EPA-600/7-89-015d
December 1989
FIELD EVALUATION OF LOW-EMISSION COAL BURNER
TECHNOLOGY ON UTILITY BOILERS
VOLUME IV
Alternate Concepts for SO^, NO^,
and Particulate Emissions Control From a
Fuel-Rich Precombustor
J. F. LaFond, J. A. Cole, W. C. Li, E. C. Moller, R. Payne
(Energy and Environmental Research Corporation)
and
P. W. Waanders
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH 44203
EPA Contract 68-02-3130
EPA Project Officer: P. Jeff Chappell
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/7-89-0l5d
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Field Evaluation of Low-Emission Coal Burner Tech-
nology on Utility Boilers; Volume IV. Alternate Con-
cepts for SOx, NOx, and Particulate Emissions*
5. REPORT DATE
December 1989
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J. F. LaFond, J. A. Cole, W. C. Li, E. C. Moller, and
R. Pavne(EERC); and P. W. Waanders (B/W)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Energy and Environmental Research Corporation
18 Mason
Irvine, California 92718
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3130**
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 9/78 - 6/86
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notes AEERL project officer is P. Jeff Chappell, Mail Drop 63, 919/541-
3738. (*) Control from a Fuel-Rich Precombustor. (**) Contract with Babcock and
Wilcox, P. O. Box 351, Barberton, OH 44203-0351.
i6. abstract rep0rf- gives results of a study of the use of precombustors for the si-
multaneous control of S02, NOx, and ash emissions from coal combustion. In Phase
1, exploratory testing was conducted on a small pilot scale-~293 kW (million Btu/hr)-
- pulverized-coal-fired precombustor to identify critical operating parameters. The
results from this testing raised several questions regarding the viability of control-
ling S02 emissions by injecting calcium-based sorbent materials, under conditions
simultaneously conducive to NOx control, and to the rejection of coal ash as a mol-
ten slag. In Phase 2, key elements of the sulfur capture process, under the fuel-rich
precombustor conditions necessary to control NOx formation, were investigated. De-
tailed experimental studies were conducted at bench aid laboratory scales to: (l) in-
vestigate the formation of stable sulfides in the entrained flow region of a precombus-
tor, using calcium-based sorbents; (2) study the evolution of sulfur from coal under
entrained flow combustion conditions; and (3) investigate the stability of sulfur spe-
cies in molten slag layers. Study results indicated that the sulfidation reactions be-
tween CaO and H2S or COS are fast and, under optimum conditions, can remove a
high fraction of the gas-phase sulfur species in a fuel-rich precombustor. s:
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. cosati Field/Group
Pollution Calcium
Coal Electric Utilities
Combustion
Sulfur Dioxide
Nitrogen Oxides
Particles
Pollution Control
Stationary Sources
Particulate
Precombustors
13 B
21D
2 IB
07B
14G
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
254
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
i
-------
ATTENTION
PORTIONS OF THIS REPORT ARE NOT LEGIBLE.
HOWEVER, IT IS THE BEST REPRODUCTION
AVAILABLE FROM THE COPY SENT TO NTIS.
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
ii
-------
ABSTRACT
This report describes the results of a study related to the use of
precombustor concepts for the simultaneous control of S02> N0X and ash
emissions from coal combustion. In a first phase of the study, exploratory
testing was conducted on a small pilot scale (293 kW — 1 x 10® Btu/hr)
pulverized-coal-fired precombustor to identify critical operating parameters.
The results from this testing raised several questions regarding the
viability of controlling SO2 emissions through the injection of calcium-based
sorbent materials, under conditions simultaneously conducive to N0X control,
and to the rejection of coal ash as a molten slag.
In a second phase of the study key elements of the sulfur capture
process, under the fuel-rich precombustor conditions necessary to control N0X
formation, were investigated. Detailed experimental studies were conducted
at bench and laboratory scales in order to: 1) investigate the formation of
stable sulfides in the entrained flow region of a precombustor using calcium
based sorbents, 2) study the evolution of sulfur from coal under entrained
flow combustion conditions, and 3) investigate the stability of sulfur
species in molten slag layers.
Results from the study indicated that the sulfidation reactions between
CaO and H2S or COS are fast, and under optimum conditions can remove a high
fraction of the gas phase sulfur species in a fuel-rich precombustor.
However, typical coal ash and ash/CaO mixtures were found to have a low
capacity for retaining large amounts of sulfur under equilibrium conditions
in a precombustor environment. In order to achieve significant fuel-rich
sulfur capture, the implication for practical precombustor systems is that
rapid slag drainage systems are required, or that slag fluxing additives may
be necessary.
The work described in this report has been supported by the United
States Environmental Protection Agency through Contract No. 68-02-3130 to
Babcock and Wilcox (B&W) and to Energy and Environmental Research Corporation
under B&W subcontract No. 940962NR.
i i i
-------
TABLE OF CONTENTS
Section Page
1.0 PROGRAM SUMMARY 1-1
1.1 Introduction 1-1
1.2 Scope 1-3
1.3 Experimental Systems 1-4
1.4 Results 1-4
1.5 Conclusions 1-21
2.0 INTRODUCTION 2-1
2.1 Precombustor Concept. . 2-1
2.2 Sulfur Control 2-3
2.3 Program History 2-8
3.0 BACKGROUND 3-1
3.1 Precombustor Experience 3-1
3.2 Review of Slag Related Industry and Research. . . . 3-15
3.3 Calcium Sulfidation 3-30
4.0 OBJECTIVES AND APPROACH 4-1
5.0 EXPERIMENTAL SYSTEMS 5-1
5.1 Isothermal Reactor (ITR) 5-1
5.2 Control Temperature Tower (CTT) 5-5
5.3 High Temperature Oven (HTO) 5-9
5.4 Continuous Flue Gas Sampling System 5-13
5.5 Reduced Sulfur Species Sampling System 5-17
5.6 Solid Sampling Techniques 5-17
5.7 Sorbent Feeding 5-21
5.8 Temperature Measurement 5-21
5.9 Test Material Compositions 5-21
v
-------
TABLE OF CONTENTS (Continued)
Section Page
6.0 ENTRAINED FLOW SULFIDATION RESULTS 6-1
6.1 Isothermal — Gas-Fired 6-1
6.1.1 Fuel Effects 6-1
6.1.2 Impact of H2S Injection Location 6-3
6.1.3 Time-Resolved Capture 6-7
6.1.4 Stoichiometry Effects 6-7
6.1.5 Temperature Effects . 6-9
6.1.6 Impact of Sorbent Type 6-9
6.1.7 Gas Phase Versus Solids Sampling 6-14
6.1.8 Comparison of Fuel-Rich to
Fuel-Lean Capture 6-14
6.2 Slow Quench Rate ~ Coal-Fired 6-14
6.2.1 Stoichiometry Effects 6-16
6.2.2 Load Effects 6-22
6.2.3 Impact of Sorbent Type and Feed Rate .... 6-22
6.2.4 Effect of Injection Temperature 6-27
7.0 SLAG SULFUR CHEMISTRY RESULTS 7-1
7.1 Sulfur Solubility and Equilibrium 7-1
7.2 Impact of Slag Fluxing Additives 7-10
7.3 Slag Desulfurization Rates 7-14
7.4 In-Situ Sulfur Capture in Slag 7-17
8.0 QUALITY CONTROL EVALUATION 8-1
8.1 Instrument Accuracy and Precision 8-1
8.2 System Uncertainties and Data Completeness 8-1
9.0 CONCLUSIONS 9-1
10.0 RECOMMENDATIONS 10-1
vi
-------
TABLE OF CONTENTS (Concluded)
Section Page
11.0 ACKNOWLEDGEMENTS 11-1
12.0 REFERENCES 12-1
APPENDIX A — QUALITY ASSURANCE/QUALITY CONTROL (QA/QC) .... A-l
A. 1 EXPERIMENTAL SYSTEMS A-l
A. 1.1 Gas Species Sampling A-l
A. 1.2 Solids Sampling A-6
A. 1.3 Temperature Measurements A-6
A. 1.4 Gas Flow Metering A-16
A. 1.5 Sorbent Feed Rate A-16
A. 2 ANALYTICAL A-17
A.2.1 Continuous Gas Species Analysis . A-17
A.2.2 Gas Chromatography Analysis . . . A-17
A. 2.3 Solids Analysis A-17
A. 2.4 Data Representativeness and
Completeness A-22
A. 3 CALIBRATION PROCEDURES A-23
A. 4 DATA REDUCTION A-33
A.4.1 Calcium Utilization A-33
A.4.2 Sulfur Capacity A-35
A.4.3 Equilibrium Sulfur Capture by
Slag A-36
A.4.4 Furnace Residence Time A-36
APPENDIX B — EXAMPLES OF RAW DATA. B-l
B.l ITR Data Example B-2
B.2 CTT Data Example B-7
B.3 HTO Data Example. B-24
vii
-------
LIST OF FIGURES
Figure Page
1-1 Common cyclonic precombustor in staged combustion
configuration attached to boiler furnace wall . 1-2
1-2 Sulfur speciation in the ITR at 1100°C, SR=0.83 when
either mixing 5000 ppm H2S with the fuel (C2H4) or
injecting it downstream of the flame 1-5
1-3 Calcium utilization in the ITR using Linwood atmospheric
hydrate, measured by both gas and solid analysis 1-6
1-4 Calcium utilization in the ITR at SR=0.83 when
injecting 5000 ppm H2S downstream of a C2H4 flame 1-7
1-5 Calcium utilization in the ITR under fuel-rich and
fuel-lean conditions for both Linwood atmospheric
hydrate and Vicron 45-3 limestone 1-9
1-6 Sulfur speciation in the CTT firing Illinois coal
at 17.6 kW (60,000 Btu/hr) 1-10
1-7 Calcium utilization in the CTT firing Illinois coal
at 17.6 kW (60,000 Btu/hr), injecting Linwood
atmospheric hydrate at about Ca/S = 1.0 1-11
1-8 Species evolution from Illinois coal in the CTT
at 17.6 kW (60,000 Btu/hr) 1-12
1-9 Capture of total sulfur in the CTT by the injection
of sorbents when firing Illinois coal at 23.4 kW
(80,000 Btu/hr), SR=0.6 1-13
1-10 Influence of temperature and molar basicity on
sulfur capacity for several slag mixtures 1-14
1-11 Equilibrium sulfur removal by slag for SR=0.5,
T=1200°C, and an Illinois coal with 3.97% sulfur. ..... 1-16
1-12 Equilibrium sulfur removal by slag for SR=0.5,
T=1400°C, and an Illinois coal with 3.97% sulfur 1-17
1-13 Fluid and T250 temperatures for Illinois coal
ash/CaO mixtures 1-18
1-14 Influence of fluxing additives on the reducing fluid
temperature of Illinois coal ash and CaO mixtures 1-19
1-15 Desulfurization of CaS and Illinois coal ash slag
mixtures, showing the influence of calcium content
and mixedness 1-20
vi i i
-------
LIST OF FIGURES (Continued)
Figure Page
2-1 Common cyclonic precombustor in staged combustion
configuration attached to boiler furnace wall 2-2
2-2 Equilibrium stability of solid calcium-sulfur compounds . . 2-4
2-3 Calcium sulfate formation at elevated injection
temperatures 2-5
2-4 Schematic diagram of some of the important
mechanisms occurring in a fuel-rich precombustor
system with sorbent addition 2-7
3-1 Precombustor arrangement and refractory detail 3-3
3-2 Mean gas temperatures and bulk residence times in
the precombustor and furnace (SRi = 0.5, 0.3 MWt) 3-4
3-3 Equilibrium distribution of sulfur species under the
reducing conditions of the precombustor for Illinois
coal with and without calcium present 3-5
3-4 Stack N0X emissions as a function of SRi for
Illinois coal and natural gas (SRy = 1.2) 3-6
3-5 Sulfur capture when firing natural gas (0.3 MWt)
and injecting Vicron 45-3 and dolomite with the
primary air at Ca/S = 2 (SRj = 1.2) 3-8
3-6 SO2 capture based on gas phase stack measurements
and potential SO2 capture based on calcium
utilization of solids samples 3-9
3-7 Effect of increased solids removal on sulfur
capture for two fuels. Vicron 45-3 injected
with primary air (SRi = 0.5, 0.3 MWt) 3-10
3-8 Effect of sorbent additive (5% Fe203 by weight) on
sulfur capture, sorbent injected with primary air
(natural gas, SR^ = 0.5, 0.3 MWt) 3-12
3-9 Cross-section of the VCC prototype combustor (0.9 MWt). • • 3-13
3-10 Recent ash retention efficiency measurements and
system pressure drop versus firing rate
(La Fond et al., 1986) in the VCC * . 3-14
3-11 N0X control by combustion staging and natural gas
reburning in the VCC; Illinois coal, SRy =1.2 3-16
ix
-------
LIST OF FIGURES (Continued)
Figure Page
3-12 Effect of stoichiometry and Ca/S molar ratio on
sulfur capture when injecting sorbent into the
VCC under staging conditions 3-17
3-13 Cross-sectional view of blast furnace 3-18
3-14 Iron desulfurization process schematic 3-20
3-15 Ladle positions in the prpduction of steel by the
basic oxygen furnace. . 3-21
3-16 Sulfur capacities, at 1650°C, in Ca0-Si02-Al2O3 melts . . . 3-23
3-17 Desired slag compositions in the Armco top slag
desulfurization process 3-24
3-18 Ternary diagram for CaO-Al203-SiO2 melts;
temperature in °C 3-25
3-19 Ternary diagram for FeO-Al203-SiO2 melts; scales
in weight percent and temperature in °C 3-26
3-20 Ternary diagram for Fe0-Na20-Si02 melts; scales
in weight percent and temperature in °C 3-29
3-21 Sulfur capacities Cs, in various slag systems at 1650°C . . 3-31
3-22 Comparison of sulfur capture capability for four
different sorbent materials under fuel-lean conditions. . . 3-33
4-1 Program approacn flow-chart 4-2
5-1 Schematic of the isothermal reactor (ITR) 5-2
5-2 Schematic detail of the precombustor/sliding
burner assembly 5-3
5-3 Cross-sectional views of the CTT configured
with a fuel-rich zone 5-6
5-4 Multi-layer refractory design of the CTT 5-7
5-5 Premixed burner 5-8
x
-------
LIST OF FIGURES (Continued)
Figure Page
5-6 Temperature profiles measured in the fuel-rich zone
of the CTT for three different operating conditions
when burning Illinois coal 5-10
5-7 High temperature oven test facility with gas
blending capability 5-11
5-8 High temperature oven facility temperature distributions. . 5-14
5-9 Phase discrimination probe tip 5-15
5-10 Continuous monitoring flue gas sampling system 5-16
5-11 SO2 continuous monitoring and batch sampling system .... 5-18
5-12 Isokinetic solids sampling probe with upper and lower
cooling jackets used in the ITR 5-19
5-13 Solid sampling system used on the CTT 5-20
5-14 Fine-wire thermocouple probe for gas temperature
measurements in the ITR 5-22
5-15 Gas temperature measurement system for the CTT 5-23
5-16 Comparison of sorbent size distributions as
measured by x-ray sedigraph technique 5-25
6-1 Sulfur speciation in the ITR at 1100°C, SR = 0.83 when
doping CH4 and C2H4 with approximately 5000 ppm H2S .... 6-2
6-2 Calcium utilization in the ITR using Linwood atmospheric
hydrate, measured by both gas and solids analysis 6-4
6-3 Sulfur speciation in the ITR at 1100°C, SR = 0.83
when either mixing 5000 ppm H2S with the fuel (C2H4)
or injecting it downstream of the flame 6-5
6-4 Calcium utilization in the ITR using Linwood atmospheric
hydrate, measured by both gas and solid analysis 6-6
6-5 Sulfur speciation with and without sorbent present
in the ITR at 1100°C, SR - 0.83 6-8
6-6 Sulfur speciation in the ITR at 1100°C when
injecting 5000 ppm H2S downstream of the C2H4 flame .... 6-10
xi
-------
LIST OF FIGURES (Continued)
Figure Page
6-7 Sulfur speciation in the ITR at sr = 0.83 when
injecting 5000 ppm H2S downstream of the C2H4 flame .... 6-11
6-8 Calcium utilization in the ITR at SR = 0.83 when
injecting 5000 ppm H2S downstream of a C2H4 flame 6-12
6-9 Calcium utilization in the ITR at 1100°C, SR = 0.83 with
5000 ppm H2S injected downstream of the C2H4 flame 6-13
6-10 Calcium utilization in the ITR under fuel-rich and
fuel-lean conditions for both Linwood atmospheric
hydrate and Vicron 45-3 limestone 6-15
6-11 Sulfur speciation in the CTT firing Illinois coal at
17.6 kW (60,000 Btu/hr) . 6-17
6-12 Calcium utilization in the CTT firing Illinois coal
at 17.6 kW (60,000 Btu/hr), injecting Linwood
atmospheric hydrate at about Ca/S = 1.0 6-18
6-13 Species evolution from Illinois coal in the CTT
at 17.6 kW (60,000 Btu/hr) 6-19
6-14 Comparison of measured gas phase sulfur speciation
from an SR = 0.8 Illinois coal flame to equilibrium
speciations calculated at conditions which correspond
to the local measured stoichiometry and temperature .... 6-21
6-15 Calcium utilization in the CTT firing Illinois coal
at SR = 0.6 injecting Linwood atmospheric hydrate
at about Ca/S = 1.0 6-23
6-16 Species evolution from Illinois coal in the
CTT at SR = 0.6 6-24
6-17 Sulfur speciation in the CTT firing Illinois
coal at SR = 0.6 6-25
6-18 Capture of total sulfur in the CTT by the injection
of sorbents when firing Illinois coal at 23.4 kW
(80,000 Btu/hr), SR = 0.6 6-26
6-19 Capture of gas phase sulfur in the CTT by the
injection of sorbents when firing Illinois coal
at 23.4 kW (80,000 Btu/hr), SR = 0.6 6-28
xii
-------
LIST OF FIGURES (Continued)
Figure Page
6-20 Calcium utilization as a function of Linwood atmospheric
hydrate injection temperature in the CTT firing Illinois
coal at 23.4 kW (80,000 Btu/hr), SR = 0.6 6-29
7-1 Slag exposure time requirements for Illinois coal ash/
CaO slag corresponding to a Ca/S = 2 at 1400oC, SR =0.5. . 7-3
7-2 Influence of temperature and molar basicity on
sulfur capacity for several slag mixtures 7-5
7-3 Sulfur capacity for iron silicate melts at various
initial slag molar basicities 7-6
7-4 Equilibrium sulfur removal by slag for SR=0.5, T=1200°C,
and an Illinois coal with 3.97% sulfur 7-7
7-5 Equilibrium sulfur removal by slag for SR = 0.5,
T = 1400°C, and an Illinois coal with 3.97% sulfur 7-9
7-6 Fluid and T250 temperatures for Illinois coal
ash/CaO mixtures 7-11
7-7 Influence of fluxing additives on the reducing fluid
temperature of Illinois coal ash and CaO mixtures ..... 7-12
7-8 Desulfurization of CaS and Illinois coal ash slag
mixtures, showing the influence of calcium content
and mixedness 7-15
7-9 Desulfurization of CaS/Illinois coal ash/additive
slag mixtures 7-16
7-10 Desulfurization plot of CaS and Illinois coal
ash mixtures plotted to obtain the rate constants
presented in Table 7-3 7-18
A-l SO2 continuous monitoring and batch sampling system .... A-2
A-2 Continuous monitoring flue gas sampling system A-4
A-3 SO2 continuous monitoring and batch sampling system .... A-5
A-4 Depletion of sulfur species within a tedlar sample
bag as a function of storage time A-7
A-5 Isokinetic solids sampling probe with upper and lower
cooling jeckets used in the ITR A-8
xiii
-------
LIST OF FIGURES (Concluded)
Figure Page
A-6 Solid sampling system used on the CTT A-9
A-7 Fine-wire thermocouple probe for gas temperature
measurements in the ITR A-10
A-8 Gas temperature measurement system for the CTT A-14
A-9 Suction pyrometer aspirator pressure calibration for CTT. . A-15
A-10 Example of daily standard checks for the gas
chromatograph during one calibration period A-20
A-ll H2S calibration curves for 70, 50 and 30 1
syringe injection A-26
A-12 SO2 calibration curves for 70, 50 and 30 1
syringe injection A-27
A-13 COS calibration curve for 70 1 syringe injection A-28
A-14 CS2 calibration curve for 70 1 syringe injection A-29
A-15 Gas chromatograph printout example A-30
A-16 Gas chromatograph printout example. . A-31
A-17 Gas chromatograph printout example A-32
xiv
-------
LIST OF TABLES
Table Page
3-1 Effect of B2O3 Substitutions on Melting Time
for the Na20-Ca0-B203-Si02 System 3-28
5-1 Temperature Variation in the ITR at a Nominal
900°C Isothermal Condition 5-4
5-2 Feed Gas Compositions 5-12
5-3 Properties of Vicron Limestone and Linwood
Atmospheric Hydrate 5-24
5-4 Illinois Coal Composition and Heating Value 5-27
5-5 Mineral Analysis of Illinois Coal Ash 5-28
5-6 Compositions of CaO/Ash and Fe203/Ash Mixtures 5-29
7-1 Feed Gas Compositions 7-2
7-2 Influence of Slag Additives on Sulfur Capacity
Ca/S = 1; log Cs Determined at 1400°C, SR = 0.5 7-13
7-3 Desulfurization Rate Constants for CaS/Illinois
Coal Ash Mixtures 7-19
8-1 Precision and Accuracy for Analysis Instruments
and Techniques 8-2
A-l Radial Temperature Change in the ITR at a Nominal
900°C Isothermal Condition A-13
A-2 CTT Instrumentation A-18
A-3 Precision, Accuracy and Completeness Data for
Continuous Monitoring Gas Species Instruments A-19
A-4 Solid Analyses Precision and Accuracy Data A-21
A-5 Calibration Procedures A-24
A-6 Calibration Gases A-25
xv
-------
1.0 PROGRAM SUMMARY
1.1 Introduction
The potential for simultaneous control of ash, N0X and S0X emissions
from coal-fired boilers and heaters by combustion exterior to the furnace
entry has made precombustor development an area of great interest and effort.
A precombustor burns coal in a chamber outside the normal furnace region. An
example of a simple precombustor scheme is presented in Figure 1-1.
Aerodynamic separation and slag drainage removes a majority of the coal
mineral matter before entry into the furnace. Also, staged combustion and
reburning have been shown to be effective means of controlling N0X emissions.
It has been proposed that the use of calcitic sorbent or possibly other
additives in a fuel-rich precombustor can produce significant reductions in
overall SO2 emissions. Successful control of all three pollutants would
allow coal users to circumvent expensive exhaust stream clean-up equipment
and help avoid derating in oil or gas retrofit applications.
The issue of sulfur capture under fuel-rich conditions has been an area
of uncertainty and much recent interest. The fuel-rich reactions of:
H2S + CaO ~ CaS + H2O or
COS + CaO ~CaS + CO2 ,
are theoretically more effective at capturing gas phase sulfur than the well
studied fuel-lean reaction:
SO2 + CaO + 1/2 O2 - CaSO^ ,
both from a thermodynamic and kinetic standpoint. However, at the time this
program was initiated the operating conditions which promote these fuel-rich
reactions had not been fully investigated. In addition, the presence of a
liquid slag in the reactor was thought to be a potential source of sulfur
capture or regeneration which required additional research.
1-1
-------
Figure 1-1. Common cyclonic precombustor in staged combustion
configuration attached to boiler furnace wall.
1-2
-------
Efforts in this program have focused on both entrained flow sorbent
sulfidation by dry powder injection and the interaction of sulfur, coal ash
and additives in molten slags under fuel-rich conditions.
The work described in this report has been supported by the United
States Environmental Protection Agency through Contract No. 68-02-3130 to
Babcock and Wilcox (B&W) and to Energy and Environmental Research (EER) under
B&W subcontract No. 940962NR.
1.2 Scope
The program involved two phases. The first phase consisted of
exploratory testing of a pilot scale coal precombustor (293 kW (1 x 10^ Btu/
hr)) to help identify critical operating parameters for precombustor systems
with N0X control and the potential for sulfur control.
The pilot-scale testing indicated that there was significant sulfur
captured as calcium sulfide (CaS) by suspended sorbent particles in the fuel-
rich combustion zone of the precombustor. However, there was evidence that
sulfur was released from the CaS when exposed to a fuel-lean flame front.
There was also concern that sulfur in the slag layer was evolving back into
the gas phase. A combination of high solids carryover, complexities of
sampling in the pi lot-scale precombustor, and a growing number of fundamental
questions concerning fuel-rich sulfur capture led to the second phase of this
program which is the main topic of this final report.
The objective of this second phase of work was to make a detailed
investigation of several key elements in the fuel-rich sulfur capture
process, including: 1) the formation of stable sulfides in the entrained
flow region of a precombustor using calcium-based sorbents, 2) the evolution
of sulfur from coal in an entrained flow process, and 3) the stability of
sulfur in molten slag layers.
1-3
-------
1.3 Experimental Systems
Three facilities were utilized in the program: 1) the Isothermal
Reactor (ITR) to investigate particular sulfidation reactions and to identify
key operating conditions in a wel1-controlled gas-flame environment, 2) the
Control Temperature Tower (CTT) for investigation of sulfur evolution from
coal and sulfidation in a coal combustion environment, and 3) the High
Temperature Oven (HTO), a new facility, for the study of sulfur and slag
interaction.
Two calcitic sorbents were tested in the entrained flow studies, Linwood
atmospheric hydrate (Ca(OH)2) and Vicron carbonate (CaC03). Methane and
ethylene doped with H2S were used in the ITR studies while natural gas and
Illinois coal were used on the CTT. The slag studies used mixtures of
Illinois coal ash and pure CaO and CaS.
1.4 Results
On the ITR facility the impact of fuel type, H2S doping location (to
change the reactor sulfur speciation), stoichiometry, temperature and sorbent
type were investigated at isothermal conditions. Sulfur capture was
determined using both gas phase and solids measurements. The gas phase
sulfur speciation was determined by gas chromatography over a variation in
residence time of about 0.25 to 1.0 seconds.
Figures 1-2 and 1-3 show how drastically the sulfur speciation is
effected by moving the dopant H2S downstream of the reactor flame and the
corresponding increase in calcium utilization when more H2S is available for
reaction.
When operating at temperatures more favorable for slag formation, in the
range of 1300°C (2372°F), calcium utilization dropped significantly compared
to capture results at 1100°C (2012°F), as shown in Figure 1-4.
