there may be some influence of the tube wall. Rate
determinations were made by measuring the steady state
concentrations of H2S in the effluent gas and calculating
rate of formation using the total gas flow rate. Based on
the reactor volume inside the reactor furnace held at
1000°F, the rates of formation were found to be 0.054 Ib.
mol/hr./ft.3 and 0.038 Ib. mol/hr./ft.3 for Runs I and II,
respectively, under the conditions listed in Table 47.
These results may be compared to the rates derived from
kinetic data of Aynsley, Pearson and Robinson^ and of
Norrish and Rideal5, In each case these authors' data
yielded good straight line plots according to the
expression
dt
- k[S]1/2[H2S]
(50)
where k =
when the calculated values of In k from their data were
plotted against 1/T over their experimental range of 550-
650°F. Using values of k so extrapolated to 1000°F, reac-
tion rates were calculated for the present experimental
conditions and are compared in Table 48.
Table 48. COMPARISON OF H2S FORMATION RATES
FROM LITERATURE AND EXPERIMENTAL
DATA (HOMOGENEOUS)
Source
Any s ley
Norrish
Experimental
Rate at Inlet Conditions at
at 1000°F (Ib. mol/hr./ft. 3)
Run I
0.00024
50.6
*
0.054
Run II
0.00017
35.5
0.038
•^•••••••^^•^^^••^•^•^^^^-^••••i • •• .*«^^fc^«j^^»Maa
k
23
4,800,000
5,024 (I)
5,270 (II)
•MK*^^— ^^^wi*B*a^bv^^^^^^^^^^^M
191
-------
As seen, the present experimental data fell between the
widely separated literature data although somewhat closer
to that of Aynsley. No adjustment of the present experi-
mental data for the difference between inlet and outlet
concentrations or the residence time in zones below 1000 F
in the reactor can account for the differences observed.
In the latter case, the inlet and outlet lines for the
reactor were heated to below 700°F, and the temperature
coefficients for reaction determined from the literature
indicate that reaction in these zones should be 500 to
1,000 times slower than at 1000°F and, therefore, negligible
Effect of Temperature, H2S .and S Concentrations
on the Heterogeneous Reaction To Form H2S with
a Catalyst of Activated Carbon
The heterogeneous reaction was carried out in an identical
reactor tube containing a small bed of activated carbon
sized to approximate the gas residence time used in 4"
pilot H2S generation work. The rate of formation of
H2S in the bed was taken as the difference between the
total H2S production from the catalytic reactor and that
found for the homogeneous reaction. At 1000°F where a
considerable amount of H2S was formed by homogeneous
reaction, this treatment may not be completely accurate
but the present results are at least illustrative.
The rates of H2S formation found were 1.5 Ib. mole/hr. /ft. •*
for Run I and 1.9 Ib. mol/hr. /ft.J for Run II under the
conditions listed in Table 47.
Table 49. COMPARISON OF HOMOGENEOUS AND
HETEROGENEOUS REACTION RATES FOR
SIMILAR INLET CONCENTRATIONS
Homogeneous
Heterogeneous
Rate H2S Formation
(Ib. , mol/hr. /ft. 3)
Run I
0.029
1.5
Run II
0.025
1.9
192
-------
A comparison to the homogeneous reaction may be made by
calculating the homogeneous rate at the reactant concen-
trations corresponding to those at the inlet to the carbon
bed which is located near the exit of the reactor tube
assuming rate = 5100[S]^[H2].
It is seen, therefore, that the H2S production rate
increased 50 - 75 times in the presence of carbon. However,
from these data it is still not clear whether the carbon
has any particular catalytic influence. For example, the
surface area of the reactor tube is approximately equal to
the external area of the carbon particles in the reactor,
so that on the basis of H£S formed per unit of exposed
surface, the "catalytic" rate is no more than twice the
"homogeneous" rate. In spite of this, it still will
be possible to obtain rates applicable to reactions
within a carbon bed as employed in practice. It should
be noted that the sulfur concentrations influent to the
carbon in these experiments are equivalent to the equilib-
rium concentration over carbon containing adsorbed sulfur
in the 10-11 Ibs. S/100 Ibs. C loading range. Such
concentrations should be typical of vapor phase sulfur
obtained during H2S generation after most of the sulfur
has been stripped off thermally.
Following the initial experiments a second set of experi-
ments was performed to further compare the homogeneous
and heterogeneous reactions, test the applicability of
the sample rate expression rj^S = k[H2][S]l'2 and deter-
mine the effect of temperature on reaction rates.
Table 50 lists the experimental conditions used. It
is noted that in the case of the heterogeneous reaction,
reactant concentrations influent to the carbon bed were
assumed to be equal to the effluent concentration from
the empty reactor. These were calculated based on the
observed conversion in the homogeneous reaction.
The experimental data was treated by calculating, for each
run, the values of the rate constant k from the observed .
rate of production of H2S, the rate equation r^S = k[H][S]_]
and the calculated averages of -the hydrogen and sulfur
concentrations between inlet and outlet. This implies
differential conditions at the average concentrations,
which is not entirely accurate.
If the assumed rate equation is valid, values of k should
be equal at a given temperature and In k should change
linearly with 1/T. Table 51 shows the values of average
193
-------
Table 50. EXPERIMENTAL CONDITIONS FOR SERIES HS-2
Uniform Conditions
Carbon Type:
Bed Depth:
Linear Veloc. at Exper. Temp
Empty Reactor Volume:
Carbon Bed Volume:
Carbon Weight:
WV-W Loaded to 24% S Run H
0.2 inches
0.22 ft./sec.
0.00273 ft.3
0.0000908 ft.3
2.255 grams
Temperatures and Concentrations
Conditions
Temperature, °F
Run
1
900
2
900
3
1000
4
1000
5
1000
6
1100
7
1100
HOMOGENEOUS REACTION
% H2 Inlet Cone.
% Si Inlet Cone.
30.1
10.0
20.0
10.0
30.1
10.1
20.1
10.1
30.2
14.5
30.0
9.93
19.9
9.93
HETEROGENEOUS REACTION
% H2 Inlet Cone.
% Si Inlet Cone.
28.7
8.29
19.0
8.84
26.6
5.73
17.4
7.07
25.7
9.07
23.2
1.20
14.0
3.35
194
-------
Table 51. COMPARISON OF RATE CONSTANTS
AT DIFFERENT TEMPERATURES
Run
No.
1
2
3
4
5
6
7
Temperature ,
°F
900
900
1000
1000
1000
1100
1100
kavg (Horn . )
xlO-3
2.07
2.05
5.40
5.45
5.68
12.2
15.1 '
kavg (Het.)
xlO-3
124
92
309
340
340
1,040
810
195
-------
k calculated for each of the runs made in the homogeneous
and heterogeneous reactors. Agreement between values of
k tends to support the assumption of the rate equation for
both homogeneous and heterogeneous cases.
Figure 64 shows plots of In k vs. 1/T for these data and
compares the results of these experiments with those
extrapolated from calculations based on the work of
Norrish and Rideal, and Aynsley, et al. The experimental
data is seen to yield straight line plots, the slopes of
which are approximately equal for both the homogeneous
and heterogeneous reactions. Activation energies calcu-
lated from these slopes are 40.9 Kcal for the homogeneous
case and 44.7 Kcal for the heterogeneous reaction. The
similarity of these activation energies indicates that while
the presence of carbon apparently increases the reaction
rate per unit volume of reactor, carbon does not have a
specific catalytic effect, except perhaps by virtue of its
surface area. Table 52 compares experimental reaction
rates and rate constants based on the reactor tube surface
area in the homogeneous case and the external carbon
surface area in the heterogeneous case. This latter area
was estimated as the surface area of spheres having the
mean particle diameter of the carbon sample. On this basis
the rate constant for the "heterogeneous" reaction was
only about twice that for the "homogeneous" reaction. This
is probably within the error of the external particle
surface area estimate.
It may also be reactor surface area which accounts for
the wide difference in results for the two literature
investigations noted. In any case it is apparent that one
aspect of future investigations must be to establish the
proper reactor size basis for rate constant calculations.
Effect of Mass Transfer on H2S Generation Kinetics -
Following H2S generation experiments reported previously,
the experimental program has continued to investigate:
1) "homogeneous" reaction kinetics with integral analysis
of data, 2) gas film resistance and the effect of linear
velocity on reaction over activated carbon.
Homogeneous Reaction - Empty Reactor -
Analysis of earlier experimental data was made assuming
that differential reaction conditions were present but it
was realized at the time that this could not be true since
conversion of sulfur was in some cases almost complete.
These data did, however, indicate that the rate expression
196
-------
Figure 64. Arrhenius plots for experimental and
literature data
ixio6 -J
ixio
id
OL
1x10
IxlO3
.50
.60 .70 .80
1/T, ("R)"1 x 103
.90 1.0
197
-------
Table 52. COMPARISON OF RATES AND RATE CONSTANTS BASED
ON REACTOR SURFACE AREA FROM SERIES HS-2
Run
No.
1
2
3
4
5
6
7
1
Rate (Horn.) x 1(P
# mol/hr. meter 2
6.14
4:10
14.3
10.0
18.2
25.5
22.1
avg k (Horn . )
470
466
1,230
1,240
1,290
2,770
3,450
Rate (Het.) x 10 J
# mol/hr. meter 2
11.2
5.91
21.2
16.1
27.2
26.6
21.8
avg k (Het.)
1,020
761
2,550
2,810
2,810
8,560
6,690
k(Hom)/kHet
2.17
1.63
2.07
2.27
2.18
3.09
1.94
00
-------
rH2S = k[H2][S] (51)
was likely candidate for further tests. To this end, the
plug flow equation was integrated using an analogous rate
expression as follows.
Derivation of Integrated Rate Equation -
Assuming reaction
H2 + 1/2 82 —*H2S (52)
then:
(53)
where N = moles reactant
No = initial moles reactant
NS2 -
where X = conversion of reaction.
Assuming the rate expression
-r = K C C 1/r2 (55)
where C = concentration.
Dividing (54) and (55) by V, and substituting:
199
-------
-rH = K[CH(l-XH)]-[Cs-l/2 CHXH]1/2 * (56)
Assuming the plug flow equation
-XH
V - / **
H>o J0 -'H
/
V V U
where u = a + bx, a = Cs, b = -1/2 CH
v = C + dx, c = CH, d = -CH
k = ad - be = Cn2/2 -
(FH)
where (F)o = initial molar flow rate
V = reactor volume
and substituting for rjj
V _ 1^ C dXfl (58)
(FH)O K JQ (cH - CHXH) (cs -1/2 cHxH)1/2
This is integral of form
(59)
Because for present data CH > 2 GS, kd < o, therefore,
proper integrated form is
I =; arctan [ " ] (60)
y v /u = / -kd • -kd
-?— arctan -CH/Cs-1/2 CHXH
XH
Ut *. Vr I.U1.1. " " ' "•
kCH
*Subscript molecular numbers are dropped, H = H2 etc.
200
-------
[ (-arctan CH^Cs-1/2 CHXH}
- (-arctan
(arctan B - arctan A)
where B = CH
A = cH 'CS/ -1/2 CHXR/ /kci
Then
arctan (, ~AB) - = (61)
where K = rate constant
k = CH2/2 - CHCS
A = CH /Cs-l/2 CHXH/ /kCH
B = CH
Confirmation of the integrated rate equation, (61) , is
obtained if the experimental data plotted as V(FH2)o vs•
the right hand expression (Z) gives a straight line pass-
ing through the origin. The reaction rate constant can
then be calculated from the slope of this line.
Experimental -
Experiments using the empty reactor tubes have now shown
that the equation does properly describe the experimental
results. In the course of these runs, however, it was
201
-------
discovered that the two matched reactor tubes, one of which
was to be used for the reaction over carbon and the other
which was to be left empty for reference did not produce
the same conversion when both were run empty. Nevertheless
the data obtained allowed testing of the proposed integrated
rate equation.
Experimental conditions were:
Temperature: 1000°F
Sulfur Cone.: 15% as Si
Hydrogen Cone.: 30% H2
Reactor Volume: 0.00273 ft 3 without inserts
0.00176 ft.3 with inserts.
The results are shown in Figure 65. Run HS-3 was for the
empty reference reactor and Run HS-4 was made using the
empty reference reactor and the empty carbon reactor, both
with Vycor inserts. This figure illustrates three findings:
1) The lines are straight and pass through the
origin, indicating a correct rate equation.
2) There is a pronounced difference in the rates
obtained using the reference reactor as com-
pared to the empty carbon reactor although
they have the same internal dimensions.
3) There is almost coincidence between the data
taken with the reference reactor with and
without the Vycor insert.
It is noted that this coincidence is related to using
reactor volume at the reactor size factor in the term
V/(FH2)O- Figure 66 shows that agreement is not as
good if the reactor size factor is in terms of internal
surface area, A, instead of volume, V.