1-4
-------
H2S Added to C2H4
H2S Injected Downstream
of C2H4 Flame
I
cn
100 -
ZJ
=3
OO
4->
o
+->
c
O)
0
S-
-------
H2S Added to C2H4
H2S Injected Downstream
of C2H4 Flame
100
80
60
40
20
1 I
1
1
0
*¦
0
.
1 r
u
1 1
1
*
O SR « 0.83
~ SR = 0.63
Open Symbols
- Gas Analysis
Closed Symbols
- Solid Analysis
0.25 0.50 0.75
Residence Time (s)
0.25 0.50 0.75 1.0
Residence Time (s)
Figure 1-3. Calcium utilization in the ITR using Linwood atmospheric hydrate, measured
by both gas and solid analysis. ITR held at 1100°C when doping the fuel
(C2H4) with 5000 ppm H2S or injecting it downstream of the flame
(Ca/S = 1.0 for each case).
-------
100 -
+¦>
c
-------
In general, these utilization results are much greater than fuel-lean
results at similar temperature and injection conditions using the same
sorbents (see Figure 1-5). Also, Figure 1-5 shows that the relative
reactivity of Linwood hydrate and Vicron carbonate is similar under fuel-rich
conditions as compared to fuel-lean conditions. This suggests that
information gained is studying sorbent properties and reactivity under fuel-
lean conditions can be applied to the fuel-rich capture problem.
Sorbent injection while firing Illinois coal on the CTT at
stoichiometric ratios of 0.6 and 0.8 produced drastically different sulfur
speciation and utilization levels as seen in Figures 1-6 and 1-7. Figure 1-7
shows how successful entrained flow sulfidation can be for controlling sulfur
under coal-firing conditions at controlled stoichiometry and temperature.
However, solid sampling revealed that these same conditions were not
favorable for entrained flow carbon burnout (see Figure 1-8).
Test results on the CTT again showed that Linwood was much more
effective than Vicron under fuel-rich conditions. Figure 1-9 shows that at
least 60 percent of the total input sulfur can be removed by the injection of
Linwood atmospheric hydrate at a Ca/S molar ratio of 1.5-2.0. This 60
percent value does not take into consideration the sulfur that remains in
char form. Capture would undoubtedly be higher if all the fuel-bound sulfur
evolved immediately into the gas phase.
Sulfur solubility in coal ash slags was tested for a variety of CaO/ash
mixtures in the HTO facility. The solubility is characterized by the sulfur
capacity, defined as:
Cs = (wt % S) (PO2/PS2)1/2
where: Cs = sulfur capacity
wt % S = percent sulfur in coal
PO2 = oxygen partial pressure
PS2 = Gaseous sulfur species partial pressure
Figure 1-10 shows results for a number of slag mixtures.
1-8
-------
Linwood Hydrate
Vicron 45-3
100 -
0.25 0.50 0.75
Residence Time (s)
0.25 0.50 0.75 1.0
Residence time (s)
O SR = 0.83
1100°C
A SR = 1.11
1000°C
Open Symbols:
- Gas Analysis
Closed Symbols:
- Solid Analysis
Figure 1-5. Calcium utilization in the ITR under fuel-rich and fuel-lean conditions
for both Linwood atmospheric hydrate and Vicron 45-3 limestone.
-------
SR = 0.8
SR - 0.6
0.5 1.0 1.5 2.0
Residence Time (s)
0.5 1.0 1.5
Residence Time (s)
~ H2S
A $02
O COS
q Char S
Figure 1-6. Sulfur speciation in the CTT firing II1inois coal at 17.6 kW (60,000 Btu/hr).
-------
100 -
CD
O
s-
Q)
Q_
c
o
<0
N
rd
e_>
~ SR = 0.6
A SR = 0.8
Open Symbols:
- Gas Analysis
Closed Symbols:
- Solids Analysis
0.25 0.5 0.75 1.0 1.25
Sorbent Residence Time (s)
Figure 1-7.
Calcium utilization in the CTT firing Illinois coal at
17.6 kW (60,000 Btu/hr), injecting Linwood atmospheric
hydrate at about Ca/S = 1.0.
-------
SR = 0.8
0.5 1.0 1.5 2.0
Residence Time (s)
SR = 0.6
0.5 1.0 1.5 2.0
Residence Time (s)
k Hydrogen
~ Nitrogen
O Carbon
A Sulfur
Figure 1-8. Species evolution from Illinois coal in the CTT at 17.6 kW (60,000 Btu/hr).
-------
100
I
CO
0)
s-
=3
4->
Q.
fd
o
u
3
<4-
0)
CJ
J-
-------
¦ Iron Silicate, molar basicity = 1.65, Fincham & Richardson (1954)
# Iron Silicate, molar basicity = 3.03, DeYoung (1984).
A Fe203/Coal Ash (Fe/S = 4.4), molar basicity = 3.01, EER.
~ Fe203/Coal Ash (Fe/S = 4.0), molar basicity = 2.75, EER.
Fe203/Coal Ash (Fe/S = 2.55 + 15.1% MgO), molar basicity = 3.08,
EER.
^ CaO/Coal Ash (Ca/S = 2), molar basicity = 2.77, EER.
O CaO/Coal Ash (Ca/S = 1), molar basicity = 1.48, EER.
0 CaO/Coal Ash (Ca/S = 0.5), molar basicity = 0.83, EER.
T(°C)
1600
1400
1200
00
o
CTi
O
1000
1/T x 104 (°K_1)
Figure 1-10. Influence of temperature and molar basicity on sulfur
capacity for several slag mixtures (solid symbols -
melted, open symbols - not melted).
1-14
-------
These slag equilibrium values can be used along with gas phase
equilibria and a sulfur mass balance to predict the best possible capture due
to sulfur solubility of the slag. Figures 1-11 and 1-12 show examples of
these best case equilibrium calculations. It is evident that calcium-based
slags are not very soluble to sulfur and are of limited value in a slagging
system due to the resulting rise in the mixture fusion temperature.
Iron-based slags are much more soluble to sulfur but only when
unacceptably large quantities of iron is added. Also, the kinetics of the
gas-slag sulfur capture process for iron-based slags have been shown to be
slow by other researchers.
What the results of Figures 1-11 and 1-12 suggest is that: 1) large
amounts of calcium in the slag can produce non-fluid slag mixtures at typical
precombustor temperatures, and 2) the CaS that forms in the entrained flow
portion of the reactor will tend to regenerate sulfur as the CaS dissolves
into the slag.
Figure 1-13 shows the impact of CaO addition to Illinois coal ash on
fluid temperature under both oxidizing and reducing conditions. For this
high sulfur coal, an addition of CaO in excess of Ca/S = 1 causes a sharp
increase in slag fluid temperature.
Fluxing additives such as P2O5, Fe203, CaF2 and B2O3 were added in small
quantities to CaO/ash slag mixtures corresponding to Ca/S = 1 and Ca/S = 2.
Figure 1-14 shows that B2O3 was the most effective fluxing agent under
reducing conditions.
To identify how quickly sulfur regenerates from slag when present at
levels in excess of equilibrium, CaO was replaced with pure CaS and the rate
of sulfur loss was measured. Figure 1-15 shows the relative rates of sulfur
regeneration for a number of slag mixtures. In general, the higher the slag
sulfur capacity and the higher the slag fluid temperature the slower the
sulfur regeneration.
1-15
-------
# Ca/S = 0.5, molar basicity = 0.83, melted
~ Ca/S = 1.0, molar basicity - 1.48, not melted
O Ca/S = 2.0, molar basicity = 2.77, not melted
Slag Mass, g/g Coal
Figure 1-11. Equilibrium sulfur removal by slag for SR=0.5,
T=1200°C» and an Illinois coal with 3.97% sulfur.
1-16
-------
^ Fe/S = 4.4, molar basicity = 2.75, melted.
A Fe/S = 4.0, molar basicity = 3.01, melted.
O Ca/S = 2.0, molar basicity = 2.77, not melted.
¦ Ca/S = 1.0, molar basicity = 1.48, melted.
Slag Mass (g/g coal)
gure 1-12. Equilibrium sulfur removal by slag for SR=0.5,
T=1400°C, and an Illinois coal with 3.975I5 sulfur.
1-17
-------
I
f—»
09
3000
2800
a>
u
Z3
+J
(O
i-
O)
a.
E
a)
2600
2400
2200
2000
Maximum
Measurable Temperature
0.5
_L
1600
1500
O Measured - Oxidizing
<2> Measured - Reducing
A T250» Sage & Mcllroy (1959)
A T250. Dolomite %
JL
C_J>
o
O)
1400 3
+->
-------
2800
2700 _ _ \ V _
Solid Symbols
Ca/S = 2
2600 -
Maximum
Measurable
Temperature
CO
c
o
•r—
TD
C
o
o
O)
c
•r-
U
3
"O
0)
cxi
^*
2100 -
Open Symbols
Ca/S = 1
\
\
\
\
\
V
ZQ CaF2
^ Fe203
2000 -
1900 -
A.
— B2O3
1800
1200
<0
S-
Q)
o.
e
a
¦o
•r-
3
1100
- 1000
10
Percent Additive (by wt.)
Figure 1-14.
Influence of fluxing additives on the reducing fluid
temperature of Illinois coal ash and CaO mixtures (open
symbols correspond to Ca/S = 1 mixtures while solid
symbols represent Ca/S = 2 mixtures).
1-19
-------
O Ca/s = 0.5
~ Ca/S = 1; CaS and Illinois coal ash well mixed
O Ca/S = 1; CaS layered on Illinois coal ash
O Pure CaS
A Ca/S = 2
Residence Time (Minutes)
Figure 1-15. Desulfurization of CaS and Illinois coal ash slag
mixtures, showing the influence of calcium content
and mixedness (open symbols represent completely
unmelted samples, solid symbols correspond to
completely melted samples).
1-20
-------
1.5
Conclusions
The scope of this program was focused on the area of fuel-rich sulfur
capture since little more than theoretical predictions and a few uncertain
test results were available. Several obstacles face the successful use of
calcitic sorbents in fuel-rich coal-fired precombustors. The most important
issues were investigated in this program, including: extent of entrained
flow sulfidation, the speciation of sulfur as it evolves from coal under
fuel-rich conditions, the equilibrium solubility of sulfur in molten slags,
the impact of fluxing additives on slag fluidity and sulfur solubility, and
the rate in which sulfur regenerates from slags containing a super-
equilibrium level of sulfur.
The major conclusions from this program are summarized below:
• The sulfidation reactions between CaO and H2$ or COS are fast and
under optimum conditions can capture a majority of the gas phase
sulfur in a fuel-rich precombustor.
• The conditions which favor fuel-rich sulfur capture (deep
substoichiometric operation and moderate temperatures) can result
in poor carbon burnout and low slag fluidity.
• Typical molten coal ash and mixtures of coal ash and CaO are
incapable of holding large amounts of sulfur in a coal precombustor
environment when at equilibrium.
• Coal ash/CaO slags quickly desulfurize from super-equilibrium
levels of sulfur at typical precombustor temperatures and gas
compositions, indicating that rapid slag drainage designs are
required.
• Slag fluxing additives, such as B2O3, can extend operating
conditions which will make fuel rich sulfur removal a possibility
over a wider range of coal types and precombustor systems.
1-21
-------
Future work at large scale is now required to determine the practical
limitations of sulfur capture in systems that have been tested and optimized
for maximum ash rejection, continuous slag drainage, and minimal N0X
emissions. The results of this program provide a major portion of the
necessary background information before proceeding in the final development
of a precombustor system that can simultaneously control ash, N0X and S0X
emissions.
1-22
-------
2.0 INTRODUCTION
2.1 Precombustor Concept
Conversion of gas- and oil-fired combustion systems to coal firing is
made difficult by the presence of mineral matter, nitrogen and sulfur in
coal. Not only are fly ash emissions a health hazard, but the presence of
mineral matter can impair operation of equipment that has not been designed
for coal use. Deposition on heat transfer surfaces can cause reduced
efficiency and increased corrosion and erosion. Nitrogen and sulfur in coal
can be released into the environment as N0X and S0X, which are both
precursors of acid rain. Cost of precleaning coal or post-combustion clean-
up are prohibitively expensive in most applications at this time.
One concept which offers the potential for simultaneous removal of ash
and reduction of N0X and S0X emissions at the burner prior to a boiler or
heater furnace is coal-fired precombustion. A precombustor burns coal in a
chamber exterior to the normal furnace region (see Figure 2-1). Mineral
matter in the coal (that does not vaporize) can be separated from the exhaust
gases aerodynamically and drained as molten slag from the precombustor. Ash
separation over 90 percent has been reported from existing systems.
Operation under staged combustion conditions (i.e., fuel-rich in the
precombustor with fuel-lean burnout in the retrofitted furnace) has been
shown to reduce N0X emissions by over 60 percent. Research efforts at Energy
and Environmental Research Corporation (EER) (La Fond and Payne, 1985) and
elsewhere (Dykema, 1985; LeCrew and White, 1985; Mattsson and Stankevics,
1985; and Zauderer and Fujimura, 1985) have been directed at developing
precombustor systems which optimize ash separation and N0X reduction.
This program was directed at the key issue that has not been adequately
addressed by previous research—whether sulfur can be controlled internally
in precombustor systems.
2-1
-------
Figure 2-1. Common cyclonic precombustor in staged combustion
configuration attached to boiler furnace wall.
2-2
-------
2.2 Sulfur Control
In recent years, much attention has been given to the use of calcium-
based sorbents (calcitic and dolomitic carbonates and hydrates) for the
removal of gas phase SO2 by reaction with CaO to form CaS03 or CaS04.
Thermodynamically, sulfur is more stable as CaS under fuel-rich conditions
than CaS04 in a fuel-lean environment. Figure 2-2 shows the effect of
stoichiometry and temperature on equilibrium solid calcium-sulfur compounds
for a Pittsburgh coal. Under fuel-rich conditions CaS formation is favored
and is thermodynamically stable even at fairly high combustion temperatures.
Under fuel-lean conditions CaS04 is the favored species; however, above
approximately 1300°C (2372°F), CaSO^ is not thermodynamically stable. Figure
2-3 presents data by Silcox et al. ( 1986) that shows the sharp drop in
calcium utilization as the reactor temperature was raised from 1150°C
(2102°F) to 1300°C (2372°F). Figure 2-2 also shows that the stability of
either solid sulfur species is low at conditions slightly below
stoichiometric.
The reactions that take place to form CaS under substoichiometric
conditions include:
CaO + H2S CaS + H20.
CaO + COS - CaS + CO2.
Borgwardt et al. (1984) have shown that the kinetics of the H2S and COS
reactions are faster than the fuel-lean sulfur reaction of:
CaO + SO2 + 1/2 O2 CaSO^.
Borgwardt's studies were performed in a laboratory scale system under ideal
conditions and low temperatures. The concept has been extended to testing in
pilot scale fuel-rich precombustor systems by various researchers with either
unfavorable or unsubstantiated results.
2-3
-------
Pittsburgh HVA
Ca/S = 1
2% Sulfur
Figure 2-2. Equilibrium stability of solid calcium-sulfur
compounds (England et al., 1986).
-------
Ca/S = 1
Linwood Atmospheric
Hydrate
Residence Time:
~ 0.55 sec
O 0.35 sec
A 0.17 sec
'2000
2400
-vh
1100
2200
Temperature (°F)
1200 1300
Temperature (°C)
1400
Figure 2-3. Calcium sulfate formation at elevated injection
temperatures (Silcox et al., 1986).
2-5
-------
The process of calcium-based sorbent injection into fuel-rich
precombustors is a complicated sequence of coal pyrolysis and oxidation,
sulfur species evolution, sorbent calcination, sulfidation and slag
formation. Figure 2-4 shows schematically some of the important physical
transformations that may be occurring within a fuel-rich precombustor system.
Upon injection, the solids are in an entrained flow phase. Coal enters
the precombustion chamber and begins to pyrolyse, giving off volatiles,
sulfur and nitrogen compounds and forming char. The coal in various stages
of pyrolysis may reach a wall where it can continue to devolatilize or
possibly become encapsulated in a liquid slag layer. At the same time, if
sorbent material such as limestone (CaC03) or calcium hydroxide (Ca(0H)2) is
injected into the chamber for the purpose of sulfur removal, CO2 or H2O is
driven off the sorbent to form CaO where it can react with sulfur compounds
in the gas phase. The calcium-sulfur solids that are formed (CaS under
strongly fuel-rich conditions) can reach the slag covered precombustor walls
and be flushed out of the system through the slag drain or regenerate sulfur
back into the gas phase. The calcined sorbent can also reach the slag layer
without completely reacting where it can continue to react with gas phase
sulfur, combine with liquid phase sulfur, or become deactivated by
interaction with mineral matter.
Successful application of sulfur control in fuel-rich precombustors will
depend on better understanding of the following key issues:
• Required conditions for sulfidation of calcium-based sorbents;
0 Compatability of sulfidation conditions to practical requirements
of precombustors, such as carbon burnout and slag fluidity;
• Means of removal of sulfur bearing solids, including sulfur
solubility and stability in coal ash slag; and
• Impact of additives on altering or extending operating conditions
to reach all performance goals simultaneously.
2-6
-------
CO2 or H2O
Sulfur
Volatiles
CO
CO2
SO2
H2S
COS
NO
N2
Exhaust
to
Furnace
¦¦vm
Slag
Drainage
Figure 2-4. Schematic diagram of some of the important mechanisms occuring
in a fuel-rich precombustor system with sorbent addition.
-------
2.3 Program History
This program was organized in two phases. Phase I included a review of
available literature on fuel-rich sulfur capture and cyclonic coal combustors
and design and testing of a 293 kW (1 x 10® Btu/hr) coal-fired precombustor.
The precombustor was designed based on previous experience with low-N0x
systems and the literature review. It was tested on EER's fire tube package
boiler simulator facility to establish conditions which favored simultaneous
control of N0X, S0X, and particulate matter and to identify areas that
required further investigation.
The Phase I work was carried out under the direction of EPA's LIMB
Branch, with Chuck Masser serving as Project Officer. A summary of the
results of the Phase I work can be found in the background portion of this
report (Section 3.0).
Although the Phase I testing showed some evidence of fuel-rich sulfur
capture and provided valuable design and operational experience, the scale of
the test apparatus along with sampling, analysis and operational difficulties
limited the amount of useful information that could be generated.
The Phase I testing led to the formulation of a second phase test plan,
which is the prime topic of this final report. The Phase II testing was
structured to investigate a broader range of issues in smaller, well-
controlled experimental facilities in order to generate a more complete set
of basic precombustor design data. The Phase II efforts were directed by
EPA's Combustion Research Branch, with Joe McSorley serving as Project
Officer.
2-8
-------
3.0 BACKGROUND
There were three main sources of information that provided background
information before commencement of Phase II of this program: 1) testing
experience on the Phase I pilot-scale precombustor system and EER's Vortex
Containment Combustor (VCC); 2) review of information available from the
steel and glass industries on sulfur solubility in slags and slag fluxing
additives; and 3) literature review of fundamental research on the subject of
calcium sulfidation.
3.1 Precombustor Experience
As mentioned earlier, Phase I of this program involved construction and
testing of a precombustor system. The testing was exploratory in nature, but
provided enough information to structure more detailed measurements in Phase
II. The Phase I tests focused on screening the effects of operational
parameters on sulfur capture by sorbent injection under fuel-rich conditions.
At the same time, the reduction of N0X and particulate emissions was
monitored to establish operating conditions for simultaneous control of all
three pollutants. Parameters varied were the first stage stoichiometric
ratio (SRi), type of sorbent, type of fuel, and method of primary air
injection (axial vs. tangential). Sorbent enhancement by addition of a small
amount of Fe203 was also tested. The furnace configuration was modified to
add slag baffles at the end of the precombustor for improved solids
rejection.
The concept of sulfur capture in fuel-rich precombustors was first
suggested by work done at EER by Case et al. (1985) that indicated that
improved sulfur capture may occur in the fuel-rich environments found under
low-N0x combustion conditions. There were also three theoretical factors
that suggested improved sulfur capture under fuel-rich conditions compared to
fuel-lean conditions. First, lower fuel-rich flame temperatures could reduce
sintering of the sorbent surface, thus increasing sorbent reactivity for
sulfidation reactions. Second, by forming calcium sulfide (CaS) under
reducing conditions rather than calcium sulfate (CaS04), the internal pore
3-1
-------
blockage which hinders sulfation reactions is reduced, allowing free
diffusion of H2S and COS into the heart of the CaO particles. And third, CaS
is thermodynami cal ly more stable under fuel-rich conditions at combustion
temperatures than CaS04 under fuel-lean conditions.
In practice, potential difficulties in achieving high sulfur capture
under fuel-rich conditions include: 1) the oxidation of CaS to form CaO and
SO2 if exposed to a hot fuel-lean environment, 2) the ability to reject ash
in either a dry or molten form while simultaneously obtaining both high
carbon burnout and high sorbent reactivity in a precombustor system, and 3)
regeneration of sulfur from molten ash/CaS mixtures.
A sketch of the precombustor system tested in Phase I is presented in
Figure 3-1. The precombustor was designed to fire with either axial or
tangential primary air injection and could accommodate baffles at the
precombustor exit to increase slag rejection.
Initial characterization of the system included temperature profiling
within the fuel-rich and fuel-lean zones of the system. Figure 3-2 presents
the profile measured at a stoichiometric ratio of 0.5. Estimated
temperatures at varying first stage stoichiometry (based on adiabatic flame
temperatures and calculated heat loss) were used to calculate equilibrium
sulfur speciation with and without sorbent addition. The results of these
equilibrium computations are presented in Figure 3-3. These plots suggest
that conditions which favor H2S formation are favorable for calcium
sulfidation. Conversion to CaS is favored at stoichiometric ratios of 0.6 or
less and improves at lower temperatures.
N0X emissions were measured as a function of first stage stoichiometric
ratio (SRi). Measurements were made for SRi between 0.5 and 0.9, holding the
overall stoichiometric ratio (SRj) constant at 1.2. Emissions from both
natural gas- and Illinois coal-firing were measured and are presented in
Figure 3-4. The natural gas N0X levels were below 100 ppm for every case.
The N0X emissions for the coal-fired runs were below 115 ppm for SRi between
0.6 and 0.8. Fine tuning of the precombustor and second stage air injection
3-2
-------
Natural Gas
(+H2S .when
used)
Prtaury
Transport
Mr
fIretube
Furnace
Simulator
6EHERAL AftftAHiiEHtWT
CO
OJ
Snap I*
t
Injection
Ports
Hultt-Liyar
CiStlblt
Mfractory
Tangential
Injection
Section B->
CROSS SECTION REFRACTORY DETAIL
REFRACTORY LEGEND
Greencast 94+
LW-26
Q
Figure 3-1. Precombustor arrangement and refractory detail.
-------
2500 _
2000
1500 -
1000 -
500
0 1
3 4 5 6
Residence Time (Seconds)
Figure 3-2. Mean gas temperatures and bulk residence times in the
precombustor and furnace (SR1 = 0.5, 0.3 MW^).
-------
c*>
tn
u
CD
CL
OO
c
o
~r~
4->
O
(d
S-
U
rd
s-
O)
o
0.0
SRi 0.5
T (°F) 2300
T (°C) 1260
(a) Illinois Coal Without CaC03 Input
(b) Illinois Coal With CaC03 Input (Ca/S=2)
Figure 3-3, Equilibrium distribution of sulfur species under the reducing conditions
of the precombustor for Illinois coal with and without calcium present.
-------
0 0.5 0.6 0.7 0.8 0.9
Precombustor Stoichiometric Ratio (SR-|)
Figure 3-4. Stack N0X emissions as a function of SRi for
Illinois coal and natural gas (SRj = 1.2).
3-6
-------
could result in lower N0X emissions over a broader range of first stage
stoichiometric ratios.
The exposure of pulverized sorbent material to sulfur species under
staged combustion conditions resulted in higher overall sulfur capture than
when exposed solely to fuel-lean conditions. Figure 3-5 shows that injection
of H2S into the fuel-rich first stage with either Vicron 45-3 or dolomite
gave higher overall exhaust SO2 reductions than when SO2 was injected into
the fuel-lean second stage. The sorbent was injected along with the primary
air for both cases to assure the same thermal environment and thus, similar
sorbent reactivity. Although these tests showed that fuel-rich sulfur
capture could potentially improve sorbent-based SO2 reduction in a staged
combustion system, they did not address such problems as loss or
"regeneration" of sulfur species from calcium-sulfur compounds in the fuel-
lean stage or differentiate between fuel-rich sulfidation and fuel-lean
sulfation reactions.
Solids sampling from the exit of the fuel-rich precombustor provided
important information that was masked by the previous measurements. A
calcium utilization (sulfur capture at Ca/S = 1) of greater than 30 percent
was measured for both natural gas/H2S and Illinois coal flames using Vicron
45-3 as the sorbent. Assuming a linear increase in capture as the Ca/S ratio
increases, this would correspond to sulfur capture levels of more than 60
percent at Ca/S = 2. However, gas phase SO2 measurements made in the system
exhaust indicated sulfur capture of only 20 and 25 percent under natural gas/
H2S and coal firing, respectively, indicating that regeneration of sulfur
species was occurring in the fuel-lean second stage. Results from the
natural gas testing are shown in Figure 3-6. Analysis of the solids also
indicated that under extremely rich first stage conditions (stoichiometric
ratio of 0.5) the sulfur was captured as calcium sulfide (CaS), as expected.
The use of slag baffles to help retain solids in the precombustor
resulted in rejection of 28 percent of the solids. As Figure 3-7 shows, an
increase in sulfur capture was recorded when the slag baffles were in use
indicating that some captured sulfur from the first stage (in the form of
3-7
-------
~ H2S premixed w/natural gas
S02 WITH SECONDARY AIR
SR
1
SORBENT
0.5
DOLOMITE
0.8
DOLOMITE
Figure 3-5. Sulfur capture when firing natural gas (0.3 MWt) and
injecting Vicron 45-3 and dolomite with the primary
air at Ca/S = 2 (SRy = 1.2).
3-8
-------
ion
80
Q)
S-
3
+J
CL
rO
O
S-
3
3
CO
~
I
-* Solids Measurements,
.4^1 Precombustor Exit
Gas Phase Stack
Measurements
2 3
Ca/S Molar Ratio
Figure 3-6. SO? capture based on gas phase stack measurements
ana potential SO2 capture based on calcium utiliza-
tion of solids samples. Vicron 45-3 in a fuel-rich
natural gas/h^S flame (SR^ = 0.5, SRy = 1.2, 0.3 MW^).
3-9
-------
With Baffles
Without
Baffles
1 2 3
Ca/S Molar Ratio
(a) Natural Gas + H^S
O)
S-
Z5
4->
CL
fO
c_>
s-
ZJ
zs
00
1 2 3
Ca/S Molar Ratio
(b) Illinois Coal
Figure 3-7. Effect of increased solids removal on sulfur capture for two fuels.
Vicron 45-3 injected with primary air (SR^ = 0.5, 0.3 MW^).
-------
CaS) was indeed being lost to regeneration in the fuel-lean second stage.
Also, the incremental increase in sulfur capture corresponded well to the
expected capture level based on solids measurements of calcium utilization.
The solids analyses for the coal-fired tests also indicated that a
substantial portion of the exiting particles remained as char (carbon burnout
of 61 to 69 percent at SRi = 0.5).
The addition of 5 percent by weight of iron oxide (Fe203) to the sorbent
resulted in an improvement in overall sulfur capture from 24 percent to 34
percent at Ca/S = 2 for the natural gas/H2S case (see Figure 3-8). This
testing showed the potential of mineral matter additives for the enhancement
of fuel-rich sulfur capture but was incapable of providing insight into the
mechanisms of the sulfur capture process.
Other precombustor experience that impacted the formation of the Phase
II work plan and interpretation of the results included testing of EER's
Vortex Containment Combustor (VCC) (La Fond et al., 1985 and 1986). The VCC
is an advanced cyclonic combustor which features:
• Reverse vortex flow for efficient particle separation.
• Long fuel residence times in an aerodynamic suspension.
• Selective ash rejection from the combustion zone.
• High radial velocity component of the inlet air to minimize wall
deposition of ash and char.
As part of a Department of Energy program (Contract number: DE-
AC22-85PC80256) the 0.9 MWt (3 x 10& Btu/hr) precombustor shown in Figure 3-9
was tested under fuel-rich conditions for ash retention efficiency, N0X
reduction by staging and reburning, system pressure drop, slag drainage,
carbon utilization, and SO2 reduction.
Figure 3-10 shows that over 80 percent of the coal ash could be
separated from the exhaust stream over a wide range of firing rates and at
low system pressure drops. N0X was effectively controlled by over 60 percent
3-11
-------
Ca/S Molar Ratio
(a) SO^ with secondary air
Q)
U
0
U
o-
cO
u
u
3
rH
3
CO
2 3
Ca/S Molar Ratio
(b) with Natural Gas
Figure 3-8. Effect of sorbent additive (5% Fe203 by weight) on sulfur capture, sorbent
injected with primary air (natural gas, SR1 = 0.5, 0.3 MWt).
-------
BtrL (12
Air
Plenum
KAOTAB 95
1900°F Insul-Board
KAOLITE 2800
Figure 3-9. Cross-section of the VCC prototype combustor (0.9 MW^).
3-13
-------
Firing Rate (10 Btu/hr)
Figure 3-10. Recent ash retention efficiency measurements and system pressure drop
firing rate (La Fond et al., 1986) in the VCC.
-------
for staged operation and by over 85 percent when natural gas reburning was
applied (see Figure 3-11).
A limited number of exploratory tests were performed to investigate SO2
reduction by fuel-rich sorbent injection. Figure 3-12 shows the results of
tests when Vicron 45-3 was injected into the VCC through auxiliary fuel
injectors at varied stoichiometric ratio and Ca/S molar ratio. The results
were discouraging, however, upon re-evaluation with the information gained
from Phase II of this program these VCC results could be expected. The
choice of operating stoichiometry, sorbent injection location and even
sorbent type were not optimal.
The VCC testing experience provided a great deal of insight into the
operational difficulties that would be faced in trying to apply sulfur
control technology to high firing density precombustor systems.
3.2 Review of Slag Related Industry and Research
Both the steel and glass industries have faced slag related problems
over their histories. In the making of steel, slag additives have been used
to help remove sulfur impurities from iron ore and steel melts. Additives
have been used in the glass industry to lower slag melting times for
production of glass. Review of this information was useful in preparing the
program test plan and in utilizing the generated data.
The desulfurization process is a crucial step in the making of iron and
steel. Sulfur is considered a contaminant that reduces the ductility and
impact strength of iron and steel, and many years of research have been
directed at its removal. In the production of iron, early desul furization
efforts concentrated on layering limestone into blast furnaces along with
coke and iron ore (see Figure 3-13).
Wilson and McLean (1980) have pointed out that current practice
concentrates on sulfur treating the pig iron after removal from the blast
furnace. A schematic example of an iron desulfurization station is presented
3-15
-------
CO
I
CTl
1.2
1.0
0.8
0.6
0.4
0.2
1
1 1 1 T 1
^ Unstaged at SR = 1.2
Baseline
.
\
/
Staged
1
I 1
i i I
0.5 0.6
_L
0.7 0.8 0.9 1.0
VCC Stoichiometric Ratio
_L
_L
i
1.1 1.2
_L
2.75 2.38 2.05 1.83 1.65 1.5
Firing Rate (10® Btu/hr)
(a) Combustion Staging
1.2 -
1.0 _
0.8 -
0.6 -
0.4 -
0.2 "
0.5
"I 1 1 1 1 r-
.Coal Firing at SR = 1.2
Baseline
_L
Coal with Natural
Gas Reburning
_L
0.6 0.7 0.8 0.9 1.0 1.1
Reburning Zone Stoichiometric Ratio
1.2
_L
_L
_L
_L
_L
0.45 0.36 0.27 0.18 0.09 0.0
~ , „ .. /Btu Reburni
Reburning Fuel Ratio ^ Btu tota
nq Fuel\
fat /
(b) Natural Gas Reburning
Figure 3-11. N0X control by combustion staging and natural gas reburning in the VCC;
Illinois coal, SRj =1.2 (La Fond et al., 1985).
-------
Fuel: 111inois Coal
Nominal Firing Rate: 2x106 Btu/hr
SRy = 1.2
Ca/S = 0.50
JL
_L
X
15
£
O
O
=3
TD
-------
Top Hopper
Hopper
Bell
Coke-
Ore
400°F (200°C)
Limestone'
430°F (220°C)'
900°F (480°C)
Hot Blast'
Cast House
3500°F (1930°C)-
il«III«I ««l««l III IIIiMllllf
w,v;.v.v,v/Avmv±'
Pro-
cess
of
Reduc-
tion
Process
of Heat
Absorption
Process
of Fusion
, Hot Blast
} Pro-
cess of Combustion
Molten
Slag
Tuyere
o
Figure 3-13. Cross-sectional view of blast furnace.
-------
in Figure 3-14. Additives for the removal of sulfur include: magnesium
(pure powder, salt coated granules, mixtures with coke, mixtures with lime),
rare earth metals (Cerium or Lanthanum), CaC2» ^003, CaO, and CaF2- By
processing the iron separate from the blast furnace, more economical
operation of the blast furnace and improved desulfurization has been
achieved. Several of these additives do not remove significant amounts of
sulfur upon initial injection, but float to the top of the molten iron bath
to form basic slags. Sulfur is then absorbed into the slag by ionic transfer
between the two liquid materials.
The desulfurization process during the making of steel follows very
similar patterns to the iron desulfurization process. Once the steel has
been decarburized the same kinds of additives are used as were previously
mentioned for iron making along with a few others: nickel-4.5 percent
magnesium, CaSi, CaO + CaF2 + A1 mixtures. A critical step in the steel
making desulfurization process is the reduction of the oxygen content to as
low a level as possible. Si and A1 additives help scavenge oxygen, making
slag sulfide formation as CaS or MgS more favorable (Na2C03 is seldom used
today due to its corrosive attack on the furnace refractory linings).
Similar desulfurization methods can be applied to open hearth furnaces,
ladles, or induction ladle furnaces.
Desulfurization techniques were investigated more scientifically in the
1950's as the demand for higher quality steels increased. Of primary
importance to these efforts was the establishment of sulfur capacity
equilibrium for slag mixtures containing CaO (with some investigation of MgO
based slags). Fincham and Richardson (1954), St. Pierre and Chipman (1956),
and Carter and Macfarlane (1957) are examples of this early work. Their
objective was to find compositions of Ca0-Si02-Al2O3 based slags that had
liquidus temperatures within the range required for steel making (1500 to
1650°C) and establish the sulfur capacity of those slags. Figure 3-15
presents some of the results of Fincham and Richardson (1954) establishing
the solubility of sulfur in Ca0-Si02-Al2O3 slag mixtures at 1650°C. These
findings have been shown to hold true in practical steel desulfurization
applications in studies by Sussman et al. (1982), as can be seen in
3-19
-------
Figure 3-14. Iron desulfurization process schematic.
3-20
-------
Si02
mol. %
Values refer to C$ x 10^
where: Cg = (wt % S)
Pa = Partial pressure oxygen
p = Partial pressure gas-
s2 eous sulfur species
Figure 3-15. Sulfur capacities, at 1650°C, in CaO-SiC^-AI^O., melts.
Crosses are experimental points and broken lines are
liquid composition limits at this temperature.
3-21
-------
Figure 3-16 and Herrera et al. (1983). Unfortunately, ternary phase diagrams
(from Levin et al., 1979) show that the liquidus temperature of the most
reactive slags are high compared to conditions in coal combustion systems
(see Figure 3-17).
Recently, DeYoung (1984) has completed an investigation of lower melting
slag mixtures, including investigation of Fe0-Al203-Si02 based slags. As
Figures 3-18 and 3-19 indicate, slag mixtures with liquidus temperatures
below 1200°C can be generated. Measurement of sulfur capacities in slags
produced by mixtures of West Virginia coal ash and additives were made
showing much greater equilibrium levels of sulfur in FeO based slags than CaO
or Na20. Calculation of possible precombustor sulfur capture based on
equilibrium gas phase concentrations were made for a range of sulfur
capacities. The result of these equilibrium calculations along with a sulfur
mass balance gave potential sulfur removal values as a function of the
percentage of slag being injected into a precombustor system. These
calculations showed that high sulfur capture in the slag was possible with
FeO-Al203-SiO2 based slags. However, no investigation of the kinetics of
sulfur evolution and sulfur mass transfer into the slag was performed and
extremely high levels of FeO were required.
Glass is defined by Doremus (1973) as an amorphous solid; it has no
order in its structure, no crystallization. Glass is typically formed by
rapidly cooling a material from the normal liquid state. The three main
categories of glass are:
• Oxide • Chalcogenide • Ionic
Of interest to this program is the role that B2O3 and P2O5 play in glass
making and their influence on silicate formation when mixed with coal ash.
In general, B2O3 and P2O5 have much lower viscosities than SiO2 the same
Si02
b2°3
Ge02
P2O5
Sulfur
Seleni um
Tellurium
Hal i de
Nitrate
Sulfate
Carbonate
3-22
-------
SiO,
0.1 0.2 0.3 0.4 0.5 A1203
-CaO-A^Og-S-^-System at 1650°C.
-—Sulfide Capacities x 10^ at 1650°C.
Good Desulfurization
Figure 3-16. Desired slag compositions in the Armco top
slag desulfurization process (Sussman et al., 1982).
3-23
-------
Si02
1723*
Crystalline Phases
Notation Oxide Formula
Cristobolite I
Tridymite ]
Pseudowollostonite
Ronkinite
Lime
Corundum
Mullite
Anorthite
Gehlenite
Si02
CqO' S1O2
3CaO 2SiOz
CoO
AlgOs
3AI2O3 2Si02
CaO-AyJs-ZSiOz
2CoO AI2O3 SiOg
Tomporaftrts up to appra»nm*ly I550*C
v« on the Gtophyttcal Laboratory
Scolt; fflot* abovt 1350*C or* on the
1948 Intemehonoi Scolt.
CoO - Si 02
1544'
Ronkinite
i<
14
3CaO-2SiOz
1460*,
1464V
2CaO- SiOg
~2I30*
~ 2050
~ 2070'
3CaO Si Oa
3AI2O3 2Si02
'1850*
1535*
3Ca0 AljOj
CQijA^Oj
I2C007ALP,
1455*
CoO AUO,
~ 1605*
CQO-2AIP3 CoOSAIjOj
~1750*
Figure 3-17. Ternary diagram for CaO-A^Og-SiC^ melts; temperature
in °C. Reprinted by permission of the American Ceramic
Society (Levin, et al. 1964).
3-24
-------
Fe0-Al203-Si02
Two
Liquids
Crystalline Phases
Notation
Cristobolite ]
Tridymite ]
Fayalite
Wiistite
Hercynite
Corundum
Mullite
Iron Cordierite
Oxide Formula
SiOz
2Fe0Si02
"FeO"
FeO ¦ AI2O3
AI20j
3AI20y2Si02
2FeO • 2AIj03- 5Si02
Temperatures up to approximately I550°C
are on the Geophysical Laboratory
Scale; those above I550t are on the
1948 International Scale.
2FeO- Si02
1205°
, 2SiO,
1850°
FeO • AI2O3
~I780»
ai2o3
j 2020°
Figure 3-18. Ternary diagram for FeO-A^Og-SiC^ melts; scales in weight per-
cent and temperature in °C. Reprinted by permission of ..the
American Ceramic Society (Levin et al. 1964).
3-25
-------
Na20-Fe0 SiC>2 Si02
[1713 *5°)
(l3B0i5°)
Figure 3-19. Ternary diagram for FeO-^O-SiC^ melts; scales
in weight percent and temperature in °C.
Reprinted by permission of the American Ceramic
Society (Levin, et al. 1969).
3-26
-------
temperature. The addition of either of these oxides to silica lowers the
mixture viscosity, with B2O3 being particularly effective.
The available literature suggests that the replacement of SiO2 wit*1 B2O3
and P2O5 in slag mixtures can lower the fluid viscosity. According to Pye et
al. (1972): "Small amounts of B2O3 or fluorides are also claimed to be
helpful (in the glass melting process), probably by producing lower-melting
phases initially or by decreasing viscosity over the entire temperature range
of the melting process." Doyle ( 1979) has reported the time required for
batches of Na20-Ca0-B203-Si02 glass to melt at 1427°C (which is related to
the batch melting temperature) as B2O3 is used to replace other components.
As can be seen in Table 3-1, replacement of SiO2 with B2O3 causes an initial
lowering in the batch melting time.
The replacement of SiO2 with B2O3 and P2O5 can also reduce the formation
of silicates. It is the formation of silicates that appears to be the most
damaging factor in reducing the sulfur capacity of a slag mixture. However,
if B2O3 or P2O5 is used as an additive, the addition of more glass making
material may cause additional reductions in sulfur capacity due to borate or
phosphate formation. Data by Abraham and Richardson (1960) shows that the
addition of 7 percent P2O5 to both Ca0-Al203 and Ca0-Si02 mixtures caused a
lowering of the sulfur capacity and that mixtures of Ca0-P20s had lower
sulfur capacities than either the Ca0-Al203 or Ca0-Si02 systems of the same
CaO fraction (see Figure 3-20).
The only available data on sulfur capacity when adding B2O3 was
performed by DeYoung (1984). DeYoung found that by replacing 5 percent of
the Si02 with B2O3 in a FeO-Al 2O3-SiO2 melt the sulfur capacity was
unchanged. A similar experiment using P2O5 resulted in a reduction in the
sulfur capacity of approximately 20 percent.
From the available literature it appeared that there may be some
benefits to the use of B2O3 and P2O5 as slag additives in a coal-fired
precombustor system.
3-27
-------
TABLE 3-1. EFFECT OF B203 SUBSTITUTIONS ON MELTING
TIME FOR THE Na20-Ca0-B203-Si02 SYSTEM
(Information from Doyle, 1979).
Composition, %
Time (min) to obtain
batch-free glass at
B203
Na20
CaO
Si02
1,427°C
10
14
12
74
13
0.25
13.75
12
74
11
0.50
13.50
12
74
12
B2O3 for Na20
1
13
12
74
15
3
11
12
74
36
5
9
12
74
83
0
14
12
74
13
0.5
14
11.5
74
11.5
B203 for CaO
1
14
11
74
12.5
3
14
9
74
20
5
14
7
74
31
(0
14
12
74
13
0.5
14
12
73.5
10.5
B203 for Si02
1
14
12
73
10.5
3
14
12
71
10.5
(5
14
12
69
10.5
(0
14
12
74
13
0.5
13.75
12
73.75
12
B203 for Na20 + Si02
1
13.50
12
73.50
14
3
12.50
12
73.50
24
\5
11.50
12
71.5
43
3-28
-------
Hole Fraction Base x(Ca0>Hg0 or MnQ)
Figure 3-20. Sulfur capacities C§, in various slag
systems at 1650°C (information from
Abraham and Richardson, 1960).
3-29
-------
The review of slag related literature did not provide examples of slag-
sulfur systems that were exactly comparable to the environment expected
within coal-fired precombustor systems. However, the measurement and
computational procedures developed by the metallurgical industry to predict
expected equilibrium sulfur capture levels in slags proved to be very
valuable in the development of precombustor slag testing performed in this
program. Also, the glass industries use of B2O3 and P2O5 gave insight into
possible slag fluxing agents for use in precombustor systems.
3.3 Calcium Sulfidation
The kinetics of calcium sulfide formation due to reactions between
calcined limestone particles (CaO) and H2S and COS has been studied
extensively at the laboratory scale by Borgwardt et al. (1984a). Borgwardt
also provides a good review and critique of previous kinetic studies of CaS
formation. This work was performed in an isothermal differential reactor
fabricated of quartz glass. The experiments were performed at low
temperature with pure gases and CaO of known specific surface area.
By extrapolation of Borgwardt's kinetic data it was predicted that CaO
with a specific surface area of 40 m^/g reacting with 5000 ppm H2S would have
40 percent conversion to CaS in only 0.04 seconds at 1250°C (2282°F). The
addition of hydrogen (H2) or chlorine (as CaCl2) reduced conversion by 10 to
20 percent (relative) which for a Ca/S molar ratio of 2 would still result in
about 65 to 70 percent sulfur capture under these conditions; that is,
assuming capture increases linearly with Ca/S ratio.
It would be expected that sorbents that are most reactive under fuel-
lean sulfation conditions are most reactive under fuel-rich sulfidation since
reactivity has been shown by Borgwardt et al. (1984a and 1984b) to be
dominated by specific surface area for both the sulfation and sulfidation
reactions. A comparison of SO2 capture by sulfate formation for two calcium
carbonates and two calcium hydroxides is presented in Figure 3-21 (from
Overmoe et al., 1986). Thus, the use of the most reactive sorbent, Linwood
calcium hydroxide, would provide the clearest measure of sulfidation and
3-30
-------
Tinj S 1260°C(2300°F)
A Linwood Ca(0H)2
O Colton CaCOH)^
A Linwood CaCO^
# Vicron CaCO-
1 2 3
Ca/S Molar Ratio
Figure 3-21. Comparison of sulfur capture capability for four
different sorbent materials under fuel-lean
conditions (Overmoe et al., 1986).
3-31
-------
would most easily identify optimal operating conditions during parametric
testing.
There was evidence during the Phase I testing that the sulfur captured
in the fuel-rich precombustor was all in the form of CaS. Also, the data
suggested that a majority of the CaS oxidized to CaO and SO2 in the fuel-lean
burnout region of the precombustor system. This problem of fuel-lean sulfur
generation was investigated briefly by Cole et al. (1985a) in an isothermal
drop-tube reactor. As Figure 3-22 shows, at typical combustion temperatures
under fuel-lean conditions, CaS decomposes within 150 ms. Note, however,
that the degree of regeneration reaches an asymptotic value of 80 percent.
This may explain some of the Phase I results. The improvement in SO2 capture
when sulfur was introduced with the sorbent into the fuel-rich precombustor
compared to when the sulfur was added with the secondary air can be accounted
for by the carryover of CaS that did not decompose by exposure to the
secondary flame.
3-32
-------
100
OJ
Co
-p
c
CD
O
s-
O)
Q_
c
o
4->
fd
5-
-------
4.0 OBJECTIVES AND APPROACH
The objective of this program was to reach a better understanding of
fuel-rich sulfur control mechanisms and their relationship to other key
precombustor operational processes such as carbon burnout, particulate
emissions control, N0X emissions control, and continuous slag drainage. This
was done by carrying out a detailed investigation of several key process
elements in the fuel-rich capture process. These elements included: 1) the
formation of stable sulfides in the entrained flow region of a precombustor
using calcium-based sorbents, 2) the evolution of sulfur (and other species)
from coal in an entrained flow process, and 3) the solubility and stability
of sulfur in molten slag layers.
Figure 4-1 presents a flow chart of the overall program approach. Phase
I was a concept screening phase which resulted in a demonstration of the
potential of simultaneous control of N0X, S0X and particulate emissions from
a precombustor at the pilot scale and identified several key issues to be
addressed in the Phase II testing.
Three facilities were utilized in Phase II of the program. The
Isothermal Reactor facility (ITR) was used to investigate the formation of
sulfides from the injection of calcium-based sorbents into the exhaust of
gaseous fuel flames doped with H2S. The ITR provided a well controlled
reaction atmosphere (in terms of temperature and residence time) for the
investigation of particular reactions and identification of key operating
conditions.
The evolution of sulfur from coal and its subsequent reaction with
sorbent was investigated in the Control Temperature Tower facility (CTT).
The CTT was larger and less fragile than the ITR so as to accommodate coal
firing. Gas phase sampling and analysis by gas chromatography allowed
identification of sulfur speciation. Solids sampling and analysis identified
the evolution of the major elements and provided a cross-check of gas phase
sulfidation results.
4-1
-------
B & W Contract 940962NR
EPA Contract 68-02-3130
TASK 2.1
Burner
Evaluation
1
X
I
TASK 2.2
Alternate Burner
Concepts
PHASE I
Precombustor
Screening
I
Precombustor Design
& Test Preparation
Phase I
Test Results
Precombustor Screening
Tpstina
N0X Control
Experience
Phase II
Work Plan
Figure 4-1. Program approach flow-chart.
4-2
-------
The third facility utilized was constructed specifically for this work
and is referred to as the High Temperature Oven facility (HTO). The system
consisted of a high.temperature tube furnace coupled to a gas blending
system. Tests were performed to determine the equilibrium level of sulfur
obtained in various slag-CaO mixtures, the rate at which CaS decomposes in
molten slags, and the influence of fluxing additives on the capacity of the
slags to hold sulfur and the fluidity of slags.
4-3
-------
5.0 EXPERIMENTAL SYSTEMS
Three experimental facilities were used in this program: a laboratory-
scale Isothermal Reactor (ITR); a bench-scale Control Temperature Tower
(CTT); and a laboratory-scale High Temperature Oven (HTO).
5.1 Isothermal Reactor (ITR)
A cross-sectional view of the ITR is shown in Figure 5-1. The reaction
chamber was centered in a 21 kW (72,000 Btu/hr) Lindberg furnace with maximum
operating temperature of 1500°C (2700°F). The heated zone was 0.9 m (35
inches) long and could hold a 10 cm (4 inches) diameter reaction tube. Three
independently controlled heating zones, separated by baffles, ensured that a
flat temperature profile was maintained over the heated length.
The ITR was downfired by a precombustor assembly coupled with a sliding
burner, as shown in Figure 5-2. There are two main features to this
assembly: a sliding burner and a venturi. The sliding burner was a
premixed, water cooled, flat-flame burner in a square stainless steel
chimney. Due to heat losses through the chimney, variance of the burner
position allowed downstream temperature control independent of the gas
velocity and burner stoichiometry. The venturi, located immediately
downstream of the sorbent injector/flow straightener, created turbulence
which enhanced sorbent dispersion in the flame gases. The gradually
divergent section after the venturi throat prevented backmixing and promoted
a flat velocity profile.
The sorbent injector utilized swirl ports near the injector outlet to
produce additional dispersion of the sorbent as it mixed with the reaction
gases. Solid or gas sampling probes were inserted axially at the exhaust end
of the reactor. Different reaction residence times were achieved by varying
the elevation of a probe tip in the reactor.
Table 5-1 presents temperature data taken radially along the ITR length
to demonstrate the level of temperature uniformity throughout the reactor.
5-1
-------
Sorbent Injector
Figure 5-1. Schematic of the isothermal reactor (ITR).
5-2
-------
Sorbent Injection
Injector $>ryi
Sheath
Flat Flame on Water-Cooled
Sintered Bronze Plug
3
Ceramic Flow
Straightner
Sorbent Dispersion Venturi
Isothermal Reactor
Sorbent Dispersion Pattern
Figure 5-2. Schematic detail of the precombustor/sliding burner assembly.
-------
TABLE 5-1. TEMPERATURE VARIATION IN THE ITR AT A NOMINAL 900°C
ISOTHERMAL CONDITION
Distance from
Burner, cm
Axi s
Temperature,
N S
°C
E
U
107
897
895
895
899
895
94
909
910
911
910
910
905
906
904
904
900
81
904
907
909
906
910
903
905
906
902
905
69
903
902
908
904
906
904
906
911
910
913
56
903
902
912
901
907
909
908
916
913
912
43
903
902
910
909
906
902
904
906
903
909
30
900
898
896
902
898
5-4
-------
5.2
Control Temperature Tower (CTT)
An overall diagram of the CTT furnace is presented in Figure 5-3. The
reactor entry (top section) includes a 45.7 cm (18 inches) long quarl that
diverges from 5.1 cm (2 inches) to 20.3 cm (8 inches). The long quarl
provided flame stabilization as well as established a one-dimensional flow
pattern. Numerous ports located along the axis of the reactor enabled
supplementary equipment; i.e., zone separation chokes, fuel and air
injectors, and cooling coils, to be placed in the reactor.
The CTT had the capability of manipulating the time/temperature profile
by firing two sets of back-fired heating sections (second and fourth sections
from the top). The back-fired sections employed natural gas burners which
fired in the direction opposite to the main combustion gas. The back-fired
effluent passed through channels in the refractory walls surrounding the main
combustion zone (see the radial cross-sectional view in Figure 5-3).
Backfiring in this manner minimized the temperature decay of the combustion
products in the main combustion zone.
Each section of the CTT was fabricated from a 91.4 cm (36 inches)
diameter steel shell (0.32 cm (1/8 inch) thick), two layers of castable
insulation and two layers of castable refractory. The multi-layer design of
the CTT is illustrated in Figure 5-4.
To best simulate the gas phase thermal environment of a precombustor
system, the CTT was operated with both backfired sections on to minimize heat
loss and maintain as close to an isothermal reaction zone as possible. The
fuel-rich and secondary burnout stages were separated by a 10.2 cm (4 inch)
ID choke to prevent backmixing of burnout air into the fuel-rich reaction
zone. The choke was located 155 cm (61 inches) from the burner entry.
The cross-sectional layout of the premixed burner used on the CTT is
shown in Figure 5-5. The primary air and fuel (natural gas or coal) were
injected radially outward through six 0.32 cm (13/16 inch) holes into a
mixing chamber. The preheated secondary air is injected radially inward into
5-5
-------
Primary Fuel/Air
Figure 5-3. Cross-sectional views of the CTT configured with a fuel-rich zone.