With regard to the results of experiments in the reference
reactors it would be tempting to conclude that there was
relatively little contribution to reaction due to wall
area and that a simple homogeneous rate constant could be
best derived based on reactor volume. Yet the other
reactor tube used in Run HS-4 which was dimensionally
identical to the reference tube produced a significantly
lower rate of reaction. Comparative reaction rate constants
based on volume for the tubes are as follows:
202
-------
Figure 65. Test of integral rate equation (61), Z as
function of tube volume (Runs HS-3 and -4
reference reactor)
/ /
•/
HS-3 (Ref) -v'
/
4
/
/ ,^HS-4 (Ref)
x X
/ /
/ '
'/
.4*
5-4 (React)
_j
3
V/F
203
-------
Figure 66. Test of integral rate equation (61), Z as
function of tube area (Runs HS-3 and -4
reference reactor)
,-y
X
X
X
x
' X
,-t
* •
X
X
x — x
204
/ HS-3
X
X
X
X
X
X
X
X
X
X
'1
X
X
' HS-4
40 80 120 160 200 240
A/(FH2)0
-------
Run Reference Tube Carbon Tube
HS-3
HS-4
ll.OxlO3,
10.4xl03
4.60xl03
Whatever other conclusions are drawn from this, it is
obvious that in studies of the reaction over carbon, it is
not possible to account for the "homogeneous" contribution
to total conversion using parallel data taken with this
reference tube. Thus, the only alternative is to account
for homogeneous reaction by means of the empty tube data
for the reactor to be used in experiments with carbon.
/
investigation of Mass Transfer Effects -
In the heterogenous reaction over carbon, the rate constant
derived from bench scale data is to be applied to the pilot
fluid bed case. Since there is a significant difference
in gas linear velocity in the two cases it becomes necessary
to determine the extent of velocity dependence, since such a
dependence indicates a rate limiting resistance due to the
diffusion rate of reactants and products through the gas
film surrounding the carbon particles. An investigation
was made by measuring conversion of hydrogen as a function
of gas flow rate for three different carbon bed sizes.
Experimental conditions are as noted in Table 53.
Table 53. EXPERIMENTAL CONDITIONS
FOR RUNS HS-4 TO HS-7
Carbon Type:
Temperature:
Inlet H2 Concentration:
Inlet Si Concentration:
Carbon Bed Weights:
Carbon Bed Volumes:
Total Gas Flow Rates:
Space Velocity Range:
Linear Velocity Range:
SG-32, 24% S Loading
1000°F
30%
15%
0, 1, 2, 4 gms
0, 4.58xlO~5, 9.16xlO~5, 18.3xlO~5 ft.3
1.96 to 7.84 ft.3/hr. @ 1000°F
10,700 to 171,000 hr."1 @ 1000°F
0.173 to 0.692 ft./sec.
205
-------
The results of these experiments are shown in Figures
67 and 68. Figure 67 shows conversion of H£(XH2) as
a function of bed volume/H2 feed rate [V/(FH2)ol- If Sas
film diffusion were unimportant the data should all lie on
a single curve. It is obvious that this is not the case
here and that this resistance affected reaction rates at
all velocities used. Figure 68 shows the conversion as
a function of gas flow rate for nearly constant inlet condi-
tions and space velocity. In the absence of film resistance,
conversion would be constant with increasing linear velocity.
Since the conversion continually increases,it is clear that
these data are affected by mass transfer effects.
H2S Generation Kinetic Model -
The kinetics of H£S generation from the reaction of hydrogen
and sulfur vapor over a catalyst of activated carbon have
been studied. The effects on the reaction rate of H2S
formation, of temperature, of hydrogen concentration, of
sulfur vapor concentration, and of mass transfer effects
have been measured. This has led to a combination of all
the variables into a kinetic model relating these variables
to the rate of H2S formation as given by Equation (62) :
rH2S = k e-^^/1 (Si)J-/^ (H2) (62)
where rjj£g = rate °f ^2S formation, moles/hr.-m^
of C surface area
2.08(ig9)(y)0-5
k = (299)(T)(e-30645/T) +(v)0.5 '
moles/hr./m^ of C surface area.
5.2.5 Combined Sulfur Stripping/H2S Generation
The existing 8 stage, 4" diameter reactor was shown to
be suitable for integral operation as a combined H2S
generator/sulfur stripper. The conclusion is based on
the facts that 86% of the stripped sulfur was converted
to H2S (75% required), 81% of the H2 was converted to
H2S (90% required), and the S02 activity of the carbon
was 106% of the virgin precursor. Although 817o of the
H2 was converted to H2S, 94% of the inlet H2 was
utilized. The unaccounted for H2 was postulated as
being chemisorbed onto the carbon or reacted with chetni-
sorbed oxygen.
206
-------
Figure 67. Variation of conversion with residence
time for three bed volumes
0.4,
Run HS-7
Gas Velocity Increase
0.1
0.2
0.3
0.4
0.5
207
-------
0.4
0.3
Figure 68. Effect of flow rate on conversion
at constant residence time
HS-7
CM
0.2
HS-6
0.1
5-5
At V/FH2 = 0.087
0
8
12
(FH2),
16
20
24
208
-------
Preliminary experiments suggested that the gas/solid con-
tact time for the H2/sulfur reaction should be increased.
This equipment change was made, problems encountered in
the initial runs were rectified, and the above mentioned
objectives were met.
The overall results indicated that for sufficient sulfur
removal, for carbon regeneration for S02 pickup, and for
H2/sulfur conversion to H2S, the operating temperature
would be 1000 to 1200°F, the inlet H£ concentration would
be 20 to 40 volume %, and the carbon residence times of
6 to 13 minutes were sufficient.
It was recommended that the 8 stage, 4" diameter regenerator
be used in integral operation as a combined H2S generator/
sulfur stripper and be integrated with the 18"0 S02 sorber.
It was further recommended that runs be made on bench scale
to verify the phenomena of H2 chemisorption and that addi-
tional runs in the 4"0 unit be considered to evaluate lower
operating temperatures. The unit was used as a combined
reactor in integrated operation with the 18"0 S02 sorber.
Also, some preliminary bench scale work on H2 chemisorption
was completed, as discussed in a previous section.
Combined Regeneration Process Concept -
In the Westvaco Process, S02 is removed from the flue gas
by activated carbon. During the removal, S02 is converted
to sulfuric acid which remains as the sorbed species on
the outlet carbon. The sorbed acid is subsequently con-
verted to elemental sulfur by the reaction with hydrogen
sulfide. The next step is to recover one-fourth of the
sulfur as elemental sulfur and the remaining three-fourths
as hydrogen sulfide by reaction with hydrogen by Reactions
63 and 64:
4 S (Sorbed) Activated^ s + 3 S (Sorbed) (63)
Carbon
3 H2 + 3 S (Sorbed) Activated». 3 H2S (64)
Carbon
The hydrogen sulfide produced is used in the first step of
regeneration. The reactivated carbon is recycled to the
S02 sorber.
209
-------
Results and Discussion -
For integral operation, both regeneration steps of sulfur
stripping and H2S generation can be carried out in the
existing 8 stage, 4" diameter regenerator. The objectives
of the task were met with 86% of the stripped sulfur con-
verted to H2S, 947<> utilization of the inlet hydrogen, and :•
the SC>2 activity of the product was 6% higher than the
virgin precursor.
The conversion of H2 and sulfur to H2S increases with
increasing temperature and with decreasing space velocity
(increasing gas/solid contact time). The S02 activity of
the carbon and removal of sulfur from carbon increases
with increased residence time of the carbon in the reactor
and with increased temperature.
The experimental results for all runs made to evaluate
combined sulfur stripping/H2S generation are summarized
in Tables 54, 55 and 56 with detailed data given in
Appendix A-18. Tables 54 and 55 give the average measure-
ments for the carbon and gas phase, respectively. Table
56 summarizes the major response variables calculated
from the experimental data.
Effect of Temperature on Sulfur Removal from Carbon -
As the temperature increases, the vapor pressure of sulfur
over carbon increases. The presence of hydrogen in the gas
phase enhances the driving force for the sulfur removal,
because of the reaction of the hydrogen with sulfur in the
gas phase and in the pores of carbon. The effect of tempera-
ture on the removal of the sulfur from the carbon with H2
present is given in Figure 69. As expected, as the
temperature increased, the percentage of sulfur stripped
from the carbon in the 8 stage reactor increased. An
operating temperature of 1000 to 1200°F is indicated from
the data to provide sufficient removal rates of sulfur.
Effect of H2 concentration and Carbon Residence
Time on Sulfur Removal from Carbdn
The effect of H2 concentration and carbon residence time
on sulfur removal from carbon at 1200°F is given in Figure
70. As the hydrogen concentration increases the sulfur
removal increases as expected. Also as the carbon residence
time increases the sulfur removal increases. Carbon resi-
dence times of 6 to 13 minutes for an inlet H2 concentra-
tion of 20 to 40 vol. % appears sufficient at 1200°F to
provide sufficient removal of sulfur.
210
-------
Table 54. EXPERIMENTAL CONDITIONS FOR EVALUATION OF COMBINED
SULFUR STRIPPING/H2S GENERATION
Expt.
Number
SHG-1
SHG-2
SHG-3
SHG-5
SHG-7
SHG-8
SHG-9
SHG-10
SHG-11
Carbon
Inlet
% S
20.70
20.30
20.35
19.78
20.16
5.08
20.12
19.81
19.81
Mat'l
Rate
#/hr.
40.5
35.1
30.9
30.9
30.9
25.2
30.9
35.0
35.0
C
Rate
#/hr.
32.2
28.0
24.6
24.8
24.7
23.9
24.7
28.1
28.1
Carbon
Residence
Time
minutes
10
11
13
13
13
13
13
10
10
Outlet
% S
15.79
4.85
2.82
4.91
2.80
2,90
2.64
4.21
8.27
Mat'l
Rate
#/hr.
37.3
29.1
23.85
25.1
25.0
24.9
24.6
28.9
30.8
C
Rate
#/hr.
31.4
27.8
23.2
23.9
24.3
24.2
24.0
27.7
28.2
Cyclone
% S
28.9
42.7
25.4
31.2
Mat'l
Rate
#/hr.
0.0055
0.0036
0.0033
0.015
0.021
Temperature, °F
Stage Number
# 1
768
942
1130
1144
1135
1132
1140
1132
1070
# 2
771
970
1168
1169
1175
1168
1173
1135
1035
# 3
781
982
1178
1179
1180
1178
1186
1072
1040
# 4
778
967
1147
1140
1192
1150
1165
899
925
# 5
801
1001
1148
1170
1198
1178
1187
1120
1085
# 6
799
970
1115
1139
1025
1178
1164
1219
1207
# 7
1024
1170
# 8
734
840
914
951
910
938
918
N>
-------
Table 55. EXPERIMENTAL CONDITIONS FOR EVALUATION OF COMBINED
SULFUR STRIPPING/H2S GENERATION
Expt.
Number
SHG-1
SHG-2
SHS-3
SHG-5
SH6-7
SH6-8
SHG-9
SHG-10
SHG-11
Gas
Inlet
Linear
Gas
Velocity
ft./sec.
1.9
2.0
1.9
2.1
1.8
1.8
2.0
2.0
2.0
Tntal
Gas
Flow,
CFH A
70° F
254
229
191
214
176
176
202
206
202
H2
Flow,
CFH @
70°F
66
67
52
22
63
63
53
' 79
0
H2
Cone. ,
Vol.
%
26.8
32.0
29.8
5.4
37.7
36.7
27.3
36.1
0
Outlet
Gas Concentration, Vol. %
H2
16.2
10.0
3.1
0
5.2
27.7*
3.06
2.0
0
H2S
4.1
13.4
15.2
3.0
18.6
2.6
13.0
27.3
1.32
S02
0.20
0
0
0.07
0
0
0.007
0
0.82
H20
6.30
8.50
5.90
5.40
6.50
0.42
5.73
9.0
9.0
CO
0
0
0
0
0
0.01
0
0
0
C02
1.0
1.25
1.80
1.40
1.70
0
1.40
1.80
2.00
N2
68.3
68.5
70.7
88.0
64.7
69.4
71.7
60.0
88.0
Sulfur
Cone.
# Si/hr.
Temoerature. °F
Stage Number
# 1
800
997
1203
1212
1210
1210
1210
1215
1200
# 2
1182
T150
9 3
800
997
1200
1204
1207
1200
1195
1198
1150
# 4
992
1020
# 5
806
997
1205
1208
1222
1210
1202
1204
1135
1 6
1234
1290
* 7
801
990
1192
1201
1192
1208
1197
1200
1215
# 8
1248
1130
Avg.
802
995
1200
1206
1208
1207
1201
1184
1161
to
*Used Stage 7 analysis (stage below gas outlet).
-------
Table 56. EXPERIMENTAL RESULTS FROM EVALUATION OF COMBINED
SULFUR STRIPPING/H2S GENERATION
Expt.
Number
^fft^fm^mf^m^^f^m^mm
SHG-1
SHG-2
SHG-3
SHG-5
SHG-7
SHG-8
SHG-9
SHG-10
SHG-1 1
Space
Velocity,
Vol. Gas/
Vol. C/
hr.
1,650
1,490
1,240
1,390
1,150
1,150
1,320
1 ,530***
1,530
Hydrogen
Balance
In/Out
1.31
1.34
1.72
1.72
1.73
1.32
1.80
1.15
Sulfur
Balance
In/Out
1.03
# mole S
Stripped/
hr.
^g^gggMgggMMM,,^^
0.079
0.179
0.176
0.152
0.173
0.017
0.174
t
0.179
0.129
# mole H2
Avail.(H2+S)/
hr.
mti^f^mitiiimiiitttimtitartimimimmttmmmmitmmiimiimMmi^
0.178
0.198
0.153
0.028
0.176
0.170
0.145
0.186
—
# mole S
Stripped/
# mole He
Available
•••{•••^•^••^••••••••••••••••••••••••••i
0.44
0.90
1.15
5.43
0.98
0.10
1.20
0.96
—
Sulfur
Stripped
% of
Inlet
mftntumfmiiimiiiii^mimi^^^miffffimm
30
80
89
80
89
44
90
82
62
*
H2
Utiliz.