-------
J—20.3—^
91.4
^^2144 K Castable
Refractory
Klv.i 1922 K Castable
Refractory
I7-7-J1477 K Castable
' ^ ^ Insulation
P7T11144 K Castable
Insulation
Dimensions in cm.
Figure 5-4. Multi-layer refractory design of the CTT.
5-7
-------
Coal/Primary Air
Figure 5-5. Premixed burner.
5-8
-------
the mixing chamber where the air and fuel are uniformly mixed. The premixed
burner produced essentially a one-dimensional, plug-flow flame with no
backmixing. Preignition of the fuel/air mixture in the mixing chamber was
prevented by the cooling matrix located at the exit of the burner.
The furnace operating conditions for the test series were as follows:
• Primary Zone Stoichiometric Ratio (SRi) - 0.8, 0.6.
• Overall Stoichiometric Ratio (SR2) - 1.25.
• Firing Rate - 23.4 kW (80,000 Btu/hr), 17.6 kW (60,000 Btu/hr).
• Primary Zone Residence Time - 1.4 to 2.0 seconds.
Figure 5-6 presents measured temperature profiles in the CTT for each
operating condition. The sorbent injection locations are designated. For
each case, injection was performed through the same port, located 57.2 cm
(22.5 inches) downstream from the burner.
Sorbent was transported into the furnace using pure nitrogen. A four
hole, 45° injector was used to achieve rapid mixing between the sorbent and
the furnace gas flow. The burnout air was introduced with an air ring that
injected the air radially inward.
5.3 High Temperature Oven (HTO)
To study the interaction of sulfur with molten slags, the HTO facility
presented in Figure 5-7 was designed and assembled. Gas mixtures of N£, CO,
CO2 and H2S were premixed and fed to the oven in a mixture approximately
corresponding to Illinois coal flue gas at stoichiometric ratios of 0.5 and
0.7 (according to the NASA equilibrium code). The blended gas compositions
are presented in Table 5-2. The electrical tube furnace had silicon-carbide
heating elements to allow operation up to 1500°C (2732°F) throughout a 5.1 cm
(2 inches) diameter, 30.5 cm (12 inches) long heated zone.
A high purity alumina tube was used for the reaction chamber. The tube
extended out of the furnace to provide a cool-down region. Inside the
5-9
-------
CD
U
=3
+->
(C
S-
17.6 kW, SR = 0.8
800
-
A 17.6 kW, SR = 0.6
600
400
-
-
200
-
0
«
• i ¦ • i i
1600
1400
1200
1000
800
600
400
200
0
o
o
Q)
U
3
4->
to
5-
0)
Q.
E
0)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Residence Time (sec)
Figure 5-6. Temperature profiles measured in the fuel-rich zone
of the CTT for three different operating conditions
when burning Illinois coal.
5-10
-------
Zinc
Acetate
Solution For
(20% in N2)
Figure 5-7. High temperature oven test facility with gas blending capability.
-------
TABLE 5-2. HTO FEED GAS COMPOSITIONS
Gas Composition, Vol.
Percent
Stoichiometric Ratio
CO
C02
H2S
0.5
68.25
26.35
4.90
0.50
0.7
74.96
13.35
11.19
0.50
5-12
-------
furnace tube was a dee-shaped alumina tube which formed a platform for
ceramic sample boats to slide upon. The inside of the dee-shaped tube also
provided a channel for preheating of the inlet gas before exposure to the
slag samples. Nitrogen was used as the purge gas so that the samples cooled
in an oxygen-free atmosphere.
Each ceramic sample boat held two smaller platinum boats made from 0.76
mm (0.003 inches) foil. The use of platinum was mandatory to prevent
interaction between the slag samples and the constituents of the ceramic
boats. The platinum was not completely inert. Iron in the slags attacked
the platinum leaving a brittle, sometimes porous boat after a few tests.
The furnace temperature was maintained by a proportional-type
controller. Temperature profiles were measured with and without a nitrogen
flow of 2000 scc/min (4.2 scfh) through the reactor and are presented in
Figure 5-8. The samples were positioned within a span of 12.7 cm (5 inches)
in the reactor to assure a uniform temperature to within +10°C (+18°F).
Slag samples were prepared by preweighing required amounts of Illinois
coal ash (ashed by ASTM procedures) CaO, CaS, or other additives.
5.4 Continuous Flue Gas Sampling System
Flue gas sampling was conducted in both the ITR and the CTT with a
stainless-steel, water-cooled probe (jacketed with water cooling). The probe
was capable (by inertial separation) of splitting the sorbent-laden sample
gases into two streams. One stream was nearly particle-free to prevent
additional sulfur reactions with sorbent within the probe. The other
particle-laden stream was filtered downstream of the probe. The tip of this
probe, called a phase discrimination probe, is shown in Figure 5-9.
The particle-laden sample stream was used to monitor O2, CO, CO2 and N0X
emissions. As shown in Figure 5-10, the sample stream was filtered, dried by
passage through condensation traps and pumped to the analysis console. The
gas species were continuously measured by the method indicated:
5-13
-------
Distance from Reactor Exhaust (cm)
Figure 5-8. High temperature oven facility temperature
distributions.
5-14
-------
Dimensions: Inches
Figure 5-9. Phase discrimination probe tip.
-------
To
Exhaust
System
Figure 5-10. Continuous monitoring flue gas sampling system.
-------
• O2 - paramagnetic
§ CO/CO2 - nondispersive infrared
• N0X - chemiluminescent
The particle-free sample stream was used to monitor SO2 emissions. The
sample was transferred through a heated sample line and filter to a
permeation tube dryer before being fed to a nondispersive ultraviolet
analyzer (see Figure 5-11). The dryer selectively removed water from the gas
stream by contacting the wet gas with a tubular membrane which was permeable
to water. Sampling moist gases in this manner enabled a dry sample stream to
be produced without the possibility of water condensation and sulfuric acid
formation.
5.5 Reduced Sulfur Species Sampling System
To measure reduced sulfur species such as H2S, COS and CS2 (as well as
SO2) by gas chromatography, the same sample train system was used as for
continuous SO2 sampling except that the dry sample stream was fed to a tedlar
sample bag rather than the ultraviolet SO2 analyzer (see Figure 5-11). The
collected sample was then transported to the EER chemistry lab for immediate
(within two hours) analysis in a Perkin Elmer £2 Gas Chromatograph equipped
with a flame photometric detector and a Supelco Chromosil 310 column.
5.6 Solid Sampling Techniques
On the ITR, solid samples were collected with the probe presented in
Figure 5-12. The probe was designed for isokinetic operation in the ITR
facility. The sample was rapidly quenched upon entry into the probe in order
to prevent further sulfidation reactions to occur. The design includes two
water-cooled zones for better quench control without formation of
condensation within the probe.
Solids sampling on the CTT did not require as long a probe as on the ITR
since sampling was performed through side ports rather than through the
exhaust. The system used is shown in Figure 5-13. Very high sample rates
5-17
-------
Phase Discrimination Probe
Sample —E
tn
I
CO
Filter •
High Flowrate
(Particulate-Laden Stream)
-Low Flowrate (Particulate-Free Stream)
Purge
Air Out
t
LTL
Perma-Pure
Drier
Heated Oven
Purge!
Air1
In
Sample
Pump
Air
Drier
Tedlar
Sample
Bag
Final
Filter
High Low S02
SO2 Calibration
Calibration Gas
Gas
\
S02
Analyzed
\
i
Figure 5-11. SO2 continuous monitoring and batch sampling system.
-------
Water
Out In
Cooling Water
Figure 5-12. Isokinetic solids sampling probe with upper and lower cooling jackets used
in the ITR.
-------
Sample
Stream
1/2" Stainless Steel Probe
CJ1
I
ro
o
Compressor
Air
Figure 5-13 • Solid sampling system used on the CTT.
-------
were used through a hot probe to avoid long probe residence times and to
maintain a hot sample stream. A thermocouple behind the sample filter was
used to assure moderate temperatures that remained above the dew point.
5.7 Sorbent Feeding
For both the ITR and CTT facilities, sorbent was fed using an Accurate
Model 300 solids displacement feeder. Features of the feeder include
variable speed replaceable augers and a flexing vinyl hopper to avoid
material bridging. The dry sorbent was transported to each furnace using dry
nitrogen.
5.8 Temperature Measurement
Gas temperatures were measured in the ITR using a fine-wire thermocouple
to minimize radiation losses. The thermocouple was made with 0.025 mm
diameter type-S thermocouple wire, butt-welded across a forked support.
Figure 5-14 presents a diagram of the fine-wire thermocouple.
In the CTT, where more room was available, gas temperatures were
measured with a B-type thermocouple in a suction pyrometer jacket as shown in
Figure 5-15. The pyrometer was placed in the flue gas stream with the
opening in the alumina sheath pointing downstream. A sufficient gas sample
was drawn around the thermocouple tip such that the convective heat transfer
of the gas to the thermocouple dominated over the radiation heat loss of the
thermocouple tip to the surroundings.
5.9 Test Material Compositions
For the entrained flow sulfidation tests, the sorbents used were Linwood
atmospheric hydrate and Vicron 45-3 carbonate. Compositions of the two
sorbents are presented in Table 5-3 and their size distributions are
presented in terms of cumulative weight percent in Figure 5-16. Both of
these sorbents are commercially available and have been investigated in fuel-
lean sulfidation studies.
5-21
-------
Copper/Alloy 11
extension wires
To readout
Figure 5-14. Fine-wire thermocouple probe for gas temperature
measurements in the ITR.
5-22
-------
Temperature
Readout
Regulator Ejector
Figure 5-15. Gas temperature measurement system for the CTT.
-------
TABLE 5-3. PROPERTIES OF VICRON LIMESTONE AND
LINWOOD ATMOSPHERIC HYDRATE
Sorbent
Vicron
Limestone
Li nwood
Atmospheric
Hydrate
Surface Area (m^/gm)
0.7
16.0
Median Particle Size i/im)
9.8
2.2
Ca (weight %)
39.2
51.6
Mg
0.33
0.97
Na
0.01
0.00
K
0.01
0.00
Fe
0.06
0.27
Si
0.09
0.66
A1
0.02
0.18
5-24
-------
Equivalent Spherical Diameter, nm
Figure 5-16. Comparison of sorbent size distributions as measured by x-ray sedigraph
technique.
-------
The Illinois coal used for the entrained flow tests is a high sulfur,
high slagging, eastern coal commonly used in cyclone combustors. Its
composition is presented in Table 5-4.
For the slag tests performed in the static oven facility, Illinois coal
ash was obtained from the same Illinois coal as used in the entrained flow
studies. The coal was ashed by ASTM specified procedure (in air at 750° C
(1382°F)). Table 5-5 presents the mineral analysis of the ash.
Additives used to mix with the ash included CaO (97.3 percent pure), CaS
(97.0 percent), MgO (95.0 percent), Fe203 (100 percent), P2O5 (99.7 percent),
B2O3 (98.9 percent), and CaF2 (99 percent). Table 5-6 presents the
calculated mineral distribution for mixtures corresponding to coal firing
with Ca/S molar ratios of 0.5, 1.0 and 2.0 and Fe/S of 1.0, 2.0, 4.0, 2.55
(with 15.1 percent MgO added) and 4.4 (the last two were tested for direct
comparison to reported results by DeYoung, 1984).
5-26
-------
TABLE 5-4. ILLINOIS COAL COMPOSITION AND HEATING VALUE
Ultimate Analysis (wt%, dry)
Carbon
69.52
Hydrogen
4.98
Nitrogen
1.33
Sulfur
3.84
Oxygen
10.71
Chlorine
0.12
Ash
9.50
Proximate Analysis (wt%, as received)
Moisture 7.92
Ash 8.75
Volatiles 38.46
Fixed Carbon 44.87
Higher Heating Value (dry)
29,602 kJ/kg (12,726 Btu/lb)
5-27
-------
TABLE 5-5. MINERAL ANALYSIS OF
ILLINOIS COAL ASH
Mineral (wt %, dry)
Si02
47.93
A12O3
16.73
Fe203
17.37
MgO
1.11
CaO
4.67
Na20
1.02
K20
2.21
HO2
0.94
P2O5
0.14
S03
4.11
MnO
0.03
Others
3.74
Total
100.00
Molar Basicity
0.27
5-28
-------
TABLE 5-6. COMPOSITIONS OF CaO/ASH AND Fe203/ASH MIXTURES
Ca/S (molar ratio)
0.5
1.0
2.0
_
Fe/S (molar ratio)
~
—
—
1.0
2.0
4.0
2.55
4.4
CaO/Ash (wt. ratio)
0.31
0.66
1.37
--
—
--
—
~
Fe203/Ash (wt. ratio)
—
--
—
0.83
1.84
3.85
2.39
4.26
Molar Basicity
0.83
1.48
2.77
0.8
1.45
2.75
3.08
3.01
SiOe (wt %)
36.43
28.55
19.92
26.14
16.87
9.87
12.31*
9.11
A1203
12.72
9.96
6.95
4.12
5.89
3.45
4.38
3.18
Fe203
13.20
10.35
7.22
54.93
70.91
82.98
65.94
84.30
MgO
0.84
0.66
0.46
0.61
0.39
0.23
13.12
0.21
CaO
26.89
42.14
58.81
2.55
1.64
0.96
1.20
0.89
Na20
0.78
0.61
0.42
0.56
0.36
0.21
0.26
0.19
K2O
1.68
1.32
0.92
1.21
0.78
0.46
0.57
0.42
Ti02
0.71
0.56
0.39
0.51
0.33
0.19
0.24
0.18
P2O5
0.11
0.08
0.06
0.08
0.05
0.03
0.04
0.03
SO3
3.12
2.45
1.71
2.24
1.45
0.85
1.06
0.78
MnO
0.02
0.02
0.01
0.02
0.01
0.00
0.00
0.00
Others
3.5
3.3
3.13
2.04
1.32
0.77
0.96
0.71
Total
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
*15.1% by weight of Fe203/ash mixture added as MgO
-------
6.0 ENTRAINED FLOW SULFIDATION RESULTS
To investigate the free-stream reaction of calcium-based sorbents with
reduced species of sulfur (H2S and COS), sorbent was injected into an
isothermal, fuel-rich exhaust stream in the ITR facility. The sulfur
concentration was controlled by the doping of H2S either with the gaseous
fuel or downstream of the flame zone. This facility was ideal for precise
control of stoichiometry, temperature and gas phase residence time. Gas
phase sampling and analysis using gas chromatography allowed determination of
the sulfur speciation.
Gradual sulfur evolution from coal firing adds another level of
complexity to the process. The evolution of sulfur from coal and its
subsequent reaction with calcium-based sorbents was investigated in the CTT
facility. The CTT was operated with a slow temperature quench rate to
approximately simulate the free-stream thermal environment within a
precombustor. Again, gas chromatography was used to establish gas phase
sulfur speciation as the sulfur evolved from the coal. Solid sampling and
analysis allowed determination of specie (C, S, H, N) evolution rates and
provided an independent check on sulfidation.
6.1 Isothermal — Gas-Fired
6.1.1 Fuel Effects
Before attempting to determine the effectiveness of calcium sulfidation
as a sulfur removal process from sulfur laden, fuel-rich exhaust streams, it
was of critical importance to establish the sulfur speciation within the ITR
reactor environment. It was found that the use of different gaseous fuel
types impacted the sulfur speciation and indirectly effected the measured
calcium utilization. Two fuels were tested, methane (CH4) and ethylene
(C2H4). Each was doped with enough H2S to generate 5000 ppm of sulfur in the
reactor. Figure 6-1 shows the difference in sulfur speciation for the two
fuels at a temperature of 1100°C (2012°F) and stoichiometric ratio of 0.83.
The methane flame exhaust sulfur speciation was mostly H2S (about 50
6-1
-------
Fuel: CH4/H2S
Fuel: C2H4/H2S
0.25 0.50 0.75 1.0
Residence Time (s)
100 -
0.25 0.50 0.75
Residence Time (s)
Figure 6-1. Sulfur speciation in the ITR at 1100°C, SR = 0.83 when doping
CH4 and C2H4 with approximately 5000 ppm H2S.
-------
percent), but had substantial amounts of SO2 and CS2» The ethylene flame
provided about the same amount of H2S but substantially higher COS and little
measurable SO2 or CS2.
As shown in Figure 6-2, the sulfur speciation had an impact on calcium
utilization. The larger amount of H2S and COS present in the ethylene flame
case resulted in a slightly higher calcium utilization due to the
availability of the sulfur compounds which support sulfidation.
The sulfur speciation for both of these cases were far from expected
based on equilibrium calculations. It is believed that under fuel-rich
operating conditions on the ITR that a delayed flame front exists. The high
peak temperatures and large radical specie concentrations that exist in the
flame front produces a non-equilibrium sulfur speciation at the entrance of
the reactor. As the exhaust gases travel through the reactor, the sulfur
speciation slowly moves towards equilibrium (see Figure 6-1).
Ethylene is a "faster" fuel than methane producing a distribution of
sulfur species at the reactor entrance which were closer to equilibrium. In
addition, a new burner was utilized with the ethylene which more effectively
held the flame to the burner.
There was, in fact, some visual evidence that the methane flame
sporadically licked into the reactor zone. This also produced unsteadiness
in the sulfur speciation.
6.1.2 Impact of H2S Injection Location
To avoid the problems of shifting sulfur speciation, the dopant H2S was
injected downstream of the ethylene flame. This method produced a very high,
steady H2S concentration throughout the reactor. Figure 6-3 compares sulfur
speciation when H2S was premixed with the ethylene to downstream H2S doping.
Figure 6-4 shows that the capture levels using Linwood atmospheric hydrate
are higher for the case of higher H2S concentration. Even though the total
reduced sulfur concentration (H2S-plus-C0S) is the same, it is believed that
6-3
-------
Fuel: CH4/H2S
T
Fuel: C2H4/H2S
100 -
0.25 0.50 0.75 1.0
Residence Time (s)
0.25 0.50 0.75
Residence Time (s)
O SR = 0.83
~ SR = 0.63
O SR = 0.50
Open Symbols:
Gas Analysis
Closed Symbols:
Sol ids
Analysis
Figure 6-2. Calcium utilization in the ITR using Linwood atmospheric hydrate, measured by
both gas and solids analysis. ITR held at 1100°C when doping CH4 and C2H4 with
5000 ppm H2S (Ca/S =*1.0 for each case).
-------
CT>
I
cn
100 -
u
13
nS
4->
O
4->
C
cu
u
5-
-------
H2S Added to C2H4
H2S Injected Downstream
of C2H4 Flame
100 -
0.25 0.50 0.75
Residence Time (s)
O SR = 0.83
~ SR = 0.63
Open Symbols:
Gas Analysis
Closed Symbols:
Solids
Analysis
0.25 0.50 0.75 1.0
Residence Time (s)
Figure 6-4. Calcium utilization in the ITR using Linwood atmospheric hydrate, measured
by both gas and solid analysis. ITR held at 1100°C when doping the fuel
(C2H4) with 5000 ppm H2S or injecting it downstream of the flame
(Ca/S =* 1.0 for each case).
-------
the kinetics of sulfidation and sorbent sintering favors the higher H2S
concentration. Borgwardt et al. (1984a) have shown that the reaction rate
for H2S and CaO reaction is about three times as high as the COS and CaO
reaction. Within similar residence times, this would favor the higher H2S
concentration. In addition, Borgwardt has shown that the reactivity of the
sorbent is reduced with time due to thermal sintering. Therefore, the case
of higher COS concentration would provide more time for thermal sintering to
occur resulting in lower capture levels.
One other complexity that must be considered is the shifting of sulfur
species upon the depletion of H2S. As Figure 6-5 shows, the measured
relative speciation when sorbent was injected was similar to that measured
before sorbent was injected. This suggests that there was a shift in
speciation from COS to H2S as the H2S pool was depleted.
6.1.3 Time-Resolved Capture
The capture data presented in Figures 6-2 and 6-4 were obtained at
various reactor residence times by adjusting the gas phase sample probe along
the length of the reactor. In all cases, the capture appears to be dominated
by events that occur early in the reactor. It is believed that the sorbent
is most reactive shortly after calcination and quickly losses reactivity as
available reaction sites are used up and as the thermal sintering process
reduces specific surface area. Once an external product layer is produced,
further reaction is limited by product layer diffusion. These results
emphasize the importance of the thermal conditions and sulfur speciation that
exist early in a precombustor.
6.1.4 Stoichiometry Effects
In the ITR facility, the sulfur was introduced as a gaseous dopant so
that sulfur evolution was not an issue. The speciation was dominated by non-
equilibrium influences, such as proximity of the flame front to the reactor
entrance. Because of these factors, the flame stoichiometry only had a
strong impact on sulfur speciation when it altered flame conditions for the
6-7
-------
S-
3
<4-
3
cn
+->
o
M-
O
c
O)
o
s-
-------
cases where H2S was doped with the fuel. For the cases when H2S was injected
downstream of the flame, the concentration distribution did not vary with
stoichiometry, as shown in Figure 6-6. Figures 6-2 and 6-4 showed earlier
that sulfur capture levels were similar as stoichiometry was varied from SR =
0.83 to 0.63.
6.1.5 Temperature Effects
An important issue facing the use of calcium-based sorbents in practical
precombustor systems is the allowable temperature range in which the process
is effective. Figure 6-7 shows that an increase in reactor temperature from
1100°C (2012°F) to 1300°C ( 2372°F) only slightly shifted the sulfur
speciation towards lower H2S and higher COS concentrations. However, calcium
utilization was significantly reduced at the elevated temperature as shown in
Figure 6-8. It is believed that the high temperature significantly increases
sorbent sintering so that a much less reactive material is available for
sulfidation. At temperatures above 1100°C, the specific surface area of
sorbents has been shown to diminish significantly (Cole et al., 1985b).
Thus, operating temperatures in high firing density, slagging precombustors
may be too high to achieve optimum calcium utilization.
6.1.6 Impact of Sorbent Type
Linwood atmospheric hydrate was chosen as the baseline sorbent due to
its high reactivity so that the impact of changing parameters could be
identified by the capture data. From an economic standpoint the performance
of an inexpensive calcium carbonate (CaC03) is of interest. A common
limestone, Vicron 45-3, was tested in the ITR giving the capture results
presented in Figure 6-9. Calcium utilization for the limestone was
significantly lower than for the Linwood hydrate but high compared to Vicron
when tested under fuel-lean, sulfation conditions. The relative capture
levels when comparing Linwood to Vicron (i.e., their reactivity ranking) is
similar under these fuel-rich sulfidation conditions to fuel-lean results on
this facility and others (Cole et al., 1985a, EER, 1986, and Overmoe et al.,
1986). This suggests that the ultimate entrained flow sulfidation will be
6-9
-------
cr»
i
s-
13
rO
4->
O
+->
C
d)
u
s-
D_
SR = 0.63
0.25 0.50 0.75 1.0
Residence Time (s)
100
~ H2S
O COS
0.25 0.50 0.75 1.0
Residence Time (s)
Figure 6-6. Sulfur speciation in the ITR at 1100°C when injecting
5000 ppm H2S downstream of the C2H4 flame.
-------
1100°C (2012°F)
1300 C (2372 F)
o>
I
s-
13
oo
as
+->
o
-M
C
(U
(J
i-
a>
cu
0.25 0.50 0.75
Residence Time (s)
~ H2S
O COS
0.25 0.50 0.75 1.0
Residence Time (s)
Figure 6-7. Sulfur speciation in the ITR at SR = 0.83 when injecting
5000 ppm H2S downstream of the C2H4 flame.
-------
1 1 "T
O
-1 - 1
O—
~
/0^
—0—
•
. /
1 -J
1 , 1
O 1100°C (2012°F)
o 1300°C (2372°F)
Open Symbols:
Gas Analysis
Closed Symbols:
Solids Analysis
0.25 0.50 0.75 1.0
Residence Time (s)
gure 6-8. Calcium utilization in the ITR at SR = 0.83 when
injecting 5000 ppm H2S downstream of a C2H4 flame.
Sorbent was Linwood atmospheric hydrate with
Ca/S =*1.0 for each case.
6-12
-------
c
O)
u
S-
0)
CL
c
o
4-5
(0
N
r0
O
/\ Linwood Atmospheric
v Hydrate
^ Vicron 45-3
Limestone
Open Symbols:
Gas Analysis
Closed Symbols:
Solids Analysis
0.25 0.50 0.75
Residence Time (s)
Figure 6-9. Calcium utilization in the ITR at 1100°C,
SR = 0.83 with 5000 ppm H2S injected down-
stream of the C2H4 flame. Both sorbents
fed at Ca/S =*1.0.
6-13
-------
limited in application by the reactivity of the sorbent. Other studies on
sorbent reactivity for fuel-lean sorbent utilization (Silcox et al., 1986)
can provide useful information on the ranking of sorbent types and the
optimum thermal environment for maximum calcium utilization.
6.1.7 Gas Phase Versus Solids Sampling
Under nearly every test condition, both gas phase and solid sampling was
performed at the reactor exit to help validate the capture results. Figures
6-2, 6-4 and 6-8 show that the results were typically within 10 percent of
each other; close enough to give support to any conclusions drawn from these
results. The fact that the solids results are consistently lower than the
gas phase results may be due to a small build-up of sorbent on the walls of
the reactor which would cause a slight increase in gas phase capture but
would not impact the sorbent that passed straight through the reactor and was
collected by the solids sampling probe.
6.1.8 Comparison of Fuel-Rich to Fuel-Lean Capture
Figure 6-10 presents a comparison of calcium utilization for both
Linwood hydrate and Vicron carbonate when tested under fuel-lean and fuel-
rich conditions at similar temperatures. Even though the batch sampling
technique used under fuel-rich conditions is less precise than the continuous
SO2 sampling system used for fuel-lean capture testing the substantial
improvements in calcium utilization under fuel-rich conditions is
unmistakable. Certainly the process holds promise as an effective sulfur
reduction technique if the proven conditions can be reproduced in practical
precombustor systems.
6.2 Slow Quench Rate — Coal-Fired
To investigate the potential of fuel-rich sulfur capture under more
realistic conditions, a coal-fired test series was initiated on the CTT. The
CTT facility provided the opportunity to test carbon burnout and sulfur
6-14
-------
Linwood Hydrate
a%
cn
+->
c
-------
evolution from coal under fuel-rich conditions at the same time as evaluating
the sulfidation reaction.
6.2.1 Stoichiometry Effects
The CTT was operated at two stoichiometric ratios, 0.8 and 0.6. A
stoichiometric ratio of 0.8 was considered optimum for practical precombustor
operation in terms of N0X control and favorable carbon utilization. A level
of 0.6 was established as a reasonable lower limit for operation of a
precombustor system.