% of
Inlet
39
68
90
100
87
18
89
94
—
**
H£ Coriv.
to H2S,
% of
Inlet H2
^••^•.•••.••.^•••••••HWIIIIIIIIIIM
16
43
48
58
45
7
45
81
—
**
S Conv.
to H2S,
% of S
Stri pped
••••••••••••••••••••aiiiiiHHA^Bv-"*
35
47
42
10
55
64
38
86
—
S02
Activity
Rel. to
Vi rgi n
Carbon
0.98
—
—
1 . 1 5****
—
—
1.06
—
to
*Goal £
**Goal *
***Effective space velocity for H2 + S reaction in vapor phase «3,000 hr.'
****S02 activity of carbon sample from Stages 4 and 5 was about 0.94.
-------
Figure 69. Effect of temperature on sulfur
stripping with H2 present
-------
Figure 70. Effect of H£ concentration and gas/solid
contact time on sulfur stripping at
1200°F
-------
Effect of Carbon Residence Time on S02 Activity -
From previous work, it is known that the S02 activity is a
function of the final sulfur loading, regeneration
temperature, and time of exposure to a reducing atmosphere.
The S02 activity increases with increasing temperature and
exposure time to a reducing atmosphere. At a fixed tempera-
ture the activity increases with decreasing sulfur loading
on the regenerated carbon. In the present experiments at
1200°F, in particular Runs SHG-7 and -10, the carbon resi-
dence time of 13 and 6 minutes, respectively, is sufficient
to provide a product carbon more active than the precursor.
The S02 activity for Run SHG-7 is 1.15 relative to the
virgin precursor and for Run SHG-10 is 1.06 relative to
the virgin carbon.
Sulfur Removal Profile -
The sulfur removal stage profile for SHG-7 (a typical run)
is given in Figure 71. As seen in this figure,
made at 1200°F, almost all of the sulfur is stripped
off in two stages. This means that even though the space
velocity for sulfur removal from the carbon is about 1,150
hr.~l, the effective gas/solid contact time for the hydrogen
and sulfur vapor to form H2S is about 2 stages (space
velocity about 4600 hr.~l). This suggested that the gas/
solid contact time for the H2 sulfur reaction over carbon
should be increased,as was done in Runs SHG-10 and -11.
Effect of Temperature on Sulfur Conversion to H2S -
The effect of temperature (SHG-1, -2 and -3) is given in
Figure 72. The sulfur conversion to H2S increases with
increasing temperature, as does the percent sulfur removed.
Assuming the unaccounted H2 is from inleakage of air, then
the curve given in the figure results. The results indi-
cate that higher temperatures near 1200°F are necessary in
the present 8 stage reactor at the space velocity of about
5,000 hr.-l for the H2/sulfur reaction. In Run SHG-10 the
space velocity for the H2/sulfur reaction was decreased to
about 3,000 hr."1 by feeding most of the carbon at the
middle of the column. This increased the conversion of the
sulfur to H2S and indicates that the temperature may be
decreased below 1200°F and still allow the goals to be
met.
Carbon Burn-off Rate -
The CO? produced during the runs was necessarily indicative
of carbon burn-off. Previous experiments indicated that
first cycle burn-off could be as high as 92 Ibs. C/Ton
sorbed S02- Burn-off then appeared to decrease to as low
216
-------
Figure 71. Sulfur removal from activated carbon in an
8 stage, 4" diameter regenerator
0.28
u>
c
TJ
-------
Figure 72. Effect of temperature on the conversion
of sulfur to hydrogen sulfide
100
80-
TJ
OO
OJ
1C
HI
O
S_
H-
l/>
60-
40-
20-i
[Effective Space Velocity for H2 + S
Reaction, 6,600 to 13,200 hr.-']
O Experimental
i
X Corrected Assuming Leak
in Outlet Gas Sample
Line
700
800
900 1000 1100
Average Temperature, °F
1200
1300
218
-------
as 2 Ibs. C/Ton S02 after 8 cycles. The data from Run
SHG-10 (1.8 vol. % C02) was equivalent to a burn-off of 67
Ibs. C/Ton sorbed S02 for a carbon loss of 0.42 wt. %.
Recent experimental and literature studies indicate that
oxygen may be chemisorbed on the carbon. This is then
evolved as CO and C02 at elevated temperatures in the
presence of oxygen-free gas. Since chemisorbed oxygen
would be depleted as C02 is evolved, burn-off should
decrease with cycling, as shown by the previous data.
Runs SHG-1 to -9 -
After Runs SHG-1 to -9 had been made a number of equipment
problems were found and subsequently corrected. The first
problem was a leak in the gas sample lines. A second
problem was that the chromatographic gas analysis of
hydrogen had been specified by the manufacturer to be
linear, but the calibration was subsequently found to be
non-linear. The hydrogen concentrations were adjusted
accordingly with the resultant values given in Table 55.
For a typical run, SHG-7, to see the effect of the leak in
the outlet gas line, if the assumption is made that all
of the unaccounted H2 was due to the inleakage of air into
the outlet gas sample line, then the outlet H2 and H2S
concentrations would be increased accordingly to 7.3 and
26.2 volume %, respectively, from 5.2 and 18.6 (see
Table 55) . This would then mean 81% conversion of the
stripped sulfur to H2S and 79% H2 utilization; however,
as discussed later, there is another factor affecting this
assumption which increases the H2 utilization to form HgS.
Also the source of the H20 in the vapor phase will be dis-
cussed later.
Regardless of the problems encountered in these first
runs, the data suggested that additional runs should be
made to increase the gas/solid contact time (decrease
space velocity) for the reaction of H2 with sulfur and
that 4 stages would be sufficient for the necessary sulfur
removal.
Runs SHG-10 and -11 -
Therefore, the equipment was modified to allow the carbon
to be fed both at the inlet and at Stage 4 (number from
bottom to top). These runs (SHG-10 and -11) were made at
1200°F with 85% of the carbon fed to Stage 4 and 15% fed
to Stage 7. This increased the effective number of stages
for H2 sulfur contact by three,.but decreased the number
for reactivation for S02 pickup by about four.
219
-------
Run SHG-11, with only N£ in the inlet gas, showed that the
water in the gas phase came from water sorbed on the carbon.
Subsequent moisture determinations in bench scale equipment
verified the findings in this run. The water on the carbon
probably came from sorption of H20 from the air during
handling.
In Run SHG-10, 86% of the stripped sulfur was converted to
H2S, 95% of the H£ was utilized (81% converted to H£S),
and the outlet carbon had an S02 activity 6% higher than
the virgin precursor. The discrepancy of H£ utilization
(94 vs. 81%) is tied up in the hydrogen material balance,
namely 0.37 Ib. H2/hr. into the reactor and 0.32 Ib. H2/
hr. out of the reactor. Based on literature data on 02
and H2 chemisorption^, 7, 8 and work Westvaco has com-
pleted on 02 chemisorption, it is felt that the
unaccounted hydrogen is chemisorbed. Calculations show
that only about 0.1 mole of H2/100 gms. C would have to
chemisorb on the carbon to account for the H2. The
literature data taken on similar activated carbonaceous
material showed the chemisorption of 02 and H2 are similar
on a mole/gram basis. This fact,combined with Westvaco
data on 02 chemisorption at 530°F of about 0.03 mole/100
gms. C with actual plant produced carbon containing about
0.1 mole 02/100 gms. C, makes a strong case for the H2
chemisorption as a possible mechanism for accounting for
the H2- The literature data indicates that once the H2
is chemisorbed, then it is irreversible unless the tempera-
ture is increased to about 1800°F. This means the phenom-
ena might be expected only in the first few cycles. If
the H2 chemisorption mechanism is accepted then the effect-
ive H2 conversion for the formation of H2S is increased
to 94%. Loss of hydrogen by leakage in these runs is
discounted, based on the fact that the sample lines were
tested with known gases at the reaction temperature.
5.2.6 Elemental Sulfur Recovery
A sulfur condenser was required for operation of the
integral pilot plant on a closed loop cycle using H2S pro-
duced internally with the process. A condenser was
designed, installed, and tested prior to use in the integral
run. This pilot development of the condenser resulted in
smooth operation of sulfur condensation and recovery during
the integral operation. The pilot development and results
for the sulfur condenser prior to the integral run are given
below.
220
-------
Sulfur Condenser Operation -
Initial Sulfur Condenser Testing -
Initial sulfur condenser tests were made without a recir-
culating stream of liquid sulfur for scrubbing purposes.
These runs, discussed below, provided a basis for improved
condenser design and operation. The run conditions for
three initial runs that were made are summarized in
Table 57.
Table 57. OPERATING CONDITIONS FOR SULFUR
CONDENSER TESTING SYSTEM
Liquid Sulfur Reservoir Temperature
Liquid Sulfur Reservoir Pressure
Liquid Sulfur Flow Rate to Vaporizer
N£ Carrier Flow Rate to Condenser
Sulfur Vaporizer Temperature
/
Condenser Inlet Gas Temperature
Condenser Outlet Gas Temperature
Condenser Steam Jacket Temperature
Condenser Exit Gas Sample Flow
Rate Through Trap
- 290-300°F
- 2-5 PSIG
- 6 Ibs./hr.
- 225 SCFH
- 1000-1200°F
- 1000°F
- 250°F
- 250°F
- 2-3 SCFH
The results of running at these conditions are summarized
in Table 58. The first run showed a 61% recovery of
sulfur, i.e., of the sulfur which left the reservoir, 61%
was recovered from the condenser as a liquid. The rest
was carried out with the nitrogen off-gas. For the second
and third runs a mist eliminator, consisting of a roll of
221
-------
Table 58. SULFUR CONDENSER TEST RUNS
Run
1
2
3
Duration
minutes
260
180
104
N2 Flow
Rate
SCFH
230
230
230
Avg. Sulfur
Flow Rate
Ibs./hr.
2.2
3.0
3.0
Flow Rate
Range
Ibs./hr.
1 - 5
1 - 7
2.5 - 3.5
Avg. Sulfur
Vapor Cone.
Ibs. S/lb. N2
0.13
0.18
0.18
Recovery
%
61
88
88
tightly wound wire mesh, was placed inside the condenser,
increasing recovery to 88% in both subsequent runs.
However, there is some question as to the validity of
these percentages, because the possibility exists that some
liquid sulfur may have bypassed the vaporizer; i.e., it
may have passed through unvaporized and reached the con-
denser as a liquid. If this condition did exist, then the
actual condenser efficiency would be lower than it appears.
In order to achieve 99% recovery of the sulfur, it was
necessary to provide some degree of scrubbing action.
One approach would be to circulate a stream of liquid
sulfur through the condenser, as was originally intended
when the condenser was designed. Another method would be
to further process the off-gas by bubbling it through a
column of liquid sulfur.
Sulfur Condenser Operating Results -
All of the experiments performed prior to integral testing
on the recovery of sulfur from regeneration off-gases by
condensation are summarized in Table 59. As can be seen
from the data, the original baffled exchanger without sulfur
recirculation or a mist eliminator removed about 61% of the
desired sulfur. Addition of a mist eliminator increased
recovery to 88%; however, it was found that the scrubbing
action of recirculating liquid sulfur was necessary to
raise sulfur recovery to the desired 99+%. Addition of 29%
H2S to the gas contacting the recirculating sulfur caused
no apparent changes in sulfur viscosity, as indicated by the
constant pump electrical load over a six hour period. The
jacket cooling fluid was changed from steam to hot water
to facilitate temperature control and improve heat transfer
rates. The recirculating sulfur condenser system was
222
-------
Table 59. SULFUR CONDENSER OPERATION
Ron
Ha.
1
2
3
4
S
Comments
Much S Vapor/Hist
in Off Gas
Installed Mist
Eliminator
Tried Unsuccess-
fully to analyze
off gas S cone.
Achieved better
, control of S flow
to vaporizer.
Exoerienced prob-
lem controlling
jacket temp.
Liquid S recircu-
lation installed.
Also new steam-
life cooling
system.
Pump enclosed in
heated box. Pump
motor current
constant ? 7.5
amps. Observed no
change in liquid
S viscosity.
Gas
Comoosition
Sulfur*
18
13
15
13
0
H2S
0
0
0
0
29
j
»2
82
87
85
87
71
Total**
Gas
Flow,
CFH 9 70°F
164
236
226
225
231
Sulfur
Recirculation
Rate,
GPM
0
0
0
•
•
0.5
0.5
Jacket
Coolant
Steam
Steam
Steam
Water
Water
Temperatures, °F
Jacket
Inlet
280*10
280110
290110
27015
26015
Jacket
Outlet
290*10
295110
305110
27015
265i5
Cond.
Inlet
1000*50
1000+50
1000150
890130
875
Cond.
Outlet
—
--
31615
280HO
.-
290tlO
Run
Time,
hours
4.3
3.0
1.7
4.0
6.0
Sulfur
Recovered,
X
61
88
88
99.6
.
Off-Stl
Sulfur
Cone..
mole I
—
—
_
-
0.06
<*»•
ro
N>
U>
*As Si-
"Based on Sulfur as
-------
tested for 14 hours of intermittent operation without
serious problems and was used as the basis for design of
a system for installation in the pilot plant. Because
of the unusual viscosity characteristics of sulfur,
temperature control will be a major design consideration
in order to keep all parts carrying liquid sulfur in the
range 260-315°F.
The feasibility was shown experimentally of a sulfur con-
denser system using recirculation of liquid sulfur as a
scrubbing media. The anticipated improvement in the con-
densing efficiency to 99.6 weight % was realized.
The run conditions for the best sulfur performance are
schematically shown in Figure 73. The sulfur concentra-
tion in the outlet gas from the condenser was determined
by weight pickup by passing a slipstream through a tube
packed with glass wool at room temperature. The slip-
stream volume was measured with a wet test meter. The
sulfur on the glass wool was determined by extraction
with carbon disulfide. The run, which lasted four hours,
operated with an outlet gas concentration near equilibrium.
From equilibrium the sulfur rate in the off-gas from the
condenser is predicted to be 0.0088 and 0.0132 Ib./hr. at
260 and 270°F, respectively. In the run made, the actual
temperature was 275 * 5°F with a sulfur rate of 0.011
Ib. S/hr. measured. Even with a predicted accuracy of
5070 for the off-gas sulfur concentration analysis, the
condenser is operating near equilibrium conditions. The
success of the run is strengthened by the material balance
of 11.3 Ibs. sulfur vaporized into the system over the
four hour period and 11.7 Ibs. sulfur recovered from the
system in the outlet streams.
The safe lower operating outlet gas temperature is about
260°F because the liquid sulfur solidifies at 238°F and a
safety margin of about 20°F is recommended. Although
there were problems with the sulfur circulating pump during
start-up, it functioned well during the run.
It has been demonstrated that the sulfur condenser can
achieve the desired condensing efficiency of greater than
99%. The improved performance is the direct result of the
liquid sulfur recirculation system, which provides the
scrubbing action needed to condense the sulfur mist.
224
-------
Figure 73. Operating conditions for sulfur condenser
Cooling HgO In
270° F
Scrubbed Gas
0.011 Ib. S/hr.
275 * 5°F
Liquid ST Circulation Rate =0.5 GPM
d
Sulfur Laden Gas
221 CFH @ 70°F
2.82 Ibs. S/hr.
890 t 30°F
Cooling H20 Out
265°F
Recovered
-*• Liquid
Sulfur
2.9 Ibs. S/hr.
Circulation
Pump
SULFUR MATERIAL BALANCE
Sulfur In
Sulfur Out
- TOTAL IN
- Liquid
- Gas
- TOTAL OUT
11.3 Ibs.
11.7 Ibs.
.04 Ibs,
11.7 Ibs.
225
-------
5.2.7 Fluidizing Mechanics
There are numerous types of reactors for contacting acti-
vated carbon with gases. From the standpoint of smaller
equipment, of the resulting decreased investment, of high
heat transfer rates, and good solids flow characteristics,
multistage fluidized bed reactors were chosen as the type
of gas/solid contactor. There are many important parameters
to be considered in design of the fluid bed reactor.
The gas/solid residence time requirements have been dis-
cussed in previous sections as related to the process
chemistry.
Westvaco has developed considerable technology in fluid bed
design and operation. In addition to the basic process
development of fluid bed design data, which had been
developed prior to the present contract, some aspects
relating to the fluidizing mechanics have been studied
under the present contract. Those points are discussed
below.
Carbon Fluidization Requirements -
As presently conceived,S02 sorption and sulfur generation
steps will be run in fluidized beds of activated carbon.
These fluidized beds operate at linear gas velocities
between the minimum fluidizing (UMf) and entrainment
velocities (Ut). Kunii and Levelspiel^ present equations
for calculating these bed characteristics, i.e.
Minimum fluidizing velocity, UMf
UMf = -- [(33.7)2 + 4.08(10-2) W^'" P8)g]l/2 . 33.7 (65)
Terminal velocity,
(Ps-Pg)dp]1/2 500 < Rep < 200, 000 (66)
Pg
where Rep
226
-------
The calculation of minimum fluidizing velocity and entrain-
ment velocity for Westvaco granular carbon, which has an
average particle size of 0.1 cm and a particle density of
1.0 g/enP, is shown in Table 60.
Table 60. CALCULATED VALUES FOR MINIMUM FLUIDIZING
VELOCITY AND ENTRAINMENT VELOCITY FOR
WESTVACO GRANULAR ACTIVATED CARBON
Gas
Air
Air
Air
Hydrogen Sulfide
Flue Gas**
Temp . ,
°F
70
200
300
300
200
Minimum
Fluidizing
Velocity,
ft. /sec .
0.84
0.87
0.75
0.86
0.82
Entrainment
Velocity*,
ft. /sec.
10.7
—
_ _
--
w •_
*For particle size of 0.042 cm.
**For gas composition: 76.6% N2, 3.4% 02,
6.0% H20, 14% C02
As can be seen from the table, there is very little differ-
ence between the calculated values for air over the range
70-300°F. In addition there is little difference between
the minimum fluidizing velocities calculated for air, H2S
and flue gas. Thus fluidization characteristics with air
at 70°F can be used to closely approximate fluidization
parameters for S02 sorption and sulfur generation. Actual
experimental measurement of the minimum, fluidizing velocity
with Westvaco granular carbon is shown in Figure 74. The
measured value, taken as the maximum in the curve as
suggested by Levenspiel9 is approximately 0.5 ft./sec.
or about 40% below the calculated value. This is slightly
outside the range of ±34% in which Levenspiely says the
values normally fall. The difference is probably due to
the deviation of the irregular carbon particles from a true
spherical shape.
227
-------
Figure 74.
Experimental determination of minimum
fluidizing velocity for Westvaco
granular carbon
o
CM
CL
o
o
3
in
t/1
t
a.
•o
01
CO
!.*»
1.2
1.0
0.8
0.6
0.4
0.2
0.1
*
?
_.
/
~
/
/
*>
y-
/
/
/
/
•
/
Gas
Tern
Avg
Car
Minimum Fluidizing Velocity
A
\
\
\
~»j
•~~.
€
— -
-
oW£
: Air
perature: 70°F
. Carbon Size-: 1.0 mm
ion Particle Density = 1.
.... ..Ill 1
~}J$i5
.--
0 g/cc
0.1
Air Velocity, ft./sec.
1.0
228
-------
Also shown in Table 60 is the calculated entrainment
velocity of 10.7 ft./sec. for a 0.42 mm particle which is
the smallest particle of significant percentage in the
carbon sample. Thus, for the Westvaco granular carbon used
in S02 removal, the range of operating velocities is the
0.5 ft./sec.'(experimental value) minimum fluidizing
velocity and the approximately 10.7 ft./sec. entrainment
velocity. In the S02 sorber, it is desirable to run as high
a velocity as practical in order to keep the adsorber
cross-section as small as possible for treating the large
volumes of flue gas. Since some smaller particles will be
generated by the bed action, a value of about 8 times the
minimum fluidizing velocity, or about 4 ft./sec., is used
for the sorber. Since the sulfur generator will be much
smaller than the sorber, a value of 4 - 6 times the minimum
fluidizing velocity (^2.5-3.0 ft./sec.) should be suffi-
cient to maintain good contact and minimize elutriation.
S02 Sorber Gas Distributor Plate Design -
One of the equipment modifications specified before integral
operation was the installation of new gas distributor
plates in the 18"0 S02 adsorber. The goals for the plates
were to minimize carbon attrition while still maintaining
proper fluidization for a high S02 recovery efficiency.
A number of drilled distributor plates were evaluated in a
batch, one stage, 18"0 fluid bed unit. The purpose was to
provide the operating characteristics of the plates to
achieve the goals given above. The plates originally in the
18"0 unit were 3.2% open area plates with l/8"0 orifices
on a 2/3" equilateral triangular pitch. The proposed plates
to reduce carbon attrition were 8% open area. In line with
this, a 2.1% open area plate with l/8"0 orifices was used
to approximate the operating conditions of the 3.270 open area
plates. In addition, two plates with 87<> open area and 1/4"
and 3/16" orifices were fabricated to test the effects of
the higher open area plates on the S02 removal efficiency.
One operating parameter of prime importance is the pres-
sure drop characteristics of the drilled distributor
plates. The pressure drop characteristics of the 2.170and 870
open area plates were measured as a function of the super-
ficial linear gas velocity. The pressure drop increases
with increasing velocity as shown in Figure 75. As the
open area increases the pressure drop decreases and as the
orifice diameter decreases for a particular open area the
pressure drop also decreases. The experimental data compares
favorably with data taken from the literature. In terms of
what can be expected in the 18"0 S02 sorber, the pressure
drop across the 3.2% open area plate is about 3" H20 at
229
-------
Figure 75. Pressure drop characteristics of distributor
plates to be used in an 18" diameter S02 sorber
100
80
o
CXJ
1C
O)
-C
u
c
o.
o
o
1/1
VI
40
20
10
-------
3 ft./sec. and with an 8% open area plate at 3 ft./sec. a
pressure drop near 0.6" H20 would be realized or a total
pressure drop reduction across the plates of a factor or 5
or more.
The S02 sorption characteristics of the plates were deter-
mined by loading the bed with 35 Ibs. virgin carbon, then
at time zero the simulated flue gas (2,000 ppm S02, 200
ppm NO. 2 to 2.4% H20, balance air) was introduced into
the fluid bed at 3 ft./sec. The total sulfur analysis of
the carbon was then followed as a function of time. The
run was stopped when the acid loading was above 0.22 Ib.
acid/lb. C. The results of the runs are given in
Table 61. The S02 removal efficiency was 0.24, 0.22,
and 0.15 for the 1/8", 3/16", and 1/4" hole diameter
plates, respectively. An unanticipated problem occurred,
however, with what is called carbon weepage during
fluidization. This phenomena is carbon flow through the
distributor plate, which in a continuous unit would mean
possible short-circuiting of the carbon which could
result in an inefficiency of the S02 sorber. The phenomena
was only observed, at least to an appreciable extent, for
the high open area plates (870) , with larger orifices.
Nevertheless, on a qualitative basis it is felt that the
S02 sorption efficiency is comparable in light of this
carbon removal phenomena occurring continuously as a
function of time.
Since the carbon weepage appeared important in the gas
distributor design for the 18"0 unit, the above plates and
a number of other ones were evaluated for carbon weepage.
The carbon weepage rate was investigated as a function
of the open area, orifice diameter, superficial linear gas
velocity, and carbon bed loading. The six plates which
were evaluated are given in Table 62. The plates varied
from 2.17. to 8.37o open area with 3/32"0 to l/4"0 orifices.
The final carbon weepage rate data was subjected to a
multiple regression analysis. The data was found to
correlate well as given by Equation (67):
A - 8.4(10~n) e23.2 do el-95 Ao(-i-)1.3 eQ.12 L (67)
r\
where A = carbon weepage rate, Ib. C/hr.-ft.
do = orifice diameter, in.
Ao = gas distributor open area, 70
v = superficial linear gas velocity,
ft./sec.
L = carbon bed loading, Ibs. C
231
-------
Table 61. OPERATING CHARACTERISTICS DISTRIBUTOR PLATES TO BE USED
IN AN 18" DIAMETER S02 SORBER
Run
No.
SA-42
SA-43
SA-44
PLATE CHARACTERISTICS II COLUMN CONDITIONS
Overal 1
Dia.,
in.
18
18
18
Thick.
in.
1/8
1/8
1/8
Triang.
Pitch
in.
0.84
0.82
0.625
Hole
Dia.
in.
1/8
1/4
3/16
No.
of
Holes
440
415
750
Open
Area
%
2.1
^c^nkiri
8.0
8.2
Hole
Velocity
ft. /sec.
142
37
36
Gas
Flow,
CFH
9
70°F
278
278
271
Super-
ficial
Gas
Veloc.
FPS
3.0
3.0
2.9
Gas Concentration
S02
PPM
2,000
2,000
2,000
NO
PPM
200
200
200
H20
Vol. %
2.0
2.2
2.4
Air
Bal.
Bal.
Bal.
Carbon
Type
Virgin
Virgin
Virgin
Carbon
Loaded
Ibs.
35
35
35
Run
Time
rain.
240
240
180
Carbon
End
on
Plate,
Ibs.
30.5
10.5
10.3
C in*
Cyclone
Ibs.
2.2
0.6
0.7
C in**
Plenum
Ibs.
***
23.9
24.0
Final
Acid Load
#Acid/#C
0.235
0.253.
0.244
S02
Removal
Efficiency
0.24
0.15
0.22
1TT-W*
Csrsor
Weecsce ;.ate
*/hr.
6.0
8.0
*/hr.-ft.2
...
3.4
4.5
N>
W
N)
*Cyclone dust given on total weight basis; was not analyzed for moisture and acid content.
**By difference since no analysis was made of carbon in plenum.
***Was not measured but discrepancy in material balance is believed due to start-up problems in which carbon may have bypassed cyclone.
****Taken as constant over time period of run. , •
*****Used to approximate 3.21 open area plate.
-------
Table 62. GAS DISTRIBUTOR PLATE CHARACTERISTICS
EVALUATED FOR CARBON WEEPAGE DURING
FLUIDIZATION
Plate
Diameter,
inches
18
18
18
18
18
18
Orifice
Diameter ,
inches
1/8
1/8
5/32
3/32
3/16
1/4
Triangular
Pitch,
inches
0.840
0.667
0.667
0.310
0.820
0.625
Plate
Open Area,
%
2.1
3.2
5.4
8.3
8.2
8.0
233
-------
The correlation coefficient (R) for the equation was 0.95,
which indicates a good fit of the model to the experi-
mental data. Although larger orifice diameters up to
l/4"0 were considered, the operating difficulties during
start-up and shutdown make the 3/16"0 and l/4"0 orifices
unsuitable for the present pilot plant equipment.
Therefore for the present plates, the orifice diameter was
taken to be 1/8".