Figure 6-11 presents a comparison of sulfur speciation along the reactor
length. Clearly operation at SR = 0.8 was not acceptable since less than 30
percent of the input sulfur was in the form of H2S and COS over a majority of
the reactor residence time. However, operation at SR = 0.6 made
approximately 70 percent of the gas phase sulfur available as H2S and COS for
reaction with CaO.
As expected, capture levels for the SR = 0.6 case are substantially
higher than at SR = 0.8 due to the availability of the reduced sulfur species
for reaction with the sorbent. Figure 6-12 shows the substantial benefits of
operation at SR = 0.6. Practical precombustor systems which promote fast and
efficient evolution of sulfur from coal along with speciation that favors H2S
and COS would be more likely to give higher sulfur capture using calcium-
based sorbents.
The evolution of carbon and other species for both stoichiometrics are
shown in Figure 6-13. Evolution of hydrogen, nitrogen and sulfur were
similar for the two cases while consumption of carbon lagged behind
noticeably for the SR = 0.6 case. At the richer stoichiometry, only 70-75
percent of the carbon evolved while around 90 percent of the carbon evolved
at SR = 0.8 by the end of the reaction zone. This emphasizes one of the
important practical issues facing sulfur control in precombustors - can high
carbon utilization be achieved at the extreme fuel-rich stoichiometrics
required for sulfur capture? Extended residence times by deposition and
6-16
-------
100
80
60
40
20
SR = 0.8
0.5 1.0 1.5 2.0
Residence Time (s)
SR = 0.6
0.5 1.0 1.5 2.0
Residence Time (s)
~ H2S
A S02
O COS
q Char S
Figure 6-11. Sulfur speciation in the CTT firing Illinois coal at 17.6 kW (60,000 Btu/hr).
-------
CT>
00
-------
0.5 1.0 1.5
Residence Time (s)
SR = 0.6
0.5 1.0 1.5 2.0
Residence Time (s)
k Hydrogen
~ Nitrogen
O Carbon
A Sulfur
Figure 6-13. Species evolution from Illinois coal in the CTT at 17.6 kW (60,000 Btu/hr).
-------
combustion of coal on precombustor walls or by aerodynamically suspended
combustion are potential requirements for effective precombustor designs.
For the case of higher stoichiometry (i.e., more fuel-lean), the H2S and
COS levels were reduced while the SO2 concentration rose down the length of
the reactor. From these results alone it is difficult to determine how the
sulfur evolved from the coal, particularly considering the incomplete char
burnout. Figure 6-14 shows, along with measured gas phase sulfur speciation
data, the calculated equilibrium distribution of sulfur species at each
sample point. The equilibrium values were calculated based on the local gas
phase stoichiometry and measured temperature. At each sample point, the
local stoichiometry is higher than the overall calculated stoichiometry due
to slow carbon utilization.
The results show that the excess O2 conditions found early in the coal
flame thermodynamically favor the formation of S02- The fact that the
measured sulfur species were relatively low in SO2 and high in H2S and COS in
the early portion of the flame indicates that a significant amount of the
sulfur must have evolved in the form of H2S and/or COS. In other words, if
the sulfur evolved as SO2, equilibrium considerations would indicate that no
H2S or COS should be present early in the coal flame. The measured sulfur
speciation began to shift towards the equilibrium levels by the end of the
reaction zone. The direct evolution of reduced sulfur species favors fast
reaction with calcium based sorbents. However, slowly evolving carbon can
cause higher local stoichiometrics which favor higher SO2 concentrations and
lower sulfur capture by the sorbent in the early portion of the reactor.
Figure 6-12 presents the capture data as a function of residence time.
Capture levels rose to about 80 percent of the ultimate capture level within
approximately 0.4 seconds. Just as was found with the ITR data, the initial
reactor conditions in which the sorbent is injected into appear to be crucial
to the ultimate capture achieved. This suggests that in a practical
precombustor system, the sorbent be separated from the incoming fuel. This
would allow the sulfur to evolve and speciate into desirable sulfur species
before the sorbent is injected and experiences its peak reactivity.
6-20
-------
r
t r
100
80
60
40
20
.-J-•-r'. i
0.5 1 .0 1 .5 2.0 00
Residence Time (s)
A so2
~ h2s
O COS
Open Symbols:
Measured
Speciation
Solid Symbols:
Calculated
Equilibrium
Speciation
Figure 6-14. Comparison of measured gas phase sulfur
speciation from an SR = 0.8 Illinois coal
flame to equilibrium speciations calculated
at conditions which correspond to the local
measured stoichiometry and temperature.
6-21
-------
Solid sample analysis for calcium utilization was calculated based on
the assumption that the sulfur evolved from the coal at the same rate when
sorbent was present as when it was not present. The solid sample capture
results corresponded well with gas phase capture data (see Figure 6-12)
indicating once more the validity of the sampling techniques and supporting
the capture levels.
6.2.2 Load Effects
Two reactor loads were tested to help determine the sensitivity of the
process to thermal environment. The high load case of 23.4 kW (80,000 Btu/
hr) gave a slightly higher sorbent injection temperature (see Figure 5-6) and
a higher average reactor temperature than the 17.6 kW (60,000 Btu/hr) case.
In addition, at the same stoichiometry the high load case produced a 33
percent higher exhaust flowrate. This caused a corresponding reduction in
reactor residence time. As Figure 6-15 shows, these changes in reactor
conditions had little impact on reaction rates or ultimate capture. The rate
and extent of species evolution and the speciation of evolved sulfur
compounds was similar for both cases as shown in Figures 6-16 and 6-17,
respectively. These results indicate that the process is relatively
insensitive to furnace load conditions; however, the CTT facility did not
show a great change in thermal environment as load was changed. The data
obtained at these two loads is more a confirmation of the repeatability of
the results than a test of the impact of thermal environment.
6.2.3 Impact of Sorbent Type and Feed Rate
Both Linwood AH and Vicron 45-3 were tested at the 23.4 kW (80,000 Btu/
hr) load, SR = 0.6 furnace condition. Figure 6-18 presents the percent
reduction of total sulfur input to the system as a function of Ca/S molar
ratio. The Ca/S molar ratio is a way of expressing the changing sorbent
feedrate relative to the amount of sulfur present. As expected from the ITR
results and results from fuel-lean sulfation tests, the capture ability of
Linwood hydrate was substantially higher than Vicron carbonate (a factor of 2
to 3 times higher).
6-22
-------
O 23.4 kW
(80,000 Btu/hr)
~ 17.6 kW
(60,000 Btu/hr)
Open Symbols:
- Gas Analysis
Solid Symbols:
- Solids Analysis
0.25 0.5 0.75 1.0 1.25 1.5
Sorbent Residence Time (s)
Figure 6-15. Calcium utilization in the CTT firing Illinois coal at
SR = 0.6 injecting Linwood atmospheric hydrate at about
Ca/S =1.0.
-------
17.6 kW (60,000 Btu/hr)
23.4 kW (80,000 Btu/hr)
100 -
0.5 1.0 1.5 2.0
Residence Time (s)
100 -
0.5 1.0 1.5 2.0
Residence Time (s)
(\, Hydrogen
~ Nitrogen
A Sulfur
O Carbon
Figure 6-16. Species evolution from Illinois coal in the CTT at SR = 0.6.
-------
17.6 kW (60,000 Btu/hr)
23.4 kW (80,000 Btu/hr)
o>
i
ro
cn
3
oo
4->
O
4->
C
(D
(J
S_
-------
cr»
&
(D
S-
3
-M
CL
rd
O
s-
3
M-
=5
OO
4->
C
-------
One reason that the capture level appears to begin to plateau near 55 to
60 percent capture for the Linwood case is that a substantial portion of the
sulfur is not available for capture. Around 25 percent of the sulfur is
still bound to the char while approximately 10 percent of the gas phase
sulfur is in the form of SO2 (see Figure 6-17). When the sulfur capture is
calculated based only on the gas phase sulfur (along with the determination
of Ca/S), the capture levels are substantially higher, as shown in Figure
6-19. This representation demonstrates the importance of evolving as much of
the sulfur from the coal char as possible to make it available to the sorbent
for capture. Carbon burnout and sulfur evolution would be expected to be
near 100 percent in a practical precombustor system.
6.2.4 Effect of Injection Temperature
An important factor in application of this technology is defining a
range of operating temperatures which will accommodate the various process
requirements in an acceptable fashion. For instance, high slag fluidity and
carbon burnout in practical precombustor systems require very high
temperatures (up to 1400°C (2550°F)) while sorbent tends to become less
reactive above about 1200°C (2192°F) due to thermal sintering. With the
reactor at 23.4 kW (80,000 Btu/hr) and SR = 0.6, the injector location was
moved to produce different injection temperatures. Figure 6-20 shows that
the calcium utilization is optimum below about 1300°C (2372°F) and is still
quite high at a temperature of about 1420°C (2588°F). The optimum injection
temperature corresponds well to results on the ITR facility considering that
the optimum may be shifted slightly to higher temperatures on the CTT since
the sorbent is injected with cool carrier gas. The heating rate of the
sorbent injection stream in conjunction with the furnace quench rate produces
a lower peak sorbent particle temperature than the measured injection
temperature. This phenomena is not found on the ITR since the reactor is
i sothermal.
All of the results presented so far have shown that entrained flow
reactions between reactive CaO and H£S or COS can be fast (less than 400 ms)
and substantial (up to 60 percent calcium utilization for an atmospheric
6-27
-------
a>
i
ro
oo
CL
rO
O
S-
3
3
OO
c
-------
Injection Temperature (°C)
1200 1300
2000 2100 2200 2300 2400
Injection Temperature (°F)
2500
2600
Figure 6-20. Calcium utilization as a function of Linwood
atmospheric hydrate injection temperature in
the CTT firing Illinois coal at 23.4 kW
(80,000 Btu/hr), SR = 0.6.
6-29
-------
hydrate). It has also been shown (see Section 3-3) that CaS is not stable
when passing through oxidizing flame zones so that the reacted sorbent must
be removed from the fuel-rich zone of a precombustor. This would most easily
be done by drainage through the molten slag drain. The mixture of CaS with
molten coal ash presents a new set of critical issues:
• Is sulfur stable in molten mixtures of CaS and coal ash?
• If not stable, what is the rate of sulfur regeneration?
• How does the presence of increased calcium effect slag fluidity?
t Are there additives which can enhance the fluidity or sulfur
solubility of the slag?
These important issues are addressed in the following section.
6-30
-------
7.0 SLAG SULFUR CHEMISTRY RESULTS
In order to determine the impact of calcium-based sorbents and other
additives on slag sulfur solubility and fluidity, mixtures of Illinois coal
ash and additives were placed in the high temperature oven described in
Section 5.3 and exposed to a thermal and gaseous environment that simulated a
fuel-rich slagging precombustor.
7.1 Sulfur Solubility and Equilibrium
The equilibrium level of sulfur in a liquid slag when exposed to a fuel-
rich, sulfur bearing gas stream can be characterized by the "sulfur
capacity", defined by Fincham and Richardson (1954) as:
Cs = (wt % S) (po2/ps2) !/2
where p02 and ps2 are the partial pressures of oxygen and sulfur,
respectively. The value of Cs can be determined by equilibrating a slag at a
known temperature and interfacial gas composition. The amount of absorbed
(or released) sulfur can be determined by measure of sulfur concentration in
the slag before and after a test. Illinois coal ash and mixtures of CaO/coal
ash and Fe203/coal ash were tested at 1200°C (2192°F) and 1400°C (2552°F) in
gas environments that simulated the exhaust gases of an Illinois coal flame
at SR = 0.5. The synthetic exhaust gas mixing system was described in
Section 5.3 and its inlet composition is presented in Table 7-1.
A sample exposure time of at least 12 hours was established as the
minimum testing time to assure an equilibrium level of sulfur in the slag
sample. Figure 7-1 shows an example of the rate of sulfur gain for a mixture
of CaO and Illinois coal ash (Ca/S = 2.0) at 1400°C. The rate of sulfur
capture was extremely slow for three reasons: 1) the CaO was well mixed with
the coal ash so that a majority of the sorbent was not in ready contact with
the sulfur-laden gas stream, 2) the CaO used in these tests was much less
reactive than the freshly calcined CaO derived from injection of Ca(0H)2 or
CaC03 as in the entrained flow studies, and 3) the heating rate was probably
7-1
-------
TABLE 7-1. HTO FEED GAS COMPOSITIONS
Gas Composition
, Vol.
Percent
Stoichiometric Ratio
N2
CO
C02
h2s
0.5
68.25
26.35
4.90
0.50
0.7
74.96
13.35
11.19
0.50
7-2
-------
Exposure Time (HRS)
Figure 7-1. Slag exposure time requirements for Illinois coal ash/
CaO slag corresponding to a Ca/S = 2 at 1400°C, SR = 0.5.
-------
faster than the sulfur uptake rate, so that the sulfur capture mechanism was
predominantly between gas phase sulfur and molten slag rather than reaction
directly with CaO particles.
Values of log Cs are presented in Figure 7-2 versus temperature. Data
by DeYoung (1984) and Fincham and Richardson (1954) have been included to
show the consistency of EER results with the literature and to compare to
reported trends. Note that mixtures of Fe2C>3 and coal ash gave results
similar to the literature, giving validity to the testing and analysis
techniques used. Figure 7-3 presents log Cs as a function of molar basicity
for iron silicate mixtures to again show that the EER experimental system
produces trends similar to what has been presented in the literature.
High sulfur capacities were obtained with CaO/coal ash mixtures of high
molar basicity (defined by the ratio of total moles of base constituents to
acid constituents). However, the mixture fluidity was reduced at high
calcium contents. The mixtures of CaO and coal ash were established based on
what would be expected when burning Illinois coal at Ca/S molar ratios of
0.5, 1.0 and 2.0. Therefore, the most effective slag mixtures in terms of
sulfur retention are too refractory for slag drainage out of a practical
precombustor system.
The relationship between the sulfur capacity and temperature varied with
molar basicity, increasing with higher temperature for high molar basicity
but decreasing with higher temperature at low molar basicity. There is no
clear explanation for this behavior at this time.
By combining the equilibrium sulfur capacity of the slag mixture with
expected gas phase equilibrium composition and a mass balance on the amount
of slag that accumulates, it was possible to predict the equilibrium amount
of sulfur captured by the slag. Figure 7-4 presents families of curves that
correspond to varying slag sulfur capacities and amounts of accumulated slag
(in grams of slag per gram of injected coal). In other words, the curves
identify the required sulfur capacity to get a desired sulfur removal
percentage for a given ash-plus-additive slag mass.
7-4
-------
B Iron Silicate, molar basicity = 1.65, Fincham & Richardson (1954).
# Iron Silicate, molar basicity = 3.03, DeYoung (1984).
A Fe203/Coal Ash (Fe/S = 4.4), molar basicity = 3.01, EER.
~ Fe203/Coal Ash (Fe/S = 4.0), molar basicity = 2.75, EER.
fek Fe203/Coal Ash (Fe/S = 2.55 + 15.1% MgO), molar basicity = 3.08,
EER.
^ CaO/Coal Ash (Ca/S = 2), molar basicity = 2.77, EER.
^ CaO/Coal Ash (Ca/S = 1), molar basicity = 1.48, EER.
% CaO/Coal Ash (Ca/S = 0.5), molar basicity = 0.83, EER.
-1
-2
-3
I/)
CJ>
CD
O
-4
-5
-6
5 6 7 8
1/t x 104 (°r1)
Figure 7-2. Influence of temperature and molar basicity on sulfur
capacity for several slag mixtures (solid symbols -
melted, open symbols - not melted).
T( C)
7-5
-------
# Iron Silicate, DeYoung (1954), melted.
Q Iron Silicate, DeYoung, two phases.
¦ Iron Silicate, DeYoung, melted.
+ Fe/S = 4.4, EER, melted.
+ Fe/S = 4.0, EER, melted,
fc Fe/S = 2.55 + 15.1% MgO, EER, melted.
# Fe/S = 2.0, EER, melted.
£ Fe/S = 1.0, EER, melted.
Calculated Initial Molar Basicity
Figure 7-3. Sulfur capacity for iron silicate melts at
various initial slag molar basicities.
7-6
-------
# Ca/S = 0.5, molar basicity = 0.83, melted
~ Ca/S = 1.0, molar basicity = 1.48, not melted
O Ca/S = 2.0, molar basicity = 2.77, not melted
"O
o
E
0)
en
Z3
t
-------
The data points on Figures 7-4 and 7-5 correspond to Illinois coal ash
combined with different amounts of added CaO or Fe203. The greater the
amount of additives, the greater the relative slag mass. For this coal, no
additives correspond to the ash content of the coal - 9.5 percent or 0.095 g
slag/g coal. CaO addition of 23.5, 39.8 and 57.78 percent by weight
(totaling 26.89, 42.14 and 58.81 percent CaO in the slag mixture when
considering the CaO originally present in the ash) corresponds to burning the
coal at Ca/S molar ratios of 0.5, 1.0 and 2.0, respectively.
At best, the equilibrium level of sulfur held in the slag corresponds to
25 percent of the total input sulfur for CaO/coal ash mixtures. Note, also
that at a temperature of 1400°C (2252°F) the slag mixtures with higher CaO
content were not melted. Even without considering the kinetic limitations of
a slag sulfur capture process, large amounts of sulfur were not stable in the
slag and the required temperatures for complete slag fluidity were higher
than available under typical combustor operating conditions.
At first look, the addition of Fe203 as an additive looks promising. A
large amount of sulfur was captured and the slag remained molten at 1400°C.
However, there is no fast kinetic mechanism for getting sulfur into the slag
in conventional precombustor designs. Also, an excessive amount of Fe203
additive was required (a ratio of Fe/S = 4 or a weight percent of 82.98
percent corresponds to an addition of 3.85 times the total ash content of the
coal).
Researchers at Alcoa (Stewart et al ., 1984) have reported the
difficulties of capturing sulfur in a fuel-rich, slagging precombustor by
forming an iron-based slag layer. They reported sulfur capture levels below
15 percent when relying on a mixture of iron ore and sand as the slag
additive (at a level of approximately 25 percent of the coal feed) while
burning out only 68-87 percent of the carbon.
The preceding results demonstrate that using the slag layer as a
mechanism for sulfur capture and retention is not a promising scheme since
equilibrium solubility is low. Also, the problem of slag fluidity upon
7-8
-------
t Fe/S = 4.4, molar basicity = 2.75, melted.
A Fe/S = 4.0, molar basicity = 3.01, melted.
<3> Ca/S = 2.0, molar basicity = 2.77, not melted.
¦ Ca/S = 1.0, molar basicity = 1.48, melted.
TD
Q)
>
O
E
(1)
C£
S-
=3
ZS
CO
0.2 0.3 0.4
Slag Mass (g/g coal)
Figure 7-5. Equilibrium sulfur removal by slag for SR=0.5,
T=1400°C, and an Illinois coal with 3.97% sulfur.
7-9
-------
mixture with large amounts of calcium compounds has been identified as a
practical problem in slagging precombustors. Since the entrained flow
reaction of CaO and H2S/COS was found to be a promising sulfur capture
mechanism (see Section 6.0), the following subsections address issues that
would make the entrained flow process viable under slagging conditions, such
as: 1) means of improving slag fluidity for operation at lower temperature
or higher calcium contents, and 2) determination of the rate in which slag
desulfurization takes place when CaS comes into contact with molten slag.
7.2 Impact of Slag Fluxing Additives
The effect of CaO on the fluid temperature of Illinois coal ash is shown
graphically in Figure 7-6. The fluid temperatures were determined by the
ASTM ash fusion procedure. Small amounts of added CaO, up to levels
corresponding to coal firing at a molar ratio of Ca/S = 1, produced melting
temperatures below that of the raw ash. Beyond Ca/S = 1 the melting
temperature rose to higher and higher levels. However, Illinois coal is a
particularly high sulfur coal so that at Ca/S = 1 a relatively large amount
of calcium was added. Other coal slags may withstand much higher Ca/S levels
before reaching unacceptable fluid temperatures.
Four additives were investigated for their influence on CaO/coal ash
slags: Fe203, CaF2, B2O3 and P2O5. These additives were selected based on
their use in the glass and steel industries and by review of ternary phase
diagrams (Levin et al., 1979). For each additive, both the fluid temperature
and the sulfur capacity under reducing conditions were measured. Figure 7-7
and Table 7-2 present the results of these additive tests. The most
effective additive for reducing the slag fluid temperature for both Ca/S = 1
and Ca/S = 2 mixtures was boric oxide (B2O3), causing a drop in fluid
temperature of about 150°C (270°F) down to a temperature of 1055°C (1930°F)
by an addition of 10 percent by weight to a Ca/S = 1 slag mixture. Iron
oxide (Fe203) was the most effective additive in increasing the sulfur
capacity, changing log Cs from -3.95 to -3.83 when a 10 percent by weight
addition was made to a Ca/S = 1 slag mixture at 1400°C (2552°F). However,
7-10
-------
3000
2800
0)
s-
fd
&-
d>
Q-
E
Dolomite %
JL
X
o
o
a)
i-
ra
i-
O)
Q.
E
a)
1200
1.0 1.5
Ca/S Molar Ratio
2.0
2.5
J 1100
3.0
_L
X
42.14
58.81
CaO (wt%)
Figure 7-6. Fluid and T250 temperatures for Illinois coal ash/CaO mixtures.
-------
2800
2700 _ _ \ \ _
Solid Symbols
Ca/S = 2
2600 -
Maximum
Measurabl
Temperature
1S\
£
O
c
o
o
o>
c
X5
0)
oz
oj
j-
3
4->
rd
J-
«*
2100 -
Open Symbols
Ca/S = 1
\
\
\
\
\
N.
ZQ Cap2
Fe203
2000 -
1900 -
A.
"——A B2O3
1800
1200
10
s-
-------
TABLE 7-2. INFLUENCE OF SLAG ADDITIVES ON SULFUR CAPACITY
Ca/S = 1; log Cs determined at 1400°C, SR = 0.5
Additives
None
Fe203
CaF2
b2°3
P2O5
log Cs
(5% Additives by Wt.)
-3.95
-3.89
-3.81
-4.09
-4.09
log Cs
(10% Additives by Wt.)
-3.95
-3.83
-4.16
-4.32
-4.26
7-13
-------
this rise in Cs of 32 percent will not make a large difference in the total
sulfur solubility of the slag.
7.3 Slag Desulfurization Rates
So far, the most promising scenario for reducing sulfur emissions from
slagging precombustors is as follows: calcium-based sorbent is injected into
the precombustor to capture sulfur through entrained flow formation of CaS;
deposition of the CaS onto the molten slag covered walls; and rapid slag
rejection before the CaS can dissolve into the slag, releasing the sulfur
back into the gas phase. The issue that had not been investigated was the
rate at which the sulfur was regenerated.
The desulfurization rate of slags that hold a super-equilibrium amount
of sulfur were tested by replacing the CaO in previously tested slag mixtures
with CaS. The slag mixtures were then exposed to a 1400°C environment for
12, 30 and 60 minutes. A blended gas stream which simulated an equilibrium
coal flame effluent stream at SR = 0.5 was passed continuously over the
samples, just as was done in establishing the slag sulfur capacities reported
in Section 7.1. Figure 7-8 shows that the rate of desulfurization varies for
different slag mixtures. The desulfurization process was faster for cases in
which the CaS and ash mixtures were better mixed and for slag mixtures having
lower sulfur capacities. Notice that pure CaS was stable under these fuel-
rich, high temperature conditions.
Figure 7-9 shows the impact of fluxing additives on the desulfurization
of CaS/coal ash mixtures. An addition of 10 percent by weight of Fe203 had
little impact on a Ca/S = 1 slag mixture, while 10 percent additions of B2O3
to both Ca/S = 1 and 2 slag mixtures caused significant increases in the
desulfurization rate. It is apparent that mixtures with lower fluid
temperatures have higher rates of desulfurization.
The data presented in Figure 7-8 can be used to determine a rate
constant according to the desulfurization rate expression developed by
Turkdogan and Pearce (1963):
7-14
-------
O Ca/S = 0.5
~ Ca/S = 1; CaS and Illinois coal ash well mixed
O Ca/S = 1; CaS layered on Illinois coal ash
O Pure CaS
A Ca/S = 2
7-8. Desulfurization of CaS and Illinois coal ash slag
mixtures, showing the influence of calcium content
and mixedness (open symbols represent completely
unmelted samples, solid symbols correspond to
completely melted samples).
7-15
-------
A Ca/S = 2
O Ca/S = 1
~ Ca/S =2+10% B2O3
O Ca/S =1+10% Fe203
O Ca/S =1+10% B2O3
Figure 7-9. Desulfurization of CaS/Illinois coal ash/
additive slag mixtures (open symbols rep-
present completely unmelted samples, closed
symbols correspond to completely melted
samples).
7-16
-------
pC02
log [(^S)eqU1*-|. - %S] = 1 (k2)t + I>
(2.303)1
where PC02 is the partial pressure of carbon dioxide, 1 is the slag layer
thickness, k2 is the rate constant for the reaction:
S2- + C02 —1/2 S2 + CO + 02"
and I is an integration constant. Figure 7-10 shows the desulfurization data
on a semi logarithmic plot to allow the establishment of rate constants for
the three well-mixed slag mixtures in Figure 7-8. Taking the sample
thicknesses to be 1 mm and the partial pressure of C02 to be 0.049 atm, the
rate constant for each mixture can be calculated and is presented in Table
7-3. In an actual precombustor system the slag layer thickness may be large
enough to significantly slow the desulfurization process.
These results would indicate that fast slag drainage (less than 10
minutes) would be a desirable precombustor design feature. Another design
option would be to avoid the formation of slag altogether by operating at
lower temperatures and separating the ash and sorbent in a dry mode, but this
approach has several practical drawbacks and is incompatible with most
conventional slagging precombustors under development.
7.4 In Situ Sulfur Capture in Slag
The previous slag tests were carried out in a static oven facility in
which premixed mixtures of Illinois coal ash and calcium compounds were
exposed to a synthetic mixture of gases. A short test series was performed
to investigate sulfur stability under an actual coal-fired slagging
environment.
These tests were performed in the CTT at a firing rate of 23.4 kW
(80,000 Btu/hr) using Illinois coal at a stoichiometric ratio of 0.6,
identical to the test conditions used in establishing the entrained flow
calcium utilization levels due to sulfidation presented earlier. Char, ash
7-17
-------
O Ca/S = 0.5
~ Ca/S = 1
A Ca/S = 2
Residence Time (Minutes)
Figure 7-10. Desulfurization plot of CaS and Illinois
coal ash mixtures plotted to obtain the
rate constants presented in Table 7-3.