The carbon weepage rate for the present plates in the 18"0
unit (L = 30 Ibs., Ao = 3.2%, do = 1.8", and v = 3 ft./sec.)
was predicted to be 6.7(10-6) Ibs. C/hr.-ft.2 or for the
18"0 unit, 1.1(10-5) Ibs. C/hr., which is effectively ;zero.
For the proposed 8% open area plates (1 = 30 Ibs., AQ =
8%, d0 = 1/8", and v = 3 ft./sec.) the predicted carbon
weepage rate is 0.14 Ib. C/hr. This weepage rate is accept-
able in the upper stages of the 18"0 unit (about 0.5%
of the total carbon flow rate through the unit), but is
felt to be less desirable for the bottom stage in extended
integral runs. This would correspond to 65 Ibs. C which
would be collected in the inlet gas plenum for the 18"0
unit over a 20 day operation.
To minimize this carbon handling a 6.15% open area plate
with l/8"0 orifices was considered and the predicted
carbon weepage rate was 0.004 Ib. C/hr. or about 2 Ibs. C
over a 20 day period. This is an acceptable quantity of
carbon but is at the expense of a slightly higher carbon
attrition rate,since the carbon attrition is a direct
function of the per cent open area of the plate.
All of the experimental data led to the design of four
distributor plates in the upper stages of 87o open area
with l/8"0 orifices on a 0.42" equilateral triangular
pitch. Two of the top uppermost stages are of aluminum
and the other two are of 316 stainless steel. The dis-
tributor plate for the bottom stage was specified to be
6.15% open area with l/8"0 orifices on a 0.48" equilateral
triangular pitch. The specifications for these distributor
plates are given in Table 63.
234
-------
Table 63. GAS DISTRIBUTOR PLATES SPECIFICATIONS
DESIGNED FOR MINIMIZING CARBON
ATTRITION IN THE 18"0 S02 SORBER
Plate
Diameter ,
inches
18
18
Orifice
Diameter ,
inches
1/8
1/8
Triangular
Pitch,
inches
0.42
0.48
Plate
Open Area,
%
8.0
6.15
235
-------
SECTION 6
1,000 MW UTILITY BOILER FLUE GAS CLEAN-UP
6.1 INTRODUCTION
The Westvaco Process has been developed over the past
eight years, the last three under joint support from EPA.
The process is designed to adsorb S02 out of flue gas
streams with activated carbon and to produce elemental
sulfur as a by-product. Pilot tests of the complete
process treating 20,000 cfh of flue gas from an oil fired
boiler have recently been completed under this EPA
contract, and results are contained in this report.
With the Westvaco Process, sulfur dioxide is removed from
waste gases with activated carbon acting as a catalyst
and adsorbent in the reaction:
S02 + 1/2 02 + H20 e»• H2S04 (Sorbed) [150-300°F] (5)
The sorbed acid is then converted to elemental sulfur
by reaction with H2S, with the carbon again acting as a
catalyst and adsorbent:
H2S04 -I- 3 H2S * 4 S + 4 H20 [200-300°F] (6)
The sorbed elemental sulfur is then thermally stripped
from the activated carbon at 800-1000°F and condensed as
product. In utility applications where H2S is unavail-
able for conversion of sorbed acid to sulfur, hydrogen
is added to the gases during the thermal stripping step
to produce the necessary amount of H2S by the reaction:
H2 + S (Sorbed) ^ ^ [800-1200°?] (68)
A schematic of the process is shown in Dwg. 2563
(Figure 76).
236
-------
Figure 76. Westvaco S02 Recovery Process schematic flowsheet
N5
SOj
SOj LADEN FLUE 6AS •
CLEAK TtlJE
GAS TO STACK
so,
REHOVAL
HYDROGEN
REGENERATED
CARBON
5Hj3 * HgSC4 3°°*F-1 S I »
SULFUR
CONDENSER
SPENT GAS
TO BOILER
SULFUR
PRODUCT
• Westvaco
CHARLESTON RESEARCH CENTER
f. 0. MX SM7 WMTH CM1UILESTOM. S. C.
WESTVACO SOz .RECOVERY .PROCESS
SCHEMATIC. .FLOWSHEET
I JM*> [WO NO
'^oiro vi
lDv»G. No.
-------
6.2 GENERAL DESIGN BASIS
6.2.1 Scope
The full scale 1,000 MW installation will consist of an
integral closed loop system continuously recycling granular
carbon for removal of S02 from approximately 1,750,000 cfm
of actual power plant flue gas for a coal fired boiler and
recovering elemental sulfur as a product. The flue gas is
treated after an electrostatic precipitator then returned
to the boiler stack.
The 1,000 MW installation consists of the following major
processing steps:
1) S02 removal from flue gas
2) Sulfur production from recovered S02
3) Thermal stripping of sulfur product and
production of hydrogen sulfide
4) Condensation and recovery of sulfur product
5) Production of chemical reducing gas (hydrogen).
Pilot plant experiments have established the technical
feasibility of the process and supplied the design basis.
Operation of a prototype unit 0-^15 MW) should finalize
scale-up data needed prior to the full scale installation.
238
-------
6.2.2 Boiler Operating Characteristics
Coal Composition -
Component, Ash:
Sulfur:
Hydrogen:
Carbon:
Nitrogen:
Oxygen:
H20 in Coal:
Excess Air:
Carbon Burned in Coal:
Heating Value in Coal:
Sulfur to S02:
Sulfur to 803:
S02 Removal:
Flue Gas Temp. Out
of Air Preheater:
Minimum Gas Temp. to
Stack after Treatment:
Gas Velocities in Ducts:
Plant Size:
Heat Rate:
Flue Gas Quantity:
Flue Gas Molecular Wt.:
Flue Gas Weight:
Flue Gas Composition -
Component, S02:
803:
Nitrogen:
Oxygen:
Water:
Fly Ash:
Water in Combustion
Air (60% RH @ 80°F)
15.2% Dry Wt. %
3.5 Dry Wt. %
5.0 Dry Wt. %
67.2 Dry Wt. %
1.6 Dry Wt. %
7.5 Dry Wt. %
4.8 Ibs./lOO Ibs
207o
Wet
100%
Wet - 11,980 BTU/lb.
Dry - 12,580 BTU/lb.
98%
2%
90%
300°F
200°F
6.0 ft./sec.
1,000 MW
9,000 BTU/KWHR.
293,000 moles/hr.
29.54
8,655,000 Ibs./hr.
0.00261
0.00005
0.13657
0.74087
0.03276
0.08714
77,900
mole fraction
mole fraction
mole fraction
mole fraction
mole fraction
mole fraction
Ibs./hr.
0.0212 mols/mol dry air
751,252 Ibs./hr.
Coal Consumption
Rate (Wet):
Gas Ducts 300°F or Higher To Be Insulated
239
-------
6.2.3 Product
Type: Elemental Sulfur
Purity: 99.8% or Better - Bright Yellow
Form: Liquid at 250°F
6.2.4 Process Conditions
SQ2 Sorber -
Reaction: S02 + % 02 + H20 -»• H2S04
Heat of Reaction: Exothermic = -117,000 BTU/mole S02
Contact Mode: Stagewise Gas/Solid Fluidized Bed
Fluidizing Gas Velocity: 4 FPS (At Base Load of . i
1,000 MW)
Fluidized Bed Temperature: 150-300°F
S02 Overall Recovery Efficiency: 907o
S02 Rate (At Base Load), from Coal = 48,943 # S02/hr,
from Recycle = 7,782 # S02/hr,
TOTAL = 56,725 # S02/hr,
Acid Loading on Carbon: 0.22 Ib. H2S04/lb. C
Number of Stages: 5
Space Velocity: 2,350 hr.'1
Carbon Bed Depth/Stage: 11 inches
Carbon Rate: 43 Tons Carbon/Hr. (At Base Load)
S02 Concentration, Inlet (Coal) = 2,610 PPM
(Recycle) = 415
TOTAL = 3,025 PPM
. Outlet: 245 PPM
Boiler Feed H20 Spray Rate: 183 GPM
Sulfur Production -
Reaction: H2S04 + 3 H2S + 4 S + 4 H20
Heat of Reaction: Exothermic = -41,107 BTU/mole H2S04
Contact Mode: Stagewise Gas/Solid Fluidized Bed
Fluidizing Gas Velocity: 3 FPS (At Base Load)
Fluidized Bed Temperature: 200-325°F
Inlet H2S Concentration: 2 Vol. %
Conversions, H2S Utilization: 99.9% of Inlet H2S
S02 Recycle: 15% of Sorbed H2S04
Sulfur Formation: 70% of Sorbed H2S04
240
-------
Number of Stages: 11
Space Velocity: 460 hr.~^
Carbon Bed Depth/Stage: 17 Inches
Carbon Rate: 43 Tons/Hr. (At Base Load)
Outlet H2S Concentration: 270 PPM
Sulfur Stripping/H2S Generation -
Reaction: a) H2 + S -> H2S
b) H2S04 + 3 H2S -»• 4 S + 4 H20
Heat of Reaction: a) Exothermic = -8,667 BTU/mole H2S04
b) Exothermic = -41,107 BTU/mole H2S04
Contact Mode: Stagewise Gas/Solid Fluidized Bed
Fluidizing Gas Velocity: 3 FPS (At Base Load)
Fluidized Bed Temp.: a) Carbon Preheater = 710°F
b) H2S Gen./S Strip. = 1000-1200°F
Inlet H2 Concentration: 19.5%
Inlet H2 Requirement: 3.3 moles H2/mole S02 recovered
Carbon Residence Time: 21 minutes
Design Rates a) S Stripping = 796 moles S/hr.
(At Base Load): b) H2S Formation = 1,910 moles H2S/hr.
Number of Stages: a) Carbon Preheater = 1
b) H2S Formation = 2
c) Sulfur Stripping = 4
Space Velocity: 1600 hr."1
Sulfur Recovery -
Type: Shell and Tube Condenser
Duty: 796 moles S/hr.
Temperature: a) Inlet Gas = 1040°F
b) Outlet Gas - 250°F
c) S Liquid Prod. = 250°F
Efficiency: 99.9% of Inlet Sulfur
Carbon Cooler -
Contact Mode: Gas/Solid Fluidized Bed
Fluidizing Gas Velocity: 3 FPS
Fluidized Bed Temperature: a) Inlet C = 1040°F
b) Outlet C = 300°F
c) Outlet Gas = 300°F
241
-------
Gas Type: a) Recycled Inert Gas
b) 5% Inert Gas Make-up/Cycle
c) 5% Boiler Feed H20 Make-up/Cycle
Number of Stages: 1
Gasifier -
Type: Bituminous Coal Feed
Product Gas: 28 cf (H2 + CO)/lb. Coal
65 cf Total Gas/lb. Coal
6.2.5 Activated Carbon Characteristics
Type: Coal Based
Size: 8x30 M [Nominal; 1.5 MM (Avg.
Particle Size)]
Density: 40 - 43 Ibs./ft.3
SC-2 Number: 75 (Minimum)
Attrition No.: 30 (Maximum)
6.3 CONCEPTUAL DESIGN
The conceptual design flowsheet for the process is shown
in Figure 77 (Dwg. 2572) and described by the process
description below.
6.3.1 Process Description
S02 Sorber - Boiler Flue Gas (FB-101-A, B, C and D) -
i
The boiler flue gas «^p> is desulfijrized by counter-current
contact with regenerated carbon <^> in stagewise fluid bed
reactors (FB-101-A, B, C and D). The total flue gas
volume of 105 million SCFH is split among four 55 diameter
mild steel sorbers containing 5 stages of fluidized carbon
for SOX removal. The bottom stage removes 803 at 300°F to
prevent acid condensation. The gas temperature is then
lowered to 170°F for more efficient S02 removal by direct
evaporation of water sprayed into the second fluid bed
stage. The remaining 4 fluid bed stages, containing 13.5
inches of activated carbon eachlower the S02 concentra-
tion to 245 ppm at the outlet ^£>. Reaction heat liberated
during S02 removal reheats the gas temperature to 200°F.
The fluidized carbon flows down the column by gravity
through overflow weirs and downcomers becoming progres-
sively loaded with sulfuric acid and leaves the column
at 300°F containing 22% of its weight in adsorbed acid.
242
-------
243
-------
feOiltLR. Ftr-> wATP.lt
t4> i fiueL e»L
. , ^H
Owm tn i ^r)
d
243-A
-------
Figure 77. Westvaco S02 Process flowsheet for 1,000 MW
unit (250 MW typical module shown)
M/ec.7
TV Pt-tti 4 W
t/.i'Di* « if. ii srff* flvioiut of a
e*s. DKf nmiit fV.ua. '/t'Ufi.m.t
• U t CIMTtXi,
/«.« MB •MI &ncve 'Jwc 0*0
243-B
-------
A blower (F-101) is included to overcome the gas pressure
drop in each adsorber. A start-up heater (A-101) is
included for pre-heating the sorber.
Sulfur Generator (FB-102-A. B. C and D) -
The acid laden carbon <^> from the S0£ sorber is contacted
with H£S at 300°F in four, 11 stage fluidized bed reactors
(FB-201-A, B, C and D) for conversion of the acid
elemental sulfur. The H2S required for
comes from H2S produced during the stripping step. The
of f-gas <^> containing possibly S02 evolved or traces of
H2S is recycled to the boiler for conversion to 502 easily
removed in the S02 sorber. The four reactors are each
11.3' diameter constructed of Alonized 304 stainless steel
throughout. Activated carbon flows downward from stage to
stage through overflow weirs and downcomers and leaves the
reactor at 300°F <^£> containing approximately 24% of its
weight in elemental and about 3% of its weight in uncon-
verted acid.