7-18
-------
TABLE 7-3. DESULFURIZATION RATE CONSTANTS FOR CaS/
ILLINOIS COAL ASH MIXTURES
Ca/S
Wt.S Ca
(as CaO)
mm
k2 ( )
min - atm
0.5
26.89
1.88
1.0
42.14
1.22
2.0
58.81
0.218
7-19
-------
and/or slag samples were collected by inserting two semicircular mullite
channels (3.8 cm (1-1/2 inch) in diameter) positioned adjacent to each other
across the 20.3 cm (8 inch) furnace duct. The concave portion of the ceramic
channels were pointed upwards. The channels provided a projected cross-
sectional area of 154.8 cm? (24 in^). The channels were inserted at a
distance of 86.4 cm (34 inches) downstream of the burner inlet. When sorbent
was introduced into the furnace it was injected at a distance of 21.6 cm (8.5
inches) above the ceramic channels.
Results of sampling in the CTT under similar conditions presented
earlier showed that approximately 75 percent of the carbon burned out at this
location in the furnace. By placing the ceramic channels across the flow, a
sample of char and ash gradually deposited. A layer of material about 1.3 cm
(1/2 inch) thick deposited after an operating time of approximately 10
minutes. The light weight, fine particle samples were easily disrupted upon
removal from the furnace by exhausting gases and thermal currents. To avoid
this, the char was burned off before extraction from the furnace, leaving a
solid ash deposit.
To simulate exposure to a fuel-rich, sulfur-laden exhaust stream without
adding more particulate matter to the sample, the furnace was quickly
switched from coal firing to an H2S-doped natural gas flame. It was found
that an exposure of from 2 to 3 minutes to the particle-free exhaust stream
was sufficient to burn off the remaining carbon and leave ash and sorbent
behind.
These results suggest that if the char material is deposited onto the
walls of a precombustor chamber at similar stoichiometry and temperature,
typical slag residence times would be sufficient for the remainder of the
carbon to evolve.
At the sample location, previous testing had determined that the calcium
utilization due to gas phase sulfidation reaction was on the order of 45
percent using Linwood atmospheric hydrate. By building up a slag sample on
the ceramic channels with sorbent present and exposing the sample long enough
7-20
-------
to evolve all the carbon, an in situ measure of sulfur stability in the ash/
slag was made.
First, tests were performed to establish required sampling conditions to
assure complete burnout of the carbon. Preliminary tests were run without
sorbent present and the following sampling procedure was established:
• 20 minutes firing Illinois coal
• 5 minutes firing H2S-doped natural gas
• 20 minutes firing Illinois coal
• 5 minutes firing H2S-doped natural gas
The sampling required two cycles in order to generate enough ash sample
for elemental analyses. Without sorbent present, the carbon level was only
0.30 percent by weight while sulfur was 0.03 percent by weight.
The same procedure was then followed while sorbent was being injected
(at Ca/S = 0.85). As expected from the slag fluidity testing, the samples
with sorbent addition were more fluid than the ash sample without sorbent
addition. However, there was a portion of the ash/sorbent sample that
deposited near the sample port entrance that did not appear completely
molten. This was due to cooler temperatures near the furnace wall and the
high thermal conductivity of the mullite channels. There was a sufficient
amount of the glass-like slag sample from the middle of the channels and the
more metallic-like sample near the furnace wall to be analyzed separately.
Elemental analyses of both samples were as follows:
Glass-Like
Metallic-Like
Carbon (wt%)
Hydrogen
Sulfur
Calcium as CaO
0.67
0.05
1.05
29.57
0.10
0.00
1.45
36.44
7-21
-------
The calcium utilization calculated for these two samples was 6.21 and
6.96 percent, respectively. These levels are far from the gas phase
utilization levels determined earlier in the program, indicating that the
sulfur regeneration observed in the static oven experiments also occurred in
the more realistic conditions of the CTT. Note that the utilizations were
fairly close to one other. This suggests that regeneration can take place in
ash/CaS mixtures that are not completely fluid or glass-like.
7-22
-------
8.0 QUALITY CONTROL EVALUATION
This section will summarize the results of the quality control/quality
assurance (QA/QC) effort that was carried out throughout the performance of
this program. Due to the nature of this program, it was classified as
requiring a category IV QA/QC program. Experimental systems and sampling and
analysis techniques were developed specifically for this program due to its
unique requirements. A detailed discussion of the system sampling control
measurements, sample analysis, calibration procedures and data reduction
procedures are presented in Appendix A.
8.1 Instrument Accuracy and Precision
Table 8-1 presents accuracy and precision summaries for the continuous
monitoring analyzers, the solids analysis instruments and techniques, and the
gas chromatograph used in this program.
8.2 System Uncertainties and Data Completeness
As a measure of the representativeness of the data and system
uncertainty, a series of sulfur balance experiments were performed on the ITR
and the CTT facilities. To two standard deviations the balances varied 27
percent on the ITR and 15 percent on the CTT. These variances were used as
combined uncertainty levels for the systems considering flow rate
fluctuations, sampling system integrity, and analysis accuracy. Data that
lay outside these uncertainty bounds were thrown out, resulting in data
completeness of 90 percent for the ITR and 93 percent for the CTT.
The calcium utilization data presented in Section 6.0 are plotted along
with approximate curve fits. All the data points lie within 27 percent and
15 percent of the plotted curves for the ITR and CTT data, respectively.
The relatively large uncertainty in the results is an indication of the
difficulties involved in measuring sulfur species under reducing conditions.
Several factors contributed to the ITR's increased uncertainty:
8-1
-------
TABLE 8-1. PRECISION AND ACCURACY FOR ANALYSIS INSTRUMENTS AND TECHNIQUES
Measurement
Precision [%)
Accuracy {%)
Goal
Actual
Goal
Actual
Continuous Monitoring:
N0X
2
0.3
20
12.3
CO
2
0.5
20
No Std.
co2
2
0.1
20
2.5
02
2
0.2
20
13.1
so2
2
0.3
20
10.2
Sorbent and Fuel Composition:
Carbon
2
0.24
2
0.8
Hydrogen
5
1.38
5
0.6
Nitrogen
10
2.13
10
1.9
Ash
10
0.33
10
1.42
Moisture
10
2.55
10
8.71
Sulfur
10
2.36
10
0.55
Calcium
10
1.0
10
0.2
Gas Chromatography:
H2S, S02> COS, CS2
None Est.
4.8
None Est.
6.1
8-2
-------
The ITR's small scale caused greater flow control uncertainty.
• The GC was located on a different site than the ITR, making sample
storage time longer for the ITR samples.
• The ITR testing was performed prior to the CTT testing so that
experience gained in sampling reduced sulfur species helped improve
data quality during the CTT testing.
The HTO was, for the most part, a static facility. That is, samples
were preloaded and sat stationary in the oven for long time intervals as
opposed to the dynamic injection and sampling tests performed on the ITR and
CTT. As a consequence, data uncertainty was dominated by the accuracy and
precision of the solids analysis instruments and techniques which are
presented in Table 8-1. Completeness for the HTO data was 100 percent.
8-3
-------
9.0 CONCLUSIONS
The investigations described in this report have been aimed at
understanding the sulfur capture process using calcium-based sorbents under
fuel-rich precombustor conditions and how the other precombustor processes
(such as carbon burnout, particulate emissions and N0X control) are impacted
by the sulfur capture process. The major findings of this program can be
summarized as follows:
• Under fuel-rich conditions, CaO reacts very quickly with H2S and to
a much higher level than CaO and SO2 under corresponding fuel-lean
conditions.
t At fuel-rich conditions that favor high carbon burnout (high
temperature, near stoichiometric), gas phase sulfur is present
mainly as SO2.
• Substantial concentrations of H2S can be generated when firing coal
under deep substoichiometric conditions (SR = 0.6).
• Sulfur capture is enhanced by injection of sorbent into lower
temperature regions (less than 1300°C (2372°F)), probably due to
less sorbent sintering when exposed to lower temperatures.
• Typical molten coal ash and mixtures of coal ash and CaO are
incapable of holding large amounts of sulfur in a coal precombustor
environment when at equilibrium.
• Large amounts of Fe203 can help reduce the melting temperature of
slag and increase the slag sulfur capacity but may be technically
difficult to implement and economically impractical.
• Coal ash/CaO slags rapidly desulfurize from super-equilibrium
levels of sulfur at precombustor temperatures and gas compositions.
9-1
-------
• B2O3 is an effective slag fluxing additive for coal ash CaO
mixtures.
The data collected within this program has indicated that there are
several imposing obstacles to overcome before sulfur control can be achieved
in conventional precombustor systems. However, there is much cause for
optimism that with additional work at the pilot scale a feasible precombustor
system with sulfur control can be designed. According to the information
generated within this program, new precombustor designs for the control of
sulfur emissions should promote the following:
• High levels of H2S and COS to feed the fast fuel-rich reaction with
CaO, potentially by deep substoichiometric operation and longer
residence times.
• Sorbent injection into lower temperature regions (i.e., avoidance
of flame fronts) of the combustor to help generate a very reactive
sorbent.
• Long fuel residence times to achieve high carbon burnout.
• The use of slag fluxing additives to extend the operable
temperature range within slagging precombustors.
• Slag drainage designs which promote quick removal of the sulfur
laden slag before extensive desulfurization can occur.
• Alternatively, low temperature operation with dry solids rejection
to avoid slag formation and slag desulfurization.
9-2
-------
10.0 RECOMMENDATIONS
The testing performed in this program has provided important process
information that can be used in applying fuel-rich sorbent injection to
existing precombustors and other combustion systems. The results have been
favorable enough to merit further process evaluation and investigation at the
pilot scale. It is recommended that a test sequence be performed on existing
precombustor systems (such as EER's 3 MW^ (10 x 10^ Btu/hr) VCC system).
Testing should include variations in stoichiometry, load (affecting
residence time), temperature, injection location, coal type, and sorbent
type. Also, use of slag fluxing additives and variations in slag drainage
design should be investigated.
Measurements to be made should include:
• Internal sulfur speciation;
• Downstream SO2 concentration;
• Calcium utilization (measured internally and by downstream SO2
change);
• Carbon burnout;
• Slag drainage consistency;
• Particle separation efficiency; and
• Slag and fly ash composition.
The potential exists that significant entrained flow sulfur capture
through calcium sulfidation can occur in precombustor systems while still
maintaining fluid slag and high carbon burnout. Of less certainty is the
fate of the CaS as it makes its exit from the precombustor. If the CaS
particles come into contact with a slag layer it may be drained fast enough
to avoid significant desulfurization. This process is very difficult to test
on bench scale facilities and is likely to vary substantially with
precombustor design.
10-1
-------
The information generated by this program also allows other full scale
testing on precombustor systems to be evaluated properly in terms of analysis
techniques and operating conditions. For some time there has been questions
as to the validity of some of the claims made by precombustor researchers -
claims made without proper substantiation. It is recommended that those
claims be evaluated in light of the results of this program and that any
further testing proceed in an informed fashion.
10-2
-------
11.0 ACKNOWLEDGMENTS
This program was supported by the United States Environmental Protection
Agency (EPA) through contract No. 68-02-3130 to Babcock and Wilcox (B&W) and
to Energy and Environmental Research (EER) under B&W subcontract No.
940962NR. Phase I of this program was under the supervision of Chuck Masser
at EPA and John Fairbanks at B&W. Phase II was under the guidance of Joe
McSorley at EPA and Peter Waanders at B&W.
The authors gratefully acknowledge those people at EER who have
contributed to the completion of this program, including: Mac McComis, Tony
Sciola, Brian Jacobs, Mike Hughes and Bill Cox for technical support; Mahin
Talebi, Tim Grogan and Jody Mclnnerny for analytical analyses; and Betty
Murphy, Chong Park and Valerie Kennedy for preparation of this text.
The review and guidance of Terry Johnson and Steve Lanier of EER and
Professor Jost Wendt of the University of Arizona is greatly appreciated.
In addition, the authors would like to thank Dave DeYoung of Alcoa
Research. Discussions with Dave were instrumental during the early planning
of the slag solubility testing.
11-1
-------
12.0 REFERENCES
Abraham, K. P. and F. D. Richardson. Journal of the Iron and Steel Institute,
Vol. 196, 1960. pp. 313-317.
Amstead, B. H., P. F. Ostwald and M. L. Begeman. Manufacturing Processes.
John Wiley and Sons, Inc., 1977. 739 pp.
Borgwardt, R. H., N. F. Roache and K. R. Bruce. Surface Area of Calcium Oxide
and Kinetics of Calcium Sulfide Formation. Environmental Progress,
Vol. 3, No. 2, 1984a. pp. 129-135.
Borgwardt, R. H., K. R. Bruce and j. Blake. "EPA Experimental Studies of the
Mechanisms of Sulfur Capture by Limestone." In: Proceedings: First
Joint Symposium on Dry SO2 and Simultaneous S02/N0x Control Technologies,
Volume 1, EPA-600/9-85-020a (NTIS PB85-232353), U.S. Environmental
Protection Agency, July 1985.
Carter, P. T. and T. G. Macfarlane. Thermodynamics of Slag Systems. Journal
of the Iron and Steel Institute, Vol. 185, 1957. pp. 54-62.
Case, P. L., L. Ho, W. D. Clark, E. Kan, D. W. Pershing, R. Payne, and M. P.
Heap. Testing of Wall-Fired Furnaces to Reduce Emissions of NO^ and S0x>
Volumes 1 and 2, EPA-600/7-85-026a and -026b (NTIS PB85-224632 and
-224640), U.S. Environmental Protection Agency, June 1985.
Cole, J. A., W. D. Clark, M. P. Heap, J. C. Kramlich, G. S. Samulesen, and
W. R. Seeker. "Fundamental Studies of Sorbent Calcination and Sulfation
for SO^ Control from Coal-Fired Boilers." EPA-600/7-85-027 (NTIS
PB85-221729), June 1985a.
12-1
-------
Cole, J. A., J. C. Kramlich, W. R. Seeker, M. P. Heap, and G. S. Samuelsen.
Activation and Reactivity of Calcareous Sorbents Toward Sulfur Dioxide.
Environmental Science and Technology, Vol. 19, No. 11, 1985b. pp.
1065-1072.
DeYoung, D. H. Sulfur Solubility in Slags for Cyclone Coal Combustors. Fuel
Chemistry Division, Vol. 29, No. 4. ACS Meeting, Philadelphia, PA,
1984. pp. 117-128.
Doremus, R. H. Glass Science. John Wiley and Sons, Inc., 1973.
Doyle, P. J. Glass Making Today. Portcullis Press, 1979.
Dykema, 0. W. Development of a Low N0x/S0x Burner. Proceedings: 1985
Symposium on Stationary Combustion N0X Control, Vol. 2,
EPA-600/9-86-021b (NTIS PB86-225059), July 1986.
EER. Kinetics of Nitrogen and Sulfur Reactions in Combustion Systems.
Quarterly Report No. 2. D0E/TC/70771-T2, 1986.
England, G. C., J. F. La Fond, and R. Payne. Coal-Fired Precombustors for
Simultaneous N0X, S0X and Particulate Control. Proceedings: 1985
Symposium on Stationary Combustion N0X Control, Vol. 2,
EPA-600/9-86-021b (NTIS PB86-225059), July 1986.
Fincham, C. J. B. and F. D. Richardson. The Behavior of Sulphur in Silicate
and Aluminate Melts. Proceedings Royal Society Series A223, 1954. pp.
40-62.
Herrera, 0. B., R. C. Sussman, J. S. Smith, and G. M. Taylor. Results of
Ladle Desul f urization of Plate Steel at ARMC0. Electric Furnace
Proceedings, Vol. 40, 1983. pp. 161-167.
12-2
-------
La Fond, J. F. and R. Payne. Development of a Coal Burning Vortex
Containment Combustor. Second Annual Pittsburgh Coal Conference, 1985.
pp. 785-793.
La Fond, J. F., T. R. Johnson, K. Kurucz, H. Nguyen, and R. Payne.
Development of a Vortex Containment Combustor: Pilot Scale Studies.
Quarterly Report No. 2. DOE Contract No. DE-AC22-85PC80256, Energy and
Environmental Research Corporation, Irvine, CA, 1985.
La Fond, J. F., A. Abele, T. R. Johnson, Y. Kwan, and R. Payne. Development
of a Vortex Containment Combustor: Pilot Scale Studies. Quarterly
Report No. 3. DOE Contract No. DE-AC22-85PC80256, Energy and
Environmental Research Corporation, Irvine, CA, 1986.
LeCrew, R. T. and D. J. White. The Burning of Coal-Water Slurry Fuels in a
Two-Stage Slagging Combustor. Proceedings: Gas Turbine Conference and
Exhibit, Houston, TX, 1985. ASME No. 85-GT-206.
Levin, E. M., C. R. Robbins, and H. F. McMurdie. "Phase Diagrams for
Ceramists," Reser, M. K., Ed., 1954: The American Ceramic Society,
Columbus, OH. pp. 219, 241.
Levin, E. M., C. R. Robbins, and H. F. McMurdie. "Phase Diagrams for
Ceramists, Volume II," M. K. Reser, Ed., 1969: The American Ceramic
Society, Columbus, OH, p. 120.
Mattsson, A. C. J. and J. 0. A. Stankevics. Development of Retrofit External
Slagging Coal Combustor System. Proceedings: Second Annual Pittsburgh
Coal Conference, Sept. 16-20, 1985. Pittsburgh, PA. University of
Pittsburgh and U.S. DOE Pittsburgh Energy Techn0logy Center.
Overmoe, B. J., J. M. McCarthy, S. L. Chen, W. R. Seeker, and D. W. Pershing.
Reactivity Study of SO2 Control with Atmospheric and Pressure Hydrated
Sorbents. EPA-600/7-86-040 (NTIS PB87-129250), November 1986.
12-3
-------
Pye, L. D., H. J. Stevens, and W. C. LaCourse. Introduction to Glass
Science. Plenum Press, 1972.
Sage, W. L. and J. B. Mcllroy. Combustion, 31(5), November, 1959. pp.
41-48.
St. Pierre, G. R. and J. Chipman. Transactions AIME, Vol. 206, 1956. pp.
1474-1483.
Silcox, G. D., S. L. Chen, W. D. Clark, J. C. Kramlich, J. F. La Fond, J. M.
McCarthy, D. W. Pershing, and W. R. Seeker. Status and Evaluation of
Calcitic SC>2 Capture: Analysis of Facilities Performance.
EPA-600/7-87-014 (NTIS PB87-194783), May 1986.
Stewart, P. L., Jr., L. E. Dostolfo, Jr. and D. H. DeYoung. Pulverized Coal
Firing of Aluminum Melting Furnaces. Final Report, DOE Contract DE-
AC07-78CS40037, (NTIS No. DE-84013093), 1984.
Sussman, R. C. and A. M. Smillie. Progress in Hot Metal and Steel De-
Sulfurization by Injection at Armco. Chinese Iron and Steel Society
Conference on Injection Metallurgy, Shanghai, China, 1982.
Turkdogan, E. T. and M. L. Pearce. Kinetics of Sulfur Reaction in Oxide
Melt-Gas Systems. Transactions of the Metallurgical Society of AIME,
Volume 227, August 1963. pp. 940-049.
Wilson, W. G. and A. McLean. Desulfurization of Iron and Steel and Sulfide
Shape Control. The Iron and Steel Society of AIME, Warrendale, PA,
1980.
Zauderer, B. and K. S. Fujimura. N0X Control in an Air Cooled Cyclone Coal
Combustor. Proceedings: 1985 Symposium on Stationary Combustion N0X
Control, Vol. 2, EPA-600/9-86-021b (NTIS PB86-225059), July 1986.
12-4
-------
APPENDIX A
QUALITY ASSURANCE/QUALITY CONTROL (QA/QC)
The collection and analysis of quality data has been of upmost importance
to this program. In order to ensure that the reported data is accurate, many
quality assurance measures have been taken in all phases of the program. This
program was classified as requiring a category IV QA/QC program due to its
exploratory research nature. Most of the procedures and measurements used for
data acquisition haven been standardized by EPA or ASTM and those which have
not are described in detail in this appendix.
Three facilities at EER were involved in obtaining the data presented in
this report; the ITR, CTT and HTO (discussed in detail in Section 5.0).
Unless stated otherwise, all of the quality assurance information presented in
this appendix applies to all three facilities.
A. 1 Experimental Systems
A.1.1 Gas Species Sampling
Exhaust gas-phase species composition was measured using the ITR and CTT
laboratory sampling systems. These systems were constructed in strict
conformance with EPA performance specifications and their use helped ensure
data quality. The species that were measured continuously included: NO, NO^,
SO^, CO, and CO2. Batch gas (GC) samples were collected for analysis of
H^S, COS, SO2, and CS^ by gas chromatography.
Figure A-l shows a schematic diagram of the continuous monitoring system
for NO, NO^, CO, CO2, and 0^ used on the CTT. The system was similar for both
the ITR and CTT. All components in contact with the sample were stainless
steel or Teflon to ensure sample integrity and corrosion resistance. The gas
sample was provided to the instruments by a sample conditioning system
consisting of a pump, moisture condenser and particulate filters.
A-l
-------
Sample Probe
Sample
M
Jacket
h2o
Glass
Wool
Filter
0 Q
Sample Pump
Recycle Valve
tHS1-
sr4e
Moisture Condensers
in Ice Bath
Bypass
Rotameter
Sample
Regu-
lator
Filter
co2/co
Anarad
(\R-600ri
CO Span
|NU-NUX
TECO
series
10
To
Exhaust
System
Sample
Manifold
-02
~ ~\
Low High
NO NO
Span Span
Vacuum Pump
^ Analysis Console j
Figure A-1. Continuous monitoring flue gas sampling system.
-------
Gas phase measurements were made routinely at the exhaust of the ITR and
CTT to assure steady operating conditions. Phase discrimination probes (see
Figure A-2) were used in both facilities for measurement of sulfur species in
the exhaust. The probes supplied two gas streams simultaneously. The main,
particle-laden stream was used for continuous monitoring of O2» CO, CO2, N0X,
and the nearly particle-free side stream was used for SO2 measurement and for
collection of batch samples for GC analysis.
Figure A-3 shows a diagram of the continuous monitoring system for
exhaust SO2 measurements. The SO2 gas sample in the nearly particulate-free
stream was transported in a heated Teflon line maintained above 100°C under
vacuum to a heated filter followed by a permeation tube drier. The
permeation tube drier selectively removed water vapor from the gas sample.
Dry air was used as the dry gas medium in the permeation drier. The sample
was then pressurized and delivered to the analyzer.
The same system was used for collection of samples for GC analysis.
Once again, the permeation tube drier was used to prevent contact between the
exhaust gases and condensated water. An investigation by Cassidy (1974)
showed that the tubing material was not permeable to either SO2 or H2S.
Tedlar sample bags were used to collect the gaseous sample and for storage
until analysis in the GC. A septum fitting on the sample bag allowed removal
of the necessary sample by syringe. Syringe injection was selected over use
of a flow valve for introduction of the sample to the GC to allow wider
calibration ranges and to minimize contacting with stainless steel to avoid
adsorption of sulfur species.
The tedlar bags were leak checked before testing began. After several
usages some bags were filled with pure N2 and submitted for GC analysis as
blanks. All of these cases failed to show measurable levels of sulfur
species. Besides testing the GC with a blank sample, this also indicated
that negligible amounts of sulfur species were absorbing onto the bags and
released later.
A-3
-------
Dimensions: Inches
Figure A-2. Phase discrimination probe tip.
-------
Phase Discrimination Probe
Figure A-3'. SO^ continuous monitoring and batch sampling system.
-------
Figure A-4 shows results of a test to establish the maximum storage time
in a sample bag before analysis. A conservative value of two hours was used
throughout the testing on the CTT and four hours on the ITR (the ITR is
located at a different location than the GC and the CTT).
A.1.2 Solids Sampling
Solids were collected in the ITR facility with the system shown
schematically in Figure A-5 and in the CTT facility with the system shown in
Figure A-6. Both systems were operated with the capability of monitoring the
gas sample temperature to assure that no condensation occurred within the
probe or filter housing. To test whether additional sulfur was captured by
the sampled sorbent while sitting on the sample filter, an H2S laden, fuel-
rich gas stream was pulled over a preloaded precalcined sorbent (with a
surface area of 32 m^/g) and a raw calcium hydroxide sample in the solids
sampling probe. For both the ITR and CTT, less than one percent of the
calcium was utilized through reaction with sulfur when tested at typical
sample duration times. Each solids analysis was performed using all the
collected sample to avoid the question of sample representativeness.
A.1.3 Temperature Measurements
High temperature gas measurements within furnaces are subject to large
inaccuracies due to the effects of thermal couple radiation loss. Therefore,
in-furnace gas temperatures were measured using a fine-wire thermo-couple on
the ITR and a suction pyrometer on the CTT.
The fine-wire thermocouple was made with 0.025 mm diameter type-S
thermocouple wire, butt-welded across a forked support. Temperature readings
in degrees Celsius were obtained with an electronic readout containing an
internal icepoint reference. The specified accuracy of the readout was +
2.3°C. Figure A-7 presents a diagram of the fine-wire thermocouple.
A-6
-------
Q.
Q_
C
o
•p—
+->
<0
s-
4J
c
0)
o
c
o
o
_ 6000
5000
E
4000 £
c
o
-p
rtf
S-
~ h2s
O cos
A so2
O CS2
- 3000
- 2000
1000
C
Q)
O
C
o
o
40 60
Storage Time (Hrs)
Figure A-4. Depletion of sulfur species within a tedlar
sample bag as a function of storage time.
A-7
-------
Water
Out In
Cooling Mater
Figure A-5.
Isokinetic solids sampling probe with upper and lower
cooling jackets used in the ITR.
-------
Sample
Stream
1/2" Stainless Steel Probe
3»
10
Compressor
Air
Figure A-6. Solid sampling system used on the CTT.
-------
Figure A-7. Fine-wire thermocouple probe for gas temperature
measurements in the ITR.