H2S Generator/Sulfur Stripper (FB-103-A, B, C and D) -
The sulfur laden carbon <^> from the sulfur generator is
contacted with hydrogen rich gas <^>in the H2S generator/
sulfur stripper (FB-jU)3-A, B, C and D) to thermally strip
the sulfur product ^2^>and to produce a part of the H2S
feed <^>for the sulfur generator. The reactor is 20'6"
diameter, contains 7 fluidized stages and is constructed
of Alonized 304 stainless steel throughout. The top
stage of the reactor is used to preheat the carbon to
710°F and is 18'7" diameter. The remaining 6 stages, at
1000-1050°F, are for sulfur stripping and H2S conversion.
The carbon flows down through the vessel from stage to
stage by gravity through overflow weirs and downcomers.
The regenerated carbon at 1050°F leaves the vessel
containing 4.1% of its weight in residual adsorbed sulfur
which remains constant throughout the whole loop.
j
The hydrogen rich gas <4£^is produced from a coal gasifier
[G-102] in series with a water-gas shift converter (CV-101-
A, B, C and D) for conversion of CO to hydrogen. Tars are
removed prior to the shift converter. After CO conversion
the gas is heated to 1040°F in a fired heater [GH-102].
The sulfur product ^^>is separated from the recycle H2S
stream <3£>in a shell and tube condenser (SC-101) ,
filtered for traces of carbon, and sent to a sulfur
storage pit (SP-101).
244
-------
Carbon Cooler (FB-105-A, B, C and D) -
After regeneration the carbon is cooled to 300°F by direct
evaporation of water sprayed into the fluidized bed cooler
(FB-105-A, B, C and D). Recycle inert gas is used as the
fluidized gas. The 14.5' diameter cooler contains one
stage and is of Alonized 304 stainless steel construction.
The H20 is condensed from the recycle gas and both are
reused.
Carbon Handling and Storage -
Activated carbon is recirculated through the process loop
at a rate of 172 tons per hour. Storage for 6 hours of
carbon recirculation is provided for both the acid laden
and regeneration loop to accommodate variation in waste
gas rates. The storage vessels (S-101 and -102) are
coned bottom tanks 28*0x60' tall of mild steel
construction. The total carbon inventory is 2,000 tons
and make-up carbon is added at a rate of 275 Ibs./hr.
Horizontal and vertical conveying are accomplished by belt
conveyors and bucket elevators, respectively. These are
not detailed on the flow sheet but are described in the
equipment lists in Appendix K.
6.4 HEAT AND MATERIAL BALANCES
The heat and material balances were made for the overall
process just described. The overall sulfur balance is
shown in Table 64. The overall sulfur balance reiterates
the 90% sulfur recovery from the flue gas taken as a basis
for this particular evaluation.
Table 64. OVERALL SULFUR BALANCE FOR
1,000 MW POWER PLANT
SULFUR IN
Stream
Flue Gas
Gasifier Gas
TOTAL IN
Ibs. S/hr.
24,940
896
= 25,836
SULFUR OUT
Stream
Flue Gas
Elemental S Product
IDS. S/hr.
2,494
23,342
TOTAL OUT = 25.836
245
-------
The overall energy balance is given in Table 65. About
60% of the total heat input originates with the flue gas,
indicating the relative energy requirement of the regene-
ration system.
Table 65.
OVERALL ENERGY BALANCE FOR
1,000 MW POWER PLANT
ENERGY INPUT
Stream
Flue Gas
Spray H20
AHrxn - SOX Removal
AHSoln - H20-H2S04
AHrxn - S Gen.
AHrxn - H2S Gen,
Sulfur Cond.
Gasifier Gas
Steam for Shift Run
Reducing Gas Preheat
Reducing Reheat
Reducing H20 Cond.
Reducing Reheat
Carbon Cooler H20 Cond.
MM BTU/Hr.
554
57
96
17
24
31
4
1
12
126
60
98
14
74
%
48.7
2.4
8.4
1.5
2.1
2.7
0.4
0.1
1.1
11.1
5.3
8.6
1.2
6.5
TOTAL IN = 1,138 MM BTU/Hr.
ENERGY OUTPUT
Stream
Flue Gas
Spray H20 Vap.
Reducing Gas
AHsoln H20-H2S04
H20 Vap. - H20-H2S04
Sulfur Vap.
AHrxn " H2S Generator
Sulfur Product
Sulfur Condenser
Carbon Cooler - H20 Vap.
Carbon Cooler - H20 Rec.
H20 Condenser
H20 Condensed from R.G.
Heat Losses - Carbon Storage
MM BTU/Hr.
284
367
•13
17
16
4
6
2
103
83
96
112
3
32
%
25.0
32.2
1.1
1.5
1.4
0.4
0.5
0.2
9.1
7.3
8.4
9.8
0.3
2.8
TOTAL OUT = 1,138 MM BTU/Hr.
The detailed heat content and stream compositions are
given in Tables 66-75.
246
-------
Table 66. STREAM CONDITIONS
N>
-P-
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
STREAM 1
Ibs./hr.
1,753
330
+40
554
moles/hr.
844
9,612
25,520
217,112
40,140
16
293,284
STREAM 2
Ibs./hr.
10
258
268
1,850
200
+10
284
moles/hr.
78
9,208
43,388
217,112
40,140
309,926
STREAM 3
Ibs./hr.
13,664
341,556
355,220
—
. 77
—
0
moles/hr.
STREAM 4
Ibs./hr.
78,944
14,276
13,654
341 ,298
448,222
—
300
—
26
moles/hr.
-------
Table 67. STREAM CONDITIONS
ro
-P-
oo
S02
~H2S ~
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
STREAM U-l
Ibs./hr.
364,960
364,960
—
150
___
27
moles/hr.
^ -^•••^^^^-^^•^h^V^bM
STREAM 5
Ibs./hr.
^^MW^H^^»^H^^^»_M
11.7
325
59
15.9
moles/hr.
1,850
680
3,036
1,872
132
160
28
7,846
STREAM 6
Ibs./hr.
• ^iii
13.5
325
4
12.5
moles/hr.
102
2.1
3,688
3,036
1,872
132
160
28
9,037
STRE/'M 7
Ibs./hr.
11.916
77.769
2,264
13,654
341,287
446,890
—
300
—
26.4
nnles/hr.
-------
Table 68. STREAM CONDITIONS
-P-
vo
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
STREAM 8
Ibs./hr.
75.824
13,654
341 ,287
430,765
— _
710
—
76.4
moles/hr.
STREAM 9
Ibs./hr.
13,654
341,281
354.934
—
1040
—
112
moles/hr.
STREAM 10
Ibs./hr.
79.9
1300
109
139
moles/hr.
28
5.496
3,036
1 ,872
2.608
132
160
28
13,360
STREAM 1-1
Ibs./hr.
81.8
710
105
70.0
moles/hr.
182
28
5^3J__,
3.036
1,872
2,608
132
160
28
13,664
-------
Table 69. STREAM CONDITIONS
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H?
$63
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press. , in. H20
Heat, MM BTU/Hr.
STREAM 1-2
Ibs./hr.
81.8
1200
91
130
moles/hr.
182
28
5,616
3,036
1,872
2.605
132
160
28
13,664
STREAM 1 1
Ibs./hr.
23.344
0.2
6
78.7
1040
67
119
moles/hr.
1.850
5,984
3,036
1,872
132
160
28
13.150
STREAM 12
Ibs./hr.
78.7
250
63
18.4
moles/hr.
1,850
5,984
3,036
1,872
132
160
28
13,150
STREAM 13
Ibs./hr.
^_
46.9
110
61
2.1
moles/hr.
1,850
680
3,036
1,872
132
160
28
7,846
-------
Table 70. STREAM CONDITIONS
to
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
STREAM 1-3
Ibs./hr.
23,344
1
25,457
___
250
—
1.6
moles/hr.
STREAM 14
Ibs./hr.
23,344
25.456
___
250
—
1.6
moles/hr.
STREAM 15
Ibs./hr.
13.653
341,281
354.934
300
—
19.4
moles/hr.
STREAM 16
Ibs./hr.
3
275
278
—
77
—
0
moles/hr.
I
_ .
!
s
-------
Table 71. STREAM CONDITIONS
to
Ul
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH^
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press. , in. H20
Heat, MM BTU/Hr.
STREAM 17
lbs./hr.
2.0
77
2
0
moles/hr.
12
258
64
334
STREAM 18
lbs./hr.
3.676
3.676
__ _
150
—
0.3
moles/hr.
STREAM 1-4
lbs./hr.
77.0
150
9
7.1
moles/hr.
3,7fiQ
7,280
1,812
12.860
STRE/'M 19
Ibs./hr.
101
300
4
29.5
molfis/hr.
7,855
7,280
1.812
16,947
-------
Table 72. STREAM CONDITIONS
fo
Ul
SO 2
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
Tons Coal/Hour
STREAM 20
Ibs./hr.
69,864
..
69.864
—
160
_ __
5.6
moles/hr.
STREAM 21
17.7
STREAM 22
Ibs./hr.
23.5
—
___
—
moles/hr.
824
3,100
3,924
STREAM 23
Ibs./hr.
38.3
100
2
1.0
moles/hr.
28
90
3,036
326
1,062
1,678
160
28
6,408
-------
Table 73. STREAM CONDITIONS
N>
Ln
S02
H2S
H2504
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press. , in. H20
Heat, MM BTU/Hr.
STREAM 24
Ibs./hr.
125,136
125,136
41.6
280
—
ll.fi
moles/hr.
STREAM 25
Ibs./hr.
79.9
200
166
12.0
moles/hr.
28
7,042
3.036
326
1,062
1,678
160
28
13,360
STREAM 26
Ibs./hr.
79.9
675
155
63.2
moles/hr .
28
7,042
3.036
326
1,062
1,678
160
28
13,360
STRE/'M 26
Ibs./hr.
79.9
710
150
68.6
nn^es/hr.
28
5,496
3.036
1,872
2,608
132
160
128
13,360
-------
Table 74. STREAM CONDITIONS
to
Ul
m
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03 '
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
1 No. 2 Fuel Oil/Hr.
STREAM 28
1.027
STREAM 29
Ibs./hr.
4.0
77
2
0
moles/hr.
24
516
128
668
STREAM 30
Ibs./hr.
4,146
moles/hr.
STREAM 31
3,592
-------
Table 75. STREAM CONDITIONS
S02
H2S
H2S04
Elemental S
02
H20
N2
C02
H2
S03
CO
CH4
C2H6
Residual S
Carbon
TOTAL
M SCFM
Temp., °F
Press., in. H20
Heat, MM BTU/Hr.
#No. 2 Fuel Oil/Hr.
STREAM 32
.
*
4,847
STREAM
Ibs./nr.
-
moles/hr.
STREAM
Ibs./hr.
moles/hr.
STREAM
!bs./hr.
molejj/Jir.
-------
6.5 COSTS OF 1,000 MW CONCEPTUAL DESIGN INSTALLATION
6.5.1 Cost Summary
Based on the conceptual design flowsheet, heat and
material balances, the costs estimated for installation
of the Westvaco Process on a 1,000 MW boiler are
summarized below:
Table 76. COST SUMMARY
Capital Cost
£ Million
35
S/KW
35
Operating Cost
' $ Million/Yr
14.6
Mil/KWH
2.0
Details of the estimate are discussed in the following
sections and back-up information is contained in
Appendix K.
6.5.2 Capital Costs
The conceptual design flowsheet was used to estimate
the cost of installing a Westvaco Process in a 1,000 MW
power boiler. The summary is given in Table 77.
6.5.3 Equipment Costs
Purchased equipment cost estimates were based on actual
vendor quotes when available. When quotes were not
available, standard engineering estimating procedures
were used. This led to a basic purchased equipment cost
The total direct installation cost was then obtained by
factors given by Miller10. The detailed estimate is
given in Appendix K.
257
-------
Table 77. CAPITAL COST SUMMARY
Basic Equipment Costs $ 7,278,100
Equipment Installation Costs 7,350,900
Cost of Additional Battery Limit Items 5,611,800
TOTAL DIRECT COST - BATTERY LIMIT 20,240,800
Auxiliary Costs (Storage, Auxil., Serv.) 1,619.200
TOTAL BATTERY LIMIT + AUXILIARIES 21,860,000
Catalyst Costs (Carbon + Shift Catalyst) 1.703,200
TOTAL DIRECT COST $23.563.200
Engineering & Supervision Cost $ 3,298,800
Construction Cost 2,356,300
Contractor's Fee 942,600
Contingency 4,712,600
TOTAL INDIRECT COST $11,310,300
TOTAL INSTALLED COST =
$34,873,500
6.5.4 Indirect Costs
The indirect costs are made up of engineering, construc-
tion, contractor's fee, and an estimating contingency.
Factors for obtaining these costs are given by Peters and
TimmerhausH. The factor used for contingency was more
than twice that suggested by Peters and Timmerhaus.
6.6 OPERATING COSTS OF 1,000 MW CONCEPTUAL DESIGN INSTALLATION
The annual operating costs for the Westvaco Process at a
1,000 MW installation is about $14.6 million/year or 2.08
mills/KWH. The operating utilities estimates were based
on operation of the installation on an 8070 yearly basis.
These operating basis and capital cost change factors were
the same as those used by M. W. Kellogg in an earlier
comparison of the process to other S02 removal processes.
258
-------
The annual operating cost is given in Table 78. The
total direct costs were $7.3 million/yr. if no sulfur
credit is taken and $5.5 million/yr. if minimal credit
were taken. The indirect cost was about $1 million/yr.
and the fixed cost was $6.3 million/yr. This led to gross
operating costs of $14.6 million/yr. or 2.08 mills/KWH if
no sulfur credit is taken. The net operating cost is
$12.8 million/yr. or 1.83 mills/KWH if minimal sulfur
credit is taken.
259
-------
Table 78. ANNUAL OPERATING COSTS
PLANT SIZE (MW): 1,000
FIXED CAPITAL INVESTMENT (FCI): $34,873,500
STREAM TIME (HRS./YR.): 7,000
DIRECT COST
$/Year
$40/Lb. Carbon
1. Operating Labor (4 Men/Shift @ $5.50/Hr.) 154,000
2. Supervision - 15% of Item 1 23,100
3. Maintenance, Labor & Materials - 4% of FCI 1,394,900
4. Plant Supplies - 15% of Item 3 209,200
5. Utilities
a. Cooling Water - 1,430 GPM $0.1Z/Gal. 1,898,400
e. Coal - 17.6 TPH P $12/Ton 1,478,400
6. Chemicals & Raw Materials
a. Activated Carbon - 275 Ibs./hr. @ $0.40/lb. 770,000
7. SUBTOTAL Direct Cost (Excl. Credits) 7,303,200
8. Credits
a. Sulfur - 305 TPD @ $20/Short Ton -1,779.200
9. TOTAL DIRECT COST 5,524,000
INDIRECT COST
10. Payroll Overhead - 20% of (1+2) 35,400
11. Plant Overhead - 50% of (1+2+3+4) 890,600
12. TOTAL INDIRECT COST 926,000
FIXED COST
13. Capital Charges - 18.22% of FCI (Includes Depreciation Interim 6,354,000
Replacements, Insurance, Taxes and Cost of Capital)
TOTAL OPERATING COST
14. Net Production Cost - Items (9+12+13) 12,804,000
UNIT PRODUCTION COST
15. Gross - Items (7+12+13) 14,583,200
a. Mills/KWH 2 08
b. $/Ton Sulfur Not Emitted 163!g
16. Net - Items (9+12+13)
a. Mills/KWH ! 83
b. $/Ton Sulfur Not Emitted 143*9
260
-------
SECTION 7
15 MW DESIGN AND COST
7.1 INTRODUCTION
The technical feasibility of the Westvaco Process has
been demonstrated in a 20,000 cfh integral pilot unit.
A^detailed assessment of the process indicated economic
viability and competitiveness with other systems proposed
for sulfur gas control. The following section outlines
the program and cost to test the process on a large scale
in an actual utility. The test program is designed to
generate operating data for more detailed process assess-
ment and to demonstrate reliability and compatability
with utility operation.
7.2 SCOPE OF THE PROTOTYPE PROGRAM
The technical feasibility of the Westvaco Process has been
demonstrated in integral pilot plant tests. In order to
obtain engineering information for a detailed economic
assessment and for process scale-up, a prototype program
is'proposed for testing at the 15 MW (30,000 cfm) level.
The scope of the program is to assess the process perform-
ance and to obtain engineering data on the major questions
related to:
1. Compatability with the boiler interface
a. turndown, upsets and outage in boiler
b. reliability
c. fuel feed variations
d. safety
2. Control of process chemistry
a. response to upsets and variations
b. long term stability
3. Performance of activated carbon
a. mechanical
b. chemical
4. Performance of large scale fluid bed vessels
a. solid distribution
b. gas distribution
c. ease of operation.
261
-------
In attaining these objectives the scope of this program
includes:
1. Definition of boiler operating characteristics
2. Preparation of detailed prototype test program
3. Preparation of prototype design specifications
4- Detailed engineering design and bid evaluation
5. Construction
6. Start-up
7. Operation
8. Data evaluation and process technical and
economic assessment.
As an adjunct to the prototype program the scope of work
will also include, as necessary, additional testing in the
current pilot and other equipment to refine the prototype
design as seems necessary.
7.3 DESCRIPTION OF PROTOTYPE PLANT AND OPERATION
7.3.1 General
The prototype plant will consist of an integral closed loop
system continuously recycling granular carbon for removal
of S02 from approximately 30,000 cfm of actual power plant
flue gas from a coal fired boiler and recovering elemental
sulfur as a product.
The prototype size is equivalent to about 15 MW of the
boiler capacity with design capabilities to handle 10 - 20 MW.
The flue gas will be withdrawn between an electrostatic
precipitator and the boiler stack. The desulfurized flue
gas from the prototype plant will be returned to the stack.
The prototype plant consists of the following major proces-
sing steps:
1. S02 removal from flue gas.
2. Sulfur production from recovered S0£
3. Thermal stripping of sulfur product and produc-
tion of hydrogen sulfide
4. Condensation and recovery of sulfur product
5. Production of chemical reducing gas (hydrogen).
262
-------
Pilot plant experiments have established the technical
feasibility of the process and supplied the data base for
the prototype design basis. In the general design basis
that follows, specifications are given to accommodate the
range of conditions anticipated in the test program and
demonstration run.
7.3.2 General Design Basis
Boiler Characteristics -
Size Equivalent: 15 MW (30,000 scfm equivalent)
Type of Fuel: Coal
Sulfur Content: 1-4%
Load Variation: 10 - 20 MW
Flue Gas: 20,000-40,000 scfm; 800-3100 ppm S02
Process Conditions -
S02 Sorber -
Reaction: S02 + % 02 + H20 -> H2S04
Heat Release: 117,000 BTU/mol S02
Contact Mode: Gas/Solid Fluidized Bed, Stagewise
Fluidizing Gas Velocity: 2-4 ft./sec.
Space Velocity: 2,350 hr."1 (Design Based on
Experimental Rate Model)
Fluidized Bed Temperature: 150 - 300°F
Carbon Feed Rate: 4.4-8.8 M Ibs./hr.
SC-2 Rate: 250-1,000 Ibs./hr.
Sulfur Production -
Reaction: 3 H2S + H2S04 + 4 S + 4 H20
Heat Release: 41,107 BTU/mol S
Contact Mode: Fluidized Bed, Stagewise
Inlet Gas: 50,800 scfh with 22% H2S
Gas Velocity: 3 ft./sec.
Space Velocity: 475 hr."1 (Design Based on
Experimental Rate Model)
Temperature: 250 - 325°F
Carbon Feed Rate: 6,600 Ibs./hr. (Average)
263
-------
Carbon Residence Time: 40 minutes
Sulfur Formation Rate: 1,328 Ibs./hr.
Sulfur Stripping/H2S Formation -
Reaction: a) H2 + S •> H2S
b) 3 H2S + H2S04 •* 4 S + 4 H20 (Overall)
Heat Release: a) 8,667 BTU/mol S
b) 41,107 BTU/mol S
Contact Mode: Fluidized Bed, Stagewise
Inlet Gas: 92,600 scfh with 19.4% H£
Gas Velocity: 3 ft./sec. @ 1000°F
Space Velocity: 1600 hr.'r for C preheat and S stripping
3800 hr."1 for H2S formation
Temperature: 715°F (Preheat); 1000°F (Stripping)
Carbon Feed Rate: 6600 Ibs./hr. (Average)
Carbon Residence Time: 21 minutes [3.5 min. (Preheat) +
17.5 minutes (Stripping)]
Design Rates: 419 Ibs./hr. (S Stripping)
187 cfm (H2S Formation)
Sulfur Recovery -
Type: Condenser, Shell and Tube
Duty: 500 Ibs. S/hr.; 2 MM BTU/hr.
Temperature: 940°F, Inlet Gas
260°F, Outlet Gas
260°F, Liquid Product
Efficiency: 99.5% of Inlet Sulfur
Gasifier -
Type: Coal Feed with Steam/Air Blast
Feed: 716 Ibs./hr. Bituminous Coal (Maximum)
716 Ibs./hr. Anthracite Coal (Maximum)
Product Gas: 28 cf H2 + CO/lb. Coal
Activated Carbon Characteristics -
Size: 8x30 Mesh (Nominal; 1.5 MM Avg. Particle Size)
Density: 40-43 Ibs./ft.3
S02 Number: 75 (Minimum)
Attrition No.: 30 (Maximum)
264
-------
Instrumentation -
Adequate instrumentation is included, based on pilot plant
experience, to control and monitor temperature, pressure,
gas and carbon flow, and gas composition. Additional
instrumentation is also included in the design to assess
the various methods of overall process control under the
varying modes of utility boiler operation.
In addition to the major equipment items, auxiliary equip-
ment such as blowers, heat exchangers, conveyors, dust
collectors, etc. are included in the final design to
accommodate the range of operating conditions defined for
the major equipment items. Details of these items are
listed in Appendix L.
7.3.3 Process Description - Prototype Plant (Dwg. 2573, Fig. 78)
Flue gas from the precipitator at 300°F passes first
through the flue gas blowers (F-101) where the pressure is
boosted and then into the S0£ adsorber (FB-101). The
adsorber is 16 feet in diameter and contains five fluidized
stages of activated carbon. Sulfur trioxide is removed
from the hot gases in the bottom stage of the adsorber.
Water sprays above the second stage cool the gas to 170°F
prior to completion of the 90% sulfur oxide removal in
this and the remaining three stages. Conversion of sulfur
dioxide to sulfuric acid during sorption by the activated
carbon reheats the flue gas <3> to 200°F. A cyclone
collector (M-101) removes entrained carbon dust from the
flue gas prior to its return to the stack. The normal
dust loading of the clean flue gas from the cyclone would
be about .01 gr/scf with a maximum expected of .02 gr/scf.
A baghouse is included to measure the efficiency of the
cyclone and evaluate the total attrition rate. Recycled
activated carbon is fed to
the sulfur generator (FB-102) by a bucket elevator (V-101)
and feeder (V-201). In the sulfur generator (FB-102), a
3' diameter, 11 stage fluid bed, the sorbed sulfuric acid
is converted to elemental sulfur by reaction with hydrogen
sulfide<|>at 300°F. The heat of reaction is dissipated
by interstage water sprays as necessary. The activated
carbon becomes progressively loaded with elemental sultur
265
-------
TTMf ^ Bottflt Tl/tM- \ /
AevMtf 5l#HSLATt»ti \ /
S GENERATOR. I
STRIPPER
266
-------
Figure 78. Westvaco S02 Process flowsheet for 15 MW
prototype unit
n 10, me**t* •*
CHARLESTON RESEARCH CENTER
f. 0. BOX 5207 NORTH CHARLESTON. S. C.
IVfSrttVO SO, PROCESS FLOtvSHeCT
fcue ifMH f>*oTOTrr>e UNIT
26 6-A
-------
as it flows by gravity through the sulfur generator and
exits with a loading of 20 Ibs. S/100 Ibs. C<7>. The
additional sulfuric acid not converted to sulfur in this
reactor is converted in the H2S generator/sulfur stripper.
The elemental sulfur product is recovered by thermally
stripping concurrently with H2S formation in the H2S
generator/sulfur stripper (FB-104). The H2S generator/
sulfur stripper is a 70" diameter, 7 stage fluid bed with
1 stage for carbon preheat, 2 stages for H2S formation,
and 4 stages for stripping. The carbon is preheated with
the reducing gas at 1300°F containing about 20 vol. % H2-
The carbon flows from the preheat stage at 715°F, bypas-
sing two stages for H2S formation, to four stages for
sulfur removal at 940 - 1000°F. The off-gas from the
carbon preheater is reheated to 1300°F and passed to the
bottom of the H2S generator/sulfur stripper where it con-
tacts counter-currently the carbon to thermally strip the
sulfur. The hydrogen and sulfur subsequently are con-
verted to the H2S necessary for sulfuric acid conversion
in the sulfur generator. Final conversion of any remain-
ing acid to sulfur also occurs in the sulfur stripper/H2S
generator. The gas leaves the H2S generator/sulfur
stripper at 940°F and passes to a shell and tube sulfur
condenser to recover the elemental sulfur. The liquid
sulfur at 270°F is filtered of any dust before being
solidified by a sulfur flaker. The cooled gas at 270°F is
effectively free of sulfur. The regeneration gas is
supplied by a gas producer capable of gasifying anthracite,
coke, charcoal, or bituminous coal. The gas from the
gasifier passes through a shift converter before being
used in the carbon preheater-H2S generator/sulfur stripper.
Regenerated carbon <^> from the H2S generator at 1000°F
is cooled to 300°F in the carbon booler (E-501). Cooling
is by evaporation of water sprayed over the single bed of
fluidized carbon in the 70" diameter unit. Superheated
steam used as the fluidizing gas passes through a cyclone
and is exhausted.
The regenerated and cooled carbon
-------
7.3.5 Start-up and Initial Operation
The operating schedule for the prototype plant calls for a
three month period of start-up, a three month test program
and a six month demonstration run.
The start-up period will be used to work through the plant
putting all units into operation, checking their opera-
bility over the specified temperature, flow and pressure
ranges, and making any required adjustments, modifications
or replacements. At the conclusion of the start-up period,
the plant will be capable of accepting flue gas and circu-
lating carbon through the adsorption and regeneration
equipment.
The initial operating period will be devoted primarily to
establishing the operating characteristics of the process.
A material and heat balance will be obtained around the
plant which will be checked against calculated values to
determine if any significant deviations are occurring.
The S02 removal capability on both a once-through basis and
as a function of the number of cycles for a limited number
of cycles will be determined. The process control charac-
teristics, particularly stability and turndown capability,
will be tested. These data will be analyzed as they are
obtained. At the conclusion of the initial operating
period, the data will be reviewed and any changes needed in
the program for demonstration operation will be made.