A-10
-------
The fine wire thermocouple was utilized in order to minimize the
uncertainties of radiation corrections. Radiation corrections to the
thermocouple readings were determined by balancing the equations for heat
transfer away from the thermocouple wire:
Pconv = ty) A (T-tc " Tgas) (A—1)
Qrad = - T^env) (A—2)
where:
^conv = convective heat loss
Qrad = radiative heat loss
h0 = convective heat transfer coefficient
A = surface area per unit length of t.c. wire
Twire = t.c. wire temperature, K
Tgas = 9as temperature, K
Tenv = temperature of surroundings, K
e = emissivity of platinum, e= 0.18
o = Stefan-Boltzman constant
The heat transfer coefficient was determined by the equation:
(A-3)
l^o D0 /Cp^A®*^
~ = \kTf
/PV D0\°-52'
0.35 + 0.56 y—J
where:
D0 = t.c. wire diameter
kf = gas thermal conductivity
CpM
— = Prandtl number, 0.7
kf
PV D0
= Reynolds number, 0.1
A-11
-------
Solving these equations yielded h0 = 4.4 x 10-2 cal •cm_2*s-l*K-l. Setting
equations A-l and A-2 equal to each other:
(T^wire " ^env^
Tgas = Vire + r;—^ (A-4)
1.79 x lOH
where Tenv is taken as the wall temperature of the ITR. The correction is
significant (>5 K) only when the thermocouple temperature differs from the
ITR temperature by more than about 75 K. Which was never the case during
this test sequence.
Radial temperature gradients in the ITR were small while operating under
isothermal conditions. Table A-l shows how the temperature varied from the
center of the ITR at 900°C under fuel-lean conditions. The reactor was
divided into four quadrants which were labeled by their closest compass
points: N, S, E, and W. The temperature was measured "near the wall" in
each direction away from the ITR axis. In all but two cases, duplicate
measurements were made.
The temperature of the ITR itself was monitored by a programmable
control unit using four type R (Pt/Pt-13 percent Rh) thermocouples. These
thermocouples are normally stable and were at no time subjected to the
reactive flame gases. Therefore, they were not checked for deterioration
during the course of the experimental work.
The suction pyrometer system used on the CTT is presented in Figure A-8.
The suction pyrometer consisted of a high-temperature thermocouple (B type)
in a porous ceramic radiation shield. The pyrometer was placed in the gas
stream with the opening in the alumina sheath pointing downstream. A high
flow rate of furnace gas was drawn through the sheath and over the
thermocouple to increase convective heat transfer to the thermocouple and
reduce radiation loss. As the gas flow rate was increased, the gas
temperature increased to a constant value, indicating no additional reduction
in radiation loss. Figure A-9 shows the results of the ejector pressure (a
measure of flow rate through the pyrometer) calibration for the CTT. An
A-12
-------
TABLE A-l. RADIAL TEMPERATURE CHANGE IN THE ITR AT A NOMINAL 900°C
ISOTHERMAL CONDITION
Distance
from
Temperature,
°C
Burner,
cm
Axis
N
S
E
W
107
897
895
895
899
895
94
909
910
911
910
910
905
906
904
904
900
81
904
907
909
906
910
903
905
906
902
905
69
903
902
908
904
906
904
906
911
910
913
56
903
902
912
901
907
909
908
916
913
912
43
903
902
910
909
906
902
904
906
903
909
30
900
898
896
902
898
A-13
-------
Temperature
Readout
Regulator Ejector
Figure A-8. Gas temperature measurement system for the CTT.
-------
Motive Side Pressure (psig)
Figure A-9. Suction pyrometer aspirator pressure
calibration for CTT.
A-15
-------
operating pressure of 40 psig was adopted for all the temperature
measurements for the CTT system.
A.1.4 Gas Flow Metering
All gas flow rates were measured by suitably sized rotameters. The
rotameters were calibrated using air with venturi and laminar flow element
flow meters. Flow rates were corrected for pressure and specific gravity
(relative to air) using the formula:
Q = true flow rate
Q° = measured flow rate
P = outlet rotameter pressure, kPa
S.G. = gas specific gravity
H2S dopant flow rate was compared to the measured SO2 concentration at
the exhaust when operating fuel-lean to assure a proper initial dopant feed
rate.
A.1.5 Sorbent Feed Rate
The sorbent feed rate was measured by feeding sorbent directly into a
tared sample jar and weighing the accumulated sorbent after a measured
discharge time. This was done before and after each test to assure a steady
sorbent feed. If the feed rate calibration varied more than 10 percent from
before to after, the test was repeated. Typically sorbent feed rates
fluctuated less than five percent.
(A-5)
where:
A-16
-------
A. 2 Analytical
A.2.1 Continuous Gas Species Analysis
Table A-2 lists the specific instruments and full-scale ranges for the
continuous monitoring system. The instrument ranges were selected to exceed
the maximum expected concentrations by a small margin.
Table A-3 gives the precision, accuracy, and completeness information
for the continuous monitoring instruments used on the ITR and CTT. In every
case, the program goals for precision and accuracy were met.
A.2.2 Gas Chromatography Analysis
Gas chromatograph {GC) analysis of fuel-rich exhaust gases for
measurement of H2S, COS, SO2 and CS2 was performed with a Perkin Elmer 2
gas chromatograph using a Supelco Chromasil 310 column and a flame
photometric detector.
The GC was checked at least twice daily with span gases of known specie
concentration. Figure A-10 presents the variation of this standard check
over a single calibration period. The precision and accuracy of the data (to
two standard deviations) was established to be 4.8, and 6.1 percent,
respectively for the GC analyses.
A.2.3 Solids Analysis
The composition of solid samples was generally determined by EER's
Analytical lab though some samples were sent to an independent laboratory.
Carbon, hydrogen and nitrogen were analyzed on a Perkin-Elmer 240B and solid
sulfur species were determined from a Leco SC32 Analyzer. Calcium analysis
was done with a Chelometric titration method which is covered under ASTM D
2795. Table A-4 gives the precision and accuracy data for these instruments
and methods.
A-17
-------
TABLE A-2. CTT INSTRUMENTATION
Gas Measured
Type of Instrument
Manufacturer
Model No.
Range
Reference
NO, NOx
Chemiluminescent
Thermo Electron Corp. (TECO)
10 Ar
1000 RPM
Teco Model 10 Operating Manual
CO
Nondispersive Infrared
(NDIR)
Anarad, Inc.
AR600R
10 Percent
Anarad AR 600 Operating Manual
C02
Nondispersive Infrared
(NDIR)
Anarad, Inc.
AR600R
40 Percent
Anarad AR 600 Operating Manual
°2
Paramagenetic
Taylor Servumex
0A250
100 Percent
Taylor 0A25O Operating Manual
S02
UV Absorption
DuPont
400
5000 ppm
DuPont 400 Operating Manual
-------
TABLE A-3. PRECISION, ACCURACY AND COMPLETENESS DATA FOR
CONTINUOUS MONITORING GAS SPECIES INSTRUMENTS
Gas Measured
Precision (%)
Accuracy (%)
Completeness {%)
Goal
Actual
Goal
Actual
Goal
Actual
NOx
(Chemi1umi nescent)
2
0.3
20
12.3
90
100
CO
(NDIR)
2
0.5
20
No
Std.
90
100
C02
(NDIR)
2
0.1
20
2.5
90
100
02
(Paramagentic)
2
0.2
20
13.1
90
100
S02
(UV Absorption)
2
0.3
20
10.2
90
100
-------
D.
CL
C
o
rO
S-
C
-------
TABLE A-4. SOLID ANALYSES PRECISION AND ACCURACY DATA
Measurement
Precision (%)
Accuracy [%)
Actual
Goal
Actual
Goal
Carbon
CSJ
•
2
0.8
2
Hydrogen
1.38
5
0.6
5
Nitrogen Perkin-Elmer 240 B
2.13
10
1.9
10
Ash
0.33
10
1.42
10
Moisture
2.55
10
8.71
10
Sulfur - Leco SC32
2.36
10
0.55
10
Calcium - ASTM D2795
1.0
10
0.2
10
A-21
-------
Ash fusion temperatures were determined by an independent laboratory
(determined by competitive evaluation between three laboratories) with a
method covered under ASTM D 1857. The laboratory participates in a monthly
round robin quality control service. The fluid temperature precision and
accuracy were determined to be 0.6 and 2.1 percent, respectively for the
months in which samples from this program were analyzed by the lab.
Slag samples were first pulverized before analysis in the sulfur
analyzer. An oxidizing additive (vanadium pentoxide) was added to the test
slags as recommended by Leco to assure that all the sulfur evolved from the
slag.
A.2.4 Data Representativeness and Completeness
The previous subsections have discussed instrument accuracy and
precision. As a measure of total uncertainty in the data associated not only
with instrument uncertainty but also system fluctuations, sulfur balances
were made for each test condition. System uncertainties included dopant
sulfur feed rate; coal, air and sorbent feed rates; sample train leakage;
sample bag leakage; instrument reading fluctuations; etc. Also, data
completeness has been determined based on the number of reported data points
compared to the number of data points taken (excluding all data that was
rejected due to identified feeding, sampling or analysis problems).
Sulfur balances on the ITR varied 27 percent (two standard deviations),
and completeness of the data was 90 percent. On the CTT, sulfur balances
when firing coal were within 15 percent (again, two standard deviations) and
completeness was 93 percent. The data presented in Sections 6.0 all lie
within these sulfur balance uncertainties. That is, the curves drawn through
the data points are within 27 and 15 percent of the plotted data points for
the ITR and CTT, respectively. Data points that laid outside of these
established uncertainty ranges were thrown out and contribute to the reported
completeness levels of 90 and 93 percent.
A-22
-------
The HTO facility system uncertainty was very low since the samples were
preloaded and sat stationary for a long furnace exposure time. The largest
uncertainties were associated with the solids sampling accuracy and precision
reported in Section A.2.3. Completeness of the HTO data was 100 percent.
A.3 Calibration Procedures
Calibration procedures and frequency for all standard measurement
systems are listed in Table A-5. As shown in the table, standard calibration
procedures were used for each system. Each system was calibrated at the
frequency shown in the table to ensure that the accuracy of the measurement
could be traceable to the calibration standards.
Each of the continuous monitoring instruments was calibrated with three
gases: a zero gas and two span gases. Table A-6 lists the span gases used
to calibrate instruments for the CTT sample system.
Calibration of the GC was performed by evacuating an 11 liter glass
chamber, injecting a known quantity of pure compound (H2S, SO2, COS or CS2)»
and filling up the remainder of the chamber with ambient air. Then, by
injecting different volumes of gas into the GC, the detector response was
recorded. In this way, the optimum instrument sensitivity could be achieved
for each sulfur specie. This was repeated at several concentrations ranging
up to 5000 ppm for H2S and SO2, 1500 ppm for COS, and 1000 ppm for CS2-
Figures A-11 through A-14 show the calibration curves used for the majority
of the data presented in this report.
Calibrations were performed immediately before each major testing
sequence and were checked daily to assure accuracy.
Figures A-15 through A-17 present examples of the gas chromatograph
readout. The peaks are sharp and clean. The flattened peaks of Figure A-15
are due to the attenuation setting of the printout equipment and no.t due to
oversaturation of the sulfur detector. The integrated values for the peaks
A-23
-------
TABLE A-5. CALIBRATION PROCEDURES
Measurement Parameter
Calibration
Procedure
Calibration
Frequency
Calibration
Standard
Reference
Fuel Composition:
\
Carbon
Hydrogen
j
\ Known quantity of
/ 100% acetanilide
Ni trogen
Ash
> Standard Samples
Daily
)
Perkin Elmer 240 B
Operating Manual
Moisture
J
Sulfur
Standard Sample
Daily
Certified sulfur
standards
Leco SC 32
Operating Manual
Exhaust Gas Composition
NO, N02
\
02
1
CO
C02
S02
f Compare to stan-
f dard compressed
( gas
Daily
EPA Protocol 1
Gases or custom
gas blends. See
Section 10.0
EPA EMSL Protocol
No. 1, June 1, 1978
Hydrocarbons
;
Sorbent Surface Area
a
Daily
Known quantity of
n2
Quantasorb Surface
Area Analyzer
Operating Manual
Sorbent Flowrate
Weigh feed over
measured time with
independent balance
Before and after
each run
Standard weights
-------
TABLE A-6. CALIBRATION GASES
Gas
Concentration
Percent of
Full-Scale
Balance
Gas
Grade
NO
836 ppm
84
Nitrogen
Cert
fied
Standard
NO
166 ppm
16
Ni trogen
Cert
fied
Standard
S02
2915 ppm
73
Nitrogen
Cert
fied
Standard
S02
437 ppm
11
Nitrogen
Cert
fied
Standard
CO
1227 ppm
82
Nitrogen
Cert
fied
Standard
C02
12.3%
31
Nitrogen
Cert
fied
Standard
C02
5%
13
Nitrogen
Cert
fied
Standard
02
10%
10
Ni trogen
Cert
fied
Standard
02
4%
4
Nitrogen
Cert
fied
Standard
CO
10%
100
Nitrogen
Cert
fied
Standard
CO
2%
20
Nitrogen
Cert
fied
Standard
A-25
-------
900
800
700
600
500
400
300
200
100
0
0.
H2S Concentration (ppm)
Figure A-11. H2S calibration curves for 70, 50 and 30 pil syringe injection
-------
SO2 Concentration (ppm)
Figure A-12. SO2 calibration curves for 70, 50 and 30 jul syringe injection (5/86).
-------
COS Concentration (ppm)
Figure A-13. COS calibration curve for 70 pi 1 syringe injection (5/86).
-------
CS2 Concentration (ppm)
Figure A-14. CS2 calibration curve for 70 /il syringe injection (5/86).
-------
pun
262 SULFUR SAMPLES
SENSITIVITIES
Ii4.
r:r?
50
4
5
4
&
4
cos
)
1
a !i-
*)
cs
2. 76
Hi5
Figure A-15. Gas chromatograph printout example.
-------
INST 1 METH 10 FILE
29
CO
RUN
271 SULFUR SAMPLES
SENSITIUITIES
"Hi!
¥:§8
3 2,2T
2.78
4.59
-¦END
50
G9S
r«9
H^S
1. 64
Figure A-16. Gas chromatograph printout example.
-------
INST 1 flETH 10 FILE 27
RUN
273 SULFUR SRMPLES
SENSITIVITIES
"^If^ Wx *§
Figure A-17. Gas chromatograph printout example.
-------
were not impacted by the clipping of the printed peaks. Identification of
the sulfur species associated with each peak was easily achieved.
A.4 Data Reduction
Equations used to calculate results for each of the measurements are
discussed below. Many of the results are obtained directly from the
measurements and thus do not require calculations. The main results which
required calculation for presentation included calcium utilization by capture
of sulfur species, sulfur capacity of slags, equilibrium capture of sulfur in
slags, and furnace residence time.
A.4.1 Calcium Utilization
The efficiency of the sulfur capture process has been presented in most
cases in this program by calculation of the calcium utilization. This has
been determined through both gas phase measurements and solids analysis
results on the ITR and CTT facilities.
Gas phase determination of the calcium utilization was performed by
measurement of the molar change in total sulfur compared to the molar rate of
calcium addition:
[S]n- - [S]c [S]T
Calcium Utilization(totai s) = — * f—=r x (A~6)
where:
[S]-j = the molar sulfur concentration measured before injection of
sorbent at location x,
[S]c = the molar sulfur concentration measured during injection of
sorbent at location x,
A-33
-------
[S]j = the equivalent total molar sulfur concentration at location x
(CS]j = [S]-,- for gas doping cases, but [S]j is greater than [S]-j
for coal firing if not all the sulfur evolves from the coal),
[Ca] = the equivalent total molar calcium concentration based on the
sorbent feed rate.
Nearly all the tests performed in this program were run at a Ca/S molar
ratio near 1.0. Higher Ca/S ratios would result in lower calcium
utilizations due to reduced sulfur availability.
For cases where coal was fired, not all the sulfur evolved from the
coal. In this case, equation A-6 would result in a lower calculated sulfur
capture than when only the evolved gas phase sulfur is considered. When
based on the gas phase sulfur, the calcium utilization becomes:
[S]n- - [S]c [S]i
Calcium Utilization(gas phase S) = —PTI ' x 100% (A~7)
LSJ-j LCaj
Figures A-18 through A-19 show the difference in sulfur capture results when
calculated on these two different bases.
Solid sample determination of calcium utilization is simply the molar
ratio of the measured sulfur to calcium:
S MWCa0
Calcium Utilization = — • (A-8)
MWS CaO
where,
S = weight percent of sulfur in the solid,
CaO = weight percent of calcium oxide in the solid,
MWS = molecular weight of sulfur, 32,
MWCaO = molecular weight of calcium oxide, 56.
A-34
-------
For cases in which coal was fired, a distinction must be made between
sulfur bound to char in the solid sample and sulfur which is present as CaS.
A calculation procedure was developed to account for the char sulfur by
considering the extent of sulfur evolution from a sample without sorbent
injection. Equation A-8 can be used with the substitution for the term below
of S in determination of the calcium utilization:
Ss " C(1 " T5o) (A*-Ca0) (A~)]
SCa = (A—9)
SE Sc
[1 - 2.5 (1 - —) (-)]
100 Ac
where,
Ac = wt % ash in the original coal,
As = wt % ash in the sample reported by 240 B analysis, which
includes coal ash, CaO and CaS04,
CaO = wt % of Ca as CaO in sample,
MWCaO = molecular weight of calcium oxide, 56,
MWs = molecular weight of sulfur, 32,
Sc = wt % sulfur in the original coal,
Sq3 = wt % sulfur associated with the calcium in the sample,
Ss = wt % sulfur in the sample,
SE = sulfur evolution based on analysis of a solid sample without
sorbent present.
A.4.2 Sulfur Capacity
As presented in Section 7.1, the sulfur capacity is defined as:
/P02\1/2
Sulfur Capacity = Cs = (S) —J (A-10)
A-35
-------
where,
S = weight percent of total sulfur in the slag,
P02 = partial pressure of oxygen, and
PS2 = partial pressure of sulfur above the slag layer.
p02 and PS2 are determined by equilibrium calculations using the test
condition inlet gas mixture and temperature. The weight percent sulfur was
determined by solids analysis after the slag reached equilibrium.
A.4.3 Equilibrium Sulfur Capture by Slag
It was stated earlier that the sulfur capacity can be used to calculate
the amount of sulfur capture in an actual coal combustion system. First, the
equilibrium partial pressures of oxygen and sulfur are determined for a
number of gas phase sulfur concentrations (simulating different amounts of
sulfur capture) at a given temperature and combustion stoichiometry. Then,
for a given slag mixture (determined by the Ca/S ratio) the corresponding
weight percent sulfur is calculated using the experimentally obtained sulfur
capacity. A mass balance of sulfur can then be performed to determine the
actual percentage of gas phase sulfur that would be removed.
A.4.4 Furnace Residence Time
Residence times are calculated by assuming one-dimensional flow in the
furnace. Based on the measured temperatures, the bulk mean residence time of
a specified furnace zone (r2) was calculated by using the following equation:
r 72 T° , %
2 = t (A-U)
V T|Tl2
where:
V2 = volume of furnace zone 2, ft^,
Y = volumetric flow rate of gas, SCFS,
A-36
-------
T0 = 293°K,
^2 = bulk mean gas temperature
= 1/2 (Ti + T0)2.
T-j and T0 are center line temperatures measured at the inlet and the exit of
the specific furnace zone, respectively.
A-37
-------
APPENDIX B
EXAMPLES OF RAW DATA
B-1
-------
B.1 ITR Data Example
B-2
-------
94
/r? 5
fxj€t £ '"c
¦£jvr£cT~£6l /Jo'v+J SfreT**^,
C^JH£rc'A( l'~«JU/OO0 MydCt'tz'.
T~M /^S M). /oloT?
jOq1"AZ< £nr6T/! |A\
Jst J.
pf - Z'Q-
//vk£=
rttjrAJdfl $r~£L
jPr'^tiry : 20 c/^N
Se^trj ^2—
y?**s/0ezi- . 6
-£L
M<
T
^i£j£a*± bostz/2- L/j?s
T5" ;
&, &y>/ti. / ^CA
Z*£ ^"'4?
Sxyfe- . 7s~y j
3
f-£<$d fab
/"¦ a/
H —r-
B-3
-------
B-4
-------
\
X
a
s
5 j
i I i«
U'i
III!
? *
$ * § ^
V tv ft
8"^
i a & a
^ V!
§ $ % $
Vts ^ £ N
^ «J ^ «!
Cos
r/w
\P ft ^ JC1
V ^0 $ §
Hi
r
$
Vt (V JO
-------
&
V
•K
5
'N
-•v.
B-6
-------
CTT Data Example
-------
(p /2)9(f : _C]]__ *Y\\dk ?<^ ^ - ( 2. /O sr~ rf
W* nZ Deyf'Vj » none- jr-''
hjfi ¦
QxeUfcreQ - "/~
Si^mes* * f^<^rvr'^eS) Fue/' ^2~//,-**&
i ^
/-?S -h I (j-/l/no~*, /f-ff r (-I'T 2 )
C f u>e/r ~7U jofhe*^ rf&P ^oc,/! &J$o,
G> CL r~&-*7 srh&o-hfll^. ~77e •£>*<. /-^ p*-/-'7
Z"1' '1 ^ 1 r\ /Vie- /C~t-—> ^v> C^'O- / ^ '^^£-0 cuo^.
n^p*+J^Qm J J
7^tf"A>Lj l^>&5 £h/le- cS^A ^ <£>/C-
UCv'C
•I
Ic-st <2. (3T/U?*, LrKuja^ ///, /^f- H)
"^FTe. -PJt) t~>e-!/. ~7^< <~>*//' *£*, V^<.
5^ymly,
USH^r-^ /44=^-,
T> Qe.~/c*~ih*_ /)yy ,
^5 A%- / ^,-«.
B-8
-------
fc
a
dOr,
( fortEA*;
%o?
i err
^zT>
f« /i/2
D; !*+>*'t
Re&d*.
%
Uh-*^
* £&x)
(
i*L£
r~ "-1?
C. r9
%?5Z>JZ
c
c-c&ir .,,
Hi ^
^/75^
/-gy
^3f
££tG) ^
M.2
3-S
/-2.S
•?-3f3
&A
AS? (iVt
?//7^5
£<^C? 'T
^ ,>,uJo cD /Ko
J-f
H-f
3.Z
t^y
*// 752>
7
& tr
dL
^inszti
<%
x^t;
{rtT5
&f )
233 J)'F - (
"Y~oiK
il3'f
K77
SM
B-9
-------
<*/2/& C/ )" •
/est (^U~T f/[W^-7 y /_iVt KJ6t£) Ah{ ^ fjrT +3- \
' S"' ti-<--
).zs
y^r
S,TO
h3
7&>
J2.3^
Vf*"
3-?^
oAT
urr
/^>,3
353
^3-«7
J-"Z.~
fl- Pr5"3"
5^ -- 2 5~3^>
-5/. £,
- 2520
f^kr(P
5^z =¦
<£. 3'2
,5^2- "" 25~2>V
^&e£U*ica5)
hx
- ¦2"e7'3/^ ^
H.n
ia*se&£) -. rn
H, n
Lj,d-f
•V
c
3r^5/^,V
B-10
-------
B-11
-------
I/)
¦ h
I
£
O
c J
<1$
I
cV"
u
Cx?
§
X*
£
l~
<*\
b
¦2.
¦8
uj
£yr
f5
TT
Z*
or
P
<
6^
^r-
rv^
O
' CS»
eo
V^
J
*
V \> v'
> ^
d\
^ /t
irOA.
^r-
¦^o
e^-
&
-J
M
l^N
L V ^ \v ^
-c
-SL
V ^
^r.
vT\
e<
^5
ri
r\
rV
¦_>
CnTS;
~T.
V X£
-r;. ^ . \
c^v
Vj"-
*r\
X-
-N.
£
.csry
Vr:
rx
¦sJ-
04
VP
\ \fe
\\ ?r\
\W x-
*v- r\.
X
j^p>'
'7*~
rf\
cO
<5o
e\>
B-12
-------
IX-+«- : le / X I
&sn&i~-hvrx3 - I~/lrx^s> Fil - roecrz, 67^/Kr- *g, -' £>¦ C=>
/fK Tr/yj '¦ ^HGC>aF {fir*-*)
Sasnfle *
cds
MzS
CS-.
Sc>
7Z-uf
C«.
Ys
Z c*p
ft^
°J'$
g*L>Jj *«¦
c^s-.bi?
tirrb !S~
)!75~C> (
// 3 /
2.ZSIO
2-^r
6>C>
^030
?LrTf j5"
)\1SZ>3
3 7/
h/
SN /
A2^
Ljl.d
&*-+- /-5"
//732>3
//??
<23
£"Sz>
mio
'(?*+ I5~
)I7S2?Y
Piri-
// 7
l^+T-F"
?3H
) SLO
5"7?
2
mo
/&
(O
?o
3 to
$13
351a
L%£>
$v.a--
?,o.
Jto.S
B-13
-------
.4,» ".
CcrrJ£Lz
Sajnffe *
)
- U./?.!'
+n>n3 1
cds
yC,
£^lh'rt*rC5
^5
/4/r
C52-
, ? $OOOc
! x/vj
3 fc7W/^r
- 2[iS~F
TZ-ut
50?,
/f
-C3-C,
»' +
Z Cy
fk+.9<
B+jJ) *
o/$*u
>
&
//75z>7
St> /
33 ?>p
HI 7^
Pirrt- /.?
7/7
3/7
//7JT
-?
60
/^5-f
L>2rC?
uc
--
'
!¦
B-14
-------
LAD SAMPLE PROJECT HO. CHARGE NO.
TRACKING REPORT , ^KINFFR
«™Est m.y# J mmK,
X t>° ||V W vX™
SAMPLE*
1 NO.
SAMPLE
DESCRIPTION
ANALYSIS REQUIRED
SAMPLE
PREPARATION
DATE
DELIVERED
DATE
COMPLETED
COMMENTS
*£jh-*t>b , 5*. =<£>-<£•
ttTU/z/s'
1
^y-<2 2 <»¦«,
•
/£>fr35"7
5k ,*p\e<£ /c
¦Yl-C?fiJ2-
r2*
5/^/^
I
S=
-V
'(
... >t
*«
/Vf /
I.
(t
It
1
1 /,Yf ~}b 2
L r*. 4H
5^di><0 ^3
/ ,v< /// •="*
r r r*- /ib
(, /
i
r<
»l
V
/ /
'A12CA
£>(.-0
/ v -
&, 5/ 2 <5/ ^STi^s,
<>¦3
57a
-------
ULI lHHin.
Date of Analysis: M3-- k/9
Analyzed By:
Sample Preparation: I>Eie& 2 at @ lOtft.
Sample Description: -StiLib Awm -Wifie
X By Weight
Analysis No.
Dry
Dry
Dry
Analytical
Method *
IOB3b4
•
Carbon
5 (*%
Hydrogen
054
Oxygen
Nitrogen
OA1
Sulfur
a>.m
> s~.r
Ash -V
Mm
W.<33
55.30.
Moisture
Total
~Unless otherwise specified, analytical methods
are as follows:
Carbon, Hydrogen, Nitrogen, Ash: Perkln Elmer
Model 240 Elemental Analyzer.