7.3.6 Demonstration Operation
The demonstration operating period is intended to show the
capability of the Westvaco Process to operate reliably
under actual industrial conditions. The level of staffing
of the plant will be reduced to that anticipated
commercially. Boiler operating personnel will be used to
the greatest extent possible. The primary responses being
monitored during this period will be catalyst activity and
attrition characteristics, process stability and control.
Process operating information will be obtained for use in
scale-up design.
The equipment will be inspected and photographed. The
condition of all critical components will be noted.
7.3.7 Technical and Economic Review of Operation
After the initial and demonstration runs in the prototype
plant, a technical and economic review will be made of the
process. All pertinent information will be incorporated
into a final report covering the technical data and present-
ing an economic evaluation of the process.
268
-------
7.4 TECHNICAL APPROACH
7.4.1 General
The overall intent of the program is to assess the Westvaco
Process on a 15 MW slipstream of a coal fired boiler.
Prior to the engineering of the unit, Westvaco will prepare
a detailed program for the equipment operation and process
design specifications for the unit. These will be used by
an applications-engineering firm to prepare the detailed
design of the system and obtain bids for fabrication and
erection. After the acceptance of bids,both Westvaco and
the engineering company will monitor fabrication and
erection. Prior to start-up, Westvaco will train technicians
for operation during the test program and operators to be
supplied by the utility who will operate the unit during
the demonstration period. Data reduction,evaluation and
process assessment will be performed jointly by Westvaco
and the applications-engineering firm.
During the construction of the prototype unit, tests will
be made to define fully the operating characteristics of
the boiler as they would affect process operation.
As an adjunct to the prototype design, tests will be made
as necessary to evaluate proposed control modes, operating
ranges and design features of the prototype unit. Input
from these will be used to modify the prototype design as
required.
The prototype program is conceived to operate the process
under actual boiler conditions for sufficient cycles to
obtain information for process assessment and design of a
large scale unit.
7.4.2 Description of Program Elements
Based on pilot plant data"and boiler operating character-
istics a detailed test program was prepared for prototype
operation as given in Figure 79. The test program will
include the assessment of the effects of flue gas composition,
reducing gas composition temperature, boiler turndown, etc. on
the operating characteristics of the unit. Data obtained will
be used to modify design procedures as necessary for
process scale-up and assessment and to supply the input
for designing additional tests as deemed necessary. The
program will also include a demonstration run to assess
longer term reliability.
269
-------
Figure 79. Prototype program schedule
c
1. PREPARE DETAILED TEST PROGRAM.
2. PREPARE PROCESS DESIGN SPECIFICATIONS.
3. DETAILED ENGINEERING DESIGN & BIDS
H. PROCUREMENT AND CONSTRUCTION
5. DEFINITION OF BUILER OPR. CHARACTERISTICS
.6. OPERATION
A, OPERATOR TRAINING
B. START-UP
c, TEST PROGRAM
D. DEMONSTRATION RUN
7. DATA EVALUATION AND PROCESS ASSESSMENT
MONTHS
1 2 4 6.8 10 12 14 16 18 20 22 24 26 28 30 32 34 36
1 1 I
N-
•
-
i i i i i
•4
i i i i i
•
i i i i i
, ...
M
i i i i i
—
•_
*B
*B
i i i i i
— '
P *
ro
-vj
o
-------
Based on*the proposed test program and the pilot plant data
available, detailed process design specifications for the
prototype unit will be prepared by the process developer.
This will include the general design basis, process descrip-
tion, heat and material balances, process equipment specifi-
cations, instrument list, proposed layout and utilities
requirements. This will be assembled in report form for use
by the applications-engineering firm to prepare a detailed
design.
Using the process design specifications report from Westvaco,
the applications-engineering company will prepare a detailed
design of the prototype unit and secure bids for its fabri-
cation and construction.
The engineering contractor will procure and construct the
prototype unit.
During the construction phase an analytical program will be
conducted to define the characteristics of the boiler
operation. This will include: coal composition, flue gas
composition and temperature and load variation. Tests will
be conducted at intervals in order to establish trends.
During the construction phase operating manuals will be pre-
pared and training programs conducted to familiarize
development support personnel and selected plant operators
with the operation of the recovery process.
The start-up will include sequential testing of the proto-
type equipment to achieve design operating characteristics.
Modifications will be made as necessary. Operators will be
given actual experience in equipment operation.
The test program will evaluate steady state and transient
behavior of the system with variations in flue gas composi-
tion, reducing gas composition, temperature, boiler load,
etc. An anticipated 100 - 125 cycles will be completed
during the test program.
A six month demonstration run will be conducted to define
the long term reliability of the process. Over the six
month period 300 - 360 process cycles would be completed
and there would be 1.25-1.5 complete inventory turn-overs.
Beginning with the test program there would be continuing
analysis of data. This would be incorporated in a
detailed assessment at the end of the test program which
would be updated after the demonstration run.
271
-------
7.5 COST OF PROTOTYPE PROGRAM
The summary of costs for the prototype program is shown
below in Table 79:
Table 79. COST OF PROTOTYPE PROGRAM
1. Installed Equipment $2,411,000
2. Design Engineering 240,000
3. Manpower, R&D $303,246
Operation 78,300
Maintenance 63.288 444 , 834
4. Overhead 422,600
5. Consultants 14,400
6. Travel 14,400
7. RawMat'l., Utilities
& Supplies
TOTAL COST $3,868,994
The total prototype cost was estimated as follows:
Purchased equipment costs were estimated using general
engineering methods and budget quotes from the equipment
lists contained in Appendix L . The total installed
equipment cost was derived from the purchased equipment
cost based on the factors given by Miller10. The cost
breakdown is shown in Appendix L.
Engineering costs were estimated from factors presented
in Peters and Timmerhaus11.
The required manhours for R&D, operators and maintenance
were derived using the development schedule and assigning •
manhours based on experience in Westvaco's prior development
work and plant experience. The manhour breakdown and cost
breakdown are shown in Appendix L.
A provisional overhead on manpower costs was estimated at
95% of direct labor costs based on past experience.
Raw materials and utilities were determined from energy
and material balances for the prototype unit and applied
to the number of operating hours at rates shown in
Appendix L.
272
-------
SECTION 8
BIBLIOGRAPHY
1. Russel, J. H., Town, J. W., and Kelly, H. J., U. S.
Bureau of Mines Report of Investigation 7415 (August
1970) .
2. Juza, R. and Blanke, W., Z. Anorg. Chem. 210, 81 (1933).
3. Tuller, W. N., Ed., THE SULFUR DATA BOOK, McGraw-Hill
Book Co., New York, 1954.
4. Aynsley, E. D., Pearson, T. G., and Robinson, P. L.,
J. Chem. Soc.. 58-68 (1935).
5. Norrish, R. G. W., and Rideal, E. K., J. Chem. Soc. 123,
1668-1704 (1923).
6. Hart, P. J. , Vastola, F. J. and Walker, P. L. , Jr'. ,
Carbon 5, 363-371 (1967).
7. Bansal, R. C., Vastola, F. J. and Walker, P. L., Jr.,
Carbon 5, 185-192 (1971).
8. Puri, B. R., Singh, D. D., Nath, J. and Sharma, L. R.,
Industrial and Engineering Chemistry 50(7):1071-1074
(1958). ~~~
9. Kunii, D., and Levenspeil, 0., FLUIDIZATION ENGINEERING,
John Wiley & Sons, Inc., New York (1969).
10. Miller, C. A., Chemical Engineering 72, 21-29 (Sept. 13,
1965).
11. Peters, M. S., and Timmerhaus, K. D., PLANT DESIGN AND
ECONOMICS FOR CHEMICAL ENGINEERS, McGraw-Hill Book Co.,
New York, 1968.
12. Brunauer, S., P. H. Emmett, and E. Teller, J. Amer.
Chem. Soc. 60, 309-319 (1938).
13. Roberts, B. F., J. Colloid and Interface Science 23,
266-73 (1967).
14. Prchlik, J., Hlinak, L. and Kartakova, C. D., "Recovery
of Sulfur by Extraction from Spent Gas Works Purification
Mass", Paliva (Czech. Journal^ 34, 298-305 (1954).
15. Russell, J. H., J. W. Town, and H. J. Kelly, Bureau of
Mines Report of Investigations 7415 (August 1970).
273
-------
17. Smisek, M. , and Cerny, S., ACTIVE CARBON, American
Elsevier Publishing Company, New York (1970).
18. Hildebrand-Jenks, J. Am. Chem. Soc. 43,2172-7 (1921).
19. Jacek, Rocz. Chem. 6. 501-9 (1926).
20. Delaplace, J. Pharm. Chim. 26, 139 (1922).
21. Burden, F. A., and W. B. S. Newling, The Inst. of Gas
Eng., No. 392, pg. 27-28 (Nov. 1951).
22. Hammick, D. L. and S. E. Holt, J. Chem. Soc. 129,
1995-2003 (1926).
23. Levenspiel, 0., CHEMICAL REACTION ENGINEERING, John
Wiley & Sons, Inc., New York (1962).
274
-------
SECTION 9
NOMENCLATURE
A - Constant
a - Constant
Ao - Gas Distributor Open Area, 70
B - Constant
b - Constant
C - Gas Concentration, Ib. moles/ft.3
D - Reactor Diameter
d0 - Orifice Diameter
E - Activation Energy
F - Molar Flow Rate
g - Function
H2S - Hydrogen Sulfide Gas Concentration
h - Carbon Bed Height
k - Rate Constant
ko - Frequency Factor for Arrhenium Equation
L - Equilibrium Sulfur Loading Sorbed Carbon or
Carbon Bed Loading
Lc - Amount of Sulfur Chemisorbed by Carbon
N - Moles
N0 - Initial Moles
P - Equilibrium Vapor Pressure of Sulfur
Ps - Saturation Vapor Pressure of Sulfur with Carbon
q - Gas Flow Rate or Differential Heat of Adsorption
of Sulfur by Carbon
R - Gas Law Constant or Carbon Flor Rate
Re - Reynolds Number
275
-------
r - Rate of Reaction
S - Sulfur Gas Concentration as S
T - Temperature
t - Time
U - Gas Velocity
V - Reactor Volume
v - Superficial Linear Gas Velocity, ft./sec.
X - Carbon Loading, weight of material/weight of carbon
- Average Number of Sulfur Atoms/Molecule
Greek Symbols
p - Density
y - Viscosity
A - Carbon Weepage Rate
Subscripts
c - Carbon
g - Of Gas Phase
H - Hydrogen
H20 - Water
H2S - Hydrogen Sulfide
j - Stage Number j
j+1 - Stage Number j+1
Mf - Minimum Fluidizing
NO - Nitric Oxide
02 - Oxygen
276
-------
p - Of Carbon Phase
S - Sulfur
S0£ - Sulfur Dioxide
T - Total
v - Sulfuric Acid
vs - Sulfuric Acid Saturation
1 - At Condition 1
2 - At Condition 2
Superscripts
A - Order of Reaction
m - Order of Reaction
n - Order of Reaction
p - Order of Reaction
q - Order of Reaction
277
-------
TECHNICAL REPORT DATA
(Please read Inalnictiunf on the reverse before completing}
1. REPORT NO.
2.
EPA- 600/2 -76-135a
4. TITLE AND SUBTITLE
Development of the Westvaco Activated Carbon
Process for SOx Recovery as Elemental Sulfur,
Volume I
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
May 1976
6. PERFORMING ORGANIZATION CODE
7.AUTHOR(S,G.N. Brown, C.M. Reed, A.J. Repik,
R.L. Stallings, andS.L. Torrence
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Westvaco
Charleston Research Center
Box 5207, North Charleston, SC 29406
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21ACX-085
11. CONTRACT/GRANT NO.
68-02-0003
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
13. TYPE OF REPORT AN!
Final: 1/71-6/74
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES pr()ject Officer fQr
Ext 2557.
report is D.A. Kemnitz, Mail Drop 62,
16. ABSTRACT The report gives results of a demonstration (in a 20,000-cfh integral pilot
plant) of an all-dry, fluidized-bed process, using activated carbon for recovering
SO2 as elemental sulfur. Granular carbon was recycled continuously more than 20
times between contact with flue gas from an oil-fired boiler and carbon regeneration
to recover sulfur. During the 315-hour run, carbon performance remained high with
essentially no chemical and low mechanical losses. Over 90% of the 2000 ppm SOx
was removed from the flue gas as sulfuric acid by catalytic oxidation and subsequent
hydrolysis within the carbon granule. In the two-step regeneration: (1) the acid was
converted to elemental sulfur at 300F with internally produced H2S, and (2) an exter-
nal source of hydrogen at 1000F was used to thermally strip the by-product sulfur
from the carbon and produce the required H2S by reaction with the remaining sulfur
on carbon. Sufficient process and design information was developed from data ob-
tained in the integral run and prior stepwise pilot equipment operation to permit
scale-up to a 15-MW prototype for a coal-fired boiler. In the preliminary design,
reducing gas is produced in a coal gasifier. An economic assessment of a 1000-MW
conceptual design for the process indicates capital and operating costs competitive
with those of other regenerable systems.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
cos AT I Field/Group
Air Pollution
Flue Gases
Activated Carbon
Sulfur Oxides
Fluidized Bed
Processing
Regeneration
(Engineering)
Fuel Oil
Sulfuric Acid
Catalysis
Oxidation
Air Pollution Control
Stationary Sources
Elemental Sulfur
Westvaco Process
Catalytic Oxidation
13B
2 IB
11G 21D
07B
07D
13H,07A 07C
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
298*
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
278
------- |