Sulfur: Leco Model SC 32 Sulfur Analyzer.
Moisture: ASTM Method D-3173
Jject No: 13 — 3-0
Project Sample No:
Lab Sample No:
Date Received:
Comments: CttQ sours:
<083tW - n
-------
Oi
&
<4]
<
O
uR
• J
UJ
a.
ui ui
I- >
< —
a —J
UJ h»
-I <
a. cz
£ <
< Q-
Cfl UJ
0£
C_
UJ H-
O-
a. —
£ OS
< O
co
vS
£
rv
jl
\/
\
N
$
«
? 5
*: t
V\
<>
«»
V. ^
^ •-
k c
tv
.i,
i
J.
3\
\j'
SL Q
f 31
v\^T
\J~
pv
0
r\
B-17
-------
C^£OC.S | \
bote
Stb. Cor^i
leu*)
AcrtwriofiS
AttH.
Enjicnoa
Aua
(tM»)
Coot. Lpit~)
i/bleu
SO*.
'fro..
1
5D
4o4. (pete
I z?5p
¦ Snx.
/Boo
4-its
1 :
SO
4oi,.n»s
Ic^ZO
lhlAi
iSflt
fooo
4-ui
?
SO "1 dw-ero&G
l-o^P
trS
cSA.
lf>Do
4-ui
1
so !
tvlo
SO-L.
. JGm.
Hi
_ so /W.W
1- ir
ieco
lf-Ul
1
so
1 7 7 i?
Til fob
SO*
i boo
4
so
AM 2&DO 1 r ~r
¦
so*
I Zoo
4-7z
<=?
SD
.3B7- P3SS
11 r'C
SO-L
l?>oo
4-no -
53)
3&4- s&ol
\1 r-o
SOT.
IfVW
412
SP
Aoy- ¦ zx>:2.
)r*T
4
— -
:
...
j
- ¦ ¦
-
—
1
1
;
1
1
J
I
I
I
I
I
I
i r
1
1
i
I
!
i
j
i
i
i
!
"i
1
1
i
1
B-18
-------
- Mouec's Suufiie, <^mpi£&
IXine
SpmPuL-
Ho-
3kx6C.Tlt>H
Qefrmoti
"Tixit CmuA
**
6« S SpW(£>
Com.,
(pr*)
A^-
C/c^Z
M&e
IIISDl
to
/.Va
554&
uisoi
10
53c. 104 5
Co*
IC*$
> \£Pp
i.rt
loyi- SledS
u>*
as».
SOr
j-ee>
25-01S&
4-11
ft. r?3/
ins°i
7 0
IM
|13L»
/• 7/
lo1\.pno
UcS
—
d. i>?
to, cXfc
ist
4-U4
o.'ivo
Spg,
**ro ^
\l1Sbl
• se
IM
3sz>.4oz3
Cob
—
brf
775. ii
jLs
J96
1A-A94I
C, s»
—
•in
owl
J<1
—
•*
lliso/
•?o
h1>i
33-?.
CcS
ho*
7V7. 5Z»^
A/.S
o208
12-il 61
tSu
-
4-1 e
PfSSI
Se>i
—
//75V/
sv
3UUi3L*b
tce-iiui
Cps
-Ml
io*.
IU*
2-04>
is-fzn
4-v/
0> Xllo
jpJ. /zsV
i/32- Z3fe>
a-trtzj?
p-ltel
W. iPleft
_£/
Iro
>
r
_u]ll&u
lnscuL
msoz
mso?-
10
7 D
-SO
/¦Vr
1-11
\is
\.4s
J'i?. ... .
j-b&
V-72>
l-M
I-U4
Mitj
B-19
-------
MpUgg/s SiLufuts
D/nt
Szmpu-
Ho-
3*054 rta>
Vpl.(^
nwe (m"A
At-iA
(.u*s)
{lAl Sptat£>
Cone.
(fT"0
blll&K
\\19o%
SE>
51. PI02
Cc?5»
/•7<7
3
7 O
Sib-ISVi
iMMv
W».5
—.
2Gt £>
2.08
IS-PtlV
£&i
It?
2
>\-t4
i*.im
Jo.-
2?ft>
II75PS
no
|.«W
zneAii
COS
l.t/1
we- smi
J2.08
Jt.Mrt
(Si
<£=>
A-1^
JO, &2&>
5o»
AIO
*
H1SP3
10
1.44
S70. 'Woo
CP5-
1.71
K7W.37I5"
li%5
-
'
Jl.t/StsZ
CSm.
u-z.
4is
M- 4e>v4
5c*.
J-^O
II1SP3
SO
i-«W
3 70- 013ir
a»
Ux.5
_Cit
—
1-7(7
lOl.OlSlu
.etUAd
\l 1&32
4-7F
13.21 U(/
—
IllsroS
so
_l.il1!
J.-70 .
2-tft
475"
hM
t-Ul
_j2-ei
4-7 U
355- ^zjq
|0-4s47
12. I«W*|
33^-^X-
. W:A%oip
.) 5. wz.^
M-llUls
l4(f. S&l
o-i 4uu
O-SI&l
>
lit*
Sl">
id
'1t>
¦
SC
C-Sv
x>*
Los
W»S
CS»
•So*
toS
HiS
CSi
5£>>
A —
L
27?
rr - 51
* 705
II.-7503
/I7src^
ID I-4H
|-7£7 .
*z-0fL .
4-77
—
B-20
-------
Houec's Souoe,
bft*-
SwOifi.
IktetnoH
VW-
fewnorf
Tiwt
fj
¦y
(x*% SpttifcS
CotiC.,
(pr*)
ufzl&v
insert
to
I'4$
ZOTA'ZO
to5
i4i
l.-ll
1^.0574
Hi
S7£>
^ in
*Ao
o.^&4
f,5*
*1-73
n. i4lft
t(75C>4
li>
lAu
/7.
Cos
.1*1
j5£>?
z. i-
^l._. .
l-V*
.>1/
s]&o
IU.&7&4
p. $4ais
ff'CAsS
hu
_fls»
3€%._
ccs
2w3.rs6£?
uh (nt
||-J«5CS
7P
h'N
MP
4os.4z.pS
W*S
^•STD
JW
Pi - 573
C=S«-
M
4-70
io. 7tt4e>
5Pt
J\0
k
insos
TO
»'4ff
ikfr,3liA I
fi
<#U0. (#45®
" li-4i3l
_ )J»S
.. _-S£?*
4l^
W-4lS2>
Zj*?172
so
i
117SC?S
\\1SPS
_l4tl
j.^
zee
4n
i<#
liz.
22Zfll<|5
\-LWbO
SlsM-jg
js\.s\43
7£?1. (75#?
J5:03>?>
£.374p
15. 434b
\oO-r>M>o
0-55S1
daO\A
- J^s
H«-S
CSz.
S>Pt
Cfi
SO*.
ACS
jfeS
£6z
5C*
^ry' lulls'
..." 2MSS"
/3o
)3t±,
HfO
/4>
(rC>
U1SOV
7 &
J. ep __ _
1-7 It
J.'A
4-11/
B-21
-------
Mpuec's Su
\
DfiTi
SewPU-
He.
ImtCnorl
ypt.
fewuM
"TVie (muA
At. fcA
(.o^)
6r« I Spt£Jki>
Cone,
tpf")
ultl&e
117 5Dlf
ip
|.2x>
to*
|£D
l-frl
iot.nsrt
Wi3
iKlP
J'Vi
0-
«»•
^ ID
f5*>
0'\b6l
12-0
103.0433
HtS
sz?n
?-8l
0-44/1
CS.t
4ui
0%
cs*
23
^~a=
Aun
3-67M
SOa.
~ \oo
\2.e>
\itse>7
no
l >fo
3SU.WjA
Cob
\u*>
II In* c/ios
Ux%
—
ae>\
-35- 8S71
C6l
H-un
So*
Ho
i
II7SD7
so
l*Ho
Cc*
—
\-uu
0*35-077/
2-t>*
13-SB^O
cs»
—
4-ttB
3-37tf
V4o
-A2^.^4o
T.WJ.37/Z
C*?5
—
1 -iti/
WiS
.
J.0A
13-3545
C-5 t.
—
4u8
3-4sh/
Bl7»-
,
Jbs.~2P\e>i-
311
)/75~
1
' 1, o
—-
1|15C7_
3C
\> 4*-
CCS
C&X
SP»
CCb
titS
C^7-
<5Cj.
UiS
CSx
&p%-
—
2>a
ic
— i
i
l'i*B
3~e>?>
4lV
Ml?
Yus
JL
4-oi
j.sq
)-US__
j-e>\
4-on
4i&.xitn
s-col^
1,4s*iq
g4-svt>4
05Z
.
I'l?n3
5s4z.
447-^eg
7-3o7g
0-leV>~l
32*
7
.
U7^1_
inspg
B-22
-------
*\ruec's SoUfafj 6
Dfiie
SwPit
Ho.
Vol.
£eiwnc*J
£m"A
AltA
(c*/>
(jui Sptut£>
Cone.,
Or*)
tiMu
WisoS
¦to
L Hi-
^,564i/
Cos
315
/ 1
i \-in
Hite.tfZX,
WaS
—
t
7. US'*)
f,S*
7
1 A-13-
O'Snu
sth:
j£- i.fO
44. m sr
toi
—
J3<>. |15g
tf»S
1170
2-6a.
3.5710
CSt
—
l7t?
O-lSTb
SP»
—
II7STP8
so MJ
Js.u&si
Cx>S
—
l£/7
WaS
II Pi?
oea
S'iBSl
Cj& V
—
' 4.11
o- mi
SO*.
5
(
.
.
—
—
- —¦ —
— —
-
.
B-23
-------
HTO Data Example
-------
ilM
j tiu
"* y$A COD^ CP bcrC£&_
cUl* {°y *T -fi>y 77>>v
^CYt/bb,^ $6mj2&s fffrty) ~tAi2
-T:( 0 Ciscjf~!\ ^>\/.S^7ys v
$ ! 1 ^"* ~/u c k >~J •>^$^r-—:
^ t/ijc ^''^y /^3- C-T-Vje. )
C c£tm*r{U\^ imX )
-f' !\JpaJLc^ fi>0{A^ ,
CfYikd]^ *>MK^-^-s, L f 0 ^ - T^-^ciX ^uX^u y ' I/fa® (H/u
i-v^l m4\ hy €e tUjtliT- /i-u n .
. | L Sn.-'m.R-^P -j/(5 > "771?w ^
fjji^tA pi 4CK
dx^X -(oY Attdtfe
B-25
-------
^ 15it|cn3*c 'i'i "
@cv/s'i ( e*p3/ (^o)
Y © u , ?'/, &K/ &<>)
* , s%, TeJi/C**)
f ® " / f/,, pLOrJf'V
7 YW?par**Aj Brc^b, <& Ofr<^»,
$ Tv*<fcfe' , £e{^ s^/go 1/
-^i 7' ItyBi C_ , /D ^//-i 2—
@ Cx/s^[ Cas ) , /vZ %* o,
0/s-\(Gji\ lo'/FeJdj
© Co,ls-~z(as>) j/of fttfj .
3 =-/) ' ''ii
/(&
2-
0> /W'J^ *U«* s£y. /^u;,w #v^ ^
^¦vWlS aj- irs^i S , ~i4-iac/c £ (-AXV C fo-stSit ol&£L ^PteS^S^ ) &> f- /y
¦IfUfU^ «r /) -^et^c(,
B-26
-------
ijfth /lewif- , AJo u-yy M} {,>^2jQ>s -fey -tc-Lxl $u£^t,(r
^ d)jch\4 ^Zlr± *i V <&"¦ >
Mu&> As b**Xs.
^ [ i/uiZ^ Co or fiAscA&st-^ &€*
~fabu&£i. &QA djl faf/ttrii+fctri* cteCtG ,
.*10"' ( I Hk -dx^^A
4; c i> (fo fiorr^ >•
J&c, ?o, 'ss
f 1'^° Ml ~b^ ']%
¦>&\-\ d-
, IHK "fov |?c4ci 'c pW^-d '
(i vst (ca *~j *> \| ^
'U
!f
cicXk C £f~ ^1
4 (J) /^0J-
'4*10
/1/
^ f fiL^HsLblJ ^
¦* Jk iL lod<
r.
B-27
-------
/*
L
IAD SAMPLE
TRACKING REPORT
WORK REQUEST
'C
PROJECT Ml/.-
ENGINFFR ' L ?
REMARKS:
CHARGE NO.
sample"
NO.
SAMPLE
DESCRIPTION
ANALYSIS REQUIRED
SAMPLE
PREPARATION
DATE
DELIVERED
DATE
COMPLETED
COMMENTS
108 7 £
foe. l-W-ft ipS-c
12/17/' •
tap.
MZi
•»
f>
li
1»b •- 0.<|S
•am
tnXatOf.. loKi*
< (-&JO
n
'/
(i
% S - \-p-S t\.3.%
'tSlty
104 fre*vx , 30'tt/K
" Wl)
•1
it
u
%S* n.s
c*/s=Z ,c+s ^
1nZB,.Oir ZomM
- e-2in
•l
«/
«)
% S 1 q-30
<»7 S""/-
» (^r/)
11
< 1
%s -- 4-18
loiltf
Ca(s-\ . & s
\n/f IJDi,
u (.Mo'/)
• t
'1
1 t
fo- i-n
mri
I, * *
^•5 4-^D /
ct7sS
'' Ho*/)l
,
1 «t_/
OO
ro
00
-------
LAB SAMPLE
TRACKING REPORT
WORK REQUEST
PROJECT NO
FMfilNFFR -fcl
REMARKS
CHARGE NO.
r n ^
. J-f aMentfts o-f e^iK ' ^g^f-iru
SAMPLE*
NO.
SAMPLE
DESCRIPTION
ANALYSIS REQUIRED
1
SAMPLE
.PREPARATION
DATE
DELIVERED
DATE
COMPLETED
COMMENTS
tosiir
C~/S* 1 .CxS
Pt)L.v€&?£o* DfMtO
io^-c.
tz/ffl'ss-
nj n/g-s
%S~- 17-7,
(OSJ66
(WS-// C~S
{r-tf/j
'1
«!
n
%S -- 17-0"
iosp7
CkS
i o/f a*Oi. t
" Wl'/.)
II
M
7»s-p.^w
tusW
Cx/S=l'
(n-HX)
V
it
'/o5» le.sft,
*
" PLEASE PUT IN ORDER OF PRIORITY FOR ANALYSIS
-------
LAD SAMPLE
TRACKING REPORT
WORK REQUEST
PROJECT MO
FNCilNFFR ' Bryx^y
REMARKS:
CHARGE NO
-&L
SAMPLE*
HO.
SAMPLE
DESCRIPTION
ANALYSIS REQUIRED
SAMPLE
PREPARATION
DATE
DELIVERED
DATE
COMPLETED
COMMENTS
'vSWt
,<-
Pot iu«* e ioi.0c.
u 11 e>s
f*S \-V\
\«tW
LKiifL
•• (rtf)
1
n
11
'Us>0.'$1,0-TS
mi(
h
'f
n
3*
v>
it
y~
¦otnz
1
~
'i
%S ' |5,£ di/.O
08775"
U(lr\. 7G&
lOfH.r*
11
i'
11
fc - ll-(* , iz.z.
l" PLEASE PUT IN ORDER OF PRIORITY FOR ANALYSIS
-------
OO
I
CO
I -
&
TaupmW
C°c)
S.P.
Runnirt^.
(i" mJL
i
cYs
r'1
l/V'i /„ x
u J
0
(0^1 +
—
—
0 Win.
its
pnSUf)
1
\f2.7Z'
—
—
iojM-70
(ifDO'C
o.f
\l mm
*i
\
ttbtf
-0.OII
Alo.
I0JW
»'
¦i
ZOMirl
u
IfZ.ltf
-o.ooG
Mo.
|M7l
it
««
bo Win
«
w&Z
-o.ooZ
/Jo.
iojsi3
j ...
—
!
i
Otnir\
i(^s)
imJ&JI
¦
zi.t7.
—
mWI
Ilfocfe
p.r
12 4Hin
it
[ittV.
-0.2ft
Hiolbn.
(OS
"
!
H i
; !
?0toVl
H
wt.
-0.&
?Ojdl*J$£
lorffl
i
i " ;
i
$6
oMV,
-1.%
Yes.
-------
p>
09
I
CO
ro
&
C°cJ
S.p.
Runnirt^.
(i mJi
ci$
W3f 1f<7 1
¦¥M
wmT*
/ fobt
c\ £"~
ow*
ts
(C+S-t
7*h)
I3-?SZ\
—
lOSMS
II
•*
[2 Wi*
»l
j
5.2%"
-P.U
Ye 5. .
!l)W|
l(
"
3^'iti
i*
-M
Yes.
logyoO
1
ll
1
<1
K
us7.
-l./z
\es.
(MS®/
1We
I
j 0.5T
Vlhirt
7*0
ca s+
ZSr.fZ
—
—
loSSOl
M
"
12 "Wirt
l\
72^Z
-o.t£
A/oy
I0?9>i
II
1
i
#;
¦
2>0 Wirl
n
fls/C
-0^2.
/Jo.
,W$'o¥
i n
I if
1
!
(pOfli*
it
/7-&
-Ml
1
-i,
- ¦ -
-------
ft
CO
CO
TeaipwW
(°cj
S-p.
Runnirwj.
(> mJL
1
cis
'
i
0
mm
1H-D0°c
OS
(U Two l&W
' "/
\IS~.4 '
s>
fivrft'dlfy
WldlM
|D5t7S"
II
^Ontlixi.
/ ¦
it
i
-P.37
1
?cu\t*oMy
UkUd -
WW#
0
ti
6D1llit1.
u
6.0f/. !
-flfr
Yes.
L lot Jit
1 If 00 c
OUh'n
(GS4
|. 4"R
8S»
i
/«#¦
—
—
Wj1$
l»
i m
11
¦
IIM
-0.70
P
-------
&¦
00
1
u>
•f*
&
—\
^ IP
o"§
S-P.
...
Runniiwj.
(imJl
Cys
1
wx7
W/( FetPi
"
wZ
—
—
lotW
•1
<<
1
(2Mn
t>
'
tWf,
-mz
Voa&Mj
TJloQtl*
ittty
rt
0
1
VI
\15'/.
Yes.
10 s 760
i
• !
;
£o#ft
\\
nW
-/.??
Yes.
Will
(Moot
1 0.
-------
CO
I
CO
cn
^xtnp&-
&
lesvipgi-oXtfe.
rcj
s.p.
Runnirtij.
( i n\JL
cYs
I
I
Sm$i
TiSi.7'
0
mnt
1 tyo'c
oS
12 Win
C
i
' i9.qz' -o./t
ahur
TtlettsA
I o$7&
\t
jooni^
~
If
f.3/ ; -OMZ
\
Ru&kM,
WettA
fa.
(°*76l
H
(?0lHirI
u
2-77%" -o.0
\
'«•
1 "
1
i
'
¦
1
i
1
1
1
1
1 — ¦ " " 1
1
1
i
i
I
(V~-
¦
-------
00
GO
Smth.
"ph^ratw
( *C)
S.R.
(%/
£r^jtc< /;i»yV'
/ imJi
*% ' Vs- i
>*%$ ' '
w!M
10.5771
~vm°t
as
IMWf
1 (C&.0+
TIImokCbJI
—
mf
-
-------
XJ3
I
CO
>sl
IcibkX Cwi-fciiw^O
SarJj-
*
fyvperc&re
C °C)
SX
(1 mJL
Co/s 1 fa/s
-7MHI5!
10 mi
I200°c
0,5"
M "iff'9
LfZHR
team HI.
Co«X^s&
10/CBM
—
i-£T
lOXltt
w
M
\i
It
icaoniL
GJIU +
[0°A &0O
?.2^t
•(Mvd
7MJ
yv
fPartiJ2L\
)asp
w
V
11
i(
2(U0fllL
Co*$.Asfi -/*
10% 8,,oO
AJt/g
' ~
'
•n
1
'
1
f
- -
.. .
1
...
. . .
-------
'"R
LAD SAMPLE
TRACKING REPORT
WORK REQUEST
PROJECT HO. /S~2Q CHARGE NO
FMfilNFFB ' L{
REMARKS = &mT gvX J/r trlaiM
SAMPLE*
NO.
SAMPLE
DESCRIPTION
ANALYSIS REQUIRED
SAMPLE
PREPARATION
DATE
DELIVERED
'(.DATE
COMPLETED
r.ten
COMMENTS
I Z-lOrb
a/s=i
F*f.,-Hr<£flS3
|?mti
)7~10'^
J2S7 >
12-IP-Z
T~us!t\4 T(RetlMU>y&tl
i) *
y\
3.093^ /zzur)
12-IQ-}1.
C*l£'l.$PJs
PmSi)*4 0,
jy) vv
\ \
71 V't -
z.os'lf A ?z>e,
If 1 J
\v n II
\ V
>v
n !'c:>
z.osts ( ?Ozo>
12-/(7-
Hi '\,^Za.K
l| M l\
%v
U
ZWWf.iZit }
12-10-6
C4i*\, toXUK
11 u «t
k
7 i f i -
Wf-IS,
(2-10-7
uts* \.W&t
l< '1 If
M
• V
{-(¦¦¦¦
wwa,
\i-(0-3.
Ui-'l,
I) v<-
H '1 *
ll
>1
>Ur3<^
CO
I
CO
CO
¦ PLEASE PUT IN ORDER OF PRIORITY FOR ANALYSIS
-------
"R
LAD SAMPLE
TRACKING REPORT
WORK REQUEST
PROJECT MO. $£($-2-0 CHARGE NO
F.1G1NFFR 1 BrtAlJZ L/v
RFMARKS; S
-------
Hazen Research, Inc.
4601 Indiana St. • Golden, Colo. 80403
Tel: (303) 279-4501 • Telex 45-860
DATE January 8, 1985
HRI PROJECT
HftI SERIES NO.
BATE RECQ.
CUST P.O.#
PROJECT #
005-17
31 7 4 6 - A
12/16/85
None F.ec'd.
8518-20
Energy St Environmental Research Corp.
Mr. Timothy C. Grogan
18 Mason
Irvine. California 92714
REPORT OF ANALYSIS
SAMPLE SAMPLE REDUCING ATMOSPHERE OXIDIZING ATMOSPHERE
NO. IDENTIFICATION IT ST HT FT IT ST HT PT
31746-1 12-10-1 2120 2165 2175 2197 2130 2169 2175 2200
./. . ftf /'
By:' /' ' ' • . /
Gerard H. Cunningham
Coal Laboratory Manager
B-40
-------
Hazen Research, Inc.
4601 Indiana St, « Golden, Colo. 80403
HAZ1N Tel: (303) 279-4501 • Telex 45-860
BATE January 8, 1985
HRI PROJECT 009-27
HRI SERIES NO. 31716-8
DATE RECD. 12/16/S5
CUST P.O.# None ftec'd,
PROJECT # 8518-r<0
Energy Si Environmental Research Corp,
Mr. Timothy C. Brogan
18 Mason
Irvine, California 92714
REPORT OF ANALYSIS
SAMPLE SAMPLE REDUCING ATMOSPHERE
NO. IDENTIFICATION IT ST HT FT
31746-2
12-10-2
2076
2140
2150
2165
-3
12-10-3
2050
2080
2090
2140
-4
12-10-4
1885
1935
1950
1975
-5
12-10-5
2135
2170
2 ISO
2193
-6
12-10-6
2114
2160
2170
2186
-7
12-10-7 .
1885
1951
1960
1982
-8
1 2-10-B
1850
1900
1930
1958
-9
12—10-f
2682
>2700
-
-
-10
12-10-10
2530
2600
2655
2675
-11
12-10-11
2675
2680
2687
2695
-12
1
1
2255
252b
2550
2580
-1J
12-10-33
2142
2264
2285
2447
v.- / /¦>'
gy I /•-<-"
Gerard H. Cunninghaffi
Coai Laboratory Manager
B-41
-------
Hazen Research, Inc.
4601 Indiana St. • Golden, Colo. 80403
HAZEN Tel: (303) 279-4501 • Telex 45-860
3ATE January 14. 1?56
HPI PROJECT
HSI SERIES NO.
DATE RECD.
CUST P.O.*
PROJECT *
OC9-2.'
317 4 6 - fi
12/1t/35
None ftec ' d.
8516-20
Energy St Environmental Research Corp,
Mr. Timothy C. Grogan
18 Mason
Irvine. California 927>4
REPORT OF ANALYSIS
SAMPLE
NO.
SAMPLE
IDENTIFICATION
DEDUCING ATMOEPHEliE
IT ST HT FT
OXIl/lilNB ATMOSPHERE
IT ST HT F;
31746-1 12-10-1
2174 2180 2186 2190
2200 2240
2276
u -' r •?<
•.-•-tr n ¦ :
B-42
-------
Hazen Research, Inc.
4601 Indiana St. • Golden, Colo. 80403
Tel: (303) 279-4501 • Telex 45-860
OflTE January 1-
HRI PROJECT
HR1 SERIES MO.
DATE RECD.
CUST P.O.#
PROJECT ¦*
1986
009-27
31746-B
12/16/35*
None Rec'd.
8516-20
Enerov Environmental Research Coro.
Mr. Vinctfiy V. Gronan
16 Mason
Irvine. California 92714
REPORT OF ANALYSIS
SflhPLt SrtnPi.E
NO. IDEMV IFI CAT I ON
REDUCING ATMOSPHERE
IT ST HT FT
31746-2
12-10-2
2115
2210
2225
2245
~ J
12-10-3
2160
21 BO
22 00
2235
-4
12-10-4
1973
1995
2005
2020
- i
12-10-5
2150
2)73
2184
2195
-13
1
0
1
2197
2393
24 30
25 2'j
B-43
-------
l/2V^ Table X
i
j
C
C.2100^*)
X«&tc'p
100^
fVity
feA
W*
Puri'-tH
B«v
w?
R^n-Ui
w3
v
r»Fj
a
AM1W5
Is/T
Rum
2 M>£-
2/W.
/?7f.
zm.
?b7S.
2 (AS.
2980.
2^?p
tM0(
fod
7MF-
Z2i$,
7420,
ztt£.
—
—
—
ZSzo
/^ewp.
zzog.
2IU.
w
21W
2675".
26fJ".
zsJp.
2?$V-
Z(8}.
ws.
/-?33.
2-M.
2*5?.
2<9£
22?s:
2$0J.
10Z
/iJii-te
'«
1st
fun
720?.
zz°t
2 ZPS".
7$ 10,
00
1
4*
-t*
Hfih.' Ale tlu dilute wsuSjts AVf r&fvrhA frnn Res&irc/v,
------- |